Boston University
OpenBU
http://open.bu.edu
Theses & Dissertations
Boston University Theses & Dissertations
2016
Identifying and modeling the
contribution of nuclear receptors to
environmental obesogen-induced
toxicity in bone
Watt, James
http://hdl.handle.net/2144/19516
Boston University
BOSTON UNIVERSITY
SCHOOL OF PUBLIC HEALTH
Dissertation
IDENTIFYING AND MODELING THE CONTRIBUTION
OF NUCLEAR RECEPTORS TO ENVIRONMENTAL
OBESOGEN-INDUCED TOXICITY IN BONE
by
JAMES DOUGLAS WATT
B.A., University of Colorado at Boulder, 2002
M.S., Boston University, 2011
Submitted in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
2016
© 2016 by
James Douglas Watt
All rights reserved except
Chapter 2 © 2015 Elsevier Publishing
Chapter 4 © 2016 Oxford Journals
Approved by
First Reader
_____________________________________________________
Jennifer J. Schlezinger, Ph.D.
Associate Professor of Environmental Health
Second Reader
_____________________________________________________
Thomas F. Webster, D.Sc.
Professor of Environmental Health
Third Reader
_____________________________________________________
Wendy Heiger-Bernays, Ph.D.
Associate Professor of Environmental Health
Outside Reader
__________________________________________________
Koren K. Mann, Ph.D.
Assistant Professor of Oncology
McGill University
Acknowledgements
I am indebted to my advisor, Jennifer Schlezinger, PhD., for her willingness to
take me on as an advisee and let me continue the research we started as part of my
Master’s degree. I’m grateful for her trust in me to branch out into the field of chemical
mixtures and develop new lines of research in the lab. Her energy and enthusiasm for
teaching and research has always been inspiring.
As a part of the Schlezinger Lab, I’ve had the privilege to work with great
students and technicians, the foremost among them being Cassie Huang, MPH and Faye
Andrews, MPH. They have been a pleasure to be around on a daily basis and without
their expertise and hard work this thesis would not have been complete. I also feel lucky
to have shared lab space and many entertaining lunches with the members of the Sherr
Lab, particularly Olga Novikov, PhD., and Elizabeth Stanford, PhD., and for the advice
of Dave “it’s a great day for science” Sherr, PhD.
I am grateful to my committee members Tom Webster, DSc., for his stimulating
collaboration on the mixtures project, and Wendy Heiger-Bernays, PhD. for her
guidance, particularly early in my Master’s degree as I was being introduced to the field
of Environmental Health. I thank the faculty and staff of the Department of
Environmental Health at BU SPH for their constant support, in particular Mike McClean,
DSc., and Madeleine Scammell, DSc. for their work as heads of the Doctoral student
program.
I am also grateful to have had the opportunity to learn alongside the many talented
and brilliant PhD students in the Environmental Health Doctoral program. It was a
iv
privilege to be a part of such a cohesive group, and always fun to hear about the creative
and varied research everyone was developing. I look forward to being your colleague for
the years to come.
I should also like to thank Jeffrey Moore, PhD., who hired me as a technician
years ago and trained me in becoming a better laboratory scientist. His encouragement
and interest in my career was a big driver of my decision to pursue a PhD.
Last but most important: I thank my family for all the unwavering support they’ve
given over the years, their belief in me, and their excitement over my following a career
in science.
v
IDENTIFYING AND MODELING THE CONTRIBUTION
OF NUCLEAR RECEPTORS TO ENVIRONMENTAL
OBESOGEN-INDUCED TOXICITY IN BONE
JAMES DOUGLAS WATT
Boston University School of Public Health, 2016
Major Professor: Jennifer J. Schlezinger, PhD., Associate Professor of Environmental
Health
ABSTRACT
Bone is a dynamic tissue, where bone forming osteoblasts and bone resorbing osteoclasts
maintain homeostasis. Research into bone toxicology has largely focused on
pharmaceutical side effects adversely affecting bone development. However, many
environmental toxicants can regulate bone homeostasis. Recently, the nuclear receptor
peroxisome proliferator activated receptor gamma (PPARγ) has emerged as an important
target of environmental toxicants. PPARγ dimerizes with the retinoid-X receptor alpha
(RXRα), is a central transcription factor in adipogenesis, and in bone can
transdifferentiate osteoblasts into adipocytes by suppressing osteogenic pathways. The
central hypothesis of this dissertation is that environmental chemicals can adversely
affect bone homeostasis by activating nuclear receptors in bone cells – particularly
osteoblasts and osteoclasts – to perturb cellular differentiation and function. Three study
aims were developed to test and refine this hypothesis. First, a set of structurally diverse
environmental PPARγ agonists were individually applied to mouse primary bone marrow
mesenchymal stromal cell cultures undergoing osteogenic differentiation. In vitro PPARγ
ligand treatment suppressed osteogenesis and stimulated adipogenesis. Organotin
vi
compounds (tributyltin, triphenyltin) in particular more efficaciously suppressed
osteogenesis. The second aim characterized the effects of in vivo tributyltin exposure on
bone microarchitecture in female C57Bl/6 mice. Tributyltin exposure resulted in a thinner
cortical bone, but significantly increased trabecular mineralization. Further analyses
suggested that tributyltin did not suppress osteoclast numbers but rather changed
osteoclast function, minimally attenuating the resorptive function and enhancing their
ability to generate osteogenesis-stimulating factors. Furthermore, tributyltin activated not
only PPARγ, but also RXR and liver X receptors. The third aim established the utility of
Generalized Concentration Addition in modeling PPARγ activation by mixtures of full
and partial PPARγ agonists. A complex mixture of multiple phthalate compounds
activated an in vitro PPARγ reporter assay, and the individual dose-responses of each
compound were used to construct modeled responses. The comparisons of empirical data
and model predictions supported the use of Generalized Concentration Addition in
modeling a complex mixture of environmental PPARγ agonists. Together, these studies
support and establish important toxicological mechanisms related to PPARγ and RXRα
activation in different aspects of bone biology and provide a basis for studying mixture
effects of PPARγ agonists.
vii
TABLE OF CONTENTS
ABSTRACT ....................................................................................................................... vi
TABLE OF CONTENTS ................................................................................................. viii
LIST OF TABLES .............................................................................................................. x
LIST OF FIGURES ........................................................................................................... xi
CHAPTER I: Introduction ............................................................................................. 1
The nuclear receptor superfamily.................................................................................... 2
Nuclear receptor activity ................................................................................................. 3
Structure and specificity in nuclear receptor activity ...................................................... 5
The PPARγ/RXRα heterodimer ...................................................................................... 6
PPARγ as a target of environmental obesogens .............................................................. 9
Obesogens and bone ...................................................................................................... 11
Relevant aspects of bone biology .................................................................................. 12
Environmental bone toxicants ....................................................................................... 22
CHAPTER II: Structurally-diverse, PPARγ-activating environmental toxicants
induce adipogenesis and suppress osteogenesis in bone marrow mesenchymal
stromal cells ..................................................................................................................... 29
Abstract ......................................................................................................................... 30
Introduction ................................................................................................................... 32
Materials and Methods .................................................................................................. 35
Results ........................................................................................................................... 40
Discussion ..................................................................................................................... 46
Tables ............................................................................................................................ 52
Figures ........................................................................................................................... 53
Supplemental material ................................................................................................... 71
CHAPTER III: Tributyltin induces distinct effects on cortical and trabecular bone
in female C57Bl/6J mice resulting from modulation of both osteoblasts and
osteoclasts......................................................................................................................... 73
Abstract ......................................................................................................................... 74
Introduction ................................................................................................................... 76
Materials and Methods .................................................................................................. 78
viii
Results ........................................................................................................................... 83
Discussion ..................................................................................................................... 91
Figures ........................................................................................................................... 98
Supplemental materials ............................................................................................... 112
CHAPTER IV: Generalized Concentration Addition modeling predicts mixture
effects of environmental PPARγ agonists ................................................................... 116
Abstract ....................................................................................................................... 117
Introduction ................................................................................................................. 118
Materials and Methods ................................................................................................ 120
Results ......................................................................................................................... 126
Discussion ................................................................................................................... 131
Tables .......................................................................................................................... 137
Figures ......................................................................................................................... 139
Supplemental material ................................................................................................. 149
CHAPTER V: Conclusions ......................................................................................... 151
Chapter 2 ..................................................................................................................... 152
Chapter 3 ..................................................................................................................... 154
Chapter 4 ..................................................................................................................... 157
Public health implications ........................................................................................... 160
BIBLIOGRAPHY ........................................................................................................... 164
CURRICULUM VITAE ................................................................................................. 191
ix
LIST OF TABLES
Table 2.1. EC50, maximal activation values, and Hill coefficients for mPPARγ
activation by selected toxicants. ................................................................................. 52
Supplementary Table S2.1. Primers used for mRNA expression analysis. ..................... 71
Supplementary Table S3.1. List of primers used for qPCR ............................................ 112
Table 4.1. Modeled parameters for phthalate compounds from dose-response fits using 3parameter Hill equation............................................................................................. 137
Table 4.2. Phthalate components of mixture rays ........................................................... 138
x
LIST OF FIGURES
FIGURE 2.1. Environmental toxicants activate PPARγ1 and PPARγ2 with differing
potencies and efficacies. ............................................................................................. 53
FIGURE 2.2. Known modulators of adipogenesis increase lipid accumulation and
expression of adipogenic genes in mouse BM-MSCs. ............................................... 55
FIGURE 2.3. Structurally distinct environmental PPARγ ligands induce lipid
accumulation and adipogenic gene expression in mouse BM-MSCs ......................... 57
FIGURE 2.4. Environmental PPARγ ligands induce perilipin protein expression in
mouse BM–MSCs. ...................................................................................................... 59
FIGURE 2.5. Rosiglitazone and TBT suppress osteogenesis in mouse BM-MSCs........ 61
FIGURE 2.6. Structurally distinct environmental PPARγ ligands suppress osteogenesis
in mouse BM-MSCs. .................................................................................................. 63
FIGURE 2.7. Structurally distinct environmental PPARγ ligands suppress osteogenic
gene expression in mouse BM-MSCs. ........................................................................ 65
FIGURE 2.8. Adipogenesis is inversely correlated with osteogenesis in PPARγ-ligand
treated mouse BM-MSCs. ........................................................................................... 67
FIGURE 2.9. Differential suppression of osteogenic Wnt10b mRNA does not explain
efficacious effect of organotins on osteogenesis. ....................................................... 69
FIGURE 3.1. TBT suppresses osteogenic differentiation pathway and mineralization
markers more potently and efficaciously in female-derived cultures compared to
male-derived cultures. ................................................................................................. 98
FIGURE 3.2. TBT reduces diaphysis cross-sectional area and cortical thickness while
increasing trabecular structure in female C57Bl/6J femurs. ..................................... 100
FIGURE 3.3. In vivo TBT exposure increases both trabeculae and marrow adipogenesis.
................................................................................................................................... 103
FIGURE 3.4. In vivo TBT exposure modifies osteoclast gene expression without
influencing osteoclast cell number............................................................................ 105
FIGURE 3.5. TBT suppresses differentiation and functional gene expression but does
not change osteoclast cell number in vitro. ............................................................... 107
FIGURE 3.6. TBT activates RXR and LXR-dependent pathways in whole bone and in
vitro osteoclast culture. ............................................................................................. 110
Supplementary Figure S3.1 ............................................................................................. 113
Supplementary Figure S3.2 ............................................................................................. 114
Supplementary Figure S3.3 ............................................................................................. 115
FIGURE 4.1. The maximal response of rosiglitazone is decreased by high doses of
nTZDpa. .................................................................................................................... 139
xi
FIGURE 4.2. GCA provides an accurate prediction of full and partial PPARγ agonist
mixture effects. ......................................................................................................... 141
FIGURE 4.3. Phthalate esters activate PPARγ with a variety of efficacies. ................. 143
FIGURE 4.4. Comparison of experimentally determined PPARγ activation, using
phthalate ray design, with model predictions. .......................................................... 145
FIGURE 4.5. Extrapolations of modeled dose-responses to high concentrations for ray
A. ............................................................................................................................... 147
Supplementary Figure S4.1 ............................................................................................. 149
xii
1
CHAPTER I: Introduction
The importance of the proper function and maintenance of the skeleton cannot be
understated, given its multiple roles in protecting the organs, providing a structure that
allows mobility, providing a niche for hematopoietic development, and acting as an
important mineral depot. However, bone as a target of environmental toxicants has
gained disproportionately little attention compared to other organ systems, and studies
characterizing toxic mechanisms in bone have largely been in the context of
pharmacological interventions. Despite the fact that numerous classes of environmental
chemicals can affect bone homeostasis through a variety of mechanisms, bone as a target
of environmental toxicants is especially understudied.
A number of these environmental toxicants have receptor-mediated mechanisms
of action. One such receptor is the peroxisome-proliferator activated receptor gamma
(PPARγ), a nuclear receptor belonging to the class of retinoid-X receptor alpha (RXRα)
heterodimers. This receptor is central in lipid homeostasis and is influential in the
differentiation of bone cells. The identification of novel environmental PPARγ toxicants
and the known role of PPARγ in bone provide the possibility that exposure to these
environmental ligands contributes to the incidence of bone pathologies.
The section below provides a background for studying environmental nuclear
receptor ligands, including PPARγ ligands, in the context of receptor-mediated
toxicology and bone physiology.
2
The nuclear receptor superfamily
The nuclear receptor superfamily provides the mechanisms by which lipophilic
hormones and hormone-like chemicals modulate gene transcription. This family of
protein receptors has a specific macromolecular structure that permits coordinate binding
of a ligand (e.g. a hormone, toxicant, or any molecule that binds a receptor) to the
receptor and the receptor to DNA to initiate transcription of target genes. These receptors
are located either in the cytoplasm or the nucleus and may be bound by co-repressor
proteins that suppress their transcriptional function until they are released.(1–3)
Most nuclear receptors exert their mechanism of action as either a homodimer (a
pairing of the same receptor) or a heterodimer (a pairing of two distinct receptors).
Examples of nuclear receptor homodimers are found among the steroid receptor
subfamily of nuclear receptors, which includes the estrogen receptor (ER) and androgen
receptor (AR), both of which bind to DNA as identical pairs (i.e ER:ER or AR:AR,
respectively) to modify reproductive physiology (among other important endpoints).(1–3)
A second class of nuclear receptor dimers is the retinoid-x receptor (RXR)
heterodimer subfamily.(4–6) The retinoid-X receptor alpha (RXRα, which belongs to a
subclass that also includes the retinoid-X receptors beta and delta) heterodimerizes with
several other structurally distinct receptors, including, but not limited to, the vitamin D
receptor (VDR),(7) the retinoic acid receptor (RAR),(5,7) thyroid hormone receptor (TR),(7)
the liver-x receptors alpha and beta (LXRα and LXRβ)(8) and the peroxisome proliferator
activated receptors (PPARs).(9–12) Pathways involving RXRα heterodimers have been
shown to be important in many tissues, including bone, where RXRα:RXRα
3
homodimerization(13) and RXRα:LXRα or RXRα:LXRβ(13,14) heterodimers suppress bone
loss and have been identified as potential therapeutic targets. On the contrary, the
RXRα:PPARγ heterodimer acts to suppress bone formation.(15,16) The contribution of
these receptor pairings in bone homeostasis will be discussed in greater detail in the
sections and chapters below.
Nuclear receptor activity
The activity of a nuclear receptor can be quantified by the expression of a target
gene, and is sensitive to the concentration of ligand present and available to bind the
receptor. When a ligand acts to increase activity of a nuclear receptor, the ligand is
generally termed an agonist; conversely, an antagonist is a ligand that acts to suppress or
inhibit the function of a receptor (Chapter IV will provide a more specific definition of
the latter, namely competitive antagonism). In either case, the trend describing the
activity of the receptor as a function of ligand concentration is known as the ligand’s
dose-response. For nuclear receptors, this function tends to be non-linear over a range of
ligand concentrations and is commonly referred to as the dose-response curve. The most
basic theory behind this mechanism is occupation theory, which makes a number of
assumptions, including, but not limited to: the binding of the drug (toxicant) to the
receptor is reversible; the observed response is proportional to the concentration of
agonist-receptor complexes; and a maximal stimulus occurs when all of the receptors are
occupied.(17)
Three important parameters describe the shape of a dose-response curve for
4
nuclear receptor activation (which is typically presented on a semi-log plot, with log
agonist concentration on the x-axis, and response on the y-axis): the maximal effect level
at which the curve plateaus (the efficacy); the concentration at which the curve reaches
half maximal efficacy (in the case of an agonist, the effective concentration of the agonist
producing 50% maximum efficacy (EC50)), which is commonly referred to as potency;
and the steepness of the curve.
The relation of these three parameters can be described by the function:
𝐸=
𝑎[𝐿]
𝐸𝐶50 + [𝐿]
Where E is the observed effect level; 𝛼 is the curve maximum (efficacy) and [L]
is the ligand concentration. The concentration of the receptor is assumed to be constant.
A detailed derivation and discussion of this equation, along with relevant assumptions
under occupation theory, can be found in Kenakin, Pharmacologic Analysis of DrugReceptor Activation, 3rd. Ed. (Lippincott-Raven, 1997).(17)
More complex models, which factor in induction of enzymes, increases in
receptor concentration, or cooperativity of multiple-ligand binding, have also been
developed.(17) However, the above equation is, in some cases, suitable for describing the
most proximal (i.e. initial) events of toxicants binding and activating a nuclear receptor
and potentially its observed downstream effects as well.
5
Structure and specificity in nuclear receptor activity
Nuclear receptors have a canonical structure comprised of subunits that show
various levels of conservation among all nuclear receptors. The two most notable
structural subunits of nuclear receptors are the ligand-binding domain (LBD) and DNAbinding domain (DBD), both of which vary in structure among nuclear receptors to
confer specificity of function.(2,18,19)
The DBD allows for specificity in gene expression via selective DNA binding.
The sequences of amino acids that comprise the DBD confer a particular structure that
permits stable interaction between the receptor and DNA at specific nucleotide sequences
termed response elements. For example, PPARγ:RXRα heterodimers bind to a pairing of
hexameric sequences of nucleotides that are separated by a single nucleotide spacer, in a
conformation designated as a DR-1 sequence (for “direct repeat” with a 1-nucleotide
spacer).(2,20,21) Genes that contain a DR-1 sequence in their promoter region are therefore
targeted for transcription by activation of PPARγ:RXRα.(22,23)
The LBD confers specificity of activation by the structure determined by the
sequence of amino acids in the ligand binding pocket. This allows the nuclear receptor to
discriminate among the ligands that it can bind. The shape of the ligand also indirectly
determines the set of co-activators recruited to the receptor:DNA complex. These coactivators replace the co-suppressors (which inhibit transcription) and subsequently
aggregate the transcriptional machinery (including RNA polymerase). This mechanism
allows for differential efficacies among ligands.(2,18,19,24,25)
6
The PPARγ/RXRα heterodimer
PPARγ has gained considerable attention in the past three decades when
Spiegelman and others discovered it to be a ligand-activated transcription factor central in
the generation of adipocytes, a process termed adipogenesis. The precise identification of
PPARγ:RXRα as the transcription factor controlling the adipocyte-specific fat enhancer
ap2 (fatty acid binding protein 4, FABP4)(26–29) came in parallel with the discovery that
PPARγ was also the target of insulin-sensitizing drugs belonging to the thiazolidinedione
class of therapeutics, which includes the compounds pioglitazone and rosiglitazone.(30)
The latter has become the experimental agonist of choice for parsing the mechanism of
action of PPARγ as it serves multiple roles as an insulin sensitizer, adipogenic
transcription factor, and determinant of mesenchymal stromal stem cell fate in bone.(31)
PPARγ is expressed as two transcript variants: PPARγ1 and PPARγ2.(32)
Although both are influential in adipocyte differentiation (and highly expressed in white
and brown adipose tissue),(27,33) PPARγ2 is more selectively expressed in adipocytes(27,29)
and its variable activity is more sensitive to the action of exogenous ligands.(34,35) The
two isoforms differ by 30 amino acids added to the proximal end of the N-terminus of
PPARγ2, which contains the DNA binding domain.(27) The N-terminal domain also
serves as a regulatory target. The proximal 120 amino acids contain multiple serine
residues that can be phosphorylated by MAP kinases to inhibit the transcriptional activity
of the receptor.(36) Recent work has shown that certain PPARγ ligands can also block
phosphorylation of these residues, and specific phosphorylation events can selectively
modify the gene targets of PPARγ.(37) This imparts selectivity in PPARγ activation, and
7
current research seeks to identify therapeutics that selectively activate PPARγ’s insulin
sensitizing effects while suppressing its adipogenic effects via kinase pathways.
Dissection of the events leading to adipocyte differentiation has established
PPARγ as an essential driver of the transition from pre-adipocytes to mature, terminally
differentiated adipocytes.(38,39) Early adipocytes express transcription factors belonging to
the basic helix-loop-helix CAAT/enhancer binding protein (C/EBP) family.(40) Two
members of this family, C/EBPβ and C/EBPδ, induce transcription of PPARγ by binding
to the PPARγ promoter.(41) PPARγ, under ligand activation, induces transcription of a
third C/EBP: C/EBPα.(42) This protein subsequently binds to the PPARγ promoter, setting
up a positive feedback loop that sustains the level of PPARγ expression.(43) C/EBPα
cannot induce adipocyte differentiation in the absence of PPARγ.(44) In contrast, the
expression of C/EBPα is not necessary for adipocyte differentiation (as it is not necessary
for the initial transcription of PPARγ).(44) This implies that the functional consequences
of PPARγ activity are supported by a feedback loop that maintains a basal level of
PPARγ in cells.(43)
However, the primary driver of PPARγ activity, and thus its downstream effects,
is ligand binding. Endogenously, prostaglandin compounds (e.g. 15-deoxy-Δ-12,14Postaglandin J2), which are long-chain fatty acids that serve primarily a paracrine
function in fat tissue, bind and activate PPARγ.(45) For exogenous ligands, much of the
research on PPARγ activity has focused on thiazolidinedione therapeutic compounds
(used in treatment of type II diabetes), such as rosiglitazone, noted above.
The C-terminal domain of PPARγ is recognized as the major region that binds
8
chemical ligands (e.g. rosiglitazone) as well as cellular proteins (e.g. RXRα and
coactivator proteins like PGC-1a).(25,31,46,47)The LBD contains a large binding pocket with
many hydrophobic residues capable of binding a variety of lipophilic ligands.(48,49)
Because co-activators also bind to this domain, the specific, dynamic interaction of the
ligand with the binding domain causes conformational shifts that expose variable residues
that interact with different co-activators.(25,50) In turn, these co-activators contribute to the
transcriptional efficiency of receptor/enzyme complex. Research into co-activator to
PPARγ/RXRα heterodimer interaction has identified variability in coactivator
recruitment with different ligands.(51,52) In combination with variable affinities for the
ligand binding domain among PPARγ agonists, PPARγ ligands therefore can vary in
efficacy and potency.
Prior to these events, it is believed that PPARγ:RXRα dimerization occurs in
solution, where interactions between the LBDs and the DBDs stabilize the dimer and
allow for high-affinity binding to DR-1 sequences along the DNA.(4,31,53) It is also notable
that Chandra et al. (2008) found that ligands for PPARγ with distinct properties (an
agonist, rosiglitazone and an inhibitor, GW9662) elicited the same conformations with
respect to domain interactions within the heterodimer.(46) This suggests that differences in
the effects of PPARγ ligands is not likely to be due to changes in PPARγ:RXRα dimer.
For RXR heterodimers, the orientation of the receptor complex is important.(18,50)
For example, RXRα is positioned upstream of the thyroid hormone receptor when the
RXRα:TR heterodimer is bound to DNA.(54) Binding of a ligand to RXRα alone does not
initiate transcription (it requires a ligand for TR), and as such the dimer is referred to as
9
“non-permissive”. Conversely, RXRα is positioned downstream of PPARγ in the
PPARγ:RXRα heterodimer.(20,46,50) As a consequence, ligand binding to either receptor
can initiate transcription; this pairing is termed a “permissive” heterodimer. This
distinction is important when considering the suite of receptors with which RXRα can
dimerize, and whether the function of an RXR partner can be induced by activation with
an RXR-specific ligand, as is the case with PPARγ:RXRα.
To summarize the above, PPARγ is an essential factor in the formation of
adipocytes, as evidenced by its controlling role in adipocyte differentiation, its expression
levels in mature fat cells, and a host of experiments demonstrating PPARγ ligand efficacy
in adipocyte accumulation. It forms an obligate, permissive heterodimer with RXRα,
which allows for binding to specific regions of DNA that control adipogenic genes.
Transcriptional activation of the PPARγ:RXRα heterodimer is accomplished by ligands
for either receptor, where variation in response is believed to be controlled, at least in
part, by ligand-selective co-activator recruitment.
PPARγ as a target of environmental obesogens
The notion that exposure to environmental toxicants contribute to obesity on a
population-wide level was proposed in 2002 in a “provocative”(55) paper by BailleHamilton.(56) Citing the concurrent increase in chemical use and obesity rates in the
United States, this hypothesis posited that commonly attributed lifestyle factors, such as
physical activity, diet, as well as genetic factors, could not fully account for the observed
increases in adipose body mass, and asserted that chemical exposure was a driving factor
10
in obesity.(56) Drawing upon mechanistic research detailing the effects of chemicals that
interact with the endocrine system to disrupt metabolism, Grun and Blumberg coined the
term “obesogen” to classify chemicals that acted to “inappropriately regulate lipid
metabolism and adipogenesis to promote obesity”.(57) Obesogens have since emerged as a
special class of endocrine disrupting chemicals that potentially contribute to obesity and
dysregulation of lipid metabolism.
Due to its role as a master regulator of adipogenesis, PPARγ was quickly
established as a putative target of obesogenic chemicals. Among the first chemicals to be
classified as such was tributyltin (TBT), which several groups showed could activate
PPARγ and induce adipogenesis in vertebrates.(58,59) In vivo, tributyltin enhances fat
formation,(59,60) even transgenerationally.(61) Importantly, organotin compounds also bind
and activate the RXRα receptor, establishing them as dual PPARγ:RXRα agonists.(62,63)
However, the obesogenic (i.e. adipogenic) effects of TBT require PPARγ.(60)
Over the past several years, numerous other environmental contaminants have
been identified as PPARγ agonists. These include phthalate metabolites (e.g. mono-ethyl
hexyl phthalate (MEHP), mono-ethyl tetra-bromo phthalate (METBP);(64–66) brominated
flame retardants (tetrabromo bisphenol-a);(67) and fungicides (triflumizole,
triphenyltin).(58,68) PPARγ is a recognized toxicological target in the EPA’s ToxCast and
Tox21 high-throughput screening initiatives, and a major component of the adipogenic
ToxPi category.(69,70) The assays in ToxCast and Tox21 are designed to identify
compounds that activate or interact with PPARγ along mechanistic pathways related to
adipogenesis, and in the case of ToxPi, prioritize compounds by potential for endocrine
11
disrupting toxicity. As the biology of PPARγ activation and chemical adipogenesis
becomes further characterized, these screening assays will be increasingly useful in
identifying important links between environmental chemical exposures and adverse
effects related to fat metabolism (though the utility of PPARγ activation screens has
recently come under criticism).(71)
Obesogens and bone
PPARγ is expressed in bone marrow mesenchymal stromal cells (BM-MSCs), as
well as bone resorbing osteoclasts.(72) Research into the consequences of PPARγ
activation BM-MSCs identified an adverse pathway of transdifferentiation, whereby
osteoblastogenesis was replaced by adipogenesis in a PPARγ-dependent manner.(15,73) In
brief, activation of PPARγ by an exogenous ligands, such as rosiglitazone, suppresses the
activity of key osteogenic transcription factors (transrepression), and upregulates
adipogenic genes. Consequently, the increased marrow adiposity and decreased
mineralization accelerate an “aging-like phenotype” of lower bone quality.(74–76) This
phenotype has been observed in numerous clinical studies of older individuals taking
glitazone drugs, and represents a significant off-target effect for diabetes therapeutics that
target PPARγ.(77–79)
With the identification of environmental PPARγ ligands such as tributyltin, bone
has been identified as a target of environmental obesogens.(60,61,80–82) This mechanism is
investigated in depth in Chapters II and III. To provide a basis for the disruption of bone
12
homeostasis by these compounds, an overview of bone biology and endogenous factors
that control bone development is presented below.
Relevant aspects of bone biology
Bone structure and turnover
Bone tissue can be divided into two types of macroscopic structure: cortical and
cancellous bone. Cortical bone is dense and found in greater proportion in long bones
(such as the femur and tibia), and makes up roughly 80% of the mass of the skeleton.
Cancellous bone is less dense and comprises only about 20% of skeletal mass, but forms
a network of branched and interconnected structures (trabeculae) that provide about 80%
of the skeletal surface area.(83) With the hard, cortical shell surrounding the network of
trabecular bone, the skeleton is able to meet and adapt to various demands – both
physical and biochemical – imposed on the body.
Bone is a dynamic tissue. Despite being composed predominately of hard,
hydroxyapatite mineral (consisting of calcium and phosphate) surrounding collagenous
protein, bone undergoes processes that alter its physical makeup: the distinct, cellmediated activities of bone growth (modeling) and coupled bone growth and resorption
(remodeling). Bone modeling is crucial in development and achieving peak bone mass, as
well as responding to physical stress to strengthen the mechanical properties of the
skeleton.(84,85)
The purpose of bone remodeling is two-fold. First, the skeleton must respond to
mechanical stress, from the microscopic to the catastrophic. The process of repairing
13
fractures requires bone first to be resorbed.(85) Second, the bone matrix serves as a storage
depot for calcium and phosphate. Release of these resources to serve various other
purposes in the body requires a dissolution of the mineralized tissue. Hence, bone
remodeling is in part under control of the endocrine system (e.g. parathyroid hormone
and steroid hormones, discussed below) and contributes to whole-body homeostasis.(86)
Bone homeostasis is a careful balance of the two processes of bone remodeling,
and pathologic states can result from an imbalance in formation and resorption. For
example, an imbalance that favors bone formation (e.g. by impairments to bone
resorption) may result in osteopetrosis (a.k.a. “marble bone disease”), which is associated
with a disproportionate number of fractures and arthritis, among other symptoms.(87,88)
Conversely, and perhaps more widely recognized, are conditions such as
osteopenia and osteoporosis, which are characterized by excessively low bone density
that can lead to fracture. This absence of bone can be caused by an impairment of bone
formation, an increase in bone resorption, or both. Accordingly, therapeutics for
osteoporosis, such as bisphosphonates and parathyroid hormone, attempt to alleviate the
loss of bone by targeting the remodeling process to shift in favor of formation.(89,90)
With advanced age (discussed below), bone remodeling inevitably shifts in favor
of resorption so that each cycle of remodeling results in a net loss of bone. Because the
balance of formation and resorption must be very tightly regulated in order to maintain
bone homeostasis throughout growth and development, and in the midst of mineral and
structural demands, control of the process is carried out locally by communication among
14
specialized, interacting cells. The three types of cells that determine bone remodeling are
osteoblasts, osteocytes, and osteoclasts.
Cellular functions in bone remodeling
Osteoblasts are responsible for generating the mineral matrix.(91,92) These cells are
derived from the mesenchymal lineage (BM-MSCs) and differentiate in response to local
signals that include soluble proteins from the Wnt family(93) as well as the bone
morphogenic protein (BMP) family.(94) The key transcription factor activated by these
signals (and other factors) is Runx2,(95) the activity of which is maintained at low levels
due to negative regulation by PPARγ.(96) Wnt signaling, for instance, can remove this
negative regulation(97,98) by upregulating the activity of beta-catenin, which allows for
Runx2-dependent transcription of osteogenic genes such as Sp7 (osterix)(99) (conversely,
PPARγ activation can transrepress Wnt signaling by preventing the nuclear localization
of beta-catenin).(100)
As the osteoblast matures, it begins to express proteins that provide conditions for
the nucleation of crystalline calcium and phosphate, as well as high expression levels of
collagen. Together, these structural elements make up the solid yet flexible bone tissue.
As the matrix develops, the mature osteoblast may line the bone surface in a quiescent
state, undergo apoptosis, or be embedded in the mineral where it fully matures into an
osteocyte.(101–103)
Osteocytes – which are by far the most prevalent cell type, making up ~90% of
bone cells – are readily identified by their distinctive shape, which features a number of
15
dendritic process that extend into the surrounding mineral.(104) This morphology allows
for osteocytes to serve as force transducers, sensing minute changes in the mechanical
stress on bones and responding to initiate remodeling. These cells orchestrate the
remodeling process via secretion of soluble factors that can promote mineralization (e.g.
Dmp1, MEPE, FGF23), or, in a negative feedback manner, inhibit formation and
stimulate resorption.(101–103,105,106)
The third type of cell – osteoclasts – are bone resorbing (dissolving) cells that
differentiate from bone marrow myeloid progenitors (of the hematopoietic lineage) and
fuse to form large, multinucleated cells. Expression of the transcription factor Nfatc1 is
necessary and sufficient to induce osteoclastogenesis, which leads to the expression of
downstream target genes specific to osteoclasts, such as Acp5/Trap (a phosphatase),(107)
Dc-stamp (involved in cell fusion),(108) and Ctsk (bone resorption),(109) along with
autoamplification of Nfatc1 to provide a positive feedback mechanism.(110–113)
Osteoclasts feature an outer membrane that tightly attaches to the bone surface to
create a sealed area. The membrane inside this sealed zone forms a ruffled border,
creating a greater surface area from which the cell can secrete the resorptive enzyme
cathepsin K,(109,114) along with protons to acidify the sealed resorption region. Dissolution
of the mineral surface produces visible resorption pits, the identification of which can be
used to assess osteoclast function both in vitro and in vivo. Digested materials, along with
the osteoclast specific protein tartrate resistant acid phosphatase 5b (TRAP5b), are
endocytosed, passed through the cell and ultimately secreted into the blood stream, where
they can be detected as biomarkers of bone turnover.(107,115) High concentrations of
16
calcium in the resorption area generated by the demineralization of the matrix leads to
osteoclast apoptosis and the cessation of bone resorption.(116)
The current paradigm of bone turnover asserts that mature osteoblasts are then
recruited to the newly formed resorption pits where mineralization can proceed. This
stimulation may come from soluble factors released from the mineral matrix during
resorption,(117) though direct communications between osteoclasts and osteoblasts also
stimulate the reversal phase of bone turnover.(118) These mechanisms and other
communication signals among the three cell types are discussed below.
Cell-cell communication in bone remodeling
Tight regulation of the remodeling process is carried out locally by
intercommunication by the three cell types. Perhaps the best characterized aspect of
turnover communication is the stimulation of osteoclast differentiation by osteoblastderived receptor activator of NFκB ligand (RANKL).(119,120) Expression of RANKL
protein in osteoblasts can be promoted by vitamin D and parathyroid hormone, among
other factors.(121) Current research suggests that membrane-bound RANKL is most
effective in inducing osteoclastogenesis (rather than soluble, secreted RANKL),
consistent with the notion that cell-cell contact is the main driver.(122) In osteoclasts,
expression of the receptor RANK is upregulated by the essential soluble macrophage
colony stimulating factor (M-CSF) (also produced by osteoblasts);(123) stimulation of
RANK by RANKL is the major upstream event that drives expression of Nfatc1.(124)
Osteoblasts also secrete an inhibitory factor to counter this communication signal:
17
osteoprotegerin (OPG).(120,125) OPG is a “decoy” receptor for RANKL, and thus interferes
directly with RANKL signaling. The ratio of RANKL expression to OPG expression
commonly is used as a measure of osteoclastogenic potential in osteoblasts, where lower
ratios are associated with larger measures of bone formation.
The matrix-entombed osteocytes can serve as negative regulators of
osteoblastogenesis via soluble factors. One such factor is sclerostin (Sost), which inhibits
the Wnt pathway by preventing a necessary dimerization between the Frizzled and LRP
receptors to which Wnt proteins bind.(106,126) Genetic deletion of Sost in mice leads to
high bone mass, and loss of sclerostin in humans may lead to multiple types of highdensity bone disorders.(127,128) The regulation of sclerostin expression provides a
mechanism by which exogenous and endogenous signals can couple with bone
metabolism: Sost expression is negatively correlated with mechanical strain in osteocytes,
linking anabolic effects of bone with stress recovery.(129) Parathyroid hormone also
downregulates Sost in osteocytes, allowing for a recovery of bone mass following
increased osteoclast activity.(130)
Osteocytes are also known to express RANKL and OPG, as well as M-CSF, to
directly control osteoclast differentiation. This occurs during osteocyte apoptosis in order
to recruit osteoclasts to the resorption surface, as well as during states of mechanical
unloading.(131,132) Thus, osteocytes serve as regulators of both bone formation and
resorption via secreted signaling mechanisms.
The communication from osteoclasts to osteoblasts is less clearly understood.
Following resorption is a “reversal” phase of bone remodeling, in which mature
18
osteoblasts replace the lost bone.(133) This requires a stimulatory signal to osteoblasts.
Conditions of osteopetrosis resulting from non-resorbing (but present) osteoclasts (i.e.
osteoclast-rich osteopetrosis) indicated that some form of pro-osteogenic signaling was
being produced by osteoclasts even in the absence of resorption.(134)
Among these signals are excreted Il-6 family cytokines which include the protein
cardiotrophin-1 (CT-1).(135,136) CT-1 acts via membrane-bound receptors on osteoblasts to
stimulate Runx2.(135) On the other hand, mature osteoclasts also express inhibitory
proteins, such as Sema4d, that act via cell-cell contact to prevent osteoblast activity.(137)
The inhibitory signaling from direct contact and stimulatory signaling from soluble
factors may serve to establish a temporal and spatial separation between resorption and
formation, as well as potentially protect osteoblasts from the highly acidified
microenvironment.
All three cell types display supportive and inhibitory communication signals
toward one another, the above mechanisms being only a few of the major examples. The
presence of redundant and feedback-responsive mechanisms highlights the importance of
careful regulation of bone remodeling to maintain homeostasis. However, even under
normal conditions of bone homeostasis, multiple physiological factors can modify this
balance.
Age and sex as co-determinants of bone homeostasis
The age and sex of an animal are highly influential in determining the rate of bone
turnover and thus macroscopic bone structure. During early development (before peak
19
bone mass is achieved), bone deposition occurs along the outside (periosteal) surface of
long bones, while resorption occurs along the inside (endosteal) surface.(138,139) Although
the thickness of the cortical shell decreases over the course of this process, the net result
is a widening of bone, providing greater structural stability and resistance to
bending.(140,141) These changes in bone geometry are generally more pronounced in
males, specifically periosteal apposition,(138,139,142–144) and are largely established during
puberty.(144) Endocortical resorption is similar between males and females during growth,
as is bone mineral density normalized to bone volume (volumetric bone mineral
density).(139,145) Thus, males tend to have stronger bones and lower rates of fracture
because of greater bone apposition producing larger bones. Decreases in the apposition of
bone on the periosteal surface during the growth phase therefore may attenuate peak bone
mass, exacerbate the consequences of bone loss, and increase the risk of fractures later in
life.(143,146)
Once peak bone mass has been achieved in young adulthood, apposition slows but
resorption continues and remodeling becomes an overall negative process. During
development, males had greater bone growth in the trabecular compartment. Thus, the
loss of trabecular bone continues for a longer period compared to females; the increased
surface area provides for more available bone to be resorbed. In males, trabeculae
generally thin with age, whereas in females, the trabeculae become perforated and
completely resorbed.(139,147) As the trabeculae disappear, bone resorption occurs to a
larger extent along the endosteal surface.(139)
In the bone marrow, the number of mature adipocytes increases with age,
20
resulting in “fatty marrow”.(148–150) This is due to a greater potential of BM-MSCs in aged
animals to differentiate into adipocytes instead of osteoblasts, and not just a
compensatory filling of space in the absence of bone.(151) Accordingly, osteoporosis in
older individuals is associated with increased marrow fat and greater efficiency of BMMSC differentiation into adipocytes.(150,152) Basal levels of adipogenic genes are higher in
BM-MSCs derived from older (20 month-old) mice compared to younger (6 month-old)
mice, as are levels of PPARγ2 mRNA.(148) Older animals are therefore more sensitive to
PPARγ-mediated suppression of bone formation, and inappropriate activation of PPARγ
results in an aged-like phenotype.(75,150) Female mice also appear to be more susceptible
to marrow adiposity, and BM-MSC cultures derived from female mice show an increased
adipogenic response to a PPARγ agonist (rosiglitazone).(153) Thus, this combination
makes older individuals more at risk for the side effects of thiazolidinedione anti-diabetic
drugs (and putative environmental chemicals) that activate PPARγ.
Sexual dimorphism in bone due to hormonal effects
The steroid hormones estrogen and androgens (e.g. testosterone) play major roles
in bone homeostasis as well, and varied levels of these hormones between males and
females have consequences for long-term bone remodeling. The importance of estrogen
loss in post-menopausal osteoporosis has been hypothesized for many decades,(154) and
indeed many of the experimental studies assessing the role of estrogen and the estrogen
receptor characterize functional losses of estrogen signaling and the subsequent
promotion of bone loss. Although these hormones have indirect effects on bone that are
21
mediated by other organ systems, only direct effects on bone turnover are discussed here.
The major effect of estrogen in bone is a slowing of remodeling, which in turn
slows the net resorptive effect of remodeling that occurs in aging bone. This mechanism
has been demonstrated by loss-of-estrogen models that show simultaneous increases in
osteoclast and osteoblast precursors.(155) Estrogen levels are negatively correlated with
levels of RANKL expression on BM-MSCs,(156) and estrogen upregulates OPG.(157) Thus,
suppression of osteoclast differentiation can be accomplished by decreasing the
RANKL:OPG ratio in osteoblasts.
Steroid hormones prolong the lifespan of mature osteoblasts(158) and can help
maintain trabecular bone density.(143) Interestingly, the estrogen receptor has estrogendependent and estrogen-independent effects on bone. Independent of its ligand, the
estrogen receptor itself is essential in osteoblast-mediated periosteal bone formation,
where it stimulates canonical Wnt pathways in osteogenesis.(159) On the other hand,
estradiol-liganded estrogen receptor can inhibit Wnt signaling and attenuate
osteoblastogenesis,(159,160) consistent with an overall suppression of bone remodeling.
Orchidectomy in rats causes increased bone turnover and bone loss that can be
ameliorated by testosterone treatment,(161) and inactivation of the androgen receptor
specifically in osteoblasts results in a decrease in trabecular structure.(162,163) More
precisely, the role of the activated androgen receptor appears to be to support periosteal
growth in differentiating – rather than mature and mineralizing – osteoblasts.(162)
Androgens also suppress RANKL (in vitro),(164) but suppress OPG as well.(165) Thus the
effect of androgens on bone is largely pro-osteoblastic in nature and serves to maintain
22
mineral content and osteoblast population.
Because of these sex- and age-linked differences, parsing the interplay of
mechanisms in bone remodeling requires careful consideration of the appropriate
experimental model, and it is often difficult to hypothesize as to the full suite of effects
that lead to an observed phenotype if the physiologic state of the animal is not controlled
for. For instance, the use of an immature mouse model may allow for observations to be
made on the effects of toxicant exposure on bone apposition that decrease peak bone
mass; mature mice may be used to study the effects of increased bone resorption,
particularly in the context of ovariectomy where the influence of estrogen is removed.
The sensitivity of a toxicant’s effect may differ between males and females, and certain
phenotypes may only be observed in light of sex-specific background phenotypes.
Regardless of the model, the tight communication within the remodeling
environment indicates that perturbation of one factor may lead to dynamic instability of
numerous other factors to upset the formation/resorption balance. It is in this context that
we study the effect of bone toxicants as adverse agents that act to not simply suppress or
stimulate one cell type, but rather to unbalance a dynamic system.
Environmental bone toxicants
The complex process of bone cell differentiation and communication provides a
number of targets for toxicants to adversely affect bone homeostasis. Several different
groups of chemicals provide examples of both naturally occurring and anthropogenic
compounds that have toxic mechanisms in bone.
23
Lead
Lead (Pb) has long been known to impact skeletal development. Because of a
chemical similarity to calcium, lead persists in the body by accumulating in bone
(particularly trabecular bone),(166,167) where its half-life is estimated to be 10–30 years.(83)
Because lead is so preferentially distributed to this tissue (accounting for ~95% of the
lead body burden),(83) cumulative lead exposure metrics typically involve non-invasive
measurements of bone lead rather than circulating blood lead levels. However, during
periods of high calcium demand, such as pregnancy, lactation or menopause,(166,168–170)
bone is resorbed and lead is mobilized into the blood.(171) Under these conditions, inverse
associations between blood lead levels and bone mineral density have been found.(166)
Mechanistically, lead exposure has been tied to suppression of osteoblast
differentiation and activity, and subsequent impairment of bone formation.(172,173) This is
accomplished by an upregulation of sclerostin and suppression of Wnt signaling, leading
to an increase in adipogenesis in the bone marrow.(174,175) Lead also suppresses the
response to osteogenic stimulation factors such as vitamin D and insulin-like growth
factor (IGF-1).(176) Interestingly, lead also has been shown to upregulate aryl hydrocarbon
receptor (AHR, discussed below) expression in osteoblasts, potentially increasing their
sensitivity to xenobiotics.(177) Ultimately, lead exposure reduces overall bone mass,(174)
leading to associations with decreases in skeletal growth that have been observed in
epidemiologic studies.(178,179) On the contrary, recent experimental data in a mouse model
have shown increases in bone volume early in life owing to a lead-induced suppression of
osteoclast resorptive activity.(180)
24
Arsenic
Epidemiologic studies of arsenic (As) toxicity have not typically focused on
specific bone structural outcomes, but rather have focused on developmental exposurerelated growth impairment and indirect effects related to bone.(181) In its inorganic
speciation, arsenic replaces phosphate in bone,(182) and the toxic effects of inorganic
arsenic on the bone marrow leading to anemia are well established.(183)
In vivo, arsenic exhibits a toxic effect on osteoblasts in rodent models where the
exposure mimics human drinking water exposure.(184) The phenotype resulting from this
exposure was characterized by decreases in bone mineral density, trabecular volume, and
cortical thickness in the tibia. This was shown to be related to suppression of Runx2 in
osteoblasts via upregulation of a specific kinase activity (shown independently to
suppress Runx2)(184) and activation of oxidative stress pathways.(185) In an in vitro
osteoclast model, inorganic arsenic induced differentiation of osteoclasts at
concentrations comparable to those observed in human biomonitoring, suggesting that
arsenic exposure could contribute to the current prevalence of low bone density
disorders.(186)
Cadmium
One of the earliest connections between cadmium (Cd) and bone disease was
found among a Japanese cohort in 1950’s suffering from itai-itai disease, a bone
pathology characterized in part by brittle and weak bones. Incidence of this disease was
linked to consumption of cadmium-contaminated rice.(187) Although in this case
25
cadmium’s critical target was the kidney, where high exposures altered vitamin D
metabolism and indirectly affected bone homeostasis, low levels of cadmium exposure
have recently been associated with low bone density and risk of osteoporosis, and animal
models have established that cadmium can act directly on bone to induce bone loss.(188–
190)
Cell culture models have shown that cadmium can act on both osteoblasts and
osteoclasts, but at different dose ranges. For instance, at higher (micromolar)
concentrations, cadmium suppresses osteoblast activity, indicated by decreases in
alkaline phosphatase activity and decreased nodule formation.(190) At nanomolar
concentrations, cadmium stimulates the differentiation and resorptive function of
osteoclasts directly.(191) Thus, cadmium is mechanistically capable of perturbing the
normal differentiation and function of both cell types, as well as their communication, in
a manner consistent with a net loss of bone.
Tungsten
Tungsten (W) can accumulate and persist in bone,(192,193) though it has only
recently been investigated as a bone toxicant.(194) In vivo, tungsten does not appear to
suppress osteogenesis, but it does increase bone marrow adipocyte content specifically in
younger male mice, a difference partially attributed to a lower baseline level of adipocyte
content in males. It is hypothesized that bone turnover facilitates accumulation of
tungsten in bone, as older male and female mice had substantially lowered tungsten
concentrations following exposure.(194)
26
Interestingly, tungsten alone does not increase adipogenesis in in vitro BM-MSC
cultures, but it does enhance PPARγ-mediated adipogenesis.(194) This suggests that
tungsten’s target molecule is upstream of PPARγ activation. Thus, co-exposures to
tungsten and PPARγ-activating compounds – a plausible scenario given tungsten’s use in
implanted medical devices(195) – could potentially exacerbate age-related bone loss.
Ethanol
Ethanol may not strictly fit the definition of environmental toxicant inasmuch as
exposure is typically intentional rather than inadvertent, but given the prevalence of
alcohol use it remains and important toxicant of note. The toxicity of ethanol to the
skeleton is somewhat underappreciated, particularly in skeletally vulnerable populations
such as women post-pregnancy,(196) though it is well-recognized that chronic alcoholism
is associated with osteoporosis and risk of fracture.(197) This loss of bone has been
attributed to both inhibition of bone formation and increases in bone resorption.
The central mechanism associated with ethanol’s effect on bone cells is oxidative
stress. Several models have shown that ethanol-induced oxidative stress effectively
inhibits Wnt signaling in mesenchymal stromal cells,(196) which is accompanied by a
reciprocal increase in PPARγ-mediated suppression of osteoblast differentiation. This can
be reversed with antioxidant treatment(196) as well as estrogen.(198) Reactive oxygen
species generated by ethanol’s mechanism of action in mature osteoblasts increases
RANKL expression, hence increased osteoclastogenesis in alcohol-exposed animals.(199)
27
Polycyclic Aromatic Hydrocarbons (PAHs)
PAHs are a class of chemicals generated from incomplete combustion and for
industrial manufacturing uses. Many of the toxic effects of PAHs are mediated by the
aryl hydrocarbon receptor (AHR), which is expressed in both osteoclasts(200) and
osteoblasts,(177) and much of the research on AHR-mediated bone effects has been driven
by the observation that cigarette smoking (a significant source of PAHs) is a risk factor
for low bone density.(201)
For example, benzo[a]pyrene (BaP) and 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD), both found in cigarette smoke, increase osteoclastogenesis and bone resorption
through a mechanism involving characteristic Cyp1 enzyme induction by AHR
activation.(202) However, it is not known as to how Cyp1 enzymes are sufficient to
stimulate osteoclast differentiation. Osteoclast-specific AHR knockout models show
resistance to ovariectomy-induced bone loss, establishing the importance of AHR in
osteoclast function and overall bone homeostasis.(203) Osteoblasts are also sensitive to
AHR-mediated effects on differentiation, as concentrations of TCDD as low as 100
femtomolar are sufficient to suppress expression of alkaline phosphatase (Alp) and the
maturation marker osteocalcin (Ocn).(204)
It is clear from these examples that bone homeostasis is sensitive to exogenous
stimuli, and a large variety of environmental chemicals can impact bone health via
different mechanisms. The general thesis of this dissertation is that environmental
chemicals can adversely affect bone homeostasis by activating nuclear receptors in bone
cells to perturb cellular differentiation and function. Three studies are presented: the first
28
two studies (Chapters II and III) investigate mechanisms of nuclear receptor-mediated
toxicity in bone, with an emphasis on PPARγ:RXRα activation, using in vitro and in vivo
approaches. The third study (Chapter IV) characterizes approaches to predicting
combinatorial effects of PPARγ ligands in complex mixtures. Last, Chapter V describes
limitations and future directions of each study, and suggests implications of this work for
public health in general.
29
CHAPTER II: Structurally-diverse, PPARγ-activating environmental toxicants
induce adipogenesis and suppress osteogenesis in bone marrow mesenchymal
stromal cells
James Watt and Jennifer J. Schlezinger
Department of Environmental Health, Boston University School of Public Health,
Boston, MA, 02118, USA
Citation:
Watt J and Schlezinger JJ. Structurally-diverse, PPARg-activating environmental
toxicants induce adipogenesis and suppress osteogenesis in bone marrow mesenchymal
stromal cells. Toxicology 2015;331:66–77.
Acknowledgements
This work was supported by the Superfund Research Program [P42ES007381] and the
National Institute of Environmental Health Science at the National Institutes of Health
[R21ES021136].
The authors would like to thank Ms. Faye Andrews for her superb technical assistance.
Abbreviations: BM-MSC: bone marrow mesenchymal stromal cell; DEHP: di-(2ethylhexyl)phthalate; DMP1: dentin matrix phosphoprotein; MEHP: mono-(2ethylhexyl)phthalate; METBP: mono-(2-ethylhexyl)tetrabromophthalate; MTT: 3-[4,5dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; OSX: osterix; PLIN1: perilipin;
pNPP: p-nitrophenyl phosphate; peroxisome proliferator activated receptor γ (PPARγ);
RN18S: 18s ribosomal RNA; RUNX2: runt-related transcription factor 2; TBBPA:
tetrabromobisphenol-a; TBT: tributyltin; TPhT: triphenyltin; WNT10b: wingless-type
MMTV integration site family, member 10b.
30
Abstract
Environmental obesogens are a newly recognized category of endocrine disrupting
chemicals that have been implicated in contributing to the rising rates of obesity in the
United States. While obesity is typically regarded as an increase in visceral fat, adipocyte
accumulation in the bone has been linked to increased fracture risk, lower bone density,
and osteoporosis. Exposure to environmental toxicants that activate peroxisome
proliferator activated receptor γ (PPARγ), a critical regulator of the balance of
differentiation between adipogenesis and osteogenesis, may contribute to the increasing
prevalence of osteoporosis. However, induction of adipogenesis and suppression of
osteogenesis are separable activities of PPARγ, and ligands may selectively alter these
activities. It currently is unknown whether suppression of osteogenesis is a common
toxic endpoint of environmental PPARγ ligands. Using a primary mouse bone marrow
culture model, we tested the hypothesis that environmental toxicants acting as PPARγ
agonists divert the differentiation pathway of bone marrow-derived multipotent
mesenchymal stromal cells towards adipogenesis and away from osteogenesis. The
toxicants tested included the organotins tributyltin and triphenyltin, a ubiquitous
phthalate metabolite (mono-(2-ethylhexyl) phthalate, MEHP), and two brominated flame
retardants (tetrabromobisphenol-a, TBBPA, and mono-(2-ethylhexyl)
tetrabromophthalate, METBP). All of the compounds activated PPARγ1 and 2. All
compounds increased adipogenesis (lipid accumulation, Fabp4 expression) and
suppressed osteogenesis (alkaline phosphatase activity, Osx expression) in mouse
primary bone marrow cultures, but with different potencies and efficacies. Despite
31
structural dissimilarities, there was a strong negative correlation between efficacies to
induce adipogenesis and suppress osteogenesis, with the organotins being distinct in their
exceptional ability to suppress osteogenesis. As human exposure to a mixture of toxicants
is likely, albeit at low doses, the fact that multiple toxicants are capable of suppressing
bone formation supports the hypothesis that environmental PPARγ ligands represent an
emerging threat to human bone health.
32
Introduction
The bone marrow is a specialized microenvironment containing both bone and
adipose cells. Because the bone tissue is undergoing near constant homeostatic
remodeling, any perturbation in the balance of adipogenesis and osteogenesis could lead
to alterations in bone maintenance. Bone ageing is associated with increased adiposity,
reduced osteoblast function, and increased osteoclast activity, giving rise to osteoporotic
pathologies such as reduced bone mass, altered bone structure, and increased risk of
fracture.(150) Bone loss that occurs with aging also is associated with an increase in
peroxisome proliferator active receptor γ (PPARγ) expression in the marrow.(148)
The plasticity of bone marrow multipotent mesenchymal stromal cells (BMMSCs) allows them to differentiate into either adipocytes or osteocytes. Adipocyte
formation is dependent on PPARγ activation(27) whereas bone formation is chiefly
controlled by runt related transcription factor 2 (Runx2).(95) MSC differentiation is
controlled by the balance of PPARγ and Runx2 transcriptional activation and coregulated by signaling through the Wnt/β-catenin pathway, which modulates the
expression and function of both PPARγ and Runx2.(96,97,100,205)
Activation of PPARγ by the therapeutic ligand rosiglitazone drives differentiation
toward adipogenesis and suppresses osteogenesis, a result observed both in vitro and in
vivo.(15,74) Conversely, decreasing expression of PPARγ (e.g. by molecular knockdown)
results in reduced adipogenesis and increased bone mass.(16) Human treatment with
therapeutic PPARγ ligands (e.g. thiazolidinediones) is associated with an increased risk
of fracture in bone.(206–208)
33
While there is a reciprocal relationship between bone formation and adipogenesis
in response to PPARγ activation by rosiglitazone, these functions of PPARγ (along with
insulin sensitization) are distinct and separable. For example, Rahman et al. (2012)
demonstrated the anti-osteogenic capability of PPARγ to be independent of its proadipogenic activity.(209) Phosphorylation of PPARγ can increase insulin sensitivity
independently of adipogenesis.(210) PPARγ ligands can selectively activate the multiples
functions of PPARγ, with the distinct abilities to activate adipogenesis and suppress
osteogenesis not necessarily being correlated.(73,211,212)
A growing number of environmental contaminants, including organotins and
phthalates, are being recognized for their ability to activate PPARγ and therefore are
members of the environmental obesogen class of toxicants.(57) While organotins
primarily have been used as antifouling agents and fungicides, their pervasive distribution
is indicated by their presence in house dust.(213) Significant human exposure is indicated
by the presence of organotins in liver and blood (0.1–450 nM).(214) Despite being
structurally distinct from several other PPARγ ligands, multiple organotins are capable of
activating PPARγ and its heterodimerization partner, retinoid X receptor, and act as
potent and efficacious adipogenic agents in pre-adipocyte and MSC models.(59,80,82)
Phthalates are well-known environmental PPARγ ligands.(51,64) Over 18 billion
pounds of phthalates are produced yearly, and humans are regularly exposed to
significant levels (~10 μg/kg bw/day) of di-(2-ethylhexyl) phthalate (DEHP).(215)
Substantial exposures occur during acute medical procedures, resulting in blood DEHP
concentrations ranging from 50–350 μM.(216) MEHP, the active metabolite of DEHP,
34
directly activates PPARγ and promotes adipogenesis in NIH 3T3L1 cells(51,64,65) and has
been measured in human blood samples at μM concentrations.(217)
Newly recognized environmental PPARγ ligands include tetrabromobisphenol-A
(TBBPA)(67) and mono-(2-ethylhexyl)tetrabromophthalate (METBP)(66). TBBPA is a
component of the mass produced brominated flame retardants (over 150,000 tons
annually),(218) is detectable in home and office dust samples,(219,220) and in human breast
milk and serum (at levels as high as 649 ng/g lipid weight).(221) Di-(2ethylhexyl)tetrabromophthalate, a component of Firemaster® 550, has been found at ppm
levels in house dust and at 260 ng/g lipid weight in humans,(222) and its metabolite
METBP is capable of increasing lipid accumulation as well as activating expression of
the PPARγ gene target fatty acid binding protein 4 (Fabp4) in NIH 3T3-L1 cells.(66)
PPARγ-mediated suppression of bone formation and the substantial human
exposure to multiple environmental PPARγ ligands highlight the need to understand their
contribution to bone loss. Furthermore, it is unknown if the pro-adipogenic effects of
these environmental PPARγ ligands are associated with or are distinct from antiosteogenic effects, an important question because divergent effects have been shown for
some natural and synthetic ligands.(73,211,212) Studies described herein were designed to
test the hypothesis that environmental PPARγ ligands, in general, selectively induce
adipogenesis at the expense of osteogenesis. Accordingly, we examined the potential for
a structurally diverse set of environmental PPARγ ligands (TPhT, MEHP, TBBPA,
METBP) to induce adipogenesis and suppress osteogenesis in mouse-derived bone
marrow MSCs (BM-MSCs) in comparison to PPARγ ligands known to suppress bone
35
formation (rosiglitazone and tributyltin (TBT)). Surprisingly, despite the structural
dissimilarities of the ligands, the efficacy of osteogenesis suppression by treatment was
strongly negatively correlated with efficacy of induction of adipogenesis. The organotins
were distinct in their exceptional ability to suppress osteogenesis. The data are consistent
with the conclusion that suppression of osteogenesis is a common toxic effect of
environmental toxicants that are capable of activating PPARγ.
Materials and Methods
Materials
Rosiglitazone was from Cayman Chemical (Ann Arbor, MI). DMSO was from
American Bioanalytical (Natick, MA). Insulin, Nile Red, p-nitrophenyl phosphate
(pNPP) reagent, TBT chloride, TPhT chloride, and TBBPA were from Sigma-Aldrich
(St. Louis, MO). MEHP was from TCI America (Portland, OR). METBP was
synthesized by AsisChem (Waltham, MA). All other reagents were from Thermo Fisher
Scientific (Suwanee, GA).
Cell culture
Bone marrow was isolated from 9-week-old male C57BL/6J mice (Jackson
Laboratories, Bar Harbor, ME). All animal studies were reviewed and approved by the
Institutional Animal Care and Use Committee at Boston University. All animals were
treated humanely and with regard for alleviation of suffering. Mice were housed 4 per
cage, with a 12 hour light cycle. Water and food (2018 Teklad Global 18% Protein
36
Rodent Diet, Irradiated, Harlan Laboratories, Indianapolis, IN) were provided ad libitum.
Animals were euthanized for collection of bone marrow two days after arrival. After
euthanasia (cervical dislocation under terminal euthanasia followed by pneumothorax),
limbs were aseptically dissected, and soft tissue was removed from the bone. Marrow
was flushed from the femur, humerus, and tibia bones, strained through a 70 um cell
strainer, and suspended in MSC medium consisting of α-MEM, 10% fetal bovine serum,
100 U/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml amphotericin-B. Cells
from 2–4 animals were pooled and plated so that each pool represented an experimental
n. Cells were plated either in 6-well (12 million cells in 2 mls per well) or 12-well plates
(6 million cells in 1 ml per well). Half of the medium was replaced 5 days after plating.
At day 7, the medium was replaced with osteoinductive medium consisting of α-MEM,
12.5 μg/ml l-ascorbate, 8 mM β-glycerol phosphate, 0.5 μg/ml insulin, and 10 nM
dexamethasone. Prior to the day 7 medium change, a naïve, undifferentiated well was
harvested for gene expression analysis as described below. Cells received no treatment
(Naïve) or were dosed with vehicle (DMSO, 0.1% final concentration) or with the
chemical of interest. Following treatment, cells were cultured for 7 days (mRNA
expression) or 10–11 days (adipocyte and bone phenotype, cell viability). During these
periods, medium was changed and the cultures were re-dosed 2 times for mRNA
expression or 3 times for phenotype analysis.
37
Cell viability assay
Viability was assessed by 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium
bromide (MTT) labeling for 3 hrs by standard methods. Absorbance measurements were
determined using a Synergy2 plate reader (BioTek, Winooski, VT). Absorbance in
experimental wells was normalized by dividing by the absorbance in untreated cultures
and reported as “Fold Change from Medium.”
Lipid accumulation
Cells were washed with one volume of PBS and incubated with Nile Red (1 μg/ml
in PBS). Fluorescence (excitation 485 nm, 20 nm bandwidth; emission 530 nm, 25 nm
bandwidth) was measured using a Synergy2 plate reader. The fluorescence in all
experimental wells was normalized by subtracting the fluorescence measured in Naïve
cells (those that received osteogenic medium alone) and reported as “RFUs.”
Osteogenesis assays
Following Nile Red staining, cells were rinsed with PBS and fixed in
paraformaldehyde (2% in PBS). To quantify alkaline phosphatase activity, cells were
incubated in pNPP solution. After quenching with NaOH (final concentration: 0.75M),
absorbance (405 nM) was measured using a Synergy2 multifunction plate reader. The
absorbance in all experimental wells was normalized by dividing by the absorbance
measured in wells that received osteogenic medium but were not treated and reported as
“Fold Change from Medium.” Following the pNPP assay, cells were stained with
38
Alizarin Red (Osteogenesis Quantitation Kit, Millipore, Billerica, MA). Cells were
extensively washed and then photographed using the UVP Bioimaging System (UVP,
Inc., Upland, CA). The resulting images were analyzed for bone nodule count using
Image-Pro Plus (MediaCybernetics, Bethesda, MD), and the number of nodules per
square cm is reported. Following image capture, Alizarin Red staining was quantified as
indicated in the manufacturer’s instructions. Absorbance (405 nM) was measured and
normalized as described above.
Gene expression analysis
Total RNA was extracted and genomic DNA was removed using the RNeasy Plus
Mini Kit (Qiagen, Valencia, CA). cDNA was prepared from total RNA using the
GoScript™ Reverse Transcription System (Promega), with a 1:1 mixture of random and
Oligo (dT)15 primers. All qPCR reactions were performed using the GoTaq® qPCR
Master Mix System (Promega). Validated primers were purchased from Qiagen (see
Supplemental Table 1). qPCR reactions (in duplicate) were performed using a 7300 Fast
Real-Time PCR System (Applied Biosystems, Carlsbad, CA): Hot-Start activation at
95°C for 2 min, 40 cycles of denaturation (95°C for 15 sec) and annealing/extension
(55°C for 60 sec). Relative gene expression was determined using the Pfaffl method to
account for differential primer efficiencies.(223) The Cq value for 18s ribosomal RNA
(Rn18s) was used for normalization. The Cq value for naïve, undifferentiated cultures
was used as the reference point, and the data are reported as “Fold Change from Naive.”
39
Protein expression
Cells were collected and lysed in Cell Lysis Buffer (Cell Signaling Technology,
Beverly, MA) followed by sonication. The lysates were cleared by centrifugation, and
the supernatants were used for protein expression analyses. Protein concentrations were
determined by the Bradford method. Total proteins (20 µg) were resolved on 10%
polyacrylamide gels, transferred to a 0.2 µm nitrocellulose membrane, and incubated
with monoclonal rabbit anti-perilipin (3470, Cell Signaling Technology (Beverly, MA)).
Immunoreactive bands were detected using HRP-conjugated goat-anti rabbit secondary
antibodies (Biorad, Hercules, CA) followed by enhanced chemiluminescence. To control
for equal protein loading, blots were re-probed with a β-actin-specific antibody (A5441,
Sigma).
Reporter assays
Cos-7 cells (in 96 well plates) were transiently transfected with vectors containing mouse
Pparg1 (plasmid 8886; Addgene, Cambridge, MA) or mouse Pparg2 (plasmid 8862;
Addgene), with human RXRA (plasmid 8882; Addgene),(27) PPRE x3-TK-luc (plasmid
1015; Addgene),(224) and CMV-eGFP using Lipofectamine 2000 (Invitrogen, Carlsbad,
CA). Transfected cultures were incubated overnight. The medium was replaced, and
cultures received no treatment (Naïve) or were treated with Vh (DMSO, 0.1%),
rosiglitazone, TBT, TPhT, METBP, MEHP, TBBPA at concentrations ranging from 0.1
nM to 400μM and incubated for 24 hrs. Cells were lysed in Glo Lysis Buffer (Promega,
Madison, WI), to which Bright Glo Reagent (Promega) was added. Luminescence and
40
fluorescence were determined using a Synergy2 plate reader. Luminescence was
normalized by the GFP fluorescence in the same well. The normalized luminescence for
each well was then divided by the normalized luminescence measured in transfected but
untreated wells to determine the “Fold Change from Naive.”
Statistical Analyses
Statistical analyses were performed with Prism 5 (GraphPad SoftwareInc., La
Jolla, CA). Data are presented as means  standard error (SE). Gene expression data
were log transformed prior to analysis. One-way ANOVAs with the Dunnett’s post hoc
test and Pearson’s correlations were performed where noted. All analyses were
performed at α = 0.05.
Results
Activation of PPARγ1 and PPARγ2
Whereas each of the toxicants being investigated in this study has been identified
as a PPARγ ligand, they have not been characterized for their potency and efficacy all in
the same study. Thus, we began by characterizing the dose-response of each individual
ligand with respect to activation of mouse PPARγ1 and PPARγ2 using a PPRE-luciferase
reporter in Cos-7 cells (Fig. 2.1). Curve maxima, EC50s, and Hill coefficients are
reported in Table 1. Rosiglitazone, the positive control, was potent and efficacious at
activating PPARγ1 and 2. The organotins, TBT and TPhT, had similar potencies and
41
efficacies as rosiglitazone. MEHP, TBBPA, and METBP each acted as partial agonists,
with lower efficacies and potencies compared to rosiglitazone.
Induction of adipogenesis in bone marrow derived MSCs
Our studies and others have demonstrated that both rosiglitazone and TBT are
potent inducers of adipogenesis in the mouse-derived bone marrow MSC lines BMS2 and
U33/γ2, as well as adipose-derived stromal cells.(60,73,80) Furthermore, rosiglitazone has
been shown to concurrently suppress osteogenesis as it induces adipogenesis.(15) Here,
we tested the hypothesis that environmental PPARγ agonists would divert differentiation
toward adipogenesis in primary bone marrow MSCs grown in osteogenic conditions.
We established the maximal adipogenic response in our system using two known
differentiation modulators, rosiglitazone and TBT. Primary bone marrow cultures were
prepared from male C57BL/6J mice and treated with Vh (DMSO, 0.1%), TBT (100 nM),
or rosiglitazone (100 nM) in the presence of osteoinductive medium. Cultures were
assessed for lipid accumulation (10–12 days) and adipogenic gene expression (7 days).
As expected, TBT and rosiglitazone induced significant lipid accumulation compared to
Vh-treated cells (Fig. 2.2a), although TBT was less efficacious than rosiglitazone.
Whereas TBT and rosiglitazone did not induce expression of Pparg1/2, they strongly
induced its target genes Fabp4 and perilipin (Plin1) (Fig. 2.2b–2.2d).
Next, we tested the ability of several structurally distinct environmental PPARγ
ligands to induce adipogenesis in BM-MSCs. Established bone marrow cultures were
treated with Vh (DMSO, 0.1%), TPhT (10, 50, 80 nM), MEHP, METBP or TBBPA (10,
42
20 μM), in the presence of osteoinductive medium. The high concentration in each case
was the maximal sub-toxic concentration in this model (Supplemental Fig. 2.1). At all
concentrations, TPhT significantly increased lipid accumulation compared to Vh-treated
cells, with an efficacy similar to 100 nM TBT (Fig. 2.3a). While there was only a trend
toward an increase in expression of Pparg1/2 mRNA, expression of the PPARγ target
genes Fabp4 and Plin1 was increased significantly by all concentrations of TPhT (Fig.
2.3a). A similar pattern of effect was observed with MEHP and TBBPA, with significant
increases occurring in lipid accumulation and PPARγ-target gene expression (Fig. 2.3b
and 2.3d). METBP showed more limited increases in lipid accumulation, Fabp4
expression and Plin1 expression (Fig. 2.3c).
Terminal adipocyte differentiation was confirmed by analyzing perilipin protein
expression. Established cultures were treated with Vh (DMSO, 0.1%), rosiglitazone (100
nM), TBT (100 nM), TPhT (50 nM), MEHP, TBBPA or METBP (20 μM each), in the
presence of osteoinductive medium for 7 days. All of the chemicals induced the
expression of perilipin protein (Fig. 2.4). Rosiglitazone alone induced formation of
multiple bands. Perilipin is a phosphoprotein, thus the multiple bands observed are likely
phosphorylated and non-phosphorylated forms of the protein.(225)
These data illustrate that structurally diverse environmental PPARγ ligands
induce adipocyte differentiation and PPARγ signaling in primary bone marrow MSCs,
but with different potencies and efficacies.
43
Environmental PPARγ ligands show differential efficacies of bone suppression in
primary bone marrow MSCs
A number of studies have demonstrated that therapeutic PPARγ agonists suppress
osteoblast differentiation and bone formation both in vitro(15,96) and in vivo.(74,75)
Similarly, TBT inhibits differentiation of osteoblast-like (ROB) cells(226) and suppresses
osteogenesis in multipotent adipose-derived stromal cells by a PPARγ-mediated
mechanism.(60) Here, we tested the hypothesis that environmental PPARγ-agonists would
suppress osteogenic differentiation in BM-MSCs.
Again, we began by establishing the effect of rosiglitazone and TBT in our model
system. Primary bone marrow cultures were prepared from male C57BL/6J mice and
treated with Vh (DMSO, 0.1%), TBT (100 nM), or rosiglitazone (100 nM) in the
presence of osteoinductive medium. Cultures were assessed for changes in alkaline
phosphatase activity, mineralization, and nodule number (10–12 days) and osteogenic
gene expression (7 days). Both TBT and rosiglitazone significantly decreased alkaline
phosphatase activity and the average number of bone nodules per square cm (Fig. 2.5a).
TBT also significantly suppressed calcium deposition (Fig. 2.5a). Suppression of
osteogenesis also was reflected in decreased expression of osteogenic genes (Fig. 2.5b).
TBT and rosiglitazone significantly suppressed expression of Runx2, an essential
transcription factor regulating osteogenesis, as well as expression of osterix (Osx), a
direct gene target of Runx2 (Fig. 2.4b). Only TBT significantly suppressed expression of
dentin matrix phosphoprotein 1 (Dmp1), a gene expressed in mineralizing osteocytes.(102)
Next, we tested the ability of the environmental PPARγ ligands to suppress
44
osteogenesis. Established bone marrow cultures were treated with Vh (DMSO, 0.1%),
TPhT (10, 50, 80 nM), MEHP, TBBPA or METBP (10, 20 μM), in the presence of
osteoinductive medium. A concentration of TPhT as low as 10 nM significantly
decreased alkaline phosphatase activity, nodule number, and mineralization (Fig. 2.6a).
The other ligands were less efficacious at suppressing osteogenesis. MEHP, METBP and
TBBPA significantly suppressed alkaline phosphatase activity, but only MEHP
significantly suppressed nodule number (Fig. 2.6b–2.6d). The treatment regimen and
concentrations did not reduce cellularity as measured by MTT labeling (Supplemental
Fig. 2.1), indicating that the effect of the ligands was specific and not the result of overt
toxicity.
Decreases in the expression of osteoblast/osteocyte-related genes accompanied
suppression of osteogenesis (Fig. 2.7). Runx2 expression was relatively resistant to
toxicant exposure and was only reduced by treatment with MEHP and TBBPA (Fig. 2.7b
and 2.7d). Osx expression was the most sensitive, with all toxicants significantly
reducing expression and TPhT, MEHP, and TBBPA reducing Osx expression at the
lowest concentration (Fig. 2.7a, 2.7b and 2.7d). TPhT and MEHP also significantly
reduced Dmp1 expression (Fig. 2.7a and 2.7b). As with the phenotypic assays, METBP
had a limited capacity to suppress osteoblast-related gene expression (Fig. 2.7d).
The data support the conclusion that suppression of osteogenesis is a common
toxic effect of environmental toxicants that are capable of activating PPARγ.
45
Effects on adipogenesis and osteogenesis are negatively correlated
A balance of transcriptional activity determines the fate of MSCs, with PPARγ
promoting adipogenesis and negatively regulating Runx2, and Runx2 promoting
osteogenesis following Wnt/β-catenin suppression of PPARγ.(97) Thus, differentiation of
MSCs has traditionally been viewed as either exclusively toward adipogenesis or
exclusively toward osteogenesis. Despite the fact that the PPARγ ligands tested here
vary widely in their structure, as well as their potency and efficacy for inducing
adipogenesis, they demonstrate strong, significant relationships (p < 0.05, Pearson’s r,
excluding organotins) between efficacy of induction of adipogenesis and efficacy of
suppression of osteogenesis (Fig. 2.8a–2.8c). However, the organotins appear to be more
efficacious at suppressing bone formation than the other toxicants. We hypothesize that
this is because they are dual RXR and PPARγ ligands, thus they were removed from the
correlation analysis, and this greatly improved the correlation. In general, these data
support the conclusion that suppression of osteogenesis accompanies induction of
adipogenesis by multiple environmental PPARγ ligands. However, the relationship of the
efficacies of the effects may be ligand-specific.
The disproportionate effect of organotins is not mediated by Wnt10b
Recently, Rahman et al. (2012) found that the adipogenic effects of PPARγ
activation are “sequestered” via β-catenin suppression, while the anti-osteogenic capacity
of PPARγ is maintained through suppression of wingless-type MMTV integration site
family, member 10b (Wnt10b), a secreted protein that activates the canonical Wnt-
46
signaling pathway of osteogenesis. Stabilization of β-catenin suppressed PPARγmediated adipogenic effects, but PPARγ-dependent suppression of Wnt10b was
maintained and resulted in an anti-osteogenic phenotype.(209) Given that our experiments
suggest a disproportionate effect of osteogenic suppression by the organotins, we
hypothesized that cells treated with organotins would show a disproportionate
suppression of Wnt10b RNA expression compared to rosiglitazone and MEHP.
To test this hypothesis, we examined the expression of Wnt10b mRNA in
osteogenic cultures treated with Vh (DMSO, 0.1%), rosiglitazone (100 nM), TBT (100
nM), TPhT (10, 50, and 80 nM) or MEHP (10 and 20 μM) for 7 days. Rosiglitazone was
the only ligand that significantly suppressed the expression of Wnt10b mRNA (Fig. 2.9a),
supporting the observation reported by Rahman et al. (2012) and Lecka-Czernik et al.
(2002).(73,209) However, neither TBT nor TPhT significantly reduced Wnt10b expression
(Figs. 2.9a and 2.9b), and the Wnt10b expression level was not qualitatively different
from the Wnt10b expression in MEHP-treated cultures (Fig 2.9c). Thus, the data do not
support a role for disproportionate effects on Wnt10b in organotin-induced suppression of
osteogenesis.
Discussion
47 million individuals in the U.S. are estimated to be at risk for osteoporosis by
2020, representing a major public health concern among the aging U.S. population.(227)
Accordingly, it is important to understand the factors that contribute to low bone density,
a significant risk factor for osteoporosis. Because precursor bone marrow cells are
47
capable of differentiation into pre-osteocytes or pre-adipocytes, a shift towards adipocyte
differentiation instead of osteocyte differentiation could have deleterious effects on bone
health. In fact, increases in lipid accumulation are associated with greater fracture risk
and the onset of osteoporosis (for review, see Rosen and Bouxsein (2006)),(150) and
treatment with therapeutic PPARγ ligands is associated with greater fracture risk among
diabetics.(206) We hypothesized that environmental PPARγ ligands would promote
differentiation of BM-MSCs into adipocytes at the expense of differentiation into
osteocytes and bone formation.
A growing number of environmental toxicants have been identified as PPARγ
ligands whose diversity of sources and high-volume production result in significant
human exposures. Here, we examined the potency and efficacy of TBT, TPhT, MEHP,
METBP, and TBBPA in activation of mouse PPARγ1 and PPARγ2 and then examined
their pro-adipogenic and anti-osteogenic effects in mouse BM-MSCs. All of the
toxicants tested activated both mPPARγ1 and γ2, with the rank potency of
Rosiglitazone=TBT=TPhT>>TBBPA>MEHP>METBP. Their rank efficacy for PPARγ1
and γ2 activation was Rosiglitazone=TBT=TPhT>MEHP>TBBPA= METBP. MEHP,
METBP and TBBPA were apparent partial agonists of PPARγ with, in the cases of
MEHP and TBBPA, EC50’s similar to values previously reported.(64,67) Similar to our
studies, others have shown TBT and TPhT have similar potency inducing lipid
accumulation in a multipotent MSC model and for activation of human PPARγ compared
to rosiglitazone and troglitazone.(58,59,80)
The potency of PPARγ activation was largely reflected in each toxicant’s ability
48
to stimulate adipocyte differentiation, as evidenced by lipid accumulation and PPARγ
target gene expression. As expected, the potent therapeutic PPARγ ligand rosiglitazone
was most efficacious at stimulating adipocyte differentiation in BM-MSCs, as has been
observed in mouse and human models of preadipocytes and multipotent MSCs.(59,60,80,82)
The organotins, while apparent full agonists in terms of PPARγ transcriptional activation,
were potent but not fully efficacious at stimulating lipid accumulation and PPARγ target
gene expression in BM-MSCs. Our lipid accumulation and PPARγ activation results
with TBT and TPhT confirm previous results showing sub-maximal efficacy in BMS2
cells.(80)
On the other hand, MEHP was an apparent partial agonist in terms of PPARγ
transactivation, but efficaciously stimulated lipid accumulation and PPARγ target gene
expression. The partial agonist nature of MEHP has been previously reported.(51) In our
study, MEHP (20 μM, 7–10 days, insulin and dexamethasone) induced lipid
accumulation in mouse BM-MSCs to an extent comparable to the full agonist
rosiglitazone, whereas Feige et al. (2007) found that NIH 3T3-L1 cells treated with
MEHP (100 μM, 10 days, insulin) accumulated lipid (triglyceride content) to only 50%
of rosiglitazone treatment.(51) Both rosiglitazone and MEHP efficaciously induced genes
important to adipogenesis in the BM-MSC and NIH 3T3 L1 models (e.g. Fabp4, Plin1
(BM-MSCs) and adiponectin (NIH 3T3-L1s)). However, other PPARγ gene targets (e.g.
glycerol kinase, oxidized low density lipoprotein receptor 1, acyl-CoA synthetase
Bubblegum 1) have been shown to only be efficaciously induced by rosiglitazone, mostly
likely because MEHP selectively recruits coactivators to the PPARγ transcriptional
49
complex.(51)
METBP, the metabolite of an emerging environmental contaminant, di-(2ethylhexyl) tetrabromophthalate, significantly stimulated lipid accumulation but was less
efficacious at stimulating PPARγ-target gene expression, in line with its marginal ability
to stimulate PPARγ transcriptional activity. These results are in accordance with a recent
study showing that METBP induced adipocyte differentiation and PPARγ activation in
NIH 3T3 L1 cells.(66) That METBP should display an adipogenic profile similar to
MEHP is not entirely surprising, given previous studies showing multiple phthalate
monoester metabolites activating PPARγ with comparable potencies and efficacies.(64)
Like MEHP, TBBPA appears to be a partial ligand of PPARγ with significant but
not maximal efficacy in inducing adipocyte differentiation. The receptor activation data
show that TBBPA is a low-potency PPARγ ligand with moderate efficacy. This is
similar to studies conducted in human PPARγ-based reporter cells, where tri- and
tetrabrominated BPA had the greatest potency and efficacy in activating PPARγ, with
decreasing substitution significantly decreasing activity.(67) Our results confirmed the
ability of TBBPA to induce adipogenesis, accompanied by significant increases in
Pparg1/2, Fabp4 and Plin1 expression.
This is the first report that TPhT, MEHP and TBBPA suppress osteogenesis. We
expected to find that those environmental toxicants that induced adipogenesis would also
suppress osteogenesis, owing to the inhibitory interactions between PPARγ and Runx2.
Mechanistically, PPARγ directly interacts with Runx2 to prevent Runx2 transcriptional
activity.(96) Additionally, PPARγ activation increases β-catenin degradation by the
50
proteasome, removing β-catenin’s support of osteogenesis and β-catenin’s Wnt-mediated
inhibition of PPARγ expression.(97,100) However, it is becoming clear that activation of
adipogenesis and suppression of osteogenesis are distinct and separable activities of
PPARγ,(210) with ligands having selective abilities to impact adipogenesis and
osteogenesis.(73,211,212) Therefore, it is important to examine environmental PPARγ
ligands for their effects on both adipogenesis and osteogenesis.
Interestingly, the effect of TBT and TPhT on osteogenesis was markedly more
severe compared to the other ligands. The results with TBT are in line with studies
demonstrating that TBT suppressed alkaline phosphatase activity in rat calvarial
osteoblasts and reduced mineralization in adipose-derived stromal cells.(60,226) The
organotins were disproportionately efficacious in suppressing bone than inducing
adipogenesis. One potential explanation for this dichotomy is that the organotins more
readily activate the anti-osteogenic functions of PPARγ than its pro-adipogenic functions.
These PPARγ functions were recently shown to be separable, and the anti-osteogenic
effects of PPARγ were shown to be dependent upon Wnt10b.(209) However, while
treatment with TBT, TPhT and MEHP appeared to suppress expression of Wnt10b, the
effect was not statistically significant. These data suggest that unique modulation of a
Wnt10b-mediated pathway by organotins does not appear to explain their selective effect
on MSC differentiation.
The toxicants studied here join a growing list of environmental contaminants that
compromise bone quality, an effect not exclusive to PPARγ ligands. Processes mediated
by aryl hydrocarbon receptor ligands (e.g. 2,3,7,8-tetrachlorodibenzo-p-dioxin, benzo-a-
51
pyrene) contribute to increased bone resorption in vivo(202) and impair MSC
differentiation in vitro.(204) Additionally, co-exposure to dioxin and TBT appears to
exacerbate the effects of exposure to each toxicant individually.(228) Heavy metals such as
lead and arsenic are toxic to the skeleton of rats, causing decreased bone density,
mineralization and strength.(174,184) It is largely unknown how complex exposures to
multiple bone-suppressive toxicants may act cooperatively to impair bone quality.
Our results show that environmental PPARγ ligands are efficacious promoters of
lipid accumulation and suppress osteogenesis in a mouse bone marrow-derived MSC
model. While the efficacies varied by chemical and endpoint, it is important to note that
effects were observed at environmentally-relevant concentrations. Additional concern
should be directed towards identifying possible interactions and how multiple,
simultaneous exposures could additively suppress bone formation in MSCs. While this
study tested exposures on an individual toxicant basis, environmental exposures to these
chemicals are likely to occur simultaneously. Further experiments should aim to assess
chemical mixtures of ubiquitous and persistent compounds in order to more accurately
characterize toxicant-induced bone loss.
Tables
Table 2.1. EC50, maximal activation values, and Hill coefficients for mPPARγ activation by selected toxicants. Values
are fit estimates from 4-parameter Hill plot using Prism (GraphPad software, Inc. La Jolla, CA). Maximal fold change in
PPARγ activation (Max) represents the difference between estimated curve maximum and minimum for each individual
chemical. n = 4–6 separate transfection experiments.
mPPARγ1
Ligand
Rosiglitazone
TBT
TPhT
MEHP
METBP
TBBPA
EC50 (M)
3.4 x 10-8
1.0 x 10-8
4.0 x 10-8
2.4 x 10-5
1.8 x 10-4
3.9 x 10-6
Max
3.9
3.6
4.8
2.4
1.3
1.5
mPPARγ2
Hill coeff.
0.9
1.9
1.2
1.3
2.1
1.2
EC50 (M)
3.0 x 10-8
1.7 x 10-8
5.0 x 10-8
2.9 x 10-5
2.0 x 10-4
7.8 x 10-6
Max
4.1
5.0
5.5
2.7
1.4
2.0
Hill coeff.
1.0
1.0
1.1
1.2
3.3
0.8
52
53
Figures
FIGURE 2.1. Environmental toxicants activate PPARγ1 and PPARγ2 with differing
potencies and efficacies.
54
Figure 2.1. Environmental toxicants activate PPARγ1 and PPARγ2 with differing
potencies and efficacies. Cos-7 cells transfected with mouse PPARγ1 or PPARγ2 and a
PPRE-luciferase reporter plasmid were treated with the indicated compounds.
Luminescence normalized to GFP fluorescence was divided by the normalized
luminescence of untreated cultures to calculate fold change from untreated. n = 4–6
independent transfections.
55
FIGURE 2.2. Known modulators of adipogenesis increase lipid accumulation and
expression of adipogenic genes in mouse BM-MSCs.
56
Figure 2.2. Known modulators of adipogenesis increase lipid accumulation and
expression of adipogenic genes in mouse BM-MSCs. Primary bone marrow cultures
were established from male C57BL/6J mice and treated with vehicle (Vh, DMSO),
rosiglitazone (Rosi, 100 nM) or TBT (100 nM) in the presence of osteoinductive media
for 7 (gene expression) or 11 days (lipid accumulation). (A) Lipid accumulation was
quantified by Nile Red staining. (B-D) mRNA expression was quantified by RT-qPCR.
Data are presented as means ± SE (n = 4–8 independent bone marrow preparations). *p <
0.05, **p < 0.01 compared to Vh-treated cultures (ANOVA, Dunnett’s). Vehicle-treated
cells showed increases in expression of adipocyte-related genes relative to
undifferentiated cells (1.9-fold for PPARγ, 7.7-fold for Fabp4, and 144.4-fold for Plin1).
57
FIGURE 2.3. Structurally distinct environmental PPARγ ligands induce lipid
accumulation and adipogenic gene expression in mouse BM-MSCs.
58
Figure 2.3. Structurally distinct environmental PPARγ ligands induce lipid
accumulation and adipogenic gene expression in mouse BM-MSCs. Primary bone
marrow cultures were established from male C57BL/6J mice and treated with Vh
(DMSO), TPhT (10–80 nM; A), MEHP (10–20 μM; B), METBP (10–20 μM; C) or
TBBPA (10–20 μM; D) in the presence of osteoinductive media for 7 (gene expression)
or 11 days (lipid accumulation). Lipid accumulation was quantified by Nile Red staining.
mRNA expression was quantified by RT-qPCR. Data are presented as means ± SE (n =
4–8 independent bone marrow preparations). *p < 0.05, **p < 0.01 compared to Vhtreated cultures (ANOVA, Dunnett’s).
59
FIGURE 2.4. Environmental PPARγ ligands induce perilipin protein expression in
mouse BM–MSCs.
60
Figure 2.4. Environmental PPARγ ligands induce perilipin protein expression in
mouse BM–MSCs. Primary bone marrow cultures were established from male
C57BL/6J mice and treated with Vh (DMSO), rosiglitazone (Rosi, 100 nM), TBT (100
nM), TPhT (50 nM), MEHP, TBBPA, or METBP (20 μM) in the presence of
osteoinductive media. Cells were harvested after 7 days. Perilipin and β-actin expression
were determined in whole cell lysates by immunoblot. Image is representative of 4
separate experiments.
61
FIGURE 2.5. Rosiglitazone and TBT suppress osteogenesis in mouse BM-MSCs.
62
Figure 2.5. Rosiglitazone and TBT suppress osteogenesis in mouse BM-MSCs.
Primary bone marrow cultures were established from male C57BL/6J mice and treated
with vehicle (Vh, DMSO), rosiglitazone (Rosi, 100 nM) or TBT (100 nM) in the
presence of osteoinductive media for 7 (gene expression) or 11 days (osteogenesis
assays). (A) Osteogenesis was assessed via alkaline phosphatase activity, alizarin
staining and bone nodule counting. (B) mRNA expression was quantified by RT-qPCR.
Data are presented as means ± SE (n = 5–8 independent bone marrow preparations). *p <
0.05, **p < 0.01 compared to Vh-treated cultures (ANOVA, Dunnett’s).
63
FIGURE 2.6. Structurally distinct environmental PPARγ ligands suppress
osteogenesis in mouse BM-MSCs.
64
Figure 2.6. Structurally distinct environmental PPARγ ligands suppress
osteogenesis in mouse BM-MSCs. Primary bone marrow cultures were established
from male C57BL/6J mice and treated with Vh (DMSO), TPhT (10–80 nM; A), MEHP
(10–20 μM; B), METBP (10–20 μM; C), or TBBPA (10–20 μM; D), in the presence of
osteoinductive media for 11 days. Osteogenesis was assessed by alkaline phosphatase
activity, alizarin staining and bone nodule counting. Data are presented as means ± SE (n
= 5–8 independent bone marrow preparations). *p < 0.05, **p < 0.01 compared to Vhtreated cultures (ANOVA, Dunnett’s).
65
FIGURE 2.7. Structurally distinct environmental PPARγ ligands suppress
osteogenic gene expression in mouse BM-MSCs.
66
Figure 2.7. Structurally distinct environmental PPARγ ligands suppress osteogenic
gene expression in mouse BM-MSCs. Primary bone marrow cultures were established
from male C57BL/6J mice and treated with Vh (DMSO), TPhT (10–80 nM; A), MEHP
(10–20 μM; B), METBP (10–20 μM; C), TBBPA (10–20 μM; D) in the presence of
osteoinductive media for 7 days. mRNA expression was quantified by RT-qPCR. Data
are presented as means ± SD (n = 4–7 independent bone marrow preparations). *p <
0.05, **p < 0.01 compared to Vh-treated cultures (ANOVA, Dunnett’s).
67
FIGURE 2.8. Adipogenesis is inversely correlated with osteogenesis in PPARγligand treated mouse BM-MSCs.
68
Figure 2.8. Adipogenesis is inversely correlated with osteogenesis in PPARγ-ligand
treated mouse BM-MSCs. Data points are representative of mean endpoint values for
individual treatments and concentrations reported in Figs 2.2, 2.3, 2.5, 2.6, 2.7. (A) Fold
change in alkaline phosphatase activity vs. lipid accumulation (Nile Red fluorescence)
(All data Pearson’s r = -0.51). (B) Bone nodule number vs. lipid accumulation (All data
Pearson’s r = -0.01). (C) Osx mRNA expression vs. Fabp4 mRNA expression (All data
Pearson’s r = -0.57). ■ – Vh (DMSO); □ – Rosiglitazone (100 nM); ● – METBP; ○ –
MEHP; Δ – TBBPA; x – TBT or TPhT. Linear fit excludes organotin data points (x). r =
Pearson’s correlation coefficient (excluding organotins), number of pairs = 8.
69
FIGURE 2.9. Differential suppression of osteogenic Wnt10b mRNA does not explain
efficacious effect of organotins on osteogenesis.
70
Figure 2.9. Differential suppression of osteogenic Wnt10b mRNA does not explain
efficacious effect of organotins on osteogenesis. Primary bone marrow cultures were
established from male C57BL/6J mice and treated with Vh (DMSO), rosiglitazone (100
nM), TBT (100 nM), TPhT (10–80 nM), MEHP (10–20 μM) in the presence of
osteoinductive media for 7 days. mRNA expression was quantified by RT-qPCR. Data
are presented as means ± SE (n = 4–10 independent bone marrow preparations). *p <
0.05, **p < 0.01 compared to Vh-treated cultures (ANOVA, Dunnett’s).
71
Supplemental material
Supplementary Table S2.1. Primers used for mRNA expression analysis.
Gene
Symbol
NCBI Reference
Sequence
Qiagen
Catalog #
Dmp1
NM_016779
QT01078210
Fabp4
NM_024406
QT00091532
Plin1
NM_001113471 QT00150360
NM_175640
Pparg 1/2
NM_001127330, QT00100296
NM_011146
Rn18s
X00686,
NR_003278
QT01036875
Runx2
NM_009820
QT00102193
Osx
NM_130458
QT00293181
Wnt10b
NM_011718
QT00113211
TPhT (nM )
M EHP (M )
M ETBP ( M )
20
10
20
2.4
2.0
1.6
1.2
0.8
0.4
0.0
10
20
2.4
2.0
1.6
1.2
0.8
0.4
0.0
10
80
50
2.4
2.0
1.6
1.2
0.8
0.4
0.0
10
TB
T
2.4
2.0
1.6
1.2
0.8
0.4
0.0
Vh
Fold Change
From M edium
2.4
2.0
1.6
1.2
0.8
0.4
0.0
TBBPA ( M )
Supplementary Figure S2.1 – Toxicity of TBT, TPhT, METBP, TBBPA and MEHP. Primary bone marrow cultures were
established from male C57BL/6J mice and treated with vehicle (Vh, DMSO), TBT (100 nM), TPhT (10–80 nM), MEHP (10–
20 μM), METBP (10–20 μM), or TBBPA (10–20 μM) in the presence of MSC medium 10 days. Cellularity was assessed by
analyzing MTT labeling. Data are presented as means ± SE (n=4–8). No statistically significant differences were observed in
experimental samples (ANOVA).
72
73
CHAPTER III: Tributyltin induces distinct effects on cortical and trabecular bone
in female C57Bl/6J mice resulting from modulation of both osteoblasts and
osteoclasts
James Watt,1 Amelia H. Baker,2 Brett Meeks,3 Paola Divieti Pajevic,4 Elise F. Morgan,3,5
Louis C. Gerstenfeld,3 and Jennifer J. Schlezinger1
1
Department of Environmental Health, Boston University School of Public Health,
Boston, MA, 02118, USA
2
Department of Medicine, Boston University School of Medicine, Boston, MA, USA
3
Department of Orthopaedic Surgery, Boston University School of Medicine, Boston,
MA, USA
4
Department of Molecular and Cell Biology, Boston University School of Dental
Medicine, Boston, MA USA
5
Department of Biomedical Engineering, Boston University, Boston, MA
Acknowledgements
This work was supported by the National Institute of Environmental Health Sciences
Superfund Research Program [P42ES007381].
The authors thank Cassie Huang and Zack Webster for their technical expertise.
Abbreviations: PPARγ: peroxisome proliferator activated receptor gamma; RXRα:
retinoid-X receptor alpha; LXR: liver-X receptor; RANKL: receptor activator of nuclear
factor kappa B ligand; OPG: osteoprotegerin; TRAP: tartrate-resistance acid phosphatase
74
Abstract
The balance of bone formation and resorption is controlled by a precise coupling between
osteoblasts and osteoclasts, respectively. Nuclear receptors transcriptionally modify
differentiation and function of both cell types, thus activation of nuclear receptors by
exogenous compounds can perturb overall bone homeostasis. The retinoid-X receptor
alpha (RXRα), peroxisome proliferator activated receptor gamma (PPARγ) and the liverx receptors (LXRs) all have been shown to influence bone homeostasis. Tributyltin
(TBT) is an environmental contaminant that is a dual RXRα and PPARγ agonist that
induces RXR, PPARγ and LXR-mediated gene transcription and suppresses osteoblast
differentiation in vitro. In these studies, 12-week old female wild-type C57Bl/6J mice
were exposed to 10 mg/kg TBT for 10 weeks via oral gavage. As expected, the femurs of
TBT-treated mice had a smaller cross-sectional area and thinner cortex. Surprisingly,
TBT induced significant increases in trabecular thickness, number, and fractional bone
volume, along with a decrease in trabecular spacing. Histological analyses showed that
the observed increase in bone was not due to a lack of osteoclasts in the trabecular space.
Analyses of whole bone RNA indicated that TBT treatment did not change the
Rankl:Opg ratio, suggesting that osteoblast stimulation of osteoclasts was not suppressed.
However, expression of cardiotrophin-1 (Ct1), an osteoblastogenic Il-6 family cytokine
secreted by osteoclasts, was increased. In TBT-exposed primary bone marrow cultures
stimulated to differentiate into osteoclasts, TBT did not inhibit the number of osteoclasts
that differentiated, despite the fact that expression of osteoclast markers Nfatc1, Acp5 and
Ctsk decreased. However, Ct1 expression remained elevated in TBT-treated cultures.
75
Additionally, TBT induced expression of LXR- and RXR-dependent genes in whole bone
and in vitro osteoclast cultures. These results suggest that TBT has a distinct effect on
cortical and trabecular bone, likely resulting from independent effects on osteoblast and
osteoclast differentiation that are mediated through multiple nuclear receptors.
76
Introduction
Bone remodeling is a coordinated sequence of bone resorption followed by bone
formation, orchestrated by the activity of osteoclasts and osteoblasts, respectively.
Maintenance of bone and mineral homeostasis requires tight regulation of this coupled
process. Where an imbalance favors the activity of osteoclasts, resorption prevails over
mineral deposition and overall bone quality decreases, potentially leading to pathologic
states of osteopenia and osteoporosis. Conversely, inherited mutations causing
impairments of osteoclast differentiation or function can result in osteopetrotic states of
overly dense and brittle bone.
Communication between the two cell types allows for local control of the
resorption-formation coupling. For example, osteoblasts produce macrophage colony
stimulating factor (M-CSF) and receptor activator of nuclear factor kappa B ligand
(RANKL).(120,125) Upon interacting with hematopoietic precursor cells, these two factors
are sufficient to promote differentiation of mature osteoclasts via downstream activation
of the transcription factor Nfatc1.(110,123,229) The RANKL signal is suppressed by the
soluble osteoprotegerin (OPG), also expressed by osteoblasts.(230) Osteoclasts reciprocally
produce and secrete cytokines (e.g. cardiotrophin-1)(135) that support the maturation of
osteoblasts to regenerate the collagenous mineral tissue removed during resorption, as
well as provide inhibitory signals (e.g. Semaphorin 4D)(137) to suppress osteoblast activity
during the resorption phase. Impairment of the function of one cell type may therefore
have indirect effects on the differentiation or function of the other.
Multiple aspects of osteoblast and osteoclast differentiation and function are
77
influenced by nuclear receptors, which respond to endogenous and exogenous ligands.
The peroxisome proliferator activated receptor γ (PPARγ) and liver-x receptors α and β
(LXRα and LXRβ) are members of the nuclear receptor superfamily that form permissive
heterodimers with the retinoid-x receptor alpha (RXRα), which can also
homodimerize.(8,26,231) All four receptor types are expressed in osteoblasts and
osteoclasts.(72)
Transactivation of PPARγ in bone marrow mesenchymal osteoblast precursors by
synthetic agonists, such as rosiglitazone, induces differentiation along the adipocyte
lineage and suppresses osteoblast differentiation.(15,16) In osteoclasts, PPARγ is essential
for differentiation via support of RANKL signaling.(232,233) In vivo exposure to
rosiglitazone results in decreased bone quality and increased fat content in the marrow
space.(74–76)
LXR agonism in vivo has been shown to attenuate osteoclast-mediated bone loss
in ovariectomized and inflammatory-induced bone loss mouse models.(234,235) However,
long-term treatment in vivo with a synthetic LXR-specific agonist appears to have no net
positive effect on bone mineral content in an intact mouse model.(236)
RXR activation plays maturation-dependent roles in osteoclasts. In osteoclast
precursors, RXR homodimer activity maintains a high level of Mafb expression, which is
necessary for a proper proliferative response to M-CSF.(13) In differentiating osteoclasts,
RXRα:LXR activation suppresses osteoclast function and differentiation by interfering
with RANKL signaling.(13,237)
The environmental contaminant tributyltin (TBT) is a dual PPARγ and RXRα
78
agonist shown to activate RXR homodimers, as well as PPARγ:RXRα(58,63,81) and
LXR:RXRα(81,238) heterodimers and is a recognized bone toxicant. In utero exposure to
TBT prevents ossification in mouse fetuses and predisposes stromal cells to favor
adipogenesis by epigenetic modifications at PPARγ target promoters.(60,226) In vitro, TBT
suppresses the osteoblastogenic transcription factors Runx2 and osterix while activating
PPARγ:RXRα. This promotes commitment of bone marrow mesenchymal stromal cells
(BM-MSCs) towards an adipocyte lineage at the expense of bone mineralization in a
manner similar to rosiglitazone.(60,80–82,239) In contrast to rosiglitazone, TBT was shown to
suppress osteoclast differentiation and resorptive capacity at nanomolar concentrations in
Raw264.7 cells;(240) however, these cells do not express significant levels of PPARγ,
RXRs or LXRs.(13) It has not been established as to which of these effects predominates
to determine a bone phenotype in adult mice exposed to TBT.
The studies herein were designed to determine the overall effect of in vivo TBT
exposure on long bones of skeletally mature, adult, female C57Bl/6 mice, and to
characterize the role of TBT, as an activator of multiple nuclear receptor pathways, in the
balance of osteoblast and osteoclast function.
Materials and Methods
Rosiglitazone was from Cayman Chemical (Ann Arbor, MI). DMSO was from
American Bioanalytical (Natick, MA). LG100268, T0901317, p-nitrophenyl phosphate
(pNPP) reagent, TBT chloride, Gil’s Hematoxilin and sodium tartrate were from SigmaAldrich (St. Louis, MO). All other reagents were from Thermo Fisher Scientific
79
(Suwanee, GA) unless noted.
In vivo studies
All animal studies were approved by the Institutional Animal Care and Use
Committee at Boston University. Female C57BL/6J mice (12 weeks of age) (Jackson
Laboratories, Bar Harbor, ME) were gavaged 3x per week for 10 weeks with vehicle
(sesame oil, 10 μl/g), or TBT (10 mg/kg dissolved in sesame oil). Mice were euthanized
four days after the last dosing. At euthanasia, the right femur was collected for microCT
and histological analyses; humeri were collected for RNA analyses. Serum was collected
at euthanasia for analysis of bone turnover markers.
Micro-Computed Tomography (micro-CT)
Sample analyses were performed according to guidelines outlined in Bouxsein et
al. (2010).(241) Samples were scanned at a resolution of 6 m/voxel using a Scanco microCT 40 system (Scanco Medical; Basserdorf, Switzerland). Trabecular bone was analyzed
in the distal femur, with the region of interest beginning 0.03 mm proximal to the growth
plate and extending 0.9 mm proximally. The trabecular compartment was manually
segmented from the cortical shell. Cortical bone was analyzed from 12 m/voxel scans of
a 0.6 mm region extending proximally from the mid-diaphysis. Treatment-specific
thresholding was used with thresholds determined by an iterative method (Scanco
Medical). Tissue mineral density was calculated with the aid of a standard curve obtained
from a scan of a hydroxyapatite phantom consisting of five different mineral densities.
80
Histology and Immunohistochemistry
Following micro-CT scans, femurs were decalcified in 14% w/v EDTA at 4°C
and embedded in paraffin. 5m slices were stained with hematoxylin and eosin or with
tartrate resistant acid phosphatase, as previously described.(242) Micrographs were
visualized on an Olympus BX51 light microscope (Olympus America Inc.; Center
Valley, PA). Osteoclast surface, osteoclast number, and adipocyte number were
determined manually within a 1.5 mm selection proximal to the growth plate.
Primary osteoblast cell culture
Bone marrow was isolated from 9-week-old male and female, C57BL/6J mice
(Jackson Laboratories). Bone marrow was flushed from the femur, tibia, and humerus,
strained through a 70 μm cell strainer, diluted in MSC media (α-MEM +10% FBS, 100
U/mL penicillin, 100 μg/mL streptomycin, 0.25 μg/mL amphotericin B). Cells were
seeded at 12 × 106 per well in a 6-well plate in 2 ml media. At 7 days, the medium was
changed to include ascorbate (12.5 μg/mL), β-glycerol phosphate (8 μM), dexamethasone
(10 nM; Sigma), and insulin (500ng/mL; Sigma), and the cultures were dosed with Vh
(DMSO, 0.1% final concentration) or TBT (10, 50nM). Naïve cells were maintained in
MSC medium, received no treatments, and were harvested at day 7 in culture. Following
initiation of differentiation, cells were cultured for 7 days (mRNA expression) or 12 days
(bone phenotype). During these periods, medium was changed and the cultures were redosed 2 times for mRNA expression or 4 times for phenotype analysis.
81
Primary osteoclast culture
Bone marrow was isolated from 12-week-old female, C57BL/6J mice (Jackson
Laboratories). Following overnight incubation in MSC media with 25 ng/ml recombinant
human M-CSF (BioLegend, San Diego, CA), suspended cells from whole bone marrow
were collected and seeded at 3 x 106 per well in a 6-well plate with 2 ml media (gene
expression) or 2.5 x 104 per well in a 24-well plate with 1 ml media (TRAP+ MuNC
counts; resorption assay) with 50 ng/ml M-CSF. On day 1, recombinant murine RANKL
(BioLegend) was added at 50 ng/ml. On days 2 and 4, media and additives were replaced
and treatments (Vh, TBT (20, 50, 80 nM), rosiglitazone (100 nM), LG100268 (100 nM),
or T0901317 (100 nM)) were added. Cells were harvested on day 6.
Tartrate-resistant acid phosphatase (TRAP) staining
Osteoclast cultures were rinsed with PBS and fixed in 4% paraformaldehyde
(Sigma). Membranes were permeabilized with ethanol/acetone (1:1) and stained with
sodium tartrate (50 mM). Cultures were counterstained with Gil’s Hematoxilin solution.
Multi-nucleated (3 or more nuclei), TRAP-positive cells were counted manually.
Mineral resorption assay
Osteoclast precursors were seeded at 2.5 x 104 per well in a 24-well Corning
OsteoAssay plate (Corning, Corning NY) and treated, as above. On day 6, cells were
removed with 10% bleach. The mineral surface was rinsed with distilled water and dried
for 3 hours at room temperature. Surfaces were visualized at 10x magnification using an
82
EVOS XL Imaging System (Thermo Fisher/Life Technologies; Carlsbad, CA). 4–5 fields
of view per well were manually thresholded and resorption pit number, size, and area
were determined using Image J (NIH; http://imagej.net). Objects smaller than 1,100 µm2
in area were excluded.
Reverse transcriptase-qPCR
Cell culture: Total RNA was extracted and genomic DNA was removed using the
RNeasy Plus Mini Kit (Qiagen, Valencia, CA). cDNA was prepared from total RNA
using the GoScript™ Reverse Transcription System (Promega), with a 1:1 mixture of
random and Oligo (dT)15 primers. All qPCR reactions were performed using the GoTaq®
qPCR Master Mix System (Promega). Validated primers were purchased from Qiagen or
synthesized by Integrated DNA Technologies (Coralville, IA) (see Table S1). qPCR
reactions were performed using a 7300 Fast Real-Time PCR System (Applied
Biosystems, Carlsbad, CA): Hot-Start activation at 95°C for 10 min, 40 cycles of
denaturation (95°C for 15 sec) and annealing/extension (55°C or 60°C for 60 sec).
Relative gene expression was determined using the Pfaffl method to account for
differential primer efficiencies.(223) The average Cq value for 18s ribosomal RNA (Rn18s)
and beta-2-microglobulin (B2m) was used for normalization. The Cq value for naïve,
undifferentiated cultures was used as the reference point, and the data are reported as
“Fold Difference.”
Whole bone: Total RNA was extracted and genomic DNA removed by double
TRIzol® extractions (Life Technologies, Grand Island, NY). RNA integrity was
83
confirmed by gel electrophoresis. cDNA preparation and RT-qPCR were performed as
above. The Cq value for naïve, 22 week old, male, C57BL/6J-derived humerus samples
(n = 5) was used as a reference point, and expression levels were normalized to the
sample-specific Cq average of Rn18s, B2m and hypoxanthine-guanine
phosphoribosyltransferase (Hprt), and the data are reported as “Fold Difference.”
Statistical analyses
Statistical analyses were performed with Prism 6 (GraphPad Software Inc., La
Jolla, CA). Data are presented as means  standard error (SE). In vitro gene expression
data were log transformed prior to analysis. One-way or two-way ANOVAs with the
Dunnett’s or Sidak’s post hoc test and unpaired t-tests were performed where noted.
Because of skewed distribution, non-parametric Mann-Whitney test was used for whole
bone RNA analyses. All analyses were performed at α = 0.05.
Results
Effect of TBT on osteogenic differentiation in vitro
We and others have previously documented TBT’s capability to suppress the in
vitro differentiation of mesenchymal stromal cells.(60,80,226,239) Initial studies herein were
designed to compare TBT’s in vitro efficacy and potency between male and female
C57Bl/6J mice. Primary bone marrow mesenchymal stromal cells (BM-MSC) were
harvested from 9-week old female and male C57Bl/6J mice and cultured in the presence
of l-ascorbate, β-glycerol phosphate, insulin and dexamethasone, and dosed with either
84
Vh (DMSO) or TBT (10 nM, 50 nM). Total RNA was harvested after 7 days to assess
relative gene expression, and analyses of bone nodule formation were conducted after 10
days.
As expected, TBT suppressed the expression of pro-osteogenic genes at
nanomolar concentrations in both male- and female-derived cultures (Figure 3.1a). Runx2
is the master regulator of osteogenesis and controls the expression of osterix (Osx), which
also is an essential osteogenic transcription factor.(99) Dentin matrix phosphoprotein 1
(Dmp1) is expressed by mineralizing osteocytes and regulates biomineralization and
mineral metabolism.(105) Levels of Runx2 expression in Vh-treated cultures were
comparable between sexes, although TBT only significantly decreased its expression in
female-derived cultures (Figure 3.1a, left panel). Osx was expressed at a higher level in
female-derived cultures. TBT reduced Osx expression in both male- and female-derived
cultures, however TBT more potently and efficaciously suppressed expression in femalederived cultures (Figure 3.1a, center panel). Levels of Dmp1 were lower overall in
female-derived cultures, and TBT significantly suppressed Dmp1 expression in both
male- and female-derived cultures. It should be noted that the relative expression level of
Dmp1 in cultures treated with 50 nM TBT was 30x lower in female-derived cultures
compared to male-derived cultures (Fold Difference: 16.4 ± 6.2 vs. 558 ± 195,
respectively) (Figure 3.1a, right panel).
The dimorphic response was more pronounced in the bone mineralization assays.
TBT significantly inhibited alkaline phosphatase activity at both 10 and 50 nM in femalederived cultures, but only at 50 nM in male cultures (Figure 3.1b). This pattern also was
85
evident in the number of bone nodules and mineral deposition (Figure 3.1b), with each
being more severely inhibited in female-derived cultures than in male-derived cultures.
From these data, we conclude that, in vitro, TBT disproportionately suppresses the
differentiation and function of BM-MSCs in female-derived cultures.
Effect of TBT on bone homeostasis in vivo
Because we observed a pronounced sensitivity to TBT in female-derived primary
bone marrow cultures, we hypothesized that in vivo exposure to TBT would produce an
overall deficit of bone tissue in the long bones of female C57Bl/6J mice. To test this
hypothesis, 12-week old female, non-ovariectomized mice were treated 3x/week for 10
weeks with either vehicle (sesame oil) or TBT (10 mg/kg bw) by oral gavage. TBT
treatment induced slightly larger increases in body weight relative to initial weight, but
thymus weight relative to body weight (a measure of systemic TBT toxicity) at
euthanasia was not affected (Figure S3.1).
At the end of the treatment period, micro-CT was used to determine TBT’s impact
on long bone microarchitecture. A femur was harvested from each mouse (n = 12 control,
11 TBT), and the cortical and trabecular compartments were analyzed at the diaphysis
and metaphysis, respectively (Figure 3.2). Overall, a slight decrease in cortical crosssectional area was evident from visual inspection of the 3D reconstructions (Figure 3.2a).
As expected from the in vitro results, cortical thickness (Ct.Th) and total bone area
(Ct.Ar) at the diaphysis were significantly decreased by TBT treatment (Figure 3.2b, top
left and top center). In TBT-treated mice, medullary area (Ma.Ar) decreased in parallel
86
with total area (Tt.Ar), resulting in a ratio of bone area to total area that was
indistinguishable from Vh-treated mice (Ct.Ar/Tt.Ar) (Figure 3.2b). At the middiaphysis, total mineral density, measured as mg of hydroxyapatite per mm2, was not
affected by treatment (Figure 3.2b, bottom right).
In the trabecular compartment, TBT treatment unexpectedly was associated with a
robust increase in trabecular mineralization that was clearly discernable in the
reconstructions (Figure 3.2c). The fraction of total area occupied by bone in the
medullary space (BV/TV) was significantly increased in the TBT-treated mice (Figure
3.2d, top left), and the tissue mineral density was increased (Figure 3.2d, top center). The
connective density (Conn.D) and number of trabeculae (Tb.N) increased with TBT
treatment (Figure 3.2d, top right and bottom left, respectively) as trabecular spacing
(Tb.Sp) decreased (Figure 3.2d, bottom center). The average trabecular thickness was not
affected (Tb.Th) (Figure 3.2d, bottom right).
Histological analyses were carried out to corroborate the structural data (Figure
3.3). An increased density of trabeculae was evident in femur sections of TBT-treated
mice (Figure 3.3a). Adipocyte number was highly variable in the TBT-treated bone
sections, and tended to be higher than in Vh (Figure 3.3b). Expression of the adipocyte
marker perilipin 1 (Plin1) RNA in whole bone was higher with TBT, although with
borderline significant (p = 0.063) (Figure 3.3c). With regard to osteoblast differentiation
and function, 10 mg/kg TBT treatment was associated with a slight but non-significant
increase in Runx2 and no change in Osx (Figure 3.3d). In accordance with the structural
data, TBT significantly increased the expression of Dmp1 (Figure 3.3d). There was no
87
change in the expression of Sost, an osteocyte-derived bone formation inhibitor (Figure
3.3d).
Increases in bone density can result from an increase in osteoblast activity, a
decrease in osteoclast number, and/or a decrease in osteoclast resorptive function.(243)
Therefore, we investigated the potential contribution of osteoclasts to the observed
phenotype. Histological sections of the distal femur were tartrate-resistant acid
phosphatase (TRAP) stained to visualize and quantify the osteoclast population (Figure
3.4a). TBT treatment did not cause a significant decrease in osteoclast coverage of
trabecular surface (Figure 3.4b). Additionally, there was no change in serum TRAP levels
between Vh- and TBT-treated mice (Figure 3.4c), suggesting that osteoclast number had
not substantially decreased with TBT treatment.
Transcript levels of Nfatc1, Acp5, and Ctsk were measured to assess TBT’s effect
on osteoclast differentiation and resorptive capacity. Nfatc1 is a transcription factor
required for osteoclast differentiation and function.(124) Trap5b, the protein product of
Acp5, is a lysosomal acid phosphatase involved in processing of non-collagenous proteins
(e.g. osteopontin) in osteoclasts during bone resorption.(244) Cathepsin K is a lysosomal
cysteine proteinase involved in bone resorption.(245) TBT treatment was associated with
significant decreases in Nfatc1 and Acp5, along with a slight but non-significant decrease
in Ctsk expression (Figure 3.4d). Osteoblast stimulation of osteoclast differentiation was
assessed by measuring relative levels of Rankl and Opg expression. Overall, TBT
increased mRNA expression levels of both proteins (data not shown), but the ratio of
Rankl:Opg was not altered from Vh-treated samples (Figure 3.4e). However, TBT was
88
associated with a significant increase in the expression of the pro-osteogenic cytokine
cardiotrophin 1 (Ct1)(135) (Figure 3.4e), suggesting a mechanism of increased osteoblast
stimulation by osteoclasts.
Together, these differences suggest diminished periosteal modeling and
potentially diminished endosteal resorption along the diaphysis associated with TBT
treatment. Measures of whole bone RNA transcripts suggest that TBT treatment perturbs
the balance between osteoblasts and osteoclasts in the trabecular compartment to favor
osteoblast activity and effect mineralization.
Effects of TBT on osteoclast differentiation in vitro
To determine whether TBT treatment resulted in dose-dependent transcription
and/or functional changes in osteoclasts specifically, primary bone marrow macrophages
from C57Bl/6J females were cultured in the presence of M-CSF and RANKL to induce
differentiation into mature osteoclasts. After 24 hours of RANKL-induced differentiation,
cells were treated with either Vh (DMSO) or TBT (20, 50, or 80 nM), and total RNA was
harvested to assess relative gene expression after 6 days. Consistent with whole bone
mRNA expression (Figure 3.4d), expression of both Nfatc1 and Acp5 were significantly
decreased by TBT relative to Vh-treated cultures (Figure 3.5a). Expression Ctsk was also
significantly decreased by all TBT concentrations (Figure 3.5a). However, the expression
of Ct1 did not decline and even showed a dose-dependent trend toward increasing with
TBT concentration (Figure 3.5b). Conversely, TBT significantly suppressed expression
of the anti-osteogenic semaphorin-4D (Sema4D)(137) (Figure 3.5b).
89
Although the in vivo results suggested no significant change in osteoclast number
with TBT treatment, the in vitro gene expression analysis suggested a suppression of
osteoclast differentiation. We therefore examined osteoclast number in vitro by staining
and quantifying TRAP-positive, multinucleated cells in culture after they were induced to
differentiate by recombinant RANKL and M-CSF in the presence of either Vh or TBT
(Figure 3.5c). After 6 days in culture, TBT treatment resulted in a slight, but nonsignificant dose-dependent increase in the number of TRAP-positive, multinucleated cells
(Figure 3.5d). Cell viability, as assessed by MTT activity, was not affected by TBT at 20
nM or 50 nM, though 80 nM showed a small but significant decrease; cell death (release
of lactate dehydrogenase) was not altered by TBT (Figure S3.2).
To assess the relative resorptive function of TBT-treated osteoclasts in vitro,
osteoclasts were differentiated on a synthetic inorganic surface under the same conditions
as the TRAP-positive cell count assay. After 6 days, resorption pits were counted and
sized. All three doses of TBT increased the average number of resorption pits, resulting
in a greater percentage of the total resorbed surface (Figure 3.5e). The average size of the
resorption pits did not differ from vehicle control (Figure 3.5e). Representative images
used for quantification are presented in Figure 3.5f.
We concluded that TBT exposure does not attenuate the population of osteoclasts
in vitro or in vivo. The gene expression data suggest that expression of the differentiation
program is impacted by TBT, but without a significant change in osteoclast cell number
or resorptive capacity. Rather, the data suggest that TBT modifies the type of osteoclast
90
that differentiates, altering expression of genes related to osteoclast-osteoblast
communication to favor osteoblastogenesis.
Activation of LXR-dependent pathways by TBT
Recent literature has highlighted the role of RXRs and LXRs in the
osteoclast.(13,234,235,237) Because TBT is a known activator of RXRα(63), and based on
previous results demonstrating TBT activation of RXR and LXR-dependent genes in
bone marrow stromal cells (Figure S3.3)(81), we hypothesized that TBT would activate
RXR and/or LXR-dependent pathways in vivo and in in vitro osteoclast cultures.
In the whole bone RNA, TBT exposure did not impact the expression of Lxra or
Lxrb (data not shown). TBT exposure was associated with a significant, 2-fold increase in
the expression of Abca1, an LXRβ gene target and cholesterol transporter, whereas
Srebp1c, an LXRα target gene and lipogenic transcription factor, was not significantly
upregulated (Figure 3.6a). Mafb, a target of either RXR homodimers or SREBP1c, was
significantly upregulated in samples from TBT-treated animals (Figure 3.6a).
In the differentiated osteoclast cultures, TBT induced a dose-dependent increase
in Abca1 expression (Figure 3.6b). The RXR-specific agonist LG100268 induced a
significant increase in Abca1 expression, implying activation of the RXRα:LXRβ
permissive heterodimer. This pattern was also seen in the expression of Srebp1c (Figure
3.6b). As expected, the LXR-specific agonist T0901317 efficaciously induced expression
of both LXR target genes. By comparison, the PPARγ-specific agonist rosiglitazone did
not induce expression of Abca1 or Srebp1c above the Vh control baseline. Mafb was
91
most efficaciously induced by TBT and LG100268, with no induction by rosiglitazone
and moderate induction by T0901317 (Figure 3.6b). These results indicate that TBT
induces activation of multiple nuclear receptors in osteoblasts and osteoclasts.
Discussion
Nuclear receptors play an important role in determining the differentiation and
function of both osteoblasts and osteoclasts. Consequently, the balance of bone resorption
and formation can be perturbed by activation of multiple nuclear receptors. The goal of
this study was to use tributyltin (TBT), which can directly activate RXRα and PPARγ(63)
and indirectly activate LXR,(81,238) to examine the in vivo outcomes and in vitro
mechanisms associated with RXRα dimer-related signaling in bone.
As a PPARγ agonist, the ability of TBT to induce adipogenesis and suppress
osteogenesis in differentiating osteoblasts has been characterized.(80,82,226,239) We began
by re-establishing this effect in primary bone marrow MSCs and found that cells derived
from female mice are more sensitive to TBT’s ability to suppress osteogenesis than those
from male mice (Figure 3.1). It is well known that mature female mice have lower
trabecular density and thinner bones compared to males, and that female-derived
osteoblasts have lower potential for differentiation.(144,246,247) We have previously shown
that female-derived BM-MSCs were more sensitive to rosiglitazone-induced suppression
of osteogenesis.(153) However, this is the first report of TBT displaying a sexually
dimorphic effect at the transcriptional level.
Here, we decided to use the increased sensitivity of female mice to TBT to probe
92
in vivo effects of TBT exposure. It has been reported that TBT suppresses ossification in
fetal mice.(226) The rationale for using skeletally mature mice was that PPARγ-mediated
bone loss is typically seen among aging individuals on thiazolidinedione therapy for type
II diabetes,(77,79) and thus mice with fully developed long bones at baseline (12 weeks of
age) would provide a more relevant model for studying this particular etiology of
osteoporosis. Furthermore, we wished to isolate the effect of TBT on ongoing modeling
and remodeling from longitudinal growth.
The thinner cortex at mid-diaphysis associated with TBT treatment was expected,
as this phenotype is consistent with the suppression of osteogenesis seen in vitro. Cortical
bone is built by periosteal bone deposition, and the smaller cortical diameter suggests that
the thinner cortex resulted from a failure of periosteal bone deposition. Importantly,
thinner (i.e. smaller cross-sectional area) bones are less structurally sound.(140) Therefore
this observed phenotype may be associated with increased susceptibility to fracture. On
the other hand, the medullary space also was smaller in the TBT-treated mice. Osteoclast
activity on the endosteal surface typically thins cortical bone. The smaller medullary
space suggests that osteoclast activity at the endosteal surface is being suppressed by
treatment with TBT.
Surprisingly, there was an increase in trabecular bone volume consistent with
increased osteoblast activity (an increase in trabecular number without an increase in
average trabecular thickness). This was corroborated by the increased expression of
Dmp1 (Figure 3.3), which is highly expressed in osteocytes.(105) This increase in bone
occurred despite the fact that TBT activates PPARγ. In vivo studies with the PPARγ
93
agonist rosiglitazone show clear reduction of trabecular bone in femur of treated animals
that is accompanied by increased marrow adipocytes.(74–76) Here, TBT-exposed bones
also showed increased adipogenesis (Figure 3.3a–c). Selective PPARγ activation has
been shown to increase marrow adiposity without affecting bone density,(211) though TBT
treatment also induced a clear increase in mineralization. These results suggest either that
TBT induces an uncoupling of the pro-adipogenic and anti-osteogenic roles of the
RXRα:PPARγ heterodimer (also suggested by Rahman et al. (2012))(209) or TBT activates
another pathway altogether.
Osteoclasts play a dual role in bone, both physically resorbing bone and signaling
to osteoblasts to initiate bone formation. Nuclear receptors have been shown to be
influential in regulating osteoclast differentiation. In vivo, osteoclast number can be
increased by a PPARγ agonist, though this effect is controversial.(232,233,248) LXR agonists
can protect against ovariectomy-induced bone loss by reducing osteoclast number.(234,235)
RXR homodimers maintain the response to M-CSF in pre-osteoclasts, but as a
heterodimer with LXR can suppress RANKL signaling in differentiating osteoclasts.(13)
The histological counts of osteoclasts and serum TRAP measurements indicated that
osteoclast number was not substantially altered by TBT in vivo, and TBT also did not
reduce osteoclast numbers in vitro. Therefore, a reduction in osteoclast number likely
does not explain the increased mineralization. However, this lack of change in osteoclast
number is inconsistent with the observed changes in gene expression from whole bone
(humerus) or osteoclast culture, where TBT suppressed osteoclast differentiation and
function markers (Figures 3.4 and 3.5, Nfatc1, Trap, Ctsk). The discrepancy between the
94
suppression of Nfatc1 in whole bone RNA, which would predict an overall decrease in
osteoclast number, and the histological data may in part be explained by assay sensitivity,
where the magnitude of Nfatc1 suppression is not sufficient to impact the population of
osteoclasts observed directly in the trabecular compartment. Additionally, in vitro
osteoclast cultures may be more sensitive to chemical perturbation due to the use of only
recombinant M-CSF and soluble RANKL rather than osteoblast-facilitated
differentiation.(122)
In the in vitro functional assay, TBT increased the number of resorption pits,
suggesting that the capacity of TBT-exposed osteoclasts to resorb inorganic bone matrix
was increased. Fuller et al. (2008) have shown that inhibition of Cathepsin K, secreted by
osteoclasts, prevents the degradation of anabolic signals (e.g. IGF-1) stored in the
inorganic matrix and increases their activity,(249) and this may provide a mechanism for
the observation of increased mineralization in cathepsin K-deficient mice.(250) Given the
suppression of Ctsk in TBT-treated cultures and whole bone RNA, it is possible that
osteoclasts in TBT-exposed mice have a decreased capacity to degrade osteogenic matrix
proteins, though not necessarily a diminished capacity to resorb inorganic material. As a
result, there may be an increase the local concentration of anabolic factors as they are
released from the matrix and not degraded, contributing to the observed increase in
trabecular mineralization.
Notably, expression of the Il-6 family cytokine cardiotrophin-1 (Ct-1) was
increased by TBT both in vivo and in vitro. Ct-1 is essential for osteoclast function, but
also acts as a stimulatory signal for osteogenesis.(135) On the other hand, TBT treatment
95
suppressed Sema4d expression in vitro, which inhibits osteoblastogenesis under normal
conditions(137) (whole bone expression of Sema4d was not significantly different between
Vh and TBT treated groups (data not shown)). Together, these expression profiles
suggest a change in the function of osteoclasts at the transcriptional level rather than an
overall change to the osteoclast population. This finding also supports the role of
osteoclast-osteoblast communication as a major determinant of TBT induced changes in
trabecular density, particularly in light of the finding that the Rankl:Opg ratio is not
decreased with TBT treatment (Figure 3.4), implying no effect on osteoblast-mediated
osteoclast differentiation.
TBT’s effect on bone has been primarily investigated in the context of PPARγ
agonism.(57) Although TBT also binds and activates RXRα with greater affinity than it
does PPARγ,(63) the role of TBT as an RXRα agonist in bone that is capable of activating
RXR:RXR homodimers and other RXRα heterodimers (e.g. RXRα:LXR) has received
less attention.(81,238,240) We confirmed TBT’s activation of RXR and LXR target genes
(previously shown in osteoblast(81) and osteoclast(238) cell lines) in whole bone and
primary osteoclast cultures, and show that it activates a pathway relevant to
osteoclastogenesis, namely the expression of Mafb. This gene has been shown to be a
target of RXR:RXR homodimers in osteoclast precursors and required for osteoclast
differentiation.(13) In maturing osteoclasts, Mafb is indirectly upregulated by
RXRα:LXRα activation via increased expression of SREBP1c, which suppresses
osteoclast maturation.(13) The TBT-induced upregulation of Mafb is consistent with the
decreases in Nfatc1 transcript levels. TBT could be increasing Mafb via RXRα
96
homodimers, RXRα:LXR heterodimers, or both. However, it is notable that the increase
in Mafb seen with TBT in vitro more closely reflects that of the RXR homodimer agonist
LG100268 rather than the LXR agonist T0901317 (Figure 3.6).
Both the LXR and RXR have been proposed as targets for interventions to combat
low bone density, given their roles in moderating osteoclast differentiation.(13,234,237) To
our knowledge, these are the first results to demonstrate an association between RXR or
LXR agonism and increased bone density in intact, adult mice rather than
ovariectomy(234) or inflammation-induced(235) models of bone loss. The results also show
for the first time that TBT’s role as a bone suppressive agent is dependent upon the level
of coupled remodeling that occurs as a given bone site. Use of TBT as a multi-functional
model compound in conjunction with the appropriate Cre-lox knockout models may
provide important insights into the relative contributions of different nuclear receptors in
determining the balance of bone resorption and formation.
We observed a unique phenotype in the long bones of female C57Bl/6J mice after
being treated with TBT. Despite an apparent slowing of bone apposition in the cortical
shell at mid-diaphysis, exposure to TBT resulted in a greater mineral content in the
trabecular space compared to Vh-treated controls. We found that that the prevalence of
osteoclasts in the trabecular space was not attenuated. Rather, the preponderance of bone
formation appears to be a result of augmented osteoclast function, in particular
osteoclast-osteoblast signaling favoring osteoblastogenesis. Future studies will determine
the overall pattern of bone growth over the course of TBT treatment and more precisely
97
characterize the role of RXRα in activating multiple nuclear receptor-mediated pathways
in bone.
98
Figures
FIGURE 3.1. TBT suppresses osteogenic differentiation pathway and
mineralization markers more potently and efficaciously in female-derived cultures
compared to male-derived cultures.
a.
Sex
ns
Sex
p=0.001
Sex
p<0.0001
Treatment
p=0.034
Treatment
p<0.0001
Treatment
p<0.001
Interaction
ns
Interaction
p=0.004
Interaction
1 .2
*
0 .4
L o g F o ld D iffe r e n c e
F o ld D iffe r e n c e
F o ld D iffe r e n c e
Dmp1
5
1 .6
0 .8
4
3
**
2
**
1
0
0 .0
Vh
10 50
TBT
Vh
ns
O sx
Runx2
10 50
TBT
Vh
nM
10 50
TBT
Vh
10 50
TBT
5
4
3
**
2
**
1
0
Vh
nM
10 50
TBT
Vh
10 50
TBT
nM
b.
Sex
p<0.0001
Sex
p=0.001
Sex
p<0.001
Treatment
p<0.0001
Treatment
p<0.0001
Treatment
p<0.0001
Interaction
p=0.038
Interaction
p=0.063
Interaction
p=0.054
B o n e N o d u le s
**
N o d u le N u m b e r /c m
F o ld D iffe r e n c e
1 .0
**
0 .5
0 .0
Vh
10
50
TBT
F e m a le
M a le
Vh
10
1 .2
**
40
nM
**
20
0
10 50
TBT
Vh
10 50
TBT
0 .4
**
Vh
nM
*
*
0 .8
0 .0
Vh
50
TBT
M in e r a liz a tio n
60
F o ld D iffe r e n c e
2
A lk a lin e P h o s p h a ta s e
1 .5
10 50
TBT
Vh
10 50
TBT
nM
99
Figure 3.1. TBT suppresses osteogenic differentiation and mineralization more
potently and efficaciously in female-derived compared to male-derived cultures.
BM-MSCs were harvested from 9-week old female and male, C57Bl/6J mice, cultured in
the presence of osteogenic media and treated with Vh (DMSO) or TBT (10 nM, 50 nM)
for 7 (gene expression) or 10 (phenotype markers) days. a.) Relative mRNA expression
of osteoblast/osteocyte differentiation markers Runx2, Osx, and Dmp1. b.) Alkaline
phosphatase activity (left), nodule count (center) and mineralization (right). Data are
presented as mean ± SE. n=4–5 independent cultures. *p < 0.05, **p<0.01, Two-way
ANOVA (Sidak’s multiple comparison test),
100
FIGURE 3.2. TBT reduces diaphysis cross-sectional area and cortical thickness
while increasing trabecular structure in female C57Bl/6J femurs.
a.
Vh
T B T 1 0 m g /k g
b.
C t .T h
C t .A r
0 .2 6
1 .1
1 .1
0 .2 0
0 .9
T B T 1 0 m g /k g
0 .7
Vh
T t .A r
T B T 1 0 m g /k g
Vh
1600
0 .5 6
0 .5 4
m g H A /c m
3
**
2 .0
%
2
0 .5 2
1 .8
0 .5 0
0 .4 8
1 .6
0 .4 6
1 .4
T B T 1 0 m g /k g
1550
1500
1450
1400
1350
0 .4 4
Vh
T B T 1 0 m g /k g
M e a n M in e r a l D e n s it y
C t.A r /T t .A r
2 .2
mm
2
0 .7
Vh
0 .9
0 .8
0 .8
0 .1 8
**
1 .0
**
1 .0
mm
0 .2 2
mm
**
2
0 .2 4
mm
M a .A r
1 .2
Vh
T B T 1 0 m g /k g
Vh
T B T 1 0 m g /k g
101
FIGURE 3.2. (cont.)
c.
Vh
T B T 1 0 m g /k g
d.
B o n e V o lu m e F r a c tio n
0 .0 4
0 .0 2
0 .0 0
150
***
3
180
160
1 /m m
m g H A /c m
%
0 .0 6
***
200
3
0 .0 8
C o n n .D
M e a n M in e r a l D e n s it y
***
0 .1 0
140
T B T 1 0 m g /k g
0
Vh
T r a b e c u la r N u m b e r
Vh
T B T 1 0 m g /k g
T r a b e c u la r T h ic k n e s s
4 .0
***
T b .S p
0 .0 5 0
3 .5
0 .3 5
mm
***
mm
0 .0 4 5
1 /m m
T B T 1 0 m g /k g
0 .4 0
0 .0 5 5
3 .0
50
120
100
Vh
100
0 .0 4 0
0 .0 3 5
0 .3 0
0 .0 3 0
2 .5
0 .0 2 5
Vh
T B T 1 0 m g /k g
0 .2 5
Vh
T B T 1 0 m g /k g
Vh
T B T 1 0 m g /k g
102
Figure 3.2. TBT reduces diaphysis cross-sectional area and cortical thickness while
increasing trabecular structure in female C57Bl/6J femurs. 12-week old female wild
type C57Bl/6J mice were treated with sesame oil (Vh) (n=12) or 10 mg/kg TBT (n=11)
via oral gavage for 10 weeks. a.) Representative micro-CT images of mid-diaphysis.
Scale bar = 400 µm. b.) Cortical bone parameters. Ct.Th: cortical bone thickness; Ct.Ar:
cortical bone area; Ma.Ar: medullary (marrow) area; Tt.Ar: total cross-sectional area;
Ct.Ar/Tt.Ar: cortical area fraction; mineral density. c.) Representative micro-CT images
of distal metaphysis. Scale bar = 400 µm. d.) Trabecular bone parameters. BV/TV: bone
volume fraction; mineral density; Conn.D: connectivity density; Tb.N: trabecular
number; Tb.Sp: mean trabecular spacing; Tb.Th: mean trabecular thickness. Data are
presented from individual mice, and the mean is indicated by a line. n=11–12 individual
mice. **p < 0.01, *** p < 0.0001 unpaired t-test.
103
FIGURE 3.3. In vivo TBT exposure increases both trabeculae and marrow
adipogenesis.
a.
T B T 1 0 m g /k g
Vh
c.
b.
P lin 1
400
1 .5
F o ld D iffe r e n c e
n u m b e r o f a d ip o c y t e s /b o n e
A d ip o c y t e c o u n t
300
200
100
p = 0 .0 6 1
1 .0
0 .5
0 .0
0
Vh
TBT
Vh
TBT
d.
O sx
2
1
0
2 .5
F o ld D iffe r e n c e
2 .5
F o ld D iffe r e n c e
F o ld D iffe r e n c e
3
2 .0
1 .5
1 .0
0 .5
0 .0
Vh
TBT
TBT
3
**
2 .0
1 .5
1 .0
0 .5
2
1
0
0 .0
Vh
Sost
Dm p1
F o ld D iffe r e n c e
Runx2
Vh
TBT
Vh
TBT
104
Figure 3.3. In vivo TBT exposure increases both trabeculae and marrow
adipogenesis. Mice were treated as described in Figure 3.2. a.) Representative images of
5 µm slices of distal femur (n=4 Vh, 5 TBT) stained with hematoxylin and eosin
(adipocytes indicated by red arrows). b.) Adipocyte count per bone section (1.5 mm from
growth plate). Whole humerus bone mRNA expression (n = 12 Vh, 11 TBT 10 mg/kg) of
c.) the adipocyte marker, perilipin (Plin1), d.) osteoblast (Runx2, Osx) and osteocyte
(Dmp1) differentiation markers, and e.) the osteocyte signaling protein Sost. Data are
presented as mean ± SE. *p < 0.05, **p < 0.01, Mann-Whitney.
105
FIGURE 3.4. In vivo TBT exposure modifies osteoclast gene expression without
influencing osteoclast cell number.
a.
b.
O c /B .P e r .
T r a b . P e r im e t e r
20
15
m
# T r a p + c e lls /n m
10
5
10
5
0
0
Vh
TBT
c.
Vh
TBT
d.
S e ru m T ra p
F o ld D iffe r e n c e
U n it s /m l
4
2 .5
1 .5
1 .0
0 .5
2 .0
**
1
0
Vh
TBT
C t1
4
F o ld D iffe r e n c e
*
3
0 .6
2
0 .4
1
0
0 .0
Vh
TBT
2
1
0 .5
R a n k l /O p g
0 .2
*
2
e.
0 .8
3
1 .0
TBT
1 .0
C ts k
3
1 .5
0 .0
0 .0
Vh
Acp5
N fa tc 1
2 .0
Vh
TBT
0
Vh
TBT
Vh
TBT
106
Figure 3.4. In vivo TBT exposure modifies osteoclast gene expression without
influencing osteoclast cell number. Mice were treated as described in Figure 3.2. a.)
Representative images of 5 µm slices of distal femur stained for TRAP (red shading; n=4
Vh, 5 TBT). b.) Oc/B.Per.: number of TRAP-positive osteoclasts per nm of trabecular
perimeter, excluding cortex. Trab. Perimeter: total perimeter of trabecular bone (um). c.)
Quantification of serum TRAP by ELISA (n = 12 Vh, 11 TBT 10 mg/kg). Whole
humerus bone mRNA expression (n = 12 Vh, 11 TBT 10 mg/kg) of d.) Osteoclast
differentiation markers Nfatc1, Apc5 and Ctsk, and e.) Intercellular communication
proteins Rankl/Opg (osteoblast to osteoclast) and Ct1 (osteoclast to osteoblast). Data are
presented as mean ± SE. *p < 0.05 **p < 0.01, Mann-Whitney.
107
FIGURE 3.5. TBT suppresses differentiation and functional gene expression but
does not change osteoclast cell number in vitro.
a.
A cp5
N fa tc 1
10000
**
**
**
F o ld D if f e r e n c e
C ts k
600
10
8
400
6
5000
4
200
2
0
0
0
Vh 20 50 80
Vh 20 50 80
Vh 20 50 80
TBT
TBT
TBT
(n M )
(n M )
(n M )
b.
Sem a4D
C t1
2 .5
F o ld D if f e r e n c e
2 .0
n .s .
2 .0
1 .5
1 .5
n .s .
1 .0
1 .0
0 .5
*
0 .5
0 .0
0 .0
Vh 20 50 80
Vh 20 50 80
TBT
TBT
(n M )
(n M )
c.
Vh
TBT 50 nM
TBT 20 nM
TBT 80 nM
108
FIGURE 3.5. (cont.)
e.
A vg. C ount
80
60
40
20
0
**
20
10
0
Vh 20 50 80
TBT
% area
5
*
A v g . S iz e
p =
10
0 .0 6 4
4
3
2
30
# o b je c t s /f ie ld
T r a p + , M u N C /w e l l
100
mm
T r a p + c e ll c o u n t s
% r e s o r b e d s u r f a c e /f ie l d
d.
2
5
1
0
0
Vh 20 50 80
Vh 20 50 80
Vh 20 50 80
T B T (n M )
T B T (n M )
T B T (n M )
(n M )
f.
Vh
TBT 20 nM
TBT 50 nM
TBT 80 nM
109
Figure 3.5. TBT suppresses differentiation and functional gene expression but does
not change osteoclast cell number in vitro. Primary bone marrow macrophages were
isolated from 9-week-old, female C57BL/6J mice, induced to differentiate to osteoclasts
with M-CSF and RANKL (see Methods), and after 24 hrs treated with Vh (DMSO) or
TBT at indicated concentrations for 6 days. mRNA expression of a.) Osteoclast
differentiation (Nfatc1, Acp5) and functional markers (Ctsk), and b.) Osteoclast signaling
molecules Ct-1 and Sema4d, c.) Representative images of TRAP-positive, multinucleated
(3 or more nuclei) cells (Scale bar = 400 µm). d.) Average number of TRAP-positive,
multinucleated cells per well (n=4 separate experiments). e.) Average number of
resorption pits per field of view, percent of mineral surface area covered by resorption
pits, and average size of resorption pits (excluding pits < 1,100 µm2). f.) Representative
images of resorption pits, following thresholding for image analysis (resorption pits in
red, scale bar = 400 µm). n=5–7 independent cultures. Data are presented as mean ± SE.
*p < 0.05, **p < 0.01 compared to Vh, one-way ANOVA (Dunnett’s).
110
FIGURE 3.6. TBT activates RXR and LXR-dependent pathways in whole bone and
in vitro osteoclast culture.
a.
Abca1
M a fb
0 .8
0 .6
0 .4
0 .2
1 .0
F o ld D iffe r e n c e
1 .0
**
F o ld D iffe r e n c e
F o ld D iffe r e n c e
1 .0
S re b p 1 c
0 .8
0 .6
0 .4
0 .2
0 .0
0 .0
Vh
**
0 .8
0 .6
0 .4
0 .2
0 .0
TBT
Vh
TBT
Vh
TBT
b.
A bca1
S re b p 1 c
**
4 .0
p =
**
0 .0 9 2
3 .0
2 .0
1 .0
**
40
F o ld D if f e r e n c e
F o ld D if f e r e n c e
5 .0
35
30
25
8
*
6
4
2
0 .0
0
3
7
8
100 nM
1
6
F o ld D if f e r e n c e
T
.
8
7
6
.
100 nM
1
2
si
(n M )
3
G
o
TBT
T
L
R
80
2
0 .0
si
0 .5
50
G
o
1 .0
20
L
**
1 .5
Vh
80
R
7
8
M a fb
*
TBT
(n M )
100 nM
2 .5
2 .0
50
1
6
.
(n M )
20
3
si
2
o
TBT
Vh
T
80
G
50
L
20
R
Vh
111
Figure 3.6. TBT activates RXR and LXR-dependent pathways in whole bone and in
vitro osteoclast culture. a.) Mice were treated as described in Figure 3.2. Whole humerus
bone mRNA expression of LXR-dependent genes Abca1, Srebp1c, and the RXRdependent gene Mafb. Data are presented as mean ± SE. n = 11–12 individual mice.
**p<0.01, Mann-Whitney b.) Osteoclasts cultures were prepared as described in Figure
3.5 and treated with Vh (DMSO), TBT (20, 50, or 80 nM), rosiglitazone (PPARγ agonist,
100 nM), LG100268 (LG268, RXRα agonist, 100 nM), or T0101317 (T317, LXRα/β
agonist, 100 nM) for 6 days. mRNA expression of LXR- and RXR-dependent genes.
Data are presented as mean ± SE. n=6 independent cultures. *p<0.05 **p<0.01 compared
to Vh, one-way ANOVA (Dunnett’s). TBT treatments were compared to Vh separately
from control comparisons.
112
Supplemental materials
Supplementary Table S3.1. List of primers used for qPCR.
Primer
Manufacturer Sequence/Catalog #
F: CTTCCCACATTTTTGCCTGG
Abca1
IDT
Acp5
IDT
B2m
IDT
Ct1
IDT
Ctsk
IDT
Dmp1
IDT
Lxra
Qiagen
IDT
Lxrb
IDT
Mafb
IDT
Nfatc1
IDT
Opg
IDT
Osx
Qiagen
Plin1
IDT
Rankl
IDT
Rn18s
IDT
Runx2
Qiagen
IDT
Sema4D
Qiagen
IDT
Sost
IDT
Srebp1c
IDT
R: AAGGTTCCGTCCTACCAAGTCC
F: CCATTGTTAGCCACATACGG
R: ACTCAGCACATAGCCCACAC
F: CTGCTACGTAACACAGTTCCACCC
R: CATGATGCTTGATCACATGTCTCG
F: GGAGGGAAGTCTGGAAGACC
R: TGTTGCTGCACGTATTCCTC
F: GTTGTATGTATAACGCCACGGC
R: CTTTCTCGTTCCCCACAGGA
F: CAACTGGCTTTTCTGTGGCAA
R: TGGGTGCGCTGATGTTTGCT
Cat # QT01078210
F: GAGTTGTGGAAGACAGAACCTCAA
R: GGGCATCCTGGCTTCCTC
F: CCCCACAAGTTCTCTGGACAC
R: TGGCGGAGGTACTGGGC
F: CGCGTCCAGCAGAAACATC
R: AGCTGCTCCACCTGCTGAAT
F: CCCTTTAAAAATGAGGACAATAGCTTT
R: TTGCTGCCCTTTCACTGATG
F: TTGCCTGGGCTGCAGAGACG
R: CCAGGAGCACCAGGAGTGCG
Cat # QT00293181
F: GGGACCTGTGAGTGCTTCC
R: GTATTGAAGAGCCGGGATCTTTT
F: GCTGGGCCAAGATCTCTAAC
R: GTAGGTACGCTTCCCGATGT
F: GTAACCCGTTGAACCCCATT
R: CCATCCAATCGGTAGTAGCG
Cat # QT01036875
F: TTTAGGGCGCATTCCTCATC
R: TGTCCTTGTGGATTAAAAGGACTTG
Cat # QT00102193
F: CCTGGGAACATGGAGAGGTA
R: GGGCGCCTACATACAGAGTG
F: GCCTCATCTGCCTACTTGTG
R: CTGTGGCATCATTCCTGAAG
F: GGAGCCATGGATTGCACATT
R: GGCCCGGGAAGTCACTG
113
Supplementary Figure S3.1
BW % change
% C h a n g e in B W (g )
30
**
20
10
0
-1 0
b.
T h y m u s /B o d y
0 .2 5
T h y m u s /B o d y W e ig h t
a.
0 .2 0
0 .1 5
0 .1 0
0 .0 5
0 .0 0
Vh
T B T 1 0 m g /k g
Vh
T B T 1 0 m g /k g
Figure S3.1. 10 mg/kg TBT via oral gavage does not cause overt systemic toxicity in
vivo. 12 week old female C57Bl/6J mice were treated with either vehicle (Vh, sesame oil,
10 µg/L) or TBT (10 mg/kg bw in sesame oil) via oral gavage 3x/week for 10 weeks
(n=12 for both groups). Body weight was measured at each dosing. a.) Percent change in
body weight (g) from first weekly average and last weekly average. b.) Thymus weight as
a percentage of body weight at euthanasia. n=11 Vh (1 thymus discarded due to
dissection error), n=12 TBT. **p<0.01, unpaired t-test.
114
Supplementary Figure S3.2
Figure S3.2. TBT does not cause overt toxicity at concentrations used in vitro.
Primary bone marrow cells were isolated from 12 week old female C57Bl/6J mice and
allowed to differentiate with MCSF (naïve) or MCSF and RANKL (see Methods). The
MSC medium was replaced, and the cultures were treated with Vh (DMSO, 0.1%) or
TBT (20, 50, 80 nM, 2 µM) and incubated for 5 days. a.) Cellularity was assessed by 3[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) labeling for 3 hrs by
standard methods. b.) Necrosis was assessed by measuring dead cell protease release
(CytoTox-Glo™ Cytotoxicity Assay, Promega). Absorbance and luminescence in
experimental wells was normalized by dividing by that measured in naive wells. Data are
presented as means ± SE of 4 independent bone marrow preparations. A 2hr treatment
with 2 μM TBT was used as a positive control.
115
Supplementary Figure S3.3
Figure S3.3. TBT activates LXR-dependent pathways in BM-MSCs. Primary bone
marrow cells were isolated from 9 week (n=3) and 12 week (n=3) female C57Bl/6J mice
and allowed to differentiate for 7 days. The media was replaced with osteogenic media
(see Methods) and either Vh (DMSO, 0.1%) or TBT (10, 50 nM). After 7 days, total
RNA was harvested for relative gene expression with qPCR. Relative gene expression
was determined using the Pfaffl method to account for differential primer efficiencies.(223)
The average Cq value for 18s ribosomal RNA (Rn18s) and beta-2-microglobulin (B2m)
was used for normalization. The Cq value for naïve, undifferentiated cultures was used as
the reference point, and the data are reported as “Fold Change from Naive.” * p<0.05,
ANOVA (Dunnett’s).
116
CHAPTER IV: Generalized Concentration Addition modeling predicts mixture
effects of environmental PPARγ agonists
James Watt, Thomas F. Webster, Jennifer J. Schlezinger
Department of Environmental Health, Boston University School of Public Health,
Boston, MA, 02118, USA
Citation: Watt J, Webster TF, Schlezinger JJ. Generalized Concentration Addition
modeling predicts mixture effects of environmental PPARγ agonists. Toxicol. Sci. Jun 2.
[Epub ahead of print].
Acknowledgements
This work was supported by the National Institute of Environmental Health Sciences
Superfund Research Program [P42ES007381].
Abbreviations: AHR: aryl hydrocarbon receptor; BBP: butyl benzyl phthalate; DEHP: diethyl hexyl phthalate; DR-1: direct repeat sequence with 1 nucleotide spacer; ES: Effect
Summation; GCA: Generalized Concentration Addition; IA: Independent Action; MBP;
mono benzyl phthalate; MBzP: mono benzyl phthalate; MEHP: mono-2-ethyl hexyl
phthalate; nTZDpa: non-thiazolidinedione partial agonist; PPARγ: peroxisome
proliferator activated receptor gamma; RXRa: retinoid-X receptor alpha; TEF: Toxic
Equivalency Factor
117
Abstract
The vast array of potential environmental toxicant combinations necessitates the
development of efficient strategies for predicting toxic effects of mixtures. Current
practices emphasize the use of concentration addition to predict joint effects of endocrine
disrupting chemicals in co-exposures. Generalized Concentration Addition (GCA) is one
such method for predicting joint effects of co-exposures to chemicals and has the
advantage of allowing for mixture components to have differences in efficacy (i.e., doseresponse curve maxima). PPARγ is a nuclear receptor that plays a central role in
regulating lipid homeostasis, insulin sensitivity, and bone quality and is the target of an
increasing number of environmental toxicants. Here, we tested the applicability of GCA
in predicting mixture effects of therapeutic (rosiglitazone and non-thiazolidinedione
partial agonist) and environmental PPARγ ligands (phthalate compounds identified using
EPA’s ToxCast database). Transcriptional activation of human PPARγ1 by individual
compounds and mixtures was assessed using a PPRE-driven luciferase reporter. Using
individual dose-response parameters and GCA, we generated predictions of PPARγ
activation by the mixtures, and we compared these predictions to the empirical data. At
high concentrations, GCA provided a better estimation of the experimental response
compared to three alternative models: Toxic Equivalency Factor, Effect Summation and
Independent Action. These alternatives provided reasonable fits to the data at low
concentrations in this system. These experiments support the implementation of GCA in
mixtures analysis with endocrine disrupting compounds and establish PPARγ as an
important target for further studies of chemical mixtures.
118
Introduction
Exposure to chemicals is rarely on a single-chemical basis, requiring risk
assessments to address chemical mixtures. Because testing all mixture combinations is
experimentally impossible, risk assessors and toxicologists employ models that predict
mixture effects from minimal data that may vary chemical-to-chemical, such as doseresponse curves characterizing a compound’s potency (e.g. EC50) and efficacy (maximum
response) when acting on a specific biological target.
Current chemical mixtures analysis for endocrine disrupting compounds
emphasize concentration addition (also called dose addition),(251) whereby individual
component concentrations producing a given mixture effect are scaled by their proportion
in the mixture and summed (for example, see Rajapakse et al. (2001)).(252) This method is
generally preferred for chemicals that act by similar mechanisms over models such as
Independent Action or Effect Summation. For example, concentration addition may be
applicable in cases of mixtures with components that act via the same molecular pathway,
such as mixtures of dioxin-like polychlorinated biphenyls (PCBs) that act through the
aryl hydrocarbon receptor (AHR), a ligand-activated transcription factor with one ligandbinding site.
Toxic equivalency factor (TEF) is a special case of concentration addition
whereby the relative potency of a mixture component is used to generate an equivalent
dose of a well-characterized reference compound.(253) In cases where the dose-response
curves of all the components differ only in potency (same shape and maximal effect),
TEFs are generated by comparing the potency of each component to a reference
119
compound, e.g., 2,3,7,8-TCDD for the dioxin-like compounds.(253) TEF uses the toxic
equivalencies to calculate an equivalent total dioxin dose. The effect of the mixture is
then calculated from the dose-response curve of dioxin. This model has the advantage of
easy interpretation and implementation where reference chemical data are available, but
crucially relies on the assumption that the dose-response curves of component chemicals
differ only in potency; the assumptions for use of TEFs are violated when compounds
differ in efficacy.
Howard and Webster (2009) introduced Generalized Concentration Addition
(GCA) to allow for differences in efficacy among mixture components.(254) This model
uses inverse mathematical functions allowing for some components to have submaximal
efficacy. They demonstrated the utility of this model toward a mixture of AhR agonists
with different efficacies.(255) GCA accurately fit the experimental data of a full and a
partial AhR agonist mixture, whereas TEF did not predict the competitive effect of a
partial agonist at high concentrations in a mixture.
Current research has identified peroxisome proliferator-activated receptor gamma
(PPARγ) as an important target molecule in endocrine disruption and developmental
biology. PPARγ is a nuclear receptor that acts as a metabolic sensor and forms an
obligate, permissive heterodimer with the retinoid X receptor alpha (RXRα).(27) As the
list of environmental toxicants that activate PPARγ grows in number and structural
variety,(64,66–68,256) there is increasing concern over the influence of environmental
chemical exposure on PPARγ-related endpoints, namely lipid homeostasis (for review
see Feige et al., 2006),(257) obesity,(61,258) and bone homeostasis.(15,80,81,239) The extent of
120
exposures and toxicological importance of PPARγ ligands further support the need for
the development of flexible models to predict mixture effects.
Here, we tested the applicability of GCA in modeling mixture effects of full and
partial PPARγ agonists and compared GCA’s performance against the TEF, Independent
Action (IA) and Effect Summation (ES) models. IA and ES are arguably inappropriate
here: IA is intended to be applied when compounds act by different modes of action,
while ES is generally considered inappropriate by mixtures toxicologists.(259) However,
investigators do not always know the mechanism of action and it is not always clear how
similar mechanisms need to be for CA to apply.(260) It is therefore useful to see how these
different models compare with GCA.
Using human PPARγ1 and RXRα and a PPRE response element-dependent
luciferase reporter, we generated dose-responses for individual ligands. We predicted
mixture effects using the different additivity models, and compared these predictions with
experimental mixture results. We establish GCA as a model for predicting mixture effects
of synthetic PPARγ agonists and show it to be a reasonable additive model of PPARγ
activation by mixtures, including therapeutic drugs and phthalates that we identified as
PPARγ agonists from the ToxCast data set.
Materials and Methods
Chemical Suppliers
Rosiglitazone was obtained from Cayman Chemicals (Ann Arbor, MI).
Nonthiazolidinedione partial agonist (nTZDpa) was from Tocris Biochemicals (Bristol,
121
UK). Di-(2-ethylhexyl) phthalate (DEHP), mono-(2-ethylhexyl) phthalate (MEHP),
mono-butyl phthalate (MBP), butyl-benzyl phthalate (BBP), and mono-benzyl phthalate
(MBzP) were obtained from Sigma (St. Louis, MO).
Reporter Assay
Cos7 cells were plated in 96-well plates and transiently transfected with human
PPARG1 (provided by V.K. Chatterjee, University of Cambridge, Cambridge, UK) and
human RXRΑ (plasmid 8882, Addgene, Cambridge, MA)(27) expression vectors, with
peroxisome proliferator response element 3x-TK-Luc (plasmid 1015, Addgene)(224) and
cytomegalovirus-enhanced green fluorescent protein reporter using Lipofectamine 2000
(Invitrogen, Carlsbad, CA). After overnight incubation at 37C, the media was replaced
and cells received no treatment (medium only), Vh (dimethyl sulfoxide, 0.1%), or PPARγ
ligands at a range of concentrations either alone or in combination (rosiglitazone 1 x 10-10
M – 1 x 10-6 M; nTZDpa 1 x 10-9 M – 5 x 10-6 M; DEHP 1 x 10-7 M – 2 x 10-4 M; MEHP
1 x 10-7 M – 1 x 10-4 M; BBP 1 x 10-7 M – 5 x 10-4 M; MBP 1 x 10-7 M – 5 x 10-4 M;
MBzP 1 x 10-7 M – 5 x 10-4 M). Concentrations of individual chemicals were chosen
using pilot studies designed to identify the highest, non-toxic dose. After 24h of
treatment, cells were lysed with GloLysis buffer and mixed with BrightGlo Luciferase
buffer (Promega Inc., Madison, WI). Luminescence and fluorescence were determined
using a Synergy2 plate reader (Biotek Inc., Winooski, VT). Luminescence of each well
was normalized to green fluorescent protein fluorescence, and the resulting values were
normalized to medium-only wells to obtain fold change in PPARγ activation.
122
ToxCast PPARγ agonist identification
ToxCast assay and chemical data were downloaded from EPA’s ToxCast portal
(http://www2.epa.gov/chemical-research/toxicity-forecaster-toxcasttm-data) in May of
2015. Chemicals were selected for inclusion that showed dose-response data fit with a
monotonic Hill function in the Attagene PPRE_cis_up assay and the Attagene
PPARγ_trans_up assays. To limit the selection to chemicals that did not also activate
RXRα, these candidates were limited to those that fit RXRα-specific assays Odyssey
Thera NURR1_NURR1RXRa_1440, Attagene RXRa_TRANS_up, and Odyssey Thera
NURR1_NURR1RXRa_0480 with a constant (i.e. no response) model. Of the resulting
88 chemicals from this screening process, we identified five phthalates for testing, two
parent phthalate compounds (di-ethyl hexyl phthalate (DEHP) and butyl-benzyl phthalate
(BBP)) and three phthalate metabolites (mono-ethyl hexyl phthalate (MEHP), monobutyl phthalate (MBP), and mono-benzyl phthalate (MBzP)).
Experimental Fits
GCA requires specification of mathematical dose response functions that are
invertible yielding real numbers for all effect levels of interest. As in our previous
work,(255) we chose the function based on a biologically-based model for binding of a
ligand to a receptor at a single binding site followed by activation, where the degree of
activation can depend on the ligand; the compounds we are investigating bind PPARγ,
not RXRα. This model yields Hill functions with a Hill coefficient of 1. Individual
chemical dose-response data were therefore fit to the following equation:
123
f([A]) = min + (max-min)([A])/(KA+ [A])
(1)
where KA is the macroscopic equilibrium constant for agonist A (EC50, equal to the dose
producing a 50% maximal effect), min and max are the curve minimum and maximum
effect levels, respectively, and [A] is the dose of ligand A. For mixtures experiments, data
points were adjusted by subtracting the response value of Vh-only treated wells, allowing
the curves to be fit with a two-parameter Hill function: f([A]) = αA([A])/(KA + [A]), where
αA is the maximum effect level.
Toxic Equivalency Factor (TEF)
TEF is a special case of concentration addition (and GCA) requiring parallel doseresponse curves, including the same efficacy. When efficacies differ, TEFs will not in
principle work. However, relative potencies may still be computed and may provide
approximate answers in some cases,(255) though results will depend on the choice of
reference compound and its efficacy. For a reference compound, we used rosiglitazone,
the compound with the highest potency and efficacy, and traditionally used as a positive
control for PPARγ activity. The relative potency (γi) is determined by γi =KA/Ki, where A
is the reference compound rosiglitazone. The total effect of the mixture, estimated for n
components, is:
𝛼𝐴 ([𝐴]+∑𝑛
𝑖=1 𝛾𝑖 [𝑋𝑖 ])
𝐸𝑇𝐸𝐹 = 𝑓𝐴 ([𝐴] + ∑𝑛𝑖=1 𝛾𝑖 [𝑋𝑖 ]) = 𝐾
𝑛
𝐴 + ([𝐴]+∑𝑖=1 𝛾𝑖 [𝑋𝑖 ])
(2)
124
Generalized Concentration Addition (GCA)
GCA relaxes the requirement of equal dose-response maxima. The general GCA
equation for n components is:
[𝑋 ]
1 = ∑𝑛𝑖=1 𝑓 −1𝑖(𝐸)
(3)
𝑖
where fi -1 (E) is the mathematical inverse of the individual chemical dose-response
function and E is the mixture effect level. The inverse of the two-parameter Hill function
above is fi -1 (E) = EKi / (a1 - E) , a function which yields real numbers for positive values of
αi, Ki and E. Substituting this equation into (3) allows for easy calculation of the effect
level at a given mixture of n components with efficacies αi and EC50s given by Ki:
𝐸𝐺𝐶𝐴 = 𝑓𝑚𝑖𝑥𝑡𝑢𝑟𝑒 ([𝑋1 ], … , [𝑋𝑛 ]) =
𝛼𝑖 [𝑋𝑖 ]
∑𝑛
𝑖=1
𝐾𝑖
[𝑋 ]
𝑛
1+ ∑𝑖=1 𝑖
𝐾𝑖
(4)
While (3) is the general definition of GCA, the form of (4) also depends on the specific
dose-response function being modeled and may not apply more generally.
Effect Summation (ES) and Independent Action (IA)
ES and IA are response-additive models that sum the individual responses, rather
than the doses, of the mixture components. For ES
125
𝐸𝐸𝑆 = 𝑓𝑚𝑖𝑥𝑡𝑢𝑟𝑒 ([𝑋]1 , … [𝑋]𝑖 ) = ∑𝑛𝑖=1 𝑓𝑖 ([𝑋]𝑖 )
(5)
IA is generally used for compounds that act by different mechanisms. IA assumes that
compounds have the same maximal effect of one:
𝐸𝐼𝐴 = 𝑓𝑚𝑖𝑥𝑡𝑢𝑟𝑒 ([𝑋]1 , … [𝑋]𝑖 ) = 1 − ∏
𝑛
(1 − 𝑓𝑖 ([𝑋]𝑖 ))
𝑖=1
When efficacies differ, the standard formula needs to be modified, and the best procedure
is not clear. Following Payne et al (2001),(261) we scaled the response of individual
components to the maximum efficacy, αmax , then scaled the overall results to the same
maximum:
𝐸𝐼𝐴 = 𝑓𝑚𝑖𝑥𝑡𝑢𝑟𝑒 ([𝑋]1 , … [𝑋]𝑖 ) = 𝛼𝑚𝑎𝑥 (1 − ∏𝑛𝑖=1 (1 −
𝑓𝑖 ([𝑋]𝑖 )
𝛼𝑚𝑎𝑥
))
(6)
The estimated maximum efficacy here was for MBzP, 7% larger than that of
rosiglitazone (Figure 4.3b). Scaling components by their own maximums (αi) produces
predictions that are larger and closer to those provided by the TEF model (data not
shown).
126
Statistics and software
Individual dose-response curves were fit and relevant parameters were estimated
using Prism (version 6, GraphPad). Combination dose-response surfaces were generated
in R using the wireframe() function. The R code for generating the additive model
predictions is available at the Boston University Superfund Research Program website
(www.busrp.org). Model extrapolations were performed in Mathematica (version 9.0.1.0,
Wolfram Alpha). Differences in experimental distributions and model predicted
distributions were tested using the non-parametric Wilcoxon Rank Sum and
Komolgorov-Smirnov tests.
Results
We began by establishing the dose-responses of two synthetic PPARγ ligands,
rosiglitazone (full agonist) and nTZDpa (partial agonist). Cos-7 cells were transfected
with expression vectors for human RXRα and PPARγ1 and a luciferase reporter gene
under transcriptional control of a PPRE response element; cells were treated with each
chemical separately and in combination. As shown in Figure 4.1a, the individual dose
response data were well fit by Hill functions with a Hill coefficient of 1 (the “S” shape of
the data in the figure is partly due to plotting concentration on the log scale). nTZDpa and
rosiglitazone stimulate PPARγ activation with similar potencies (EC50s of 2.2 x 10-8 M
and 7.1 x 10-8 M, respectively), which were comparable to established estimates (4.3 x
10-8 M for,(30) 5.7 x 10-8 M for nTZDpa(262)). nTZDpa is a significantly less efficacious
PPARγ ligand, with a response maximum of only ~17% of rosiglitazone (Figure 4.1a). A
127
PPARγ-dominant negative construct suppresses luciferase gene expression stimulated by
rosiglitazone in this system,(263) indicating that the increased luminescence as a function
of concentration is due specifically to PPARγ binding and activation by the agonists.
We hypothesized that mixtures of rosiglitazone and nTZDpa would show a
reduction in overall PPARγ activation at effect levels above the efficacy of nTZDpa,
owing to the behavior of a partial agonist to act in the manner of a competitive antagonist
in a mixture. Figure 4.1b demonstrates that at high effect levels, the response to
rosiglitazone is antagonized and the overall activation of PPARγ by the mixture
decreases: note that the 1 x 10-11 nTZDpa curve (solid line, black circles) closely
approximates the rosiglitazone curve in 1a, but increasing concentrations of nTZDpa
reduce the curve closer to that of nTZDpa in Figure 4.1a. These data support the
conclusion that nTZDpa is outcompeting rosiglitazone for PPARγ binding and that a
higher proportion of partial agonist-bound receptors is attenuating the overall effect.
To compare the empirical dose response data with those predicted by different
additive models, three-dimensional response surfaces of PPARγ activation by the
rosiglitazone/nTZDpa combinations were plotted (Figure 4.2). Marginal (i.e. individual)
dose-response curves lie along the front edges of the box (in bold) and combination
effects comprise the interior surface. The antagonistic effect of nTZDpa is evident in the
dose response surface, with the surface dropping downward at high concentrations of
nTZDpa and rosiglitazone (Figure 4.2a). Using individual-chemical fit parameters (listed
in the embedded table in Figure 4.2), we generated predicted response surfaces under
different additive models (ES, TEF, IA and GCA).
128
The effect summation (ES) model, generated from equation 5, sums the individual
effect levels of each combination of rosiglitazone and nTZDpa. At low concentration
combinations, the ES surface approximates the empirical surface (Figure 4.2b). However,
at high concentration combinations, the modeled surface implies supra-maximal
activation, as the effect of nTZDpa is added to the effect of rosiglitazone.
Independent Action (IA, equation 6), assumes the observed individual effects
result from independent mechanisms of action. Like ES, at low concentrations the model
approximates the empirical surface (Figure 4.2c). Higher concentration combinations
begin to show the characteristic competitive antagonism effect of the full and partial
ligand mixture, but the model fails to fully capture the reduction in overall effect.
Toxic Equivalency Factor (TEF), defined in equation 2, assumes that the
efficacies of rosiglitazone and nTZDpa are equal, and that nTZDpa is a diluted form of
the more potent rosiglitazone. The modeled response surface generated by TEF shows no
antagonistic effect of nTZDpa at high concentrations (Figure 4.2d).
Because GCA was formulated to allow for differences in efficacy among mixture
components,(254) we hypothesized that the surface generated by GCA would most
accurately resemble the empirical surface. The GCA-modeled surface (equation 4)
captures the attenuating effect of high concentrations of nTZDpa (Figure 4.2e). The nonparametric Wilcoxon Rank Sum and Komologorov-Smirnov tests indicate that the
distribution of GCA-modeled estimates is not significantly different from empirical
results (p > 0.05), whereas the TEF and ES distributions do differ significantly from the
experimental data (p < 0.01 for both comparisons). Generation of modeled distributions
129
using the upper and lower bounds of the experimental standard errors resulted in minimal
changes to the modeled estimates (Supplementary Figure S4.1). From these results, we
concluded that GCA is a viable model for predicting combination effects of two PPARγspecific agonists that have different efficacies.
Next, we tested the applicability of GCA in modeling PPARγ activation
stimulated by mixtures of environmental ligands. From EPA’s ToxCast data set, we
identified potential PPARγ agonists, including five phthalate compounds: di-ethyl hexyl
phthalate (DEHP), butyl-benzyl phthalate (BBP), mono-ethyl hexyl phthalate (MEHP),
mono-butyl phthalate (MBP), and mono-benzyl phthalate (MBzP). Individually, all but
DEHP activated PPARγ in our assay. Table 1 lists the efficacy (as maximum-minimum)
and potency (EC50) derived from the dose-response curve of each parent (Figure 4.3a)
and metabolite (Figure 4.3b) compound. Though the potencies were mostly within the
same order of magnitude (with MBzP being slightly less potent than the other phthalates),
the efficacies varied roughly 4-fold between the least (MBP) and most efficacious
(MBzP).
We began by testing an array of rosiglitazone (full agonist) and MEHP (partial
agonist) combinations. Similar to rosiglitazone and nTZDpa mixtures, MEHP increases
PPARγ activation at low effect levels (left side of Figure 4.3c), but antagonizes the
PPARγ activation stimulated by rosiglitazone at higher effect levels, reducing the mixture
activation of PPARγ to the response maximum of MEHP (right side of Figure 4.3c). The
predictions of GCA fit the observed data somewhat better than the TEF model at high
dose combinations where a small amount of antagonism by MEHP becomes evident
130
(Figure 4.3d, 4.3e and 4.3f). The difference between the two models is not as stark as that
for the rosiglitazone/nTZDpa mixture (Figure 4.2). The reason is that the estimated
efficacy for MEHP is 80% that of rosiglitazone, far closer in maximal effect compared to
nTZDpa.
Next, we tested a complex mixture of phthalates. We hypothesized that, given the
different efficacies among the phthalates, GCA would again provide the most accurate fit,
while the response-additive models and TEF would overestimate the mixture effect. To
test this hypothesis, we constructed two mixtures of multiple phthalates. Because of the
logistical complication of testing mixtures with more than two components, we used ray
designs in which each component is in constant proportion to one another across mixture
dilutions. We excluded DEHP from the rays because of its lack of activity in the PPARγ
activation assay. The starting solution for the first ray (A) consisted of BBP, MEHP,
MBP, and MBzP at concentrations just below their respective EC10 values, as determined
by their individual dose-responses (Table 4.2). The starting solution for the second ray
(B) combined the four phthalates at concentrations that produced the same effect level as
10% of rosiglitazone’s maximum effect (Table 4.2). Each mixture was diluted by factors
of 2, 5, 10, 20, 50, and 100, and the dilution series were applied in the PPARγ activation
assay. Using the individual dose response parameters listed in Table 4.1, mixture
responses also were predicted using the ES, TEF, GCA, and IA models.
The starting concentrations of mixtures A and B produced effect levels that were
7% and 30% higher, respectively, than the predicted individual effect of the most
efficacious component (MBzP) at the concentration at which it was added to each
131
mixture (1.6 x 10-5 M, Table 2). By the second and fourth dilutions of ray A and B,
respectively, the activation of PPARγ by the mixture was indistinguishable from baseline
(Figure 4.4a and 4.4b). At low concentrations, all four models — GCA, TEF, ES and IA
— adequately fit the empirical data. At higher concentrations, the TEF predictions
strongly diverge from the observed data, while GCA may provide a slightly better fit than
ES and IA. Figure 4.4c provides a closer comparison of the starting mixture
concentrations and the modeled values.
To examine the behavior of these models at high concentrations, we used the
same fitting parameters to extrapolate models to high dose levels for ray A. As shown by
Figure 4.5, the models diverge with increasing mixture dose. At high mixture doses, the
ES model vastly exceeds the other models. Both the TEF and IA models overestimate
effects compared to GCA; this occurs in part because only GCA takes into account the
antagonistic effect of partial agonists.
Discussion
Current methods for estimating risk from chemical exposure predominately focus
on single chemical exposure instead of the more realistic scenario of co-exposure to
multiple chemicals. At the same time, several agencies are recognizing the need to focus
research on low dose exposures to endocrine disrupting chemicals that act via
inappropriate activation of nuclear receptors.(70,264,265) One such nuclear receptor is
PPARγ, which plays a central role in toxic mechanisms related to lipid homeostasis, bone
quality, and other endpoints (for review, see Ahmadian et al., 2013).(266) Given its status
132
as a target of endocrine disruption by a variety of environmental agonists, PPARγ is an
important model for which to study chemical co-exposures and develop ways to predict
toxicity of chemical mixtures. Here, we established the utility of Generalized
Concentration Addition (GCA)(254) in modeling joint effects of chemical co-exposures
that activate PPARγ.
The experimental design used to generate experimental mixture data and test the
additive models was similar to that of Howard et al. (2010),(255) who used GCA to model
mixture effects of aryl hydrocarbon receptor (AhR) agonists. In that study, they showed
experimentally that the partial AhR agonist (galangin) could decrease the overall mixture
effect when combined with a full AhR agonist (PCB126). This effect of galangin
resembled that of a competitive antagonist. The difference is that a partial agonist
increases the mixture effect at levels below its efficacy, but decreases the mixture effect
at levels above its efficacy.
We began by testing the utility of GCA in predicting the activation of PPARγ by a
mixture containing a full (rosiglitazone) and partial (nTZDpa) agonist. At high combined
concentrations, the activation of PPARγ was reduced by nTZDpa, and this was predicted
by GCA (Figure 4.2d). Because of overt toxicity and the comparable potencies between
rosiglitazone and nTZDpa, nTZDpa could not be added to a level that fully reduced the
activity of the mixture to that of the maximum effect of nTZDpa. However, GCA was
able to model the decrease in mixture effect, whereas the TEF model, in assuming equal
efficacy between the two components (and using rosiglitazone as the reference
compound), inflated the effect of high concentrations of nTZDpa (Figure 4.2c). ES and
133
IA proved inadequate as well (Figure 4.2a). These results indicate that GCA is a viable
model in the case of two chemicals designed specifically as PPARγ ligands.
Phthalates comprise a group of chemicals designed for use in products and
manufacturing but also show biological activity. Manufactured at high volume for use as
plasticizing agents and flame retardants, phthalates can act on numerous targets,
especially in the male reproductive system.(267–269) Specific phthalate esters (MEHP,
MBzP, MBP, and BBP) have also been shown to activate PPARγ.(64,270) Exposure to
phthalates is widespread and detected in human biomonitoring data, including National
Health and Nutrition Examination Survey.(271,272) Parent phthalate compounds such as
DEHP and BBP undergo biotransformation to multiple metabolites following absorption,
making phthalate mixtures biologically relevant.(273,274) Typically, phthalate metabolites
are more biologically active than the parent compounds with regard to activation of
PPARs.(64,275)
To identify phthalates that activate PPARγ, we used EPA’s ToxCast database (US
EPA, 2014). Because our assay uses a transfected luciferase reporter construct under the
transcriptional control of a peroxisome proliferator response element, we selected
compounds that activated an equivalent model assay in ToxCast. To increase the
specificity for identifying PPARγ ligands, we included chemicals that also activated a
PPARγ ligand binding domain fused with a Gal-4 reporter, and excluded chemicals that
showed any activity in assays for PPARγ’s permissive heterodimer partner, RXRα.
Contrary to the ToxCast screen, DEHP did not activate PPARγ transcriptional activity in
our assay. We speculate that this is due to differences in the source of the cell line
134
(human liver in ToxCast, African Green monkey for Cos-7 cells in our assay), though
others have shown that DEHP does not activate human PPARγ in vitro.(64)
In a binary mixture, MEHP reduced the activation of PPARγ by rosiglitazone at
higher concentrations of both compounds, which was predicted by GCA (Figure 4.3c and
4.3f). GCA also accurately predicted the activation of PPARγ by a more complex
phthalate mixture. However, ES, IA and TEF all approximated GCA and fit the observed
data at low concentrations of both rays. In contrast to our results, Silva et al. (2002)
demonstrated that ES and IA do not sufficiently predict mixture effects of estrogen
receptor full agonists at low concentrations; this is caused by the non-linear shape of the
dose-response curves for these compounds at low doses.(259) Our dose-response data are
nearly linear at low doses (equation 1). As a result, the predictions of ES and IA
approximate those of GCA. One would not expect the dramatic “Something from
Nothing” effect that can occur with mixtures that have sub-linear dose-response
curves.(259)
When we extrapolate these models to higher dose levels of the phthalate mixture
(where the dose-response curve departs from linearity) we see considerable divergence
among the models (Figure 4.5). We also see this pattern in the effect summation
prediction of the mixture of rosiglitazone and nTZDpa (Figure 4.2a), where effect
summation overestimates the effect of the mixture at high concentrations.
GCA requires that the user specify a dose-response function that can be inverted
yielding real numbers for effect levels of interest. Based on the biological mechanism of
PPARγ, where ligands bind the receptor at a single site, we used a Hill function with a
135
Hill coefficient of 1.(254,255) This function meets the invertibility requirement. Some
scenarios may warrant fitting dose-response data with different dose-response functions.
For example, ligands for receptors that homodimerize (e.g., estrogen receptor) would be
expected to be non-linear at low dose. We would not expect such data to be adequately fit
by a Hill function with a Hill coefficient of 1; on the other hand, use of a Hill function of
2 would violate the invertibility requirement needed for applying GCA to partial agonists.
We are currently working on appropriate biologically-based models for these types of
systems. GCA may not be applicable to all mixtures of full and partial agonists.
Accordingly, careful consideration should be placed on properly specifying the individual
dose-response functions.
As an alternative to GCA, Scholze et al. (2014) introduced the Toxic Unit
Extrapolation model for predicting the joint effect of a mixture that includes partial
agonists, and applied it to ligands of the estrogen receptor.(276) However, toxic unit
extrapolation does not account for the antagonistic effect of partial agonists at high doses;
this effect may in principle occur even when there are two binding sites as in a
homodimer. We attempted to limit our biological mechanism exclusively to PPARγspecific binding via our selection of candidate compounds. Future work will address
whether concentration additive models are appropriate for cases involving agonist
mixtures that activate separate components of permissive heterodimers, or whether it is
better fit by an IA model.
Although TEF overestimates the mixture response, the model does have the
advantage of easy interpretability and use, as each component is treated as a dilution of a
136
well-characterized reference compound (for PPARγ, we used rosiglitazone). For
regulatory purposes, TEF would provide conservative estimates that may be of use to risk
assessors, where applicable. On the other hand, the model prediction is dependent on the
choice of reference compound and its efficacy; the model overestimation will increase
with increasing efficacy of the reference compound. Given the variety of sources,
structures, and pharmacodynamics (e.g. as shown by Tan et al., 2012)(277) of PPARγ
agonists, the choice of a ‘correct’ reference compound for PPARγ agonism is not clear.
Realistic exposures to environmental endocrine disrupting compounds tend to be
at low concentrations, where GCA, TEF, IA and ES produced similar results for PPARγ
ligands. However, GCA produced estimates quite different from the other models when
the mixture concentration was extrapolated to high doses. For PPARγ ligands, GCA may
be particularly useful in determining effects of interactions between environmental
exposures and therapeutic treatments. PPARγ already serves as a relevant model for this
type of mixture, given its well-established role as a target for insulin-sensitizing drugs.
For example, Cmax estimates of rosiglitazone in human blood during therapeutic use range
from 200nM to 1.2 uM,(278) concentrations that fall within the drug’s maximal efficacy
for PPARγ activation. When co-exposure to low efficacy environmental PPARγ ligands
occurs, the efficacy of the administered therapeutic may be reduced. Given the breadth of
PPARγ’s biologic influence and the increasing number of identified ligands in the built
environment, future studies should characterize mixture effects on relevant endpoints
under control of PPARγ activation, such as insulin sensitivity and lipid accumulation, to
better inform the biology behind predictive models.
Tables
Table 4.1. Modeled parameters for phthalate compounds from dose-response fits using 3-parameter Hill equation (with
Hill coefficient of 1) (Figure 4.4a and 4.4b).
EC50 (M)
(95% CI)
DEHP
N.D. *
MEHP
2.2 x 10-5
(9.8 x 10-6,
4.7 x 10-5)
1.3
(1.0, 1.6)
BBP
1.7 x 10-5
(7.9 x 10-6,
3.6 x 10-5)
1.0
(0.89, 1.2)
MBP
5.1 x 10-5
(1.5 x 10-5,
1.7 x 10-4)
0.61
(0.42, 0.79)
MBzP
1.6 x 10-4
(8.3 x 10-5,
3.0 x 10-4)
2.5
(1.9, 3.1)
Span (max–min)
(95% CI)
N.D.
Efficacy relative to
Rosiglitazone maximum (2.3fold change (95% CI: 2.0, 2.6))
R2
N.D.
0.56
0.43
0.26
1.1
N.D.
0.87
0.88
0.71
0.90
DEHP: di-ethyl hexyl phthalate; MEHP: mono-ethyl hexyl phthalate; BBzP: butyl benzyl phthalate; MBP: mono-butyl
phthalate; MBzP: mono-benzyl phthalate. * N.D. (not detected): Because of the minimal efficacy and poor statistical fit, DEHP
was considered to be inactive for the purpose of the mixture experiments. 95% CI: 95% Confidence Interval associated with fit
estimate. Span: difference in estimated curve maximum and curve minimum (see Equation 1).
137
Table 4.2. Phthalate components of mixture rays. Ray A: phthalates mixed at initial concentrations below calculated
EC10 from individual dose-response curves. Ray B: phthalates mixed at initial concentrations producing the equivalent
effect as rosiglitazone EC10 from control dose-response curve (Figures 4.4a and 4.4b).
Ray A
Ray B
Concentration
(M) in stock
mixture
Effect
level % max
Concentration
(M) in stock
mixture
Effect
level % max
MEHP
BBzP
MBP
MBzP
2.2 x 10-6
1.7 x 10-6
4.3 x 10-6
1.6 x 10-5
9.3
9.2
7.7
9.5
3.8 x 10-6
3.7 x 10-6
1.5 x 10-5
1.6 x 10-5
15
18
23
9.4
Stock mixture concentration
(total)
2.4 x 10-5 M
3.9 x 10-5 M
Note: DEHP excluded from rays (see footnote in table 1)
138
139
Figures
FIGURE 4.1. The maximal response of rosiglitazone is decreased by high doses of
nTZDpa.
140
Figure 4.1. The maximal response to rosiglitazone is decreased by high
concentrations of nTZDpa. a. Individual (marginal) dose-response curves of Rosiglitazone
(full agonist) and nTZDpa (partial agonist). Cos-7 cells were transfected with peroxisome
proliferator response element-dependent 3x luciferase reporter as described in Methods and
dosed with Vh (dimethyl sulfoxide) or the indicated concentrations of agonist. Green
fluorescent protein fluorescence-normalized luminescence was normalized to that of
untreated culture wells. EC50 rosiglitazone: 2.2 x 10-8 M; EC50 nTZDpa: 7.1 x 10-8 M. n = 4
separate transfections. b. Dose-response curves of rosiglitazone in combination with
increasing partial agonist concentration (error bars: SEM), (n = 4).
141
FIGURE 4.2. GCA provides an accurate prediction of full and partial PPARγ
agonist mixture effects.
142
Figure 4.2. GCA provides an accurate prediction of full and partial PPARγ agonist
mixture effects. a.) Experimental joint dose-response surface of Rosiglitazone and nTZDpa
dose combinations. Vertical axis: fold change in PPARγ activation over untreated cultures
(linear). Marginal curves correspond to individual dose-response curves shown in Figure 4.1.
b.) Effect Summation (ES), c.) Independent Action (IA), d.) Toxic Equivalency Factor (TEF),
and e.) GCA model fits based on parameters (with Hill coefficient constrained at 1.0) listed in
embedded table and formulations listed in Methods.
FIGURE 4.3. Phthalate esters activate PPARγ with a variety of efficacies.
143
144
Figure 4.3. Phthalate esters activate PPARγ with varying efficacies. Cos-7 cells were
transfected with peroxisome proliferator response element-dependent 3x luciferase reporter
as described in Methods and dosed with Vh (DMSO) or the indicated concentrations of
agonist. GFP fluorescence-normalized luminescence was normalized to that of untreated
culture wells. Positive control data (rosiglitazone) included for comparison. n = 4 separate
transfections. a.) Dose-response curves of parent phthalates DEHP and BBP. b.) Doseresponse curves of metabolite phthalate compounds MEHP, MBP, and MBzP. c.)
Rosiglitazone dose-response curves with increasing concentrations of MEHP. Error bars
omitted for clarity. All curves fit with a 3-parameter Hill equation (with Hill coefficient
constrained at 1). Dose response parameters for phthalate components are listed in Table 1.
d.) Experimental dose-response surface e.) TEF-predicted surface f.) GCA-predicted surface
of rosiglitazone/MEHP mixture. Interior values predicted using parameters derived from 3a.
(Rosiglitazone EC50: 2.3 x 10-8 M) and Table 1. Rosiglitazone dose-response curves by
increasing MEHP are labeled i. – iv. to facilitate comparison to 4.3c.
FIGURE 4.4. Comparison of experimentally determined PPARγ activation, using phthalate ray design, with model
predictions.
145
146
Figure 4.4. Comparison of experimentally determined PPARγ activation (fold
change over untreated cells), using phthalate ray design, with model predictions. a.)
Phthalates combined at starting concentrations below individual EC10. b.) Phthalates
combined at starting concentrations producing effect levels equal to 10% of
rosiglitazone’s maximal effect. c.) Comparison of rays and models at highest
concentrations. TEF: Toxic Equivalency Factor model; ES: Effect Summation model; IA:
Independent Action model; GCA: Generalized Concentration Addition model. Ray
design specifications listed in table 2. Experimental data n = 4 separate transfections,
mean +/- SEM. Baseline: Experimental curve fit estimated minimum (experimental curve
fit omitted for clarity).
147
FIGURE 4.5. Extrapolations of modeled dose-responses to high concentrations for
ray A.
148
Figure 4.5. Extrapolations of modeled dose-responses to high concentrations for ray
A. Vertical line: maximum mixture concentration tested (figure 4.4a).
149
Supplemental material
Supplementary Figure S4.1
150
Supplementary Figure S4.1. GCA provides an accurate prediction of full (rosiglitazone)
and partial (nTZDpa) PPARγ agonist mixture effects. a.) 3D surface plots of experimental
data depicting mean values (left column), mean plus SEM (center column) and mean
minus SEM (right column) values (n = 4 separate experiments). GCA (b.), TEF (c.), IA
(d.) and ES (e.) surface plots were constructed by inputting the mean, mean-plus-SEM or
mean-minus-SEM values of marginal curves (GCA, TEF) or response (ES, IA) to models
as described in methods. SEE: standard error of estimate (model distribution versus
corresponding experimental distribution) defined as
𝑆𝐸𝐸 = √
∑(𝑒𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡𝑎𝑙 𝑣𝑎𝑙𝑢𝑒 − 𝑚𝑜𝑑𝑒𝑙𝑒𝑑 𝑣𝑎𝑙𝑢𝑒)2
𝑁
Where N = number of paired values = 36. GCA consistently provides the best fit (i.e.
lowest SEE) and appears the least sensitive to changes in input values.
151
CHAPTER V: Conclusions
This research was designed to determine the mechanisms of toxicant-induced
bone outcomes specifically as they relate to coordinate activation of nuclear receptors in
bone cells. We used murine in vitro and in vivo models to experimentally identify
important changes in gene expression and phenotype resulting from exposure to
environmental toxicants that bind and activate the peroxisome proliferator activated
receptor gamma (PPARγ). The first study characterized an important negative correlation
between adipogenesis and osteogenesis in developing bone cells treated with
environmental PPARγ ligands, and found that the dual PPARγ/RXRα agonist organotins
disproportionately suppress osteogenesis (Chapter 2). In the second study, we exposed
adult, female C57Bl/6 mice to tributyltin (TBT) and observed a unique phenotype
characterized by a suppression of both cortical bone apposition and trabecular bone
resorption, and determined that a perturbation of normal osteoclast differentiation and
function contributed to an increased osteoblast stimulus where bone turnover is prevalent
(Chapter 3). Last, we contributed to the field of toxicological mixtures by establishing the
utility of Generalized Concentration Addition in modeling the additive effects of
environmental PPARγ agonists (demonstrated with phthalate parent and metabolite
compounds) (Chapter 4).
152
Chapter 2: Structurally-diverse, PPARγ-activating environmental toxicants induce
adipogenesis and suppress osteogenesis in bone marrow mesenchymal stromal cells
This study used in vitro murine, bone marrow mesenchymal stromal (BM-MSC)
cell cultures to characterize the changes in gene expression and bone and fat formation
that are indicative of an adipogenic skewing of BM-MSC differentiation in response to
PPARγ activation. Under osteogenic conditions, primary BM-MSCs were exposed to a
set of diverse environmental toxicants already shown to activate PPARγ. With varying
efficacies and potencies, these compounds suppressed expression of osteogenic genes
Runx2, Sp7, and Dmp1 concomitant with increases in adipogenic genes Fabp4 and Plin1.
These changes were accompanied by proportional suppression of alkaline phosphatase
activity, bone nodule formation, and mineralization in parallel with increased lipid
accumulation. Together, these experiments reinforced the current paradigm asserting that
there is a trade-off between adipogenesis and osteogenesis in differentiating BM-MSCs,
and that PPARγ activation can instigate this trade-off. However, the organotin
compounds (TBT and triphenyltin) suppressed bone formation with a distinct efficacy.
We speculated that this difference was, in part, attributable to the dual PPARγ:RXRα
agonist nature of these compounds. These studies highlighted an important mechanism by
which PPARγ activation, particularly by environmental agents, can regulate bone quality.
The primary limitation to this study was the in vitro methodology. Although this
setup allowed for precise determination of gene transcription and differentiation markers
in osteoblasts, the bone marrow is a complex environment consisting of numerous cell
types that intercommunicate. As such, these influences were not taken into account.
153
Chapter 3 addressed this limitation by using an in vivo model, however, the conclusions
from Chapter 2 – suppression of osteoblast differentiation and bone formation – had led
us to hypothesize that overall bone quality would be decreased with in vivo exposure to
TBT, especially given that TBT appears to be disproportionately efficacious in this
regard.
Although we characterized novel toxic effects for several compounds in this study
(e.g. TBBPA and METBP), this in vitro design limitation points to the fact that in
determining a critical effect of a toxicant, an in vitro model is not sufficient on its own
(though identifying new critical effects was not our stated intent). PPARγ activation does
effectively suppress osteoblastogenesis in favor of adipogenesis to decrease bone quality,
and this is supported by in vivo data using rosiglitazone. However, we surmised that
additional nuclear receptor activity, namely RXRα activation, can modify this effect.
Thus, these outcomes may be affected by even the most proximal events of differential
nuclear receptor activation. Building on these data, more precise mechanisms involving
the RXR will be investigated in primary osteoblast BM-MSC cultures to determine if the
dual ligand nature of organotins is indeed consequential. In particular, three questions
may be asked: does activation of an RXR homodimer alone by TBT sufficiently suppress
osteoblast differentiation? Is there a simultaneous, dual binding of PPARγ and RXRα by
TBT that functions additively or synergistically? Or is this an effect of RXRα dimerizing
with another nuclear receptor (e.g. LXR) to activate an independent pathway that also
suppresses osteoblast differentiation? We have generated preliminary (Chapter 3
supplemental) and published(80) data to establish the activation of additional pathways by
154
TBT, though their relative contributions to TBT’s disproportionate efficacy in
suppressing osteoblast differentiation specifically has not been determined.
Chapter 3: Tributyltin induces distinct effects on cortical and trabecular bone in female
C57Bl/6J mice resulting from modulation of both osteoblasts and osteoclasts
Here, 12-week old, female, wild-type C57Bl/6 mice were exposed to 10 mg/kg
TBT (or a sesame oil vehicle) for 10 weeks via oral gavage. Long bone samples were
collected for micro architectural structure, histological analysis, and whole bone gene
expression. Micro-computed tomography (micro-CT) analysis identified that the middiaphysis of the treated animals was smaller in cross-sectional area, though the medullary
space decreased in size with the cortical thickness; the percentage of bone area to total
area was maintained. The trabecular compartment at the distal metaphysis was more
heavily mineralized in TBT-treated animals compared to controls, with an increase in
bone density, trabecular number, and thickness, and a decrease in inter-trabecular
spacing. Histological analyses showed that trabecular osteoclast coverage was not
affected by treatment with TBT, suggesting that the function of the osteoclasts, and not
prevalence, had been compromised to decrease bone resorption. Whole bone mRNA
analyses showed significant decreases in osteoclast differentiation and functional markers
Nfatc1, Trap, and Ctsk. Additionally, there was a significant increase in expression of Ct1, which codes for a pro-osteoblast cytokine, cardiotrophin-1. These results were
interpreted as a TBT-induced shift in the osteoclast population to one that is functionally
deficient in resorption but still capable of stimulating osteoblastogenesis in an area of
155
high bone turnover (the trabeculae), as is seen in some models of osteopetrosis. In vitro
primary osteoclast cultures supported this interpretation, recapitulating the pattern of
gene expression with TBT treatment seen in whole bone mRNA. Furthermore, TBT
treatment did not reduce the number of TRAP-positive, multinucleated cells cultured with
RANKL and MCSF compared to vehicle (DMSO) treatment. However, TBT does still
seem capable of suppressing osteoblast activity in an area of low bone turnover, as
evidenced by the decreased bone apposition at the midshaft.
Because the micro-CT data is representative of a single time point, we are forced
to only hypothesize that bone apposition is not occurring at the periosteal surface in
treated mice (as opposed to bone loss). In comparison to untreated, 3 month-old
animals,(153) there was a modest increase in total cross-sectional area in the TBT-treated
mice after 10 weeks of treatment, though it was smaller than the increase observed in
vehicle-treated mice. This implies a stunted outward growth of cortical bone in TBTtreated mice. The use of dynamic histomorphometry to obtain a time-lapse of bone
deposition over the course of a 10-week treatment regime will definitively establish
whether periosteal bone apposition – and thus suppression of osteoblast activity – is in
fact an endpoint of TBT exposure. This would corroborate findings such as those of
Chapter 2, as well as help parse the apparent paradoxical observation of bone suppression
in one compartment (cortical shell) and bone deposition in another (trabecular
compartment). We hypothesize that in the absence of osteoclasts (such as along the
periosteum), the predominant effect of TBT is suppression of osteoblast differentiation;
in the presence of osteoclasts, indirect osteogenic stimulation predominates.
156
An important finding of this study was that TBT activates multiple nuclear
receptor pathways in osteoclasts. The in vitro data confirmed that TBT can suppress
normal osteoclast function, however, the correlation in these results with other nuclear
receptor-specific ligands (e.g. LG100268 (RXRα) or T0901317 (LXRα and LXRβ) is
circumstantial without the use of antagonist or knockout models. Antagonism of the
LXR, for instance, may remove some of this inhibitory effect, as others have shown that
LXR activation plays a suppressive role in osteoclast differentiation.(13,235,279) RXRα
activation has been shown to modify osteoclast differentiation as well.(13) RXR
heterodimer-specific ligands have been developed, as have LXRα and LXRβ knockout
mice from which bone marrow can be cultured. Future studies will take advantage of
these resources to more precisely determine the contribution of RXR heterodimers to
osteoclast function.
The use of ex vivo osteoclast cultures from TBT-treated mice may also be
informative, as they would provide an assessment of the potential for TBT-exposed preosteoclasts to differentiate under normal conditions. A recent study showed that RXR
homodimer activation plays a role in the early differentiation of osteoclasts (i.e. a proper
response to MCSF), whereas RXR:LXR activation is influential in osteoclast response to
RANKL.(13) The present in vitro studies only applied nuclear receptor ligands to
osteoclasts after an initial RANKL exposure. Thus, the mechanisms of action of these
agonists as they relate to events upstream of RANK activation and Nfatc1 expression are
not fully investigated. These experiments would also help identify whether the
hypothesized shift of osteoclasts to a subtype that sustains osteoblast signaling despite
157
losing resorptive capacity is established early in osteoclast differentiation or later, after
the osteoclast has matured.
In spite of these limitations, this study provided further evidence that RXR is an
important factor in osteoclast-mediated bone homeostasis, demonstrated for the first time
(to our knowledge) a pro-osteogenic effect of an RXR or (indirect) LXR ligand in an
intact mouse model, and suggested the possibility that obesogenic compounds may have
wider ranging influences in perturbing normal bone turnover beyond PPARγ-specific
effects. As the list of PPARγ-activating chemicals grows, attention should be focused on
the permissive aspect of the PPARγ:RXRα heterodimer so as to not preclude the
activation of alternate pathways. For instance, the use of PPARγ:RXRα heterodimers in
in vitro screening assays has been criticized for the potential for non-specific, off-target
effects.(71) However, this study demonstrates that such results may in fact be informative
and help expand a given chemical’s known mechanism of action, as shown here.
Chapter 4: Generalized Concentration Addition modeling predicts mixture effects of
environmental PPARγ agonists
Generalized Concentration Addition (GCA) is an additive model that was
developed to address a limitation inherent to other concentration additive models (such as
the toxic equivalency factor model, TEF): the joint effects of two (or more) chemicals
cannot be accurately predicted at a level greater than the efficacy of the least efficacious
component. Due to receptor-ligand pharmacodynamics, the efficacies of mixture
components tend to differ, making GCA a useful model for predicting high mixture effect
158
levels. Though GCA’s application to PPARγ was plausible in theory, it had not been
demonstrated that this modeling strategy could predict the combined effects of a mixture
of PPARγ agonists. In these experiments, dose-response curves of individual PPARγ
ligands were generated using a luciferase reporter gene under control of a DR-1 response
element in Cos-7 cells. A binary (rosiglitazone and the partial agonist nTZDpa) and a
complex mixture (phthalate parent compounds and metabolites identified as PPARγspecific agonists in ToxCast) of PPARγ-activating chemicals were each applied to the
luciferase assay. Using parameters estimated from the individual component doseresponse curves, mixture effects were predicted under different additive models and
compared to the empirical mixtures data. Generalized Concentration Addition was able to
accurately model the empirical data despite the variety of efficacies among the mixture
components, whereas TEF could not, and the effect summation and independent action
models were only accurate at low doses where the individual curves were approximately
linear. Because an increasing number of environmental PPARγ agonists are being
identified, thus increasing the likelihood of PPARγ ligand mixture exposures, this study
provides an important basis for determining the most appropriate model to use. These
results support the use of Generalized Concentration Addition in modeling the effect of a
mixture of PPARγ-specific ligands, provided certain assumptions of PPARγ-specific
activity are true.
The design of the complex mixture studies relied on the assumption that the
compounds only bound and activated PPARγ, and not RXRα additionally. Under this
assumption, the mixture should show attenuated effects at levels where a partial agonist
159
would antagonize the receptor:full agonist complex.(17) We were only able to obtain data
to observe this effect at the highest concentration of our ray design, as each component
was added to the mixture at sub-maximal concentrations (at or less than the individual
EC10). However, a binary mixture with rosiglitazone and one phthalate (MEHP) did
show the competitive antagonism effect expected under our assumption of one binding
site (i.e. PPARγ specificity), and the selection of compounds was in part based on criteria
that RXRα-activation assays were null.
However, it is known that TBT can activate both members of the PPARγ:RXRα
heterodimer.(63) In this scenario, a second binding site is presented, which may
circumvent the competitive antagonism effect of a second, PPARγ-specific partial
agonist, such as nTZDpa. Preliminary data (not shown) have suggested that this is the
case, as the reduction in response element activation is not observed at high combination
doses of TBT and nTZDpa as it is with a rosiglitazone/nTZDpa mixture. The
permissiveness of the PPARγ:RXRα heterodimer allows for activation by either agonist:
the low-efficacy PPARγ agonist should not therefore act as an antagonist to TBT on
RXRα. Because TBT can also bind and activate an RXR homodimer, the observed
response could also be due to a combination of PPARγ:RXRα heterodimer and
RXRα:RXRα homodimer activity. In either case, we hypothesize that such situations
would not be GCA. Use of heterodimer-specific RXR agonists and antagonists, as in
Chapter 3’s future studies, will again provide important insights into the behavior of the
PPARγ:RXRα heterodimer and potentially other receptor pairings and their activation by
chemical mixtures.
160
Public health implications
A well-maintained balance of bone formation and resorption is essential for long
term bone health. Our results with TBT suggested that the chemical induces an
osteopetrosis-like phenotype in the trabecular compartment. Osteopetrosis as a disorder
encompasses a number of different pathologies that are generally characterized by failure
of osteoclast function.(243) These disorders are inherited,(88) and exogenous factors that
significantly contribute to the onset of osteopetrosis are not established in the literature. It
may be that TBT’s mechanism of action to suppress osteoclast function represents a
potential risk for decreased bone resorption. As this balance is upset in the high-turnover
trabecular space, there are potential complications for proper calcium and phosphate
homeostasis. For instance, women may lose up to 10% of bone mineral density during
pregnancy and nursing to provide calcium for the developing fetal and infant
skeleton.(170,280) Exposure to TBT or a similar-acting compound may severely attenuate
the mobilization of minerals during this important window of susceptibility.
Because TBT appears to increase trabecular density, a diagnosis of osteoporosis
in exposed individuals would not seem likely. However, the effect of TBT to thin the
bone, via decreasing the cross-sectional area at the mid-shaft, represents a potentially
serious pathology. Thinner bones are more susceptible to fracture,(140) and bone fractures
have substantial associations with mortality, particularly in the elderly population. For
example, in some populations the excess age-adjusted mortality rate in the first year after
a hip fracture has been found to be as high as 20% in women, and higher in men.(281)
Women who experience a hip fracture may have up to five times the odds of mortality
161
within the first year following the fracture.(282) The risk of fracture is significantly
increased with a prior fracture,(283) even if asymptomatic.(281) The mortality rates decline
by year following the fracture, eventually returning to background rates,(282) though
quality of life is unquestionably impacted.(284,285) It should be noted that a number of
these studies showing the most drastic effects have focused exclusively on white women,
thereby limiting the generalizability to other populations. However, the world population
as a whole is ageing and the incidence of fractures is accordingly expected to
increase.(286)
The exact contribution of environmental exposures to the incidence of fractures
on a population level is, at this point, not quantifiable, although some studies are
beginning to identify important biological covariates in predicting fracture risk.(287) Even
if epidemiologic evidence is unattainable in the short-term, mechanistic investigations of
bone modifying agents are valuable. Importantly, a diagnosis of low bone density is
usually not given until after a traumatic event. Osteoporosis has been termed a “silent”
disease, in that it tends to not be recognized until a fracture occurs, and its prevalence in
the United States alone is expected to exceed 14 million individuals by 2020, when it is
estimated that one in two individuals will be at risk.(227) Hence, studying the etiology of
low bone density and decreased bone apposition is important in identifying preventative
measures. In light of the mechanisms related to environmental PPARγ/RXRα agonist
exposures, it is reasonable to speculate that any exogenous modifiers of bone that either
predispose individuals to fractures later in life (i.e. by suppressing bone apposition during
development and reducing peak bone mass) or diminish proper bone maintenance and
162
turnover in mature bone will have a realistic influence on fracture risk.
Environmental levels of TBT have decreased in the past decade thanks to
regulatory efforts.(288) With environmental exposures and biomarker levels generally
decreasing, tributyltin is becoming less relevant as an environmental toxicant to
humans.(213,256,289,290) However, these studies show a potential utility of TBT to serve as a
model ligand to probe the mechanisms of the PPARγ:RXRα heterodimer. The use of a
toxicant to study mechanisms of disease is not a novel concept (for example, aryl
hydrocarbon receptor-activating compounds are frequently used to identify important
pathways in cancer).(291,292) Using appropriate knockout mouse models in conjunction
with TBT as a multi-functional agonist may help in elucidating the mechanisms behind
RXRα-, LXR-, or PPARγ-mediated mechanisms in bone, ultimately contributing to the
development of therapeutics.
Although TBT’s relevance as an environmental threat to humans may be waning
due to successful regulatory measures, PPARγ has gained attention recently as a target of
an increasing number of environmental obesogen compounds. Recent work has
investigated PPARγ’s role in changing the adipogenic ‘setpoint’ during development,
predisposing individuals to obesity later in life when exposed perinatally.(60,61) Our
studies confirm its influence in suppression of osteoblastogenesis and increases in bone
marrow fat content, which also may be considered a targeted obesity outcome that can
have lifelong effects from early life exposures. Accordingly, the mechanistic
consequences of PPARγ and/or RXRα activation in bone disease should be taken into
consideration in developing toxicological screens for environmental ligands. These data
163
can further inform mixture analyses that rely on experimentally based and precise
mechanistic information to increase their predictive ability. PPARγ and its ligands sit at a
union of increasing environmental prevalence and toxicological relevance, and thus
provide a robust example of the important role of nuclear receptors in environmental
health.
164
BIBLIOGRAPHY
1.
McEwan IJ. Sex, drugs and gene expression: signalling by members of the nuclear
receptor superfamily. Essays in Biochemistry. 2004;40:1–10.
2.
Aranda a, Pascual a. Nuclear hormone receptors and gene expression.
Physiological Reviews. 2001 Jul;81(3):1269–1304.
3.
Huang P, Chandra V, Rastinejad F. Structural overview of the nuclear receptor
superfamily: insights into physiology and therapeutics. Annual Review of
Physiology. 2010 Dec;72(6):247–272.
4.
Mangelsdorf DJ, Evans RM. The RXR heterodimers and orphan receptors. Cell.
1995 Dec 15;83(6):841–850.
5.
Mangelsdorf DJ, Ong ES, Dyck J a, Evans RM. Nuclear receptor that identifies a
novel retinoic acid response pathway. Nature. 1990;345(6272):224–229.
6.
Lefebvre P, Benomar Y, Staels B. Retinoid X receptors: Common
heterodimerization partners with distinct functions. Trends in Endocrinology and
Metabolism. 2010;21(11):676–683.
7.
Kliewer S a, Umesono K, Mangelsdorf DJ, Evans RM. Retinoid X receptor
interacts with nuclear receptors in retinoic acid, thyroid hormone and vitamin D3
signalling. Nature. 1992;355(6359):446–449.
8.
Willy PJ, Umesono K, Ong ES, Evans RM, Heyman R a., Mangelsdorf DJ. LXR, a
nuclear receptor that defines a distinct retinoid response pathway. Genes &
Development. 1995;9(9):1033–1045.
9.
Issemann I, Green S. Activation of a member of the steroid hormone receptor
superfamily by peroxisome proliferators. Nature. 1990 Oct 18;347(6294):645–650.
10.
Issemann I, Prince RA, Tugwood JD, Green S. The retinoid X receptor enhances
the function of the peroxisome proliferator activated receptor. Biochimie. 1993
Jan;75(3–4):251–256.
11.
Kliewer SA, Umesono K, Noonan DJ, Heyman RA, Evans RM. Convergence of 9cis retinoic acid and peroxisome proliferator signalling pathways through
heterodimer formation of their receptors. Nature. 1992 Aug 27;358(6389):771–
774.
12.
Rosen ED, Spiegelman BM. PPARgamma : a nuclear regulator of metabolism,
differentiation, and cell growth. Journal of Biological Chemistry. 2001 Oct
12;276(41):37731–37734.
13.
Menéndez-Gutiérrez MP, Rőszer T, Fuentes L, Núñez V, Escolano A, Redondo
JM, Clerck N De, Metzger D, Valledor AF, Ricote M. Retinoid X receptors
165
orchestrate osteoclast differentiation and postnatal bone remodeling. Journal of
Clinical Investigation. 2015;125(2):809–823.
14.
Robertson KM, Windahl SH, Hultenby K, Ohlsson C. Cholesterol-sensing
receptors, liver X receptor α and β, have novel and distinct roles in osteoclast
differentiation and activation. Journal of Bone and Mineral Research. 2006;
21(8):1276–1287.
15.
Lecka-Czernik B, Gubrij I, Moerman EJ, Kajkenova O, Lipschitz DA, Manolagas
SC, Jilka RL. Inhibition of Osf2/Cbfa1 expression and terminal osteoblast
differentiation by PPARgamma2. Journal of Cellular Biochemistry. 1999 Sep 1;
74(3):357–371.
16.
Akune T, Ohba S, Kamekura S, Yamaguchi M, Chung U, Kubota N, Terauchi Y,
Harada Y, Azuma Y, Nakamura K, Kadowaki T, Kawaguchi H. PPARgamma
insufficiency enhances osteogenesis through osteoblast formation from bone
marrow progenitors. Journal of Clinical Investigation. 2004 Mar;113(6):846–855.
17.
Kenakin T. Pharmacologic Analysis of Drug–Receptor Interaction. 3rd ed.
Lippincott-Raven; 1997.
18.
Rastinejad F, Huang P, Chandra V, Khorasanizadeh S. Understanding nuclear
receptor form and function using structural biology. Journal of Molecular
Endocrinology. 2013 Dec;51(3):T1–21.
19.
Bain DL, Heneghan AF, Connaghan-Jones KD, Miura MT. Nuclear Receptor
Structure: Implications for Function. Annual Review of Physiology. 2007;69:201–
220.
20.
IJpenberg a., Jeannin E, Wahli W, Desvergne B. Polarity and Specific Sequence
Requirements of Peroxisome Proliferator-activated Receptor (PPAR)/Retinoid X
Receptor Heterodimer Binding to DNA: A Functional Analysis of the Malic
Enzyme Gene PPAR Response Element. Journal of Biological Chemistry. 1997
Aug 8;272(32):20108–20117.
21.
Dreyer C, Krey G, Keller H, Givel F, Helftenbein G, Wahli W. Control of the
peroxisomal β-oxidation pathway by a novel family of nuclear hormone receptors.
Cell. 1992;68(5):879–887.
22.
Harris PK, Kletzien RF. Localization of a pioglitazone response element in the
adipocyte fatty acid-binding protein gene. Molecular Pharmacology. 1994;
45(3):439–445.
23.
Schoonjans K, Peinado-Onsurbe J, Lefebvre AM, Heyman RA, Briggs M, Deeb S,
Staels B, Auwerx J. PPARalpha and PPARgamma activators direct a distinct
tissue-specific transcriptional response via a PPRE in the lipoprotein lipase gene.
EMBO Joournal. 1996;15(19):5336–5348.
166
24.
Wurtz JM, Bourguet W, Renaud JP, Vivat V, Chambon P, Moras D, Gronemeyer
H. A canonical structure for the ligand-binding domain of nuclear receptors.
Nature Structural Biology. 1996;3(1):87–94.
25.
Zoete V, Grosdidier A, Michielin O. Peroxisome proliferator-activated receptor
structures: Ligand specificity, molecular switch and interactions with regulators.
Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids.
2007;1771(8):915–925.
26.
Tontonoz P, Graves R a, Budavari a I, Erdjument-Bromage H, Lui M, Hu E,
Tempst P, Spiegelman BM. Adipocyte-specific transcription factor ARF6 is a
heterodimeric complex of two nuclear hormone receptors, PPAR gamma and RXR
alpha. Nucleic Acids Research. 1994 Dec 25;22(25):5628–5634.
27.
Tontonoz P, Hu E, Graves RA, Budavari AI, Spiegelman BM. mPPAR gamma 2:
tissue-specific regulator of an adipocyte enhancer. Genes & Development. 1994
May 15;8(10):1224–1234.
28.
Graves RA, Tontonoz P, Spiegelman BM. Analysis of a tissue-specific enhancer:
ARF6 regulates adipogenic gene expression. Molecular and Cellular Biology.
1992;12(3):1202–1208.
29.
Tontonoz P, Hu E, Spiegelman BM. Stimulation of adipogenesis in fibroblasts by
PPAR gamma 2, a lipid-activated transcription factor. Cell. 1994 Dec 30;79(7):
1147–1156.
30.
Lehmann JM, Moore LB, Smith-Oliver TA, Wilkison WO, Willson TM, Kliewer
SA. An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome
proliferator-activated receptor gamma (PPAR gamma). Journal of Biological
Chemistry. 1995 Jun 2;270(22):12953–12956.
31.
Tontonoz P, Spiegelman BM. Fat and beyond: the diverse biology of
PPARgamma. Annual Review of Biochemistry. 2008 Jan;77:289–312.
32.
Zhu Y, Qi C, Korenberg JR, Chen XN, Noya D, Rao MS, Reddy JK. Structural
organization of mouse peroxisome proliferator-activated receptor gamma (mPPAR
gamma) gene: alternative promoter use and different splicing yield two mPPAR
gamma isoforms. Proceedings of the National Academy of Sciences of the United
States of America. 1995;92(17):7921–7925.
33.
Sears IB, MacGinnitie MA, Kovacs LG, Graves RA. Differentiation-dependent
expression of the brown adipocyte uncoupling protein gene: regulation by
peroxisome proliferator-activated receptor gamma. Molecular and Cellular
Biology. 1996;16(7):3410–3419.
34.
Werman A, Hollenberg A, Solanes G, Bjorbaek C, Vidal-Puig AJ, Flier JS.
Ligand-independent Activation Domain in the N Terminus of Peroxisome
Proliferator-activated Receptor γ (PPARγ): Differential activity of PPAR-1 and -2
167
isoforms and influence of insulin. Journal of Biological Chemistry. 1997;272(32):
20230–20235.
35.
Ren D, Collingwood TN, Rebar EJ, Wolffe AP, Camp HS. PPAR gamma
knockdown by engineered transcription factors : exogenous PPAR gamma 2 but
not PPAR gamma 1 reactivates adipogenesis. Genes & Development. 2002;
16(734):27–32.
36.
Hu E, Kim JB, Sarraf P, Spiegelman BM. Inhibition of adipogenesis through MAP
kinase-mediated phosphorylation of PPARgamma. Science. 1996 Dec 20;
274(5295):2100–2103.
37.
Choi JH, Banks AS, Estall JL, Kajimura S, Bostrom P, Laznik D, Ruas JL,
Chalmers MJ, Kamenecka TM, Bluher M, Griffin PR, Spiegelman BM. Antidiabetic drugs inhibit obesity-linked phosphorylation of PPARgamma by Cdk5.
Nature. 2010;466(7305):451–456.
38.
Fajas L, Fruchart JC, Auwerx J. Transcriptional control of adipogenesis. Current
Opinion in Cell Biology. 1998 Apr;10(2):165–173.
39.
Farmer SR. Transcriptional control of adipocyte formation. Cell Metabolism. 2006
Oct;4(4):263–273.
40.
Cao Z, Umek RM, McKnight SL. Regulated expression of three C/EBP isoforms
during adipose conversion of 3T3-L1 cells. Genes & Development. 1991;5(9):
1538–1552.
41.
Clarke SL, Robinson CE, Gimble JM. CAAT / Enhancer Binding Proteins Directly
Modulate Transcription from the Peroxisome Proliferator- Activated Receptor γ 2
Promoter. Biochemical and Biophysical Research Communications. 1997;240(1):
99–103.
42.
Zuo Y, Qiang L, Farmer SR. Activation of CCAAT/enhancer-binding protein
(C/EBP) α expression by C/EBPβ during adipogenesis requires a peroxisome
proliferator-activated receptor-γ-associated repression of HDAC1 at the C/ebpα
gene promoter. Journal of Biological Chemistry. 2006;281(12):7960–7967.
43.
Wu Z, Rosen ED, Brun R, Hauser S, Adelmant G, Troy AE, McKeon C,
Darlington GJ, Spiegelman BM. Cross-regulation of C/EBPα and PPARγ controls
the transcriptional pathway of adipogenesis and insulin sensitivity. Molecular Cell.
1999;3(2):151–158.
44.
Rosen ED, Hsu C, Wang X, Sakai S, Freeman MW, Gonzalez FJ, Spiegelman BM.
C/EBPa induces adipogenesis through PPARg: a unified pathway. Genes &
Development. 2002;16:22–26.
45.
Forman BM, Tontonoz P, Chen J, Brun RP, Spiegelman BM, Evans RM. 15Deoxy-delta 12, 14-prostaglandin J2 is a ligand for the adipocyte determination
168
factor PPAR gamma. Cell. 1995 Dec 1;83(5):803–812.
46.
Chandra V, Huang P, Hamuro Y, Raghuram S, Wang Y, Burris TP, Rastinejad F.
Structure of the intact PPAR-gamma-RXR- nuclear receptor complex on DNA.
Nature. 2008 Nov 20;456(7220):350–356.
47.
Nolte RT, Wisely GB, Westin S, Cobb JE, Lambert MH, Kurokawa R, Rosenfeld
MG, Willson TM, Glass CK, Milburn M V. Ligand binding and co-activator
assembly of the peroxisome proliferator-activated receptor-gamma. Nature. 1998;
395(6698):137–143.
48.
Zhang L, Ren X-M, Wan B, Guo L-H. Structure-dependent binding and activation
of perfluorinated compounds on human peroxisome proliferator-activated receptor
γ. Toxicology and Applied Pharmacology. 2014;279(3):275–283.
49.
Knouff C, Auwerx J. Peroxisome proliferator-activated receptor-gamma calls for
activation in moderation: lessons from genetics and pharmacology. Endocrine
Reviews. 2004 Dec;25(6):899–918.
50.
DiRenzo J, Söderstrom M, Kurokawa R, Ogliastro MH, Ricote M, Ingrey S,
Hörlein A, Rosenfeld MG, Glass CK. Peroxisome proliferator-activated receptors
and retinoic acid receptors differentially control the interactions of retinoid X
receptor heterodimers with ligands, coactivators, and corepressors. Molecular and
Cellular Biology. 1997 Apr;17(4):2166–2176.
51.
Feige JN, Gelman L, Rossi D, Zoete V, Métivier R, Tudor C, Anghel SI,
Grosdidier A, Lathion C, Engelborghs Y, Michielin O, Wahli W, Desvergne B.
The endocrine disruptor monoethyl-hexyl-phthalate is a selective peroxisome
proliferator-activated receptor gamma modulator that promotes adipogenesis.
Journal of Biological Chemistry. 2007 Jun 29;282(26):19152–19166.
52.
Gelman L, Feige JN, Desvergne B. Molecular basis of selective PPARgamma
modulation for the treatment of Type 2 diabetes. Biochimica et Biophysica Acta.
2007 Aug;1771(8):1094–1107.
53.
Feige JN, Gelman L, Tudor C, Engelborghs Y, Wahli W, Desvergne B.
Fluorescence imaging reveals the nuclear behavior of peroxisome proliferatoractivated receptor/retinoid X receptor heterodimers in the absence and presence of
ligand. Journal of Biological Chemistry. 2005;280(18):17880–90.
54.
Rastinejad F, Perlmann T, Evans RM, Sigler PB. Structural determinants of
nuclear receptor assembly on DNA direct repeats. Nature. 1995. p. 203–11.
55.
Heindel JJ. Endocrine disruptors and the obesity epidemic. Toxicological Science.
2003;76(2):247–249.
56.
Baillie-Hamilton PF, Phil BSD. Chemical Toxins : A Hypothesis to Explain the
Global Obesity Epidemic. Journal of Alternative and Complementary Medicine.
169
2002;8(2):185–192.
57.
Grün F, Blumberg B. Environmental obesogens: organotins and endocrine
disruption via nuclear receptor signaling. Endocrinology. 2006 Jun;147(6 Suppl):
S50–55.
58.
Kanayama T, Kobayashi N, Mamiya S, Nakanishi T, Nishikawa J. Organotin
compounds promote adipocyte differentiation as agonists of the peroxisome
proliferator-activated receptor gamma/retinoid X receptor pathway. Molecular
Pharmacology. 2005 Mar;67(3):766–774.
59.
Grün F, Watanabe H, Zamanian Z, Maeda L, Arima K, Cubacha R, Gardiner DM,
Kanno J, Iguchi T, Blumberg B. Endocrine-disrupting organotin compounds are
potent inducers of adipogenesis in vertebrates. Molecular Endocrinology. 2006
Sep;20(9):2141–2155.
60.
Kirchner S, Kieu T, Chow C, Casey S, Blumberg B. Prenatal exposure to the
environmental obesogen tributyltin predisposes multipotent stem cells to become
adipocytes. Molecular Endocrinology. 2010 Mar;24(3):526–539.
61.
Chamorro-García R, Sahu M, Abbey RJ, Laude J, Pham N, Blumberg B.
Transgenerational inheritance of increased fat depot size, stem cell
reprogramming, and hepatic steatosis elicited by prenatal exposure to the obesogen
tributyltin in mice. Environmental Health Perspectives. 2013;121(3):359–366.
62.
Nakanishi T, Nishikawa J, Hiromori Y, Yokoyama H, Koyanagi M, Takasuga S,
Ishizaki J, Watanabe M, Isa S, Utoguchi N, Itoh N, Kohno Y, Nishihara T, Tanaka
K. Trialkyltin compounds bind retinoid X receptor to alter human placental
endocrine functions. Molecular Endocrinology. 2005 Oct;19(10):2502–2516.
63.
le Maire A, Grimaldi M, Roecklin D, Dagnino S, Vivat-Hannah V, Balaguer P,
Bourguet W. Activation of RXR-PPAR heterodimers by organotin environmental
endocrine disruptors. EMBO Reports. 2009 Apr;10(4):367–373.
64.
Hurst CH, Waxman DJ. Activation of PPARalpha and PPARgamma by
environmental phthalate monoesters. Toxicological Science. 2003 Aug;74(2):297–
308.
65.
Bility MT, Thompson JT, McKee RH, David RM, Butala JH, Vanden Heuvel JP,
Peters JM. Activation of mouse and human peroxisome proliferator-activated
receptors (PPARs) by phthalate monoesters. Toxicological Science. 2004
Nov;82(1):170–182.
66.
Springer C, Dere E, Hall SJ, McDonnell E V, Roberts SC, Butt CM, Stapleton
HM, Watkins DJ, McClean MD, Webster TF, Schlezinger JJ, Boekelheide K.
Rodent thyroid, liver, and fetal testis toxicity of the monoester metabolite of bis(2-ethylhexyl) tetrabromophthalate (tbph), a novel brominated flame retardant
present in indoor dust. Environmental Health Perspectives. 2012 Dec;
170
120(12):1711–1719.
67.
Riu A, Grimaldi M, le Maire A, Bey G, Phillips K, Boulahtouf A, Perdu E, Zalko
D, Bourguet W, Balaguer P. Peroxisome proliferator-activated receptor γ is a
target for halogenated analogs of bisphenol A. Environmental Health Perspectives.
2011 Sep;119(9):1227–1232.
68.
Li X, Pham HT, Janesick AS, Blumberg B. Triflumizole is an obesogen in mice
that acts through peroxisome proliferator activated receptor gamma (PPARγ).
Environmental Health Perspectives. 2012 Dec;120(12):1720–1726.
69.
Dix DJ, Houck K a., Martin MT, Richard AM, Setzer RW, Kavlock RJ. The
toxcast program for prioritizing toxicity testing of environmental chemicals.
Toxicological Science. 2007;95(1):5–12.
70.
Reif DM, Martin MT, Tan SW, Houck K a, Judson RS, Richard AM, Knudsen TB,
Dix DJ, Kavlock RJ. Endocrine profiling and prioritization of environmental
chemicals using ToxCast data. Environmental Health Perspectives. 2010;
118:1714–1720.
71.
Janesick AS, Dimastrogiovanni G, Vanek L, Boulos C, Chamorro-García R, Tang
W, Blumberg B. On the utility of ToxCastTM and ToxPi as methods for identifying
new obesogens. Environmental Health Perspectives. 2016;(January). DOI:
10.1289/ehp.1510352
72.
Imai Y, Youn M-Y, Inoue K, Takada I, Kouzmenko A, Kato S. Nuclear Receptors
in Bone Physiology and Diseases. Physiological Reviews. 2013;93(2):481–523.
73.
Lecka-Czernik B, Moerman EJ, Grant DF, Lehmann JM, Manolagas SC, Jilka RL.
Divergent effects of selective peroxisome proliferator-activated receptor-gamma 2
ligands on adipocyte versus osteoblast differentiation. Endocrinology. 2002 Jun;
143(6):2376–2384.
74.
Rzonca SO, Suva LJ, Gaddy D, Montague DC, Lecka-Czernik B. Bone is a target
for the antidiabetic compound rosiglitazone. Endocrinology. 2004 Jan;145(1):401–
406.
75.
Lazarenko OP, Rzonca SO, Hogue WR, Swain FL, Suva LJ, Lecka-Czernik B.
Rosiglitazone induces decreases in bone mass and strength that are reminiscent of
aged bone. Endocrinology. 2007 Jun;148(6):2669–2680.
76.
Ali AA, Weinstein RS, Stewart SA, Parfitt AM, Manolagas SC, Jilka RL.
Rosiglitazone causes bone loss in mice by suppressing osteoblast differentiation
and bone formation. Endocrinology. 2005 Mar;146(3):1226–1235.
77.
Dormuth CR, Carney G, Carleton B. Thiazolidinediones and fractures in men and
women. Archives of Internal Medicine. 2009;169(15):1395–1402.
171
78.
Grey a, Bolland M, Gamble G, Wattie D, Horne a, Davidson J, Reid IR. The
Peroxisome Proliferator-Activated Receptor- Agonist Rosiglitazone Decreases
Bone Formation and Bone Mineral Density in Healthy Postmenopausal Women: A
Randomized, Controlled Trial. Journal of Clinical Endocrinology & Metabolism.
2007 Apr;92(4):1305–1310.
79.
Billington EO, Grey A, Bolland MJ. The effect of thiazolidinediones on bone
mineral density and bone turnover: systematic review and meta-analysis.
Diabetologia. 2015;58(10):2238–2246.
80.
Yanik SC, Baker AH, Mann KK, Schlezinger JJ. Organotins are potent activators
of PPARγ and adipocyte differentiation in bone marrow multipotent mesenchymal
stromal cells. Toxicological Science. 2011 Aug;122(2):476–488.
81.
Baker AH, Watt J, Huang CK, Gerstenfeld LC, Schlezinger JJ. Tributyltin
Engages Multiple Nuclear Receptor Pathways and Suppresses Osteogenesis in
Bone Marrow Multipotent Stromal Cells. Chemical Research in Toxicology. 2015;
150513143140005.
82.
Carfi’ M, Croera C, Ferrario D, Campi V, Bowe G, Pieters R, Gribaldo L. TBTC
induces adipocyte differentiation in human bone marrow long term culture.
Toxicology. 2008 Jul 10;249(1):11–18.
83.
Holz JD, Sheu T, Drissi H, Matsuzawa M, Zuscik MJ, Puzas JE. Environmental
agents affect skeletal growth and development. Birth Defects Research. Part C,
Embryo Today: Reviews. 2007 Mar;81(1):41–50.
84.
Seeman E, Delmas PD. Bone quality--the material and structural basis of bone
strength and fragility. New England Journal of Medicine. 2006;354(21):2250–
2261.
85.
Eriksen EF. Cellular mechanisms of bone remodeling. Reviews in Endocrine &
Metabolic Disorders. 2010;11(4):219–227.
86.
Carmeliet G, Dermauw V, Bouillon R. Vitamin D signaling in calcium and bone
homeostasis: A delicate balance. Best Practice & Research. Clinical
Endocrinology & Metabolism 2015;29(4):621–631.
87.
Coudert AE, De Vernejoul MC, Muraca M, Del Fattore A. Osteopetrosis and its
relevance for the discovery of new functions associated with the skeleton.
International Journal of Endocrinology. 2015; Article ID 372156.
88.
Stark Z, Savarirayan R. Osteopetrosis. Orphanet Journal of Rare Diseases.
2009;4(1):5.
89.
Raisz LG. Pathogenesis of osteoporosis: concepts, conflicts, and prospects. Journal
of Clinical Investigation. 2005 Dec;115(12):3318–3325.
90.
Pietschmann P, Rauner M, Sipos W, Kerschan-Schindl K. Osteoporosis: An age-
172
related and gender-specific disease - A mini-review. Gerontology. 2009;55(1):3–
12.
91.
Titorencu I, Pruna V, Jinga V V., Simionescu M. Osteoblast ontogeny and
implications for bone pathology: An overview. Cell and Tissue Research. 2014;
355(1):23–33.
92.
Long F. Building strong bones: molecular regulation of the osteoblast lineage.
Nature Reviews. Molecular Cell Biology. 2011;13(1):27–38.
93.
Gaur T, Lengner CJ, Hovhannisyan H, Bhat RA, Bodine PVN, Komm BS, Javed
A, van Wijnen AJ, Stein JL, Stein GS, Lian JB. Canonical WNT signaling
promotes osteogenesis by directly stimulating Runx2 gene expression. Journal of
Biological Chemistry. 2005 Sep 30;280(39):33132–33140.
94.
Lee KS, Kim HJ, Li QL, Chi XZ, Ueta C, Komori T, Wozney JM, Kim EG, Choi
JY, Ryoo HM, Bae SC. Runx2 is a common target of transforming growth factor
beta1 and bone morphogenetic protein 2, and cooperation between Runx2 and
Smad5 induces osteoblast-specific gene expression in the pluripotent mesenchymal
precursor cell line C2C12. Molecular and Cellular Biology. 2000;20(23):8783–
8792.
95.
Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G. Osf2/Cbfa1: a
transcriptional activator of osteoblast differentiation. Cell. 1997 May 30;
89(5):747–754.
96.
Jeon MJ, Kim JA, Kwon SH, Kim SW, Park KS, Park S-W, Kim SY, Shin CS.
Activation of peroxisome proliferator-activated receptor-gamma inhibits the
Runx2-mediated transcription of osteocalcin in osteoblasts. Journal of Biological
Chemistry. 2003 Jun 27;278(26):23270–23277.
97.
Kang S, Bennett CN, Gerin I, Rapp L a, Hankenson KD, Macdougald O a. Wnt
signaling stimulates osteoblastogenesis of mesenchymal precursors by suppressing
CCAAT/enhancer-binding protein alpha and peroxisome proliferator-activated
receptor gamma. Journal of Biological Chemistry. 2007 May 11;282(19):14515–
14524.
98.
Bennett CN, Ross SE, Longo K a, Bajnok L, Hemati N, Johnson KW, Harrison
SD, MacDougald O a. Regulation of Wnt signaling during adipogenesis. Journal of
Biological Chemistry. 2002 Aug 23;277(34):30998–31004.
99.
Nakashima K, Zhou X, Kunkel G, Zhang Z, Deng JM, Behringer RR, de
Crombrugghe B. The novel zinc finger-containing transcription factor osterix is
required for osteoblast differentiation and bone formation. Cell. 2002 Jan 11;
108(1):17–29.
100. Moldes M, Zuo Y, Morrison RF, Silva D, Park B, Liu J, Farmer SR. Peroxisomeproliferator-activated receptor gamma suppresses Wnt/beta-catenin signalling
173
during adipogenesis. Biochemical Journal. 2003 Dec 15;376(Pt 3):607–613.
101. Dallas SL, Bonewald LF. Dynamics of the transition from osteoblast to osteocyte.
Annals of the New York Academy of Sciences. 2010;1192:437–443.
102. Bonewald LF. The amazing osteocyte. Journal of Bone and Mineral Research.
2011;26(2):229–238.
103. Capulli M, Paone R, Rucci N. Osteoblast and osteocyte: Games without frontiers.
Archives of Biochemistry and Biophysics. 2014;561:3–12.
104. Boyce BF, Rosenberg E, de Papp AE, Duong LT. The osteoclast, bone
remodelling and treatment of metabolic bone disease. European Journal of Clinical
Investigation. 2012;42(12):1332–1341.
105. Rios HF, Ye L, Dusevich V, Eick D, Bonewald LF, Feng JQ. DMP1 is essential
for osteocyte formation and function. Journal of Musculoskeletal & Neuronal
Interactions. 2005;5(4):325–327.
106. Poole KES, van Bezooijen RL, Loveridge N, Hamersma H, Papapoulos SE, Löwik
CW, Reeve J. Sclerostin is a delayed secreted product of osteocytes that inhibits
bone formation. FASEB Journal 2005;19:1842–1844.
107. Oddie GW, Schenk G, Angel NZ, Walsh N, Guddat LW, de Jersey J, Cassady a. I,
Hamilton SE, Hume D a. Structure, function, and regulation of tartrate-resistant
acid phosphatase. Bone. 2000;27(5):575–584.
108. Yagi M, Miyamoto T, Sawatani Y, Iwamoto K, Hosogane N, Fujita N, Morita K,
Ninomiya K, Suzuki T, Miyamoto K, Oike Y, Takeya M, Toyama Y, Suda T. DCSTAMP is essential for cell-cell fusion in osteoclasts and foreign body giant cells.
Journal of Experimental Medicine. 2005;202(3):345–351.
109. Drake FH, Dodds R a, James IE, Connor JR, Debouck C, Richardson S, LeeRykaczewski E, Coleman L, Rieman D, Barthlow R, Hastings G, Gowen M.
Cathepsin K, but not cathepsin B, L, or S, is abundantly expressed in human
osteoclasts. Journal of Biological Chemistry. 1996;271(21):12511–12516.
110. Asagiri M, Takayanagi H. The molecular understanding of osteoclast
differentiation. Bone. 2007;40(2):251–264.
111. Miyamoto T. Regulators of osteoclast differentiation and cell-cell fusion. Keio
Journal of Medicine. 2011;60(4):101–105.
112. Boyle WJ, Simonet WS, Lacey DL. Osteoclast differentiation and activation.
Nature. 2003 May 15;423(6937):337–342.
113. Charles JF, Aliprantis AO. Osteoclasts: More than “bone eaters.” Trends in
Molecular Medicine; 2014;20(8):449–459.
174
114. Lecaille F, Brömme D, Lalmanach G. Biochemical properties and regulation of
cathepsin K activity. Biochimie. 2008;90(2):208–226.
115. Väänänen K. Mechanism of osteoclast mediated bone resorption - Rationale for
the design of new therapeutics. Advanced Drug Delivery Reviews. 2005;
57(7):959–971.
116. Lorget F, Kamel S, Mentaverri R, Wattel A, Naassila M, Maamer M, Brazier M.
High extracellular calcium concentrations directly stimulate osteoclast apoptosis.
Biochemical and Biophysical Research Communications. 2000;268(3):899–903.
117. Tang Y, Wu X, Lei W, Pang L, Wan C, Shi Z, Zhao L, Nagy TR, Peng X, Hu J,
Feng X, Van Hul W, Wan M, Cao X. TGF-beta1-induced migration of bone
mesenchymal stem cells couples bone resorption with formation. Nature Medicine.
2009;15(7):757–765.
118. Martin TJ, Sims NA. Osteoclast-derived activity in the coupling of bone formation
to resorption. Trends in Molecular Medicine. 2005;11(2):76–81.
119. Matsuzaki K, Udagawa N, Takahashi N, Yamaguchi K, Yasuda H, Shima N,
Morinaga T, Toyama Y, Yabe Y, Higashio K, Suda T. Osteoclast differentiation
factor (ODF) induces osteoclast-like cell formation in human peripheral blood
mononuclear cell cultures. Biochemical and Biophysical Research
Communications. 1998;246(1):199–204.
120. Yasuda H, Shima N, Nakagawa N, Yamaguchi K, Kinosaki M, Mochizuki S,
Tomoyasu a, Yano K, Goto M, Murakami a, Tsuda E, Morinaga T, Higashio K,
Udagawa N, Takahashi N, Suda T. Osteoclast differentiation factor is a ligand for
osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to
TRANCE/RANKL. Proceedings of the National Academy of Sciences of the
United States of America. 1998;95(7):3597–3602.
121. Udagawa N, Takahashi N, Jimi E, Matsuzaki K, Tsurukai T, Itoh K, Nakagawa N,
Yasuda H, Goto M, Tsuda E, Higashio K, Gillespie MT, Martin TJ, Suda T.
Osteoblasts/stromal cells stimulate osteoclast activation through expression of
osteoclast differentiation factor/RANKL but not macrophage colony-stimulating
factor. Bone. 1999;25(5):517–523.
122. Itoh K, Udagawa N, Matsuzaki K, Takami M, Amano H, Shinki T, Ueno Y,
Takahashi N, Suda T. Importance of membrane- or matrix-associated forms of MCSF and RANKL/ODF in osteoclastogenesis supported by SaOS-4/3 cells
expressing recombinant PTH/PTHrP receptors. Journal of Bone and Mineral
Research. 2000;15(9):1766–1775.
123. Kodama H, Nose M, Niida S, Yamasaki A. Essential role of macrophage colonystimulating factor in the osteoclast differentiation supported by stromal cells.
Journal of Experimental Medicine. 1991;173(5):1291–1294.
175
124. Takayanagi H, Kim S, Koga T, Nishina H, Isshiki M, Yoshida H, Saiura A, Isobe
M, Yokochi T, Inoue JI, Wagner EF, Mak TW, Kodama T, Taniguchi T. Induction
and activation of the transcription factor NFATc1 (NFAT2) integrate RANKL
signaling in terminal differentiation of osteoclasts. Developmental Cell.
2002;3(6):889–901.
125. Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T, Elliott R,
Colombero a., Elliott G, Scully S, Hsu H, Sullivan J, Hawkins N, Davy E,
Capparelli C, Eli a., Qian YX, Kaufman S, Sarosi I, Shalhoub V, Senaldi G, Guo
J, Delaney J, Boyle WJ. Osteoprotegerin ligand is a cytokine that regulates
osteoclast differentiation and activation. Cell. 1998;93(2):165–176.
126. Li X, Zhang Y, Kang H, Liu W, Liu P, Zhang J, Harris SE, Wu D. Sclerostin binds
to LRP5/6 and antagonizes canonical Wnt signaling. Journal of Biological
Chemistry. 2005;280(20):19883–19887.
127. Li X, Ominsky MS, Niu Q-T, Sun N, Daugherty B, D’Agostin D, Kurahara C, Gao
Y, Cao J, Gong J, Asuncion F, Barrero M, Warmington K, Dwyer D, Stolina M,
Morony S, Sarosi I, Kostenuik PJ, Lacey DL, Simonet WS, Ke HZ, Paszty C.
Targeted deletion of the sclerostin gene in mice results in increased bone formation
and bone strength. Journal of Bone & Mineral Research. 2008;23(6):860–869.
128. Balemans W, Ebeling M, Patel N, Van Hul E, Olson P, Dioszegi M, Lacza C,
Wuyts W, Van Den Ende J, Willems P, Paes-Alves a F, Hill S, Bueno M, Ramos
FJ, Tacconi P, Dikkers FG, Stratakis C, Lindpaintner K, Vickery B, Foernzler D,
Van Hul W. Increased bone density in sclerosteosis is due to the deficiency of a
novel secreted protein (SOST). Human Molecular Genetics. 2001;10(5):537–543.
129. Robling AG, Niziolek PJ, Baldridge LA, Condon KW, Allen MR, Alam I, Mantila
SM, Gluhak-Heinrich J, Bellido TM, Harris SE, Turner CH. Mechanical
stimulation of bone in vivo reduces osteocyte expression of Sost/sclerostin. Journal
of Biological Chemistry. 2008;283(9):5866–5875.
130. Keller H, Kneissel M. SOST is a target gene for PTH in bone. Bone. 2005;37(2):
148–158.
131. Kennedy OD, Laudier DM, Majeska RJ, Sun HB, Schaffler MB. Osteocyte
apoptosis is required for production of osteoclastogenic signals following bone
fatigue in vivo. Bone. 2014 Jul;64(7):132–137.
132. Cabahug-Zuckerman P, Frikha-Benayed D, Majeska RJ, Tuthill A, Yakar S, Judex
S, Schaffler MB. Osteocyte apoptosis caused by hindlimb unloading is required to
trigger osteocyte RANKL production and subsequent resorption of cortical and
trabecular bone in mice femurs. Journal of Bone & Mineral Research. 2016 Feb 6.
doi: 10.1002/jbmr.2807.
133. Delaisse J-M. The reversal phase of the bone-remodeling cycle: cellular
prerequisites for coupling resorption and formation. BoneKEy Reports. 2014;
176
3(April):561.
134. Karsdal MA, Neutzsky-Wulff A V., Dziegiel MH, Christiansen C, Henriksen K.
Osteoclasts secrete non-bone derived signals that induce bone formation.
Biochemical and Biophysical Research Communications. 2008;366(2):483–488.
135. Walker EC, McGregor NE, Poulton IJ, Pompolo S, Allan EH, Quinn JMW,
Gillespie MT, Martin TJ, Sims N a. Cardiotrophin-1 is an osteoclast-derived
stimulus of bone formation required for normal bone remodeling. Journal of Bone
& Mineral Research. 2008;23(12):2025–2032.
136. Sims N a., Vrahnas C. Regulation of cortical and trabecular bone mass by
communication between osteoblasts, osteocytes and osteoclasts. Archives of
Biochemistry and Biophysics. 2014;561:22–28.
137. Negishi-Koga T, Shinohara M, Komatsu N, Bito H, Kodama T, Friedel RH,
Takayanagi H. Suppression of bone formation by osteoclastic expression of
semaphorin 4D. Nature Medicine. 2011;17(11):1473–1480.
138. Ruff CB, Hayes WC. Subperiosteal expansion and cortical remodeling of the
human femur and tibia with aging. Science. 1982 Sep 3;217(4563):945–948.
139. Seeman E. Pathogenesis of bone fragility in women and men. Lancet. 2002;
359(9320):1841–1850.
140. Bouxsein ML, Karasik D. Bone geometry and skeletal fragility. Current
Osteoporosis Reports. 2006;4(2):49–56.
141. Seeman E. Clinical review 137: Sexual dimorphism in skeletal size, density, and
strength. Journal of Clinical Endocrinology and Metabolism. 2001 Oct;86(10):
4576–4584.
142. Duan Y, Turner CH, Kim BT, Seeman E. Sexual dimorphism in vertebral fragility
is more the result of gender differences in age-related bone gain than bone loss.
Journal of Bone & Mineral Research. 2001;16(12):2267–2275.
143. Laurent M, Antonio L, Sinnesael M, Dubois V, Gielen E, Classens F,
Vanderschueren D. Androgens and estrogens in skeletal sexual dimorphism. Asian
Journal of Andrology. 2014;16(2):213–222.
144. Callewaert F, Sinnesael M, Gielen E, Boonen S, Vanderschueren D. Skeletal
sexual dimorphism: relative contribution of sex steroids, GH-IGF1, and
mechanical loading. Journal of Endocrinology. 2010 Nov;207(2):127–134.
145. Lu PW, Cowell CT, LLoyd-Jones SA, Briody JN, Howman-Giles R. Volumetric
bone mineral density in normal subjects, aged 5–27 years. Journal of Clinical
Endocrinology and Metabolism. 1996 Apr;81(4):1586–1590.
177
146. Tabensky a, Duan Y, Edmonds J, Seeman E. The contribution of reduced peak
accrual of bone and age-related bone loss to osteoporosis at the spine and hip:
insights from the daughters of women with vertebral or hip fractures. Journal of
Bone & Mineral Research. 2001;16(6):1101–1107.
147. Aaron JE, Makins NB, Sagreiya K. The microanatomy of trabecular bone loss in
normal aging men and women. Clinical Orthopaedics and Related Research. 1987
Feb;(215):260–271.
148. Moerman EJ, Teng K, Lipschitz D a, Lecka-Czernik B. Aging activates adipogenic
and suppresses osteogenic programs in mesenchymal marrow stroma/stem cells:
the role of PPAR-gamma2 transcription factor and TGF-beta/BMP signaling
pathways. Aging Cell. 2004 Dec;3(6):379–389.
149. Moore SG, Dawson KL. Red and yellow marrow in the femur: age-related changes
in appearance at MR imaging. Radiology. 1990;175(1):219–223.
150. Rosen CJ, Bouxsein ML. Mechanisms of disease: is osteoporosis the obesity of
bone? Nature Clinical Practice. Rheumatology. 2006 Jan;2(1):35–43.
151. Shen W, Chen J, Gantz M, Punyanitya M, Heymsfield SB, Gallagher D, Albu J,
Engelson E, Kotler D, Pi-Sunyer X, Gilsanz V. MRI-measured pelvic bone
marrow adipose tissue is inversely related to DXA-measured bone mineral in
younger and older adults. European Journal of Clinical Nutrition. 2012;66(9):983–
988.
152. Rodriguez JP, Montecinos L, Ros S, Reyes P, Martnez J. Mesenchymal stem cells
from osteoporotic patients produce a type I collagen-deficient extracellular matrix
favoring adipogenic differentiation. Journal of Cellular Biochemistry. 2000; 79(4):
557–565.
153. Bragdon B, Burns R, Baker AH, Belkina AC, Morgan EF, Denis G V., Gerstenfeld
LC, Schlezinger JJ. Intrinsic sex-linked variations in osteogenic and adipogenic
differentiation potential of bone marrow multipotent stromal cells. Journal of
Cellular Physiology. 2015;230(2):296–307.
154. Forbes AP. Fuller Albright. His concept of postmenopausal osteoporosis and what
came of it. Clinical Orthpaedics and Related Research. 1991 Aug;8(269):128–141.
155. Jilka RL, Takahashi K, Munshi M, Williams DC, Roberson PK, Manolagas SC.
Loss of estrogen upregulates osteoblastogenesis in the murine bone marrow.
Evidence for autonomy from factors released during bone resorption. Journal of
Clinical Investigation. 1998 May 1;101(9):1942–1950.
156. Eghbali-Fatourechi G, Khosla S, Sanyal A, Boyle WJ, Lacey DL, Riggs BL. Role
of RANK ligand in mediating increased bone resorption in early postmenopausal
women. Journal of Clinical Investigation. 2003;111(8):1221–1230.
178
157. Hofbauer LC, Khosla S, Dunstan CR, Lacey DL, Spelsberg TC, Riggs BL.
Estrogen stimulates gene expression and protein production of osteoprotegerin in
human osteoblastic cells. Endocrinology. 1999 Sep;140(9):4367–4370.
158. Kousteni S, Bellido T, Plotkin LI, O’Brien CA, Bodenner DL, Han L, Han K,
DiGregorio GB, Katzenellenbogen JA, Katzenellenbogen BS, Roberson PK,
Weinstein RS, Jilka RL, Manolagas SC. Nongenotropic, sex-nonspecific signaling
through the estrogen or androgen receptors: dissociation from transcriptional
activity. Cell. 2001 Mar 9;104(5):719–730.
159. Almeida M, Iyer S, Martin-Millan M, Bartell SM, Han L, Ambrogini E, Onal M,
Xiong J, Weinstein RS, Jilka RL, O’Brien C a., Manolagas SC. Estrogen receptorα signaling in osteoblast progenitors stimulates cortical bone accrual. Journal of
Clinical Investigation. 2013;123(1):394–404.
160. Di Gregorio GB, Yamamoto M, Ali AA, Abe E, Roberson P, Manolagas SC, Jilka
RL. Attenuation of the self-renewal of transit-amplifying osteoblast progenitors in
the murine bone marrow by 17 beta-estradiol. Journal of Clinical Investigation.
2001 Apr;107(7):803–812.
161. Wakley GK, Schutte HD, Hannon KS, Turner RT. Androgen treatment prevents
loss of cancellous bone in the orchidectomized rat. Journal of Bone & Mineral
Research. 1991 Apr;6(4):325–330.
162. Chiang C, Chiu M, Moore AJ, Anderson PH, Ghasem-Zadeh A, McManus JF, Ma
C, Seeman E, Clemens TL, Morris HA, Zajac JD, Davey RA. Mineralization and
bone resorption are regulated by the androgen receptor in male mice. Journal of
Bone & Mineral Research. 2009;24(4):621–631.
163. Notini AJ, McManus JF, Moore A, Bouxsein M, Jimenez M, Chiu WSM, Glatt V,
Kream BE, Handelsman DJ, Morris H a, Zajac JD, Davey R a. Osteoblast deletion
of exon 3 of the androgen receptor gene results in trabecular bone loss in adult
male mice. Journal of Bone & Mineral Research. 2007;22(3):347–356.
164. Kawano H, Sato T, Yamada T, Matsumoto T, Sekine K, Watanabe T, Nakamura
T, Fukuda T, Yoshimura K, Yoshizawa T, Aihara K-I, Yamamoto Y, Nakamichi
Y, Metzger D, Chambon P, Nakamura K, Kawaguchi H, Kato S. Suppressive
function of androgen receptor in bone resorption. Proceedings of the National
Acacemy of Sciences of the United States of America. 2003;100(16):9416–9421.
165. Hofbauer LC, Hicok KC, Chen D, Khosla S. Regulation of osteoprotegerin
production by androgens and anti-androgens in human osteoblastic lineage cells.
European Journal of Endocrinology. 2002;147(2):269–273.
166. Nash D, Magder LS, Sherwin R, Rubin RJ, Silbergeld EK. Bone density-related
predictors of blood lead level among peri- and postmenopausal women in the
United States: The third national health and nutrition examination survey, 1988–
1994. American Journal of Epidemiology. 2004;160(9):901–911.
179
167. Pounds JG, Long GJ, Rosen JF. Cellular and molecular toxicity of lead in bone.
Environmental Health Perspectives. 1991;91:17–32.
168. Pavel OR, Popescu M, Novac L, Mogoantă L, Pavel LP, Vicaş RM, Trăistaru MR.
Postmenopausal osteoporosis — clinical, biological and histopathological aspects.
Romanian Journal of Morphology and Embryology. 2016;57(1):121–130.
169. Silbergeld EK, Schwartz J, Mahaffey K. Lead and osteoporosis: Mobilization of
lead from bone in postmenopausal women. Environmental Research. 1988;47(1):
79–94.
170. Silbergeld EK. Lead in bone: Implications for toxicology during pregnancy and
lactation. Environmental Health Perspectives. 1991;91:63–70.
171. Rosen JF, Wexler EE. Studies of lead transport in bone organ culture. Biochemical
Pharmacology. 1977 Apr 1;26(7):650–2.
172. Carmouche JJ, Puzas JE, Zhang X, Tiyapatanaputi P, Cory-Slechta DA, Gelein R,
Zuscik M, Rosier RN, Boyce BF, O’Keefe RJ, Schwarz EM. Lead exposure
inhibits fracture healing and is associated with increased chondrogenesis, delay in
cartilage mineralization, and a decrease in osteoprogenitor frequency.
Environmental Health Perspectives. 2005 Mar 14;113(6):749–755.
173. Klein RF, Wiren KM. Regulation of osteoblastic gene expression by lead.
Endocrinology. 1993;132(6):2531–2537.
174. Beier E, Maher J. Heavy metal lead exposure, osteoporotic-like phenotype in an
animal model, and depression of Wnt signaling. Environmental Health
Perspectives. 2013;121(1):97–104.
175. Beier EE, Sheu T, Dang D, Holz JD, Ubayawardena R, Babij P, Puzas JE. Heavy
Metal Ion Regulation of Gene Expression. Journal of Biological Chemistry. 2015;
290(29):18216–18226.
176. Angle CR, Thomas DJ, Swanson SA. Lead inhibits the basal and stimulated
responses of a rat osteoblast-like cell line ROS 17 2.8 to 1α,25-dihydroxyvitamin
D3 and IGF-I. Toxicology and Applied Pharmacology. 1990;103(2):281–287.
177. Ryan EP, Holz JD, Mulcahey M, Sheu T-J, Gasiewicz T a, Puzas JE.
Environmental toxicants may modulate osteoblast differentiation by a mechanism
involving the aryl hydrocarbon receptor. Journal of Bone & Mineral Research.
2007;22(10):1571–80.
178. Ballew C, Khan LK, Kaufmann R, Mokdad a, Miller DT, Gunter EW. Blood lead
concentration and children’s anthropometric dimensions in the Third National
Health and Nutrition Examination Survey (NHANES III), 1988–1994. Journal of
Pediatrics. 1999;134(5):623–630.
180
179. Kafourou A, Touloumi G, Makropoulos V, Loutradi A, Papanagiotou A, Hatzakis
A. Effects of lead on the somatic growth of children. Archives of Environmental
Health. 2009;52(5):377–383.
180. Beier EE, Holz JD, Sheu T-J, Puzas JE. Elevated lifetime lead exposure impedes
osteoclast activity and produces an increase in bone mass in adolescent mice.
Toxicological Sciences. 2016 Feb;149(2):277–288.
181. Ahmed S, Rekha RS, Ahsan K Bin, Doi M, Grander M, Roy AK, Ekstrom EC,
Wagatsuma Y, Vahter M, Raqib R. Arsenic exposure affects plasma Insulin-Like
Growth Factor 1 (IGF-1) in children in rural Bangladesh. PLoS One. 2013;8(11):
1–12.
182. Lindgren A, Vahter M, Dencker L. Autoradiographic studies on the distribution of
arsenic in mice and hamsters administered 74As-arsenite or -arsenate. Acta
Pharmacologica et Toxicologica. 1982 Sep;51(3):253–265.
183. Feussner JR, Shelburne JD, Bredehoeft S, Cohen HJ. Arsenic-induced bone
marrow toxicity: ultrastructural and electron-probe analysis. Blood. 1979 May;
53(5):820–827.
184. Wu C-T, Lu T-Y, Chan D-C, Tsai K-S, Yang R-S, Liu S-H. Effects of arsenic on
osteoblast differentiation in vitro and on bone mineral density and microstructure
in rats. Environmental Health Perspectives. 2014 Jun;122(6):559–65.
185. Hu YC, Cheng HL, Hsieh BS, Huang LW, Huang TC, Chang KL. Arsenic trioxide
affects bone remodeling by effects on osteoblast differentiation and function.
Bone. 2012;50(6):1406–1415.
186. Szymczyk KH, Kerr BAE, Freeman TA, Adams CS, Steinbeck MJ. Involvement
of hydrogen peroxide in the differentiation and apoptosis of preosteoclastic cells
exposed to arsenite. Biochemical Pharmacology. 2006;72(6):761–769.
187. Nordberg GF. Historical perspectives on cadmium toxicology. Toxicology and
Applied Pharmacology. 2009;238(3):192–200.
188. James KA, Meliker JR. Environmental cadmium exposure and osteoporosis: A
review. International Journal of Public Health. 2013;58(5):737–745.
189. Wallin M, Barregard L, Sallsten G, Lundh T, Karlsson MK, Lorentzon M, Ohlsson
C, Mellström D. Low-Level Cadmium Exposure Is Associated With Decreased
Bone Mineral Density and Increased Risk of Incident Fractures in Elderly Men:
The MrOS Sweden Study. Journal of Bone & Mineral Research. 2015;31(4):732–
741.
190. Bhattacharyya MH. Cadmium osteotoxicity in experimental animals: Mechanisms
and relationship to human exposures. Toxicology and Applied Pharmacology.
2009;238(3):258–265.
181
191. Wilson AK, Cerny EA, Smith BD, Wagh A, Bhattacharyya MH. Effects of
cadmium on osteoclast formation and activity in vitro. Toxicology and Applied
Pharmacology. 1996 Oct;140(2):451–460.
192. Guandalini GS, Zhang L, Fornero E, Centeno JA, Mokashi VP, Ortiz PA,
Stockelman MD, Osterburg AR, Chapman GG. Tissue distribution of tungsten in
mice following oral exposure to sodium tungstate. Chemical Research in
Toxicology. 2011;24(4):488–493.
193. Kelly ADR, Lemaire M, Young YK, Eustache JH, Guilbert C, Molina MF, Mann
KK. In vivo tungsten exposure alters B-cell development and increases DNA
damage in murine bone marrow. Toxicological Sciences. 2013;131(2):434–446.
194. Bolt AM, Grant MP, Wu TH, Flores Molina M, Plourde D, Kelly ADR, Negro
Silva LF, Lemaire M, Schlezinger JJ, Mwale F, Mann KK. Tungsten Promotes
Sex-Specific Adipogenesis in the Bone by Altering Differentiation of Bone
Marrow-Resident Mesenchymal Stromal Cells. Toxicological Sciences. 2016;
150(2):333–346.
195. Bolt AM, Sabourin V, Molina MF, Police AM, Negro Silva LF, Plourde D,
Lemaire M, Ursini-Siegel J, Mann KK. Tungsten Targets the Tumor
Microenvironment to Enhance Breast Cancer Metastasis. Toxicological Sciences.
2015;143(1):165–177.
196. Chen JR, Lazarenko OP, Shankar K, Blackburn ML, Badger TM, Ronis MJ. A
role for ethanol-induced oxidative stress in controlling lineage commitment of
mesenchymal stromal cells through inhibition of Wnt/β-catenin signaling. Journal
of Bone & Mineral Research. 2010;25(5):1117–1127.
197. Turner RT. Skeletal response to alcohol. Alcoholism, Clinical and Experimental
Research. 2000;24(11):1693–1701.
198. Almeida M, Han L, Ambrogini E, Bartell SM, Manolagas SC. Oxidative stress
stimulates apoptosis and activates NF-kappaB in osteoblastic cells via a
PKCbeta/p66shc signaling cascade: counter regulation by estrogens or androgens.
Molecular Endocrinology. 2010;24(10):2030–2037.
199. Chen J, Lazarenko OP, Shankar K, Blackburn ML, Lumpkin CK, Badger TM,
Ronis MJJ. Inhibition of NADPH oxidases prevents chronic ethanol-induced bone
loss in female rats. Journal of Pharmacology and Experimental Therapeutics.
2011;336(3):734–742.
200. Ilvesaro J, Pohjanvirta R, Tuomisto J, Viluksela M, Tuukkanen J. Bone resorption
by aryl hydrocarbon receptor-expressing osteoclasts is not disturbed by TCDD in
short-term cultures. Life Sciences. 2005;77(12):1351–1366.
201. Ward KD, Klesges RC. A meta-analysis of the effects of cigarette smoking on
bone mineral density. Calcified Tissue International. 2001 May;68(5):259–270.
182
202. Iqbal J, Sun L, Cao J, Yuen T, Lu P, Bab I, Leu NA, Srinivasan S, Wagage S,
Hunter C a, Nebert DW, Zaidi M, Avadhani NG. Smoke carcinogens cause bone
loss through the aryl hydrocarbon receptor and induction of Cyp1 enzymes.
Proceedings of the National Academy of Sciences of the United States of America.
2013 Jul 2;110(27):11115–11120.
203. Yu TY, Kondo T, Matsumoto T, Fujii-Kuriyama Y, Imai Y. Aryl hydrocarbon
receptor catabolic activity in bone metabolism is osteoclast dependent in vivo.
Biochemical and Biophysical Research Communications. 2014;450(1):416–422.
204. Korkalainen M, Kallio E, Olkku A, Nelo K, Ilvesaro J, Tuukkanen J, Mahonen A,
Viluksela M. Dioxins interfere with differentiation of osteoblasts and osteoclasts.
Bone. 2009 Jun;44(6):1134–1142.
205. Bennett CN, Longo K a, Wright WS, Suva LJ, Lane TF, Hankenson KD,
MacDougald O a. Regulation of osteoblastogenesis and bone mass by Wnt10b.
Proceedings of the National Academy of Sciences of the United States of America.
2005;102(9):3324–3329.
206. Aubert RE, Herrera V, Chen W, Haffner SM, Pendergrass M. Rosiglitazone and
pioglitazone increase fracture risk in women and men with type 2 diabetes.
Diabetes Obesity & Metabolism. 2010 Aug;12(8):716–721.
207. Bilik D, McEwen LN, Brown MB, Pomeroy NE, Kim C, Asao K, Crosson JC,
Duru OK, Ferrara A, Hsiao VC, Karter AJ, Lee PG, Marrero DG, Selby J V,
Subramanian U, Herman WH. Thiazolidinediones and fractures: evidence from
translating research into action for diabetes. Journal of Clinical Endocrinology and
Metabolism. 2010 Oct;95(10):4560–4565.
208. Schwartz A V, Sellmeyer DE, Vittinghoff E, Palermo L, Lecka-czernik B,
Feingold KR, Strotmeyer ES, Resnick HE, Carbone L, Beamer BA, Park SW,
Lane NE, Harris TB, Cummings SR. Thiazolidinedione use and bone loss in older
diabetic adults. Journal of Clinical Endocrinology and Metabolism. 2006
Sep;91(9):3349–3354.
209. Rahman S, Czernik PJ, Lu Y, Lecka-Czernik B. β-catenin directly sequesters
adipocytic and insulin sensitizing activities but not osteoblastic activity of PPARγ2
in marrow mesenchymal stem cells. PLoS One. 2012 Jan;7(12):e51746.
210. Choi S-S, Kim ES, Koh M, Lee S-J, Lim D, Yang YR, Jang H-J, Seo K-A, Min SH, Lee IH, Park SB, Suh P-G, Choi JH. A novel non-agonist peroxisome
proliferator-activated receptor γ (PPARγ) ligand UHC1 blocks PPARγ
phosphorylation by cyclin-dependent kinase 5 (CDK5) and improves insulin
sensitivity. Journal of Biolical Chemistry. 2014 Sep 19;289(38):26618–26629.
211. Lazarenko OP, Rzonca SO, Suva LJ, Lecka-Czernik B. Netoglitazone is a PPARgamma ligand with selective effects on bone and fat. Bone. 2006 Jan;38(1):74–84.
183
212. Kolli V, Stechschulte LA, Dowling AR, Rahman S, Czernik PJ, Lecka-Czernik B.
Partial agonist, telmisartan, maintains PPARγ serine 112 phosphorylation, and
does not affect osteoblast differentiation and bone mass. PLoS One. 2014;9(5):1–
13.
213. Kannan K, Takahashi S, Fujiwara N, Mizukawa H, Tanabe S. Organotin
compounds, including butyltins and octyltins, in house dust from Albany, New
York, USA. Archives of Environmental Contamination and Toxicology. 2010
May;58(4):901–907.
214. Antizar-Ladislao B. Environmental levels, toxicity and human exposure to
tributyltin (TBT)-contaminated marine environment. A review. Environment
International. 2008 Feb;34(2):292–308.
215. Koch H, Drexler H, Angerer J. An estimation of the daily intake of di (2ethylhexyl) phthalate (DEHP) and other phthalates in the general population.
International Journal of Hygiene and Environmental Health. 2003;206(2):77–83.
216. Tickner J a., Schettler T, Guidotti T, McCally M, Rossi M. Health risks posed by
use of di-2-ethylhexyl phthalate (DEHP) in PVC medical devices: A critical
review. American Journal of Industrial Medicine. 2001;39(1):100–111.
217. Li L-X, Chen L, Meng X-Z, Chen B-H, Chen S-Q, Zhao Y, Zhao L-F, Liang Y,
Zhang Y-H. Exposure levels of environmental endocrine disruptors in mothernewborn pairs in China and their placental transfer characteristics. PLoS One.
2013 Jan;8(5):e62526.
218. de Wit CA, Herzke D, Vorkamp K. Brominated flame retardants in the Arctic
environment—trends and new candidates. Science of the Total Environment. 2010
Jul 1;408(15):2885–2918.
219. Ali N, Harrad S, Goosey E, Neels H, Covaci A. “ Novel”brominated flame
retardants in Belgian and UK indoor dust: Implications for human exposure.
Chemosphere. 2011 May;83(10):1360–1365.
220. D’Hollander W, Roosens L, Covaci A, Cornelis C, Reynders H, Campenhout K
Van, Voogt P De, Bervoets L. Brominated flame retardants and perfluorinated
compounds in indoor dust from homes and offices in Flanders, Belgium.
Chemosphere. 2010 Sep;81(4):478–487.
221. Cariou R, Antignac J-P, Zalko D, Berrebi A, Cravedi J-P, Maume D, Marchand P,
Monteau F, Riu A, Andre F, Le Bizec B. Exposure assessment of French women
and their newborns to tetrabromobisphenol-A: occurrence measurements in
maternal adipose tissue, serum, breast milk and cord serum. Chemosphere. 2008
Oct;73(7):1036–1041.
222. He S, Li M, Jin J, Wang Y, Bu Y, Xu M, Yang X, Liu A. Concentrations and
trends of halogenated flame retardants in the pooled serum of residents of Laizhou
184
Bay, China. Environmetnal Toxicology and Chemistry. 2013;32(6):1242–1247.
223. Pfaffl MW. A new mathematical model for relative quantification in real-time RTPCR. Nucleic Acids Research. 2001 May 1;29(9):e45.
224. Kim JB, Wright HM, Wright M, Spiegelman BM. ADD1/SREBP1 activates
PPARgamma through the production of endogenous ligand. Proceedings of the
National Academy of Sciences of the United States of America. 1998 Apr 14;
95(8):4333–4337.
225. Greenberg AS, Egan JJ, Wek SA, Garty NB, Blanchette-Mackie EJ, Londos C.
Perilipin, a major hormonally regulated adipocyte-specific phosphoprotein
associated with the periphery of lipid storage droplets. Journal of Biological
Chemistry. 1991;266(17):11341–11346.
226. Tsukamoto Y, Ishihara Y, Miyagawa-Tomita S, Hagiwara H. Inhibition of
ossification in vivo and differentiation of osteoblasts in vitro by tributyltin.
Biochemical Pharmacology. 2004 Aug 15;68(4):739–746.
227. H.H.S. Bone Health and Osteoporosis: A Report of the Surgeon General. US
Department of Health and Human Services. 2004.
228. Koskela A, Viluksela M, Keinanen M, Tuukkanen J, Korkalainen M. Synergistic
effects of tributyltin and 2,3,7,8-tetrachlorodibenzo-p-dioxin on differentiating
osteoblasts and osteoclasts. Toxicology and Applied Pharmacology. 2012 Sep 1;
263(2):210–217.
229. Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C, Morony S,
Oliveira-dos-Santos a J, Van G, Itie A, Khoo W, Wakeham A, Dunstan CR, Lacey
DL, Mak TW, Boyle WJ, Penninger JM. OPGL is a key regulator of
osteoclastogenesis, lymphocyte development and lymph-node organogenesis.
Nature. 1999;397(6717):315–323.
230. Simonet W., Lacey D., Dunstan C., Kelley M, Chang M-S, Lüthy R, Nguyen H.,
Wooden S, Bennett L, Boone T, Shimamoto G, DeRose M, Elliott R, Colombero
A, Tan H-L, Trail G, Sullivan J, Davy E, Bucay N, Renshaw-Gegg L, Hughes T.,
Hill D, Pattison W, Campbell P, Sander S, Van G, Tarpley J, Derby P, Lee R,
Boyle W. Osteoprotegerin: A novel secreted protein involved in the regulation of
bone density. Cell. 1997;89(2):309–319.
231. Zhang XK, Lehmann J, Hoffmann B, Dawson MI, Cameron J, Graupner G,
Hermann T, Tran P, Pfahl M. Homodimer formation of retinoid X receptor
induced by 9-cis retinoic acid. Nature. 1992 Aug 13;358(6387):587–591.
232. Wan Y, Chong L-W, Evans RM. PPAR-gamma regulates osteoclastogenesis in
mice. Nature Medicine. 2007;13(12):1496–1503.
185
233. Wei W, Wang X, Yang M, Smith LC, Dechow PC, Wan Y. PGC1β mediates
PPARγ activation of osteoclastogenesis and rosiglitazone-induced bone loss. Cell
Metabolism. 2010;11(6):503–516.
234. Kleyer A, Scholtysek C, Bottesch E, Hillienhof U, Beyer C, Distler JH,
Tuckermann JP, Schett G, Krönke G. Liver X receptors orchestrate
osteoblast/osteoclast crosstalk and counteract pathologic bone loss. Journal of
Bone & Mineral Research. 2012 Dec;27(12):2442–2451.
235. Kim H-J, Yoon K-A, Yoon H-J, Hong JM, Lee M-J, Lee I-K, Kim S-Y. Liver X
receptor activation inhibits osteoclastogenesis by suppressing NF-κB activity and
c-Fos induction and prevents inflammatory bone loss in mice. Journal of
Leukocyte Biology. 2013;94(July):99–107.
236. Prawitt J, Beil FT, Marshall RP, Bartelt A, Ruether W, Heeren J, Amling M, Staels
B, Niemeier A. Short-term activation of liver X receptors inhibits osteoblasts but
long-term activation does not have an impact on murine bone in vivo. Bone. 2011;
48(2):339–346.
237. Robertson Remen KM, Henning P, Lerner UH, Gustafsson JÅ, Andersson G.
Activation of liver X receptor (LXR) inhibits receptor activator of nuclear factor
κB ligand (RANKL)-induced osteoclast differentiation in an LXRβ-dependent
mechanism. Journal of Biological Chemistry. 2011;286(38):33084–33094.
238. Cui H, Okuhira K, Ohoka N, Naito M, Kagechika H, Hirose A, NishimakiMogami T. Tributyltin chloride induces ABCA1 expression and apolipoprotein AI-mediated cellular cholesterol efflux by activating LXRalpha/RXR. Biochemical
Pharmacology. 2011;81(6):819–824.
239. Watt J, Schlezinger JJ. Structurally-diverse , PPAR-activating environmental
toxicants induce adipogenesis and suppress osteogenesis in bone marrow
mesenchymal stromal cells. Toxicology. 2015;331:66–77.
240. Yonezawa T, Hasegawa S ichi, Ahn JY, Cha BY, Teruya T, Hagiwara H, Nagai K,
Woo JT. Tributyltin and triphenyltin inhibit osteoclast differentiation through a
retinoic acid receptor-dependent signaling pathway. Biochemical and Biophysical
Research Communications. 2007;355(1):10–15.
241. Bouxsein ML, Boyd SK, Christiansen B a., Guldberg RE, Jepsen KJ, Müller R.
Guidelines for assessment of bone microstructure in rodents using micro-computed
tomography. Journal of Bone & Mineral Research. 2010 Jul;25(7):1468–86.
242. Gerstenfeld LC, Cho TJ, Kon T, Aizawa T, Tsay A, Fitch J, Barnes GL, Graves
DT, Einhorn TA. Impaired fracture healing in the absence of TNF-alpha signaling:
the role of TNF-alpha in endochondral cartilage resorption. Journal of Bone &
Mineral Research. 2003 Sep;18(9):1584–92.
186
243. Thudium CS, Moscatelli I, Flores C, Thomsen JS, Bruel A, Gudmann NS, Hauge
EM, Karsdal MA, Richter J, Henriksen K. A comparison of osteoclast-rich and
osteoclast-poor osteopetrosis in adult mice sheds light on the role of the osteoclast
in coupling bone resorption and bone formation. Calcified Tissue International.
2014;95(1):83–93.
244. Suter A, Everts V, Boyde A, Jones SJ, Lüllmann-Rauch R, Hartmann D, Hayman
a R, Cox TM, Evans MJ, Meister T, von Figura K, Saftig P. Overlapping functions
of lysosomal acid phosphatase (LAP) and tartrate-resistant acid phosphatase
(Acp5) revealed by doubly deficient mice. Development. 2001 Dec;128(23):4899–
4910.
245. Gowen M, Lazner F, Dodds R, Kapadia R, Feild J, Tavaria M, Bertoncello I,
Drake F, Zavarselk S, Tellis I, Hertzog P, Debouck C, Kola I. Cathepsin K
knockout mice develop osteopetrosis due to a deficit in matrix degradation but not
demineralization. Journal of Bone & Mineral Research. 1999;14(10):1654–1663.
246. Somerville JM, Aspden RM, Armour KE, Armour KJ, Reid DM. Growth of
C57Bl/6 Mice and the Material and Mechanical Properties of Cortical Bone from
the Tibia. Calcified Tissue International. 2004;74(5):469–475.
247. Zanotti S, Kalajzic I, Aguila HL, Canalis E. Sex and genetic factors determine
osteoblastic differentiation potential of murine bone marrow stromal cells. PLoS
One. 2014;9(1):1–13.
248. Zhao D, Shi Z, Warriner AH, Qiao P, Hong H, Wang Y, Feng X. Molecular
mechanism of thiazolidinedione-mediated inhibitory effects on osteoclastogenesis.
PLoS One. 2014;9(7):1–13.
249. Fuller K, Lawrence KM, Ross JL, Grabowska UB, Shiroo M, Samuelsson B,
Chambers TJ. Cathepsin K inhibitors prevent matrix-derived growth factor
degradation by human osteoclasts. Bone. 2008;42(1):200–211.
250. Li CY, Jepsen KJ, Majeska RJ, Zhang J, Ni R, Gelb BD, Schaffler MB. Mice
lacking cathepsin K maintain bone remodeling but develop bone fragility despite
high bone mass. Journal of Bone & Mineral Research. 2006;21(6):865–875.
251. US EPA. Guidance for Conducting Health Risk Assessment of Chemical Mixtures.
2000.
252. Rajapakse N, Ong D, Kortenkamp A. Defining the impact of weakly estrogenic
chemicals on the action of steroidal estrogens. Toxicological Sciences. 2001;
304:296–304.
253. Safe SH. Hazard and risk assessment of chemical mixtures using the toxic
equivalency factor approach. Environmental Health Perspectives. 1998;106 Suppl
(August):1051–1058.
187
254. Howard GJ, Webster TF. Generalized concentration addition: a method for
examining mixtures containing partial agonists. Journal of Theoretical Biology.
2009 Aug 7;259(3):469–477.
255. Howard GJ, Schlezinger JJ, Hahn ME, Webster TF. Generalized concentration
addition predicts joint effects of aryl hydrocarbon receptor agonists with partial
agonists and competitive antagonists. Environmental Health Perspectives. 2010
May;118(5):666–672.
256. Fang M, Webster TF, Stapleton HM. Activation of human peroxisome
proliferator-activated nuclear receptors (PPARγ1) by semi-volatile compounds
(SVOCs) and chemical mixtures in indoor dust. Environmental Science &
Technology. 2015;150730154949006.
257. Feige JN, Gelman L, Michalik L, Desvergne B, Wahli W. From molecular action
to physiological outputs: Peroxisome proliferator-activated receptors are nuclear
receptors at the crossroads of key cellular functions. Progress in Lipid Research.
2006;45(2):120–159.
258. Janesick A, Blumberg B. Minireview: PPARγ as the target of obesogens. Journal
of Steroid Biochemistry and Molecular Biology. 2011 Oct;127(1–2):4–8.
259. Silva E, Rajapakse N, Kortenkamp A. Something from “nothing”— eight weak
estrogenic chemicals combined at concentrations below NOECs produce
significant mixture effects. Environmental Science & Technology. 2002 Apr 15;
36(8):1751–1756.
260. Webster TF. Mixtures of endocrine disruptors: How similar must mechanisms be
for concentration addition to apply? Toxicology. 2013;314(2–3):129–133.
261. Payne J, Scholze M, Kortenkamp A. Mixtures of four organochlorines enhance
human breast cancer cell proliferation. Environmental Health Perspectives.
2001;109(4):391–397.
262. Berger JP, Petro AE, Macnaul KL, Kelly LJ, Zhang BB, Richards K, Elbrecht A,
Johnson B a, Zhou G, Doebber TW, Biswas C, Parikh M, Sharma N, Tanen MR,
Thompson GM, Ventre J, Adams AD, Mosley R, Surwit RS, Moller DE. Distinct
properties and advantages of a novel peroxisome proliferator-activated protein
[gamma] selective modulator. Molecular Endocrinology. 2003 Apr;17(4):662–76.
263. Pillai HK, Fang M, Beglov D, Kozakov D, Vajda S, Stapleton HM, Webster TF,
Schlezinger JJ. Ligand binding and activation of PPARγ by firemaster 550: Effects
on adipogenesis and osteogenesis in vitro. Environmental Health Perspectives.
2014;122(11):1225–1232.
264. WHO/UNEP. State of the science of endocrine disrupting chemicals. 2013.
188
265. Carlin DJ, Rider C V., Woychik R, Birnbaum LS. Unraveling the health effects of
environmental mixtures: An NIEHS priority. Environmental Health Perspectives.
2013;121(1):6–8.
266. Ahmadian M, Suh JM, Hah N, Liddle C, Atkins AR, Downes M, Evans RM.
PPARγ signaling and metabolism: the good, the bad and the future. Nature
Medicine. 2013 May;19(5):557–566.
267. Parks LG, Ostby JS, Lambright CR, Abbott BD, Klinefelter GR, Barlow NJ, Gray
LE. The plasticizer diethylhexyl phthalate induces malformations by decreasing
fetal testosterone synthesis during sexual differentiation in the male rat.
Toxicological Sciences. 2000;58(2):339–349.
268. Dalgaard M, Nellemann C, Lam HR, Sorensen IK, Ladefoged O. The acute effects
of mono(2-ethylhexyl)phthalate (MEHP) on testes of prepubertal Wistar rats.
Toxicology Letters. 2001;122(1):69–79.
269. Howdeshell KI, Wilson VS, Furr J, Lambright CR, Rider C V., Blystone CR,
Hotchkiss AK, Gray LE. A mixture of five phthalate esters inhibits fetal testicular
testosterone production in the Sprague-Dawley rat in a cumulative, dose-additive
manner. Toxicological Sciences. 2008;105(1):153–165.
270. Pereira-Fernandes A, Demaegdt H, Vandermeiren K, Hectors TLM, Jorens PG,
Blust R, Vanparys C. Evaluation of a screening system for obesogenic compounds:
screening of endocrine disrupting compounds and evaluation of the PPAR
dependency of the effect. PLoS One. 2013 Jan;8(10):e77481.
271. Hatch EE, Nelson JW, Stahlhut RW, Webster TF. Association of endocrine
disruptors and obesity: Perspectives from epidemiological studies. International
Journal of Andrology. 2010;33(2):324–331.
272. Buser MC, Murray HE, Scinicariello F. Age and sex differences in childhood and
adulthood obesity association with phthalates: Analyses of NHANES 2007–2010.
International Journal of Hygiene and Environmental Health. 2014;217(6):687–694.
273. Zota AR, Calafat AM, Woodruff TJ. Temporal trends in phthalate exposures:
Findings from the national health and nutrition examination survey, 2001–2010.
Environmental Health Perspectives. 2014;122(3):235–241.
274. Langer S, Bekö G, Weschler CJ, Brive LM, Toftum J, Callesen M, Clausen G.
Phthalate metabolites in urine samples from Danish children and correlations with
phthalates in dust samples from their homes and daycare centers. International
Journal of Hygiene and Environmental Health. 2014;217(1):78–87.
275. Cornu MC, Lhuguenot JC, Brady AM, Moore R, Elcombe CR. Identification of
the proximate peroxisome proliferator(S) derived from DI (2-ethylhexyl) adipate
and species differences in response. Biochemical Pharmacology. 1992;43(10):
2129–2134.
189
276. Scholze M, Silva E, Kortenkamp A. Extending the applicability of the dose
addition model to the assessment of chemical mixtures of partial agonists by using
a novel toxic unit extrapolation method. PLoS One. 2014;9(2): e88808.
277. Tan Y, Muise ES, Dai H, Raubertas R, Wong KK, Thompson GM, Wood HB,
Meinke PT, Lum PY, Thompson JR, Berger JP. Novel transcriptome profiling
analyses demonstrate that selective peroxisome proliferator-activated receptor γ
(PPARγ) modulators display attenuated and selective gene regulatory activity in
comparison with PPARγ full agonists. Molecular Pharmacology. 2012;82(1):68–
79.
278. Rx List. Avandia (Rosiglitazone maleate) Drug information [Internet]. 2014 [cited
2016 Feb 11]. Available from: http://www.rxlist.com/avandia-drug/clinicalpharmacology.htm
279. Robertson Remen KM, Lerner UH, Gustafsson J -A., Andersson G. Activation of
the liver X receptor potently inhibits osteoclastogenesis from lipopolysaccharideexposed bone marrow-derived macrophages. Journal of Leukocyte Biology. 2012;
93(January):71–82.
280. Sanz-Salvador L, Garcia-Perez MA, Tarin JJ, Cano A. Bone metabolic changes
during pregnancy: A period of vulnerability to osteoporosis and fracture. European
Journal of Endocrinology. 2015;172(2):R53–65.
281. Cummings SR Melton LJ. Epidemiology and outcomes of osteoporotic fractures.
Lancet. 2002;359(9319):1761–1767.
282. LeBlanc ES, Hillier TA, Pedula KL, Rizzo JH, Cawthon PM, Fink HA, Cauley JA,
Bauer DC, Black DM, Cummings SR, Browner WS. Hip fracture and increased
short-term but not long-term mortality in healthy older women. Archives of
Internal Medicine. 2011;171(20):1831–1837.
283. Klotzbuecher CM, Ross PD, Landsman PB, Abbott TA, Berger M. Patients with
prior fractures have an increased risk of future fractures: a summary of the
literature and statistical synthesis. Journal of Bone & Mineral Research 2000;
15(4):721–739.
284. Fink HA, Ensrud KE, Nelson DB, Kerani RP, Schreiner PJ, Zhao Y, Cummings
SR, Nevitt MC. Disability after clinical fracture in postmenopausal women with
low bone density: The fracture intervention trial (FIT). Osteoporosis International.
2003;14(1):69–76.
285. Willig R, Keinänen-Kiukaaniemi S, Jalovaara P. Mortality and quality of life after
trochanteric hip fracture. Public Health. 2001;115(5):323–327.
286. Felsenberg D, Silman AJ, Lunt M, Armbrecht G, Ismail AA, Finn JD, Cockerill
WC, Banzer D, Benevolenskaya LI, Bhalla A, Bruges Armas J, Cannata JB,
Cooper C, Dequeker J, Eastell R, Felsch B, Gowin W, Havelka S, Hoszowski K,
190
Jajic I, Janott J, Johnell O, Kanis JA, Kragl G, Lopes Vaz A, Lorenc R, Lyritis G,
Masaryk P, Matthis C, Miazgowski T, Parisi G, Pols HA, Poor G, Raspe HH, Reid
DM, Reisinger W, Schedit-Nave C, Stepan JJ, Todd CJ, Weber K, Woolf AD,
Yershova OB, Reeve J, O’Neill TW. Incidence of vertebral fracture in europe:
results from the European Prospective Osteoporosis Study (EPOS). Journal of
Bone & Mineral Research. 2002;17(4):716–24.
287. Barbour KE, Lui L-Y, Ensrud KE, Hillier TA, LeBlanc ES, Ing SW, Hochberg
MC, Cauley JA. Inflammatory markers and risk of hip fracture in older white
women: The study of osteoporotic fractures. Journal of Bone & Mineral Research.
2014;29(9):2057–2064.
288. Matthiessen P. Detection, monitoring, and control of tributyltin: An almost
complete success story. Environmental Toxicology and Chemistry. 2013;
32(3):487–9.
289. Airaksinen R, Rantakokko P, Turunen AW, Vartiainen T, Vuorinen PJ,
Lappalainen A, Vihervuori A, Mannio J, Hallikainen A. Organotin intake through
fish consumption in Finland. Environmental Research. 2010 Aug;110(6):544–547.
290. Rantakokko P, Turunen A, Verkasalo PK, Kiviranta H, Mannisto S, Vartiainen T.
Blood levels of organotin compounds and their relation to fish consumption in
Finland. Science of the Total Environment. 2008;399(1–3):90–95.
291. Currier N, Solomon SE, Demicco EG, Chang DLF, Farago M, Ying H,
Dominguez I, Sonenshein GE, Cardiff RD, Xiao Z-XJ, Sherr DH, Seldin DC.
Oncogenic signaling pathways activated in DMBA-induced mouse mammary
tumors. Toxicologic Pathology. 2005;33(6):726–737.
292. Hall JM, Barhoover M a, Kazmin D, McDonnell DP, Greenlee WF, Thomas RS.
Activation of the aryl-hydrocarbon receptor inhibits invasive and metastatic
features of human breast cancer cells and promotes breast cancer cell
differentiation. Molecular Endocrinology. 2010;24(2):359–369.
191
CURRICULUM VITAE
NAME:
James Douglas Watt
BIRTH YEAR:
1980
ADDRESS:
Department of Environmental Health
Boston University School of Public Health
715 Albany Street, Talbot 4 West
Boston, MA, 02118
PHONE:
617-833-8786
EMAIL:
jwatt@bu.edu
EDUCATION
2012–2016
2008–2011
1998–2002
Ph.D. Environmental Health (September 2016)
Boston University School of Public Health, Boston, MA
Dissertation Title: Identifying and Modeling the Contribution of
Nuclear Receptors to Environmental Obesogen-Induced Toxicity in
Bone
Defense Date: June 13, 2016
M.S. Environmental Health
Boston University School of Public Health, Boston, MA
B.A. Kinesiology and Applied Physiology
University of Colorado, Boulder, CO
AWARDS AND HONORS
2016
3rd place, Society of Toxicology New England Chapter Student
Travel Award, Society of Toxicology Annual Meeting, New Orleans,
LA
2016
Graduate Student Travel Award, Society of Toxicology Annual
Meeting, New Orleans, LA
2015
2nd place, Society of Toxicology New England Chapter Fall Meeting
Poster Competition, Boston, MA
2014
Dean’s Award, Boston University Scholar’s Day Poster Exhibition
2013
1st place, Ph.D. Student Category, Boston University School of Public
Health Research Day Poster Competition
192
PROFESSIONAL EXPERIENCE
2003–2012
Lab technician, Moore Laboratory, Boston University School of
Medicine, Department of Physiology and Biophysics, Boston, MA
2010–2012
Lab technician, Gabel Laboratory, Boston University School of
Medicine, Department of Physiology and Biophysics, Boston, MA
2011
Summer Intern, Massachusetts Department of Environmental
Protection, Office of Research and Standards, Boston, MA
2002–2003
Lab technician, Yarus Laboratory, University of Colorado at Boulder,
Department of Molecular, Cellular and Developmental Biology,
Boulder, CO
2002
Undergraduate Intern, Human Cardiovascular Research Laboratory,
University of Colorado at Boulder, Department of Kinesiology and
Applied Physiology, Boulder, CO
TEACHING ACTIVITIES
2013–2015
Teaching Assistant/Guest Lecturer, Introduction to Toxicology (Dr.
Jennifer Schlezinger), Boston University School of Public Health,
Boston, MA
2014–2015
Teaching Assistant/Guest Lecturer, Quantitative and Analytical
Methods in Environmental Health (Dr. Wendy Heiger-Bernays, Dr.
Patricia Janulewicz), Boston University School of Public Health,
Boston MA
2011, 2013
Teaching Assistant/Guest Lecturer, Introduction to Environmental
Health (Dr. Madeleine Scammell, Dr. Junenette Peters), Boston
University School of Public Health, Boston, MA
SELECTED PEER-REVIEWED PUBLICATIONS
2016
Watt J, Webster TF, Schlezinger JJ. Generalized Concentration
Addition predicts mixture effects of environmental PPARgamma
agonists. Toxicological Sciences (in press)
2015
Watt J, Schlezinger JJ. Structurally diverse, PPARgamma-activating
environmental obesogens induce adipogenesis and suppress
osteogenesis in bone marrow mesenchymal stromal cells. Toxicology
331: 66–77. doi: 10.1016/j.tox.2015.03.006
2015
Baker AH, Watt J, Huang C, Gerstenfeld LC, Schlezinger JJ.
Tributyltin engages multiple nuclear receptor pathways and
suppresses osteogenesis in bone marrow multipotent stromal cells.
Chemical Research in Toxicology 28(6): 1156–1166. doi:
10.1021/tx500433r
SCIENTIFIC MEETING PRESENTATIONS
2016
Watt J, Baker A, Meeks B, Morgan E, Gerstenfeld L, Schlezinger J.
Tributyltin engages the RXR and LXR nuclear receptor pathways to
modify osteoblast/osteoclast crosstalk and enhance bone deposition in
193
2016
2015
2015
2015
2014
2014
2013
2013
2012
2008
a sex-dependent manner. Society of Toxicology Annual Meeting,
New Orleans, LA. Poster presentation.
Watt J, Webster T, Schlezinger J. Strengths and limitations of
Generalized Concentration Addition in modeling PPARγ activation
by endocrine-disrupting compounds. Society of Toxicology Annual
Meeting, New Orleans, LA. Poster presentation.
Watt J, Webster T, Schlezinger J. “Generalized Concentration
Addition Modeling of PPARgamma Activation by Phthalate
Compounds in the ToxCast Data Set.” NIEHS Superfund Research
Program Annual Meeting, San Juan, Puerto Rico. Oral presentation
Watt J, Baker A, Meeks B, Morgan E, Gerstenfeld L, Schlezinger J.
The environmental contaminant tributyltin engages the RXR and
LXR nuclear receptor pathways and enhances bone formation in
female C57Bl-6J mice. Society of Toxicology New England Regional
Chapter Fall Meeting, Boston MA. Poster presentation.
Watt J, Webster TJ, Schlezinger J. Predicting Joint Effects of
PPARgamma Ligands Using Generalized Concentration Addition.
Society of Toxicology Annual Meeting, San Diego, CA. Poster
presentation.
Watt J, Webster TJ, Schlezinger J. Predicting Joint Effects of
PPARgamma Ligands Using Generalized Concentration Addition.
NIEHS Superfund Research Program Annual Meeting, San Jose, CA.
Poster presentation.
Watt J, Andrews FV, Webster TJ, Schlezinger J. Emerging
Toxicants Induce Adipogenesis and Suppress Osteogenesis in Mouse
Mesenchymal Stromal Cells. Society of Toxicology Annual Meeting,
Phoenix, AZ. Poster presentation.
Watt J, Andrews FV, Webster TJ, Schlezinger J. Building Fat in
Bone: The Role of Environmental Obesogens. NIEHS Superfund
Research Program Annual Meeting, Baton Rouge, LA. Poster
presentation.
Watt J, Andrews FV, Schlezinger J. Environmental Obesogens: Are
They Bad to the Bone? Elsevier Environmental Health Conference,
Boston, MA. Oral presentation.
Baker AH, Watt J, Meeks BD, Gerstenfeld LC, Schlezinger J.
Environmental Obesogens: Are They Bad to the Bone? NIEHS
Superfund Research Program Annual Meeting, Raleigh, NC. Poster
presentation.
Pant K, Watt J, Greenberg MJ, Jones M, Szczesna-Cordary D,
Moore JR. Removal of the cardiac myosin regulatory light chain
increases isometric force production. Biophysical Society Meeting,
Long Beach, CA. Poster presentation.