The National Institute of Dental and Craniofacial Research, a division of the National Institutes of Health, reported that temporomandibular joint (TMJ) disease is the second most common musculoskeletal disease in the United States, with 10.8 million people suffering from TMJ problems at any given time. Eighty percent of individuals seeking treatment for TMJ disorders (TMDs) are females of childbearing age. Because of the high prevalence of TMDs in women of reproductive age, it has been postulated that estrogen may make an individual susceptible to TMD.1–9 Estrogen modulates multiple biological processes within the TMJ region including inflammation,10 matrix metalloproteinase activity,11 and pain modulation.12 However, none of these have been able to adequately explain the gender predilection for TMD, suggesting that other estrogen-dependent mechanisms are involved.
Estrogen inhibits the growth of the mandibular condyle. Addition of estrogen to rat mandibular condylar organ cultures results in a decrease in condylar cartilage thickness and proliferation.13 Numerous studies have shown that ovariectomy causes an increase in the thickness of the mandibular condylar cartilage in young rats,14, 15 newborn mice,16 and young mice17; and this is reversed by the administration of exogenous estrogen.15 Not only does lack of estrogen (ovariectomy) increase the condylar cartilage thickness but it also increases the width of the mandibular condylar head, which includes both the condylar cartilage and the subchondral bone, in young mice.17 Furthermore, estrogen negatively regulates condylar cartilage differentiation in vivo. In 4-month-old mice, ovariectomy has been shown to cause an increase in collagen type X expression after 1 to 8 weeks.18 The ability of estrogen to decrease growth and differentiation of the mandibular condylar cartilage is consistent with the literature on growth plate cartilages; however, estrogen is believed to have a chondroprotective effect on articular cartilages (for review, see Tanko and colleagues19). The distinct roles of estrogen in the different cartilages might help explain why the peak incidence of TMD occurs when estrogen levels are the highest; as opposed to the peak incidence of osteoarthritis in other joints, which occurs when estrogen levels are at the lowest (eg, menopausal women).
The classical estrogen receptors, alpha and beta, have distinct roles in estrogen's regulation of skeletal growth. ERβ-deficient mice have normal estrogen levels20 and skeletal axial growth is affected only in adult female mice. Specifically, ERβ-deficient adult female mice (70–240 days old) have increased axial growth (crown-rump length) compared with the matched wild-type (WT) mice.20, 21 On the other hand, estrogen receptor alpha (ERα) seems to promote skeletal linear growth in male mice, but its effect on female mice is unclear.20–22 Because ERβ inhibits skeletal growth exclusively in females, the goal of this study was to examine whether a similar mechanism occurs in the TMJ. We hypothesized that female ERβ-deficient mice would have larger mandibular condyles than female WT mice. Greater understanding of estrogen signaling within the TMJ is critical to deciphering the gender predilection of the disease process.
Subjects and Methods
Breeding pairs of C57Bl/6 wild type (Cat# 000664) and ERβ KO mice (homozygous male, heterozygous female, Cat# 004745) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). For the studies with ERβ KO mice, only homozygous females were used. Seven-day-old to120-day-old female C57Bl/6 WT (total n = 42) and ERβ KO mice (total n = 43) were used for the study (see Table 1 for details). At either 3 or 16 hours prior to euthanasia, mice were injected intraperitoneally with 0.1 mg bromodeoxyuridine (BrdU) per gram body weight. All experiments were performed in accordance with animal welfare based on an approved IACUC protocol# AAAD0950 from Columbia University animal care committee.
µCT = micro–computed tomography; n = number of mice; BrdU = bromodeoxyuridine; ERβ KO = estrogen receptor beta knockout; IGF-1 = insulin-like growth factor 1.
These mice were also used for measuring body weight, serum IGF-1 level in 49-day-old mice, and condylar cartilage thickness.
RNA extraction and PCR amplification
For each mouse the mandibular condylar heads (left and right), containing both the mandibular condylar cartilage and subchondral bone, were carefully harvested and all the soft tissues were removed under a dissecting microscope. Total RNA was obtained from the condylar head and extracted with TRIzol Reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) following the manufacturer's protocol. Total RNA was reverse-transcribed into cDNA using the ABI High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, USA) following the manufacturer's protocol. Real-time PCR was performed, for expression of different genes, in separate wells (single-plex assay) of 96-well plates with reaction volume of 20 µL. Gapdh was used as an endogenous control. Three replicates of each sample were amplified using Assays-on-Demand Gene Expression (Applied Biosystems) for the particular gene of interest, with predesigned unlabeled gene-specific PCR primers and TaqMan MGB FAM dye-labeled probes. The PCR reaction mixture (including 2× TaqMan Universal PCR Master Mix, 20× Assays-on-Demand Gene Expression Assay Mix, 50 ng of cDNA) was run in an Applied Biosystems ABI Prism 7300 Sequence Detection System instrument using universal thermal cycling parameters. For the genes in which the efficiencies of target and endogenous control amplification were approximately equal, relative expression in a test sample was compared to a reference calibrator sample (ΔΔCt method) and used for data analysis. For the genes that were not amplified with the same efficiency as the endogenous control the Relative Standard Curve method, in which target quantity is determined from the standard curve and divided by the target quantity of the calibrator, was used. Gene expression was performed for proteoglycan 4 (Prg4), parathyroid hormone related protein (Pthrp), SRY-box containing gene 9 (Sox9), collagen type II (Col2a1), Indian hedgehog (Ihh), collagen type X (Col10a1), vascular endothelial growth factor (Vegf), insulin-like growth factor 1 (Igf-1), receptor activator for nuclear factor κ B ligand (Rankl), and osteoprotegerin (Opg).
Histology and immunohistochemistry
Whole mouse heads were sectioned into halves, fixed in 10% formalin for 4 days at room temperature and decalcified in 14% EDTA (pH 7.1) (Sigma, St. Louis, MO, USA) for 10 days and 14 days in 49-day and 120-day samples, respectively. Subsequently, the samples were processed through progressive concentrations of ethanol, cleared in xylene and embedded in paraffin. The TMJ was sagittally serially sectioned into 5-µm sections by Microm HM 355s microtome (Thermo Fisher Scientific, Waltham, MA, USA), and every fifth section was stained with hematoxylin and eosin (H&E).
Histomorphometry measurements were made in a blinded, nonbiased manner using the BioQuant computerized image analysis system (BioQuant, Nashville, TN, USA). Mandibular condylar cartilage thickness measurements were performed on H&E sagittal sections corresponding to the mid-coronal portion of the mandibular condylar head. The entire mandibular condylar cartilage area was divided into three layers, superficial layer (layer S), flattened layer (layer F), and hypertrophic layer (layer H) based on the cartilage cell size, shape, and staining (Fig. 2A).The entire anterior-posterior condylar cartilage area was measured for each section and the average of at least three sections was calculated as the cartilage thickness for that mouse. Measurements included cell number, thickness, and area of each layer. Six to eight mice from each age and genotype group were analyzed.
For immunohistochemistry, tissue sections were deparaffinized with xylene and rehydrated with decreasing concentrations of ethanol. Following rehydration, the sections were treated with 3% peroxide to block endogenous peroxidase activity and digested for 60 minutes with pepsin for unmasking (Lab Vision, Fremont, CA, USA; Cat# AP-9007-006). All sections were blocked with Protein Block Serum-Free (DakoCytomation, Carpinteria, CA, USA; Code# X0909). Immunohistochemical staining was performed using the LSAB + System-HRP Kit (DakoCytomation; Code# K0690) following the procedure recommended by the manufacturer. Primary antibodies used in this study were collagen type II antigen (Millipore; MAB8887, 1:200 dilution), Tieg1 (Santa Cruz Biotechnologies, Santa Cruz, CA, USA; sc-23159, 1:100 dilution), p57 (Santa Cruz; sc-8298, 1:200 dilution) and IGF-1 (Abcam; ab40657, 1:100 dilution). BrdU immunohistochemical analysis was completed using a BrdU staining kit following the manufacturer's instructions (Zymed Laboratories-Invitrogen Corporation, Carlsbad, CA, USA). Negative controls were prepared by omitting the primary antibody step and incubating with blocking solution. To quantify BrdU, Tieg1, and p57 staining, the labeling index (number of BrdU-, p57-, or Tieg1-positive cells divided by the total number of cells) was calculated. Three to five sections, corresponding to the same anatomical area (mid-sagittal), were counted for each animal and the average of these sections was used for the labeling index.
Osteoclastic cell measurement were performed on tartrate-resistant acid phosphatase (TRAP)-stained sections by TRAP kit (Sigma Aldrich, St. Louis, MO, USA). The measurements were made in the bone region beneath the mandibular condylar cartilage. The thickness of the measured region was designated to the approximate cartilage thickness of the correlated region. Osteoclast number, osteoclastic surface, and bone surface in this region were analyzed with BioQuant computerized image analysis system.
Micro–computed tomography analysis
The subchondral bones of the mandibular condyles from C57Bl/6 WT and ERβ KO female mice were analyzed. The three-dimensional morphology of the subchondral bone was evaluated by the micro–computed tomography (µCT) facility at the University of Connecticut Health Center. The analysis included bone volume, total volume, trabecular number, trabecular spacing, and trabecular thickness.
Mouse serum was collected from 49-day-old WT and ERβ KO female mice. Quantitative assay was performed with IGF1 Mouse ELISA Kit (Abcam; ab100695) following the manufacturer's protocol. The serum was diluted 1:250 in Assay Diluent A (from the Kit).
Statistical significance of differences among means was determined by one-way analysis of variance (ANOVA) with post hoc comparison of more than two means by the Bonferroni method using GraphPad Prism (San Diego, CA, USA).
At 49 days of age WT and ERβ KO female mice weighed the same, but by 120 days of age, ERβ KO mice weighed significantly more than age-matched WT mice (Fig. 1A). There was a significant increase in the female ERβ KO mandibular condylar cartilage thickness at 49 and 120 days of age compared with age-matched female WT mice (Fig. 1B).
Sagittal sections of the mandibular condylar cartilage from 49-day-old and 120-day-old female ERβ KO and WT mice were stained with H&E (Fig. 2A). The mandibular condylar cartilage can be divided into four zones: articular, polymorphic, flattened, and hypertrophic. In the mandibular condylar cartilage from the 49-day-old ERβ KO mice, cells in the polymorphic and flattened zones were vertically stacked, reminiscent of the proliferating zone of the growth plate cartilage. The vertically stacked cells were not apparent in the mandibular condylar cartilage from WT mice of the same age. As a result of the difficulty in delineating articular from polymorphic zones in 49-day-old and 120-day-old mice, cell counts were done for the layer S, which includes both the articular and polymorphic zones; the layer F, which includes cells in the flattened zone; and the layer H, which includes hypertrophic chondrocytes not embedded in the subchondral bone. We found significant increases in the total number of cells and in the number of cells in the flattened layer in 49-day-old and 120-day-old ERβ KO mice compared with the WT mice (Fig. 2B, C). Thickness of the F layer was also significantly increased in 49-day-old and 120-day-old KO mice (Fig. 2B, C).
Gene expression analysis revealed significant increases in Prg4, Pthrp, Col10a1, and Opg mRNA expression and significant decreases in Sox9, Rankl, Vegf, and Ihh mRNA expression in the 49-day-old ERβ KO compared with age-matched WT mice (Fig. 3B). Immunohistochemistry of collagen type 2 expression revealed similar levels of expression between ERβ KO and WT mice (Fig. 3A), which was consistent with the gene expression analysis results. However, when gene expression assay was performed with 120-day-old WT and ERβ KO mice, there were no significant differences in any of the genes examined (Fig. 3C).
Cell cycle exit
Proliferation was measured by BrdU immunohistochemistry in the mandibular condylar cartilage from female WT and ERβ KO mice injected with BrdU 3 hours prior to euthanasia. We found no significant difference in the number of BrdU-labeled cells compared to the total number of cells in the mandibular condylar cartilage from 49-day-old (Fig. 3B) and 120-day-old (data not shown) female ERβ KO compared to age-matched female WT mice. In order to determine if the increase in cell numbers in the ERβ KO mice was due to a delay in the number of cells exiting the proliferative pool, we examined markers of cell-cycle exit. Immunohistochemistry for BrdU was performed on the mandibular condylar cartilage of mice that received BrdU either 3 hours or 16 hours prior to euthanasia. There was a significant decrease in the percentage of BrdU-labeled cells after 16 hours compared with 3 hours in WT mice; but no significant difference was evident in the ERβ KO mice (Fig. 4). This decrease in BrdU labeling represents cells that have exited the proliferative pool of the mandibular condylar cartilage.23 Immunohistochemistry was also performed for two regulators of cell-cycle exit, Tieg1 and p57. Both were predominantly localized to the area of transition between flattened and hypertrophic zones in the mandibular condylar cartilage. We also found that there was a significant decrease in p57 and Tieg1 in the female ERβ KO compared with age-matched WT mice (Fig. 5).
Local and serum IGF-1
ELISA assay showed a tendency toward increased serum IGF-1 levels in 49-day-old ERβ KO mice compared with WT. However, the difference is not statistically significant (Fig. 1C, p = 0.052). Real-time PCR detected no significant difference in condylar head mRNA levels (Fig. 3C) between WT and ERβ KO mice, in either the 49-day-old or 120-day-old groups. Similarly, immunohistochemistry showed no difference in the IGF-1 protein distribution pattern in the condylar cartilage of WT and ERβ KO female mice (data not shown).
µCT analysis revealed a significant increase in the total volume and a significant decrease in the bone density (BV/TV) and trabecular thickness in 49-day-old ERβ KO compared with age-matched WT mice (Fig. 6). µCT of 120-day-old WT and ERβ KO mice revealed no significant differences in any of the measurements. TRAP staining was performed to quantitate the number of osteoclasts and the ratio of osteoclast surface/bone surface (OcS/BS) in 49-day-old and 120-day-old groups. We found that in 49-day-old mice, there were significant decreases in both the number of osteoclasts and OcS/BS in the subchondral bone of ERβ KO mice compared with age-matched WT mice (Fig. 7).
The mandibular condylar cartilage is unique compared with other cartilages in the body in a number of ways. For example, other articular cartilages are composed of hyaline cartilage, whereas the mandibular condylar cartilage is composed of fibrocartilage (reviewed in Benjamin and Ralphs24) displaying distinct layers of cells at various stages of differentiation. Furthermore, the mandibular condylar cartilage undergoes endochondral ossification and is derived from periosteum.25 Interestingly, estrogen acting via the ERβ pathway has been shown to inhibit endochondral ossification of the femur,26 mechanical load-induced periosteal bone formation,27 and fibrocartilage maturation of the fracture callus.28
In our study we found that the female ERβ KO mice had increased mandibular condylar cartilage thickness and an increased number of cells in the flattened zone at 49 and 120 days of age compared with age-matched female WT mice. We believe that this increase may be due to a delay in cell-cycle exit in the cells of the flattened zone. In support of this claim, we found a significant decrease in the number of BrdU-labeled cells in female WT mice that were injected with BrdU 16 hours prior to euthanasia when compared with those injected with BrdU 3 hours prior to euthanasia. This reduction in BrdU labeling signifies cells have exited the flattened zone and undergone hypertrophic maturation. In contrast, ERβ KO female mice exhibited no significant differences in BrdU labeling when BrdU was injected 16 hours prior to euthanasia compared to 3 hours prior to euthanasia. A significant decrease in the expression of the cell-cycle regulators, p57 and Tieg1, was found in the mandibular condyle of female ERβ KO compared with female WT mice. In growth plate cartilage, p57 has been shown to inhibit cyclin-dependent kinases, and its activation is associated with cells exiting the proliferative pool and undergoing hypertrophic maturation.29, 30 Tieg1 has been shown to act downstream of ERβ signaling.31 Furthermore, in tenocytes32 and myoblasts33 Tieg1 deficiency caused a delay in cell-cycle exit.
Another possible explanation for the increased mandibular condylar cartilage thickness is that ERβ KO mice have been shown to have increased serum IGF-1 levels,20 and exogenous IGF-1 treatment causes increased mandibular condylar cartilage thickness.34 However, we do not believe this is the major mechanism involved because we did not find any significant changes in serum IGF-1 levels or in the local expression of IGF-1 mRNA within the mandibular condylar cartilage of female ERβ KO compared with wild-type mice. Additionally, IGF-1 treatment was recently shown to cause increased mandibular condylar cartilage proliferation,35 which was not part of our ERβ KO mandibular condylar cartilage phenotype.
A more pronounced phenotype was present in the mandibular condylar cartilage of ERβ KO female mice at 49 days of age than at 120 days of age. For example, 49-day-old female ERβ KO mice presented with decreased osteoclast numbers, increased chondrocyte maturation markers, and a larger mandibular condylar total volume compared with the age-matched WT mice. At 2 to 4 months of age, the long bones of ERβ KO female mice exhibit an increased femur length compared with age-matched WT mice; this difference is not evident in 8-month-old mice.21, 26 Unlike the femur, the majority of growth of the mandibular condyle occurs before puberty,36–38 which may explain the more pronounced female ERβ KO mandibular condylar phenotype observed at 49 days of age.
There were decreased osteoclast numbers in the subchondral bone of the mandibular condylar cartilage in female ERβ KO mice, which is consistent with what has been reported in ERβ KO long bones.39 The decrease in the number of osteoclasts could be due to either decreased RANKL expression in the hypertrophic chondrocytes of the mandibular condylar cartilage40 or decreased Tieg1 expression, which has been shown to repress OPG promoter activity in female ERβ KO mice.41 We speculate that the increase in Col X expression in the mandibular condylar cartilage of ERβ KO mice is due to a decrease in osteoclasts (as seen in Fig. 7). In support, accumulation of hypertrophic chondrocytes in the mandibular condyle has been reported in models of defective osteoclasts.42
Decreased levels of estrogen by ovariectomy resulted in increased mandibular condylar cartilage thickness and increased osteoclast numbers compared with female sham-operated controls.16, 43 The fact that we see decreased osteoclast numbers in the mandibular condyles of the ERβ KO mice suggests that other estrogen-mediated pathways are involved. For example, ERα may be also involved; estrogen through the ERα has been shown to cause increased osteoclast apoptosis.44
In this study we show that deletion of ERβ reduces osteoclast surface/bone surface and delays the number of cells exiting the flattened zone, resulting in a larger mandibular condyle at 49 days of age. In the growth plate cartilage, there is believed to be a finite number of cells capable of proliferation.45 We have previously shown that mechanical loading also causes increased mandibular condylar cartilage proliferation.46 Therefore, it is possible that high estrogen levels during puberty may deplete the number of proliferative cells in the mandibular condylar cartilage, making postpubertal females less responsive to mechanical load–induced TMJ remodeling. A similar phenomenon occurs in periosteal bones, where prepubertal females exhibit a greater increase in load-induced periosteal bone formation than postpubertal females.47 In our future studies we plan to investigate the effects of ERβ deficiency in decreased versus increased TMJ loading models.
All authors state that they have no conflicts of interest.
This work was supported by NIH 1R01DE020097 (to SW and HD). We thank Dr. Ilona Polur (Columbia University College of Dental Medicine) for proofreading the manuscript.
Authors' roles: Study design: HD and SW. Study conduct, data collection, and analysis: YK, JC, AU, MX, and TC. Data interpretation: JC, HD, and SW. Drafting manuscript: JC, HD, and SW. Revising manuscript content: JC, HD, and SW. YK, JC, MX, and SW take responsibility for the integrity of the data analysis.