Estrogen receptor Alpha in human knee articular cartilage of healthy and osteoarthritic females

Abstract

Objective

Estrogen and osteoarthritis association has previously been proposed. This study evaluates the presence of estrogen receptors (ER)-α in articular cartilage, and its association.

Methods

A prospective cohort study of women undergoing anterior cruciate ligament reconstruction (controls) or total knee arthroplasty (cases) was performed. Cartilage samples were processed and ER-α expression was quantified.

Results

Twenty patients were included: 12 cases and 8 controls. ER-α expression was higher in the case subjects.

Conclusion

Increased ER-α expression is associated with age, and degeneration. This suggests estrogen deficiency is a risk for osteoarthritis and is inversely related to proliferative looking chondrocytes.

1. Introduction

Arthritis is the most common cause of disability in the U.S. with a prevalence of 52.5 million adults.¹ Osteoarthritis (OA), the most common, is a multi-factorial arthritic condition² characterized, among others, by articular cartilage breakdown. Hip and knee joints are the joints most commonly affected.^3,4 Current treatments for early to moderate OA primarily target symptomatic relief and rely on anti-inflammatory medications and physical therapy. Although these treatments are somewhat effective at providing pain relief and functional improvement, they are ineffective at preventing initiation or halting the progression of OA. The lack of effective regenerative or disease-modifying therapies is in part due to a gap in knowledge regarding suspected risk factors (e.g., genetics, obesity, and hormonal changes on estrogen, estrogen receptors) and the roles they play in disease onset and progression.

It is well known that women are at higher risk for developing hip and knee OA with marked increases in prevalence and worsening disease severity occurring after menopause.^2,5,6 The relationship of estrogen in degenerative arthritis was originally proposed in 1925,⁷ and remains a focus of further investigation.^8,9 The presence of estrogen receptors have been documented in chondrocytes and within the synovia.^10,11 Estrogen receptors are also very well-recognized in bone cells and play a key role in regulating bone turnover, which is relevant to OA pathophysiology. To the purpose of this study, while it is known that estrogen receptors exist in articular cartilage, the effects of varying blood estrogen concentrations, various life cycle processes, disease states, and drug use on their articular cartilage expression remains unknown.¹² Additionally, there is limited research in humans correlating articular cartilage degenerative changes to the expression of estrogen receptors.¹²

Current literature regarding the pathophysiological mechanism of the relationship between estrogen and OA development is contradictory. Whereas some studies have proposed a chondroprotective estrogenic effect (e.g., inhibition of collagen II catabolism, enhanced glycosaminoglycan synthesis, antioxidative effect, decreased metalloproteinase synthesis, tightly binding receptor polymorphisms),^8,13, 14, 15, 16, 17, 18, 19 others have found a chondrodestructive relationship (e.g., increased chondrocyte apoptosis, inhibition of proteoglycan synthesis, increased pro-inflammatory factors, increased metalloproteinase synthesis).20, 21, 22, 23, 24, 25

Before we can understand better and appropriately treat OA, we must quantify the changes that occur in estrogen and its receptors with respect to varying degrees of disease. In this prospective, controlled cohort study, we hypothesize that increased expression of estrogen receptor Alpha (ER-α) in articular cartilage directly correlates with age, menopausal status, decreasing estradiol levels and increasing severity of OA. The purpose of this study was to document and evaluate the presence of ER-α in articular cartilage, and its association with the severity of OA as determined clinically, radiographically and histologically. In addition, a brief literature review was performed regarding different types of estrogen receptors present in the articular cartilage and the theoretical implications in osteoarthritis.

2. Methods

2.1. Study design and research methods

A prospective, single-institution, controlled cohort study was performed with Colorado Multiple Institutional Review Board approval (COMIRB) 14–1971. Patients were recruited from a consecutive series of individuals seen at the University of Colorado Hospital at Anschutz Medical Campus under informed consent. Inclusion criteria: age ≥18 years-old; female sex; surgery scheduled for primary knee arthroscopic anterior cruciate ligament (ACL) reconstruction for acute ACL tear (controls) or primary total knee arthroplasty (TKA) for severe knee OA (cases). Exclusion criteria: age <18 years-old; revision surgery; pregnancy; current lactation; history of liver disease, connective tissue disease, rheumatoid arthritis (RA), septic arthritis, crystal arthropathies, and other inflammatory poly-arthropathies. Cartilage samples obtained were processed and ER- α expression was quantified as described below.

2.1.1. Study questionnaires

Upon enrollment, the included patients were asked to fill out our Original Patient Questionnaire and the Knee injury and Osteoarthritis Outcome Score Questionnaire (KOOS) prior to collecting blood and articular cartilage samples. Electronic medical records and radiographic images of enrolled patients were analyzed.

The Original Patient Questionnaire consisted of basic demographics (i.e., age, sex, body mass index, and ethnicity/race: Caucasians, Hispanic or Other), medications, history of menarche, menopausal status, use of hormone replacement therapy (HRT), birth control mechanism (e.g., oral contraceptive pills, patch, etc.), oophorectomy, and history of arthritis and joint disease. The KOOS Questionnaire consisted of questions related to knee symptoms, pain, activities of daily living, sport and recreation, and quality of life. The KOOS score is a percentage score ranging from 0 (extreme problems) to 100 (no problems).²⁶

2.1.2. Data and sample collection

Medical history, images, laboratory results, and other data were obtained from the electronic medical records at the University of Colorado Hospital at Anschutz Medical Campus. Records were reviewed by two authors (MLH and FRF) for the following: laboratory tests pertaining to hormone levels and joint health/arthropathies; history inflammatory polyarthropathies; prior and current medication history; history of prior musculoskeletal (MSK) imaging or surgery; cancer and urological history; and major diagnosed medical diseases. De-identified, pre-operative radiographs of the knee obtained from the records were reviewed and graded for OA severity by two blinded orthopedic surgeons at University of Colorado Hospital using the Kellgren-Lawrence (K-L) grading system (grades: 0, none; 1, doubtful; 2, minimal; 3, moderate; 4, severe).^27,28

A venous peripheral blood sample of approximately 5 mL was collected the same day of surgery in a 5 mL heparinized tube and centrifuged at 1800×g for 20 min at 4 °C, followed by storage of plasma (1 ml) at −80 °C until all samples could be assayed for plasma estradiol concentration. Cartilage tissue was removed sharply with use of a blade, curette or grasper for histological analysis during surgery. All samples were labeled with de-identified study ID numbers. In the case of the control subjects (healthy), cartilage tissue was taken from the site where the tunnel would be drilled for anchoring the ACL graft at the femoral notch area. In the case subjects (OA), the cartilage tissue was taken from the femoral notch of the femoral explants. The harvested cartilage was placed and fixed in 4 % paraformaldehyde for 24 h. Samples measuring approximately 0.5–1 cm² were then washed with phosphate buffered saline (PBS) and decalcified in Immunocal for 48 h. Followed by gradual ethanol-based dehydration and paraffin embedding for sectioning.

2.2. Laboratory analyses

2.2.1. Plasma estradiol analysis

The collected plasma was measured for estradiol concentration by the University of Colorado's Clinical and Translational Research Center (CTRC) Core Lab using the Beckman Coulter assay according to manufacturer's instructions.

2.2.2. Immunohistochemistry (IHC) for ER-α

Histological articular cartilage sections 5-μm thick were prepared and stained with standard Hematoxylin-Eosin (H&E) and Safranin-O/fast green protocols. To assess expression of ER-α, 5-μm thick sections were placed on charged slides and coated with Poly-l-Lysine (Sigma, P4707). Endogenous peroxidase activity was blocked using 3 % hydrogen peroxide in 100 % methanol for 10 min. Tissue sections then underwent antigen retrieval in 10 mMol/L citrate buffers (pH 6.0) at 95 °C for 10 min. Sections were then incubated with 10 % normal goat serum in PBS with 1 % Bovine Serum Albumin (BSA, Fisher BP-1600) for 15 min at room temperature (RT), followed by overnight incubation at 4 °C with antibodies against ER-α (Abcam, ab16660, 1:250 dilution). After rinsing with PBS with 1 % Tween, sections were incubated with biotinylated goat anti-rabbit (1:500 dilution) IgG in PBS/BSA for 30 min. Negative controls were prepared by omitting the primary antibody. Sections were then incubated in ABC Elite complex (Vectastain ABC kit, Vector Cat#PK-6100) for 30 min at RT. Staining was detected with Dako DAB substrate solution for 10 min, followed by rinsing in distilled water. The slides were counterstained with Harris-Hematoxylin 2 solution for 2 min, dehydrated in graded ethanol, cleared in xylene, and mounted with Permount medium.

2.2.3. Semi-quantitative analysis of ER-α staining intensities

The intensity of ER-α immunoreactivity was evaluated semi-quantitatively and percentual. Staining was graded (1 through 4; 1 for low, 4 for high) and final grades were determined by taking the mean scores of deidentified samples by two blinded observers (MLH and FRF) and accuracy of grading was then confirmed by the Chair of Pathology at the University of Colorado Hospital at Anschutz Medical Campus. Chondrocytes are quiescent cells which maintain the cartilage architectural homeostasis. From the histological view, they are found in a spectrum range of different states that goes from a resting or senescent to a proliferative (cascade or cluster pattern), to a hypertrophic looking like cell and then apoptosis. For the purpose of this study, we divided chondrocytes in two looking like groups: normal (senescent) and reactive (metabolic, proliferative active).²⁹ The percentage of nuclei staining positive with a grade 3 or 4 for ER-α was determined for both normal (senescent) and reactive (metabolic, proliferative active) looking chondrocytes.

2.3. Statistical analysis

The analysis of variance (ANOVA) was applied including evaluation of the distribution of the studied parameters, mean values and standard deviation for continuous variables, and rates (%) for categorical variables. The Wilcoxon none-parametric test was used to define a statistical significance of differences between the studied groups in continuous variables; the Fisher exact test was used for comparison of the categorical variables. The correlation between the studied continuous variables was assessed by the Pearson's correlation coefficient (r). Impact of potential confounders was assessed by stratification. The JMP 7.0.1 software (SAS Institute Inc.; http://www.jmp.com) was used for analysis. P ≤ 0.05 was considered as a statistically significant.

3. Results

Twenty female patients were included: 12 cases (arthritis) and 8 controls (healthy). The following population characteristics had no intergroup difference: ethnicity, body mass index (BMI), menarche, Tamoxifen use and KOOS score for symptoms. Some population characteristics were different between study groups including: Age, 65.2 years-old (SD, 11.9) in cases, versus 28.9 years-old (SD, 3.6) in controls (P < 0.0001); menopause status, 11 (91.67 %) post-menopause in the cases, versus 0 in the controls; HRT, 50 % of the cases vs 0 % of controls (P = 0.005) (Table 1).

Table 1.

Comparative analysis of the study groups by demographic, radiographic, clinical, and the studied characteristics.

Characteristic		Statistical index	Case (arthritis), N = 12	Control (healthy), N = 8	P-value
Sex	Female	N (%)	12(100)	8(100)	NA
Age	Year	Mean (SD)	65.2 (11.9)	28.9 (3.6)	<0.0001
Ethnicity	Caucasians	N (%)	12(100)	6(75.0)	0.13
	Hispanic	N (%)	0(0)	1(12.5)
	Other	N (%)	0(0)	1(12.5)
BMI	Conventional units	Mean (SD)	30.1(6.8)	25.5(2.7)	0.09
Age at first menstruation	10–12 year	N (%)	8(66.67)	5(62.5)	0.89
	13–14 year	N (%)	2(16.670	2(25.0)
	15–17 year	N (%)	2(15.0)	1(12.5)
Menopause status	Pre-menopause	N (%)	1(8.33)	8(100)	<0.0001
	Post-menopause	N (%)	11(91.67)	0(0)
Hormone replacement	Yes	N (%)	6(50)	0(0)	0.005
	No	N (%)	6(50)	8(100)
Tamoxifen use	Yes	N (%)	1(8.33)	0(0)	0.30
	No	N (%)	11(91.67)	8(100)
Oral contraceptive use	Yes	N (%)	6(50)	7(87.5)	0.072
	No	N (%)	6(50)	1(12.5)
Severity of osteoarthritis by radiographic images (K-L OA score)		Mean (SD)	3.6(0.60)	1.1(0.79)	<0.0001
Estrogen Receptor Alpha (ER-α) expression	% of normal chondrocytes staining positive for ER-α	Mean (SD)	57.9(29.5)	23.3(28.5)	0.006
	% of reactive chondrocytes staining positive for ER-α	Mean (SD)	80.8(21.9)	10(13.2)	0.0002
Plasma Estradiol level	pg/ml	Mean (SD)	24.3(17.2)	150.1(163.6)	0.045
KOOS Sub-Scores (lower number = greater issue) PAIN	Conventional units	Mean (SD)	39.2(16.7)	70.1(18.7)	0.0037
KOOS Sub-Scores (lower number = greater issue) SYMPTOMS	Conventional units	Mean (SD)	48.8(15.9)	65.2(23.6)	0.095
KOOS Sub-Scores (lower number = greater issue) DAILY LIFE ACTIVITIES)	Conventional units	Mean (SD)	47.9(18.3)	78.8(15.4)	0.0026
KOOS Sub-Scores (lower number = greater issue) SPORTS/RECREATION	Conventional units	Mean (SD)	14.6(20.1)	46.3(28.3)	0.0045
KOOS Sub-Scores (lower number = greater issue) QUALITY OF LIFE	Conventional units		22.3(21.4)	41.4(23.1)	0.048

Note: N, number of patients; NA, non-applicable; SD, standard deviation; BMI, body mass index; K-L, Kellgren-Lawrence; OA, osteoarthritis; KOOS, Knee injury and Osteoarthritis Outcome Score Questionnaire.

As was expected given the nature of the grouping, OA severity by K-L grading was higher in the case subjects at 3.6 (SD, 0.6), versus 1.1 (SD, 0.79) (P < 0.0001). KOOS scores were higher in the controls for pain (P = 0.0037), daily life activities (P = 0.0026), sports/recreation (P = 0.0045) and quality of life (P = 0.048). Also, as anticipated, plasma estradiol was decreased in the case subjects, 24.3 pg/ml (SD, 17.2), compared to control subjects, 150.1 pg/ml (SD, 163.6). With respect to our primary question, ER-α expression in normal and reactive chondrocytes was respectively higher, 57.9 (SD, 29.5) and 80.8 (SD, 21.9) in the case subjects, versus 23.3 (SD, 28.5) and 10 (SD, 13.2) in the control subjects (P = 0.006 and P = 0.0002, respectively) (Table 1).

Upon histologic study, the cases displayed fewer glycosaminoglycans and a higher percentage of reactive chondrocytes as compared to the controls. In some cases, fibrocartilage was evident upon H&E staining (Fig. 1, Fig. 2). IHC for ER-α displayed a percentage of positive stain for the normal and reactive chondrocytes types in the cases as compared to the controls. This increase in ER- α staining in case subject cartilage interestingly showed positive correlation with the increased severity of OA (Fig. 2).

Fig. 1.

Representative images of Safranin-O stain for articular cartilage in both groups. Magnification 40x, and scale as stated on the image, 50μm. Case (left), shows less glycosaminoglycans' content (red stain) and less cellularity in cascade like pattern; as compared to the control (right) with more cluster like cell pattern.

Fig. 2.

Representative images of Hematoxylin-Eosin (H&E) and Immunohistochemistry (IHC) for estrogen receptor Alpha in both normal and reactive chondrocytes; and in none (controls), moderate and severe OA (cases). Magnification 40x, and scale as stated on the image, 30μm. None (left column), shows less positive IHC stain for both normal and reactive chondrocytes. Moderate and Severe (middle and right columns), show increasing positivity of estrogen receptor Alpha. Fibrocartilage and low cellularity were evident in some cases, as evident in the right column for normal chondrocytes. In addition, an overall presence of cascade and cluster like orientation were was noticed in the degenerative cartilage.

The correlation (r) of radiographic OA severity to previously described linked risk factors (i.e., age and BMI) and suggested hormonal factors (i.e., ER-α and estradiol)¹² was calculated. OA severity was strongly positively correlated to age (r = 0.78, P < 0.0001), BMI (r = 0.46, P = 0.039) and to the percentage of reactive chondrocytes staining positive for ER-α (r = 0.8, P < 0.0001) (Table 2 and Fig. 2). Plasma estradiol level was moderately negatively correlated to age (r = −0.49, P = 0.026), percentage of reactive chondrocytes staining positive for ER-α (r = −0.44, P = 0.05). Percentage of normal chondrocytes staining positive for ER-α was moderately positively correlated to age (r = 0.49, P = 0.026) and % of reactive chondrocytes staining positive for ER-α (r = 0.59, P = 0.006). Percentage of reactive chondrocytes staining positive for ER-α was weakly positively correlated to age (r = 0.074, P = 0.0002) (Table 2). The KOOS scores were not found to correlate with OA severity nor to the percentage of chondrocytes staining positive.

Table 2.

Correlation (r) of severity of osteoarthritis by the radiographic characteristics with the parameters which likely have linkage with this pathology.

		Severity of osteoarthritis by radiographic images (K-L OA score)	Plasma Estradiol level (pg/ml)	Estrogen Receptor Alpha (ER-α) Expression		Age (year)	BMI
				% of normal chondrocytes staining positive for ER-α	% of reactive chondrocytes staining positive for ER-α
Severity of osteoarthritis by radiographic images (K-L OA score)		1.0	−0.3 (P = 0.2)	0.28 (P = 0.22)	0.8 (P < 0.0001)	0.78 (P < 0.0001)	0.46 (0.039)
Plasma Estradiol level (pg/ml			1.0	−0.18 (P = 0.43)	−0.44 (P = 0.053)	−0.49 (P = 0.026)	−0.06 (0.79)
ER-α expression	% of normal chondrocytes staining positive for ER-alpha			1.0	0.59 (P = 0.006)	0.49 (P = 0.026)	0.25 (P = 0.29)
	% of reactive chondrocytes staining positive for ER-alpha				1.0	0.074 (P = 0.0002)	0.29 (P = 0.21)
Age (year)						1.0	0.24 (P = 0.30)
BMI							1.0

Note: BMI, body mass index; K-L, Kellgren-Lawrence; OA, osteoarthritis; ER-α, estrogen receptor alpha.

When analyzing the impact of potential confounders on radiographic OA severity, post-menopausal status was significantly linked to higher K-L grades (N = 11, K-L 3.6 ± 0.6; P < 0.001) (Table 3). Moreover, post-menopausal status was significantly linked to higher percentage of normal and reactive chondrocytes staining positive for ER-α, respectively, 54.1 ± 27.6 (P = 0.043) and 79.1 ± 22.1 (P = 0.003); whereas estradiol level was unsurprisingly significantly linked to pre-menopausal status (N = 9, 141.6 ± 155.2; P = 0.018) (Table 4).

Table 3.

Impact of other potential confounders on the radiographic severity of osteoarthritis.

Potential confounders	Subgroups	N	Severity of osteoarthritis by radiographic images (K-L OA score)
			Mean	SD	P-value
Menopause status	Pre-menopausal	9	1.4	1.1	<0.001
	Post-menopausal	11	3.6	0.6
Post-menopausal + Hormone replacement	Yes	6	3.7	0.4	0.87
	No	5	3.6	0.89
Tamoxifen use	Yes	1	4.0	NA	0.25
	No	19	2.5	1.4
Use of oral contraceptive	Yes	13	2.3	1.5	0.29
	No	7	3.1	1.2
Age at first menstruation (years)	10–12	13	2.5	1.4	0.72
	13–14	4	2.9	1.3
	15–17	3	2.7	2.3
Ethnicity	Caucasians	18	2.8	1.3	0.17
	Hispanic	1	1.0	NA
	Other	1	1.5	NA

Note: N, number of patients; SD, standard deviation; K-L, Kellgren-Lawrence; OA, osteoarthritis.

Table 4.

Impact of menopause status on the studied estrogen characteristics.

Estrogen characteristics		Menopause status	N	Mean	SD	P-value
Estrogen Receptor Alpha (ER-α) Expression	% of normal chondrocytes staining positive ER-α	Pre-menopausal	9	31.7	36.9	0.043
		Post-menopausal	11	54.1	27.6
	% of reactive chondrocytes staining positive ER-α	Pre-menopausal	9	20.0	32.4	0.003
		Post-menopausal	11	79.1	22.1
Plasma estradiol level (pg/ml)		Pre-menopausal	9	141.6	155.2	0.018
		Post-menopausal	11	19.8	8.1

Note: N, number of patients; SD, standard deviation.

4. Discussion

The purpose of this study was to evaluate the presence of ER-α in articular cartilage and to investigate its association with the severity of OA as determined clinically, radiographically and histologically, taking into consideration potential confounders. We hypothesized that increased expression of ER-α in articular cartilage directly correlates with age, menopausal status, decreasing estradiol levels and increasing severity of OA. As stated above, chondrocytes are quiescent cells which maintain the cartilage architectural homeostasis, and account for 1–5% of total tissue volume.²⁹ From the histological perspective, they are found in different stages. In a dynamic representation, from senescence stage, they can become actively proliferating. These can be found in cascade like and/or a cluster like pattern (Fig. 1, Fig. 2). The end stage, hypertrophic chondrocytes, are enlarged produce mineralized ECM and eventually apoptosis or cell calcification, which leaves lacunar emptying.²⁹ The most important finding was that both, percentage of reactive ER-α positive chondrocytes as well as age, strongly correlate with KL score. This expression was higher in post-menopausal patients for normal and reactive chondrocytes; and was inversely related to estradiol levels. This suggests that increased ER-α activity, proliferative/reactive cells, and decreased plasma estradiol levels are linked to increasing OA severity. The relationship of whether this is a potential cause-for or effect-of increased OA severity is yet to be determined. Seems, there is a cellular proliferation attempt to maintain the homeostasis of the articular hyaline cartilage with an up regulation of ER-α. The epidemiological common denominator suggests that estrogen deficiency is a risk for OA. Hence, HRT, whether systemic or local, could potentially benefit a group of selected patients.30, 31, 32 While it is known that ER-α and ER-β exist in articular cartilage,³³ the effects of varying blood estrogen concentrations, various life cycle processes, disease states, and hormone affecting drug use on the expression of these receptors in articular cartilage remains unknown. Additionally, there is conflicting data in the literature in human subjects correlating degenerative changes within articular joint cartilage to expression of estrogen receptors and variations in plasma estradiol levels.

The synovial tissue is responsible for synovial fluid production, and for the origin of pain and joint inflammation.³⁴ The healthy human synovium has predominantly ER-β.¹¹ Estrogens are known to maintain healthy epithelial surfaces through balanced cellular and secretory homeostasis. Metka et al.³⁵ suggested that menopausal arthralgia could be partly explained by the atrophy and reduced synovial fluid production secondary to decreased estrogen levels.³⁵ This is partially supported by other studies showing that HRT may improve joint symptoms due to its effects on cartilage height,^36,37 joint space^33,38 and risk of OA.^39,40 However, when analyzing the complexity of OA and endocrinological factors, the immune system should be considered. In this direction, Capellino et al.⁸, studying a group of severe-OA, found glucocorticoid receptors expression density to be significantly positively correlated with ER-α, ER-β, and androgen receptors in the synovium (coupled expression). Such coupled expression was affected in RA patients due to the presence of higher pro-inflammatory cytokines secretion (IL-6, Interleukin-6; and IL-8, Interleukin-8) in the synovial fluid.⁸ The local estrogen levels for OA and RA patients were previously reported to be non-different.⁴¹ Hence, the presence of local cytokines (i.e., IL-6 and IL-8) seems to strongly modulate the expression and uncouple such expression, leading to a higher unbalanced pro-inflammatory state in RA.^8,42 The coupled phenomenon expression of glucocorticosteroids receptors and ER in OA suggests a more favorable balance. Although, pathophysiologically different, it sheds light in better understanding steroid receptor and cytokine dynamics, as well as the role they play.

While androgens and glucocorticosteroids exert an anti-inflammatory effect; the behavior of estrogens seems dual.^43,44 Schmidt et al.⁴⁵, found that estrone sulfate transporters in the synovial tissue allowed the passage and downstream conversion of estrone (main estrogen in post-menopause) and estradiol to hydroxylated(OH) isoforms. While the 16-α–OH–estrone and 16-α–OH–estradiol are pro-inflammatory and have no inhibitory effect on tumor necrosis factor (TNF), the 2- and 4-OH-estrogens exert an anti-inflammatory effect through TNF inhibition in patients with OA.⁴⁵ An increase in 16-α–OH–estrogens relative to 2- and 4-OH-estrogens leads to a pro-inflammatory state. Estradiol administration as HRT can rapidly increase estrone sulfate, leading to a pro-inflammatory state. Such dynamics are different in pre-menopausal women due to cyclical estradiol levels.⁴⁴ This adds controversy to previous favorable HRT outcomes.^30,36, 37, 38, 39, 40

The same concentration dependent dual phenomenon was seen in other in vitro and in vivo studies. In these cases, estradiol was beneficial at physiological doses or detrimental at supraphysiological doses.³³ Maneix et al. found estradiol to enhance glycosaminoglycans (GAGs) synthesis through ER-α in cultured rabbit chondrocytes.¹³ Another study found that ovariectomized rats had an accelerated collagen II degradation, and could be countered with early estrogen therapy.¹⁴ Additional studies found estradiol to impair cyclooxygenase 2 expression in bovine chondrocytes which is suspected to protect the chondrocytes from oxidative damage.^15,16 Moreover, another study found estradiol to also inhibit the synthesis of metalloproteinases (MMPs).¹⁷ However, Ye et al. found supraphysiological estradiol doses to accelerate temporomandibular OA in a rat model through collagen II degradation, loss of GAGs and chondrocyte apoptosis.²⁰

In addition to the varying levels and effects of estrogen metabolites, the presence of ER-α polymorphisms may also lead to favorable or unfavorable outcomes and can even differ between ethnic groups.²² Specifically, Pvu II, was found in Caucasians to be inversely correlated with OA.⁴⁶ However, Pvu II and Xbal were linked to severe OA47, 48, 49; and another study found no association.⁵⁰ Notwithstanding, recent recognition of another receptor, ER-γ, was found as a novel catabolic regulator in OA, due to enhanced MMPs synthesis and collagen II and aggrecan degradation.²¹

Although polymorphism was not investigated in the present study, genetics are fundamental in better understanding OA for future patients. Clinically, some patients taking HRT report improvement on their joint symptoms, while others show no improvement or progression of their disease symptoms. This could be partly explained by the presence of ER upregulation (α, β and γ) and polymorphisms within the post-menopausal population that respond differently to estrogens. Therefore, in the future, analyzing OA patient's ER polymorphisms could possibly aid clinicians in determining which patients could benefit from HRT.

This study acknowledges some limitations. It is limited by a small sample size, the inability to age-match controls to cases, yet this is due to the nature of the study design as well and the cartilage samples were taken from the same anatomical area. In addition, OA pain is also strongly linked with subchondral bone remodeling, a site where ER-α could also play a significant role. Evaluation of bone and subchondral bone would have been of interest and may have provided some valuable information if similar changes in OA versus ACL patients was not seen in bone as it is presented in cartilage.

5. Conclusion

Estrogens possibly have an immunomodulatory and anabolic effect on pre-menopausal articular cartilage. Homeostasis in the maintenance of healthy cartilage tissue is lost post-menopause. This is possibly due to the decreasing estradiol and increasing estrone levels. Although the upregulation in ER-α is likely a compensatory response to decreasing estradiol levels in the post-menopausal state, our study is too small to determine whether upregulation of ER-α in the joint space could be occurring prior to menopausal changes. It is thus possible that ER-α upregulation is detrimental to the cartilage due to the higher passage and downstream hydroxylation of estrone. As discussed, the presence of ER-γ and/or the presence of other unfavorable polymorphisms may also have an impact on cartilage homeostasis and further investigation is warranted.

Author contributions

-Marissa Hughbanks, M.D.: study conception and design, IRB application drafting and adherence, patient consenting, tissue collection and processing, data analysis and interpretation, manuscript drafting and editing, final approval of the article.

-Francisco Rodriguez-Fontan, M.D.: patient consenting, tissue collection and processing, data analysis and interpretation, drafting, editing, final approval of the article.

-Christopher J. Kleck M.D.: Study conception and design, manuscript drafting and editing, final approval of the article.

-Evalina Burger-Van der Walt M.D.: Study conception and design, manuscript drafting and editing, obtaining funding, final approval of the article.

Funding

This work was supported by Anschutz Foundation Grant to the University of Colorado Spine Center's research fund. In addition, The Colorado Clinical and Translational Sciences Institute and Clinical and Translational Research Center Core Lab that performed plasma analysis was supported by NIH/NCATS Colorado CTSA Grant Number UL1 TR002535 and UL1 TR001082. Contents are the authors' sole responsibility and do not necessarily represent official NIH views.

Declaration of competing interest

None.