Relationship of Relaxin Hormone and Thumb Carpometacarpal Joint Arthritis

Abstract

Background

The female predominance in thumb carpometacarpal (CMC) joint arthritis has led to speculation that reproductive hormones or hypermobility are responsible. Evidence shows that patients with pathologic laxity have a higher rate of thumb CMC arthritis. Relaxin hormone increases laxity in the pelvic ligaments through upregulation of matrix metalloproteases (MMPs). It is thus a hormone of interest in the development of thumb CMC arthritis.

Questions/purposes

Our goals were to identify demographic and hormonal factors associated with joint laxity in patients with CMC arthritis and to evaluate the relationship among serum relaxin, relaxin receptors, and MMPs in the anterior oblique ligament (AOL) of the thumb. We hypothesized that serum relaxin was correlated with joint laxity as well as with relaxin receptors and MMPs in the AOL.

Methods

Forty-nine patients undergoing thumb CMC arthroplasty underwent laxity examination, blood draw, and AOL sampling. Ligaments were analyzed for relaxin receptor and MMPs 1 and 3 using quantitative reverse-transcriptase polymerase chain reaction.

Results

Women demonstrated more joint laxity than men (p < 0.001). RNA analysis confirmed relaxin receptors in the AOL as well as MMPs 1 and 3. There was a significant correlation between serum relaxin and MMP-1 (p = 0.04). Detectable serum relaxin was negatively correlated with relaxin receptors in the AOL (p = 0.02).

Conclusions

Further studies are needed to evaluate the role of laxity and sex hormones in thumb CMC arthritis.

Clinical Relevance

Relaxin hormone may play a role in the development of arthritis at the thumb CMC joint.

Level of Evidence

Level I, prognostic study. See Guidelines for Authors for a complete description of levels of evidence.

Introduction

Among the general population, the prevalence of radiographic thumb carpometacarpal (CMC) arthritis ranges from 7% to 15% with symptomatic thumb CMC arthritis having a prevalence of 2% to 6% among elderly white patients [19]. The female predominance seen in both symptomatic [30] and radiographic [40] thumb CMC arthritis has promoted speculation that this disparity is secondary to differences in joint laxity and levels of reproductive hormones. Sodha et al. [40] noted a female-to-male ratio of 6:1 for radiographic thumb CMC arthritis in a series of 615 patient radiographs.

Joint hypermobility is defined as greater than average motion at multiple joints [2]. In patients with increased joint laxity, studies have shown a higher correlation with anterior cruciate ligament (ACL) rupture, shoulder instability [4], and ankle sprain [17]. In addition, laxity is associated with a higher rate of knee arthritis [38], implying that greater joint mobility leads to abnormal biomechanical stresses on the joint. Jonsson et al. [22] noted a higher prevalence of thumb CMC arthritis in Icelandic patients with joint laxity compared with those without hypermobility. However, Kraus et al. reported that joint hypermobility was associated with a lower incidence of hand arthritis, particularly in the proximal interphalangeal joints in a large cohort [25].

The differences in the prevalence of multiple musculoskeletal diseases between men and women have led to theories that reproductive hormone variants might account for these disparities. More women than men sustain ACL tears in soccer and basketball [35]. Cooley et al. [11] noted a higher rate of overall hand arthritis in women with a greater lifetime exposure to reproductive hormones, defined as an earlier onset of menarche or later onset of menopause. In studies of the impact of hormones in these joints, it has been shown that ACL tears occur most frequently during the ovulatory phase of menstruation in women [49]. Relaxin, a hormone produced by the corpus luteum in pregnancy that loosens the pelvic ligaments in preparation for childbirth, has been proposed as a specific target hormone for the ligament injury and development of arthritis as a result of attenuation of the supporting joint ligaments [45]. Relaxin is a member of the insulin superfamily, produced by the uterine endometrium in nonpregnant women [15, 27] and in the prostate in men [44] and has been demonstrated in postmenopausal women [3] as well. As a potential mediator of laxity in the stabilizing ligaments of the thumb CMC joint, relaxin has been shown to be present in both males and females in a circulating form [31] and a previous study showed receptors present in the anterior oblique ligaments of women [26]. With the theory that relaxin has an impact on the thumb ligaments, we wanted to investigate the potential downstream impacts of relaxin at the molecular level, specifically the relationship of relaxin and matrix metalloproteases (MMPs) in the thumb ligaments. Relaxin exerts its effect through upregulation of MMPs, which are enzymes that degrade the extracellular matrix [41]. These include collagenase (MMP-1), which breaks down collagen, stromelysin (MMP-3), and others [24]. Estrogen and relaxin receptors have been described in the ACL [16], and preliminary studies also showed estrogen receptors at the thumb volar beak ligament [33]. Kapila et al. [24] noted that estrogen and relaxin caused dose-dependent matrix degradation of temporomandibular fibrocartilage explants.

To perform activities of daily living, the human thumb has evolved to combine mobility and stability. Generalized laxity has been shown to translate to laxity in the ligaments surrounding the thumb CMC joint [47]. These ligaments are potentially affected by their exposure to different levels of reproductive hormones in their environment. With evidence that relaxin has an effect on the thumb ligaments, we wanted to investigate the potential downstream impacts of relaxin at the molecular level, specifically whether relaxin affected MMPs in the thumb ligaments.

We therefore hypothesized that there is a relationship between circulating relaxin and generalized joint laxity as well as relaxin receptors in the anterior oblique ligaments as well as between the molecules upregulated by relaxin, the MMPs. The purposes of this study were to identify demographic and hormonal factors associated with joint laxity in patients with CMC arthritis and to examine the correlation among serum relaxin, relaxin receptors, and MMPs in the anterior oblique ligament of the thumb CMC joint in patients with arthritis.

Patients and Methods

We obtained institutional review board and research committee approval before initiating this study. Forty-nine consecutive patients were prospectively enrolled before undergoing surgical treatment for thumb CMC joint arthritis by one of two fellowship-trained hand surgeons (JW, FS) at two institutions, one of which was a Veterans Administration medical center. Forty-eight of 49 underwent trapeziectomy with ligament reconstruction and tendon interposition [8] and one underwent ligament reconstruction for persistent pain and laxity, as described by Eaton and Littler [14]. Exclusion criteria included inflammatory arthritis, immunologic disease, and concurrent surgical procedures. Each subject was evaluated for generalized joint laxity using the Beighton-Horan index [5] (Table 1), underwent venipuncture before surgery for analysis of serum relaxin levels, and their most recent preoperative radiographs were evaluated by a fellowship-trained musculoskeletal radiologist (BP) and rated on the Eaton scale [14]. Blood draws were performed in the morning when possible to exclude variability resulting from the diurnal variation of relaxin secretion [7]. Radiographs were available in 47 of 49 subjects, and those without these data were not included in analysis or correlations using Eaton scores. The study cohort consisted of 49 patients (30 females, 19 males) with an average age of 62 years (range, 43–78 years). Of the 30 women (age range, 43–77 years), 27 were postmenopausal and two were being treated with hormone replacement. The 19 men were between 58 and 78 years of age, and none had a history of prostate cancer or thyroid disease.

Table 1.

The Beighton-Horan score: an assessment of joint mobility*

Joint measurement	Points given (maximum score of 9)
Passive extension of the small finger > 90°	1 for right hand 1 for left hand
Ability to passively appose the thumb to the forearm with the wrist flexed	1 for right side 1 for left side
Hyperextension of elbow > 10°	1 for right side 1 for left side
Hyperextension of the knee > 10°	1 for right side 1 for left side
Ability to place the hands flat on the floor when flexed from the waist with straight knees	1 point

* Beighton P, Horan F. Orthopaedic aspects of the Ehlers-Danlos syndrome. J Bone Joint Surg Br. 1969;51:444–447.

Measurement of serum relaxin was performed using a Quantikine Human Relaxin-2 Immunoassay kit (R&D Systems, Inc, Minneapolis, MN, USA), which used a monoclonal antibody with a mean minimum detectable dose of 1.0 pg/mL (range, 0.26–4.57 pg/mL). Relaxin-2 is the primary circulating form of relaxin in humans [39]. Each assay was performed in duplicate. Results were analyzed at 450 nm wavelength using a microplate reader (ThermoElectron Multiskan Plus, San Jose, CA, USA). Linear regression of eight serially diluted calibrators, 0 to 500 pg/mL, was used to determine the concentration of relaxin-2, the predominant circulating form of relaxin in humans, in each serum sample.

Surgical specimens of the anterior oblique ligament were evaluated for human relaxin receptor (RXFP-1), MMP-1, and MMP-3 using quantitative real-time reverse-transcriptase polymerase chain reaction (qRT-PCR). On entering the thumb CMC joint, the deep anterior oblique ligament was identified deep to the capsule on the volar ulnar aspect of the joint with visualization of ligamentous fibers used as confirmation and a 0.5- to 1-cm sample harvested. Each ligament was snap-frozen in liquid nitrogen and stored at −80 °C. Next, the ligaments were batch-powdered and RNA was extracted and purified (RNeasy Lipid Tissue Mini Kit; QIAGEN, Valencia, CA, USA). Reverse transcription on 1 μg total RNA was used to prepare complementary DNA (cDNA) with a HC cDNA RT kit with RNase inhibitor (Applied Biosystems, Foster City, CA, USA). Expression of the genes of interest was analyzed with qRT-PCR using the TaqMan system and the ABI prism 7500 sequence detector (Applied Biosystems, Foster City, CA, USA) and reported as a ratio using GAPDH as the internal control. Samples were run in duplicate. Primers and probes for MMP-1, MMP-3, and relaxin receptor-1 were designed using specialized sequencing software (Primer Express; Applied Biosystems) (Table 2).

Table 2.

Accession number and sequences for the analyzed genes

Gene name Accession number	Sequence type	Sequence	Optimized molar concentration (nm)
GAPDH NM_002046.3	Forward	GGG CTG GCA TTG CCC TC	900
	Reverse	TTG CTG TAG CCA AAT TCG TTG TCA TA	900
	Probe	VICACG ACC ACT TTG TCA AGC TCA TTT CCT GTAMRA	200
MMP-1 NM_002421.3	Forward	CCT CGC TGG GAG CAA ACA	900
	Reverse	TTG GCA AAT CTG GCG TGT AA	900
	Probe	6FAMCTG ACC TAC AGG ATT GAAMGBNFQ	200
MMP-3 NM_002422.3	Forward	CAA GCC CAG GTG TGG AGT TC	900
	Reverse	CGG GAT GCC AGG AAA GGT	900
	Probe	6FAMTGA TGT TGG TCA CTT CAGMGBNFQ	200
RXFP-1 NM_021634.2	Forward	TTT GCC AAG GTC TGG AGC TT	300
	Reference	ATT TGA AGA AAC CGA TGG AAC AG	900
	Probe	6FAMACT GTG ATG AAA CCA ATT TMGBNFQ	200

An a priori power analysis was performed. We assumed that the majority of surgical patients were older than 60 years old. The analysis was based on the enrollment of both men and women into this cohort. To ensure that a sufficient number of subjects would be enrolled to assess the association between relaxin levels and ligament receptor levels, the analysis was based on a medium effect size (r = 0.30) [9]. To detect r = 0.30, given alpha = 0.05 and power = 0.80, approximately 52 subjects were required.

Descriptive summaries, bivariate analysis using the t-test for normally distributed data and the Mann-Whitney test for nonparametric data, correlation analysis using Pearson’s rho for normally distributed data and Spearman’s rho for nonparametric data as well as multivariate regression were performed using SPSS (released 2010, IBM SPSS Statistics for Windows, Version 19.0; IBM Corp, Armonk, NY, USA). Log transformations were used for interval data that were not normally distributed. Statistical significance was set at p < 0.05.

Results

Analysis of serum relaxin showed there was no difference between men and women in mean relaxin levels (4.03 versus 3.29 pg/mL, p = 0.44). Women demonstrated significantly greater generalized joint laxity, based on the Beighton score, than men (3.32 versus 0.35, p < 0.001). There was no correlation between serum relaxin level and generalized joint laxity (p = 0.72). Thumb CMC arthritis in this series was more advanced in the male patients than in the female patients (3.56 versus 3.07, p = 0.048) (Table 3).

Table 3.

Summary of overall means and mean differences by sex

Factor	Mean ± SD (number)	Female, mean ± SD (number)	Male, mean ± SD (number)	p value
Age (years)	61.76 ± 8.72 (50)	59.83 ± 9.33 (30)	64.65 ± 6.96 (19)	0.055*
Serum relaxin (pg/mL)	3.73 ± 3.10 (49)	4.03 ± 3.12 (30)	3.29 ± 3.10 (19)	0.436†
Total comorbid conditions§	2.52 ± 1.59 (50)	2.60 ± 1.69 (30)	2.40 ± 1.47 (19)	0.668*
VAS pain score	5.48 ± 2.57 (50)	5.23 ± 2.13 (30)	5.85 ± 3.14 (19)	0.448*
Beighton score	2.08 ± 2.22 (48)	3.32 ± 2.11 (28)	0.35 ± 0.67 (19)	< 0.001†
Eaton staging	3.26 ± 0.77 (47)	3.07 ± 0.84 (29)	3.56 ± 0.51 (18)	0.048†
MMP-1	0.022 ± 0.037 (50)	0.026 ± 0.043 (30)	0.017 ± 0.027 (19)	0.601‡
MMP-3	0.318 ± 0.724 (50)	0.389 ± 0.926 (30)	0.213 ± 0.155 (19)	0.901‡
RXFP relaxin	0.00052 ± 0.00051 (50)	0.00044 ± 0.00030 (30)	0.00065 ± 0.00070 (19)	0.287‡

* t-test comparing means for normally distributed data; ^†Mann-Whitney test comparing medians for nonparametric distribution; ^‡t-test comparing means performed on log transformation; ^§comorbid conditions included hypertension, depression, lung disease, thyroid disease, regular tobacco use, gastroesophageal reflux or other gastrointestinal disease, and vascular problems; VAS = visual analog scale; MMP = matrix metalloprotease.

RNA analysis confirmed relaxin receptor transcription as well as MMP-1 and MMP-3 message using quantitative RT-PCR in the anterior oblique ligament (Fig. 1). With the numbers available, there were no differences between males and females in quantitative relaxin receptor transcript (0.00052 ag versus 0.00044 ag, p = 0.29) or in MMP-1 (0.01702 versus 0.02622, p = 0.60) and MMP-3 (0.21279 versus 0.38868, p = 0.90) amounts in the anterior oblique ligament.

Fig. 1.

Relaxin receptor, MMP-1, and MMP-3 expression shown from qRT-PCR.

Transcription of both MMPs 1 and 3 likewise was noted in the anterior oblique ligament. A multivariate regression model was used to identify predictors of MMP-1. Because MMP-1 was not normally distributed in the sample, a model was developed using the log of MMP-1 (p = 0.024). After controlling for the number of comorbid conditions and sex, a significant positive association was found for serum relaxin (p = 0.039) and Eaton score (p = 0.017). Additionally, in a separate analysis of patients with detectable relaxin levels (n = 32), there was a significant negative correlation (Pearson R = −0.41, p = 0.019) between serum relaxin and the log of the relaxin receptor, RXFP-1 (Table 4).

Table 4.

Variable correlations with serum relaxin and Eaton score

Variable	Serum relaxin correlations (Pearson coefficient)		Eaton score and means comparison by analysis of variance
	R value	p value	1 (n = 1), mean (SD)	2 (n = 6), mean (SD)	3 (n = 20), mean (SD)	4 (n = 20), mean (SD)	p value
MMP-1 log	0.269	0.061	−5.26718	−5.70012 (1.964283)	−4.76018 (1.882068)	−5.00459 (1.539675)	0.716
MMP-3 log	0.136	0.353	−3.76329	−0.95774 (1.561555)	−2.0236 (0.83105)	−1.86238 (0.914764)	0.035
RXFP1 log	−0.147	0.314	−8.70266	−7.73843 (0.631929)	−7.91814 (0.707693)	−7.83225 (0.66887)	0.604

MMP = matrix metalloprotease.

Discussion

The purposes of this study were to identify demographic and hormonal factors associated with joint laxity in patients with thumb CMC arthritis and to examine the correlation among serum relaxin, relaxin receptors, and MMPs in the anterior oblique ligament of the thumb CMC joint in patients with arthritis. Relaxin hormone was posited as a target hormone causing thumb CMC laxity, leading to osteoarthritis, and our goal was to characterize its systemic presence as well as potential impact at the thumb ligaments. The extent of joint laxity and its impact on musculoskeletal disease is not clear, because laxity varies by sex and likely by joint as well.

The limitations of this study include a small group of subjects and lack of sex-matching for exact comparisons. Although we identified female subjects’ hormonal status and use of supplemental hormones, we do not know the impact of these factors on circulating relaxin. We were able to measure serum relaxin in generally low levels, and not all subjects had measurable relaxin levels. Dragoo et al. noted detectable relaxin levels in 46 of 128 female athletes (36%) with an average relaxin level of 5.7 pg/mL in those without ACL tear and 12.1 pg/mL in those with tears [13]. Physiologic levels of circulating relaxin have been defined based on the pregnant female, but there is limited information on normal levels in postmenopausal women or older men, who made up the majority of our cohort. In a large study of volunteers [48], we noted detectable relaxin levels in 121 of 287 subjects (42%) with an average overall serum relaxin level of 1.8 pg/mL with average detectable relaxin levels of 2.4 pg/mL in women 60 years and older with 2.6 pg/mL in the same age group in men. It is interesting to note that serum relaxin was higher on average in our group with surgically treated arthritis than in a similar age group of volunteers. Finally, we do not have information about the distribution of relaxin receptors in the ligament or comparisons to nonarthritic control tissues.

Based on studies in subjects with extremes of joint laxity [18] as well as normative populations, there is evidence that joint laxity impacts the thumb CMC joint and leads to arthritis. In a study of asymptomatic blood bank donors, 5% had benign hypermobility [21], and among 550 university students, Didia et al. showed a 17% prevalence of hypermobility features among females compared with 8% in males [12]. In patients diagnosed with joint hypermobility, studies have shown a higher correlation with ACL tear [28] and shoulder instability [4]. Joint hypermobility has also been associated with a disturbance in proprioception [20]. The combination of decreased proprioception in addition to abnormal biomechanical stresses believed to result from increased joint mobility [38] is theorized to contribute to the development of arthritis. There is some limited evidence that increased knee laxity predisposes to knee arthritis with greater degrees of laxity seen in women based on a study of varus-valgus laxity in 164 patients with knee arthritis [38]. One difficulty in evaluating the impact of laxity on arthritic development is that the joints measured by the Beighton-Horan score have overall decreased mobility with age, and there is no alternative to measure joint mobility in older persons.

Pellegrini et al. and others postulated that the development of thumb CMC arthritis was the result of attenuation and laxity of the supporting ligaments of the joint [10, 34]. We have previously shown that radiographic laxity at the thumb CMC joint is correlated with generalized joint laxity with a significantly higher degree of both generalized laxity and radiographic laxity in women [47]. In the present study, we also noted a significant difference in joint laxity between men and women but were unable to show a correlation between serum relaxin and laxity in a group of older adults undergoing surgery for thumb arthritis. Additionally, although women had slightly higher circulating relaxin levels than men, this difference was not significant. These findings are consistent with previous studies [1, 46] although the current population is far older than those studied previously.

Using RT-PCR, we demonstrated quantitative transcription of relaxin receptors in the thumb anterior oblique ligament, adding to the data of Lubahn et al. [26] who reported immunohistochemical staining for receptors previously. Our study differed in that we enrolled men as well as women and interestingly noted no differences in qRT-PCR transcript amounts between sexes. We noted a negative correlation between the level of serum relaxin and the quantitative amount of relaxin transcript, suggesting that relaxin downregulates its receptor. This mechanism of action of relaxin on its receptor has been previously shown in rat models [32]. Because we were unable to test equivalent control tissue in young, nonarthritic individuals, we do not know whether the relaxin receptors are dynamically regulated during reproductive years or continue to be activated during the development of arthritis. In addition, in dermal fibroblasts, relaxin directly decreases collagen synthesis and secretion [43], indicating that relaxin may attenuate ligaments directly by decreasing their ability to produce collagen.

In this study, we evaluated transcription of MMPs 1 and 3 in the thumb ligament and showed transcript levels at one and two magnitudes higher, respectively, than relaxin expression. The effect of relaxin to increase expression of MMPs has been shown previously in vitro in fibrochondrocytes cultured from the temporomandibular joint [23, 29]. We noted a correlation between serum relaxin and MMP-1, suggesting a potential connection with binding to relaxin receptors in the ligament with subsequent upregulation of MMP-1 transcription. Advanced Eaton stage of arthritis was correlated with MMP-3 levels, suggesting that MMP-3 may be contributory to ligament attenuation in worsened arthritis. However, the correlation between MMP-3 expression in the ligaments and the Eaton stage of arthritis may be confounded by the fact that most of our patients undergoing surgery had an advanced degree of arthritis with fewer at earlier stages for statistical comparison. There is some preliminary evidence to suggest that circulating MMP-3 is a biomarker for severity in early-stage osteoarthritis along with MMP-9, n-kappaβ ligand (RANKL), and nitric oxide. Additionally, Brophy et al. [6] showed increased levels of multiple MMPs and other catabolic markers in the menisci of patients with ACL and meniscal tears under 40 years of age, suggesting increased risk of the development of arthritis in this age group with injury.

A model of relaxin knockout has been established in the mouse, and this phenotype shows progressive fibrosis with increased interstitial collagen deposition in the kidney, lung, and liver [37]. This model suggests that relaxin is a naturally occurring inhibitor of collagen deposition [36]. Thus, relaxin, through upregulation of MMPs [42], may act to loosen the stabilizing ligaments of the thumb CMC joint and prevent reparative processes through a direct decrease in collagen synthesis once these ligaments are damaged.

The findings of the present study suggest that relaxin may play a role in ligament attenuation in the genesis of thumb CMC joint arthritis. If this is confirmed, relaxin may be a potential target for intraarticular blockade to prevent or lessen the effects of thumb CMC joint osteoarthritis.