Correlation of Synovial Tissues Protein Abundance With Age and Menopause in Osteoarthritis and Knee Joint Injury of Female Patients

Knee injury and osteoarthritis (OA) are common articular disorders that affect numerous people of different ages, particularly menopausal women. Epidemiological surveys conclude that ladies gain a higher prevalence of OA than men, the menopausal women even get higher morbidity. However, studies about OA animal models suggest that the animals could remain a good joint function regardless of aging and menopause; age and menopause seem to have little effect on joint function. Thus, the study aimed to explore whether age and menopause will inuence the clinical recovery of a knee joint in OA and injury women and the relevant mechanism by analyzing the proteome prole of synovial tissue (ST) samples. of aging and menopause on the knee joint and throw light on the mechanism through which ST might affect the functions of the knee joint.


Background
Osteoarthritis (OA) is a common complaint in young and old people, causing the patients many inconveniences in daily life and work. Therefore, epidemiological studies and controlled trials are carried out by many researchers. Based on the previous surveys, age, estrogen, sex, dietary intake et al are important risk factors for the initiation and development of OA [1,2,3,4,5]. Besides, menopause is also suggested to be associated with the onset and progression of OA in women [6]. The involvement of all of the above risk factors in the progression of osteoarthritis requires further research.
Synovial tissues (ST) have also been shown to be involved in remaining the joint function and the incidences of many joint diseases [7]. Essential components expressed in ST are of great signi cance in regulating the functions of the knee joint and the responses to multifarious joint disorders. For example, the prolactin receptor (PRLR), expressed by macrophages, is detected in ST samples from both OA and rheumatoid arthritis (RA) patients, stimulating the activation of macrophages with other pro-in ammatory cytokines [8, 9,10]. β1,4-galactosyltransferase-I (β1,4-GalT-I) was also reported to be elevated in ST of OA patients [11]. However, many other proteins that might also contribute to the incidence and the development of OA in ST are not fully identi ed. Therefore, in the study, we purpose to explore the association of aging, menopause and OA by characterized ST proteomic expression patterns together with clinical outcomes.

Participants
Synovial tissue samples were obtained from female patients who suffered from knee OA or knee injury and underwent surgical therapy from Nov. of 2018 to April of 2019 in Sichuan Provincial People's Hospital. To be speci c, patients who tted the following criterion were included: (1) female patients with knee OA or knee injury, (2) aged at 20~70y, (3) without other maladies of bone and articular, or any pre-therapy history of the knee OA or knee injury, (4) signed the study consent and was willing to be follow-up. Although 21 patients were included, only samples from 8 patients were applied for proteomic analysis regarding their comparability.
Detection of blood plasma hormone levels The peripheral blood used for estrogen testing was obtained on the morning of the operative day by nurses and was sent to the Hospital's testing center immediately.

Protein preparation and digestion
The synovial tissue (ST) samples were obtained during the surgery and immediately stored in liquid nitrogen. Protein extraction was performed according to the requirements of MS analysis. Brie y, 20mg of ST sample was ground and sonicated (Q800R sonicator, QSONICA, Newton, Connecticut, USA) in 1ml lysis buffer (7M urea, 4% SDS, 1mM phenylmethane sulfonyl uoride (PMSF), 2mM EDTA, 10mM dithiothreitol (DTT), 1×protease inhibitor cocktail (Sigma-Aldrich, St. Louis, USA) in 30mM HEPES, pH 7.4). The insoluble content in the lysate was discarded by centrifugation at 13000g and 4 °C for 10min. The protein concentration of the supernatant was determined with a BCA Protein Assay Kit (Thermo Fisher Scienti c, Bonn, Germany) and the quality of the samples was evaluated via the SDS-PAGE method. The samples were then precipitated with 15% trichloroacetic acid for 2 h at 4 °C and washed with acetone. The precipitates were redissolved in triethylammonium bicarbonate buffer. The solutions were then digested with trypsin (Promega, Madison, WI, USA) at an enzyme/ substrate ratio of 1:50 at 37 °C for 16h.
TMT labeling, HPLC fractionation The digested peptide mixtures were rst desalted using a Sep-Pak C18 column (100mg, 1cm 3 ; Waters Corporation, Milford, MA, USA). TMT labeling was performed according to the manufacturer's instructions. Brie y, anhydrous acetonitrile was mixed with TMT Reagents preequilibrated to room temperature. The mixture was added to each sample, and the puri ed samples were chemically labeled with the TMT10plex™ Isobaric Label Reagent Set (Thermo Scienti c, Rockford, CA, USA) for 2h at RT. The samples were then pooled and dried by vacuum centrifugation. The dried samples were acidi ed with 0.1% tri uoroacetic acid and desalted using Sep-Pak C18 (100mg, 1cm 3 ; Waters Corporation, Milford, MA, USA) and fractionated through a BEH C18 column (2.1×150mm, 1.7μm, 130Å) using ACQUITY UPLC® H-Class System (Waters Corporation, Eschborn, Germany) at a ow rate of 250μl/min.

Evaluation of clinical outcomes
Just the day before and one month after the surgery, joint functions of patients were evaluated by scores of Visual Analogue Score (VAS), Lysholm, The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). All the evaluations were conducted by the same group of doctors.

The analysis of reactions and pathways among proteins
The immune or endocrine system-related proteins which were insigni cant abundances between old OA patients and young OA patients were selected out by the ClueGO tool of Cytoscape (v3.7.2) and were analyzed with Reactions and Pathways of Reactome program provided online (https://reactome.org/)

Statistical analysis
The proteomic raw data was initially processed with Proteome Discovery 1.4 (Thermo Fisher Scienti c, Waltham, MA, USA). KOBAS 3.0 (http://kobas.cbi.pku.edu.cn/) [12], a web server for gene/protein functional annotation and functional gene set enrichment, was used for pathway and Gene Ontology functional category enrichment analysis. The false discovery rate (FDR) was calculated by the method described by Benjamin. Signi cant results were determined based on FDR-adjusted p-value ≤ 0.05. SPSS version 17.0 (SPSS Inc., Chicago, IL) was used for the data analysis, and P value < 0.05 was considered as statistically signi cant. The data of BMI, Mean age, VAS, Lysholm, WOMAC scores were shown as the mean ± standard deviation. The student's t-test was applied for comparing the means of two independent groups and one-way ANOVA analysis was used to compare the means among three groups. Pearson's correlation and linear regression analysis were used to evaluate the correlation between the concentration of estrogen and the VAS, Lysholm as well as WOMAC scores.
The effects of aging and menopause on the clinical outcomes of knee OA and injury women The basic information of the participants whose ST samples were analyzed as well as the estrogen levels, the pre-surgery and post-surgery evaluated scores were summarized in Table 1. Aging, menopause, or estrogen levels were essential factors that would result in a change of postsurgery function outcomes. Even though all participants got a better score after a month of operative therapy, young patients exhibited more agreeable outcomes compared to the aging ones regardless of menopause and disease types (FDR<0.05); on the other hand, premenopausal ladies also enjoyed more satisfying function outcomes than menopausal ones (FDR<0.05). However, no differences in post-surgery function recovery were detected comparing women with higher levels of estrogen to women whose estrogen levels were much lower.
Characteristics of the proteome pro le of ST from knee OA and knee injury female patients Around 4000 proteins were detected in the proteome and nally, 3465 proteins were used for the analysis after quantifying (FDR<0.01). The proteins were composed of different cellular components and were involved in various biological processes (Fig.1), according to GO and KEGG analysis.

Proteins pro le in ST from OA and knee injury female patients
Our results displayed that ST protein expression patterns were affected by age, menopause, and disease types which were grouped into knee OA group and knee injury group. The average counts of the top 20 candidate proteins which would be most affected by age, menopause, or disease types estrogen levels were listed in Table 2.
To better understand whether the protein pro les were affected by estrogen, the estrogen levels were also tested. The differently expressed proteins were as shown in (FDR<0.01) Table 2. The relationship between estrogen levels and protein pro les was identi ed by Pearson correlation analysis. Yet, no signi cant correlation in protein expression patterns and estrogen levels was detected.
The protein abundances of ST were affected by aging. The protein pro les varied: some proteins were upregulated and the expression levels of some proteins decreased from ST in young female OA and knee injury patients while compared to aging participants ( Fig.2a and b).
The protein pro les were also in uenced by menopause. In comparisons of menopausal knee OA and knee injury patients to the premenopausal ones, some protein levels such as DCTN5, AKR7A2, and NHP2 were signi cantly increased, while proteins like HDGFL3, MYCBP2 and PRS15 were markedly decreased in ST samples (Fig.2c).
Interestingly, the protein pro les of ST from female knee OA patients also differed from those of knee injury women (Fig.2d). Female knee OA patients exhibited a higher level of proteins such as TRADD, CTSZ and LZIC; yet a lower level of proteins such as ZFYVE26, LRRC25 and QPRT compared with the knee injury ones.

Network analysis of candidate proteins
Candidate proteins that would be affected by aging, menopause, disease types and estrogen levels were further analyzed by network analysis. The protein-protein interaction network was displayed in Fig.3: candidate proteins that mostly affected by aging, as well as their interacted partners, were showed in Fig.3a, and then the candidate molecules which signi cantly in uenced by menopause (Fig.3b), the ones that were correlated with disease types (Fig.3c) and estrogen levels (Fig.3d).
The reactions and pathways analysis of candidate proteins analyzed by the Reactome program The ST proteome data were used for further analysis by the Reactome program. Because the development of OA is associated with the immune system and endocrine system, candidate proteins that were indifferent abundances in ST proteome between menopausal and premenopausal participants were analyzed and those related to the immune or endocrine system were selected. Then those selected proteins were analyzed by Reactions and Pathways in the Reactome program. The potential reactions and pathways among these proteins were shown in Fig.4, which might provide insight into the mechanism by which menopause contributes to the development of OA.

Discussion
OA is a leading cause of joint pain and disability, affecting a substantial proportion of people worldwide [13]. Therefore, more and more investigators focus on proteomics studies investigating the componence and potential roles of synovial uid [14,15,16,17,18]. Animal models of osteoarthritis are also used in recent studies to understand the pathophysiology of OA and develop successful treatment of OA [19,20]. However, the role of ST of the knee joint in the onset and development of OA remains largely unknown. In this study, we explored the relationship between the recovery of OA after surgery, age, menopause and ST proteome data, aiming at providing insights for mechanisms of the onset and development of OA.
As shown in the results above, aging is the key factor that may in uence the post-surgery recovery of knee OA or injury patients, but menopause is not associated. No statistical differences were observed while in comparison to participants with different estrogen levels. The reason might be that estrogen levels may vary according to the period of studied women; the whole picture of their regular mean estrogen levels could not be represented by several tests. Namely, our data of the analyzed estrogen levels were not comprehensive. Since those data might not comprehensively re ect the actual estrogen levels, we were unable to gure out whether estrogen contributes to the remaining joint functions or not. Yet, the role of estrogen in OA initiation and development was inconclusive. According to some surveys, estrogen did contribute to the recovery of OA and could markedly reduce the prevalence of OA [21. In vivo data also suggested that estrogen played an anti-in ammation role in OA animal models [22]. By inhibiting the development of joint swelling and reducing the release of in ammatory cytokines, estrogen could effectively relieve much joint pain and improve the joint swelling situation [22]. Therefore, some estrogen drugs were applied for OA treatment. As one of the OA therapy strategies, estrogen drugs were evaluated by many investors, yet the results and conclusions were inconsistent among the clinical surveys [21,23,24]. Such a phenomenon might be a consequence of various causes.
We demonstrated that the protein pro les vary with aging, menopause, as well as disease types. The proteomic data revealed that the protein composition of old OA women differed from those of young OA women. Our data is consistent with the former report [3]. Unfortunately, perplexed by how age affected human life for centuries, scientists are still unable to explain the aging problem. According to the current technology, we cannot explain how aging leads to the difference in protein mass spectrum? We are also not clear whether such differences are a cause of aging or are a consequence of aging.
It is well established that menopause leads to a signi cant change of expression levels of some special genes. However, to our best knowledge, similar in uences caused by estrogen levels in ST have not been well studied. According to our data, there were no remarkable differences in ST protein expression levels between patients with different estrogen levels at comparable ages, which is quite adverse to the results of a signi cant association between ST protein pro le and menopause. This paradox may be the result of the dynamic uctuation of estrogen levels in premenopausal women. Women with perimenopausal estrogen levels vary greatly, estrogen levels can be very high.
Similarly, it is also reported that compared to normal or ACL patients, the RNA and protein expression levels demonstrated statistical differences in OA patients in samples such as synovial uid, synovium and peripheral blood [25]. The different expression genes are supposed to contribute to the initiation and OA and have aroused universal concern, which are optical potential biomarkers of joint diseases. For instance, a report suggests that relevant biomarkers of joint disease are identi ed by metabolomic pro ling of synovial uid [26].
Then the differently expressed proteins in different groups were further analyzed (Fig.3). Although the proteins that might be changed due to aging, menopause and OA were identi ed, the pathways via which they might contribute to the joint function remaining or degeneration were still unclear. Protein-protein interaction might provide important clues in studying the molecular mechanisms that might be involved.
The reactions and pathways among immune or endocrine-related proteins which varied signi cantly in old OA women compared to young OA women were analyzed and visualized by the Reactome program (Fig.4), which might provide insight into the molecular mechanism of aging contributing to the development of OA.

Conclusions
In the study, protein expression pro les of synovial tissues from female knee OA and knee injury patients were analyzed. Around 4000 proteins were identi ed; 3465 proteins were analyzed. Our results revealed that protein expressions in ST were affected by both aging and menopause. Similarly, aging and menopause also in uenced the post-surgery joint functional outcomes of these patients. Therefore, aging and menopause might affect the development of OA as well as the post-surgery recovery of the knee joint through changing the protein expression patterns of ST, we supposed. However, the exact effect of the protein components in ST on the occurrence, development, and rehabilitation of OA remains unclear. More studies are needed to further elucidate the in uences of aging and menopause on the knee joint and throw light on the mechanism through which ST might affect the functions of the knee joint. *Estrogen levels of postmenopausal women are all less than 11.8.   The protein-protein interaction network analyzed by STRING. a The protein-protein interaction network of the proteins most affected by aging from synovial tissues (ST) of knee osteoarthritis (OA) women. b The protein-protein interaction network of the proteins varied greatly with menopause in synovial tissues (ST) of knee osteoarthritis (OA) women. c The protein-protein interaction network of the proteins statistical different between knee OA and injury women. d The protein-protein interaction network of the proteins statistical different in ST samples from OA women with high and low estrogen level.