Cartilage Targets of Knee Osteoarthritis Shared by Both Genders

As the leading cause of disability, osteoarthritis (OA) affects people of all ages, sexes, and races. With the increasing understanding of OA, the sex differences have attracted specific attention as the burden of OA is greater in women. There is no doubt that gender-specific OA management has great potential for precision treatment. On the other hand, from the marketing aspect, a medication targeting the OA-responsive biomarker(s) shared by both genders is more favorable for drug development. Thus, in the current study, a published transcriptome dataset of knee articular cartilage was used to compare OA and healthy samples for identifying the genes with the same significantly different expression trend in both males and females. With 128 genes upregulated and 143 genes downregulated in both OA males and females, 9 KEGG pathways have been enriched based on the current knowledge, including ‘renal cell carcinoma,’ ‘ECM-receptor interaction,’ ‘HIF-1 signaling pathway,’ ‘MicroRNAs in cancer,’ ‘focal adhesion,’ ‘Relaxin signaling pathway,’ ‘breast cancer,’ ‘PI3K-Akt signaling pathway,’ and ‘human papillomavirus infection.’ Here, we explore the potential impacts of these clusters in OA. We also analyze the identified ‘cell plasma membrane related genes’ in-depth to identify the potential chondrocyte cell surface target(s) of OA management.


Introduction
Osteoarthritis (OA) is the most common form of arthritis and a leading cause of disability [1], resulting in a tremendous burden on the affected individuals, healthcare systems, and socioeconomic costs. A recently published epidemiological report reveals that, globally, the age-standardized point prevalence and annual incidence rate of OA in 2017 were 3754.2 and 181.2 per 100,000, a boost of 9.3% and 8.2% from 1990, respectively [1]. With increased life expectancy and the inflated global population, OA-responsive financial disaster is foreseeable escalating worldwide problem [1].
Due to the fact that females have a higher prevalence of OA than males [1,2], 'sex differences' [2][3][4][5][6][7] and 'gender-specific' [8][9][10][11] started to attract more and more attention in the field of OA [12]. Women also experience more significant functional disability than men. Besides, compared with men, women are more likely to seek knee OA-related healthcare, such as narcotic or nonnarcotic analgesics, corticosteroid administration, hyaluronic acid injection, knee magnetic resonance imaging, physical therapy, and occupational therapy preceding total knee arthroplasty [13]. No doubt, fully understanding the sex-related differences of OA could significantly aid in comprehensively dissecting the underlying mechanism of the disease and eventually developing safe and efficient therapeutic approaches to OA treatment.
Despite the efforts made in the last several decades, there are currently no approved disease-modifying osteoarthritis drugs (DMOADs) that can prevent, stop, or even restrain OA progression [14][15][16]. The current treatments in clinical settings, such as the analgesics and corticosteroids mentioned above, are palliative and, unfortunately, end with the painful procedure of joint replacement. Thus, the Osteoarthritis Research Society International (OARSI) still recognizes OA as an incurable condition [15]. What is more, multiple adverse side-effects in the musculoskeletal, cardiovascular, and gastrointestinal systems [17][18][19] challenge the long-term usage of non-steroidal anti-inflammatory drugs (NSAIDs) and glucocorticoids as the safe treatment options for OA. Recently, the pro-chondrogenic growth factors turned out to be the hot points for seeking new OA therapeutics [20][21][22][23][24][25][26]. However, the prospect of strategies using well-known pro-chondrogenic growth factors as treatments for arthritis does not appear to be optimistic either. For example, a recent double-blind, phase 3 clinical trial shows that the application of transforming growth factor (TGF)β1-overexpressing chondrocytes does not have statistically significant effects for OA treatment in comparison with a placebo, even with the sacrifice of inducing multiple adverse events including peripheral edema (9%), arthralgia (8%), joint swelling (6%), and injection site pain (5%) [27]. Therefore, there is an emerging demand for novel DMOADs that simultaneously inhibit inflammation and promote cartilage regeneration.
As mentioned previously, multiple studies have highlighted that OA pathogenesis, epidemiology, and disease characteristics are different between males and females; thus, when developing novel treatment targets of OA, we should also consider the gender-related factors. However, no matter if the studies were conducted in cell cultures in vitro or in animal models in vivo, most of the investigations intending to uncover OA mechanisms have not considered 'Sex as a Biological Variable' (SABV) to date [2]. For instance, from 2011 through 2019, only 32-40% of publications concerning knee OA evaluated sex-specific outcomes [12]. With the unignitable weight of sex, the current failure to identify an efficient OA treatment strategy might consciously exist until we take the sex into consideration. There is no doubt that gender-specific OA management has great potential for precision treatments. On another hand, from the marketing aspect, a medication targeting the OAresponsive biomarker(s) shared by both genders is more favorable for drug development. Therefore, publicly available transcriptome datasets of knee articular cartilage derived from healthy and OA donors were analyzed in the current study to gain insight into developing the new generation of DMOADs that could potentially benefit both genders.

Initial Evaluation of the Cartilage RNA-seq Data Set
Using the keywords 'osteoarthritis' and 'cartilage' in the https://www.ncbi.nlm.nih. gov/gds database with the selection of 'Homo sapiens' under the column of 'Top Organisms' and 'Expression profiling by high throughput sequencing' under the column of 'study type,' 31 series were identified. After reviewing all these datasets to check if they provided the sex information of the donors, one series (GSE114007) containing transcriptome data of human knee cartilage from 18 healthy (5 females, 13 males) and 20 OA (11 females, 9 males) samples was included in the current study. The sample information is listed in Supplementary Table S1.
Both the multidimensional scaling (MDS) and principal component analysis (PCA) on the transcriptome profiling could distinguish the healthy and OA samples (Figure 1), while only MDS exhibits a trend to separate male and female samples (MDS2; Figure 1A), which highlights the difficulty for developing gender-specific OA-treatments.

Identification of the Common OA-Responsive Genes in Both Males and Females
The heatmap of the differential expression genes (DEGs) in the articular cartilage of the knee derived from male and female donors visualized the gene expression changes between healthy and OA samples ( Figure 2). Overall, the healthy samples could be easily distinguished from OA samples, but there were no distinct male and female clusters. Comparing the gene alternation in the cartilage of each gender, females display more gene alteration in response to OA than males: 923 DEGs (541 increased, 382 decreased) were identified when comparing OA female samples to their healthy counterparts versus 419 OA-responsive DEGs (186 increased, 233 decreased) in male samples (Figure 3, Supplementary Tables S2 and S3). Moreover, we recognized 128 upregulated and 143 downregulated DEGs shared by both genders (Figure 3, Supplementary Tables S4 and S5). Besides, no gene exhibited different expression trends in males and females in response to OA ( Figure 3).

Pathway Enrichment
Firstly, pathway enrichment was conducted by the STRING networks [28] to gain insight into the potential underlying regulation pathways. Two hundred and thirty of the total 271 DEGs could be recognized by the STRING server (https://string-db.org/cgi/input?sessionId=bSVIOAB7JrVf&input_page_show_search=on; Figure 4), and 9 KEGG pathways were successfully enriched with a false discovery rate (FDR) smaller than 0.05 ( Table 1). The 'ECM-receptor interaction' pathway was recognized as expected. In addition, 'HIF-1 signaling pathway,' 'Focal adhesion,' 'Relaxin signaling pathway,' and 'PI3K-Akt signaling pathway' were identified; these should be considered high priority targets for the further investigation as targets of OA shared by both genders. Surprisingly, 'renal cell carcinoma,' 'MicroRNAs in cancer,' 'breast cancer,' and 'human papillomavirus infection' were also enriched.

Identification of the Common OA-Responsive Genes in Both Males and Females
The heatmap of the differential expression genes (DEGs) in the articular cartilage of the knee derived from male and female donors visualized the gene expression changes between healthy and OA samples ( Figure 2). Overall, the healthy samples could be easily distinguished from OA samples, but there were no distinct male and female clusters. Comparing the gene alternation in the cartilage of each gender, females display more gene alteration in response to OA than males: 923 DEGs (541 increased, 382 decreased) were identified when comparing OA female samples to their healthy counterparts versus 419 OAresponsive DEGs (186 increased, 233 decreased) in male samples ( Figure 3, Supplementary  Tables S2 and S3). Moreover, we recognized 128 upregulated and 143 downregulated DEGs shared by both genders (Figure 3, Supplementary Tables S4 and S5). Besides, no gene exhibited different expression trends in males and females in response to OA (Figure 3).

Pathway Enrichment
Firstly, pathway enrichment was conducted by the STRING networks [28] to gain insight into the potential underlying regulation pathways. Two hundred and thirty of the total 271 DEGs could be recognized by the STRING server (https://string-db.org/cgi/ input?sessionId=bSVIOAB7JrVf&input_page_show_search=on; Figure 4), and 9 KEGG pathways were successfully enriched with a false discovery rate (FDR) smaller than 0.05 ( Table 1). The 'ECM-receptor interaction' pathway was recognized as expected. In addition, 'HIF-1 signaling pathway,' 'Focal adhesion,' 'Relaxin signaling pathway,' and 'PI3K-Akt signaling pathway' were identified; these should be considered high priority targets for the further investigation as targets of OA shared by both genders. Surprisingly, 'renal cell carcinoma,' 'MicroRNAs in cancer,' 'breast cancer,' and 'human papillomavirus infection' were also enriched.
genders, there are 128 upregulated DEGs and 143 downregulated DEGs shared by both genders. No gene exhibited different expression trends in males and females in response to OA. Figure 4. The protein-protein interaction network generated on the SPRING server against the Homo sapiens gene database. The STRING server identified 230 genes among the input of the 271 common DEGs in cartilage shared by both genders. The network nodes represent proteins. The empty nodes display the proteins without a known or predicted 3D structure, while the filled nodes illustrate those with a known or predicted 3D structure. The edges represent protein-protein associations: light blue-known interactions from curated databases; burgundy-experimentally determined known interactions; green-predicted interactions by gene neighborhood; orange-predicted interactions by gene fusions; dark blue-predicted interactions by gene co-occurrence; yellow-text mining; black-co-expression; purple-protein homology. . The protein-protein interaction network generated on the SPRING server against the Homo sapiens gene database. The STRING server identified 230 genes among the input of the 271 common DEGs in cartilage shared by both genders. The network nodes represent proteins. The empty nodes display the proteins without a known or predicted 3D structure, while the filled nodes illustrate those with a known or predicted 3D structure. The edges represent protein-protein associations: light blue-known interactions from curated databases; burgundy-experimentally determined known interactions; green-predicted interactions by gene neighborhood; orange-predicted interactions by gene fusions; dark blue-predicted interactions by gene co-occurrence; yellow-text mining; black-co-expression; purple-protein homology.

Cellular Component Enrichment
From a drug development stand point, the molecules located in the cell surface should be the first line targets, as they play critical roles in cell-cell interactions, nutrient and metabolite transportation, the translation of extracellular signals into intracellular response and vice versa, and, most importantly, due to their accessibility [29]. Thus, the enrichment of cellular components was also performed by STRING in the current study.

Symbol
Gene Name Expression Function in Uniprot

RAMP3
Receptor activity-modifying protein 3 plasma membrane adenylate cyclase-activating G protein-coupled receptor signaling pathway; adrenomedullin receptor signaling pathway; amylin receptor signaling pathway; angiogenesis; calcium ion transport; cellular response to estradiol stimulus; cellular response to hormone stimulus; cross-receptor inhibition within G protein-coupled receptor heterodimer; G protein-coupled receptor signaling pathway; G protein-coupled receptor signaling pathway involved in heart process; intracellular protein transport; positive regulation of calcium ion import across plasma membrane;   activation of protein kinase activity; angiogenesis; artery morphogenesis; basophil chemotaxis; branching involved in blood vessel morphogenesis; branching morphogenesis of an epithelial tube; camera-type eye morphogenesis; cardiac muscle fiber development; cardiac vascular smooth muscle cell development; cell maturation; cell migration involved in sprouting angiogenesis; cellular response to hypoxia; cellular response to vascular endothelial growth factor stimulus; cellular stress response to acid chemical; commissural neuron axon guidance; coronary artery morphogenesis; coronary vein morphogenesis; cytokine-mediated signaling pathway; dopaminergic neuron differentiation; endothelial cell chemotaxis; epithelial cell differentiation; eye photoreceptor cell development; heart morphogenesis; induction of positive chemotaxis; in utero embryonic development; kidney development; lactation; lung development; lymph vessel morphogenesis; macrophage differentiation; mammary gland alveolus development; mesoderm development; monocyte differentiation; motor neuron migration; negative regulation of adherents junction organization; negative regulation of apoptotic process; negative regulation of blood-brain barrier permeability; negative regulation of cell-cell adhesion mediated by cadherin; negative regulation of cysteine-type endopeptidase activity involved in apoptotic process; negative regulation of establishment of endothelial barrier; negative regulation of gene expression; negative regulation of transcription by RNA polymerase II; nervous system development; outflow tract morphogenesis; Table 3. Cont. promoter in response to hypoxia; response to hypoxia; sprouting angiogenesis; surfactant homeostasis; tube formation; vascular endothelial growth factor receptor-2 signaling pathway; vascular endothelial growth factor receptor signaling pathway; vascular endothelial growth factor signaling pathway; vascular wound healing; vasculogenesis; VEGF-activated neuropilin signaling pathway Table 3. Cont.

WNT5A
Protein Wnt-5a Extracellular region or secreted activation of GTPase activity; activation of JUN kinase activity; activation of MAPK activity; activation of protein kinase B activity; anterior/posterior axis specification, embryo; atrial septum development; axon guidance; canonical Wnt signaling pathway; cartilage development; cell fate commitment; cellular protein localization; cellular response to calcium ion; cellular response to interferon-gamma; cellular response to lipopolysaccharide; cellular response to retinoic acid; cellular response to transforming growth factor beta stimulus; cervix development; chemoattraction of serotonergic neuron axon; chemorepulsion of dopaminergic neuron axon; cochlea morphogenesis; convergent extension involved in axis elongation; convergent extension involved in organogenesis; embryonic skeletal system development; epithelial cell proliferation involved in mammary gland duct elongation; epithelial to mesenchymal transition; establishment of epithelial cell apical/basal polarity; establishment of planar polarity; excitatory synapse assembly; face development; genitalia development; heart looping; hematopoietic stem cell proliferation; hindgut morphogenesis; hypophysis morphogenesis; inhibitory synapse assembly; keratinocyte differentiation; kidney development; lateral sprouting involved in mammary gland duct morphogenesis; lens development in camera-type eye; lung development; male gonad development; mammary gland branching involved in thelarche; melanocyte proliferation; membrane organization; mesenchymal-epithelial cell signaling; mesodermal to mesenchymal transition involved in gastrulation; midgut development; negative regulation of apoptotic process; negative regulation of axon extension involved in axon guidance; negative regulation of BMP signaling pathway; negative regulation of canonical Wnt signaling pathway; negative regulation of cell proliferation in midbrain; negative regulation of epithelial cell proliferation; negative regulation of fat cell differentiation; negative regulation of fibroblast growth factor receptor signaling pathway; negative regulation of melanin biosynthetic process; negative regulation of mesenchymal cell proliferation; negative regulation of prostatic bud formation; negative regulation of synapse assembly; negative regulation of transcription, DNA-templated; neuron differentiation; non-canonical Wnt signaling pathway; non-canonical Wnt signaling pathway via JNK cascade; notochord morphogenesis; olfactory bulb interneuron development; optic cup formation involved in camera-type eye development; paraxial mesoderm formation; planar cell polarity pathway involved in axis elongation; planar cell polarity pathway involved in axon guidance; planar cell polarity pathway involved in cardiac muscle tissue morphogenesis; planar cell polarity pathway involved in cardiac right atrium morphogenesis; planar cell polarity pathway involved in gastrula mediolateral intercalation; planar cell polarity pathway involved in midbrain dopaminergic neuron differentiation; planar cell polarity pathway involved in neural tube closure; planar cell polarity pathway involved in outflow tract morphogenesis; planar cell polarity pathway involved in pericardium morphogenesis; positive regulation of angiogenesis; Table 3. Cont.    It is worth noting that one of the plasma membrane expressed proteins, ITGB8, is associated with 4 of the 9 enriched KEGG pathways, including 'ECM-receptor interaction,' 'focal adhesion,' 'PI3K-Akt signaling pathway,' and 'human papillomavirus infection' (Table 1). This finding indicates that ITGB8 might be a key regulator for several cellular events and pathways in the cartilage during OA for both genders. Indeed, ITGB8 is already known to be associated with cartilage development (Table 3). Thus, ITGB8 might be considered one of the main targets for further OA treatment strategy investigations.

Symbol
On the other hand, among the identified plasma membrane expressed proteins, ITGB8 is the only protein known to function in cartilage development or (re)generation based on the Uniprot database. As inflammation is a major event during OA, it is no surprise to recognize several proteins in the current study that participate in the inflammatory response, such as ANPEP (which has a function in neutrophil degranulation), ITGB2 (which has a function in leukocyte migration in the inflammatory response), KIT (which has a function in B cell differentiation and T cell differentiation), and NOD2 (which has a function in the cytokine production involved in the immune response; interleukin (IL)-1mediated signaling pathway; macrophage apoptotic process; B cell activation; as well as IL10, IL17, IL1β, IL6, and IL8 production) (Table 3).
Unexpectedly, several proteins characterized as important neural event regulators were also displayed in the enrichment. For example, CDH2 regulates brain and cerebral cortex development, glial cell differentiation, and neuroligin clustering in postsynaptic membrane assembly; GRIA2 and GRIN2A participate in chemical synaptic transmission and regulation of NMDA receptor activity; and NCAM1 is important in commissural neuron axon guidance (Table 3). Since cartilage has no nerve intervention [31], the function of these neural related proteins on cartilage, especially during OA progression, warrants further investigation.

OA-Responsive Genes with Different Fold Change in Females and Males
It is possible that, although some genes have a similar alteration trend in both genders, their expression fold change in response to OA may be significantly different in males and females. As listed in Table 4, we identified five genes that upregulated in both male and female OA cartilage tissues, including collagen type I alpha 1 chain (COL1A1), collagen type I alpha 2 chain (COL1A2), matrix remodeling associated 5 (MXRA5), THY1, TNF alpha induced protein 6 (TNFAIP6), which exhibited significantly higher OA-responsive increases in females than that found in males. For instance, both COL1A1 and COL1A2 are significantly higher in females than males, aligning with the observation that female patients tend to have more advanced cartilage degradation than male patients during OA [32]. Also, based on the functional information collected in the Uniprot database, both THY1 and TNFAIP6 participate in the inflammatory response (Supplementary Table S6), indicating that the female patients may have more intense inflammation in the cartilage than male patients during OA. The potential interaction of these five genes was also analyzed by the STRINGs server, and four of them were enriched in a biological process termed the 'immune system process' (Figure 5). This result further emphasizes the differently aroused levels of the immune response in males and females during OA, which potentially contributes to the phenomenon that female patients tend to have more severe OA symptoms than males [33]. Therefore, these genes and their encoded proteins may represent another set of targets for gender-specific OA-therapies.  Figure 5. The STRING server displayed the protein-protein interaction network among the five common OA-responsive DEGs, which have significantly different multiples of change in male and female cartilage. The network nodes represent proteins, while filled nodes are for the proteins with known or predicted 3D structures. Edges represent protein-protein associations. The list contains identified pathways with a false discovery rate (FDR) of less than 0.05.

Discussion
In the current study, several commonly used OA biomarkers have been detected in the cartilage of both genders, including the upregulated a disintegrin and metalloproteinase with thrombospondin-like motifs 7 (ADAMTS7) [34], ADAMTS14 [34], COL1A1, COL1A2, and matrix metallopeptidase 13 (MMP13) [35,36], as well as the downregulated sex-determining region Y box 9 (SOX9) [37]. In addition, four of eight hub genes recently identified by Zhang et al. [38], including Periostin (POSTN), Collagen Type III Alpha 1 Chain (COL3A1), COL1A1, and COL1A2, were identified in the current study. These results demonstrate the reliability of the current study.
It is worth noting that female cartilage demonstrated a significantly higher elevation of COL1A1 and COL1A2 in response to OA than male cartilage. Clinically, when patients seek treatment, females have more debilitating pain and experience more advanced OA stages than their male counterparts [32]. Our current finding proved that, at least for the donors included in the current study, women tend to have more fibrocartilage containing Figure 5. The STRING server displayed the protein-protein interaction network among the five common OA-responsive DEGs, which have significantly different multiples of change in male and female cartilage. The network nodes represent proteins, while filled nodes are for the proteins with known or predicted 3D structures. Edges represent protein-protein associations. The list contains identified pathways with a false discovery rate (FDR) of less than 0.05.

Discussion
In the current study, several commonly used OA biomarkers have been detected in the cartilage of both genders, including the upregulated a disintegrin and metalloproteinase with thrombospondin-like motifs 7 (ADAMTS7) [34], ADAMTS14 [34], COL1A1, COL1A2, and matrix metallopeptidase 13 (MMP13) [35,36], as well as the downregulated sex-determining region Y box 9 (SOX9) [37]. In addition, four of eight hub genes recently identified by Zhang et al. [38], including Periostin (POSTN), Collagen Type III Alpha 1 Chain (COL3A1), COL1A1, and COL1A2, were identified in the current study. These results demonstrate the reliability of the current study.
It is worth noting that female cartilage demonstrated a significantly higher elevation of COL1A1 and COL1A2 in response to OA than male cartilage. Clinically, when patients seek treatment, females have more debilitating pain and experience more advanced OA stages than their male counterparts [32]. Our current finding proved that, at least for the donors included in the current study, women tend to have more fibrocartilage containing type I collagen than men during OA. Simultaneously, female OA cartilage displayed a more severe increase in TNFAIP6, THY1, and MXRA5. TNFAIP6 (also known as TSG-6), which has been demonstrated to be highly expressed in damaged articular and meniscal cartilage and cytokine-treated chondrocytes [39]. At the same time, THY1 (also known as CD90), a well-known inflammation regulator [40] and fibroblast-like cell surface marker [41], is also proved to be highly expressed in OA cartilage and could be induced by inflammatory cytokine interleukin-1β [42]. The reports about MXRA5 in OA cartilage are limited, but it has been identified in the synovial fluid of patients with OA [43]. Thus, the female OA cartilage has a more robust OA-responsive arousal of fibrocartilage-related and inflammation-related biomarkers than the male OA cartilage, which may partially explain why female patients tend to have more severe OA symptoms than males [33].
Besides, the enriched biological processes include 'response to transforming growth factor beta' based on proteins MXRA5, COL1A1, and COL1A2 ( Figure 5). In fact, transforming growth factor-beta (TGF-β) family members have been intensively studied for OA therapeutic intervention [44,45]. However, the clinical trial results of TissueGene-C (TG-C), a cell and gene therapy for knee OA consisting of non-transformed and retrovirally transduced TGF-β1-overexpression chondrocytes (3:1), are not promising [27]. Considering the findings in the current study, we infer that the efficiently effective dose of TGF-β might be different in females and males.
Interestingly, we also observed several tumor-related clusters, including 'renal cell carcinoma,' 'microRNAs in cancer,' and 'breast cancer' from the enrichment of the common DEGs shared by female and male OA cartilage. Indeed, a recent cohort study reported that elderly patients with knee or hip OA have higher risks of melanoma, renal cell cancer, and cancer of the bladder, breast, uterus, and prostate [46], supporting our current discovery. This association between OA and tumorigenic risk may be attributed to the chronic usage of NSAIDs; however, further investigation is needed to uncover the full story.
In addition, 'Relaxin signaling pathway' is one of the enriched clusters. Through binding to its receptor in various tissues and via medication by signaling pathways, Relaxin has regulatory effects on the musculoskeletal and other systems [56]. The Relaxin signaling pathway has many biological actions, including anti-fibrotic, vasodilatory, angiogenic, anti-inflammatory, anti-apoptotic, and organ protective effects across various tissues [57]. Specific to cartilage, Relaxin appears to decrease knee articular cartilage stiffness [56]. However, knowledge about the Relaxin signaling pathway in OA is still largely lacking.
As mentioned above, the surface molecules could be the primary targets for drug development because of their critical roles and, more crucially, their accessibility [29]. ITGB8 is a cell plasma membrane protein that matched with four of the nine enrichment pathways, including 'ECM-receptor interaction,' 'focal adhesion,' 'PI3K-Art signaling pathway,' and 'human papillomavirus infection.' The integrin family members play a major role in mediating the interaction between cells and the extracellular matrix (ECM) [66]. Using multiple chondrogenic models, a previous in vitro investigation demonstrated that ITGB8 was consistently upregulated in human mesenchymal stem cells (MSCs) during chondrogenic differentiation [67]. On the other hand, knocking down ITGB8 markedly impacted MSCs' chondrogenic differentiation [67]. A recent publication also indicates that ITGB8 plays a key role in chondrocyte sheet attachment and the in ECM synthesis in cartilage [68]. In combination, ITGB8 seems to have a beneficial effect on cartilage. Interestingly, we noticed the upregulation of ITGB8 in both male and female OA cartilage, suggesting upregulation of ITGB8 may have a protect/repair function on the cartilage in response to OA progression. However, the detailed functions of ITGB8 and whether it can be the master target that regulates several important signaling pathways during OA treatment are still open questions that remain to be answered.
Finally, several neural cell membrane proteins' levels were significantly altered in response to OA in both male and female cartilage. Since cartilage has no nerve intervention [31], the OA-responsive adaption of gene expression is more than likely from chondrocytes. Meanwhile, accumulating studies demonstrate that the crosstalk between skeletal and neural tissue is critical for skeletal health [68,69]. For example, there is bidirectional communication between the sensory neurons and osteoblasts through glutamate, substance P, and ATP [70]. Unfortunately, since the skeletal-neuron crosstalk evaluation is still in its infancy, most investigations have focused on bone and muscle as these two tissues have nerve innervation [71,72]. So far, based on the authors' thorough literature search with the keywords 'cartilage or chondrocyte + nerve or neural + crosstalk,' there is no publicly available literature dissecting the interaction between articular cartilage and the neural system. The only available report we could access revealed the elevation of Chemotactic cytokine ligand 2 (CCL2) and nerve growth factor (NGF) in the joint during OA mediate inflammation and pain [73]. Based on the current findings that cartilage expresses neural-related cell membrane proteins and that the expression of these proteins amended in response to OA, whether cartilage-nerve crosstalk exists or the neural-related proteins have some unrevealed functions in cartilage generation and maintenance are definitely interesting topics to be investigated in-depth.
Although the main objective of this investigation is to analyze OA-related gene expression changes in cartilage shared by both genders, we are aware that both age and body mass index (BMI) are also important confounders. Unfortunately, although healthy and OA donors' body mass is not statistically different, individual BMI information was not provided in the original study [74]. We also noticed that both male and female OA donors were older than their healthy counterparts (Supplementary Table S1), which agrees with the clinical observation that OA's incidence increases along with age [75]. On the other hand, this age difference may lead to the false-positive identification of DEG to some degree, as discussed in the original investigation [74]. Considering that the OA samples were collected in the clinic for a single study, the sample size is insufficient for us to conduct multivariate analysis involving the donor's age and BMI for result adjustment. It is potentially a major limitation of the original investigation, as well as our current study, and highlights the importance of worldwide collaboration on sample collection and complete data sharing.

Materials and Methods
The sequence read archive (SRA) data of healthy and OA human knee cartilage (GEO accession number: GSE114007) RNA-seq data were downloaded from https://www.ncbi. nlm.nih.gov/sra. Specifically, normal human knee cartilage tissues were procured by tissue banks from donors without a history of joint disease or trauma and processed within 24-48 hours post mortem. OA-affected cartilage was harvested from the tissue removed during knee replacement surgery. Cartilage was stored at −20 • C in Allprotect Tissue Reagent (Qiagen, Valencia, CA, USA) immediately after harvest until RNA extraction [74]. Data analyses were performed on the Galaxy platform (UseGalaxy.org, [76]). The FASTQC RNAseq reads were aligned to the human genome (GRCh38) using the HISAT2 aligner (Galaxy Version 2.1.0+galaxy 5) with default parameters [77]. Raw counts of sequencing read for the feature of genes were extracted by featureCounts (Galaxy Version 1.6.4+galaxy1) [78]. Then, the limma package (Galaxy version 3.38.3+galaxy3) was used to identify DEGs with its voom method [79,80]. The expressed genes were selected as their counts per million (CPM) value were not less than one in at least two samples across the entire experiment, while lowly expressed genes were removed for the flowing analyses. Quasi-likelihood F-tests (ANOVA-like analysis) were achieved to identify DEGs [81]. Genes with a fold change (FC) of more than 2 and FDR less than 0.01 were assigned as DEGs. Heatmap, MDS, PCA, and Venn diagram construction were conducted in R (version 3.6.3, Free Software Foundation, Boston, MA, USA) [82] with the packages pheatmap (version 1.0.12), vegan (version 2.56), ggplot2 (version 3.3.0) and VennDiagram (version 1.6.20). The pathway enrichment of identified DEGs was performed against the STRING network [28]. In addition, the summary of the known biofunctions for these genes was searched in the Uniprot database [30] for manually functional annotation.

Conclusions
In summary, our current study identified several novel genes and signaling pathways as the potential targets for further OA investigations, which will benefit both genders. In addition, several neural-related cell surface proteins whose expression is altered in response to OA in both genders' cartilage were also recognized, indicating the potential neuro-cartilage interaction. Last but not least, genes which have a significantly different fold change in male and female cartilages during OA were also evaluated, providing a potential explanation for different OA symptom severity and treatment responses in male and female patients. Overall, we hope our work could help direct the further development of DMOADs, which could truly benefit both male and female OA patients.
Supplementary Materials: Supplementary materials can be found at https://www.mdpi.com/1422 -0067/22/2/569/s1. Table S1. The information of the samples included in the current study. Table S2. The list of the identified genes statistically significantly changed in OA cartilage of males. Table S3. The list of the identified genes statistically significantly changed in OA cartilage of females. Table S4. The list of the identified genes statistically significantly upregulated in OA cartilage of both genders. Table S5. The list of the identified genes statistically significantly downregulated in OA cartilage of both genders. Table S6. The list of the identified genes statistically significantly different between genders.
Author Contributions: Conceptualization, C.L.; methodology, C.L. and Z.Z.; validation, C.L. and Z.Z.; formal analysis, Z.Z.; writing-original draft preparation, C.L.; writing-review and editing, Z.Z. Both authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are contained within this article and Supplementary materials.

Conflicts of Interest:
The authors declare no conflict of interest.