Sortilin Is Upregulated in Osteoarthritis-Dependent Cartilage Calcification and Associated with Cellular Senescence

Osteoarthritis (OA) is a chronic joint disease characterized by articular cartilage calcification, loss of articular cartilage, bone changes, pain, and disability. Cartilage calcification is one hallmark of OA and is predominantly caused by basic calcium crystals formed due to an imbalance of the pyrophosphate pathway. Sortilin is a transmembrane protein that contributes to vascular calcification in atherosclerosis by externalizing alkaline phosphatase (ALP)-containing vesicles. Calcification in atherosclerosis and osteoarthritis has been associated with cellular senescence. The aim of this study was to investigate the potential role of sortilin and senescence in osteoarthritis-dependent cartilage calcification. Osteoarthritic cartilage from human knee joints was collected after joint replacement, and samples were analyzed by immunohistochemistry and quantitative RT-PCR analysis. Human chondrocytes were treated with osteogenic medium for up to 21 days to induce calcification. Western blots for sortilin and ALP, as well as an ALP activity assay, were performed. Human chondrocytes were treated with mitomycin C to induce senescence, and sortilin expression was quantified at the protein and gene levels. Sections of knee joints from a murine model of osteoarthritis were stained for sortilin and p16 and analyzed by immunohistochemistry. Treatment of wild-type chondrocytes using an osteogenic medium similar to human chondrocytes was performed. Osteoarthritic cartilage from mouse and human knee joints showed an increased number of sortilin and p16-positive chondrocytes compared to healthy cartilage. This observation was corroborated by increased gene expression of sortilin and p16 in mild and moderate osteoarthritic cartilage samples. To investigate the mechanism of sortilin regulation, human chondrocytes were treated with osteogenic medium to induce calcification. Sortilin protein levels and expression were increased after 7 days of stimulation, whereas ALP levels and activity were upregulated after 21 days of stimulation. Similar observations were made in a murine osteoarthritis model. Mechanistically, senescent chondrocytes induced by mitomycin C showed an upregulation of sortilin and ALP gene expression compared to non-senescent chondrocytes. Our data indicate that sortilin and ALP are upregulated during cartilage calcification, which is associated with chondrocyte senescence and thus might contribute to the pathogenesis of osteoarthritis. Cellular senescence seems to induce sortilin expression.


Introduction
Osteoarthritis (OA) is the most common chronic disease of the musculoskeletal system. It causes pain and loss of function in the affected joint, predominantly affecting the knee and hip joints [1][2][3]. This leads to impairments in the patients' quality of life [4]. Age, genetic, metabolic, mechanical, and traumatic factors play a crucial etiological role in regulating the onset of OA [5].

Sortilin and Senescence Are Increased in OA Mouse Knees
We investigated the expression of sortilin and cellular senescence in natural OA modelaged mice. Representative sections of safranin Orange staining ( Figure 1A) and immunofluorescence (IF) staining of sortilin ( Figure 1B) and senescence marker p16 ( Figure 1C) are shown. The safranin Orange staining clearly shows the hallmarks of OA in the cartilage of old mice. A severe loss of proteoglycans as well as an irregular cartilage surface is present in the old mice (45 weeks), which is absent in the young mice (8 weeks). Old mice significantly showed more abundant signs of OA compared to young mice (p < 0.0001). Sortilin-positive cells were significantly more abundant in old mice compared to young . IgG staining was used as a negative control for the specific antibody staining. The statistical evaluation was performed using an unpaired t-test. *** p < 0.001,**** p < 0.0001.

Sortilin and Senescence Are Increased in Human OA Cartilage
27 patients were included in this study. The patients were ranked in four groups according to the radiological severity of OA as assessed by the Kellgren-Lawrence Score (KL score). The patients were divided into mild (KL score 1-2; N = 9); moderate (KL score 3; N = 9); and severe (KL score 4; N = 9) OA. Knee cartilage from three healthy donors who had The white square indicates the amplified area in the picture below. Quantification of p16-positive chondrocytes (young: 28.42 ± 14.88, old: 41.04 ± 20.46, p = 0.0006, N < 8). IgG staining was used as a negative control for the specific antibody staining. The statistical evaluation was performed using an unpaired t-test. *** p < 0.001,**** p < 0.0001.

Sortilin and Senescence Are Increased in Human OA Cartilage
27 patients were included in this study. The patients were ranked in four groups according to the radiological severity of OA as assessed by the Kellgren-Lawrence Score (KL-score). The patients were divided into mild (KL-score 1-2; N = 9); moderate (KLscore 3; N = 9); and severe (KL-score 4; N = 9) OA. Knee cartilage from three healthy donors who had died without a history of OA was used as a control. Representative X-ray images, sortilin immunostaining, and the respective isotype control are shown in Figure 2A.
Sortilin-positive cells are about two-fold increased at all stages of OA severity ( Figure 2B). In line with this observation, sortilin gene expression was also significantly increased in mild and moderate OA compared with controls ( Figure 2C). In severe OA, sortilin was less upregulated, which was also reflected in fewer sortilin-positive chondrocytes in the immunostaining. Age did not correlate with the number of sortilin-positive chondrocytes in the present cohort (Supplementary Figure S1B: Pearson correlation: r = 0.301; p = 0.08). Additionally, no correlation of the patient's gender with sortilin-positive cells in knee joint cartilage was observed (Supplementary Figure S1B: unpaired t-test, p = 0.26). Therefore, age and sex do not seem to correlate with sortilin expression, but the main effect can be attributed to OA changes in cartilage. However, to test whether senescence is increased in OA cartilage, human OA cartilage samples (KL-4) were stained for p16, and the number of positive chondrocytes was counted. The number of p16-positive chondrocytes was significantly increased in OA cartilage ( Figure 2D), as was the p16 expression compared to healthy cartilage ( Figure 2E). died without a history of OA was used as a control. Representative X-ray images, sortilin immunostaining, and the respective isotype control are shown in Figure 2A. Sortilin-positive cells are about two-fold increased at all stages of OA severity ( Figure 2B). In line with this observation, sortilin gene expression was also significantly increased in mild and moderate OA compared with controls ( Figure 2C). In severe OA, sortilin was less upregulated, which was also reflected in fewer sortilin-positive chondrocytes in the immunostaining. Age did not correlate with the number of sortilin-positive chondrocytes in the present cohort (Supplementary Figure S1B: Pearson correlation: r = 0.301; p = 0.08). Additionally, no correlation of the patient's gender with sortilin-positive cells in knee joint cartilage was observed (Supplementary Figure S1B: unpaired t-test, p = 0.26). Therefore, age and sex do not seem to correlate with sortilin expression, but the main effect can be attributed to OA changes in cartilage. However, to test whether senescence is increased in OA cartilage, human OA cartilage samples (KL-4) were stained for p16, and the number of positive chondrocytes was counted. The number of p16-positive chondrocytes was significantly increased in OA cartilage ( Figure 2D), as was the p16 expression compared to healthy cartilage ( Figure 2E). . IgG staining was used as a negative control for the specific antibody staining. For statistical evaluation, either a one-way ANOVA in the case of comparing more than two groups with Dunett's post-hoc test or a Mann-Whitney test was performed. The data were analyzed for normal distribution using a Shapiro-Wilk normality test. * p < 0.05, ** p < 0.01, *** p < 0.001.

Sortilin and ALP Are Associated with Increasing Calcification of Human Chondrocytes
As sortilin function has been associated with tissue calcification and ALP activity, we investigated its role in chondrocytes under calcifying conditions. Therefore, we stimulated isolated human chondrocytes with osteogenic medium to induce calcification. Alizarin Red S was used to quantify the calcification ( Figure 3A). We observed a significant increase in Alizarin Red staining upon treatment with osteogenic medium (OM) at days 7 and 21 (p < 0.0001). However, a slight spontaneous increase in alizarin red staining over time was also observed in the untreated samples.
The analyses of p16 gene expression using qRT-PCR corroborated this result (Mann-Whitney test: Median Healthy: 0.004; OA: 0.57; p = 0.04). All pictures were taken at 400× magnification (scale bar 50µm). IgG staining was used as a negative control for the specific antibody staining. For statistical evaluation, either a one-way ANOVA in the case of comparing more than two groups with Dunett's post-hoc test or a Mann-Whitney test was performed. The data were analyzed for normal distribution using a Shapiro-Wilk normality test. * p < 0.05, ** p < 0.01, *** p < 0.001.

Sortilin and ALP Are Associated with Increasing Calcification of Human Chondrocytes
As sortilin function has been associated with tissue calcification and ALP activity, we investigated its role in chondrocytes under calcifying conditions. Therefore, we stimulated isolated human chondrocytes with osteogenic medium to induce calcification. Alizarin Red S was used to quantify the calcification ( Figure 3A). We observed a significant increase in Alizarin Red staining upon treatment with osteogenic medium (OM) at days 7 and 21 (p < 0.0001). However, a slight spontaneous increase in alizarin red staining over time was also observed in the untreated samples. To investigate the regulation of sortilin and ALP protein expression during this process, we performed a western blot. Both sortilin and ALP increase over time with OM For statistical evaluation, a two-way ANOVA with Sidak's post-hoc test was performed. ns: not significant; * p < 0.05, ** p < 0.01, **** p < 0.0001.
To investigate the regulation of sortilin and ALP protein expression during this process, we performed a western blot. Both sortilin and ALP increase over time with OM ( Figure 3B). Sortilin shows a significant increase after treatment with OM at day 7 compared with the CM control group (p = 0.04) ( Figure 3C). ALP is significantly increased in the OM group at day 21 (p = 0.02) ( Figure 3D). At the same time, we observed an increase in ALP activity on day 21 of osteogenic differentiation (p < 0.0001) ( Figure 3E). Similar results were observed using murine neonatal chondrocytes (Supplementary Figure S2).

Sortilin Is Upregulated in Senescent Cells
As our previous experiments have linked sortilin to ALP and calcification, we aimed to investigate the regulation of sortilin expression. Cellular senescence is associated with osteoarthritis and might therefore be a trigger for sortilin expression. Furthermore, p16 and p21 were used as senescence markers to verify the induction of senescence using mitomycin C. Interestingly, we observed a time-dependent upregulation of sortilin upon stimulation with mitomycin C (day 5: p = 0.02 and day 10: 0.003) ( Figure 4A). After 10 days of mitomycin C stimulation, p16 expression was upregulated (p = 0.02) ( Figure 4B), and after 5 and 10 days of mitomycin C stimulation, p21 expression was significantly increased (day 5: p < 0.0001 and day 10: 0.0004) ( Figure 4C). Interestingly, ALP was also significantly upregulated during senescence induction after days 5 and 10 (day 5: p = 0.0095; day 10: p = 0.0109) ( Figure 4D). However, in immunofluorescence staining, only a trend for sortilin and p16 increases was observed, which did not reach statistical significance ( Figure 4E), although a doubling in the number of senescent cells was observed by p16 stainings ( Figure 4F).  Figure 3B). Sortilin shows a significant increase after treatment with OM at day 7 compared with the CM control group (p = 0.04) ( Figure 3C). ALP is significantly increased in the OM group at day 21 (p = 0.02) ( Figure 3D). At the same time, we observed an increase in ALP activity on day 21 of osteogenic differentiation (p < 0.0001) ( Figure 3E). Similar results were observed using murine neonatal chondrocytes (Supplementary Figure S2).

Sortilin Is Upregulated in Senescent Cells
As our previous experiments have linked sortilin to ALP and calcification, we aimed to investigate the regulation of sortilin expression. Cellular senescence is associated with osteoarthritis and might therefore be a trigger for sortilin expression. Furthermore, p16 and p21 were used as senescence markers to verify the induction of senescence using mitomycin C. Interestingly, we observed a time-dependent upregulation of sortilin upon stimulation with mitomycin C (day 5: p = 0.02 and day 10: 0.003) ( Figure 4A). After 10 days of mitomycin C stimulation, p16 expression was upregulated (p = 0.02) ( Figure 4B), and after 5 and 10 days of mitomycin C stimulation, p21 expression was significantly increased (day 5: p < 0.0001 and day 10: 0.0004) ( Figure 4C). Interestingly, ALP was also significantly upregulated during senescence induction after days 5 and 10 (day 5: p = 0.0095; day 10: p = 0.0109) ( Figure 4D). However, in immunofluorescence staining, only a trend for sortilin and p16 increases was observed, which did not reach statistical significance ( Figure 4E), although a doubling in the number of senescent cells was observed by p16 stainings (Figure 4F).

Discussion
After SORT1, the gene coding for sortilin, was linked to increased cardiovascular risk by genome-wide association studies (GWAS) as a possible gene locus, sortilin came into focus as an influencing factor in the development process of age-associated diseases, such as dyslipidemias, atherosclerosis, diabetes mellitus type 2, and also Alzheimer's disease [12,[36][37][38][39]. Sortilin is involved in the secretion and transport of disease-related proteins in the corresponding tissues [17,38,40]. This study reveals that sortilin expression, induced by senescence, impacts the development of OA. Sortilin thus becomes another common factor in the pathogenesis of atherosclerosis and OA.
Previous studies demonstrated a close association between arteriosclerosis and significantly elevated levels of sortilin [17,41]. We observed the same effect for sortilin in OA. An increase in sortilin at protein and gene expression levels was observed in human and murine OA cartilage (Figures 1 and 2), although the severity of OA did not appear to influence sortilin expression.
A previous atherosclerosis study mimicked vascular calcification using human vascular smooth muscle cell (hSMC) stimulation with OM. As the calcification of hSMC increased, the level of sortilin also increased [17]. Similarly, human and murine chondrocytes were treated with OM in this study. An association between sortilin expression and calcified OA chondrocytes was demonstrated in this study. The increase in sortilin expression was accompanied by an increase in ALP activity (Figure 3, Supplementary Figure S2). Human chondrocytes exhibited more rapid calcification and sortilin and ALP increases compared with murine chondrocytes (Figure 3 and Supplementary Figure S2). The human chondrocytes were derived from OA cartilage and therefore originated from an already calcified environment, whereas the murine chondrocytes were isolated from neonatal knee joints. This explains the comparatively faster response of human OA chondrocytes to OM. Already on day 7, a significantly increased level of sortilin at the protein level was measurable in the human chondrocytes (Figure 3). After 21 days of calcification induction, this was followed by significantly increased ALP protein levels and activity (Figure 3). Sortilin is temporally upregulated upstream of ALP. Sortilin may thus be an inducer of ALP in the calcification process of OA. Sortilin-dependent upregulation of ALP has already been demonstrated in atherosclerosis. Sortilin induces both ALP and the calcification of hSMCs and the ejection of calcifying extracellular vesicles [17].
Previous studies demonstrated a close association between cellular senescence and atherosclerosis as well as OA [24,26,[28][29][30]42]. One study demonstrated that cellular senescence induces arterial calcification. Increased expression of ALP was found in the senescent hSMC of vessels [31]. In this study, senescent human chondrocytes showed upregulation of sortilin ( Figure 4). Sortilin is part of the calcification processes in OA, as described above (Figure 3).
The existing common features of extracellular matrix calcification in atherosclerosis and OA could be extended by this study. Senescence seems to be involved in the calcification process in both diseases. Sortilin provides an important contribution to calcification in OA and atherosclerosis. The exact relationship between sortilin and ALP induction in OA should be the subject of further research.
Our data indicates that sortilin and ALP are upregulated during cartilage calcification, which is associated with chondrocyte senescence and thus might contribute to the pathogenesis of osteoarthritis. Cellular senescence seems to induce sortilin expression.

Mouse Model
Knees of 8-week-old (young) and 45-week-old (old) mice were fixed overnight in 4% formaldehyde (Otto Fischar GmbH & Co. KG, Saarbrücken, Germany) at 4 • C, subsequently washed three times with PBS, and decalcified in EDTA at RT for at least 5 weeks. The tissue was embedded in paraffin, and 4-micrometer-thick sections were cut at the microtome.

Human Cartilage
Samples of knee cartilage were collected during the implantation of total knee arthroplasty as well as in unicondylar joint replacement at the Department of Orthopaedic Surgery of the University Hospital Magdeburg. The study was reviewed and approved by the Institutional Review Board (IRB) of the Medical School, Otto-von-Guericke University, Magdeburg (IRB No. 28/20). The patients/participants provided their written informed consent to participate in this study. OA severity was determined radiologically using the Kellgren-Lawrence score (KL-score). The subjects were ranked into mild (KL-score 1-2), moderate (KL-score 3), and severe (KL-score 4) OA. Knee cartilage from people who had died without a history of OA was used as a control. The absence of OA was assessed histologically by the OARSI score.

Chondrocyte Isolation and Cell Culture
Chondrocytes were isolated from the knee joints of 4-5-day-old C57Bl6 wt/wt mice and from human cartilage. Chondrocytes were isolated from murine cartilage tissue by digestion using 3 mg/mL collagenase D (type IV) (Worthington Biochemical Corporation, Lakewood, NJ, USA) for 45 min at 37 • C. The collagenase solution was then diluted 1:6 and incubated with the cartilage pieces over night at 37 • C. Human cartilage samples of patients undergoing knee replacement (KL 3-4) were cut into small pieces and incubated in 1 mg/mL pronase (Sigma-Aldrich, St. Louis, MO, USA) at 37 • C for 30 min. After removal of pronase, overnight digestion was performed with 1 mg/mL collagenase D (type IV) at 37 • C. The suspensions were applied to a cell strainer the next day and centrifuged at 400× g for 10 min (Heraeus, Megafuge 16R, Thermo Fischer Scientific, Waltham, MA, USA). The cells were cultured in chondrocyte medium (DMEM High Glucose (Sigma-Aldrich, St. Louis, MO, USA), +10% FCS, +1% penicillin/streptomycin, +1% sodium pyruvate (100 mM)) according to the planned experimental protocol.

RNA Analysis
For RNA extraction, human cartilage samples were minced and dissolved in 1 mL of Trizol (Qiagen, Hilden, Germany). The RNeasy Micro Kit (Qiagen, Hilden, Germany) was used according to the manufacturer's protocol. The mitomycin C-treated chondrocytes were harvested, and RNA extraction was performed using 500 µL of Trizol according to the manufacturer's protocol.
The amount and purity of the isolated RNA were determined photometrically, and 500 ng of RNA was reverse-transcribed into cDNA via reverse transcription using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fischer Scientific, Waltham, MA, USA). Gene expression was assessed by quantitative real-time polymerase chain reaction (RT-PCR) using the Applied Biosystems QuantStudio6 Flex Real-Time PCR System (Thermo Fischer Scientific, Waltham, MA, USA). RT-PCR was performed using PowerTrack SYBR Green Master Mix (Thermo Fischer Scientific, Waltham, MA, USA) according to the manufacturer's protocol and the following primers: Sortilin (forward AATGGCCTGTGGGTCCAA and reverse AGGTCAGCTTTGCAGGAGCC), p16 (forward CAACGCACCGAATAGTTACG and reverse ACCAGCGTGTCCAGGAAG), p21 (forward GGAGACTCTCAGGGTCGAAA and reverse CTTCCTGTGGGCGGATTA), and GAPDH served as housekeeper (forward CCCACTCCTCCACCTTTGAC and reverse AGGTCAGCTTTGCAGGAGCC). Absolute quantification was carried out using standard curves. Target gene expression was normalized to Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH).

Alkaline Phosphatase
ALP activity was measured in cell lysates using a colorimetric assay according to the manufacturer's protocol (ab83369, Abcam, Cambridge, MA, USA). Cells were harvested from sterile 24-well plates on days 1, 7, and 21 of treatment, resuspended in 50 µL assay buffer, and stored at −80 • C until activity determination. p-nitrophenyl phosphate (pNPP) was used as the enzyme substrate, which appears yellow after dephosphorylation by ALP. The assay was performed on a 96-well plate. After 60 min of light-protected incubation at 25 • C, the reaction was stopped by adding Stop-Solution, and absorbance was measured at 405 nm (Infinite F200 Pro, Tecan Group AG, Switzerland).

Statistics
All data were presented as mean ± SD. Unpaired t-tests were used for analyzing data comparing two groups for statistical significance. Data with more than two groups were analyzed by a repeated measures one-way analysis of variance (ANOVA) or a oneway ANOVA followed by a Dunnett's test as a post hoc test in the case of a statistically significant ANOVA result. Data that were investigated for more than one parameter were analyzed using a two-way ANOVA followed by a Sidak's test as a post hoc test in the case of a statistically significant ANOVA result. A Shapiro-Wilk normality test was performed to identify parametric or non-parametric data distributions. GraphPad Prism V.6.00 for Windows (GraphPad 8 Software, La Jolla, CA, USA) was used. Statistical significance was determined at a level of p ≤ 0.05.