Deletion of DYRK1A Accelerates Osteoarthritis Progression Through Suppression of EGFR-ERK Signaling

Dual-specificity tyrosine phosphorylation regulated kinase 1A (DYRK1A) signaling is involved in the dynamic balance of catabolism and anabolism in articular chondrocytes. This study aimed to investigate the roles and mechanism of DYRK1A in the pathogenesis of osteoarthritis (OA). The expressions of DYRK1A and its downstream signal epidermal growth factor receptor (EGFR) were detected in the cartilage of adult wild-type mice with destabilized medial meniscus (DMM) and articular cartilage of patients with OA. We measured the progression of osteoarthritis in chondrocyte-specific knockout DYRK1A(DYRK1A-cKO) mice after DMM surgery. Knee cartilage was histologically scored and assessed the effects of DYRK1A deletion on chondrocyte catabolism and anabolism. The effect of inhibiting EGFR signaling in chondrocytes from DYRK1A-cKO mice was analyzed. Trauma-induced OA mice and OA patients showed downregulation of DYRK1A and EGFR signaling pathways. Conditional DYRK1A deletion aggravates DMM-induced cartilage degeneration, reduces the thickness of the superficial cartilage, and increases the number of hypertrophic chondrocytes. The expression of collagen type II, p-ERK, and aggrecan was also downregulated, and the expression of collagen type X was upregulated in the articular cartilage of these mice. Our findings suggest that DYRK1A delays the progression of knee osteoarthritis in mice, at least in part, by maintaining EGFR-ERK signaling in articular chondrocytes.


INTRODUCTION
Osteoarthritis (OA) is the most common disabling joint disease, affecting about 10% of male and 18% of female among population over 60 years old.
important to further elucidate the pathological mechanism of OA [1,2].
When the epidermal growth factor receptor (EGFR) is activated by its ligand, EGFR forms homodimers with itself or heterodimers with other members of the ErbB family. Those dimers regulate various cellular functions, such as adhesion, survival, proliferation, differentiation, and migration [3,4]. Cartilage (Col2-Cre), specific EGFR knockout mice have fewer superficial chondrocytes, less boundary lubrication secretion, and weaker mechanical strength at cartilage surface than wild-type (WT) mice [5].
In the presence of defects in the articular cartilage surface, these mice significantly accelerated progression of knee OA following OA injury. Wei et al. further demonstrated that activating the EGFR signaling pathway can promote the expression of anabolic molecules in chondrocytes [6]. Pozo et al. found that dual-specificity tyrosine phosphorylation regulated kinase 1A (DYRK1A) prevent the endocytosis and degradation of EGFR. The inhibition of DYRK1A can promote the degradation of EGFR [7]. This suggests that DYRK1A is critical for the stability of EGFR.
DYRK1A, also named as Minibrain (MNB) due to its Drosophila melanogaster ortholog, is located on human chromosome 21 (HSA21). The gene includes 151 kb and 15 exons and encodes two main protein isoforms which compose 763 and 754 amino acids. In mammals, the DYRK subfamily members include DYRK1A, DYRK1B, DYRK2, DYRK3, and DYRK4. DYRK1A is generally expressed in human and mouse tissues [8,9].
Deshmukh et al. found that after chondrocytes were induced by interleukin (IL), knockdown of DYRK1A by siRNA resulted in decreased expression of IL-6, tumor necrosis factor-α, IL-1β, and MMP9. These results show inhibitory effect on cell inflammation [10]. However, Forristal et al. [11] found that the assembly of DREAM requires DYRK1A to phosphorylate LIN52 to complete. Mice lacking DREAM showed downregulation of cartilage proliferation and early death. Using drugs to inhibit expression of DYRK1A caused a similar downregulation of cartilage proliferation.
On the one hand, experiments at the cellular level cannot fully simulate the changes of diseases in vivo. On the other hand, DYRK1A is expressed in a variety of cells. Although articular cartilage is the main involved tissue in osteoarthritis, it includes subchondral bone and synovium. All joint components are involved in the pathogenesis of OA, so it is difficult to exclude the role of tissues other than articular cartilage in the pathogenesis of OA. More evidence is needed to convincingly demonstrate a direct role of DYRK1A in articular cartilage maintenance and its underlying mechanisms.
In this study, we conditionally induced DYRK1A deletion in chondrocytes from adult mice. We found that deletion of DYRK1A may inhibit the antagonism of endocytic degradation of EGFR, thereby inhibiting the promotion of EGFR to downstream extracellular signalregulated kinases (ERK) [12] and chondrocyte anabolic molecules such as Collagen II (Col II) and aggrecan, ultimately aggravating the degeneration of articular cartilage in adult mice.

Animals and Moudel of OA
DYRK1A Flox/Flox (DYRK1A fl/fl ) mice were purchased from The Jackson Laboratory. Col2a1-CreER T2 mice were purchased from Cyagen. DYRK1A fl/fl mice were crossed with DYRK1A fl/fl ; Col2a1-CreER T2 mice to obtain DYRK1A fl/fl ; Col2a1-CreER T2 mice (DYRK1Aconditional-knockout mice), and DYRK1A fl/fl mice (Cre-negative control mice). Male DYRK1A conditional knockout mice and Cre-negative mice were intraperitoneally injected with tamoxifen (Sigma-Aldrich) (1 mg/10 g/day for 5 days) at 2 months of age, then induction of OA in the right knee meniscus by destabilization of the medial meniscus (DMM) [13][14][15]. The left knee joints, which were received sham surgery with medial capsulotomy only, were in control group. The 2-monthold WT mice were subjected to DMM surgery on the right knee joint, and only the medial capsulotomy was performed on the left knee joint. Human articular cartilage samples were obtained from OA patients undergoing total knee replacement. The mean age of the patients was 60.5 years, and the baseline information of clinical data was provided (Supplemental Fig. 5 M). We got cartilage from the femoral condyle and cut it into 2 mm cubes for histology. All human and mouse protocols were approved by the Institutional Animal Care and Utilization Committee of the Second Affiliated Hospital of Chongqing Medical University. Experiments were performed according to approved guidelines. Written informed consent was obtained from all the subjects.

The μCT Analysis
In this experiment, a Swiss Scanco viva 40 μCT scanner was used to perform computed tomography scanning on mouse knee joint specimens. The scanning parameters were set as voltage 70 kV, current 113 mA, and scanning thickness 15 μm. All scans were performed under the same conditions. After the scan, the Scanco viva system analysis software was used to perform 3D reconstruction of the knee joint [16]. The reconstruction parameters were set to 0.8/1/220, and the part of the region was used for statistical analysis.

Histological Assessment
Mice were sacrificed; knee joints were fixed with 4% paraformaldehyde, decalcified with 4% EDTA, and embedded in paraffin. A continuous sagittal plane of the entire joint was obtained by collecting 5 mm sections at 50-mm intervals. Sections were stained with Fast green/Safranin O and hematoxylin and eosin (HE) for histological analysis. The scintillation intensity of femoral and tibial growth plates was used as the intraassay control. Immunohistochemical analysis of intermediate sections was performed. Using a modified Mankin Score (mRS) system, the histological changes of the knee joints stained with H&E were scored on a scale of 0-14. The highest score for the same sample was calculated to assess the severity of cartilage destruction. Histological changes in knee joints stained with Fast green/Safranin O were scored on a scale of 0-6 as recommended by the Osteoarthritis Institute International (OARSI), and the sum of the 4 highest scores for each specimen was calculated. Assess the severity of cartilage destruction. Scoring was performed by 3 independent investigators.
Using liquid DAB substrate chromogen system (Dako) to developed Immunoreactions. Then, using hematoxylin stained nuclei and saturated lithium carbonate returned to blue. Using the ImageJ Version 4 software (Media Cybernetics) counted the number of positive cells in articular cartilage central areas.

Bioinformatic Data
In the GEO database, osteoarthritis was used as the keyword, and Homo sapiens was selected as the species for retrieval. The experimental grouping and the number of samples were combined, and finally, GSE117999 was selected for analysis.

Statistical Analysis
This study used the GraphPad Prism version 8 analyzed data. Results are expressed as a mean ± SD. Differences in means were assessed by Student's t-test. P values less than 0.05 were considered significant.

DYRK1A Negatively Correlated with the Pathological Progression of OA in Adult Mice and Humans
To explore the role of DYRK1A in OA, DMM surgery was performed on the knee joints of 2-month-old WT mice to simulate OA model and mice received sham surgery as control (Sham) [17]. After DMM operation, paraffin sections were performed on the knee joints of mice, and then, HE and Fast green/Safranin O staining were performed which were evaluated by the modified Makin score and OARSI score, respectively [18]. The results showed that compared with the sham operation group, the damage degree of the DMM group was significantly increased, indicating the successful establishment of OA model (Supplemental Fig. 5A-D). The expression changes of DYRK1A in the knee joints of WT mice at 1 month, 2 months, and 3 months after DMM surgery were detected by immunohistochemistry staining. At 2 months and 3 months after DMM surgery, the expression of DYRK1A in the OA cartilage was significantly downregulated compared with sham surgery group. These differences show statistical significance. The expression of DYRK1A also showed differences with the progression of OA. Compared with the knee joint of mice at 1 month after operation, the expression of DYRK1A in the knee joints of mice at 2 months and 3 months after surgery was further reduced and had a statistical difference ( Fig. 1A, B). We collected the articular cartilage of patients requiring joint replacement surgery due to OA (most involved) and segment of relatively complete OA cartilage as control group (least involved) [19]. The results of hematoxylin and eosin (HE) (Supplemental Fig. 5E, F), Fast green/Safranin O staining, the modified Makin score, and OARSI score confirmed more severe cartilage degradation than that of the least involved group (Fig. 1C, D). We found that DYRK1A was significantly downregulated in the most involved group compared to the least involved group (Fig. 1E, F).
Our data showed that DYRK1A was downregulated during the pathogenesis of OA and may be negatively correlated with disease progression.

Accelerated Articular Cartilage Destruction in Adult Mice with Chondrocyte-Specific Deletion of DYRK1A
Five days after tamoxifen injection, the expression of DYRK1A protein in articular chondrocyte specific deletion of DYRK1A (DYRK1A-cKO) mice was significantly reduced (Fig. 1G-I), indicating that the DYRK1A gene was effectively deleted. In adult DYRK1A-cKO mice, μCT scans were performed 3 months after DMM surgery (Fig. 1J). Results of µCT analysis showed that compared with Cre-negative mice, DYRK1A-cKO mice exhibited more obvious joint deformity with osteophyte formation, subchondral sclerosis, and joint space narrowing of the knee joint.
It was observed that OA was significantly aggravated in adult mice with DYRK1A-cKO compared to Crenegative mice. However, the subchondral bone plate bone volume/tissue volume (BV/TV), trabecular separation (Tb.Sp), mean of osteophyte size, and trabecular thickness (Tb.Th) were statistically analyzed, and there was no significant difference (Supplemental Fig. 5G-K). The effect of DYRK1A deficiency on the development of osteoarthritis was subsequently assessed. One month after DMM, DYRK1A-cKO mice had early symptoms of OA, such as unevenness, small pits on the joint surface, and loss of cartilage and proteoglycan, which could be detected by HE staining or by Fast green/Safranin O staining. These phenotypes were more severe than that in Cre-negative mice. Two months after DMM surgery, compared with Cre-negative mice, DYRK1A-cKO mice showed more extensive destruction of most areas of articular cartilage and partial exposure of subchondral bone. Three months after DMM surgery, compared with Cre-negative mice, DYRK1A-cKO mice had further aggravated articular cartilage destruction and extensive exposure of subchondral bone ( Fig. 2A). Histological analysis showed that the average thickness of articular cartilage and the percentage of superficial chondrocytes in DYRK1A-cKO mice were smaller than those in Cre-negative mice at 1, 2, and 3 months after surgery, and the number of hypertrophic chondrocytes was higher than that of Crenegative mice (Fig. 2C, E, F). At 2 and 3 months after DMM surgery, using modified Makin score and OARSI score, the degree of joint damage in DYRK1A-cKO mice was more severe than that in Cre-negative mice, and DYRK1A-cKO mice after DMM surgery arthritis progression worsened over time (Fig. 2B, D).
These results suggest that DYRK1A-cKO mice exacerbates the DMM surgery-induced osteoarthritis phenotype.

Chondrocyte-Specific Deletion of DYRK1A Disrupts the Balance of Anabolism and Catabolism in Mouse Articular Cartilage
The first line of defense against OA initiation is the superficial layer. The superficial layer spans the first 10-20% of the full thickness articular cartilage and contains dense collagen fibrils and low levels of aggrecan [20,21]. It produces lubricant proteins, harbors cartilage precursor cells, resists shear stress, acts as a sliding surface, and plays an important role in protecting chondrocytes. In OA, degenerative changes begin with decreased cellular anabolism and increased catabolism [5]. DYRK1A has the effect on promoting cell proliferation and cell anabolism. DYRK1A is abundantly expressed in superficial chondrocytes, and deletion of DYRK1A aggravates the progression of OA [22]. To elucidate the mechanism of accelerated OA in DYRK1A-cKO mice, we detected the protein levels related to anabolism and catabolism in chondrocytes by immunohistochemistry. Collagen (Col II) and aggrecan are important anabolic markers in the extracellular matrix of cartilage [23]. Compared with Cre-negative mice, DYRK1A-cKO mice showed lower levels of immunoreactivity for Col II in articular cartilage and aggrecan (Figs. 3A, B; 4C, D). Collagen (Col X) is a marker of chondrocyte hypertrophy, and MMP13 can promote catabolism [24]. Compared with Cre-negative mice, the immunoreactivity of Col X and MMP13 (Fig. 3C-F) in articular cartilage of DYRK1A-cKO mice was respectively increased. This was associated with reduced Fast green/Safranin O staining of articular cartilage in mutant mice. These results suggest that DYRK1A deficiency may directly promote catabolic activity and inhibit anabolic activity of articular chondrocytes. This disrupts the homeostasis of articular cartilage in adult mice.

Fig. 2 A, B
In mice with cartilage-specific deletion of DYRK1A, OA damage was aggravated and positively correlated with time. Hematoxylin and eosin (H&E) staining was used to detect the degree of joint damage in Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. B Modified Mankin score of H&E staining in Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. C, D The hypertrophic chondrocytes of mice OA-injured joints increased, and the superficial cartilage decreased. Fast green/Safranin O staining to detect the degree of superficial cartilage damage and hypertrophic chondrocytes in Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. The parts selected by the dotted boxes show higher-magnification views. D OARSI score of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, 3 months after OA modeling which received Fast green/Safranin O staining. E Average thicknesses of total articular cartilage in Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. F The percentage of chondrocytes in the superficial layer of cartilage. Horizontal lines and error bars show the mean ± SD (n = 3 mice per group). *P < 0.05.

Decreased EGFR-ERK Signaling and Aggrecan Expression Mediate the Development of DYRK1A-Deficient Osteoarthritis
EGFR-ERK signaling is important for maintaining the normal function of superficial cartilage during skeletal development. Furthermore, EGFR-ERK signaling is thought to be an anabolic mediator in articular cartilage destruction [5]. Pozo et al. observed that DYRK1A can prevent the endocytic degradation of EGFR to maintain the activity of EGFR and downstream signaling [7]. DYRK1A inhibits the endocytic degradation of EGFR and maintains the phosphorylation level of downstream ERK by EGFR to promote the expression of p-ERK, aggrecan, Col II, and other chondrocyte anabolic factors [25,26]. Using immunohistochemistry to detect EGFR, EGFR was downregulated in knee cartilage of DYRK1A-cKO mice compared with Cre-negative mice at 3 months after DMM modeling (Fig. 3G, H). The expression of EGFR in the knee joints of WT mice at 3 months after DMM modeling was significantly lower than that of mice in the Sham group (Fig. 4A, B). We found that EGFR was downregulated in the most involved group compared to the least involved group in human cartilage (Fig. 4E,  F). The cartilage of DYRK1A-cKO mice 3 months after DMM was selected for Western blot detection of EGFR expression. We found that the expression of EGFR in DYRK1A-cKO mice was decreased compared with that of Cre-negative mice, but there was no statistical significance. This may be related to the fact that other tissues besides cartilage, such as synovium and meniscus, also express EGFR (Fig. 4K, L). Using GEO2R analysis, it was found that EGFR in cartilage of osteoarthritis patients was significantly downregulated compared with normal human cartilage (Fig. 4I, J). Therefore, we investigated whether EGFR-ERK signaling mediates the aggravation of OA caused by DYRK1A deficiency. We found that p-ERK levels were significantly downregulated in knee cartilage of DYRK1A-cKO mice compared with Cre-negative mice at 3 months after DMM surgery (Fig. 4G, H), and aggrecan expression was also significantly decreased (Fig. 4C, D). Similarly, we selected the cartilage of DYRK1A-cKO mice at 3 months after DMM surgery to detect the expression of p-ERK by Western blot. We found that the expression of p-ERK was reduced in DYRK1A-cKO mice compared to Cre-negative mice, but that was not statistically significant. Besides articular cartilage, experimental samples contain multiple components. In the future, cartilage should be extracted for verification (Fig. 4N, O). This suggests that DYRK1A deletion may attenuate the expression of cellular anabolic markers such as aggrecan and Col II through EGFR-ERK signaling. Therefore, we propose a possible mechanism for this phenomenon (Fig. 4M).

DISCUSSION
Osteoarthritis (OA) used to be considered as a wear and tear disorder [27]. Now people understood it is a more complex pathological process affecting multiple joints and whole joint structures [28].
DYRKs belong to the CMGC family of cyclindependent protein kinases (CDKs), mitogen-activated protein kinases (MAPKs), glycogen synthase proteins (GSKs), and CDC2-like proteins (CLKs). Although the relationship between the signaling pathways of the CDK and MAPK families and cellular anabolism has been extensively studied, little is known about how DYRK is linked to its OA pathological process [29,30].
Recently, DYRK1A signaling was found to be involved in the pathogenesis of OA by affecting chondrocyte metabolism [31]. Deshmukh et al. found that lorecivivint alleviated arthritis progression by inhibiting CLK2 and DYRK1A [10]. Forristal et al. found that drug inhibition of DYRK1A downregulated chondrocyte proliferation [11]. Although these findings are controversial, DYRK1A has been shown to regulate articular cartilage metabolism. This suggests that targeting DYRK1A in articular cartilage is a good strategy for the treatment of OA if we accurately understand the role of DYRK1A in anabolic and catabolic balance in adult articular cartilage. We therefore investigated the direct role of DYRK1A in the maintenance of adult articular cartilage using mice with a specific deletion of DYRK1A in chondrocytes (DYRK1A-cKO). We found that loss of DYRK1A aggravated articular cartilage destruction mainly by attenuating superficial chondrocyte stability and chondrocyte anabolic activity.
Articular cartilage is an anisotropic viscoelastic tissue which can be functionally and structurally divided into four zones (superficial, medium, deep, calcified). These differ in terms of extracellular matrix (ECM) composition, density, collagen fiber aggregation, phenotype, and chondrocyte activity. The surface layer provides a smooth sliding surface, ensures lubrication and low wear of the joint, and exhibits lower tensile strength and hardness than the deep layer in terms of compressive properties. However, damage to the surface area may lead to rapid wear of articular cartilage and subsequent cartilage rupture [32,33]. The surface layer of cartilage contains cartilage precursor cells. The cells are responsible for forming the rest of the articular cartilage during development [34]. In our study, we found that DYRK1A can stabilize cartilage surface cells and promote the expression of cellular anabolic activity markers such as Col II and aggrecan. These results suggest that DYRK1A delays the development of OA by stabilizing the surface cells of articular cartilage. Damage to articular cartilage may also be associated with chondrocyte catabolic activity [35]. In the progression of OA, preventing chondrocyte hypertrophy can effectively alleviate cartilage damage [36]. The definition of OA chondrocyte hypertrophy is mainly based on molecular characteristics, especially the expression of key genes such as Col X genes. This molecule is also a marker of chondrocyte catabolic activity. Hypertrophic chondrocytes located in different cartilage regions are exposed to an inappropriate stromal and endocrine/paracrine environment that is even less favorable for survival [37]. We found that in the absence of DYRK1A, the expression of Col X and MMP13 was downregulated in articular chondrocytes, suggesting that the loss of DYRK1A leading to reduced anabolic effects and may relatively promote chondrocyte hypertrophy and the catabolism of articular cartilage, thereby aggravating articular cartilage degenerate. However, the detailed mechanism of the anabolic effects of DYRK1A signaling on chondrocytes is unclear.
Recently, it was reported that DYRK1A can act on EGFR and maintain the stability of EGFR [38,39]. We used the GEO database and selected the GSE117999 dataset for pre-analysis [40,41]. After normalized comparative analysis of differential genes, it was found that compared with normal human cartilage, EGFR was significantly downregulated in cartilage of patients with OA. In order to further verify the correlation between EGFR and OA, we found similar phenomena in WT mouse and human cartilage tissue after DMM surgery comparing with normal tissue. These findings suggest that EGFR plays an important role in the development of OA. EGFR can activate the downstream ERK signaling pathway [42]. The ERK1/2 signaling pathway plays a key role in chondrocyte anabolism, and it is activated to produce cellular anabolic markers such as p-ERK and aggrecan to alleviate arthritis progression [25]. In addition, transgenic mice with overactivated EGFR signaling reduced articular cartilage damage and increased anabolic markers such as p-ERK in chondrocytes [6]. In contrast, depletion of EGFR in chondrocytes aggravated DMM surgery-induced the development of OA [5]. Our study found that in the absence of DYRK1A, the stability of EGFR was also lost. Also, the downregulation of EGFR led to the downregulation of downstream chondrocyte synthesis markers, such as p-ERK and aggrecan and the cellular catabolic markers MMP13 raised. These results indicate that DYRK1A can enhance chondrocyte anabolism and attenuate chondrocyte catabolism in part by stabilizing the expression of EGFR molecules and continuously activating downstream p-ERK signaling. Ultimately it showed to alleviate the progression of osteoarthritis.
There are conflict on the relationship between DYRK1A and OA progression [10,11]. On the one hand, the environment of DYRK1A deletion may not be well simulated due to in vitro experiments and inhibitor interference. On the other hand, experimental mice and cells may not fully mimic the human disease environment. Therefore, our study selected chondrocytespecific knockout DYRK1A mice and some human cartilage tissues as the research objects. However, our experiments did not further investigate the effect of specifically activating DYRK1A in chondrocytes on osteoarthritis, so further studies on this molecule are required. In addition, the EGFR-ERK signaling pathway may have two functions of promoting decomposition and synthesis. Studies have found that intra-articular administration of EGFR inhibitors inhibits the progression of OA in a rabbit meniscectomy model [43]. However, another study pointed out that mice with cartilagespecific EGFR deficiency can accelerate the progression of OA, and cartilage-specific activation of EGFR can resist cartilage degeneration and other signs of OA [6]. Therefore, we speculate that there may be a relative balance in the influence of EGFR-ERK signaling pathway on the decomposition and synthesis of chondrocytes under physiological conditions. At different stages of the disease, one of the abilities will be relatively stronger, Fig. 3 A, B Col II expression was also further reduced in deletion of DYRK1A mice after OA modeling. Immunohistochemical detection of Col II expression in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. B Percentage of Col II-positive cells in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. C, D Col X expression was also further increased in deletion of DYRK1A mice after OA modeling. Immunohistochemical detection of Col X expression in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. D Percentage of Col X-positive cells in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. E, F MMP13 expression was also further increased in deletion of DYRK1A mice after OA modeling. Immunohistochemical detection of MMP13 expression in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. F Percentage of MMP13-positive cells in articular cartilage of Crenegative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. G, H EGFR expression was also further reduced in deletion of DYRK1A mice after OA modeling. Immunohistochemical detection of EGFR expression in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. H Percentage of EGFR-positive cells in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 1 month, 2 months, and 3 months after OA modeling. Horizontal lines and error bars show the mean ± SD (n = 3 mice per group). *P < 0.05. resulting in the overall signaling pathway showing a role in promoting or inhibiting the progression of OA. Therefore, the study of different stages of OA progression is helpful to better understand the occurrence and development of the disease. Taken together, by using both methods of human tissue and induced chondrocyte-specific inactivation, we found that DYRK1A deficiency in chondrocytes exacerbates the development of DMM-induced OA. These findings strongly suggest that DYRK1A may have a Fig. 4 A, B EGFR expression was reduced in WT mice with OA. Immunohistochemical detection of EGFR expression in articular cartilage of WT mice at 3 months after sham surgery or OA modeling, respectively. B EGFR-positive cell rate in articular cartilage of WT mice 3 months after sham operation or OA modeling. C, D In DYRK1A-deficient arthritic mice, aggrecan expression was further reduced. Immunohistochemical detection of aggrecan expression in articular cartilage of Cre-negative mice and DYRK1A-cKO mice 3 months after OA modeling. D Percentage of aggrecanpositive cells in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 3 months after OA modeling. E, F In human cartilage, OA injury downregulates the expression of EGFR. Immunohistochemical detection of EGFR expression in cartilage of control group and OA patients. F Percentage of EGFR-positive cells in control and OA patient cartilage. G, H The p-ERK expression was also further reduced in DYRK1A-deficient arthritic mice. Immunohistochemical detection of p-ERK expression in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 3 months after OA modeling. H Percentage of p-ERK positive cells in articular cartilage of Cre-negative mice and DYRK1A-cKO mice at 3 months after OA modeling. I, J GEO data analysis showed that compared with normal human cartilage, EGFR expression was decreased in chondrocytes of OA patients. GSE117999 was analyzed by the GEO 2R analysis software to obtain differential gene (I) and EGFR expression (J) in normal human cartilage and cartilage of OA patients. K, L EGFR expression may also be reduced in DYRK1A-deficient arthritic mice. Western blot detection of EGFR expression in knee joints of Cre-negative mice and DYRK1A-cKO mice at 3 months after OA modeling. L After 3 months of OA modeling, Cre-negative mice and DYRK1A-cKO mice had relatively denstiy of EGFR in knee joints. M Hypothesis map of the mechanism by which DYRK1A affects OA progression. N, O p-ERK expression may also be reduced in DYRK1A-deficient arthritic mice. Western blot detection of p-ERK expression in knee joints of Cre-negative mice and DYRK1A-cKO mice at 3 months after OA modeling. L After 3 months of OA modeling, Cre-negative mice and DYRK1A-cKO mice had relatively density of p-ERK in knee joints. Horizontal lines and error bars show the mean ± SD (n = 3 mice per group). *P < 0.05. chondroprotective role in adult mice and human tissues, which will help us to understand the pathogenesis of OA and develop effective therapeutic strategies to against this common disease.