Abstract

   Osteoarthritis (OA) is a degenerative joint disease characterized by
   the metabolic dysfunction of chondrocytes. A promising therapeutic
   strategy for OA involves suppressing the catabolism of the chondrocyte
   and promoting its anabolism to restore joint homeostasis. Here, it is
   demonstrated that Catalpol, a natural compound, can promote chondrocyte
   anabolic and proliferation, while inhibiting the catabolic activities
   and oxidative stress, thereby maintaining the dynamic balance of the
   extracellular matrix and alleviating inflammation‐induced cartilage
   damage. Mechanistically, it has been discovered that Catalpol acts as a
   direct inhibitor of heat shock protein 90β (Hsp90β), and the amino
   acids ASP88, THR179, ASP49, and ASN46 of N‐terminal domain‐Hsp90β are
   confirmed as the binding sites for Catalpol. Knockdown of Hsp90β in
   primary chondrocytes demonstrates a similar biological effect as
   Catalpol treatment. Moreover, to develop a nanoparticle‐based
   interventional platform for OA management, biodegradable mesoporous
   silica nanoparticles (bMSN) are prepared to load Catalpol (Ca‐bMSN).
   The engineered Ca‐bMSN is able to penetrate into the chondrocytes,
   prolong retention in the joint space, and mitigate OA progression.
   These findings shed light on a potential mechanism by which Catalpol
   modulates chondrocyte metabolism, offering a promising therapeutic
   strategy for OA treatment.

   Keywords: bMSN, Catalpol, chondrocyte metabolism, Hsp90β,
   osteoarthritis
     __________________________________________________________________

   The schematic diagram delineates the chondroprotective effects of
   Ca‐bMSN in OA models. The upper portion of the diagram depicts the
   synthesis process of Ca‐bMSN. The lower section illustrates the
   post‐injection of Ca‐bMSN in joint cavity, catalpol released and
   subsequently mediates the reduction of oxidative stress, ECM
   catabolism, apoptosis, and inflammation, while promoting ECM anabolism,
   thereby contributing to cartilage protection.

   graphic file with name ADVS-12-2503909-g005.jpg

1. Introduction

   Osteoarthritis (OA), a degenerative joint disease characterized by
   cartilage breakdown, affects an estimated 250 million people worldwide.
   Notably, a staggering 80% of individuals over the age of 75 in China
   are reported to be diagnosed with OA.^[ [52]^1 ^] The typical clinical
   features of OA include joint pain, swelling, stiffness, and limited
   range of motion of joint, significantly impacting patient quality of
   life. As a degenerative disease, OA is usually caused by ageing,
   chronic strain, and trauma. The excessive wear of articular cartilage
   surfaces induced by chronic strain was the most common predisposing
   risk factor for OA. Nonsteroidal anti‐inflammatory drugs (NSAIDs) are
   currently the mainstay of treatment for OA. However, they have not been
   shown to stop cartilage degeneration or promote its repair.^[ [53]^2 ^]
   In addition, their use is often linked to significant kidney and heart
   problems. Therefore, there is a critical need for a safe and effective
   treatment that can slow the progression of OA.

   Chondrocytes represent the sole cellular component within the articular
   cartilage, enclosed by an extracellular matrix (ECM) abundant in
   proteoglycans and type II collagen to maintain joint function.^[ [54]^3
   ^] The excessive degradation contributes to the breakdown of the ECM is
   a hallmark feature of OA. Chondrocytes are responsible for maintaining
   cartilage health through a delicate balance of synthesis and
   degradation of the ECM, achieved by regulating anabolic and catabolic
   activities.^[ [55]^4 ^] Anabolic processes primarily focus on the
   synthesis of ECM components by upregulating cartilage anabolic factors
   like COL2A1 and ACAN in chondrocytes. Conversely, the catabolic
   phenotype is characterized by the abnormal production of ECM‐degrading
   enzymes, such as matrix metalloproteinase 3, 9, 13 (Mmp3, Mmp9, Mmp13)
   and a disintegrin and metalloproteinase with thrombospondin motifs 4, 5
   (Adamts4, Adamts5), by chondrocytes. During OA progression,
   pro‐inflammatory cytokines like interleukin‐1β (IL‐1β) and tumor
   necrosis factor‐α (TNF‐α) emerge as key players. These mediators induce
   chondrocyte dedifferentiation and promote oxidative stress, ultimately
   disrupting the balance between anabolic and catabolic processes. This
   imbalance can lead to chondrocyte apoptosis, further accelerating
   cartilage degeneration.^[ [56]^5 ^] Therefore, therapeutic strategies
   that enhance chondrocyte anabolic activity while inhibiting catabolic
   processes hold significant promise for managing OA.

   Catalpol, an iridoid glycoside derived from the root of Rehmannia,
   exhibits significant therapeutic potential for various diseases due to
   its anti‐inflammatory, antioxidant, and anti‐apoptotic properties.^[
   [57]^6 ^] An increasing body of evidence suggests that Catalpol
   significantly reduces Mmp3, Mmp9, and Adamts5 and increases
   mitochondrial biogenesis, suggesting a potential anti‐cartilage
   degradation effect in ATDC5 cells and rat chondrocytes.^[ [58]^7 ,
   [59]^8 , [60]^9 ^] However, these studies have only focused on the
   effects of Catalpol on the breakdown of cartilage in chondrocytes.
   Further confirmation is needed to determine whether Catalpol can also
   promote cartilage growth. In addition, the exact mechanism and target
   of Catalpol in OA treatment remain unclear.

   The intra‐articular (IA) injection of small molecule drugs is an
   effective way to treat knee OA. However, this method suffers from rapid
   clearance from the joint cavity due to interaction with synovial fluid
   constituents, significantly reducing therapeutic efficacy. To address
   this, drug‐delivery systems were developed. Mesoporous silica
   nanoparticles (MSN) are promising drug delivery carriers due to their
   adjustable pore size, high porosity, excellent biocompatibility, and
   convenient drug loading.^[ [61]^10 ^] MSNs protect drugs from premature
   degradation before release and provide sustained release, enhancing
   therapeutic effect and reducing side effects.^[ [62]^11 ^] However, the
   skeleton structure of “─Si─O─Si─” is stable, and it may be extremely
   concentrated in organs of body, such as liver, spleen and bladder,
   which may cause a serious problem such as inflammatory reaction,
   oxidative damage, and organ fibrosis. Biodegradable MSN (bMSN) not only
   retains the advantage of drug release performance, but also can be
   degraded rapidly in vivo, which can avoid the accumulation of
   cytotoxicity. Although studies have shown that bMSNs may have the
   potential to be used as drug delivery vehicles of drugs for OA therapy,
   the evidence of their effectiveness in the OA model is still lacking.^[
   [63]^12 ^]

   In this study, bMSN was synthesized to load Catalpol (Ca‐bMSN). The
   effects of Catalpol, bMSN, and Ca‐bMSN on chondrocytes were then
   investigated in both OA chondrocytes and OA animal models. In addition,
   the molecular targets and underlying mechanism of Catalpol on
   regulating OA were investigated by a series of experiments such as drug
   affinity responsive target stability (DARTS), cellular thermal shift
   assay (CETSA), isothermal titration calorimetry (ITC), molecular
   dynamics (MD), and small interfering RNAs (siRNAs) assays. This study
   offered the definitive mechanism for the cartilage protective effects
   of Catalpol and provided a promising combinatory approach for OA.

2. Results

2.1. Catalpol Regulates Chondrocyte Proliferation and Metabolism

   Chondrocytes, the only cell type residing in cartilage tissue, play a
   crucial role in regulating ECM homeostasis. Therefore, we initially
   assessed the influence of Catalpol (Figure [64]1A) on the proliferation
   of IL‐1β induced primary mouse chondrocytes at 24, 48, and 72 h using
   varying Catalpol concentrations. As shown in Figure [65]1B, Catalpol
   significantly promoted chondrocyte proliferation, with the most
   effective concentration being 20 × 10^−6 m at 24 and 48 h. We further
   evaluated the effect of Catalpol on chondrocyte anabolism and
   catabolism. In an OA cell model induced by 10 ng mL^−1 IL‐1β, Catalpol
   at concentrations of 10, 20, and 40 × 10^−6 m effectively stimulated
   the expression of anabolic marker genes (Col2a1, Acan, Comp)
   (Figure [66]1C). Conversely, Catalpol reduced the expression of
   catabolic marker genes (Mmp9, Mmp13, Adamts4) (Figure [67]1C). This
   effect was confirmed at the protein level, with Catalpol increasing
   COL2A1 and ACAN expression while decreasing MMP13 and ADAMTS4
   (Figure [68]1D,E). Similar results were observed in human chondrocytes
   (Figure [69]S1A,B, Supporting Information). Alcian blue staining
   indicated that Catalpol treatment improved the expression of acidic
   polysaccharides in IL‐1β‐induced OA chondrocytes compared to the
   control group (Figure [70]1F). These findings collectively suggest that
   Catalpol exerts a protective effect on chondrocytes by facilitating
   their proliferation and anabolic activity while inhibiting their
   catabolic processes.

Figure 1.

   Figure 1
   [71]Open in a new tab

   Catalpol promotes anabolism in primary chondrocytes and inhibits
   catabolism. A) Structure of Catalpol. B) The effect of different
   concentrations Catalpol on IL‐1β induced primary chondrocytes viability
   was detected by CCK‐8 assay at 24, 48, and 72 h, respectively (n = 5,
   *p < 0.05 and **p < 0.01). C) qRT‐PCR analysis of the expression level
   of anabolism marker genes (Col2a1, Acan, Comp) and catabolism marker
   genes (Mmp9, Mmp13, Adamts4) in chondrocytes treated with IL‐1β or
   IL‐1β combined with different concentrations of Catalpol at 24 h (n =
   3, *p < 0.05, ## or **p < 0.01, ###p < 0.001 when compared with control
   group and IL‐1β group, respectively). D) After treatment with IL‐1β or
   IL‐1β and combined with different concentrations of Catalpol for 24 h,
   the expression of COL2A1, ACAN, MMP13, and ADAMTS4 in the primary
   chondrocytes was analyzed using Western blot assay. The quantification
   result is shown at the bottom (n = 3, ## or **p < 0.01 when compared
   with control group and IL‐1β group, respectively). E) After treatment
   with different concentrations of Catalpol for 24 h, the expression of
   COL2A1 in the primary chondrocytes was analyzed using
   immunofluorescence assay. The quantification result is shown at right.
   Scale bar = 400 µm (n = 3, *p < 0.05, ##p < 0.001 when compared with
   control group and IL‐1β group, respectively). F) The level of acidic
   polysaccharides in the primary chondrocytes under the treatment with
   different concentrations of Catalpol for day 7 were analyzed by Alcian
   blue staining. The quantification result is shown at right (n = 3, *p
   < 0.05, ##p < 0.01, ***p < 0.001 when compared with control group and
   IL‐1β group, respectively). Data are shown as mean ± SEM. P‐values were
   obtained by one‐way ANOVA with multiple comparisons.

2.2. Anti‑oxidative Effects of Catalpol on the Osteoarthritic Chondrocytes

   Oxidative stress and apoptosis are the key factors in the decrease of
   chondrocytes and the loss of cartilage. Therefore, we investigated
   whether Catalpol possesses anti‐oxidative effects in OA cells. As shown
   in Figure [72]2A,B, Catalpol treatment significantly reduced excessive
   reactive oxygen species (ROS), malondialdehyde (MDA), and glutathione
   disulfide (GSSG) levels in the OA cell model at concentrations of 20 ×
   10^−6 and 40 × 10^−6 m, while significantly increasing the level of
   antioxidant enzyme superoxide dismutase (SOD). Notably, Catalpol
   displayed the most significant increase in total antioxidant capacity
   at 20 × 10^−6 m. Similarly, Catalpol inhibited IL‐1β‐induced ROS
   increase in human chondrocytes (Figure [73]S2, Supporting Information).
   In addition, Catalpol treatment maintained mitochondrial membrane
   potential and increased ATP production compared to the untreated OA
   cell model (Figure [74]2C,D). We further investigated the influence of
   Catalpol (10, 20, 40 × 10^−6 m) on OA cell apoptosis. Flow cytometric
   analysis results revealed that Catalpol reversed IL‐1β‐induced
   apoptosis (16.53%) in a dose‐dependent manner, with reduced percentages
   at 10 × 10^−6 m (9.96%), 20 × 10^−6 m (6.91%), and 40 × 10^−6 m (5.87%)
   (Figure [75]2E). Collectively, these results indicate that Catalpol has
   the ability to suppress the generation of ROS and decrease chondrocyte
   apoptosis, thereby enhancing chondrocyte proliferation, anabolism and
   ultimately promoting cartilage injury repair.

Figure 2.

   Figure 2
   [76]Open in a new tab

   Protective effect of Catalpol on chondrocytes against oxidative stress
   injury induced by IL‐1β. A) ROS production in chondrocytes was assessed
   through the utilization of the ROS probe DCFH‐DA under the treatment
   with IL‐1β or different concentrations of Catalpol. Fluorescence
   microscopy was used to capture images, while flow cytometric analysis
   was conducted to quantify the results. Scale bar = 400 µm. B) The level
   of SOD, MDA, GSSG, and total antioxidant capacity in chondrocytes after
   treated with IL‐1β or different concentrations of Catalpol were
   detected using specific kits (n = 3, *p < 0.05, ## or **p < 0.01, ###
   or ***p < 0.001 when compared with control group and IL‐1β group,
   respectively). C) The concentration of ATP in the in chondrocytes under
   the treatment with IL‐1β or different concentrations of Catalpol was
   detected by ATP assay kit (n = 3, *p < 0.05, **p < 0.01, ### or ***p
   < 0.001 when compared with control group and IL‐1β group,
   respectively). D) Mitochondrial membrane potential. Red, JC‐1
   aggregates; green, JC‐1 monomer. Scale bar = 100 µm (n = 3, *p < 0.05,
   **p < 0.01, ### or ***p < 0.001 when compared with control group and
   IL‐1β group, respectively). E) Cell apoptosis was detected by flow
   cytometry under the treatment with IL‐1β or different concentrations of
   Catalpol (n = 3, ### or ***p < 0.001 when compared with control group
   and IL‐1β group, respectively). Data are shown as mean ± SEM. P‐values
   were obtained by one‐way ANOVA with multiple comparisons.

2.3. Preparation and Physicochemical Characterization of Ca‐bMSN

   Next, we attempted to evaluate the protective effect of Catalpol on OA
   in vivo. Injecting drugs into the joint cavity is a highly effective
   clinical treatment strategy with fewer side effects. However, the
   efficacy of drugs may be reduced due to their rapid clearance from the
   joint cavity. Thus, to prolong the action time of Catalpol in vivo, the
   bMSNs—large surface area, excellent biocompatibility, and high drug
   loading capacity—were used to deliver Catalpol. Figure [77]3A
   schematically illustrates the bMSN synthesis process. Transmission
   electron microscopy (TEM) imaging revealed well‐distributed, uniformly
   sized, and spherical bMSNs (Figure [78]3B). Fourier transform infrared
   (FTIR) spectroscopy confirmed the presence of the characteristic
   Si─O─Si bond in bMSNs (Figure [79]3C). The bMSNs exhibited a particle
   size of ≈200 nm and a zeta potential of −12.42 mV (Figure [80]3D).
   Notably, the lower zeta potential of Ca‐bMSN (31.3 mV) compared to bMSN
   (Figure [81]3E) indicated successful loading of Catalpol onto the
   bMSNs. High‐performance liquid chromatography (HPLC) analysis
   determined a drug loading efficiency (LE) of 35% (Figure [82]S3,
   Supporting Information).

Figure 3.

   Figure 3
   [83]Open in a new tab

   Preparation and physicochemical characterization of bMSN and Ca‐bMSN.
   A) Schematic of bMSN synthesis processes. B) TEM images of bMSN. Scale
   bar = 200 nm. C) FTIR spectra of bMSN. D) Size of bMSN and Ca‐bMSN. E)
   Potential of bMSN and Ca‐bMSN.

2.4. Assessment of the Cellular Uptake and Chondroprotective Effect of
Ca‐bMSN In Vitro

   Before in vivo experiments, we incubated chondrocytes with fluorescein
   isothiocyanate (FITC)‐modified bMSN and Ca‐bMSN. Laser confocal
   microscopy revealed that both bMSN and Ca‐bMSN were efficiently taken
   up by mouse chondrocytes that were treated with IL‐1β within 4 h, as
   well as in mouse and human chondrocytes (Figures [84]4A and [85]S4A,
   Supporting Information). This suggests these nanoparticles could be
   taken up by chondrocytes. Macromolecules and nanoparticles typically
   enter cells via endocytosis, followed by trafficking to lysosomes for
   degradation. Lysosome co‐localization experiments demonstrated that
   only a minimal amount of internalized bMSN and Ca‐bMSN (5.16%, 6.22%)
   signals co‐localized with lysosomes in the mouse chondrocytes that were
   treated with IL‐1β (Figures [86]4B and [87]S4B, Supporting
   Information). This suggests that bMSN and Ca‐bMSN could bypass the
   traditional endocytic pathway and evade degradation within lysosomes.
   Furthermore, bMSN at concentrations of 20, 50, and 100 µg mL^−1
   exhibited no toxic effects on primary chondrocytes and Ca‐bMSN at 20
   and 50 µg mL^−1 significantly promoted chondrocyte proliferation
   (Figure [88]4C). In addition, both bMSN and Ca‐bMSN treatments
   maintained normal cell morphology and high ECM synthesis ability,
   suggesting no obvious toxic effects on the cells (Figure [89]4D). We
   further investigated the expression levels of Col2a1, Acan, and Mmp13
   in primary chondrocytes treated with IL‐1β, IL‐1β + Catalpol, and IL‐1β
   + Ca‐bMSN (using a concentration of Ca‐bMSN equivalent to 20 × 10^−6 m
   Catalpol). The results demonstrated that 20 µg mL^−1 Ca‐bMSN displayed
   effects comparable to 20 × 10^−6 m Catalpol, promoting Col2a1 and Acan
   expression while inhibiting Mmp13 expression (Figure [90]4E).

Figure 4.

   Figure 4
   [91]Open in a new tab

   Ca‐bMSN promotes anabolism in primary chondrocytes and inhibits
   catabolism. A) Ca‐bMSN and bMSN labeled with FITC observed by confocal
   laser microscope (left). Uptake of bMSN and Ca‐bMSN by IL‐1β induced
   primary chondrocytes. Nuclei, blue; Cell membrane, red (right). Scale
   bar = 5 µm. B) Fluorescent visualization (right) co‐localization
   between bMSN or Ca‐bMSN and lysosome in IL‐1β induced primary
   chondrocytes. Ca‐bMSN or bMSN, green; Nuclei, blue; Lysosome, red. The
   quantification result is shown at left. Scale bar = 5 µm. C) The effect
   of bMSN or Ca‐bMSN on primary chondrocytes viability was detected by
   CCK‐8 (n = 3, *p < 0.05, **p < 0.01, ### or ***p < 0.001 when compared
   with control group and IL‐1β group, respectively). D) Effects of bMSN
   and Ca‐bMSN on cell morphology and the synthesis of proteoglycan in
   chondrocytes. Scale bar = 50 µm. E) The expression level of Col2a1,
   Acan, and Mmp13 in the primary chondrocytes, which were stimulated with
   IL‐1β combined with Catalpol, bMSN, or Ca‐bMSN treatment for 24 h (n =
   3, *p < 0.05, **p < 0.01, ### or ***p < 0.001 when compared with
   control group and IL‐1β group, respectively). Data are shown as mean ±
   SEM. P‐values were obtained by one‐way ANOVA with multiple comparisons.

2.5. The In Vivo Chondroprotective Efficacy of Catalpol and Ca‐bMSN

   Degradable MSNs are generally considered preferable to nondegradable
   ones for medical applications due to their lower bioaccumulation and
   higher biocompatibility. To investigate whether Ca‐bMSN could deliver
   drugs to cartilage and undergo biodegradation, we monitored the signal
   from bMSN and Ca‐bMSN in rat knee joints after successful injection
   (Figure [92]S5A, Supporting Information). MSNs have been reported to be
   as CT and optical imaging agents, thus the degradation of bMSN and
   Ca‐bMSN in rat knee joints was monitored at day 0, 1, 3, 5, and 7 by
   micro‐CT. The results showed that both bMSN and Ca‐bMSN resided in the
   joint for ≈7 d before degrading (Figure [93]S5B, Supporting
   Information). TEM images revealed a noticeable alteration in the
   morphology of bMSNs, with a rapid decrease in nanoparticle diameter and
   minimal residual material detectable by day 7. In contrast, the
   morphology of MSNs remained largely unchanged from day 0 to day 7
   (Figure [94]S5C, Supporting Information). Furthermore, HPLC analysis
   was used to determine the retention of Catalpol in OA articular fluid
   at predetermined time points. This analysis revealed that the Ca‐bMSN
   group exhibited significantly higher Catalpol retention compared to the
   Catalpol‐alone group even at 24 h. This suggests that Ca‐bMSN
   effectively achieves sustained release of Catalpol within the articular
   fluid of OA rats (Figure [95]S5D, Supporting Information).

   Behavioral tests were undertaken to assess the cartilage protection
   efficacy of Catalpol and Ca‐bMSN at different concentrations (1, 5,
   10 mg mL^−1) on OA model rats (Figure [96]5A). Hyaluronic acid (HA), a
   common clinical treatment strategy for OA, was served as positive
   control. Gait analysis revealed a significant increase in paw area for
   rats treated with Catalpol alone, all Ca‐bMSN concentrations (1, 5,
   10 mg mL^−1) and HA compared to the OA model group and the free bMSN
   group (Figure [97]5B). Treatment with Catalpol, various Ca‐bMSN
   concentrations, and HA significantly reduced the differences in bearing
   area (Δarea), bearing weight (Δweight), and swinging time (Δposture
   duration) between the left and right hind limbs of OA model rats
   compared to the model group, free bMSN did not demonstrably affect
   Δarea, Δweight, or Δposture duration in OA model rats
   (Figure [98]5C–F). Furthermore, we compared the walking length
   discrepancies between the left and right hind limbs of OA rats. The
   findings indicated that as the treatment progressed from week 1 to week
   4, the walking length difference tended to disappear, particularly in
   the high‐dose Ca‐bMSN group, which exhibited a superior therapeutic
   effect compared to Catalpol, low‐dose Ca‐bMSN, medium‐dose Ca‐bMSN and
   HA groups, bMSN alone did not have any significant impact on the
   walking length of OA model rats (Figure [99]5G). These results
   collectively demonstrate that Ca‐bMSN offers significantly greater
   protective effects against OA symptoms like pain and limited joint
   mobility compared to bMSN alone, Catalpol alone, or HA.

Figure 5.

   Figure 5
   [100]Open in a new tab

   Behavioral effects of Catalpol, bMSN, and Ca‐bMSN treatment on OA rat.
   A) The schematic diagram depicts the process of creating a rat model of
   OA and the experimental design used to assess the protective effects of
   Catalpol, bMSN, and Ca‐bMSN. HA served as positive control. B) Gait
   analysis images of Sham, Model, bMSN, Catalpol, Ca‐bMSN, and HA groups.
   LF, left front paw; RF, right front paw; LR, left rear paw; RR, right
   rear paw. Δ Area, Δ Weight, and Δ Posed duration of sham, model
   Catalpol, bMSN, Ca‐bMSN, and HA groups at C) week 1, D) week 2, E) week
   3, and F) week 4 (n = 5, *p < 0.05, **p < 0.01, ### or ***p < 0.001
   when compared with sham group and model group, respectively). G) Δ
   Stride length of sham, model Catalpol, bMSN, Ca‐bMSN, and HA groups (n
   = 5) is shown during a gait cycle. # versus Sham, * versus Model, *p
   < 0.05, ## or **p < 0.01, ### or ***p < 0.001 (n = 5, *p < 0.05, ## or
   **p < 0.01, ### or ***p < 0.001 when compared with sham group and model
   group, respectively). Data are shown as mean ± SEM. P‐values were
   obtained by one‐way ANOVA with multiple comparisons.

   The subchondral bone undergoes structural changes during OA
   development. Administration of Catalpol and various concentrations of
   Ca‐bMSN effectively inhibited abnormal bone remodeling within the
   subchondral bone (Figure [101]6A). Compared to the OA model group,
   treatment groups exhibited significantly lower values for bone mineral
   density (BMD), bone volume fraction (BV/TV), bone surface fraction
   (BS/TV), and trabecular thickness (Tb.Th) (Figure [102]6B).
   Histological analysis using hematoxylin and eosin (H&E) and
   safranin‐O‐fast green staining revealed a thinner cartilage layer with
   a coarse surface in the OA model group. In contrast, the Catalpol and
   Ca‐bMSN groups displayed a smooth cartilage surface with well‐organized
   chondrocytes, particularly in the high‐dose Ca‐bMSN groups (5 and 10 mg
   mL^−1) (Figure [103]6C). Consequently, Ca‐bMSN at all concentrations
   (1, 5, and 10 mg mL^−1) had lower OARSI scores compared to the free
   Catalpol, bMSN, HA, and OA model groups (Figure [104]6D). Furthermore,
   Catalpol, Ca‐bMSN, and HA treatment reversed the increase in
   inflammatory mediators (IL‐1β, IL‐6, and TNF‐α) observed in the joint
   fluid of OA rats compared to other groups (Figure [105]6E). Notably,
   bMSN alone did not cause an inflammatory reaction in joint fluid.
   Immunohistochemistry (IHC) and immunofluorescence analyses revealed
   that Catalpol and Ca‐bMSN treatment decreased the expression of MMP9,
   MMP13, and ADAMTS4, while increasing the expression of COL2A1 and ACAN
   in OA rats compared to other groups (Figure [106]S6A,B, Supporting
   Information).

Figure 6.

   Figure 6
   [107]Open in a new tab

   In vivo effect of Catalpol, bMSN, and Ca‐bMSN treatment on OA rat. A)
   3D images of the knee joints by micro‐CT at 4 weeks of treatment with
   Catalpol, bMSN, Ca‐bMSN, and HA groups. Arrows indicate the lesion
   points of joint. B) Quantitative analysis of BMD, BV/TV, BS/TV, and
   Tb.Th (n = 5, *p < 0.05, **p < 0.01, ### or ***p < 0.001 when compared
   with sham group and model group, respectively). C) H&E staining and
   safranin‐O‐fast green staining (SO/FG) of knee joints treated with
   Catalpol, bMSN, Ca‐bMSN, and HA at 4 weeks. Scale bar = 500 µm. D) The
   severity of OA in the sham, model Catalpol, bMSN, Ca‐bMSN, and HA
   groups was analyzed by grading histological sections using the
   Osteoarthritis Research Society International (OARSI) score system (n =
   5, *p < 0.05, ## or **p < 0.01, ### or ***p < 0.001 when compared with
   sham group and model group, respectively). E) The concentration of
   IL‐1β, IL‐6, and TNFα in joint fluid of sham, model, Catalpol, bMSN,
   Ca‐bMSN, and HA groups (n = 5, *p < 0.05, ## or **p < 0.01, ### or ***p
   < 0.001 when compared with sham group and model group, respectively).
   Data are shown as mean ± SEM. P‐values were obtained by one‐way ANOVA
   with multiple comparisons.

   To further determine whether the administration of Catalpol and
   different concentrations of Ca‐bMSN (1, 5, 10 mg mL^−1) have the same
   cartilage protective effect in the OA mouse model as in the OA rat
   model, gait analysis, micro‐CT, and histological analyses were
   performed. As shown in Figure [108]S7A–E (Supporting Information),
   Catalpol, various Ca‐bMSN concentrations, and HA significantly reduced
   the differences in bearing area (Δarea), bearing weight (Δweight), and
   swinging time (Δposture duration) between the left and right hind limbs
   of OA mouse model compared to the model group. However, bMSN alone did
   not have a significant impact on gait parameters. Micro‐CT analysis
   revealed an increase in knee joint bone hyperplasia and osteophyte
   formation in the OA mouse model group, along with elevated subchondral
   bone parameters such as BMD, BV/TV, BS/TV, and Tb.Th, compared to the
   sham group. Treatment with Catalpol and varying concentrations of
   Ca‐bMSN effectively inhibited hyperplasia, osteophyte formation, and
   subchondral bone abnormal remodeling, as illustrated in Figure
   [109]S7F,G (Supporting Information). Furthermore, Catalpol, Ca‐bMSN,
   and HA administration mitigated polysaccharide loss and cartilage
   surface damage, as evidenced by Safranin O‐Fast Green (SO/FG) staining,
   shown in Figure [110]S7H (Supporting Information). These results
   suggested that Catalpol and Ca‐bMSN could significantly ameliorate the
   degree of joint cartilage degradation and improve joint function in the
   OA mouse model, as in the OA rat model.

   To evaluate the potential toxicity of bMSN and Ca‐bMSN,
   histopathological analysis was conducted on liver, spleen, kidney,
   lung, and heart tissues obtained from rats subjected to intra‐articular
   injections of bMSN and high‐concentration Ca‐bMSN. No histopathological
   abnormalities were detected in these organs (Figure [111]S8, Supporting
   Information). In addition, the hemolysis percentage of bMSN and Ca‐bMSN
   was determined to be less than 5% (Figure [112]S9A,B, Supporting
   Information). The biochemical indicators, including ALT, AST, ALP, and
   CREA, were performed to assess the potential toxicity towards the major
   organs of bMSN and Ca‐bMSN. As shown in Figure [113]S9C,D (Supporting
   Information), bMSN and Ca‐bMSN have no potential liver and kidney
   toxicity in rat and mouse. Taken together, these data demonstrate that
   Ca‐bMSN has the therapeutic potential to improve OA progression in knee
   OA animal model without obvious side effects.

2.6. Identification of Potential Molecular Mechanisms Regulated by Catalpol

   To illuminate the underlying therapeutic mechanism of Catalpol, RNA‐seq
   was performed. The differential gene expression profile was visualized
   using a volcano plot (Figure [114]7A), revealing seven significantly
   upregulated genes and 184 significantly downregulated genes in the
   IL‐1β versus IL‐1β + Catalpol group. KEGG pathways enrichment observed
   17 pathways were significantly enriched, mainly involved in PI3K‐Akt,
   cellular senescence, MAPK, NF‐κB, etc. It is noteworthy that PI3K/Akt
   and MAPK pathways were involved in the activation of NF‐κB
   (Figure [115]7B). Furthermore, the gene set enrichment analysis (GSEA)
   further proved that the NF‐κB signaling pathway may be regulated by
   Catalpol (Figure [116]7C). The heat map in Figure [117]7D demonstrated
   noteworthy differences genes in the NF‐κB signaling pathway in the
   IL‐1β versus IL‐1β + Catalpol group. Therefore, the NF‐κB pathway was
   more critical. Moreover, transcriptome analysis confirmed that Catalpol
   regulates the genes expression level involved in cell apoptosis,
   oxidative stress, and ECM synthesis and degradation (Figure [118]S10,
   Supporting Information). While NF‐κB signaling pathway could regulate
   cell apoptosis, oxidative stress, and ECM degradation. These results
   indicated that Catalpol exerts anti‐OA effects that may be mediated by
   the modulation of the NF‐κB signaling pathway.

Figure 7.

   Figure 7
   [119]Open in a new tab

   Hsp90β was identified as the target of Catalpol. A) Volcano plot
   representation of the different expression genes. Green dots:
   downregulated genes, red dots: upregulated genes, blue dots: none. B)
   KEGG pathway enrichment analysis. C) GSEA enrichment analysis indicated
   that Catalpol regulates NF‐κB signaling pathway. D) Heat map of gene
   expression associated with NF‐κB signaling pathway. E) The DARTS
   samples were subjected to silver staining, revealing that Catalpol
   provided protection to the band with a molecular weight of ≈180 kDa. F)
   Mass spectrogram of Hsp90β protein. G) Total protein from primary
   chondrocytes were treated with varying doses of protease with or
   without Catalpol, followed by the detection of Hsp90β levels through
   Western blot analysis. H) Mouse primary chondrocytes lysate was
   incubated with PBS or Catalpol for 30 min at 25 °C, and then denatured
   under various temperatures. The protein level of Hsp90β was assayed by
   Western blot. I) The affinity between Hsp90β and Catalpol was measured
   by ITC. J) Domain structure of HSP90 (732 aa) and N‐terminal domain of
   HSP90 (N‐Hsp90, 1–220 aa). K) N‐terminal of Hsp90β (1–220 aa) was
   treated with varying doses of protease with or without Catalpol or
   Geldanamycin, followed by the detection of Hsp90β levels through
   Western blot analysis. L) The affinity between the N‐terminal of Hsp90β
   (1–220 aa) was measured by ITC. M) N‐terminal of Hsp90β (1–220 aa) was
   incubated with PBS, Catalpol, or Geldanamycin for 30 min at 25 °C, and
   then denatured under various temperatures. The protein level of
   N‐terminal of Hsp90β (1–220 aa) was assayed by Western blot. N) The
   affinity between the N‐terminal of Hsp90β (1–220 aa) and Geldanamycin
   was measured by ITC.

2.7. HSP90β Is a Novel Target of Catalpol

   To identify the target of Catalpol in chondrocytes, we performed a drug
   affinity responsive target stability (DARTS) assay on primary
   chondrocyte protein extracts, followed by analysis using SDS‐PAGE. A
   comparison of Catalpol and control lysate lanes on the gel revealed a
   band with a molecular weight of ≈180 kDa that exhibited different
   intensities (Figure [120]7E). These bands were excised for unbiased,
   high‐throughput protein identification by mass spectrometry
   (Figure [121]7F). Mass spectrometry identified a total of 202 proteins.
   Hsp90β displayed the highest significance of difference between
   Catalpol and control groups with high confidence, suggesting it as a
   potential target candidate. To confirm Hsp90β as Catalpol's target, we
   conducted DARTS and Western blot assays. Chondrocyte total protein was
   digested with varying protease concentrations, with or without Catalpol
   incubation. Our results suggest that Catalpol protects Hsp90β from
   enzymatic degradation (Figure [122]7G). A cellular thermal shift assay
   (CETSA) was performed to further validate the interaction between
   Hsp90β and Catalpol. Compared to the PBS group, Catalpol significantly
   increased Hsp90β accumulation at temperatures of 61, 64, 67, 70, and 73
   °C, suggesting that Catalpol enhances Hsp90β stability
   (Figure [123]7H). Isothermal titration calorimetry (ITC) confirmed
   direct binding between Catalpol and Hsp90β with a dissociation constant
   (K [D]) of 1.83 × 10^−6 ± 433 × 10^−9 m (Figure [124]7I). Since Hsp90β
   inhibitors typically target the ATP‐binding site of the N‐terminal
   domain, the N‐terminal domain of Hsp90β (N‐Hsp90β, amino acids 1–220)
   was produced and used to assess its binding capacity with Catalpol
   (Figure [125]7J). Geldanamycin, a known Hsp90 inhibitor targeting the
   ATP‐binding domain, was used as a positive control. DARTS and CETSA
   assays demonstrated the protective effects of both Catalpol and
   Geldanamycin on Hsp90β against enzymatic digestion and
   high‐temperature‐induced degradation or denaturation
   (Figure [126]7K,M). In addition, ITC analysis determined the binding
   affinities between N‐Hsp90β and Catalpol or Geldanamycin to be ≈2.05 ×
   10^−6 and 2.13 × 10^−6 m, respectively (Figure [127]7L,N). Taken
   together, these findings demonstrate that Catalpol could directly bind
   to Hsp90β, particularly through its interaction with the N‐terminal
   domain.

2.8. Molecular Docking, MM‐GBSA Calculation, and Molecular Simulation

   A docking simulation was performed to gain a deeper understanding of
   the interaction between Catalpol and the N‐terminal domain of Hsp90β
   (N‐Hsp90β). This simulation aimed to predict Catalpol's binding
   affinity and identify the critical amino acids forming the binding
   site. As shown in Figure [128]8A, the N‐Hsp90β‐Catalpol complex
   displayed water‐mediated hydrogen bonds between residues GLY92, ASN101,
   ASP49 of N‐Hsp90β and the hydroxyl groups of Catalpol's sugar moiety.
   Notably, Catalpol exhibited a stronger binding affinity (−11.09 kcal
   mol^−1) toward N‐Hsp90β compared to the original ligand (−8.81 kcal
   mol^−1), likely due to this water‐mediated hydrogen bond network. Prime
   MM‐GBSA calculations were used to accurately predict the binding energy
   of the docked pose. The binding energy (Gbind) of Catalpol was
   −59.56 kcal mol^−1, which was higher than the original ligand
   (−71.71 kcal mol^−1). To further validate the binding free energy
   prediction, a MD simulation was conducted. The root‐mean‐square
   deviation (RMSD) of the N‐Hsp90β‐Catalpol complex remained within a
   stable range of 1–3 Å throughout the simulation (Figure [129]8B),
   suggesting rapid stabilization around 1.5 Å within 5 ns. This indicated
   that Catalpol has a strong stabilizing effect on N‐Hsp90β. The RMSD
   analysis supports the accuracy of the docking results and suggests that
   the complex achieved a stable conformation during the simulation.
   Furthermore, root mean square fluctuation (RMSF) analysis revealed
   significantly low fluctuations for residues (ASP88, GLY92, THR179)
   located around the binding pocket (Figure [130]8C). This suggests these
   residues are crucial for intermolecular interactions, thereby
   stabilizing the protein conformation. Intermolecular interaction
   analysis identified four types of contacts between Catalpol and
   N‐Hsp90β: hydrogen bonds, hydrophobic interactions, ionic bonds, and
   water bridges. The hydroxyl group at the C6 position of Catalpol's
   sugar moiety acted as a hydrogen bond donor with ASN46
   (Figure [131]8D). In addition, water bridges contributed to contacts
   between ASP49 and the C6 hydroxyl group, which represented a vital
   functional group for Catalpol's inhibitory activity. The hydroxyl group
   on the side chain of the iridoid scaffold also interacted with N‐Hsp90β
   in two ways: as a hydrogen bond donor with ASP88 and through a water
   bridge with THR179. Interestingly, the epoxy group in the iridoid
   structure also interacted with the crucial amino acid ASP88 via a water
   bridge. Based on the interaction statistics obtained during the 500 ns
   simulation, it is highly conceivable that ASN46, ASP88, and THR179 play
   significant roles in forming a stable binding conformation
   (Figure [132]8E). In conclusion, the findings from the MD simulation
   are consistent with the ITC experiments, suggesting that Catalpol has
   the potential to act as an Hsp90β inhibitor.

Figure 8.

   Figure 8
   [133]Open in a new tab

   Hsp90β molecular simulations. A) Binding model of Catalpol into the
   N‐terminal ATP‐binding site of Hsp90β and original ligand redocking. A
   diagram illustrating the precise atomic interactions between the ligand
   and Hsp90β residues over a 500 ns molecular dynamic simulation. Only
   the interactions that occurred for more than 30.0% of the simulation
   time within the chosen trajectory are depicted. B) RMSD plots of
   protein C‐alpha, backbone, side chain, heavy atoms throughout the
   500 ns MD simulations. C) RMSF plots of Hsp90β‐Catalpol during the MD
   simulations. D) The statistical graph illustrates protein–ligand
   interactions between Hsp90β and Catalpol. The value indicates the
   percentage of time the specific interaction is maintained during the
   simulation. Values above 1.0 signify that the relative residue
   establishes multiple contacts with Catalpol. E) The timeline
   representation of the interactions and contacts. The upper panel
   illustrates the cumulative count of distinct interactions established
   by Hsp90β with the ligand throughout the trajectory. Conversely, the
   lower panel delineates the specific residues engaged in interactions
   with Catalpol in each frame of the trajectory. Notably, these residues
   exhibit multiple specific contacts with Catalpol, as indicated by the
   darker shade of orange on the color scale provided alongside the plot.

2.9. Catalpol Regulates Chondrocytes Metabolism and Oxidative Stress via
Inhibition of Hsp90β/NF‐κB Signaling Pathway

   To explore the potential correlation between Hsp90β and chondrocyte
   apoptosis, oxidative damage, anabolism/catabolism and NF‐κB signaling
   pathway in knee OA patients, the publicly accessible transcriptomic
   data of aged male OA chondrocytes and healthy male chondrocytes sourced
   from the Gene Expression Omnibus (GEO) database (GSE 246425) was
   subjected to reanalysis. The volcano plots revealed that 5537 genes
   were significantly upregulated and 2061 genes were downregulated in old
   OA male chondrocytes versus OA health male chondrocytes group, and
   Hsp90ab1 is highly expressed in OA patients (Figure [134] 9A and Figure
   [135]S11A, Supporting Information). Gene ontology (GO) annotations
   described the biological roles of differential genes sorted as three
   categories including biological process, cellular component, and
   molecular function, mainly including extracellular matrix organization,
   extracellular structure organization, collagen containing
   extracellular, actin binding process, and so on (Figure [136]9B). Gene
   correlation analysis revealed significant positive correlations between
   Hsp90ab1 and IKKα, as well as Adamts4, and a negative correlation with
   Bcl2. In addition, IKKα exhibited positive correlations with Adamts4,
   IKKβ, and Bax, while Bax displayed negative correlations with Sod1,
   Sox9, and Comp (Figure [137]9C). The findings suggest a positive
   correlation between Hsp90ab1 and marker genes associated with ECM
   catabolism, apoptosis, oxidative damage, and the NF‐κB signaling
   pathway, as well as a negative correlation with marker genes related to
   ECM anabolism, anti‐apoptosis, and anti‐oxidation.

Figure 9.

   Figure 9
   [138]Open in a new tab

   Hsp90β is required for Catalpol regulation of chondrocyte metabolism
   and ROS production induced by IL‐1β through the NF‐κB signaling
   pathway. A) Volcano plot representation of the different expression
   genes in patents chondrocytes and health chondrocytes. B) Differential
   expression genes GO enrichment. C) Correlation analysis to analyze the
   correlation between target gene Hsp90ab1 and chondrocytes apoptosis
   (Bax, Bcl2), oxidative stress (Sod1), ECM metabolism (Sox9, Col2a1,
   Acan, Comp), and NF‐κB signaling pathway (IKKα, IKKβ) marker genes.
   Red: positive correlation; Blue: negative correlation. D) The changes
   of NF‐κB signaling pathway between different groups using Western blot
   assay. The quantification result is shown at right (n = 3, ## or **p
   < 0.01, ### or ***p < 0.001 when compared with control group and IL‐1β
   group, respectively). E,F) qRT‐PCR analysis of the expression level of
   anabolism marker genes (Col2a1, Acan) and catabolism marker genes
   (Mmp13, Adamts4) in chondrocytes with or without IL‐1β, Catalpol, or
   Hsp90ab1‐siRNA (SI) (n = 3, *p < 0.05, **p < 0.01, ###p < 0.001 when
   compared with control group and IL‐1β group, respectively). G) ROS
   production in chondrocytes under the treatment with IL‐1β or IL‐1β
   combined with Catalpol, SI+ Catalpol, and SI was determined by DCFH‐DA.
   Fluorescence microscopy was used to capture images (left). ROS
   production was determined by the DCF fluorescence intensity was
   determined by microplate reader (right). Scale bar = 400 µm (n = 3, **p
   < 0.01, ###p < 0.001 when compared with control group and IL‐1β group,
   respectively). Data are shown as mean ± SEM. P‐values were obtained by
   one‐way ANOVA with multiple comparisons.

   During OA development, IL‐1β stimulation promotes oxidative stress,
   degradation of ECM, and through the activation of the NF‐κB pathway. To
   investigate whether Catalpol protects chondrocytes by targeting Hsp90β,
   leading to the inhibition NF‐κB signaling pathway, protein–protein
   interactions among the Hsp90β, IKKβ, IKKα, NF‐κB (p65), MMP13, ADAMTS4,
   COL2A1, and ACAN were first predicted with the STRING tool and revealed
   that Hsp90β served as a hub protein and had the strongest interaction
   with IKKβ, IKKα, and NF‐κB (p65), and then regulated other genes
   associated with ECM metabolism (Figure [139]S11B, Supporting
   Information). Subsequently, we treated an OA chondrocyte model with
   Catalpol alone, Catalpol combined with siRNA targeting Hsp90β
   (siRNA‐Hsp90ab1, SI), or SI alone. Hsp90ab1 gene silence efficacy by
   siRNA is shown in Figure [140]S11C (Supporting Information). Western
   blot analysis was then performed to assess the total protein expression
   of IKKα, IKKβ, and P65. As shown in Figure [141]9D, silencing the
   expression of Hsp90ab1 gene in the primary chondrocytes was accompanied
   by the decreased expression of IKKα, IKKβ, and NF‐κB (P65) protein.
   Meanwhile, Catalpol treatment, either alone or combined with SI,
   significantly suppressed the expression levels of IKKα, IKKβ, and P65
   compared to untreated controls. To validate the working mechanism of
   Catalpol in vivo, 6 nmol siRNA‐Hsp90ab1, processed with
   ago‐modification, were injected into OA rat joint cavity weekly.
   Hsp90ab1 gene silence efficacy by siRNA is shown in Figure [142]S11D
   (Supporting Information). The expression level of IKKα, IKKβ, and NF‐κB
   (P65) was detected by Western blot. As shown in Figure [143]S12
   (Supporting Information), the same as in vitro results above were also
   obtained in in vivo experiments. In addition, Catalpol treatment alone
   or combined with SI significantly upregulated the expression of
   anabolic marker genes (Col2a1 and Acan) and downregulated the
   expression of catabolic marker genes (Mmp13 and Adamts4)
   (Figure [144]9E,F). Meanwhile, ROS levels were significantly reduced in
   chondrocytes treated with Catalpol, Catalpol combined with SI, or SI
   alone (Figure [145]9G). Notably, knockdown of Hsp90ab1 in primary
   chondrocytes demonstrates a similar biological effect as Catalpol
   treatment on cell metabolic activity and ROS level, and the combined
   administration of Catalpol and SI does not result in a more pronounced
   regulatory effect than SI alone, potentially due to the absence of
   Hsp90β protein. These findings demonstrate that Catalpol exhibits
   protection effects on chondrocytes specifically by targeting Hsp90β,
   leading to the inhibition of the NF‐κB signaling pathway.

3. Discussion

   OA is the predominant chronic degenerative condition affecting
   articular cartilage, characterized by chronic pain, functional
   limitations, instability, and deformity, resulting in a significant
   decline in quality of life. Globally, OA is a leading cause of
   disability among older individuals.^[ [146]^13 ^] Due to its high
   prevalence and associated disability, OA imposes a significant burden
   on both individuals and society. Restore the normal architecture and
   function of a damaged joint by the chondrogenic differentiation of stem
   cells and the synthesis of cartilage matrix by chondrocytes are the
   effective strategies for OA treatment. Currently, several effective
   protein drugs have been developed for this strategy. LNA043, a
   recombinant version of the human angiopoietin‐like 3 (ANGPTL3) protein,
   was reported to facilitate the chondrogenic differentiation of MSCs and
   enhanced the anabolic activities of cartilage, that is currently in a
   phase 2b trial ([147]NCT04864392) in patients with knee OA.^[ [148]^14
   ^] Sprifermin (recombinant human FGF18, rhFGF18) can not only promote
   the proliferation of chondrocytes and the synthesis of ECM, but also
   inhibit the activity of proteolytic enzymes and remarkedly reduce the
   degeneration of cartilage, which is currently in phase III clinical
   trial.^[ [149]^15 ^] TCM has been shown to be effective and safe for
   treating OA. There is a growing interest in exploring compounds derived
   from natural plants or traditional medicinal practices as potential
   treatments for OA due to their demonstrated clinical efficacy and
   minimal side effect profile. However, the development of traditional
   Chinese medicine monomer still lags behind that of protein drugs.

   Catalpol emerged as a candidate with the ability to upregulate the
   expression of anabolic marker genes (Col2a1, Acan, Comp) and
   downregulate the expression of catabolic marker genes (Mmp9, Mmp13,
   Adamts4) in human and mouse chondrocytes. In vivo administration of
   Catalpol demonstrated efficacy in limiting OA development, alleviating
   OA‐associated pain sensitivity, and partially reversing cartilage
   degradation in OA rat models. To elucidate the mechanisms underlying
   Catalpol's protective effect, we used a comprehensive suite of
   techniques including DARTS assays, proteomics, CETSA, ITC, molecular
   dynamics simulations, and RNA‐seq experiments. These investigations
   revealed Catalpol treatment correlated with NF‐κB signaling pathway and
   Hsp90β as the direct target of Catalpol, with a binding affinity of
   1.83 × 10^−6 ± 433 × 10^−9 m.

   Hsp90β, a member of the large heat shock protein (HSP) family,
   functions as a molecular chaperone. It plays a critical role in protein
   folding, stabilization, and the refolding of denatured proteins.^[
   [150]^16 , [151]^17 ^] Previous studies have demonstrated that
   Celastrol, a triterpenoid compound, inhibits the activity of Hsp90β and
   suppresses the production of MMPs, inducible nitric oxide synthase
   (iNOS), and cyclooxygenase‐2 (COX‐2) in chondrocytes isolated from
   osteoarthritic patients.^[ [152]^18 ^] Furthermore, Hsp90β has been
   shown to reduce the levels of NO production and chondrocyte death
   induced by IL‐1β or TNF‐α.^[ [153]^19 ^] The present study confirmed
   that knockdown of Hsp90ab1 expression could significantly upregulate
   chondrocyte anabolic activity. In addition, our findings demonstrated
   that the chondroprotective effect of Catalpol was dependent on
   Hsp90ab1. In conjunction with these observations and previous reports,
   we propose that Hsp90ab1 represents a promising drug target for the
   treatment of OA. Hsp90β exhibited a high degree of evolutionary
   conservation across species. The amino acid sequences of Hsp90β were
   100% identical between mice and rats, and only three amino acid
   residues differ between humans and these rodents (Figure [154]S13,
   Supporting Information). This high degree of conservation suggests the
   potential for cross‐species application of therapeutic strategies
   targeting Hsp90β.

   Hsp90β is an ATP‐dependent homodimeric chaperone protein consisting of
   three highly conserved domains: the N‐terminal domain (NTD; ≈25 kDa),
   the middle domain (M domain; ≈35 kDa), and the C‐terminal domain (CTD;
   ≈12 kDa).^[ [155]^20 ^] The N‐terminal domain harbors the ATP‐binding
   site, which also serves as a target for Hsp90 inhibitors. This domain
   has been extensively explored as a target for cancer therapy with drugs
   such as geldanamycin, its derivatives, and radicicol. These antitumor
   agents exhibit potent and specific inhibition of Hsp90 with nanomolar
   affinities.^[ [156]^21 , [157]^22 , [158]^23 ^] In this study, the
   N‐terminal domain of Hsp90β (N‐Hsp90β, amino acids 1–220) was produced
   and used to assess its binding capacity with Catalpol and Geldanamycin.
   DARTS, CETSA, and ITC assays revealed that N‐Hsp90β displayed binding
   affinities of ≈2.05 × 10^−6 and 2.13 × 10^−6 m for Catalpol and
   Geldanamycin, respectively. To gain further insight into the
   interaction between Catalpol and N‐Hsp90β, we used molecular docking
   and molecular dynamics simulation techniques. These techniques revealed
   the presence of water‐mediated hydrogen bonds between residues GLY92,
   ASN101, ASP49 of N‐Hsp90β and the hydroxyl groups of the sugar moiety
   in Catalpol. Furthermore, Catalpol displayed a significantly stronger
   binding affinity (−11.09 kcal mol^−1) to N‐Hsp90β compared to the
   original ligand redocking result (−8.81 kcal mol^−1). These findings
   suggest that water‐mediated hydrogen bonding plays a crucial role in
   the interaction of Catalpol with N‐Hsp90β, consistent with observations
   from previously crystallized protein complexes (PDB ID: 1UYM).^[
   [159]^24 , [160]^25 ^] Intermolecular interaction analyses identified
   the hydroxyl group at the C6 position of the sugar moiety in Catalpol
   as a hydrogen bond donor with ASN46. RMSF analysis revealed that
   residues ASP88, GLY92, and THR179, located within the binding pocket,
   exhibited minimal fluctuations, suggesting their critical role in
   intermolecular interactions and protein conformation stabilization.
   Based on interaction statistics obtained during the 500 ns simulation,
   it was inferred that ASN46, ASP88, and THR179 played significant roles
   in forming a stable binding conformation. Collectively, these findings
   suggest that amino acids ASP88, THR179, ASP49, and ASN46 within the
   N‐terminal domain of Hsp90β may be critical for its interaction with
   Catalpol.

   To investigate the mechanism by which Catalpol regulates OA
   progression, we performed transcriptomic profiling of OA chondrocytes
   and Catalpol treated OA cells by RNA sequencing. The result revealed
   that Catalpol treatment correlated with NF‐κB signaling pathway.^[
   [161]^26 ^] By analyzing the transcriptome data of aged male OA
   chondrocytes and healthy male chondrocytes, it was found that there was
   a significant positive correlation between Hsp90ab1 and IKKα and
   Adamts4, but a negative correlation between Hsp90ab1 and Bcl2. In
   addition, IKKα was positively correlated with Adamts4, IKKβ, and Bax,
   while Bax was negatively correlated with Sod1, Sox9, and Comp. In
   addition, Hsp90 is known to regulate the enzymatic activity and
   biogenesis of IKKs, key molecules in the NF‐κB signaling pathway. Our
   results showed that inhibiting Hsp90β with either Catalpol or
   siRNA‐Hsp90ab1 led to decreased expression of IKKα, IKKβ, and P65 and
   accompanied by increased expression of anabolic marker genes (Col2a1
   and Acan) and decreased expression of catabolic marker genes (Mmp9,
   Mmp13, and Adamts4). It is well‐established that NF‐κB signaling
   promotes the expression of inflammatory cytokines and catabolic factors
   in chondrocytes under oxidative stress or inflammatory conditions.^[
   [162]^27 ^] Our findings support the notion that Catalpol acts as an
   inhibitor of Hsp90β, thereby regulating the anabolic and catabolic
   processes of chondrocytes through modulation of the NF‐κB pathway. This
   is consistent with previous reports demonstrating that Catalpol
   inhibits the inflammatory response and catabolism in IL‐1β‐induced rat
   chondrocytes via suppression of the NF‐κB signaling pathway.^[ [163]^8
   ^]

   Intra‐articular administration remains the preferred approach for
   treating symptomatic OA. However, the limited potency of small‐molecule
   therapeutics hinders their efficacy due to rapid clearance and short
   residence time within the joint. To address this challenge, emerging
   nanocarriers such as MSNs are being investigated as effective drug
   delivery systems. Since their discovery in 2001, MSNs have shown
   promise in treating various diseases, including bone and tendon tissue
   engineering, diabetes, inflammation, and cancer.^[ [164]^28 ^] Studies
   have shown that MSNs loaded with Oltipraz
   [4‐methyl‐5‐(2‐pyrazinyl)‐1,2‐dithiole‐3‐thione (OL) exhibited
   chondroprotective potential after intra‐articular injection in OA
   models, with MSN‐OL persisting in the joint for at least 21 d.^[
   [165]^29 ^] However, it is reported that prolonged nanoparticle
   accumulation may lead to inflammation, oxidative damage, and organ
   fibrosis, hindering clinical translation. In addition, MSNs with an
   average size of 25–100 nm exhibit a genotoxic effect in HT‐29 cells,
   whereas 30–300 nm do not demonstrate any harmful effects on the HeLa
   cells, suggesting that larger MSN particle sizes are associated with
   lower toxicity levels.^[ [166]^30 ^] Conversely, smaller MSNs are
   considered more toxic due to their increased cellular uptake and higher
   number of silanol groups interacting with the cell membrane. As a
   result, in this study, we developed degradable MSNs with an average
   size of 200 nm. These nanoparticles maintain drug release functionality
   while biodegrading and metabolizing rapidly within the body, thus
   avoiding inflammatory toxicity. Furthermore, we did not observe
   increased inflammation in the joint fluid of OA rats treated with bMSN
   or Ca‐bMSN by detecting inflammatory mediators (IL‐1β, IL‐6, and
   TNF‐α). It is noteworthy that the administration of bMSN alone may
   result in a modest amelioration of OA symptoms. This improvement may
   potentially be attributed to the nanoparticles’ capacity to inhibit
   inflammatory processes and establish an immune microenvironment
   conducive to bone regeneration, achieved through the activation of
   immune cells and the stimulation of cytokine release.^[ [167]^31 ^] In
   vitro studies demonstrated efficient uptake of bMSNs and Ca‐bMSN by
   mouse and human chondrocytes, and their ability to evade lysosomal
   degradation, suggesting excellent intracellular trafficking behavior.
   Furthermore, Ca‐bMSN promoted the expression of Col2a1 and Acan while
   inhibited the expression of Mmp13, with no toxic effects on primary
   chondrocytes.

   Fluorescent dyes are widely used to mark nanoparticles for fluorescence
   imaging in vivo, which usually have a limited half‐life. Meanwhile,
   MSNs have been reported to be as CT and optical imaging agents.
   Therefore, the in vivo degradation of bMSNs was assessed using
   micro‐CT, revealing almost complete degradation within 7 d,
   significantly reducing potential accumulation in the joint.
   Furthermore, HPLC analysis showed that bMSNs achieved sustained drug
   release compared to free Catalpol, thereby prolonging its duration of
   action. Following successful in vivo delivery, Ca‐bMSNs effectively
   suppressed the expression of MMP9, MMP13, and ADAMTS4 while increasing
   COL2A1 and ACAN levels in OA rat model, leading to improved OA
   symptoms. These findings demonstrate the potential of Ca‐bMSNs for OA
   treatment and provide an effective delivery and treatment strategy.

   This study holds importance from multiple angles: (1) The
   identification of Catalpol as a promoter of chondrocyte proliferation
   and anabolism, along with the utilization of a delivery system for
   sustained release of Catalpol to chondrocytes, presents a novel
   approach for improving chondrocyte function and cartilage integrity.
   (2) Hsp90β was identified as a novel target of Catalpol, which enhances
   our comprehension of Catalpol's mechanisms of action and provides a
   basis for further exploration of the Catalpol/Hsp90β interaction in
   diverse conditions. The study acknowledges its limitations and suggests
   further research, including in vivo characterization and pharmaceutic
   kinetic assays of the drug in OA, as well as investigations with
   preclinical animal models and human clinical trials. While the study
   has shown that Hsp90β is the direct target of Catalpol, further
   exploration of the underlying mechanisms is warranted.

4. Conclusions

   In summary, this study demonstrates the potential of Catalpol as a
   therapeutic agent for OA. Our findings reveal that Catalpol stimulates
   chondrocyte anabolism, inhibits catabolism, oxidative stress,
   inflammation, and cell apoptosis, leading to reduced cartilage loss and
   alleviation of OA‐associated pain in vivo. In addition, we identify
   Hsp90β as a novel target of Catalpol, providing valuable insights into
   its mechanism of action and the underlying basis for its
   chondroprotective effects. Furthermore, the use of bMSNs as drug
   carriers enables prolonged and controlled release of Catalpol within
   the joint, significantly improving its in vivo efficacy. Overall, this
   study not only identifies an effective drug and target for promoting
   chondrocyte function and maintaining cartilage integrity but also
   presents a promising delivery strategy, contributing significantly to
   the development of targeted therapeutic candidates for OA.

5. Experimental Section

Cell Culture

   Primary chondrocytes were isolated from the rib cartilage of neonatal
   mice, as previously described. Briefly, the rib cartilage was
   infiltrated with 3 mg mL^−1 collagenase B (Sigma, USA) for 45 min at 37
   °C, followed by overnight digestion with 0.5 mg mL^−1 collagenase B at
   37 °C. The digestion was terminated with Dulbecco's Modified Eagle
   Medium/Ham's F12 (DMEM/F12) (Gibco, USA) supplemented with 5% fetal
   bovine serum (FBS) (BI, USA). The cells were collected by
   centrifugation (300 g, 5 min) and cultured in DMEM/F12 supplemented
   with 5% FBS (BI, USA) and 1% Penicillin/Streptomycin (Sigma, USA) at 37
   °C and 5% CO[2].^[ [168]^32 ^] Human chondrocyte cell line C28I2 was
   cultured in DMEM with 5% FBS, and 1% Penicillin/Streptomycin at 37 °C
   and 5% CO[2].

Quantitative Real‐Time Polymerase Chain Reaction (qRT‐PCR)

   Total RNA was isolated from primary mouse chondrocytes, C28I2 cells,
   and cartilage tissues following the Total RNA Extraction Kit protocol
   (Tiangen, China). Subsequently, cDNA synthesis was carried out using
   the PrimeScript RT reagent Kit (TaKaRa, USA) with 1 µg of total RNA.
   qRT‐PCR was performed on the CFX96 Real‐Time System (Bio‐Rad, USA)
   utilizing the TB Green Premix Ex Taq II kit (TaKaRa, USA) along with
   specific primers (Table [169]S1, Supporting Information). The relative
   gene expression level was determined using the 2^−ΔΔCT method, with
   Ribosomal Protein L4 (RPL4) serving as the internal control gene.

CCK‐8 Analysis

   The activity of small molecules was assessed using a combination of
   CCK‐8 assays. Briefly, primary mouse chondrocytes (5 × 10^3 cells/well)
   were seeded in a 96‐well plate. Small molecules were individually added
   at final concentrations of 10, 20, 40, 60, 80, and 100 × 10^−6 m to
   separate wells. The plates were then incubated for 24 h at 37 °C and 5%
   CO[2]. After incubation, 10 µL of CCK‐8 reagent (per well) was added
   and incubated for 1 h at 37 °C. Cell proliferation was measured by
   absorbance at 450 nm using a multifunctional microplate reader (Tecan,
   Switzerland).

Western Blot Analysis

   To prepare protein samples for Western blot analysis, chondrocytes or
   cartilage tissues were lysed using protein lysis buffer. The lysates
   were then loaded onto an SDS‐PAGE gel and electrophoretically
   transferred to a PVDF membrane (Millipore, USA). The membrane was
   blocked with 5% non‐fat milk in TBST for 2 h at room temperature.
   Primary antibodies (COL2A1, ACAN, MMP13, ADAMTS4, and RPL4;
   Proteintech, China) were incubated with the PVDF membrane overnight at
   4 °C. Following washes, the membranes were incubated with the
   appropriate horseradish peroxidase (HRP)‐conjugated secondary
   antibodies for 1 h at room temperature. Protein bands were visualized
   using enhanced chemiluminescence (ECL) substrate (BeyoECL Plus,
   Beyotime, China) and detected with a ChemiDocXRS+ system (Bio‐Rad,
   USA). The dilutions for the primary antibodies were: COL2A1, ACAN,
   MMP13, and ADAMTS4 (1:1000) and RPL4 (1:2000).

Immunofluorescent Assay

   Cells were immobilized by incubation in 4% paraformaldehyde (Beyotime,
   China) for 10 min. Following three washes with PBS, cells were
   permeabilized with 0.5% Triton X‐100 (Solarbio, China) for 20 min.
   Subsequently, cells were incubated with anti‐COL2A1 antibodies
   (Proteintech, China; 1:500 dilution) for 2 h at 37 °C. To visualize the
   antigen–antibody complex, cells were treated with FITC‐labeled goat
   anti‐rabbit IgG (Proteintech, China; 1:1000 dilution) for 30 min. In
   addition, nuclei were stained with DAPI staining solution (Beyotime,
   China). Finally, images were captured using a cell imager (Thermo,
   USA).

Alcian Blue Staining

   OA chondrocytes were treated with varying concentrations of Catalpol
   (10, 20, and 40 × 10^−6 m) for 7 d, the glycosaminoglycan content was
   quantified using an Alcian blue staining assay. Briefly, chondrocytes
   were stained with Alcian blue solution (Solarbio, China) for 2 h at
   room temperature, following the protocol described in a previous study.
   After staining, cells were washed, and the Alcian blue dye was
   extracted with 6 m guanidine hydrochloride (Macklin, China). The amount
   of extracted dye was quantified by measuring the optical density (OD)
   at 620 nm using a multifunctional microplate reader.^[ [170]^33 ^]

Evaluation of ROS Levels

   To assess the level of reactive oxygen species within chondrocytes, the
   Reactive Oxygen Species Detection Kit (Beyotime, China) was used
   following the manufacturer's instructions. Briefly, cells were
   incubated with the kit's probes for 20 min at 37 °C. Following
   incubation, fluorescence intensity was measured using a flow cytometer
   (Beckman Coulter, USA) to quantify ROS levels.

Analysis of SOD, MDA, GSSG, and ATP Level

   Primary chondrocytes were treated with IL‐1β alone or in combination
   with Catalpol (10, 20, or 40 × 10^−6 m) for 24 h. After that, cells
   were washed with precooled PBS and lysed. The supernatant was obtained
   by centrifuging and used for SOD, MDA, GSSG, total antioxidant, and ATP
   detection using specific assay kits (Beyotime, China) according to the
   manufacturer's protocol.

Mitochondrial Membrane Potential Determination

   Primary chondrocytes were treated with IL‐1β and IL‐1β + Catalpol (10,
   20, 40 × 10^−6 m). After 24 h of culture, mitochondrial membrane
   potential was detected by Mitochondrial membrane potential assay kit
   with JC‐1 (Beyotime, China) according to the experimental protocol. In
   detail, 0.3 mL of JC‐1 staining solution was added to each well of
   24‐well plate, and incubated in cell incubator at 37 °C for 20 min.
   Images were obtained by cell imager. Quantification was performed on a
   multifunctional microplate reader.

Analysis of Cell Apoptosis

   Mouse primary chondrocytes were divided into control, model (cells
   treated with 10 ng µL^−1 IL‐1β) and Catalpol (10, 20, 40 × 10^−6 m +
   10 ng µL^−1 IL‐1β) groups. After treatment with IL‐1β or IL‐1β +
   Catalpol, chondrocytes were prepared by incubation with 300 µL binding
   buffer including 5 µL Annexin V, followed by 5 µL propidium iodide (PI)
   for 5 min at 37 °C in the dark. Analysis was conducted using a flow
   cytometer.

RNA Sequencing and Data Analysis

   Mouse primary chondrocytes were divided into control, model (cells
   treated with 10 ng µL^−1 IL‐1β) and Catalpol (20 × 10^−6 m + 10 ng
   µL^−1 IL‐1β) groups and processed for RNA extraction. The RNA
   sequencing and data analysis were performed by Novogene Co., Ltd. In
   addition, differential expression analysis in human chondrocytes RNA
   sequencing data (GEO: GSE 246425) was performed with Voom and Limma
   methods (Limma package). The GO annotations and correlation between
   target genes and were conducted via a free online platform for data
   visualization and graphing ([171]https://www.bioinformatics.com.cn).

Preparation of bMSN and bMSN‐FITC

   The synthesis of bMSNs was performed using a modified procedure. A
   three‐necked flask was utilized to stir 1.2 g of cetyl trimethyl
   ammonium bromide (CTAB), 0.30 g of triethanolamine (TEA), and 100 mL of
   deionized water at 80 °C for a duration of 30 min. Subsequently, a
   solution containing 1800 µL of TEOS and 200 µL BTES was gradually added
   to the surfactant solution over a period of ≈0.5 h.

   The resulting mixture was then stirred at 80 °C for an additional 4 h
   at a stirring speed of 1000 rpm. The products were obtained through
   centrifugation (11000 rpm, 5 min), followed by three washes with
   ethanol. Finally, the collected products were refluxed in an ethanol
   solution of HCl (2% w/v) for a duration of 6 h. The bMSNs were
   collected, washed, and dried. To attach amine groups to the MSNs, 20 mg
   of MSNs was suspended in 20 mL of dimethylformamide. 10 µL of
   3‐aminopropyl triethoxysilane (APTES) was added and stirred for 24 h at
   room temperature. The resulting MSN‐NH[2] was then washed three times
   with EtOH to remove any remaining APTES and dried. For the fluorescent
   modification of bMSN, 0.5 mg of FITC was added to 10 mL of bMSN‐NH[2]
   solution (1 mg mL^−1 in absolute ethanol). The mixture was stirred
   overnight in the dark. The FITC‐modified bMSNs were collected via
   centrifugation and subsequently washed using ethanol.

Physicochemical and Ex Vivo Characterization of Nanoparticles

   The morphology and structure of MSNs were examined using a TEM (Hitachi
   HT‐7700, Japan). The TEM was operated at 100 kV. The particle size and
   zeta potential were assessed using Zetasizer (Malvern, UK).^[ [172]^34
   ^] To analyze the drug loading efficiency of bMSNs, the wetness
   impregnation method was used to load Catalpol into the bMSNs. Briefly,
   1 mL solution of Catalpol (1 mg mL^−1) was added into deionized water
   containing 1 mg of bMSNs with gentle continuous stirring and ultrasonic
   mixing for 24 h, then centrifuged at 11000 rpm, 5 min at 4 °C. The
   supernatant was finally collected. The Catalpol drug content in the
   loaded MSNs was determined using Agilent 1260 HPLC with monitoring of
   Catalpol absorbance at 203 nm.^[ [173]^35 ^] The encapsulation
   efficiency (EE%) was calculated as follows: EE (%) = (Weight of
   Catalpol in loaded MSNs/Weight of feeding Catalpol) × 100. The weight
   of Catalpol in loaded MSNs was determined by subtracting the weight of
   Catalpol in the supernatant from the weight of feeding Catalpol. For
   analysis of the drug stability and release in the synovial fluid, the
   rats were anesthetized and the synovial fluid was taken from articular
   joint. Synovial fluid was divided into two groups, and Catalpol (20 ×
   10^−6 m) and Ca‐bMSNs (10 mg mL^−1) were added, respectively. Samples
   were gently mixed in the shaker at 37 °C, and then collected at
   predetermined time intervals (0, 1, 3, 12, and 24 h). Concentrations in
   the release medium were determined by HPLC.

Absorption and Degradation of bMSNs and Ca‐bMSNs In Vitro and In Vivo

   Primary mouse chondrocytes and human chondrocytes (C28I2) were used for
   investigating the cellular uptake of bMSNs and Ca‐bMSNs. Briefly,
   chondrocytes were incubated in glass bottom cell culture dish at a
   density of 5 × 10^5 cells per well and cultured for 1 d. The bMSNs and
   Ca‐bMSNs were labelled with FITC and then added into plates at a
   concentration of 10 µg mL^−1 to assess intracellular efficiency. After
   incubation of chondrocytes with bMSNs or Ca‐bMSNs for 4 h, the cell
   samples were washed three times with PBS. Then, the cells were stained
   with Hoechst 33342 (Beyotime, China) for 20 min at 37 °C to label
   nuclei. The images were captured and analyzed by laser confocal
   microscopy (Leica Germany).^[ [174]^36 ^] The degradation of bMSNs and
   Ca‐bMSNs in rats joint was detected by Micro‐CT (Perkin Elmer, USA)
   after receiving an IA injection of 100 µL bMSNs and Ca‐bMSNs (200 µg
   mL^−1) at days 0 (after IA injection immediately), 1, 3, 5, and 7.

Hemolytic Assays

   Blood was collected from the abdominal aorta and preserved in
   anticoagulant tubes for the isolation of red blood cells (RBCs). The
   RBCs underwent five washes with a PBS solution, followed by
   centrifugation at 4000 rpm for 10 min at 4 °C to remove the
   supernatant. A 5% stock suspension of RBCs in PBS was then prepared for
   use in hemolytic assays. Ca‐bMSNs and bMSNs (1, 5, 10, 100 mg mL^−1)
   were incubated with 500 µL of the RBC stock suspension for 4 h at room
   temperature. Following incubation, the tubes were centrifuged for
   10 min, and the supernatants were carefully transferred to a clean
   96‐well plate for detection using a multifunctional microplate reader
   set at 485–700 nm. ddH[2]O[2] served as positive control. Saline was
   used with negative control. The percentage of hemolysis = [optical
   density − negative control optical density/(positive control optical
   density − negative control optical density)*100.

Intracellular Trafficking of bMSNs and Ca‐bMSNs in Chondrocytes

   To demonstrate whether bMSNs and Ca‐bMSNs could avoid lysosomal
   elimination in chondrocytes, chondrocytes were seeded at a density of 5
   × 10^5 cells per dish and cultured for 24 h. The medium was replaced
   with fresh medium containing bMSNs or Ca‐bMSNs (labeled with FITC) at a
   concentration of 10 µg mL^−1 and incubated for an additional 4 h.
   Subsequently, the lysosomes were labeled with LysoTracker Green (200 ×
   10^−9 m) for 30 min. In addition, the nuclei were stained with Hoechst
   33342 for 20 min. The images were then acquired and examined using
   laser confocal microscopy.^[ [175]^36 ^]

Establishment of OA Model and Administration of Drugs

   The animal protocol of this study was approved by the Institutional
   Animal Care and Use Committee of Changchun University of Chinese
   Medicine (No. 2023029). Adult male Sprague‐Dawley (SD) rats
   (8‐weeks‐old) were randomly divided into sham group (n = 10), model
   group (n = 10), bMSN group (10 mg mL^−1, n = 10), Catalpol group (20 ×
   10^−6 m, n = 10), Ca‐bMSN group (1, 5, 10 mg mL^−1, n = 10, 10, 10),
   and HA group (10 mg mL^−1, n = 10). The OA model was made through
   destabilization of the medial meniscus (DMM) combined with running
   motion (25 m min^−1, 80 min) (Xinruan, China) for four weeks.^[
   [176]^37 , [177]^38 ^] Rats with OA were injected with different
   components (bMSN, Catalpol, bMSN + Catalpol) in the knee joint cavity
   weekly (100 µL/time). Adult male C57BL/6 mice (8‐weeks‐old) were
   randomly divided into sham group (n = 10), model group (n = 10), bMSN
   group (10 mg mL^−1, n = 10), Catalpol group (20 × 10^−6 m, n = 10),
   Ca‐bMSN group (1, 5, 10 mg mL^−1, n = 10, 10, 10), and HA group (10 mg
   mL^−1, n = 10). Rats and mice were sacrificed after four weeks of
   continuous treatment administration.

Behavioral Assessment

   Gait analysis of rats in the sham group, model group, bMSN group,
   Catalpol group, and Ca‐bMSN group was performed on gait analyzer (Mouse
   Specifics Inc., USA). The rats were placed on a transparent conveyor
   belt at a running speed of 20 m s^−1. DigiGait video imaging
   acquisition system (Mouse Specifics Inc., USA) was used to observe each
   paw area (mm^2), weight (%), posed duration (s), and stride length
   (cm).

Micro‐CT, Histology, and IHC

   The complete knee joint of the rats and mice in the different group was
   scanned using a micro‐CT system (Perkin Elmer, USA). The surface
   smoothness of cartilage and the degree of subchondral bone sclerosis
   were detected in coronal and sagittal positions. Reconstruction of 3D
   images was obtained with N‐Recon software, BMD, BV/TV, BS/TV, and Tb.Th
   were measured by Analyze 12.0 software (AnalyzeDirect, Inc., USA).^[
   [178]^32 ^] For histology and IHC analysis, knee joints were embedded
   in wax blocks followed by fixed in 4% paraformaldehyde and then were
   sectioned at 4 µm thickness. Each section was stained with H&E and
   safranin‐O‐fast green to observe the cartilage, or antibodies including
   Col2a1, Acan, Mmp13, and Adamts4 to detect specific protein expression
   level. The images were captured using a digital scanning microimaging
   system M8 (Precipoint, Germany).

Drug Affinity Responsive Target Stability Assay and Mass Spectrometry

   The DARTS assay was carried out according to previous reports.^[
   [179]^39 , [180]^40 ^] In brief, total protein was obtained from mouse
   primary chondrocytes that lysed with RIPA lysis buffer (Beyotime,
   China), then centrifuged at 12000 rpm. 50 µL of total protein was
   incubated with PBS or Catalpol (500 × 10^−6 m) for 20 min at 25 °C. The
   mixture was digested by protease (Sigma, USA, 1:100, 1:200) for 30 or
   40 min. Samples were loaded on SDS‐PAGE, and stained with silver to
   detect target protein band. Target band was then analyzed by mass
   spectrometry performed by Novogene Co., Ltd.

Expression and Purification of Full Length and Truncated Hsp90β Protein

   Full‐length Hsp90β and its N‐terminal domain (N‐Hsp90β, residues 1–200)
   were cloned into the pET‐28a (+) vector, each with an N‐terminal
   His‐tag. The constructs were then transformed into Escherichia coli
   BL21 (DE3) cells (Tiangen, China). Single colonies were grown in
   culture at 37 °C with shaking at 200 rpm until the OD600 reached 0.6.
   Protein expression was induced by adding 1 × 10^−3 m isopropyl
   β‐d‐thiogalactoside (IPTG) and incubating the cells for 24 h at 16 °C
   with shaking at 180 rpm. Following induction, cells were harvested by
   centrifugation at 4000 rpm, resuspended in PBS, and lysed by sonication
   for 1 h (50% power, 2 s on, 3 s off) using a sonicator (Scientz,
   China). Cell debris was removed by centrifugation at 12000 rpm. The
   supernatant was loaded onto a nickel affinity column (Cytiva, Sweden)
   and washed with 100 mL of washing buffer (PBS with 10 × 10^−3 m
   imidazole) to remove unbound proteins. His‐tagged Hsp90β or N‐Hsp90β
   protein was then eluted with elution buffer (PBS with 250 × 10^−3 m
   imidazole). The eluted proteins were dialyzed against PBS and
   concentrated using ultrafiltration with a 10 kDa molecular weight
   cutoff membrane (Millipore, USA). Protein purity was assessed by
   SDS‐PAGE with Coomassie Brilliant Blue staining.^[ [181]^41 ^]

Cellular Thermal Shift Assay

   The mouse primary chondrocytes lysates were diluted with PBS and
   divided into two groups, with one group treated with PBS, another
   aliquot with Catalpol (500 × 10^−6 m). After 20 min incubation at 25
   °C, the lysates were divided into 20 µL/tube and heated individually at
   gradient temperatures (46–79 °C) for 3 min followed by cooling for
   3 min at 4 °C. Samples were detected by immunoblot with Hsp90β
   antibodies (Selleck, USA, 1:1000) or His‐Tag antibodies (Proteintech,
   China, 1:1000).^[ [182]^42 , [183]^43 , [184]^44 ^]

Isothermal Titration Calorimetry Analysis

   ITC analysis was conducted at 25 °C, utilizing the MicroCal PEAQ‐ITC
   instrument (Malvern, Sweden) with the titration of 13 injections of 2
   µL of Catapol (200 × 10^−6 m) or Geldanamycin into the reaction cell
   containing 20 × 10^−6 m of purified full length Hsp90β or N‐Hsp90β
   protein solution. The Catapol and protein were prepared with same
   buffer. Titration of Catapol to the same buffer was used as control to
   calibrate the equipment. Binding isotherms were fitted using nonlinear
   regression with the MicroCal PEAQ‐ITC analysis software.

Ligand, Protein Preparation, and Receptor Grid Generation

   The 3D molecule model of Catalpol was constructed and geometrically
   optimized using ChemSketch. LigPrep was further used to ionize the
   ligand based on the Epik module. The Hsp90β N‐terminal domain (PDB ID:
   5uc4) underwent processing via Protein Preparation Wizard, which is
   part of the Schrodinger Suite in Maestro.^[ [185]^25 ^] The water
   molecules that were not involved in ligand interactions and were
   situated at a distance greater than 5 Å from the ligand were eliminated
   from the original protein crystal structure. The missing loop or side
   chains in the crystal structure were reconstructed using Prime.
   Subsequently, hydrogen bonds were established with all amino acid
   residues at pH 7.0, considering their least ionized states. Energy
   minimization was performed using the OPLS‐3 force field until a RMSD of
   0.30 was attained, effectively reducing steric hindrance. The grid box
   was created to characterize the centroid of the binding site for
   docking purposes. The box is generated using the Glide grid generation
   wizard, and its size was set to match that of the crystallized ligand.

Molecular Docking and Binding Free Energy Calculations Based on MMGBSA Method

   After optimizing Catalpol using LigPrep, it was subjected to docking
   within a recently established grid box using Glide's extra‐precision
   mode, without any constraints.^[ [186]^45 ^] To overcome the steric
   hindrance, the OPLS‐3 force field was utilized to minimize energy until
   a RMSD of 0.30 was attained. The activate site was formed by applying
   the coordinates of preexisting co‐crystals protein–ligand complex (PDB
   ID: 5uc4), while a centroid grid box was used to create the binding
   pocket. The resulting grid box was used for docking Catalpol in Glide's
   “extra‐precision mode.” Subsequently, the docking pose of Catalpol was
   selected for the computation of Gibbs free energy, using the OPLS‐3
   force field, the VSGB solvent model, and the rotamer search tools.

Molecular Dynamic Simulation

   A 500‐ns MD simulation was performed to assess the stability of the
   Catalpol–Hsp90β complex after docking. To investigate this complex in
   an explicit solvent system, the Desmond module of the Schrodinger suite
   was used, applying the OPLS3 force field. The molecular system
   underwent solvation with the TIP3P model and was subjected to truncated
   octahedral periodic boundary conditions. The elimination of overlapping
   water molecules and neutralization of the system were accomplished
   through the introduction of sodium counterions. The system consists of
   24183 atoms and 6947 water molecules. Stable conditions were ensured at
   300 K and 1 bar using a Nose–Hoover thermostat and barostat ensemble
   (NPT). An energy minimization method, consisting of a hybrid technique
   involving 1000 steps of steepest descent and the Broyden‐Fletcher
   Goldfarb‐Shanno (LBFGS) algorithm, was used. The Smooth Particle Mesh
   Ewald method was used to facilitate long‐range electrostatic
   interactions, while short‐range van der Waals and Coulomb interactions
   were limited to a cut‐off radius of 9 Å. The dynamics of bonded,
   near‐bonded, and far‐bonded interactions was examined using multiple
   the time step RESPA integration (reference system propagator
   approaches) at intervals of 2, 2, and 6 frames s^−1, respectively. Data
   were collected every 200 ps, and the computational trajectory was
   assessed through simulation event analysis.

Silencing of Hsp90ab1 in Primary Chondrocytes and in OA Rats

   For the experiment of Hsp90ab1 silencing in chondrocytes, the
   incubation mixture comprising 10 pmol Hsp90ab1 silencer (Genepharma,
   China) or negative control siRNA (Genepharma, China), 2 µL
   Lipofectamine (Thermo, USA) and 47 µL OPTI medium (Gibco, USA) was
   added to a 24‐well plate, and each well contained 1 × 10^5 cells
   growing without adherence. The cells were grown in a 5% CO[2], 95%
   atmospheric air incubator at 37 °C. For the experiment of Hsp90ab1
   silencing in OA rats, siRNA‐Hsp90ab1 was modified by ago for
   intra‐articular injection in rats. The sequence of ago‐siRNA‐Hsp90ab1
   or negative control siRNA was designed and constructed by GenePharma
   Co., Ltd. 6 nmol siRNA‐Hsp90ab1 in 10 µL PBS were injected into OA rat
   joint cavity. OA rats received injections once weekly for four weeks.^[
   [187]^46 ^] Silencing efficiency was then detected by qRT‐PCR using
   specific primers listed in Table [188]S1 (Supporting Information). The
   information of siRNA is listed in Table [189]S2 (Supporting
   Information).

Protein–Protein Interaction (PPI) Network

   Protein–protein interaction network was constructed and downloaded from
   the STRING database ([190]https://www.string‐db.org/).

Statistical Analysis

   The numerical data presented in the figures were analyzed by SPSS
   software and presented as the mean ± SEM (n ≥ 3). Significant
   differences between the two groups were determined to use unpaired
   Student's t‐tests. ANOVA was used for multiple comparisons. The data
   were tested for normality of distribution using Shapiro–Wilk test
   before analysis. P‐values < 0.05 were considered statistically
   significant.

Conflict of Interest

   The authors declare no conflict of interest.

Ethical Statement

   This study was approved by the Ethics Committee of Changchun University
   of Traditional Chinese Medicine (No. 2023029).

Author Contributions

   Designed and drafted the manuscript: H.D., H.S., Z.Z., X.L., and D.Z.
   Methodology: Z.Z., H.S., H.D., and X.L. Chondrogenesis‐related
   experiments: B.Z., L.L., J.Y., and Z.Z. Identified the drug target:
   Z.Z., H.D., B.Z., and L.L. Drug screening: H.S., Y.W., and Z.Z.
   Synthesis and characterization of bMSNs and Ca‐ bMSNs: C.L. and Y.W.
   Micro‐CT analysis: J.Y. and H.Z. Expression and purification of
   protein: Y.W., Z.J., and Z.P.

Supporting information

   Supporting Information
   [191]ADVS-12-2503909-s001.docx^ (13.1MB, docx)

Acknowledgements