Therapeutic Delivery of rAAV sox9 via Polymeric Micelles Counteracts the Effects of Osteoarthritis-Associated Inflammatory Cytokines in Human Articular Chondrocytes

Osteoarthritis (OA) is a prevalent joint disease linked to the irreversible degradation of key extracellular cartilage matrix (ECM) components (proteoglycans, type-II collagen) by proteolytic enzymes due to an impaired tissue homeostasis, with the critical involvement of OA-associated pro-inflammatory cytokines (interleukin 1 beta, i.e., IL-1β, and tumor necrosis factor alpha, i.e., TNF-α). Gene therapy provides effective means to re-establish such degraded ECM compounds by rejuvenating the altered OA phenotype of the articular chondrocytes, the unique cell population ubiquitous in the articular cartilage. In particular, overexpression of the highly specialized SOX9 transcription factor via recombinant adeno-associated viral (rAAV) vectors has been reported for its ability to readjust the metabolic balance in OA, in particular via controlled rAAV delivery using polymeric micelles as carriers to prevent a possible vector neutralization by antibodies present in the joints of patients. As little is known on the challenging effects of such naturally occurring OA-associated pro-inflammatory cytokines on such rAAV/polymeric gene transfer, we explored the capacity of polyethylene oxide (PEO) and polypropylene oxide (PPO)-based polymeric micelles to deliver a candidate rAAV-FLAG-hsox9 construct in human OA chondrocytes in the presence of IL-1β and TNF-α. We report that effective, micelle-guided rAAV sox9 overexpression enhanced the deposition of ECM components and the levels of cell survival, while advantageously reversing the deleterious effects afforded by the OA cytokines on these processes. These findings highlight the potentiality of polymeric micelles as effective rAAV controlled delivery systems to counterbalance the specific contribution of major OA-associated inflammatory cytokines, supporting the concept of using such systems for the treatment for chronic inflammatory diseases like OA.


Introduction
Osteoarthritis (OA) represents a prevalent, chronic, and deteriorating joint affliction that is the leading cause of impaired function and disability [1]. OA is characterized by multiple functional and structural cartilage tissue and cell shifts, such as the progressive and permanent degradation of the articular cartilage matrix (loss of type-II collagen and of proteoglycans), the restructuration of the subchondral bone, and the formation of osteophytes [2,3] due to defective homeostasis [4,5].

Preparation of Micellar Copolymer Solutions Containing rAAV-FLAG-hsox9 Vectors
Copolymer solutions (PF68 or T908) were prepared in 10% sucrose aqueous solution at 4 • C, mixed with rAAV-FLAG-hsox9, and maintained in ice-water bath for 30 min prior to their use as previously described [21][22][23]. The final micellar concentration into the culture medium was 2%. Effective interaction between the vectors and the polymeric micelles was confirmed by dynamic light scattering and electron microscopy [21][22][23].
Expression of SOX9 was monitored by immunocytochemistry using and anti-SOX9 specific primary antibody, a biotinylated secondary antibody, with the ABC method with diaminobenzidine (DAB) as previously described [23,29]. To control for secondary immunoglobulins, OA chondrocytes in monolayers cultures were assayed with exclusion of the primary antibody. All cultures were inspected under light microscopy (Olympus CKX41).

Histological and Immunocytochemical Analyses
Chondrocytes in monolayer cultures were harvested after 1 and 10 days and fixed in 4% formalin [21][22][23] prior to the immunocytochemical analyses. Expression of SOX9 and type-II collagen was detected using specific primary and biotinylated secondary antibodies, and the ABC method with DAB chromogen, with examination under light microscopy (Olympus CKX41) [21,22]. Alcian blue Nanomaterials 2020, 10, 1238 4 of 13 staining was involved to detect matrix proteoglycans [21,22,34,35]. Briefly, fixed monolayer cultures were stained with Alcian blue (1% in HCl 1 N) and excess stain was washed with double distilled water. The staining was solubilized by overnight incubation in 6 M guanidine hydrochloride and the absorbance at 595 nm was quantified with a GENios spectrophotometer (Tecan Crailsheim, Germany).

Histomorphometry
The mean intensities of SOX9 and type-II collagen immunostaining (ratio of positively stained surface to the total surface) were assessed at four randomized locations for each replicate condition as previously described [21][22][23]. Analyses were accomplished by using SIS AnalySIS (Olympus, Hamburg, Germany) and Adobe Photoshop (Adobe Systems Software CS2, Unterschleissheim, Germany) [21,23].

Evaluation of Cell Proliferation and Viability
Proliferation of chondrocytes in monolayer cultures was estimated using the Cell Proliferation Reagent WST-1, with optical density (OD) values proportional to the cell numbers [22,23,30]. Controls included the same conditions depicted in 2.5. ODs at 450 nm were registered using a GENios spectrophotometer (Tecan) and the percent's of cell viability were calculated as follows: (1)

Statistical Analysis
Each condition was tested in duplicate in four independent experiments using all patients. The values registered are depicted as mean ± standard deviation (SD). A t-test was employed, with p < 0.05 being considered statistically significant.

Efficacy of rAAV-Mediated sox9 Overexpression in Conditions of Inflammation upon Vector Delivery via Polymeric Micelles
We first evaluated whether the presence of pro-inflammatory cytokines may alter the overexpression of SOX9 in human OA chondrocytes monolayer cultures.
In agreement with our previous observations [23], effective SOX9 overexpression was noted in the cells via rAAV-FLAG-hsox9 transduction (up to a 1.4-fold increase relative to the negative control in the absence of cytokines, p = 0.021) ( Figure 1A,B). Similarly, supply of rAAV in polymeric micelles led to the most intense SOX9 immunoreactivity (up to a 1.8-fold increase when compared with the negative control in the absence of cytokines on day 1, p = 0.015), leading to more sustained levels of expression over time (up to a 1.4-fold difference with respect to the cell control in the absence of cytokines on day 10, p = 0.009) ( Figure 1A,B).
Treatment with IL-1β did not alter the levels of SOX9 expression early on (p = 0.450 compared with the negative control in the absence of cytokines on day 1) ( Figure 1A,B versus Figure 1C,D) while a reduction was noted after 10 days (up to a 1.1-fold difference with respect to the negative control in the absence of cytokines, p = 0.106) ( Figure 1A,B versus Figure 1C,D). Interestingly, the treatment with rAAV-FLAG-hsox9 promoted a significant enhancement in SOX9 expression levels following IL-1β treatment (up to a 1.5-fold difference with respect to the cell control in the presence of IL-1β, p = 0.024), especially upon vector delivery via micellar systems (up to a 1.7-fold difference when compared with the cell control in the presence of IL-1β on day 1, p = 0.040) ( Figure 1C,D). Such effects were also maintained over the time of evaluation (up to a 1.5-fold increase with respect to the cell control in the presence of IL-1β on day 10, p = 0.045) ( Figure 1C,D).
Administration of TNF-α did not affect the levels of SOX9 expression, regardless of the time points evaluated (p = 0.090 compared with the cell control in the absence of cytokines) ( Figure 1A,B versus Figure 1E,F). Overexpression of sox9 via rAAV led to increased levels of SOX9 expression Nanomaterials 2020, 10, 1238 5 of 13 (up to a 1.4-fold increase with respect to the cell control in the presence of TNF-α on day 10, p = 0.048) ( Figure 1E,F). Notably, delivery of the vector via micellar carriers resulted in the highest levels of SOX9 expression (up to a 1.7-fold difference with respect to the cell control in the presence of TNF-α on day 1, p = 0.021) ( Figure 1E,F).
Concomitant IL-1β/TNF-α application led to a decrease in the levels of SOX9 expression (up to a 1.2-fold difference with respect to the negative control in the absence of cytokines on day 10, p = 0.150) ( Figure 1A,B versus Figure 1G,H). Significantly increased SOX9 levels were noted either using free rAAV-FLAG-hsox9 form (up to a 1.5-fold difference relative to the cell control in the presence of IL-1β/TNF-α on day 10, p = 0.019) ( Figure 1G,H), or via delivery in PF68 or T908-based micelles (up to a 1.6-fold increase with respect to the cell control in the presence of IL-1β/TNF-α on day 1, p = 0.020) ( Figure 1G,H). Administration of TNF-α did not affect the levels of SOX9 expression, regardless of the time points evaluated (p = 0.090 compared with the cell control in the absence of cytokines) ( Figure 1A,B versus Figure 1E,F). Overexpression of sox9 via rAAV led to increased levels of SOX9 expression (up to a 1.4-fold increase with respect to the cell control in the presence of TNF-α on day 10, p = 0.048) ( Figure 1E,F). Notably, delivery of the vector via micellar carriers resulted in the highest levels of SOX9 expression (up to a 1.7-fold difference with respect to the cell control in the presence of TNF-α on day 1, p = 0.021) ( Figure 1E,F).
Concomitant IL-1β/TNF-α application led to a decrease in the levels of SOX9 expression (up to a 1.2-fold difference with respect to the negative control in the absence of cytokines on day 10, p = 0.150) ( Figure 1A,B versus Figure 1G,H). Significantly increased SOX9 levels were noted either using free rAAV-FLAG-hsox9 form (up to a 1.5-fold difference relative to the cell control in the presence of IL-1β/TNF-α on day 10, p = 0.019) ( Figure 1G,H), or via delivery in PF68 or T908-based micelles (up to a 1.6-fold increase with respect to the cell control in the presence of IL-1β/TNF-α on day 1, p = 0.020) ( Figure 1G,H).

Effects of rAAV-FLAG-hsox9/Polymeric Micelle Delivery on the Anabolic Activities of Human OA Chondrocytes in Inflammatory Conditions
We next investigated the effects of SOX9 overexpression on the deposition of type-II collagen and proteoglycans following rAAV-FLAG-hsox9 gene transfer via micellar vehicles in human OA chondrocytes monolayer cultures maintained in conditions of inflammation.
Administering of rAAV-FLAG-hsox9 significantly incremented type-II collagen deposition in the cells (up to a 1.3-fold increase with respect to the cell control in the absence of cytokines, p = 0.017) (Figure 2A,B). These levels increased over time, chiefly by delivery of the vectors via PF68 micelles (up to a 1.5-fold difference with respect to the negative control in the absence of cytokines on day 10, p = 0.040) (Figure 2A,B). Of note, these levels were higher than those achieved with free vector administration (up to a 1.2-fold difference compared with free rAAV-FLAG-hsox9 application on day 10, p = 0.011) (Figure 2A,B). Treatment with IL-1β decreased type-II collagen deposition (up to a 1.1-fold difference relative to the negative control in the absence of cytokines on day 1, p = 0.300) (Figure 2A,B versus Figure 2C,D). rAAV-FLAG-hsox9 application to IL-1β-treated chondrocytes significantly increased type-II collagen deposition, especially when using micelle-guided vector delivery (up to a 1.5-fold difference with respect to the cell control in the presence of IL-1β on day 10, p = 0.011; up to a 1.2-fold difference compared with free vector applying in the presence of IL-1β on day 10, p = 0.006) ( Figure 2C,D). A similar tendency was noted when applying TNF-α alone or combined as a IL-1β/TNF-α co-treatment, showing modest decreases in type-II collagen deposition compared with cells kept in culture in the absence of cytokines (p = 0.290) (Figure 2A,B versus Figure 2E-H). Similarly, overexpression of SOX9 significantly increased type-II collagen deposition over time, particularly when providing rAAV-FLAG-hsox9 in micellar carriers (up to a 1.4-fold difference with respect to the cell control in the presence of TNF-α alone or as an IL-1β/TNF-α combination on day 10, p = 0.040) ( Figure 2E-H).

Effects of rAAV-FLAG-hsox9/Polymeric Micelle Delivery on the Anabolic Activities of Human OA Chondrocytes in Inflammatory Conditions
We next investigated the effects of SOX9 overexpression on the deposition of type-II collagen and proteoglycans following rAAV-FLAG-hsox9 gene transfer via micellar vehicles in human OA chondrocytes monolayer cultures maintained in conditions of inflammation.
Administering of rAAV-FLAG-hsox9 significantly incremented type-II collagen deposition in the cells (up to a 1.3-fold increase with respect to the cell control in the absence of cytokines, p = 0.017) (Figure 2A,B). These levels increased over time, chiefly by delivery of the vectors via PF68 micelles (up to a 1.5-fold difference with respect to the negative control in the absence of cytokines on day 10, p = 0.040) (Figure 2A,B). Of note, these levels were higher than those achieved with free vector administration (up to a 1.2-fold difference compared with free rAAV-FLAG-hsox9 application on day 10, p = 0.011) (Figure 2A,B). Treatment with IL-1β decreased type-II collagen deposition (up to a 1.1fold difference relative to the negative control in the absence of cytokines on day 1, p = 0.300) ( Figure  2A,B versus Figure 2C,D). rAAV-FLAG-hsox9 application to IL-1β-treated chondrocytes significantly increased type-II collagen deposition, especially when using micelle-guided vector delivery (up to a 1.5-fold difference with respect to the cell control in the presence of IL-1β on day 10, p = 0.011; up to a 1.2-fold difference compared with free vector applying in the presence of IL-1β on day 10, p = 0.006) ( Figure 2C,D). A similar tendency was noted when applying TNF-α alone or combined as a IL-1β/TNF-α co-treatment, showing modest decreases in type-II collagen deposition compared with cells kept in culture in the absence of cytokines (p = 0.290) (Figure 2A,B versus Figure 2E-H). Similarly, overexpression of SOX9 significantly increased type-II collagen deposition over time, particularly when providing rAAV-FLAG-hsox9 in micellar carriers (up to a 1.4-fold difference with respect to the cell control in the presence of TNF-α alone or as an IL-1β/TNF-α combination on day 10, p = 0.040) ( Figure 2E-H).  Overexpression of SOX9 in rAAV-FLAG-hsox9-transduced chondrocytes significantly increased the accretion of ECM-proteoglycans compared with untransduced cells (up to an 1.8-fold difference with respect to the negative control in the absence of cytokines on day 10, p = 0.030) ( Figure 3A,B). Of note, delivery of rAAV-FLAG-hsox9 via micellar systems led to the highest proteoglycan deposition (up to a 1.2-fold increase with respect to free vector administering on day 10, p = 0.030) and proliferative index (up to a 1.4-fold increase with respect to the negative control in the absence of cytokines) ( Figure 3A,B). Strikingly, treatment with IL-1β significantly decreased the deposition of proteoglycans and the cell proliferation ratio (up to a 1.2-fold difference with respect to the negative control in the absence of cytokines on day 1, p = 0.030) ( Figure 3A,B versus Figure 3C,D). Additionally, rAAV-FLAG-hsox9-mediated transduction of IL-1β-treated chondrocytes prompted the restoration of proteoglycans, an effect more marked over time (up to a 1.7-fold increase when compared to the control in the presence of IL-1β on day 10, p = 0.038), exhibiting higher cell proliferation. Interestingly, providing rAAV-FLAG-hsox9 in micellar carriers led to the highest proteoglycan deposition (up to a 2.1-fold increase relative to the cell control in the presence of IL-1β on day 10, p = 0.006), reaching values that were higher than those reached with the free vector administration (up to a 1.3-fold difference with respect to free rAAV-FLAG-hsox9 application in the presence of IL-1β on day 10, p = 0.046) ( Figure 3C,D). Treatment with TNF-α also decreased the deposition of proteoglycans (up to a 1.2-fold difference with respect to the negative control in the absence of cytokines on day 10, p = 0.203) and the cell proliferation index ( Figure 3A,B versus Figure 3E,F). Transduction of TNF-α-treated chondrocytes with rAAV-FLAG-hsox9 significantly increased the cell proliferation and proteoglycan deposition, especially when the vectors were delivered via micellar systems (up to a 2-fold difference compared with the cell control in the presence of TNF-α on day 10, p = 0.010) ( Figure 3E,F). Simultaneous IL-1β/TNF-α administration significantly decreased the deposition of proteoglycans (up to a 1.1-fold difference with respect to the negative control in the absence of cytokines on day 1, p = 0.011) and the cell proliferation rates ( Figure 3A,B versus Figure 3G,H). Again, SOX9 overexpression increased the deposition of proteoglycans following IL-1β/TNF-α treatment, especially when the vectors were transferred via micellar vehicles (up to 2-fold difference relative to the cell control in the presence of IL-1β/TNF-α on day 10, p = 0.043; up to a 1.5-fold difference compared with free vector administration in the presence of IL-1β/TNF-α on day 10, p = 0.009) ( Figure 3G,H). Likewise, genetic modification of chondrocytes via rAAV-FLAG-hsox9 resulted in an increased proliferation index (up to a 1.8-fold relative to the cell control in the presence of IL-1β/TNF-α on day 10, p = 0.001).

Effects of rAAV-FLAG-hsox9/Polymeric Micelle Delivery on the Viability Processes in Human OA Chondrocytes in Inflammatory Conditions
We finally examined the effects of SOX9 overexpression on the cell viability processes following rAAV-FLAG-hsox9 gene transfer via micellar systems in human OA chondrocytes monolayer cultures maintained in conditions of inflammation.

Effects of rAAV-FLAG-hsox9/Polymeric Micelle Delivery on the Viability Processes in Human OA Chondrocytes in Inflammatory Conditions
We finally examined the effects of SOX9 overexpression on the cell viability processes following rAAV-FLAG-hsox9 gene transfer via micellar systems in human OA chondrocytes monolayer cultures maintained in conditions of inflammation.
In concordance with our previous observations [23], no cytotoxic effects from none of the gene transfer procedures (polymeric vehicles, free vector supply) were noticed with respect to the control condition (p = 0.130) ( Figure 4A). A similar tendency was evidenced when providing copolymer solutions in the absence of vector treatment (not shown). Moreover, while separate cytokine treatment resulted only in slight decreases in cell viability (~90%) ( Figure 4B,C), concomitant administration of both cytokines led to higher toxicity especially in untransduced cells (~75% cell In concordance with our previous observations [23], no cytotoxic effects from none of the gene transfer procedures (polymeric vehicles, free vector supply) were noticed with respect to the control condition (p = 0.130) ( Figure 4A). A similar tendency was evidenced when providing copolymer solutions in the absence of vector treatment (not shown). Moreover, while separate cytokine treatment resulted only in slight decreases in cell viability (~90%) ( Figure 4B,C), concomitant administration of both cytokines led to higher toxicity especially in untransduced cells (~75% cell viability on day 10 in the negative control in the presence of IL-1β/TNF-α) ( Figure 4D). Strikingly, overexpression of SOX9 led to higher cell viability indices in the presence of both cytokines (~100% compared with the cell control in the presence of IL-1β/TNF-α on day 10, p = 0.045) ( Figure 4D). Nanomaterials 2020, 10, 1238 9 of 13 viability on day 10 in the negative control in the presence of IL-1β/TNF-α) ( Figure 4D). Strikingly, overexpression of SOX9 led to higher cell viability indices in the presence of both cytokines (~100% compared with the cell control in the presence of IL-1β/TNF-α on day 10, p = 0.045) ( Figure 4D).

Discussion
A potential means to counterbalance the disrupted cartilage homeostasis altered during OA disease is based on the correction of specific chondrocyte gene expression patterns [19]. Herein, transcription factors are critical mediators of cartilage metabolism prompting chondrogenesis in both physiologic and pathologic conditions [19]. Among them, SOX9 plays vital roles in the settlement of skeletal and cartilage formation [36] and the differentiation of chondrocytes [37]. Several studies have reported a decline in SOX9 expression in OA pathology [38,39]. Therefore, genetic adjustment of the levels of SOX9 expression may constitute a valuable strategy for re-equilibrating the disturbed balance characteristic of OA cartilage towards the synthesis of ECM compounds, affording the rescue of a native articular cartilage surface [19]. rAAV vectors are convenient carriers for efficiently and steadily targeting human OA chondrocytes [11,19] and avoiding the shortcomings and/or risks inherent to other types of vectors (short-term nonviral vectors, immunogenic adenoviral vectors, potentially tumorigenic retro-/lentiviral vectors) [40,41]. However, clinical administration of rAAV for OA treatments in patients may be hindered by the prevalence of circulating anti-AAV capsid antibodies in the subjects [42], especially in the synovial fluid from patients affected with joint disorders [20]. To overcome this hurdle, we evidenced the capability of PEO-PPO-PEO-based polymeric micelles (PF68 and T908) to efficiently and durably deliver rAAV vectors with increased stability and bioactivity to chondrocytes and mesenchymal stem cells (MSCs), affording protection against neutralizing antibodies [21,31]. Equally important, rAAV-mediated gene transfer of sox9 via polymeric micelle delivery resulted in the remodeling of OA cartilage, with increased proteoglycan accumulation and cell proliferation in OA chondrocytes relative to free vector administration [23].
In light of these observations, the goal of the present study was to test the potentiality of these micellar nanocarriers to deliver the rAAV-FLAG-hsox9 vector to human OA chondrocytes in an

Discussion
A potential means to counterbalance the disrupted cartilage homeostasis altered during OA disease is based on the correction of specific chondrocyte gene expression patterns [19]. Herein, transcription factors are critical mediators of cartilage metabolism prompting chondrogenesis in both physiologic and pathologic conditions [19]. Among them, SOX9 plays vital roles in the settlement of skeletal and cartilage formation [36] and the differentiation of chondrocytes [37]. Several studies have reported a decline in SOX9 expression in OA pathology [38,39]. Therefore, genetic adjustment of the levels of SOX9 expression may constitute a valuable strategy for re-equilibrating the disturbed balance characteristic of OA cartilage towards the synthesis of ECM compounds, affording the rescue of a native articular cartilage surface [19]. rAAV vectors are convenient carriers for efficiently and steadily targeting human OA chondrocytes [11,19] and avoiding the shortcomings and/or risks inherent to other types of vectors (short-term nonviral vectors, immunogenic adenoviral vectors, potentially tumorigenic retro-/lentiviral vectors) [40,41]. However, clinical administration of rAAV for OA treatments in patients may be hindered by the prevalence of circulating anti-AAV capsid antibodies in the subjects [42], especially in the synovial fluid from patients affected with joint disorders [20]. To overcome this hurdle, we evidenced the capability of PEO-PPO-PEO-based polymeric micelles (PF68 and T908) to efficiently and durably deliver rAAV vectors with increased stability and bioactivity to chondrocytes and mesenchymal stem cells (MSCs), affording protection against neutralizing antibodies [21,31]. Equally important, rAAV-mediated gene transfer of sox9 via polymeric micelle delivery resulted in the remodeling of OA cartilage, with increased proteoglycan accumulation and cell proliferation in OA chondrocytes relative to free vector administration [23].
In light of these observations, the goal of the present study was to test the potentiality of these micellar nanocarriers to deliver the rAAV-FLAG-hsox9 vector to human OA chondrocytes in an environment similar to that in OA, i.e., in the presence of pro-inflammatory IL-1β and TNF-α cytokines [4,5,24]. First, and in good concordance with our previous findings [23], the data indicate that the transfer of rAAV-FLAG-hsox9 to human OA chondrocytes via polymeric micelles led to enhanced levels of SOX9 expression over time relative to free vector treatment. Of note, rAAV-FLAG-hsox9 transduction of chondrocytes prompted elevated and sustained levels of SOX9 expression in cells treated with IL-1β, especially when the vectors were carried by the polymeric micelles. A similar trend was observed in the presence of TNF-α alone or combined with IL-1β (IL-β/TNF-α condition), showing that delivery of the vectors via micellar systems to the highest levels of SOX9 expression. Likewise, rAAV-FLAG-hsox9-mediated treatment in the presence of IL-β/TNF-α increased the levels of SOX9 expression. These results are in agreement with previous work reporting an increased rAAV-mediated modification of fibroblast-like synoviocytes in conditions of inflammation [43].
The results next indicate that rAAV sox9 treatment led to significantly higher levels of type-II collagen deposition compared with untransduced controls, most particularly when the vectors were delivered via polymeric micelles, concordant with our previous work when providing rAAV-FLAG-hsox9 to experimental human osteochondral defects [23] and with the pro-anabolic properties of this transcription factor [44,45]. Of further note, while administration of IL-1β to chondrocyte cultures decreased the levels of type-II collagen deposition, in agreement with previous findings [7], transduction with rAAV-FLAG-hsox9 reversed such undesirable effects by increasing type-II collagen deposition in IL-1β-treated chondrocytes especially when providing the construct via polymeric micelles, expanding earlier work using lentiviral delivery of sox9 [8]. In this regard, the use of rAAV provides strong advantages for clinical translation, as they do not carry the risk of insertional mutagenesis inherent to lentiviruses [9]. Similar observations were made following TNF-α treatment (alone or combined with IL-1β) and genetic modification via rAAV-FLAG-hsox9, with increased type-II collagen deposition especially using polymeric micelle-guided rAAV gene transfer. Also remarkably, SOX9 overexpression via rAAV was capable of reverting the inhibitory effects of the cytokines upon the deposition of proteoglycans and the proliferation index [7], especially when delivering the therapeutic construct in polymeric micelles, again expanding work with lentiviral gene delivery of sox9 [8], and concordant with the pro-anabolic activities of the transcription factor [45]. Moreover, no detrimental effects were noted, regardless of the gene transfer method adopted, as previously described with rAAV [23], and SOX9 overexpression was again capable to counteract the cytotoxic effects of the cytokines by preserving the viability of the OA chondrocytes, in agreement with work highlighting the role of SOX9 to preserve chondrocyte survival [8].

Conclusions
The present study shows the potentiality of polymeric micelles as powerful rAAV controlled delivery systems to counteract the specific contribution of major OA-associated inflammatory cytokines in chondrocyte cultures. Here, we provide concrete evidence that encapsulation of an rAAV vector carrying a sox9 sequence in such systems promotes significant SOX9 expression levels capable of increasing the deposition of major ECM components (type-II collagen, proteoglycans) and the cell survival processes in human OA chondrocytes while reversing their downregulation afforded by OA cytokines. While this work evidence the utility of such micellar systems to tackle the OA phenotype in chondrocytes in a 2D environment, work is currently ongoing to broaden this investigation at longer time points and to support the present findings when cells are embedded in their own pericellular matrix using an experimental model of osteochondral defect in situ [22,23] where the chondrocytes may also be influenced by interplay with subchondral bone cells that have key roles in OA development and progression. Overall, such observations show the effectiveness of polymeric micelles as rAAV controlled delivery systems in an inflammatory environment, making them attractive tools for the treatment for chronic inflammatory diseases like OA.