Gelatin hydrogels with eicosapentaenoic acid can prevent osteoarthritis progression in vivo in a mouse model

Eicosapentanoic acid (EPA) is an antioxidant and omega‐3 polyunsaturated fatty acid that reduces inflammatory cytokine production. Gelatin hydrogel can be used as a carrier of a physiologically active substance that release it gradually for an average of ~3 weeks. Therefore, this study aimed to clarify the effect of EPA‐incorporating gelatin hydrogels on osteoarthritis (OA) progression in vivo. Ten‐week‐old male C57BL/6J mice were randomly divided into six groups (n = 6): Sham, destabilization of the medial meniscus (DMM), Corn: DMM + 2 µL corn oil, EPA injection alone (EPA‐I): DMM + 2 µL corn oil + 125 μg/μL EPA, Gel: DMM + gelatin hydrogels, and EPA‐G: DMM + 125 μg/μL EPA‐incorporating gelatin hydrogels. The mice were euthanized at 8 weeks after DMM or Sham surgery, and subjected to histological evaluation. Matrix‐metalloproteinases‐3 (MMP‐3), MMP‐13, interleukin‐1β (IL‐1β), p‐IKK α/β, CD86, and CD163 protein expression in the synovial cartilage was detected by immunohistochemical staining. F4/80 expression was also assessed using the F4/80 score of macrophage. Histological score was significantly lower in EPA‐G than in EPA‐I. MMP‐3‐, MMP‐13‐, IL‐1β‐, and p‐IKK α/β‐positive cell ratio was significantly lower in EPA‐G than in EPA‐I. However, CD86‐ and CD163‐positive cell ratio was not significantly different between EPA‐I and EPA‐G. The average‐sum F4/80 score of macrophage in EPA‐G was significantly lower than that in EPA‐I. EPA‐incorporating gelatin hydrogels were shown to prevent OA progression in vivo more effectively than EPA injection alone. Our results suggested that intra‐articular administration of controlled‐release EPA can be a new therapeutic approach for treating OA.


| INTRODUCTION
Interleukin-1β (IL-1β) induces pro-inflammatory effect by activating the protein kinase cascades involved in the nuclear factor-κB (NF-κB) signaling pathways. 1 The induction requires phosphorylation of IKKα/β (p-IKKα/β), which phosphorylates and degrades IκB molecules. IκB release results in the nuclear localization and binding of NF-κB to specific DNA sites, inducing the transcription of target genes, such as matrix-metalloproteinases (MMPs). 2,3 MMP-1 and MMP-13 induce the digestion of type II collagen, a major collagen in hyaline cartilages, whereas MMP-3 plays a role in extracellular matrix degradation by degrading proteoglycans. 4 Thus, IL-1β plays an important role in osteoarthritis (OA) etiology by disrupting the catabolic and anabolic processes as well as causing cartilage matrix loss. Therefore, pharmacological treatments that inhibit the effect of IL-1β are useful strategies for OA treatment.
Eicosapentaenoic acid (EPA) is an antioxidant and a type of n-3 polyunsaturated fatty acid (PUFA) contained in fish oil. 5 Currently, n-3 PUFAs are well-known for their anti-inflammatory and immunomodulatory properties as well as for their efficacy in reducing inflammatory cytokine production. 6 n-3 PUFAs exert their antiinflammatory effects through multiple mechanisms, including inhibition of arachidonic acid conversion to pro-inflammatory eicosanoids, synthesis of anti-inflammatory products, and downregulation of proinflammatory gene expression. 7,8 In recent years, the effects of n-3 PUFAs have been extensively studied in cancer, inflammatory bowel disease, and other autoimmune diseases, such as rheumatoid arthritis. 9,10 Regarding OA, we previously reported that EPA treatment inhibits chondrocyte apoptosis and MMPs expression induced by oxidative stress in vitro as well as prevents OA progression in mice. 11 However, in that previous study, weekly injection of EPA into the knee joint was required.
Gelatin hydrogels have been developed as a potentially safe drug delivery system that contains physiologically active substance and releases the substance gradually for an average of less than 3 weeks. 12 It was reported that the incorporation of a gelatin hydrogel sheet with growth factors and its subsequent adsorption using electrostatic force preserved the biological activity of the growth factors. 13 Therefore, the preserved bioactive growth factors can be sustainably released until the gelatin hydrogel sheet is degraded, and the release periods can be controlled by changing the water content of the hydrogels. 14 Thus, a gelatin hydrogel sheet has an excellent use of containing and delivering medications to the applied area as a drug delivery system, providing mechanical support owing to its structural and physiological properties. [13][14][15] The benefits of gelatin hydrogels are also recognized in orthopedic surgery. 16,17 Oka et al 16 reported that simvastatin-conjugated gelatin hydrogels promote tendon-bone healing by affecting both angiogenesis and osteogenesis. Another study reported that a combination of bone morphogenetic protein-2 (BMP-2)-releasing gelatin/β-tricalcium phosphate (β-TCP) sponge is a promising technique to induce bone regeneration in bone marrow for treating segmental bone defects induced by X-ray irradiation. 17 These findings on the antioxidant action of EPA and the excellent effect of gelatin hydrogels as an innovative drug delivery system suggested that degradable gelatin hydrogels containing EPA may be more effective than single injection of EPA in preventing OA progression. We hypothesized that the in vivo effect of gelatin hydrogels containing EPA is more potent than that of single EPA injection. Therefore, this study aimed to investigate the effect of gelatin hydrogels containing EPA on OA progression in vivo.

| MATERIALS AND METHODS
This study was performed in strict accordance with the recommendations from the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (Bethesda, MD). All procedures were approved by the Animal Studies Committee of our institute (permit number: P130108). Thirty 10-week-old male C57BL/6J mice were used to investigate the sustained release of EPA in vivo, and 72 mice were used for histological evaluation in vivo.

| Gelatin hydrogel incorporating EPA
EPA was provided by Sigma-Aldrich (St Louis, MO). Gelatin hydrogels containing EPA were produced according to a previous report. 18 Briefly, in the first step, L-lactic acid oligomers (LAo) with a mean molecular weight of 1000 were synthesized from L-lactide monomers through ring-opening polymerization with stannous octoate as a catalyst and 1-dodecanol as an initiator. LAo-grafted gelatin was synthesized by activating the hydroxyl groups of LAo and mixing them with gelatin solution. EPA/ethanol solution was added to the LAo-grafted gelatin aqueous solution. The reaction mixture was freeze-dried to obtain EPA water-solubilized by LAo-grafted gelatin micelles (EPA micelles). EPA-micelle aqueous solution and gelatin hydrogel aqueous solution were mixed and adjusted to prepare the desired amount of EPA. Second, 2-μL drops of the mixed solution were placed on a parafilm. Pellets of gelatin hydrogels containing EPA micelles were prepared through dehydrothermal crosslinking (140°C, 24 hours) of gelatin and ethylene oxide gas sterilization. The obtained solid small pellets of gelatin hydrogels containing EPA micelles were used in this study. The small pellets shape was hemispherical with a diameter of approximately 1.5 mm. EPA content of the pellets was 125 μg/μL. The gelatin hydrogels were designed to biodegrade over a period of~3 weeks under in vivo conditions. 19 2.2 | Release test of EPA from gelatin hydrogel incorporating EPA The supernatant was collected and frozen at −80℃ until analysis.
The amount of EPA in the supernatant was measured according to previous report using high-performance liquid chromatography (HPLC). 20 The mass of EPA obtained by HPLC at each time-point was measured, and the ratio of the mass compared with that at time-point when gelatin hydrogel was completely dissolved was calculated and this assay was repeated five times. To investigate the sustained release of EPA in vivo, 10-week-old male C57BL/6J mice (n = 30) were obtained from the CREA Japan (Tokyo, Japan). Gelatin hydrogels were placed into the knee joint cavity. These mice were sacrificed, and the gelatin hydrogels were taken from the knee joint cavity at 1, 3, 7, 14, 21, and 28 days after surgery. EPA concentration of gelatin hydrogels was measured by the same procedure as in vitro.

| In vivo study
Ten-week-old male C57BL/6J mice (n = 72) were obtained from the CREA Japan. The animals were bred in a mouse house with automatically controlled lightening (12 hours on and 12 hours off) and a stable temperature of 23°C, and they received food and water ad libitum. The mice were anesthetized by intraperitoneal injection of medetomidine (0.3 mg/kg), midazolam (4 mg/kg), and butorphanol (5 mg/kg), and destabilization of the medial meniscus (DMM) surgery was performed on the right-knee joint through transection of the anterior attachment of the medial meniscotibial ligament, as described previously. 21,22 The control group was subjected to Sham surgery, in which skin and capsule incision and suture were performed on skin and capsule of the right-knee joints where the ligaments were intact. We randomly divided 72 mice into six groups: Sham, DMM, Corn (DMM and injection of corn oil), EPA-I (DMM and injection of corn oil and EPA), Control (DMM and gelatin hydrogels), and EPA-G (DMM and gelatin hydrogels containing EPA). Corn oil (2 μL) alone or corn oil (2 μL) plus EPA (125 μg/μL) was injected once into the knee joint immediately after DMM surgery, and capsule was closed. Control gelatin hydrogels (2 μL) or EPA (125 μg/μL)containing gelatin hydrogels (2 μL) were placed into the knee joint cavity during DMM surgery, and capsule was closed. One gelatin hydrogel was placed in each case. The amount of EPA was determined according to a previously reported method. 11 All mice were sacrificed and knee joints analyzed at 1 and 8 weeks after DMM or Sham surgery, and subjected to histological evaluation. Six mice were analyzed in each group.

| Histological evaluation of cartilage degeneration and synovitis
Mouse knee joints were fixed in 4% paraformaldehyde for 24 hours, decalcified with 14% ethylenediaminetetraacetic acid for 7 days, and embedded in paraffin. Coronal histological sections were obtained through the joint at 80 μm intervals. Cartilage degeneration was evaluated by Safranin-O and Fast Green staining. Histological evaluation of OA was performed using the cartilage OA histopathology scoring system of the Osteoarthritis Research Society International (OARSI). 23 Histological scores were measured in four quadrants (medial femoral condyle, medial tibial plateau, lateral femoral condyle, and lateral tibial plateau) of the knee joints at all sectioned levels (eight sections per knee) to obtain summed OA scores. Summed scores were calculated from all four quadrants for all sections that represented whole-joint changes. 23 Synovitis was also evaluated using an established scoring system by Lewis et al 24 and hematoxylin-eosin staining. The average score for changes in synovial lining thickness and cellular density in the synovial stroma (maximum site score 6) of each compartment in the coronal slice was obtained as synovitis score. Score 0 was the best, and score 6 was the worst.

| Immunohistochemistry
Deparaffinized sections were digested with proteinase (Dako, Glostrup, Denmark) for 10 minutes and treated with 3% hydrogen peroxide (Wako Pure Chemical Industries, Osaka, Japan) to block endogenous peroxidase activity. We assessed F4/80 expression using a previously reported scoring system for immunohistochemistry analysis of immune and inflammatory cell markers. 25 M1 macrophage is pro-inflammatory and M2 macrophage is anti-inflammatory. 26 CD86 was used as an M1 macrophage marker, CD163 was used as an M2 macrophage marker, and F4/80 was used as a pan macrophage marker. 27  Second, the chondrocytes were then cultured in three settings: under treatment with 10 ng/mL recombinant human IL-1β for the duration that showed the highest p-IKKα/β level in the above procedure; under the same stimulation for the same duration after pretreatment with 30 μg/mL EPA (Sigma-Aldrich) for 30 minutes; and without stimulation. The chondrocytes were subsequently subjected to the same procedure as that described above.
Finally, the chondrocytes were subjected to five different treatments: treatment with 10 ng/mL recombinant human IL-1β for 24 hours; the same treatment for 24 hours after pretreatment with EPA (Sigma-Aldrich) at 10 (low dose), 30 (medium dose), or 50 μg/mL (high dose) for 30 minutes; or without treatment. Next, the chondrocytes were subjected to the same procedure as that described above, except that membranes were incubated with antibody against anti-MMP-13 (catalog no. ab39012; Abcam) instead of antibody against anti-p-IKKα/β. The expression of α-tubulin protein was detected using rabbit anti-α-tubulin polyclonal antibody (catalog no. ab4074; Abcam) as a primary antibody.
Protein expression was determined through semi-quantification of F I G U R E 1 Time profiles of eicosapentaenoic acid (EPA) release in vitro and in vivo (A) time profiles of eicosapentaenoic acid (EPA) release from gelatin hydrogels containing EPA micelles in phosphate-buffered saline containing collagenase. The gelatin hydrogels in this study were designed to biodegrade over a period of~3 weeks under in vivo conditions. B, Time profiles of EPA release from gelatin hydrogels containing EPA micelles in mouse joints. The gelatin hydrogels were also biodegraded over a period of~4 weeks in vivo 2160 | proteins in digitally captured images using ImageJ software (National Institutes of Health, MD, http://imagej.nih.gov/ij/).

| Statistical analysis
Statistical analysis was performed using one-way or two-way analysis of variance with Tukey's post hoc test for multiple comparisons of paired samples. The Mann-Whitney U test was used to compare between two groups. Results are presented as mean values with standard deviation (SD). Differences in mean values were considered significant at P < .05.

| RESULTS
3.1 | Change of EPA concentration by release of EPA from gelatin hydrogels containing EPA micelles Figure 1A shows the time profiles of EPA release from gelatin hydrogels containing EPA micelles in PBS with collagenase. In the presence of collagenase, EPA was released with time from the gelatin hydrogels.
The all gelatin hydrogels in this study were biodegraded over a period of 3 weeks in vivo, in accordance with their design. Figure 1B shows the time profiles of EPA release from gelatin hydrogels containing EPA micelles in mouse joints. As shown in this figure, the gelatin hydrogels were biodegraded over a period of~4 weeks in vivo.   F I G U R E 3 Synovitis evaluations using OARSI-recommended scoring systems Synovitis in hematoxylin and eosin-stained sections was evaluated using OARSI-recommended scoring systems. The average score of each compartment in the coronal slice was obtained from synovitis scores at 1 and 8 weeks after surgery. The average sum scores at 1 week were significantly different between the Sham and DMM groups, Corn and EPA-I groups, and Control and EPA-G groups, and not significantly different between the EPA-I and EPA-G groups. The average sum scores at 8 weeks were significantly different between the Sham and DMM groups, Control and EPA-G groups, and EPA-I and EPA-G groups, and not significantly different between the Corn and EPA-I groups, and Sham and EPA-G groups. Synovitis evaluation revealed a significant decrease of OARSI-recommended scoring systems in OA after EPA treatment, especially in the group treated with EPA-incorporating gelatin hydrogels at 8 weeks after surgery. Control, DMM and gelatin hydrogels; Corn, DMM and injection of corn oil; DMM, destabilization of the medial meniscus; EPA, eicosapentaenoic acid; EPA-G, DMM and gelatin hydrogels containing EPA; EPA-I, DMM and injection of corn oil and EPA; OA, osteoarthritis; OARSI, Osteoarthritis Research Society International [Color figure can be viewed at wileyonlinelibrary.com] 5.7 ± 0.47, 5.2 ± 0.69, 6.5 ± 0.96, and 2.7 ± 0.74, respectively ( Figure 4D), with significant difference between the Sham and DMM groups (P = .004), Control and EPA-G groups (P = .004), and EPA-I and EPA-G groups (P = .004), and no significant difference between the Corn and EPA-I groups (P = .262) (Figures 4F and 4H).

| EPA prevented macrophage infiltration in synovial tissues
There was significantly different between the Sham and DMM groups (P = .004), Corn and EPA-I groups (P = .004), and Control and EPA-G groups (P = .004), and not significantly different between the EPA-I and EPA-G groups (P = .749) in the ratio of CD86-positive cells at 1 week after surgery ( Figure 5A). In addition, the ratio of F I G U R E 4 Immunohistochemical analysis using F4/80 score EPA was disposed to prevent macrophage infiltration in synovial tissues. Immunohistochemical analysis showed the average-sum F4/80 score of synovitis at 1 and 8 weeks after surgery. The average sum scores at 1 week were significantly different between the Sham and DMM groups, Corn and EPA-I groups, and Control and EPA-G groups, and not significantly different between the EPA-I and EPA-G groups. The average sum scores at 8 weeks were significantly different between the Sham and DMM groups, Control and EPA-G groups, and EPA-I and EPA-G groups, and not significantly different between the Corn and EPA-I groups. Immunohistochemical evaluation of F4/80 revealed a significant decrease of F4/80 score in OA after EPA treatment, especially in the group treated with EPA-incorporating gelatin hydrogels at 8 weeks after surgery. Control, DMM and gelatin hydrogels; Corn, DMM and injection of corn oil; DMM, destabilization of the medial meniscus; EPA, eicosapentaenoic acid; EPA-G, DMM and gelatin hydrogels containing EPA; EPA-I, DMM and injection of corn oil and EPA; OA, osteoarthritis [Color figure can be viewed at wileyonlinelibrary.com] groups, and not significantly different between the EPA-I and EPA-G groups ( Figure S1 and S2). The ratios of IL-1β-, p-IKKα/β-, MMP-3-, and MMP-13-positive cells at 8 weeks after surgery were significantly different between the Sham and DMM groups, Control and EPA-G groups, and EPA-I and EPA-G groups, and not significantly different in the Corn and EPA-I groups (Figure 7; Figure S2).
Primary antibody-negative controls (without primary antibody) used for immunohistochemical analysis are shown in Figure S4. In this study, no adverse effects caused by the gel placement itself were observed.

| Enhanced expression of inflammatory transcription factors in response to IL-1β stimulation in vitro
To investigate the role of EPA in signal transduction, p-IKKα/β in NHAC-kn treated with 10 ng/mL IL-1β was assessed. The results confirmed time-dependent IL-1β-induced p-IKKα/β. Particularly, p-IKKα/β level markedly increased at 15 minutes after treatment ( Figure 8A). Therefore, NHAC-kn was treated with IL-1β for 15 minutes in the subsequent experiment. Western blotting results showed that the level of p-IKKα/β, which is a kinase necessary for the activation of nuclear factor NF-κB, was significantly higher following treatment without EPA than following treatment with EPA ( Figure 8B). Moreover, MMP-13 level following treatment without EPA was significantly higher than that following treatment with EPA.
Furthermore, the effect was dose-dependent; the higher the dose was, the more remarkable the effect was ( Figure 8C). These results indicated that EPA suppressed IL-1β-induced p-IKKα/β, and thus prevented the activation of the NF-κB signaling pathway.

| DISCUSSION
The most important finding in this study was that the in vivo effect of degradable gelatin hydrogels containing EPA was more potent compared with a single EPA injection, which supported our study F I G U R E 5 Immunohistochemical analysis using the average-sum ratio of CD86-positive cells Immunohistochemical analysis showed the average-sum ratio of CD86-positive cells at 1 and 8 weeks after surgery. The ratio of CD86-positive cells at 1 week in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 3.3 ± 2.7, 23.3 ± 5.4, 19.9 ± 2.3, 6.6 ± 5.7, 21.1 ± 4.2, and 6.8 ± 3.9, respectively, with significant difference between the Sham and DMM groups, Corn and EPA-I groups, and Control and EPA-G groups, and no significant difference between the EPA-I and EPA-G groups. The ratio of CD86-positive cells at 8 weeks in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 6.2 ± 1.7, 13.2 ± 1.2, 12.1 ± 4.2, 10.6 ± 1.3, 12.9 ± 3.6, and 7.8 ± 2.9, respectively, with significant difference between the Sham and DMM groups, and Control and EPA-G groups, and no significant difference between the Corn and EPA-I groups, EPA-I and EPA-G groups, and Sham and EPA-G groups. Immunohistochemical evaluation of CD86 revealed a significant decrease of the average-sum ratio of CD86-positive cells in OA after However, the inflammation inhibitory effect of single EPA injection was diminished at 8 weeks after surgery. In contrast, the effect of EPAincorporating hydrogels was prolonged for up to 8 weeks after surgery.
Although we previously reported that EPA treatment prevents OA progression in vivo, 11 weekly injection of EPA into the knee joint was required. This weekly injection was thought to be highly invasive in clinical applications; thus, we examined the effects of single intervention. The MMP-13 level of chondrocytes following treatment with IL-1β in the presence of EPA depended on the EPA dose in this study.
Furthermore, several papers have reported on the possibility that increasing EPA intake may have systemic beneficial effects. 32,33 Therefore, the effect of EPA was considered to be dose-dependent in both local and systemic environments, and it is necessary to carefully examine the optimal amount for clinical application.
Matsuzaki et al 18 reported that gelatin hydrogels incorporating rapamycin gradually release rapamycin within approximately 2 days in vitro, but they release rapamycin for approximately 10 weeks in vivo. 18 Furthermore, gelatin hydrogels incorporating simvastatin have been reported to exhibit sustained release for at F I G U R E 6 Immunohistochemical analysis using the average-sum ratio of CD163-positive cells Immunohistochemical analysis showed the average-sum score ratio of CD163-positive cells at 1 and 8 weeks after surgery. The ratio of CD163-positive cells at 1 week in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 2.8 ± 2.6, 11.9 ± 3.7, 9.6 ± 3.3, 8.0 ± 4.4, 9.3 ± 2.0, and 8.1 ± 1.4, respectively, with significant difference between the Sham and DMM groups, and no significant difference between the Corn and EPA-I groups, Control and EPA-G groups, and EPA-I and EPA-G groups. The ratio of CD163-positive cells at 8 weeks in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 4.7 ± 1.3, 13.3 ± 3.1, 12.5 ± 3.0, 9.7 ± 2.2, 11.7 ± 3.4, and 8.3 ± 2.0, respectively, with significant difference between the Sham and DMM groups, and no significant difference between the Corn and EPA-I groups, Control and EPA-G groups, and EPA-I and EPA-G groups. Immunohistochemical evaluation of CD163 revealed no significant decrease of the average-sum ratio of CD163-positive cells in OA after EPA treatment. Results are presented as mean values with standard deviation. Control, DMM and gelatin hydrogels; Corn, DMM and injection of corn oil; DMM, destabilization of the medial meniscus; EPA, eicosapentaenoic acid; EPA-G, DMM and gelatin hydrogels containing EPA; EPA-I, DMM and injection of corn oil and EPA; OA, osteoarthritis [Color figure can be viewed at wileyonlinelibrary.com] least 3 weeks in vitro and 4 weeks in vivo. 19,34 The present study used EPA-incorporating hydrogel formulation solubilized in water into gelatin micelles. In vitro sustained release test showed that EPA was released after the decomposition of the gelatin hydrogels.
This indicated that the decomposition of gelatin was a rate-limiting step for the release of EPA from hydrogels, allowing EPA to be released slowly over a long period in vivo. Consequently, the EPAincorporating gelatin hydrogels prevented OA progression in vivo more effectively than did a single injection of EPA did.
There are two opposite responses elicited by activated macrophages. 35 Macrophages are known to induce pro-and anti-inflammatory responses that mediate matrix destruction or deposition. 36 The pro-inflammatory M1 macrophages produce cytokines, such as tumor necrosis factors, IL-1β and IL-6; chemokines such as monocyte chemoattractant protein-1; and growth factors, such as vascular endothelial growth factor. 37 They also enhance oxidative stress promoters, such as reactive oxygen species and reactive nitrogen species, which in turn activate the NF-κB signaling pathway, causing destruction of cartilage matrix. 38 In contrast, the anti-inflammatory M2 macrophages secrete high levels of anti-inflammatory mediators, such as IL-10, which are necessary to regulate inflammation. 39 F I G U R E 7 Immunohistochemical analysis using the average-sum ratio of IL-1β-, p-IKKα/β-, and MMP-13-positive cells Immunohistochemical analysis showed the average-sum ratio of IL-1β-, p-IKKα/β-, and MMP-13-positive cells at 8 weeks after surgery. The ratio of IL-1β-positive cells at 8 weeks in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 15.6 ± 0.66, 35.6 ± 3.6, 33.8 ± 6.8, 27.8 ± 3.0, 34.1 ± 3.3, and 20.7 ± 2.1, respectively. The ratio of p-IKKα/β-positive cells at 8 weeks in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 9.4 ± 2.3, 30.6 ± 4.4, 27.1 ± 1.8, 25.3 ± 4.9, 27.6 ± 2.0, and 15.9 ± 1.5, respectively. The ratio of MMP-13-positive cells at 8 weeks in the Sham, DMM, Corn, EPA-I, Control, and EPA-G groups were 18.0 ± 1.5, 40.1 ± 2.3, 32.0 ± 2.6, 25.9 ± 2.7, 38.3 ± 2.8, and 18.7 ± 1.9, respectively. The ratios of IL-1β-, p-IKKα/β-, and MMP-13-positive cells at 8 weeks were significantly different in the Sham and DMM groups, Control and EPA-G groups, and EPA-I and EPA-G groups, not significantly different in the Corn and EPA-I groups. Immunohistochemical evaluation of IL-1β, p-IKKα/β and MMP-13 revealed a significant decrease of the average-sum ratio of IL-1β-, p-IKKα There were some limitations in this study. First, in this study, surgical administration of gelatin hydrogels into mouse knee joint was performed. Therefore, future studies should use intra-articular injection of gelatin hydrogels after altering the hardness of the hydrogels. Second, the differences in effects between weekly injection of EPA and one-time placement of EPA-incorporating gelatin hydrogels were not investigated. Finally, the systemic effects of EPA-incorporating gelatin hydrogels placed into the knee joint were not investigated in this study.

| CONCLUSIONS
We investigated the effect of gelatin hydrogels containing EPA on OA progression in vivo. The EPA-containing gelatin hydrogels could prevent OA progression more effectively than a single injection of EPA in vivo in a DMM mouse model. Our results suggested that intraarticular administration of controlled-release EPA can be a new therapeutic approach for treating patients with OA.

ACKNOWLEDGMENTS
We thank Ms Kyoko Tanaka, Ms Minako Nagata, and Ms Maya Yasuda for their technical assistance.
F I G U R E 8 Enhanced expression of inflammatory transcription factors in response to IL-1β stimulation in vitro (A) phosphorylation of p-IKKα/β in NHAC-kn treated with 10 ng/mL IL-1β was assessed. The results confirmed the time-dependent change in the phosphorylation of p-IKKα/β following IL-1β stimulation, and that the phosphorylation levels of p-IKKα/β were markedly elevated at 15 minutes after treatment. X-axis showed that the time of incubation time with IL-1β stimulation. B, The IL-1β-induced expression of p-IKKα/β after incubation without EPA was significantly higher compared with that after incubation with EPA. X-axis showed p-IKKα/β expressions among three groups; without IL-1β stimulation and EPA incubation, with IL-1β stimulation and without EPA incubation, and with IL-1β stimulation and EPA incubation. C, The IL-1β-induced expression of MMP-13 after incubation without EPA was significantly higher than that after incubation with EPA. Furthermore, the effect was dose-dependent; the higher the dose, the more remarkable the effect. X-axis showed MMP-13 expressions among five groups; without IL-1β stimulation and EPA incubation, with IL-1β stimulation and without EPA incubation, with IL-1β stimulation and low-dose EPA incubation, with IL-1β stimulation and medium-dose EPA incubation, and with IL-1β stimulation and high-dose EPA incubation. EPA, eicosapentaenoic acid EPA(+), 10 μg/mL (low-dose); EPA ( + + ), 30 μg/mL (medium-dose); EPA ( + + + ), 50 μg/mL (high-dose); IL-1β, interleukin-1β; MMP-13; matrix metallopeptidase-13 TSUBOSAKA ET AL. | 2167