2.1 Characterizations of MRC-PPL
PPL copolymers self-assembled to nano-micelles (120.3 ± 22.81 nm), as revealed by transmission electron microscopy (TEM, Figure 2a). Similarly, the hydrodynamic diameter reported by dynamic light scattering (DLS) is 121.5 ± 26.1 nm (Figure 2b). And loading micelles with PSO molecules using thin-film hydration method didn’t obviously alter the micelle size (Figure 2a). Zeta-potential measurements (Figure 2c) indicate that micelles consisting of only PPL are negatively charged (-9.25 mV), while MRC-PPL micelles are less negative (-5.38 mV). PSO is highly negatively charged (-8.75 mV) which hinders its cellular uptake since cell membrane is also negatively charged[21, 26]. Desirably, the negative charge of MRC-PPL@PSO micelles is significantly lesser (-5.75 mV). The successful construction of MRC-PPL was further confirmed by the blue shift induced by incorporation of BHQ-3 and Cy5.5 in MRC-PPL in UV-vis spectra (Figure 2d).
Because of the energy resonance transfer to BHQ-3, the fluorescence intensity from Cy5.5 in MRC-PPL is much weaker than the equal amount of free Cy5.5 molecules (Figure 2e). As demonstrated in Figure 2f, introduction of MMP-13 enzyme to MRC-PPL dispersion led to gradual increases of fluorescence (becoming ~3 times brighter after 30 min), whereas the presence of MMP-13 inhibitor prevented this phenomenon. This observation confirms the MMP-13 specific (or OA-specific) fluorescence response, resulting from the cleavage of GPLGVRGC peptide in the probe and subsequent escape of Cy5.5 from quenching by BHQ-3.
2.2 Drug release studies
The strong hydrophobicity of PSO has significantly prevented its application in OA. But through hydrophobic interaction it can be readily loaded into the core of MRC-PPL micelle via self-assembly. The loading percentage of PSO in the micelles is as high as 16.9%. The release of PSO from MRC-PPL@PSO at pH=7.4 and pH=6.5 was determined by high performance liquid chromatography (HPLC) in vitro. As shown in Figure 2g, release of PSO was significantly accelerated by low pH (6.9% at pH 7.4 vs. 21.1% at pH 6.5 within 48 h). Evidently, MRC-PPL@PSO enables effective and sustained drug release under the acidic OA condition.
Figure 2. Characterization of MRC-PPL nano-micelles. (a) TEM images of MRC-PPL and MRC-PPL@PSO micelles. Scale bare = 500 nm. (b) Size distribution of MRC-PPL micelles based on dynamic light scattering. (c) Zeta potentials of PSO, PPL, MRC-PPL micelles and MRC-PPL@PSO micelles. (d) UV-Vis absorbance. (e) Fluorescence intensity of MRC-PPL micelles and Cy5.5. (f) Fluorescence intensity of MRC-PPL micelles, without or with MMP-13 (0.01 μM), in the absence or presence of MMP-13 inhibitor (0.45 μM). (g) In vitro release of PSO from MRC-PPL micelles in PBS (pH 6.5 and 7.4) with 0.1% Tween 80 (mean ± SD, n = 3).
2.3 In vitro cytotoxicity assessment
The cytotoxicity of micelles to chondrocytes isolated from C57BL6/J mice was investigated in vitro using CCK8 assay. Even at a high concentration (200 μg/mL) of MRC-PPL, the cell viability is >85% after 48 hours of incubation, indicating its good biocompatibility (Figure 3a). Then, the possible cytotoxicity of PSO-loaded MRC-PPL micelles was determined as the function of PSO concentration. Figure 3b shows that MRC-PPL@PSO micelles carrying 15 μM PSO didn’t exert any cytotoxicity. Therefore, this dosage was chosen for the following experiments.
High level of interleukin 1 beta (IL-1β) is a marker for OA and is commonly used to induce inflammation of chondrocytes[27-29]. As showed in Figure 3c, IL-1β stimulation (10 ng/mL) for 48 h caused 40% decrease of cell viability while MRC-PPL@PSO treatment largely rescued IL-1β induced cell death. We then used quantitative real-time PCR (qRT-PCR) to detect the gene expression levels of TNF-α, MMP-3 and MMP-13 with promoting cartilage destruction effect in each treatment group, as well as the chondroid-specific marker, Col2a1. As shown in Figure 3d, the expressions of TNF-α, MMP-3 and MMP-13 are elevated by 115.72, 3.33 and 2.59 times respectively after IL-1β stimulation whereas the expression of Col2a1 was decreased to 23.78%. The upregulation of inflammatory factors and down-regulation of Col2a1 were significantly reversed by MRC-PPL@PSO treatment.
As shown in hematoxylin and eosin (HE) staining (Figure 3e), chondrocytes with spindle shape and circular nucleate should be significantly transformed into elongated fibroblast-like cells after treatment of IL-1β for 24 h. MRC-PPL@PSO was able to restore the morphology of most IL-1β pre-treated chondrocytes. MMP-13 and TNF-α that play important roles in OA development were detected by immunofluorescence staining. As shown in Figure 3e, positive staining in immunofluorescence staining (MMP-13 or TNF-α) in IL-1β-stimulated cells showed intense green fluorescence in IL-1β group. However, the fluorescence intensity was obviously decreased after treated by MRC-PPL@PSO.
Taking the observations in Figure 3c-e together, we provide the in vitro evidence that MRC-PPL@PSO can efficiently rectify inflammation and rescue cartilage degradation in OA. PSO and MR-PPL @ PSO also play an anti-inflammatory role, but the effectiveness of PSO is much lower. MR-PPL@PSO micelles performed better than free PSO molecules because the engineered micelles can more effectively carry the hydrophobic drug into cells. MRC-PPL@PSO micelles outperformed MR-PPL@PSO micelles because it is equipped with cartilage targeting peptide.
Figure 3. In vitro study on chondrocytes isolated from C57BL6/J mice. (a and b) Cell viability after treatment with MRC-PPL or MRC-PPL@PSO. (c) Cell viability after various treatments to IL-1β-stimulated chondrocytes. (d) Relative mRNA levels of Col2a1, TNF-α, MMP-3 and MMP-13 on IL-1β-stimulated chondrocytes with various treatments. (e) HE staining and immunofluorescence images. The nuclei were counterstained with DAPI (blue), and MMP-13 or TNF-α positive staining was stained with FITC (green). Scales bar: 400 μm. Each data point represents mean ± s.d (n=3). *, # indicate p < 0.05; **, ## indicate p < 0.01; ***, ### indicate p<0.001.
2.4 Anti-inflammatory Effect of MRC-PPL@PSO on IL-1β-Induced Chondrocytes via the MAPK, NF-κB, and PI3K/Akt Signaling Pathways.
During the progression of osteoarthritis, IL-1β levels are elevated, which activates the nuclear factor NF-κB pathways to induce the expression of matrix metalloproteinases (MMPs) in cultured chondrocytes, leading to ECM degradation, abnormal bone metabolism and inflammatory disease[30, 31]. In addition, activated P38 MAPK has been reported to promote nuclear translocation of NF-κB[32]. Previous studies have found that PSO suppresses the expression of pro-inflammatory cytokines and chemokines as well as the expression of MMPs, which are the key regulators of cartilage destruction[33]. The PI3K/AKT signaling pathway also plays a crucial role in OA[34]. To explore the potential molecular mechanisms underlying the nano-platform treatment, the involvement of the MAPK, NF-κB and PI3K/Akt signaling pathways was investigated by Western blots. In vitro study showed that IL-1β-induced increase of phospho-P38/P38 (p-P38/P38), phospho-Akt/Akt (p-Akt/Akt) and NF-κB expression was diminished by each treatment group. (Fig. 4a). Among all the treatment groups, MRC-PPL@PSO exhibited the greatest reduction of related proteins expression, superior to the PSO and MR-PPL@PSO groups. These results indicate that released PSO inhibits the production of inflammatory mediators in OA possibly mediated by the regulation of the PI3K/AKT pathway or MAPK cascades, leading to NF-κB inactivation.
Figure 4. Molecular mechanism of MRC-PPL@PSO. The expression of PI3K/AKT, MAPK and NF-κB signaling pathway proteins p-P38, P38, p-Akt, Akt and NF-κB was determined by (a) western blot and (c–d) quantification analysis. Each data point represents mean ± s.d (n=3). *, # indicate p < 0.05; **, ## indicate p < 0.01; ***, ### indicate p<0.001.
2.5 Cellular uptake of MRC-PPL
Cellular uptake and MMP-13 responsive behaviors of MRC-PPL micelles were investigated after 24 h incubation with normal chondrocytes and IL-1β stimulated chondrocytes. As shown in Figure 5a, after administration of MRC-PPL, IL-1β stimulated cells exhibit much stronger fluorescence intensity (red) than normal chondrocytes cells, indicating that MRC-PPL micelles are efficient in response to excessive MMP-13 induced by IL-1β to release cy5.5 due to the grafting of MMP-13 peptide substrate in MRC-PPL. Moreover, most red fluorescence emitted from cy5.5 overlaps with the positive dye type II collagen (green fluorescence), suggesting MRC-PPL can successfully target to type II collagen produced by chondrocytes compared with MR-PPL (Figure 5a). The results demonstrated that MRC-PPL is highly responsive to MMP-13 and can effectively target to type II collagen.
Figure 5. In vitro cellular uptake of MRC-PPL or MR-PPL micelles. (a) Immunofluorescence staining in chondrocytes to co-localized with collagen type II in the presence of MMP-13 or its inhibitor. The nuclei were counterstained with DAPI (blue) and collagen type II was stained with FITC (green). Scales bar: 400 μm. (b) Fluorescence quantification of Cy5.5 after uptake of MRC-PPL or MR-PPL micelles by cells. Scale bars: 40 µm. (n = 3; mean ± s.d; *, # indicate p < 0.05, **, ## indicate p < 0.01, ***, ### indicate p<0.001.)
2.6 Living imaging of animals
Two weeks after IA injection of papain solution, OA developed in the knee joint of mice, showing mild swelling and deformity. As demonstrated by in vivo fluorescence imaging (Figure 6a) and quantitative fluorescence intensity profile over time (Figure 6b), MRC-PPL micelles had more accumulation and better retention in the OA joints than MR-PPL micelles, benefitting from the cartilage targeting peptide. In contrast, MRC-PPL micelles poorly accumulated in normal joints or OA joints or OA joints treated with MMP-13 inhibitor, due to insufficient presence of MMP-13 enzymes to activate the theranostic nanoplatforms. At day 21, fluorescence was still obvious in the MRC-PPL treated OA joint, but not in other organs including lung, spleen, heart, liver, kidney (Figure 6c). The results of in vivo imaging are consistent with those of in vitro imaging. Taken together, we provide both in vitro and in vivo evidence that MRC-PPL micelles can specifically target on cartilage and label OA joint.
Figure 6. In vivo fluorescence imaging in normal or OA knees of mice. (a) Fluorescence imaging of OA joints injected with MRC-PPL, MRC-PPL+MMP-13 inhibitor, or MR-PPL at different time post-IA injection, as well as normal joint injected with MRC-PPL (excitation =630 nm, emission = 700 nm). (b) The corresponding fluorescence intensity from OA joints at different times. (c) Ex vivo fluorescence imaging of heart, liver, spleen, lung, kidney, left knee and right knee at day 21 post-injection. (n = 5, mean ± s.d.).
2.7MRC-PPL@PSO nano-micelles attenuate the progression of OA
After the mice receiving different treatments and being sacrificed at week 2 or 6, the femoral condyle and tibial plateau were collected and evaluated according to the criteria described by Lydia Wachsmuth et al for the depth of erosion in articular cartilage[35]. Compared with the control (healthy) group, OA features represented by cartilage erosion and osteophyte formation and deterioration were found in PBS group (Figure7a). Osteophyte and surface lesion were significantly reduced at week 2 and 6 after administrating MRC-PPL@PSO, with OARSI score reduction of 78.59% and 89.42% respectively (Figure 7b). Although with much lesser degrees, PSO and MR-PPL@PSO were also able to reduce the scores by 47.38% and 63.19% respectively at week 6.
Figure 7. The macroscopic observation (a) and macroscopic score (b) of cartilage after IA-injection with PBS, PSO, MR-PPL@PSO and MRC-PPL@PSO for 2 and 6 weeks. Each data points represents mean ± s.d. (n=5). *, # indicate p < 0.05; **, ## indicate p < 0.01; ***, ### indicate p < 0.001.
The cartilage tissues were then evaluated by hematoxylin-eosin (H&E) staining and Safranin O-fast green staining. As shown in Figure 8a, the cartilage layer with surface roughness, vertical cracks, erosion, denudation and deformation were observed in the PBS group, which were consistent with the characteristics of OA. Compared with the PBS group, all three treated groups showed different degrees of improvement in morphological change, matrix staining and tidemark integrity. Noteworthy, MRC-PPL@PSO can reduce the erosion and deformation of the cartilage layer surface, as well as the proliferation of tissue cells. Therefore, it is an effective method to maintain the columnar structure of cartilage. Furthermore, MRC-PPL@PSO group showed more intensive Safranin O staining (red) than other groups, indicating more secretion glycosaminoglycan (Figure 8a). This phenomenon indicates that MRC-PPL@PSO facilitates deposition of glycosaminoglycan and attenuates cartilage matrix depletion and cartilage thinning. As presented in Figure 8b, the OARSI scores based on histological analysis in all the treatment groups decreased to some extent compared with the PBS group, and the MRC-PPL@PSO group showed the lowest score with about 42.3% and 64.7% reduction at week 2 and 6, respectively. The protein expression level of MMP-13 was also assessed by immunohistochemistry staining in the cartilage at week 2 or 6. The MMP-13 positive staining in chondrocytes was observed to be dark brown in PBS group (Figure 8c). On the contrary, the expression level of MMP-13 decreased after different treatments, following the order of PSO > MR-PPL@PSO > MRC-PPL@PSO. Noteworthy, the expression of MMP-13 in the MRC-PPL@PSO treated joints was nearly identical to the healthy control.
Figure 8. Histological analyses of different treatments for 2 or 6 weeks. (a) H&E (upper) and safranin-O/fast green staining (lower) of cartilage sections after the treatments with PBS, PSO, MR-PPL@PSO and MRC-PPL@PSO. Scale bar = 1mm. (b) OARSI scores of articular cartilage after the treatments. Each data points represents mean ± s.d. (n=5). *, # indicate p < 0.05; **, ## indicate p < 0.01; ***, ### indicate p < 0.001. (c) Immunohistochemical staining of MMP-13 on cartilage sections after the treatments. Scale bar = 300 μm.