Cell viability and proliferation inside the hydrogel
To assess the suitability of the CH-GP-HEC to support cell survival, hMSCs (Figure 1) and rMSCs (Figure 2) were encapsulated in CH-GP-HEC hydrogels and viability monitored by propidium iodide-fluorescein diacetate (PI-FDA) staining of cell-hydrogel constructs over indicated time-points. Cell-free hydrogels were used as controls (Figure 2). Cells encapsulated inside CH-GP-HEC hydrogels were viable and maintained a round morphology. Consistently, Kim et al. (2011) showed that seeded cells have a spindle-like shape, but encapsulated cells have a round morphology inside chitosan gels (Kim et al., 2011). CH-GP-HEC hydrogels can maintain a good overall viability of MSCs for 28 days of culture, although the encapsulation process may still result in cell death at early time-points (day 0; immediately after encapsulation).
Figure 1. Live/dead fluorescent staining of encapsulated hMSCs with concentration of 2 × 105 cells in 500 µL CH-GP-HEC hydrogels after 0, 1, 7, 14 and 28 days in culture. Using PI-FDA staining live cells are indicated by green colour, while dead cells can be observed in red.
Download figure to PowerPoint
Figure 2. PI-FDA staining of rMSCs seeded at 2 × 105 cells in 500 µL scaffold per well in 12-well plates. Directly after seeding, some dead cells (red) can be observed. This may result from the encapsulation process of cells into the scaffold and the long experimental time. At day 7, the majority of cells were alive within the scaffold (green).
Download figure to PowerPoint
In the next step, proliferation of encapsulated human and rat MSCs in chitosan hydrogel was monitored in expansion media at different time-points over 28 days by MTS assay (Figure 3A and B). Both cells showed significantly higher metabolic activity at day 7–28 compared to day 0 (P < 0.001), indicating that the cells could grow and proliferate within CH-GP-HEC hydrogels. However, changes in cell number were inconsistent for rMSCs (Figure 3B) compared to hMSCs (Figure 3A). A possible explanation for this is that MSCs from different species and donors were heterogeneous and had variable proliferation rates with time. Other possibilities may be inhomogeneous distribution of cells or non-optimal initial cell density so that after sometime cells face surface area limitation and undergo differentiation or cell death. However, fold change of cell number at day 28 compared to day 0 was significantly increased (P < 0.001) for all donors from both specious by ∼5-, ∼7.5- and ∼5-fold for hMSCs (donors 1, 2 and 3, respectively) and by ∼4- and ∼13-fold for rMSCs (donors 1 and 2, respectively). These results are consistent with results from Ahmadi and de Bruijn (2008) and demonstrate that chitosan hydrogels are not only non-toxic, but can stimulate MSC proliferation, representing a suitable biomaterial for cell encapsulation.
Figure 3. Number of encapsulated hMSCs (A) and rMSCs (B) in 500 µL of CH-GP-HEC hydrogels after 0, 7, 14 and 28 days in expansion media. Proliferation of cells was measured by MTS assay. Blank hydrogel was used as control for background subtraction. Results are represented as means ± SD (n = 3). *** denotes significant difference at all time points as compared to day 0 of culture (P < 0.001).
Download figure to PowerPoint
Chitosan hydrogels with high salt (GP) concentration produce ionic strengths unsuitable for cells and inhibit cell growth (Ahmadi and de Bruijn, 2008; Cho et al., 2008). In order to reduce the cytotoxicity of chitosan-based injectable hydrogels and improve cell viability, some authors have used less toxic cross-linkers, such as genipin, glyoxal and HEC, responsible for the solidification with low concentrations of GP (Li and Xu, 2002; Hoemann et al., 2007; Moura et al., 2011; Wang and Stegemann, 2011).
Insulin release kinetics
The in vitro release of insulin from the CH-GP-HEC hydrogels was determined over 8 days, the cumulative release being shown in Figure 4. The CH-GP-HEC hydrogels containing a higher concentration (1,000 µg/500 µL scaffold) of insulin released significantly greater amounts compared with hydrogels containing a lower concentration (100 µg/500 µL scaffold) (Figure 4A) within the examination time. However, the percentage cumulative insulin release from the differently loaded gels showed no obvious difference (Figure 4B). For example, the CH-GP-HEC hydrogels containing a higher concentration of insulin (1,000 µg/500 µL scaffold) exhibited average cumulative release up to 60.7 ± 1.6% at day 2 similar to average cumulative release of lower concentration (100 µg/scaffold) up to 61.9 ± 2.1% at day 2. The results show that the initial loading concentration does not play an important role in the release profile of insulin from the CH-GP-HEC hydrogels; both concentrations had similar percentage release profiles (Figure 4B) and complete release was achieved after 8 days. Thus insulin seems to have a sustained release from the CH-GP-HEC hydrogels over this period of time and is independent of the initial loading concentrations.
Figure 4. Cumulative in vitro insulin release from CH-GP-HEC hydrogel. (A) Cumulative released amount as micrograms of two different insulin loading concentrations per 500 µL hydrogel during indicated times. (B) Cumulative release percent. Results are presented as means ± SD of five replicates. No significant differences were observed between the release percentages of two different concentrations.
Download figure to PowerPoint
Chitosan-based delivery systems have gained popularity because of their ability to release drugs at a local target site in a controlled manner (Ruel-Gariepy et al., 2000; Bae et al., 2006; Yu and Ding, 2008; Bhattarai et al., 2010; Kim et al., 2010). Generally, drug release from biodegradable polymeric systems is governed by initial drug loading, diffusion, drug and polymer interactions, drug solubility and polymer degradation (Gunatillake and Adhikari, 2003; Bhattarai et al., 2010). In this study, the incorporation of insulin into CH-GP-HEC hydrogels showed a sustained release of insulin over 8 days. This is in contrast with the findings of other studies (Ruel-Gariepy et al., 2000; Wu et al., 2006) as it demonstrated that insulin encapsulation into a CH-GP-HEC formulation allows a prolonged release of drugs. Interestingly, percentage insulin release was not affected by initial drug loading. Conversely, Ruel-Gariepy et al. (2004) found that the initial paclitaxel loading substantially affected the release percentage from chitosan-GP hydrogel, 92% cumulative release for the 6.4 mg/mL-loaded gel rather than a 43% cumulative release for the 64 mg/mL-loaded gel after one month (Ruel-Gariepy et al., 2004).
Due to the high water content of hydrogels, their release mechanisms are different from other drug delivery systems comprised of less hydrophilic or hydrophobic polymers. Some studies have predicted that the release of an active agent from a hydrogel, as a function of time, is controlled by diffusion, swelling or chemical reactions (Amsden, 1998; Bhattarai et al., 2010; Lamberti et al., 2011). Considering the nature of hydrogels and their high water content, the mechanism behind the insulin release seems to be concentration gradient-driven diffusion. In this manner, the insulin was incorporated into the aqueous environment of the CH-GP-HEC hydrogels and its in vitro release was governed by its diffusion from the gel matrix. However, other studies suggest that the delivery of drugs and growth factors from injectable chitosan hydrogels is not only initially by diffusion, but later by degradation of the hydrogel (Ruel-Gariepy et al., 2000; Nair et al., 2007; Bhattarai et al., 2010). However, we did not observe obvious degradation of CH-GP-HEC hydrogels in vitro after 8 days (data not shown). In addition, the initial release rate can be controlled by molecular weight, degree of deacetylation of chitosan and concentration of GP (Ruel-Gariepy et al., 2000; Bae et al., 2006; Wu et al., 2006; Nair et al., 2007). Li and Xu (2002) investigated in vitro release profiles of CH-GP in combination with three different concentrations of HEC, showing that the release mechanism of drugs from CH-GP-HEC hydrogels is independent of HEC content in the gels and is governed by Fickian diffusion, although HEC is one of the principle components of the network formation (Li and Xu, 2002).
We explored the possibility of the CH-GP-HEC hydrogel for delivery of insulin as a potent factor that prolongs the cell survival and enhances the matrix production of chondrocytes. Indeed, the feasibility of using the CH-GP-HEC gels as a stem cell carrier vehicle as well as an insulin delivery vehicle has been shown. Others found insulin to enhance strongly the redifferentiation of chondrocytes, which inevitably dedifferentiate during cell expansion, to increase the cell viability in the middle part of the constructs and induce the in vivo cartilage regeneration (Ko et al., 2011). In our previous study, we showed that the insulin released from PLGA microspheres effectively promoted cell viability and matrix synthesis (Andreas et al., 2011). Another study suggests the efficacy of insulin in increasing the growth rate and glycosaminoglycan fraction of engineered cartilage (Kellner et al., 2001; Malafaya et al., 2010).
Chondrogenic differentiation of the encapsulated MSCs
The chondrogenic induced cell-CH-GP-HEC hydrogel constructs were examined by histological methods to identify the location of the entrapped cells and the distribution of the retained proteoglycans. The Alcian Blue/Nuclear Fast Red stainings were used to detect the anionic GAG chains of the proteoglycans (Figure 5A). Additionally, Safranin O staining was performed (Figure 5B) to confirm the results. A non-induced construct as a control experiment showed less or minimum staining for proteoglycans/GAGs in chondrogenic control medium (without TGF-β3) at day 28 (Figure 5C). Analysis of the cell-encapsulating hydrogel by haematoxylin staining revealed a relatively uniform distribution of cells throughout the scaffold (Figure 5D). Consistent with our data, others have shown that hydrogels allow uniform encapsulation of cells (Hoemann et al., 2005; Amini and Nair, 2012; Rao et al., 2012).
Figure 5. Histological staining of hydrogels. Staining of proteoglycans/GAGs produced by hMSCs encapsulated in CH-GP-HEC hydrogel with Alcian Blue/Nuclear Fast Red (A) and Safranin O (B) cultured in chondrogenic induction medium on days 14 and 28. Cryosections prepared from constructs cultured in control medium without TGF-β3 at day 28 were used as non-induced control, stained with Alcian Blue/Nuclear Fast Red (C). Haematoxylin staining (D) was performed for observation of cell distribution.
Download figure to PowerPoint
The CH-GP-HEC hydrogel supported GAG production. Histological staining for proteoglycans showed the differences in the matrix deposition over time, more visible at day 14, indicating the time in which MSCs started to differentiate towards chondrocytes. The GAG density in the scaffold increased considerably from day 14 to 28 (Figure 5A and B). The data suggest that MSCs encapsulated in CH-GP-HEC hydrogel expressed elevated levels of proteoglycans compared to the control group, indicating that addition of TGF-β3 shifted the MSCs toward the chondrogenic phenotype. Others have demonstrated enhanced MSC chondrogenesis following delivery of TGF-β3 from alginate microspheres within hyaluronic acid hydrogels in vitro and in vivo (Bian et al., 2011). Even transient exposure (7 days) to a very high level of TGF-β3 (100 ng/mL) improved chondrogenesis of MSC-loaded hyaluronic acid hydrogel (Kim et al., 2012). Hence, it is reasonable to conclude that CH-GP-HEC constructs with a sustained release profile of 8 days may be used as scaffolds to transiently induce chondrogenesis of encapsulated MSCs by incorporating TGF-β3 and/or insulin for cartilage engineering. Administration of insulin into atelocollagen hydrogel enhances the in vivo cartilage regeneration (Ko et al., 2011).
Biochemical analyses of the constructs showed significant accumulation of ECM (assessed as GAG) in thermo-sensitive chitosan hydrogels at days 14 and 28 compared to control and to day 1 (Figure 6). Increased GAG content indicates that encapsulated hMSCs maintained their differentiated phenotype. Irrespective of GAG release into the medium, a statistically significant increase in GAG accumulation in the gels occurred between 14 and 28 days for all three donors. However, there is a significant difference between donor 3 relative to donor 2, but not to donor 1 (Figure 6), indicating MSC-donor dependency of results. Similarly, heterogeneity of MSCs originated from different donors has already been reported (Karp and Leng Teo, 2009). Increase of proteoglycan content has gained particular note, because it has been shown to be important for functional compressive and mechanical properties (Kisiday et al., 2002; Mizuno et al., 2006), although the role of collagen type II is more prominent (Huang et al., 2008). It is important to consider that a low cell seeding density of 1 × 107 cells/mL was used here to specifically examine the effect of the biomaterial on cell behaviour. Typically, cell-encapsulating hydrogel constructs for cartilage tissue engineering are seeded with 20 × 106 to 60 × 106 cells/mL (Huang et al., 2008; Hao et al., 2010; Kim et al., 2012). Thus, our low initial seeding density may have contributed to the low levels of ECM production (Hoemann et al., 2005).
Figure 6. Total GAG content present in the supernatant of constructs after papainase digestion of hydrogels. Normalisation of samples was done by wet weight of the constructs at each indicated time point of culture from 3 different MSC donors (n = 3). Blank hydrogel digest was used as control for background subtraction. CM assigned for chondrogenic control medium. Data are represented as means ± SD (n = 3). # denotes significant difference relative to CM at each time point (P < 0.001). * denotes significant difference of each donor compared with day 1 of culture (P < 0.001). @ denotes significant difference of donor 3 relative to donor 2 at same time point (P < 0.05).
Download figure to PowerPoint
Visual observations showed no obvious changes (shrinkage or swelling) of CH-GP-HEC hydrogels with time (data not shown). Many hydrogels suffer from shrinkage or swelling in solution that finally lead to structural disruptions (Li and Xu, 2002; Bhattarai et al., 2010; Bertolo et al., 2012). Our results were consistent with other reports (Hannouche et al., 2007; Kurdi et al., 2010) indicating that CH-GP-HEC hydrogels are stable in solution and keep their structural integrity for 4 weeks in vitro. After 28 days, cellular controls also remained translucent while the cell-loaded constructs had an opaque appearance.
Consistent with retained GAGs, release of GAGs into medium significantly increased over 28 days (Figure 7). Only ∼26% of the proteoglycans produced by MSCs was retained by the constructs at day 28 and remaining proteoglycans (∼74%) were released into the culture medium. The specific hypothesis of our study was that the cationic chitosan would form an ideal environment in which large quantities of newly synthesised anionic proteoglycan/sulphated-GAG could be entrapped in the hydrogel. In contrast to Roughley et al. (2006), cell-encapsulating CH-GP-HEC hydrogels showed that the majority of proteoglycans produced by encapsulated MSCs were released into the culture medium rather than retained within the hydrogels. A possible explanation could be that the cationic chitosan in our formulation (CH-GP-HEC) may be fully neutralised by GP (Hoemann et al., 2007). Irrespective of the mechanism of retention, which may result from chemical or physical properties of the hydrogel (Roughley et al., 2006), it remains to be established whether a proteoglycan concentration similar to that observed in vivo can ever be obtained. In the present work, proteoglycan contents of the scaffold (Figure 6) never reached that of native tissue. Prolonging the culture period and increasing cell density may improve the proteoglycan content (Kisiday et al., 2002; Hoemann et al., 2005).
Figure 7. Total net GAGs present in culture media from three different MSC donors (each in triplicate) at the end of each time period. Blank hydrogel supernatant was used as control for background subtraction. CM assigned for chondrogenic control medium. Data are represented as means ± SD. # denotes significant difference relative to CM at each time point (P < 0.001). * denotes significant difference of each donor compared with day 1 of culture (P < 0.001). @ denotes significant difference of donors relative to others at same time point (P < 0.05).
Download figure to PowerPoint
Regarding potential applications, autologous MSCs can be recovered from the patient's bone marrow or adipose tissue, expanded, chondrogenically differentiated and combined with the hydrogel to repair cartilage defects (Moreau and Xu, 2009). MSCs provide an ideal cell source for cartilage engineering because they are relatively easy to access, can differentiate into chondrogenic cells, possess little to no immunogenic and tumorigenic ability, and their use is not complicated by ethical and legal controversies (Nejadnik et al., 2010; Tang et al., 2012; Maumus et al., 2013). MSC-laden hydrogels can achieve tensile properties that are comparable to chondrocyte-seeded constructs (Huang et al., 2008), confirming the utility of this alternative cell source in cartilage tissue engineering.
The development of scaffolds is a central topic in MSC-based cartilage engineering. To date, hydrogels and other polymeric systems have been examined for cell delivery and cartilage engineering (Hannouche et al., 2007; Park et al., 2007; Hao et al., 2010; Bian et al., 2011). The application of hydrogels as 3D scaffolds has recently gained attention because they can mimic key features of the ECM, such as their 3D nature and high water content. A wide range of polymers can be used for their fabrication, including (but not limited to), collagen-derived scaffolds (i.e., atelocollagen and collagen sponges), gelatin hydrogels, alginate hydrogels, hyaluronic acid hydrogels and chitosan hydrogels (Hoemann et al., 2005; Roughley et al., 2006; Hannouche et al., 2007; Park et al., 2007; Cho et al., 2008; Tan et al., 2009; Hao et al., 2010; Bian et al., 2011; Ko et al., 2011; Kim et al., 2012). Consistent with others (Lahiji et al., 2000; Roughley et al., 2006; Hao et al., 2010), our results show that CH-GP-HEC hydrogels can support chondrogenic differentiation. Richardson et al. (2008) studied the temperature-sensitive hydrogel CH-GP, seeded with hMSCs and cultured for 4 weeks in standard medium. They demonstrated differentiation of MSCs to a phenotype which showed similarities to both articular chondrocytes and nucleus pulposus cells, suggesting that MSC-seeded CH-GP gels could be used clinically for the regeneration of human intervertebral disc. The development of an injectable, biodegradable hydrogel composite of oligo(poly(ethylene glycol) fumarate) (OPF) with encapsulated rabbit MSCs and gelatin microparticles loaded with transforming growth factor-β1 for cartilage tissue engineering applications has been described (Park et al., 2007). Their results indicate that encapsulated rabbit MSCs remain viable over the culture period and differentiate into chondrocyte-like cells, thus suggesting the potential of OPF composite hydrogels as part of a novel strategy for localised delivery of stem cells and bioactive molecules. Tan et al. (2009) examined the potential of the injectable in situ forming biodegradable chitosan-hyaluronic acid based hydrogel for cartilage tissue engineering. They found that this composite hydrogel supported cell survival and retained chondrogenic morphology. Additional experiment demonstrated enhanced chondrogenic differentiation of murine embryonic stem cells in hydrogels with glucosamine, an amino monosaccharide found in chitin, glycoproteins and GAGs, such as hyaluronic acid, chondroitin sulphate and heparin sulphate (Hwang et al., 2006). Hannouche et al. (2007) demonstrated that in alginate-poly glycolic acid/MSCs constructs, cell growth phase and the chondrogenic differentiation of MSCs occurred during the first 3 weeks. In addition, cells remained round in the hydrogel, thickness of alginate-polyglycolic acid (PGA)/MSCs constructs increased substantially over time, no shrinkage was seen, and composite hydrogel-PGA scaffold supported the in vitro growth of implantable cartilaginous structures.