Cartilage ECM contains sGAGs that are critical for the shock-absorbing functions of cartilage. In this study, we directly compared the efficacy of different sGAG molecules for supporting MSC chondrogenesis using methacrylated CS and HS. Our results showed that CS is a much more potent biochemical cue than HS in enhancing 3D MSC chondrogenesis, as shown by enhanced total sGAG and collagen production (Figs. 2 and 4). Generally, soft hydrogels are more desirable to facilitate neocartilage deposition, whereas increasing hydrogel stiffness to 36 kPa inhibited MSC proliferation and restricted neocartilage deposition to only pericellular regions (Figs. 2 and 4). Soft hydrogels (~ 7.5 kPa) containing an intermediate dose of CS (5% (w/v)) were found to be the optimal hydrogel formulation within the tested range for supporting MSC-based cartilage regeneration.

Previous literature has highlighted the important role of HS in cartilage development during embryogenesis, mostly indirectly via serving as a binding reservoir for soluble factors [18, 30, 31]. Soluble HS has also been shown to be able to enhance chondrogenesis in the presence of TGF-β [32]. However, how HS directly modulates stem cell chondrogenesis in comparison to other ECM molecules such as CS remains largely elusive. Furthermore, the effects of varying HS and CS dosage on 3D MSC chondrogenesis have not been well characterized. Our study addresses these unanswered questions by comparing the efficacy of CS and HS side-by-side for supporting MSC chondrogenesis across a wide range of dosages up to 10% (w/v). Importantly, the stiffness of hydrogels containing varying doses of ECM molecules was kept constant using bioinert polymer PEG. While both HS and CS supported MSC chondrogenesis in a dose-dependent manner, CS was more potent in maximizing cartilage deposition than HS (Figs. 2 and 4). Increasing CS concentration to an intermediate dosage (5% (w/v)) enhanced MSC chondrogenesis in soft hydrogels (~ 7.5 kPa), and further increases led to a decrease in neocartilage formation (Fig. 2b and c), suggesting that an intermediate dosage of CS was optimal (Figs. 2 and 4). Furthermore, soft hydrogels containing 5% (w/v) CS were the leading group among all hydrogel formulations tested, as shown by significantly upregulated cartilage gene expressions of type II collagen and aggrecan (Additional file 5: Figure S1A and B), and downregulations of the hypertrophy marker type X collagen (Additional file 5: Figure S1D). In addition, biochemical data showed highest cell proliferation and neocartilage deposition in this hydrogel group as compared with other hydrogel compositions (Fig. 2). One possible explanation for the observed inhibitory effects of CS at higher doses is the increased negative charge associated with CS molecules, which may interfere with the bioactivity of positively charged growth factors including TGF-β through charge repulsion [33, 34]. Previous studies report that that MMP13 upregulation may result in the breakdown of type II collagen [35]. The high MMP13 gene expression observed in HS-containing hydrogels may explain the decreased type II collagen deposition after 21 days (Fig. 4, and Additional file 5: Figure S1E).

A difference that comes with varying sGAG types is the charge density; HS monomer contains three sulfate groups while CS monomer contains only one sulfate group. As such, at a comparable concentration, HS groups would have higher charge density than CS groups. Therefore, if the observed differences are due to charge density, then CS with a higher dosage should perform comparably to the HS group with similar charge density. However, our data show that this is not the case. For example, while the 2% (w/v) HS group has theoretically comparable charge density with the 5% (w/v) CS group, soft hydrogels containing 5% CS led to a much higher cell proliferation and cartilage matrix deposition (Fig. 2). These results indicate that charge density alone is not the main contributor to the observed differential cellular responses.

It is worth noting that the GAG doses in this study refer to the amount of GAG initially used for hydrogel formation, and the actual GAG content in hydrogels may differ depending on the incorporation efficiency of GAG in PEG hydrogels. In a recent study, we have performed diffusion assays to measure the amount of sGAGs that leached out from hydrogels containing GAG modified with different degrees of methacrylation over time [12]. Our data suggest that unbound sGAGs were washed out within the first few hours, and the majority of sGAGs remain stably incorporated inside hydrogels over time. [12]. When cells are incorporated in the GAG-containing hydrogels, cell-secreted enzymes can potentially accelerate the degradation and release of initially incorporated GAG while depositing new cartilage matrix.

In addition to biochemical cues, mechanical cues such as matrix stiffness have also been shown to play an important role in regulating stem cell fate [28, 36]. A recent report has shown that MSC chondrogenesis was promoted on soft substrates when cultured in two dimensions [37]. However, how matrix stiffness modulates 3D MSC chondrogenesis remains elusive. In our study, we compared ECM containing hydrogels with two stiffnesses (~ 7.5kPa and ~ 36 kPa), representing soft and stiff microenvironments, respectively. For CS-containing hydrogels, increasing hydrogel stiffness resulted in a substantial decrease in neocartilage deposition as shown by biochemical assays and histology (Figs. 2, 4, and 6), while neocartilage formation was restricted largely to pericellular regions. All cell-laden stiff hydrogels showed a loss of mechanical properties due to degradation and disconnected neocartilage nodules (Fig. 3b). In contrast, while soft hydrogels have an initial lower mechanical modulus, cell-laden soft hydrogels containing 5% (w/v) CS exhibited a ~ 109% increase in Young’s Modulus after 21 days of culture (Fig. 3a). Consistent with this observation, more interconnected and homogenous neocartilage depositions were observed in some of the soft hydrogels (Figs. 4, 5 and 6), resulting in large increases in the mechanical properties of engineered cartilage over time (Fig. 3a). In contrast, the acellular hydrogel of the same composition underwent substantial degradation and resulted in almost a complete loss of mechanical moduli by day 21 (Additional file 6: Figure S2A). Taken together, our data suggest that soft hydrogels provide a more permissive environment for supporting MSC-based neocartilage formation, likely due to the less physical restriction with lower crosslinking densities [38, 39]. Moreover, these results confirm that the increase in the mechanical property of engineered cartilage was contributed to by the neocartilage deposited by the cells [40]. Since an important criterion for selecting a scaffold to enhance stem cell-based cartilage regeneration is enabling of new matrix deposition with increased mechanical properties of engineered cartilage tissues over time [8, 41, 42], choosing hydrogels with lower initial matrix stiffness would be beneficial.

The presence of CS in the scaffold is critical for enabling the observed improvement in cartilage function, as soft hydrogels without CS did not show any increase in mechanical stiffness (Fig. 3a) compared with day 1 (Fig. 1), and deposited neocartilage was restricted to pericellular regions only (Figs. 4, 5 and 6). Our observation is in line with previous reports that demonstrated that scaffolds that facilitate homogenously distributed and interconnected neocartilage are critical for improving the mechanical properties of the engineered cartilage over time [43]. Although this study focuses on the effects of hydrogel stiffness and concentration of sGAG, one confounding factor is that hydrogel degradation is also a variable since varying PEG concentration was used to keep the hydrogel stiffness constant. The fact that acellular hydrogel of the leading group (soft hydrogels containing 5% CS) also exhibits fast degradation suggests that the enhanced cartilage formation may be a collective result of low stiffness and fast degradation.

While our leading group (soft hydrogels containing 5% CS) supported extensive type II collagen deposition, a desirable matrix for articular cartilage (Fig. 4), immunostaining also showed high levels of type I collagen (Fig. 5). Our observation is in line with previous reports that MSC-based cartilage regeneration is often associated with high level of type I collagen [9, 44, 45]. Type I collagen is a fibrocartilage marker, which is undesirable for articular cartilage. To reduce the undesirable type I fibrocartilage phenotype, future studies may employ gene silencing approaches such as using shRNA to minimize type I collagen deposition [46].