Mesenchymal stem cell‐inspired microgel scaffolds to control macrophage polarization

Abstract There is a desire in regenerative medicine to create biofunctional materials that can control and direct cell function in a precise manner. One particular stem cell of interest, human mesenchymal stem cells (hMSCs), can function as regulators of the immunogenic response and aid in tissue regeneration and wound repair. Here, a porous hydrogel scaffold assembled from microgel subunits was used to recapitulate part of this immunomodulatory behavior. The scaffolds were used to culture a macrophage cell line, while cytokines were delivered exogenously to polarize the macrophages to either a pro‐inflammatory (M1) or alternatively activated (M2a) phenotypes. Using a cytokine array, interleukin 10 (IL‐10) was identified as one key anti‐inflammatory factor secreted by hMSCs in pro‐inflammatory conditions; it was elevated (125 ± 25 pg/ml) in pro‐inflammatory conditions compared to standard medium (6 ± 10 pg/ml). The ability of hMSC laden scaffolds to reverse the M1 phenotype was then examined, even in the presence of exogenous pro‐inflammatory cytokines. Co‐culture of M1 and M2 macrophages with hMSCs reduced the secretion of TNFα, a pro‐inflammatory cytokine even in the presence of pro‐inflammatory stimulatory factors. Next, IL‐10 was supplemented in the medium or tethered directly to the microgel subunits; both methods limited the secretion of pro‐inflammatory cytokines of encapsulated macrophages even in pro‐inflammatory conditions. Cumulatively, these results reveal the potential of biofunctional microgel‐based scaffolds as acellular therapies to present anti‐inflammatory cytokines and control the immunogenic cascade.


| INTRODUCTION
Hydrogels provide a useful platform for regenerative medicine, capable of recapitulating microenvironmental cues or presenting biochemical stimuli to delivered cells. [1][2][3][4] In particular, the ability of scaffolds to manipulate the regenerative potential of specific stem cell types is desirable in the context of creating materials based therapies for wound healing and tissue regeneration. In this regard, human mesenchymal stem cells (hMSCs) are increasingly used in clinical trials for their reparative and regenerative potential 5,6 due to their ability to differentiate into multiple cell types. 7,8 Their role in mitigating the inflammatory response during wound healing through their secretory properties [9][10][11] is particularly attractive for regenerative therapies. During periods of inflammation, endogenous hMSCs can migrate to wounds and repolarize resident inflammatory cells from inflammatory to tissue remodeling phenotypes. 12,13 Upon exposure to inflammatory cytokines (e.g., Interferon gamma (IFNγ), Interleukin 1 beta (IL-1β), Tumor necrosis factor alpha (TNFα)), which are factors typically secreted by macrophages, hMSCs release a variety of antiinflammatory cytokines (e.g., TNF-inducible gene 6 protein (TSG-6), Transforming growth factor beta (TGFβ), Leukemia inhibitory factor (LIF), Interleukin 1 receptor antagonist (IL-1RA)) 14, 15 and small molecules (e.g., Prostaglandin E2 (PGE-2), Indoleamine 2,3-dioxygenase (IDO), nitric oxide). 14,16 Recently, this effect has been leveraged for cell-based therapies, where hMSCs are activated ex vivo by proinflammatory conditions, a procedure termed "licensing." Licensing is performed with the goal of manipulating hMSC secretory properties to upregulate anti-inflammatory cytokines and ultimately, their regenerative potential. 16,17 A better understanding of the secretory profile of hMSCs and its effect on inflammatory cell function can aid in the design of acellular, functionalized materials capable of mimicking the immunomodulatory properties of hMSCs.
When designing material platforms to promote tissue regeneration, it is vital to consider the inflammatory microenvironment that exists upon injury. 18 The inflammatory cascade is perpetuated by a variety of immune cells (e.g., neutrophils, macrophages, T cells) that arrive over different time scales. 19 hMSCs have been demonstrated to inhibit T-cell proliferation, 20 modulate dendritic cell activation, 21 limit B-cell maturation, 22 and direct macrophage polarization and function. 23,24 Macrophages are an innate immune cell type that help maintain tissue homeostasis with a variety of different polarization states. 25,26 However, during an acute injury response, macrophages become "activated" to an inflammatory (M1) phenotype, where they secrete a variety of pro-inflammatory cytokines (e.g., IFNγ, TNFα, Interleukin 6 (IL-6)) that perpetuate the immune response. 27,28 As a cell-based therapy, hMSCs have the ability to control directly macrophage polarization, where they can reprogram M1 macrophages to a variety of different regenerative polarization states that comprise the broader M2 phenotype. 29,30 This role further aids in the resolution of chronic inflammation and improves wound healing. 31 While the M2 phenotype encompasses several distinct phenotypes, "M2a" macrophages have both been implicated in tissue regeneration and antiinflammatory activity. 32,33 As macrophages infiltrate and interact with implanted biomaterials, 34,35 understanding how biomaterials can be used to modulate macrophage polarization is important for the design of effective biomaterials for in vivo applications.
While recent research has revealed several key molecules involved in the crosstalk between macrophages and hMSCs, the extent that biomaterial design can be used to alter the behavior of local and delivered cell behavior is not fully known. Clearly, biomaterial structure and composition can be modified in unique ways to influence cell behavior in the context of regenerative medicine. For example, the Geissler group demonstrated that porous materials (pore diameter 120 μm) improve hMSC clustering and influence their regenerative effects on myoblast function. 36 Our group expanded on this concept, by embedding hMSCs in microgel scaffolds designed to cluster hMSCs by varying the average pore size of the porous scaffold (e.g., 10 and 200 μm), 37 where larger pore sizes upregulate the secretion of numerous trophic factors. Cell-matrix interactions and microenvironmental structure also impact immune cell function. For example, the Bryers group demonstrated that scaffold porosity can directly impact dendritic cell function, with faster maturation occurring in scaffolds with smaller diameter (20 μm) pores compared to larger ones (90 μm). 38 Collectively, these results point to the need to better understand the complex interplay and effects of porosity and spatial confinement on not cell function and their regenerative properties.
Beyond structural features, bioactive hydrogel scaffolds can be created through the tethering of known biochemical cues to influence cell behavior locally. For instance, Garcia et al. tethered IFNγ to hydrogel scaffolds and altered hMSC cytokine secretion, increasing the release of several immunomodulatory moieties, such as IDO and macrophage colony-stimulating factor (M-CSF). 39 Conditioned medium from hMSCs cultured in IFNγ-functionalized gels can also influence immune cell function, significantly downregulating T-cell proliferation and dendritic cell differentiation. While stimulation (or "priming") of MSCs is an effective way to boost their regenerative properties, the influence of this priming is often short-lived, which can limit effectiveness. The design of acellular biomaterials to mimic this immunomodulatory potential may prove advantageous. In one early investigation, Hume et al. tethered TGF-β1 to hydrogel scaffolds to reduce the maturation of dendritic cells in vitro. 40 The design of similar materials that draw inspiration from the hMSC secretory profile could be highly effective in altering regeneration in vivo.
With these studies in mind, experiments herein focused on the development of porous microgel scaffolds to mimic the immunomodulatory activity of hMSCs through the direct inclusion of bioactive cues.
Assembled microgel scaffolds were first designed for macrophage culture, and the macrophages were polarized to pro-or anti-inflammatory phenotypes by treatment with lipopolysaccharide (LPS) and IFNγ or interleukin 4 (IL-4) and interleukin 13 (IL-13), respectively. hMSC immunomodulatory properties were then assessed in a context relevant to tissue regeneration in vivo, where hMSCs were co-cultured with M1 macrophages and/or placed in pro-inflammatory conditions (i.e., conditioned medium from M1 macrophages) to identify key immunomodulatory factors secreted by hMSCs. These factors were quantified using cytokine arrays and enzyme-linked immunosorbent assays (ELISAs) for specific proteins. The anti-inflammatory cytokine IL-10 was significantly upregulated by hMSCs cultured in proinflammatory conditions. hMSCs in microgel scaffolds were then cocultured with macrophages in separate microgel scaffolds to assess their ability to control macrophage polarization status. Based on these findings, it was investigated whether a microgel scaffold could be designed to "mimic" this MSC immunomodulatory behavior. It was hypothesized that tethering IL-10 directly to the microgel platform would limit macrophage M1 activity, even in the presence of exogenous pro-inflammatory cytokines. The creation of biofunctional materials that can mimic the modulatory behavior of hMSCs would be of exceptional interest to the fields of regenerative medicine.
Finally, the anti-inflammatory protein IL-10 was assessed, as it is indicative of an M2 macrophage phenotype. IL-10 was significantly Human mesenchymal stem cells (hMSCs) secrete the anti-inflammatory protein IL-10 in response to inflammatory conditions. (a) hMSC laden microgel networks were cultured in a proinflammatory environment (conditioned medium [CM] from M1 macrophages) to assess their secretory properties. (b) After 72 h, a cytokine array demonstrated a global decrease of many potent pro-inflammatory factors compared to gels in standard culture conditions. (c) Analysis of several antiinflammatory proteins revealed that IL-10 was highly upregulated compared to in growth medium.

| Effects of IL-10 on macrophage polarization
Given IL-10 was upregulated by hMSCs cultured in inflammatory conditions and that hMSC co-cultures limited M1 macrophage activity, the effect of local presentation of IL-10 on macrophages was subsequently investigated. First, a dose screen was performed using THP1 macrophages cultured on tissue culture polystyrene (TCPS), and 10 ng/ml of IL-10 significantly reduced TNFα secretion ( Figure S2). Based on these data, an azide-modified IL-10 was synthesized via an NHS coupling reaction. The modified IL-10 was tethered directly to the microgel subunits during fabrication at 10 ng/ml and confirmed by immunostaining ( Figure S3). THP1 cells were then encapsulated within the IL-10 modified microgel scaffolds ( Figure 5(a)). Compared to unmodified scaffolds, TNFα secretion from macrophages was significantly decreased in the IL-10-functionalized scaffolds for all Co-culturing human mesenchymal stem cells (hMSCs) with macrophages can limit pro-inflammatory activity of M1 macrophages in multiple environments. (a) Macrophages were encapsulated in microgel networks and polarized to the M1 phenotype as normal over 96 h. Medium was then replaced with one of: standard THP1 medium, M1 induction medium, or M2a induction medium. Additionally, hMSCs were encapsulated in separate microgel networks and added to half of the conditions via a transwell insert. After 72 h, medium was collected and assessed for M1 and M2 specific markers. (b) TNFα secretion was significantly decreased in all M1 conditions upon coculture with hMSC gels, independent of medium conditions. (c) IL-1β secretion was decreased in THP1 and M1 induction media in hMSC-M1 co-culture conditions compared with M1s alone and slightly upregulated in M2a medium. (d) The anti-inflammatory cytokine IL-10 was elevated in M2a treated conditions and was most elevated in hMSC coculture conditions; *p < 0.05, **p < 0.01, ***p < 0.001, n.s. = not significant (p > 0.05) conditions (THP1 medium: from 705 ± 67 to 428 ± 174 pg/ml, proinflammatory: from 874 ± 70 to 590 ± 57 pg/ml, and anti-inflammatory: 581 ± 37 to 219 ± 141 pg/ml) ( Figure 5(b)). Similarly, IL-1β levels were decreased in IL-10 tethered scaffolds (THP1 medium: from  Figure 5(c)).. IL-6 levels were significantly decreased in tethered IL-10 gels compared to unmodified gels in both THP1 medium (from 111 ± 21 to 42 ± 8 pg/ml) and pro-inflammatory conditions (from 153 ± 37 to 105 ± 22 pg/ml), and nonsignificantly decreased in antiinflammatory conditions (from 35 ± 10 to 12 ± 7) ( Figure 5(d)). The Finally, it was assessed whether the effects of tethered IL-10 on macrophage pro-inflammatory activity were comparable with those of soluble IL-10 ( Figure 6(a)). Specifically, levels of macrophage secreted TNFα and IL-1β were assessed by macrophages in either scaffold in the presence of M1 stimulatory cytokines. TNFα secretion was significantly downregulated for macrophages in M1 media when exposed to IL-10, but there was no significant difference between those treated solubly with IL-10, azide-modified IL-10, or those in IL-10 tethered microgel scaffolds (Figure 6(b)). Similar results were observed for IL-1β, with no significant change in macrophage secretion when exposed to IL-10 in each of the three forms in M1 media (Figure 6(c)). This indicates that neither the modification of IL-10 with the azide handle nor   52,53 In this study, hMSC laden microgels were cultured in the presence of prepolarized M1 macrophages and found to limit activation to the M1 phenotype, even in the presence of pro-inflammatory stimuli. Specifically, hMSCs downregulated the macrophage's production of the proinflammatory cytokines TNFα and IL-1β. Our modular cell culture platform allows us to investigate the immunomodulatory potential of F I G U R E 5 Tethered IL-10 microgel scaffolds reduces the secretion of pro-inflammatory activity by THP1a. (a) THP1s were then encapsulated within unmodified or IL-10 tethered microgel scaffolds and polarized to pro-or anti-inflammatory conditions as normal. (b) TNFα secretion was significantly decreased in all IL-10 tethered network conditions compared to the corresponding unmodified networks. (c) IL-10 networks significantly reduced IL-1β in THP1 medium and proinflammatory conditions, and nonsignificantly reduced in tethered IL-10 networks in antiinflammatory conditions. (d) IL-6 was also significantly reduced in THP1 medium and proinflammatory conditions, and nonsignificantly reduced in tethered IL-10 networks in antiinflammatory conditions. (e) IL-8 was significantly reduced in proinflammatory conditions and nonsignificantly reduced in tethered networks in both THP1 medium and anti-inflammatory conditions. (f) Finally, IL-10 secretion by macrophages was significantly elevated in IL-10 gels compared to their unmodified counterparts. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, n.s. = not significant (p > 0.05) for one-way ANOVA with multiple comparisons between groups hMSCs in response to pro-inflammatory cell activity in a 3D culture environment. Deeper investigation of the immune-regulatory function of hMSCs in response to high levels of pro-inflammatory factors 14 can be used to design and engineer acellular biomimetic scaffolds for regenerative therapies.

| DISCUSSION
As the immunomodulatory potential of hMSCs is modulated by environmental stimuli, we set out to understand the effects of proinflammatory macrophages on the hMSC secretome. Co-culture models can complicate determination of the source of cytokine production from the two cell types, as hMSCs can also secrete a wide variety of cytokines. 48 To circumvent this, hMSCs were first cultured in M1 conditioned medium (pro-inflammatory conditions) to isolate their secretory properties from that of the macrophages. The hMSC secretory properties changed dramatically in pro-inflammatory conditions, downregulating a variety of pro-inflammatory cytokines and upregulating IL-10, as measured by a broad screen cytokine array. As we observed a decrease in pro-inflammatory activity of macrophages co-cultured with hMSCs, we focused on the specific role of the main Azide-modified and tethered IL-10 result in similar changes in macrophage phenotype as unmodified IL-10. (a) Microgel scaffolds were treated with M1 induction media along with 10 ng/ml IL-10 or IL-10-N 3 and compared to IL-10 tethered microgel scaffolds. (b) TNFα secretion was reduced in all IL-10 conditions with no significant difference between the soluble, azide modified, and tethered groups. (c) The secretion of the pro-inflammatory cytokine IL-1β was significantly decreased upon treatment with 10 ng/ml IL-10, IL-10-N 3 and in IL-10 tethered microgel scaffolds in M1 induction conditions. **p < 0.01, ***p < 0.001, n.s. = not significant (p > 0.05) scaffolds to modulate macrophage polarization. While cell-based therapies provide promise for regulating regeneration, there are also significant drawbacks (e.g., variability in therapeutic efficacy, regulatory issues, low cell survivability). [64][65][66] Acellular biomaterials platforms that could replicate aspects of the immunomodulatory behavior of hMSCs may prove advantageous for the future of regenerative medicine. 67,68 We demonstrate that IL-10 is able to reduce macrophage proinflammatory activity, both when introduced solubly in the medium and when tethered directly to microgel scaffolds using a SPAAC reaction.
Macrophages encapsulated in IL-10 functionalized scaffolds had reduced pro-inflammatory activity, similar to macrophages exposed to soluble IL-10, supporting the notion that the protein is stable after conjugations. Future studies should further investigate this concept, specifically considering changes in the duration of IL-10 signaling to cells. It has been documented that the context of a biomolecules presentation (i.e., soluble or tethered) can greatly impact its stability or cellular signaling. 39,69,70 Any extension in the lifetime of cue signaling would highlight the potential regenerative impact of these bioactive acellular materials.
To fully explore the potential of these scaffolds, translation to an in vivo setting is needed. The in vitro studies presented here provide a highly specified environment for understanding cell-material interactions in a precise manner. The demonstration that an acellular platform can provide similar anti-inflammatory stimuli to macrophages as hMSCs is a significant demonstration and holds promise for future translational applications. In total, we anticipate that the development of our in vitro platform that allows for probing of immune cell polarization and reprogramming in response to matrix interactions in precise manners should aid in the design of scaffolds for regenerative therapies.

| Microgel scaffold formation and cell encapsulation
Cell-laden microgel scaffolds were formed as previously described. 37

| Macrophage polarization analysis
Macrophage medium was collected at the end of polarization or repolarization and saved for specific protein analysis. ELISAs (R&D systems "Duo" kits for TNFα, IL-1β, IL-6, IL-8, IL-10) were performed according to the manufacturer's protocols.

| Statistical analysis
Statistical analysis and interpolation for specific ELISAs was conducted using GraphPad Prism software. Statistical significance was determined using one-way ANOVAs with multiple post hoc comparisons (Tukey correction). All data represent three independent biological replicates unless noted otherwise. All data are presented as the mean value plus/minus SD unless otherwise stated.

| CONCLUSION
In the presented work, microgel assembled scaffolds were designed to modulate macrophage polarization directly. THP-1 cells were encapsulated with microgel networks and polarized to M1 and M2a phenotypes. Macrophage phenotype was reprogrammed through the introduction of exogenous cytokines, and this repolarization was further improved through hMSC co-culture conditions. Further investigation of the immunomodulatory potential of hMSCs revealed a distinct secretory phenotype in inflammatory conditions and revealed IL-10 as a key mediator of the reprogramming of macrophages. Finally, IL-10 was introduced into macrophage culture, both solubly and tethered to microgel subunits, to direct macrophage polarization. These results can be used to further inform biomaterial design in regenerative medicine, both for cell transplantation efforts and acellular biomaterial implantation.