Engineering the MSC secretome: A hydrogel focused approach

Abstract

The therapeutic benefits of exogenously delivered mesenchymal stromal/stem cells (MSCs) has been largely attributed to their secretory properties. However, clinical translation of MSC-based therapies is hindered due to loss of MSC regenerative properties during large-scale expansion and low survival/retention post delivery. These limitations might be overcome by designing hydrogel culture platforms to modulate the MSC microenvironment. Hydrogel systems could be engineered to: (i) promote MSC proliferation and maintain regenerative properties (i.e., stemness and secretion) during ex vivo expansion, (ii) improve MSC survival, retention, and engraftment in vivo, and/or (iii) direct the MSC secretory profile using tailored biochemical and biophysical cues. Herein we review how hydrogel material properties (i.e., matrix modulus, viscoelasticity, dimensionality, cell adhesion, and porosity) influence MSC secretion, mediated through cell-matrix and cell-cell interactions. In addition, we highlight how biochemical cues (i.e., small molecules, peptides, and proteins) can improve and direct the MSC secretory profile. Lastly, we provide our perspective on future work towards the understanding of how microenvironmental cues influence the MSC secretome, and designing the next generation of biomaterials, with optimized biophysical and biochemical cues, to direct the MSC secretory prolife for improved clinical translation outcomes.

Graphical Abstract

Hydrogel culture platforms can be engineered, with optimized material properties in combination with specific biochemical cues, to modulate the MSC microenvironment to promote global secretion or direct a specific secretory profile. Utilizing these biomaterials to control MSC secretion has the potential for improved clinical translation outcomes.

1. Introduction

Originally isolated from the bone marrow in the 1970s, mesenchymal stromal cells (MSCs) were thought to be a new stem cell line capable of differentiating into multiple lineages.^[1] In 2006, the International Society for Cellular Therapy (ISCT) defined a set of characteristics to identify MSCs based on their adherence to plastic, their spindle-like morphology, cell surface marker profile, and their trilineage differentiation potential (i.e., chondrogenic, osteogenic, and adipogenic).^[2] Since these early discoveries, MSCs have been found in many tissues, including adipose, muscle, dental pulp, Wharton’s jelly, and umbilical cord, and have been differentiated into additional pathways, such as cardiogenic, neurogenic, and myogenic.^[3,4]

Due to their ease of isolation, proliferation capacity, and multipotency, MSCs have been widely explored for use in various cell-based therapies.^[4,5] However, studies tracking in vivo cell fate found the therapeutic benefits of exogenously delivered MSCs were not primarily related to their differentiation, but instead their secretory properties.^[6–8] Specifically, MSC secrete factors that can signal to endogenous cells and influence proliferation, resolve inflammation, deposit matrix, and heal wounds.^[6–8] As a result, recent academic and clinical efforts have focused on MSC-secreted trophic factors and understanding their role as potent modulators of tissue regeneration.^{[4,6,7,9–11]}

1.1. MSC secreted factors are key regulators of regeneration

MSCs secrete cytokines, chemokines, inflammatory factors, growth factors, exosomes and microvesicles filled with proteins and genetic material, collectively referred to as the MSC secretome.^[4,9,11] These factors signal to endogenous cells in a manner that can increase proliferation,^[12] direct migration,^[13] initiate differentiation,^[14] and even modulate activation^[15] or polarization^[16,17] of immune cells. Based on these findings, in 2019, the ISCT updated the definition of MSCs to include functional assays, such as the trophic factor secretion, modulation of immune cells, and promotion of angiogenesis.^[18]

The potent, yet versatile secretion profiles of MSCs, has led to their use in treating a diverse range of diseases. For example, since 2010, ten MSC therapies have been approved around the world to treat Graft Versus Host disease (e.g., Canada, New Zealand, Japan), critical limb ischemia (e.g., India), and complex perianal fistulas to treat Chron’s Disease (e.g., Europe).^[5] In the United States, a query of clinical trials (NIH, clincaltrials.gov) mentioning mesenchymal stem cells found over 1200 trials. Approximately 25% of trials focused on treating musculoskeletal diseases, ~20% on autoimmune diseases, and ~10% on diseases related to the cardiovascular, neurological, or respiratory systems (Figure 1a). Of further note, the MSCs were predominantly sourced from bone marrow, followed by isolation from adipose tissue and the umbilical cord (Figure 1b).

Figure 1. Mesenchymal stem cell-based clinical trials in the United States between 2010–2020.

a) MSC clinical trials based on the organ system affected by the disease. b) Source of MSCs used in clinical trials. c) Number of clinical trials using MSCs over time. d) MSC clinical trials categorized by phase. e) Methods of delivering MSCs in clinical trials. Local refers to direct injection into the targeted tissue; systemic refers to intravenous infusion; and biomaterial refers to its combination with MSC delivery. These data were obtained from clinicaltrials.gov and searching for “Mesenchymal Stem Cell” in the “Other” category, limited to the United States. Data was collected on October 11, 2020.

The use of MSCs as a therapeutic continues to grow, as evidenced by the growth in the number of US clinical trials utilizing MSCs over the past ten years (Figure 1c). In fact, the largest increase occurred in 2020, with over 230 new trials documented through October (many associated with the COVID-19 pandemic). While the number of early Phase (Phases 1–2) clinical trials has consistently increased, many trials fail to progress to Phase 3 or beyond (Figure 1d). This significant drop is just one indication of the many challenges to ensure the efficacy of MSC-based therapies. Looking how MSCs have been delivered, >90% of trials directly inject MSCs into the body, either locally or systemically (Figure 1e), but evidence and intuition has shown that this leads to low MSC survival and retention upon delivery. As a result, several research groups have been engineering scaffolds for MSC delivery, but to date, only 6% of clinical trials combined MSCs with biomaterials (Figure 1e). Clearly, there are many opportunities to improve on MSC therapies and advance the technology to latter stages of clinical trials and regulatory approval.

1.2. Challenges hindering MSC clinical success

A large discrepancy remains between the number of MSCs that can be easily and reliably isolated from the bone marrow, and the number of cells needed for a clinical dose. For example, one milliliter of bone marrow contains only 10–100 MSCs; however, most clinically relevant treatments require 1–200 million cells per dose. These calculations highlight the fact that after their isolation, MSCs must be expanded ex vivo.^[5,19,20] Large-scale MSC expansion is typically performed in bioreactors that rely on materials such as tissue culture plastic (TCPS) (e.g., multilayered flasks) or silica and dextran microspheres (e.g., microcarriers).^[20,21] However, expansion of MSCs on these substrates, as well as enzymatic passaging methods, can bias MSCs towards an osteogenic fate,^[22] cause loss of multipotency (e.g., chondrogenic differentiation),^[22,23] hinder DNA repair,^[24] induce replicative senescence,^[22,25,26] and even decrease expression of surface markers (e.g., CD105, CD90, CD73) that are associated with the MSC undifferentiated phenotype. Others have also shown that two-dimensional (2D) culture is not as effective in maintaining MSC secretory properties compared to three-dimensional (3D) culture methods.^[26,27]

Following expansion, MSCs can be delivered locally, at the site of injury, or systemically throughout the body.^[28–30] However, upon injection into a tissue site, many MSCs are washed away, phagocytosed, or necrose.^[29,31] In fact, studies have shown that < 5% of administered MSCs are present in the tissue a few hours after transplantation.^[32] Similarly, systemic administration of MSCs, typically through intravenous injection, can lead to accumulation in the lungs and clearance by monocytes within 24 hours.^[33–36] This rapid clearance requires multiple doses of MSCs for clinical efficacy, resulting in greater reliance on in vitro expansion.^[37]

Despite their high innate paracrine activity, the method used to expand MSCs ex vivo can further enhance or alter their secretory function. For example, biochemical priming methods (e.g., exposure to specific biochemical factors), hypoxic microenvironments,^[38,39] pro-inflammatory cytokines,^[40] small molecules,^[41,42] and various growth factors^[5,43] have all been shown to enhance MSC secretory properties. Depending on the disease context, specific factors secreted can be important (e.g., angiogenic factors for vascularization); however, an overall increase in the secretion of all factors may also be detrimental (e.g., pro-inflammatory factors that exacerbate inflammation in rheumatoid arthritis treatments). In the end, while priming methods provide promising strategies to alter MSC’s secretory profiles, the altered profile can be short lived, as the cells revert back to their steady state once the cue is removed.^[44]

Another critical challenge to address towards their clinical use is the hemocompatibility of MSCs and MSC based products.^[45] Previous studies have shown that MSC incompatibility with the innate immune cascade of the blood has resulted in adverse events such as embolism and thrombotic complications.^[46–49] Furthermore, incompatibility of MSCs can be due to the expression of tissue factor (CD142), a glycoprotein which instigates the extrinsic pathway of the blood coagulation cascade through thrombin initation.^[45,50]

1.3. Designing hydrogels to improve MSC clinical translation

Advances in MSC biology and bioengineering have led to the identification of strategies which have the potential to address many of the limitations related to MSC-based therapies. Specifically, researchers have focused on two general approaches. The first alters MSCs directly, using strategies such as genetic engineering, cell surface modifications, or intracellular nanoparticle delivery to alter MSC functions. For example, advances in gene editing technologies have been used to increase MSC secretion of specific therapeutic factors, especially those relevant for treatment in a specific disease scenario where one factor can make a large therapeutic impact. Cellular approaches, though, do not directly improve the in vivo retention time of MSCs and their protection from immune clearance in vivo. Thus, a second, complementary approach is to control the MSCs environment using biomaterials, during expansion and/or after delivery. Biomaterials can provide matrix interactions to improve MSC survival, local retention, or even influence secretory properties via outside-in signaling (e.g., mechanosensing).

Biomaterials play critical roles in current MSC-based engineering approaches, as they allow user control of the biophysical and biochemical extracellular matrix signal that can influence a cell’s behavior and function. To date, many biomaterials have been investigated to control MSC functions, ranging from natural to synthetic materials, polymers, ceramics, metals, including those with complex 2D and 3D architectures. While a vast assortment of biomaterials exists to date this perspective focuses on hydrogels, and their wide applications for MSC expansion and delivery. We direct the reader to other reviews which carefully evaluate the broader range of biomaterials used for tissue engineering applications.^[51–53]

Hydrogels can be synthesized from synthetic or natural homopolymers, copolymers, or macromolecular monomers that readily dissolve in water, but are physical or chemically crosslinked to render them insoluble.^[54,55] The resulting network imbibes large amounts of water, but the crosslinks impart structural integrity along with unique material properties.^[56,57] Depending on the chemistry, hydrogels can be engineered to degrade via cell-directed mechanisms (e.g., enzyme cleavable crosslinkers), environmental mechanisms (e.g., hydrolysis, pH changes), user-directed mechanisms (e.g., light), or a combination of the three. The use of hydrogels offers a highly tunable material platform which allows for temporal and spatial control over cell-matrix and cell-cell interactions. In addition, cells can be expanded on the surface of hydrogels (i.e., 2D culture) or encapsulated within a hydrogel (i.e., 3D culture). To promote MSC attachment, proliferation, differentiation, and secretory properties, hydrogels can be functionalized with integrin-binding peptides,^[58,59] degradable peptide sequences,^[60] small molecules,^[61,62] nanoparticles,^[63] or even chemokines and growth factors.^[40,64]

Some of the barriers to the success of clinical therapies might be overcome by designing hydrogel culture platforms capable of promoting MSC secretory properties during expansion and delivery (Figure 2). Specifically, hydrogels systems could be engineering to: (i) promote MSC proliferation and maintain regenerative properties (i.e., stemness and secretion) during ex vivo expansion, (ii) improve MSC survival, retention, and engraftment in vivo, and/or (iii) direct the MSC secretory profile using tailored biochemical and biophysical cues.

Figure 2. Culture systems to expand and deliver MSCs.

A variety of material platforms are used to expand and deliver MSCs, including tissue culture polystyrene (TCPS), hydrogels (either on 2D surfaces or 3D encapsulation), microspheres, porous scaffolds, and or multi-cellular spheroids. Each of these systems has the ability to influence MSC proliferation, secretory properties, and survival upon delivery. Strategies that encapsulate cells (e.g., hydrogels, microspheres) lead to higher levels of cell-matrix interactions compared to 2D surfaces. Porous scaffolds and multi-cellular spheroids lead to more cell-cell interactions. As a qualitative assessment, minus sign (−) indicates a system that does not improve the corresponding property, while the single up arrow (↑) indicates a slight improvement and the double arrow (↑↑) indicates a higher level of improvement.

This perspective highlights some of the current strategies being developed in the biomaterial community to understand and direct MSC secretion, with a significant focus on influencing the cytokine, chemokines, inflammatory factors, and growth factors secreted by MSCs. While examples herein focus specifically on the aforementioned MSC secretions, we acknowledge that microvesicles and exosomes are also key components of the MSC secretome. However, the influence of cell culture methods on MSC exosomes and microvesicles has been thoroughly reviewed elsewhere.^[65–67] Instead, this contribution reviews how material properties, such as the matrix modulus, viscoelasticity, dimensionality, cell adhesion, and porosity, influence MSC secretion. Emphasis is placed on how hydrogel culture platforms can be engineered to control MSC cell-matrix and cell-cell interactions and increase overall secretion. In addition, examples are selected to highlight how biochemical cues, such as peptides, small molecules, and proteins, can improve and direct the MSC secretory profile. Finally, we posit on future directions to fill gaps with respect to understanding how microenvironment can influence the MSC secretome and designing the next generation of biomaterials, with optimized biophysical and biochemical cues, to direct the MSC secretome for improved clinical translation outcomes.

2. Tailoring hydrogel network properties to direct MSC secretion

Designing biomaterials to direct MSC secretory profiles first requires a basic understanding as to how cells interact with their matrix. In the body, cells are surrounded by an extracellular matrix (ECM) that provides a structural basis for the tissue, providing a foundation for cell adhesion and an environment for cell-cell and cell-matrix signaling interactions. The composition of the ECM is tissue specific and can vary greatly; however, its main components are structural proteins, such as collagen and branched glycosaminoglycans. In addition to providing binding sites for cell adhesion, these proteins sequester bioactive molecules making the ECM capable of directing cell behaviors.^[68] Countless studies have confirmed the crucial role of the ECM in directing cell growth, differentiation, and disease progression.^[68–70] Cell-matrix interactions are largely facilitated through integrins; receptors on the cell surface that are internally connected to the cytoskeleton. Integrin receptors span the cell membrane and are comprised of an alpha and beta subunit. In mammals, 18 distinct alpha domains and eight distinct beta domains have been identified. In humans, combinations of these subunits result in 24 integrin receptors that each bind to specific amino acid sequences found in various ECM proteins. For example, the sequence RGD (arginine – glycine – aspartic acid), present in fibronectin, laminin, and vitronectin, is bound by α5β1, α8β1, αVβ1, αVβ3, αVβ5, αVβ6, αVβ8, and αIIbβ3 integrins. Upon binding, multiple integrin receptors cluster together with other cytosolic proteins to initiate the formation of a focal adhesion complex, which ultimately mature and facilitate cell spreading and migration.^[71]

With this information in mind, biomaterial researchers often incorporate ECM components into engineered matrices to promote cell attachment and to ensure cell responsive behaviors due to changes in matrix mechanics. Incorporation of ECM components into biomaterials can be accomplished by using specific ECM proteins as scaffolding materials, such as in collagen-based gels,^[72,73] decellularized tissues,^[74] or by including whole length ECM proteins.^[75] Additionally, bioactive peptide sequences, such as RGD (a fibronectin mimic) or GFOGER (a collagen mimic), can be incorporated into scaffolds created with bioinert polymers (Figure 3a).^[76–78] While these proteins are critical to ensuring cell adhesion, ECM proteins can also independently participate in initiating downstream cell signaling, specifically by influencing the MSC secretory profile.

Figure 3. Matrix composition and physical properties influence MSC secretion.

a) To promote cell adhesion, bioactive cell adhesion molecules are often incorporated in hydrogel formulations. Integrins, present on the cell surface, bind to amino acid sequences found in ECM adhesion proteins. For example, RGD and GFOGER peptides, fibronectin and collagen mimics respectively, have been shown to differentially influence MSC secretory profiles. Similarly, peptides derived from N-cadherins (e.g., HAVDI) can mimic cell-cell interactions and influence MSC secretory properties. b) Bulk hydrogel properties (e.g., stiffness, viscoelasticity) influence MSC interactions and global secretory properties. Porosity and degradation properties can direct MSC clustering and promote secretion through increased cell-cell contacts.

2.1. ECM composition

In 2014, De Lisio et al. showed that expression of paracrine factors, such as Interleukin (IL)-6, granulocyte-macrophage colony-stimulating factor (GM-CSF), vascular endothelia growth factor (VEGF), transforming growth factor beta-1 (TGF-b1), and tumor necrosis factor-alpha (TNF-a), were altered between murine MSCs (mMSCs) cultured on plates coated with collagen or laminin, in which different integrins mediate cell attachment.^[72] Generally, the gene expression of the aforementioned growth factors was downregulated when mMSCs were grown on collagen. Abdeen et al. compared the effects of fibronectin, collagen, and laminin coatings on human MSC (hMSC) angiogenic factor secretion.^[73] Human microvascular endothelial cells (HMVECs) treated with hMSC conditioned media from cells cultured on fibronectin coated polyacrylamide hydrogels (E~40 kPa) resulted in increased tubulogenesis compared to HMVECs treated with media from hMSCs cultured on collagen I and laminin coated gels. However, this effect was fully dependent on the stiffness of the gels, and no differences in HMVEC tubulogenesis was observed with hMSC conditioned media treatment from 0.5 kPa and 10 kPa gels coated with different ECM proteins. In another study, monocyte chemotaxis towards MSC conditioned media was enhanced by MSC culture on hydrogels containing hyaluronic acid.^[79]

With respect to ECM mimicking peptide incorporation, in vitro encapsulation of hMSCs in poly(ethylene glycol) (PEG) hydrogels modified with the peptide GFOGER, a collagen I mimic that binds to integrin α5β1, improved secretion of inflammatory factors, IL-8 and IL-6, and chemotactic factor monocyte chemoattractant protein-1 (MCP-1) compared to hydrogels modified with RGD alone.^[76] When hMSCs in GFOGER gels were delivered to a murine radial segmental defect, microCT revealed significantly higher levels of new bone formation was observed at both 4 and 8 weeks, compared to defects treated with hMSCs in the absence of GFOGER.^[76] Collectively, these results show that ECM composition can be a powerful tool to direct MSC secretion.

2.2. Matrix modulus

In addition to the biochemical components present in the ECM, the mechanical cues present in a cell’s surroundings can also influence their activity. Utilizing biomaterial culture platforms, researchers can easily modify several bulk material properties, such as stiffness, viscoelasticity, porosity, and degradability, all of which can direct MSC fate and secretory behavior (Figure 3b).

For example, increased matrix stiffness can lead to increased cytoskeletal tension, resulting in a more open nucleus and the translocation of key transcriptional regulators, which initiate downstream gene expression. Specific to hMSCs, stiff and soft mechanical cues are known to direct cell fate, proliferation, and apoptosis. Prolonged exposure of hMSCs to soft or stiff matrices can result in an irreversible commitment of the cells to the lineages specified by the matrix elasticity of their substrate, despite contradictory soluble differentiation cues; a phenomenon dubbed mechanical memory. This memory is often thought to be determinantal in the context of culturing, as serial passaging of MSCs is known to cause a loss of multipotency, increased DNA damage, and eventual senescence.^[22,24]

Compared to differentiation studies using MSCs, a smaller number of investigations have explored the influence of matrix stiffness on MSC cytokine secretion. hMSCs cultured on 2D polyacrylamide gels (E~40 kPa) coated with fibronectin (i.e., α3β1, αVβ1) resulted in increased VEGF and insulin-like growth factor (IGF) secretion compared to hMSCs on soft (E~500 Pa) gels with the same coating. To determine if hMSC secreted factors could cause changes in a functional cellular output, conditioned media from hMSCs was used to promote human microvascular endothelial cell (HMVEC) tubulogenesis.^[73] Results from this study revealed the highest tube area in cultures with media collected from hMSCs cultured on 40 kPa gels.^[73]

Expanding on the influence of matrix stiffness on MSC secretory factors, Ogle et al. cultured MSCs on 30 kPa and 100 kPa on PEG-diacrylate (PEGDA) hydrogels functionalized with peptides targeting integrins or cadherins or on PEGDA hydrogels containing hyaluronic acid or heparin.^[79] Regardless of the functionalization, MSCs cultured on 30 kPa gels exhibited an overall abundance of immunomodulatory factors relative to MSCs cultured on 100 kPa PEGDA gels or TCPS.^[79] Further, conditioned media from MSCs cultured on 30 kPa gels enabled promotion of vessel network formation in human umbilical cord endothelial cells (HUVECs). However, the 30 kPa gels were not able to promote MSC proliferation as much as stiffer hydrogel conditions relative to TCPS, indicating that a combination approach might be necessary to achieve high numbers of secretory MSCs.

A potential combinatorial approach would be to use hydrogel culture post TCPS expansion to rescue MSC secretory phenotypes. Rao et al. observed that hMSCs serially passaged on TCPS lose their secretory properties over time.^[26] In late passage hMSCs, a significant decrease was observed in the secretion of factors related to cell proliferation and differentiation, such as TGF-β1, VEGF, glial cell-derived neurotrophic factor (GDNF), and epidermal growth factor (EGF). However, simply transferring hMSCs to soft hydrogels, at either early or late passages, can restore the secretion of key chemokines, growth factors, and inflammatory factors. Collectively, these results indicate that hydrogel interventions could be employed during ex vivo expansion to instruct MSCs to secrete factors before they are delivered in vivo. Additionally, factors from MSCs cultured on 2D biomaterial substrates could be collected and delivered cell-free into patients.

Although tuning matrix stiffness might prove advantageous when harvesting MSC-secreted factors or priming MSCs before intravenous delivery, embedding MSCs in biomaterial matrices for in vivo delivery is often efficacious. For example, Won et al. suggested that a soft ECM may enhance the effect of the inflammatory stimuli, TNF-α, on MSC secretion. hMSCs were encapsulated within soft (E~2 kPa) and stiff (E~35 kPa) RGD-functionalized alginate hydrogels.^[58] hMSCs in soft matrices secreted higher levels of chemokines involved in monocyte recruitment (CCL2, IL-6) upon treatment with TNF-α, which was attributed to an increased clustering of TNF receptors and redistribution of actin polymerization mediated by lipid rafts. Both led to NF-κB activation and upregulation of downstream genes (e.g., CCL2 and IL-6).^[58]

2.3. Viscoelasticity

Viscoelastic hydrogels are synthesized using covalent adaptable linkers, hydrophobic interactions that increase with temperature, and guest-host interactions. Viscoelastic hydrogels are of growing interest within the field of tissue engineering, as they can better recapitulate aspects of the mechanical properties of soft tissues and are often injectable. To investigate the influence of a viscoelastic material on the MSC secretome, Liu et al. cultured MSCs on polydimethylsiloxane (PDMS) surfaces with varying shear storage moduli (~1–100 kPa) and viscoelastic properties ((G”/G’) = tan (d) ~0.2–1.2).^[80] MSCs cultured on low stiffness (~1 kPa) substrates with the highest tan (d) (>1) had a >3-fold increase in osteoponin expression, relative to other conditions. Similar increases were observed for IL-8, MCP-1, IL-21, brain-derived neurotrophic factor (BNDF), and stromal cell-derived factor (SDF)-1a for MSCs on compliant substrates.

Viscoelastic, shear-thinning hydrogels have also been developed as injectable MSC delivery system and to test the influence of viscoelasticity on the MSC secretome.^[81]. Human adipose derived stem cells (hASCs) cultured in gels with intermediate stiffnesses (100–300 Pa) and relaxation time constants (12–13 seconds) resulted in increased secretion of angiogenic factors (VEGF, ANG, HGF) compared to cells cultured in gels with lower or higher stiffnesses and relaxation times. While further secretory studies need to be conducted using MSCs, viscoelastic materials show promise as injectable cell delivery systems for clinical use.

Biophysical and biochemical cues delivered to MSCs through their matrix environment, such as stiffness, viscoelasticity, and ECM proteins, can direct both their constitutive factor secretion and their responsive factor section. However, it is clear that the influence of matrix stiffness on MSC secretome is factor dependent and thus, clear conclusions cannot be drawn as is possible with matrix stiffness and MSC differentiation. Additionally, other types of mechanical stimuli, such as strain, compression, or tension, and other material properties, such as viscoelasticity and topography, need further experimentation to fully determine their influence on MSC secretion.

3. Material design strategies to increase MSC cell-cell interactions

Beyond cell-matrix interactions, MSCs are also dependent on cell-cell signaling cues, mediated in part through N-cadherins present on the cell surface. N-cadherins contain both an extracellular domain, that dimerizes with N-cadherins present on neighboring cells, and an intracellular domain, linked to actin cytoskeleton which is capable of facilitating downstream signaling. Specifically, the cytoplasmic domain binds to b-catenin, a transcriptional regulator involved in Wnt and NF-κB pathways. In vivo, bone marrow derived MSCs reside in N-cadherin expressing clusters in their native niche which help to maintain their stemness.^[4] In vitro, N-cadherin expression has been confirmed on MSC cell membranes,^[82] and its expression has been shown to be elevated in MSC clusters.^[77]

Relevant for MSC therapies, N-cadherin signaling can directly upregulate MSC secretion. MSCs residing in cell aggregates or clusters, with increased N-cadherin expression, also exhibit increased secretion of growth factors.^[82] Further validating this observation, several groups have observed a loss of MSC secretory abilities when cells are cultured in the presence of N-cadherin blocking antibodies.^[27,83] In one example, an N-cadherin blocking antibody was administered to aggregated MSCs (~40 cells/clusters) in a porous PEG-based microgel scaffold with average pore diameters of 200 mm. Over a ten-fold decrease was observed for 45% of the measured cytokines when N-cadherin signaling was blocked for three days in MSC clusters.^[77] This points to increased cellular contacts and N-cadherin signaling as a valuable tool to sustainably increase MSC secretion.

3.1. Culture of MSCs as spheroids

To increase MSC cell-cell contacts and better mimic their physiological niche, researchers have employed the use of MSC spheroids, or large cell aggregates, produced by hanging drop culture or centrifugal aggregation techniques.^[84] Spheroid culture has been shown to improve MSC survival, multipotency, and secretory potential. Specifically, Leach has shown that MSCs in spheroid culture have distinct transcriptomes compared to adherent cells, with a marked increase in genes associated with ontologies for wound healing and inflammatory responses.^[59,84] Bartosh et al. found that MSCs in spheroids (25,000 cells/spheroid) secreted high levels of potent anti-inflammatory factors, such as tumor necrosis factor-α-stimulated protein 6 (TSG-6) and stanniocalcin-1 (STC-1), and were able to decrease activation of macrophages in vitro, and inflammation in vivo using a peritonitis mouse model.^[85]

In addition, spheroid size can affect MSC secretion. MSCs in large spheroids (40,000 cells/spheroid) have elevated secretion of several important cytokines involved in inflammatory signaling, including growth-regulated oncogene (GRO), interferon gamma (IFN-γ), and IL-10, compared to MSCs in smaller spheroids (10,000 cells/ spheroid).^[86] However, apoptosis and necrosis can increase at the center of very large spheroids (>100,000 cells) due to nutrient deficiencies.^[85]

While spheroid culture alone can improve MSC secretion, encapsulating aggregates in biomaterials may further improve and control secretion. Even though matrix interactions are limited to the peripheral cells, these signals can be amplified throughout the aggregate. For example, MSC spheroids encapsulated in RGD modified alginate hydrogels resulted in elevated secretion levels of VEGF, TGF-b, GRO, and EGF compared to spheroids in unmodified gels.^[59] Once encapsulated, the stiffness of the matrix can also influence cell secretion. MSC spheroids encapsulated in stiff fibrin gels (G’~1200 Pa) had higher VEGF secretion, while spheroids in soft gels (G’~100Pa) secreted higher levels of PGE2.^[87] Using multifactorial statistical analysis, the researchers were able to a predict an optimal intermediate gel formulation (G’~ 400Pa) in which both VEGF and PGE2 levels were highest.^[87]

3.2. 3D porous biomaterials

In lieu of spheroid culture to increase MSC cell-cell contacts, recent studies have focused on the use of porous biomaterial environments to cluster MSCs. Porous scaffolds can be synthesized using various methods, including dissolution of embedded porogens by solvents,^[88,89] in situ degradation of soft materials,^[90] cell-mediated degradation of bulk single-phase hydrogels,^[91] 3D printing and extrusion,^[92] and lyophilization.^[27] Additionally, individual hydrogel building blocks on micron length scales, termed “microgels”, have been used to create porous scaffolds. Monodisperse microgels have been synthesized using microfluidic devices^[93,94] or on a bulk scale using suspension polymerization^[83] techniques. Once synthesized, the microgels can then be assembled in the presence of cells to create 3D porous networks where cells reside in the void spaces.

Compared to bulk scaffolds where cells are embedded in hydrogels with nanometer-sized pores, Qazi et al. observed that rat MSCs (rMSCs) embedded in lyophilized alginate scaffolds (average pore size of 122 ± 29 μm) secreted higher levels of cytokines and regenerative factors (e.g., hepatocyte growth factor (HGF), fibroblast growth factor-2 (FGF-2), and IGF) compared to those encapsulated in bulk alginate gels.^[27]

In porous systems, a material characteristic (i.e., pore size) can be used to direct cell cluster size, albeit in much smaller numbers than those used in typical spheroid cultures. Caldwell and Rao et al. used microgels networks to control cluster size.^[77] MSC clusters (~40 cells/cluster) in the 200 μm scaffolds expressed elevated cytokine secretion as measured by cytokine array. In another example, extrusion printing was used to fabricate gelatin scaffold with three pore sizes: ~200 μm, ~300 μm, and ~400 μm. After 3 days of culture, hMSCs in the ~300 μm scaffolds had significantly higher expression of angiopoietin (ANGPT) and HGF, compared to MSCs in other pore sizes. In addition, increased expression of VEGF and FGF was observed after 7 days of MSC culture in scaffolds with pore sizes of ~300 μm, relative to results obtained in 200 μm and 400 μm pores.

In a similar manner used to incorporate integrin binding peptide epitopes, biomaterial scientists have also included N-cadherin peptide mimics into scaffolds to mimic cell-cell interactions. For example, the amino acid sequence HAVDI (Histidine – Alanine – Valine – Aspartic Acid – Isoleucine) binds to N-cadherins,^[82,95] and has led to a >10-fold increase in GNDF and IGF by MSCs. In one study, inclusion of HAVDI resulted in increased secretion of 96% of all measured cytokines by clustered MSCs relative to RGD only conditions. Most notably, the inclusion of HAVDI elevated secretion of non-clustered cells to levels similar to MSCs in clusters.^[77,78] Just as the inclusion of ECM peptide mimics drastically improves cellular adhesion to synthetic biomaterials, the inclusion of HAVDI can be used to radically increase MSC secretion in biomaterials where MSC clustering cannot be induced, thereby decreasing the numbers of MSCs needed per dose.

3.3. Microencapsulation

Another method to achieve hydrogel encapsulated MSC aggregates is through microencapsulation, where cells are encapsulated within individual microgels using a microfluidic device. In one example, mMSCs were encapsulated within alginate microgels, which was later crosslinked with poly-D-lysine and resulted in aggregates of 2–7 cells. The multicellular aggregates had increased expression of anti-inflammatory genes (e.g., IL-10, TSG-6, and TGF-b1) relative to mMSCs on TCPS.^[96] Intravenous delivery of the MSC microspheres intravenously results in an in vivo half-life of over 50 hours, a 5x increase compared the half-life of unencapsulated MSCs.

In addition to increased constitutive factor expression, MSCs in multicellular aggregates respond to pro-inflammatory stimulants more effectively than their single cell counterparts. This heightened response has been observed by MSCs in microporous environments or in spheroids; all of which promote cell-cell contacts.^[85,97] Exogeneous IGF-1 amplified MSC paracrine secretion in a microporous environment (pore size ~120 mm) relative to MSCs in nanoporous gels (pore size~10 nm).^[27] These effects were abrogated in the presence of an N-cadherin blocking antibody, confirming that cell-cell contacts are necessary for MSC reactive responses. Similarly, MSCs in spheroids have increased immunomodulatory paracrine section in the presence of IFN-γ and TNF-a relative to disassociated cells.^[85,98]

Overall, material properties significantly influence MSC secretion, mediated both by cell-matrix and cell-cell interactions. However, these factors are highly interdependent, making it difficult to reach definitive conclusions as to how each individual material properties can be used to direct MSC secretion. Generally, properties, such as matrix stiffness and viscoelasticity in materials with integrin binding domains, instruct MSC secretion though direct cell-matrix interactions. Soft substrates can broadly increase secretion of MSCs cultured on them, but receptor clustering in 3D can be different (e.g., allowing MSCs to respond more effectively to TNF-a treatment). Numerous studies have reported the effectiveness of cell-cell interactions on MSC secretion in an N-cadherin dependent manner. Complementary, material properties, such as porosity, can be used to direct MSCs clustering, thereby increasing secretion. Lastly, as bioactive components are incorporated into materials (e.g., integrin- or cadherin-biding peptides), cell-matrix interactions can further direct the MSC secretion profile, or simply increase total factor secretion, without relying on large cell numbers or exogenous delivery of biochemical factors. While the field must continue to elucidate the underlying mechanisms responsible for outside-in signaling and MSC-matrix interactions on the secretory profile, the role of biochemical compounds are equally important.

4. Biochemical compounds to direct the MSC secretome

MSC culture within hydrogels enables modulation of the local cellular environment to maintain MSC phenotypes during cell expansion, provide protection upon delivery, and promote paracrine secretion. However, the sole use of a material environment is not always sufficient to sustain MSC secretion. In the tissue engineering field, combining hydrogels with biochemical compounds in cell culture is ubiquitous. Numerous biochemical compounds (i.e., small molecules, peptides, and proteins) are known to influence MSC functions (e.g., adhesion, migration, proliferation, and differentiation), and these molecules can be introduced into hydrogels through bulk adsorption or matrix-immobilization (Figure 4a). A subset of these factors has already have been shown to influence and direct the MSC secretory profile. For this reason, combining biochemical priming methods with optimal hydrogel properties that predictably direct MSC secretion may prove especially beneficial in the translation of cell-based therapies.

Figure 4. Biochemical microenvironmental modifications to influence MSC secretion.

a) Biochemical compounds such as small molecules, peptides, or proteins, can be incorporated into hydrogels by a variety of methods focused on either bulk adsorption or immobilization to the matrix. b) Release profiles of biochemical compounds will vary depending on how the compound is incorporated into the hydrogel and can be tailored for immediate or prolonged release and cell exposure over time. c) Immobilization of biochemical compounds to hydrogel matrices enables slower release profiles to be obtained, compared to bulk adsorption, which results in the need for lower biochemical doses. Bulk adsorption of biochemical factors is the simplest method for incorporation; however, burst release profiles often result, necessitating higher concentrations of the bioactive factor.

4.1. Directing MSC secretion using small molecules

Small molecules (<1000 Da) are advantageous for cell-therapy applications because their characteristic properties (i.e., small size, high stability, non-immunogenicity, and low cost) minimize, and even overcome, many of the downsides associated with protein-based biofactors (such as, high cost, poor shelf-life, and recombinant manufacturing considerations).^[99,100] Specific to regenerative medicine, advances in stem cell biology have shown the promise of small molecules to control MSC fate. Examples include: purmorphamine (for osteogenic differentiation),^[101] phenamil (for osteogenic differentiation),^[102] and kartogenin (for chondrogenic differentiation) for MSCs.^[103]

Purmorphamine was discovered by Wu et al. who demonstrated the use of this molecule to differentiate mouse embryonic mesoderm fibroblasts into an osteoblast lineage.^[101] Later, it was revealed that delivery of adsorbed purmorphamine throughout porous calcium phosphate beads resulted in increased trabecular bone formation when implanted in a chick embryo femur.^[104] Park et al. identified phenamil as a molecule that induced osteogenic differentiation and mineralization of mMSCs.^[61] Delivery of phenamil in vitro has been achieved either by adsorption or entrapment of the drug in biodegradable poly(lactide-co-glycolide acid) (PLAGA) scaffolds.^[102,105] In another study, Fan et al. used a combination of phenamil and BMP-2 with PLAGA scaffolds to induce in vitro osteogenesis of MSCs and regenerate bone in a mouse calvarial defect.^[62] To promote chondrogenic differentiation of hMSCs, Johnson et al. identified kartogenin, which also demonstrated chondroprotective effects in vitro.^[103]

While the aforementioned small molecules were identified primarily for inducing MSC differentiation, control of other MSC functions is often necessary in therapeutic interventions. For example, adhesion, migration, and homing of cells to their target organ for regeneration and repair is highly sought-after; however, poor homing to disease sites is often observed when MSCs are systemically infused.^[106] For the first time, a study by Levy et al. used a screen to identify small molecules to improve targeting of systemically infused MSCs.^[42] Specifically, 9,000 signal-transduction modulators were screened to identify hits that increase MSC surface expression of homing ligands, such as CD11a, that bind to intercellular adhesion molecule 1 (ICAM-1). When MSCs were treated with Ro-31–8425 (an identified hit from this screen), increased cell adhesion to an ICAM-1-coated substrate was observed in vitro. Targeted delivery of systemically administered MSCs to inflamed sites in vivo was also achieved in an ICAM-1-binding domain-dependent manner. Pre-treatment of MSCs with Ro-31–8425 prior to delivery in vivo resulted in an increased anti-inflammatory response through decreased expression of TNF-α at the site of inflammation. This use of Ro-31–8425 represents a new paradigm for engineering MSC homing to enhance their therapeutic efficacy.

Modulating the immune response upon MSC delivery is another desirable aspect to control for improving patient outcomes and decreasing adverse effects. For this reason, Yang et al. developed a high-throughput screening method which evaluated a library of 1402 FDA-approved bioactive compounds to activate the secretion of PGE2, an inflammatory mediator secreted by MSCs that can reduce foreign body responses after implantation.^[41] The authors identified tetrandrine (a calcium channel blocker) as a potential candidate to increase MSC secretion of PGE2 through the NF-κB/COX-2 signaling pathway. When co-cultured with murine macrophages, tetrandrine-primed MSCs diminished the level of TNF-α secreted by the macrophages, and when delivered into a murine ear skin inflammation model, a significant reduction in TNF-α levels were observed. These results indicate that tetrandrine-primed MSCs, with enhanced secretion of PGE2, achieved stronger immunosuppressive effects in vivo compared to unprimed MSCs. In addition, the study highlights how small molecule priming can be utilized to increase anti-inflammatory signaling by MSCs. Further identification of small molecules that perform similar functions are needed to develop a library of compounds that influence the MSC secretome.

Other molecules known to have significant effects on many cell functions are hormones. For example, the hormone estrogen can influence cell growth, metabolism, and differentiation in various tissues via estrogen receptors (ER)-α and ER-β, both of which MSCs possess. Hong et al. reported that supplementation of 17-β estradiol (E2), a form of estrogen, significantly increased the proliferation of hMSCs in vitro; however, the dose range over which MSCs responded varied by donor sex.^[107] More specifically, a wider range of E2 concentrations (10⁻⁸ to 10⁻¹² M) was observed to significantly increase male MSC proliferation compared to female MSCs (10⁻⁸ to 10⁻¹⁰ M). E2 supplementation maintained the native MSC phenotype during in vitro expansion by expression of MSC surface markers and their ability to differentiate into osteogenic and adipogenic lineages. These results demonstrate that estrogen supplementation may play an important role in maintaining hMSC phenotype during expansion in vitro, which may help produce the large numbers of undifferentiated MSCs often required for cell-based therapies. The effect of E2 supplementation on the MSC secretome has yet to be investigated.

Another avenue which warrants further exploration is examination of small molecules that are known inhibitors of pathways involved in MSC secretion. For example, inhibitors of NF-κB, TGF-β, and Wnt/β-catenin signaling pathways. Although these pathways regulate multiple cell functions, targeted inhibition would identify specific regulatory proteins that may control crucial MSC secreted factors. For example, the small molecules ML-10B, an inhibitor of NF-kB signaling, was able to suppresses TNF-a induced expression of CCL2 and IL-6 in MSCs, indicating that this signaling pathway is required for MSC pro-inflammatory factor priming.^[58] Further exploration of small molecule inhibitors would provide valuable insight towards developing methods to modulate the MSC secretome mechanistically.

The molecules discussed herein are a small sample of recently identified compounds that target specific MSC functions namely, cell expansion, homing, differentiation, and anti-inflammatory secretion. Most of the studies described include incorporation of small molecules into the cell culture media to modulate specific cell functions in vitro. It is important to note that MSC priming is often short-lived (on the scale of hours to several days).^[5] After delivered in vivo, MSCs frequently lose their directed secretion that was previously obtained in vitro. However, to overcome such challenges, hydrogels can be utilized as delivery vehicles to provide sustained release of biochemical compounds (Figure 4b and 4c), in addition to controlling the local MSC environment.

4.2. Use of pro-inflammatory cytokines to direct the MSC secretory profile

Many studies have demonstrated that MSCs possess a broad range of immunoregulatory abilities that influence both the adaptive and innate immune responses.^[108] In addition, MSCs expanded ex vivo have been shown to suppress the activity of many immune cells, such as, macrophages, T cells, B cells, dendritic cells, and various white blood cells.^[109] Though it is clear that MSCs exhibit immunosuppressive effects, the underlying cellular and molecular mechanisms responsible for such actions have yet to be fully elucidated. However, much evidence points to secretion of soluble factors by MSCs to be the culprit for select immunomodulation functions.

During the inflammatory phase of the wound healing process, neutrophils and macrophages are chemoattracted to the site of injury by bioactive compounds (i.e., various growth factors and cytokines). Once the immune cells arrive, they secrete pro-inflammatory factors such as TNF-a, IL-1b, or IFN-γ to induce inflammation. While this inflammatory step is crucial to initiate the wound healing process, it is quickly succeeded by a more regenerative stage, where endogenous cells deposit matrix and ensure vascularization. A prolonged inflammatory stage can lead to longer wound healing times and can even be responsible for the development of various diseases.

Secretion of anti-inflammatory cytokines by MSCs is highly sought-after towards the development of cell-based therapies. MSC soluble factors, which have been shown to suppress select immune cell functions and transition cells from pro- to anti-inflammatory polarizations, include TGF-b1, HGF, PGE2, IL-6, and IL-10.^[108,109] In addition, MSCs have been shown to alter the cytokine secretion profile of select immune cells by upregulating regulatory cytokines (i.e., IL-10) and downregulating inflammatory cytokines (i.e., IFN-γ, IL-12, and TNF-α), inducing greater anti-inflammatory effects.^[109,110] This characteristic of MSCs makes them extremely powerful to potentially mitigate inflammation and reduce adverse effects upon cellular delivery in vivo.^[110–112] However, this anti-inflammatory factor secretion does not occur naturally for MSCs, rather, they must be probed with pro-inflammatory factors (such as, IFN-γ, TNF-α, IL-1α, and IL-1β) from their surrounding environment in order to direct their secretion profile.^[113–115]

The most common immunomodulatory agents used to direct specific MSC factor secretion are TNF-α and IFN-γ.^[85,116] IFN-γ is often studied because preliminary activation of MSCs by immune cells in vivo can be accomplished by secretion of IFN-γ alone, or in combination with additional cytokines.^[111,117] Alone, or in combination with IFN-γ, TNF-α (a pro-inflammatory cytokine produced by macrophages/monocytes during acute inflammation) is responsible for a diverse range of cell signaling events.^[118] In a study by Chinnadurai et al., MSC treatment with IFN-γ (50 ng/mL) inhibited proliferation of activated T cells and blocked cytokine production by T cells (specifically, IFN-γ, IL-2, and TNF-α).^[64] This inhibition of T cell effector function was found to be through upregulation of programmed cell death-1 ligands (PDL-1). Using media from mMSCs primed with IFN-γ (10 ng/mL), Vigo et al. found the immunosuppressive properties of these cells to be mediated by early phosphorylation of signal transducer and activator of transcription (STAT) 1 and STAT3, as well as inhibition of mTOR activity, leading to inhibition of T cell proliferation.^[40] These results provide insight, demonstrating IFN-γ mediated manipulation of MSCs, and providing an understanding of the intracellular pathways affected by IFN-γ. It is important to note, however, INF-γ stimulation of MSCs is often dose dependent (requiring doses > 10 ng/mL) and short-lived (on the scale of hours to days). MSC priming with TNF-α is known to promote upregulation of select immunoregulatory factors, specifically, PGE2, CCL2, and HGF.^[43] Combinatory preconditioning of MSCs with TNF-α and IFN-γ can increase factor H (a regulator protein in the alternative complement pathway) production by MSCs, thus inhibiting complement activation in both dose and time dependent manners.^[119] In a study by François et al., MSCs primed with TNF-α and IFN-γ suppressed T cell proliferation in vitro due to IFN-mediated indoleamine 2,3-dioxygenase (IDO) upregulation.^[120] This increase in IDO activity in MSCs led to the differentiation of monocytes into IL-10 secreting M2 immunosuppressive macrophages. These M2 macrophages were then responsible for suppression of T cell proliferation in an IL-10-independent manner. These results showcase the immunosuppressive properties of TNF-α and IFN-γ primed MSCs.

In another study, Redondo-Castro et al. pre-treated hMSCs with inflammatory cytokines to prime the cells towards an anti-inflammatory and pro-trophic phenotype in vitro.^[121] hMSCs from three different donors were cultured in vitro and treated with either IL-1α, IL-1β, TNF-α, or IFN-γ. MSCs primed with either IL-1α or IL-1β resulted in increased trophic factor secretion of granulocyte-colony stimulating factor (G-CSF) mediated through an IL-1 receptor type 1 (IL-1R1) mechanism.^[121] To further confirm the anti-inflammatory potential of MSCs, immortalized mouse microglial cells (a population of macrophages found in the central nervous system) were treated with bacterial lipopolysaccharide and exposed to conditioned media of IL-1-primed MSCs. The authors showed that IL-1-primed MSC conditioned media added to inflamed microglial cells resulted in decreased secretion of inflammation markers (specifically, IL-6, G-CSF and TNF-α), and an increase in the microglial-derived anti-inflammatory mediator IL-10. These results highlight the ability of primed MSCs to orchestrate other cells to induce a more effective anti-inflammatory response, demonstrating the potential use of priming inflammatory treatments to enhance the beneficial actions of MSCs for future stroke therapies.

5. Directing MSC secretion using hypoxic culture conditions

A common element of tissue injury is the presence of hypoxia, a reduction in oxygen to levels of less than 5%. This reduction in oxygen tension leads to activation of several factors and chemoattractants (such as, stromal cell-derived factor 1, secreted by endogenous stromal cells), which cause MSCs to migrate to areas of hypoxia. Upon MSC migration, it has been demonstrated that production of various therapeutic paracrine mediators (i.e., VEGF, FGF-2, and IL-6 by MSCs) are increased.^[122,123] These in vivo phenomena can be recapitulated in vitro through the use of hypoxic culture conditions (<5% O₂). Many studies have demonstrated that hypoxic conditioning of MSCs results in secretion of various angiogenic (i.e., VEGF, FGF-2, HGF, and IGF-1) and anti-apoptotic factors (BCL-2 mediated) from MSCs isolated from various sources (bone marrow, adipose tissue, and placenta).^[110,124] Hypoxic culture conditions not only increase growth factor secretion from MSCs, but also promote MSCs to retain their stemness and an undifferentiated cell phenotype.^[125] Collectively, hypoxic conditioning of cultured MSCs may result in increased production and secretion of trophic factors and augmentation of angiogenic effects from the conditioned cells relative to normoxic (~21% O₂) culture conditioning.^[122]

Kim et al. demonstrated that MSCs cultured in hypoxic conditions of 3% O₂ for 5 days show enhanced stemness and immunomodulatory functions. Specifically, hypoxic-conditioned MSCs were resistant to passage-dependent senescence mediated by the monocyte chemoattractant protein-1 (MCP-1) and p53/p21 cascade, and secreted large amounts of proangiogenic and immunomodulatory factors, resulting in suppression of T cell proliferation in vitro.^[38] Administration of MSCs primed with hypoxia in a humanized rat model of graft-versus-host disease significantly augmented symptoms and improved survival outcomes.^[38]

In a study by Antebi et al., human and porcine bone marrow MSC functions were evaluated after short (48 hours) and long term (10 days) exposure to hypoxic environments.^[39] Specifically, MSCs were evaluated for their metabolic activity, proliferation, viability, clonogenicity, gene expression, and secretory capacity. The authors demonstrated that hypoxia augments the therapeutic characteristics of both porcine and human MSCs. Short-term (48 hours) hypoxia (2% O₂) offered the greatest benefit overall, exemplified by the increase in proliferation, self-renewing capacity, and modulation of key genes (i.e., VEGF, HMGB1 and NANOG) and the inflammatory milieu (i.e., IFN-γ and IL-18) as compared to normoxia (21% O₂). These results are important indications that hypoxic conditioning of MSCs augments cellular functions desired for clinical applications.

It is well known that cell-cell contacts can increase MSC survival and trophic factor secretion, as showcased by MSCs cultured in spheroids compared to dispersed cells (as discussed in Section 3.1). In addition to promoting cellular contacts, it is hypothesized that spheroids improve MSC secretion by the formation of a hypoxic core.^[126,127] However, in a study by Murphy et al., it was found that while a small (<10%) gradient of oxygen tension was observed in spheroids of approximately 350 μm, the enhanced function of MSC spheroids is not oxygen mediated at this size.^[126] In a follow up to this study, the same group investigated short-term hypoxic preconditioning of MSCs prior to spheroid formation to increase cell viability, proangiogenic potential, and resultant bone formation. Ho et al., exposed hMSCs in a monolayer either to 1% O₂ or ambient air for 3 days prior to spheroid formation of varying cell densities and encapsulation in alginate hydrogels. Hypoxia-preconditioned MSC spheroids were more resistant to apoptosis, secreted increased levels of VEGF compared to ambient air controls, and high cell density spheroids (15,000 cells) exhibited the greatest osteogenic potential in vitro.^[127] When hypoxia-preconditioned MSC spheroids in gels were transplanted into a rat critical-sized femoral segmental defect, increased bone healing was observed compared to gels containing preconditioned individual MSCs or acellular gels. These results demonstrate that hypoxic preconditioning enhances the therapeutic potential of MSC spheroids for tissue engineering applications. Further investigations of hypoxic environments on the MSC secretory profile are necessary to elucidate the underlying mechanisms responsible for the increased cell secretion.

6. Engineering precision hydrogels to direct MSC secretion

6.1. Where we have been and where we are going?

Over the last two decades, the increased understanding of MSCs has caused a shift in their therapeutic use. Once primarily focused on identifying ideal culture conditions to induce MSC differentiation, researchers are now transitioning towards investigating microenvironments that can direct the MSC secretome. Initial reliance on TCPS and mechanical signaling cues has given way to expanding and delivering MSCs in highly controlled 2D and 3D microenvironments containing multiple biophysical and biochemical stimuli. While MSC culture on TCPS provides a simple method for expansion and characterization, other studies have revealed that continuous culture on stiff substrates can bias MSCs towards an osteogenic fate, cause loss of multipotency, and decrease their stemness.^[26] In addition to identifying specific stiffness ranges, experimenters have identified specific transcription factors, genes, growth factors, and media cocktails to promote MSC commitment towards a particular lineage or pathway. These early investigations were essential for the advancement of MSC biology and identifying methods to harness the vast potential of MSCs for end use in a clinical environment.

Later, bioengineers and materials scientist began to use hydrogels as 2D and 3D culture platforms for MSCs, in some cases aiming to recapitulate the in vivo tissue environment. Hydrogel matrices allow control of the cell’s surrounding microenvironment, enabling end-users to tailor material properties such as matrix modulus, viscoelasticity, porosity, and degradability. Published work demonstrates the individual effects of the aforementioned hydrogel properties on MSC functions (see Sections 2–3). In relation to the MSC secretome, more recent studies have demonstrated the direct influence of matrix stiffness and viscoelasticity on global secretion. Additionally, hydrogel porosity and degradation have been used to direct MSC clustering and promote secretion by increasing cell-cell contacts. However, the underlying mechanisms responsible for increased MSC secretion due to specific hydrogel properties have yet to be elucidated.

To continue to understand the biophysical regulation of MSC secreted factors, sophisticated hydrogel culture platforms must be designed and employed. One critical material characteristic of the ECM found in many tissues is the viscoelasticity, or the ability to deform under constant applied strain. Viscoelastic properties in a material can direct MSC cell spreading, mechanotransduction, and cell fate.^[128] However, the effects of viscoelasticity on the MSC secretome has yet to be fully explored. Additionally, the influence of temporal changes in the matrix, both locally and macroscopically, on MSC properties should be more thoroughly investigated, and many dynamic biomaterials matrices could prove beneficial. Materials allowing for in situ control over stiffening, softening, or degradation should be employed to study how temporal changes in matrix modulus and degradability can influence the MSC secretory profile. In addition, platforms utilizing photo or chemical patterning techniques to create gradients of mechanical properties or of chemical signals may prove useful to optimize the mechanical and/or chemical dosages needed to achieve a specific MSC secretory profile. As MSC secreted factors are known to influence other cell types in vivo, the development of more sophisticated in vitro co-culture platforms that allow control of the spatial proximity of cells and control of cell-cell interactions will provide insight as to how secreted factors influence cellular crosstalk. Finally, culturing methods to test the influence of other material properties known to direct cellular behavior, such as topography, surface charge, and surface roughness, on MSC secretion should be investigated. By thoroughly understanding how specific material properties and matrix interactions influence MSC secretion, researchers will be able to design improved expansion and delivery platforms capable of improving clinical translation of MSC-based therapies.

Furthermore, MSC responses to biophysical cues are not limited to mechanical cues alone. MSCs have been shown to respond to electrical stimulation resulting in changes in cell morphology and fate.^[129,130] Emloyed use of conductive substrates with hydrogels,^[131] conductive composites,^[132] and conductive polymers^[133] for MSC culture has shown great promise for applications in cardiac, bone, and neural tissue engineering. While the effects of mechanical and electrical stimulation on the MSC secretome has not been explored, combinatorial approaches must be utilized to elucidate cellular responses to muliple biophysical cues to recapitulate the in vivo environment.

Another advantage of using hydrogels as MSC culture environments is the ease of systematic incorporation of biochemical compounds (i.e., small molecules, peptides, growth factors, cytokines, etc.) into the material system.^[5,43,134] Bio-orthogonal conjugation methods and/or traditional adsorption and affinity methods allow MSCs to receive both biophysical and biochemical stimuli simultaneously. Receptor dynamics, local concentrations, etc. can be highly dependent on the matrix environment. Previous studies have investigated the influence of various bioactive factors such as, cytokines (i.e., IFN-γ and TNF-α), growth factors (i.e., BMP-2, TGF-β1, and VEGF), and small molecules (i.e., estradiol, purmorphamine, and kartogenin) on MSC cell fate and secretion profile. In regard to influencing MSC secretion, cell priming with biochemical cues (e.g., pro-inflammatory) can be advantageous to induce a specific secretory profile (e.g., anti-inflammatory) or increase secretion of a specific factor (e.g., IL-10). However, the altered secretion profile is often short-lived; once the appropriate cue is removed, MSCs can revert to their steady state profile. To overcome such challenges, a combinatorial approach requiring the use of hydrogels with biochemical treatments may promote sustainable alterations to the MSC secretory profile (Figure 5).

Figure 5. Design considerations for the next generation of hydrogels to direct the MSC secretome.

Prior to hydrogel use, MSC donor characteristics, such as, gender, age, health, and genetics, must be evaluated. A hydrogel platform with specific material properties can then be utilized for MSC culture. Material characteristics should be chosen based on whether the hydrogel will be injected or implanted, in addition to how these properties will affect the MSC secretome. Furthermore, various biochemical compounds can be incorporated into the material system. Specific release profiles of bioactive factors can be obtained depending on how the molecules are incorporated into the hydrogel system.

Methods, including nanoparticle delivery, bulk adsorption, or direct tethering of bioactive factors into hydrogels, could be used to direct the secretion of encapsulated MSCs for longer time scales than what can be achieved with traditional priming methods. Several molecular signaling pathways have been implicated in governing MSC secretory activity, including NF-kB, Smad, VEGF, and Beta-catenin. Small molecules, genetic components, proteins, or peptide fragments could be used to either inhibit or enhance specific pathways of interest. Inclusion of biochemical compounds by matrix-immobilization (i.e., direct tethering of a molecule or loaded nanoparticle) to hydrogels would enable sustained presentation to MSCs, whether internalized or through receptor binding. A combination of drug loading strategies, such as bulk adsorption and molecule conjugation, could be used to deliver secretory enhancing molecules sequentially or in parallel and affect not only MSCs within the hydrogel, but surrounding endogenous cells once delivered in vivo. The combination of biochemical signals with material design strategies has the potential to improve both the innate and altered MSC secretions.

While the focus of this perspective has been on measuring final outputs, i.e., proteins that MSCs are secreting, more work is needed to further understand the molecular pathways dictating the composition and concentration of the MSC secretome. Multiple studies have shown that material stiffness influences MSC secretion, which begs the question: are the often-studied rheostats governing MSC fates, specifically YAP and TAZ, involved in regulating their secretory properties? Or is it a different mechanotransduction mechanism all together? Are there explicit stiffness ranges, ECM ligands, viscoelastic time scales, or cell cluster sizes that reliably lead to specific secretory profiles, as is observed with the specific material properties needed to direct MSC proliferation, migration, and differentiation? Additionally, multiple research groups have determined that N-cadherins are key regulators of MSC paracrine activity. However, further studies are needed to identify the downstream regulators of N-cadherin signaling that lead to improved secretory outcomes. Furthermore, are other pathways capable of facilitating cell-cell interactions, such as E-cadherins dimerization or gap junctions? This warrants further investigation pertaining to their potential involvement in regulating MSC secretion. Understanding the synergistic effects of mechanical and biochemical cues on the MSC secretome will be key in the rational design of the next generation of hydrogel platforms.

6.2. Importance of secretome measurement techniques

When investigating the secretome, it is important to note the differences in methods for detection. Proper characterization of the MSC secretome will aid in clinical utility once a specific profile is obtained. The identification of protein components is commonly obtained using two proteomic approaches: immunological based or shot-gun based (Figure 6).

Figure 6. Proteomic approaches used to detect and quantify MSC secretory components.

Shot-gun based proteomics (primarily mass spectrometry based) are used for the detection of unknown or unique proteins. Once these proteins are identified, publicly available databases and bioinformatics can be used for pathway analyses to determine specific protein roles (extensive analyses). Immunological-based assays are used for testing a broad range of known proteins. These assays are often user-friendly involving minimal sample preparation and analysis methods resulting in quantitative or semi-quantitative outcomes.

Immunological-based assays offer high specificity, sensitivity, and reproducibility. These aspects are critical as secreted proteins are often present at low levels in media, ranging from picograms to nanograms. Examples of immunological-based assays include: enzyme-linked immunosorbent assays (ELISA), multi-plex antibody bead-based assays, microarrays, western blotting, and cytokine antibody arrays. While these assays are highly specific, they are limited to the detection of known proteins. However, while some of these systems, such as cytokine arrays, are advantageous to broadly characterize the differences between experimental conditions, they often do not provide the level of quantification needed to compare results between research groups.

The shot-gun based proteomics approach is more exploratory and used for identification of unknown or unique secreted proteins. Examples of such techniques include: liquid chromatography with tandem mass spectrometry, matrix-assisted laser desorption/ionization-time of flight, quadrupole time-of-flight mass spectrometry, and 2D gel electrophoresis. Once specific proteins are identified, publicly available databases and bioinformatics tools can be used for pathway analyses to determine protein roles of biological and clinical relevance. Compared to immunological-based assays, shot-gun based proteomics approaches require more specialized instrumentation and intensive downstream data processing.

6.3. Precision hydrogels for MSC-based therapies

To move towards the translation of MSC-based therapies, it is important to realize a “one-size-fits-all” approach is clinically outdated. Donor/recipient characteristics such as gender, age, genetics, and overall health status (e.g., severity of disease) should be considered when developing material strategies for clinical translation. From this perspective, the evolution of ‘precision biomaterials’, tailored to deliver MSCs at the right dose with the right secretome for the tissue or disease, might be envisioned for personalized cell therapies.^[135] The high tunability and versatility of hydrogels should enable progress towards this goal. To develop personalized MSC therapies, hydrogels could incorporate customized material chemistries, bioactive components, and patient-specific factors to promote tissue regeneration to treat disease. Once delivered, these hydrogel-based MSC therapies should acclimate to the patient’s diseased microenvironment. For example, a change in the temperature, pH, or concentration of a cytokine in vivo might trigger release of a drug to stimulate MSC trophic factor secretion. In addition, in cases of autologous transplantation, precision hydrogel use during MSC expansion and delivery may have the ability to correct an undesired patient specific MSC secretory profile.

Overall, MSCs represent a cell population that can serve as an incredible source of bioactive factors. By exerting control over their microenvironment, the field is beginning to unravel the influence of both biophysical material properties and biochemical signals on the composition and concentration of MSC trophic factor secretion. However, challenges remain with respect to the clinical translation of MSC-based therapies. MSCs lose many of their regenerative properties, such as stemness and secretion, during expansion on TCPS. Once delivered via injection, MSCs have low survival, retention, and engraftment rates in vivo. Additionally, the MSC secretion profile is rarely, intentionally directed before their clinical use. To address these challenges, we propose a precision biomaterial-based approach utilizing hydrogels with optimized biophysical properties, in combination with biochemical compound incorporation, to enhance and direct the MSC secretory profile. Once the ideal material design parameters are fully elucidated, this next generation of hydrogel based MSC-therapies will have immense clinical potential.