Growth Factor and Cytokine Delivery Systems for Wound Healing

Abstract

Skin wound healing is a highly coordinated process involving multiple tissue-resident and recruited cell types. Cells within the wound microenvironment respond to key secreted factors such as pro-proliferative growth factors and immunomodulatory cytokines to repair the skin and promptly restore its essential barrier role. Therefore, recombinant growth factors and cytokines are promising therapeutics for skin wounds, in particular for large acute wounds such as burns, or wounds associated with underlying pathologies such as nonhealing chronic and diabetic wounds. However, translation of growth factors and cytokines into clinically effective treatments has been limited. Short half-life, poor stability, rapid diffusion, uncontrolled signaling, and systemic side effects are currently the key challenges to developing efficient growth factor- and cytokine-based therapies. To overcome these limitations, novel delivery systems have been developed to improve the regenerative potential of recombinant growth factors and cytokines. In this review, we discuss biomaterial and protein engineering strategies used to optimize the delivery of growth factor and cytokine therapeutics for skin wound treatment.

Timely and efficient wound healing is paramount to restoring the skin's vital barrier function following injury. In vertebrates, acute wounding initiates a complex, highly coordinated response that is generally categorized into the four overlapping phases of hemostasis, inflammation, proliferation, and tissue remodeling. Hemostasis occurs within seconds to restrict blood loss and is mediated by vasoconstriction and coagulation (Gurtner et al. 2008; Eming et al. 2021). The inflammatory phase begins within minutes of injury, lasting several days, and involves dynamic recruitment of multiple immune cell types over the course of healing (Larouche et al. 2018). As inflammation begins to resolve, the proliferative phase is entered where dermal fibroblasts and epidermal cells proliferate to replace lost tissue, while angiogenesis occurs to support the wound microenvironment. Following wound closure, the repaired tissue continues to undergo remodeling, where extracellular matrix (ECM) and newly formed vascular networks are reshaped over a period of several months to ultimately improve tissue integrity (Rodrigues et al. 2019). Despite these sophisticated physiological mechanisms, additional therapeutic intervention may be required, for example, in human patients suffering severe burns or where underlying pathologies result in chronic, nonhealing wounds.

Burns and chronic wounds cause considerable physical and psychological impacts to patients, such as scarring, infection, amputation, and depression, with increased long-term rates of mortality reported for these individuals (Sen et al. 2009; Jeschke et al. 2020). In addition, burns and chronic wounds represent a significant and increasing economic burden worldwide, with more than US$25 billion per year estimated to be spent on wound treatment in the United States alone (Sen et al. 2009). Accordingly, substantial efforts are being made to develop novel therapeutics for the treatment of cutaneous wounds.

Given their comprehensively studied role in tissue development and homeostasis, growth factors are a rational candidate for promoting proliferation and wound closure in skin injuries. Similarly, given that excess inflammation is a significant contributor to the pathophysiology of chronic wounds, treatment with immunomodulatory cytokines is an attractive therapeutic strategy. Indeed, biologics such as recombinant growth factors and cytokines have been developed and explored for their potential therapeutic benefit in wound healing; however, these have often met with limited clinical success (Nurkesh et al. 2020; Ren et al. 2020).

To date, the only recombinant growth factor approved by the U.S. Food and Drug Administration (FDA) for chronic wound treatment is platelet-derived growth factor (PDGF)-BB, with its use indicated for patients suffering from diabetic neuropathic ulcers (Baldo 2014). Other treatments based on basic fibroblast growth factor (FGF-2), epidermal growth factor (EGF), granulocyte macrophage colony-stimulating factor (GM-CSF), and vascular endothelial growth factor A (VEGF-A), among others, have been approved for use in chronic wounds or burn injury treatment in jurisdictions outside the United States, are currently used in the clinic as off-label treatments, or are undergoing clinical trials (Barrientos et al. 2014; Ren et al. 2020). Collectively, these treatments aim to promote wound closure by stimulating proliferation of dermal and epidermal cells, suppressing inflammation, and promoting angiogenesis to restore oxygen supply and nutrient delivery to the wound microenvironment (Barrientos et al. 2014).

Despite their initial promise, the use of recombinant growth factor and cytokine treatments for wound therapy has encountered significant limitations that have since precluded their widespread adoption in the clinic. High levels of endogenous proteolytic activity in skin wounds, often coupled with pathological inflammation in wounds requiring additional therapeutic intervention, ultimately lead to impaired stability and short half-life for most delivered biologics. In addition, locally delivered growth factors or cytokines may diffuse rapidly from injury sites, making it difficult to achieve therapeutically relevant concentrations within the wound environment and increasing the risk of undesired systemic effects (Mitchell et al. 2016; Ren et al. 2020). Consequently, supraphysiological and repeated doses are often required to achieve significant improvements in wound healing, leading to prohibitive treatment costs and significant adverse effects (Ren et al. 2020; Berry-Kilgour et al. 2021). For example, despite being effective for the treatment of diabetic ulcers, the use of recombinant human PDGF-BB (commonly known as becaplermin or Regranex) has been associated with an increased risk of malignancy (Papanas and Maltezos 2010). Moreover, the immunogenicity of recombinant protein-based therapeutics has been acknowledged as an important concern, with resulting side effects such as autoimmunity, myelosuppression, and neurological consequences exacerbated by repeated and/or high doses (Tovey and Lallemand 2011; Baldo 2014; Sauna et al. 2018). Besides adverse effects, a number of biologics including recombinant forms of VEGF-A, interleukin 10 (IL-10), and transforming growth factor β3 (TGF-β3) have failed to progress through clinical trials as wound therapies after being unable to demonstrate significant efficacy (Berry-Kilgour et al. 2021).

These shortcomings in recombinant protein therapeutics underscore the need to develop effective delivery systems that increase their efficacy and safety. This can be achieved by improving protein stability, restricting activity to the wound environment, and increasing potency by augmenting signaling activity, while minimizing side effects, doses, and costs (Fig. 1). Here, we provide a review of the biomaterial- and protein engineering-based strategies that have been developed in an effort to address the current limitations in cytokine and growth factor therapeutics, within the context of skin wound healing. In particular, we discuss methods used for the spatial and temporal control of delivery, strategies to improve half-life, and novel techniques to modulate the signaling of therapeutic growth factors and cytokines.

Figure 1.

Key principles to optimize the delivery of recombinant growth factors and cytokines to wounds. The translation of recombinant growth factors and cytokines into effective therapeutics for skin wounds is limited in the absence of suitable delivery systems. Poor delivery systems require high doses of recombinant protein, have minimal ability to control their release, lead to rapid degradation, do not maintain effective signaling, and ultimately leads to increased side effects and poor efficacy (top panel). Optimal delivery systems are able to minimize the required dose of recombinant protein by restricting their release within the wound environment, increasing their half-life, and maintaining an optimal signaling activity. This has demonstrated improved wound healing outcomes in animal models with minimal adverse effects (bottom panel).

BIOMATERIAL ENGINEERING FOR CONTROLLED GROWTH FACTOR AND CYTOKINE DELIVERY

To date, growth factor and cytokine delivery for wound therapy has been predominantly achieved by topical application or local injection of unmodified recombinant proteins. The simplest delivery strategies include topical sprays applied directly to wounded skin, with several recombinant proteins trialed using this method for the treatment of burns, pressure ulcers, and diabetic foot ulcers (Berry-Kilgour et al. 2021). These treatments deliver human recombinant growth factors such as FGF-2, EGF, or PDGF-BB in antimicrobial or inert vehicle solutions at amounts ranging from 1 to 10 µg per cm² of wound area, often using repeat doses over the course of several weeks (Robson et al. 1992; Mustoe et al. 1994; Akita et al. 2008; Park et al. 2018). Intradermal, subcutaneous, or intralesional injections have also been used as simple delivery methods for recombinant growth factors (e.g., TGF-β3, GM-CSF, EGF) or cytokines (e.g., IL-10) for treatment of chronic wounds or surgical scarring in humans, with treatment regimens varying widely (Da Costa et al. 1999; Ferguson et al. 2009; Kieran et al. 2013; Gomez-Villa et al. 2014). Treatments using these delivery methods have found very limited success in clinical trials, with the majority failing to demonstrate efficacy for the improvement of chronic wound healing or scar appearance (Berry-Kilgour et al. 2021). In addition to patient and wound diversity, study design, or patient compliance, the intrinsic properties of growth factors and cytokines pose considerable challenges for their translation into effective wound therapies. Given their short half-life and propensity to rapidly diffuse away from wound sites, recombinant growth factor and cytokine delivery using biomaterials has been adopted in an effort to overcome these limitations (Mitchell et al. 2016; Nurkesh et al. 2020).

Entrapment and Absorption of Growth Factors and Cytokines within Biomaterial Scaffolds

Among the most common categories of biomaterial explored to date for growth factor and cytokine delivery in skin wounds are decellularized ECM scaffolds, protein- or polysaccharide-based sponges, and hydrogels (Fig. 2A; Berry-Kilgour et al. 2021). Whereas decellularized ECM- and sponge-based systems have shown promising results for therapeutic protein delivery to skin wounds, the potential to tailor their physicochemical characteristics to suit various biologics and wound environments is perhaps limited (Saghazadeh et al. 2018). Hydrogels are a particularly attractive delivery system in this context given their ability to conform to the shape of the wound and mimic ECM, while allowing fluid exchange for hydration and drug delivery (Kharkar et al. 2013; Park et al. 2017; Saghazadeh et al. 2018). Furthermore, given that wound hydration is known to play an important role in healing outcome (Ousey et al. 2016), the high content of water within hydrogels presents an advantage over other scaffold-based delivery systems. These characteristics establish hydrogels as an ideal biomaterial for the treatment of skin wounds.

Figure 2.

Examples of strategies to improve growth factor and cytokine delivery to wounds. (A) Commonly used biomaterial systems used to deliver growth factors and cytokines to skin wounds. (B) Functionalization strategies for biomaterials or endogenous extracellular matrix (ECM) to control growth factor and cytokine retention and release. (C) Protein engineering strategies to improve growth factor and cytokine stability. (D) Approaches to modulate growth factor and cytokine signaling.

A multitude of hydrogel systems have been developed to achieve spatial and temporal control over therapeutic protein release, particularly where prolonged delivery is required (e.g., chronic wounds). Derived from natural or synthetic polymers, hydrogels are cross-linked, hydrophilic, and highly biocompatible materials that can be readily loaded with recombinant proteins and applied directly to skin wounds (Kharkar et al. 2013). The most commonly used polymers for hydrogel preparation are natural and include polysaccharides such as carboxymethylcellulose, chitosan, hyaluronic acid and alginate, as well as proteins such as collagen, fibrin, and gelatin (Park et al. 2017; Berry-Kilgour et al. 2021). Although hydrogels based on these polymers have shown some success in delivering biologics to improve wound healing both in animal models and human trials, there is limited capacity to control their physical, chemical, and mechanical properties, which are largely dictated by the type of polymer used (Uebersax et al. 2009; Berry-Kilgour et al. 2021). Synthetic polymers such as polyethylene glycol (PEG) and poly(lactic co-glycolic acid) (PLGA) have been widely adopted for hydrogel preparation given their readily modifiable properties and biocompatibility (Aswathy et al. 2020; Pan et al. 2021). In addition, novel zwitterionic hydrogels composed of sulfated poly(sulfobetaine methacrylate) have shown promise for delivering growth factors to promote wound healing in murine models compared to more established hydrogel formulations (Wu et al. 2018; Xiao et al. 2021).

Although used for small molecule delivery, other novel hydrogel systems have been developed for wound therapy that may be adapted for growth factor and cytokine delivery. For example, an injectable hydrogel prepared using multi-arm thiolated PEG cross-linked with silver nitrate was shown to have antibacterial properties and was used to deliver a proangiogenic drug, deferoxamine, to diabetic rat skin wounds where significantly improved healing outcomes were observed. Notably, these hydrogels were self-healing and able to resist external mechanical forces while within the wound, suggesting they may be of particular use where extended treatment periods are required (Chen et al. 2019). Further examples of novel hydrogels for wound therapy include “smart” hydrogels with pH-, enzyme-, and temperature-responsive characteristics that are able to adapt to changes in wound infection and inflammation (Abbasi et al. 2020; Guan et al. 2020; Preman et al. 2020; Hu et al. 2021). The responsive characteristics of smart hydrogels can be exploited to enable sequential release of appropriate factors to modulate the distinct phases of wound healing (Saghazadeh et al. 2018; Wang et al. 2021), making them ideal delivery systems for pathological wound environments.

To enable greater control of therapeutic protein release via modulation of hydrogel degradation rate, several systems have been developed using combinations of natural and/or synthetic polymers and various cross-linking methods (Lu et al. 2018; Op ‘t Veld et al. 2020). Alternatively, hydrogel degradation rate can be manipulated by simply varying the ratio of constituent polymers and cross-linking times. This has been demonstrated in an in vivo photocross-linkable hydrogel formulation consisting of the synthetic polymer Pluronic F127 and chitosan-derived chitooligosaccharide that was used to deliver recombinant human EGF to murine skin wounds. In this hydrogel system, higher concentrations of chitooligosaccharide and shorter cross-linking times were found to increase the rate of EGF release (Choi and Yoo 2010).

Instead of manipulating cross-linking, control of hydrogel degradation can be achieved by functionalizing polymers with protease-sensitive peptides to facilitate their degradation within wounds and subsequently release loaded proteins. This strategy has been used to cross-link matrix metalloprotease (MMP) substrate sequences within PEG hydrogels and control growth factor delivery within skin wounds (Ehrbar et al. 2007; Martino et al. 2013). Hydrogel degradation rate can also be controlled by incorporating protease inhibitors. For instance, covalent cross-linking of the fibrinolysis inhibitor, aprotinin, into fibrin hydrogels was shown to inhibit degradation in a concentration-dependent manner. This results in a highly tunable system to control hydrogel degradation rate and, thus, growth factor release (Sacchi et al. 2014).

The wide range of polymers and cross-linking methods available for hydrogel preparation offers myriad possibilities for the design of growth factor and cytokine delivery systems and enables adaptability for various wound environments. Despite their utility, greater control of delivery may be required for challenging pathological wound settings, which can be achieved by hydrogel functionalization (discussed later herein).

Particle-Based Delivery Systems

Encapsulation of protein therapeutics within particle-based systems such as liposomes, microspheres, and other biocompatible nanoparticles is a highly effective alternative, or complement, to hydrogel-based delivery methods. For growth factor and cytokine delivery, these systems may offer higher loading capacities, improved stability, environmental responsiveness, and greater control of release compared to nonfunctionalized hydrogel-only systems (Park et al. 2017; Wang et al. 2017; Gonçalves et al. 2020).

Liposomes are amphiphilic lipid bilayer vesicles that can be used to encapsulate biologics within their aqueous core, while their hydrophobic exterior provides protection for these proteins against enzymatic degradation and adverse conditions within the wound environment (e.g., pH and temperature variations) (Wang et al. 2017). Liposomes have been used to deliver growth factors and cytokines such as FGF-2, stromal cell-derived factor 1 (SDF-1; CXCL12), and EGF topically to murine burn and diabetic wounds, with significant improvements in healing outcome observed compared to the delivery of free growth factor solutions (Alemdaroğlu et al. 2008; Xiang et al. 2011; Olekson et al. 2015). These improvements are attributed to attenuated degradation and prolonged release of growth factors facilitated by liposome encapsulation (Alemdaroğlu et al. 2008; Xiang et al. 2011; Olekson et al. 2015). Although demonstrating important advantages over delivery of free growth factors, liposomes are more prone to degradation and burst release kinetics compared to other particle-based systems, given their lipid composition (Park et al. 2017; Wang et al. 2017). Consequently, to increase stability and modulate release kinetics further, liposomes have been engineered to contain hydrogel cores or have additional polysaccharide coatings (Xu et al. 2017; Parchen et al. 2020).

In contrast to liposomes, microspheres are produced using natural or synthetic polymers or proteins including alginate, PLGA, polycaprolactone (PCL), and gelatin (Dong et al. 2008; Liu et al. 2017; Turner et al. 2017; Shamloo et al. 2018). Microspheres encapsulating growth factors have been successfully demonstrated to improve wound healing in both diabetic and nondiabetic mouse models, with steady and sustained release of growth factors being a typical feature of these systems (Dong et al. 2008; Dogan et al. 2009; Huang et al. 2016; Liu et al. 2017; Shamloo et al. 2018). Microspheres can be delivered to wounds directly in suspension or after incorporation into hydrogels, providing an even greater opportunity to control release by also exploiting hydrogel characteristics.

Nanoparticle delivery systems based on materials such as lipids, natural polysaccharides, synthetic polymers, and metals have been developed to encapsulate or conjugate growth factors and cytokines (Ezhilarasu et al. 2020; Gonçalves et al. 2020). Lipid nanoparticles, including solid lipid nanoparticles (SLNs) and nanostructured lipid carriers (NLCs), demonstrate high stability in vivo and have been used to treat diabetic mouse skin wounds with human EGF (Gainza et al. 2014; Wang et al. 2017). Significantly improved closure and reepithelialization was observed in wounds treated with EGF-loaded SLNs and NLCs compared to those treated with free EGF. Notably, this study also compared lipid nanoparticles to PLGA-alginate microspheres and observed wound healing effects for SLNs and NLCs loaded with 10–20 µg of EGF to be comparable to microspheres loaded with 75 µg of EGF (Gainza et al. 2014). A later study used human EGF-loaded NLCs to treat porcine full-thickness excisional wounds and showed accelerated healing compared to control (Gainza et al. 2015).

Nonlipid nanoparticles used to deliver encapsulated biologics to skin wounds include PLGA- and chitosan-based nanoparticles (Chu et al. 2010; Parajó et al. 2010; Losi et al. 2013). Delivery of single or dual growth factor-encapsulating nanoparticles alone or as hybrid systems within scaffolds has been reported to significantly improve wound healing in nondiabetic and diabetic murine models (Chu et al. 2010; Losi et al. 2013; Tanha et al. 2017). In contrast to encapsulation, growth factors have been conjugated to the surface of gold nanoparticles (AuNPs) for delivery to diabetic wounds in mice (Li et al. 2019; Wei et al. 2019). AuNPs are readily modifiable and capable of loading multiple factors, thus presenting a highly customizable delivery system (Liu et al. 2015; Liu and Peng 2017). For example, AuNPs conjugated with both VEGF-A and an antimicrobial compound were shown to significantly improve the healing of murine diabetic wounds infected with methicillin-resistant Staphylococcus aureus (Wei et al. 2019).

Whether delivered alone or as a hybrid system with other biomaterials, particle-based delivery systems offer remarkable flexibility for the development of skin wound therapeutics. Careful consideration of material selection is required when using these delivery systems to achieve the appropriate balance between loading efficiency and release kinetics, which is largely dictated by biodegradability within the wound environment. Particle aggregation (e.g., liposome aggregation) may also be a limitation of these systems, affecting the distribution and release of delivered proteins (Wang et al. 2017). Although showing promising results for growth factor delivery in murine and larger animal skin wound models, to our knowledge, human trials to treat wounds using these delivery systems have not been conducted to date. In addition, although used in other contexts, there has been limited investigation of immunomodulatory cytokine delivery for skin wound healing using these technologies (Gonçalves et al. 2020). Nevertheless, these systems are substantial improvements on the delivery of free biologics and show considerable potential to be effective therapeutics for human wound therapy. Notably, virus- and lipid-based particle systems have also been used to deliver DNA/mRNA encoding growth factors as gene therapy for wound healing; however, these systems are beyond the scope of this review and have been comprehensively discussed elsewhere (Branski et al. 2007; Eming et al. 2007; Laiva et al. 2018).

Microneedles as Delivery Systems for Hypertrophic Scars and Keloids

In contrast to slow or nonhealing wounds, aberrant repair can also result in wounds that are “over-healed.” Although scarring is an inevitable consequence of most skin wounds occurring in postnatal mammals, excessive fibroblast proliferation and ECM deposition can result in pathological scars such as hypertrophic scars and keloids. Hypertrophic scars can occur following deep wounds such as surgical incisions or burns and are restricted to the original injury site, whereas keloids can occur following very minor wounds and generally extend beyond the initial injury margins (Gauglitz et al. 2011).

Microneedles have gained attention as a promising, minimally invasive treatment for pathological scarring. Microneedles are a transdermal delivery system composed of an array of needles commonly measuring several hundred microns in length and fabricated using various materials including naturally occurring polysaccharides and synthetic polymers (Lee et al. 2008; Yeo et al. 2017; Lin et al. 2019). For pathological scar treatment, microneedle arrays are applied topically as a patch and, depending on their composition, can be used to simply create microchannels within the scar tissue or to also deliver therapeutic agents (Lee et al. 2008; Yeo et al. 2017; Lin et al. 2019; Waghule et al. 2019; Juhasz and Cohen 2020; Park and Kim 2021). “Drug-free” microneedling has been suggested to improve scar characteristics by disrupting collagen and reinitiating the wound healing and remodeling process, with microneedle patches also suggested to prevent and effectively treat hypertrophic scars (Yeo et al. 2017; Alster and Li 2020). However, the current evidence supporting drug-free microneedles as an effective therapeutic for pathological scarring is limited as a result of varying treatment regimens and small sample sizes; therefore, additional investigation is required to confirm their efficacy. In contrast, microneedle patches as a delivery system for therapeutics have shown promising results in pathological scar treatment.

Dissolvable microneedle arrays have been demonstrated to effectively deliver single or combinations of therapeutics to hypertrophic scars. For example, delivery of the corticosteroid, triamcinolone acetonide, using cyclodextrin/hyaluronic acid-based microneedle patches was shown to reduce hypertrophic scar elevation, as well as levels of collagen and profibrotic TGF-β1 within the scar (Lin et al. 2019). In another study, a bilayer dissolvable microneedle array was used to deliver triamcinolone acetonide and 5-fluorouracil, with unique polymer compositions used within each layer to control the rate of release of each therapeutic. These patches were shown to significantly reduce hypertrophic scar elevation, fibroblast proliferation, and collagen deposition (Yang et al. 2021). Nondissolvable microneedle patches have also been used to deliver chitosan-based nanoparticles loaded with 5-fluorouracil to human keloid fibroblasts in vitro, where they were shown to inhibit their proliferation; however, the effectiveness of this combination delivery system in vivo requires additional investigation (Park and Kim 2021).

Given their ability to effectively deliver therapeutics within pathological scars and the capacity to manipulate their physicochemical characteristics for controlled release, the use of microneedles as delivery systems for recombinant growth factors and cytokines may be an effective treatment strategy in this context (Fig. 2A). For example, recombinant IL-10 and TGF-β3 have been shown in previous studies to both reduce scar formation following intradermal injection at surgical incision sites and improve scar appearance after their formation (Ferguson et al. 2009; So et al. 2011; Kieran et al. 2013, 2014). As described earlier, these biologics were unable to progress through clinical trials due to limited efficacy; however, the use of improved delivery systems such as microneedles may be an effective method to augment their availability and activity within pathological scars.

Additional advantages of microneedle patches include minimal pain upon application and the ability for patients to self-administer treatments; however, given their limited size, consideration should be given to the amount of protein that can be effectively loaded and delivered to scars that are often relatively large in size (Vora et al. 2021). Overall, microneedles as delivery systems for biologics warrant further investigation in the context of pathological wound healing and scar treatment.

Biomaterial Functionalization with Growth Factor and Cytokine-Binding Motifs

Hydrogels are valuable delivery systems for local application of biologics such as growth factors and cytokines. However, due to the high water content nature of hydrogels, the release of these therapeutics by diffusion is usually very rapid when in contact with interstitial fluids. To achieve greater control over delivery and ultimately signaling, growth factors and cytokines have been immobilized within hydrogels functionalized with functional groups or motifs that can attach and retain these therapeutics within hydrogel matrices (Fig. 2B; Briquez et al. 2015; Abune et al. 2021).

Many growth factors and cytokines that have “heparin-binding” proprieties naturally bind a family of ECM proteins called proteoglycans. These highly glycosylated proteins contain glycosaminoglycans, which are carbohydrate polymers with a negative charge that can store growth factors and cytokines within the ECM. Therefore, hydrogels functionalized with glycosaminoglycan-like molecules such as heparin and heparan sulfate have been used extensively as delivery systems for growth factors (Sakiyama-Elbert 2014; Briquez et al. 2015; Zhang et al. 2019). For example, synthetic hydrogels composed of covalently cross-linked chondroitin sulfate and PEG diacrylate with or without heparin were used to deliver human FGF-2 into full-thickness skin wounds of diabetic mice (Liu et al. 2007). Earlier in vitro characterization of these hydrogels found that in the absence of heparin, most of the loaded FGF-2 was released after 35 days, whereas hydrogels containing 5% (w/w) heparin released ∼50% of loaded FGF-2 (Cai et al. 2005). Using the same hydrogel formulations, the degree of wound closure at 2 wk postinjury was found to be significantly increased in wounds treated with FGF-2-loaded hydrogels containing heparin compared to those treated with nonheparinized hydrogels also loaded with FGF-2 (Liu et al. 2007). In a separate study, the “pro-healing” cytokine IL-4 was incorporated into hydrogels composed of star-shaped PEG (starPEG) functionalized with heparin. The addition of heparin was shown to significantly retain IL-4 within starPEG hydrogels, with ∼11% released after 24 h compared to over 62% released by nonheparin-functionalized hydrogels in the same period. Heparin-functionalized starPEG hydrogels continued to steadily release IL-4 for at least 14 days in vitro and improved the thermal stability of IL-4 while inhibiting its proteolytic degradation. Furthermore, IL-4 delivered using this system retained its ability to polarize macrophages to the pro-repair/anti-inflammatory “M2” state, demonstrating potential for treating chronic wounds (Schirmer et al. 2016). Although only investigated using in vitro models, heparinized starPEG hydrogels have also been used to deliver TGF-β1 to human dermal fibroblasts (Watarai et al. 2015), and VEGF-A165 and FGF-2 to human endothelial cells (Zieris et al. 2010), both of which are relevant models for wound healing. These studies demonstrated the capacity for this delivery system to enable controlled release of growth factors in a readily tunable manner.

The release of growth factors from heparin-functionalized hydrogels can be controlled even further by manipulating ratios of polymer, cross-linkers, and heparin, or through chemical modification of polymers or heparin (Nie et al. 2007). For example, the release of VEGF-A165 from heparinized PEG-based hydrogels could be modulated through selective removal of sulfate groups from heparin, with VEGF-A165 release demonstrated to be inversely related to the level of heparin sulfation. Using low sulfate heparin, functionalized VEGF-A165-loaded hydrogels were shown to improve granulation tissue formation and angiogenesis in diabetic mouse skin wounds compared to VEGF-A165-loaded hydrogels containing highly sulfated heparin (Freudenberg et al. 2015).

Although the incorporation of heparin within hydrogels has demonstrated significant advantages compared to nonheparinized systems, its derivation from animal sources and anticoagulant activity may present as limitations to widespread clinical use (He et al. 2019). Therefore, alternative methods of functionalization have also been explored.

Besides heparin-like molecules, hydrogels have been functionalized with ECM protein domains or sequences that have the ability to bind various growth factors and cytokines. For instance, ECM proteins such as fibronectin, vitronectin, tenascin C, fibrinogen, and laminin naturally regulate growth factor and cytokine activity as well as presentation (Briquez et al. 2015). Domains present within these ECM proteins have been shown to bind growth factors promiscuously, as well as a subset of cytokines (Martino and Hubbell 2010; Martino et al. 2013; Tortelli et al. 2013; Ishihara et al. 2018). For example, several growth factors belonging to the FGF, TGF-β, PDGF, and VEGF families have been demonstrated to bind with strong affinity to the 12th–14th type III repeats of the ECM protein fibronectin (FN III12-14) (Martino and Hubbell 2010). In light of these findings, a fibrin-based hydrogel containing covalently linked FN III12-14 fused with the integrin-binding 9th–10th type III repeats of fibronectin (FN III9-10) was developed to achieve both controlled release of growth factors and synergistic signaling with integrins (discussed later herein). Treatment of diabetic mouse wounds with growth factors delivered via the fibronectin domain-functionalized hydrogel resulted in significantly faster wound closure and increased granulation tissue formation compared to all other conditions, including fibrin-only hydrogels loaded with growth factors (Martino et al. 2011). Similarly, another study found that several growth factors including FGF-2 and placenta growth factor (PlGF)-2 are able to bind the heparin-binding domain (HBD) in the β chain of fibrinogen (Fg β15–66) (Martino et al. 2013). In this study, Fg β15–66 was dimerized to mimic its in vivo configuration and incorporated into a synthetic fibrin-mimetic hydrogel based on PEG. When loaded with FGF-2 and PlGF-2, Fg β15–66-functionalized fibrin-mimetic hydrogels displayed more gradual release of growth factors and were able to significantly improve wound healing in diabetic mice compared to nonfunctionalized PEG hydrogels (Martino et al. 2013). In addition to growth factor binding, fibronectin and fibrinogen have been shown to bind to several cytokines with varying affinity (Tortelli et al. 2013). In particular, C-X-C motif chemokine 11 (CXCL11) has been described as important for skin wound healing in mice and binds with relatively high affinity to fibronectin and fibrinogen (Yates et al. 2008, 2009; Tortelli et al. 2013). When delivered within a fibrin matrix functionalized with a cytokine-binding domain from fibronectin (type I repeats 1–5), CXCL11 significantly improved skin wound healing outcome in diabetic mice compared to treatment conditions with CXCL11 lacking fibronectin functionalization (Tortelli et al. 2013). Last, the HBD of laminin was also shown to bind to various growth factors promiscuously and was used to functionalize fibrin matrices. When used to deliver VEGF-A165 and PDGF-BB to diabetic mouse skin wounds, laminin HBD-functionalized matrices significantly improved wound closure, granulation tissue formation, and angiogenesis compared to nonfunctionalized fibrin matrices (Ishihara et al. 2018). Although hydrogel functionalization using ECM-derived motifs is an attractive way to control the delivery of cytokines and growth factors, the efficiency of these systems is somewhat constrained by the naturally occurring affinity between the ECM motif and the growth factor or cytokine. In addition, the use of multiple recombinant proteins (ECM motif and growth factor or cytokine) may complicate the regulatory path for approval and increase costs.

To enable greater flexibility for biomaterial-based biologics delivery, novel approaches such as functionalization with aptamers have been investigated. Aptamers are single-stranded synthetic oligonucleotide sequences chosen from large nucleic acid libraries using an in vitro selection technique known as systematic evolution of ligands by exponential enrichment (SELEX). This process results in the selection of aptamers with high specificity and affinity for their target ligands. Owing to these properties, aptamer-functionalized hydrogels are emerging delivery systems that offer the potential to specifically bind and retain any protein to facilitate their controlled release (Abune et al. 2021). To date, fibrin- and PEG-based hydrogels have been successfully functionalized with anti-FGF-2, -PDGF-BB, and -VEGF-A aptamers and have shown promising abilities to specifically sequester their target growth factors and enable their release in a steady and sustained manner (Battig et al. 2014; Abune et al. 2019; Zhao et al. 2019a,b). In addition, codelivery of two growth factors was able to be achieved by simply incorporating separate aptamers into hydrogel matrices (Abune et al. 2019; Zhao et al. 2019b), demonstrating the potential adaptability of this system for the modulation of complex microenvironments such as skin wounds. Notably, aptamer-functionalized fibrin-based hydrogels were successfully used to deliver VEGF-A to murine skin wounds, resulting in significantly increased reepithelialization, angiogenesis, and hair follicle regeneration compared to control conditions (Zhao et al. 2019a).

Aptamer functionalized hydrogels provide advantages over other methods of hydrogel functionalization where the protein to be delivered is not strongly retained within matrices (e.g., lacking heparin-binding properties). However, additional work is required to understand the degradation of aptamer functionalized hydrogels, in particular by nuclease activity, and how this affects release of bound proteins. In addition, modification of aptamers could be required for their effective incorporation into hydrogels and improvement of stability, and achieving this without affecting their ability to bind target proteins may be challenging (Gačanin et al. 2020; Abune et al. 2021). Previous work has identified methods of improving aptamer stability in serum through chemical modification, and these techniques may be exploited for future aptamer-based hydrogel development (Kratschmer and Levy 2017). The relatively high costs of suitable aptamer identification and large-scale synthesis compared to other functionalization methods are also important considerations for future hydrogel delivery system development. Notably, aptamer-based therapies for a variety of conditions are currently undergoing clinical trials, with one treatment achieving FDA approval, demonstrating the potential for these systems to be adopted for clinical use in human wound therapy (Kaur et al. 2018).

Together, these functionalized hydrogel systems provide myriad options for controlling the delivery of growth factors and cytokines to skin wounds. When selecting a delivery system, consideration should be given to wound characteristics and underlying etiology. For example, acute wounds without underlying pathologies may benefit from shorter, burst release treatments, whereas chronic ulcers complicated by inflammation and infection may require prolonged, gradual release of one or several factors. In addition, understanding the biochemical properties of the biologic to be delivered (e.g., affinity for ECM or heparin) is important to determine the most appropriate delivery system.

PROTEIN ENGINEERING FOR CONTROLLED GROWTH FACTOR AND CYTOKINE DELIVERY

Biomaterial-based delivery systems for therapeutic proteins have been important advancements for wound therapy, particularly where prolonged and controlled release has been a significant challenge. As an alternative or complement to biomaterial engineering, strategies that involve modifying the protein itself to augment binding to biomaterials and ECM, improve stability or control signaling have been increasingly explored. Engineered recombinant growth factors and cytokines have shown promise for regenerative medicine applications, including the treatment of skin wounds (Ren et al. 2020).

Engineering Strategies to Increase Affinity to Biomaterials and Extracellular Matrix

Diffusion of biologics away from the wound environment increases the risk of systemic side effects, reduces their half-life, and consequently limits their therapeutic efficacy (Mitchell et al. 2016; Ren et al. 2020). In addition to biomaterial-based systems, protein engineering strategies have been developed to restrict recombinant protein localization to wounds and prolong their activity following delivery. The most prevalent approach has been to exploit the affinity of known ECM-binding domains and conjugating these motifs to the therapeutic protein of interest to enhance their interaction with the wound ECM or biomaterial delivery system (Fig. 2B). For example, given that collagen is the most abundant ECM protein within the skin, growth factors fused with collagen-binding domains (CBDs) have been engineered and delivered with and without biomaterials to skin wounds (Addi et al. 2017). Commonly used CBDs are derived from human von Willebrand factor, fibronectin, or collagenase, all of which demonstrate a strong affinity for wound ECM and/or collagen-based biomaterial matrices (de Souza and Brentani 1992; Andrades et al. 2001; Ishikawa et al. 2001; Sun et al. 2007; Yan et al. 2010). Growth factors fused with CBDs for wound healing applications include EGF, FGF-2, VEGF-A, and PDGF-BB, with these demonstrating prolonged retention within wounds and significant improvements in wound closure compared to treatment with growth factors in their native form (Andrades et al. 2001; Ishikawa et al. 2001; Sun et al. 2007; Yan et al. 2010). CBDs have also been fused to immunomodulatory cytokines for cancer immunotherapy; however, to our knowledge, these have not been investigated as wound therapeutics (Mansurov et al. 2021).

In addition to CBDs, other domains capable of promiscuously binding ECM proteins have been identified and used for growth factor engineering. In particular, the HBD of PlGF-2 was found to bind strongly to several ECM proteins and was fused to VEGF-A165 and PDGF-BB to create engineered growth factors with up to 100-fold higher affinity for ECM proteins compared to their wild-type versions. These “super-affinity” growth factors retained their biological activity and, when delivered alone or within fibrin matrices, were able to significantly improve skin wound healing in diabetic mice (Martino et al. 2014). Notably, the doses of super-affinity VEGF-165 and PDGF-BB used were between 40- and 250-fold lower than those shown to be effective in previous studies, with significantly improved re-epithelialization, granulation tissue formation, and angiogenesis reported compared to equivalent doses of unmodified recombinant protein or untreated/fibrin-only controls (Galiano et al. 2004; Chan et al. 2006; Martino et al. 2014). Importantly, the effectiveness of super-affinity VEGF-A165 at low doses was able to abrogate the common VEGF treatment–induced side effect of increased vascular permeability by up to 90% (Martino et al. 2014). These findings demonstrate the versatility of this system and its potential to be adopted for any growth factor or cytokine that shows promise for wound therapy.

To enable greater control of release from fibrin-based delivery systems, growth factors have been engineered to facilitate covalent attachment within hydrogel matrices. This has been achieved by fusing the substrate sequence for factor XIIIa, the transglutaminase involved in fibrin cross-linking, to growth factors resulting in covalent attachment within fibrin matrices during polymerization. This substrate sequence, known as α₂-plasmin inhibitor_1–8 (α₂-PI_1–8) has been fused or covalently attached to a number of growth factors including VEGF-A, PDGF-AB, insulin-like growth factor 1 (IGF-1), and keratinocyte growth factor (KGF/FGF-7) for delivery to various injury models including skin wounds (Geer et al. 2005; Sacchi et al. 2014; Mittermayr et al. 2016; Vardar et al. 2018). Initially demonstrated in studies of angiogenesis, α₂-PI_1–8-fused VEGF-A121 was shown to retain its activity and significantly enhance endothelial cell proliferation when delivered in fibrin matrices compared to native VEGF-A121 (Zisch et al. 2001; Ehrbar et al. 2004). In a wound healing study also delivering another member of the VEGF family, α₂-PI_1–8 in combination with an MMP cleavage site was fused to the amino terminus of VEGF-C, enabling its release from fibrin by plasmin- and MMP-mediated degradation within skin wounds. This resulted in significantly increased lymphangiogenesis compared to delivery of free VEGF-C, with subsequent improvements in granulation tissue formation and ECM deposition in murine diabetic skin wounds (Güç et al. 2017). Using an alternative method, α₂-PI_1–8 has also been covalently conjugated to KGF prior to incorporation within fibrin. Although rates of wound healing were comparable between KGF-treated groups, it was found that α₂-PI_1–8-KGF was released more slowly from fibrin compared to unmodified KGF, suggesting this strategy is more likely to demonstrate benefits in the context of delayed healing rather than normal wounds (Geer et al. 2005).

Engineered growth factors with increased affinity for biomaterials and ECM proteins have clearly demonstrated their potential to improve wound healing in animal models. Similar engineering strategies should be explored for cytokines known to be important for modulating wound healing responses, such as IL-10 (King et al. 2014).

Engineering Strategies to Improve Protein Stability

Improving the stability of biologics within the inflammatory and protease-rich wound environment is essential to maintaining their therapeutic activity. To this end, protein engineering strategies have been developed to protect recombinant cytokines and growth factors from proteolytic degradation and increase their thermal stability (Fig. 2C).

The site-specific covalent conjugation of PEG to therapeutic growth factors and cytokines in a process known as PEGylation is a well-established approach to improving the half-life of biologics. PEGylation is suggested to increase protein stability by masking proteolytic cleavage sites and limiting protein unfolding (Roberts et al. 2002; Huang et al. 2011). PEGylation of acidic fibroblast growth factor (FGF-1) was shown to increase in vivo half-life by 4.6-fold compared to native FGF-1 and accelerated wound closure in a diabetic rat model compared to controls (Huang et al. 2011). Increasing the molecular weight of conjugated PEG was shown to improve the stability of recombinant human EGF in rat wound extracts, with a 37.7-fold increase in half-life reported for EGF conjugated with 20 kDa PEG compared to unconjugated EGF (Hee Na et al. 2006). Importantly, PEGylation can impair bioactivity, thus PEGylation sites should be carefully selected to maintain protein function (Ramos-de-la-Peña and Aguilar 2020). Using a structure-guided approach, a recent study described the production of several PEGylated FGF-2 variants that demonstrated ∼7- to 14-fold increased stability in mouse acute wound fluid compared to unmodified FGF-2. It was shown that PEGylation at a site distal to both the receptor- and HBDs of FGF-2 maintained the highest pro-healing activity in vivo, whereas PEGylation close to the HBD resulted in the greatest decrease in activity (Sun et al. 2020).

Recently, a protein found in silkworm hemolymph, 30Kc19α, was reported to increase the stability and skin permeability of FGF-2 when fused to its amino terminus. In vitro, 30Kc19α–FGF-2 demonstrated significantly increased stability from 6 h compared to unmodified FGF-2. Treatment of murine full-thickness excisional wounds with 30Kc19α–FGF-2 significantly enhanced wound closure, collagen deposition, and angiogenesis compared to control and unmodified FGF-2-treated wounds (Lee et al. 2021). Although the exact mechanisms remain elusive, it is suggested that 30Kc19α fusion increases protein stability via shielding effects arising from hydrophobic interactions. Importantly, daily administration of 30Kc19 to mice for 14 days did not identify any toxic or immunogenic effects, indicating its suitability for use as an in vivo delivery system (Park et al. 2012).

As an alternative to polymer conjugation or protein fusion, increased stability can be conferred to growth factors and cytokines by genetic modification to precisely edit their amino acid sequences. For example, site-directed mutagenesis of a plasmin cleavage site within VEGF-A165 was able to increase its resistance to proteolytic degradation. An R110A or R110Q substitution, as well as an A111P substitution was able to inhibit plasmin-mediated proteolysis of VEGF-A165 in vitro, while maintaining its mitogenic activity. Furthermore, A111P-substituted VEGF-A165 displayed an ∼2.5-fold increase in stability compared to wild-type VEGF-A165 after 1 h of incubation in human chronic wound fluid (Lauer et al. 2002). In a separate study, a double mutation of FGF-1 (C117P and K118V) was shown to increase resistance to trypsin-mediated degradation by 100-fold compared to wild-type FGF-1, while maintaining mitogenic effects on murine fibroblasts in vitro (Kobielak et al. 2014).

Targeted amino acid substitutions to introduce additional disulfide bonds within a protein's structure has been used successfully to enhance stability. EGF and FGF-2 stabilization using this strategy resulted in an 11°C and 4°C increase in melting temperature, respectively, compared to their unmodified versions. These stabilized growth factors were able to significantly improve diabetic wound healing when delivered in a hyaluronate-collagen matrix compared to matrix-only control, indicating that alteration of their structure using this strategy did not affect their activity (Choi et al. 2018).

Another strategy to improve growth factor and cytokine stability involves the production of dimerized proteins. For example, a dimerized form of IL-10 produced by using an internal Gly-Ser flexible linker resulted in increased thermal and pH stability compared to the native form of IL-10. In a murine LPS-induced skin inflammation model, dimerized IL-10 was shown to be more effective at suppressing inflammation compared to native IL-10 (Minshawi et al. 2020). Fusion of the Fc region of immunoglobulin to proteins and the resulting dimerization has also been shown to significantly improve the stability of therapeutics in vivo (Czajkowsky et al. 2012). For example, treatment with Fc-fused interleukin 22 (IL-22-Fc) has been reported to significantly improve diabetic wound healing in mice (Kolumam et al. 2017). Interestingly, in addition to topical application, systemic delivery of IL-22–Fc was able to improve wound healing in this study, highlighting the remarkable stability of Fc-fused proteins. Fc-fused variants of growth factors relevant to wound healing, such as FGF-1, VEGF, and EGF have also been developed (Dikov et al. 1998; Ogiwara et al. 2005; Yu et al. 2012); however, to our knowledge, their application for wound therapy has not been reported to date.

The delivery of biologics with increased stability has clearly been demonstrated to result in significant improvements in wound healing outcome. However, these improvements further necessitate the restriction of these proteins to the wound environment, as reduced degradation and clearance may increase the risk of side effects if permitted to diffuse away from the wound. Complementing carefully selected concentrations of stabilized proteins with optimal delivery systems may be an effective strategy to minimize the risk of unwanted systemic effects.

Engineering Strategies to Modulate Signaling

The efficiency of a therapeutic protein is ultimately dependent on its capacity to induce and maintain appropriate activity within its target cells. We have seen that it is imperative to control growth factor and cytokine concentration over time to trigger a sustained, “drop by drop,” signaling within the wound. This can be achieved to some extent with the various delivery systems described above. However, controlled signaling may also be achieved by engineering interaction of the growth factors and cytokines with their cell-surface receptors and coreceptors (Fig. 2D).

For example, it was reported that fusion of a syndecan-binding domain (SBD) from laminin subunit α1 to the carboxyl terminus of VEGF-A121 and PDGF-BB was able to alter the signaling characteristics of these growth factors (Mochizuki et al. 2020). Syndecans are transmembrane heparan sulfate proteoglycans that can function as cell surface docking receptors for growth factors (Kwon et al. 2012). Using in vitro assays, SBD-fused VEGF-A121 (VEGF-A-SB) and PDGF-BB (PDGF-BB-SB) were shown to induce a lower-intensity but sustained form of signaling known as “tonic signaling” compared to the rapid but short-lived burst signaling triggered by wild-type growth factors. This was attributed to a lower level of receptor internalization and degradation induced by VEGF-A-SB and PDGF-BB-SB compared to their wild-type forms. Subsequently, these growth factors were fused with an amino-terminal α₂-PI_1–8 and MMP cleavage site and delivered to injury models via covalent attachment within fibrin matrices. Compared to wild-type and non-syndecan-binding covalently linked versions, VEGF-A-SB and PDGF-BB-SB were able to significantly improve diabetic wound healing and bone regeneration, respectively. Importantly, side effects were reduced using these engineered growth factors, with PDGF-BB-SB shown to have no effect on tumor growth in vivo and VEGF-A-SB inducing less vascular permeability compared to non-syndecan-binding versions (Mochizuki et al. 2020).

Another strategy to modulate growth factor activity involves the stimulation of synergistic signaling responses. Several studies have reported that interactions between integrins and growth factor receptors can potentiate signaling induced by each receptor, because they form large receptor clusters and share intracellular signaling molecules (Comoglio et al. 2003). As described earlier, when cross-linked into fibrin matrices, a fusion protein consisting of integrin-binding domain FNIII 9–10 and the growth factor-binding domain FNIII 12–14 was able to bind VEGF-A165 and PDGF-BB and enhance their pro-healing effects when delivered to diabetic mouse skin wounds. Besides improving growth factor retention within the fibrin matrix, the binding of integrin α₅β₁ by FNIII 9–10/FNIII 12–14 fusion protein resulted in synergism with growth factor receptors on target cells in vitro. Codelivery of FNIII 9–10/FNIII 12–14 with VEGF-A165 and PDGF-BB resulted in increased and prolonged phosphorylation of VEGF receptor 2 and PDGF receptor β, with subsequent increases in the activity of downstream signaling pathways. This synergistic effect likely occurs due to the proximity of growth factor receptors and integrins enabled by the structure of the FNIII 9–10/FNIII 12–14 fusion protein (Martino et al. 2011).

Stimulating optimal signaling responses with minimal doses of therapeutic protein is an important step toward overcoming the limitations of recombinant growth factor- and cytokine-based therapies. The use of lower doses minimizes the risk of undesirable side effects while maintaining efficacy and improving cost-effectiveness. These are key features to achieve to successfully translate these novel strategies into the clinic for wound healing applications.

CONCLUDING REMARKS

To develop highly effective skin wound treatments, understanding the most important factors involved at each stage of the wound healing process and the underlying mechanisms by which healing is impaired or totally prevented is essential. A comprehensive understanding of these mechanisms enables the development of novel strategies that can counteract key contributors to pathological healing such as excessive and prolonged inflammation or promote pathways that stimulate cell proliferation and minimize scarring.

As a result of their critical roles in wound healing, growth factor and cytokine therapeutics offer the potential to augment the intrinsic reparative and regenerative mechanisms present within tissues, in particular when those are impaired due to chronic diseases or aging. These treatments capitalize on our ever-expanding knowledge of these molecules and the mechanisms underlying normal and pathological tissue repair. To date, several growth factors and immunomodulatory cytokines have been identified as important positive regulators of wound healing (Werner and Grose 2003; Larouche et al. 2018); however, their translation into effective therapeutics in humans has been limited. Nevertheless, developing optimal delivery systems and engineering strategies for those proteins holds substantial promise for future therapeutic development.

Given the complex, multifaceted nature of wound healing, it is likely that delivering combinations of growth factors and cytokines will result in the most beneficial outcomes for patients. However, the use of multiple biologics or complex biomaterial systems may hinder the wide adoption of these therapeutic strategies, due to regulatory path hurdles and costs. Simple and effective therapeutic strategies based on growth factors and cytokines may require the development of novel systems and engineering methods to ensure that the most appropriate growth factors and cytokines are made available at relevant stages of healing. Furthermore, new delivery systems should be developed to ensure that growth factor or cytokines target the desired cell type at the optimal moment.

Finally, human wound healing is highly diverse and can be influenced by underlying pathologies and genetic variation (Tipton et al. 2020). Therefore, current and future delivery strategies for recombinant growth factors and cytokines must allow a high degree of adaptability for varying wound contexts to maximize treatment success.