From structure to application: Progress and opportunities in peptide materials development

Abstract

For over 20 years, peptide materials in their hydrogel or soluble fibril form have been used for biomedical applications such as drug delivery, cell culture, vaccines, and tissue regeneration. To facilitate the translation of these materials, key areas of research still need to be addressed. Their structural characterization lags compared to amyloid proteins. Many of the structural features designed to guide materials formation are primarily being characterized by their observation in atomic resolution structures of amyloid assemblies. Herein, these motifs are examined in relation to peptide designs identifying common interactions that drive assembly and provide structural specificity. Current efforts to design complex structures, as reviewed here, highlight the need to extend the structural revolution of amyloid proteins to peptide assemblies to validate design principles. With respect to clinical applications, the fundamental interactions and responses of proteins, cells, and the immune system to peptide materials are still not well understood. Only a few trends are just now emerging for peptide materials interactions with biological systems. Understanding how peptide material properties influence these interactions will enable the translation of materials towards current and emerging applications.

Introduction

Decades of research in the structure, formation, function, and biological effects of amyloid proteins have uncovered the main factors driving their assembly into supramolecular structures [1-4]. Many of the emerging structural features observed in amyloid proteins validate general design concepts that have historically been used in peptide materials, a field in which high-resolution structure determination is rare. For example, the same non-covalent interactions such as the hydrophobic effect, van der Waals forces, and hydrogen bonding drive assembly and provide structural specificity in protein and peptide assemblies. Several early classes of self-assembling peptide systems, from which much has been learned, take advantage of these overarching biophysical forces, including aromatic dipeptides, such as diphenylalanine (FF) and its Fmoc derivatives, and 10 - 20 residue amphiphilic or ionic self-complementary oligopeptides [5, 6]. However, beyond these general biophysical concepts, emerging amyloid structures provide insight into distinct inter-residue and main-chain interactions, and this, coupled with the diversity of natural and non-natural amino acid building blocks, should inspire creative new assembly designs.

Peptides prone to assembly can do so when initially dissolved in an aqueous solution or can be designed to form fibrils when triggered by different stimuli such as pH, temperature, and concentration. Depending on the concentration, peptide fibrils can either stay in solution or, at higher concentration, entangle to form physically crosslinked three-dimensional networks called hydrogels. β-sheet peptide materials bear attractive properties for biomedical applications either in their soluble fibril form or as hydrogels. The design of these peptides at the monomer level directly translates into the materials’ structure and properties, facilitating the incorporation of chemical diversity, bioactive cues, and motifs for further functionalization. Peptide materials are typically biocompatible and do not elicit adverse immune response unless designed to do so. Also, they are easily biodegradable, and their fibrous structure can be designed to resemble the natural extracellular matrix environment [7]. These properties facilitated the use of peptide materials in numerous applications, including vehicles for localized and controlled delivery of protein and nucleic acid therapeutics, small molecule drugs, and cell delivery [8-12], scaffolds for 2D and 3D cell culture for in vitro models [13, 14], vaccine adjuvants [15, 16], as surgical hemostatic or anastomotic agents [17-20], and materials for promoting wound healing and tissue regeneration [21, 22] (Figure 1).

Figure 1.

a) The common structural features of amyloid proteins such as hydrophobic packing, π-π, hydrogen bonding, and ionic interactions have been used to develop β-sheet assembling peptides. b) Examples of peptide systems that rely on these features are MAX1, diphenylalanine derivatives, Ac-I₃QGK-NH₂, and RADA16-I, to name a few. c) Self-assembling peptides form materials such as fibrillar hydrogels and soluble fibrils used in biomedical applications. The Protein Data Bank (PDB) ID of amyloid-β [32], FUS [57], α-synuclein [39], and MAX1 [53] are shown in parenthesis. The other peptide structures are models or representations of proposed structures. FmocFF model adapted from [58]. © 2019, Ji et al. under CC BY license. Ac-I₃QGK-NH₂ adapted from [59] © Wang et al. under CC BY 4.0 license. RADA16-I structure image adapted from [54] © 2013, American Chemical Society.

To date, much effort has been dedicated to designing peptides capable of assembly, establishing the rules governing self-assembly, and developing applications for the resulting materials, mostly in drug delivery. However, our knowledge of the underlying structure of these systems at the molecular and atomic level greatly lags in contrast with the recent surge of amyloidogenic protein structure determination. Detailed structural knowledge of the diverse β-sheet peptide materials being developed may uncover new self-assembly principles and improve our ability to control the formation of well-defined nanostructured materials. Furthermore, there is still a void in our mechanistic understanding of the assembly process and the influence that the fibril structure and chemical composition has on the biological behavior. The ability to predict the cellular or immune responses to these fibrillar materials will inform the design of new materials and aid in the selection of existing materials for targeted applications. Despite an extensive exploration of peptide biomaterials, only a few self-assembling systems have achieved commercialization as scaffolds for cell culture and 3D printing, and only one has been cleared by the FDA for clinical use as a medical device for hemostasis and wound healing [23]. This review discusses the recent advances and trends in the design of biomaterials composed of β-sheet peptides directly derived from amyloid proteins or de novo designs. We focus on applications and observations on the cellular and immune responses to these systems and highlight areas in need of study using a few peptide materials as examples. The reader can find other excellent reviews that comprehensively cover the plethora of published peptide-material systems [6, 24-28].

Structural insights of β-sheet self-assembling peptide nanofibers

There has been a recent explosion of work elucidating the structure of amyloidogenic proteins and unraveling amyloid formation mechanisms, which can potentially uncover therapeutic liabilities of amyloid-associated diseases. Multiple structures from amyloid-beta (Aβ) [29-37], α-synuclein [38-40], Tau [41-43], FUS [44], among other proteins [45-48], have been obtained thanks to advances in crystallography, Cryo-EM, NMR, and computation. These studies revealed several common structural features that define amyloid proteins. First, they adopt a cross-β structure in which β-strands align perpendicular to the fibril growth axis while forming hydrogen bonds along the peptide backbone. Amyloid fibrils also form the so-called “steric zipper” composed of buried, mostly hydrophobic residues that drive aggregation by the hydrophobic effect, see amyloid-β in Figure 1a. These general features (main-chain H-bonding and hydrophobic collapse) have been inferred for many years based on biophysical studies, modeling, scattering, and microscopy. Although these proteins can form polymorphic fibrils depending on environmental conditions, these underlying structural features are always conserved. However, the exact β-fold of the main chain and the highly repetitive aromatic and polar interactions between residue side chains are illuminating. Inter-residue interactions are typically local in globular proteins, but in amyloid assemblies, they become repetitive long-range interaction highways that permeate the entire length of a given fibril. Great examples are the π-π aromatic interactions and H-bonding made between the tyrosine side chains in FUS and the salt bridge network found in α-synuclein, Figure 1a. These polar interactions often provide folding/assembly specificity. In addition to these contributions from the structural biology community, the field of crystal engineering has characterized similar intermolecular interactions that effect crystal packing and can be utilized in materials design [49, 50].

Amyloid and other natural proteins have been an inspiration for the development of peptide materials. Assembling sequences can be derived directly from protein motifs or created de novo utilizing the common structural features of amyloid protein. Zhang et al. provided one of the first examples by discovering ionic complementary self-assembling peptides derived from Zuotin protein [50]. An extensive body of work studying this system led to RADA-16-I, the only FDA-approved peptide gel, Figure 1b. Since then, various peptide designs have been reported that rely on hydrophobic packing, π-π interactions, ionic and polar interactions to control assembly and fiber formation. Most of these systems' structural characterization is limited to a few biophysical techniques that show the expected β-sheet fibrillar structure, such as circular dichroism, infrared spectroscopy, and nominal resolution TEM. High-resolution structures cannot be determined by these techniques, so in most cases, only schemes of predicted or desired models are provided. Although there has been a recent effort in determining the structures of model β-sheet forming peptides [51, 52], to our knowledge, only two structures have been determined for peptide fibrils, namely MAX1 [53] and RADA16-I [54]. In general, high resolution structural characterization of more systems is required to confirm existing design principles and uncover new ones leading to new materials, including ones with dynamic structures and function. To this end, although solid-state NMR continues to be extremely useful for self-assembled systems, recent advances in using cryo-EM to elucidate fibrillar structures should catalyze materials design. Further, computation can be a powerful tool in designing peptide and protein materials especially when couple with high-resolution structure determination as was demonstrated for the generation of filaments comprised of self-assembled helical protein [55]. Although computation has recently been applied to sheet-containing systems [56], to our knowledge no high-structures have been solved.

Trends in β-sheet peptide materials design

A current interest is to create materials with complex structures, hierarchical architectures, multiple functionalities, and dynamic properties. Typically, peptide materials are homogenous systems composed of a single self-assembling peptide, which provides a well-defined chemical composition, ease in design and preparation, and controlled physicochemical properties. One emerging strategy to increase structural and functional complexity involves designing multi-component systems [60] where different peptides and/or other molecules are mixed and allowed to form co-assembled [61, 62] or self-sorted structures [63] (Figure 2a). In co-assembled systems, individual fibers contain all the components available for assembly. Co-assembly can be directed by charge complementarity [62], hydrophobic interactions [58, 64], or stereocomplexation [65-68]. Co-assembly generates materials with enhanced or different properties than their single-component counterpart [69, 70]. For instance, Fmoc-protected diphenylalanine (Fmoc-FF) can undergo a structural transition from β-sheet nanoribbons to helical twisted fibrils and have responsiveness to light and redox conditions when co-assembled with bipyridine derivatives [58, 64]. Co-assembly can also be used to enhance peptide hydrogel stiffness either by stereocomplexation [66, 71] or inducing crosslink formation in a purely supramolecular manner [72]. In particular, stereocomplexation can be a predictive and useful tool, especially in the formation of rippled β-sheet assemblies as recently reviewed [73]. In self-sorting systems, individual components self-assemble independently, creating a mixture of coexisting but separate structures. For example, Webber et al. showed that a mixture of DWDW and KWKW tetrapeptides formed DWDW-based fibrillar gel containing KWKW-based spherical nanoparticles [63]. Self-sorting can also be achieved by mixing peptides with other molecules such as lipids [74], polysaccharides, and polymers [75, 76]. The development of these more complex systems should evolve together with biophysical and structural analysis of the resulting materials and strategies to favor the formation of selective structures [62, 77, 78].

Figure 2.

Trends in β-sheet peptide materials design and emerging applications in drug delivery and tissue regeneration. a) Multicomponent peptide systems can coassemble or self-sort into fibrillar structures. b) Strategies to create hierarchical assembly and fiber alignment by molecular crowding, polar interactions, and phase merging. c) Different strategies used for controlled drug delivery of small molecules, nucleic acids, and proteins. Peptide materials are also used to promote tissue regeneration. Structure of Ac-I₃QGK-NH₂ representing lateral aggregation by side chain interactions was adapted from [59] © Wang et al. under CC BY 4.0 license.

Another strategy to introduce molecular complexity and structural diversity is creating hierarchical structures by inducing fiber alignment (Figure 2b). In nature, it is common to see complex and hierarchical structures, for example, collagen bundles within the extracellular matrix. The hierarchical assembly of collagen that results from post-translational modification of collagen peptides, including glycosylation and enzymatic crosslinking, is required to provide mechanical integrity and guide tissue formation [79]. Although peptide materials have not yet achieved the hierarchal complexity of tissues, progress has been made in inducing fiber-fiber interactions leading to hierarchical assembly. For example, the bundling of fibrils formed by the glycosylated peptide Q11 (QQKFQFQFEQQ) can be enhanced by inducing molecular crowding with polymeric additives, Figure 2b [80]. Xu et al. showed that fibril lamination could be facilitated by the incorporation of uncharged polar amino acids. Adding glutamine, serine, or asparagine to the self-assembling peptide Ac-I₃XGK-NH₂ (X = Q, S, R) favors the formation of wide ribbons comprised of laminated fibers, suggesting that favorable polar interactions are formed between individual fibers in the final laminated structure [59]. This phenomenon is also observed in heterochiral ld or dl FKFEFKFE peptides, which form wide helical tapes instead of typical nanofibers [81]. These examples highlight the promise of hierarchical peptide assembly and the opportunity to discover the structural principles guiding their formation and function. Lastly, aligned nanofiber bundles can be fabricated by physical methods (Figure 2b). Self-assembly of the dipeptide FF at the interface of hexafluoroisopropanol and water leads to the formation of fibers that can be physically pulled from the interface affording micrometer-long bundles that have electromechanical properties [82]. Further exploration of strategies to create hierarchical structures could ultimately result in materials that mimic anisotropic and highly ordered tissues.

Emerging biomedical applications of fibrous peptide materials

Peptide materials have been used extensively as scaffolds for the controlled delivery of proteins, nucleic acids, and small-molecule drugs. Generally, two main mechanisms dictate the release rate of a drug from a peptide gel network. The first is simple diffusion which depends on the size of the drug and the network's mesh size. As the size of the drug approaches the mesh size, steric interactions impact its release. The second factor involves the reversible formation of specific or non-specific interactions between the drug and the fibril network. Depending on the identity of the drug and the gel network, one or both factors can be at play. A current focus involves designing alternative strategies that enhance control over delivery rates, offer regiospecific control over drug action, and create generic hydrogel materials capable of delivering a broad range of therapeutics, eliminating the need to design custom gels for individual therapies.

Self-assembling prodrugs are simply peptides that contain a drug ligated via a cleavable linker (Figure 2c). The peptide portion is responsible for the assembly and formation of fibrils that display drugs from their surface. Drug release is dependent on linker degradation, typically the result of enzymatic action. These systems offer localized and sustained release [83]. Drug-peptide conjugates are a similar class of assembler [84-89]. Here, drugs are also appended to the peptide but can participate in the self-assembly process and become incorporated into the fibril's core architecture [84-86, 90]. For example, Hartgerink et al. designed a nanofibrous hydrogel called L-NIL-MDP, where the lysine residues of the self-assembling peptide K₂(SL)₆K₂ were substituted by the small-molecule drug N⁶-(1-iminoethyl)-L-lysine (L-NIL). Gels prepared from this peptide retained native L-NIL bioactivity, inhibiting the production of inducible nitric oxide synthase in vitro and in vivo [84].

Peptide-drug conjugates can also be utilized to selectively deliver a therapeutic agent either at the cell surface or intracellularly by a process called enzyme-instructed self-assembly pioneered by Xu et al. [91-93] (Figure 2c). An enzymatically cleavable moiety is incorporated into the peptide to endear solubility, thus inhibiting self-assembly until a specific enzyme is encountered. Enzyme-mediated removal of the water-solubilizing moiety triggers peptide assembly in a regiospecific manner [87, 88, 94-96]. These enzyme-susceptible peptides can be functionalized with a drug to concentrate the therapy at desired cellular organelles through enzyme-triggered fibrillization. This strategy was used by Zhong et al. to deliver Lonidamine (LND) and inhibit glycolysis in the mitochondria of cancer cells [88]. In general, the use of fiber-forming peptides conjugated with drugs, whether self-assembling prodrugs or peptide-drug conjugates, provides several unique advantages. First, fiber formation can be triggered by environmental factors, such as pH, redox conditions, or enzymatic activity, to name a few, allowing regiospecific material formation. The resulting fibers can display the drug at high density, increasing its local concentration. Appending ligands to the self-assembling peptide to target unique cell surface receptors can deliver therapeutic material to select cells, potentially reducing side effects [97, 98]. The drug release profiles of these materials depend on their environment, for example, the presence of enzyme or a change in pH, affecting on-demand delivery. Lastly, some can form injectable hydrogels facilitating implantation [84] and be designed to co-deliver multiple components for combination therapy [86, 99].

Other recent strategies aim to produce general peptide-based delivery platforms that can be used for an entire class of cargo, eliminating the need to custom design a new hydrogel for each drug one wishes to deliver. One example involves protein therapeutics whose delivery can be challenging due to the possibility of protein-material adsorption, which can cause protein denaturation. Schneider et al. developed a protein delivery platform that limits direct contact between globular protein domains and the hydrogel fibrillar network (Figure 2c). Their approach uses complementary electrostatic interactions made between a suite of designed interaction domains (IDs), ligated to the terminus of a protein of interest, and a negatively charged hydrogel network. The gel network repels the globular domain of like-charged proteins with only the ID binding to the fibrillar network. Desorption energies can be modulated by the choice of ID to control protein release kinetics [100].

A general hydrogel has also been designed for the delivery of hydrophilic small molecules. For example, Tofacitinib, an immunosuppressant drug, is notoriously difficult to deliver because of its hydrophilicity. The drug quickly diffuses from most materials resulting in burst delivery. However, Tofacitinib forms microcrystals when dissolved together with hydrogel-forming peptides [101]. Triggered peptide assembly leads to the formation of hydrogel containing the soluble drug. Tofacitinib then crystallizes in the nanofibrous network forming a multiphase hydrogel where the two components coexist independently. Importantly, since the gel is non-interacting, the drug's release rate depends on crystal dissolution [101] (Figure 2c). In general, the approaches described in this section demonstrate the potential of finely controlled drug release. Future research might focus on achieving complete control over the on-demand drug release in vivo, where materials are exposed to the immune system and susceptible to fouling and degradation. Understanding in vivo response to these systems is important because it will ultimately affect these materials' efficacy and the resulting therapeutic effect, as will be described in the following sections.

Effects of peptide-based materials on cell responses

Peptide-based coatings and hydrogels have been widely used for two-dimensional and three-dimensional cell culture and as vehicles for exogenous cell delivery, commonly demonstrating high cytocompatibility [6, 9, 13, 25, 102]. Material cytocompatibility is assessed by quantifying viable cells and evaluating their metabolic activity, attachment (for adherent cells), and proliferation over time. Although experiments reported across the literature are performed with different cell types, highly diverse experimental conditions, and varying peptide designs, there are a few emerging trends in the cell response to this class of materials (Figure 3). For example, the hydrogel's charge state seems to be an important factor in determining cell behavior. Hydrogels prepared from negatively charged peptides seem to support cell viability. However, adherent cells do not adopt their native spread morphology and instead clump together. This phenomenon has been observed for human dermal fibroblasts cultured on gels formed from the β-hairpin peptide Ac-VES3 [9] (Figure 3a) and 3T3 fibroblasts seeded on glass coated with the decapeptide E1Y9 (Ac-EYEYKYEYKY) [103]. Similarly, viable fibroblasts do not spread homogenously and clump when cultured on coatings prepared from the zwitterionic self-assembling peptide QQKFQFQFEQQ (Q11) [80] or FEFEFKFK hydrogel [104] (Figure 3h). However, the incorporation of integrin-binding motifs or doping with other natural proteins facilitates cell adhesion on these peptide gels and coatings.

Figure 3.

Factors that play a role in cellular responses to β-sheet-rich peptide materials. a) Human Dermal Fibroblast (HDF) 2D cultured on the surface of 0.5 wt% AcVES3 and AcVES-GGGGRGDV for four days in medium without serum. Cells do not grow and spread on the negatively charged peptide gel unless an integrin-binding motif is incorporated. Green: calcein AM, red: ethidium homodimer I. Adapted with permission from [9]. © 2019, American Chemical Society. b) 2D cultured HDF cells on 0.5 wt% MAX1 gel with serum-free or serum-supplemented media for four days. Green: calcein AM, red: ethidium homodimer I. Adapted with permission from [105]. © 2021, American Chemical Society. c) 3D cultured NIH 3T3 fibroblasts on positively charged multidomain peptide (MDP) K₂ gel, which is cytotoxic initially, but viable cells proliferate over time or d) on neutral MDP O₅ gel, cells are highly viable but do not proliferate. Green: calcein AM or Phalloidin, red: ethidium homodimer I. Adapted with permission from [106]. © 2019, American Chemical Society. e) Adhesion of 2D cultured NIH 3T3 fibroblasts on glycosylated and non-glycosylated Q11 coatings in the presence of serum. Cells do not adhere to glycosylated Q11, but a few adhere and clump on the non-glycosylated material. Adapted with permission from [80]. © 2019, Restuccia et al. under CC BY 4.0 license. f) Oligodendrocyte precursor cells cultured on AYFIL gel for four days remain viable, and some neural processes are observed. Adapted with permission from [107]. © 2019, American Chemical Society. g) 2D culture of chondrocytes and NIH 3T3 fibroblasts on Fmoc-FF hydrogel for two days. High cell death is observed in fibroblasts in contrast to chondrocytes indicating that different cell types react differently to the same material. Green: calcein AM, red: ethidium homodimer I. Adapted with permission from [108]. © 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. h) 3D cultured human mammary fibroblast on FEFEFKFK gels with and without collagen. Cells spread and attach when collagen is present. Adapted with permission from [104]. © 2019, Ashworth et al. under CC BY license. The peptide charge at pH 7 is indicated in the parenthesis for each example.

In contrast, materials prepared from positively charged peptides sometimes present higher cytotoxicity, as has been observed for gels prepared from β-hairpin peptides MAX1 (Figure 3b), MAX8, and HLT2 [105], the multidomain peptide K₂(SL)₆K₂ [106] (Figure 3c), pentapeptide KYFIL [107] among others. However, the cytotoxicity of these systems is highly dependent on the culture conditions, and alternative hydrogel preparation methods can be used to enhance cell survival. For example, the Schneider group demonstrated that the inherent cytotoxicity displayed by highly positively charged peptide hydrogels could be significantly reduced by priming the hydrogels with serum, a common element in cell culture media, or even single protein additives. They demonstrated that positively charged gels adsorb serum proteins via electrostatic interactions that passivate the surface charge and mediate cell signaling, resulting in the activation of both the PI3/Akt and MAPK/ERK pathways. These pathways are known to facilitate resistance to stress-induced apoptosis and overall cell survival [105] (Figure 3b). Thus, for both positive and negatively charged gels, protein adsorption seems to play an important role, but the specifics are not yet clear.

The case for materials prepared from neutral and non-ionic peptides is even less understood and divergent responses have been observed (Figure 3d-g). For example, in some systems such as SFFSF [72] and AYFIL (Figure 3f) [107], cells grow, proliferate, and differentiate; other peptides such as Fmoc-FF (Figure 3g) do not support fibroblast cell viability [8, 108]; and others such as glycosylated Q11 (Figure 3e) [80], and O₅(SL)₆O₅ (O represents hydroxyproline, Figure 3d) [106, 109] do not induce cell attachment or proliferation when cells are encapsulated or seeded on the surface of the peptide hydrogels or peptide-coated glass. These observations suggest that other material attributes beyond charge are certainly at play.

Self-assembly pathways leading to fibrillization may also influence cell behavior, as is the case in amyloid formation. Early on, many thought that amyloid deposits were responsible for degenerative pathologies, but now there is strong evidence that oligomeric species formed during fibril formation are the toxic culprits [110]. In fact, several research groups are developing methods to catalyze the conversion of oligomers to non-toxic fibrils as a potential therapy [67, 111-113]. Oligomers might also form during the preparation of peptide materials. Using model peptides derived from prion proteins, Ventura et al. showed that cytotoxic oligomers were formed during the fibrillization process, which took seven days. The authors suggested that oligomer formation should also be assessed during the preparation of peptide-based materials [114]. With that said, these studies were performed under conditions that allowed slow fibrillization and concomitantly long persistence times for the oligomers. In contrast, most de novo designed material-forming peptides assemble quickly once triggered, where the persistence time of any oligomer would be short-lived if they are on-pathway to fibril formation. Certainly, it will be important in future studies to investigate the ramifications of self-assembly pathways and how material preparation methods influence cell behavior. Advancements in techniques such as liquid-cell TEM [115], solid-state NMR [116], and Cryo-EM will improve these efforts. For instance, Gelain and Weingarth utilized solid-state NMR to characterize the structure and assembly of several material-forming peptides for neural stem cell culture [116]. They observed a correlation between the degree of material formation, mechanical stiffness, and homogeneity on stem cell behavior, primarily cell differentiation. Heterogenous hydrogels were defined as those having a mixture of fibrillar networks and soluble "small assemblies". Heterogenous gels were characterized by low mechanical rigidities and poor cell differentiation potential. In contrast, homogenous gels, comprised mostly of fibrils, were mechanically stiff and better potentiated differentiation [116, 117]. These examples highlight that more mechanistic work is warranted and that the structural analysis of peptide materials can help define material attributes that influence cell behavior.

The different peptide-based materials developed over the years present distinct amino acid compositions, charge, self-assembling pathways, physical and mechanical properties, comprise numerous fiber morphologies, and often contain additional bioactive motifs to increase their cytocompatibility [8, 9, 109, 118]. Further, the effects of cell-induced peptide modifications or enzymatic degradation products have yet to be explored. Determining if and how each of these factors affects cell behavior is a daunting but necessary task. The result will be principles that globally unite the myriad of published observations and improve future material designs.

Immune response to β-sheet peptide materials: a relevant factor and target for materials design

Advances in immunology and immunoengineering have emphasized the immune system’s role in determining the overall response to materials and how material properties can be designed to modulate and induce specific immune responses [119-121]. Most of these studies used natural or synthetic polymers. With respect to β-sheet peptide materials, there has been incredible progress in their design, biophysical and mechanical characterization, and their assessment in vitro. In vivo studies have largely been carried out to evaluate their performance in affecting an intended function, for example, drug delivery or tissue engineering. However, the immune response to the materials was often not studied in any depth. In fact, efforts to define the immune response to peptide materials significantly lag behind those dedicated to commonly used polymers.

Early in the development of peptide materials, many outside the field believed that materials prepared from peptides would, a priori, elicit adverse immune responses precluding their biomedical use. Unfortunately, our community had not yet collected enough data to refute those concerns. Gratifyingly, over the last several years the field has begun to examine immune responses in earnest. As expected, responses vary and depend on the exact material, but adverse reactions in animal models have been rare.

Several independent observations of tissue responses to implanted peptide hydrogels show that these materials can be infiltrated by cells with different profiles depending on the peptide sequence and stereochemistry, hydrogel cargo such as cells or drugs, and the tissue type [21, 22, 84, 100, 101, 106, 109, 122] (Figure 4a). For example, gels prepared from the positively charged β-hairpin peptide MAX8, when injected subcutaneously in mice, were mildly infiltrated initially by neutrophils that gave way to macrophages after a few days. The gel persists for over a month but is ultimately degraded by macrophage-dependent phagocytosis and remodeled to native tissue [101]. A separate study suggested that stereochemistry may influence the immune response. Gels prepared from the negatively charged β-hairpin peptide AcVES3 had high macrophage infiltration after a few days in stark contrast to a gel prepared from its enantiomer d-AcVES3, which had significantly fewer infiltrating cells [100]. The basis for this stereochemical response is not yet known. Concerning the hairpin gels studied to date, neutrophils and macrophages largely represent the infiltrating milieu, with little representation of B- or T-cells. However, these cells can be recruited if certain cytokines are encapsulated in the gel (unpublished). Conversely, the global cellular response can be dampened by adding immunosuppressants to the hydrogel [101].

Figure 4.

Immune and inflammatory responses to β-sheet peptide materials. a) Examples of peptide hydrogel sections from subcutaneous injection of materials stained with Hematoxylin and Eosin. Cell nuclei are stained purple, cell cytoplasm, collagen, and hydrogel are stained in different shades of pink. Distinct cell morphology, infiltration degree, and cell distribution is observed depending on the peptide sequence, stereochemistry, immunotherapy drug cargo, and time. L- and D- AcVES3 images were adapted from [100] with permission from the American Chemical Society. Images of MDP K₂ and R₂ were adapted from [122] and MDP K₂ with immunotherapy drug from [123]. © 2020 and 2019 Elsevier Ltd. All rights reserved. Image from MDP O₅ was adapted from [106]. © 2019, American Chemical Society. The peptide charge at pH 7 is indicated in the parenthesis. b) When peptide hydrogels are implanted in injured nerve tissue such as spinal cord or sciatic nerve, macrophage infiltration is observed. RADA16-I image adapted from [21]. © 2020 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. MDP K₂ nerve image was adapted from [109]. © 2020 Elsevier Ltd. All rights reserved. c) Effects of peptide charge and structure in the efficacy of antibody production of epitope-bearing peptide fibers.

Hartgerink et al. compared the host immune response to the β-sheet peptide gels X₂(SL)₆X₂, where X can be lysine, arginine, glutamic acid, or aspartic acid residues [122] (Figure 4). They found that the X-residue identity and the charge dictated the immune response to those materials. The negatively charged gels were infiltrated by tissue-resident macrophages that do not induce proinflammation and cleared the material within ten days. In contrast, the positively charged peptide gels provoked higher cellular infiltration of inflammatory monocytes, macrophages, and granulocytes and presented higher levels of pro-inflammatory cytokines, particularly the arginine-based peptide. Further, the clearance was much slower for the positively charged gels. The immune response to the lysine-based peptide can be modified when immunotherapy drugs are loaded into the gel [123]. The neutral peptide from this family containing hydroxyproline in the flanking domains is also infiltrated only by macrophages and is easily degraded within two weeks [106] (Figure 4a). In injury models, hydrogels prepared from zwitterionic RADA16-I or positively charged MDP K₂ elicit robust macrophage responses in nerve tissues (Figure 4b) [21, 109].

Although the immune response varies depending on the exact peptide material, nearly all the materials reported to date are eventually biodegraded and replaced with native tissue [100, 101, 106, 122]. Similar to the discussion earlier on in vitro cellular responses to peptide materials, the factors influencing the immune response are not completely understood. The chemical composition, charge, shape, size, mechanical properties, and tissue type could all influence the response to these materials.

The most comprehensive work in analyzing immune responses to peptide nanofibers (as opposed to hydrogels) has been done in vaccine development. Peptide nanofibers bearing different immunogenic epitopes have been used to induce strong humoral immune responses without the use of adjuvants. The efficacy of these materials is due to their ability to present a high density of epitopes on the nanofiber surface, which can be easily customized during peptide synthesis and self-assembly [16, 124-126]. It is important to note that the fibers, by themselves, are non-immunogenic and require the addition of the epitope to elicit an adaptive immune response to specific targets. However, the peptide sequence does influence the vaccines’ effectiveness, although it is not understood exactly how. Collier and coworkers compared the antibody responses against OVA, an ovalbumin epitope, when presented on soluble nanofibers prepared from negatively versus positively-charged peptides derived from the parent peptide Q11 [127]. T-cell and B-cell responses were raised by OVA-Q11 and lysine modified-Q11 (K_nSGSG-Q11) but not by the nanofibers containing negative residues (E_nSGSG-Q11) (Figure 4c). When comparing the efficacy of peptide vaccines composed of the β-sheet-rich Q11 fibrils versus an α-helical peptide fiber, formed from the peptide Coil29, Wu et al. observed that the α-helical fibers elicited a greater antibody response suggesting that fibril morphology also influences activity [128] (Figure 4c). The immune and inflammatory responses to β-sheet peptide fibers and hydrogels represent the foundation of how these materials interact with living systems. However, it is still unclear what factors and material properties are responsible for evoking specific immune responses. To better guide efforts in the clinical translation of peptide materials, there is a need to better understand how their individual attributes (e.g. sequence composition, charge, biophysical/mechanical properties, etc.) influence the immune system. Further, a significant opportunity exists in the immunoengineering of these materials to control and inform the immune system.

Conclusion

Over the last 20 years, research on β-sheet peptide fibrils, coatings, and gels has evolved from understanding and controlling the mechanisms and forces driving assembly to their use in a broad range of biomedical applications. Novel designs shift from single to multiple component peptide systems and hierarchical assembly to incorporate structural and functional diversity into the fibrillar materials. The fabrication of more complex systems requires a deeper understanding of their structure to enhance our ability to favor the formation of specific and well-defined fibril morphologies and resulting materials. Among the most explored applications of peptide materials is their use as vehicles or scaffolds for drug delivery, cell culture, regenerative medicine, and vaccines. Significant advances have been made with the design of self-assembling prodrugs, peptide-drug conjugates, and hydrogels for controlled drug delivery, which represents a new frontier in therapeutics. The development of these materials can potentially impact the challenges associated with drug stability, localization, and side effects. Furthermore, β-sheet peptide materials display promising properties for delivering cell therapy and promoting tissue regeneration and wound healing. However, there is still a lack of fundamental understanding of the peptide material interactions with cells and organisms, particularly the immune system. Exploring the influence of material properties on biological systems will complement our ability to design new systems and guide the selection of existing materials for a particular application. Interdisciplinary collaboration could expedite efforts to establish rules that connect peptide composition, sequence, and structure with material biophysical/mechanical attributes, and cellular/immune responses in vivo. We intended to highlight needed research efforts throughout the review that should help catalyze the translation of academic discoveries through preclinical evaluation and into clinical use.