Affinity Hydrogels for Protein Delivery

Abstract

Proteins have been studied as therapeutic agents for treatment of various human diseases. However, the delivery of protein drugs into the body is challenging. The purpose of this review is to summarize and highlight the progress in developing affinity hydrogels (i.e., hydrogels functionalized with protein-bound ligands) for controlled protein release. Contrary to traditional hydrogels that release proteins mainly through diffusion, affinity hydrogels stably retain and sustainably release proteins mainly based on diffusion coupled with a binding reaction. These hydrogels can also be modulated to release proteins in response to defined molecules in a triggered manner. Future research efforts may focus on the development of intelligent affinity hydrogels to mimic the properties of human tissues in sensing different environmental stimuli for on-demand release of single or multiple proteins (i.e., biomimetic intelligence for protein delivery).

Hydrogel, protein delivery and molecular recognition

Hydrogels are a three-dimensional network of hydrophilic polymers with the ability to absorb a large amount of water. Hydrogels can be fabricated from various polymers in different shapes (e.g., rectangular, cylindrical, spherical) and formats (e.g., coatings, and slabs). Because of these features, hydrogels have attracted significant attention for protein delivery.[1–3]

Proteins are essential molecules involved in nearly all biological processes. But they can be produced in vitro in a large scale with a relatively low cost thanks to advances in recombinant DNA technologies, cell culture and biological separation.[4] Thus, proteins are promising agents to treat various human diseases. However, in vivo protein delivery has been challenging due to their fragile structures, short in vivo half-lives and potentially high toxicity.[5] Thus, great efforts have been made over the past few decades in developing polymeric delivery systems, particularly hydrogels, for protein delivery.[1–3]

Human tissues can produce, retain and release signaling molecules at desired times with the right doses for the right duration. The fundamental mechanism governing these actions is molecular recognition. Molecular recognition is the specific noncovalent bonding between two or more molecules, which is primarily determined by electrostatic effects, hydrogen bonding, hydrophobic forces, and van der Waals forces.[6] Typical examples include protein-polysaccharide complexation, ligand-cell receptor association, enzyme-substrate binding, nucleic acid hybridization and nucleic acid-protein complexation, to name only a few.[7] For example, heparan sulfate (HS) in tissues can bind a large number of proteins for protein sequestration and release.[8] The application of molecular recognition for the development of hydrogels for protein delivery has received significant attention over the past two decades.

Traditional hydrogels for protein delivery

Traditional hydrogels are designed without the function of molecular recognition.[1–3] The first hydrogel developed for protein delivery was reported by Davis in 1972.[9] This hydrogel was made of polyacrylamide (PAA) and loaded with insulin to mitigate the symptom of diabetes. Davis also found that the diffusivity (see Glossary) of proteins decreased with increased molecular weights and exhibited an inversely logarithmic dependence on the polymer concentration.[10] Protein release could be prolonged for more than 5 weeks by increasing the PAA concentration from 5% to 40%.[10] Consistent with Davis’s works, Langer and Folkman found that over 50–90% of loaded proteins were released from hydrogels within the first several days.[11] Langer and Folkman prepared hydrogels by dissolving polymers in organic solvents and casting the mixture of polymer and protein in a mold. The protein-loaded hydrogel matrix could be further coated with additional layers of polymers to form a multilayered hydrogel implant with a ‘sandwich’ structure.[11]

These pioneering studies have clearly shown the promise of using hydrogels for protein delivery. However, they also show the challenges of protein delivery, including the fast protein release kinetics and the involvement of harsh factors in the procedure of preparing hydrogels.[10,11] To solve these problems, various covalent and physical crosslinking methods have been studied to synthesize hydrogels under mild reaction conditions. A typical example is a click reaction including thiol–ene reaction,[12] copper-free azide–alkyne cycloaddition,[13] Diels–Alder reaction, etc.[14] An alternative strategy of preparing hydrogels is physical crosslinking that generally does not involve toxic initiators, catalysts and/or organic solvents as it is achieved through ionic interactions, hydrogen bonds and hydrophobic interactions or a combination of these mechanisms.[15] For instance, alginate can form a hydrogel in the presence of calcium ions through ionic interactions.[16] Traditional hydrogels synthesized under these mild crosslinking conditions have been widely studied for protein delivery.[1–3]

Fundamental mechanisms of protein release from hydrogels

Multiple mechanisms account for protein release from hydrogels. For most hydrogels, they do not undergo instantaneous and significant swelling after transferred to a release environment.[17] In this scenario, diffusion is the dominant mechanism to control protein release from hydrogels (i.e., diffusion-controlled protein release) (Figure 1A).[18,19] Fick’s laws of diffusion can be used to describe the procedure of protein release reasonably well. However, some hydrogels can swell quickly and significantly after exposed to an aqueous environment. A typical example is hydroxypropyl methylcellulose-based hydrogel systems.[20] Drugs initially exist in an immobile state before the swelling of the hydrogels. However, once the hydrogels start a glass-to-rubbery phase transition during the swelling, drugs start to rapidly diffuse. The release rate of entrapped drugs is largely dependent on the degree of the swelling (i.e., swelling-controlled release). In addition to physically entrapping proteins within hydrogels, proteins can in principle be chemically conjugated to hydrogel chains. The degradation of hydrogel chains will be the rate-limiting step for protein release from these hydrogels (i.e., degradation-controlled release).[21]

Figure 1.

Mechanisms of protein release. (A) Protein release from traditional hydrogel is primarily controlled via diffusion. The abundance of free proteins in the hydrogel compared to the environment results in a high concentration gradient that drives the fast release of proteins from hydrogel. (B) Protein release from affinity-based hydrogel is controlled via the coupling of diffusion and reaction. The binding reaction between proteins and affinity ligands lowers the abundance of free proteins, which leads to a reduced concentration gradient and a show release rate of proteins.

Contrary to traditional hydrogels lacking the function of protein recognition, affinity hydrogels control the release of proteins primarily based on a reaction-diffusion mechanism (Figure 1B). These hydrogels are designed with specific ligands for protein recognition and physical immobilization (Table 1).[22–24] The commonly used ligands are heparin (Hp) and peptides. Recently, aptamers have started to get great attention. Certainly, hydrogels can also be designed with molecularly imprinted structures or functionalized with metal ions (e.g., Ni²⁺) for binding tagged proteins.[25,26] As proteins bind to their corresponding ligands that are chemically conjugated to the hydrogel chains, they will have both free and bound states in hydrogels. This binding reaction is a binding equilibrium reaction. The concentrations of free and bound proteins are governed by the equilibrium constant or the binding affinity between proteins and their ligands. As the concentration gradient of proteins is directly determined by their free state, the binding reaction basically reduces the concentration gradient of free proteins. Thus, molecular recognition or the binding reaction decreases the concentration gradient of proteins in the hydrogels. The driving force for protein release is reduced accordingly. As a result, protein release from affinity hydrogels can be slowed in comparison to traditional hydrogels that are without the function of molecular recognition.

Table 1.

Release mechanisms, advantages and disadvantages of hydrogels for protein delivery.

Type	Characteristics	Mechanism	Advantages	Disadvantages	References
Traditional hydrogels	• Proteins are physically retained in the matrix • No specific binding sites for protein drugs	Diffusion	• Simple	• Poor protein sequestration • Significant burst release	[10,11]
Heparin-based hydrogels	• Heparins serve as binding site for proteins • Binding affinity can be adjusted by varying heparin concentration and sulfation pattern	Reaction-diffusion	• Low burst release • Promiscuous interaction with a variety of proteins	• Safety concern due to the derivation from animal tissues • Low specificity	[33,44]
Peptide-based hydrogels	• Peptides serve as binding site for proteins • Binding affinity can be adjusted by varying peptide affinity and concentration	Reaction-diffusion	• Reduced burst release • High biocompatibility	• Low binding affinity • High peptide to protein ratio	[67,71]
Aptamer-based hydrogels	• Aptamers serve as binding site for proteins • Binding affinity can be adjusted by varying aptamer affinity and concentration	Reaction-diffusion	• Low burst release • High affinity and specificity • High biocompatibility • Spatiotemporal release control	• Limited availability of aptamers	[93,94]

Certainly, while the reaction-diffusion mechanism is dominant to determine protein release from affinity hydrogels, other mechanisms such as swelling and degradation of the hydrogel networks can affect protein release. During swelling and degradation, multiple factors such as mesh size, pore interconnectedness, water content, polymer volume fraction, and release boundaries are all changed. These factors are not independent. The change of one factor can affects the others. For instance, the gradual swelling of hydrogels will lead to the gradual increase of mesh size and pore interconnectedness. These factors and their changes essentially determine the freedom and random motion of proteins in a porous medium domain, i.e., apparent diffusivity of proteins in a dynamic hydrogel environment. It is also important to note that as a binding reaction is an equilibrium reaction, the equilibrium will be shifted to favor the dissociation of bound proteins from their ligands during the procedure of protein release. Moreover, when a competitive inhibitor is introduced to the system to change or inactivate the binding function of ligands, the equilibrium can be shifted more to accelerate protein dissociation.[27,28] With the enhancement of protein dissociation, the concentration gradients of free proteins can be increased for faster protein release.

Heparin-mediated protein release

Hp and HS are linear sulfated polysaccharides.[29] Hp is produced primarily by mast cells.[29] HS is a component of the extracellular matrix found in most tissues.[29] These polysaccharides bind a large number of proteins that ranges from extracellular matrix components to enzymes and growth factors mainly through electrostatic interactions.[30] Because of their capabilities of binding proteins, they have been applied to develop hydrogels for protein delivery.[30–32] For the simplicity, we use Hp to represent both Hp and HS in this review.

Edelman et.al first developed an Hp-functionalized hydrogel system for prolonged release of basic fibroblast growth factor (bFGF) that is a highly positively charged protein with an isoelectric point of ~10.[33] This hydrogel system was prepared by physically dispersing Hp-Sepharose beads in an alginate hydrogel. Nearly all bFGF was released from alginate hydrogels within the first two days. By contrast, only 20% of bFGF was released from the hybrid bead-alginate hydrogel during the first seven days. Notably, over 85% of Hp-bound bFGF maintained bioactivity at any time point during the release study whereas bFGF released from traditional hydrogels lost 99% of bioactivity.[33] This study successfully demonstrated that negatively charged Hp-Sepharose beads can bind positively charged proteins for sustained protein release, and that molecular binding can significantly help maintain protein bioactivity.

Hp can be directly incorporated into the hydrogel networks.[34–37] Hp is rich with hydroxyl and carboxylate groups.[38] These functional groups can be chemically modified for crosslinking with monomers or macromers. For instance, the hydroxyl groups can react with methacrylic anhydride to acquire methacrylate groups via esterification.[35] Methacrylated Hp can copolymerize with other molecules bearing methacrylate groups to form hydrogels through free radical polymerization.[35] Chemically Hp-functionalized hydrogels were found to release bFGF slowly over a 5-week time period, with nearly zero-order release kinetics after an initial burst release.[35] However, free radical polymerization involves toxic initiators, free radicals, and catalysts that can inactivate proteins. Mild reaction conditions are needed to avoid these factors. The carboxylate groups of Hp can be converted into thiol groups.[36] When thiolated Hp and proteins are mixed with macromers with diacrylate (e.g., PEGDA), Hp-functionalized hydrogels are formed via Michael addition.[36] Recently, Limasale et.al. synthesized hydrogel based on maleimide-bearing heparin and thiol-star-shaped poly (ethylene glycol) peptide conjugates crosslinked by a Michael type addition.[39] The heparin containing hydrogels were able to retain more than 97% of the loaded VEGF throughout the two weeks’ release period. The release of VEGF was further adjusted by varying heparin concentration. Increasing heparin concentration from 500 µM to 1500 µM allowed to decrease the overall release of VEGF by a factor of three.[39] Moreover, Hp can be chemically functionalized with azide, alkyne or dopamine groups for the formation of hydrogels based on click chemistry or dopamine-mediated self-crosslinking. [40,41]

Hp can also be incorporated into hydrogels via physical crosslinking.[42–44] The physical crosslinking will not only avoid harsh factors commonly involved in many chemical crosslinking methods, but also provide an opportunity to develop responsive, reversible and injectable protein delivery systems. As human antithrombin III binds to Hp, Hp-binding peptides (HBPs) have been derived from antithrombin III.[45] HBP-functionalized macromers can physically bind Hp to form a hydrogel. Hp immobilized within HBP-functionalized fibrin hydrogels has been demonstrated to control the release of nerve growth factor (NGF).[44] By designing HBP with different affinities and varying HBP to Hp ratios, one can tune Hp-HBP hydrogels for desired rheological behavior and protein release kinetics.[46] As Hp is highly charged, they can also form hydrogels with cationic polypeptides via electrostatic interactions.[43] Biodegradable ester bonds can be further incorporated into the backbone of the polypeptides to improve the biocompatibility and degradation of the hydrogels.[43] Heparin-mimicking molecules have also been studied to fabricate hydrogels. Many of them are sulfated polymers such as poly (vinyl sulfonate), poly-vinylsulfonic acid, poly-4-styrenesulfonic acid, sulfated alginate, sulfated hyaluronan and sulfated chitosan. [47–51] Heparin-mimicking peptides are another candidate to synthesize hydrogels for controlled protein delivery.[52,53] The peptides with a high density of sulfonate, hydroxyl and carboxylate groups can bind proteins via electrostatic interaction similar to Hp.[54]

Indeed Hp-functionalized hydrogels can sequester proteins and have been widely studied for protein delivery in regenerative medicine applications such as tendon and bone regeneration,[55] skin wound healing,[56] nerve regeneration,[57] spinal cord repair,[58] cardiac regeneration,[59] and tumor chemotherapy.[60] However, several issues need to be considered. The first issue is safety as they are derived from animal tissues (e.g., porcine intestines or bovine lungs). The second issue is relevant to its potency of anticoagulation. When Hp is used to synthesize hydrogels, its total amount is similar to that of monomers or macromers. The degradation of hydrogels will release a large quantity of Hp, which may cause local or systemic bleeding and thrombocytopenia.[61,62] The third concern is binding specificity. Hp sequesters proteins primarily based on electrostatic interactions.[30] Thus, Hp does not differentiate proteins. It is challenging to use Hp to control the release of multiple proteins with distinct kinetics. Moreover, Hp may not sequester neutrally or negatively charged proteins. It is also important to note that when they are implanted into a tissue, they may nonspecifically adsorb and retain inflammatory cytokines from the surrounding tissue and release these molecules gradually. It will reduce the biocompatibility of an implanted hydrogel. While the use of Hp is associated with these concerns, Hp-functionalized hydrogels are promising biomaterials for sustained protein release in various applications.

Peptide-mediated protein release

Specific protein-protein binding (e.g., antibody-antigen recognition) is common in nature. The binding between two proteins does not need the entire surface of the proteins but relies on a small area or a small portion of amino acids.[63] In principle, a peptide can be designed or discovered to replace one whole protein for recognizing the other one. Thus, in addition to mimicking heparin for binding proteins with electrostatic interactions,[52,53] peptides can be rationally designed for binding target proteins with specificity. Peptides can be derived from parent proteins either experimentally or through calculation and simulation.[64] Peptides can also be screened from peptide libraries against a specific protein target through phage display screening.[65] Phage display screening is performed via consecutive incubation, washing, amplification, and re-selection.[65] This reiteration increases the binding specificity and affinity of peptides against their cognate proteins. These peptides can be integrated into hydrogels for protein delivery via diffusion coupled with molecular recognition.

As amino acids have amino, carboxyl or thiol groups, numerous chemical methods are available for conjugating peptides with a hydrogel matrix. Free amines can react with polymers with reactive esters groups.[66] Thiol groups can react with polymers bearing vinyl sulfone, acrylate, or methacrylate groups via Michael addition or thiol-acrylate photopolymerization.[67] For instance, peptide-functionalized PEG hydrogels via thiol-acrylate photopolymerization have been studied for bFGF delivery.[67] The carboxyl or amine groups of peptides at the N- or C-terminus can also be chemically modified to acquire different functional groups.[68] The modified peptides can be crosslinked with macromers with corresponding reactive groups via mild reactions such as the dies-alder reaction [69] and the copper-catalyzed azide-alkyne cycloaddition (CuAAC) click reaction.[70] Protein release kinetics can be controlled by varying numerous parameters. One of the most important parameters is the peptide to protein ratio. Increasing this ratio can slow protein release from hydrogels due to the equilibrium shift to the formation of peptide-protein complexes. For instance, Delplace et al. studied the effect of peptide concentration on the release of chondroitinase ABC(ChABC), an enzyme that degrades the glial scar, from peptide-functionalized methylcellulose hydrogels.[69] The effective diffusivity of the loaded protein was reduced by two to three times with 50 to 100 times molar excess peptides.[69] In general, the ratio of protein: peptide often has to be high to achieve a slow release,[67] presumably because of steric hindrance that interferes with peptide-protein binding. Using linear polymers as a spacer has been studied to reduce this ratio and lower potential steric effects for stronger peptide-protein binding and stable protein sequestration.[67] Of course, the protein release kinetics can also be tuned by using peptides with different affinities.[71]

Peptides can be integrated into a therapeutic protein via genetic engineering to achieve peptide-mediated protein release. A typical example is the design of collagen-binding fusion proteins.[72] These proteins have two functional domains including a therapeutic protein and a collagen-binding peptide. When the fusion proteins are loaded into collagen hydrogels, the peptide binds to collagen for retaining the fusion proteins in the hydrogels.[73] This method has been applied to deliver various proteins such as stromal cell-derived factor-1α (SDF-1 α) for cardiac regeneration and bFGF for traumatic tympanic membrane and facial nerve reparation. [73–75] Fusion proteins with Src homology (SH3) domain are another interesting example. A recent study has shown that ciliary neurotrophic factor (CNTF) with the SH3 domain can be slowly released from hydrogels functionalized with SH3 binding peptides in the mouse retina. [76] The advantage of this method is that peptides do not have to be chemically incorporated into a hydrogel system. However, peptides may affect the functions of therapeutic proteins.

Peptides can also be integrated into hydrogels through self-assembly of peptide amphiphiles.[77] Peptide amphiphiles can self-assemble into nanofibers or nanofiber-based hydrogels under physiological conditions mainly driven by electrostatic repulsions and hydrogen bonding among charged amino acid residues and hydrophobic interactions of the alkyl tails.[78,79] When a protein-binding epitope is conjugated with peptide amphiphiles, the self-assembly leads to the formation of hydrogels with the protein-binding potential.[77] A recent study has shown that this type of hydrogel could sequester more proteins than the control hydrogel without the protein-binding epitope.[77] However, the protein release from the hydrogels with and without the protein-binding epitope was nearly the same.[77] As this protein-binding epitope itself does not participate in the self-assembly, the similar release patterns may result from the hiding of the epitope within the nanofibers or the low binding strength between the epitope and its cognate protein due to steric hindrance. Similar to this peptide assembly, protein-binding peptides can be incorporated into fibrin hydrogels during fibrinogen assembly for controlled protein release.[80] Peptides can be engineered to contain a substrate for enzymatic coupling. For instance, laminin peptides containing an additional factor XIIIa substrate sequence NQEQVSPL at the amino terminus were incorporated into fibrin hydrogels via transglutaminase activity of factor XIIIa during coagulation.[81]

In addition to self-assembly, peptide-based hydrogels can be developed through polyvalent peptide complexation.[82,83] For instance, the WW domain (i.e., a protein module with forty amino acids) binds to the proline-rich peptide specifically.[83] These peptides can be linked together to form a linear polymer with multiple repeating segments. When the linear polymers are mixed together, they can instantaneously react with each other to form a peptide-based hydrogel. During this polyvalent peptide complexation, a protein or peptide drug fused with a proline-rich motif can be incorporated into the hydrogel network through the same recognition mechanism.[83] This type of hydrogel has been studied for releasing a VEGF-mimicking helical peptide fused with a proline-rich motif. [83]

Peptide-functionalized hydrogels have been studied to deliver therapeutic proteins for various applications including bone regeneration,[84] cardiac recovery,[85] bladder regeneration,[86] and wound healing.[87] While encouraging results have been acquired, it is important to note that as peptides are selected segments of amino acids or derived from parent proteins. They lack the complex tertiary structures of natural proteins. Thus, their Kd values typically range from the micromolar to millimolar levels, suggesting their low binding affinities in comparison to their natural counterparts.[88] Peptides may also be hidden in the hydrogel network, which may reduce peptide-protein binding strength due to steric hindrance.[67] It may require a high concentration of peptides to compensate for their low affinities, which may bring a problem with cost.

Aptamer-mediated protein release

Aptamers are synthetic oligonucleotides selected from DNA/RNA libraries using a technique called systematic evolution of ligands by exponential enrichment (SELEX).[89,90] Aptamers have numerous merits for the development of affinity hydrogels.[91] They are synthesized chemically, which avoids batch-to-batch variation and derivation from cells or tissues.[92] Aptamers can be modified with different functional groups. They can maintain their binding functionality under harsh conditions for chemical modification or conjugation.[92] They bind to cognate proteins with high affinities and specificities similar to or even better than antibodies while they are one order of magnitude smaller than antibodies in size.[92]

Our group showed the first evidence of sustained protein release from hydrogels functionalized with aptamers as pendant branches for molecular recognition.[93] Native hydrogels released over 60% of loaded platelet-derived growth factor (PDGF-BB) during the first 12 hours.[93] By contrast, aptamers decreased the initial burst release from 60% to 10%. After this initial release, the daily release from aptamer-functionalized hydrogels was ~1–2%.[93] While this study was encouraging, we found several critical problems. We could detect PDGF-BB but could not detect any VEGF either from the release medium or within the hydrogels. It suggests that nearly all VEGF molecules were denatured presumably because free radicals denatured VEGF. It highlights the importance of avoiding free radicals or other harsh conditions during the preparation of hydrogels. It also highlights the necessity of measuring the bioactivity of released proteins instead of only the amount. In addition, while the amount of aptamers was several orders of magnitude lower than monomers, the efficiency of aptamer incorporation was not 100%.[93] This incompleteness limits effective protein loading. It may account for the presence of free proteins and the initial burst release.

An ideal hydrogel system needs to be an off-the-shelf biomaterial without proteins. It can absorb a freshly prepared protein solution right before an application. Such a hydrogel system also has a specific advantage, i.e., decoupling the synthesis of hydrogels from the loading of proteins. Thus, even if hydrogels are synthesized under a harsh condition (e.g., the presence of free radicals) as mentioned above, it does not affect proteins loaded at a later stage. Moreover, even if aptamers are not completely incorporated into the hydrogel network, they can be washed away before protein loading. We developed an off-the-shelf aptamer-functionalized superporous hydrogel using free radical polymerization coupled with gas formation.[94] The hydrogel had a pore size of 50 to 100 µm.[94] The superporous structure enabled the instantaneous absorption of PDGF-BB solutions. More importantly, when a high-affinity aptamer was used, the hydrogel released only 0.3% of loaded PDGF-BB during a 24-h incubation.[94] Notably, the hydrogel was not treated by pre-washing before the sequestration and release test. Thus, the aptamer-functionalized superporous hydrogel has a capability of sequestering proteins by virtually 100% and reducing the initial burst release to the level of nearly zero. When a low-affinity aptamer was used, the hydrogel released 60% of PDGF-BB in a sustained manner in two weeks.[94] Our findings have been confirmed by other research groups recently.[95–98] For instance, Enam et al. showed that aptamer-functionalized PEG hydrogels could sequester and passively release endogenous chemokines for recruitment of anti-inflammatory immune cells into sites of injury.[98]

Aptamers have shown great promise for the development of hydrogels. However, the scientific community has two major concerns with using aptamers to develop hydrogels or other biomaterials. One is availability. The sequences of many aptamers are not disclosed due to the consideration with intellectual properties. Researchers interested in one specific aptamer have to select aptamers by themselves. The other is stability. Similar to peptides, nucleic acids can be quickly degraded in biological fluids.[99] However, aptamers can be chemically modified either internally or at the 3’ and 5’ ends with a capability of significantly resisting nuclease degradation.[99] It is also important to note that the environment of a biological fluid is different from that of a hydrogel. Aptamers conjugated to the hydrogel can be protected by the hydrogel with a much longer half-life (unpublished), which is against the common sense with the instability of free and unmodified nucleic acids.

Hydrogels with biomimetic intelligence for protein release

Human tissues (e.g., islets) have the function of releasing signaling molecules based on the need of the body. An ideal hydrogel needs to be designed with this ability to mimic the intelligence of human tissues and release proteins at a right time with a right dose for a right duration (i.e., biomimetic intelligence for protein release). It also needs to have the capability of controlling the release of multiple proteins as mounting evidence suggests that treatment of human diseases such as tissue loss or cancer require more than one drug. [100,101] To meet this need, great efforts have been made to develop hydrogels for on-demand protein release in response to environmental stimuli.

Hydrogels can be stimulated to undergo degradation, swelling or shrinkage,[3] which leads to the change of structures of the hydrogels and subsequently the change of the apparent diffusivity and release rates of proteins. However, while these physical changes can be applied to manipulate protein release from hydrogels, it does not have the function of differentiating proteins if multiple proteins are loaded. In another word, different types of proteins in the same hydrogel would be released in the same or similar manner. It is also important to emphasize that a prerequisite for effective on-demand release is an ability to sequester proteins stably within the hydrogel network before the release needs to be started. Otherwise, automatic protein release would significantly compromise the effectiveness of on-demand release. Hydrogels functionalized with Hp, peptide and aptamer can stably sequester proteins in comparison to traditional hydrogels. These affinity hydrogels can be further designed to release proteins driven by environmental stimuli.

Hp or peptide can be hydrolyzed by enzymes existing in human tissues. This hydrolysis can accelerate protein release from hydrogels in response to the progression of a disease, which is promising for the treatment of various diseases. [33] Hydrogels have been designed to be proteolytically sensitive by the incorporation of a Matrix metalloproteinases (MMPs) sensitive peptide sequence (Figure 2A).[102,103] The release of protein from these hydrogel networks can be modulated due to the presence of MMPs.[103]

Figure 2.

On-demand protein release from hydrogels with biomimetic intelligence. (A) Protein release via enzymatic degradation of peptide-functionalized hydrogel networks. Adapted from ref.[102,103] (B) Protein release from aptamer-functionalized hydrogels triggered by complementary sequences (CSs). Adapted from ref.[94,107] (C) Sequential patterning of hydrogel with FAM-labeled anti-PDGF-BB and TAMRA-labeled anti-VEGF aptamers for spatiotemporal PDGF-BB and VEGF release triggered by CSs. Adapted from ref.[108]. (D) Traction force-triggered protein release. The 5’ end of a nucleic acid aptamer is conjugated to the hydrogel matrix and the 3’ end is conjugated with a cell-adhesive peptide (e.g., RGD) (a). The stretching induces the change of the aptamer’s conformation (b). Adapted from ref.[111]

Aptamers can hybridize with their complementary sequences in addition to binding their target molecules.[104–106] This hybridization brings a competitive inhibition mechanism into the procedure of controlling the release of aptamer-bound proteins (i.e., competition-driven release) (Figure 2B). Our group has shown that the daily protein release rate could be increased from less than 1% to ~10–20% via this molecular competition.[94,107] Moreover, multiple proteins such as VEGF and PDGF-BB can be triggered to release by the complementary sequences of aptamers from dual aptamer-functionalized hydrogels at defined time points with the right amounts.[28]

Aptamer-functionalized hydrogels have been further applied to spatiotemporally controlled protein release.[108] The Du group sequentially patterned anti-VEGF and anti-PDGF-BB aptamers to chitosan hydrogels by a photomask using UV light for capturing VEGF and PDGF-BB (Figure 2C).[108] The release of the protein drugs was triggered by adding the corresponding complementary sequences of the aptamers. This competitive inhibition can be further modulated to release proteins by biophysical stimuli (e.g., light).[109] Our group and the Almquist group also integrated aptamers and RGD peptides together to develop dual-affinity hydrogels.[110,111] The Almquist group showed that instead of using triggering complementary sequences, on-demand protein release could be realized through the dynamic change of cellular traction forces during the procedure of wound healing (Figure 2D).[111] These elegant studies have significantly enriched the development of hydrogels with biomimetic intelligence for controlled protein release.

Conclusions and Future Perspectives

Affinity hydrogels with ligands such as Hp, peptides and aptamers have a great potential to mimic human tissues in stably sequestering proteins for sustained protein release based on the diffusion-reaction release mechanism. Moreover, these affinity hydrogels can be further triggered by stimuli such as enzymes and complementary sequences of ligands for on-demand protein release at a defined time for a right duration. However, while encouraging results have been acquired, more efforts are needed in several research directions.

First, each ligand has its unique characteristics. Thus, the potential integration of different ligands such as aptamers and peptides together will enrich the development of multifunctional hydrogels for protein release (see Outstanding Questions). Second, triggering complementary sequences and enzymes are currently the main stimuli used to achieve on-demand protein release. However, human tissues and organs can sense a variety of environmental or metabolic stimuli for different and programmable changes. Biochemical and physical stimuli such as light, metabolites, pH, magnetic fields and traction forces are promising to regulate the binding functions of the ligands and the structures of hydrogels. Thus, it is possible to integrate these stimuli into affinity hydrogels to advance the control of protein release. For instance, protein release can be shifted from a reaction-diffusion mechanism to a reaction-diffusion-swelling mechanism when hydrogels are stimulated to swell quickly, which is largely unexplored. Third, human tissues and organs have multiple functions in addition to producing, retaining and releasing proteins. A hydrogel with biomimetic intelligence may need other functions (e.g., dynamic regulation of cell-hydrogel interactions) beyond the scope of only controlling protein release. Fourth, many elegant works were accomplished only based on chemical synthesis and/or in vitro cell studies. More works are needed to examine the therapeutic efficacy of affinity hydrogels in animals. More data from in vivo studies will convincingly accelerate the development of this field. Fifth, as most hydrogels designed for protein delivery are typically biodegradable, more studies are really needed to understand how degradation affects the reaction-diffusion procedure during protein release. This understanding requires not only more experimental studies but also mathematical modeling efforts. Definitely, imagination should not be limited to the five examples discussed herein. With more efforts made into this field, more hydrogels with unique and biomimetic functionality will be developed for protein delivery for real-world applications in the near future.

Outstanding Questions.

• What will happen if we integrate different types of ligands such as aptamers and peptides together for the development of multifunctional hydrogels for protein release?

• How does the degradation and swelling of affinity hydrogels affect protein’s diffusion and release kinetics?

• What is the difference in controlling protein release from affinity hydrogels between an in vitro release medium and an in vivo release environment?

• What is the bioactivity of released proteins in vitro and in vivo?

• Can we design protein-loaded affinity hydrogels with the ability to undergo dynamic changes in response to multiple environmental stimuli?

Highlights.

• Affinity hydrogels are developed by the functionalization of hydrogels with affinity ligands such as heparin, peptides and aptamers.

• Affinity hydrogels have the capability of sequestering proteins stably due to strong binding of proteins to their ligands.

• Affinity hydrogels control protein release mainly through the mechanism of diffusion coupled with a binding reaction.

• Affinity hydrogels can be designed to acquire biomimetic intelligence for on-demand protein release using specific triggering molecules.

Glossary

Affinity hydrogels

hydrogels functionalized with affinity ligands such as heparin, peptide, or aptamer for protein recognition

Aptamers

single-stranded DNA or RNA oligonucleotides capable of binding to targets with high affinities and specificities

Binding affinity

the binding strength of a biomolecule (e.g. protein) to its ligand (e.g. receptor, DNA)

Binding equilibrium

a state of binding reaction in which the rate of the forward association equals the rate of reverse dissociation

Binding reaction

the formation of a molecular complex due to a molecular interaction between two molecules such as a protein and a ligand

Biomimetic intelligence

referred to as designing specific molecules, structures, materials, devices or even tissues that can resemble or exceed their natural counterparts in certain functions

Click reaction

a class of reactions that are fast, versatile, and high yield, which are commonly used to conjugate a small biomolecule to a substrate of choice

Covalent crosslinking

the process of forming covalent bonds between polymer chains

Diffusion-controlled protein release

protein release is mainly governed by the diffusivity and concentration gradient of proteins and the release behavior can be well described using Fick’s laws of diffusion

Diffusivity

also known as diffusion coefficient that is a measure of the rate of solute transport due to the random thermal movement of solutes

Physical crosslinking

the process of forming a reversible linking between polymer chains via non-covalent interactions such as ionic interaction, hydrogen bonding, hydrophobic interactions and electrostatic interactions

Reaction-diffusion mechanism

protein release is mainly governed by the binding equilibrium of proteins and ligands and the procedure of passive diffusion