Innovations in Disease State Responsive Soft Materials for Targeting Extracellular Stimuli Associated with Cancer, Cardiovascular Disease, Diabetes and Beyond

Abstract

Recent advances in polymer chemistry, materials sciences, and biotechnology have allowed the preclinical development of sophisticated programmable nanomedicines and materials that are able to precisely respond to specific disease-associated triggers and microenvironments. These stimuli, endogenous to the targeted diseases, include pH, redox-state, small molecules and protein upregulation. Herein, we describe recent advances and innovative approaches in programmable soft materials capable of sensing the aforementioned disease associated stimuli and responding via a range of dynamic processes including morphological and size transitions, changes in mobility and retention, as well as disassembly. In this field generally, the majority of ongoing and past research effort has focused on oncology. Given this interest, we have chosen examples of the latest innovative approaches to chemo- and immunotherapy treatment strategies for cancer. Moreover, as the field broadens its attention, we highlight applications of programmable materials in other diseases, with a special focus on cardiovascular disease and diabetes mellitus, where limited attention has been paid by the field, but where many promising avenues exist with high potential impact.

Graphical Abstract

Over the past decade, the field of programmable soft materials for biomedical applications has rapidly expanded. Herein, we describe recent advances that show promising results in preclinical animal models for the treatment of cancer, cardiovascular disease and diabetes. Efforts have focused on materials capable of sensing the disease associated endogenous stimuli and responding via a range of dynamic processes.

1. Introduction

Nature has an established mastery over programmed, cascade and closed-loop pathways enabling living organisms to sense and respond to a vast array of endogenous and exogenous stimuli. This allows precise control over processes occurring at multiple time and length scales. Inspired by nature, a myriad of small molecules, nano- and micro-sized injectable materials, along with locally applied materials have been developed to respond to biological signals for safer and more efficacious therapeutics delivery.

At the small molecule level, the concept of programmable functional groups has been exploited for many years in the form of simple prodrugs. With this approach, drugs are chemically modified to inactive forms and then undergo programmed enzymatic or chemical transformations as a response to metabolic or physical and chemical signals to expose the active functionalities at the target site in vivo.^[1] This strategy often enhances the solubility, reduces the toxicity and improves the bioavailability of the parent drug.^[1] As an example, sulfasalazine, a disease-modifying antirheumatic drug, relies on bacterial metabolism in the colon to cleave the azo linkage and release the active sulfapyridine and 5-aminosalicylic acid.^[2]

This concept has been more recently adapted at the nano- and micrometer scale in the development of more complex materials. The resulting programmable responsive systems are a class of materials capable of dynamic transitions and function as a response to one or multiple triggers. While many materials have been designed to respond to exogenous, or applied stimuli such as light, ultrasound, and magnetic fields,^[3] or to intracellular low pH or reductive environments,^{[3a, 3b, 4]} responsive nanomedicine and materials that are able to recognize signals associated with specific disease states remain relatively underdeveloped. Such approaches offer extraordinary opportunities for reducing off-target effects and enhancing therapeutic indices. In the development of pathologies, diseased tissues exhibit specific biological features, which can be classified as physical, chemical and biomolecular differentiators that are associated with the disease state. Common disease associated physical cues include changes in pressure (shear force)^[5] and permeability^[6]. Chemical hallmarks can involve changes in pH and redox state.^[7] Biomolecular features can be extracellular (upregulated enzymes and biomolecule production)^[8] and intracellular (ATP and nucleic acids, among others)^[9]. Such markers help to differentiate diseased from healthy tissues and are useful in the design of disease-targeted programmable materials.

Herein, we will highlight innovations in programmable soft materials that respond to extracellular signals (e.g. pH, redox state, overexpressed proteins and biomarkers) endogenous to the diseased tissue. We will focus on recent advances on materials that span length scales, from injectable nanoparticles to locally applied materials that include those that are implantable, sprayable or wearable (Figure 1). In each case they have been designed to sense extracellular cues in the diseased tissue and respond through various changes in situ. This may include morphological and size transitions, changes in mobility and retention, assembly and disassembly. Given the generous amount of literature in the field we will limit our discussion to soft matter systems, and while the vast majority of examples comprise synthetic polymeric materials, special attention is given to novel biomolecule-based approaches and to the use of biohybrid material systems.

Figure 1.

Programmable responsive polymeric, biomolecular and biohybrid nano- and micrometer scale materials targeting extracellular disease-associated stimuli and their routes of administration. Icon for “Proteins”: Reproduced with permission.^[10] Copyright 2019, American Chemical Society. Icon for “Nucleic Acids”: Reproduced with permission. ^[11] Copyright 2018, Springer Nature. Icon for “Hydrogels/Scaffolds”: Reproduced with permission under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/). ^[12] Copyright 2019, The Authors, published by Springer Nature. Icon for “Implantable Materials”: Reproduced with permission under a Creative Commons Attribution-NonCommercial License 4.0 (http://creativecommons.org/licenses/by-nc/4.0/).^[13] Copyright 2017, The Authors, exclusive licensee American Association for the Advancement of Science. Icon for “Sprayable Materials”: Reproduced with permission.^[14] Copyright 2018, The Authors, published by Springer Nature.

By far, the majority of research efforts and focus in this field, has been on cancer. Therefore, we seek to cover some of the most innovative approaches for this disease. In addition, we include efforts to explore other important maladies especially cardiovascular disease (CVD) and diabetes. Although far less studied, it is clear that responsive materials have a role to play in these diseases. Hence, in the pages that follow, we have divided our discussion based on disease type, starting with cancer, and branching out. Where possible, we have avoided overlap between approaches, and have attempted to highlight examples that cover the wide gamut of materials in this space with a focus on novel, innovative strategies, in which therapeutic efficacy has been shown in vivo using relevant animal models. We aim to provide insight into recent developments in stimuli-responsive materials for biomedical applications and to highlight latest successes in preclinical settings. Future perspectives and challenges towards clinical translations are also critically assessed.

2. Programmable Materials for Cancer Therapy

According to the World Health Organization (WHO), cancer is the second leading cause of death worldwide, accounting for an estimated 9.6 million deaths in 2018 and an incidence of about 18 million cases worldwide, which are expected to rise to about 29.5 million by 2040.^[15] Despite significant advancements in diagnostic, precision surgery and radiation, chemotherapy has remained the leading cancer treatment over the past decade. Drawbacks associated with such therapy (e.g. off-target-derived side effects, toxicity and multidrug resistance processes) continue to limit clinical outcomes for cancer patients.^[16] For these reasons, the field of programmable responsive materials has been centered around efforts in oncology, with significant focus on establishing novel material-based approaches to selectively target tumors.

Nanotechnology is emerging as a promising tool for the selective delivery of cancer treatments. The potential advantages of nanotechnology, including tumor targeting and overcoming multidrug resistance, have most commonly been demonstrated by exploiting tumor accumulation in solid tumors and through the use of intracellular stimuli to achieve controlled release of chemotheraputics.^{[4a, 17]} The more recently characterized interplay between cancer cells and neighboring cells, including stromal and immune cells and the discovery of biochemical and physical alterations of the tumor microenvirnment (TME), suggested novel therapeutic targets, which point toward the design of novel active targeting approaches.^[18] Programmable polymeric, biomolecular and biohybrid materials have been developed to harness stimuli present in the TME to bias conventional chemotherapeutics and novel immunotherapeutics to localize in the TME.^{[19], [20]} There are a number of different physical and chemical features in the TME that can be exploited to this purpose (Figure 1). These include low pH,^{[7b, 21]} hypoxia,^[22] high enzyme concentrations,^[8] fibrosis,^[23] and the presence of reactive oxygen species (ROS).^[7a] Among these, acidic pH, enzymatic upregulation and ROS production have gained the most attention.^[24] The extracellular pH of tumor tissues is generally acidic, and acidic metabolites including lactic acid produced via anaerobic glycolysis in hypoxia, are the main cause.^{[7b, 21]} High concentrations of enzymes are also observed in the TME, in particular, matrix metalloproteinases (MMPs), which represent the most prominent family of protease associated with tumorigenesis, and are responsible for several key functions such as tissue remodeling, inflammation, tissue invasion, and metastasis.^{[8, 25]} ROS, including superoxide, hydrogen peroxide and hydroxyl radicals, are generated by the mitochondria and/or exogenous sources. They affect tumor immunity and promote the tumorigenic environment.^[7a]

While myriad nanomaterials have been designed to respond to such stimuli and promote enhanced cellular internalization via ligand exposure, drug release and charge switching,^{[3b, 26]} the remaining of this section provides an overview of the most recently developed programmable materials able to change size, morphology or mobility as a response to TME stimuli. In particular, novel emerging approaches to responsive discrete polymeric, biomolecular and biohybrid nanomaterials, as well as locally applied implantable, sprayable, or wearable materials that target and remodel the TME will be described. For each class of material, several promising platform technologies will be highlighted.

2.1. Programmable Polymeric Nanomaterials Targeting the TME

The rapidly growing field of controlled and functional group tolerant polymerization methods makes synthetic polymers the most exploited class of programmable materials for the design of novel cancer therapies.^[27] Such polymers are available in a wide range of compositions with readily adjustable structures and properties. Myriad approaches exploiting TME cues to trigger the release of therapeutics, as well as to activate nanoparticles and prodrugs prior cellular uptake have been proposed.^[26] Herein we take the opportunity to highlight some recent advances on the design of size-tunable strategies.^[24] In particular, we will focus our attention on polymeric materials undergoing a size and morphology switching as a response to endogenous TME stimuli, with the most exploited being acidic pH and the presence of overexpressed enzymes. Such materials can be classified into two different classes: 1) shrinkable/disassembling materials, that upon TME localization reduce in size favoring further tumor penetration, and 2) accumulation/assembling materials, that upon TME localization change size and morphology, promoting longer retention in the TME and slow release of therapeutics.^[24]

The concept of multistage shrinkable materials was first introduced using a protein-hybrid nanomaterial (see “Programmable Biomolecular and Biohybrid Nanomaterials Targeting the TME” section).^[28] This approach involves 100–200 nm nanocarriers, which can passively and safely accumulate in the TME, and that are able to disassemble into smaller sub-20 nm structures as a response to TME-associated cues. The smaller fragments can then penetrate efficiently into tumors and enhance the therapeutic index of the delivered drug. The acidic pH of the TME offers the opportunity to tailor pH-responsive polymer nanocarrier shrinkage and to promote deeper tumor penetration. For example, smart polymeric clustered nanoparticles (iCluster, Figure 2) with an initial size of∼100 nm, were able to trigger the discharge of platinum prodrug-conjugated poly(amidoamine) dendrimers (PCL-CDM-PAMAM/Pt, diameter∼5 nm) by cleavage of 2-propionic-3-methylmaleic anhydride (CDM)-based acid sensitive linker (Figure 2A-B).^[29] This structural alteration facilitated tumor penetration (Figure 2C) and cell internalization of the therapeutics. The superior in vivo antitumor activity of iCluster compared to the non-shrinkable (pH stable) assemblies, was validated in tumor models including poorly permeable pancreatic cancer, drug-resistant cancer (Figure 2D), and metastatic cancer, demonstrating the versatility and broad applicability of this method across tumor types with varying levels of aggressiveness.^[29]

Figure 2. pH-shrinkable, Pt-containing polymeric clustered nanoparticles (iClusters/Pt) enable deeper tumor penetration.

(A) Chemical structure of PCL-CDM-PAMAM/Pt and nanoparticles building blocks. (B) Self-assembled iCluster/Pt and subsequent structural change as a response to tumor acidity and intracellular reductive environment. (C) Real-time distribution of RhBiClusterFlu in BxPC-3 tumor at 10, 90, and 240 min post intravenous (i.v.) injection as determined by confocal fluorescence microscopy. Scale bar: 100 μm. PAMAM was labeled with fluorescein (green), whereas the core of the nanoparticles was labeled with Rhodamine B (red), and blood vessels were marked with platelet endothelial cell adhesion molecule 1 (PECAM-1) and CFL-647 secondary antibody (yellow). (D) Inhibition of tumor growth in a A549R cisplatin-resistant human lung cancer model. Mice were i.v. administered an equivalent platinum dose of 1.5 mg/kg on days 0, 3, and 6. “Cluster/Pt” pH stable assemblies were used as a control. Reproduced with permission.^[29] Copyright 2016, The Authors, published by National Academy of Sciences.

Polymer directed assembly of platinum-prodrug conjugated polyamidoamine (PAMAM) dendrimers, in which the amphiphilic polymer contains ionizable tertiary amine groups for a more rapid pH-responsiveness was also developed by the same group.^[30] These superstructures had an initial size of ∼80 nm at neutral pH, but underwent a rapid dissociation into dendrimer building blocks (less than 10 nm in diameter) upon a narrow pH drop (less than 0.1–0.2 pH units) in the TME. The rapid size-switch facilitated drug penetration into the poorly permeable BxPC-3 pancreatic tumor models, improving therapeutic efficacy of the delivered chemotherapeutics. Moreover, this concept served as the foundation for the development of more complex hybrid materials. As an example, pClusters, which rely on polymer protonation in the lower pH of the tumor to release the smaller anticancer peptide-Au nanohybrids (approximately10 nm in size), showed successful cellular internalization and inhibition of tumor growth and metastasis in several animal models. pClusters were also shown to synergize with the programmed cell death-1/programmed death-ligand 1 (PD-1/PD-L1) checkpoint blockade immunotherapy.^[31] Additionally, other pH-triggered shrinkable inorganic composite delivery systems have been recently developed, which description goes beyond the aim of this review.^[32]

Besides pH, MMPs overexpression in the TME has also been used to trigger changes in nanoparticles size and morphology upon accumulation in the TME.^[33] While MMPs have generally been used to disassemble delivery systems or cleave prodrugs,^{[25b, 26]} enzymatic digestion has also been used to trigger the opposite effect. Specifically, enzymatic digestion of the hydrophilic component of block-copolymer-assembled micelles promoted aggregation and formation of intratumoral drug depots, leading to selective TME targeting, prolonged retention and sustained release of the carried therapeutics following systemic administration (Figure 3).^[34] This strategy is referred to as enzyme-directed assembly of particle therapeutics, or EDAPT. These materials are prepared through the assembly of amphiphilic diblock copolymers to generate spherical micellar nanoparticles. The polymers are designed such that the hydrophobic block, which forms the core of the spherical micelle, covalently carries the therapeutic of interest. The hydrophilic polymer block, forming the shell of the micelle, is comprised of MMP-responsive peptides. The spherical shape and small size (around 10–20 nm) of the nanoparticles allow safe intravenous administration. Consequent diffusion of the particles into tumor tissue, followed by enzymatic cleavage of the peptides in the TME, results in a transition from small spherical nanoparticles to micrometer-scale scaffolds that can be retained for weeks by means of interaction with the surrounding extracellular matrix, where the loaded therapeutics can be slowly released (Figure 3A).^[34]

Figure 3. EDAPT targeting and retention of nanoparticles (NPs) in the TME.

(A) Targeting and retention of NPs in the TME by action of the overexpressed MMPs as compared to nonresponsive NPs and healthy tissues. (B) Structure of the amphiphile di-block copolymer where the TLR-7 (1V209, in red) is incorporated into the hydrophobic block and the hydrophilic block is composed of the MMP-responsive peptide (purple) and the fluorescent dye Cy5.5 (green), as well as schematic representation of the MMP-responsive NPs obtained by polymer assembly in water. (C) Transmission electron microscopy (TEM) images showing the morphology of the NPs before and after enzymatic treatment. Scale bars: 5 μm. (D) Plasma concentration of IL-6, MCP-1, and IP-10 cytokines at 2 h post-(i.v.) injection of responsive (1V209-L-NP), non-responsive (1V209-D-NP), free drug (1V209) and vehicle (PBS). (E) Timeline of the in vivo experiment using the 4T1 tumor model and (F) number of metastatic tumor foci deriving from the 4T1 murine syngeneic breast cancer model per analyzed H&E stained lung section on day 29 post-injection. Mice were injected intravenously with either vehicle, free 1V209 or with MMP-responsive NPs, either empty, (L-NP_F) or containing 1V209 (1V209-L-NP_F). Reproduced with permission.^[34e] Copyright 2019, The Authors, published by Wiley-VCH.

This concept has been demonstrated for the delivery of the anticancer drug paclitaxel (PTX) in a subcutaneous murine xenograft tumor model of human cancer,^[34c] for the delivery of an oxaliplatin analogue,^[34d] and it has been recently exploited for the systemic delivery and targeting of a small molecule immunotherapeutic.^[34e] Incorporation of the small molecule Toll-Like receptor agonists 7 (TLR-7), 1V209, into the hydrophobic block of the copolymer, resulted in the formation of 10–20 nm nanoparticles able to respond to MMPs both in vitro (Figure 3B-C) and in vivo (Figure 3D-F). Intravenous administration of such formulation prevented systemic immune activation, but maintained efficacy in a syngeneic orthotopic 4T1 breast cancer model and significantly reduced lung metastases formation.^[34e]

2.2. Programmable Biomolecular and Biohybrid Nanomaterials Targeting the TME

While synthetic polymers represent the widest class of responsive materials for drug delivery, they are often associated with the risks to induce inflammatory/immunogenic responses. Therefore, naturally exsiting biopolymers are emerging as promising nanocarriers because of their inherent biocompatibility and in some cases biodegradability. A series of protein-, polysaccaride- and oligonucleotide-based delivery vehicles have been reported as novel nanomaterials for cancer therapy.^[35] In the following section, a few representative examples of programmable biopolymer and biohybrid delivery systems that respond to local TME stimuli with either shrinkage/disassembly or accumulation/assembly and selectively delivery of the carried therapeutics are discussed in details.

The abundance of synthetic and bioconjugation tools and appealing properties including long in vivo half lives, biological stability and targeting ability, make proteins one of the most exploited natural delivery vehicles.^[36] Among the different examples reported, gelatin stands out owing to its reduced cost and its synthetic versatility in manufacturing.^[37] Moreover, as the protein derivatives of collagen, gelatin is a good substrate for MMPs and therefore can serve as a building block for the design of TME responsive nanomaterials. Indeed, quantum dot conjugated gelatin nanoparticles (QDGelNPs) were the first reported example of multistage nanoparticles used for cancer treatment.^[28a] In this pioneering work, 100-nm QDGelNPs, which favor long blood circulation and extravasation through tumor vascular fenestrations, successfully released 10-nm quantum dots from their surface upon TME accumulation and MMP digestion. The small released quantum dots were shown to better penetrate into dense tumor collagen matrices.^[28] When gold nanoparticles decorated with the chemotherapeutic agent doxorubicin were delivered using this carrier, enhanced penetration into the tumor mass resulted in successful tumor growth inhibition in both a glioma tumor model,^[38] as well as in the 4T1 breast cancer model.^[39] This work provided the basis for the development of diverse polymerc pH shrinkable delivery systems, some of which are discussed in the previous section of this review.^{[29], [30]}

Serum Albumin (SA) is another versatile protein, whose biocompatibility, non-immunogeneicity and extended blood circulaton makes it an ideal candidate for drug delivery applications.^[40] Furthermore, albumin-mediated accumulation in tumors has been exploited for the delivery of several chemotherapeutic agents.^{[40b, 41], [10]} Among the reported human serum albumin (HSA)-based delivery systems is the FDA approved Abraxane, which uses the protein as an excipient and a non-stimuli responsive disperse micro- to nanoscale PTX formulation.^[42] Similar to gelatin shrinkable delivery systems, SA based carriers able to change size upon accumulation in the TME have been reported.^[43] In one example, a SA-drug conjugate was complexed with magnesium oxide nanoparticles to form a redox and pH responsive hybrid delivery systems. SA was pre-conjugated with either chlorine e6 (Ce6), a photosensitizer for photodynamic therapy (PDT), or c,t,c-[Pt(NH₃)₂-(O₂CCH₂CH₂COOH)(OH)Cl₂] (cis-Pt(IV)SA), a pro-drug of cis-platinum. Such constructs were then used as the template to induce the formation of MnO₂ nanoclusters through biomineralization in alkaline conditions, resulting in the synthesis of a hybrid HSA-MnO₂-Ce6&Pt (HMCP) nanoparticles. Upon accumulation in the TME, the nanoparticles simultaneously generated O₂ by reacting with endogenous tumor-associated H₂O₂ and disassembled into the SA-drug conjugates of 10 nm when exposed to the low pH of the TME. While O₂ generation is promising to avoid hypoxia associated resistance to PDT, the small size of the released SA-drug conjugates allowed successful tumor penetration. Overall, the superior tumor penetration achieved using this material improved the combination treatment outcome in the 4T1 breast cancer model.^[43]

In contrast to shrinkable materials able to promote tumor penetration, assembly and retention of nanomaterials at the tumor site can also be used as a strategy to slowly release therapeutics of interest and modulate conditions within the TME. While several examples of tumor accumulation and retention of polymer nanomedicines in the TME are available (see “Programmable Polymeric Nanomaterials Targeting the TME” section), the first example of protein-bound therapeutics targeting the tumor extracellular matrix has only recently been reported. This approach was initially developed to target and retain immunotherapy to the TME upon systemic administration.^[44] A similar concept was then developed to localize and retain chemotherapy in the TME.^[45] In this work, targeting was achieved using a biohybrid delivery vehicle composed of two proteins, SA and the collagen binding domain (CBD) of von Willebrand factor (Figure 4).^[45] SA was recombinantly fused with the CBD to bind within the tumor stroma after extravasation in the tumor extracellular matrix. Doxorubicin (Dox) was conjugated to the CBD-SA via a pH-sensitive linker, which allowed rapid tumor accumulation 2 h post-injection and subsequent sustained Dox release from the formed collagen-bound depots in the acidic TME as well as tumor cell internalization of the Dox-CBD-SA (Figure 4A-B). Tumor targeting and retention of Dox-CBD-SA significantly decreased doxorubicin-associated adverse effects and suppressed tumor growth compared to both Dox-SA and doxorubicin treatment in a mouse model of breast cancer. Moreover, this approach efficiently stimulated antitumor immunity in the MC38 colon carcinoma when used in combination with anti-PD-1 (αPD-1) checkpoint inhibitor (Figure 4C-E).^[45] The ability to use natural and therefore biocompatible and biodegradable carriers to target the abundant and accessible collagen matrix in tumors open new avenues in the field of programmable materials for cancer therapy, with many potential applications still unexplored.

Figure 4. Biohybrid material composed of SA and the CBD of von Willebrand factor promote SA binding to the extracellular matrix, tumor retention and sustained Dox release.

(A) Synthetic scheme of Dox-CBD-SA, in which the anticancer drug (in red) is linked to the fused proteins (black line) via a pH sensitive linker (in blue). (B) Representative images of tumor tissue upon intravenous administration of dye-labeled SA and CBD-SA at one hour post-(i.v.) injection. DyLight 488–labeled SA and DyLight 488–labeled CBD-SA (green) were injected intravenously in MMTV-PyMT tumor-bearing mice. One hour post-injection, tumors were harvested and fluorescence was analyzed by confocal microscopy. Tissues were also stained with 4′,6-diamidino-2-phenylindole (DAPI, blue) and anti-CD31 antibody (red). Scale bars: 100 μm. (C) Timeline of in vivo experiment using the MC38 tumor model. Mice were injected intravenously with aldoxorubicin or Dox-CBD-SA (5 mg/kg on a Dox basis) on days 6, 9, and 12. αPD-1 was also injected intraperitoneally on days 10 and 13. (D) Tumor volume (mean ± SEM) and (E) survival rate. Reproduced with permission.^[45] Distributed under the terms of the Creative Commons Attribution-NonCommercial License 4.0 (http://creativecommons.org/licenses/by-nc/4.0/). Copyright 2019. The Authors, published by AAAS.

Besides proteins, polysaccharides and their derivatives, such as alginate, chitosan, hyaluronic acid (HA), and dextran (DEX) are commonly used in drug delivery for the preparation of nanocarriers.^[46] These biopolymers undergo spontaneous enzymatic and hydrolytic degradation in vivo, resulting in safe degradation products that can either be reused or cleared by the body.^[46a] Additionally, these biopolymers consist of a large number of hydroxyl and other hydrophilic groups, such as carboxyl groups in alginate and amino groups in chitosan, which make them water soluble and can serve as biorecognition sites. Such functionalities can be easily reacted with drugs, synthetic polymers or with responsive moieties to generate programmable materials. While different TME responsive functionalities have been incorporated into the polysaccharides via chemical modifications, in particular pH-responsive groups,^[46a] we include here an interesting example describing the use of pH responsive polysaccharide triggering a change in TME morphology, therefore enhancing tumor penetration and upregulating the therapeutic response of established treatments. In this example, the enzyme hyaluronidase (HAase) was modified with DEX via a pH-responsive traceless linker.^[47] The formulated pH-responsive DEX-HAase nanoparticles showed enhanced enzyme stability, reduced immunogenicity and prolonged blood half-life after intravenous injection. Upon efficient passive tumor accumulation, DEX-HAase within the acidic TME dissociated to release native HAase, which in turn triggered the breakdown of HA to loosen the extracellular matrix (ECM) structure. This process promoted the therapeutic effect of nanoparticle-based photodynamic therapy (PDT) owing to enhanced tumor penetration. Moreover, the combination of PDT and anti-PD-L1 (αPD-L1) checkpoint blockade immunotherapy resulted in remarkable synergistic effects and inhibited the growth of both primary and distant tumors.^[47]

Oligonucleotides are another class of biopolymer and both DNA and RNA have been extensively incorporated into drug delivery vehicles owing to their potential therapeutic applications.^[48] On the contrary, their function as drug delivery vehicles have been underutilized due to their poor in vivo stability and complex chemistry. However, significant efforts have been made recently in drug delivery utilizing DNA suprastructures for cancer targeting. In the past couple of decades, the advent of DNA origami has allowed the implementation of complex three-dimensional shapes from a single-stranded DNA molecule.^[49] Harnessing the high fidelity of Watson–Crick DNA pairs and advanced computational methods, rationally designed nano-constructs were fabricated to precisely target cancer tissues.^[50] Although this field is still in its infancy and biomedical applications are still challenged by the limited structural integrity of DNA in complex biological fluids,^[51] recent combinations of DNA origami and other responsive components highlight the potential of this novel field in the preparation of next generation programmable materials.^{[50b, 52]} DNA origami delivery vehicles able to sense the surrounding environment and respond with changes in morphology have been recently developed (Figure 5).^{[11, 53]} In this work, DNA origami was used to prepare an autonomous DNA “robot” programmed to transport payloads and present them specifically to tumors. This tubular protein-DNA nanorobot contained a DNA aptamer that binds nucleolin, a protein specifically expressed on tumor-associated endothelial cells, on the outside and carried a blood coagulation protease thrombin within its inner cavity. The nucleolin-targeting aptamer served both as a targeting domain to promote accumulation and retention of the nanodevice in the TME, and as a molecular trigger for the mechanical opening of the DNA nanorobot (Figure 5A-B).

Figure 5. DNA origami, so-called “nanorobot” responds to the TME-associated protein nucleolin triggering tumor necrosis.

(A) Schematic illustration of the construction of thrombin-loaded nanorobot (nanorobot-Th) and its transition to a rectangular DNA sheet as a response to nucleolin binding. (I) Single-stranded M13 phage genomic DNA is linked by predesigned staple strands, leading to the formation of a rectangular DNA sheet. (II) Thrombin is loaded onto the surface of the DNA sheet structure by hybridization of poly-T oligonucleotides conjugated to thrombin molecules with poly-A sequences that extend from the surface of the DNA sheet. (III) Addition of the fasteners and targeting strands results in the formation of thrombin-loaded, tubular DNA nanorobots with additional targeting aptamers at both ends. (IV) The tube nanocarrier opens as a response to the presence of nucleolin and exposes the encapsulated thrombin. (B) Schematic representation of the therapeutic mechanism of nanorobot-Th within tumor vessels. Upon i.v. administration, the nanorobot-Th binds to the vascular endothelium of the tumor by recognizing nucleolin and opens to expose the encapsulated thrombin, inducing localized thromboses, tumor infarction and finally cell necrosis. (C) MDA-MB-231 tumors harvested before and 24 and 72 h after administration of nanorobot-Th were stained with H&E to detect necrosis (necrotic tissues are denoted by N). Scale bars: 200 μm. Reproduced with permission.^[11] Copyright 2018, The Authors, published by Springer Nature.

Intravenous delivery of such programmed DNA materials successfully localized thrombin in the tumor associated blood vessels and induced intravascular thrombosis, resulting in necrosis (Figure 5C), which in turn inhibited the growth of the tumor in the MDA-MB-231 tumor model. Moreover, the nanorobot proved to be effective in a B16F10 syngeneic melanoma tumor model as well as in poorly vascularized subcutaneous xenografts generated from human ovarian cancer cells SK-OV3. Although in this case the inhibitory efficiency of the DNA nanorobot was not as remarkable as in the melanoma model, the results indicate that limited tumor permeability due to poor vascularization does not prevent the nanorobot from exerting its antitumor activity.

2.3. Programmable Polymeric, Biomolecular and Biohybrid Implantable, Sprayable and Wearable/Transdermal Materials for Localized Cancer Treatment

So far, we have highlighted examples of discrete polymeric and biomolecular nanocarriers capable of sensing cancer-associated cues and undergoing size/morphological transitions. Indeed, a great portion of research in the field of responsive cancer drug delivery systems relies on these types of micro- and nanomaterials. On the other hand, locally applied drug reservoirs offer a simple alternative to overcome the weaknesses of conventional therapeutic methods, such as lack of selectivity and inefficient local drug concentrations. When possible, local application of biodegradable materials results in targeted, localized drug depots able to assure constant high concentrations of the therapeutics of interest and TME modulation.^{[54], [55]} Such devices can be broadly categorized as implantable materials, such as preformed or injectable scaffolds and hydrogels,^{[55a, 56]} sprayable,^[14] and wearable devices like microneedle patches for transdermal delivery.^[57]

The first reported examples of implantable materials relied on biodegradability as a way to release cargo, with synthetic polymers such as poly(lactic-co-glycolic acid) (PLGA), poly(ε-caprolactone) (PCL), poly(glycerol monostearate-co-ε-caprolactone) (PGC-C18),^[55] polyethylene glycol (PEG) and Pluronic F-127 (Poloxamer 407) as the most exploited building blocks.^[55b] Biologically-derived scaffolds and hybrid systems comprising proteins (i.e. gelatin and collagen), polysaccharides (i.e. DEX, chitosan and alginate) and aminoglycans such as HA and derivatives have also been widely used owing to their inherent biocompatibility and in some cases biodegradability.^[55b] Many literature reviews have summarized the use of these materials to create chemotherapeutic depots able to degrade in vivo and promote sustained drug release into the TME.^{[55, 58]} However, programmable materials have a role to play in this approach, as the ability to tune drug release kinetics in response to TME stimuli can significantly impact treatment performance. In a recent study, acetylated dextran (Ace-DEX) nanofibrous scaffolds were used to generate implantable PTX depots characterized by either fast, medium, and slow degradation. This resulted in 14.1%, 2.9%, and 1.3% PTX release per day in a glioblastoma tumor model. Critically, only the fast release, which is achieved through the use of acid-sensitive Ace-DEX moieties, displayed treatment efficacy in the distant metastasis model of glioblastoma. This approach enabled faster degradation upon exposure to the lower pH conditions associated with the tumor, releasing more PTX in vivo and resulting in higher antitumor activity as compared to nonacid sensitive scaffolds.^[59]

While initial efforts on the design of stimuli responsive local treatments was centered around chemotherapy delivery, their focus has recently shifted to the modulation of the TME using chemoimmunotherapy combinations as well as immunotherapy. Several recent reviews provide detailed and comprehensive descriptions of different engineered implantable (and injectable) materials developed for this purpose.^{[55a, 60],[20b, 61], [83]} Here we will provide some innovative examples of devices for local delivery of chemo-immunotherapy, which respond to triggers associated with the TME. In a recent paper, a hydrogel was designed to locally release both the anticancer drug gemcitabine (GEM) and an αPD-L1 blocking antibody upon injection into the tumor site as a response to the highly abundant ROS within the TME.^[62] The ROS-responsive hydrogel was obtained by crosslinking poly(vinyl alcohol) (PVA) with the ROS-labile linker N1-(4-boronobenzyl)-N3-(4-boronophenyl)-N1,N1,N3,N3-tetramethylpropane-1,3-diaminium (TSPBA). Peritumoral injection of the hydrogel in the B16F10 melanoma mouse model resulted in in vivo oxidation and hydrolysis of TSPBA upon H₂O₂ exposure in the TME with consequent GEM and αPD-L1 release. Confocal imaging of ex vivo tumor sections revealed that the fluorescein signal associated with GEM was evident after one day post injection, whereas the signal corresponding to αPD-L1 increased gradually within 3 days. Control over release kinetics facilitated the desired sequential effects. The faster release of GEM promoted an immunogenic tumor phenotype resulting in enhanced PD-L1 expression on tumor cells and therefore enhanced the therapeutic effect given by the slower local release of the αPD-L1 antibody. This result was further demonstrated in the 4T1 breast tumor model and was additionally validated by the ability to prevent tumor recurrence after primary tumor resection.^[62] A similar effect was also achieved using an injectable peptide progelator system for the local release of the αPD-1 antibody in combination with the anticancer drug camptothecin (CPT).^[63] This combination induced a similar immune-stimulating phenotype boosting the effect of the PD-1 blockade immune response. The system relied on the use of an amphiphilic progelator obtained by conjugating a hydrophilic iRGD, a peptide sequence facilitating tumor tissue penetration, to two hydrophobic CPT molecules through a matrix metalloproteinase 2 (MMP-2) responsive linker (PLGLAG peptide). The two CPT moieties were attached to the PLGLAG peptide through a reducible intracellular disulfanyl-ethyl carbonate (etcSS) linker. This diCPT-PLGLAG-iRGD prodrug spontaneously assembled into supramolecular micrometer scale nanotubes in aqueous environments. Codelivery of the αPD-1 antibody was achieved by simply mixing the antibodies with the diCPT-PLGLAG-iRGD before gelation. Upon local injection, the release of the drug combination by action of the TME overexpressed MMPs resulted in 100 % tumor regression for all treated mice in both the GL-261 brain cancer and CT 26 colon cancer models.^[63] A similar concept has been reported by the same group in the co-delivery of CPT in combination with a cyclic di-nucleotide STING agonist with the aim of stimulating both adaptive and innate immunity.^[64]

Another novel, local approach for immunotherapy delivery relies on the use of sprayable materials to modulate the TME post-surgery and maintain a tumor suppressive environment. A biohybrid protein system has been recently designed to prepare a sprayable pH responsive immunotherapeutic gel that inhibits local tumor recurrence after surgery and development of distant tumors.^[14] In this work, CaCO₃ nanoparticles were incorporated into a fibrin gel, formed by the interaction of fibrinogen and thrombin and served as a release reservoir of immunomodulatory therapeutics as well as a proton scavenger to modulate the acidity of the tumor environment (Figure 6).^[14] The fibrinogen solution containing anti-CD47 antibody-loaded CaCO₃ nanoparticles (αCD47@CaCO3) and thrombin solution could be quickly sprayed and mixed within the tumor resection cavity after surgery to form an immunotherapeutic reservoir (Figure 6A-C). Within the acidic TME, the incorporated CaCO₃ nanoparticles gradually dissolved and released the encapsulated αCD47, promoting the activation of M1-type tumor associated macrophages (TAMs) and inducing macrophage phagocytosis of cancer cells via blockade of the CD47 and SIRPα interaction as well as boosting antitumor T cell responses. The ability of this system, to awaken the host innate and adaptive immune systems while reducing the toxic effects associated with the systemic administration of αCD47, was demonstrated in a B16F10 tumor model resulting in both inhibition of both local tumor recurrence post-surgery and potential metastatic spread (Figure 6D-E).

Figure 6. A biohybrid, pH responsive immunotherapeutic gel inhibits local tumor recurrence after surgery.

(A) Schematic representation of the hybrid fibrinogen, thrombin and αCD47@CaCO₃ nanoparticle building blocks and (B) in situ sprayed bioresponsive fibrin gel containing αCD47@CaCO₃ nanoparticles within the post-surgery tumor bed. αCD47@CaCO₃ nanoparticles encapsulated in fibrin scavenged the H⁺ in the tumor bed with subsequent release of the αCD47. This promotes both polarization of TAMs to an M1-like phenotype and blockade of the “don’t eat me” signal in cancer cells. (C) Representative cryo-scanning electron microscope (SEM) images of gel loaded with αCD47@CaCO₃ nanoparticles. Scale bar: 1 μm. (D) Systemic immune system activation was assessed by investigating the effect on distal B16F10 melanoma tumor. Tumors on the right side of the animals were designated as ‘primary tumors’ with αCD47@CaCO₃@fibrin treatment, and those on the left side were designated as ‘distant tumors’ without any treatment. (E) Representative photographs of mice at day 22 highlight the inhibition of the growth of both primary and distal tumors (blue arrows). Reproduced with permission.^[14] Copyright 2018, The Authors, published by Springer Nature.

Finally, a recently emerging, minimally invasive approach for the delivery of immunotherapy is the use of materials acting as wearable transdermal biomaterials.^{[57, 65]} This delivery approach is nearly painless and it is effective for delivering drugs directly to the upper epidermis or dermis. These microneedle (MN) arrays can contain hundreds of micrometer-long needles which can be made from a variety of materials including metals, polymers, glass, and ceramics, and which are applied using patches to create micrometer-scale insertions through the skin.^[65] While the majority of MN-based approaches have been used to deliver drugs via biodegradation/biodissolution in physiological environments,^{[57, 65–66]} or less frequently upon exogenous stimuli,^{[57, 65]} one recently reported example of a pH responsive MN might pave the way for the development of next generation cancer-responsive hybrid transdermal devices. In this work, a pH responsive system was designed for the treatment of skin cancer.^[67] Glucose oxidase (GOx), which converts blood glucose to gluconic acid was incorporated into microneedle delivered pH responsive NPs. Catalase (CAT) was also used to assist glucose oxidation via the generation of O₂ and consumption of undesired hydrogen peroxide. Specifically, these nanoparticles (NPs) were formed from four components: acid-degradable biopolymer matrix, polyelectrolyte-based surfactant, GOx/CAT enzymatic system, and αPD-1 antibody. The oligosaccharide DEX was chosen as the matrix component of the NPs and the hydroxyl groups were reacted into pendant acetals. The resulting dextran derivative was soluble in organic solvents, enabled the encapsulation of the antibody during the formation of NPs, and was responsive to acidic environments. The anionic polysaccharide alginate was further incorporated as surfactant to form a negatively charged surface coating. The NPs were then embedded in the polymer-based MN matrixes made from cross-linked HA. Once in contact with the melanoma tissue, blood access to the needles gave rise to the local acidic environment. This promoted the self-dissociation of NPs and subsequently resulted in a substantial release of αPD-1. To evaluate anticancer efficiency of the MN patch, a syngeneic B16F10 mouse model (melanoma) was used to mimic clinical metastatic melanoma. The MN patch loaded with the antibody and with or without GOx was given via a single local application on the tumor site. Mice receiving α-PD1 delivered by MN patch (with GOx) exhibited significant tumor inhibition, with 40 % of mice still alive 40 days after treatment. Further co-delivery of anti-CTLA4 (αCTLA4) antibody showed a remarkable synergistic effect and resulted in long-term disease-free survival in about 70 % of the treated mice over a period of 60 days.^[67]

3. Programmable Materials Beyond Cancer

With a rise in the variety of approaches and materials capable of responding to triggers of all kinds, cancer remains the most investigated target for stimuli-responsive materials. While materials were initially designed to passively target cancer with the goal of maximizing intracellular concentrations of anticancer drugs, novel approaches relying on active responses to TME stimuli and to the delivery of immunotherapeutics and combinations have become more prevalent. Many other diseases, including CVD and arthritis, involve inflammatory processes that share similar pathological cues as cancer.^[68] Therefore, several of the platforms and strategies described in the previous sections, originally developed for cancer therapy, have been adapted in some form in settings outside of cancer. In the following sections we will discuss some recent, innovative programmable materials for applications beyond cancer, with a focus on CVD and diabetes as increasingly important human diseases globally. The following examples are described in sections divided by disease type and are classified based on the biological disease-associated triggers exploited to exert the specific therapeutic function.

3.1. Programmable Materials for Cardiovascular Disease

Cardiovascular disease (CVD) is the leading cause of death worldwide, claiming more lives than cancer.^[69] Myocardial infarction (MI), which is characterized by ischemic damage resulting from the build-up of atherosclerotic plaques in the coronary arteries, is one of the major contributors to the development of CVD.^[69] Despite technical advances and well-established clinical protocols to restore blood flow, early intervention approaches to preserve cardiac shape and function are lacking. Bulk hydrogels have been reported for the delivery of growth factors and cells to improve cardiac function by providing mechanical support and cell adhering surfaces.^[70] However, clinical translation is difficult for these systems at early stages following MI because of the requirement for invasive local injections. Therefore, nanomedicine allowing remote injection and targeting has become increasingly of interest for therapeutic delivery post MI. Small particle sizes, usually less than 200 nm, would enable nanoparticles to successfully extravasate from the leaky vasculature formed after MI and localize in the heart via noninvasive or minimally invasive administration.^[71] The disadvantage of this approach is that this extravasation is countered by rapid clearance caused by continuous dynamic muscle contraction of the heart even with the aid of active ligand targeting, for example, the overexpressed CD34 and angiotensin II type 1 (AT1) receptors on the surface of injured cardiomyocytes.^[71–72] Therefore, programmable materials that change their physiochemical properties in response to signals in the infarcted heart have been developed to address this challenge. The most commonly explored programmable materials for MI are locally applied smart hydrogels that respond to pH^[73] and temperature.^[74] These materials rely on stimuli triggered gelation for prolonged retention in the heart for gradual therapeutic release. As for biomolecular stimuli, MI shares biomarkers that are similar to those found in the TME, including overexpressed disease associated proteolytic enzymes (MMPs) and the production of ROS. Below, we highlight several innovative designs that have harnessed the overexpression of MMPs, and high concentrations of ROS involved in CVD development for long-term retention and local drug release.

3.1.1. MMP Responsive Programmable Materials

Following MI, a series of inflammatory responses are triggered, including MMP overexpression that leads to excessive degradation of the extracellular matrix (ECM), heart tissue remodeling and ultimately deterioration of cardiac function.^[75] MMP upregulation initiated as a part of the response to injury has been exploited as a key endogenous stimulus as well as being a therapeutic target for small molecules and smart hydrogels. For example, injectable hydrogels have been designed to slowly degrade in response to MMPs and providing controlled accumulation and diffusion of a therapeutic. In 2014, Burdick and coworkers reported an injectable biopolymeric hydrogel that responds to MMPs for on-demand release of a recombinant tissue inhibitor of MMPs (rTIMP-3) post MI (Figure 7).^[76] The hydrogel was composed of modified HA and dextran sulphate (DS) with MMP cleavable peptide as the crosslinks (Figure 7A). The negatively charged sulphate-functionalized polymer backbone binds to rTIMP3 through electrostatic interaction, significantly reducing passive drug release. The researchers demonstrated that the hydrogel served as a depot and gradually degraded upon exposure to MMPs, exposing rTIMP3 for inhibition (Figure 7B-C). Following local injection to the heart in a pig MI model, they observed significant decrease in interstitial MMP activity, attenuation of post-MI tissue remodeling and improvement in cardiac function as compared to the untreated or non-drug loaded hydrogel controls (Figure 7D-E). Using the same platform, the researchers then demonstrated on-demand delivery of a small interfering RNA (siRNA) to inhibit MMP2 translation post MI.^[77] Similar MMP degradable hydrogels that are fabricated from 1) glutathione (GSH)-modified collagen for growth factor delivery^[78] and 2) peptide amphiphile nanomatrices for cell therapy^[79] have been independently reported by different research groups. However, the main obstacle for the translation of these smart hydrogels remains the need for invasive local injections, which are not suitable for use in the acute early stages following MI. Compatibility of these materials with less invasive catheter-based administration is an area of need, as are intracoronary injections or entirely remote administration routes.

Figure 7. MMP-degradable injectable hydrogel.

(A) Structures of modified hyaluronic acid and dextran sulphate. The MMP-cleavable sequence GGRMSMPV is underlined. (B) Hydrogel crosslinking was achieved through hydrazone bond formation under physiological condition. Upon MMP exposure, the crosslinks were cleaved, allowing the rTIMP3 to expose, release and bind to MMP for inhibition. (C) Hydrogels with and without encapsulated rTIMP-3 were incubated with or without rMMP-2. As indicated by the green arrow, rMMP-2 was refreshed every two days to maintain enzyme activity. Encapsulated rTIMP-3 attenuated rMMP-2-mediated hydrogel degradation, confirming activity of rTIMP-3 across the 14-day study. (D) Local injection of hydrogel into the myocardium at nine equally spaced sites in a pig MI model. (E) Injection of rTIMP-3 encapsulated MMP-degradable hydrogel significantly inhibited the interstitial MMP activity as compared to shame (healthy), MI (nontreated), and hydrogel alone groups at 14 days post MI. Reproduced with permission.^[76] Copyright 2014,. The Authors, published by Springer Nature.

Other than MMP-degradable hydrogels, polymeric nanoparticles and biomolecular microgels that change morphology in response to MMPs have also been investigated. In 2015, our laboratory reported MMP responsive nanoparticles that can target and be retained in the infracted heart for up to 28 days following MI.^[80] These nanoparticles were designed based on the aforementioned EDAPT approach, in which nanoparticles are composed of a hydrophobic core and an MMP cleavable peptide shell. Following intravenous injection, the nanoparticles extravasated into the heart through the leaky vasculature and underwent a morphological switch to form microscale assembles upon MMP exposure only in the infarcted region. This morphological change physically trapped the material in the infarcted heart for long-term retention. Recently, this type of strategy was developed further to deliver an entirely biocompatible and degradable peptide-based gel. Here, cyclic, MMP responsive peptide progelators were designed for catheter and intracoronary injection (Figure 8).^[12] These peptides contain a substrate recognition sequence for MMP and elastase, a self-assembling peptide sequence (KFDF)₃ for gelation, and a disulfide bridge for cyclization (Figure 8A). The initial cyclized peptides had little resistance to flow due to the conformational constraint and could be easily administered by catheter (Figure 8B). Upon exposure to MMPs and elastase, the progelators were cleaved, linearized and self-assembled into entangled fibrous structures for gelation (Figure 8C-D). Local injection of the progelator in a rat MI model confirmed the enzyme triggered gelation without inducing an inflammatory response or cardiomyocyte apoptosis (Figure 8E). As compared to bulk hydrogels and nonresponsive particles, this morphological switchable platform provides advantages in terms of delivery. First, its initial small size enables non-invasive administration that is more compatible with the treatment of acute phase MI. Second, the MMP-triggered microscale assembly prevents the material from being extruded from the injection site through continuous dynamic muscle contraction, enhancing the precision in targeting, retention and accumulation.

Figure 8. MMP-responsive cyclic peptide based progelator.

(A) The progelator was composed of MMP and elastase cleavable sequence (red), gelation sequence (green), rhodamine label (pink) and cyclized via disulfide bond. (B) Schematic demonstration of enzyme triggered cyclic progelator self-assembly into bulk fibrous structure. (C) Predicted β-sheet re-orientation of the linearized KFDF peptide via computational modeling. (D) Progelator incubated with active and denatured enzymes. Material aggregation and settling from solution were observed upon active enzyme treatment. TEM confirmed the formation of fibrous structures. Scale bars: 100 nm. (E) Activation and gelation of the cyclic progelator in the infarcted heart following local injection in vivo. Reproduced with permission.^[12] Distributed and licensed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/). Copyright 2019, The Authors, published by Spring Nature.

3.1.2. ROS Responsive Programmable Materials

Excessive ROS production is another feature of the inflammatory response during the progression of CVD. It plays an important role in the pathogenesis of atherosclerosis, the most common cause of MI that is associated with the accumulation of fatty deposits and inflammatory cells within the inner walls of artery vessels.^[81] Recently, Scott and coworkers reported on a ROS responsive filamentous hydrogel depot (FM-depot) for sustained delivery of anti-inflammatory nanocarriers as cardiovascular immunotherapy (Figure 9).^[82] The major component of the hydrogel is poly(ethylene glycol)-block-poly(propylene sulfide) (PEG-b-PPS) capped with either methoxy or vinyl sulfone (Figure 9A). The copolymers were formulated into cylindrical filomicelles and loaded with an anti-inflammatory agent 1, 25-dihydroxyvitamin D3 (aVD) via the thin-film hydration method, followed by crosslinking with 8-arm PEG thiol to afford the FM-depots. Upon oxidation, the hydrophobic PPS block was modified to give hydrophilic sulfoxide and sulfone, decreasing the surface tension and triggering the transition from cylindrical filomicelles to spherical micelles (Figure 9B). This budding process was confirmed by thermodynamic modeling and characterized via both cryogenic transmission electron microscopy (cryo-TEM) (Figure 9C) and small-angle light scattering (SAXS). Following subcutaneous injection in atherosclerotic mice (Figure 9D), it was shown that the sustained delivery of aVD via FM-depots could significantly induce proliferation, expansion and homing of regulatory T cells (Foxp3+ Tregs) in lymphoid organs and aorta as compared to PBS and free drug controls (Figure 9E). This novel strategy demonstrates the potential of programmable materials as anti-inflammatory immunotherapeutics for atheroprotection. A potential problem for further investigation would be the ability of this system to distinguish between basal and elevated level of ROS.

Figure 9. ROS responsive filamentous hydrogel depot (FM-depot).

(A) Structures of PEG-b-PPS block copolymers and 1, 25-dihydroxyvitamin D3 (aVD). (B) Schematic demonstration of the backbone oxidation, crosslinking and cylinder to micelle transition. (C) Cryo-TEM micrographs of aVD-FM and formation of micelles. (D) Timeline of animal study. (E) In vivo aVD-FM injection elicited significant Treg generation in lymph nodes and spleen as compared to untreated and free drug controls. Reproduced with permission.^[82] Distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/). Copyright 2020, The Authors, published by Frontiers Media.

3.2. Programmable Materials for Diabetes

Responsive materials that undergo programmed changes for precise insulin delivery to treat diabetes is another rapidly growing field. Diabetes is a global health threat, impacting 422 million people worldwide.^[83] Despite technological advances in insulin replacement therapy, including insulin pumps and continuous glucose monitors, it is still challenging to achieve precise drug dosage control through self-administration, leading to hyperglycemia, hypoglycemia as well as diabetic complications.^[84] As such, efforts have been made in the design of self-regulated insulin delivery systems that release insulin in response to blood glucose levels. So far, three major types of glucose sensors have been harnessed for the preparation of glucose-responsive insulin release systems: 1) glucose oxidase (GOx), 2) phenylboronic acid (PBA), and 3) glucose binding molecules. For a detailed review on recent advances on these “smart insulin” systems, we guide the readers to Gu. et al.^[85] In the following sections, we intend to highlight several platforms that involve innovative stimuli-response systems to place them in context with other systems described in this review.

3.2.1. GOx Based Programmable Materials

GOx has high specificity for glucose. It catalyzes glucose oxidation in the presence of O₂, resulting in local pH drop, H₂O₂ production and hypoxia. These signals have been exploited for the fabrication of hydrogels, nanoparticles, liposomes and MN patches that undergo triggered conformational transformations,^[84–85] including swelling^[86] and degradation^[87] to release insulin. One recent example by Gu and coworkers utilized a painless MN patch integrated with GOx containing hypoxia-sensitive HA-based vesicles (GRV) for insulin delivery (Figure 10).^[87b] HA was modified with hydrophobic 2-nitroimidazoles, and assembled into nanoscale vesicles, encapsulating GOx and insulin (Figure 10A). In the presence of high blood glucose levels, O₂ was consumed via GOx catalyzed glucose oxidation, creating a hypoxic microenvironment where 2-nitroimidazoles were reduced into hydrophilic 2-aminoimidazoles, triggering disassembly of the vesicles and subsequent insulin release (Figure 10B). The insulin release kinetics was tunable based on the glucose concentration as indicated by a pulsatile release pattern (Figure 10C). These vesicles were loaded into the tips of the MN-array fabricated from crosslinked HA (Figure 10D) and showed a rapid insulin release to achieve normal glucose level (< 200 mg/dL) following transdermal treatment in type 1 diabetic mice (Figure 10E). The MN patch holds potential for clinical applications due to the ease of use, painless administration route, and low risk of foreign body response as compared to bulk hydrogel and implantable devices. However, the rate of responsiveness of the GOx based materials varies with temperature and oxygen concentration. Long-term enzyme stability and activity are other problems that may hamper clinical translation.

Figure 10. GOx based glucose responsive vesicle (GRV) loaded-microneedle array patch for insulin release.

(A) Structures of modified HA before and after hypoxia induced reduction. (B) Schematic demonstration of the glucose induced dissociation and subsequent insulin release from the assembled vesicles. TEM micrographs showed the vesicles post 400 mg/dL glucose incubation. Scale bars: 200 nm. (C) Pulsatile insulin release pattern following alternative glucose concentration. (D) SEM image of MN-array. Scale bar: 200 μm. (E) Change in blood glucose level post MN-array treatment in vivo. ½ E indicated for half of the insulin loading as E. Reproduced with permission.^[87b] Copyright 2015, The Authors, published by National Academy of Sciences.

3.2.2. PBA-Based Programmable Materials

As compared to GOx based systems, enzyme-free PBA incorporated materials are more suited to long-term and continuous use. For PBA based system, dynamic glucose binding to PBA shifts the equilibrium of the noncharged PBA towards the anionic form and disintegrates the existing PBA-diol crosslink in the carriers. Myriad PBA based systems that experience structural changes such as swelling,^[88] shrinkage^[89] and disassembly,^[88c] in response to high glucose concentration have been fabricated for controlled insulin release. In 2017, Matsumoto, Tanaka and coworkers reported an implantable catheter-based device combined with PBA containing polymeric gel for on demand insulin delivery (Figure 11).^[13] The device was designed so that the surface of the gel functions as a “gate”, only allowing insulin permeation when the gel is at the hydrated state upon glucose binding to PBA (Figure 11A). The gel was coated on the inner layer of silicon catheters, which had pores along the axis for communication between the insulin encapsulated gel and glucose in the outer interstitial fluid (Figure 11B). After implanting the device under the skin of type 1 diabetic mice (Figure 11C), the researchers observed rapid insulin release and significant decrease in glucose level as compared to the PBS control (Figure 11D). PBA moieties have also been incorporated to fabricate glucose-responsive modified insulin that potentially bind to human serum albumin for enhanced circulation half-life^[90] and composite MN patch for hypoglycemia-triggered glucagon release^[83]. However, PBA has nonspecific binding towards cis diols, not limited to glucose. To translate the current platforms into a clinically relevant setting, functionalized PBA systems that have higher glucose binding specificity need to be developed.

Figure 11. PBA containing polymeric gel-based insulin release device combined with catheter.

(A) Structure of PBA containing polymeric gel and schematic demonstration of the “gate” mechanism. Under low glucose level, the polymeric gel was at dehydrated state and the outer surface serves as a skin layer, preventing insulin release. In the presence of high glucose concentration, the gel shifted to hydrated state due to glucose binding to PBA, allowing insulin permeation. (B) Schematic representation of glucose triggered insulin release from the polymeric gel combined catheter device. (C) Figure of combined device and implanting in diabetic mice. (D) In vivo insulin release profile and the change in blood glucose level. Reproduced with permission.^[13] Distributed under a Creative Commons Attribution NonCommercial License 4.0 (http://creativecommons.org/licenses/by-nc/4.0/). Copyright 2017, The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works.

3.2.3. Glucose Binding Molecules-Based Programmable Materials

As compared to PBA, glucose binding molecules usually have higher specificity for glucose over other diols.^[84] Concanavilin A (ConA) remains the most utilized glucose binding protein. This has been used in release systems where competitive glucose binding to ConA weakens the crosslinks of ConA within a saccharide in a gel, leading to expansion, disassembly or phase switching.^[91] Other molecules including glucose transporter inhibitor are emerging in sensor design in the preparation of responsive materials for on demand release of modified insulin.^[92] Long-term safety and stability of the proteins remains the main challenge for clinical translation.

3.3. Programmable Materials for Other Diseases

For the diseases beyond cancer, CVD and diabetes, there has been relatively little work in the area of stimuli-induced targeting or release from designed materials. The platforms that have been developed so far mostly utilize inflammation-targeting systems that harness many of the same aforementioned pathological abnormalities including local acidification, enzymatic upregulation and activity, and ROS as triggers. For example, our laboratory has demonstrated the targeted delivery and long-term retention of the EDAPT platform in ischemic skeletal muscle.^[93] Karp and coworkers have reported a triglycerol monostearate based, MMP-disintegrable hydrogel for arthritis flare-responsive drug delivery, as well as immunosuppressive therapy in a vascularized composite allograft.^[94] Similar inflammation targeting hydrogel microfibers have also been demonstrated in inflammatory bowel disease.^[95] Proof of concept work that capitalizes on ROS responsive hollow microspheres which contain anti-inflammatory drugs, acid precursors for Fenton reaction, and bubble generating agents, has shown protective effects against joint destruction in a mouse model for osteoarthritis.^[96] A boronic ester containing, micelle encapsulated hydrogel that disassembles upon exposure to low pH and ROS has been leveraged for the delivery of anti-bacterial and anti-inflammatory drugs to promote wound healing.^[97]

It’s worthwhile noting that several examples harness physical and biological cues other than the inflammatory markers have been recently reported for the design of programmable materials targeting other diseases. For instance, abnormal shear pressure caused by vascular narrowing has been used by Ingber and coworkers to generate shear-activated nanotherapeutics for thermolytic drug delivery. ^[5] In response to the high shear stress in occluded blood vessels, the PLGA based microaggregates disassembled into nanoparticles (~180 nm in diameter), improving their adherence to the adjacent vessel walls for drug release. Following intravenous injection, this formulation induced rapid clot dissolution in a mouse mesenteric injury model and normalized pulmonary artery pressure and increased survival in an otherwise fatal mouse pulmonary embolism model. Thrombin, a proteinase involved in blood coagulation, has been utilized as the trigger in the fabrication of size-shrinkable nanoparticles that can cross the blood-brain barrier for stroke therapy^[98] and degradable MN patches that can be applied to anticoagulation therapy.^[99] In addition to the native physiological signals, programmable materials that are responsive to bacteria associated toxins, lipases and cell wall constituents have been developed to fight against bacterial infection.^[100] A recent example involves a self-assembling, human defensin-6 (HD6) mimic peptide (HDMP) that targets the lipoteichoic acid on Gram-positive bacteria for antibacterial therapy. The HDMP initially self-assembled into spherical nanoparticles. Upon recognizing and binding to the bacterial cell wall via ligand-receptor interaction, HDMP transformed into nanofibers, trapping the bacteria to prevent bacterial invasion.^[101]

4. Conclusions and Future Perspectives

Over the past decade, the field of programmable responsive soft materials for biomedical applications has rapidly expanded as testified by the enormous innovations and variety of approaches reported in the literature. In this noncomprehensive review, we have highlighted some recent examples that show promising results in preclinical animal models for the treatment of cancer, CVD and diabetes. Efforts have focused on exploiting endogenous stimuli associated with specific disease states to program materials to undergo morphological and size transitions, disassembly, and changes in mobility and retention, with the final aim to exert a more selective and efficacious therapeutic function. Synthetic polymers, naturally derived biopolymers, and hybrid materials have all been utilized to fabricate these responsive materials. Different delivery routes, spanning from noninvasive intravenous delivery to local delivery (injection, implant, spray, and transdermal patch, among others) have been demonstrated. Overall, the employment of disease state associated chemical, physical and biological fingerprints has proven to be a valid starting point for the development of responsive materials as novel disease treatments. Such delivery vehicles have the potential to enhance stability, tissue specificity and therefore efficacy of the delivered therapeutics, reducing off-target side effects and decreasing frequency of dosing.

However, despite the evident progress in the field, concerns regarding efficacy, safety and fabrication impede clinical implementation of these programmable materials. It is indeed often premature to evaluate the performance of the reported systems with only animal models of human diseases. Given the difference between the pharmacokinetics and biodistribution across animal species and humans, it is not uncommon that a formulation with promising results in animal models exhibits low performance/efficacy in clinical trials. A critical example is the enhanced tumor permeability and accumulation of nanomedicine in the TME, which so far has only been proven in certain animal models and evidence in humans is disappointing or limited.^[102] In addition, initial animal studies rarely provide information on long-term efficacy, biodistribution and toxicity, which are critical in the human clinical setting.

Undefined biocompatibility, biodegradability and the potential immunogenicity of the employed materials remain another obstacle to clinical translation. In the case of systemically injectable nanomaterials, inability to obtain prolonged circulation time while bypassing liver and spleen accumulation, as well as poor targeting of circulating or disseminated diseased cells, for instance in the case of metastatic cancer or non-solid cancer, constitute additional drawbacks. Biopolymeric materials, which are biocompatible and often biodegradable as compared to the synthetic versions can potentially obviate some of these challenges. However, such systems generally require more expensive and sophisticated synthetic preparations to achieve efficient responses towards endogenous stimuli. Hybrid materials are a compromise that can combine the upsides of polymeric materials, such as straightforward synthesis and facile incorporation of responsive moieties, with the excellent biocompatibility of biomolecules. In the context of hybrid systems, cell-mediated delivery of nanomaterials have been recently proposed as a novel strategy.^[103] Indeed, cells can be engineered as biocompatible and biodegradable vehicles, can spontaneously home to infected or inflamed tissues and have the potential to respond to endogenous stimuli.^[103c] While a variety of cell-based approaches employing red blood cells, immune cells and bacteria have been exploited in the past decade, predominantly in the context of cancer therapy, most of these approaches are still in their infancy and efficient response to endogenous triggers has not been fully explored.^{[103a, 103c, 104]} Moreover, while designing programmable responsive materials, we should keep in mind that some of the most exploited hallmarks, from enzymes overexpressed during inflammation to redox state and low pH, are common across different diseases. While such features offer exceptional convenience and can be beneficial in adapting approaches across different disease types, they may also result in low selectivity and safety in human patients with comorbidities. Other endogenous stimuli, for example biomolecules that are unique to cardiomyocytes and that are released into the extracellular environment post cell death, as well as downstream inflammatory markers,^[68a] should be screened to enhance precise targeting and accumulation of programmable materials in the case of heart disease. A deeper understanding of disease-associated microenvironments and pathophysiology would allow for the design of more selective materials that can, for instance, respond to disease specific markers or multiple pathologic triggers, to further improve targeting, duration of administration and efficacy.

Finally, technical features play an important role in the clinical implementation of programmable materials. The fabrication needs to be straightforward and modular. The resulting materials need to be easy to manipulate and stable during handling. Additionally, batch-to-batch variations must be avoided. All of these aspects need to be taken into account in the design of a novel material. A balance between sophisticated chemistry design, and ease in manufacturing needs to be met to develop more translatable materials. Careful multidisciplinary efforts are fundamental to direct novel programmable responsive systems toward clinical validation and to enable valuable breakthroughs in the field of biomedical materials with real world impact.