Novel Protein Therapeutics Created Using the Elastin-Like Polypeptide Platform

Abstract

Elastin-like polypeptides (ELPs) are bioengineered proteins that have a unique physical property, a thermally triggered inverse phase transition, that can be exploited for drug delivery. ELP-fusion proteins can be used as soluble biologics, thermally targeted drug carriers, self-assembling nanoparticles, and slow-release drug depots. Because of their unique physical characteristics and versatility for delivery of nearly any type of therapeutic, ELP-based drug delivery systems represent a promising platform for biologics development.

Introduction

Elastin-like polypeptides (ELPs) are a class of engineered, repetitive proteins derived from a motif found in human tropoelastin (1). The most common repeat motif used in ELPs is (VPGxG)_n, where “x” is a guest residue that can be any amino acid except proline and “n” indicates the number of pentapeptide repeats. The defining feature of ELPs is their inverse phase transition, where the proteins undergo a temperature-induced secondary structure change that exposes hydrophobic surface area, leading to coacervation of ELP molecules into micrometer-sized aggregates (2). The inverse phase transition of ELPs occurs at a specific temperature for a given ELP, known as its transition temperature (T_t), and the aggregation process is fully reversible, leading to complete resolubilization of the ELP when the solution is returned to a temperature below the T_t. The T_t is also tunable, as it is dependent on modifiable features including the size of the ELP (3) and the hydrophobicity of the guest residue (2). This reversible phase transition can be exploited in multiple ways. First, it allows for nonchromatographic purification of ELPs and ELP-fusion proteins by a process known as inverse transition cycling (4), a simple procedure that involves purification of ELPs from contaminating proteins by centrifugation above the T_t. This purification process is economically favorable since it eliminates time-consuming chromatography steps, and it is scalable from small batches up to industrial-scale production. In addition to its usefulness as a purification tool, the ELP phase transition can also be utilized for drug delivery purposes, creating soluble biologics, thermally targeted carriers, nanoparticles, or even subcutaneous drug depots.

In addition to the phase transition, other properties of ELPs contribute to their utility as drug carriers. Because they are proteins, they can be produced by recombinant expression (5). This allows for production of very defined species, giving ELPs a major advantage over other synthetic polymers that must be produced by chemical synthesis, which inherently results in a polydisperse product. The proteinaceous nature of ELPs imparts another advantage over some synthetic polymeric drug carriers in that ELPs are degradable in vivo and the degradation products are natural amino acids, reducing the risk for toxicity (6). The proteinaceous nature of ELPs also makes them ideal drug carriers, as peptide- or protein-based therapeutic fusion proteins can be made at the gene level, and the chimeric therapeutic proteins can be produced recombinantly (7). The precise control over the ELP sequence can also be utilized to engineer reactive sites for attachment of small-molecule cargo. Another advantage of ELPs is their low immunogenicity (8), resulting from their simple sequence that is present natively in human elastin. Collectively, these features make ELPs a promising platform on which to design drug delivery systems.

ELP-Fusion Proteins as Soluble Biologics

The simplest application for ELP-based drug delivery is using the proteins as soluble carriers for peptide, protein, nucleic acid, or small-molecule drug cargo. This is achieved by selecting ELPs with T_t well above the 37°C body temperature (FIGURE 1A). The phase transition can still be exploited for purifying these ELP-fusion proteins by changing the buffer conditions (e.g., increasing the salt concentration) to lower the T_t in vitro and allow for purification by inverse transition cycling. However, when used in vivo, the ELPs’ high T_t ensures that the proteins will not form coacervates in the blood or in the tissues in which they ultimately deposit. There are several characteristics of ELPs that make them well suited as soluble carriers for therapeutics. First is their amenability to modification. Both the NH₂ and COOH termini can be modified with either targeting agents or therapeutic cargo (FIGURE 1B), multiple reactive sites can be engineered for carrying a large number of small molecules if desired (9, 10), and even the guest residue in the pentapeptide repeat can be engineered with cysteine or lysine residues to introduce reactive thiol or primary amine groups, respectively, that can be used for drug attachment. Targeting agents can be peptide or protein based and encoded at the DNA level by cloning their coding sequences in frame with the ELP coding sequence (11), or they can be small molecules designed to bind to specific receptors in vivo. Small-molecule attachment can be via cleavable linkers if desired, and varying types of cleavable linkers can be used to achieve drug release under desired conditions (12). Another feature of ELPs that lends to their utility as soluble macromolecular carriers is the ability to tune their size by altering the number of peptide repeats used in the ELP moiety. This allows for tunability of plasma pharmacokinetics, as the clearance rate of soluble ELPs is inversely proportional to their size (13). Soluble ELP-fusion proteins have been utilized in a large variety of disease applications spanning from kidney disease (FIGURE 1C) to immune disorders to diseases of pregnancy (FIGURE 1D).

FIGURE 1.

Soluble elastin-like polypeptides (ELPs) for drug delivery A: when used as soluble drug carriers, ELPs with a transition temperature (T_t) well above body temperature are chosen. B: ELPs can be modified at both the NH₂ and COOH termini with cargo. Because of the modular nature of ELP carriers and the ease of modifying the protein sequence by simple molecular biology techniques, modifications can include targeting peptides, cell-penetrating peptides, therapeutic peptides or proteins, drug-binding proteins, or reactive sites for small-molecule drug attachment. C: ELPs accumulate in the kidneys, and renal targeting can be improved further with kidney-targeting peptides. Areas of ELP accumulation are shown highlighted in yellow with red borders. Renally targeted ELPs are being used for delivery of renoprotective therapeutics, including agents for therapeutic renal angiogenesis and agents to block renal inflammation. D: ELP carriers and ELP-bound cargo do not cross the placental barrier. ELPs accumulate in the placenta on the maternal side but do not cross the syncytiotrophoblast barrier or enter fetal circulation (areas of accumulation highlighted with red borders). Therefore, ELPs are powerful tools to improve the safety of therapeutics used during pregnancy and as a platform for drug development for disorders of pregnancy, including preeclampsia.

Renal Drug Delivery with Soluble ELP-Fusion Proteins

ELPs are cleared from circulation by renal filtration, and ELPs naturally accumulate at high levels in the kidney (11, 13, 14). This fact can be exploited for renal drug delivery. Additionally, with kidney-targeting peptides (KTPs) selected by phage display to have affinity for the vasculature of the kidney (15), ELP renal levels and specificity of renal targeting can be increased (11). Within the kidney, ELPs can be detected in the vasculature and in the tubular system, with biodistribution being dependent on the size of the ELP and on its cargo (11, 13). In a series of studies, ELP was fused with the angiogenic cytokine vascular endothelial growth factor (VEGF) at its COOH-terminal modification site to create a novel biologic for therapeutic angiogenesis to treat renal disease. A hallmark feature of chronic kidney disease (CKD), regardless of whether the etiology is uncontrolled hypertension, diabetes, or other insults such as renovascular disease (RVD) or acute kidney injury, is loss of renal microvasculature (16–23). The loss of functional blood vessels results in impaired renal blood flow (RBF) and glomerular filtration rate (GFR), and no currently approved therapies for CKD address the underlying microvascular rarefaction. Because of its high renal deposition, the ELP-VEGF fusion protein was tested for its ability to protect or restore the renal microvasculature in models of RVD and CKD. The ELP-VEGF fusion protein was a chimeric molecule with a total molecular mass of 74 kDa (7) and a transition temperature well above body temperature, ensuring that it remains soluble in vivo. Importantly, the ELP-VEGF chimera maintained its ability to stimulate the VEGF receptor, as ELP-VEGF and free VEGF had identical potency in cellular assays of endothelial cell proliferation, matrix invasion, and tube formation (7, 24). When administered intrarenally via an interventional catheter in a swine model of RVD, significant amounts of ELP-VEGF were retained in the injected kidney, and the remaining protein cleared the body with a half-life of 13.5 h, with significant accumulation in the contralateral, noninjected kidney (24). Intrarenal ELP-VEGF led to increased microvascular density in the injected kidney, which was associated with improved RBF and GFR and attenuated renal fibrosis. Additionally, ELP-VEGF was more effective than free VEGF for the restoration of renal function. In follow-up studies, ELP-VEGF administered systemically in the swine RVD model also homed to the kidney and improved renal function and renal microvascular density and reduced renal fibrosis (25). Additionally, ELP-VEGF was tested in a coadjuvant application with percutaneous transluminal renal artery angioplasty (PTRA) and stenting in the swine RVD model (26). ELP-VEGF was administered via the intraluminal catheter directly into the renal artery after placement of the stent to resolve the arterial stenosis. When ELP-VEGF was combined with PTRA, renal microvascular density, RBF, GFR, and plasma and histological markers of renal injury were all improved relative to PTRA alone. Similar results were obtained with intrarenal ELP-VEGF in a swine model of CKD induced by bilateral renal artery stenosis combined with a dyslipidemia (27). The ELP-VEGF molecule was also modified with a KTP to further improve renal targeting. KTP-ELP-VEGF displayed improved renal delivery relative to ELP-VEGF in a mouse model (28), and intrarenal KTP-ELP-VEGF induced recovery of renal function and renal microvascular density in a swine model of RVD (29). Mechanistic studies pointed to two possible pathways for the prolonged improvements in renal function following ELP-VEGF therapy. Markers of endothelial progenitor cells were increased in ELP-VEGF-treated kidneys weeks after treatment, supporting a role for these progenitor cells in microvascular protection and recovery (26). Also, resident renal macrophages were found to be shifted from an M1 phenotype in control injured kidneys to a proangiogenic M2 phenotype, which was directly observed to produce VEGF, in ELP-VEGF-treated kidneys (27, 29), supporting a possible positive feedback loop of angiogenesis activated by a single ELP-VEGF treatment. A comprehensive dose-escalating toxicology study was performed using a single ELP-VEGF intravenous administration in a rat model, beginning at therapeutic doses and increasing to 200 times the highest therapeutic dose (30). ELP-VEGF exhibited a very good safety profile, having no effects on body weight, blood pressure, proteinuria, GFR, plasma markers of renal and liver function, or histological markers of renal injury in the therapeutic dosing window. Only at 100 to 200 times the highest therapeutic dose were any adverse events observed, which included transient hypotension, elevated GFR, and microvascular rarefaction. ELP-VEGF also caused no change in tumor progression or tumor weight in a mouse xenograft model of triple-negative breast cancer. However, ELP-VEGF did induce a dose-dependent increase in tumor vascular density, suggesting possible contraindication for its use in patients with malignancy (30).

Improved Delivery of Immune Suppressors by Fusion to Soluble ELP Carriers

Cyclosporine A is a cyclic peptide that functions as an immunosuppressant by inhibiting calcineurin signaling and blocking the production of inflammatory cytokines in T lymphocytes (31). Cyclosporine is lipophilic, and its therapeutic window is limited by its potential to induce hypertension and kidney, liver, and neurotoxicity. For delivery of cyclosporine A, its receptor, cyclophilin A (CypA), was fused to a 74-kDa soluble ELP carrier at the ELP’s NH₂ terminus (32). The CypA-ELP fusion protein formed a dimer in solution that bound to cyclosporine A with an affinity of 189 nM. Binding to the ELP-fusion protein improved the solubility of cyclosporine A, and the drug was released from the carrier with a slow 957-h half-life. The ELP-bound cyclosporine A was bioavailable after subcutaneous administration, and it cleared from mice with half-lives of 29.2 and 22.4 h after intravenous and subcutaneous administration, respectively. The drug’s efficacy was tested in a mouse model of Sjögren’s syndrome, an autoimmune disease that results in dry eyes due to inflammation in the lacrimal glands. When administered every other day for 14 days, the ELP-delivered cyclosporine A increased tear production with efficacy similar to free cyclosporine, delivered as the clinically approved comparator Sandimmune. Importantly, mice treated with the ELP-delivered cyclosporine A had no increase in blood urea nitrogen, which was increased in the free drug comparator group, indicating that the ELP-delivered drug had reduced indications of nephrotoxicity (32).

A similar strategy was used for the delivery of an alternative immunosuppressant, the small-molecule mechanistic target of rapamycin (mTOR) inhibitor rapamycin. Both the NH₂ and COOH termini of a 74-kDa soluble ELP were modified with the rapamycin binding protein FKBP12, creating a fusion protein capable of binding two molecules of rapamycin (33). The ELP-FKBP12-bound rapamycin was also bioavailable after subcutaneous administration, and delivery of rapamycin via this protein carrier partially prevented the drug’s sequestration by FKBP proteins in red blood cells. In a mouse model of Sjögren’s syndrome, the ELP-delivered rapamycin suppressed lymphocyte infiltration into the lacrimal glands, reduced lacrimal gland fibrosis, and reduced expression of inflammatory cytokines in the lacrimal glands. The ELP-delivered rapamycin did, however, cause a transient increase in plasma glucose levels that was not present with free rapamycin, raising concern for blood glucose control in patients with diabetes. Therefore, in a follow-up study, one of the FKBP12 moieties was replaced with a peptide that binds to the cell surface molecule Intercellular adhesion molecule-1 (ICAM-1), creating a protein that contained an ICAM targeting domain at the ELP’s NH₂ terminus and the rapamycin-binding FKBP12 domain at the COOH terminus (34). ICAM-1 is expressed in inflamed tissues and is involved in lymphocyte migration and B- and T-cell activation. The ICAM-targeted protein retained high-affinity rapamycin binding and improved deposition in the lacrimal glands of Sjögren’s syndrome mice. However, the relative efficacy and safety compared with the nontargeted iteration were not tested.

ELPs can also be used for delivery of antibody-based therapeutics. A single-chain monoclonal antibody derived from the Camelidae heavy chain with specificity for tumor necrosis factor alpha (TNF-α) was fused to a 100-repeat ELP carrier (35). ELP fusion to the antibody did not significantly affect its affinity for TNF-α. However, the half-life of the small, nonhumanized antibody was increased from 28 min to 11.4 h via the ELP fusion. In a demonstration of potential efficacy in a model of septic shock, mice treated with the ELP-fused anti-TNF antibody survived a lipopolysaccharide challenge that was lethal in untreated mice.

ELP for Drug Delivery during Pregnancy

Soluble ELPs have also been adapted for drug delivery during pregnancy. Since ELPs are not substrates for the neonatal Fc receptor (FcRn), the protein responsible for mediating antibody transport across the placental barrier, or for any other placental protein transporter, ELPs do not cross the placenta (36). The lack of placental transport even holds for prolonged infusion of ELPs and for ELPs modified with cell-penetrating peptides (CPPs), short peptide sequences capable of increasing uptake of protein cargo across cellular plasma membranes (37), when assessed in a rat pregnancy model (36). The plasma clearance rate of ELPs is dependent on the ELP size in pregnant rats. Although biodistribution and intraplacental deposition of ELPs are dependent on ELP size, even ELPs as small as 25 kDa do not cross the placental barrier (38). This feature can be used to prevent fetal exposure to therapeutics by administering them as fusions with the ELP carrier. Small-molecule placental transfer is inhibited by ELP fusion, which has promise for improving the safety of drugs during pregnancy.

Additionally, ELP-fusion proteins have been developed to treat pregnancy-specific disorders such as preeclampsia. Preeclampsia is a pregnancy-specific syndrome that is hallmarked by increased inflammation (39) and increased circulating levels of a VEGF antagonist protein called sFlt-1 (40, 41). ELPs fused to members of the VEGF family have been used to target the characteristic angiogenic imbalance that is present in preeclampsia and is driven by overabundance of sFlt-1 (42, 43). When administered to pregnant rats in a surgically induced model of placental ischemia (44), ELP-VEGF was confined to the maternal circulation (45). ELP-VEGF sequestered the circulating sFlt-1 in placental ischemic rats and prevented the rise in blood pressure associated with placental ischemia in a dose-dependent manner. Adverse effects including ascites production and aberrant angiogenesis around the peritoneal minipump were observed and were associated specifically with continuous intraperitoneal delivery. Follow-up studies revealed that these effects can be ameliorated by using intravenous or subcutaneous dosing (unpublished data). Also, ongoing work is utilizing other less angiogenic isoforms of the VEGF family, including VEGF-B and Placental Growth Factor (PlGF) ELP fusions, to antagonize sFlt-1 without directly agonizing the proangiogenic VEGF receptor Flk-1.

Another strategy to treat preeclampsia is to target the maternal inflammation that is a hallmark of the syndrome. Preeclamptic patients have elevated levels of circulating inflammatory cytokines, including interleukins 1 and 6 (IL-1 and IL-6) and tumor necrosis factor alpha (TNF-α) (46, 47). These cytokines contribute to systemic endothelial dysfunction and ultimately to the maternal syndrome of preeclampsia. The transcription factor Nuclear factor kappaB (NF-κB) is the central mediator of inflammation. The NF-κB pathway is activated by multiple inflammatory stimuli, including the TNF receptor and several interleukin receptors (48), and it activates the production of inflammatory cytokines. Peptide inhibitors of the NF-κB pathway have been described that block its activation at several steps (49–52). However, free peptides are poorly suited as therapeutics because of their short plasma half-life, poor cell penetration, and lability to degradation. An ELP-delivered NF-κB inhibitory polypeptide was developed by modifying ELP at its NH₂ terminus with a CPP and at its COOH terminus with a peptide that blocks the nuclear translocation of NF-κB following its activation (53). When administered to pregnant rats, the ELP-delivered NF-κB inhibitory peptide had improved plasma pharmacokinetics relative to the free peptide, and the ELP-delivered peptide was maternally sequestered whereas the free peptide passed through the placental barrier (54). In the placental ischemia model of preeclampsia, the ELP-delivered NF-κB inhibitory peptide partially ameliorated the hypertension and reduced placental TNF-α levels, while showing no signs of toxicity.

Thermally Targeted Drug Delivery Using ELP Carriers

The temperature-induced phase transition of ELPs can be used to implement thermally targeted drug delivery. For these applications, ELPs are chosen that have a T_t slightly above body temperature, in the 39–41°C range (FIGURE 2A). As with the soluble ELP-delivered therapeutics, these ELPs can also be fused to small-molecule drugs or to proteinaceous therapeutics, with or without targeting agents or CPPs (FIGURE 2B). When administered systemically, they will remain soluble in circulation at body temperature. However, coacervation of the ELP moiety can be induced by externally applied mild hyperthermia at a desired site (FIGURE 2, C AND D), usually in the range of 40–42°C. These temperatures are warm enough to induce ELP coacervation but not hot enough to cause significant tissue necrosis. Within the heated tissue, the ELP will coacervate and accumulate (FIGURE 2E, LEFT), as has been observed by direct observation in heated tumor tissue with intravital microscopy (55, 56). Thermal targeting efficiency is improved by cycling between bouts of hyperthermia and normothermia in targeted tissue (57). This heat cycling protocol induces ELP aggregation and accumulation in the vasculature of the target tissue, and then reduction of the temperature allows the ELP to resolubilize, when it then moves down the newly established concentration gradient from the blood into the tissue parenchyma (FIGURE 2E, RIGHT). Mild, focused hyperthermia has been achieved by water bath immersion or illumination with infrared light in animal models (56, 58), but technology is available clinically to achieve targeted mild hyperthermia in human patients with image-guided high-intensity focused ultrasound (59–62). Thermal targeting applications of ELP-fusion proteins have traditionally been used for targeting solid tumors, although these applications are potentially useful for any disease state in which targeting of a large area is desired, provided the target area is amenable to externally applied hyperthermia. Seminal studies demonstrated that ELPs, engineered with a T_t around 40°C and radiolabeled to allow in vivo tracking, could be thermally targeted to tumors in mouse xenograft models. A 59-kDa ELP demonstrated an 8.37-h half-life in mice, very slow degradation in mouse plasma (63), and accumulation in heated tumors to levels 1.8-fold higher than in nonheated tumors (64).

FIGURE 2.

Elastin-like polypeptides (ELPs) for thermally targeted drug delivery A: when designed for thermal targeting, ELPs with a transition temperature (T_t) slightly above body temperature, in the range of 39–41°C, are chosen. B: as with the soluble ELP drug carriers, thermally targeted ELPs can be modified at both the NH₂ and COOH termini with cargo. For cancer drug delivery, cargo has historically been cell-penetrating peptides or targeting peptides and antiproliferative therapeutic peptides or reactive sites for small-molecule drug conjugation. C: thermally targeted ELPs are engineered with a T_t slightly above body temperature, so they will remain soluble in systemic circulation but aggregate and accumulate in heated tumors. D: focused, mild hyperthermia can be applied clinically by adapting techniques traditionally used for ablative hyperthermia, including high-intensity focused ultrasound and radiofrequency ablation. E: systemically circulating ELPs will aggregate and accumulate in heated tumor vasculature and then resolubilize and extravasate upon return to normothermia (soluble ELP monomers not drawn to scale).

Thermally Targeted Delivery of Small-Molecule Chemotherapeutics

ELPs have been utilized for thermally targeted delivery of small-molecule chemotherapeutics, including doxorubicin (12, 65) and paclitaxel (66). The chemotherapeutic agents were attached at reactive sites engineered onto the termini of the ELP, and a variety of cleavable linkers, from acid labile to protease cleavable, have been utilized to mediate drug release upon cellular uptake (10, 12, 65). An ELP construct containing an NH₂-terminal Tat CPP and a single molecule of doxorubicin attached at the COOH terminus following a cathepsin-cleavable linker demonstrated antiproliferative activity in uterine sarcoma cells, and the antiproliferative potency was enhanced 20-fold when treatment was combined with mild hyperthermia (65). The ELP-delivered doxorubicin also overcame P-glycoprotein-mediated export and drug resistance in multidrug-resistant cancer cells (67). A similar construct utilizing an ELP modified NH₂-terminally with the SynB1 CPP and COOH-terminally with up to three doxorubicin molecules, fused via an acid-labile cleavable linker, showed improved potency in breast cancer cells (10). Importantly, this ELP delivery system slowed the plasma clearance relative to free doxorubicin when delivered intravenously in mice bearing orthotopic breast tumors, and the ELP-delivered doxorubicin accumulated in the tumor in a manner that was enhanced twofold by thermal targeting. In contrast to free doxorubicin, the ELP-delivered doxorubicin was not detectable in the heart, and the ELP-delivered drug consequently had a higher maximum tolerated dose. When combined with hyperthermia, the ELP-delivered doxorubicin effectively slowed tumor progression, and the effect was optimized, completely preventing tumor growth, when the maximum tolerated dose was used in combination with tumor hyperthermia applied immediately after ELP-Dox injections (10).

Thermally Targeted Delivery of Therapeutic Peptides via ELP Fusion

Because of the ability to edit the ELP coding sequence, the system is especially well suited for delivery of therapeutic peptides. Chimeric proteins containing targeting peptides or CPPs at one terminus and therapeutic peptides at the other terminus have been produced for a variety of applications, including thermally targeted versions developed for cancer therapy. An ELP-fused peptide inhibitor of the oncogenic transcription factor c-Myc was developed and tested with a variety of CPPs. A version delivered using the penetratin CPP blocked c-Myc-activated transcription and inhibited proliferation in breast cancer cells, and the cellular uptake and antiproliferative effect were enhanced when ELP coacervation was induced with mild hyperthermia (68). Potency and nuclear targeting of this ELP-delivered peptide were improved by iterating the CPP moiety (69), demonstrating that CPP choice can be used to achieve optimized cellular uptake and to drive intracellular distribution. This improved protein, utilizing the Bac-7 CPP, was tested in rodent models of breast cancer and glioma. In an orthotopic breast cancer model, the Bac-ELP-delivered c-Myc inhibitory peptide accumulated in the tumor after systemic intravenous or intraperitoneal injection, and tumor levels were enhanced by focused tumor hyperthermia (58). The protein extravasated in the tumor parenchyma and was detectable in tumor cells. Tumor progression and final tumor mass were reduced in mice treated with the ELP-delivered peptide plus tumor hyperthermia, and the protein was well tolerated, with no body weight loss. The same chimeric protein was also tested in a rat model of glioma achieved by intracranial implantation of tumor cells (14). It was cytotoxic in three different glioma cell lines in a heat-enhanced manner, the protein escaped the tumor vasculature and accumulated in intracranial glioma tumors, and tumor levels were enhanced nearly fourfold by focused hyperthermia. Tumor progression was followed with contrast-enhanced MRI, and the Bac-ELP-delivered inhibitory peptide combined with targeted hyperthermia induced an 80% reduction in tumor volume relative to untreated tumors. Rats treated with the protein plus tumor hyperthermia also had a 72% reduction in neurological motor deficits relative to rats bearing untreated tumors, and survival was significantly improved by treatment, from 20 days to >36 days after tumor implantation (14).

The thermally targeted ELP drug delivery system has been used to deliver several other therapeutic peptides. A CPP-ELP-delivered peptide that activates the p21 cyclin-dependent kinase inhibitor pathway caused cell cycle arrest and thermally enhanced inhibition of proliferation in a number of cancer cell lines in vitro (37, 70). A related protein showed potency against pancreatic cancer cells and accumulated in pancreatic tumor xenografts in a thermally targeted manner (71). A CPP-ELP-delivered antiproliferative peptide derived from bovine lactoferrin caused mitochondrial membrane depolarization, caspase activation, and inhibition of cell proliferation that was enhanced 30-fold by hyperthermia in pancreatic cancer cells (72). An ELP-delivered peptide that targets assembly of splicing factors inhibited proliferation in cancer cells (73), and an ELP-delivered proapoptotic peptide exhibited improved cellular uptake, improved induction of apoptosis, and enhanced antiproliferative potency with hyperthermia in estrogen receptor-positive and estrogen receptor-negative breast cancer cells (74). A CPP-ELP-delivered protein that acts as a dominant-negative inhibitor of Notch signaling killed glioma cells in a thermally targeted manner when used as monotherapy (75), and it accumulated in heat-targeted breast tumor xenografts and sensitized them to paclitaxel treatment when used in a combination therapy (76). In an application outside of the cancer field, a CPP-ELP-delivered peptide that inhibits the activation of the receptor for advanced glycation end products (RAGE) via S100B was capable of blocking S100B-induced oxidative damage in cultured neurons, and the protein was effectively delivered to the cerebellum via thermal targeting (77, 78).

Self-Associating ELP-Based Nanoparticles

ELPs can be engineered to self-assemble into nanoparticles with precisely defined structural properties. Two major strategies have been employed to produce ELP nanoparticles. The first is the use of di-block ELP copolymers. Di-block ELPs incorporate two distinct regions in one fusion protein. The first region contains a hydrophilic amino acid in the repeat motif’s “guest residue” position, imparting a high T_t to that domain. The second region incorporates a hydrophobic guest amino acid, giving that domain a T_t that is below body temperature. These di-block copolymers undergo a biphasic phase transition in which the hydrophobic domain collapses first and nucleates the formation of a micellar nanoparticle (FIGURE 3, A AND B) (79). At very high temperatures, well above body temperature, the hydrophilic domain will also phase transition, leading to the formation of micrometer-sized coacervates. The size of the nanoparticles and the critical micelle concentration are a function of the length and the hydrophobicity of the two domains (80). This core di-block copolymer can be modified with reactive sites for drug attachment or produced as a fusion protein with peptide or protein cargo to achieve nanoparticles with targeting or therapeutic moieties expressed on their corona (FIGURE 3C) (81–83).

FIGURE 3.

Elastin-like polypeptide (ELP) nanoparticles A: di-block ELP nanoparticles utilize a hydrophobic block with a transition temperature (T_t) below body temperature and a hydrophilic block with a T_t above body temperature, giving them a biphasic aggregation profile. B: the hydrophobic domain incorporates guest residues such as valine, isoleucine, and phenylalanine, which coacervate and stimulate micelle formation at body temperature. C: di-block ELP nanoparticles are organized as micelles with the hydrophobic domain in the core and the hydrophilic domain in the shell, with any cargo attached at the terminus of the hydrophilic domain displayed on the corona. D: a second strategy to drive ELP nanoparticle formation is to attach multiple copies of a hydrophobic small molecule at one of the protein’s termini. E: when a critical number of attached small molecules is reached, these proteins self-assemble into micelles with the hydrophobic drug at the core.

The second strategy that has been employed for ELP nanoparticle formation is to fuse multiple copies of a hydrophobic small molecule at one terminus of an ELP with a neutral or slightly hydrophilic guest residue (FIGURE 3D) (9). The high concentration of hydrophobic small molecules attached at one end of the ELP drives micelle formation (FIGURE 3E). By incorporating a cleavable linker in the drug attachment chemistry or by dissolving more hydrophobic free drug cargo in the micelle, this strategy has been used to deliver chemotherapy agents for tumor-targeted cancer therapy (9, 84).

ELP Nanoparticles Made from Di-block ELP Copolymers

In one application that was a modification of the soluble carrier for rapamycin described above, di-block ELP nanoparticles were engineered using a 48-pentapeptide repeat hydrophobic block containing an isoleucine or phenylalanine guest residue and a 48-pentapeptide repeat hydrophilic block containing a serine residue with the FKBP12 protein attached in-frame at the NH₂ terminus (85). These fusion proteins formed 50-nm particles at body temperature. Rapamycin was bound by the FKBP12 protein that assembled on the particle’s corona, and additional drug was dissolved in the hydrophobic core of the micelle. The release rate of the drug from the core depended on the hydrophobicity, with the phenylalanine core nanoparticle releasing drug more slowly than the isoleucine core nanoparticle. About 70% of the drug was bound in the core, and 30% was bound to FKBP12 in the corona, resulting in biphasic release kinetics with 1.9-h and 57.8-h half-lives. When tested in an aggressive breast cancer xenograft model, the nanoparticle-delivered rapamycin totally prevented tumor growth and was more tolerable than free drug, inducing no weight loss versus animals in the free drug group that had to be removed from the study because of weight loss (85). An additional advancement was made on this strategy by taking advantage of the ELP phase transition-driven nanoparticle formation to generate complex nanoparticles formed from a mixture of two different di-block ELPs. The FKBP12-fused di-block ELPs were mixed with similar di-block ELPs fused to an integrin-binding RGD tripeptide motif (86). At body temperature, the two mixed ELPs assembled into a complex nanoparticle containing both FKBP12 and RGD at the corona. These bifunctional nanoparticles assembled and released rapamycin similarly to the parent nanoparticles, but the addition of the RGD targeting domain resulted in enhanced tumor targeting and effective tumor growth suppression at a threefold lower dose than the first-generation nanoparticles.

As described above, the mTOR inhibitor rapamycin is also a treatment for the dry-eye sequelae of Sjögren’s syndrome. The same particles were also tested in a mouse model of Sjögren’s syndrome, where they exhibited a 62.5-h terminal half-life after systemic administration, an increase of fivefold over the half-life of free drug (87). Systemic delivery of these nanoparticles suppressed lymphocyte infiltration into the lacrimal gland with less drug-induced toxicity than free drug. A similar nanoparticle was engineered with the same di-block structure but with the mitogenic protein lacritin fused and displayed on the corona for treatment of corneal infection or injury (88). When administered topically as eye drops, these nanoparticles promoted corneal wound healing, with improved integrity of the corneal epithelium, demonstrating the diversity of the system for delivery of multiple active cargoes via multiple administration routes.

ELP Nanoparticles Made Using Asymmetrically Attached Hydrophobic Small-Molecule Drugs

Nanoparticle formulations for delivery of the anthracycline doxorubicin were created for cancer therapy by fusing multiple copies of doxorubicin to the COOH terminus of ELP. An ELP construct with a T_t well above body temperature and a series of cysteine residues at the COOH terminus in the form of eight glycine-glycine-cysteine tripeptide (GGC) motifs was utilized. When no drug was conjugated, this protein remained a soluble monomer. However, when a doxorubicin analog, containing a maleimide reactive group to bond to the cysteine thiol and an acid-labile linker, was conjugated at the string of terminal GGC motifs, the protein-drug conjugate spontaneously assembled into a 40-nm nanoparticle with doxorubicin at its core (9). The nanoparticle released doxorubicin at pH 5 but not at neutral pH. When systemically administered in a mouse model of colon cancer subcutaneous xenografts, the nanoparticle cleared from circulation with a terminal plasma half-life of 9.3 h, and it accumulated in tumors at levels 3.5 times higher than free doxorubicin. Importantly, since cardiotoxicity is dose limiting for doxorubicin, the nanoparticle formulation accumulated in the heart 2.6-fold less than free drug and had a four times higher maximum tolerated dose. A single bolus dose of the nanoparticle eliminated the C26 colon cancer tumors, significantly outperforming free drug. These nanoparticles also reduced metastasis burden, time to onset, and extended life in mouse models of highly metastatic breast cancer and Lewis lung carcinoma (89). In a follow-up study, the doxorubicin delivery nanoparticle was iterated by changing the ELP domain to select one that exhibited a phase transition in the 39–41°C thermal targeting range (90, 91). These ELPs were conjugated with an average of 5.4–6.3 doxorubicin molecules per protein molecule, and they made similar 40-nm nanoparticles. However, the nanoparticles using this ELP moiety were thermally responsive at just above physiological temperature and maintained their potency under hyperthermia (90). Plasma clearance rate and biodistribution were dependent on ELP chain length, with larger ELP sizes having reduced off-target accumulation in the heart and kidney. Intravital microscopy revealed that, upon application of cycles of focused hyperthermia, the nanoparticles coacervated into micrometer-sized particles in the tumor vasculature, a process that increased tumor targeting by 2.6-fold (91).

A similar strategy utilized a hydrophilic hexapeptide ELP with 70 repeats containing 32 cysteine residues at one end for delivery of the small-molecule ionophore salinomycin for cancer therapy (84). Salinomycin was fused to the cysteine residues, inducing micelle formation, and then additional free drug was dissolved in the hydrophobic core. To optimize particle diameter, loading, and release, a complex particle was created from this base micelle by addition of α-tocopherol, resulting in a final particle of 180-nm diameter with a 4.1-h salinomycin release half-life. When delivered via this nanoparticle system, the in vivo half-life of the drug was extended 5.2-fold and it accumulated in tumors at 2.4-fold higher levels than free drug, with less accumulation in the heart. The nanoparticle drug was more effective than free drug at reducing the progression of the highly aggressive 4T1 breast cancer model.

ELP Coacervate Controlled-Release Depots

Using ELPs that phase transition below body temperature, it is possible to create depots for slow-release drug delivery. The T_t of ELP is easily engineered by modifying the polypeptide length and guest residue hydrophobicity to create ELP-fusion proteins with transition temperatures above room temperature but below body temperature (FIGURE 4A). When administered in vivo, these ELP-fusion proteins will form coacervates that persist from days to weeks. These polypeptides can then be fused with drugs or therapeutic peptides/proteins, incorporating protease cleavable linkers as desired (FIGURE 4B), or even labeled with radioactive isotopes. Prominent examples of drug delivery applications taking advantage of this strategy include intra-articular delivery of coacervate ELP gels carrying anti-inflammatory therapeutics (92–94), subcutaneous delivery of coacervate ELP gels that slowly release bioactive peptides (FIGURE 4C) for therapy of diabetes and other chronic conditions (95–97), and even intratumoral delivery of radiolabeled ELP gels for brachytherapy (64, 98, 99).

FIGURE 4.

Elastin-like polypeptide (ELP) subcutaneous slow-release depots A: when designed for subcutaneous depot formation, ELPs with a transition temperature (T_t) below body temperature are chosen. B: the ELP can be modified at the termini with multiple copies of a bioactive peptide, each separated by proteolytic cleavage sites. C: the fusion proteins are soluble at room temperature in a syringe, but they form subcutaneous coacervates upon injection that can persist for weeks. The coacervates are slowly cleaved by extracellular proteases, or they slowly redissolve and leach monomers. The released cargo are absorbed into circulation, resulting in an extended-release formulation.

ELP Coacervates as Anti-inflammatory Agents for Orthopedic Injuries

Initial work with ELP coacervate-mediated drug delivery was conducted for orthopedic applications. ELPs were fused to anti-inflammatory proteins, including interleukin 1 receptor antagonist (IL1-RA) (92) and a soluble tumor necrosis factor (TNF) receptor (94). The ELP coacervation had dramatic effects on in vivo longevity. The in vivo half-life of the ELP coacervate was 7-fold longer than the half-life of a soluble ELP when injected into the L₅ dorsal root ganglion (DRG) in rats (93) and 25-fold longer when given intra-articularly in the knee joint of rats (100). When used to deliver a soluble TNF receptor to the DRG, the treatment decreased markers of inflammation and prevented metabolic stress when challenged with TNF (94). ELP fusions with soluble TNF receptor or IL1-RA were also tested via intra-articular injection, where they slowly released cargo over a 7-day period and reduced synovial inflammation in a tibia fracture model (101). In a related application, ELP coacervates were used to deliver the natural product curcumin in a mouse model of radiculopathy (102). Curcumin was covalently conjugated to glutamate residues engineered into the ELP repeat. This created ELP monomers each with 5.9–7.2 cargo molecules evenly distributed throughout. The conjugation eliminated the negative charge on the glutamate residues, leaving the other hydrophobic valine and isoleucine guest residues to dominate and decrease the transition temperature. When administered intramuscularly proximal to the sciatic nerve, the ELP-bound curcumin was more soluble than free drug, which precipitated in vivo. These results highlight the promise of ELP coacervates, given as injections into joints or into tissue adjacent to sites of injury, to deliver anti-inflammatory agents for treatment of orthopedic injuries.

Subcutaneous ELP Coacervates for Controlled Release of Bioactive Peptides

A potential limitation of soluble ELP-fusion proteins is that they must be parenterally administered because their proteinaceous nature limits their oral bioavailability. To circumvent this limitation for chronic conditions such as diabetes, ELP coacervate controlled-release depots have been used to achieve slow release of active peptides for long-term blood sugar control. Using what they termed “protease operated depots,” original work in this area utilized an ELP with a T_t below body temperature that was fused to six copies of glucagon-like peptide 1 (GLP-1), each separated by a protease cleavage site (95). GLP-1 is a 30- to 31-residue peptide that functions as an incretin, inducing insulin production and lowering blood sugar levels (103). When administered into the subcutaneous space, the ELP-fused GLP-1 formed a depot, and extracellular proteases slowly cleaved the linking residues, releasing active GLP-1 peptides into circulation (FIGURE 4C). The original depot formulation persisted for 5 days in the subcutaneous space, whereas a soluble ELP-GLP-1 comparator fusion protein was cleared within 72 h of injection. Release rate was also dependent on the size of the ELP moiety. A version containing a 25-kDa ELP successfully reduced plasma glucose levels for a period of 48 h after injection, and a 100-kDa ELP-GLP-1 depot reduced plasma glucose for 120 h (95). With a version with a slightly modified cleavage site and a 50-kDa ELP, the depot persisted for 7 days in the subcutaneous space and achieved 144 h of mean plasma glucose reduction and improved glucose tolerance following a glucose challenge, even 102 h after implantation (96). In a demonstration of the versatility of the ELP carrier, the GLP-1 depot strategy was optimized by iterating both the hydrophobicity of the ELP guest residue and the length of the ELP domain (97). This work demonstrated that ELPs with a T_t above body temperature exhibit bolus release and ELPs with a T_t far below body temperature exhibit minimal release, with optimal release of cargo from the depot occurring for ELPs with a T_t of ∼30°C. The optimized ELP-GLP-1 construct was tested in three mouse models: C57BL/6 mice fed a high-fat diet for 11 wk, leptin-deficient ob/ob mice, and leptin receptor-deficient db/db mice. The treatment effectively lowered blood glucose for at least 240 h in all models and induced weight loss in the diet-induced obesity and ob/ob models, with no evidence of histopathological damage to major organs. When given as a once-weekly subcutaneous injection in ob/ob mice, the ELP-GLP-1 depot reduced hemoglobin A1c relative to control by day 28, and it remained lowered at day 56. Similarly, in diet-induced obesity, treated mice gained less weight and had lower hemoglobin A1c after 56 days of once-weekly injections. Translational studies in cynomolgus monkeys revealed that the depot release resulted in constant GLP-1 plasma levels for 10 days after subcutaneous injection, with levels still detectable for 17–21 days, suggesting that once-per-month dosing is potentially possible in humans (97).

ELP Coacervates for Brachytherapy and Tumor Drug Delivery

ELPs with low T_t can also be administered by direct intratumoral injection, where they will form coacervates for sustained drug delivery. If the ELP molecules themselves are conjugated with strong gamma-emitting radioisotopes, this strategy can be used for brachytherapy. ¹³¹I-labeled ELPs with a T_t either below or above body temperature were administered by direct injection into 4T1 breast tumors. The proteins with low T_t that were able to form coacervates were retained significantly longer in the tumor than the soluble ELPs, with half-lives of 44.2 versus 8.3 hours, respectively, and 30% of the coacervate still present in the tumor 72 h after injection (98). These intratumor coacervates slowed progression of the breast tumors by delivering localized ionizing radiation. In a follow-up study, this brachytherapy approach led to growth inhibition in FaDu and PC-3 xenografts and complete tumor remission in 67% of animals (99). A related strategy utilized a hydrophobic ELP with a COOH-terminal tail of seven repeating glycine-tyrosine (GY) dipeptides that were labeled with ¹³¹I (104). With the asymmetric GY tail, these ELPs formed micelles at room temperature that then aggregated into large coacervates upon intratumoral injection. The coacervates had a 7- to 7.5-day half-life in orthotopic prostate and pancreatic cancer models. Some protein spillover was detected in off-target organs 3 days after implantation, but it was quickly cleared. Within 9 days of implantation, this brachytherapy approach induced significant reduction in prostate tumor size, and tumors were reduced to 5% of their original size by 26 days after implantation. Although some tumor relapse was observed after clearance of the depot, this brachytherapy application and other examples of delivery of fused therapeutics (105, 106) demonstrate the promise of ELPs for intratumor drug delivery, or for delivery of therapeutics to the site of a tumor after resection in order to inhibit recurrence (107).

Conclusions

ELPs represent a versatile platform on which to build drug delivery technology. The proteinaceous nature, low immunogenicity, customizable size, and phase transition temperature, combined with the ability to carry nearly any type of cargo from small molecules to fusion proteins, make ELPs unique and ideal carriers for a wide variety of therapeutic applications. As demonstrated by the diverse studies reviewed here, ELPs can be used in their soluble form as fusion-protein biologics with tunable pharmacokinetics, engineered for thermally targeted drug delivery, formulated into nanoparticles with a wide array of applications, and engineered to form controlled-release drug depots for treatment of more chronic conditions. Currently, several ELP-based depot formulations are in clinical testing for diabetes, heart failure, and pulmonary artery hypertension (108–111), and our own soluble ELP-VEGF fusion biologic is undergoing primate testing for a preeclampsia indication and toxicology testing for renal indications. These translational efforts make it highly likely that ELP-fusion proteins will soon be available as clinically approved therapeutics, and the powerful platform will likely continue to be used to create even more clinically useful agents in the future.