Engineering growth factor ligands and receptors for therapeutic innovation

Abstract

Growth factors signal through engagement and activation of their respective cell surface receptors to choreograph an array of cellular functions, including proliferation, growth, repair, migration, differentiation, and survival. Due to their vital role in determining cell fate and maintaining homeostasis, dysregulation of growth factor pathways leads to the development and/or progression of disease, particularly in the context of cancer. Exciting advances in protein engineering technologies have enabled innovative strategies to redesign naturally occurring growth factor ligands and receptors as targeted therapeutics. Here, we review growth factor protein engineering efforts, including affinity modulation, molecular fusion, and the design of decoy receptors, dual-specific constructs, and vaccines. Collectively, these approaches are catapulting next-generation drugs to treat cancer and a host of other conditions.

Growth factors: Key components of cellular communication pathways

Growth factors represent a diverse and essential class of proteins, intricately programmed to regulate a range of cellular activities including proliferation, growth, repair, migration, differentiation, and survival. These signaling molecules, which are either secreted from cells or expressed as membrane-bound precursors that are then released into the extracellular space via protease cleavage, orchestrate critical functions vital for development, healing, and maintenance of tissue homeostasis. Through binding to their specific cognate receptors on target cells, growth factors trigger downstream signaling cascades that effect transcriptional programs to induce distinct cellular outcomes [1,2].

At the heart of cellular communication, growth factor ligand/receptor interactions serve as the cornerstones of complex signaling networks. Growth factors can be categorized into several families, including the vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), nerve growth factor (NGF), and insulin-like growth factor (IGF) families [3]. The different families of growth factors exhibit structural and functional diversity, although growth factors also show a high degree of pleiotropy and redundancy. Pleiotropy indicates that a single growth factor can exert multiple effects across different receptors, cell types, or tissues (Figure 1A, Key figure), while redundancy points to the shared functions and common receptors between different growth factors (Figure 1B). Moreover, some growth factor receptors are activated independent of ligands, through heterodimerization with other receptors, showcasing the intricate and functionally diverse nature of growth factor signaling pathways.

Figure 1. Key figure. Overview of growth factor signaling.

(A) Growth factor pathways exhibit pleiotropy, wherein a single growth factor can bind to multiple different receptors to elicit diverse cellular responses across various cell types, enhancing cellular complexity and functional versatility. (B) Redundancy describes the ability of different growth factors to bind a common receptor to initiate similar cellular responses, conferring robustness to biological systems. (C) Mechanisms of activation for growth factors involve specific interactions between growth factor ligands and their cognate receptors, leading to receptor dimerization and triggering downstream signaling cascades. Figure was created with Biorender.com.

The structural properties of growth factor ligands and their receptors are paramount to their function. Growth factor ligands are generally small (≈6-35 kDa), soluble proteins comprised of a single domain and often feature conserved sequences or structural motifs [3–6]. They are secreted from cells or proteolytically cleaved from membrane-bound precursors as either monomers or oligomers, and their unique 3-dimensional orientations are critical for their specific interactions with cognate receptors. Growth factor receptors typically exhibit a single-pass transmembrane configuration with an extracellular ligand-binding domain and an intracellular domain responsible for signal transduction [7]. These structures are characterized by convergent features, such as extracellular immunoglobulin (Ig) domains and dimerization-facilitating cysteine-rich domains, as well as intracellular kinase domains, which mediate signaling.

Growth factor signaling is initiated when extracellular growth factor ligand associates with membrane-expressed receptors. This typically leads to dimerization of growth factor receptors in one of two categories: (1) homodimerization, wherein both receptor subunits are identical; and (2) heterodimerization, wherein the receptor subunits are distinct [8]. Dimerization of receptor extracellular domains leads to interaction between the intracellular kinase domains on each receptor monomer, resulting in activation [9]. In the specific case of the EGFR system, structural evidence has revealed that receptor homodimers are asymmetric; the N-terminus of the “acceptor” kinase is positioned adjacent to the C-terminus of the “donor” kinase, meaning that the resulting activation is orientation dependent [9,10]. Growth factor signaling triggers various intracellular cascades, the most common of which are Janus kinase/signal transducer and activator of transcription (JAK/STAT), mitogen activated protein kinase (MAPK), and phosphoinositide 3-kinase (PI3K) pathways (Figure 1C) [11–14]. In the JAK/STAT signaling pathway, the endogenous portion of each receptor is constitutively associated with JAK proteins, which phosphorylate one another upon receptor dimerization, leading to phosphorylation of key intracellular residues on the receptor that serve as docking sites for activation of STATs [12].

Once phosphorylated, STATs dimerize and translocate to the nucleus to alter gene expression [12,15]. The MAPK and PI3K pathways use guanine nucleotide exchange factors (GNEFs) to catalyze the exchange of guanidine diphosphate (GDP) for guanidine triphosphate (GTP) to activate the adaptor protein Ras, leading to a phosphorylation cascade involving multiple intracellular kinases [11,16]. The cytoplasmic kinase proteins extracellular signal-regulated kinase (ERK) and AKT are the respective final targets of the MAPK and PI3K pathways, and these molecules phosphorylate transcription factors, which manipulate the expression of genes that regulate cell fate [17,18].

Due to their critical functions in controlling cell behavior, growth factor ligands and receptors have been leveraged and targeted in several biomedical contexts, including early development, tissue engineering, wound healing, regenerative medicine, aging, cardiovascular disease, and cancer [19–21]. In particular, their promotion cell survival, proliferation, and differentiation makes growth factors desirable engineering targets for repairing damaged tissue [19]. Conversely, in cancer, elevated growth factor levels can lead to dysregulation of growth and apoptosis, resulting in uncontrolled cell expansion [21]. Thus, by engineering small molecule inhibitors, antagonist proteins, and neutralizing antibodies, cancer-associated signaling dysfunction can be corrected for effective disease treatment [22,23].

While growth factor-inspired treatments show promise for many disease types, several intrinsic properties of these proteins counteract their efficacy as drugs. First, growth factor signaling is pleiotropic, and the receptors involved can elicit a variety of often contradictory outcomes. Additionally, the pleiotropic nature of growth factors predisposes them to off-target effects, which can be harmful. The potential for growth factors to interact with unintended tissues not only can lead to unwanted cellular proliferation but also decreases the dose delivered to the intended target tissue [24]. Finally, due to their small size, growth factors are rapidly cleared from circulating blood by the kidneys, making long-term persistence in the body a challenge [25]. Collectively, these challenges have limited the safety and efficacy of growth factor-inspired drugs in the clinic.

To overcome the limitations of natural growth factor ligands and receptors as therapeutics, emerging approaches to engineer these molecules have led to significant improvements in stability, affinity, specificity, and safety. By employing innovative molecular engineering techniques, researchers have crafted modified growth factors with prolonged half-lives, increased tissue specificity, selective signaling pathway activation, and fewer side effects. These strategies have led to new therapies for cancer and other diseases, as well as advancing technologies in regenerative medicine. In this review, we describe the biology of growth factor systems and explore recent breakthroughs in the engineering of growth factor systems, emphasizing developments in the VEGF, EGF, NGF, FGF, and IGF families that set the stage for designing novel medical interventions, with a particular focus on developments relevant to oncology.

The vascular endothelial growth factor (VEGF) system

The VEGF system, initially identified in the 1980s [26–28], plays a crucial role in angiogenesis, the formation, growth, and permeabilization of vasculature, which is essential for tissue repair, but also drives disease progression, for example in the context of cancer. The human VEGF system includes ligands VEGF-A through D. The closely related placental growth factor (PlGF)-1 and PlGF-2 also share receptors with the VEGF ligands [29,30]. All VEGF ligands present as dimers, which self-assemble through cysteine-mediated disulfide bonds to activate their biological functions [31]. VEGF ligands play critical roles in activating vascular endothelial cells, serving as crucial mediators of tissue repair, but also promoting pathogenesis through the growth of aberrant, leaky, and disease-sustaining blood vessels in the context of cancer and neovascular diseases.

VEGF growth factor ligands interact with 3 different surface receptors: VEGF receptor (VEGFR)-1, VEGFR-2, and VEGFR-3 [32]. VEGFRs comprise 3 functional fragments: (1) an extracellular portion containing 7 Ig-like subunits; (2) a lipophilic transmembrane domain; and (3) an intracellular domain with tyrosine kinase activity [33,34]. As members of the receptor tyrosine kinase (RTK) superfamily, VEGFRs undergo autophosphorylation of their kinase domains upon homo- or heterodimerization following recognition of specific ligands: VEGFR-1 binds to VEGF-A, VEGF-B, and PlGF-1 and 2; VEGFR-2 binds to VEGF-A, as well as the post-proteolytically processed forms of VEGF-C and VEGF-D; and VEGFR-3 binds to the full VEGF-C and VEGF-D ligands [33,34].

Early therapeutic strategies targeting the VEGF pathway focused on small molecule and antibody inhibitors targeting VEGF ligands or receptors [35]. A noteworthy example of continued work in this field is the development of bispecific antibodies, designed to simultaneously target a VEGF protein and a second molecular target, such as a different growth factor or immune checkpoint protein. This dual-targeting approach aims to couple angiogenesis inhibition with other therapeutic modalities [36]. These targeted interventions have been reviewed elsewhere [35,37–40] and this review will focus on strategies for engineering VEGF ligands or receptors themselves.

VEGF receptor engineering

VEGFR-based antagonists represent a strategic approach to blocking the VEGF pathway in order to inhibit angiogenesis. This approach is particularly applicable in cancer therapy, as limiting tumor blood supply can in turn limit tumor growth and metastasis [28]. Unfortunately, the effectiveness of anti-VEGF therapies in cancer is hindered by challenges of tumor permeability, drug delivery, and resistance [40,41]. In particular, tumors can develop resistance to VEGFR-based antagonists by activating alternative angiogenic pathways, altering tumor metabolism to reduce dependency on angiogenesis, or by forming abnormal vasculature that decreases drug delivery to the tumor [40–42]. Furthermore, the tumor microenvironment, including stromal and immune cells, can adapt to support tumor growth despite inhibition of growth factor signaling. Newer approaches to overcome resistance include the development of combination and multi-targeted therapies that disrupt two or more pathways simultaneously to enhance efficacy [40]. For example, agents that can block VEGF while also inhibiting other pro-angiogenic factors and/or pathways involved in tumor growth and metastasis are being explored. These bispecific antagonists aim to boost the affinity and specificity of the therapy, while also preventing resistance by limiting the tumor’s ability to compensate for the blocked pathway through alternate routes [43].

Due to the limitations of engineered VEGF pathway antagonist strategies in cancer, recent approaches have shifted towards addressing eye conditions, due to the critical role VEGF plays in abnormal angiogenesis that leads to vision loss in prevalent diseases such as neovascular age-related macular degeneration (nvAMD) and diabetic retinopathy (DR) [41,44]. One of the most successful VEGFR-based antagonist strategies to date has been the design of decoy receptors, which are engineered soluble proteins that mimic natural VEGFRs to sequester VEGF ligands, thereby preventing these molecules from activating their cognate cell surface receptors to inhibit angiogenesis. This decoy receptor strategy, also known as ligand trapping, has been employed for a variety of growth factors and other soluble effector proteins, allowing for targeted antagonism of disease-driving pathways. A successful clinical development of decoy receptors includes aflibercept (Eylea^®, Regeneron Pharmaceuticals), which is comprised of VEGF-binding domains (VEGFR-1 domain 2 and VEGFR-2 domain 3) fused to a dimerizing human immunoglobulin (Ig) G1 fragment crystallizable (Fc) domain (Figure 2A) [45,46]. Aflibercept binds to VEGF-A isoforms, VEGF-B, PlGF-1, and PlGF-2, but does not engage VEGF-C or VEGF-D [45]. Due to its anti-angiogenic activities, aflibercept has been clinically approved since 2011 to treat neovascular eye diseases, including nvAMD and DR [47].

Figure 2. Engineering strategies for the vascular endothelial growth factor (VEGF) system.

(A) An engineered VEGF receptor (VEGFR)-based decoy protein known as aflibercept fuses the VEGF-binding domains of VEGFR (VEGFR-1 domain 2 and VEGFR-2 domain 3) to the human IgG1 Fc domain to form a homodimer. Aflibercept binds the PlGF-1, PlGF-2, VEGF-A, and VEGF-C growth factors with higher affinity than surface-bound VEGFR-1 and VEGFR-2, preventing activation of downstream signaling. (B) An engineered VEGF ligand-based antagonist protein derived from a single-chain VEGF contains one molecule that is engineered for enhanced binding affinity to VEGFR-2, while the molecule is modified to prevent VEGFR dimerization, inhibiting subsequent signaling. Figure was created with Biorender.com.

Recent advancements in the field have concentrated on enhancing affinity and specificity of VEGFR-based antagonists, with the goal of reducing treatment frequency to mitigate side effects and improve patient compliance. The yeast surface display platform was applied for directed evolution [48,49] to design a novel engineered decoy receptor based on aflibercept that strengthened the binding affinity for VEGF-A through a decrease in dissociation rate [50]. This molecule was successfully delivered via suprachoroidal nanoparticle-based gene delivery with persistent expression in the choroidal space of rat eyes [50], illustrating a powerful synthesis between molecular engineering and gene therapy approaches. Advancements in VEGFR engineering have profound implications improving the specificity, binding, and delivery of engineered VEGFRs in oncology applications. Indeed, receptor decoys such as aflibercept have been employed as cancer therapeutics, for instance Ziv-aflibercept (Zaltrap^®, Sanofi-Aventis and Regeneron Pharmaceuticals), which targets metastatic colorectal cancer while sharing the same molecular structure as aflibercept with higher osmolarity [40,46,51].

VEGF ligand engineering

In addition to VEGF receptor engineering approaches, VEGF ligand engineering strategies have also been used to modulate angiogenesis for therapeutic purposes. These approaches have focused on modifying the VEGF protein to fine-tune its interaction with receptors, in order to improve its safety and efficacy as a therapeutic intervention. Several approaches have been implemented, including: (1) design of VEGF-derived antagonist proteins; and (2) development of stabilized VEGF ligands for more robust and sustained activity.

VEGF ligands naturally occur as dimers, with each protomer engaging a receptor molecule to induce dimerization. Ligand variants have been engineered to interfere with VEGF-induced dimerization in order to prevent downstream signaling and consequent angiogenic responses. VEGF-A isoforms are strongly implicated in neovascularization, and this collection of proteins reflects the natural diversity of VEGF ligands with respect to receptor engagement, both in terms of equilibrium binding and kinetics [52]. A previous study developed a single-chain VEGF (scVEGF) variant by linking two VEGF-A 121 monomers linearly through a 14 base pair flexible peptide linker, allowing for precise control over the molecular configuration and functionality of the resulting protein [53]. To create a dominant negative inhibitor antagonist ligand, site-directed mutagenesis was conducted on one of the receptor binding sites of the scVEGF to prevent receptor dimerization and activation, while retaining the ability for the other VEGF monomer to bind VEGFR-2. The resulting scVEGF molecule interfered with a specific VEGF pathway, while also minimizing the unintended interactions that often accompany broad-spectrum VEGF inhibitors.

In a similar vein, another study applied directed evolution using yeast surface display platform on the scVEGF molecule to create a dual-specific scVEGF variant [54]. The resulting molecule presents one VEGF monomer that contains 4 point mutations conferring higher affinity for VEGFR-2 compared to the natural bivalent VEGF ligand, and a second monomer that presents a binding loop for αvβ3 integrin, a receptor known to synergize with VEGFR-2 in promoting angiogenesis. Functional assays, including surface plasmon resonance studies and cellular binding tests, demonstrated that these dual-specific proteins could engage both receptors simultaneously, leading to robust inhibition of VEGF-mediated endothelial cell functions and angiogenesis, both in vitro and in mouse models [54]. Further building on the scVEGF platform, yeast surface display technology was applied to enhance the scVEGF affinity towards VEGFR-2, resulting in a distinct variant that binds the receptor with 14-fold higher affinity compared to native VEGF [55]. This monomer was then paired with a mutated VEGF monomer that cannot bind VEGFR-2, creating a dominant negative antagonist (Figure 2B). The resulting protein bound to human endothelial cells with a higher affinity than wild-type VEGF and was found to inhibit pathologic neovascularization in a mouse model [55].

Another approach to VEGF engineering entails the design of stabilized ligand variants. These proteins are designed to resist rapid degradation and maintain persistent bioactivity, thereby enhancing stability and bioavailability compared to the natural ligand. These engineered VEGF ligands have led to improved angiogenesis for applications such as wound healing and cardiovascular repair [56]. Previous studies developed therapeutic chimeras that fused VEGF-A and VEGF-B to elastin-like polypeptides (ELPs) as carriers for the treatment of preeclampsia [57,58]. They demonstrated ELP fusion to be an effective carrier strategy for VEGF ligands, as capillary formation was enhanced in mouse models in the presence versus the absence of ELP, highlighting the potential for ELP modification to augment the therapeutic activity of VEGF ligands.

Overall, innovative strategies in the VEGF space reflect the dynamic and rapidly evolving field of growth factor ligand/receptor engineering. Their clinical success, particularly in targeting ocular neovascularization, demonstrates the powerful potential for utilizing growth factor-inspired molecules as therapeutics. Moreover, as these strategies improve, it may be possible to leverage similar approaches to create inhibitors relevant to VEGF and other cancer-driving growth factors in the tumor microenvironment, such as EGF, PDGF, and IGF [21,59]. Moreover, multispecific designs targeting tumor-associated antigens as well as tumor-driving growth factor pathways will enable safe and effective disease targeting.

The epidermal growth factor (EGF) system

EGFR is an RTK that belongs to the human epidermal growth factor receptor (EGFR/HER/ErbB) family, which comprises four members: EGFR (also called ErbB1/HER1); HER2/ErbB2; HER3/ErbB3; and HER4/ErbB4 [60]. EGFR, which is one of the most broadly studied RTKs, is a 170 kDa transmembrane protein consisting of extracellular ligand-binding and dimerization regions, a hydrophobic transmembrane region, and an intracellular region [61]. The HER/ErbB family of receptors is activated by several distinct ligands that bind extracellularly, creating a complex ligand/receptor network. EGF, transforming growth factor alpha (TGF-α), epigen, and amphiregulin bind exclusively to EGFR; heparin binding-EGF (HB-EGF), epiregulin and betacellulin bind to both EGFR and HER4; neuregulin 1 (NRG1) and NRG2 bind to HER3 and HER4; and NRG3 and NRG4 bind exclusively to HER4 [62]. Interestingly, there is no ligand that interacts with HER2. Ligand binding induces conformational changes that result in receptor homo- or heterodimerization, leading to autophosphorylation of tyrosine residues on the intracellular domains of these receptors, which in turn phosphorylate various effector or adaptor proteins to activate downstream signaling pathways [61]. These signaling cascades induce a variety of cellular outcomes, including survival (JAK/STAT pathway, MAPK pathway), proliferation (MAPK pathway, PI3K/AKT pathway, phospholipase C (PLC)/protein kinase C (PKC) pathway), angiogenesis (PLC/PKC pathway, PI3K pathway), and metastasis (PLC/PKC pathway, JAK/STAT pathway) [61].

Due to their critical role in cell growth and development, the expression and activity of EGFR family receptors are tightly regulated to maintain homeostatic balance. Dysregulation of EGFR leads to pathological fates, such as cancer development. In particular, overexpression and/or mutation of EGFR is frequently observed in tumor cells and has been shown to strongly correlate with cancer progression and poor patient outcomes [60]. Point mutations and large deletions within the EGFR gene are commonly observed in cancer patients and can drive ligand-independent constitutive signaling. An example of this is the frequently observed variant III deletion (EGFRvIII), which removes exons 2–7, resulting in a truncated EGFR extracellular domain that does not engage ligands, but allows for constitutive dimerization to perpetually activate kinase activity [63]. Due to the ligand-independent activity of EGFRvIII, ligand-blocking interventions are ineffective against cancers that express this EGFR mutant, making it a particularly challenging drug target. Because of the prevalence of EGFR dysregulation in cancer and its association with disease progression, intensive efforts have focused on developing therapeutic interventions targeting the EGF system.

Targeting the EGF system for cancer treatment

Thanks to structural and biochemical insights into the EGF system, several successful strategies have been developed targeting the EGF/EGFR network. Currently, 2 approaches are prevalent in clinical practice: monoclonal antibodies which bind to the extracellular domain of EGFR family receptors; and small molecule drugs known as tyrosine kinase inhibitors (TKIs), which inhibit the phosphorylation of the intracellular tyrosine kinase domains of EGFR family receptors by competitively binding to the ATP binding pocket, thus preventing the phosphorylation of tyrosine residues at the C-terminal tails of these receptors and thereby inhibiting downstream signaling cascades [60]. Monoclonal antibodies and TKIs have shown success in cancer treatment; however, mutations in the target receptors, upregulation of ligands (including through autocrine production by cancer cells), and activation of compensatory signaling pathways in tumor cells impede the efficacy of these drugs and lead to therapeutic resistance [60,63]. Furthermore, on-target, off-tumor binding of administered drugs to EGFR in healthy tissues can lead to adverse effects, which most frequently manifest as skin rashes or gastrointestinal disorders [63].

EGF receptor engineering

To overcome the limitations of clinical monoclonal antibodies and TKIs, new approaches to EGF pathway inhibition are being explored, including receptor engineering strategies. As in the VEGF system, decoy receptors have been developed for the EGF system to block signaling, and these soluble antagonists have shown efficacy in cancer therapy (Figure 3A). A specific advantage of this approach is that through depletion of EGF family ligands, these molecules can potentially overcome ligand-mediated escape pathways established in response to monoclonal antibodies and TKIs. For instance, RB200, a purified heterodimer consisting of EGFR and HER3 separately fused to human IgG1 Fc fragments, traps multiple HER ligands, including TGF-α, HB-EGF, NRG1-α, and NRG1-β3 [64]. Through inhibiting tyrosine phosphorylation of HER family proteins, RB200 has been shown to antagonize tumor cell proliferation in both cellular and animal models. Moreover, RB200 synergistically inhibited tumor cell growth when administered in combination with EGFR and HER2 TKIs [64]. Similarly, TRAP-Fc, an EGF family ligand trap, which incorporates the extracellular ligand-binding domains of both EGFR and HER4, efficiently sequesters all 11 EGF family ligands [65,66]. TRAP-Fc inhibited the growth of several cancer cell lines and reduced tumor growth and metastasis across multiple human tumor xenograft and transgenic mouse models [65,66]. Furthermore, TRAP-Fc was found to sensitize tumor cells to cytotoxins, resulting in synergy with chemotherapeutic agents [66].

Figure 3. Engineering strategies for the epidermal growth factor (EGF) system.

(A) EGFR-based decoy receptors comprise extracellular domains of EGF family receptors (EGFR is shown here, but other family members have also been employed), which are separately fused to the human IgG1 Fc domain to form homo- or heterodimers. These decoy receptors can sequester multiple EGF family ligands, thereby preventing receptor dimerization and downstream signaling on tumor cells. (B) The vaccine Rindopepimut^® (CDX-110, Celldex) comprises a 14-mer peptide of the aberrant EGFRvIII truncation mutant of EGFR conjugated to an immunogenic carrier protein. The vaccine is uptaken by antigen-presenting cells (APCs) and presented on major histocompatibility complexes (MHCs), which promotes the secretion of EGFRvIII-specific antibodies from B cells to selectively eliminate mutant receptor-expressing tumor cells. (C) Immunotoxin (TGF-PE40) fuses the EGFR ligand TGF-α to a toxin that can directly induce cell death following internalization by a tumor cell. Figure was created with Biorender.com.

While most reported decoy receptors for the EGF system employ EGF family receptors, several recent studies have developed bispecific decoy receptors capable of simultaneously targeting two signaling pathways as more robust strategies for cancer treatment. For example, ED2 is a fusion protein comprising domain 2 of VEGFR-1 fused to the C-terminus of the anti-EGFR monoclonal antibody cetuximab (Erbitux^®, Eli Lilly) [67]. ED2 can therefore concurrently engage both VEGF and EGFR, sequestering soluble VEGF, blocking EGFR signaling, and inducing internalization of EGFR for lysosome-mediated degradation in tumor cells. ED2 exhibited more potent antitumor activity in mouse models compared the anti-VEGF antibody bevacizumab and cetuximab alone or in combination [67]. Another bispecific decoy receptor, denoted VEGFR-EGFR/Fc, is a homodimeric fusion protein that contains domains 1-3 of VEGFR-1 and the full extracellular domain of EGFR fused to the human IgG1 Fc domain [68]. VEGFR-EGFR/Fc efficiently captures EGF, TGF-α, and VEGF, and this molecule exhibited promising anti-tumor efficacy in preclinical mouse models [68]. As with monospecific EGF family decoy receptors, bispecific decoy receptors also have potential to synergize with chemotherapy drugs.

An alternative approach to receptor-inspired therapeutic design for the EGF system is the development of vaccines that present EGFR family receptors in order to train the immune system to destroy cancer cells. Use of vaccine strategies also introduces the possibility of targeting particular EGFR mutants that are cancer-specific, reducing toxicity to healthy tissue. Moreover, activating immune cells against EGFR peptides could dampen cancer resistance through epitope spreading and provide long-lasting antitumor immune responses that prevent disease recurrence and metastases through effective immune surveillance [69]. Rindopepimut^® (CDX-110, Celldex) is a vaccine designed to treat EGFRvIII-positive glioblastoma multiforme (GBM) [70]. It comprises a 14-mer peptide that presents the aberrant amino acid sequence found in the EGFRvIII truncation mutant of EGFR, which is highly prevalent in GBM, conjugated to an immunogenic carrier protein called keyhole limpet hemocyanin (KLH) (Figure 3B). Vaccination with Rindopepimut^® has demonstrated efficacy in increasing B cell production of EGFRvIII-specific antibodies and selectively eliminating cells that express EGFRvIII [70]. Supportive clinical studies led to Rindopepimut^® receiving Breakthrough Designation for EGFRvIII-positive glioblastoma in 2015 [71]. However, it was later withdrawn due to lack of efficacy in subsequent trials. In another vaccine design strategy, an immunogenic 19 amino acid peptide that represents a region of the EGFR extracellular domain 4 with an isoaspartate modification was generated [69]. The investigators found that this peptide induced strong T cell responses against EGFR, which led to inhibition of tumor growth in immunized mice. Moreover, the antibodies generated by the host immune system were also capable of recognizing homologous regions of other EGFR family receptors, further enhancing antitumor activity [69].

EGF ligand engineering

In addition to receptor engineering strategies, several EGF family ligand engineering approaches have emerged. An exemplary ligand engineering strategy is the immunotoxin TGF-PE40 (Merck), which fuses the EGFR ligand TGF-α to Pseudomonas exotoxin A as a cytotoxic agent (Figure 3C) [63]. This construct selectively targets cells expressing EGFR and directly induces cell death, rather than inhibiting EGFR signaling. Consequently, TGF-PE40 is less susceptible to tumor evasion via mutation or upregulation of alternative signaling pathways. Preclinical studies demonstrated the high cytotoxicity of this TGF-α-based immunotoxin in cellular models, and upon systemic delivery in multiple mouse tumor xenograft models, the molecule effectively inhibited tumor growth while eliciting only mild toxicities [72–74]. Motivated by this promising preclinical work, a phase I clinical trial was initiated with TGF-PE40, which demonstrated that the immunotoxin was well tolerated without noticeable toxicities in patients with refractory superficial bladder cancer. However, the molecule failed to induce sustained therapeutic responses in patients, and further clinical development was thus abandoned [72,75].

Other Growth Factor Engineering Strategies

In addition to VEGF and EGF, many other growth factors harbor promise for cancer treatment, as well as in other therapeutic areas. NGF, first cited in the early 1950s as a signaling molecule responsible for neurite outgrowth in the peripheral nervous system, signals through 2 different receptors: tropomyosin receptor kinase A (TrkA) homodimers and p75 neurotrophin receptor (p75NTR) homodimers [22,76]. Signaling through TrkA promotes survival and differentiation of peripheral nerves, whereas signaling through p75NTR leads to increased cell death through cell cycle arrest and suppression of mitosis [76,77]. These contrasting effects dependent on receptor engagement make NGF an attractive engineering target, both to expand nervous system-associated cell populations and to reduce excessive cell proliferation [77].

Another growth factor family with therapeutic promise, the FGF family, has been shown to induce pro-mitogenic signaling in a wide variety of cell types, most notably fibroblasts and endothelial cells [78,79]. The 22 FGF ligands expressed in humans signal through 4 RTKs denoted FGF receptors 1-4 (FGFR1-4) and can be divided into 2 subsets: (1) secreted FGFs that bind to and induce signaling through FGFRs, which are composed of 3 Ig-like extracellular binding domains and two intracellular tyrosine kinase domains; and (2) intracellular FGFs that most commonly act as cofactors for voltage-gated sodium ion channels, which are necessary for regulating activity of sensory neurons [80,81]. Additionally, secreted FGFs can be partitioned into the endocrine and paracrine categories. Endocrine FGFs generally regulate cellular metabolism of macromolecules as well as bile acid and phosphate, whereas paracrine FGF signaling modulates processes involved in cell fate, including differentiation, proliferation, and mitosis [5,82]. In humans, FGF-encoding genes have been arranged into 7 subfamilies based on their degree of phylogenetic relatedness and similarity of their downstream effects (Fgf1, Fgf4, Fgf7, Fgf8, Fgf9, Fgf11, and Fgf19) [80]. Thus, the FGF system possesses remarkable tunability that can be leveraged through engineering approaches.

Lastly, the growth factor IGF, isolated in 1976 from human serum, was named due to its similar structure to the human hormone precursor, proinsulin. There exist two IGF isoforms, IGF-I and IGF-II, which differ in the receptors with which they associate [83]. Both IGF-I and IGF-II signal through either the homodimeric IGF-I receptor (IGF-IR), the homodimeric insulin receptors A and B, or heterodimers between IGF-IR and either IR-A or IR-B. IGF-II can also signal through the monomeric IGF-IIR (also called mannose 6 phosphate receptor) [83]. IGF-IR signaling promotes survival by upregulating genes responsible for proliferation and differentiation, whereas IGF-IIR activation results in IGF clearance and cell migration, opposing IGF-I signaling effects by decreasing concentration of IGF-I in the blood [83,84]. As with the PDGF system, the intricate nature of the IGF network provides an inviting framework for engineering. Here, we discuss recent design approaches targeting the NGF, PDGF, and IGF growth factor pathways for applications in cancer as well as other conditions.

The nerve growth factor (NGF) system

The strong correlation between NGF signaling and stem cell polarization towards neurons and glia invites opportunities to harness NGF activity to address brain and peripheral nerve damage [85]. A recent study attempted to leverage NGF to repair damaged nerve cells following traumatic brain injury [86]. A modified version of NGF consisting of 3,4-dihydroxyphenylalanine (DOPA) linked to the N-terminus of the growth factor was formulated to improve the adhesion of NGF to poly lactic-co-glycolic acid (PLGA) spherical microcarriers [86]. The resulting scaffold promoted the adhesion and proliferation of human umbilical cord mesenchymal stem cells (HUC-MSCs) through TrkA signaling, leading to improvements in the motor and learning abilities of mice following traumatic brain injury (Figure 4A) [86]. These results demonstrate that modification of growth factors to enable adhesion to polymeric scaffolds allows for specific and localized activity to manipulate cellular outcomes.

Figure 4. Engineering strategies for additional growth factor systems.

(A) Fusion of 3,4-dihydroxyphenylalanine (DOPA) to nerve growth factor (NGF) improves adhesion onto poly-lactic-co-glycolic acid (PLGA) microcarriers. When exposed to human umbilical cord mesenchymal stem cells (HUC-MSCs), DOPA-fused NGF induces TrkA signaling, leading to improved neurite outgrowth and motor function recovery following traumatic brain injury. (B) A decoy receptor termed LEVI-04, consisting of the p75NTR receptor fused to the Fc domain of human IgG1 to form a homodimer, was developed to neutralize excess NGF produced by breast cancer cells. (C) Cyclic oligomers synthesized using helical repeat proteins as building blocks were fused to the computationally designed minibinder mb7, which binds to the c-isoform of FGFR1 (FGFR1c). The resulting minibinder-oligomer fusion proteins induced clustering of FGFR1c on the surface of induced pluripotent stem cells (iPSCs), leading to phosphorylation of the tyrosine kinase domains on adjacent receptors and consequent induction of endothelial cell differentiation. In the presence of soluble mb7, FGF2 is blocked from associating with the ligand binding domain of FGFR1c. Thus, iPSCs treated with co-delivered mb7 and FGF2 are activated through the b-isoform of FGFR1 (FGFR1b) and primarily differentiate into perivascular cells. (D) The soluble insulin-like growth factor I (IGF-I) decoy receptor IGF-Trap is comprised of the extracellular domain of IGF receptor I (IGF-IR) fused to the Fc domain of human IgG1 to form a homodimer. By neutralizing soluble IGF-I and IGF-II, IGF-Trap reduces signaling through IGF-IR on the surface of tumor cells, leading to reduced tumor growth and metastasis. Figure was created with Biorender.com.

With respect to cancer applications, NGF is currently being studied in the context of neuropathic cancer pain. Sufferers of this phenomena experience burning, tingling, and electrical sensations in their peripheral appendages due to excessive production of NGF by tumor cells, which induces upregulation of pain-sensing nociceptors on their surface [87]. Neuropathic cancer pain is particularly common in breast cancer, wherein NGF secretion is frequently upregulated [88]. The chimeric fusion protein LEVI-04, consisting of the NGF receptor p75NTR fused N-terminally to the Fc domain of human IgG1 for dimeric and sustained presentation, was designed as a decoy receptor with the goal of sequestering NGF to block its pathogenic activities (Figure 4B) [89]. This injectable protein drug is currently in Phase II clinical trials and shows promise in treating chronic pain, including neuropathic cancer pain [89].

The fibroblast growth factor (FGF) system

Aberrant FGF2 expression by tumor cells can lead to excessive expansion of endothelial cells through activation of the receptors FGFR1, FGFR2, FGFR3, and FGFR4, increasing blood flow to the tumor microenvironment [90]. For this reason, developing treatment strategies which block this important angiogenic pathway is an attractive approach to deprive malignant cells of the nutrients they require in order to inhibit tumor growth. A recent study designed cyclic oligomers with 4, 5, 6, 7, and 8 subunits constructed using de novo computational design by creating various permutations of a pool of helical repeat proteins [91]. Selected cyclic oligomer designs were expressed in Escherichia coli, representing a breadth of valencies and orientations [91]. The investigators sought to use these engineered oligomers as a scaffold to manipulate dimerization and signaling in the FGFR system. To this end, they fused 6 designed oligomers (denoted C2-58, C4-71, C6-71, C6-79, and C8-71) to mb7, a de novo minibinder protein that targets the c-isoform of FGFR and inhibits FGF signaling, at either the N- or C-terminus of each subunit (Figure 4C) [91,92]. Interestingly, of the 10 minibinder-oligomer fusion proteins that were tested, 5 induced signaling of Chinese hamster ovary (CHO) cells expressing the c-isoform of FGFR1, demonstrating that these engineered fusion proteins could induce FGF pathway signaling through higher-order clustering and consequent intracellular phosphorylation of FGFR1 [91]. The investigators further showed the ability to bias FGF ligand activity towards cells expressing the b-isoform of FGFR1 by co-delivering FGF2 with the untethered c-isoform-inhibiting mb7 protein [91]. Functionally, the authors showed that their modulation of FGFR signaling could alter cell fate by treating induced pluripotent stem cells (iPSCs) with either the native FGF2 ligand or minibinder-oligomer fusion proteins. Whereas iPSCs treated with either FGF2 or minibinder-oligomer fusion proteins separately (to induce activation of the c-isoform of FGFR1) resulted in differentiation towards endothelial cells, iPSCs treated with co-delivered FGF2 and mb7 (to induce activation of the b-isoform of FGFR1) biased differentiation towards perivascular cells [91]. These novel findings, combined with observations of c-isoform FGFR1 upregulation in solid carcinomas [93], implicate mb7 as a potential candidate for blocking FGF-mediated angiogenesis in the tumor microenvironment by skewing stem cell differentiation away from endothelial cell lineages. Overall, these results illustrate the capacity to elucidate and manipulate the FGF pathway through computational design of both agonist and antagonist ligands.

The insulin-like growth factor (IGF) system

Due to its importance in insulin processing, IGF has been targeted for treatment of diabetes-related health conditions. IGF has also been identified as having an important role in cancer progression, upregulating cell survival genes through IGF-IR signaling and contributing to metabolic dysregulation in tumors [23]. Current monoclonal antibody drugs targeting IGF and its receptors are largely ineffective at preventing tumor cell expansion, necessitating development of alternative treatments to block the IGF system [23]. To this end, a decoy receptor (denoted IGF-Trap) comprised of the extracellular region of IGF-IR (which binds to both IGF-I and IGF-II) fused N-terminally to the human IgG1 Fc domain was developed [23]. This fusion protein was shown to sequester extracellular IGF and prevent pathogenic signaling (Figure 4D) [23]. It was further found that IGF-Trap is safe and potent in murine models of colon cancer and lung carcinoma, decreasing both primary tumor growth and metastasis [23]. Collectively, studies with the NGF, PDGF, and IGF systems demonstrate how a variety of growth factors offer promise as treatments for applications including tumor suppression. As the nuances of growth factor signaling mechanisms become better understood, technologies harnessing and manipulating their functional activities can be more efficiently and intentionally developed.

Concluding Remarks and Future Perspectives

Growth factors including VEGF, EGF, NGF, FGF, IGF, and others play crucial roles in coordinating diverse cellular responses, and the dysregulation and/or mutation of these growth factors correlate strongly with tumor development and progression. These properties render growth factors promising targets for therapeutic intervention in oncology. Despite successful advancements in the design of small molecule and antibody drugs targeting growth factor pathways, emerging strategies are leveraging ligand/receptor binding interactions as a starting point for therapeutic engineering. Growth factor ligands and receptors have been engineered into various formulations with bespoke properties, including decoy receptors, stabilized and/or biased ligands, immunotoxins, vaccines, and PLGA microcarriers. There are also emerging strategies looking to combine molecular engineering approaches with innovative cell engineering technologies, for example, chimeric antigen receptor (CAR) T cell therapy. One interesting recent approach incorporates a CAR into T cells that contains an antibody fragment which selectively engages the activated state of EGFR, which is aberrantly prevalent on tumor cells [94]. This strategy has potential to significantly decrease the on-target/off-tumor toxicities associated with targeting the ubiquitously expressed EGFR antigen. Taken together, these innovative strategies open up new avenues in protein engineering for treating cancer, as well as other conditions.

Engineering ligands or receptors offers several advantages over traditional therapeutic approaches. Firstly, ligands and solubilized receptors are smaller in size relative to antibodies, which enables better drug penetration into tumors and can lead to improved outcomes in treating solid tumors [63]. To enhance the pharmacokinetics of ligand and growth factor therapies, several strategies have been employed, such as increasing circulation persistence through Fc fusion [95] or serum albumin conjugation [96] as well as modifying proteins to increase negative charge to repel adhesion to the basement membrane of the kidney and avoid renal clearance [96]. Another potential advantage for ligand/receptor engineering technologies is that human proteins generally have low immunogenicity when delivered to patients and possess favorable physicochemical properties, thus minimizing the risk of developing undesired immune responses [97]. Finally, the natural binding affinity of ligand/receptor interactions may exceed that of antibodies, and engineering techniques such as directed evolution via phage display or yeast display can be used to further strengthen these binding interactions. These same platforms can be used to improve the target specificity of natural growth factor molecules, overcoming their pleiotropic nature, and resulting in more selective proteins that evoke favorable therapeutic outcomes.

Despite these advantages, ligand/receptor engineering approaches face several challenges (see Outstanding Questions). First and foremost, the mitogenic nature of some growth factor ligands triggers proliferation of various cell types, which can be highly undesirable for cancer therapy and can conversely promote disease pathogenesis [98]. Thus, it is necessary to select growth factors that lack these undesirable effects or to carefully engineer growth factors to circumvent their mitogenic properties to advance the development of cancer therapeutics [99]. Secondly, ligand-based antagonists, receptor decoy molecules, and other engineered growth factor reagents may lack efficacy when targeting aggressive and/or advanced tumors due to pathway redundancy and inherent mechanisms of resistance. Additionally, acquired resistance can develop when using these targeted therapeutic agents, for example through transcriptional changes and/or pathway dysregulation. Hence, developing multispecific proteins capable of targeting two or more pathways presents a more attractive approach for tumor elimination. Furthermore, combining therapeutic strategies, such as chemotherapy or immunotherapy, with ligand/receptor engineering can yield synergistic effects by recruiting complementary mechanisms for cancer cell targeting and killing [100]. Another potential challenge for growth factor engineering approaches in cancer therapy is the heterogeneous nature of tumor cells, which can hinder the effectiveness and limit the applicability of targeted treatments. Indeed, these “one size fits all” approaches are not always optimal for oncology applications [101]. Personalized medicine approaches are emerging as potential alternatives, wherein matching specific predictive biomarkers in patients with appropriate therapies can lead to higher efficacy and improved clinical results [100,101]. Overall, synergistic advances in growth factor engineering as well as therapeutic design promise to promote the development of more effective cancer treatments.

Outstanding Questions Box:

• What strategies can be used to overcome the limitations of natural growth factor ligands and receptors as therapeutics, such as their pleiotropic nature and poor pharmacological properties?

• How can protein engineering platforms be employed to further enhance the performance of engineered growth factor ligands and receptors as cancer drugs?

• What approaches can help mitigate the risk of acquired resistance to growth factor-based targeted therapies?

• Are there ways to counteract the complications of pathway redundancy and compensatory adaptations in growth factor biology for disease treatment?

• How can we strategically combine therapeutic strategies to create effective treatment regimens that incorporate growth factor-based therapies?

• Can we identify convenient methodologies to identify specific predictive biomarkers in patients to address disease heterogeneity and allow for customized treatment of tumors?

Highlights:

• Growth factors play pivotal roles in controlling cellular activities in the context of health and disease, serving as a crucial therapeutic target in cancer

• Engineering natural growth factor ligands or receptors represents a promising and emerging approach within the realm of molecular design

• Several growth factor pathways have been engineered for therapeutic development, including the vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), nerve growth factor (NGF), platelet-derived growth factor (PDGF), and insulin-like growth factor (IGF) systems

• Strategies for engineering growth factor receptors include the design of decoy receptors to sequester growth factor ligands and vaccines incorporating mutated growth factor receptor fragments

• Strategies for engineering growth factor ligands include molecular fusions, dual-specific formulations, and affinity modulation