Technologies for the discovery of G protein-coupled receptor-targeting biologics

Abstract

G protein-coupled receptors (GPCRs) are important pharmaceutical targets, working as entry points for signaling pathways involved in metabolic, neurological and cardiovascular diseases. Although small molecules remain the major GPCR drug type, biologic therapeutics, i.e. peptides and antibodies, are increasingly found among clinical trials and FDA approved drugs. Here, we review state-of-the-art technologies for the engineering of biologics that target GPCRs as well as proof-of principle technologies that are ripe for this application. Looking ahead, inexpensive DNA synthesis will enable the routine generation of computationally pre-designed libraries for use in display assays for the rapid discovery of GPCR binders. Advances in synthetic biology are enabling the increased throughput of functional GPCR assays to the point that they can be used to directly identify biologics that modulate GPCR activity. Finally, we give an overview of adjacent technologies that are ripe for application to discover biologics that target human GPCRs.

INTRODUCTION

G-protein coupled receptors (GPCRs) play an important role in many biological functions, including metabolism, blood sugar regulation, and immunity¹. Thus, GPCRs are a key pharmacological target with > 30% of FDA approved drugs on the market². Although small molecules remain the major GPCR drug type, encompassing 88% of the FDA approved drugs in 2018³, biologic therapeutics have steadily gained ground. Biologics, such as peptides, antibodies and mRNA, provide higher affinity and specificity as well as fewer off-target effects than small molecule drugs. Indeed, six of the top ten drugs by sales in 2022 were biologics⁴. Among GPCRs, between 2017 and 2020, 27% of drug candidates in clinical trials targeted Class A GPCRs- the receptors most targeted due to their wide physiological function- were peptides⁵.

GPCRs that naturally bind peptide hormones, i.e. peptide GPCRs, are involved in muscle contraction, appetite control, and digestion⁶, and are associated with diabetes, asthma, cardiovascular, rheumatoid arthritis and cancer². Humans code for 91 peptide GPCRs, 76 rhodopsin-like Class A GPCRs and 15 secretin-like Class B GPCRs⁷ (Figure 1). As of 2020, 48 Class A GPCRs and 14 Class B GPCRs are targeted by 137 FDA approved drugs. Although the number of drugs targeting peptide GPCRs is increasing, the majority of them target the same diseases or conditions previously targeted by other drugs (Table 1). To date,19 druggable peptide GPCRs remain orphan, i.e. with no known ligand. Such is the case for adenylate cyclase-activating polypeptide type I receptor (PACR), which plays an important role in neurotransmission, learning and memory⁸, making it a key target for Alzheimer's, Parkinson’s and Huntington’s Disease. Further, there are whole peptide GPCR families that remain orphan, such as the Bombesin and Galanin receptors.

Figure 1.

Peptide GPCRs by class and drugs that target them. A. Class A peptide GPCRs. B. Class B peptide GPCRs. Information for the graphs was obtained from GPCRdb⁷.

Table 1.

FDA-approved biologic therapeutics targeting peptide GPCRs approved from 2020-2023

			Year approved
Category	Condition	GPCR	2020	2021	2022	2023
Cancer	Prostate cancer	Gonadotropin Releasing Hormone Receptor 1	Relugolix
Diagnosis	Diagnosis Agent	Somatostatin Receptor 2	64CU oxodotreotide
Endocrine	Menopause hot flashes	Tachykinin Receptor 3				Fezolinetant
	Obesity	Melanocortin 4 Receptor	Setmelanotide
	Severe Hypoglycemia	Glucagon Receptor		Desiglucagon
	Type 2 Diabetes	Glucagon-like peptide-1 receptor			Tirzepatide
		Glucose-dependent insulinotropic polypeptide receptor		Tirzepatide
Neurologic	Episodic Migraine	Calcitonin gene-related Peptide receptor		Atogepant
	Insomnia	Orexin 1 receptor			Daridorexant
		Orexin 2 receptor		Daridorexant
	Migraine	Calcitonin gene-related peptide receptor	Rimegepant
Renal	Hepatorenal syndrome	Vasopressin 1 Receptor			Terlipressin
	Primary immunoglobulin A nepheopathy	Angiotensin Receptor 1				Sparsentan
		Endothelien Receptor A				Sparsentan
	Priuritus from Chronic Kidney Disease	Kappa-Opioid Receptor		Diflikefalin

Peptide hormones produced by endocrine and neuronal tissues activate peptide GPCRs¹. Endogenous peptide hormones, such as vasopressin, a blood pressure regulator, and oxytocin, a uterine contraction stimulant, have been used as therapeutics since the 1950s⁹. The poor pharmacokinetics of peptides¹⁰, including short half-life, rapid clearance, and low gastrointestinal stability requiring injection for administration, led to the development of small molecules to target peptide GPCRs since the 1970s, including oxycodone to target opioid receptors¹¹. Starting in the mid-1980s, synthetic peptides started to be used to target GPCRs, such as Leupolide, a synthetic 9-residue peptide analogue of gonadotropin releasing hormone¹². Indeed, peptides’ low toxicity, high affinity, selectivity and potency often make them the only or front line of treatment. Peptide modification strategies, such as introduction of N-terminal acetylation or methylation, peptide cyclization and stapling, peptide display on a stabilized scaffold, and use of unnatural and β-amino acids have been successful at improving the pharmacokinetics of peptide drugs¹.

Sidestepping the challenges faced by peptides, antibodies have been engineered to target GPCRs. By 2023, the FDA had approved four GPCR-targeting antibodies: dulaglutide (diabetes), erenumab (migraine), mogamulizumab (oncology) and talquetamab (oncology). The large size of antibodies often precludes them from binding the GPCR’s orthosteric site buried within the transmembrane domains. Antibodies that target GPCRs often exert their activity by: 1) binding the large extracellular N-terminal domain present in some GPCRs¹³, such as in Class B GPCRs (erenumab) and CC chemokine receptors (mogamulizumab), 2) stabilizing a known peptide agonist¹⁴ (dulaglutide), 3) triggering antibody dependent cellular toxicity¹⁵ (mogamulizumab) or 4) bringing together receptors on different cells¹⁶ (talquetamab). Key challenges in engineering antibodies against GPCRs include, their conformational flexibility and their short extracellular loops, especially in Class A GPCRs, which limit antibody engagement.

Here, we review the literature for state-of-the-art technologies that have either been used to engineer biologics to target GPCRs or are ripe to be adapted for this application. Specifically, we focus on biologics that bind GPCRs on the extracellular side, and thus can serve as therapeutics. The now standard GPCR small molecule drug discovery pipeline that relies on virtual docking of millions of compounds on a GPCR structure followed by functional GPCR assays¹⁷ may not be easily translatable to the discovery of biologics. Biologics have greater degrees of freedom than small molecules, and there is a more limited number of structures of GPCRs bound to biologics. Therefore, in this review, we focus on approaches to engineering biologics that target GPCRs that are, at their core, experimental. We find that discovery of biologics that modulate GPCRs is a nascent field, with most of the literature still at the proof-of-principle assay development stage. However, the ease and inexpensiveness of coding for and synthesizing libraries of biologics when compared to small molecule libraries, and the large throughput enabled by screens and selections of functional GPCR-based assays, point towards biologics becoming the future of GPCR therapies.

STATE-OF-THE-ART TECHNOLOGIES FOR THE DISCOVERY OF BIOLOGICS THAT TARGET GPCRS

Animal immunization for GPCR antibody discovery.

Animal immunization followed by hybridoma or phage display screening is the most widely used methodology to identify antibodies against GPCRs¹⁸. The multipass membrane protein nature of GPCRs coupled with their conformational flexibility, their buried orthosteric site, and their short extracellular loops makes it challenging to generate effective antigens to raise anti-GPCR antibodies. Some of the antigen presenting strategies applied to GPCRs include 1) fusing a GPCR extracellular region to a soluble carrier protein, 2) using membrane fractions expressing GPCRs, 3) in vivo electroporation of DNA coding for GPCRs, and 4) GPCRs embedded in nanodisks¹⁸. Because the antibodies are only searched for GPCR binding, they often do not show modulation of GPCR activity, i.e. agonism or antagonism. Nevertheless, these highly specific binders are an excellent starting point for the generation of antibody drug conjugates, in particular for cancer applications¹⁹.

Binding assays for the discovery of biologics that target GPCRs.

Display technologies including phage, yeast and ribosome display are widely used to engineer biologics²⁰. High-throughput display technologies are used as a first step in the identification of GPCR antibody binders, which are subsequently assayed for modulation of GPCR activity in functional assays. In one of the most recent examples, in 2021, Liu et al. used a phage-displayed library to identify GLP1-R agonists for blood glucose regulation. A 10¹⁰ GPCR-focused antibody library was generated by leveraging GPCR binding motifs found among known GPCR ligand interactions, including protein/peptide ligands, peptide mimetic, and GPCR N-terminal extracellular domains/loops²¹. Phages were panned against mammalian cells expressing fluorescently-tagged GLP1-R (Figure 2A). After five rounds of panning, 100 isolated antibody sequences were expressed in mammalian cells and evaluated for GLP1-R binding. Although none of the 13 sequences that showed specific GLP-1R binding were agonists, eight of them were GLP1-R antagonists. Interestingly, two of the antagonist sequences had GPCR binding motifs not previously identified in nature. The low cost of DNA synthesis allowed the generation of 100% computationally pre-designed library, unlike DNA mutagenesis strategies for protein library generation that result in a portion of the library being superfluous. Although only a single antibody was thoroughly characterized, based on the quality of the antibody library and next generation sequencing analysis performed after each panning round, further data analysis may reveal additional novel GPCR binding motifs and shed light into GPCR antibody evolution. Identification of GPCR binding motifs not previously observed in nature could be used to uncover overlooked endogenous peptides that may interact with GLP1-R.

Figure 2.

High-throughput assays for the identification of biologics that target GPCRs. Left: binding assays. A. Phage-displayed antibody library panned against fluorescently labelled GPCR expressed in mammalian cells²¹. B. Yeast-displayed nanobody library screened against GPCR expressed in mammalian cells carrying an intracellular dye²². C. Yeast-displayed nanobody library outcompetes low affinity agonist bound to a solubilized GPCR²³. Right: functional assays. D. Mammalian cell co-expresses a nanobody library and a GPCR on the cell surface. Activation of the GPCR is linked to β-arrestin recruitment and ultimately reporter gene expression²⁷. E. Yeast cell and mammalian cell are encapsulated in a droplet. The yeast cell secretes a peptide to the medium. The peptide binds a GPCR on the mammalian cell surface, which upon activation results in reporter gene expression²⁹.

The size of yeast cell surface-displayed protein libraries are smaller than phage-displayed libraries due to limitations in the transformation efficiency of DNA into yeast. Nevertheless, yeast is amenable to fluorescent activated cell sorting (FACS), thus enabling the separation of biologics that bind GPCRs based on affinity rather than relying in washing steps, as it is the case with phage-displayed protein libraries. In 2023, Krohl et al. displayed in yeast a c-myc labelled nanobody library and used it to identify binders for GPCRs expressed on the surface of mammalian cells carrying a Celltrace intracellular dye (Figure 2B) ²². Nanobody-GPCR interactions are identified by co-incubating yeast and mammalian cells followed by labelling of the nanobodies with a fluorescent anti-myc antibody. Successful nanobody-GPCR interactions resulted in yeast-mammalian complexes that showed both Celltrace and anti-myc antibody fluorescence. Key to the success of this strategy was the use biopanning, i.e. incubation of the yeast-displayed nanobody library with an adherent monolayer of target-null mammalian cells, to remove non-specific binders prior to positive selection with magnetic activated cell sorting (MACS) and FACS. Using this technology, Krohl and coworkers identified nanobodies able to bind four GPCRs: chemokine receptors 2 and 4 (CXCR2, CXCR4), GLP1-R, and the glucagon receptor. The nanobody against CXCR2 displaced its cognate chemokine, IL-8, while the nanobodies identified for the other three GPCRs had dissociation constants in the low nM range. Although the nanobodies were not assessed for their ability to modulate GPCR activity, this is a rapid and generalizable technology to identify GPCR binders.

Yeast-displayed libraries have also been used in conjunction with solubilized GPCRs to identify GPCR antagonists. In 2020, McMahon et al. used a yeast-displayed nanobody library to identify antagonists to angiotensin II type 1 receptor (AT1R), to control hypertension²³. First, the nanobody library was incubated with solubilized AT1R to enrich for intracellular and extracellular AT1R binders using MACS. Next, enrich for antibodies that bound to the extracellular side of AT1R, the nanobody library was incubated with solubilized AT1R and known high and low affinity AT1R ligands. The high and low affinity ligand were labelled with two different fluorophores. Nanobodies that successfully displaced the low affinity ligand, but not the high affinity ligand could be identified via FACS (Figure 2C). After two rounds of MACS and one round of FACS, two AT1R antagonists were identified. Error prone PCR of the best AT1R antagonist followed by two additional rounds of FACS resulted in an antagonist able to reduce hypertension in mice. The fluorescent probe displacement strategy should be generalizable for the identification of biologics that modulate GPCRs as long as the targeted GPCR already has known low and high affinity ligands.

Functional assays for the discovery of biologics that modulate GPCR activity.

Most experimental approaches to discovering biologics that modulate GPCR activity have relied on binding assays to first identify GPCR binders followed by functional assays to determine their ability to modulate GPCR transduction²⁴. GPCR binding assays, however, often identify biologics that do not bind at the correct site (neutral binders) or block access to the orthosteric site (blockers) rather than engaging with the orthosteric site to modulate GPCR activity (agonism/antagonism) ^23,25,26. The throughput of GPCR functional assays has increased over the last couple of years, and functional assays are starting to be used as the first step in the identification of biologics that modulate GPCR activity.

In 2020, Ren et al. displayed an apelin receptor (APJ) camelid immune nanobody library on the cell surface of a Tango β-arrestin reporter mammalian cell that also expressed APJ on the cell surface (Figure 2D)²⁷. In this assay, nanobody activation of APJ results in β-arrestin recruitment and ultimate expression of the lacZ gene coding for β-galacosidase that converts a substrate (530nM) into a product (460nM), both of them retained inside the cell and trackable via FACS²⁸. To search for APJ antagonists, cells were sorted for high substrate/low product intensities. After three rounds of sorting, several antagonists were identified, with a higher percentage of antagonists among the hits compared to a solely phage displayed driven screen²⁶. To search for agonists, cells were sorted for high product/low substrate intensity. After three rounds of sorting, one agonist was identified, which was validated using a G_αs activation assay and a GPCR internalization assays. Of note, as the DNA transformation efficiency limits the library size achievable in mammalian cells (~10⁶), prior to the functional assay, the nanobody library was enriched for APJ binders using a phage displayed panning step. This panning step likely biased the nanobody library as it may have removed agonists with weaker binding affinity. The authors observed that constitutive expression of APJ agonists led to gene loss and apoptosis, thus, inducible nanobody or GPCR expression may increase the number of agonists identified. Finally, the use of β-arrestin recruitment for GPCR activation detection makes this a general technology to identify modulators of GPCR activity.

Cell display technologies restrict the conformations biologics can attain, thus limiting their evaluation as modulators of GPCR activity. To address this challenge, in 2019, Yaginuma et al. used droplet microfluidics to co-culture 1) Saccharomyces cerevisiae secreting variants of the known GLP1-R peptide agonist Exendin-4, and 2) mammalian cells expressing GLP1-R and engineered to link GLP1-R activation to lacZ expression (Figure 2E)²⁹. Despite a small library size (~1,000 members), the assay identified one peptide variant resulting in a 40% increase in GLP1-R signal after activation. Importantly, co-culturing yeast and mammalian cells did not result in significant toxicity. Although Yaginuma and coworkers did not capitalize on the ultrahigh-throughput capabilities afforded by droplet microfluidics coupled to FACS (5,000 droplets/second³⁰), in the future, this technology may enable the search of much larger peptide libraries. Mathematical modeling of droplet parameters, including droplet size and the number of yeast cells encapsulated per droplet, is needed to isolate peptides with the desired EC₅₀. Investigating the effect of yeast:mammalian cell encapsulation ratio will also improve the utility of the system. On the experimental side, swapping the LacZ reporter system with a fluorescent reported with higher quantum yield may allow for better differentiation among peptide hits.

FUTURE TECHNOLOGIES FOR THE DISCOVERY OF BIOLOGICS THAT TARGET GPCRS

Functional assays for the discovery of biologics that modulate GPCR in model organisms.

To date, only functional assays in mammalian cells have been used to identify biologics that target GPCRs. GPCR-based assays in S. cerevisiae offer an interesting alternative as yeast has a shorter doubling time (1.5 hrs vs. 24 hrs), codes for only two GPCRs rather ~100 in mammalian cells, and functionally expresses human GPCRs (hGPCRs) on the cell surface; further, GPCR activation has been successfully coupled to the yeast machinery to control reporter gene expression^31,32. Briefly, upon hGPCR activation on the yeast cell surface, G_α dissociates from G_βγ, which triggers the activation of the mating pathway, ultimately resulting in reporter gene expression. Yeast-based GPCR functional assays have been applied to the discovery of small molecule GPCR agonists (serotonin receptor 4³³, hydroxycarboxylic acid receptor 3³⁴, and the cannabinoid receptor³⁵) and antagonist (histamine receptor 2³⁶).

Application of yeast-based GPCR functional assays to the discovery of biologics is a nascent area. In 2020, Rowe et al. co-expressed somastatin receptor 5 (SSTR5) in the yeast membrane and its peptide agonist (SRIF-14) trapped in the yeast cell wall pointing towards SSTR5 (Figure 3A) ³⁷. First, Rowe and colleagues augmented a yeast GPCR functional assay by replacing the yeast G_α, GPA1, with yeast/human G_α chimeras where the last five amino acids of GPA1 were replaced with those of 10 different human G_αs. As a proof-of-principle, the GPCR/trapped peptide system was used to determine that SSTR5 preferentially interacts with the GPA1/hG_α15 chimera. In the future, this technology could be used to express a peptide library (>10⁶) and evaluate its effectiveness to modulate GPCR activity. Of note, yeast-expressed hGPCRs do not have mammalian posttranslational modifications, such as proper glycosylation or phosphorylation. Thus, it is possible that biologics that modulate hGPCR activity in yeast will not result in hGPCR modulation when expressed in mammalian cells. Therefore, secondary assays in mammalian cells, ideally in the appropriate cell line, leading to the appropriate second messenger changes, are needed to corroborate the ligands’ effectiveness. To date, small molecule ligands identified via yeast assays have translated well to GPCR activation in mammalian cells^33-36. The ease of work and speed achievable in yeast assays versus their mammalian counterparts makes yeast a compelling platform for the future identification of peptide agonists for GPCRs.

Figure 3.

Future technologies for the discovery of biologics that target GPCRs. A. Yeast expresses and traps a peptide on the cell wall pointing towards the membrane. A GPCR is expressed on the yeast membrane and upon activation by the peptide triggers fluorescent reporter expression³⁷. B. Yeast is engineered to secrete a peptide that activates a GPCR expressed on a different yeast³⁸. C. Cell-free protein expression system (CFPS) produces peptides that can be attached to a bead and evaluated for protein binding⁴⁰.

Yeast has a carbohydrate-rich cell wall, yet small peptides (~13 amino acids) can penetrate that wall and be secreted from the cell. Indeed, in the fungal world, peptides work as pheromones, binding GPCRs to initiate mating. Therefore, engineering yeast to both secrete a peptide and detect it via a GPCR functional assay is a potentially viable strategy to identify peptides that modulate hGPCR. In proof-of-principle work using fungal GPCRs and their cognate peptides, Billerbeck et al demonstrated that S. cerevisiae can secrete and detect eight orthogonal fungal GPCR-peptide pairs (Figure 3B)³⁸. Further, using exogenous synthetic peptides, the authors showed that GPCR activation can be modulated by peptide sequence. Although applied to fungal GPCR-peptide pairs, as hGPCRs are functionally expressed in S. cerevisiae, the secreted peptide-GPCR activation workflow coul be adapted, in the future, to secrete peptide libraries and evaluate their activity against hGPCRs.

Technologies for the rapid synthesis of biologics.

A technology that has not yet been applied to the discovery of biologics that target GPCRs, yet may be ripe for this application, is the use of cell-free protein synthesis (CFPS) for biologics synthesis. Briefly, in CFPS, the non-living unpurified bacteria transcription/translation cell machinery is used for gene expression. A key advantage of CFPS for the synthesis of biologics is the lack of cell membrane³⁹, which reduces the downstream processing steps necessary for biologics purification. In CFPS, biologics purification can be as simple as adding labeled magnetic beads to the reaction mixture and using magnets to separate the biologics. In 2023, Thames et al. leveraged CFPS for the synthesis of thirteen protein allergens (Figure 3C)⁴⁰. One of the purified allergens, dust mite allergen, was detected by monoclonal antibodies that recognized the allergen in a bead-based Elisa assay. Importantly, the purified allergen was used in a mammalian-based assay and showed to trigger degranulation of mast cells. Although purified from an E. coli-derived CFPS, the endotoxin levels in the purified allergen were below the FDA limit. In the future, bacterial CFPS could provide an inexpensive and automatable route for the synthesis of biologics that can be purified en masse and then evaluated in functional GPCR assays.

OUTLOOK

The conformational flexibility of GPCRs, their short extracellular loops, and the fact that the orthosteric site is buried within their transmembrane region makes it difficult to identify biologics that modulate this medically important receptor class. Reduction in the cost of DNA synthesis allows the production of computationally pre-designed libraries. Towards discovering biologics that modulate GPCR activity, display assays are being applied to debulk the biologics library, while high-throughput GPCR functional assays now allow to evaluate larger libraries of biologics for GPCR modulation. In the future, the ability of cells to both encode a biologics library and a GPCR function assay may move the field from high-throughput screening to the directed evolution of biologics that target GPCRs.

Going forward, the higher throughput, ease of work, and reduced cost of GPCR functional assays in S. cerevisiae will open the door to evaluate biologics against known medically relevant GPCRs as well as understudied and even orphan GPCRs. Because orphan GPCRs have no known ligands, binding and functional assays may be more difficult to implement due to the lack of known GPCR ligands to use as controls. The high-risk, high-reward nature of identifying agonists for orphan GPCRs may be the ideal application for functional GPCR assays in S. cerevisiae. At a minimum, functional assays in model organisms should reduce the search space to be performed in mammalian cells. In the best-case scenario, the ligands identified in the model organisms may be translatable in mammalian cells assays and mice.

Highlights.

• Biologics are increasingly found among clinical trials and FDA approved drugs targeting GPCRs.

• Biologics that modulate GPCR's activity are difficult to discover due to GPCRs' buried orthosteric site.

• GPCR functional assays allow for direct identification of biologics that modulate GPCR activity.

• Co-encoding GPCR functional assays and biologics libraries foresees the evolution of GPCR-targeting biologics.

Data Availability

No data were used for the research described in the article