Pharmacology of Free Fatty Acid Receptors and Their Allosteric Modulators

1.1. Allosterism at G Protein-Coupled Receptors G protein-coupled receptors (GPCRs) are the largest group of membrane proteins and play key roles in various physiological and pathological processes. The main function of GPCRs is to constantly monitor the cellular environment and signal to adapt the cellular responses to biological events over short and long distances throughout the body. GPCRs are targeted by endogenous and exogenous molecules via binding to specific sites at the receptor protein, stabilizing and/or inducing conformational changes that dictate the signaling properties of the receptor-adaptor molecule interaction and thus downstream signaling. The exact mechanisms of receptor activation and transduction of information are not fully understood and remain a matter of intense research. Known modulators that can substantially change the behavior and signaling of GPCRs in vivo are, for example, (i) (sub)cellular localization of the receptors, including spatiotemporal regulation of GPCR–effector interaction, (ii) homo- or heteromeric multimerization of receptors, (iii) graded ligand efficacy, e.g., full versus partial agonism, (iv) multidimensional efficacy, i.e., signaling bias, and (v) allosteric modulation. As the latter phenomenon is the focus of this review, allosterism and its most common shapes shall be introduced. Allosteric modulation is a biological phenomenon not only restricted to GPCRs, but occurs in practically all biologically active molecules. The concept of allosterism is considered and widely used in drug discovery and development, with several molecules that act allosterically being clinically applied [ 1 , 2 ]. The essential mechanism is the transduction of information, e.g., in the form of molecular conformation or ensembles of conformations, from one location to another. If an endogenous ligand binds to one binding site, i.e., the orthosteric site, allosteric ligands can bind to a site spatially distinct to this orthosteric site—i.e., an allosteric site. It becomes clear that GPCRs harbor several potential binding sites for ligands, so that although by definition only one orthosteric site exists, several allosteric sites can theoretically be identified. The transduction of information by the interaction between orthosteric and allosteric ligands can come in various and at times complex shapes. Allosteric ligands can either increase, leave unchanged, or decrease the affinity and/or the efficacy of another ligand—e.g., orthosteric ligand. Additionally, a ligand that binds to a site distinct from the orthosteric site can exert effects independently of a second ligand. If the allosteric ligand creates intrinsic efficacy, one would consider it an allosteric agonist. In the case of a neutral or decreasing effect on efficacy, the ligand is denoted as a neutral allosteric antagonist or inverse allosteric agonist, respectively. In many cases, the nature of the interaction between allosteric and orthosteric ligands gives the ligand its name. An allosteric ligand that increases the affinity and/or efficacy of an orthosteric ligand is a positive allosteric modulator (PAM), while an allosteric ligand that decreases the affinity and/or efficacy of an orthosteric ligand is a negative allosteric modulator (NAM). The magnitude and direction of an interaction are continuous and independent characteristics of allosteric ligands. Thus, ligands also exist that do not modulate one or more parameters of an orthosteric ligand. Such a ligand can be called a silent allosteric modulator (SAM) or neutral allosteric ligand (NAL). Additionally, combinations are possible, e.g., PAM-antagonists, which increase the affinity of the orthosteric ligand but reduce the efficacy of the orthosteric ligand [ 3 ]. Ago-PAMs are agonists that are also allosteric modulators. A characteristic feature of allosteric interaction is reciprocity, which is the effect of the allosteric ligand to the orthosteric ligand, which also applies the other way around—i.e., from the orthosteric ligand to the allosteric ligand. Another key feature of allosteric communication is probe dependency—i.e., that a given allosteric ligand might increase the affinity of a certain orthosteric ligand but could show a completely different effect when cobound with another orthosteric ligand [ 4 ]. Terms such as PAM or NAM thus describe a relationship rather than a single ligand, although they are most often used to describe certain molecules. Hence, a general classification can only fall short of detailed characteristics and cannot describe the ligand features holistically. The interdependent relationship between the ligands can also be described mathematically—for instance, based on the ternary complex model where the receptor is subjected to binding of both an orthosteric ligand and an allosteric ligand. The modulation of the affinities of the orthosteric and allosteric ligands to the receptor is described by the α factor. The modulation of efficacy

2.1. Introduction In accordance with the deorphanization of FFAR1 as a long-chain fatty acid sensing receptor, focus was put on the pancreatic expression of FFAR1, especially on β-cells, and its involvement in the secretion mechanisms of insulin [ 16 , 18 ]. Over the years, the expression pattern has been uncovered to be much broader, ranging from brain cells to bone and monocytes, and intestinal cells to rodent taste buds. Due to their shared ligand space, interpreting studies using fatty acids as physiological stimuli to deduce specific FFA receptor functions is not straight forward. Discovery of selective and potent (synthetic) agonists or antagonists as well as genetically engineered systems have helped to gain greater insight into the function of each of the receptors in particular tissues. For this reason, only studies using either genetically modified systems or chemically selective compounds targeting FFAR1 have been considered in the following.

In addition to the pancreas, intestinal FFAR1 expression has been another focus of investigations around the biology of this receptor. FFA1 receptors can be found on several enteroendocrine cells, including L, K, and I cells, and are associated with glucagon-like peptide 1 (GLP-1), gastric-inhibitory peptide (GIP), and cholecystokinin (CCK) secretion, respectively [ 48 , 49 ]. Edfalk et al. showed that dietary fatty acids mediated GLP-1 and GIP secretion via the FFA1 receptor [ 48 ] and Liou et al. demonstrated that olive oil ingestions in FFAR1 expressing mice, but not in FFAR1-deficient mice, led to plasma CCK increase [ 49 ]. Interestingly, experiments using isolated perfused intestines from rats suggest that vascular but not luminal administration of endogenous and synthetic FFAR1 agonists can induce GLP-1 secretion [ 50 ]. Whether this is due to differences in apical versus basal receptor localization and associated signaling needs further elucidation. Another study followed up on this and found that chylomicrons were able to induce GLP-1 and GIP secretion from mouse and human intestinal tissues and a human peptide-secreting cell line model, suggesting that the effects of dietary fatty acids on incretin secretion found in vivo/in vitro and ex vivo could rely on their systemic or basolateral activities after absorption by the intestine and repackaging as triglycerides into chylomicrons and basolateral release [ 51 ]. The involvement of FFAR1 in this process is not entirely clear as the FFA1 receptor antagonist GW1100 could block cell line-derived incretin secretion but not murine tissue-derived incretin [ 51 ]. Additionally, the signal transduction composition downstream of FFAR1 appears to determine whether sole GSIS activity via β-cells or combined pancreatic GSIS and incretin secretion via enteroendocrine cells occur. Hauge et al. showed that agonists that induce both Gs and Gq signaling evoked GLP-1 and GIP release more effectively than agonists that induce Gq signals only, such as endogenous orthosteric ligands [ 52 ]. However, due to tissue- and cell type-dependent differences in receptor reserves as well as G protein abundancy and preference in different cell types, it is not clear whether it is actually the induction of different receptor conformations that leads to distinct signaling outcomes (i.e., signaling bias), or the graded strength of receptor activation—i.e., full versus partial agonism.

Other reports suggest an additional role of FFAR1 in bone, as it is expressed on several bone cells, including osteoclasts, osteoblasts, and osteocytes [ 56 ]. While RNA-mediated receptor knock-down experiments suggested that the FFA1 receptor mediates osteocyte apoptotic signaling of thiazolidinedione application, another study supports the notion that FFAR1 activation is protective against bone loss using FFAR1 knock-out mice together with the FFAR1/FFAR4 dual agonist GW9508 [ 57 ].

3.1. Introduction and Molecular Receptor Pharmacology The FFA2 and FFA3 receptors are activated by short-chain fatty acids (SCFAs), which are the endogenous agonists mainly produced by the gut microbiota via fermentation of indigestible carbohydrates in humans—i.e., dietary fiber. Hence, FFAR2 and FFAR3 overlap in their recognition of orthosteric ligand as both sense SCFAs up to a chain length of six carbon atoms. However, there is a slightly different order of potency for the endogenous agonists. For FFAR2, this is: C2~C3~C4 > C5 > C6~C1 for the human receptor, while C2 has higher affinity compared to C3 for the mouse FFA2 receptor and for FFAR3: C3~C4~C5 > C6 > C2 > C1 [ 85 ]. In addition to the naturally occurring SCFAs, synthetic ligands have also been generated mainly with the aim of overcoming the similar ligand recognition pattern to achieve receptor selectivity. Chemical modification led to a group of small carboxylic acids that showed preferential activation of FFAR2 over FFAR3 and thus increased the understanding of orthosteric ligand–receptor interactions of the SCFA sensing FFA receptors [ 86 ], which was followed by more selective synthetic orthosteric FFAR2 agonists [ 87 ]. A high-throughput screening with subsequent compound optimization identified selective phenylacetamide scaffold-based FFAR2 ligands that were shown to act as allosteric agonists, which both activate the receptor on their own but also enhance the signaling capacities of orthosteric SCFAs at the FFA2 but not the FFA3 receptor [ 88 , 89 ]. Although the exact binding site of the identified prototypical allosteric ligand 4-CMTB could not be unequivocally identified by extensive mutational analysis [ 90 ], Smith et al. provided some structural insights into the allosteric mechanism by showing that the extracellular loop 2 (ECL2) of the FFA2 receptor is crucial for conveying allosteric communication from the SCFA binding site to the allosteric binding site of 4-CMTB [ 91 ]. Later, Grundmann et al. found that FFA2 receptor activation by 4-CMTB actually functions in a unique manner. 4-CMTB could trigger receptor activation by both the orthosteric as well as the allosteric sites in a temporally staggered fashion, thus characterizing 4-CMTB as the first sequentially activating ligand (SEAL) [ 92 ]. It was shown that this process involves interactions with two distinct receptor sites to allow for temporal control of GPCR signaling, a manifestation of what has been denoted as temporal bias [ 93 ]. FFAR2 selective antagonists that all showed to be inverse agonists by reducing constitutive activities of FFA2 receptors were described in a Euroscreen patent, with CATPB being the best characterized antagonist [ 94 ], and later being described as structurally distinct in a high-throughput screening (HTS) campaign including 20,000 compounds (Park et al. [ 95 ]). The only FFAR2 antagonist that has reached the clinical development stage is GLPG0974, which was discovered and developed by Galapagos to treat intestinal bowel disease patients via a neutrophil activation reducing mechanism [ 96 , 97 ]. However, further clinical development was discontinued after a lack of efficacy in a phase II clinical study. While the FFA3 receptor has been reported to only signal via the Gi protein family [ 20 ], FFAR2 was found to activate G proteins from both the Gi and the Gq families [ 20 ], but induce signaling in a Gαi/o/q/11- and β-arrestin-independent fashion, which was shown to be mediated via Gα12/13 proteins [ 31 ].

FFAR2 and FFAR3 have been found to reside on pancreatic β-cells and regulate insulin secretion. While it is known that Gq protein signaling promotes and Gi protein signaling decreases GSIS of β-cells, it is not surprising that the Gq/i protein-coupled FFA2 receptor both positively and negatively regulates GSIS and that the Gi protein-coupled FFA3 receptor is generally linked to negative regulation of insulin secretion [ 103 , 104 , 105 , 106 , 107 , 108 ]. This divergence in G protein signaling downstream of FFAR2 has led to hypotheses on potential biased ligands that evoke distinct pharmacological profiles. In fact, small carboxylic acids recognized to activate FFAR2 orthosterically promote GSIS via a PLC inhibitor-sensitive Gαq/11-mediated pathway, whereas allosteric activation with 4-CMTB [ 88 ] inhibited GSIS in a PTX-sensitive and Gαi/o-dependent fashion [ 107 ]. Another allosteric activator of FFAR2 (compound 58 [ 89 ]), although active on both Gq and Gi protein pathways, showed a Gαq/11-dependent potentiation of GSIS in mouse and human islets [ 105 ]. These results indicate that balancing inhibitory and stimulatory signaling pathways downstream of FFAR2 or even fine-tuning of G protein interaction with the cellular signaling machinery upstream of the second messengers, IP 3 /Ca 2+ and cAMP, might prove useful in shaping insulinotropic function. Interestingly, allosterically activated FFA2 receptors did not desensitize after prolonged stimulation and insulin secretion was persevered even after sustained allosteric stimulation [ 105 ]. Whether biased signaling downstream of FFAR2 can be exploited for therapeutic applications by a targeted approach, e.g., rationally designed biased ligands, will have to be assessed in future studies.

Inflammation is a shared (patho)physiological mechanism in the amalgamative disease fields of metabolic dysregulation, infectious diseases, and (auto-)immune diseases, such as intestinal disorders. FFA2 receptor activation on dendritic cells by SCFAs derived from the gut microbiota triggers antibody production by B cells by increasing BAFF and ADLH1a2 expressions, which leads to elevated retinoic acid levels [ 118 , 119 ]. These mechanisms that underlie the beneficial effects on intestinal homeostasis relied on FFAR2 expression as knock-out animals did not response in a similar fashion and were reported to be independent of T cells [ 119 ]. Another study placed focus on T regulatory (T reg ) cells in the microbial SCFA-mediated immunomodulatory effects of FFAR2. SCFAs regulated the abundance and activity of this cellular population to finally protect mice against colitis induction in the T cell-transfer model [ 120 ]. Furthermore, T reg cells have be found to be immunoregulatory during kidney allograft transplantation. Wu and colleagues could show that mice receiving a kidney allograft, but not an isograft, exhibited gut microbial dysbiosis and allograft rejection when fed a normal chow diet, which was prevented by feeding a high-fiber diet or a normal chow diet supplemented with acetate. These beneficial effects were mediated by T reg cells in an FFAR2-dependent manner, as FFAR2-deficient mice were not protected [ 121 ]. In addition to hematopoietic cells, renal resident cell types such as tubular epithelial cells and podocytes have also been found to express FFAR2 and react with reduced proinflammatory and profibrotic gene expressions upon hyperglycemic insult when FFA2 receptors were stimulated. Antifibrotic, anti-inflammatory, and kidney function preserving effects of SCFAs such as acetate and butyrate or high-fiber diet via changes in the gut microbiome have been demonstrated to be mediated by FFA2 and GPR109A receptors for mice suffering from streptozotocin (STZ)-induced diabetic nephropathy [ 122 ]. Renal mesangial cells have also been shown to express FFAR2 and stimulation with orthosteric and allosteric ligands for FFAR2 led to reduced oxidative stress and proinflammatory gene transcription by downregulation of NFκB signaling in an FFA2 receptor -dependent manner. This process was suggested to underlie the protective effects of SCFA-mediated amelioration of renal dysfunction in STZ-induced diabetic nephropathy in mice [ 123 ]. Additionally, the FFAR2 specific allosteric agonist CFMB had decreasing effects on TNF-α-induced MCP-1 upregulation in cortical epithelial cells, indicating an anti-inflammatory effect of FFA2 and FFA3 receptor signaling in renal diseases [ 124 ]. Protection from westernized diet-induced cardiovascular disease by SCFA supplementation or resistant starch diet to wild-type mice but not mice deficient of the SCFA receptors FFAR2, FFAR3, and GPR109A was recently reported. Kaye et al. assessed cardiovascular endpoints such as cardiac function, hypertrophy and fibrosis, renal sodium and potassium excretion, and blood pressure and provided evidence for a prominent role of FFAR2 and GPR109A in mediating a T reg cell driven immune-modulatory effect to protect from detrimental effects of low-fiber diets on cardiovascular health [ 125 ]. Fachi et al. documented the involvement of ILC3s in the balancing act of FFA2 receptor signaling in pro- and anti-inflammatory reaction by showing how the short-chain fatty acid acetate mediates IL-22 secretion by ILC3s in an FFA2 receptor-dependent and IL-1 receptor involving way. FFA2 receptor-enhanced neutrophil sensitization and secretion of IL-1β together with the immune-modulatory function via ILC3s were shown to underlie the protective value of acetate by defending a Clostridium difficile infection [ 126 ]. A central role of the SCFA sensing receptor FFAR2 in the tight link between the gut microbiome and host health state has thus been repeatedly confirmed. Protection of host integrity from pathogen infections has also been shown for the airway system. Klebsiella pneumoniae lung infection was hindered by increased phagocytotic activity of neutrophil and FFAR2 activity of macrophages [ 127 ]. Sencio et al. investigated the role of gut microbiota in the controls of Streptococcus pneumoniae superinfection postinfluenza and found FFAR2 function on alveolar macrophages to be critical in reducing susceptibility of infected mice to bacterial respiratory superinfection [ 128 ]. Internalization of stimulated FFA2 receptors affects cellular entry and following replication of the influenza A virus. siRNA knock-down of FFAR2 or of internalization machinery components such as β-arrestin 1 or the β2-subunit of the AP-2 complex reduced receptor internalization and thus virus endocytosis [ 129 ]. While regulation of FFA2 receptor surface abundance is an elegant mechanism to confer cellular sensitivity, it will require further investigation to elucidate whether t

FFAR4, formerly called GPR120, is next to FFAR1, the second receptor activated by long-chain fatty acids. α-Linolenic acid (18:3(n-3)) (aLa) and linoleic acid (18:2(n-6)) activate FFAR4 with EC 50 values of 0.4 to 2 and 5 µM, respectively. However, there are further free fatty acids similar to stearidonic acid, eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) known to activate FFAR4 with an EC 50 below 5 µM [ 148 ]. Together with aLa, the latter are members of the ω-3 family of fatty acids that have been associated with anti-inflammatory effects [ 149 , 150 ]. Agonism at FFAR4 triggers two different signaling pathways. The receptor is known to couple, via Gαq/11 proteins, to the production of inositol trisphosphate (IP 3 ) and Ca 2+ release [ 151 , 152 ]. This particular signaling of the receptor has been linked, for instance, to its function in GLP-1 release from enteroendocrine cells in the colon [ 23 ]; the extent of the contribution by FFAR4 to GLP-1 release has been questioned [ 63 , 153 ]. FFAR4 also couples to Gαi/o and the inhibition of adenylate cyclase activity was shown by Engelstoft et al. [ 28 ] in the example of ghrelin secretion inhibition for primary gastric mucosal cells. Another mechanism employed by FFAR4 is based on binding to β-arrestin 2. The FFAR4/β-arrestin 2 complex binds transforming growth factor-β-activated kinase 1 (TAK1) binding protein 1 (TAB1) and thereby sequesters it from TAK1. This subsequently leads to a reduction in the TAB1/TAK1 complex and its proinflammatory activity at the conversion point of the toll-like receptor (TLR) and tumor necrosis factor (TNF) α receptor signaling pathways [ 154 ] in macrophages. Human FFAR4 is expressed as a short splice variant of 361 aa and a long isoform carrying an additional 16 aa in the third intracellular loop (ICL3). For the latter, it has been shown that it is, compared to the short isoform, unable to mediate agonist-stimulated Ca 2+ release, whereas β-arrestin binding and downstream internalization as well as trafficking are comparable for the two splice variants. Because the long isoform is not found in other species such as rodents and even cynomolgus monkeys [ 151 , 155 ], the physiological consequence of this species difference is so far unknown. FFAR4 expression is detected in the colon, in enteroendocrine cells, adipocytes, and macrophages. In addition, a high expression level was also found in the lung [ 23 ]. Here, expression is localized to the airway epithelium and experiments with FFAR4 knock-out mice suggest that the receptor promotes epithelial repair after lung injury [ 156 ]. The pancreas is a further organ where FFAR4 is expressed and it has been shown to be protective against palmitate-induced apoptosis. In adipocytes, FFAR4 stimulation enhances GLUT4 translocation and glucose uptake [ 154 ]; moreover, this receptor has been suggested to be involved in the differentiation and development of adipocytes [ 157 , 158 ]. Houthuijzen et al. [ 159 ] have shown that FFAR4 is expressed in splenic macrophages and is related to the development of chemotherapy resistance by regulating the production of the resistance inducing lipid mediator lysophosphatidylcholine (24:1). For murine RAW 264.7 cells and also for intraperitoneal macrophages, anti-inflammatory effects of FFAR4 activation point toward an interesting mechanism in the context of inflammation and insulin resistance, which has been closely examined by different groups [ 154 , 160 , 161 ]. Oh et al. [ 154 ] demonstrated an inhibition of the proinflammatory activities of TLR-2, -3 and -4 agonists as well as for TNFα for ω-3 fatty acid DHA and RAW 264.7 cells. This effect was abrogated by FFAR4 and β-arrestin 2 siRNA-mediated knock-down, whereas Gαq/11 and β-arrestin 1 reduction caused no effect. Co-immunoprecipitation experiments supported the above-described FFA4 receptor agonist-dependent association of β-arrestin 2 with TAB1 and the entailing reduced amount of proinflammatory TAB1/TAK1 complexes. The broad anti-inflammatory effects of FFAR4 agonism could also be seen for LPS-stimulated macrophages with phosphorylation of IKK, JNK, and MCP-1 as well as IL-6 secretion being dependent on the presence of the FFA4 receptor. Anti-inflammatory FFAR4 effects were also seen in 3T3-L1 adipocytes. For these cells, as well as in primary adipose tissue cultures, DHA-dependent translocation of GLUT4 to the plasma membrane and enhanced 2-deoxyglucose uptake could be demonstrated. In this case, FFAR4 and Gαq/11 knock-down abolished enhanced translocation and uptake, whereas the reduction in β-arrestin 1 and β-arrestin 2 levels by siRNA showed no effect [ 154 ]. The latter authors supported their results by confirming them with the FFAR1/FFAR4 agonistic tool compound GW9508, which is more of an FFAR1 agonist than an FFAR4 agonist [ 162 ], necessitating the reliance on negative expression results for FFAR1 in macrophages and adipocytes. Subsequently, Shimpukade et al. [ 16

With the reported involvement of FFARs in energy and metabolic homeostasis, they are interesting targets for pharmaceutical intervention. For designing potent and selective modulators of FFARs, elucidation and understanding of the structural foundation of ligand binding in FFARs is of major importance. Multiple allosteric binding sites of FFAR1 have been suggested by radioligand binding studies before crystallographic structural elucidation of the receptor–ligand complexes became available. A network of allosteric interactions between the different sites has been documented—e.g., positive co-operativity between partial and full allosteric agonists [ 62 , 199 , 200 ]. Positive co-operativity and synergistic effects have also been documented for the partial allosteric agonist TAK-875 and the endogenous orthosteric full agonist γ-linoleic acid in a Ca 2+ flux assay [ 70 ]. So far, four crystallographic structures of FFAR1 have been published. Structural elucidation of the remaining FFA receptors has not yet been carried out. In all published structures, FFAR1 adopted the fundamental structure of GPCRs consisting of a seven-transmembrane helices bundle. Srivastava et al. and Lu et al. reported two complexes employing the partial allosteric agonists TAK-875 and MK-8666, respectively [ 81 , 201 ]. Lu et al. furthermore provided the structure of a ternary complex consisting of a receptor, partial allosteric agonist MK-8666, and full allosteric agonist AP8, whereas Ho et al. published a binary complex of FFAR1 together with the full allosteric agonist compound 1 [ 81 , 84 ]. All crystallographic structures are based on stabilized human FFAR1 constructs. The most commonly applied changes to the sequence are a T4 lysozyme protein inserted into the third intracellular loop accompanied by four point mutations in the transmembrane array of helices (Leu-42 2.40 Ala, Phe-88 3.34 Ala, Gly-103 3.49 Ala, and Tyr-202 5.58 Phe) [ 201 , 202 ]. Several observations regarding the behavior of this mutant FFA1 receptor have been made. From comparison with the helix architecture of other known GPCR structures, Srivastava et al. deduced that these mutations seem to constrain the receptor in an inactive conformation [ 201 ]. Ho et al. found reduced binding of [ 3 H]-AM-1638 to the stabilized receptor when compared to the wild type FFAR1 [ 84 ] and Lu et al. documented a five-fold decrease in the dissociation constant Kd of the partial allosteric agonist MK-8666 when bound to the thermostabilized mutant receptor, while co-operativity with the full allosteric agonist AP8 was maintained [ 81 ]. Ternary complex crystals of the FFA1 receptor with the two allosteric ligands MK-8666 and AP8 were not obtained, with mutant receptors carrying the four point mutations. Instead, a similar construct without the Phe-88 3.34 Ala exchange was needed to produce stable crystals [ 81 ]. These four structures show two distinct allosteric binding sites and motives, which fits very well with reported pharmacological findings. A summarized view of the structural findings about the different allosteric sites at the FFA1 receptor is shown in Figure 2 . Partial allosteric agonists could be localized to site A, whereas full allosteric agonists were found to occupy site B. Srivastava et al. provided the first crystallographic structure of an FFAR1 ligand complex with the partial agonist TAK-875. Herein, TAK-875 was found to bind to FFAR1, between the transmembrane helices 3–5 and the extracellular loop 2 (ECL2) in a site, henceforth denoted as site A. The ligand bound closer to the exterior membrane surface than most ligands in other GPCRs. TAK-875 seems to enter a noncanonical binding pocket most likely via the lipid layer, resulting in an unusual method of binding. TAK-875 was pharmacologically characterized as an ago-allosteric ligand; thus, this unique binding site most likely represents an allosteric binding pocket in the receptor [ 201 ] ( Figure 3 ). Upon visual inspection of the receptor surface, two extra binding pockets in the vicinity of the TAK-875 site were described [ 201 ]. The first pocket is adjacent to TAK-875 and enlarges its binding pocket. It might allow a potential ligand to pass between transmembrane helices 4 and 5, especially as in the human FFA1 receptor the conserved proline 4.60 is replaced by glycine 4.58, which might add some flexibility to the end of transmembrane helix 4. A potential third binding pocket can be found in a location very similar to the binding site of the allosteric modulator LY2119620 in the M2 receptor [ 203 ]. So far, the roles of these sites cannot be determined by biological data. Distinct from the TAK-875 binding site, crystallographic data showed another site in a lipophilic region of the receptor occupied by a monoolein molecule. Lu et al. showed that this binding site (here site B) is involved in binding of full allosteric agonists. They solved the crystallographic structure of the trimeric complex of the h

Allosterism of GPCRs holds the promise to introduce new pharmacology and to unlock traditionally inaccessible modes of actions to eventually treat patients in need. There are a variety of possible mechanisms ranging from positive allosteric modulation to negative allosteric modulation, with all the nuances in between, whether by neutral, partial, inverse agonists, surmountable or insurmountable antagonists, and combinations thereof. Considering the depth of the GPCR biology toolbox, e.g., biased signaling or spatiotemporal control of (sub)cellular or tissue specific signaling, the possibilities awaiting exploration by academic and industry researchers are vast. The first allosteric GPCR modulators that have already reached the market contribute to the necessary learning curve on how this intricate new modality can be medically leveraged. The growing number of academic–industry and public–private partnerships is another manifestation of the recognition that much still needs to be explored in terms of the fundamental understanding of GPCR allosteric biology and of the prospect that transforming basic science into real drugs provides novel opportunities to address unmet medical need. The family of receptors responding to free fatty acids play roles in various processes and lend themselves as promising drug targets. The FFA1 receptor is probably the most studied and medically exploited member of the FFA receptor family. Reports of discontinuation or stagnation of clinical development programs, most prominently of the FFA1 allosteric modulator fasiglifam/TAK-875 from Takedadue to unexpected liver toxicity findings, are certainly discouraging, but seem to be target-independent and compounds have been shown to be amenable to optimization [ 72 , 73 , 74 ]. The growing number of scientific reports on novel FFAR1 pharmacology based on allosteric modulators, however, sparks renewed interest in the target. The next best studied member of the family is the FFA2 receptor. The orthosteric FFA2 antagonist GLPG0974 [ 216 , 217 ] was developed for the treatment of intestinal bowel disease and was found to be safe in phase 1 but nonefficacious in a proof-of-concept phase 2 trial [ 218 ]. The lack of efficacy of GLPG0974 needs to be seen in context with the complex pharmacology of FFAR2 in different disease settings. As elaborated earlier, results from genetically modified preclinical models have been conflicting and have triggered a discussion as to whether FFAR2 agonists or antagonists are the preferred approaches [ 139 ]. Novel findings from allosteric FFAR2 ligands reveal promising pharmacological traits and other promising results have generated new interest in this target considering novel allosteric modalities. GLPG1205 is an antagonist and negative allosteric modulator for GPR84 [ 195 ] that was discontinued for the development of IBD due to the negative ORIGIN phase 2 trial ( NCT02337608 ) but showed promising preclinical results in models of idiopathic pulmonary fibrosis [ 191 ] and also positive topline results from the proof-of-concept PINTA phase 2 study in IPF patients ( NCT03725852 ). Hence, allosteric targeting of receptors for small-, medium-, and long-chain fatty acids is a promising approach to address patient needs in different therapeutic areas mirroring the biological versatility of the receptor family members and their attractiveness as drug targets.