Identification and characterization of an atypical Gαs-biased β2AR agonist that fails to evoke airway smooth muscle cell tachyphylaxis

G protein–coupled receptors display multifunctional signaling, offering the potential for agonist structures to promote conformational selectivity for biased outputs. For β₂-adrenergic receptors (β₂AR), unbiased agonists stabilize conformation(s) that evoke coupling to Gαs (cyclic adenosine monophosphate cAMP] production/human airway smooth muscle HASM] cell relaxation) and β-arrestin engagement, the latter acting to quench Gαs signaling, contributing to receptor desensitization/tachyphylaxis. We screened a 40-million-compound scaffold ranking library, revealing unanticipated agonists with dihydroimidazolyl-butyl-cyclic urea scaffolds. The S-stereoisomer of compound C1 shows no detectable β-arrestin engagement/signaling by four methods. However, C1-S retained Gαs signaling—a divergence of the outputs favorable for treating asthma. Functional studies with two models confirmed the biasing: β₂AR-mediated cAMP signaling underwent desensitization to the unbiased agonist albuterol but not to C1-S, and desensitization of HASM cell relaxation was observed with albuterol but not with C1-S. These HASM results indicate biologically pertinent biasing of C1-S, in the context of the relevant physiologic response, in the human cell type of interest. Thus, C1-S was apparently strongly biased away from β-arrestin, in contrast to albuterol and C5-S. C1-S structural modeling and simulations revealed binding differences compared with unbiased epinephrine at transmembrane (TM) segments 3,5,6,7 and ECL2. C1-S (R2 = cyclohexane) was repositioned in the pocket such that it lost a TM6 interaction and gained a TM7 interaction compared with the analogous unbiased C5-S (R2 = benzene group), which appears to contribute to C1-S biasing away from β-arrestin. Thus, an agnostic large chemical-space library identified agonists with receptor interactions that resulted in relevant signal splitting of β₂AR actions favorable for treating obstructive lung disease.

Most G protein–coupled receptors (GPCRs) are now recognized as multisignal transducers (1, 2). Early concepts of agonist–receptor interactions were based on the idea that there was a single “active” receptor conformation induced by the binding of any agonist, resulting in an interaction with the heterotrimeric G protein and a universal, singular signal. Generally, the α-subunit of the G protein, upon its dissociation, was considered the primary activator (or inhibitor) of the effector, resulting in the intracellular signal. Subsequently, it became clear that multiple signaling outcomes from activation of a given GPCR can occur from a single agonist due to specific molecular determinants of the receptor triggering independent mechanisms (3–5). As these multiple functions were being identified, it was apparent that agonists with different structures could act at a given receptor to preferentially activate one signal with minimal engagement of others, a property later termed signal biasing (6–8). Biased agonists, then, could represent important advantages over nonbiased agonists due to this signal selectivity, activating a specified therapeutic pathway while minimally evoking unnecessary or deleterious signaling. The pathway selectivity of biased agonists is thought to be established by the stabilization of specific conformation(s) of the agonist–receptor complex via a set of interactions that differ from those of unbiased (also called balanced) agonists (9–12). While it is conceivable that small modifications of established cognate agonists might yield such specialized signaling, significant deviation from common agonist structures may be necessary to meet this goal (13).The signals/functions of a given GPCR that might be sought for selective activation are defined by the cell type, disease, and desired final physiologic function. In asthma and chronic obstructive pulmonary disease (COPD), active human airway smooth muscle (HASM) cellular contraction limits airflow, representing a major cause of morbidity and mortality. β₂-adrenergic receptors (β₂ARs) expressed on HASM cells are the targets for binding of therapeutically administered β-agonists, which relax the cells via a cyclic adenosine monophosphate–mediated mechanism (14). β-agonists are used for treating acute bronchospasm as well as for long-term prevention. However, the HASM bronchodilator response to acute β-agonist is attenuated by receptor desensitization (15), with typical treatments of humans, or isolated HASM cells, leading to a loss of receptor function over time (16–18), clinically termed tachyphylaxis.Agonist-promoted desensitization of β₂AR (and other GPCRs) is due to partial uncoupling of the receptor to the G protein, which is initiated by phosphorylation of intracellular Ser/Thr residues of the receptor by G protein–coupled receptor kinases (GRKs) (19, 20). The GRK-phosphorylated β₂AR recruits β-arrestin1 or β-arrestin2 to these receptors, with subsequent interactions that appear to compete with the receptor for its binding to the Gα subunit, thus attenuating the intracellular response (11, 21). Such competition has been strongly inferred for the β₂AR (22, 23) and is compelling for rhodopsin–arrestin interactions (24). In addition, β-arrestin binding to GPCRs can initiate receptor internalization and other events such as receptor activation of ERK1/2 (25) through its multiprotein adapter functions. Thus β-arrestin engagement can be considered an early “second signal” of the β₂AR as well as a desensitization initiator for attenuating the Gs signal. An agonist that is biased toward Gαs coupling (cAMP production and airway smooth muscle ASM] relaxation) and away from β-arrestin binding (desensitization) would be desirable in treating obstructive lung diseases, since efficacy would not be attenuated acutely, nor would tachyphylaxis be experienced from extended treatment. While biased agonists favoring either G protein or β-arrestin (6) signaling have been described for some GPCRs (such as μ-opioid and type 1 angiotensin II receptors), Gαs biasing has not been apparent from most studies with catecholamine-like compounds for the β₂AR. Thus, we have little information as to whether the two β₂AR pathways can be differentially activated in a selective manner by an efficacious agonist, nor is it apparent from a structural standpoint what strategy might be employed to design agonists biased in this manner for this receptor.In order to find this type of biasing for the β₂AR, we screened a 40-million-compound scaffold ranking (SR) library that was agnostic to known β₂AR agonist structures. We found a scaffold in which substitutions of certain R groups led to individual compounds that are apparently Gαs-biased agonists for β₂AR with no apparent engagement of β-arrestin in model systems. Additional studies in HASM cells revealed a lack of tachyphylaxis of the relaxation effect by the lead compound compared with the most widely utilized β₂AR agonist, albuterol. The structure of this biased agonist is very different from that of catecholamine-like agonists. To ascertain the mechanism that may underlie this biased activity, we used structural modeling and molecular simulations and studied homologous compounds with different R groups and receptor mutagenesis to predict the interaction sites with the activated β₂AR. Such studies uncovered distinct structural characteristics that may be responsible for the biasing effect.