Beyond Receptor-Proximal Uncoupling: EGFR-PKC Signaling and Arrestin Deubiquitination-Coupled Nuclear Trafficking in GPCR Signaling Regulation

Abstract

G protein-coupled receptors (GPCRs) constitute the largest family of cell-surface signaling proteins and major therapeutic targets. Their function is critically shaped by desensitization, the attenuation of receptor responsiveness during sustained or repeated stimulation. The canonical model, established in systems such as rhodopsin and the β₂-adrenergic receptor, emphasizes GRK-mediated receptor phosphorylation, arrestin recruitment, and steric inhibition of G-protein coupling, resulting in rapid and largely reversible receptor desensitization. Accumulating evidence, however, indicates that GPCR adaptation can extend beyond this receptor-proximal mechanism. Here, I propose a Two-Level Framework that incorporates an EGFR-dependent adaptive program following sustained activation of selected GPCRs. EGFR serves as an upstream integration hub linking two complementary pathways: a PKC-dependent proximal pathway that mediates receptor-level desensitization, and a Src-PDK1-Akt-USP33 pathway proposed to promote arrestin deubiquitination and subsequent nuclear translocation of arrestin-Gβγ complexes. This translocation may reduce membrane-associated Gβγ availability, thereby contributing to persistent attenuation of receptor signaling, while also influencing nuclear transcriptional activity. The persistent component is revealed by washout-rechallenge paradigms, in which signaling remains impaired despite apparent receptor recovery. Thus, Level II represents an additional layer of GPCR adaptation in which receptor desensitization is accompanied by longer-lasting remodeling of intracellular signaling capacity. Although derived primarily from D2-like and β₂-adrenergic receptor systems, this framework may apply more broadly to GPCRs capable of transactivating EGFR or related receptor tyrosine kinases, providing a broader view of desensitization that integrates receptor-level regulation with network-level adaptation.

Introduction

The canonical signaling cycle of G protein-coupled receptors (GPCRs) is well established. Agonist binding induces conformational changes that promote GDP-to-GTP exchange on the Gα subunit of heterotrimeric G proteins, resulting in dissociation of Gα-GTP from Gβγ subunits. These signaling components regulate diverse downstream effectors, including adenylyl cyclases, phospholipase Cβ, ion channels, and protein kinases, ultimately generating second messengers such as cAMP, IP3/DAG, and Ca²⁺ that coordinate a broad range of physiological responses (Gilman, 1987; Weis and Kobilka, 2018).

Because GPCR signaling must remain both precise and adaptable, mechanisms that limit receptor responsiveness are equally important. Among these, desensitization represents the most rapid and fundamental form of negative regulation and is defined as the decline in receptor responsiveness during sustained or repeated agonist stimulation (Gainetdinov et al., 2004; Hausdorff et al., 1990). Extensive studies have established the GRK-arrestin pathway as the canonical mechanism underlying acute GPCR desensitization. Following agonist stimulation, GPCR kinases (GRKs) phosphorylate activated receptors, promoting arrestin recruitment and steric blockade of productive G-protein coupling at the plasma membrane (Kohout et al., 2001; Mundell et al., 1999; Paing et al., 2002). This receptor-proximal process provides a rapid and largely reversible mechanism for limiting GPCR responsiveness and is referred to here as Level I desensitization.

Despite the success of this model, accumulating evidence indicates that receptor uncoupling alone does not fully account for all forms of desensitization. Studies employing washout-rechallenge (pulse-chase-pulse) paradigms have demonstrated that signaling responsiveness can remain attenuated even after receptor dephosphorylation, recycling, and restoration of surface expression (Cho et al., 2010; Min et al., 2013). Here, washout-rechallenge refers to an experimental protocol in which cells are first exposed to an agonist pulse, subsequently washed to allow receptor recovery, and then challenged with a second agonist pulse to determine whether signaling competence has been fully restored. In contrast, continuous-stimulation paradigms monitor signaling during uninterrupted agonist exposure, when receptor phosphorylation and arrestin recruitment dominate. The persistent signaling deficit revealed by washout-rechallenge experiments therefore suggests the existence of regulatory mechanisms that extend beyond classical receptor-proximal desensitization.

This apparent discrepancy raises a fundamental question: How is transient receptor-associated desensitization converted into a more persistent state of signaling attenuation? Increasing evidence identifies epidermal growth factor receptor (EGFR) transactivation as a central signaling hub downstream of many GPCRs. Although originally recognized as a mechanism linking GPCR activation to receptor tyrosine kinase signaling (Daub et al., 1996), EGFR transactivation is now known to engage multiple downstream pathways capable of reshaping cellular signaling networks (Kundu et al., 2024). Rather than simply amplifying GPCR signaling, EGFR may coordinate multiple downstream regulatory pathways that together establish a hierarchical adaptive response during sustained receptor activation.

Among the pathways activated by EGFR transactivation, protein kinase C (PKC) reinforces receptor-level attenuation through heterologous phosphorylation of GPCRs independently of agonist-driven GRK recruitment (Cho et al., 2007; Gainetdinov et al., 2004; Hausdorff et al., 1990). In parallel, activation of the Src-PDK1-Akt pathway engages the deubiquitinating enzyme USP33 (ubiquitin-specific peptidase 33), promoting arrestin deubiquitination and the formation of cytoplasmic deubiquitinated arrestin-Gβγ complexes that subsequently translocate to the nucleus. Although the functional significance of this redistribution remains incompletely defined, it has been proposed to reduce the availability of membrane-associated Gβγ required for efficient GPCR signaling, thereby contributing to the persistent attenuation observed in washout-rechallenge paradigms. Nuclear accumulation of deubiquitinated arrestin may additionally couple persistent desensitization to arrestin-dependent transcriptional regulation and longer-term cellular adaptation, although direct evidence supporting these downstream consequences remains limited (Cho et al., 2010; Kang et al., 2005; Min et al., 2013, 2023; Wang et al., 2003b; Witherow et al., 2004; Zheng et al., 2020).

To integrate these observations, I propose a Two-Level Framework for GPCR desensitization (Fig. 1). Level I comprises the classical receptor-proximal mechanism mediated by GRK-dependent phosphorylation, arrestin recruitment, and receptor uncoupling, producing rapid but largely reversible attenuation of GPCR signaling. Level II, in contrast, represents an EGFR-associated hierarchical adaptation program that emerges during sustained receptor activation and remodels intracellular signaling capacity beyond receptor uncoupling. Within this framework, Level II consists of two complementary functional modules: A proximal module that reinforces receptor-level desensitization and a distal remodeling module that redistributes intracellular signaling components. Together, these processes provide a plausible mechanistic basis for persistent signaling attenuation following receptor recovery, although several key steps remain to be experimentally validated.

Fig. 1.

Overview of the proposed two-level model of GPCR desensitization. This review categorizes GPCR desensitization into two mechanistically distinct processes: Level I (proximal) and Level II (distal) desensitization. The upper panel illustrates the canonical Level I pathway, in which sustained agonist stimulation induces GRK-mediated receptor phosphorylation, arrestin recruitment, and rapid receptor desensitization at the plasma membrane through disruption of productive receptor-G protein coupling. The lower panel summarizes the proposed Level II pathway, which is triggered following repeated receptor stimulation and involves EGFR transactivation, PKC activation, and USP33-dependent arrestin deubiquitination. Deubiquitinated arrestin associates with Gβγ and undergoes importin-dependent nuclear translocation. This process is proposed to contribute to persistent desensitization by reducing the membrane-proximal Gβγ pool and may additionally influence gene transcription, although these downstream consequences remain hypothetical.

Within this framework, continuous-stimulation and washout-rechallenge paradigms should be viewed as complementary rather than competing experimental approaches. Continuous stimulation primarily reflects the kinetics of Level I desensitization because receptor phosphorylation and arrestin recruitment dominate during ongoing agonist exposure. By contrast, washout-rechallenge experiments reveal Level II regulation by determining whether signaling competence has been fully restored after receptor dephosphorylation and recycling. Consequently, receptors may satisfy conventional biochemical criteria for resensitization while signaling competence remains persistently impaired because intracellular signaling networks have not yet returned to their pre-stimulation state (Cho et al., 2010; Min et al., 2013).

Although the molecular mechanisms underlying Level II desensitization are likely to differ among GPCR families, evidence from dopamine D2-like receptors and β₂-adrenoceptors suggests that persistent attenuation is not adequately explained by receptor uncoupling alone (Cho et al., 2010; Kundu et al., 2024; Min et al., 2023).

In this review, I evaluate the experimental evidence supporting this Two-Level Framework and discuss how EGFR transactivation integrates canonical receptor desensitization with secondary regulatory pathways. Particular emphasis is placed on the complementary functions of the EGFR-PKC and EGFR-Src-PDK1-Akt-USP33 pathways and on how arrestin redistribution may link receptor desensitization to longer-term adaptive signaling programs. By integrating receptor-proximal and EGFR-dependent mechanisms into a unified sequential model, I aim to provide a conceptual framework for understanding persistent GPCR desensitization, signaling plasticity, and potential therapeutic modulation of GPCR responsiveness.

The canonical GRK/arrestin model: foundation of Level I desensitization

The canonical model of GPCR desensitization provides the mechanistic foundation for Level I desensitization in the Two-Level Framework proposed in this review. In this model, receptor phosphorylation by GRKs promotes arrestin recruitment, leading to steric inhibition of G protein coupling. Accordingly, Level I represents a receptor-proximal, membrane-confined regulatory program that produces rapid and largely reversible attenuation of GPCR signaling (Benovic et al., 1987; Kohout et al., 2001) and serves as the foundation upon which the EGFR-dependent Level II program is subsequently superimposed.

Agonist-induced receptor activation promotes GRK recruitment and phosphorylation of residues within the receptor C-terminal tail and/or intracellular loops. Phosphorylation increases the affinity of the activated receptor for arrestin2/3, allowing arrestin to engage the receptor in a core-bound conformation. In this state, arrestin sterically occludes the transmembrane cavity that normally accommodates the Gα α5 helix, thereby preventing productive receptor-G protein coupling (Fig. 2). Arrestin additionally recruits clathrin and AP-2, directing the receptor into clathrin-coated pits for internalization and subsequent trafficking.

Fig. 2.

Structural basis of arrestin-mediated steric blockade and competitive exclusion at the intracellular cavity. Schematic illustration of the competitive exclusion mechanism within the GPCR intracellular core. The intracellular transmembrane cavity is formed by helices TM5 and TM6 (gray). Productive G protein signaling requires insertion of the Gα C-terminal α5 helix (blue) into this cavity. Upon recruitment in a core-engaged conformation, the arrestin finger loop (teal) occupies the same structural region. Because the Gα α5 helix and the arrestin finger loop engage overlapping interfaces within the receptor core, their binding is mutually exclusive (indicated by the crossed arrow). This steric hindrance directly prevents receptor-G protein coupling, thereby mediating receptor desensitization (TM, transmembrane helix).

The GRK family: Initiators of the desensitization program

The GRK family comprises seven mammalian members (GRK1-7) that serve as the primary initiators of the desensitization relay (Pitcher et al., 1998; Premont and Gainetdinov, 2007). GRK2, GRK3, GRK5, and GRK6 are broadly expressed and regulate the majority of non-visual GPCRs. All GRKs share a central kinase domain whose full catalytic activity requires engagement with an agonist-occupied receptor, thereby conferring intrinsic selectivity for active signaling units (Lodowski et al., 2006; Tesmer et al., 2005).

Despite this shared catalytic mechanism, GRK isoforms differ substantially in their modes of membrane recruitment, receptor engagement, and downstream signaling consequences. GRK2 and GRK3 are predominantly cytosolic proteins that are recruited to activated receptors through interactions with liberated Gβγ subunits and membrane phospholipids via their pleckstrin homology (PH) domains (Pitcher et al., 1998; Premont and Gainetdinov, 2007). In contrast, members of the GRK4 subfamily, particularly GRK5 and GRK6, are constitutively associated with the plasma membrane via lipid-binding domains and amphipathic helices, thereby making them less dependent on acute Gβγ-mediated membrane recruitment (Komolov et al., 2017; Pitcher et al., 1998). As a result, GRK5/6 can sustain receptor phosphorylation under conditions where GRK2/3 recruitment may be limiting.

Importantly, accumulating evidence indicates that distinct GRK isoforms generate different receptor phosphorylation “barcodes,” thereby stabilizing distinct arrestin conformational ensembles and promoting divergent signaling outcomes (Nobles et al., 2011; Zidar et al., 2009). In several GPCR systems, GRK2/3-dependent phosphorylation preferentially promotes transient arrestin interactions associated with rapid desensitization and receptor recycling, whereas GRK5/6-dependent phosphorylation has been linked to more stable arrestin engagement and enhanced arrestin-mediated signaling, including sustained ERK activation (Wei et al., 2003; Zidar et al., 2009). These observations support the concept that GRKs function not merely as signal terminators but as upstream encoders of signaling bias and receptor fate determination.

Beyond receptor phosphorylation, several GRK isoforms also participate in non-canonical signaling activities. GRK2 and GRK3 can modulate signaling through kinase-independent interactions with Gαq and liberated Gβγ, thereby restricting access of downstream effectors (Carman et al., 1999; Tesmer et al., 2005). GRK5, in contrast, can translocate to the nucleus during prolonged receptor stimulation and phosphorylate non-receptor substrates such as HDAC5 (histone deacetylase 5), linking GPCR activation to transcriptional regulation and longer-term cellular responses (Martini et al., 2008). Although these activities broaden the signaling repertoire of GRKs, they are mechanistically distinct from the EGFR-triggered hierarchical adaptation program (Level II) proposed in this review.

Structural basis of level I: Steric exclusion and receptor-arrestin coupling

The fundamental mechanism of Level I desensitization is steric exclusion. High-resolution structures of GPCR-arrestin complexes, beginning with rhodopsin-visual arrestin and expanding to β₁AR-arrestin2, have established a conserved structural framework for this interaction (Kang et al., 2015; Huang et al., 2020). A key feature is the formation of an intracellular cavity upon receptor activation, where the outward displacement of transmembrane helix 6 (TM6) creates a binding pocket (Rasmussen et al., 2011).

Structural superposition demonstrates that this pocket accommodates either the α5 helix of Gα (blue) or the finger loop of arrestin (teal) in a mutually exclusive manner (Garcia-Nafria et al., 2018; Lee et al., 2020). This competition at the receptor core provides direct evidence that arrestin binding sterically precludes G protein coupling (Fig. 2). Consequently, Level I is defined by this "receptor-core occlusion," which acts as a rapid biochemical brake at the plasma membrane. The stability of this interaction typically follows Class A or Class B patterns, where clustered phosphorylation sites (Class B, e.g., V₂R) promote high-affinity, long-lived complexes compared to the more transient interactions seen in Class A receptors (e.g., β₂AR) (Kim et al., 2005; Oakley et al., 2000).

It is important to distinguish canonical receptor trafficking from the persistent desensitization described in the Two-Level Framework. Canonical GPCR regulation comprises receptor phosphorylation, arrestin recruitment, internalization, dephosphorylation, and recycling, processes that are largely completed within tens of minutes following agonist exposure (Claing et al., 2002; Hausdorff et al., 1990). However, persistent attenuation of signaling can remain after receptor recycling has restored receptor competence, indicating that recovery of receptor availability is not necessarily accompanied by recovery of cellular signaling capacity (Min et al., 2017, 2023).

Furthermore, receptor internalization is not simply a mechanism for signal termination but a critical step in receptor resensitization (Cho et al., 2010; Yu et al., 1993). Internalized GPCRs are trafficked to endosomes, where phosphatases such as PP2A (protein phosphatase 2A) remove GRK-mediated phosphorylation (Little and Lee, 1995), restoring receptor competence before recycling to the plasma membrane (Cho et al., 2010; Zhang and Kim, 2017). Internalization also limits receptor accessibility to extracellular agonists (Sorkin and von Zastrow, 2009) and can support a distinct phase of endosomal G protein- or arrestin-dependent signaling (Calebiro et al., 2009; Ferrandon et al., 2009).

Structural basis of phosphorylation barcode interpretation

Although distinct GRK isoforms establish receptor-specific phosphorylation barcodes, the molecular basis by which arrestins interpret these phosphorylation patterns has only recently become clear. While steric hindrance is the physical outcome of Level I desensitization, the recruitment, orientation, and functional engagement of arrestin are governed by the “phosphorylation barcode.” This hypothesis proposes that distinct GRK-dependent phosphorylation patterns stabilize different arrestin conformational ensembles, thereby biasing arrestin toward specific functional outcomes (Butcher et al., 2011; Nobles et al., 2011).

Structural and biophysical studies support this model, demonstrating that distinct receptor phospho-patterns engage the positively charged N-domain of arrestin in different ways and promote varying degrees of tail-engaged and core-engaged receptor-arrestin complexes (Bous et al., 2022; Staus et al., 2020). In the core-engaged state, representing a fully engaged complex, the arrestin finger loop inserts deeply into the intracellular cavity created by the outward displacement of TM6, partially or fully occluding the G protein-binding interface and thereby favoring proximal desensitization (Huang et al., 2020; Kang et al., 2015) (Fig. 2). In contrast, tail-engaged complexes, functioning as partially engaged complexes, are characterized predominantly by interactions between phosphorylated receptor tails and the arrestin N-domain, with less extensive engagement of the receptor core. Crucially, tail-engaged complexes reduce steric occlusion of the G protein-binding interface and permit the formation of transient GPCR-G protein-arrestin "megacomplexes," allowing sustained G protein signaling while simultaneously supporting arrestin scaffolding and clathrin-mediated endocytosis (Kumari et al., 2016; Lee et al., 2020; Nguyen et al., 2019; Thomsen et al., 2016).

Importantly, these states should not be viewed as rigid binary conformations but rather as members of a dynamic conformational continuum whose stability and functional consequences depend on the specific phosphorylation pattern, receptor class, and cellular context. Indeed, the tail-engaged/core-engaged framework may not universally capture the full diversity of GPCR-arrestin interactions. Certain receptors, particularly those with relatively short C-terminal tails but extensive phosphorylation sites in intracellular loops, may recruit arrestins through partially distinct structural mechanisms, as suggested for several dopamine receptor family members and other class A GPCRs (Kim et al., 2001; Lan et al., 2009; Pack et al., 2018). Thus, the phosphorylation barcode is likely interpreted through receptor-specific conformational landscapes rather than through a single universally conserved binary model.

Accordingly, the phosphorylation barcode determines not only the kinetics of arrestin recruitment but also the qualitative trajectory of receptor desensitization, receptor trafficking, and arrestin-dependent signaling.

Functional diversification: Arrestin as a multi-state regulator

Building upon this framework, it is now clear that arrestin activation involves a continuum of conformational states rather than a single "on" state (Gurevich and Gurevich, 2006). These conformational states give rise to distinct functional outputs, including G protein uncoupling, clathrin-mediated endocytosis, and arrestin-dependent biased signaling (Shenoy et al., 2006).

The complexity of arrestin regulation is exemplified by closely related GPCRs that exhibit distinct arrestin behaviors despite high structural similarity. For example, the vasopressin V₂ receptor (V₂R) forms stable, long-lived complexes with arrestin (Class B), whereas the β₂AR interacts with arrestin more transiently (Class A), resulting in marked differences in receptor trafficking and signaling persistence (Nobles et al., 2011; Oakley et al., 2000). These differences reflect the ability of individual receptors to stabilize distinct arrestin conformational ensembles, thereby biasing signaling toward receptor uncoupling, endocytic trafficking, or arrestin-dependent signaling.

Nuclear roles of arrestins

In addition to their well-established cytoplasmic functions, arrestins can also translocate to the nucleus, where they regulate transcription through multiple molecular interactions. Nuclear arrestin2 has been shown to interact with the transcriptional co-activators, CREB-binding protein (CBP) and p300, thereby modulating histone acetyltransferase activity and chromatin-dependent transcription (Kang et al., 2005). Arrestin2-dependent nuclear transport of inhibitor of κB alpha (IκBα) suppresses nuclear factor κB (NF-κB)-mediated inflammatory gene expression by inhibiting transcriptional activity at target promoters (Witherow et al., 2004). In addition, arrestins interact with the E3 ubiquitin ligase Mdm2 (mouse double minute 2 homolog), linking GPCR signaling to p53-dependent transcriptional and cell-fate regulatory pathways (Wang et al., 2003a). Together, these findings establish arrestins as multifunctional regulators that extend GPCR signaling beyond the plasma membrane into the nucleus.

These observations indicate that arrestins function not only as regulators of receptor signaling at the plasma membrane but also as modulators of nuclear signaling pathways. As discussed in later sections, I propose that sustained GPCR activation engages an EGFR-triggered hierarchical adaptation program comprising two complementary downstream arms. One arm activates PKC to execute receptor-proximal reinforcement of desensitization, whereas the other engages the Src-PDK1-Akt-USP33 pathway to promote arrestin deubiquitination, cytoplasmic arrestin-Gβγ complex formation, and subsequent nuclear translocation of arrestin. This pathway may provide a potential link between persistent GPCR desensitization and longer-term adaptive transcriptional responses, although the molecular determinants governing arrestin nuclear trafficking remain to be fully defined.

Collectively, the studies discussed in this section establish the GRK-arrestin pathway as the molecular foundation of rapid receptor-proximal desensitization (Level I). Although receptor phosphorylation, arrestin recruitment, internalization, and recycling adequately explain the acute attenuation and recovery of GPCR responsiveness, they do not fully account for the persistent reduction in signaling competence revealed by washout-rechallenge paradigms after apparent receptor recovery. These findings therefore suggest that receptor-proximal regulation alone is insufficient to explain persistent loss of signaling competence. As developed in the following sections, I propose that sustained GPCR activation engages an EGFR-triggered hierarchical adaptation program (Level II), in which receptor-level reinforcement through PKC and distal remodeling through the Src-PDK1-Akt-USP33 pathway together establish persistent desensitization beyond canonical receptor uncoupling.

Receptor subtype diversity reveals the limits of the canonical desensitization model

The β₂ adrenergic receptor: Paradigm and complexity

The β₂AR has served as the primary model for the mechanistic dissection of GPCR desensitization. Rapid agonist-dependent phosphorylation by GRK2 and GRK3, followed by arrestin recruitment and receptor internalization via clathrin-coated pits, was first described for this receptor (Benovic et al., 1987; Ferguson et al., 1996; Goodman et al., 1996; Lohse et al., 1990). Quantitative and imaging studies have confirmed that β₂AR desensitization occurs on a timescale consistent with GRK2 recruitment, receptor phosphorylation, and arrestin engagement (Krasel et al., 2005; Violin et al., 2008).

Despite its status as a paradigm, β₂AR desensitization reveals complexity when examined under different conditions. β₂AR can undergo heterologous desensitization mediated by protein kinase A (PKA) and PKC, which phosphorylate the receptor at different sites than GRKs and can shift receptor coupling from Gs to Gi, a mechanism with important implications for cardiac β-adrenergic physiology (Daaka et al., 1997; Zamah et al., 2002). Additionally, work by Shenoy and colleagues demonstrated that β₂AR ubiquitination and the dynamic deubiquitination of bound arrestin by USP33 regulate the kinetics of receptor trafficking and the distinction between Class A (rapid dissociation) and Class B (sustained arrestin association) behavior (Calebiro et al., 2009; Oakley et al., 2000; Shenoy et al., 2009). β₂AR therefore illustrates that canonical GRK-arrestin-mediated desensitization can diverge into an EGFR-dependent secondary regulatory program, thereby providing an instructive bridge between Level I and Level II desensitization (Kundu et al., 2024).

μ-Opioid and δ-opioid receptors: Tolerance and phosphorylation diversity

Opioid receptors provide some of the most clinically important examples of GPCR desensitization because tolerance and loss of analgesic efficacy remain major challenges in pain management. At the μ-opioid receptor (MOR), this is exemplified by the contrasting actions of peptide agonists such as DAMGO and alkaloid agonists such as morphine. DAMGO robustly promotes receptor phosphorylation, arrestin recruitment, and efficient receptor internalization, whereas morphine, despite producing potent analgesia, induces relatively limited receptor internalization together with a more restricted phosphorylation pattern (Christie, 2008; Keith et al., 1998; Whistler and von Zastrow, 1998). This agonist-dependent divergence demonstrates that receptor activation alone does not determine the outcome of desensitization. Rather, ligand-specific phosphorylation patterns and the downstream trafficking programs they initiate shape the functional fate of activated receptors.

The differential regulation of MOR also illustrates a limitation of models in which desensitization is explained solely by arrestin recruitment. MOR trafficking is determined not only by the extent of arrestin binding but also by the quality of receptor phosphorylation and the conformational state adopted by arrestin. Compared with high-efficacy agonists such as DAMGO, morphine induces a distinct GRK-dependent phosphorylation pattern that fails to efficiently stabilize arrestin3 in a trafficking-competent conformation, thereby limiting recruitment of the endocytic machinery (Just et al., 2013; Nobles et al., 2011; Shukla et al., 2014). Consequently, morphine-bound MOR exhibits ligand-directed signaling bias in which distinct agonists encode different receptor regulatory programs rather than simply different levels of receptor activation (Huang et al., 2026).

A complementary but mechanistically distinct principle is observed at the δ-opioid receptor (DOR), where the persistence of desensitization is strongly influenced by post-endocytic trafficking and receptor recycling. Rather than being determined exclusively by receptor phosphorylation, recovery of signaling responsiveness depends on the stability of receptor-associated signaling complexes during receptor trafficking. Studies suggest that receptor-Gβγ interactions contribute to this process by influencing arrestin engagement and the efficiency of receptor resensitization following agonist exposure (Audet et al., 2012).

Together, studies of MOR and DOR demonstrate that opioid receptor desensitization cannot be explained solely by receptor phosphorylation and arrestin recruitment. Instead, receptor-specific phosphorylation patterns, arrestin conformational states, and post-endocytic trafficking properties collectively determine signaling outcomes and the extent of functional recovery. These observations are consistent with the Two-Level Framework proposed in this review, in which receptor-proximal regulation establishes acute desensitization (Level I), whereas persistent alterations in signaling competence are likely to require additional intracellular adaptive mechanisms that extend beyond receptor uncoupling (Level II).

Muscarinic receptors: GRK isoform selectivity and Gq coupling

The five muscarinic acetylcholine receptor subtypes (M₁-M₅) illustrate how GRK isoform selectivity contributes to the diversity of GPCR desensitization. Gi/o-coupled M₂ and M₄ receptors are preferentially regulated by GRK2 and GRK3, whereas Gq/11-coupled M₁, M₃, and M₅ receptors engage multiple GRK isoforms that exert distinct effects on receptor regulation and signal termination (Haga and Haga, 1992; Tsuga et al., 1994; Willets et al., 2005). These subtype-specific differences indicate that receptor-proximal desensitization is not governed by a single, uniform GRK-arrestin mechanism but instead depends on the repertoire of GRK isoforms available to individual receptors.

The M₃R provides a representative example of this complexity. GRK6 directly promotes receptor phosphorylation and uncoupling, whereas GRK3 attenuates Gq/PLC signaling through mechanisms that cannot be fully explained by receptor phosphorylation alone (Willets et al., 2001, 2005). These findings suggest that individual GRK isoforms regulate signaling at multiple levels, influencing both receptor-proximal events and downstream signaling efficiency.

Together, muscarinic receptors demonstrate that even receptor-proximal desensitization represents a heterogeneous regulatory process rather than a single canonical pathway. Differences in GRK isoform engagement and receptor subtype context create distinct patterns of signaling attenuation, highlighting the complexity of Level I regulation. This receptor-proximal diversity provides an important framework for understanding why additional regulatory mechanisms may be required to account for persistent alterations in GPCR signaling competence.

Dopamine receptors: A critical test case

Dopamine D2-like receptors (D₂R, D₃R, and D₄R) provide an especially instructive subtype comparison because they preserve broad Gi/o coupling while displaying markedly different desensitization profiles. As such, they provide a critical test of whether receptor-proximal mechanisms alone are sufficient to explain differences in desensitization among closely related GPCR subtypes.

The decisive paradox emerges from comparison of D₂R and D₃R. D₂R undergoes efficient agonist-dependent phosphorylation, robust arrestin3 translocation, and substantial receptor internalization, yet exhibits relatively limited functional desensitization. In contrast, D₃R displays profound desensitization despite limited receptor phosphorylation and comparatively weak arrestin recruitment (Cho et al., 2006, 2010; Kim et al., 2001; Min et al., 2013; Westrich and Kuzhikandathil, 2007). This mismatch is difficult to reconcile with a model in which phosphorylation-dependent arrestin recruitment directly predicts the extent of signaling attenuation. D₄R further highlights this diversity by demonstrating that closely related receptor subtypes can exhibit distinct patterns of phosphorylation, trafficking, and functional regulation (Cho et al., 2006; Min et al., 2017). The pharmacological implications of this divergence are substantial. Antipsychotic drugs targeting D2-like receptors can produce tolerance, dopamine supersensitivity, or paradoxical sensitization during chronic treatment, phenomena that are not fully explained by changes in receptor abundance alone and may reflect altered engagement of regulatory mechanisms beyond classical receptor uncoupling (Kim, 2023; Li, 2016; Sokoloff and Le Foll, 2022).

Together, the D₂R/D₃R comparison demonstrates that receptor phosphorylation and arrestin recruitment are important but insufficient predictors of persistent attenuation of signaling. These findings raise the possibility that the functional state and intracellular fate of arrestin, rather than recruitment alone, may influence whether receptor activation progresses from rapid receptor-proximal desensitization toward a more persistent adaptive state. This concept provides a mechanistic basis for exploring how downstream pathways, including EGFR-dependent regulation of arrestin function, may contribute to Level II desensitization.

Class B GPCRs: V₂R and AT₁R as examples of stable arrestin engagement

The V₂R and the angiotensin II type 1A receptor (AT₁R), which exhibit Class B-like arrestin trafficking properties, illustrate how arrestin engagement can support receptor regulation while also enabling sustained intracellular signaling. In V₂R, arrestin remains associated with the receptor following endocytosis, enabling prolonged arrestin-dependent ERK signaling from endosomal compartments (He et al., 2026; Nguyen et al., 2019). Structural studies further support the concept that arrestin engagement does not necessarily indicate complete signaling shutdown, as receptor-G protein-arrestin complexes can maintain selected signaling outputs under specific conditions (Thomsen et al., 2016).

In AT₁R, arrestin3 interaction promotes cardioprotective signaling pathways, including ERK and Akt activation, that are mechanistically distinct from Gq-mediated cardiac hypertrophic signaling (Rajagopal et al., 2006; Violin et al., 2014). Thus, AT₁R demonstrates that arrestin-mediated attenuation of G protein signaling and activation of distinct downstream signaling programs are not mutually exclusive but can represent coordinated aspects of receptor regulation. More broadly, this duality is a defining feature of many Class B receptor systems, in which stable receptor-arrestin complexes support both prolonged signaling and altered receptor trafficking behavior.

Collectively, these Class B receptor systems demonstrate that arrestin recruitment is not a binary switch between active and inactive receptor states. Instead, arrestin function depends on receptor context, trafficking properties, and the intracellular signaling environment. These examples highlight the functional versatility of the canonical GRK-arrestin pathway but also emphasize its limitation: Although arrestin can regulate receptor trafficking and initiate distinct signaling programs, persistent loss of signaling competence after receptor recovery requires additional regulatory mechanisms beyond receptor-proximal control. This concept provides the rationale for the EGFR-dependent Level II desensitization framework proposed in this review.

Beyond receptor-proximal desensitization: EGFR transactivation as a hierarchical checkpoint for Level II desensitization

Comparisons across diverse GPCR families suggest that receptor-proximal desensitization alone cannot fully explain the diversity, magnitude, or persistence of signaling attenuation. If receptor phosphorylation and steric uncoupling at the plasma membrane were sufficient to determine desensitization outcomes, receptors with similar G protein-coupling properties would be expected to exhibit comparable patterns of signal attenuation. However, this is often not observed. A notable example is the dopamine D2-like receptor family, in which D₂R and D₃R share high structural homology and couple predominantly to Gi/o proteins yet display markedly different desensitization phenotypes (Cho et al., 2007; Kim et al., 2001; Westrich and Kuzhikandathil, 2007).

These observations indicate that receptor phosphorylation and arrestin recruitment represent essential initiating events in desensitization but are insufficient to determine the magnitude, persistence, or receptor-specific variability of adaptive signaling attenuation. Within the Two-Level Desensitization Framework proposed here, progression beyond Level I requires engagement of an additional intracellular signaling checkpoint. EGFR transactivation is proposed to function as this hierarchical checkpoint, determining whether receptor attenuation remains a transient receptor-proximal event or progresses into a persistent adaptive state. Whereas receptor-proximal phosphorylation barcodes diversify the acute regulatory responses that define Level I (Nobles et al., 2011; Staus et al., 2020), EGFR-dependent signaling determines entry into the persistent adaptive program proposed for Level II (Fig. 1, Table 1).

Table 1. Comparison of Level I and Level II GPCR desensitization.

Parameter	Level I: Proximal desensitization	Level II: Distal desensitization
Primary site of control	Receptor/plasma membrane	Redistribution of signaling components across compartments
Core mechanism	GRK-dependent phosphorylation, arrestin engagement, reduced receptor-G protein coupling	Arrestin deubiquitination and nuclear translocation, potentially resulting in Gβγ sequestration and component limitation (not directly demonstrated)
Triggering condition	Acute receptor activation	Sustained stimulation with engagement of EGFR-Src-PDK1-Akt-USP33 relay
Temporal profile	Rapid	Delayed/progressive
Reversibility	Often readily reversible	More persistent and slower to reset
Principal readouts	Receptor phosphorylation, arrestin recruitment, initial uncoupling, internalization	PKC activation, Gβγ redistribution, altered subcellular localization, and durable signaling loss
Proposed predictive value for long-term attenuation	Limited	High
Conceptual interpretation	Receptor-centered attenuation	Spatial reprogramming of signaling competence

It should be emphasized that the proposed Level II pathway is derived primarily from studies of dopamine D2-like receptors and β₂-adrenoceptors. Whether analogous mechanisms operate broadly across other GPCR families remains to be established. Nevertheless, these systems provide a mechanistic framework for understanding how sustained GPCR activation can engage intracellular remodeling processes beyond classical receptor uncoupling.

Mdm2 maintains the basal signaling-permissive state of arrestin

Under basal conditions, cytoplasmic Mdm2 maintains a constitutive pool of ubiquitinated arrestin (Arr-Ub), thereby establishing an arrestin state compatible with receptor-associated signaling competence (Min et al., 2017; Shenoy and Lefkowitz, 2003) (Fig. 3). Within this basal licensing state, Arr-Ub promotes the assembly of a preassembled signaling complex comprising the receptor (R), Gβγ, and Arr-Ub. This complex remains signaling-competent, requiring only Gα recruitment for rapid activation upon agonist stimulation. Rather than functioning as a desensitizing modification, this basal ubiquitination preserves normal receptor signaling capability while preventing premature initiation of the Level II remodeling program.

Fig. 3.

Mdm2-dependent basal licensing and ubiquitination dynamics of arrestin in Level II GPCR desensitization. Schematic model illustrating the proposed role of Mdm2/USP33-regulated arrestin ubiquitination in the molecular licensing of Level II GPCR desensitization. Basal licensing state: Under basal conditions, cytoplasmic Mdm2 maintains arrestin in a ubiquitinated form (Arr-Ub), maintaining a preassembled signaling complex consisting of the receptor (R), Gβγ, and Arr-Ub. This complex remains signaling-competent and requires only recruitment of Gα for rapid activation upon agonist stimulation (red stars). Agonist activation and signaling: Following agonist binding, Gα-GDP associates with the receptor complex, undergoes GDP-GTP exchange, and dissociates into active Gα-GTP and Gβγ subunits. Subsequent GTP hydrolysis converts Gα-GTP back to Gα-GDP, generating inorganic phosphate (Pi). Desensitizing conditions and Level II desensitization: During sustained or repeated agonist exposure, cytoplasmic Mdm2 translocates to the nucleus (dotted red arrow), while USP33-mediated deubiquitination converts Arr-Ub to deubiquitinated arrestin (Arr) (Fig. 4). Deubiquitinated arrestin exhibits enhanced affinity for free Gβγ and sequesters the Gβγ dimer (red dashed arrows). This interaction prevents reassociation of Gα-GDP with Gβγ (red X), thereby preventing reassembly of the heterotrimeric G protein. Sequestration and redistribution of Gβγ away from the membrane-proximal signaling pool constitute a molecular licensing mechanism that contributes to Level II desensitization and attenuates subsequent receptor signaling. Mdm2, murine double minute 2; Ub, ubiquitin; Arr, arrestin; USP33, ubiquitin-specific-processing protease 33; Gα, G protein α subunit; Gβγ, G protein βγ dimer; R, receptor; GDP, guanosine diphosphate; GTP, guanosine triphosphate; Pi, inorganic phosphate.

Recent structural studies have demonstrated that activated GPCRs can simultaneously associate with G proteins and arrestins within receptor-G protein-arrestin "megacomplexes," proving that arrestin recruitment does not inherently terminate G protein signaling (He et al., 2026; Nguyen et al., 2019; Thomsen et al., 2016). Although these structural models did not explicitly define the ubiquitination status of arrestin, biochemical evidence indicates that ubiquitinated arrestin is fully compatible with receptor-associated signaling complexes that maintain Gβγ availability and permit rapid Gα coupling (Min et al., 2017).

A structural basis for this regulation is suggested by studies showing that Mdm2 interacts with the N-domain surface of arrestin, which partially overlaps with the receptor phosphotail-binding interface (Yun et al., 2023). Although this overlap implies potential competition between receptor- and Mdm2-bound states, the degree to which both can be simultaneously engaged remains structurally unresolved. Notably, Mdm2 preferentially binds the inactive conformation of arrestin, raising the possibility that receptor-induced conformational activation reduces Mdm2 affinity. Together, these observations support a sequential model in which Mdm2 establishes and maintains a basal pool of ubiquitinated arrestin prior to receptor engagement, rather than requiring persistent Mdm2 occupancy during active receptor signaling.

Functional evidence supporting this model comes from C147K-D₃R and K149C-D₂R mutants, which preserve receptor-associated Gβγ availability and delay progression into the USP33-dependent remodeling pathway (Min et al., 2017, 2023; Zheng et al., 2020).

Collectively, these structural and functional observations indicate that Mdm2 acts upstream of Level II remodeling by establishing a basal ubiquitinated arrestin state that remains signaling-competent (Fig. 3). Upon sustained receptor stimulation, cytoplasmic Mdm2 translocates to the nucleus, enabling USP33-dependent deubiquitination to convert Arr-Ub to deubiquitinated arrestin (Arr). Deubiquitinated arrestin then sequesters free Gβγ and prevents heterotrimeric G protein reassembly, thereby driving the transition into Level II desensitization (Fig. 3, 4).

Fig. 4.

Proposed molecular mechanism of EGFR-dependent Level II GPCR desensitization. Schematic model depicting the signaling pathways underlying distal (Level II) GPCR desensitization. ADAM-mediated EGFR transactivation: Repeated 7-TM GPCR stimulation activates an ADAM metalloproteinase, which cleaves membrane-anchored pro-HB-EGF to release soluble HB-EGF (sHB-EGF). Extracellular sHB-EGF binds to EGFR and promotes receptor dimerization and activation, thereby initiating two downstream signaling cascades. Parallel signaling cascades: Activated EGFR stimulates the PLC-PKC pathway, representing a well-established receptor-level mechanism of desensitization. Simultaneously, EGFR transactivation engages the Src-PI3K-PDK1-Akt signaling axis, leading to activation of the deubiquitinase USP33. Arrestin-Gβγ nuclear translocation and functional consequences: Active USP33 deubiquitinates arrestin (Arr), promoting formation of a stable Arr-Gβγ complex. Bound by importin, this complex translocates into the nucleus through the nuclear pore complex. Nuclear accumulation of Gβγ is proposed to influence transcriptional regulation and gene expression. Concurrently, sequestration of Gβγ within the nucleus depletes the membrane-proximal Gβγ pool, limiting reassembly of the heterotrimeric G protein and thereby contributing to sustained Level II desensitization. 7-TM GPCR, 7-transmembrane G protein-coupled receptor; ADAM, a disintegrin and metalloproteinase; pro-HB-EGF, membrane-anchored heparin-binding EGF-like growth factor; sHB-EGF, soluble HB-EGF; EGFR, epidermal growth factor receptor; PLC, phospholipase C; PKC, protein kinase C; PI3K, phosphoinositide 3-kinase; PDK1, phosphoinositide-dependent kinase-1; Akt, protein kinase B; USP33, ubiquitin-specific-processing protease 33; Arr, arrestin; Gβγ, G protein βγ dimer; Gα, G protein α subunit.

Beyond receptor-proximal desensitization: EGFR transactivation as a hierarchical gate to Level II desensitization

Repeated or prolonged GPCR stimulation is proposed to engage an upstream checkpoint that determines whether signaling attenuation remains a transient receptor-proximal event or progresses into a persistent adaptive state. Within the Two-Level Framework, EGFR transactivation functions as this threshold-setting event that enables transition from Level I to Level II desensitization (Fig. 4).

EGFR transactivation can occur through ADAM-dependent shedding of EGF-family ligands, such as heparin-binding EGF-like growth factor (HB-EGF), and through Src-mediated receptor cross-talk (Daub et al., 1996; Luttrell et al., 1999). Unlike receptor-confined G protein signaling, EGFR activation generates a broader and more sustained signaling platform that engages multiple downstream kinase networks, including PI3K/Akt, PLC/PKC, and ERK pathways. Thus, EGFR transactivation is proposed to function not merely as an additional signaling intermediate but as a molecular checkpoint that converts sustained GPCR activation into a coordinated adaptive program.

Once this EGFR-dependent threshold is exceeded, two parallel downstream pathways are engaged. EGFR-dependent PKC activation reinforces receptor-level desensitization through heterologous receptor phosphorylation, whereas activation of the Src-PDK1-Akt-USP33 cascade promotes arrestin deubiquitination and initiates the distal remodeling events characteristic of Level II desensitization.

The efficiency of Level II entry is likely determined by both receptor-specific and cellular-context-dependent factors. Receptor architecture may influence the ability of individual GPCRs to engage EGFR transactivation, whereas the abundance and activity of EGFR, ADAM metalloproteases, Src family kinases, and downstream signaling components may determine whether the signaling threshold required for Level II activation is achieved. This framework provides a mechanistic explanation for why receptor activation alone does not uniformly predict persistent desensitization.

Accordingly, receptors with efficient coupling to EGFR-dependent signaling, such as D₃R or engineered D₂R variants with enhanced EGFR engagement, may exhibit greater propensity to enter Level II desensitization, whereas receptors with limited EGFR coupling may remain predominantly regulated through the classical GRK/arrestin-dependent Level I pathway.

Src-PDK1-Akt signaling couples EGFR activation to USP33-mediated arrestin remodeling

Following EGFR threshold activation, intracellular kinase signaling converts sustained receptor stimulation into the molecular remodeling events characteristic of Level II desensitization.

Src activation downstream of both GPCR stimulation and EGFR signaling promotes propagation toward the PI3K pathway (Luttrell et al., 1997). PDK1 subsequently activates Akt, and inhibition of this pathway prevents arrestin3 deubiquitination and suppresses progression into the USP33-dependent adaptive state (Alessi et al., 1997; Min et al., 2023). Together, these findings place the Src-PDK1-Akt cascade upstream of USP33-mediated arrestin remodeling.

A potential mechanism linking Akt signaling to USP33 is suggested by the presence of a candidate Akt consensus phosphorylation motif within USP33 (¹⁴⁹KARGLT¹⁵⁴) (Blom et al., 2004; Xu et al., 2015). Functional studies indicate that Thr154 is required for USP33-dependent arrestin3 deubiquitination and adaptive desensitization responses (Min et al., 2023; Wang et al., 2021). Although direct phosphorylation of USP33 by Akt remains to be demonstrated, these observations support a model in which Akt regulates USP33 activity either directly or through phosphorylation-dependent intermediates.

Because this signaling relay requires sequential activation of multiple kinase steps, transient or weak receptor stimulation may fail to generate sufficient signaling output to activate USP33-dependent remodeling. In contrast, sustained or repeated receptor activation allows temporal accumulation and amplification of signaling through the EGFR-Src-PDK1-Akt axis, thereby enabling the arrestin remodeling program that defines Level II desensitization.

USP33-mediated arrestin deubiquitination promotes distal remodeling

Activation of USP33 removes ubiquitin from arrestin3, thereby reversing the basal ubiquitinated state maintained by Mdm2 (Shenoy et al., 2009). Within the proposed Two-Level Framework, arrestin deubiquitination should therefore be viewed not as the upstream trigger for Level II entry, but as the molecular switch that executes the distal remodeling program after EGFR-dependent signaling has been engaged.

A central feature of this model is that the ubiquitination state of arrestin is proposed to influence its preferred molecular binding partner. Under basal conditions, ubiquitinated arrestin preferentially associates with receptor-associated signaling complexes that preserve Gβγ signaling competence. In contrast, transition from the ubiquitinated to the deubiquitinated state is accompanied by enhanced association of arrestin with free Gβγ dimers in experimental systems including D₃R and β₂AR models (Min et al., 2023; Zheng et al., 2020). This state transition is therefore proposed to redirect arrestin from receptor-associated signaling complexes toward formation of arrestin-Gβγ complexes, thereby initiating distal remodeling of intracellular signaling components.

One potential consequence of this remodeling is redistribution of arrestin-Gβγ complexes away from membrane-proximal signaling assemblies, thereby reducing the intracellular availability of signaling-competent Gβγ required for efficient heterotrimeric G protein activation during subsequent receptor stimulation. Although it remains to be established whether Gβγ redistribution alone is sufficient to account for persistent desensitization, this mechanism provides a plausible explanation for how arrestin remodeling could generate adaptive changes that persist beyond receptor dephosphorylation and recycling.

Thus, USP33-mediated arrestin deubiquitination represents a critical execution step linking EGFR-dependent signaling to intracellular remodeling beyond classical receptor uncoupling. Rather than simply reinforcing receptor-proximal desensitization, this ubiquitination-state transition is proposed to redirect arrestin toward a distal adaptive program that may convert transient receptor regulation into the persistent loss of signaling competence characteristic of Level II desensitization.

Spatial irreversibility and hybrid execution of Level II desensitization

Following USP33-mediated deubiquitination, arrestin-Gβγ complexes undergo importin β1-dependent nuclear redistribution (Min et al., 2023; Zhang et al., 2020). This spatial remodeling provides a potential explanation for why signaling recovery may remain incomplete even after receptor dephosphorylation and recycling. Although receptor competence can be restored at the plasma membrane, redistribution of intracellular signaling components may delay re-establishment of signaling-competent G protein assemblies required for efficient receptor activation.

Importantly, this mechanism illustrates the hybrid nature of Level II desensitization. Initiation remains receptor-specific through homologous GPCR activation and EGFR transactivation, whereas execution proceeds through two complementary EGFR-dependent downstream arms. One arm activates PKC to reinforce receptor-proximal desensitization through heterologous receptor phosphorylation, whereas the second engages the Src-PDK1-Akt-USP33 pathway to promote arrestin deubiquitination and intracellular remodeling. Because this distal remodeling arm acts through redistribution of signaling components rather than receptor-specific uncoupling, it has the potential to influence signaling by multiple GPCRs within the same cellular environment. However, direct evidence for such cross-receptor effects remains limited.

Beyond desensitization, nuclear redistribution of arrestin-Gβγ complexes may also contribute to longer-term cellular adaptation. Nuclear Gβγ subunits have been implicated in transcriptional regulation through interactions with RNA polymerase II, AP-1-containing transcriptional complexes, and chromatin-remodeling machinery (Bhatnagar et al., 2013), whereas nuclear arrestins can regulate transcription through interactions with chromatin-modifying enzymes and transcriptional regulators (Kang et al., 2005; Wang et al., 2003a). Whether these nuclear functions directly contribute to the establishment or maintenance of Level II desensitization remains unresolved, but they provide a plausible mechanistic link between persistent signaling attenuation and long-term adaptive transcriptional responses (Table 1).

Expanding biased agonism to include desensitization bias

Biased agonism has emerged as a central concept in GPCR pharmacology, with most studies focusing on differential engagement of receptor-proximal signaling pathways, particularly G protein versus arrestin coupling (Kenakin and Christopoulos, 2013; Ma et al., 2025; Rajagopal et al., 2010; Violin and Lefkowitz, 2007). The proposed Two-Level Framework suggests that functional selectivity may extend beyond acute signaling outputs to include differential engagement of adaptive mechanisms that govern the persistence of receptor desensitization.

Within this framework, receptor-proximal arrestin recruitment and persistent desensitization are not necessarily mechanistically coupled, indicating that receptor-proximal arrestin efficacy alone is insufficient to define the long-term adaptive consequences of receptor activation. Accordingly, ligands that efficiently recruit arrestins may nevertheless produce only limited persistent desensitization if they fail to engage the coordinated Level II adaptation program. Conversely, ligands with relatively modest receptor-proximal arrestin efficacy may still induce persistent desensitization if they efficiently engage this receptor-distal adaptive program. Thus, conventional measures of signaling bias may not fully predict either the magnitude or the persistence of receptor desensitization.

I propose that ligand-dependent engagement of persistent desensitization mechanisms may constitute an additional dimension of functional selectivity, herein referred to as "desensitization bias." Whereas classical signaling bias describes preferential activation of receptor-proximal signaling pathways, desensitization bias would describe the propensity of a ligand to engage the coordinated adaptive processes that determine the magnitude and persistence of receptor desensitization. This distinction may help explain why ligands exhibiting similar acute signaling profiles can nevertheless display markedly different tolerance liabilities or long-term pharmacological effects. If validated experimentally, desensitization bias would expand the concept of biased agonism beyond selective pathway activation to include selective engagement of adaptive programs that govern the persistence of receptor desensitization.

The opioid field provides a potentially relevant example. G protein-biased agonists such as Oliceridine, PZM21, and SR-17018 were developed with the expectation that reduced receptor-proximal arrestin recruitment would preserve analgesic efficacy while minimizing receptor desensitization and associated adverse effects (DeWire et al., 2013; Manglik et al., 2016; Schmid et al., 2017). However, both preclinical and clinical studies suggest that reduced receptor-proximal arrestin recruitment does not consistently translate into proportional reductions in tolerance development (Gillis et al., 2020; Kliewer et al., 2020). Although opioid tolerance is undoubtedly multifactorial, these observations suggest that receptor-proximal signaling bias alone may be insufficient to account for long-term pharmacological adaptation and raise the possibility that receptor-distal desensitization mechanisms contribute independently to tolerance.

Viewed from this perspective, the Two-Level Framework suggests that future evaluation of biased ligands may benefit from considering not only acute signaling efficacy but also the molecular processes governing desensitization, recovery, and adaptive remodeling. If the EGFR-dependent Level II adaptation program proves to be broadly conserved across GPCR families, pharmacological uncoupling of acute signaling efficacy from persistent desensitization could represent an additional dimension of functional selectivity. Accordingly, ligands may differ not only in their receptor-proximal signaling efficacy (Level I), but also in their capacity to engage the coordinated Level II adaptation program that governs persistent desensitization. Incorporating this additional pharmacological dimension into the evaluation of biased ligands may improve prediction of long-term efficacy, tolerance liability, and therapeutic durability beyond conventional measures of signaling bias.

In summary, incorporating desensitization bias into the framework of biased agonism provides a novel conceptual dimension, wherein ligand-dependent engagement of adaptive remodeling dictates not only acute signaling efficacy but also the persistence and durability of receptor desensitization.

Tolerance, transcriptional adaptation, and the desensitization continuum

The proposed Two-Level Framework raises the possibility that receptor desensitization and cellular adaptation exist along a mechanistic continuum rather than representing entirely distinct processes. Whereas the Level I program primarily attenuates receptor signaling at the plasma membrane, the proposed Level II program extends this process beyond receptor-proximal regulation by linking persistent desensitization to intracellular remodeling. In this way, Level II provides a potential mechanistic bridge between acute signal attenuation and longer-term cellular adaptation.

A key feature of the proposed Level II program is USP33-dependent arrestin remodeling, which promotes formation and importin β1-dependent nuclear redistribution of arrestin-Gβγ complexes (Min et al., 2023; Zhang et al., 2020). Redistribution of these complexes may alter the intracellular availability of signaling-competent Gβγ pools, providing a potential explanation for incomplete signaling recovery despite receptor dephosphorylation and recycling. Although direct evidence that this redistribution mediates persistent desensitization remains lacking, it provides a plausible mechanism linking intracellular remodeling to persistent attenuation of receptor responsiveness.

Importantly, the molecular events associated with this remodeling process have also been implicated in nuclear signaling. Nuclear Gβγ subunits interact with RNA polymerase II, AP-1-containing transcriptional complexes, and chromatin-remodeling machinery (Bhatnagar et al., 2013; Saini et al., 2007), whereas arrestins can regulate transcription through interactions with multiple nuclear regulatory proteins. These observations raise the possibility that molecular components involved in persistent desensitization may also participate in transcriptional adaptation. Whether these nuclear functions contribute directly to the establishment or maintenance of Level II desensitization remains an important unresolved question.

Thus, persistent desensitization may represent a pivotal intermediate state within a broader adaptive continuum, bridging acute receptor-proximal uncoupling with spatial intracellular remodeling and longer-term transcriptional adaptation.

Conclusions and future perspectives

GPCR desensitization has traditionally been viewed as a receptor-centered process governed by receptor phosphorylation, arrestin recruitment, receptor trafficking, and heterologous regulatory inputs. Although these mechanisms provide essential control of acute signaling attenuation, they do not fully explain the diversity, persistence, or kinetic heterogeneity of adaptive responses observed across GPCR systems. In this Review, I propose the Two-Level Framework, in which persistent desensitization emerges through an additional hierarchical regulatory layer initiated by sustained receptor activation and EGFR-dependent signaling.

Within this framework, Level I desensitization represents the classical receptor-proximal response mediated by GRK/arrestin-dependent uncoupling, phosphorylation-dependent regulation, and receptor trafficking. In contrast, Level II represents an EGFR-triggered hierarchical adaptation program that extends beyond receptor regulation by linking receptor-level attenuation with intracellular signaling remodeling. Through this hierarchical organization, sustained receptor activation may generate persistent changes in signaling competence that outlast receptor-proximal regulatory events, thereby providing a mechanistic basis for prolonged receptor adaptation.

Importantly, this framework does not imply that receptor-proximal mechanisms are insufficient for desensitization or that all GPCRs engage Level II regulation. Rather, it proposes an additional regulatory layer that may help explain why closely related receptors and ligands exhibit distinct desensitization kinetics despite similar patterns of receptor activation. By integrating receptor regulation with EGFR-dependent remodeling of intracellular signaling capacity, the Two-Level Framework provides a conceptual basis for understanding how transient GPCR activation may evolve into a more durable adaptive state in which receptor recovery and restoration of signaling competence become partially uncoupled.

This perspective also unifies several concepts that have traditionally been considered separately. Within the proposed desensitization continuum, persistent desensitization may represent an intermediate adaptive state linking receptor-proximal regulation with intracellular remodeling and longer-term cellular adaptation. In this context, tolerance may emerge as the cumulative consequence of progressively engaged adaptive processes rather than as a single discrete event. Likewise, ligand-dependent differences in the engagement of these adaptive mechanisms raise the possibility that desensitization bias may constitute an additional dimension of functional selectivity, in which ligands differ not only in acute pathway activation but also in their capacity to engage mechanisms that determine the magnitude and persistence of receptor desensitization.

Although many aspects of this model remain to be experimentally validated across diverse GPCR families and physiological contexts, the Two-Level Framework provides a conceptual foundation for integrating receptor regulation, persistent desensitization, adaptive remodeling, tolerance, and biased agonism within a unified mechanistic framework. Future studies examining receptor specificity, EGFR pathway engagement, arrestin post-translational regulation, nuclear signaling functions, the reversibility of intracellular signaling redistribution, and ligand-dependent differences in Level II engagement will be essential for determining the generality and physiological significance of this mechanism.

Collectively, the Two-Level Framework suggests that GPCR desensitization should be viewed not simply as a consequence of receptor shutdown, but as a hierarchical adaptive process that links receptor-proximal regulation to persistent intracellular remodeling. Recognizing this additional level of regulation may improve our understanding of long-term receptor adaptation, tolerance development, and functional selectivity. More broadly, it integrates the proposed concepts of the desensitization continuum and desensitization bias into a unified view of GPCR adaptation, thereby opening new opportunities to predict and pharmacologically modulate persistent receptor responsiveness, long-term therapeutic efficacy, and tolerance liability.