Viral receptors as host–pathogen control nodes linking entry, signalling, and system-level pathogenesis

Abstract

Viral receptors have traditionally been viewed as docking sites that mediate viral attachment and entry. However, this perspective no longer captures the full complexity of receptor function during infection. Receptor engagement links viral binding to intracellular trafficking, signal transduction, barrier integrity, and ultimately disease outcome. Rather than acting solely as entry determinants, receptors influence endocytic routing, cellular permissiveness, tissue tropism, immune responses, and systemic dissemination. A functional distinction between attachment factors, entry receptors, co-receptors or accessory factors, and receptor-like immune sensors is essential for accurate mechanistic interpretation and translational applications. This review examines the events that follow receptor engagement, including intracellular routing, signalling, regulation of receptor availability, and host determinants of susceptibility. Particular emphasis is placed on the concept of receptor landscapes, defined as dynamic and spatially organized receptor topologies across epithelial and endothelial barriers that are shaped by development, ageing, inflammation, and comorbidities. In this view, systemic disease is viewed as the consequence of failures in receptor-defined barriers and interconnected feedback circuits across organs. Finally, we discuss the implications of this perspective for host-directed therapies, receptor decoys, and biomarker development. Reframing receptors as regulatory hubs helps explain differences in viral tropism and disease severity and highlights new opportunities for host-targeted intervention.

Viral receptors are key determinants of infection, linking viral entry to cellular responses, tissue damage, and the development of local and systemic disease.

Introduction: from docking sites to control nodes

The conceptual limitation of the entry-centric view

The receptor concept entered virology as a useful simplification: identify what the virion binds, and one has explained susceptibility. This logic remains useful, but it is incomplete in several biologically important ways. First, entry is not a single step but a sequence of state transitions in which receptor engagement is continuously interpreted by the cell. Binding can recruit endocytic machinery, partition virus–receptor complexes into specific membrane microdomains, and set the trafficking itinerary that determines where fusion and uncoating occur, with immediate consequences for innate detection and replication efficiency (Pietiäinen et al. 2005, Mercer and Helenius 2009, Mercer et al. 2010, Grove and Marsh 2011). Second, a receptor is rarely an isolated molecule. It sits in a physical and functional neighbourhood: glycocalyx, heparan sulfate proteoglycans, lipid rafts, tetraspanin-enriched microdomains, integrin platforms, proteases, and adaptor proteins. These associations can amplify avidity, prime fusion proteins, and coordinate signalling, such that the “receptor” is more realistically a modular entry-and-signalling apparatus rather than a single viral binding site (Pietiäinen et al. 2005, Mercer et al. 2010, Hantak et al. 2019).

Third, the receptor’s biology is not neutral. Many viral receptors are themselves regulatory proteins: enzymes, adhesion molecules, transporters, immunomodulatory lectins. Their engagement can trigger kinase cascades, cytoskeletal remodelling, and transcriptional programmes that feed back on antiviral immunity and tissue integrity (Mercer et al. 2010, Takeuchi and Akira 2010, Grove and Marsh 2011). This helps explain why it is unsurprising that the same virus can produce divergent phenotypes across hosts or tissues with nominally “similar” receptor expression. Expression alone is an unreliable proxy for functional availability, localization, or coupling to relevant endocytic and signalling circuits (Mercer et al. 2010, Grove and Marsh 2011).

The recent coronavirus literature made this visible to a broader community: ACE2 is not simply a viral handle; its tissue distribution, membrane context, and coupling to proteases and attachment factors help define where infection begins and how it propagates, while receptor-adjacent biology intersects with barrier damage and systemic complications (Hamming et al. 2004, Clausen et al. 2020, Hoffmann et al. 2020a, Hou et al. 2020, Sungnak et al. 2020). The point is not to privilege one virus, but to underline a general principle: receptors do not merely permit entry, they shape the entire early decision tree of infection.

Receptors as integrated determinants of cellular susceptibility, tissue tropism, immune trajectory, and clinical phenotype

Cellular susceptibility is commonly reduced to “receptor present or absent”. Yet permissiveness is better conceived as the intersection of (i) receptor binding and avidity, (ii) fusion/endocytosis licensing, (iii) productive intracellular routing, and (iv) the cell’s capacity to tolerate and respond to invasion. Receptor engagement can favour clathrin-mediated endocytosis, macropinocytosis, caveolin-dependent/raft-associated routes, or surface fusion, each exposing the virion to distinct pH environments, proteases, restriction factors, and pattern-recognition receptors (Pietiäinen et al. 2005, Mercer and Helenius 2009, Mercer et al. 2010, Takeuchi and Akira 2010). A receptor that routes a virus into compartments rich in endosomal sensors may convert the same binding event into an abortive, interferon-dominated outcome; conversely, routing that delays sensing may grant the virus a kinetic advantage (Mercer et al. 2010, Takeuchi and Akira 2010).

Tissue tropism is similarly multi-dimensional. Receptor distribution is spatially structured across barriers: nasal versus distal airway epithelium, villus versus crypt, apical versus basolateral membranes, endothelial beds with distinct glycocalyces. Single-cell atlases and spatial profiling have shown that entry factors cluster in specific epithelial subsets and are co-expressed with innate immune modules, implying that early tropism and early immune response profile are entangled properties rather than independent variables (Sungnak et al. 2020). Moreover, tissue tropism is modified by context: inflammation, smoking, metabolic disease, chronic lung remodelling, and drug exposures can alter receptor abundance, glycosylation, shedding, and membrane organization, effectively reshaping receptor availability and spatial organization (Pietiäinen et al. 2005, Mercer and Helenius 2009, Grove and Marsh 2011).

Immune responses depend on which cells are infected, which cells are merely engaged, and which sensors are triggered. Pattern-recognition receptors form a parallel layer of host sensing that translates entry-associated cues into cytokine programmes, inflammasome activation, and adaptive priming (Takeuchi and Akira 2010). Because entry routing and microdomain localization influence sensor exposure, receptor usage becomes an early determinant of the inflammatory response, not merely a gateway to replication (Mercer et al. 2010, Takeuchi and Akira 2010).

Clinical phenotype follows from the above, but not linearly. Localized replication may remain clinically mild if barrier responses contain spread, whereas receptor-enabled access to endothelium, immune cells, or neuro-invasive routes may convert similar inocula into systemic disease. Multi-organ disease can therefore arise when receptor-dependent access to different tissues coincides with inflammation, vascular dysfunction, and impaired barrier control.

A caveat is necessary here. Receptor complexes govern attachment and entry, but they do not by themselves guarantee productive infection. Once inside the cell, the virus must still complete genome replication, direct viral gene expression, assemble new particles, and achieve egress, and each of these steps depends on host factors that are largely independent of the receptor used for entry. A cell can therefore express a fully competent receptor and still fail to support infection if any of these downstream requirements is not met, just as a cell lacking the canonical receptor can sometimes be infected through alternative routes. Tropism is best understood as the product of receptor-dependent access and receptor-independent intracellular compatibility, not as a property of receptor expression alone (Schneider-Schaulies 2000, McFadden et al. 2009).

From receptor engagement to disease pathogenesis

This review is organized around a central premise: viral receptors are host–pathogen control nodes. They connect viral evolution with host barrier biology, membrane organization, and immune sensing. Receptor engagement selects entry routes, dictates intracellular trafficking, and initiates signalling and barrier responses that can either confine infection or enable dissemination (Pietiäinen et al. 2005, Mercer and Helenius 2009, Mercer et al. 2010, Takeuchi and Akira 2010, Grove and Marsh 2011). Because receptor availability changes with comorbidity, age, inflammation, and drug exposure, receptor usage also helps explain epidemiological and clinical variability.

The viruses and virus–receptor systems discussed throughout this review were selected as representative examples illustrating the major conceptual and mechanistic principles addressed, rather than as an exhaustive survey of all known virus–receptor interactions.

Accordingly, we will: (i) redefine receptor “taxonomy” in functional terms (attachment factors versus entry receptors versus co-receptors/accessory factors versus sensors), (ii) dissect mechanistic layers downstream of contact (routing, signalling, turnover, and genetic/epigenetic determinants), (iii) develop the concept of receptor landscapes across barriers and states, (iv) connect local infection to system-level pathogenesis via receptor-defined amplifiers and bottlenecks, (v) integrate receptor plasticity with viral evolution and spillover risk, and (vi) identify translational opportunities for host-directed therapy and biomarker development.

The payoff of this reframing is practical: it improves experimental inference (by forcing attention to localization, topology, and context), sharpens risk prediction (by separating compatibility from access), and supports therapeutic approaches beyond binding blockade towards rerouting, damping, and decoy strategies that preserve receptor physiology (Fig. 1).

Figure 1.

Viral entry pathways, host signaling activation, and barrier disruption across biological scales. Viruses utilize multiple entry mechanisms into host cells, including membrane fusion, macropinocytosis, and clathrin-mediated endocytosis, depending on viral and cellular context. Engagement of surface receptors initiates intracellular signaling hubs that coordinate multiple cellular responses: antiviral pathways characterized by interferon (IFN) production and establishment of an antiviral state, alongside pro-inflammatory and pro-viral signaling cascades such as PI3K and MAPK activation. Concurrently, viral infection perturbs epithelial and endothelial barrier integrity through tight junction disruption, increased permeability, and endothelial activation. These processes promote the transition from cellular injury to tissue dysfunction and ultimately systemic dissemination, contributing to disease pathogenesis.

Receptor taxonomy redefined: a functional taxonomy of viral receptors

A persistent source of confusion in virology is lexical rather than experimental: the word receptor is used to cover at least four separable functions. In the most benign form, this produces sloppy mechanistic shorthand. In the more damaging form, it drives category errors: an attachment factor is mistaken for an entry-licensing factor; a signalling hub is treated as if it were a gatekeeper; an innate sensor is rhetorically promoted to a “receptor” in the entry sense, obscuring what the perturbation changes. A functional classification is therefore not a matter of terminology alone but is essential for accurate interpretation.

The most economical starting point is operational. A host factor qualifies as an entry receptor only if its engagement is necessary to licence productive entry under physiological conditions. Determinants that merely amplify surface capture, increase avidity, or shape lateral mobility should be treated as attachment factors, even when their blockade reduces infection in vitro. Likewise, factors that are indispensable for entry but act by priming, trafficking, or complex formation are best classified as co-receptors or accessory entry factors rather than “receptors” in the narrow sense. Finally, receptor-like immune sensors must be considered as a distinct functional category: they do not licence entry, but they strongly condition downstream trajectory and clinical phenotype.

Attachment factors versus entry receptors

Attachment factors are catalysts that amplify and mediate virus/receptor interaction. They raise the effective local concentration of virions at the cell surface, facilitate lateral diffusion and clustering, and bias the probability that an entry receptor (or a receptor complex) will be engaged with a productive geometry. They commonly include glycans, proteoglycans, and lectin-like capture systems. Their defining feature is that they facilitate infection without directly determining productive entry: they accelerate encounters, but do not by themselves commit the virus to a productive entry route.

Heparan sulfate proteoglycans (HSPGs) illustrate both their power and the interpretive traps they invite. HSPGs can function as genuine attachment factors for numerous viruses, increasing binding and in some cases enabling engagement of a more specific entry receptor. This dependence is not uniform across respiratory viruses. Comparative studies have shown that glycosaminoglycan interactions may differ substantially even among human coronaviruses, suggesting that attachment-factor usage is highly context dependent and cannot be generalized across related viral taxa (Fuochi et al. 2023). Yet HS dependence can also be an artefact of cell-culture adaptation, where selection favours variants that bind abundant anionic glycans rather than the physiologically relevant drivers of tropism. Any taxonomy that collapses HS binding into “the receptor” therefore risks mistaking an in vitro convenience for a physiologically relevant determinant of infection. The practical implication is stark: HS-binding should be treated as attachment unless stringent genetic and physiological criteria establish it as an entry-licensing requirement (Cagno et al. 2019 , Liu et al. 2021). Experimental evidence obtained using SARS-CoV-2 infection models further supports the functional importance of glycosaminoglycan-mediated attachment, showing that exogenous heparan sulfate and enoxaparin can significantly impair viral entry despite not acting as canonical entry receptors (Fuochi et al. 2025b).

Sialylated glycans represent a parallel case where attachment and entry are coupled but not identical. Influenza A virus attachment is mediated by haemagglutinin recognition of sialic acid linkages, with α2,6 versus α2,3 preferences contributing to host range, tissue binding patterns, and interspecies transmission constraints. However, the initial glycan engagement does not exhaust the entry problem: avidity, multivalency, and the density/topology of sialoglycans determine whether binding becomes productive uptake and whether subsequent routing exposes the virion to permissive proteolysis or to hostile sensing compartments. Treating sialic acids as “the receptor” is often acceptable shorthand, but accurate classification requires recognizing that they operate primarily as attachment factors whose pattern of linkage, presentation, and microanatomical distribution acts upstream of route choice and immune exposure (Pekosz et al. 2009 , Imai and Kawaoka 2012).

Entry receptors, in contrast, are licence-granting factors. Their engagement commits the virus to fusion and/or a defined internalization programme, frequently by triggering conformational changes in viral fusion machinery, recruiting endocytic adaptors, or delivering the complex into the correct intracellular itinerary. Importantly, an entry receptor can be located at the plasma membrane or within an intracellular compartment. The latter is not an edge case but a useful stress test: it forces acknowledgement that “receptor” is a functional role within an entry pathway, not a synonym for a surface molecule.

Co-receptors and accessory factors

Co-receptors and accessory entry factors perform functions that are distinct from both attachment and primary receptor engagement: the virus can bind an entry receptor, yet entry fails or becomes inefficient without additional factors that prime, cluster, traffic, or stabilize the entry complex. The HIV-1 entry system is the canonical demonstration of why such distinctions matter. CD4 engagement is necessary but insufficient; productive entry requires chemokine receptor co-receptors, classically CCR5 for macrophage-tropic strains and CXCR4 for T-cell line-tropic strains. The identification of CCR5 as a major co-receptor, alongside the cloning of “fusin” (CXCR4), not only clarified cellular tropism but also exposed how co-receptor usage predicts transmission patterns and pathogenesis. A taxonomy that calls CCR5 the receptor without qualification erases the logical structure of entry: CD4 is the primary receptor; CCR5/CXCR4 are co-receptors that complete the licensing machinery and shape signalling and membrane fusion competence (Deng et al. 1996, Dragic et al. 1996, Feng et al. 1996).

Accessory factors often operate through proteolytic priming, trafficking constraints, or microdomain arrangement. Proteases can gate entry routes by determining whether fusion occurs at the surface or after endosomal maturation, thereby changing exposure to innate sensors and restriction factors. Likewise, integrins are frequently used as receptors or co-factors across diverse viruses, sometimes as bona fide entry receptors, sometimes as signalling and internalization partners that increase efficiency rather than absolute permissiveness. Their importance lies not only in binding but in their ability to organize cytoskeletal and kinase circuits that make entry mechanically and temporally plausible (Stewart and Nemerow 2007, Campadelli-Fiume et al. 2016).

A particularly instructive example of accessory factor logic is the filovirus requirement for Niemann–Pick C1 (NPC1), an intracellular receptor engaged within endosomal compartments after host priming of the viral glycoprotein. Here, receptor engagement is inseparable from routing and proteolysis: the entry pathway is staged such that binding competence is revealed only after glycan-cap removal, and receptor recognition occurs precisely where fusion must be executed. This arrangement makes plain that receptor identification is not the end of the entry story. It highlights the importance of intracellular trafficking and compartment-specific entry pathways (Miller et al. 2012, Bornholdt et al. 2016).

Box 1. Parvovirus B19: why receptor identification alone does not explain viral tropism

Parvovirus B19 (B19V) represents a useful example of the complexity underlying virus–host interactions (Qiu et al. 2017). Although the virus displays a remarkably restricted tropism for erythroid progenitor cells, the molecular basis of this selectivity cannot be explained by a single receptor. Instead, productive infection depends on a sequence of events involving attachment, internalization, intracellular trafficking, and cell-specific permissiveness (Zakrzewska et al. 2023).

A first clue is given by the interaction of the capsid shell with the neutral glycolipid globoside. Globoside is well represented on the erythroid series and was first identified as a receptor for B19V by means of binding and competition assays (Brown et al. 1993). Further, it is part of the P blood group, and its absence as in the rare p- phenotype was shown to confer resistance to infection, in the population as well as in in vitro infection experiments (Brown et al. 1994). The tissue distribution of globoside is in good accordance with the known tropism of B19V, and initially it seemed to account for its selectivity. Subsequent research indicated a more complex scenario and a shifting paradigm. Now, at least three players come into play.

Globoside is not considered anymore the primary receptor, but a docking moiety requiring mild acidic conditions for its interaction with viral capsid. Strikingly, its absolute requirement for a productive infection has been confirmed at more than thirty years from the initial report, but its role is now linked more to directing intracellular trafficking rather than favouring binding (Bieri and Ros 2019). A model of B19V—globoside interaction has been proposed on the airway epithelia, allowing for transcytosis as a first step of infection (Suter et al. 2023), conducting virus into the bloodstream and then to target cells in the bone marrow.

The presence of coreceptor(s) for B19V on target erythroid progenitor cells was very soon hypothesized; integrins and/or Ku80 received supporting evidence for such a role. A key finding was that the VP1u region of the capsid binds a moiety only present on target cells (a tentative VP1u receptor, VP1uR), thus directing B19V tropism to a selected cell population able to support its replication (Leisi et al. 2016). Identification of VP1uR is still elusive; the AXL transporter has received supporting evidence for a role in B19V internalization (Ning et al. 2023), but its wide distribution does not match the restricted one of the VP1u binder. Another candidate receptor proposed to play a key role is hTFR (CD71), a marker of erythroid differentiation, whose cellular distribution closely matches the tropism of B19V; hTFR can selectively bind VP1u, allowing viral internalization leading to a productive replicative cycle (McFarlin et al. 2026). In vitro infection experiments suggest that B19V can be internalized in cells independently of the VP1uR, but only the VP1u-VP1uR interactions allows for onset of a productive replicative cycle (Bua et al. 2023, Fasano et al. 2025). Clearly, complex mechanisms involving many molecular partners and different pathways determine the outcome of infection; in vivo, this restricts B19V replication in the erythroid compartment and determines clinical outcomes.

Finally, ensuing a productive infection in the bone marrow, a high titre viremia contributes to diffusion of the virus to peripheral tissues. Endothelial and connective cells can be infected, although not productively, while the trigger of inflammatory responses is common and of clinical relevance. In this instance, a role has been assigned to the immune response, with production of specific antibodies that may favour viral internalization, such as in endothelial cells, by means of complement C1q receptor (von Kietzell et al. 2014). In B-lymphocytes in tonsillar tissues, on the other hand internalization is mediated by IgG and can lead to viral persistence, a possible site for archival of genomes and viral reactivation (Pyoria et al. 2017).

Immune sensors as “receptor-like” control elements

Host defence systems frequently perform functions analogous to those of receptors: they detect viral nucleic acids or danger cues and convert them into interferon programmes, inflammatory cascades, and antigen-presentation trajectories. Yet conflating these sensors with entry receptors produces predictable analytical failures. If the perturbation of a sensor changes disease severity without meaningfully altering entry, calling the sensor a “viral receptor” in the entry sense distorts both biological interpretation and translation. These observations suggest that sensing represents a parallel host response pathway that is often coupled to entry routing and receptor microdomain organization, but not identical to entry licensing.

Pattern-recognition receptors exemplify this distinction: they do not grant access, but they can impose an early kinetic penalty on replication, influence the nature and magnitude of inflammatory responses, and thereby shape tissue damage and systemic symptoms. In this context, immune sensors are best handled as receptor-like elements whose activation probability depends on the entry pathway selected upstream. This preserves causal clarity: receptor usage influences the entry pathway, the entry pathway determines sensor exposure, and sensor activation contributes to disease phenotype (Kawai and Akira 2010, Takeuchi and Akira 2010).

Receptor complexes and microdomains: lipid rafts, tetraspanins, and integrin platforms

Real membranes are not well-mixed soups. Viral entry therefore occurs within the spatial organization of membrane microdomains, where receptors, co-receptors, proteases, and adaptors cluster into semi-stable platforms. Lipid rafts, despite historical controversies about operational definitions, remain a useful organizing concept: cholesterol- and sphingolipid-enriched assemblies can concentrate signalling receptors and create nanoscale environments that bias both binding geometry and downstream signalling. For viruses, this implies that the functional “receptor” may be the platform, not the molecule: disrupting microdomain organization can phenocopy receptor blockade without altering receptor expression (Simons and Ikonen 1997, Simons and Gerl 2010).

Tetraspanin-enriched microdomains provide a more protein-centric parallel. Tetraspanins assemble lateral networks that bring together receptors, co-factors, and cytoskeletal links, creating entry and egress platforms for multiple virus families. Conceptually, tetraspanins convert a list of “required factors” into a highly organized membrane platform: co-localization becomes a controllable variable, and entry efficiency becomes a function of platform integrity. This is one reason why single-factor overexpression experiments can mislead: the cell may express the receptor, but not the platform that makes it operational in the relevant compartment (Florin and Lang 2018, New et al. 2021).

Integrin platforms sit at the intersection of adhesion biology and entry. Their value to viruses is not limited to binding motifs; integrins provide a ready-made internalization and signalling apparatus, linking extracellular engagement to actin remodelling and kinase cascades that facilitate uptake, trafficking, and in some contexts, barrier traversal. As with rafts and tetraspanins, the key move is to stop treating receptors as isolated molecules and start treating them as components of spatially organized regulatory platforms.

Biochemical diversity of receptor complexes: glycolipids, glycans, and membrane lipids

Receptor complexes are not restricted to transmembrane proteins. Several viruses achieve productive engagement through non-proteinaceous moieties, showing that the functional receptor concept spans distinct biochemical classes rather than a single molecular type. SV40 attaches through the ganglioside GM1, a glycosphingolipid displayed on the plasma membrane rather than a protein receptor, and this lipid-based recognition is sufficient to license entry and intracellular routing (Tsai et al. 2003). C-type lectins such as DC-SIGN recognize glycan structures displayed on viral envelope glycoproteins rather than binding a protein epitope directly, acting as attachment and, in some cases, entry-competent receptors for multiple unrelated virus families (van Kooyk and Geijtenbeek 2003). Flaviviruses add a further layer: phosphatidylserine exposed on the viral membrane, rather than any virus-encoded protein, is recognized by TIM and TAM receptors through a mechanism termed apoptotic mimicry, extending receptor engagement to a lipid intrinsic to the virion envelope (Perera-Lecoin et al. 2013). These examples show that protein-protein, protein-glycan, and protein-lipid recognition modes coexist across virus families, and can even operate for the same virus at different stages of infection; receptor complexes must therefore be treated as biochemically heterogeneous rather than assumed to be a uniform class of transmembrane protein.

The proposed functional taxonomy of host determinants involved in viral infection is summarized in Fig. 2, while representative examples are reported in Table 1.

Figure 2.

Minimal nomenclature for host factors in viral entry and immune sensing: Host–virus interactions at the cell surface can be operationally partitioned into four functional classes. Attachment factors (AFs) increase virus-receptor binding and surface retention but are not sufficient to license entry, often facilitating subsequent engagement of specific receptors. Entry receptors (ERs) are necessary determinants whose engagement licenses productive entry, including membrane fusion and/or receptor-mediated endocytosis and trafficking. Co-receptors/accessory entry factors act in concert with ERs and are required to complete entry by enabling priming, multicomponent complex formation, spatial organization of entry complexes, or intracellular routing. In contrast, sensors (e.g. innate immune receptors) shape downstream immune trajectories without licensing entry per se.

Table 1.

Functional classes of viral receptor usage with canonical examples.

Functional class	Primary function	Typical determinants	Canonical examples	Key inference logic
Attachment factor (glycan)	Avidity amplification; surface capture	Sialoglycans	Influenza A → sialic acids	Binding/tropism correlate; entry still depends on routing and context (Pekosz et al. 2009, Imai and Kawaoka 2012)
Attachment factor (proteoglycan)	Concentration and lateral diffusion; co-engagement facilitation	HSPGs	Multiple viruses → heparan sulfate	Often enhances infection in vitro; may reflect adaptation bias (Cagno et al. 2019, Liu et al. 2021)
Entry receptor	Licences fusion/endocytosis; defines productive itinerary	Specific membrane proteins	SARS-CoV-2 → ACE2	KO/rescue logic; route coupling; tissue gates (Hoffmann et al. 2020b)
Co-receptor	Completes licensing and fusion competence	Chemokine receptors	HIV-1 → CCR5/CXCR4 (with CD4 primary)	Binding alone insufficient; co-receptor usage predicts tropism (Deng et al. 1996, Dragic et al. 1996, Feng et al. 1996)
Accessory entry factor (priming/trafficking)	Proteolytic activation; compartment gating	Proteases; trafficking factors	Filoviruses → NPC1 (endosomal)	Entry depends on staged priming and compartmental receptor engagement (Miller et al. 2012, Bornholdt et al. 2016)
Platform organizer	Spatial assembly of entry machinery	Rafts; tetraspanins; integrins	Multiple viruses → TEM/raft/integrin platforms	Disrupting organization phenocopies “receptor loss” without expression change (Simons and Ikonen 1997, Simons and Gerl 2010, Florin and Lang 2018, New et al. 2021)
Receptor-like sensor	Immune trajectory control	PRRs and signalling hubs	TLR/PRR systems	Alters cytokine/IFN set-points and pathology more than entry (Kawai and Akira 2010, Takeuchi and Akira 2010)

Mechanistic layers: what a receptor does after the virus touches it

Binding geometry, avidity, and multivalency: why affinity alone is an incomplete predictor

A persistent misconception in entry biology is that high affinity necessarily predicts high infectivity. In practice, viruses often behave as multivalent ligands that exploit avidity, rebinding kinetics, and nanoscale receptor clustering, such that effective interaction strength depends on receptor organization and interaction dynamics rather than a single equilibrium constant (Kiessling et al. 2006).

Multivalency is not merely quantitative amplification; it can be a qualitative selector of productive engagement by stabilising otherwise transient contacts and by favouring receptor microdomain partitioning and co-factor recruitment. Accordingly, interventions that disrupt clustering, lateral mobility, or local receptor density may collapse infectivity even when the canonical binding interface remains intact (Kiessling et al. 2006).

Entry routing: clathrin-mediated endocytosis, macropinocytosis, caveolin-dependent endocytosis, direct fusion

Following receptor engagement, the choice of entry pathway becomes a major determinant of infection outcome: internalization programmes impose distinct pH trajectories, proteolytic environments, trafficking itineraries, and exposure to innate sensing. The major cellular strategies exploited by animal viruses span clathrin-mediated endocytosis, macropinocytosis, caveolin-dependent/raft-associated uptake, and additional clathrin-independent routes (Doherty and McMahon 2009, Mercer and Helenius 2012, Cossart and Helenius 2014).

These routes also differ in their dependence on the GTPase dynamin: clathrin-mediated and caveolar entry both require dynamin-driven scission of the nascent vesicle, whereas macropinocytosis and several clathrin-independent carriers proceed without it, relying instead on actin-driven membrane deformation (Damke et al. 1994).

Clathrin-mediated endocytosis couples receptor–adaptor logic to comparatively stereotyped delivery into early endosomal compartments. Macropinocytosis, by contrast, is frequently driven by signalling-dependent actin remodelling and can accommodate large or irregular particles, making it a recurrent solution for diverse virus families (Doherty and McMahon 2009, Mercer and Helenius 2012). Caveolin-dependent and raft-associated routes, while mechanistically heterogeneous, highlight the importance of membrane microdomain organization and cholesterol-rich platforms as determinants of uptake and downstream signalling (Chazal and Gerlier 2003, Doherty and McMahon 2009, Lingwood and Simons 2010). For some enveloped viruses, direct fusion at the plasma membrane provides a more immediate route to genome release, but typically demands receptor-coupled conformational priming of viral fusion proteins and careful coordination with host proteolytic activation (Harrison 2015).

Current evidence argues against a strict either-or interpretation of so-called “non-canonical” or “exotic” entry pathways. Available data generally do not support the idea that viruses generate wholly new endocytic machinery. Instead, receptor engagement recruits, activates, and spatially reorganizes pre-existing host components, including kinases, adaptors, actin regulators, and membrane lipids. Nevertheless, the virion is not necessarily a passive cargo. Influenza A virus can nucleate de novo clathrin-coated pits at sites of attachment, vaccinia virus activates actin-dependent macropinocytosis through apoptotic mimicry, and multivalent SV40 binding to GM1 can directly induce membrane curvature and tubulation (Rust et al. 2004, Mercer and Helenius 2008, Ewers et al. 2010). Although these processes may appear virus-specific, they remain built from endogenous membrane-remodelling and trafficking modules. Many apparently non-canonical routes may therefore be more appropriately viewed as induced, hybrid, or incompletely characterized variants of cellular uptake pathways rather than as entirely virus-created mechanisms. Distinguishing between these possibilities requires direct visualization and orthogonal genetic perturbation of multiple pathway components, because morphology or sensitivity to a single inhibitor cannot establish a unique entry route (Cossart and Helenius 2014).

The entry routes are summarized in Fig. 3.

Figure 3.

Representative pathways for viral entry into host cells. Following receptor engagement, viruses can enter host cells through distinct uptake mechanisms depending on both viral and host determinants. (A) Membrane fusion: enveloped viruses fuse directly with the plasma membrane, releasing the viral nucleocapsid into the cytoplasm. (B) Macropinocytosis: signalling-induced actin remodelling generates large membrane ruffles that engulf virions into macropinosomes before intracellular trafficking. (C) Clathrin-mediated endocytosis: receptor engagement triggers the formation of clathrin-coated pits and vesicles that deliver virions to endosomal compartments, where entry is completed through fusion or membrane penetration. These pathways illustrate the major cellular mechanisms exploited by viruses to gain access to the intracellular environment.

Post-binding signalling: receptor-triggered kinases, interferon modulation, inflammatory skewing

Receptors are not passive. Viral engagement can exploit receptor-associated signalling pathways, triggering kinase cascades, cytoskeletal rearrangements, and transcriptional responses that may either facilitate entry or impose early antiviral pressure. A key implication is that trafficking can modify the location and duration of signalling events: receptor activation can persist from endosomal compartments, producing outputs that differ from plasma-membrane signalling (Irannejad et al. 2013).

How binding information at the cell surface is relayed into the cell remains only partly resolved. Receptor cytoplasmic tails, and the adaptor and cytoskeletal proteins they recruit, are central to this relay: they couple ligand engagement to clustering, to the choice of internalization route, and to the kinase cascades that follow, effectively translating an extracellular contact into a defined intracellular programme. Because these tails and their partners differ across receptor families, the same binding event can be read out as distinct signalling and trafficking outcomes depending on which cytoplasmic machinery is engaged, a point developed in detail by Boulant, Stanifer and Lozach (Boulant et al. 2015). This receptor-to-signal coupling step, although experimentally difficult to dissect, deserves more systematic attention as a determinant of entry efficiency and downstream fate.

Downstream, innate antiviral programmes can impose rapid constraints on replication. Thus, the receptor-determined entry route influences the extent to which viral components are exposed to innate immune sensors and interferon trajectory, with subsequent consequences for the magnitude of inflammatory responses and tissue injury (Schoggins and Rice 2011).

Receptor turnover and “availability”: shedding, internalization, recycling, glycosylation, proteolytic priming

Receptor expression is an incomplete proxy for susceptibility because availability is dynamic.

Shedding. Regulated ectodomain shedding can reduce surface receptor exposure and generate soluble forms that act as decoys or modulators. For ACE2, ADAM17-mediated shedding is a well-characterized example of receptor availability being regulated at the post-translational level (Lambert et al. 2005).

Internalization and recycling. Receptors cycle between surface and endosomal pools, with recycling kinetics and compartmental routing often determining the time-integrated probability of productive encounters (Grant and Donaldson 2009).

Proteolytic priming and fusion competence. For many enveloped viruses, receptor engagement is necessary yet insufficient: host proteases prime fusion machinery and can gate whether fusion occurs at the surface or within endosomes, thereby coupling entry efficiency to cell-type-specific protease expression. SARS-CoV-2 provides a paradigmatic demonstration of ACE2 usage coupled to TMPRSS2-dependent priming and route selection (Koch et al. 2021). The broader fusion model emphasizes conserved energetic intermediates whose completion depends on timing and location (Harrison 2015).

Genetic and epigenetic determinants: polymorphisms, expression quantitative trait loci, imprinting-like effects in tissue programmes

Host susceptibility is stratified by regulatory variation and physiological and pathological states that shape receptor abundance, localization, and interactions with trafficking and signalling pathways. Regulatory variants (including eQTLs) can influence receptor expression in a tissue- and context-specific manner, producing phenotypes that may only manifest under inflammation, ageing, or hormonal differentiation (Albert and Kruglyak 2015).

At the experimental level, these considerations sharpen an essential caution: RNA abundance is frequently an unreliable surrogate for protein abundance and even less reliable for surface localization, particularly when turnover and trafficking dominate (Liu et al. 2016).

Box 2. Common experimental traps: overexpression artefacts, permissive cell lines, “RNA ≠ protein”, and spatial mislocalization

1. Overexpression is not validation. It can enforce non-physiological localization, clustering, or adaptor recruitment, manufacturing a pathway rather than revealing one (Doherty and McMahon 2009, Grant and Donaldson 2009).

2. Permissive cell lines lie by omission. Missing restriction factors or damped sensing can inflate the apparent importance of a receptor or route (Schoggins and Rice 2011).

3. RNA ≠ protein (and definitely ≠ surface protein). Transcript maps are hypotheses; protein localization and trafficking can decouple sharply from RNA levels (Liu et al. 2016).

4. Spatial mislocalization breaks causality. A receptor can be abundant yet segregated from relevant microdomains or uptake machinery; low-abundance receptors can dominate if correctly positioned and rapidly recycled (Grant and Donaldson 2009).

5. Route inference by single inhibitors is fragile. Endocytic pathways cross-compensate; orthogonal perturbations and trafficking readouts are required for clean attribution (Doherty and McMahon 2009).

Receptor landscapes: spatial, temporal, and contextual topologies

The receptor concept becomes more informative when spatial organization is considered. “Receptor landscapes” are not a metaphor for heterogeneous expression: they are the combined spatial organization of (i) where a determinant sits (cell type, membrane domain, microdomain), (ii) when it is accessible (differentiation state, turnover, inflammation), and (iii) whether it is functionally coupled to the requisite co-factors (proteases, adaptors, lipid platforms). In this sense, landscapes define state-dependent patterns of permissiveness across barriers, identifying regions that may either favour or restrict dissemination. Large-scale atlases now allow these relationships to be investigated directly, but they also sharpen a warning already implicit in earlier entry biology: the relevant unit is rarely “receptor abundance”, and almost never “bulk tissue expression”, but rather co-localized, polarized, and state-conditioned availability in the right microanatomical niche (Grant and Donaldson 2009, Albert and Kruglyak 2015, Millet and Whittaker 2015, Liu et al. 2016, Bunyavanich et al. 2020, Muus et al. 2021, Consortium* et al. 2022). The concept of receptor landscapes and their context-dependent modulation across biological barriers is illustrated in Fig. 4.

Figure 4.

Context-dependent outcomes of viral receptor engagement. Engagement of the same host receptor can lead to different outcomes depending on the route of viral internalization and the surrounding cellular context. Following receptor binding, viruses may exploit distinct entry pathways, including membrane fusion, macropinocytosis, and clathrin-mediated endocytosis. These processes are further influenced by modulatory factors such as ageing, cellular state, immunosuppression, and therapeutic interventions. The integration of entry route and host context can favour productive infection, promote inflammatory responses, or support antiviral responses, illustrating how receptor function is dynamically shaped by the biological environment.

Tissue distribution is not static: differentiation, developmental stage, ageing

Barrier tissues are in constant flux. Differentiation programmes redistribute entry determinants across epithelial subtypes and across apical versus basolateral domains, while developmental stage and ageing remodel both the cellular census and the transcriptional state of each lineage. An important conceptual shift is to treat “tropism” as a moving intersection between viral binding constraints and host lineage dynamics. In the airway, for example, single-cell analyses have repeatedly indicated that entry-associated factors concentrate in specific epithelial subsets, while co-expression with innate modules implies that early cellular targets and early immune trajectory are entangled rather than separable variables (Bunyavanich et al. 2020, Sungnak et al. 2020). Meta-analytic integration across tissues and demographics further demonstrates that apparent “organ susceptibility” often decomposes into small cell-type fractions with context-dependent co-factor expression, making demographic effects biologically interpretable as changes in cell-state composition and coupling rather than as a simple receptor-on/receptor-off rule (Muus et al. 2021).

Ageing adds a second layer: not only does it shift the proportional representation of epithelial subsets, it can also bias repair states and inflammatory set-points, thereby altering both receptor availability and downstream tissue responses. Recent nasal epithelium work illustrates this logic cleanly: age-stratified epithelial responses can diverge in interferon dynamics and in the emergence of replication-permissive injury-associated states, suggesting that age-related changes in receptor organization and accessibility can independently influence susceptibility and disease progression, upstream of systemic disease trajectory (Woodall et al. 2024).

Barrier anatomy as receptor topology: epithelium, endothelium, basement membrane, glycocalyx, BBB

Barriers are not flat surfaces. For epithelial infections, polarity is a first-order constraint: receptors and co-factors may be enriched apically or basolaterally, and viruses can be forced into distinct entry and egress logics depending on which side is accessible. Polarized models show that entry and release are frequently vectorial, with basolateral access often correlating with deeper tissue invasion and systemic spread, whereas apical cycles can remain locally contained despite high replication (Cong and Ren 2014, Chapuy-Regaud et al. 2022). The immediate implication for receptor inference is uncomfortable but necessary: a determinant may be abundant and even functional, yet irrelevant in vivo if it is sequestered to an inaccessible membrane domain.

Beneath epithelia, the basement membrane and stromal architecture further define accessibility by acting as physical filters and by structuring immune-cell trafficking. At the vascular interface, the endothelial glycocalyx and associated proteoglycans form a charged and mechanically active layer that influences permeability, leukocyte adhesion, and coagulation. This matters for viruses in two ways: (i) it modulates the probability of endothelial engagement (by controlling surface presentation and electrostatic capture), and (ii) its disruption alters vascular permissiveness into one permissive for inflammatory amplification, leakage, and immunothrombotic loops (Tarbell and Cancel 2016, Delgadillo et al. 2021, Zhao et al. 2021, Suzuki et al. 2022). The blood–brain barrier (BBB) is an extreme case of barrier topology: tight junction architecture, specialized transport programmes, and a neurovascular unit in which endothelial glycocalyx, pericytes, and astrocytic endfeet collectively define access. Treating “neurotropism” as a receptor-only property is therefore an oversimplification; it is a barrier topology problem in which receptor availability, immune-cell trafficking, and barrier injury co-determine entry routes into neural compartments (Daneman and Prat 2015, Sweeney et al. 2018).

Inflammation and comorbidity as landscape remodellers

Inflammation is not merely a downstream consequence of infection; it can be a major determinant of receptor remodelling. Cytokines can alter receptor transcription, glycosylation, shedding, and trafficking, while also changing microdomain organization and endocytic routing. A particularly instructive case is interferon: interferon programmes can modulate ACE2-related transcripts, yet careful dissection has shown that viruses and interferons can induce a truncated ACE2 isoform (dACE2) rather than the full-length entry receptor, emphasizing that “upregulation” at the RNA level can mislead unless isoforms and surface localization are resolved (Onabajo et al. 2020, Ziegler et al. 2020). More broadly, the interferon axis exemplifies how landscape remodelling can create time windows in which entry determinants and sensing modules are simultaneously reshaped, changing not only permissiveness but also inflammatory responses and tissue tolerance (Takeuchi and Akira 2010, Schoggins and Rice 2011, Ziegler et al. 2020).

Comorbid states act through similar mechanisms but on longer timescales: chronic lung disease, metabolic disease and autoimmunity remodel epithelia and endothelium, shift baseline interferon tone, and injure the glycocalyx, collectively altering both accessibility and the likelihood that local infection will tip into vascular and systemic pathology (Tarbell and Cancel 2016, Delgadillo et al. 2021, Suzuki et al. 2022). In this framework, comorbidity is best conceptualized not as an “effect modifier” in the epidemiological sense, but as a driver of receptor remodelling that changes baseline tissue susceptibility.

Iatrogenic modulation: how drugs and polypharmacy shift susceptibility without intending to

If receptor landscapes are state-dependent, then pharmacological interventions can unintentionally influence susceptibility. Many drug classes alter barrier integrity, endocytic routing, lipid microdomains, cytokine tone, or glycocalyx composition. Cholesterol-modifying exposures, for instance, can perturb raft organization and thereby alter the assembly and signalling competence of entry platforms (Simons and Gerl 2010, Tarbell and Cancel 2016). Anti-inflammatory and immunomodulatory regimens can reshape interferon dynamics, potentially shifting the timing of epithelial sensing versus permissiveness, with consequences for early containment and later immunopathology (Takeuchi and Akira 2010, Schoggins and Rice 2011, Ziegler et al. 2020). Meanwhile, therapies that influence endothelial activation and glycocalyx shedding can, in principle, move a host from a state in which the vasculature limits dissemination to one in which it becomes a contributor to systemic pathology (Tarbell and Cancel 2016, Delgadillo et al. 2021, Suzuki et al. 2022).

The practical consequence is that “risk” can be pharmacologically produced without any direct interaction between a drug and a viral binding interface. Polypharmacy, in particular, should be treated as a broad perturbation of receptor availability and barrier function: small effects on barrier permeability, trafficking and inflammatory set-points can summate into meaningful changes in susceptibility and dissemination potential. This is not a call for therapeutic nihilism; it is a demand for mechanistic humility when interpreting drug-associated susceptibility signals.

Spatial biology: gradients, niches, and “hotspots” of permissiveness revealed by single-cell and spatial omics

Single-cell atlases established that entry determinants are sparse and cell-type-specific in many tissues. Spatial methods now add the missing dimension: where those rare permissive cells sit relative to barrier microanatomy, immune niches, and vascular interfaces. Spatial transcriptomics in the respiratory system highlights that clinically relevant biology is organized into regional programmes and niche-dependent states, such that receptor and co-factor expression can be concentrated in microdomains that are disproportionate determinants of exposure, initiation of local replication, or repair-coupled amplification (Megas et al. 2025). This is the empirical foundation for the “hotspot” concept: small spatially defined cell pools can act as control points that determine whether infection remains local, becomes disseminated, or contributes to multi-organ disease.

Importantly, spatial data also expose why simple receptor mapping fails: permissiveness emerges from co-localization of receptors, accessory proteases, endocytic competence, and immune context within specific anatomical microenvironments. In translational terms, this is where receptor-centred systems-level virology can inform risk stratification and host-directed interventions (repair programmes, inflammatory microdomains, glycocalyx integrity) rather than single molecules in isolation (Consortium* et al. 2022, Megas et al. 2025).

Box 3. Receptor landscapes in the real world: why disease states rewrite the rules of tropism

• Polarity is destiny. Apical versus basolateral receptor placement can separate local replication from invasive potential (Cong and Ren 2014, Chapuy-Regaud et al. 2022).

• The glycocalyx is not decoration. Endothelial and epithelial glycocalyces function as dynamic filters that shape access and inflammatory amplification; injury converts a protective barrier into a driver of inflammatory amplification (Tarbell and Cancel 2016, Delgadillo et al. 2021, Zhao et al. 2021, Suzuki et al. 2022).

• Inflammation reshapes receptor landscapes before infection outcomes are determined. Interferon and cytokine programmes reshape receptor-related transcripts, isoforms and trafficking, changing permissiveness and sensing in the same stroke (Onabajo et al. 2020, Ziegler et al. 2020).

• Demography often means cell-state composition. Apparent age effects can reflect shifts in epithelial states and repair-associated permissive niches rather than uniform receptor changes (Muus et al. 2021, Woodall et al. 2024).

• Drugs modulate access and routing. Microdomain organization and barrier physiology are pharmacologically plastic; susceptibility signals may reflect changes in tissue state rather than direct effects on viral binding (Simons and Gerl 2010, Tarbell and Cancel 2016).

From local infection to system-level pathogenesis: receptors as determinants of dissemination and disease severity

The passage from a focal infection to a multi-organ syndrome is rarely explained by viral load alone. It is, more often, a question of receptor distribution and accessibility: where the relevant receptors sit, on which cell states, across which barriers, and with what signalling and haemostatic consequences once engaged (Marsh and Helenius 2006). Receptors therefore influence dissemination while simultaneously modifying host physiology in ways that can shape clinical outcome (Marsh and Helenius 2006, Engelmann and Massberg 2013). This section develops a receptor-centred logic for systemic disease, treating receptor engagement not as a single step of entry, but as a multilevel biological process across barriers, immune lineages, and vascular compartments (Barabási et al. 2011).

Local replication versus systemic dissemination: receptor-dependent dissemination

At the tissue scale, dissemination is not merely a function of replication competence but of access to the cellular conduits that bridge compartments (Marsh and Helenius 2006). Viruses restricted to apical epithelial determinants may generate intense local pathology without efficient systemic spread; conversely, engagement of receptors enriched on migratory immune cells or barrier-adjacent endothelia can transform a mucosal exposure into a body-wide event (Marsh and Helenius 2006). A canonical illustration is measles virus, whose systemic phase is promoted by entry into immune cells via SLAM/CD150 (Tatsuo et al. 2000), while transmission from the host is subsequently enabled by infection of epithelial surfaces through nectin-4/PVRL4 (Noyce et al. 2011). The sequence is instructive: receptor usage is staged, with early receptors governing amplification and dissemination, and later receptors governing shedding and onward spread (Tatsuo et al. 2000, Noyce et al. 2011). The disease phenotype is thereby shaped as much by the order of receptor engagement as by intrinsic cytopathic capacity (Marsh and Helenius 2006).

Analogous logic applies across viral families. Arenaviruses, for example, can exploit α-dystroglycan, a receptor linking extracellular matrix to cytoskeleton, which provides a mechanistic bridge between broad tissue accessibility and severe systemic phenotypes in relevant infections (Cao et al. 1998). Filoviruses provide a complementary lesson: while attachment and priming are multi-step processes, a stringent intracellular receptor requirement, exemplified by NPC1, can represent a critical determinant of susceptibility at the cellular level even when exposure and uptake are permissive (Carette et al. 2011). The apparent paradox, that a virus may “reach” a cell yet fail to establish productive infection, is resolved by recognizing receptor requirements as layered constraints rather than a single lock-and-key event (Marsh and Helenius 2006, Carette et al. 2011).

In clinical terms, dissemination can be understood as a process strongly influenced by receptor distribution and cellular trafficking. Receptor expression on antigen-presenting cells, trafficking lymphocytes, or barrier-proximal stromal populations effectively determines whether the infection remains a compartmentalized mucosal event or becomes a systemic inflammatory disease (Tatsuo et al. 2000, Marsh and Helenius 2006, Barabási et al. 2011).

Endothelial engagement and vascular pathology: permeability, microthrombi, immunothrombosis

If systemic viral disease has a common final pathway, it is frequently vascular. The endothelium is both a barrier and an organ-scale signalling surface; receptor engagement on endothelial cells and platelets can therefore convert infection into haemostatic and inflammatory catastrophe (Engelmann and Massberg 2013). The concept of immunothrombosis, in which coagulation functions as an intravascular effector arm of innate immunity, provides a principled framework for understanding why receptor-proximal endothelial perturbation can culminate in microvascular occlusion, tissue hypoxia, and organ failure (Engelmann and Massberg 2013).

Endothelial receptor relevance is inseparable from surface architecture. The endothelial glycocalyx is a mechanically and biochemically active layer that regulates leukocyte adhesion, shear sensing, and permeability; its integrity therefore shapes whether vascular surfaces behave as filters, signal integrators, or inflammatory amplifiers (Reitsma et al. 2007, Weinbaum et al. 2007). From this perspective, glycocalyx damage can increase effective accessibility of endothelial determinants and potentiate downstream thrombo-inflammatory programmes, even when direct endothelial infection is limited or heterogeneous (Reitsma et al. 2007, Weinbaum et al. 2007).

Several lines of evidence across diseases support a receptor-mediated vascular axis. In COVID-19, histopathological studies reported pulmonary vascular injury with thrombosis and microangiopathy, with features consistent with dysregulated endothelial–immune crosstalk (Ackermann et al. 2020). Complementary observations suggested endothelial infection and endotheliitis as contributors to vascular dysfunction (Varga et al. 2020). Several studies have argued that endothelial activation is not a peripheral detail but a co-determinant of severe systemic disease, with the vasculature acting as a propagation surface for inflammation and coagulation (Teuwen et al. 2020). Clinical biomarker evidence further supports an endotheliopathy phenotype linked to critical illness and mortality, consistent with a vascular model of disease progression in which endothelial state helps set the system-level trajectory (Goshua et al. 2020). While the extent and uniformity of direct endothelial infection remain debated across tissues and stages, the receptor-centred inference is stable: where endothelial engagement and/or activation is permitted, the vascular compartment becomes capable of amplifying systemic inflammation and coagulation (Engelmann and Massberg 2013, Goshua et al. 2020, Teuwen et al. 2020, Varga et al. 2020).

Neutrophil extracellular traps (NETs) provide a bridge mechanism between immune activation and thrombosis, and their association with thrombotic tendencies in severe viral disease underscores how receptor engagement can propagate into haemostatic pathology via immune effector programmes rather than via cytolysis alone (Leppkes et al. 2020, Zuo et al. 2020). In parallel, viruses that exploit receptors on platelets and endothelium can directly entangle entry biology with thrombocytopenia and vascular leak. Pathogenic hantaviruses use β3 integrins to mediate cell entry, a receptor choice that is mechanistically congruent with platelet involvement and endothelial dysfunction (Gavrilovskaya et al. 1998). Further work showed that pathogenic hantaviruses can inhibit β3 integrin–directed endothelial migration, illustrating how receptor engagement can reprogramme endothelial behaviour in ways that plausibly contribute to barrier failure and impaired vascular repair (Gavrilovskaya et al. 2002). Together, these observations indicate that vascular disease in viral infection is often receptor-proximal, not a generic downstream “complication” (Gavrilovskaya et al. 1998, 2002, Engelmann and Massberg 2013, Goshua et al. 2020).

Neurotropism and neuroinvasion: barrier crossing, axonal routes, immune cell trafficking

Neuroinvasion is frequently treated as a special case, yet it can be analyzed with the same conceptual principles (Marsh and Helenius 2006). The nervous system is protected by layered barriers and privileged microenvironments; viruses that access neuronal or glial determinants can bypass constraints that limit dissemination elsewhere (Abbott et al. 2010). Rabies virus is the archetype of receptor-guided neurotropism. Evidence supports interactions with the nicotinic acetylcholine receptor (nAChR) at neuromuscular junctions (Lentz et al. 1982, Gastka et al. 1996), the neural cell adhesion molecule (NCAM/CD56) (Thoulouze et al. 1998), and the low-affinity nerve growth factor receptor p75NTR (Tuffereau et al. 1998). The conceptual unity is strong: receptor engagement at peripheral nerve terminals and along adhesion or neurotrophin pathways can couple entry to retrograde transport and central nervous system access (Lentz et al. 1982, Gastka et al. 1996, Thoulouze et al. 1998, Tuffereau et al. 1998).

For other viruses, neuroinvasion may proceed through barrier disruption rather than canonical neuronal receptor usage. Here, the blood–brain barrier (BBB) becomes a key determinant of receptor accessibility and neuroinvasion, in which endothelial and perivascular signalling states dictate permeability and immune trafficking (Abbott et al. 2010). Receptor engagement on vascular or immune compartments can therefore indirectly govern neuroinvasion by altering BBB integrity and leukocyte transmigration, even when neuronal entry receptors are not the limiting step (Abbott et al. 2010, Barabási et al. 2011).

Accessory factors further complicate neurotropism by modulating entry efficiency at specific anatomical sites. Neuropilin-1 (NRP1), for example, has been described as a host factor that potentiates SARS-CoV-2 infection following furin cleavage of spike, consistent with a role in enhancing infectivity in receptor-rich niches (Cantuti-Castelvetri et al. 2020, Daly et al. 2020). Such factors are best interpreted as accessory host factors: they may not define permissiveness in isolation, but can reshape where and how efficiently infection proceeds, thereby altering the probability of neuroinvasion via olfactory, vascular, or immune-mediated routes (Marsh and Helenius 2006, Cantuti-Castelvetri et al. 2020, Daly et al. 2020).

Immune cell infection/interaction: receptors on myeloid and lymphoid lineages as phenotype shapers

Immune cells are not passive responders. Their receptor profiles can render them targets, reservoirs, and mediators of immune signalling (Marsh and Helenius 2006). When viruses productively infect immune cells, receptor usage becomes a determinant of immunological trajectory: which compartments are depleted, which cytokine programmes dominate, and whether the net outcome is immunosuppression, hyperinflammation, or immune diversion (Barabási et al. 2011).

Measles again provides a clean demonstration. SLAM/CD150 is expressed on activated immune cells, and its use as an entry receptor coheres with systemic spread and profound immune modulation (Tatsuo et al. 2000). The later use of nectin-4 on epithelia underscores that distinct receptor engagements can correspond to distinct phenotype modules within the same infection (Noyce et al. 2011). In HIV-1, receptor usage is inseparable from disease course. CCR5 acts as a major co-receptor enabling entry into CD4 + cells and is tightly linked to early macrophage-tropic infection dynamics (Connor et al. 1997, Berger et al. 1999). The identification of fusin (CXCR4) as an entry cofactor established that alternative co-receptor usage is not a minor variation but a biological axis with implications for cell targeting and clinical progression (Bleul et al. 1996). In this setting, determinants of immune-cell targeting define which subsets are engaged and therefore which immunological functions are eroded or distorted (Bleul et al. 1996, Connor et al. 1997, Berger et al. 1999).

Multi-organ syndromes: linking receptor engagement to systemic disease

Multi-organ syndromes can be conceptualized as failures of distributed receptor networks. In this framing, the organism is a graph whose nodes include barrier surfaces, vascular endothelium, immune lineages, and organ-specific parenchymal compartments (Barabási et al. 2011). Viral receptor engagement perturbs multiple nodes; disease severity emerges when these perturbations synchronize via feedback loops (for example, endothelial activation → immunothrombosis → microvascular occlusion → hypoxia → inflammatory amplification) rather than remaining localised (Barabási et al. 2011, Engelmann and Massberg 2013, Leppkes et al. 2020, Middleton et al. 2020).

Network medicine offers a useful analogy: complex diseases arise not from single lesions but from perturbations that propagate across interconnected modules (Barabási et al. 2011). Viral receptors provide mechanistic entry points for such propagation. A virus that engages determinants in vascular and immune compartments effectively acquires a disproportionate impact on systemic disease, enabling systemic phenotypes that are not reducible to local replication alone (Barabási et al. 2011, Engelmann and Massberg 2013). Conversely, when receptor engagement is confined to a limited compartment, pathology may remain severe but localized, and systemic collapse is less likely (Marsh and Helenius 2006). This perspective suggests that predicting severe disease requires mapping not only receptor expression but also the relationship between receptor engagement and downstream physiological processes: coagulation, barrier permeability, neuroimmune signalling, and inflammatory set-points (Abbott et al. 2010, Barabási et al. 2011, Engelmann and Massberg 2013).

Receptor plasticity, evolution, and spillover: when landscapes meet selection

Host range as a receptor-compatibility problem (and its constraints)

Host range is often described as an ecological property, but it is more cleanly treated as a compatibility constraint imposed by receptor engagement and the downstream entry programme it licenses. A virus may circulate in a reservoir with high prevalence yet remain effectively “non-human” if its surface proteins cannot productively engage human entry receptors in the correct cellular context. This is why receptor usage is not merely a mechanistic footnote: it is a major evolutionary constraint that shapes the potential for cross-species transmission. For coronaviruses, the receptor-binding domain (RBD) provides a particularly stark example: lineage emergence is repeatedly accompanied by RBD changes that alter receptor compatibility and, by extension, tissue tropism and transmission routes. The modularity of spike, together with recombination, makes receptor engagement an unusually evolvable constraint within otherwise conserved replication machinery (Graham and Baric 2010, Cui et al. 2019).

Compatibility, however, is multi-dimensional. Binding must occur at the right affinity and geometry, but also within the right barrier organization: the receptor must be accessible at the relevant anatomical interface, presented in a permissive microenvironment, and coupled to an entry route that avoids immediate abortive outcomes. Structural work on ACE2 engagement illustrates how receptor compatibility is defined by a finite set of interaction hot-spots: modest differences at key residues can flip a species from permissive to restrictive, even if ACE2 is expressed (Lan et al. 2020). Conversely, adaptation studies of SARS-CoV showed that changes in spike can meaningfully tune usage of human ACE2, thereby shifting the effective host range (Li et al. 2005).

A parallel logic applies to MERS-CoV: DPP4 is the functional receptor, but interspecies restriction is strongly shaped by sequence and structural variation in DPP4, which can be overcome by viral adaptation along multiple mutational paths (Raj et al. 2013, Wang et al. 2013, Letko et al. 2018). The implication is awkwardly simple: host range depends on multiple biological constraints, of which receptor compatibility is among the most readily measurable.

Adaptive routes: mutation, recombination, convergent solutions

Once receptor compatibility becomes the limiting step, selection predictably concentrates on receptor-binding proteins and the accessory features that condition receptor engagement (glycosylation patterns, protease priming sites, and conformational dynamics). Deep mutational scanning of the SARS-CoV-2 RBD provides an empirical map of this distribution of functional constraints: most single substitutions are deleterious for folding or ACE2 binding, yet there exists a non-trivial set of tolerated or affinity-enhancing changes, including at interface residues. This matters because it formalizes a key point for spillover inference: adaptation is not “anything goes”, but neither is it vanishingly improbable when a virus is repeatedly exposed to a new receptor environment (Starr et al. 2020).

Recombination adds a second, faster route: it can import an RBD (or larger spike segments) that is already compatible with a new receptor background, effectively bypassing slow, stepwise optimization. This is not hypothetical. The coronavirus literature has long argued that recombination, reservoir diversity, and spike modularity jointly accelerate cross-species emergence, precisely because receptor-binding is both selectable and exchangeable (Graham and Baric 2010, Cui et al. 2019).

Convergence is the third route, and it is frequently observed across unrelated virus families. Across virus families, similar receptor constraints yield recurrent solutions: increased glycan engagement for attachment, altered receptor-binding specificity, or mimicry of host ligands that dock into widely expressed entry platforms. Influenza offers the canonical case: shifts in haemagglutinin (HA) receptor-binding preference (avian-type versus human-type sialylated glycans) are repeatedly implicated in interspecies transmission and changes in transmissibility and pathogenesis (Imai and Kawaoka 2012, Imai et al. 2012, de Graaf and Fouchier 2014). Even when the proximate mechanisms differ, the underlying principle is conserved: selection gravitates to the interface where a virus can expand its host range and get evolutionary advantage.

A more contentious, but illustrative, example of convergent logic is the exploration of integrin engagement by SARS-CoV-2 spike. Regardless of where the debate ultimately settles, the underlying evolutionary temptation is clear: integrins are abundant, structurally permissive docking partners used by diverse viruses, and any credible route to integrin-assisted entry would represent a classic case of exploitation of widely expressed host determinants to expand tissue access (Beaudoin et al. 2021).

Intermediate hosts and receptor stepping-stones

Spillover is rarely a clean jump from reservoir to humans. Intermediate hosts can function as receptor “stepping-stones”, not because they are mystical epidemiological waypoints, but because they provide a permissive receptor environment in which partially compatible variants can replicate long enough to explore adaptive space. This concept is most persuasive when the intermediate host offers (i) receptor compatibility that is closer to humans than the reservoirs, and (ii) an exposure interface that amplifies contact with humans.

Influenza again provides a useful organizing analogy: receptor-binding specificity is a core determinant of whether avian strains remain sporadic zoonoses or acquire attributes compatible with sustained mammalian transmission. Reviews and experimental work converge on the principle that altered receptor binding is necessary but not sufficient: it must intersect with the right replication and transmission phenotype, and the right host anatomy pathogenesis (Imai and Kawaoka 2012, Imai et al. 2012, de Graaf and Fouchier 2014).

For MERS-CoV, receptor usage is anchored in DPP4, and the intermediate-host narrative is strongly supported by mechanistic receptor data: DPP4 is the functional receptor, and structural studies define the interaction surface in detail, while experimental adaptation demonstrates how spike can traverse species variation in DPP4 via multiple mutational trajectories (Raj et al. 2013, Wang et al. 2013, Letko et al. 2018). Here, the intermediate host is not just “where the virus was found”; it is a biological environment that facilitates adaptation to receptor constraints.

Crucially, receptor stepping-stones are not unique to respiratory viruses. Canine parvovirus emergence hinged on adaptation to the transferrin receptor (TfR), a clean demonstration that a small number of capsid changes at the receptor interface can retune host range with dramatic ecological consequences (Allison et al. 2014). This is the broader lesson: intermediate hosts matter most when they reduce receptor-related barriers without fully removing them, because that is precisely when selection has something to work with.

Predicting spillover risk via receptor landscape similarity and barrier accessibility

Predicting spillover risk is a losing game if it is framed as prophecy. It becomes tractable, however, if framed as risk scoring across measurable constraints: receptor compatibility, receptor accessibility at real-world barriers, and ecological amplification. Comparative receptor analyses are one such constraint-based approach. The ACE2 “binding-score” logic used to predict SARS-CoV-2 host range across vertebrates illustrates how receptor landscape similarity can be operationalized using sequence and structural features at known binding sites, rather than vague phylogenetic proximity (Damas et al. 2020).

Yet receptor compatibility alone is not spillover. Spillover is compatibility plus exposure plus within-host opportunity. This is where barrier biology stops being a chapter heading and becomes the entire plot. The most elegant conceptual synthesis in the broader spillover literature is the argument that zoonoses must traverse a hierarchical series of barriers, from pathogen pressure at the interface through within-human susceptibility and onward transmission. Receptors sit at the nexus of these barriers: they shape susceptibility, but are themselves shaped by inflammation, comorbidity, and tissue state (Table 2) (Plowright et al. 2017).

Table 2.

Spillover-relevant variables: receptor compatibility, barrier exposure, expression plasticity, and ecological interfaces.

Variable	What it means mechanistically	Practical proxies (examples)	Why it matters for spillover
Receptor compatibility	Productive engagement of an entry receptor (geometry, affinity, activation)	Structural modelling at binding hot-spots; DMS constraint maps; pseudotype entry across receptor orthologues	Sets the first biological gate for cross-species infection (Damas et al. 2020, Starr et al. 2020)
Barrier accessibility (“access”)	Receptor is reachable at the relevant anatomical interface (airway, gut, skin, endothelium)	Tissue localization (apical vs basolateral); barrier disruption markers; interface-contact frequency	Converts compatibility into real infection probability (Plowright et al. 2017)
Expression plasticity (“state dependence”)	Receptor availability shifts with inflammation, differentiation, age, comorbidity	Single-cell/spatial atlases in health vs disease; cytokine-stimulated models; iatrogenic modulation signatures	Explains why risk is contextual, not species-fixed (Plowright et al. 2017)
Ecological interface & amplification	Contact rates + opportunities for onward transmission	Human–animal interface mapping; husbandry/trade data; viral prevalence/seasonality	Determines how many “evolutionary draws” the virus gets (Kreuder Johnson et al. 2015, Olival et al. 2017)

Finally, surveillance prioritization benefits from recognizing that some viruses are structurally predisposed to spill over because they exhibit high host plasticity, broad host breadth, or other traits associated with zoonotic potential. Trait-based models are not substitutes for receptor mechanics, but they provide an orthogonal axis: receptor compatibility, ecological exposure, and clinical impact represent complementary dimensions of spillover risk assessment (Kreuder Johnson et al. 2015, Olival et al. 2017).

Box 4. A pragmatic spillover heuristic: “compatibility × access × amplification”

A useful heuristic is to treat spillover risk as the product of three terms:

1. Compatibility: can the virus productively engage a human receptor (or a sufficiently human-like receptor) and complete entry? (Damas et al. 2020)

2. Access: does the relevant receptor landscape exist at the interface where exposure occurs, in the physiological states that actually happen outside grant applications? (Plowright et al. 2017)

3. Amplification: how often does the virus get to try, and how effectively is it amplified in reservoir/intermediate populations before encountering humans? (Kreuder Johnson et al. 2015, Olival et al. 2017)

This heuristic is not philosophy. It is a decision rule for surveillance triage: high compatibility with low access is usually a laboratory curiosity; high access with low compatibility is usually noise; high amplification makes even “improbable” compatibility pathways statistically inevitable.

Translational implications: receptors as targets, modifiers, and biomarkers

Receptors occupy an unusually attractive position for intervention because they sit upstream of replication yet downstream of exposure: they represent one of the earliest biological constraints encountered during infection, and they often couple binding to signalling, trafficking, and barrier responses that shape disease severity (Marsh and Helenius 2006, Engelmann and Massberg 2013). This makes receptor-centred strategies conceptually distinct from classical antivirals. Rather than targeting viral proteins directly, they attempt to reduce the host’s entry competence or to divert engagement into non-productive routes, ideally with variant-robustness and a wider therapeutic window than direct-acting antivirals allow (Marsh and Helenius 2006, Starr et al. 2020). Similar principles have also informed the evaluation of naturally derived compounds capable of interfering with receptor-associated stages of viral infection, including viral attachment, entry, and early host-cell responses (Gentile et al. 2021). The price is equally clear: receptors are not passive docking sites but physiological control elements, and their perturbation carries tissue- and time-dependent liabilities (Reitsma et al. 2007, Weinbaum et al. 2007). Host-directed therapy contexts therefore emphasize stratification, careful endpoint selection, and an explicit safety logic that is supported by evidence rather than by generic safety assumptions (Kaufmann et al. 2018, Kumar et al. 2020). Beyond direct receptor targeting, compounds capable of simultaneously modulating viral adsorption and inflammatory signalling may offer complementary host-directed strategies aimed at limiting both viral entry and tissue damage (Spampinato et al. 2022, 2025). The exploration of natural products and semisynthetic derivatives further illustrates the diversity of strategies currently being investigated to modulate host–virus interactions beyond conventional direct-acting antivirals (Fuochi and Furneri 2021) (Fig. 5).

Figure 5.

Therapeutic intervention points across the progression of viral infection. Host–virus interactions provide multiple opportunities for therapeutic intervention from initial attachment to systemic disease. Viral binding can be targeted using neutralizing antibodies or soluble decoy receptors, whereas subsequent entry can be inhibited by blocking membrane fusion and/or endocytic uptake. Receptor-associated intracellular signalling may be modulated through immunomodulatory strategies aimed at limiting excessive inflammatory responses. At later stages, preservation of epithelial and endothelial barrier integrity, including stabilization of tight junctions, may reduce tissue damage and dissemination. Once systemic involvement develops, antiviral and anti-inflammatory treatments can be used to limit viral replication and host-mediated pathology. Together, these strategies illustrate how receptor-centred and receptor-adjacent interventions can act at distinct stages of infection, from viral binding and entry to cellular and systemic damage.

Host-directed therapeutics aimed at receptors: monoclonals, small molecules, allosteric modulators

The cleanest clinical precedent for receptor-targeting as antiviral therapy remains HIV-1, where entry requires coordinated engagement of CD4 and chemokine co-receptors (Bleul et al. 1996, Connor et al. 1997, Berger et al. 1999). CCR5 antagonism by small molecules established that it is feasible to treat a viral infection by pharmacologically disabling a host receptor in vivo, provided that receptor usage is appropriately phenotyped and the clinical context tolerates partial immune modulation (Fätkenheuer et al. 2008). In parallel, receptor-directed monoclonals illustrate an orthogonal route: ibalizumab binds CD4 and functions as a post-attachment inhibitor, exemplifying how receptor engagement can be modulated without fully ablating receptor expression or broadly suppressing immune competence (Emu et al. 2018). Together, these programmes exemplify the key translational point: receptor interventions require phenotype-based selection (who is using which receptor, in which compartments) rather than non-selective clinical use (Kaufmann et al. 2018, Kumar et al. 2020).

A second class of “receptor-targeting” interventions operates through receptor-adjacent host factors that license entry. For SARS-CoV-2, ACE2 is necessary but not sufficient; proteolytic priming by TMPRSS2 can route entry through a cell-surface pathway that differs from endosomal entry, with consequences for inhibition strategies, tissue selectivity, and timing (Koch et al. 2021). This aligns with the broader receptor-control-node thesis: “entry” is a composite phenotype in which receptors, accessory factors, and microdomain context jointly determine success or failure (Marsh and Helenius 2006). Additional evidence from pseudoviral systems indicates that glycosaminoglycans can interfere with Spike–ACE2 engagement in a treatment-dependent manner, reinforcing the concept that receptor-associated attachment processes remain pharmacologically tractable targets for host-directed intervention (Fuochi et al. 2025a).

Decoy receptors and engineered traps: design principles, avidity engineering, safety liabilities

Decoy receptors transform the receptor concept from a cellular gate into a soluble sink. Mechanistically, they aim to neutralize virions by providing high-avidity binding surfaces that outcompete membrane receptors, ideally across variants that would evade monoclonals targeting more immunodominant viral epitopes. The ACE2 axis in COVID-19 made this strategy unusually tangible because the entry receptor is known, structurally characterized, and itself enzymatically active (Zoufaly et al. 2020). This enables a dual-action hypothesis: viral neutralization plus partial rebalancing of dysregulated renin–angiotensin signalling, even if the latter remains context-dependent and difficult to prove in heterogeneous clinical disease (Zhang et al. 2023).

The engineering literature now supports design principles rather than isolated constructs. Receptor traps can be optimized through multimerization, Fc fusion, and interface engineering to increase affinity and residence time, with in vitro neutralization potencies approaching those of high-affinity antibodies (Glasgow et al. 2020). Engineered decoys have demonstrated broad binding across SARS-related coronaviruses and retained activity against divergent variants, consistent with a key advantage over epitope-restricted antibodies: a virus cannot easily escape a decoy without simultaneously compromising endogenous receptor binding (Chan et al. 2021). In vivo work suggests that decoys can mitigate lung injury and improve survival in relevant models, strengthening the argument that decoys can act as both antiviral and pathophysiology modulators when receptor biology is itself engaged in disease amplification (Zhang et al. 2022). This observation is consistent with spillover biology: receptor-binding is both a constraint and an anchor, which is precisely why it can be exploited for variant-robust countermeasures (Graham and Baric 2010, Cui et al. 2019, Starr et al. 2020).

Clinical translation remains more complex. Soluble recombinant ACE2 has been reported in severe disease contexts and discussed as a plausible dual-action intervention, but its deployment raises predictable constraints: dosing to achieve airway-relevant concentrations, duration of effect, route of administration, and whether systemic delivery meaningfully influences the mucosal interface where early replication dominates (Zoufaly et al. 2020, Zhang et al. 2023, Bauer et al. 2024). Aerosol delivery is therefore mechanistically coherent, and early tolerability data illustrate feasibility even if definitive efficacy remains a separate question (Bauer et al. 2024).

Indirect receptor modulation: modifying receptor availability rather than blocking binding

Some of the most pragmatic receptor-centred interventions do not block binding at all. Instead, they shift the receptor landscape by altering receptor localization, proteolytic processing, or the cell-state programmes that render receptors accessible at key barriers (Reitsma et al. 2007, Weinbaum et al. 2007). This is less direct than receptor blockade, but more aligned with how barriers behave in vivo: receptor organization is dynamic, inflammation-sensitive, and pharmacologically plastic (Marsh and Helenius 2006, Weinbaum et al. 2007).

Protease inhibition exemplifies this category. Demonstrations that SARS-CoV-2 entry depends on ACE2 and TMPRSS2, and can be inhibited by a clinically used protease inhibitor, framed entry as a composite phenotype that can be disrupted without directly targeting ACE2 itself (Hoffmann et al. 2020b). Mechanistic work has reinforced that protease expression patterns dictate entry routing, implying that therapeutic success is likely to be tissue- and timing-dependent rather than uniform across disease stages (Koch et al. 2021).

A second lever is receptor shedding and turnover. ACE2 can be processed by host proteases, generating soluble forms with complex implications: soluble ACE2 can act as an inhibitor by sequestering virions, yet altered shedding may also influence membrane receptor availability and therefore local entry competence (Heurich et al. 2014, Guo et al. 2022). More broadly, the general point is stable: interventions that alter receptor processing can alter susceptibility without permanently silencing receptor function, but they demand careful validation because the same processing pathway may be protective in one compartment and harmful in another (Marsh and Helenius 2006).

Receptors as biomarkers: severity prediction, organ risk, and stratification for trials

If receptor biology contributes to disease progression, then receptor-related measurements should, in principle, carry prognostic information. The difficulty is that “receptor level” is not a single variable: it includes membrane localization, microdomain context, post-translational state, and shedding dynamics (Marsh and Helenius 2006, Weinbaum et al. 2007). Circulating ACE2 illustrates both promise and complexity. Associations between plasma ACE2 and clinical outcomes have been reported, supporting the notion that receptor-axis readouts can stratify risk and reflect systemic vulnerability in severe viral illness (Kragstrup et al. 2021). Yet interpretation is non-trivial: circulating levels may represent shedding, tissue injury, compensatory regulation, or therapeutic interference, and therefore require models that incorporate clinical and biological context rather than simple thresholds (Kaufmann et al. 2018, Kumar et al. 2020, Kragstrup et al. 2021).

A more clinically useful application is trial stratification. Receptor-centred therapeutics are, by definition, contingent on receptor biology; therefore, receptor proxies can function as enrichment biomarkers to increase signal-to-noise in interventional studies. Mechanistic endpoints should be specified upfront: receptor-axis biomarker kinetics, protease signatures, and (where feasible) barrier-relevant localization readouts, rather than post hoc “correlates” designed to look good in supplementary data.

Preparedness: receptor-informed surveillance and prioritization frameworks

Preparedness benefits from a receptor-informed approach because it converts emergence risk from narrative to measurable constraints. If a virus cannot productively engage a human receptor landscape at a relevant barrier, then its pandemic potential is limited regardless of ecological noise (Plowright et al. 2017, Damas et al. 2020). Conversely, viruses whose entry determinants are already compatible with human receptors, or whose receptor-binding is modular and evolvable (via mutation or recombination), deserve disproportionate attention (Graham and Baric 2010, Cui et al. 2019, Starr et al. 2020). This is not a replacement for ecological surveillance; it is a weighting function that makes surveillance triage less arbitrary (Kreuder Johnson et al. 2015, Olival et al. 2017).

A receptor-informed framework integrates three layers: (i) receptor compatibility evidence (structural, functional, or both), (ii) barrier accessibility in real-world physiological states, and (iii) amplification opportunity at human–animal interfaces (Kreuder Johnson et al. 2015, Olival et al. 2017, Plowright et al. 2017, Damas et al. 2020). Such a framework complements established spillover barrier models by making the first barrier, the cellular gate, experimentally addressable and therefore useful for prioritization (Plowright et al. 2017, Grange et al. 2021).

Future directions: towards a receptor-centred systems virology

The functional receptor model: beyond expression to localization, membrane organization, and state dependence

A receptor-centred systems virology will fail if it continues to treat receptors as static “expression levels” rather than dynamic control elements whose function depends on membrane topology, nanoscale organization, and state-dependent availability (Marsh and Helenius 2006, Weinbaum et al. 2007). The immediate priority is therefore to formalize a functional receptor framework: a description that integrates (i) protein abundance, (ii) subcellular localization, (iii) microdomain partitioning, (iv) post-translational state (glycosylation, shedding, proteolysis), and (v) coupling to endocytic and signalling machinery (Marsh and Helenius 2006). This is not pedantry: it is the difference between predicting permissiveness and narrating it after the fact (Marsh and Helenius 2006, Starr et al. 2020).

Two methodological shifts make this feasible at scale. First, spatial mapping approaches allow receptor and accessory-factor programmes to be interpreted in situ rather than inferred from dissociated cells, reducing the recurrent error of collapsing barrier anatomy into averaged signals (Ståhl et al. 2016, Rodriques et al. 2019). Second, multi-omic single-cell methods bridge the gap between receptor expression and receptor function by linking transcriptional state to surface protein and immune phenotype at the same cellular address, directly addressing the “RNA ≠ protein” trap that distorts receptor inference (Stoeckius et al. 2017, Stuart et al. 2019). These methods matter because receptors act as control nodes only when they are accessible, correctly localized, and coupled to the relevant entry and signalling routes (Weinbaum et al. 2007).

Practically, the field should converge on reporting standards that state explicitly: RNA detected, protein detected, spatial compartment, and assay context, because each one meaningfully alters inference about susceptibility and pathogenesis (Kumar et al. 2020). Where possible, receptor work should default to spatially resolved readouts (polarity, junctional location, raft association), since these variables are often more predictive of dissemination and systemic effects than bulk tissue expression (Weinbaum et al. 2007).

Spatial multi-omics, organoids, barrier-on-chip systems, and in vivo validation logic

If receptor engagement is a control node, then the experimental agenda must privilege models that preserve barrier physics and cell-cell context rather than convenience (Marsh and Helenius 2006). Spatial multi-omics applied to infected tissues offers a direct route from receptor distribution to downstream immune trajectory and tissue injury, helping discriminate whether a receptor is acting primarily as an entry licence, a signalling trigger, or a barrier-disruptive amplifier (Tan et al. 2023). This becomes especially relevant when systemic phenotypes hinge on vascular and immunothrombotic coupling rather than cytolysis alone (Barabási et al. 2011, Engelmann and Massberg 2013).

Barrier-on-chip systems move the field from descriptive atlases to mechanistic stress tests. The lung-on-chip paradigm demonstrated that organ-level responses can be reconstructed with controlled mechanical and cellular parameters, offering an experimental grammar for separating receptor-dependent entry from receptor-triggered injury cascades (Huh et al. 2010). For neuroinvasion, BBB platforms provide a tractable compromise between reductionism and realism: hypoxia-enhanced BBB chips support barrier function and transport behaviours relevant to CNS exposure (Park et al. 2019), and protocol-level standardization makes cross-lab comparability plausible rather than aspirational (Hajal et al. 2022). This is precisely the level at which neurotropism should be interrogated: as a receptor-plus-barrier problem embedded in physiology, not as a binary “neurotropic/non-neurotropic” label (Abbott et al. 2010, Barabási et al. 2011).

Gut-on-chip systems provide an equally important roadmap. They allow receptor landscapes to be tested under peristalsis-like mechanics and microbiome-associated perturbation, avoiding the common mistake of modelling intestinal susceptibility in static monolayers (Kim et al. 2016). Controlled co-culture of complex anaerobic microbiomes with human epithelium in chip systems adds a missing dimension: microbiome-driven mucosal chemistry and oxygen gradients can plausibly remodel glycan presentation and inflammatory tone, thereby modifying receptor accessibility at real barriers (Jalili-Firoozinezhad et al. 2019). Across barriers, the validation logic is non-negotiable: discover receptor control signatures in spatial tissue context, stress-test causality in barrier-competent systems, verify in vivo with anatomically resolved endpoints (Marsh and Helenius 2006, Kaufmann et al. 2018, Kumar et al. 2020).

Computational integration: atlas-driven models, network controllability, and risk scoring

The receptor-centred approach becomes predictive only when receptor data are treated as nodes in a controllable network rather than as isolated variables (Barabási et al. 2011). Atlas-scale efforts such as the Human Cell Atlas and HuBMAP provide the reference scaffolding needed to compare receptor programmes across tissues and states, rather than within one privileged tissue or one outbreak (Snyder et al. 2019). The objective is to model susceptibility and systemic spread as emergent properties of: where receptors are, what they are coupled to, and which barrier compartments are accessible under real-world perturbations (Marsh and Helenius 2006, Weinbaum et al. 2007).

Infection-oriented resources are now sufficient to make receptor-centred hypotheses testable across multiple datasets, viruses, tissues, and biological contexts, rather than being “validated” once in a single model. A recent example is scMOVIR, which aggregates infection and immune-response single-cell multi-omics, supporting receptor-focused and cross-context meta-analyses across viruses and tissues (Zhang et al. 2026). Integration frameworks remain essential because receptor landscapes are distributed across modalities and batches; harmonization methods allow comparability without pretending that platform artefacts are biology (Stuart et al. 2019).

Systems control theory offers a disciplined language for why certain receptors behave like system-level levers. Network controllability formalizes how complex directed systems can be steered by acting on a subset of driver nodes, providing a rigorous analogue to the importance of entry constraints (Liu et al. 2011). Target control extends this by recognizing that clinical aims are often modular (control endothelial permeability programmes; reduce neuroinvasion risk; dampen maladaptive inflammation) rather than a global system reset (Gao et al. 2014). The practical deliverable should be receptor-based risk scoring that combines compatibility, access, and amplification (as developed in spillover logic), constrained by mechanism rather than relying on poorly interpretable predictive models (Kreuder Johnson et al. 2015, Olival et al. 2017, Plowright et al. 2017, Damas et al. 2020).

A forward-looking agenda: standardized reporting, cross-virus comparability, and clinically actionable endpoints

A receptor-centred systems virology will not scale without standardization. First, receptor measurements must be reported in ways that support reuse and comparison. The FAIR principles remain the baseline, because receptor landscapes are only useful if they can be discovered, integrated, and re-analyzed across studies (Wilkinson et al. 2016). FAIRness should also be measurable rather than performative; FAIR metrics provide an operational framework for evaluating whether datasets are genuinely reusable (Wilkinson et al. 2019). FAIRsharing provides the community infrastructure linking standards, repositories and policies, which is how minimum requirements become adoptable rather than punitive (Sansone et al. 2019).

Second, cross-virus comparability requires stable ontological scaffolding. Without consistent cell-type definitions, receptor expression in myeloid cells can describe fundamentally different entities across studies, collapsing inference into semantics rather than biology. Cell type ontologies aligned with atlas-scale single-cell efforts provide a workable route to harmonized annotation and therefore generalizable receptor models (Rozenblatt-Rosen et al. 2017, Osumi-Sutherland et al. 2021).

Finally, the agenda must remain clinically anchored. The actionable endpoint is not “the receptor is expressed” but: which barrier is at risk, under which inflammatory or iatrogenic states, and which host-directed lever is most likely to shift trajectory without collateral damage (Marsh and Helenius 2006, Weinbaum et al. 2007, Engelmann and Massberg 2013, Kaufmann et al. 2018, Kumar et al. 2020). If receptors are control nodes, the task is to map the switchboard with enough precision that intervention targets the wiring, not just the plug (Marsh and Helenius 2006, Barabási et al. 2011).

Conclusions

This review has argued for a deliberately unfashionable reframing: viral receptors are not merely docking sites that permit cell entry, but host–pathogen control nodes that couple molecular recognition to barrier behaviour, immune trajectory, and ultimately clinical phenotype. The entry-centric view is conceptually tidy yet biologically incomplete. Receptor engagement initiates a chain of decisions, some executed by the virion, many executed by the host: which entry route is licensed, which signalling circuits are triggered or suppressed, which barriers are perturbed, and which compartments become accessible to dissemination.

A receptor-centred lens clarifies why tropism is rarely reducible to “where a receptor is expressed”. Tropism reflects both receptor organization and tissue state: polarity, microdomain organization, post-translational processing, and the inflammatory or iatrogenic context that reshapes receptor availability in real tissues. In that sense, receptor landscapes are not static maps but dynamic fields, continuously modified by physiology, comorbidity, and the host’s own attempts at repair. The same receptor can therefore function as a permissive determinant in one compartment, a restriction in another, and a signalling amplifier that dominates systemic outcome elsewhere.

At the organism scale, receptors behave like distributed switchboards. Engagement at barrier surfaces governs initial amplification; engagement in immune and endothelial compartments governs routing, immunothrombosis, and multi-organ synchronization. Neuroinvasion, vascular pathology, and immune diversion become legible as consequences of receptor-defined access and receptor-proximal signalling, rather than as unrelated “complications”. The resulting picture is neither purely viral nor purely host-driven: disease emerges from the interaction topology between viral binding solutions and the host’s receptor network architecture.

The evolutionary corollary is direct. Host range, adaptation, and spillover are constrained by receptor compatibility, but realized only when compatibility intersects with barrier access and ecological amplification. This is why receptor biology is intrinsically translational: it links what is selectable in viral evolution to what is modifiable in host physiology. It also explains why receptor-targeted and receptor-adjacent interventions can be variant-resilient, yet also why they carry liabilities, because the same receptors that gate entry often regulate homeostasis.

Looking forward, the field’s most valuable output will not be another list of receptors. It will be a functional receptorome construction: topology-aware, state-resolved, and validated in barrier-competent systems, coupled to computational models that treat receptors as controllable network elements rather than isolated variables. If that programme is executed with discipline, receptor-centred systems virology can move from retrospective explanation to prospective stratification: predicting which barriers fail, which organs are at risk, and which host-directed levers can shift trajectory with acceptable collateral cost.

In short, receptors are where binding becomes biology and biology becomes disease. Treating them as control nodes is not semantic ambition; it is the most parsimonious way to connect micro-scale attachment to macro-scale pathogenesis, and to convert that connection into intervention logic.

Conflicts of interest

The authors declare no conflict of interest.

Funding

This research received funding from Techdow Pharma Italy S.r.l. Cod. 20762142078.