Viral Immune Evasion

Almost every virus of medical importance has evolved elaborate, often redundant strategies to circumvent host immunity, and the sophistication of those strategies is itself the clearest measure of how dangerous the immune response is to a virus. Host cells have in turn evolved countermeasures, and the viruses countermeasures to those, producing a genetic arms race in which each side is continually selected against the other. The strategies fall into a recognisable catalogue: blocking the sensors and signalling that raise the innate alarm, neutralising the interferon response and the host restriction factors, hiding infected cells from cytotoxic T cells and antibody, and varying or concealing the antigens the adaptive response is trying to track. The catalogue is not abstract. Successful evasion is precisely what allows a virus to persist, so the same mechanisms that defeat immunity also explain latency, chronic carriage, reactivation under immunosuppression, and several of the therapeutics now used against persistent infection and virus-associated cancer.

Evading innate immunity

The innate response is fast, broadly specific, and continuously present, so a virus that cannot blunt it loses the race in the first hours of infection. Three broad tactics recur:

1. Degrading the signalling components of pattern-recognition pathways.
2. Interfering with the transcription factors those pathways activate.
3. Mimicking host proteins so as to jam the machinery from inside.

Subverting pattern-recognition signalling

Because it senses double-stranded RNA, a hallmark of viral replication, the Toll-like receptor 3 (TLR3) pathway is a frequent target. Its adaptor protein TRIF (TIR-domain-containing adaptor-inducing interferon beta) is singled out for degradation by viral proteases, an unusually effective move because losing TRIF disables both arms downstream of it: the inflammatory transcription factor NF-kappa-B and the interferon transcription factor IRF3 (interferon regulatory factor 3).

The serine protease NS3/4A of hepatitis C virus and the 3C protease of several picornaviruses both cleave TRIF and so attenuate the antiviral response. Targeting TRIF is not confined to RNA viruses: TRIF levels fall sharply during gammaherpesvirus infection. The same hepatitis C virus protease also cleaves MAVS (mitochondrial antiviral signalling protein, also called IPS-1), the adaptor on which the cytosolic RNA sensors RIG-I and MDA5 depend, so a single viral enzyme disables pattern-recognition sensing at two separate points.

Other viruses block further downstream, inhibiting the IKK (inhibitor-of-kappa-B kinase) complex that several activation pathways converge on, or interfering with ubiquitination, the protein-tagging process that controls the turnover and location of these signalling intermediates.

Blocking interferon induction and signalling

Even when sensing succeeds, viruses interrupt the interferon response at either end: the production of interferon or its action on neighbouring cells.

The non-structural protein NS4B of dengue virus blocks synthesis of interferon alpha and beta by inhibiting the JAK–STAT pathway (the Janus kinase and signal transducer and activator of transcription cascade that interferon receptors signal through), specifically by preventing phosphorylation of STAT1. The downstream consequence is failed production of the interferon-induced transmembrane proteins that would otherwise restrict viral entry.

Many viruses encode proteins that block one or another of the sequential steps of the interferon response in this way, a tactic central to persistence in hepatitis C virus and HIV but also deployed in several transient infections, including SARS, West Nile, dengue, and Ebola.

Molecular mimicry of innate sensors

A complementary strategy is to supply a counterfeit of a key signalling component. The V protein of paramyxoviruses mimics IRF3 and acts as a non-functional substrate for the kinases that would normally activate IRF3, thereby jamming the pathway triggered when TLR3 binds viral RNA.

The poxviruses are the best-studied exponents: their A49 protein blocks induction of type I interferon by abolishing activation of NF-kappa-B, the transcription factor required to switch on the interferon genes.

Large DNA viruses extend this idea to the extracellular space, secreting virokines and viral cytokine-binding proteins that imitate or mop up host cytokines and chemokines before they can organise a response.

Manipulating apoptosis and autophagy

A heavily infected cell that kills itself denies the virus its factory, so many viruses actively preserve the cells they occupy. For several viruses, maintaining and expanding the endoplasmic reticulum is essential for assembly and maturation while avoiding the stress-induced apoptosis that would otherwise follow. Dengue virus and other flaviviruses achieve this by carefully controlling the cellular unfolded protein response, the stress programme triggered by an overloaded endoplasmic reticulum, preventing cytopathicity yet permitting the membrane rearrangements that expand the reticulum into a replication platform.

Dengue virus also subverts autophagy, the pathway by which a cell digests its own components. Autophagosomes normally fuse with lysosomes to degrade their contents; dengue virus blocks that fusion and shelters within the autophagosomes, using them as protected sites for RNA replication.

Countering natural killer cells

Natural killer cells present viruses with a dilemma. Downregulating MHC class I to hide from cytotoxic T cells is exactly the missing-self signal that licenses a natural killer cell to attack. Viruses, the herpesviruses above all, resolve the dilemma with a dedicated set of counter-tactics rather than abandoning MHC downregulation.

Viral mechanisms for evading natural killer cells:

Mechanism	Examples	Effect
MHC class I homologues	Herpesviruses	Engage inhibitory receptors, switching the natural killer cell off
Selective regulation of MHC class I on the target cell	Herpesviruses, SIV	Avoid missing-self without exposing peptide to cytotoxic T cells
Proteins that disrupt activating receptor–ligand pairs	Herpesviruses, HIV, HTLV	Block natural killer cytotoxicity and interferon gamma release
Decoy cytokine-binding proteins or chemokine antagonists	Herpesviruses, papillomaviruses	Suppress interferon gamma production and cell trafficking
Direct effects on the natural killer cell	Herpesviruses, HIV, HCV	Hepatitis C virus E2 binds CD81 and dampens natural killer activity

The first two rows are the elegant centre of the problem: a viral decoy that looks like class I keeps the inhibitory receptor satisfied while the real class I molecules, carrying viral peptide, are pulled from the surface.

Host restriction factors and counter-restriction

Restriction factors are antiviral proteins expressed constitutively by many cell types that recognise an incoming virus and block a specific step of its replication. They are a defining feature of the arms race because viruses carry dedicated accessory proteins whose only job is to neutralise them, and the reciprocal evolution of the two is visible in the genomes of both host and virus.

TRIM5-alpha (tripartite-motif protein 5-alpha) is a species-specific factor that binds the incoming retroviral capsid at an early, pre-integration stage and aborts uncoating. Rhesus and cynomolgus macaque TRIM5-alpha restricts HIV-1, which is one reason the virus does not propagate in Old World monkeys, and the factor is regarded as part of the innate response to retroviral infection.

APOBEC3G (apolipoprotein B mRNA-editing enzyme catalytic polypeptide-like 3G) is a host enzyme that is packaged into budding virions and, in the next cell, deaminates cytidine to uridine on the minus strand of viral DNA as it is reverse transcribed, scattering often lethal mutations through the genome. HIV and the simian immunodeficiency viruses counter it with the accessory protein Vif (viral infectivity factor), which binds APOBEC3G and routes it for destruction through the ubiquitin-proteasome pathway before it can be packaged. The conflict is measurably ongoing: across African green monkey subspecies, simian APOBEC3G continues to evolve even in non-pathogenic infection, and the viruses adapt in turn to re-target the changing factor, a reciprocal pressure that helps explain the narrow species range of closely related retroviruses.

Tetherin, a transmembrane protein, holds fully formed enveloped virions on the cell surface so they cannot detach. HIV neutralises it with Vpu (viral protein U), which degrades and downregulates tetherin to free new virions; the same action reduces the cell’s vulnerability to antibody-dependent cellular cytotoxicity, and Vpu separately drives degradation of newly made CD4 molecules.

Evading adaptive immunity

The adaptive response brings antigen-specific recognition, antibody, cytotoxic T cells, and memory, and persistent viruses have a countermeasure for each.

Interfering with MHC and antigen presentation

Cytotoxic CD8 T cells inspect viral peptides displayed on MHC class I, and helper CD4 T cells inspect them on MHC class II, so reducing peptide display directly favours persistence. Viruses interrupt the pathway at every stage. The E1A protein of adenoviruses, the Tat protein of HIV-1, and the E5 and E7 proteins of bovine papillomavirus inhibit transcription of the genes encoding MHC components. The E7 protein of human papillomavirus type 18 downregulates transcription of TAP1, the transporter that delivers cytosolic peptides into the endoplasmic reticulum for loading. Other viral proteins block the generation of peptides (Epstein-Barr virus, human herpesvirus 8, HIV-1), their transport (herpes simplex virus, cytomegalovirus, Epstein-Barr virus), or the surface expression of MHC itself (adenoviruses, cytomegalovirus, human T-lymphotropic virus 1).

The dedicated gene cluster of human cytomegalovirus, which attacks several of these steps at once, is the reference example of the strategy, and the counter to the natural killer cell problem it creates is the decoy and inhibitory-receptor toolkit set out above.

Antigenic variation

A virus that changes the antigens the response is tracking stays one step ahead of it. In HIV, an enormous replication rate, of the order of ten billion new virions a day, combined with a reverse transcriptase that makes roughly one error in every ten thousand nucleotides copied, generates a continuous supply of escape mutants that the current antibody and T cell repertoire recognises only partly or not at all. Such mutants are frequently less fit, and do not outcompete the parent unless the immune response is holding the parent back, so variation and selection proceed together.

Influenza viruses show the population-level version of the same logic: gradual point mutation of the surface glycoproteins (antigenic drift) erodes existing immunity and forces annual vaccine reformulation, while wholesale reassortment of genome segments (antigenic shift) can introduce a glycoprotein the population has never seen and seed a pandemic.

Downregulating adhesion molecules

Recognition by a T cell depends not only on the peptide-MHC signal but on the adhesion that holds the two cells together long enough to read it. Burkitt’s lymphoma cells carrying the Epstein-Barr virus genome display reduced amounts of the adhesion molecules ICAM-1 (intercellular adhesion molecule 1) and LFA-3 (lymphocyte function-associated antigen 3), and so bind T cells with lower affinity and are correspondingly harder to engage.

Decoying antibody and complement

Antibody can be defeated by quantity and by misdirection as well as by antigenic change. The very high antibody titres typical of chronic infection include a large fraction directed at viral epitopes irrelevant to neutralisation; by binding the virion at the wrong site, these non-neutralising antibodies sterically block the access of neutralising antibody. Some viruses go further and flood the circulation with a soluble decoy: the chronic hepatitis B carrier state is marked by a vast excess of non-infectious hepatitis B surface antigen particles, which absorb neutralising antibody and may simply overwhelm the host’s capacity to produce it.

Antibody already bound to an infected cell surface can be neutralised too. Because antibodies are divalent, they can bridge adjacent viral molecules on the membrane and drive them into caps that are then internalised, stripping antigen from the surface and reducing the cell’s value as a target; this antigenic modulation is readily shown in persistently infected cell cultures and is implicated in subacute sclerosing panencephalitis, the late measles complication in which antibody also suppresses transcription of the viral genome.

Several herpesviruses additionally encode their own Fc receptors and complement- control proteins that intercept antibody and complement before they can trigger lysis.

Subverting cytokine actions

Beyond the interferon-blocking already described, viruses neutralise other antiviral cytokines. An adenoviral gene product partially shields the infected cell from tumour necrosis factor, and the secreted decoy receptors and binding proteins of the large DNA viruses extend the principle across the cytokine network, buying time before an effective response can be mounted.

Driving T cell exhaustion

Persistent antigen does not always provoke an escalating response; often it does the opposite. Under the continuous stimulation of chronic HIV, hepatitis C, and hepatitis B infection, virus-specific T cells enter a state of exhaustion, progressively upregulating inhibitory receptors such as PD-1 (programmed cell death protein 1) and losing their cytotoxic and cytokine functions. Exhaustion spares the host the immunopathology of an unrelenting response but cedes control of the virus, and it is the rationale for the checkpoint-inhibitor antibodies, which block PD-1 and its ligand, now being explored to reinvigorate exhausted T cells in chronic viral infection and in virus-associated cancer.

Establishing persistence

Most human virus infections are transient: the virus replicates, the immune response develops, and the agent is cleared, with or without symptoms along the way. Some apparent persistence is something else, a chronic bacterial superinfection such as post-viral sinusitis or bronchitis, or a succession of different viruses mistaken for one. But a growing number of viruses once thought purely acute are now known to linger in sequestered sites in trace amounts, usually silently and occasionally to clinical effect. Persistence requires two things at once: the immune system fails to eliminate the virus, and, for any virus that would otherwise kill its host cell, viral replication is restrained. The genome may persist with no complete replication cycle at all, the infection becoming evident only when a latent state reactivates later in the host’s life.

Whether persistence causes disease matters enormously to the host but often little to the virus, whose survival in nature depends on shedding and onward spread rather than on the fate of any one host. Long-term persistence is the outcome of a balance: the virus’s capacity to keep replicating in, or simply residing within, long-lived cells, set against its ability to counter both arms of immunity. Where the balance settles varies by agent. The herpesviruses and HIV persist invariably; the common respiratory viruses are almost always cleared; hepatitis B virus goes either way depending chiefly on the age at exposure. Immunosuppression shifts the whole balance toward the virus, so that infections that are normally transient may persist and infections that are normally localised or latent may disseminate.

Mechanisms of ineffective immune responses in persistent infection:

Category	Mechanism	Examples
Reduced antigen expression	Restricted viral gene expression	Herpes simplex virus (neurons); Epstein-Barr virus (B cells); HIV-1
Evasion of the response	Cell-to-cell spread by fusion	HIV-1, measles, cytomegalovirus
	Sequestration in sanctuary sites	HIV-1 (brain); cytomegalovirus and polyomaviruses (kidney); papillomaviruses (skin)
	Antigenic drift	HIV-1, visna, hepatitis C virus
Immunosuppression	Infection of effector cells; polyclonal B cell activation	HIV (CD4 T cells), Epstein-Barr virus
	Loss of macrophage function	HIV
Tolerance	Congenital or neonatal infection induces T cell non-responsiveness	Rubella, cytomegalovirus, LCM virus, hepatitis B, parvovirus B19
Reduced antigen display	Antibody stripping; MHC downregulation	Measles (SSPE); adenoviruses
	Viral Fc receptor blocks immune lysis	Herpesviruses
Antibody subversion	Non-neutralising antibody; soluble decoy antigen	Many viruses; hepatitis B virus
Cytokine evasion	Interference with interferon and tumour necrosis factor	Adenoviruses, hepatitis C virus

Several of these mechanisms operate in transient infections too, and in many cases the association with persistence is established without a proven causal link. The sections that follow take the principal categories in turn.

Restricting viral gene expression

A virus cannot persist in a cell it destroys, so long-term carriage of a potentially cytocidal virus is possible only if the genome stays wholly or partly silent, or if newly made virus can keep finding fresh uninfected cells.

Latent herpesviruses take the first route: only a few early genes are transcribed during latency, and the silence is broken, often after immunosuppression or under a cytokine or hormone that de-represses the genome, allowing the full programme to resume and the virus to reactivate. HIV illustrates the second: virus production continues in T lymphocytes and kills them, while a parallel wave of proliferation among uninfected lymphocytes supplies fresh targets for the next round.

Restricted expression has a second payoff beyond survival of the host cell: with few viral genes transcribed, few viral antigens are presented, and the infected cell becomes nearly invisible to immune surveillance. A latently infected sensory neuron carrying herpes simplex virus displays no viral antigen at all, which shields it not only from cytotoxic T cells but from antibody, complement, and antibody-dependent cellular cytotoxicity alike.

Latency in non-permissive, resting, or undifferentiated cells

A given virus may replicate productively in one cell type yet establish a quiet, non-productive infection in another, so that the second cell type becomes a reservoir that reseeds the first on reactivation. Epstein-Barr virus replicates productively in mucosal epithelium but takes up latency in B lymphocytes.

Even within one cell type, permissiveness can hinge on the cell’s state of differentiation or activation. Papillomaviruses complete their replication cycle only in fully differentiated keratinocytes; HIV replicates in CD4 T cells once a cytokine has activated them but lies latent in resting CD4 T cells, and maintains a separate, less cytopathic relationship with the monocyte and macrophage lineage, one of several reservoirs that make the infection so difficult to clear.

Non-cytocidal infection

Some viruses simply do not kill the cells they infect, which removes the main obstacle to persistence. Arenaviruses and certain retroviruses establish lifelong, non-cytocidal infection in their natural reservoir hosts, sometimes with no disease and sometimes with a carriage that wanes slowly enough to require transmission to the next generation of hosts to complete the cycle.

Junin virus, the cause of Argentine haemorrhagic fever, persists without disease in its reservoir, the vesper mouse Calomys musculinus, shedding for long periods in urine and saliva; aerosol and fomite transmission to agricultural workers, particularly at harvest, produced years of severe human outbreaks until a vaccine reduced the incidence.

Hepatitis B virus is non-cytocidal in the hepatocyte, and the extent of liver damage tracks not the virus but the immune attack on infected cells, the same immune-mediated injury that places chronic hepatitis B among the immunopathological diseases.

Cell-to-cell spread and anatomical sanctuaries

A virus that never enters the extracellular space cannot be reached by circulating antibody or, in transit, by cytotoxic T cells. Lentiviruses such as HIV, paramyxoviruses such as measles, and herpesviruses such as cytomegalovirus can fuse adjacent cells, letting the genome travel contiguously from one cell to the next while remaining shielded throughout.

Other viruses persist by retreating into sites the immune system patrols poorly. The brain is largely closed to lymphocyte traffic by the blood–brain barrier, and its neurons express very little MHC, so they are poorly lysed by cytotoxic T cells; herpesviruses, polyomaviruses, and lentiviruses all exploit this, and HIV is sheltered in the brain and in the epididymis. The kidney is the other recurrent sanctuary, harbouring cytomegalovirus and the JC and BK polyomaviruses, which are not acutely cytopathic and are released across the luminal surface of polarised epithelium, away from immune surveillance.

Papillomaviruses reach an extreme of inaccessibility: infection begins in the basal layer of the epidermis but infectious virions are assembled only in the fully differentiated outer layers, where immune surveillance is minimal. That the immune system nonetheless retains some control is shown by the increase in warts under immunosuppression and by the tendency of multiple warts to regress together when control is restored.

Antigenic drift in persistence

Within a single chronically infected host, the same error-prone, high-turnover replication that drives HIV variation at the population level generates a moving target for the immune response, so that escape mutants displace one another in step with the antibody and T cell specificities of the moment. The continual regeneration of variants the current response cannot grip is one of the central reasons HIV is never cleared.

Virus-induced immunosuppression and tolerance

Infection usually elicits a response that secures recovery, but a number of viruses turn instead to suppressing that response, either broadly across many antigens or selectively against their own. Three routes account for most of it:

1. Infecting the effector cells of the immune system.
2. Inducing tolerance to viral antigens.
3. Interfering directly with immune function through viral products, the last of which has already been met as the MHC, antibody, and cytokine evasion above.

Infection of immune effector cells

The cells of the reticuloendothelial system are tempting targets: they travel throughout the body, can seed virus to any organ, and are central to the response itself. HIV is the dramatic case, replicating in CD4 T lymphocytes, dendritic cells, and the monocyte and macrophage lineage. In macrophages the cytopathic effect is slight, but phagocytosis, antigen processing and presentation, and cytokine production are impaired; in CD4 T cells the integrated provirus replicates only once the cell is activated, and the cells are lost to apoptosis, to fusion into short-lived syncytia, and to lysis by CD8 T cells. The falling CD4 count, normally between 500 and 1500 cells per cubic millimetre, tracks this collapse clinically, and its end point, the near-total loss of CD4 help, leaves the untreated patient to die of opportunistic infection or cancer.

Other viruses suppress immunity more transiently. Epstein-Barr virus drives a polyclonal activation of B cells that diverts the response into irrelevant activity.

Measles virus replicates non-productively in T lymphocytes and suppresses the Th1 cells that mediate delayed-type hypersensitivity; the observation, made by von Pirquet in 1908, that measles depresses the tuberculin skin reaction and can reactivate latent tuberculosis remains the classic demonstration. The accompanying defects, a depressed delayed-type hypersensitivity response, reduced natural killer activity, raised plasma IgE, and a raised soluble interleukin-2 receptor, can persist for up to four weeks after the rash and account for much of the susceptibility to the secondary infections that cause most measles mortality.

Induction of immunological tolerance

The likelihood that an acute infection becomes chronic is strongly age-dependent, and congenital or perinatal transmission is the strongest single predictor of persistence, as seen with hepatitis B virus, rubella virus, cytomegalovirus, and parvovirus B19. The immaturity of the response at the time of first encounter can produce a split tolerance, in which a B cell response is mounted but the corresponding T cells remain unresponsive.

The experimental model is lymphocytic choriomeningitis virus: mice infected in utero mount no T cell response to the virus, and the antibody they do make scarcely affects its course, yet the persistent infection can be reversed by transferring sensitised CD8 T cells from an immune animal, which both demonstrates the tolerance and confirms the central role of CD8 T cells in clearing the infection.

Clinical and therapeutic significance

The evasion catalogue maps directly onto the clinic. The strength of a virus’s MHC and natural killer evasion predicts its capacity for latency and lifelong carriage, which is why the herpesviruses reactivate predictably whenever T cell control lapses, in transplantation, advanced HIV, or high-dose corticosteroid therapy. Sanctuary-site persistence explains why JC virus surfaces as progressive multifocal leukoencephalopathy when surveillance of the brain fails, and why cytomegalovirus and the polyomaviruses re-emerge from the kidney in transplant recipients.

Antigenic variation sets the tempo of clinical practice from the other direction: antigenic drift is the reason influenza vaccine is reformulated each year, and within-host variation in HIV and hepatitis C is a standing obstacle to a sterilising vaccine. T cell exhaustion, once seen only as failure, has become a therapeutic target, with checkpoint-inhibitor antibodies aiming to lift the brake on exhausted antiviral T cells.

Read the other way, the same logic warns that any intervention which weakens a given arm of immunity will release the particular viruses that arm was holding in check, which is the unifying principle behind the reactivation syndromes of the immunosuppressed.

References and recommended reading

• Burrell CJ, Howard CR, Murphy FA. Innate Immunity. In: Fenner and White’s Medical Virology, 5th edition, Chapter 5. Academic Press / Elsevier; 2017. The principal source for viral evasion of pattern-recognition signalling, the interferon system, natural killer cells (Table 5.2), and the host restriction factors and their viral counter-restriction.
• Burrell CJ, Howard CR, Murphy FA. Pathogenesis of Virus Infections. In: Fenner and White’s Medical Virology, 5th edition, Chapter 7. Academic Press / Elsevier; 2017. The principal source for the mechanisms of persistence, restricted gene expression and latency, sanctuary-site sequestration, antigenic drift, virus-induced immunosuppression and tolerance (Table 7.7).
• Iwasaki A, Schoggins JW, Hur S, et al. Innate Immunity to Viruses; The Adaptive Immune Response to Viruses. In: Fields Virology, 7th edition, Chapters 9 and 10. Wolters Kluwer; 2023. The current reference for restriction-factor biology, interferon-pathway antagonism, MHC and antigen-presentation interference, and CD8 T cell exhaustion.
• Sompayrac LM. How the Immune System Works, 6th edition. Wiley-Blackwell; 2019. Plain-language explanations of the missing-self problem and the logic of immune evasion.