Viral Immunopathology

The virus damages the body in several ways. At the cellular level, viral pathogenesis occurs when the virus dismantles the cell it infects directly, through metabolic hijacking, programmed cell death, and structural disruption. Some viruses instead disable the immune system itself, by destroying its cells (HIV depletes the CD4 T cells that coordinate the response) or through virus-induced immunosuppression, leaving the host open to the secondary opportunistic infections of the immunodeficient state. Others drive unregulated proliferation, with oncogenic viruses accounting for roughly 15 to 20 per cent of human cancers worldwide. And in immunopathology, the host’s own response to the virus, rather than the virus, does the damage.

The immune response is essential to recovery from almost every viral infection, but the infiltration of lymphocytes and macrophages, the release of cytokines, and the killing of infected cells that clear the virus also damage tissue. Many familiar features of viral illness, fever, rash, oedema and lymph-node enlargement, are immunological in origin rather than caused directly by the virus.

In practice the two are not exclusive: viral disease sits on a spectrum between direct cytopathology and immunopathology, and most infections draw on both. Three observations mark out the immunopathological end. The severity of human disease does not track the cytopathic effect a virus produces in culture: some enteroviruses lyse cultured cells dramatically yet cause only inapparent human infection, while wild-type rabies virus is barely cytocidal yet uniformly lethal. The disease is ameliorated by immunosuppression, the clearest sign that the response, not the virus, is the problem. And severity that is uncoupled from the amount of virus present points to immunopathology, whereas severity that rises with the viral burden points to direct viral injury. For a non-cytocidal virus the immune response is essentially the whole disease. Where the damage falls matters as much as how much: loss of cells in skin or striated muscle is tolerated, the same loss in heart or brain is not, and oedema that is trivial in most sites is lethal when it raises intracranial pressure or floods the alveoli.

Mechanisms of viral immunopathology:

Mechanism	Principal effector	Paradigm virus	Resulting disease
Cytokine storm and hyperinflammation	Innate cytokines, macrophages	SARS-CoV-2, avian influenza, Ebola	ARDS, endothelial leak, shock, HLH
Antibody-dependent enhancement	Sub-neutralising antibody, Fc receptors	Dengue	Haemorrhagic fever, shock
Cytotoxic T cell injury (Type IV)	CD8 T cells, macrophages, cytokines	Hepatitis B and C	Chronic hepatitis, cirrhosis
Immune complex disease (Type III)	Antigen-antibody complexes, complement	Hepatitis B and C	Glomerulonephritis, polyarteritis nodosa, cryoglobulinaemia
Molecular mimicry and autoimmunity	Cross-reactive antibody and T cells	Measles, influenza, Epstein-Barr virus	Guillain–Barré syndrome, encephalomyelitis, multiple sclerosis

Cytokine storm and hyperinflammation

The response to an infection is normally scaled to its extent and switched off once the infection resolves. In a cytokine storm it does neither. A cytokine storm, also called a systemic inflammatory response syndrome, is a massive, self-amplifying release of inflammatory cytokines, principally interleukin-6, interleukin-1, tumour necrosis factor, interleukin-8 and the interferon-induced chemokine IP-10, together with stress mediators, that overwhelms the host. It is most prominent in emerging and zoonotic infections, and is thought to have driven the exceptional lethality of the 1918 influenza pandemic, which killed healthy young adults with vigorous immune systems out of proportion to the frail. The damage takes a few recognisable forms according to where the storm is concentrated.

Severe pneumonia and acute respiratory distress syndrome

In the severe respiratory viruses, infection of alveolar macrophages and lung epithelium triggers a massive influx of immune cells whose cytokine output destroys the alveolar architecture, floods the air spaces with fluid, and produces acute respiratory distress syndrome (ARDS): diffuse alveolar damage, hyaline membranes and refractory hypoxaemia.

In susceptible people, severe COVID-19 follows this course, with raised interleukin-6 and ferritin marking the hyperinflammatory phase and a tendency to microvascular thrombosis (immunothrombosis) compounding the lung injury. The same immune mechanism, rather than direct viral replication, drives the lethal pneumonias of the highly pathogenic avian influenza strains (H5N1, H7N9), whose human case-fatality has exceeded half of those reported, and of pandemic influenza. In infants the same logic produces respiratory syncytial virus bronchiolitis, where the immune infiltrate and the mucus it provokes, not viral destruction, obstruct the narrow airways.

Systemic hyperinflammation and endothelial leak

When the storm is systemic rather than confined to the lung, it attacks the vascular endothelium, and the result is capillary leak, hypovolaemic shock and haemorrhage. A surge of pro-inflammatory mediators makes the small vessels leak plasma, while activation of the coagulation cascade consumes clotting factors and produces disseminated intravascular coagulation. This is the shared final pathway of the viral haemorrhagic fevers.

Severe dengue is the prototype, its sudden plasma leakage and shock driven by a cytokine surge, complement activation, and the viral NS1 protein acting directly on the endothelium; the antibody event that sets this off is taken up under enhancement below. The filoviruses Ebola and Marburg cripple early antiviral defence in dendritic cells and macrophages while triggering a catastrophic release of mediators and tissue factor that drives both shock and disseminated intravascular coagulation. The hantaviruses concentrate the same endothelial injury in one organ, the lung in hantavirus cardiopulmonary syndrome and the kidney in haemorrhagic fever with renal syndrome; they infect the endothelium without killing it, entering through beta-3 integrins and leaving it hyper-responsive to vascular endothelial growth factor (VEGF), so the leak is mediator-driven rather than lytic. The pathophysiology is shared across these agents, even as the target organ and the clinical picture differ.

Not every severe infection in this group is inflammatory overshoot. In Ebola and Lassa fever, lethal disease reflects a failure to mount effective adaptive immunity, with uncontrolled viral replication and high viraemia, as much as an excess of inflammation: the severity there is driven partly by too little useful immunity rather than too much.

Haemophagocytic lymphohistiocytosis

The extreme of the hyperinflammatory axis is virus-associated haemophagocytic lymphohistiocytosis (HLH), and its rheumatological counterpart the macrophage activation syndrome (MAS): a relentless activation of CD8 T cells and macrophages that begin engulfing the host’s own blood cells (haemophagocytosis) in the marrow, spleen and liver, with very high ferritin and soluble CD25, falling blood counts and hepatosplenomegaly. Epstein-Barr virus is the commonest viral trigger, infecting T or natural killer cells and driving continuous cytokine production, often in a host with an underlying defect of the perforin-dependent killing that normally terminates the response. Cytomegalovirus is another, particularly in the immunocompromised.

A distinct, delayed form of hyperinflammation appears several weeks after SARS-CoV-2 infection: the multisystem inflammatory syndrome in children (MIS-C), with fever, shock, myocardial dysfunction and mucocutaneous features resembling Kawasaki disease. It is a post-infectious immune phenomenon, separate from acute COVID-19, occurring once the acute infection has resolved.

Antibody-dependent enhancement and vaccine-associated enhanced disease

Antibody usually protects, but at the wrong concentration or specificity it can make disease worse. In antibody-dependent enhancement (ADE), antibody binds a virion without neutralising it and then, through Fc receptors, ferries the virus into monocytes and macrophages, increasing rather than reducing the infected-cell burden.

Dengue is the defining example. Four dengue serotypes circulate, and antibody raised against a first infection wanes over time to sub-neutralising levels that, on a second heterotypic infection, enhance it rather than block it. The enhanced infection of macrophages is what triggers the cytokine surge and endothelial leak of dengue haemorrhagic fever and dengue shock syndrome described above. The risk is sharply titre-dependent: severe disease peaks within a narrow window of pre-existing antibody, while high titres are protective.

The vaccine corollary is stark, because a vaccine can prime enhancement just as a first infection does. The CYD-TDV dengue vaccine (Dengvaxia) behaves like a silent primary infection: given to dengue-seronegative recipients it raised their risk of hospitalisation and severe dengue on subsequent natural exposure, while protecting those already seropositive. Dengue vaccination is therefore now gated on prior dengue serostatus.

The historical precedent is the formalin-inactivated vaccines of the 1960s. Children who received inactivated respiratory syncytial virus or inactivated measles vaccine developed worse disease than the unvaccinated on later natural infection, through a Th2-skewed, non-neutralising, complex-and-enhancement mechanism now called vaccine-associated enhanced respiratory disease. A related antibody mechanism, antibody-dependent cellular cytotoxicity, in which antibody bound to viral antigen on an infected cell recruits Fc-receptor-bearing killer cells and fixes complement to lyse it, is readily shown in the laboratory, but its contribution to human viral immunopathology remains uncertain.

Cytotoxic T cell immunopathology (Type IV)

Cell-mediated immunity clears most viral infections, but a cytotoxic T lymphocyte (CTL) kills any cell displaying viral peptide whether or not the host can spare it, so the same response that controls infection can become the disease. This is delayed-type, or Type IV, hypersensitivity directed at viral antigen, and it involves both the CTL itself and the macrophages and proinflammatory cytokines drawn to the site.

Hepatitis B and C are the great human examples, and the damage is characteristically chronic. Neither virus is cytocidal: the liver injury, the inflammation, the progressive fibrosis and the cirrhosis that follows years of infection all come from the sustained CD8 T cell attack on infected hepatocytes rather than from the virus. Transgenic mice expressing hepatitis B envelope proteins show little liver damage until they are given hepatitis B-specific CTL, which reproduce the hepatocyte necrosis and inflammatory infiltrate of viral hepatitis.

In patients the degree of liver damage is governed by the strength of the response: profound immunosuppression allows high viral loads with little injury, a strong response clears the virus, and an intermediate response produces continuing damage alongside persistence. Because the damage is the response, restoring immune competence can precipitate it, as in hepatitis B and HIV co-infected patients who flare as antiretroviral therapy partially reconstitutes their immunity.

The principle is general. The classic experimental model is lymphocytic choriomeningitis virus (LCMV) in adult mice inoculated into the brain: the virus replicates harmlessly in the meninges and choroid plexus for about a week until the CTL response matures, at which point it causes fatal meningitis, cerebral oedema, convulsions and death, an outcome that knockout and T cell transfer experiments confirm depends on the CTL. The double-edged nature recurs in influenza pneumonia in mice, where transferred CD8 CTL protect but CD4 Th1 cells accelerate death, and in Coxsackie B myocarditis, where perforin-deficient mice develop only mild disease.

Two further human examples extend the principle. In HTLV-1-associated myelopathy (tropical spastic paraparesis), cytotoxic T cells and the cytokines they release damage the long corticospinal tracts of the spinal cord, a chronic immunopathology rather than direct viral destruction. The maculopapular rash of measles is itself immune-mediated: the virus infects endothelial cells in the superficial dermis and spreads to the overlying epidermis, and the rash is the vascular dilatation and the CD4, CD8 and macrophage infiltrate that follow. Patients with deficient cellular immunity may develop measles with no rash at all, and their disease is more severe rather than milder.

Immune complex disease (Type III)

Antigen-antibody complexes drive inflammation and tissue damage, the Type III hypersensitivity reaction, and their fate depends on the ratio of antibody to antigen. In antibody excess each antigen is coated and removed by macrophages bearing receptors for the antibody Fc region. At rough equivalence, large lattices form and are cleared quickly by the reticuloendothelial system. The damaging situation is antigen excess.

Antigen excess is exactly what a persistent infection produces: viral antigen is released into the blood continuously while the antibody response is weak, of low avidity, or non-neutralising. Complexes are then deposited in small blood vessels and renal glomeruli over weeks, months or years, impairing glomerular filtration and producing chronic glomerulonephritis. The model is LCMV in mice infected in utero or as neonates: persistent antigenaemia plus small amounts of non-neutralising antibody yields complexes that coat the glomerular membranes, ending in glomerulonephritis, uraemia and death.

Deposition elsewhere explains a range of human syndromes. Complexes lodged in the small vessels of skin, joints and choroid plexus activate complement and attract macrophages, producing the prodromal rashes of the exanthematous infections; heavier deposition gives erythema nodosum; and involvement of small arteries gives polyarteritis nodosa (the older term periarteritis nodosa), occasionally seen in hepatitis B. Hepatitis C drives mixed cryoglobulinaemia, in which cold-precipitable complexes cause a small-vessel vasculitis with purpura, arthralgia, peripheral neuropathy and membranoproliferative glomerulonephritis.

The systemic effects of complex formation are familiar as the viral prodrome itself. Soluble mediators mobilised by antigen-antibody complexes produce fever, malaise, anorexia and lassitude, with the fever attributed to interleukin-1, tumour necrosis factor and the interferons. Rarely, a systemic complex reaction activates the coagulation cascade, depositing fibrin in the kidneys, lungs, adrenals and pituitary and producing disseminated intravascular coagulation with infarcts and haemorrhage.

Molecular mimicry and post-infectious autoimmunity

Some viral diseases outlast the virus that triggered them, when the response cross-reacts with the host’s own tissue. Autoantibodies are detectable in many viral infections, usually transient and at low titre, and the cross-reactivity is not rare: in one survey 4 per cent of a large panel of antiviral monoclonal antibodies also bound normal tissue, and a monoclonal against Coxsackievirus B4 bound cardiac muscle, the very tissue that virus targets in myocarditis.

The leading explanation is molecular mimicry. Viral and host proteins share identical or near-identical stretches of 6 to 10 amino acids far more often than chance predicts, so a response raised against the virus can cross-react with a self protein. Myelin basic protein, a major constituent of the nerve sheath, shares such homology with several viral proteins.

Guillain–Barré syndrome is an acute immune-mediated polyradiculoneuropathy that follows infection, classically with Campylobacter but also after a range of viral infections, among them cytomegalovirus, Epstein-Barr virus, varicella-zoster, influenza, Zika, West Nile and chikungunya, and historically with the 1976 swine-influenza vaccine. In that episode, mimicry between a peripheral-nerve myelin epitope and the influenza NS2 protein, which is normally removed during vaccine purification but was retained in some batches, may explain the excess of cases.

Post-infectious, or acute disseminated, encephalomyelitis (ADEM) is a rare demyelinating illness, roughly one case in a thousand, arising a few weeks after measles and more rarely after varicella, mumps, rubella or vaccinia. The pathology is demyelination without the neuronal degeneration that direct viral infection of the brain produces; virus cannot be recovered from the brain; and myelin basic protein, anti-myelin antibody and reactive T cells appear in the patient, all pointing to autoimmunity rather than direct viral injury.

The strongest recent evidence comes from multiple sclerosis. A large longitudinal cohort study found that the risk of multiple sclerosis rose roughly thirtyfold after seroconversion to Epstein-Barr virus, with no comparable rise after other infections including the similarly transmitted cytomegalovirus, and with a marker of neuro-axonal damage rising only after Epstein-Barr virus infection. This is the strongest evidence yet that the virus is a necessary, though not sufficient, trigger for the disease. Enteroviral infection of the insulin-producing islet cells of the pancreas is a long-standing candidate trigger for type 1 diabetes through similar mechanisms, although causation is not proven.

Mimicry is not the only route to autoimmunity, and the principal proposed mechanisms are set out below.

Potential mechanisms of virus-induced autoimmune disease:

Mechanism	How it breaks tolerance
Molecular mimicry	A viral epitope cross-reacts with an identical or similar host epitope, so the antiviral response damages self tissue
Bystander activation	Local inflammation liberates sequestered self-antigen and supplies an adjuvant effect, activating low-affinity autoreactive T cells; the response may then widen by epitope spreading
Immortalisation of autoreactive cells	A virus such as Epstein-Barr virus drives polyclonal expansion of autoreactive lymphocytes or antigen-presenting cells
Dual T cell receptors	An autoreactive T cell that should have been deleted in the thymus is activated when its second receptor meets a viral antigen
Cytokine-induced MHC expression	Interferon gamma and tumour necrosis factor induce MHC class II on cells that do not normally present antigen, such as glial cells, letting them present self proteins like myelin
Loss of regulatory T cells	The virus destroys or downregulates the regulatory T cells that normally suppress responses to self

Autoimmune damage of this kind can, in principle, continue long after the triggering virus has been cleared: once cross-reactive injury exposes further sequestered self proteins, the response can spread to them and become self-sustaining. This is the proposed link between common viral infections and the chronic human autoimmune diseases, from multiple sclerosis to rheumatoid arthritis, though for most of them definitive proof of a viral trigger is still lacking.

Clinical and therapeutic significance

Recognising that a disease is immune-mediated changes how it is read and how it is treated. Three features point to immunopathology: a severity out of proportion to the viral burden, a non-cytocidal virus producing organ damage, and a clinical response to immunosuppression. The corollary is that the treatment may need to target the response rather than, or as well as, the virus.

That logic underlies much of the management of severe viral disease. Corticosteroids reduce mortality in the hypoxaemic, hyperinflammatory phase of COVID-19, and interleukin-6 receptor blockade with tocilizumab adds benefit in the most inflamed patients; haemophagocytic lymphohistiocytosis is treated with immunosuppression and, when severe, etoposide-based protocols; intravenous immunoglobulin is the mainstay in the multisystem inflammatory syndrome in children and in Guillain–Barré syndrome, with plasma exchange an alternative in the latter. In each case the target is the immune response, not the pathogen.

The central difficulty is that the same response is also clearing the virus, so immunosuppression is a balance struck against the clock. Suppress too early or too hard, while the virus is still replicating freely, and the infection worsens; this is why corticosteroids help in late hypoxaemic COVID-19 but not in early mild disease, and why antiviral control and the timing of immunomodulation both matter. The preventive lesson runs the other way: because sub-neutralising antibody can enhance disease, a viral vaccine must elicit genuinely neutralising rather than merely binding antibody, and screening for enhanced disease, together with the serostatus-gating of dengue vaccination, is now built into vaccine development.

References and recommended reading

• 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 spectrum of disease production, cytotoxic T cell and immune-complex immunopathology, the cytokine storm, and the mechanisms of virus-induced autoimmunity (Table 7.8).
• Morrison TE, Heise MT. Pathogenesis of Viral Infection. In: Fields Virology, 7th edition, Volume 4, Chapter 8. Wolters Kluwer; 2023. The current reference for the integrated virus-host framework of disease production, the routes of immune-mediated injury (cytotoxic T cell clearance, immune complexes, antibody-dependent enhancement and autoimmunity), and the discrimination of immunopathology from direct cytopathic damage.
• Yang B, Yang KD. Immunopathogenesis of Different Emerging Viral Infections: Evasion, Fatal Mechanism, and Prevention. Frontiers in Immunology 2021;12:690976. DOI 10.3389/fimmu.2021.690976. A current synthesis of the routes to fatal emerging-virus disease, distinguishing immune deficiency with viraemia, direct cytotropic injury, augmented immunopathology, and antibody-dependent enhancement.