Cross-reactive T cells
Examples of cross-reactive epitopes exacerbating disease in humans can be found when viruses of different serotypes sequentially infect a host. For instance, the pathogenesis of dengue haemorrhagic fever (DHF) is thought to be mediated by cross-reactive T cells to different dengue virus serotypes, which exacerbate pathogenic immune responses (Sangkawibha et al., 1984; Mathew et al., 1998; Guzman et al., 2000; Mongkolsapaya et al., 2003). There is clinical evidence that, as with dengue virus serotypes, T cell cross-reactivity elicited by completely unrelated pathogens may also influence the outcome of disease. For example, T cells that react to epitopes common to influenza A virus and hepatitis C virus (HCV) influence the outcome of HCV infection. A subset of healthy, non-HCV-infected blood donors have HCV NS3-specific memory T cells that recognize a determinant of the neuraminidase protein of influenza A virus, suggesting that exposure to influenza may lead to anti-HCV memory responses in HCV-uninfected individuals (Wedemeyer et al., 2001). Interestingly, patients who develop severe acute HCV infection have unusually strong CD8 T cell responses to the HCV NS3 epitope, which cross-reacts with the influenza virus neuraminidase sequence (Urbani et al., 2005). These findings suggest that exposure to influenza leads to CD8 T cell cross-reactivity, which influences the severity of HCV-associated liver pathology during acute infection. This is consistent with the evidence from murine models showing that expansion of pre-existing cross-reactive memory CD8 T cells can modify the primary immune response and modulate the immunopathologic response to subsequent infection (Fig. 1, III).
Polarized immune responses
In general, polarized immune responses are more clearly demarcated in inbred mice than in people (Mestas and Hughes, 2004), but there are several clinical examples of human pathogens that induce polarized immune responses. For instance, patients co-infected with shistosomiasis and HCV predominantly mount type 2 immune responses and have weak HCV-specific CD4 responses compared with patients infected with HCV alone (Kamal et al., 2001a,b). These findings are associated with higher HCV viral loads and lower viral clearance, which may explain the higher incidence of cirrhosis and hepatocellular carcinoma in co-infected patients compared with those infected with HCV alone (Angelico et al., 1997; Kamal et al., 2000; 2001b).
Infection with schistosomiasis may also influence the immune response to HIV. S. mansoni co-infected adults have decreased Gag-specific CD8+ cytolytic T cell responses and increased number of Gag-specific IL-10 positive CD8+ T cells (McElroy et al., 2005). Given the deleterious effect of type 2 responses on HIV pathogenesis, it has been suggested that therapy against helminthic infections and schistosomiasis may slow HIV progression (Fincham et al., 2003). However, a recent study from Uganda showed that treatment of S. mansoni was associated with transient increases in viral load and sustained decreases in CD4+ T lymphocyte count, which were associated with enhanced post-treatment S. mansoni-specific type 2 responses (Brown et al., 2005).
Interestingly, the ability to mount polarized immune responses may be partly determined in utero. A cross-sectional study of mother–child pairs in Kenya showed that prenatal sensitization to filariae and schistosomes occurs in ∼50% of newborns and is relatively long-lasting, persisting into childhood (Malhotra et al., 1999). BCG-vaccinated children born to mothers without filariasis or schistosomiasis produced 10 times more IFN-γ in response to mycobacterial antigens than BCG-vaccinated children of helminth-infected mothers (Malhotra et al., 1999). These data suggest that prenatal sensitization can bias immunity induced by M. bovis BCG away from protective type 1 immune responses. In adults, helminthic infections are also associated with blunted IFN-γ responses to tuberculin antigen (Ferreira et al., 2002).
These findings may have implications for vaccination strategies. The efficacy of M. bovis BCG against tuberculosis correlates with distance from the equator; greater protection has been observed at higher latitudes (Palmer and Long, 1966; Fine, 1995). Higher exposure to environmental bacteria commonly found in warm climates is thought to influence subsequent responses to BCG (Fine, 1995). Similarly, helminthic infections are more prevalent in tropical climates and may downregulate effective cell-mediated immune responses to mycobacteria (Malhotra et al., 1999; Ferreira et al., 2002). Treatment of intestinal parasites prior to BCG vaccination is associated with improved proliferation and IFN-γ production by peripheral blood mononuclear cells after stimulation with tuberculin antigens (Elias et al., 2001). However, the effect of antihelminthic therapy during pregnancy and prior to immunization on the actual efficacy of BCG vaccine has not been established.
Polarization of immune responses by vaccination may influence the outcome of future infections. Epidemiologic studies have shown that immunization with live attenuated vaccines that elicit predominantly type 1 immune responses, such as M. bovis BCG and measles vaccine had a non-specific beneficial effect on childhood survival. In contrast, diphtheria-pertussis-toxoid (DPT) vaccine, which primarily elicits type 2 immune responses, had the opposite effect (Kristensen et al., 2000; Garly et al., 2003; Shann, 2004; Roth et al., 2005). Despite controlling for the vaccine preventable infections, it was difficult to prove a true causal relationship between vaccination and survival because the specific cause of death in most of the children was not well established in these studies. However, it is tempting to speculate that the persistent immunomodulatory properties of these vaccines may impact host susceptibility to infections. Type 1-polarized immune responses may be effective against many childhood pathogens commonly encountered in the developing world (Shann, 2004).
The effect of systemic polarized responses elicited by chronic infections may have implications beyond the field of infectious diseases. In particular, the ‘hygiene hypothesis’ argues that the dramatic rise in atopic disorders observed in industrialized countries over the last few decades may be related to a decline in infectious diseases resulting from improved living standards and immunization programmes (Strachan, 1989; Yazdanbakhsh et al., 2002). The relative paucity of allergic diseases in developing countries supports this concept. Susceptibility to allergic diseases may result from a combination of genetic and environmental factors. For example, it has been demonstrated that Hepatitis A virus may protect against atopy in individuals who carry a particular variant of the gene that encodes TIM-1, a glycoprotein that co-stimulates T cell activation and has been implicated in atopic diseases (McIntire et al., 2003). It has been proposed that TIM-1, which is a receptor for hepatitis A, facilitates entry in to TIM-1 cells, leading to deletion of Th2 cells and a reduction in asthma.
Other pathogens, such as mycobacteria (Shirakawa et al., 1997), measles (Shaheen et al., 1996) and gastrointestinal organisms (Matricardi et al., 2000) have been inversely correlated with atopic responses. Immune profiles of individuals with positive tuberculin skin tests show a type 1 cytokine predominance and lower IgE serum levels (Shirakawa et al., 1997). Paradoxically, helminthic infections, which induce strong type 2 immune responses, also protect against allergic disease (Lynch et al., 1987; van den Biggelaar et al., 2004). It has been proposed that persistent immune challenge from chronic infectious pathogens, such as helminths and mycobacteria, induce an anti-inflammatory regulatory network that protects against allergic diseases (Mahanty and Nutman, 1995; Yazdanbakhsh et al., 2002).
Regulatory T cells
There is mounting evidence that chronic infections induce Treg cells that can modulate the response to bystander antigenic stimulation. In addition to protecting against allergic disease (Lynch et al., 1987; van den Biggelaar et al., 2004), heavy parasitic infections can attenuate antigen-specific T cell responses. For example, individuals infected with Onchocerca volvulus have diminished antibody production, proliferative responses and IFN-γ production following vaccination with tetanus toxoid compared with uninfected vaccinated controls (Cooper et al., 1998). Furthermore, O. volvulus-infected individuals produced significantly higher levels of IL-10 compared with uninfected controls, suggesting that Treg-mediated pathways may play an immunosuppressive role. Likewise, asymptomatic patients infected with Wucheria bancrofti mount lower proliferative and IFN-γ responses after vaccination with tetanus toxoid compared with uninfected individuals (Nookala et al., 2004). Diminished immune responses appear to be mediated by IL-10, which is secreted at higher levels in peripheral blood mononuclear cells (PBMCs) from microfilaremic patients compared with controls. Although none of these studies directly evaluate the role of Treg cells, the association between helminthic infections and anti-inflammatory cytokines such as IL-10 suggests that regulatory immunosuppressive mechanisms may be playing a role in the diminished immune response to concurrent infections.
Priming innate responses
As predicted by experimental data showing enhanced HIV replication in co-infection models, opportunistic infections can accelerate the clinical progression of HIV (Perneger et al., 1995; Whalen et al., 1995). For example, infection with M. tuberculosis leads to higher HIV viral loads and faster CD4 declines (Toossi et al., 1993; Goletti et al., 1996). In general, activation of T cells, macrophages and dendritic cells is thought to enhance viral entry and promote HIV replication. Bronchoscopic samples obtained from patients co-infected with M. tuberculosis and HIV show that changes in CXCR4 expression patterns may contribute to the accelerated progression of HIV with tuberculosis. HIV strains that use the CXCR4 co-receptor (X4 strains) instead of the CCR5 co-receptor (R5 strains) for entry into T cells are associated with a more rapid decline in CD4 cells. CXCR4 expression on alveolar macrophages obtained by bronchoscopic lavage is significantly higher in patients co-infected with M. tuberculosis and HIV than in HIV-infected patients without tuberculosis (Hoshino et al., 2004). In vitro, M. tuberculosis infection of macrophages increases CXCR4 expression and entry of X4 virus (Hoshino et al., 2004), suggesting that the innate response to M. tuberculosis may favour X4 viral replication. The effect of latent tuberculosis infection on HIV progression has not been well characterized, though the implications could be important as it is estimated that one-third of the world's population is latently infected (Dye et al., 1999).
Several epidemiological studies have shown that active HSV-2 infection is associated with an increased risk of HIV acquisition (Boulos et al., 1992; Hook et al., 1992; Corey et al., 2004). Although mechanical disruption of mucous membranes is the major risk factor for increased HIV acquisition, local recruitment and activation of CD4 T cells (Cunningham et al., 1985) and macrophages (Kucera et al., 1990; Heng et al., 1994) increase the number of target cells for HIV entry. HSV promotes HIV replication in human T cells (Mosca et al., 1987) and in macrophages by inducing NF-κB activity (Moriuchi et al., 2000). Increased HIV viral shedding can also be found in subclinical non-ulcerative HSV (Mbopi Keou et al., 1999; McClelland et al., 2002) suggesting that HIV transmissibility may be enhanced without direct mucosal disruption and herpetic ulceration. Ongoing clinical trials are evaluating the role of acyclovir in reducing HIV transmission.