To Benefit from the Full Content of Nature Reviews Microbiology Every Month, Simply Take

To benefit from the full content of Nature Reviews Microbiology every month, simply take out a subscription - click here for details.
Nature Reviews Microbiology2, 401-413 (2004); doi:10.1038/nrmicro878


[1110K]

CONTROL OF HIV-1 INFECTION BY SOLUBLE FACTORS OF THE IMMUNE RESPONSE

AnthonyL.DeVico & RobertC.Galloabout the authors

Institute of Human Virology, University of Maryland Biotechnology Institute, Baltimore, Maryland21202, USA.

correspondence to:

An increasing body of evidence indicates that the immune system uses a range of soluble molecules to suppress certain viral infections without killing infected host cells. Recent studies indicate that such factors might have an especially important role in the immune response to HIV-1. Accordingly, this review uses HIV-1 as a model to explore the diversity of non-cytolytic antiviral factors and considers how these molecules might be used to develop new therapeutic and prophylactic strategies to fight viral infections.

Viruses and immune systems play microscopic games of 'hide-and-seek' during the course of an infection. The virus attempts to find and enter its host cell, replicate its genome, assemble new particles and spread to new target cells, while minimizing its exposure to the immune system. At the same time, the immune system attempts to recognize and eliminate the invading viruses as quickly as possible and without causing damage to the host. In cases of chronic viral infections, or when viruses (such as retroviruses) use cells of the immune system as their hosts, this game can be very complicated indeed.

To gain an advantage, the immune system uses overlapping effector mechanisms, which are activated by viral infection and suppress one or more steps in the virus life cycle (Fig. 1). NATURAL KILLER CELLS (NK cells)1, 2, granulocytes and, possibly, -T CELLS3-5 provide the initial line of defence when stimulated by chemical signals that are released at sites of infection. All of these cells produce immunoregulatory molecules: neutrophils produce antimicrobial defensins; and NK cells and -T cells can kill infected cells through cytolytic mechanisms1-5. Host macrophages and dendritic cells also have important roles in this early stage by taking up and presenting viral antigens and by secreting several soluble CYTOKINES and CHEMOKINES that help to amplify the immune response. These early processes are called INNATE IMMUNE RESPONSES because they can act immediately using extant receptors and do not require the induction of major histocompatibility complex (MHC) gene products or antigen presentation in the context of MHC molecules. Over time, the innate response gives way to an adaptive response, which requires viral antigen presentation by cell-surface MHC molecules. Adaptive immunity is carried out by several effector-cell subsets, including CD8+ T CELLS, CD4+ HELPER T CELLS, and B cells. These cells establish important cell–cell communication mechanisms by releasing a wide range of soluble molecules. The cellular arm of adaptive immunity is active at an early stage (1 week post-infection) and has an important role in fighting most viral infections. CELLULAR IMMUNE RESPONSES are mediated by CD8+ cytotoxic T lymphocytes (CTLs) that have been primed by dendritic cells and other cell subsets that present viral antigens in conjunction with Class I MHC molecules. Once primed by MHC–antigen complexes and co-stimulatory signals, CTL clones target infected tissues and attempt to kill infected cells before they can produce progeny viruses. To do this, the CTL must interact with viral antigens in conjunction with Class I MHC molecules on the surface of an infected cell. The CTL then delivers pro-apoptotic signals and soluble cytolytic enzymes that destroy the infected target. Several weeks after infection, the humoral arm of adaptive immunity joins the cellular response in combating infection. HUMORAL IMMUNE RESPONSES arise as a result of interactions between antigen-specific B cells and CD4+ helper T cells that have been stimulated by viral antigens in conjunction with MHC Class II antigens. Ultimately, the interacting B cells release antibodies that react with specific epitopes on the viral antigens. In most cases, antiviral antibodies prevent the intercellular transmission of virions and the reinfection of the host, although in some instances they can also contribute to the direct suppression of viral replication. Overall, the conventional view of the immune system holds that the control and prevention of viral infections rests with one or more of these defence mechanisms, depending on the virus in question.

/ /
Figure 1|A working model of an antiviral immune response that uses cytolytic cell killing, antibodies and soluble non-cytolytic factors as a means to suppress infection.
In the first phase of an infection, viruses infect susceptible host cells. In response to infection, the innate immune response mediates several antiviral mechanisms, including cytolytic cell killing by NKcells, which lyse infected cells. Several soluble factors are also released, which directly suppress infection without killing infected host cells. Infected cells or virions that escape the innate response are controlled by the humoral and cellular adaptive immune responses. However, non-cytolytic antiviral mechanisms continue to have a crucial role in suppressing viral replication. Infected host cells are shown in red and uninfected cells in green. IFN, interferon; NK, natural killer; TNF, tumour necrosis factor.

However, this view is being reconsidered. We are beginning to understand that both innate and adaptive responses to viral infections can be supplemented by NON-CYTOLYTIC SOLUBLE SUPPRESSOR FACTORS, which, remarkably, can be antiviral either by accident or design. This new perspective has emerged primarily from four lines of evidence. First, it has become clear that certain immunoregulatory cytokines, including INTERFERONS (IFNs) and tumour necrosis factor (TNF), not only induce apoptosis and necrosis of certain cell types on infection, but also activate a number of intracellular pathways that directly suppress viral replication without killing the host cell6-18. Interferons have been shown to suppress hepatitis B virus (HBV), hepatitis C virus (HCV), herpes simplex virus, vesicular stomatitis virus (VSV), vaccinia virus, picornaviruses, retroviruses, influenza viruses and other types of viruses in vitro in a non-cytolytic manner6-13, 17. These broad antiviral effects are mediated by several mechanisms that rely on receptor-mediated gene-expression pathways, including the JAK/STAT (Janus kinase/signal transducer and activator of transcription) signal-transduction cascade6-11, 18, 19. In the presence of double-stranded RNA, IFN- and IFN- mediate well-characterized antiviral mechanisms that either degrade viral RNA transcripts or inhibit viral protein synthesis. Other mechanisms inhibit the attachment, entry or assembly of certain viruses. IFN- is likely to activate overlapping pathways as well as non-redundant pathways and comparable antiviral mechanisms8, 11, 18, 19. Second, many cell subsets of both the innate and adaptive responses, including NK cells, mononuclear phagocytes, -T cells, CD4+ T cells and CTLs, have been shown to secrete non-cytolytic antiviral molecules, such as IFN and TNF-, in response to stimulation by viral antigens10. Third, a substantial amount of data from experiments with HBV transgenic mice indicate that the primary mechanism for the control and clearance of HBV infection is provided by the direct non-cytolytic antiviral activity of soluble factors that are released by CTLs and/or NK cells10, 14, 15. In this transgenic system, the animals exhibit hepatic HBV gene expression, but do not mount an immune response (anti-self) against HBV antigens. Furthermore, the data do not support cell–cell virus spread and reinfection by HBV. So, these animals are highly useful for specific analyses of effector subsets. Adoptive transfer experiments carried out in these mice showed that HBV-specific CTL clones suppress viral replication in many infected cells through non-cytolytic antiviral activities that are mediated by IFN- and, perhaps, TNF-20-23. Fourth, cell subsets other than CTLs have been shown to release soluble factors with non-cytolytic antiviral activity. CD4+ T cells have been shown to suppress influenza, VSV, HBV and vaccinia infections in mice24-30, and HIV-1 infections in human lymphocyte cultures31, through the release of soluble factors. NK cells have also been observed to suppress certain viral infections by releasing non-cytolytic antiviral factors32-37. The production of soluble suppressor factors by -T cells has been correlated with the protection of macaques from simian immunodeficiency virus (SIV)38.

Given these findings, a new model for antiviral immunity has emerged, which includes the direct antiviral effects of soluble non-cytolytic factors as an effector mechanism. In this model, the array of soluble mediator factors that are released in response to infection includes factors that diffuse through the site of infection and directly suppress virus replication without killing infected cells (Fig. 1). As will be discussed below, this working model accounts for the possibility that some of the virus suppressor molecules might also perform immunoregulatory roles. In these cases, the relative importance of antiviral versus immunoregulatory function varies depending on the infecting virus and the nature of the adaptive response.

In theory, this model presents three advantages for the infected host. First, a soluble/diffusible non-cytolytic effector function would boost the antiviral potency of a CTL beyond its destructive capacity, which is inherently limited by the frequency of effector–target-cell contacts. So, a CTL would be able to more rapidly clear an infection in tissues where it is outnumbered by infected cells. Second, NK cells and CTLs could suppress infection in vital organs without having to destroy a large number of important cells. Third, an immune response to one virus might suppress other viruses in the area by bystander effects that are mediated through the diffusion of soluble antiviral factors. For example, the release of IFN by CTLs in response to one type of virus might partially 'sterilize' the area to other sensitive viruses. In the past such a phenomenon was difficult to show in experimental viral systems. In mice, CTL clones that are specific for lymphocytic choriomeningitis virus (LCMV) were passively transferred to animals that were then co-challenged with LCMV and the related pichinde arenavirus. Although the animals were protected from LCMV, they were still infected by pichinde39. Similarly, passive transfer of CTLs that are specific for influenza haemagglutinin HA2 protected mice from influenza virus but not from concomitant challenge with influenza haemagglutinin HA1 (Ref. 40). Such specificity indicates that if soluble non-cytolytic factors are released by antiviral CD8+ T lymphocytes, they might be active only over very short distances. On the other hand, it was shown that hepatocellular HBV gene expression was potently suppressed during LCMV infection in HBV transgenic mice41. Such suppression was non-cytolytic and principally mediated by TNF- and IFN-/ produced by LCMV-infected intrahepatic macrophages41. Similarly, clinical studies have provided evidence that HBV replication is suppressed by acute hepatitis A virus (HAV)-induced production of soluble factors including IFN-42, 43. So, soluble 'bystander' suppression may indeed occur more readily in certain tissues and viral systems.

In recent years, HIV-1 has been characterized as a virus that is highly sensitive to non-cytolytic suppression by soluble factors. Indeed, the nature of soluble HIV-1 suppressor activity might reflect an intimate link between viral replication and the immune system. So, HIV-1 infection is an ideal model for appreciating the capacities of soluble factors to mediate antiviral immunity. Accordingly, the following sections focus on HIV-1 suppressor factors, their activities and their relevance to natural infection.

Soluble HIV-1 suppressor activity

The first observations of non-cytolytic HIV-1 suppressor activity were made almost two decades ago in the context of CD8+ T-cell responses44-46. At that time, it was recognized that peripheral blood mononuclear cells (PBMCs) taken from seropositive asymptomatic individuals often failed to manifest HIV-1 replication in vitro. In 1986, Walker et al.44 showed that this suppression of viral replication was linked to the presence of CD8+ T cells in the cultures. Selective removal of these cells resulted in an elevation of viral replication, whereas depletion of other cell types, such as CD16+ cells (including NK cells), had no effect44. Furthermore, reconstitution of depleted cultures with autologous CD8+ T cells re-established suppression of HIV-1 replication in a concentration-dependent manner without altering the proliferation or viability of CD4+ HIV-1 host cells45-49. Taken together, these data show that CD8+ T cells are able to block active HIV-1 replication through non-cytolytic virus-suppressive mechanisms. Later studies revealed two more significant characteristics of this activity. First, the factor(s) that are responsible for HIV-1 suppressor activity are soluble45, 46, 50. Experiments carried out in transwell chambers clearly showed that non-cytolytic suppression was achieved even when the CD8+ T cells were separated from the CD4+ host cells by semi-permeable membranes50, 51. Other experiments showed that HIV-1 suppression was mediated by filtered supernatants from cultures of CD8+ T cells that had been activated with mitogen or anti-CD3 antibody and interleukin (IL)-2 (Refs 45,51,52). Second, the soluble factor(s) are capable of suppressing many, if not all, primary HIV-1 strains46, 47. CD8+ cell supernatants suppressed infection in infectivity systems that used cell-free virus stocks, as well as in experiments that used CD4+ T cells derived from HIV-positive individuals as the source of primary virus.

Taken together, these findings indicate that the immunological control of HIV-1 might involve non-cytolytic antiviral mechanisms, such as those shown in Fig. 1. As CD8+ T cells release the factor(s) with suppressive activity, it was reasonable to suspect that this activity is most relevant to cellular responses against HIV-1. So, the soluble suppressive factor was eventually named CD8 ANTIVIRAL FACTOR, or 'CAF', on the basis of the narrow assumption that activity could be explained by a single molecule.

However, our perception of soluble HIV-1 suppressor activity has progressed beyond this simplistic concept in three significant ways. First, it is now appreciated that soluble HIV-1 suppressor activity reflects the collective action of multiple factors. This characteristic was revealed in 1995, when it was determined that the chemokines RANTES and macrophage inflammatory proteins 1 and 1 (MIP-1 AND MIP-1) are involved in the suppressor activity when released by activated CD8+ T cells53. Specifically, it was shown that neutralizing anti-chemokine antibodies completely abrogate the HIV-1 suppressor activity of activated CD8+ T cells from HIV-1 seropositive, asymptomatic individuals. Notably, antibodies against any one of the chemokines had little effect on suppressor activity against the test isolate, HIV-1Bal. However, a mixture of antibodies against all three chemokines reversed the suppression of the virus (Fig. 2). This important observation showed that RANTES, MIP-1 and MIP-1 were responsible for nearly all of the HIV-1Bal suppressor activity in the CD8+ T-cell-culture supernatants. However, more extensive testing with a wider variety of isolates revealed that although RANTES, MIP-1 and MIP-1 always suppressed what were then called MACROPHAGE-TROPIC viruses, they did not suppress T-TROPIC isolates, such as HIV-1IIIB53. This was in contrast to unfractionated CD8+ T-cell supernatants, which suppressed all strains of HIV-1 regardless of tropism. We now know that macrophage-tropic viruses are selectively suppressed by RANTES, MIP-1 and MIP-1 owing to their specific requirements for entry. To enter target cells, HIV-1 must first establish an envelope–receptor complex that includes the viral envelope glycoprotein (gp120) and cell-surface CD4 molecule. The gp120–CD4 complex that is formed by macrophage-tropic viruses then binds selectively to a seven-transmembrane-spanning, G-protein-coupled surface co-receptor called CCR5 (Refs 54–58), which is also the natural receptor for RANTES, MIP-1 and MIP-1. As a consequence of this shared receptor usage, the entry of macrophage-tropic (now called R5) HIV-1 strains is blocked by the CCR5 ligands53, 59-61. Conversely, the T-tropic (now designated X4) strains enter cells using a different chemokine receptor known as CXCR4, which is not bound by CCR5 ligands58, 62 but instead by the chemokine stromal-cell-derived factor or SDF-1 (Ref. 63). This preference renders X4 strains immune to inhibition by CCR5 ligands. So-called DUAL-TROPIC (R5X4) strains can use either co-receptor type, depending on target-cell expression patterns, but are inhibited by CCR5 ligands whenever CCR5 is the operative co-receptor. So, RANTES, MIP-1 and MIP-1 collectively account for the CD8-derived suppression of R5 viruses (and R5X4 viruses in a CCR5-dependent system), whereas other, unknown factors are responsible for inhibiting the isolates that must use the CXCR4 co-receptor (Fig. 3). In riposte, 'CAF' is now defined as the portion of CD8-derived suppressor activity that suppresses X4 HIV-1 strains (or R5X4 strains in a CCR5-minus system), or any suppressive molecule other than RANTES, MIP-1 or MIP-117, 64, 65. In general, it is now common to categorize factors according to whether they suppress R5 versus non-R5 (X4 or R5X4) isolates.