Stephen Dewhurst, Ph.D. and Linda Whetter, D.V.M., Ph.D.

Department of Microbiology and Immunology, and Cancer Center, University of Rochester Medical Center, Rochester NY

Received 2/21/97; Accepted 2/25/97; On-line 3/1/97

3. DISCUSSION

3.1. Pathogenesis of HIV-1 infection

The pathogenesis of HIV-1 infection reflects the complex interplay between virus replication, virally-induced lymphocyte killing, and the immune response of the host. The latter includes both beneficial responses, which suppress viral replication, and harmful responses which enhance replication and exacerbate cell killing. Each of these aspects of the pathogenesis of HIV-1 infection will be considered separately; HIV-1 neuropathogenesis is not discussed here but has been reviewed elsewhere (2).

3.1.1. Viral replication and viral dynamics

The typical pattern of HIV-1 infection in vivo is shown in Figure 1. It should be noted, however, that virus infection does not always conform to this representative scenario. For example, in about 5% of virus-positive individuals, infection is nonprogressive with no decline in CD4⁺ lymphocyte counts and very low levels of viral RNA (3-5). This nonprogressive state may occur as the result of both host and viral factors, such as infection by naturally-occuring, attenuated, strains of HIV-1 -- such as nef-deleted viruses (6, 7).

In most cases, the level of HIV-1 RNA remains high during all phases of virus infection, including the period of clinical latency (8-10). Indeed, the average total HIV-1 production has been estimated at approximately 1 x 10¹⁰ virions per day (10). This reflects an active, dynamic, process in which CD4 lymphocytes are being infected and killed in large numbers. Indeed, estimated total CD4⁺ cell destruction rates are on the order of 3 x 10⁹ cells per day (8). As noted by Coffin, these findings are consistent with a simple steady-state model of HIV-1 viral dynamics, in which virus infection, cell death and cell production are in balance (11, 12). Roughly 1-2% of the body's total complement of CD4⁺ lymphocytes (2 x 10¹¹ cells) are eliminated, and replaced, each day over a period of many years. This level of production presumably places a high burden on the replacement system and may ultimately prove non-sustainable -- resulting in an immune collapse.

FIGURE 1: Schematic representation of the course of HIV-1 infection in vivo.

The relevance of HIV-1 dynamics to the pathogenesis of AIDS has been convincingly demonstrated by several studies. Perhaps the most compelling data are those reported by Mellors and colleagues (13). These investigators found that, among persons with equivalent baseline CD4⁺ T cell counts, individuals with high baseline plasma HIV-1 RNA loads (>10,190 molecules/ml) died more rapidly (mean of 6.8 years) than individuals with low (< 10,190 molecules/ml) baseline plasma HIV-1 RNA loads (time to death > 10 years). A similar difference was detected in the rate of disease progression. As a result of these and other studies, plasma HIV-1 RNA determinations, which can be performed by three commercially available assays -- branched DNA (bDNA), reverse transcriptase-polymerase chain reaction (RT-PCR) and nucleic acid sequence-based amplification (NASBA) -- are likely to become increasingly important in clinical practice (14).

Interestingly, HIV-1 RNA load is relatively stable throughout much of the course of the viral infection (at least, until the onset of frank AIDS). Indeed, it appears that the steady-state, or equilibrium, level of HIV-1 RNA load is established within the first several months following virus infection. Thus, in early HIV-1 infection, a high virus burden is predictive of rapid disease progression both in adults and in children (15, 16). This suggests that the first few weeks and months following the initial HIV-1 infection may represent a key phase in the pathogenesis of AIDS, in which the virus is able to replicate to very high levels, to "seed" lymphoid organs, and to establish a state of equilibrium with its host. The "set point", at which this equilibrium occurs, may be critical to the subsequent disease progression (17). This may explain why administration of AZT during primary HIV-1 infection results in a striking improvement in the clinical outcome of infection, as measured by the rate of disease progression over time (18).

Although the level of HIV-1 RNA load in peripheral blood can be readily measured, less information has been garnered concerning viral replication in lymphoid tissues, principally due to the relative inaccessibility of these compartments. This issue has recently been addressed by Haase and colleagues, who applied a quantitative image analysis technique to analyze the viral burden in lymphoid tissues (19). A strikingly large and stable pool of extracellular virions was detected within lymphoid tissue, trapped on the surfaces of follicular dendritic cells (FDC). This FDC virus pool has been estimated to be 10 to 40 times larger than the productive virus pool, and it therefore accounts for most of the total body burden of HIV-1 RNA (roughly 10¹¹ copies of HIV-1 RNA) (19). The FDC virus pool may be important for the perpetuation and spread of viral infection within lymphoid tissue, and may complicate antiviral therapy, due to its relative stability.

3.1.2. Viral genetic variation

The rapid emergence of viral quasispecies (closely related but genetically distinct viral variants) is one of the hallmarks of lentiviral infections (20), both in humans and in primates. This property is also characteristic of RNA viruses as a whole, in large part due to the high mutation rate of RNA polymerases (21). In considering the genetic variation of HIV-1, it is instructive to note that the average viral generation time (defined as the interval between the release of a viral particle and the infection of a new host cell and release of a new burst of progeny virions) has been estimated at 2.6 days (10). Thus, HIV-1 replicates at a rate of 140 generations per year, for a period of ten or more years. It can be readily appreciated that this will lead to rapid and explosive genetic variation (11), since the mutation rate of HIV-1 is roughly 3 x 10^-5 per base per replication cycle, and the viral genome size is 10⁴ base pairs. If 10¹⁰ new viral genomes are produced each day, then on average, every mutation at each position in the viral genome will occur several times -- in just a single day. This unprecedented genetic diversification has important implications for the evolution of viral variants with resistance to antiviral drugs. One can predict that such mutations will occur inevitably and rapidly -- except, perhaps, when multiple antiviral drugs are used in combination. This issue is discussed in more detail in section 4.

Viral genetic variation may also be driven, to some extent, by the host immune response. Recent findings by two groups of investigators, suggest that HIV-1 undergoes adaptive evolution in vivo, in response to selective pressure exerted by the host immune system (22, 23). As a result, viral genetic diversity is greatest in clinically healthy individuals and much less in persons with AIDS (where the immune response has become severely impaired) (22). Thus, the extent of intrahost HIV-1 evolution is to some degree related to the length of the immunocompetent period (23). It is less clear whether HIV-1 genetic diversity plays any direct role in the pathogenesis of AIDS. Nowak and colleagues have proposed the existence of an "antigenic diversity threshold", in which the ever-expanding genetic diversity of HIV-1 eventually exhausts the capacity of the immune system to respond, resulting in an immune collapse (24). While direct support for this theory has been elusive (22, 23), it is intuitively obvious that viral variation could facilitate immune evasion. Consistent with this, Borrow et al. showed that primary virus infection was associated with the rapid clearance of the transmitted strain, followed by selection for a virus population comprised of strains bearing a mutation in the major immunodominant cytotoxic T-lymphocyte (CTL) epitope (25). CTL escape mutants have also been associated with the progression to AIDS, many years after the initial virus infection (26).

3.1.3. Protective host immune responses

In all likelihood, the pathogenesis of AIDS reflects a balance between viral replication and the immune response. Available evidence suggests that humoral immunity is probably not the dominant mode by which HIV-1 replication is controlled. Serum antibodies are of uncertain relevance, in part because primary HIV-1 isolates are relatively resistant to serum neutralization, unlike T-cell line-adapted strains (27-29). In addition, although HIV-1 infected long-term non-progressors produce vigorous serum antibody responses, virus strains isolated from these individuals were not neutralized by autologous sera (30). In contrast, the cellular immune response to HIV-1 may be critical in controlling viral infection and in determining the steady-state HIV-1 RNA load that is established following the primary infection (25, 31, 32). Pantaleo and colleagues have made the interesting observation that mobilization of a broad T cell receptor (TCR) repertoire during acute HIV-1 infection is associated with a relatively stable clinical course, while mobilization of a restricted subset of CD8⁺ T cells is associated with more rapid disease progression (32). If, as these authors suggest, the CD8⁺ T cell families which are expanded during acute infection are rapidly deleted, then it is relatively easy to understand how a broad TCR repertoire could be important in controlling the early stages of HIV-1 infection, and thereby establishing a relatively low steady-state level of HIV-1 RNA load.

Other protective aspects of the immune response may include the production of high levels of soluble inhibitors of virus infection and replication -- as first noted by Levy and colleagues (33). The most well characterized examples are the ß-chemokines, macrophage inflammatory proteins 1 (MIP-1) alpha and ß and RANTES (regulated on activation, normally T-cell expressed and secreted), which are shown in Figure 2.

FIGURE 2: Structural representation of the ß-chemokines, MIP-1ß (top) and RANTES (bottom). Red coils represent alpha-helices and orange sheets represent ß-sheets. Both chemokines are homodimers composed of identical subunits. At the amino acid level, RANTES and MIP-1ß exhibit 49% identity, but structurally the two molecules are even more closely related (see above), and both bind to the same receptor, CCR5. The above molecular representations were generated with RasMol , using the PDB database files 1HRJ.PDB (RANTES) and 1HUM.PDB (MIP-1ß).

The ß-chemokines can interfere with the ability of M-tropic HIV-1 strains to infect T cells (34), by blocking the binding of HIV-1 gp120 to the CCR5 entry cofactor, as illustrated in Figure 3 (35-40). High levels of these chemokines have been found in exposed-uninfected persons, who are relatively resistant to HIV-1 infection (41), and these chemokines may also contribute to the control of virus replication in infected individuals. However, the ß-chemokines are not effective against all HIV-1 strains or in all cell types. Indeed, these factors are effective only against infection of T cells by M-tropic strains of HIV-1. Infection of monocyte/macrophages by these same HIV-1 strains is not blocked (42, 43), and neither is the infection of T cells by T-tropic HIV-1 strains (34). The latter finding reflects the fact that T-tropic HIV-1 strains utilize a distinct entry cofactor, CXCR4 (44-46). The failure of ß-chemokines to block infection of macrophages is somewhat more surprising, since macrophages express CCR5 (43). This suggests that the process of HIV-1 entry in these cells may be different than for T cells (47, 48). This might also explain why macrophages express CXCR4 but cannot be infected by T-tropic HIV-1 strains (43).

FIGURE 3: Model for HIV-1 entry in T cells. HIV-1 gp120 first binds to cellular CD4. This results in a conformational change in gp120, allowing it to bind to the chemokine receptors CCR5 or CXCR4, thereby forming a trimolecular complex (CD4-gp120-CCR/CXCR). An excess of the natural ligands for CCR5 or CXCR4 can competitively inhibit this step of infection (as noted). After binding to the chemokine receptor, gp120 is thought to become stripped off the virion, thereby exposing a hydrophobic domain at the N-terminus of gp41, which mediates fusion of the host cell and virus membranes, thereby allowing the virus core to enter the host cell cytoplasm.

ß-chemokines are almost certainly not the only naturally occuring cytokines with anti-HIV-1 activity. For example, infection of T cells by T-tropic HIV-1 strains can be inhibited by the CXCR4 ligand, stromal cell-derived factor-1 (SDF-1) (45, 46). In addition, other soluble factors can inhibit HIV-1 infection, including the CD4-binding chemokine interleukin-16 (49, 50), and as-yet unidentified soluble factor(s) derived from CD8⁺ T cells (33).

3.1.4. Immunpathogenic mechanisms

Regulation of HIV-1 replication is influenced by the opposing effects of suppressive factors such as the ß-chemokines and inductive factors, including endogenous proinflammatory cytokines such as tumor necrosis factor (TNF) alpha and interleukin-6 (IL-6) (51, 52). Levels of these proinflammatory cytokines are often high during the acute phase of HIV-1 infection (53-55), particularly in individuals who experience a severe or symptomatic acute primary HIV-1 syndrome (54, 56). This may in part explain why extensive T cell activation during the early phases of HIV-1 infection is associated with more rapid disease progression (15, 56).

T cell activation may also contribute to the replication of HIV-1 during the post-acute phase of virus infection. For example, HIV-1 infection of cultured CD4⁺ T cells is greatly enhanced in the setting of antigen-specific immune activation (57), and virus replication increases dramatically in vivo after vaccination of HIV-1 infected persons with a variety of immunogens (58-60). T cell activation in HIV-1 infected individuals is also associated with extensive apoptosis of virus-negative T cells (61, 62). This increased susceptibility of T cells from HIV-1 infected persons to activation-induced cell death may be important to the pathogenesis of immune deficiency and may help to explain the accelerated course of HIV-1 induced disease in areas where immune activation may be persistent or chronic due to endemic parasites and other pathogens (63).

An important corollary of the relationship between immune activation and HIV-1 replication is the prediction that immune suppression should lead to a reduction in the viral burden and perhaps even to an improvement in clinical status (64). Several studies suggest that this may be the case. For example, blockade of the endogenous proinflammatory cytokines TNF-alpha, IL-6 and interleukin-1ß results in suppression of HIV-1 replication in vitro (52, 65). In addition, immune suppressive therapy has been associated with a rise in CD4⁺ cell counts in asymptomatic HIV-1 infected persons (the median CD4 increase in a cohort of 44 persons who received oral prednisolone for one year was 119 cells/µl (66)). A more aggressive immunosuppressive regimen (total lymphoid irradiation) also resulted in a transient decrease in viral burden and a lack of disease progression in simian immunodeficiency virus (SIV) infected macaques (67). These data suggest that immune suppressive therapies may have a role in the treatment of HIV-1 infection (68).