Can modulation of viral fitness represent a target for anti-HIV-1 strategies?

Prolonged use of anti-retroviral compounds in human immunodeficiency virus type 1 (HIV-1) infection selects for drug-resistant and often mutidrug-resistant viral variants. Drug-resistance mutations may also affect viral fitness. Interestingly, recent research has indicated that some of the unfit drug-resistant variants may be less pathogenic, suggesting that decreased viral fitness is beneficial for the host and may be driven by specific treatments during anti-HIV-1 infection. A second potential antiviral strategy starting with profound inteference with viral fitness aims at forcing viruses towards lethal mutagenesis (the so-called "error catastrophe"). This review summarizes the methods for addressing HIV-1 fitness in vitro and ex vivo, the current understanding of clinical implications of reduced HIV-1 fitness, and the potential use of anti-HIV-1 strategies aiming at modulating viral fitness. Finally, it is emphasized how the peculiar features of HIV-1 quasispecies (displaying two different forms of memory, a replicative and a non-replicative form) may sharply influence the design of future diagnostic methodologies for fitness analysis.     

Highly pathogenic simian immunodeficiency virus mne variants that emerge during the course of infection evolve enhanced infectivity and the ability to downregulate CD4 but not class I major histocompatibility complex antigens

The replicative, cytopathic, and antigenic properties of simian immunodeficiency virus (SIV) variants influence its replication efficiency in vivo. To further define the viral properties and determinants that may be important for high-level replication in vivo and progression to AIDS, we compared a minimally pathogenic SIVmne molecular clone with two highly pathogenic variants cloned from late stages of infection. Both variants had evolved greater infectivity than the parental clone due to mutations in nef. Interestingly, a pol determinant in one of the highly pathogenic variants also contributed to its increased infectivity. Furthermore, because replication in vivo may also be influenced by the ability of a virus to evade the cellular immune response of the host, we examined whether the variants were more capable of downregulating surface expression of class I major histocompatibility complex (MHC). Decreased MHC class I expression was not observed in cells infected with any of the viruses. Furthermore, the Nef proteins of the highly pathogenic variants only slightly reduced surface MHC class I expression in transfected cells, although they efficiently downregulated CD4. Together, these data demonstrate that mutations which can enhance viral infectivity, as well as CD4 downregulation, may be important for efficient replication of SIV in the host. However, Nef-mediated reduction of MHC class I expression does not appear to be critical for the increased in vivo replicative ability of highly pathogenic late variants.     

Importance of B-cell responses for immunological control of variant strains of simian immunodeficiency virus

The properties of three variants of cloned simian immunodeficiency virus strain 239 (SIV239) were compared. One strain (M5) lacked five sites for N-linked carbohydrate attachment in variable regions 1 and 2 (V1 and V2) of the gp120 envelope protein, one strain (DeltaV1-V2) completely lacked V1 and V2 sequences, and another (316) had nine mutations in the envelope that impart high replicative capacity for tissue macrophages. All three strains were capable of significant levels of fusion independent of CD4, and all three were considerably more sensitive to antibody-mediated neutralization than the parent strain from which they were derived. Upon experimental infection of rhesus monkeys, these three variant strains replicated to viral loads at peak height around day 14 that were indistinguishable from or only slightly less than those observed in monkeys infected with the parental SIV239 strain. Viral loads at the set point 20 to 50 weeks after infection, however, were more than 400- to 10,000-fold lower with the variant strains. Depletion of B cells around the time of infection with M5 resulted in less effective immunological control and much higher viral loads at the set point in two of three monkeys. The differences between SIV239 infection, where there is not effective immunological control, and SIVM5 infection, where there is effective immunological control, cannot be easily explained by differences in the inherent replicative capacity of the viruses; rather, they are more readily explained by differences in the effectiveness of the antibody response. These results suggest that resistance of SIV239 to antibody-mediated neutralization is very important for evading effective immunological control, for allowing continuous viral replication, for maintenance of moderate-to-high viral loads at set point, and for disease progression.     

Characteristics of a pathogenic molecular clone of an end-stage serum-derived variant of simian immunodeficiency virus (SIV(F359))

End-stage simian immunodeficiency virus (SIV) isolates are suggested to be the most fit of the evolved virulent variants that precipitate the progression to AIDS. To determine if there were common characteristics of end-stage variants which emerge from accelerated cases of AIDS, a molecular clone was derived directly from serum following in vivo selection of a highly virulent SIV isolate obtained by serial end-stage passage in rhesus monkeys (Macaca mulatta). This dominant variant caused a marked cytopathic effect and replicated to very high levels in activated but not resting peripheral blood lymphocytes. Furthermore, although this clone infected but did not replicate to detectable levels in rhesus monocyte-derived macrophages, these cells were able to transmit infection to autologous T cells upon contact. Interestingly, although at low doses this end-stage variant did not use any of the known coreceptors except CCR5, it was able to infect and replicate in human peripheral blood mononuclear cells homozygous for the Delta 32 deletion of CCR5, suggesting the use of a novel coreceptor. It represents the first pathogenic molecular clone of SIV derived from viral RNA in serum and provides evidence that not only the genetic but also the biological characteristics acquired by highly fit late-stage disease variants may be distinct in different hosts.     

Macaque long-term nonprogressors resist superinfection with multiple CD8+ T cell escape variants of simian immunodeficiency virus

Human immunodeficiency virus (HIV)-positive individuals can be superinfected with different virus strains. Individuals who control an initial HIV infection are therefore still at risk for subsequent infection with divergent viruses, but the barriers to such superinfection remain unclear. Here we tested long-term nonprogressors' (LTNPs') susceptibility to superinfection using Indian rhesus macaques that express the major histocompatibility complex class I (MHC-I) allele Mamu-B 17, which is associated with control of the pathogenic AIDS virus SIVmac239. The Mamu-B 17-restricted CD8(+) T cell repertoire is focused almost entirely on 5 epitopes. We engineered a series of SIVmac239 variants bearing mutations in 3, 4, or all 5 of these epitopes and used them to serially challenge 2 Mamu-B 17-positive LTNPs. None of the escape variants caused breakthrough replication in LTNPs, although they readily infected Mamu-B 17-negative naive macaques. In vitro competing coculture assays and examination of viral evolution in hosts lacking Mamu-B 17 suggested that the mutant viruses had negligible defects in replicative fitness. Both LTNPs maintained robust immune responses, including simian immunodeficiency virus (SIV)-specific CD8(+) and CD4(+) T cells and neutralizing antibodies. Our results suggest that escape mutations in epitopes bound by "protective" MHC-I molecules may not be sufficient to establish superinfection in LTNPs.     

CD4+ target cell availability determines the dynamics of immune escape and reversion in vivo

Infections with human immunodeficiency virus (HIV) and the closely related monkey viruses simian-human immunodeficiency virus (SHIV) and simian immunodeficiency virus (SIV) are characterized by progressive waves of immune responses, followed by viral mutation and "immune escape." However, escape mutation usually leads to lower replicative fitness, and in the absence of immune pressure, an escape mutant (EM) virus "reverts" to the wild-type phenotype. Analysis of the dynamics of immune escape and reversion has suggested it is a mechanism for identifying the immunogens best capable of controlling viremia. We have analyzed and modeled data of the dynamics of wild-type (WT) and EM viruses during SHIV infection of macaques. Modeling suggests that the dynamics of reversion and immune escape should be determined by the availability of target cells for infection. Consistent with this suggestion, we find that the rate of reversion of cytotoxic T-lymphocyte (CTL) EM virus strongly correlates with the number of CD4(+) T cells available for infection. This phenomenon also affects the rate of immune escape, since this rate is determined by the balance of CTL killing and the WT fitness advantage. This analysis predicts that the optimal timing for the selection of immune escape variants will be immediately after the peak of viremia and that the development of escape variants at later times will lead to slower selection. This has important implications for comparative studies of immune escape and reversion in different infections and for identifying epitopes with high fitness cost for use as vaccine targets.     

Coreceptor Specificity of Temporal Variants of Simian Immunodeficiency Virus Mne

The simian immunodeficiency virus (SIV) Mne envelope undergoes genetic changes that alter tropism, syncytium-inducing capacity, and antigenic properties of the emerging variant virus population during the course of an infection. Here we investigated whether the mutations in envelope of SIVMne also influence coreceptor usage. The data demonstrate that the infecting macrophage-tropic SIVMne clone as well as the envelope variants that are selected during the course of disease progression all recognize both CCR5 and Bob (GPR15) but not Bonzo (STRL33), CXCR4, or CCR3. Although it remains to be determined if there are other coreceptors specific for dualtropic or T-cell-tropic variants of SIVMne that emerge during late stages of infection, these data suggest that such SIV variants that evolve in pathogenic infections do not lose the ability to recognize CCR5 or Bob/GPR15.     

The human immunodeficiency virus type 1 nef gene can to a large extent replace simian immunodeficiency virus nef in vivo.

The nef gene of the pathogenic simian immunodeficiency virus (SIV) 239 clone was replaced with primary human immunodeficiency virus type 1 (HIV-1) nef alleles to investigate whether HIV-1 Nef can substitute for SIV Nef in vivo. Initially, two rhesus macaques were infected with the chimeric viruses (Nef-SHIVs). Most of the nef alleles obtained from both animals predicted intact open reading frames. Furthermore, forms containing upstream nucleotide substitutions that enhanced expression of the inserted gene became predominant. One animal maintained high viral loads and slowly progressed to immunodeficiency. nef long terminal repeat sequences amplified from this animal were used to generate a second generation of Nef-SHIVs. Two macaques, which were subsequently infected with a mixture of cloned chimeric viruses, showed high viral loads and progressed to fatal immunodeficiency. Five macaques received a single molecular clone, named SHIV-40K6. The SHIV-40K6 nef allele was active in CD4 and class I major histocompatibility complex downregulation and enhanced viral infectivity and replication. Notably, all of the macaques inoculated with SHIV-40K6 showed high levels of viral replication early in infection. During later stages, however, the course of infection was variable. Three animals maintained high viral loads and developed immunodeficiency. Of the remaining two macaques, which showed decreasing viral loads after the acute phase of infection, only one efficiently controlled viral replication and remained asymptomatic during 1.5 years of follow-up. The other animal showed an increasing viral load and developed signs of progressive infection during later stages. Our data demonstrate that HIV-1 nef can, to a large extent, functionally replace SIVmac nef in vivo.     

IL-4 increases Simian immunodeficiency virus replication despite enhanced SIV immune responses in infected rhesus macaques

It is widely believed that a Th1 type CD4 response is critical for enhancement of CD8 immunity and for controlling HIV-1 infection. Th2 type responses, such as what might be seen in a chronic parasitic infection, would sacrifice cellular immunity and thus benefit the virus at the expense of the host. However, there has been little direct examination of the hypothesis in a primate model system. Accordingly, the simian immunodeficiency virus (SIV) infected rhesus macaque model was used to investigate the impact of immunisation with SIV expressing DNA constructs and co-injection with IL-4 on the SIV specific immunological responses, lymphocyte cell counts, as well as the impact on viral load. IL-4 is a Th2 type cytokine, which enhances antibody production and inhibits a CD4 Th1 phenotype. Rhesus macaques were infected with 10 AID50 of SIVmac239 and treated with 9-[2-(phosphonomethoxy)propyl]adenine (PMPA) 9 weeks post-infection. During PMPA treatment, animals were immunised with plasmids that expressed the SIV proteins, env, rev, gag and pol. In addition, they were immunised with a construct that encoded the gene for IL-4. IL-4 co-immunisation increased the neutralizing antibody titres in this group. Importantly, the viral loads in animals vaccinated with IL-4 expressing plasmid increased during the immunisation regimens despite the higher neutralizing antibody titres. In addition, neutralizing antibodies did not correlate with viral set point prior to PMPA treatment, however, there was a correlation between viral loads and antibody titres following the treatment with PMPA. Antibody titres decreased following the suppression of viral load. Importantly, vaccination in the absence of IL-4 protected CD4 levels without increasing viral load. The data support the hypothesis that inappropriate immune bias toward a Th2 pathway would ultimately enhance disease progression.     

Antiretroviral therapy during primary immunodeficiency virus infection can induce persistent suppression of virus load and protection from heterologous challenge in rhesus macaques

A limited period of chemotherapy during primary immunodeficiency virus infection might provide a long-term clinical benefit even if treatment is initiated at a time point when virus is already detectable in plasma. To evaluate this strategy, we infected rhesus macaques with the pathogenic simian/human immunodeficiency virus RT-SHIV and treated them with the antiretroviral drug (R)-9-(2-phosphonylmethoxypropyl)adenine (PMPA) for 8 weeks starting 7 or 14 days postinfection. PMPA treatment suppressed viral replication efficiently in all of the monkeys. After chemotherapy ended, virus replication rebounded and viral RNA in plasma reached levels comparable to that of the controls in four of the six monkeys. However, in the other two animals, virus loads peaked only moderately after withdrawal of the drug and then declined to low or even undetectable levels. These low levels of viremia remained stable for at least 31 weeks after cessation of therapy. At this time point, these two monkeys were challenged with SIV(8980) to evaluate whether the host responses which were able to keep RT-SHIV replication under control were also sufficient to protect against infection with a highly pathogenic heterologous virus. Both monkeys proved to be protected against the heterologous virus. In one of the two animals, low levels of SIV(8980) replication were detected. Thus, by chemotherapy during the acute phase of pathogenic virus replication, we could achieve not only persistent virus load suppression in two out of six monkeys but also protection from subsequent heterologous challenge. By this chemotherapeutic attenuation, the replication kinetics of attenuated viruses could be mimicked and a vaccination effect similar to that induced by live attenuated simian immunodeficiency virus vaccines was achieved.     

Selective Expansion of Viral Variants following Experimental Transmission of a Reconstituted Feline Immunodeficiency Virus Quasispecies

Following long-term infection with virus derived from the pathogenic GL8 molecular clone of feline immunodeficiency virus (FIV), a range of viral variants emerged with distinct modes of interaction with the viral receptors CD134 and CXCR4, and sensitivities to neutralizing antibodies. In order to assess whether this viral diversity would be maintained following subsequent transmission, a synthetic quasispecies was reconstituted comprising molecular clones bearing envs from six viral variants and its replicative capacity compared in vivo with a clonal preparation of the parent virus. Infection with either clonal (Group 1) or diverse (Group 2) challenge viruses, resulted in a reduction in CD4+ lymphocytes and an increase in CD8+ lymphocytes. Proviral loads were similar in both study groups, peaking by 10 weeks post-infection, a higher plateau (set-point) being achieved and maintained in study Group 1. Marked differences in the ability of individual viral variants to replicate were noted in Group 2; those most similar to GL8 achieved higher viral loads while variants such as the chimaeras bearing the B14 and B28 Envs grew less well. The defective replication of these variants was not due to suppression by the humoral immune response as virus neutralising antibodies were not elicited within the study period. Similarly, although potent cellular immune responses were detected against determinants in Env, no qualitative differences were revealed between animals infected with either the clonal or the diverse inocula. However, in vitro studies indicated that the reduced replicative capacity of variants B14 and B28 in vivo was associated with altered interactions between the viruses and the viral receptor and co-receptor. The data suggest that viral variants with GL8-like characteristics have an early, replicative advantage and should provide the focus for future vaccine development.     

Cross-protective immune responses induced in rhesus macaques by immunization with attenuated macrophage-tropic simian immunodeficiency virus.

The simian immunodeficiency virus (SIV) macaque model of AIDS has provided a valuable system with which to investigate vaccine approaches for protection against human immunodeficiency virus type 1 (HIV-1) infection. In particular, the ability of macaques persistently infected with attenuated infectious molecular clones of SIV to resist challenge with the pathogenic parental swarm has conclusively demonstrated that protective immunity can be achieved by immunization prior to exposure. The breadth of these protective responses and the immunological correlates of protection, however, have not been identified. In addition, vaccine studies have mainly employed lymphocyte-tropic strains of HIV-1 and SIV. Recent studies have implicated macrophage-tropic strains in the transmission of HIV-1 and have suggested that these virus strains should be examined in vaccine strategies. Macrophage-tropic viruses may confer additional advantages in the induction of protective immunity by replication in antigen-presenting cells. In this study, the immune response of rhesus macaques inoculated with an attenuated macrophage-tropic recombinant of SIVmac239 (SIV/17E-Cl) was evaluated with respect to protective immunity by heterologous challenge at various times after infection. Vigorous type-specific neutralizing-antibody responses restricted to SIV/17E-Cl were evident by 2 weeks postinfection. By 7 months, however, cross-reactive neutralizing antibodies emerged which neutralized not only SIV/17E-Cl but also the heterologous primary isolate SIV/DeltaB670. Challenge of SIV/17E-Cl-infected monkeys with SIV/DeltaB670 at various times postinfection demonstrated that protective responses were associated with the appearance of cross-reactive neutralizing antibodies. Furthermore, passive transfer of sera from SIV/17E-Cl-infected animals passively protected two of four naive recipients.     

Common Themes of Antibody Maturation to Simian Immunodeficiency Virus, Simian-Human Immunodeficiency Virus, and Human Immunodeficiency Virus Type 1 Infections

Characterization of virus-specific immune responses to human immunodeficiency virus type 1 (HIV-1) and simian immunodeficiency virus (SIV) is important to understanding the early virus-host interactions that may determine the course of virus infection and disease. Using a comprehensive panel of serological assays, we have previously demonstrated a complex and lengthy maturation of virus-specific antibody responses elicited by attenuated strains of SIV that was closely associated with the development of protective immunity. In the present study, we expand these analyses to address several questions regarding the nature of the virus-specific antibody responses to pathogenic SIV, SIV/HIV-1 (SHIV), and HIV-1 infections. The results demonstrate for the first time a common theme of antibody maturation to SIV, SHIV, and HIV-1 infections that is characterized by ongoing changes in antibody titer, conformational dependence, and antibody avidity during the first 6 to 10 months following virus infection. We demonstrate that this gradual evolution of virus-specific antibody responses is independent of the levels of virus replication and the pathogenicity of the infection viral strain. While the serological assays used in these studies were useful in discriminating between protective and nonprotective antibody responses during evaluation of vaccine efficacy with attenuated SIV, these same assays do not distinguish the clinical outcome of infection in pathogenic SIV, SHIV, or HIV-1 infections. These results likely reflect differences in the immune mechanisms involved in mediating protection from virus challenge compared to those that control an established viral infection, and they suggest that additional characteristics of both humoral and cellular responses evolve during this early immune maturation.     

Inclusion of Vpr accessory gene in a plasmid vaccine cocktail markedly reduces Nef vaccine effectiveness in vivo resulting in CD4 cell loss and increased viral loads in rhesus macaques

We compared the immunogenicity of plasmid vaccines containing multiple human immunodeficiency virus (HIV) antigens and found that covaccination with plasmids expressing HIV-1 14 kDa vpr gene product profoundly reduces antigen-specific CD8-mediated cytotoxic T-cell activity (CTL). Interestingly, Th1 type responses against codelivered antigens (pGag-Pol, pNef, etc.) encoded by the plasmid vaccines were suppressed. This suggested that vpr might compromise CD8 T-cell immunity in vivo during infection. A pilot primate vaccine study was designed to test the hypothesis to compare the following groups: unvaccinated controls, animals vaccinated without simean immunodeficiency virus (SIV)-Nef antigen plasmid, and animals covaccinated with the identical plasmid antigen and a plasmid construct encoding SIV Vpr/Vpx. Animals were subsequently challenged intrarectally with pathogenic SIVmac251 after the final vaccination of a multiple immunization protocol. Control animals were all infected and exhibited high viral loads and rapid CD4+ T-cell loss. In contrast, the Nef plasmid-vaccinated animals were also infected but exhibited preservation of CD4+ T-cells and a multilog reduction in viral load compared with controls. Animals covaccinated multiple times with the Nef vaccine and pVpr/Vpx plasmid suffered rapid and profound loss of CD4+ T-cells. These results have important implications for the design of multicomponent and particle vaccines for HIV-1 as well as for our understanding of HIV/SIV pathogenesis in vivo.     

Immune response of rhesus macaques to recombinant simian immunodeficiency virus gp130 does not protect from challenge infection.

Simian immunodeficiency virus (SIV) infection of rhesus macaques is a model for human immunodeficiency virus (HIV) infection in humans. Inactivated and modified live whole-virus vaccines have provided limited protective immunity against SIV in rhesus macaques. Because of safety concerns in the use of inactivated and live whole-virus vaccines, we evaluated the protective immunity of vaccinia virus recombinants expressing the surface glycoprotein (gp130) of SIVmac and subunit preparations of gp130 expressed in mammalian cells (CHO). Three groups of animals were immunized with recombinant SIV gp130. The first group received SIV gp130 purified from genetically engineered CHO cells (cSIVgp130), the second group was vaccinated with recombinant vaccinia virus expressing SIVmac gp130 (vSIVgp130), and the third group was first primed with vSIVgp130 and then given a booster immunization with cSIVgp130. Although anti-gp130 binding antibodies were elicited in all three groups, neutralizing antibodies were transient or undetectable. None of the immunized animals resisted intravenous challenge with a low dose of cell-free virus. However, the group primed with vSIVgp130 and then boosted with cSIVgp130 had the lowest antigen load (p27) compared with the other groups. The results of these studies suggest that immunization of humans with HIV type 1 surface glycoprotein may not provide protective immunity against virus infection.     

Reversion of CTL escape-variant immunodeficiency viruses in vivo

Engendering cytotoxic T-lymphocyte (CTL) responses is likely to be an important goal of HIV vaccines. However, CTLs select for viral variants that escape immune detection. Maintenance of such escape variants in human populations could pose an obstacle to HIV vaccine development. We first observed that escape mutations in a heterogeneous simian immunodeficiency virus (SIV) isolate were lost upon passage to new animals. We therefore infected macaques with a cloned SIV bearing escape mutations in three immunodominant CTL epitopes, and followed viral evolution after infection. Here we show that each mutant epitope sequence continued to evolve in vivo, often re-establishing the original, CTL-susceptible sequence. We conclude that escape from CTL responses may exact a cost to viral fitness. In the absence of selective pressure upon transmission to new hosts, these original escape mutations can be lost. This suggests that some HIV CTL epitopes will be maintained in human populations.     

Emerging cytopathic and antigenic simian immunodeficiency virus variants influence AIDS progression

Genetic variants of human and simian immunodeficiency virus (HIV and SIV) that evolve during the course of infection and progression to AIDS are phenotypically and antigenically distinct from their progenitor viruses present at early stages of infection. However, it has been unclear how these late variants, which are typically T-cell tropic, cytopathic and resistant to neutralizing antibodies, influence the development of clinical AIDS. To address this, we infected macaques with cloned SIVs representing prototype variants from early-, intermediate- and late-stage infection having biological characteristics typical of viruses found at similar stages of HIV infection in humans. These studies demonstrate that sequential, phenotypic and antigenic variants represent viruses that have become increasingly fit for replication in the host, and our data support the hypothesis that emerging variants have increased pathogenicity and drive disease progression in SIV and HIV infection.     

Similar Impact of CD8+ T Cell Responses on Early Virus Dynamics during SIV Infections of Rhesus Macaques and Sooty Mangabeys

Despite comparable levels of virus replication, simian immunodeficiency viruses (SIV) infection is non-pathogenic in natural hosts, such as sooty mangabeys (SM), whereas it is pathogenic in non-natural hosts, such as rhesus macaques (RM). Comparative studies of pathogenic and non-pathogenic SIV infection can thus shed light on the role of specific factors in SIV pathogenesis. Here, we determine the impact of target-cell limitation, CD8+ T cells, and Natural Killer (NK) cells on virus replication in the early SIV infection. To this end, we fit previously published data of experimental SIV infections in SMs and RMs with mathematical models incorporating these factors and assess to what extent the inclusion of individual factors determines the quality of the fits. We find that for both rhesus macaques and sooty mangabeys, target-cell limitation alone cannot explain the control of early virus replication, whereas including CD8+ T cells into the models significantly improves the fits. By contrast, including NK cells does only significantly improve the fits in SMs. These findings have important implications for our understanding of SIV pathogenesis as they suggest that the level of early CD8+ T cell responses is not the key difference between pathogenic and non-pathogenic SIV infection.     

Replicative fitness of protease inhibitor-resistant mutants of human immunodeficiency virus type 1

The relative replicative fitness of human immunodeficiency virus type 1 (HIV-1) mutants selected by different protease inhibitors (PIs) in vivo was determined. Each mutant was compared to wild type (WT), NL4-3, in the absence of drugs by several methods, including clonal genotyping of cultures infected with two competing viral variants, kinetics of viral antigen production, and viral infectivity/virion particle ratios. A nelfinavir-selected protease D30N substitution substantially decreased replicative capacity relative to WT, while a saquinavir-selected L90M substitution moderately decreased fitness. The D30N mutant virus was also outcompeted by the L90M mutant in the absence of drugs. A major natural polymorphism of the HIV-1 protease, L63P, compensated well for the impairment of fitness caused by L90M but only slightly improved the fitness of D30N. Multiply substituted indinavir-selected mutants M46I/L63P/V82T/I84V and L10R/M46I/L63P/V82T/I84V were just as fit as WT. These results indicate that the mutations which are usually initially selected by nelfinavir and saquinavir, D30N and L90M, respectively, impair fitness. However, additional mutations may improve the replicative capacity of these and other drug-resistant mutants. Hypotheses based on the greater fitness impairment of the nelfinavir-selected D30N mutant are suggested to explain observations that prolonged responses to delayed salvage regimens, including alternate PIs, may be relatively common after nelfinavir failure.