Protease inhibitor-resistant HIV-1 from patients with preserved CD4 cell counts is cytopathic in activated CD4 T lymphocytes

Objective: To evaluate CD4 T-cell cytopathicity of protease inhibitor (PI)-resistant isolates from patients with preserved CD4 cell counts after long-term virologic failure.

Cited Here...: PI-resistant primary isolates from 14 patients with stable or increasing CD4 T-cell counts despite long-term virologic failure during continuous combination therapy were examined. Replication and cytopathicity were assessed in activated peripheral blood mononuclear cell cultures in the presence and absence of PI using titered stocks of primary HIV-1 isolates and during initial viral isolation. Also studied were PI-sensitive isolates from four of these patients after therapy discontinuation and reversion to PI-sensitive virus and from seven antiretroviral drug-naive patients. Coreceptor use, syncytia-inducing (SI) phenotype and protease sequences were determined by standard methods.

Cited Here...: All isolates obtained during continued therapy showed genetic markers of PI resistance and decreased phenotypic susceptibility. PI-resistant SI isolates were highly to moderately cytopathic whereas non-syncytia-inducing isolates were moderately to weakly cytopathic. PI-susceptible and PI-resistant isolates obtained after discontinuation of therapy were equally cytopathic at similar replication levels. The cytopathicity of PI-resistant isolates was not altered by PI and was similar to that of isolates from untreated subjects.

Conclusions: Primary isolates from patients showing virologic rebound without net CD4 T-cell loss during continued therapy are as cytopathic as PI-sensitive isolates with equivalent input infectious titer. As with PI-sensitive isolates, cytopathicity of PI-resistant viruses was determined primarily by coreceptor preference. These results suggest that the sustained immunologic response observed after failure of PI-containing regimens is not due to the emergence of PI-resistant strains that are intrinsically less cytopathic for activated peripheral CD4 lymphocytes.

Introduction

In HIV-infected persons, plasma viremia correlates with CD4 T-cell decline and disease progression ^[1]. Although combination antiretroviral therapy (ART) usually suppresses viremia and increases peripheral CD4 T-cell number ^[2], viral rebound is common ^[3,4].

Recent cohort studies suggest that virologic failure of protease inhibitor (PI)-based therapy may not predict clinical failure ^[3,5,6]. The role of PI in maintaining CD4 T-cell levels during active viral replication is unclear; T-cell gains are greater and more sustained than those seen with failure of reverse transcriptase (RT) inhibitors alone ^[7,8]. This apparent discordance in viral replication and immune preservation may reflect the evolution toward less virulent strains under PI selection, alterations in the host immune system allowing increased CD4 T-cell regeneration or a continued partial virologic response.

This study tested whether PI-resistant primary isolates from patients demonstrating long-term virologic rebound without significant CD4 T-cell loss replicate less cytopathically in activated peripheral blood mononuclear cells (PBMC) than PI-sensitive viruses from the same subject or from asymptomatic drug-naive patients.

Patients and methods

This is a cross-sectional study of HIV-infected adults experiencing long-term virologic failure on continued PI-based therapy ^[3]. The study eligibility criteria were a documented CD4 cell count nadir < 300 cells/μl, at least 18 months of continuous ART including a PI, and documented virologic failure for the previous 24 weeks (HIV RNA > 500 copies/ml and recently > 2500 copies/ml). Four subjects subsequently enrolled in a prospective study evaluating virologic and immunologic effects of treatment interruption ^[9,10].

Viral cocultures were initiated with equal numbers of unstimulated PBMC from study subjects and phytohemagglutanin-stimulated PBMC from HIV negative donors as described ^[11]. For each patient, cocultures were established and maintained in parallel in the presence and absence of 50 nM of the most recently failed PI (indinavir, n = 3; nelfinavir, n = 3; or ritonavir, n = 8). Cocultures from patients undergoing treatment interruption were maintained in up to 500 nM PI to suppress PI-sensitive viral outgrowth.

Isolates from asymptomatic patients naive to ART were generated from frozen archived virus stocks under conditions identical to those used to derive PI-resistant isolates. Viruses were titrated by endpoint dilution on activated PBMC ^[11]. Phytohemagglutinin-stimulated PBMC from uninfected donors were inoculated with 1 × 10³ TCID₅₀(50% tissue culture infectious dose)/10⁶ cells, and cultures were maintained at 2 × 10⁶cells/ml as described above.

The relative proportion of CD4 T-cells was determined by three-color flow cytometry (CD3, CD4, CD8) with a FACSCalibur and CELLQuest (Becton Dickinson Immunocytometry Systems, San Jose, California, USA) and FlowJo (Treestar, San Carlos, California, USA) analysis software. The percentages of CD3+ CD4+ cells from infected cultures were normalized to those in matched uninfected control cultures.

MT-2 and GHOST cl.3 (parental, CXCR4, and CCR5) cell lines were obtained from the National Institutes of Health AIDS Reagent Program (Rockville, Maryland, USA). The MT-2 cell assay was performed as described ^[11]. The GHOST-cell assay ^[12] was performed 36 h after addition of virus isolates at a two-thirds dilution in the presence of 10 μg/ml polybrene.

Protease and reverse transcriptase (RT) gene sequences were determined from frozen culture supernatants and plasma by RT-PCR of the protease and RT reading frame followed by automated cycle sequencing using either nested PCR and dye-labeled terminators ^[13-15] or the TRUEGENE HIV-1 genotyping kit (Visible Genetics, Toronto, Canada).

Wilcoxon rank-sum tests were used to assess differences in p24 concentrations and CD4 T-cell depletion between groups. Spearman rank correlation tests were used to evaluate associations between continuous variables. Analyses were performed with SAS version 6.12 (SAS Institute, Cary, North Carolina, USA).

Results

Fourteen subjects on continuous ART were studied. Four subjects subsequently discontinued ART, and virus was isolated after the predominant species switched from PI-resistant to PI-susceptible ^[9,10]. The median duration of PI-based therapy (baseline to sampling) was 136.7 weeks (range, 109.3-193.4). Median baseline CD4 T-cell count was 120 cells/μl (range, 29-251 cells/μl); median viral load was 4.8 log₁₀copies RNA/ml (range, 4.5-5.7 log₁₀copies RNA/ml). After an initial response to PI-based therapy, all subjects had virologic failure to within o1ne log₁₀ of baseline (Fig. 1a and b). Despite this rebound in plasma viremia, subjects demonstrated an overall sustained CD4 T-cell increase (median gain, 159 cells/μl; range, 30-251 cells/μl).

Fig. 1

The HIV protease and RT gene sequences were determined from all isolates and from plasma of 12 patients. Culture supernatants and plasma virus from any one patient on continued therapy showed multiple primary and secondary genetic markers of PI resistance. Primary isolates obtained after treatment interruption grown in the absence of PI selection showed no such markers. Parallel isolates grown in the presence of PI contained genetic resistance markers identical to those of the same subject obtained during therapy. Consistent with their extensive treatment histories, isolates from these subjects also had multiple genetic markers of RT inhibitor resistance.

These results indicate that isolates obtained during therapy and cultured with or without PI have genetic markers predictive of resistance to multiple classes of antiretroviral drugs and suggest limited genetic drift in vitro. Further, with respect to the protease/RT reading frame, the culture isolates are representative of the major viral quasi-species found in blood plasma.

To determine if PI-resistant primary isolates show an uncoupling of replication and virulence and to determine if such a phenotype is dependent on the presence of PI, we measured viral growth and CD4 T-cell depletion in PBMC cocultures. For each subject, isolates maintained in the absence and presence of PI showed similar kinetic profiles for up to 4 weeks (P > 0.2) (Fig. 1c), indicating phenotypic resistance to PI at 50 nM. After periods of peak viral replication, CD4 T-lymphocyte depletion levels varied widely and were not altered significantly by the presence of PI [with PI: mean, 49.5% (range, 0-87.5%); without PI: mean, 41% (range, 0-89.5%)]. For a given isolate, CD4 T-cell depletion was reproducible in multiple cocultures. Isolates with characteristic minimal, moderate, or extensive pathogenicity at similar replication levels are shown in Fig. 1c. The mean CD4 cell depletion of all paired isolates showed similar cytopathicity in the presence versus absence of PI at all time points (Fig. 1d).

The extent of CD4 T-cell depletion correlated with coreceptor phenotype. In MT-2 and GHOST cell assays, isolates from seven patents each were syncytia-inducing (SI), CXCR4-utilizing or were non-syncytia-inducing (NSI), CCR5-utilizing. When stratified by coreceptor use, SI and NSI isolates showed statistically indistinguishable growth kinetics at steady-state replication in cocultures with or without PI. However, SI/CXCR4-utilizing isolates caused significantly greater mean CD4 T-cell depletion than NSI/CCR5-utilizing isolates (%CD3+ CD4+ at week 2, 29 ± 7.5 versus 72 ± 9.1, P = 0.01).

To compare cytopathicity induced by drug-resistant versus drug-sensitive isolates from the same patient, phenotypically and genotypically distinct virus pairs were isolated after extended ART interruption in four patients. Viral growth kinetics and CD4 T-cell depletion in representative parallel cocultures of matched PI-resistant and PI-sensitive isolates are shown in Fig. 2a. For all patients at equal replication levels, T-cell depletion from the PI-sensitive strains was indistinguishable from that seen in the resistant strains.

Fig. 2

Next, activated PBMC were infected with isolates (three SI, three NSI) at equivalent titers from PI-resistant subjects and similarly derived isolates (two SI and five NSI) from seven untreated progressors showing continued CD4 T-cell depletion. Viral growth kinetics and CD4 T-cell depletion after peak viral replication were assessed (Fig. 2b and c). Consistent with our results from all 14 PI-resistant isolates, SI isolates were substantially more cytopathic at similar replication levels than NSI isolates regardless of drug susceptibility phenotype (P = 0.0043).

Discussion

Our results indicate that primary HIV-1 isolates with multiple markers of genotypic resistance and reduced phenotypic drug susceptibility from patients showing substantial long-term viremia without immunologic signs of disease progression can be highly cytopathic in activated, primary CD4 T cells, particularly SI/CXCR4-utilizing strains. These results are consistent with increased expression of CXCR4 on recently activated T lymphocytes ^[16,17], as seen in PBMC cocultures. When stratified by HIV-1 coreceptor phenotype, PI-resistant isolates were no less virulent at equivalent replication levels and with equal input infectious titers than PI-susceptible isolates from the same subject or from untreated naive patients with declining CD4 cell counts. Similarly, these PI-resistant viruses can be highly cytopathic for CD4 T cells in human tonsil organ cultures in the absence of exogenous stimulation ^[18], where the chemokine receptor distribution closely resembles that of activated PBMC ^[19]. These results do not support the hypothesis that PI-resistant viruses have evolved during long-term drug exposure to a state where replication is uncoupled from virulence in activated mature peripheral T cells.

It is possible that the decreased virulence of PI-resistant isolates from patients with sustained immunologic responses may be more evident in target tissues other than differentiated PBMC. In human fetal thymocyte infections with either paired primary isolates or chimeric viruses, PI-resistant viruses showed significantly attenuated growth profiles when compared to PI-sensitive ones ^[20] (C. A. Stoddart, T. J. Liegler, F. F. Mammano, V. D. Linquist-Stepps, M. Hayden, J. M. Harris, S. G. Deeks, R. M. Grant, F. Clavel and J. M. McCune, unpublished data). Even slightly reduced infectivity and cytopathicity in lymphoid precursor cells or non-cycling T cells could shift the balance in favor of T-cell production over destruction, leading to a net increase in CD4 T-cell count. An intriguing possibility is that PI-resistant viruses are associated with decreased clinical progression due to a restricted host tissue range, sparing cells essential for lymphoid cell regeneration.

Alternatively, CD4 T-cell preservation may simply reflect decreased absolute viremia due to partial drug suppression, rather than any qualitative change in the virus or specific cytopathicity. On average, patients with a transient response to therapy showed partial viral rebound to levels below those before PI-based therapy ^[3,8] (Fig. 1b). Interruption of ART in patients with long-term virologic failure results in rapid outgrowth of PI-sensitive virus, increased plasma viremia and decreased CD4 T-cell count, suggesting partial viral suppression and attenuated fitness of PI-resistant viruses in vivo^[9,10]. Partially suppressed plasma viremia may allow host immune cell regeneration to match or exceed virus-induced cell destruction, resulting in preserved peripheral CD4 T-cell counts.

In summary, primary isolates from subjects with continued CD4 T-cell accumulation after virologic failure replicated cyotopathically in activated PBMC. PI did not alter replication and cytopathicity, which were equivalent to the levels seen with matched PI-susceptible isolates. These results suggest that the sustained immunologic response in patients with virologic failure of PI-containing regimens may be due to decreased viral replication or reduced HIV-1-associated pathogenicity in a compartment other than mature activated T cells. Further investigation into the virologic basis of CD4 T-cell maintenance in the setting of continued viral replication using primary isolates may address fundamental questions in HIV-1 virulence and pathogenesis, as well as guide the timing of changes in antiviral therapy after virologic failure.

Acknowledgements

The authors thank B. Drews, M. Thounaojam, J. Javier and P. Wong at the Gladstone/UCSF Laboratory of Clinical Virology for expert technical help with viral cultures and HIV-1 sequencing. We also thank J. Barbour, S. Ordway, G. Howard and C. Stoddart for their critical reading of this manuscript. We are appreciative of the patients in the SFGH positive Health Program for their continued participation in research studies.