Changes in human immunodeficiency virus type 1 fitness and genetic diversity during disease progression

Abstract

This study examined the relationship between ex vivo human immunodeficiency virus type 1 (HIV-1) fitness and viral genetic diversity during the course of HIV-1 disease. Primary HIV-1 isolates from 10 patients at different time points were competed against control HIV-1 strains in peripheral blood mononuclear cell (PBMC) cultures to determine relative fitness values. Patient HIV-1 isolates sequentially gained fitness during disease at a significant rate that directly correlated with viral load and HIV-1 env C2V3 diversity. A loss in both fitness and viral diversity was observed upon the initiation of antiretroviral therapy. A possible relationship between genotype and phenotype (virus replication efficiency) is supported by the parallel increases in ex vivo fitness and viral diversity during disease, of which the correlation is largely based on specific V3 sequences. Syncytium-inducing, CXCR4-tropic HIV-1 isolates did have higher relative fitness values than non-syncytium-inducing, CCR5-tropic HIV-1 isolates, as determined by dual virus competitions in PBMC, but increases in fitness during disease were not solely powered by a gradual switch in coreceptor usage. These data provide in vivo evidence that increasing HIV-1 replication efficiency may be related to a concomitant increase in HIV-1 diversity, which in turn may be a determining factor in disease progression.

FIG. 1.

Estimation of ex vivo HIV-1 fitness and env C2V3 genetic diversity for patient K. An identical strategy was used to estimate ex vivo HIV-1 fitness and genetic diversity for additional patients. (a) Patient PBMC were used to propagate primary HIV-1 isolates for competition and were the source for clonal sequence analysis of the env C2V3 region. (b) A representative HTA analysis of competitions between patient K primary isolates (K44 and K69) and NSI/R5 control strains (A-92RW009 and B-92BR017) is shown. K44 and K69 were both able to outcompete the low-fitness control virus A-92RW009, whereas K69 replicated much more efficiently than K44 in competition with B-92BR017. All 34 primary HIV-1 isolates from 10 patients were competed against these NSI/R5 control strains, as well as two SI/X4 control strains. (c) A phylogenetic analysis of env C2V3 clonal sequences from each timepoint for patient K was performed by using PAUP* 4.0b10. Colored circles at terminal nodes indicate the sample timepoint from patient K. Significant bootstrap values are indicated on branches (, >90; , >80; , >70). Genetic divergence from the consensus sequence at the first timepoint for each patient is shown as synonymous (d_S) and nonsynonymous (d_N) mutations. The d_S/d_N ratio within each set of clonal sequences at each time point is also displayed.

FIG.2.

Trends in HIV-1 ex vivo fitness, genetic diversity, and disease progression for six ARV naive patients (panels a to f). The time scale is shown in months with the baseline (zero) time point indicating the first available sample for each patient. The number in parentheses at the baseline time point indicates the number of months from the patient's first positive HIV test to the first sample. Vertical bars indicate the total relative fitness for each HIV-1 primary isolate derived from competitions against the four control strains. Each bar is marked with the HIV-1 isolate name (by patient and month, e.g., K11 and K44) and coreceptor usage is shown by the color of the bar: light gray, R5; dark gray, R5/X4. The CD4 cell count is indicated with red closed circles, whereas the plasma viral load is indicated with green open circles. Genetic diversity and divergence are indicated in blue with squares and triangles, respectively.

FIG.3.

Trends in HIV-1 ex vivo fitness, genetic diversity, and disease progression for four ARV experienced patients (panels a to d). The time scale is shown in months, with the baseline (zero) time point indicating the first available sample for each patient. The number in parentheses at the baseline time point indicates the number of months from the patient's first positive HIV test to the first sample. Vertical bars indicate the total relative fitness for each HIV-1 primary isolate derived from competitions against the four control strains. Each bar is marked with the HIV-1 isolate name (by patient and month, e.g., U0, U11, etc.), and coreceptor usage is indicated by the color of the bar: white, R5; gray, R5/X4. The CD4 cell count in indicated with red closed circles, while the plasma viral load is indicated with green open circles. Genetic diversity and divergence are indicated in blue with squares and triangles, respectively. Patients received antiretroviral treatment for the time periods indicated by bars next to the drug name: AZT, zidovudine; 3TC, lamivudine; ddI, didanosine; ddC, zalcitabine; d4T, stavudine; IDV, indinavir sulfate; SQV, saquinavir mesylate; RTV, ritonavir; NFV, nelfinavir. Patients A and Q received AZT treatment for 39 and 10 months prior to the zero time point, respectively.

FIG. 4.

Interpatient correlates of ex vivo HIV-1 fitness. Pearson product moment correlations were determined for all 34 HIV-1 primary isolates analyzed in the present study. Ex vivo HIV-1 fitness correlated with time since first positive HIV test (a), genetic diversity (b), CD4 cell count (c), and plasma viral load (d). The 95% confidence intervals for the line are displayed.

FIG. 5.

Intrapatient correlates of ex vivo HIV-1 fitness. Slopes of the correlation between ex vivo HIV-1 fitness and genetic diversity, plasma viral load and CD4 cell count were plotted for each patient (n = 10). The significance of the mean slope of each correlation (null hypothesis = 0) was tested by using a one-sample t test. Outlier points are displayed as open circles, and analyses excluding these points are displayed.

FIG. 6.

Comparing fitness of primary isolates from the same patient in head-to-head competitions. HIV-1 isolates from patient T at 0 and 21 months (i.e., ca. 12 and 33 months after infection) were competed together in a PBMC culture as described in Materials and Methods. HTA was performed by using the C4 HIV-1 probe. Heteroduplex bands representing virus T0 and T21 in monoinfections and direct dual infections are shown in panel (a). (b) A phylogenetic analysis of env C2V3 clonal sequences from each time point for patient T was performed by using PAUP* 4.0b10. Shaded circles at terminal nodes indicate the sample time point from patient T. Significant bootstrap values are indicated on branches (, >90; , >80; , >70). The boxes highlight terminal nodes from the HIV-1 T0 and T21 sequences within the patient PBMC, i.e., the same PBMC sample used to propagate the viruses. The results of T0 and T21 (c) or I0, I10, and I21 (e) competitions against control strains are presented as individual pie charts. The gray slices represent the percentage of sample virus growth against one of the four control strains, which are represented by black slices (A-92RW009, B-92BR017, A-92UG029, and E-CMU06). Total relative fitness values (W) derived from control virus competitions are indicated in panels c and e. The results from direct competitions between T0 and T21 are shown in panel d. Competitions I0 versus I22 and I10 versus I22 are shown in panel f. Gray and black slices of each pie chart in panels d and f are as labeled in the figure.

FIG. 7.

Competitive fitness versus PSSM score. The total relative fitness of each primary HIV-1 isolate was compared to the PSSM score based on V3 sequences from patient PBMC (a) and the propagated primary HIV-1 isolate (b). The raw PSSM scores were normalized by mapping the minimum score to −1, the maximum score to 1, and intermediate scores to points between −1 and 1 reflecting their original fractional position between the minimum and maximum.