Determinant of HIV-1 mutational escape from cytotoxic T lymphocytes

Abstract

CD8+ class I-restricted cytotoxic T lymphocytes (CTLs) usually incompletely suppress HIV-1 in vivo, and while analogous partial suppression induces antiretroviral drug-resistance mutations, epitope escape mutations are inconsistently observed. However, escape mutation depends on the net balance of selective pressure and mutational fitness costs, which are poorly understood and difficult to study in vivo. Here we used a controlled in vitro system to evaluate the ability of HIV-1 to escape from CTL clones, finding that virus replicating under selective pressure rapidly can develop phenotypic resistance associated with genotypic changes. Escape varied between clones recognizing the same Gag epitope or different Gag and RT epitopes, indicating the influence of the T cell receptor on pressure and fitness costs. Gag and RT escape mutations were monoclonal intra-epitope substitutions, indicating limitation by fitness constraints in structural proteins. In contrast, escape from Nef-specific CTL was more rapid and consistent, marked by a polyclonal mixture of epitope point mutations and upstream frameshifts. We conclude that incomplete viral suppression by CTL can result in rapid emergence of immune escape, but the likelihood is strongly determined by factors influencing the fitness costs of the particular epitope targeted and the ability of responding CTL to recognize specific epitope variants.

Figure 1.

HIV-1–specific CTL clones differentially inhibit replication of HIV-1 point mutants. A panel of HIV-1 NL4–3–based viruses containing point mutations in the SL9 epitope was screened for inhibition of viral replication by the SL9-specific CTL clones 161JxA14 and 18030D23. Viral replication as assessed by quantitative p24 ELISA was measured 7 d after addition of CTL to acutely infected T1 cells. Inhibition of each virus by each clone (in log₁₀ units) was determined by comparing p24 production in the presence and absence of CTL. Plotted in this figure are the ratios of suppression of HIV-1 with each SL9 mutation relative to virus with the index SL9 sequence SLYNTVATL.

Figure 2.

Scheme for passaging HIV-1 under selective pressure by CTL. Defined starting virus was grown in T1 cells (able to present antigen to the selecting CTL clone) or a control cell line (unable to present antigen to the CTL due to HLA mismatching or antigen transport defect) in the presence of a selecting CTL clone for 7 d. After this first passage, proviral DNA and supernatant HIV-1 were harvested. These were then used for the analyses of sensitivity to CTL and epitope sequence (Figs. 3, 5, 6, and 7). The virus was also further passaged under the same conditions for a second passage, after which proviral DNA and supernatant HIV-1 were again harvested for analysis.

Figure 3.

The SL9-specific CTL clone 161JxA14 is rapidly escaped by HIV-1. HIV-1 IIIB previously passaged (Fig. 2) under selective pressure in HLA A2-matched T1 cells (bottom row) or no selective pressure in HLA mismatched H9 cells (top row) by the CTL clone 161JxA14 (recognizing the Gag epitope SL9) for 1 (first column) or 2 (second column) wk of passage was tested for susceptibility to this clone or the control clone 68A62 (recognizing an RT epitope). Replication of the control and 161JxA14-selected viruses in T1 cells in the absence (open triangles) or presence (open and closed circles) of CTL is shown. These control and CTL-selected viruses were also sequenced for the SL9 epitope at limiting dilution, and the viral sequences are indicated below each graph. Similar results (decreased susceptibility to inhibition and -----I----- mutation observed after two weeks) were obtained in three experiments using HIV-1 IIIB and one using the molecular clone NL4–3.1 (Table II, Fig. 5, and unpublished data).

Figure 4.

The SL9 variant -----I--- is poorly recognized by 161JxA14. Lysis of target cells labeled with serial dilutions of exogenously added SL9 or -----I--- variant peptides was assessed by standard chromium release assay.

Figure 5.

Another SL9-specific CTL clone 18030D23 does not rapidly select resistant virus. The experiment depicted in the Figure 3 was repeated with another clone 18030D23 (derived from another infected person, also recognizing the SL9 epitope) using HIV-1 NL4–3.1 (see also Table II). Again, control virus was passaged under no selective pressure (this time in TAP-deficient T2 cells) or with selective pressure (in HLA A2–matched T1 cells) by the selecting clone (18030D23). These viruses were tested for inhibition by 18030D23 or 68A62 as indicated, and SL9 sequences for these tested viruses are again shown. In a repeat experiment using strain IIIB, neither phenotypic resistance nor SL9 mutation was observed for 18030D23-selected virus (unpublished data).

Figure 6.

18030D23 readily inhibits the escape virus for 161JxA14. HIV-1 NL4–3.1 viruses previously passaged (Fig. 2) for 2 wk under no selective pressure in TAP-deficient T2 cells (A) or selective pressure in HLA A2–matched T1 cells (B) by 161JxA14 were tested for susceptibility to viral inhibition by clones 161JxA14 and 18030D23. Below each graph are the SL9 sequences of the tested viruses.

Figure 7.

The Nef-specific CTL clone STD11 very rapidly selects resistant virus that contains multiple mutations. HIV-1 NL4–3.1 previously passaged (Fig. 2) under selection by STD11 (recognizing the Nef epitope KEKGGLEGL) was tested for phenotypic sensitivity to STD11 or 68A62, as well as genetic changes in the Nef epitope recognized by STD11. Frameshift = frameshift mutations upstream of the epitope. Early stop = substitution mutation upstream of the epitope leading to a stop codon. Similar results were obtained in another experiment using STD11 and HIV-1 IIIB (unpublished data) and two experiments using another clone KM3 of the same specificity (also with both NL4–3.1 and IIIB). Similar results (polyclonal epitope mutations and upstream frameshifts) were also obtained with Nef-specific CTL recognizing an HLA B15-restricted CTL clone and H9 target cells (unpublished data).