Mathematical modeling of ultradeep sequencing data reveals that acute CD8+ T-lymphocyte responses exert strong selective pressure in simian immunodeficiency virus-infected macaques but still fail to clear founder epitope sequences

Abstract

The prominent role of antiviral cytotoxic CD8(+) T-lymphocytes (CD8-TL) in containing the acute viremia of human and simian immunodeficiency viruses (HIV-1 and SIV) has rationalized the development of T-cell-based vaccines. However, the presence of escape mutations in the acute stage of infection has raised a concern that accelerated escape from vaccine-induced CD8-TL responses might undermine vaccine efficacy. We reanalyzed previously published data of 101,822 viral genomes of three CD8-TL epitopes, Nef(103-111)RM9 (RM9), Tat(28-35)SL8 (SL8), and Gag(181-189)CM9 (CM9), sampled by ultradeep pyrosequencing from eight macaques. Multiple epitope variants appeared during the resolution of acute viremia, followed by the predominance of a single mutant epitope. By fitting a mathematical model, we estimated the first acute escape rate as 0.36 day(-1) within escape-prone epitopes, RM9 and SL8, and the chronic escape rate as 0.014 day(-1) within the CM9 epitope. Our estimate of SIV acute escape rates was found to be comparable to very early HIV-1 escape rates. The timing of the first escape was more highly correlated with the timing of the peak CD8-TL response than with the magnitude of the CD8-TL response. The transmitted epitope decayed more than 400 times faster during the acute viral decline stage than predicted by a neutral evolution model. However, the founder epitope persisted as a minor population even at the viral set point; in contrast, the majority of acute escape epitopes were completely cleared. Our results suggest that a reservoir of SIV infection is preferentially formed by virus with the transmitted epitope.

FIG. 1.

Escape dynamics and viral load kinetics during acute SIV infection. Dynamics for the percentage of each variant of Nef_103-111RM9 from four Mafa-A1*063-positive Mauritian cynomolgus macaques (cy0161, cy0162, cy0163, and cy0165) (left) and of each variant of Tat_28-35SL8 from four Mamu-A1*001-positive Indian rhesus macaques (rh2122, rh2124, rh2126, and rh2127) (right) are plotted as a function of time (days postinfection). The wild-type (transmitted) epitopes, Nef_103-111RM9, RPKVPLRTM, Tat_28-35SL8, and STPESANL, are shown in black, and other variants are color coded. For instance, in cy0161, at day 7 and day 14 postchallenge, a single dominant epitope, the transmitted (founder) sequence, prevails. At days 21, 28, and 56 postinfection, three variant epitopes are prominent in addition to the transmitted one. By day 140, the mutant RPQVPLRTM (denoted K3Q) was positively selected and accounted for more than 60% of total RM9 epitopes in this animal. The kinetics of plasma viral RNA also are plotted for each macaque in the second and fourth columns. Note that animals rh2122, rh2126, and rh2127 were vaccinated prior to infection with SIVmac239 (54), while all other animals were unvaccinated prior to infection (8).

FIG. 2.

Shannon entropy dynamics. The Shannon entropy is plotted over time for each infected animal. Here, all minor variants were included when we estimated the Shannon entropy. Entropy was maximal at 21 to 28 days postinfection in all animals except rh2124 and then decreased as the set point viral load was established. The entropy values at 140 days postinfection (set point) and at 28 days postinfection (viral decline after acute peak viremia) were statistically different (P = 0.00088, paired t test). The peak timing of the entropy coincided with the initial decline from peak viremia.

FIG. 3.

Fit of the mathematical model for first CD8-TL escape to experimentally acquired sequence data. The mathematical model for the first viral CD8-TL escape within two escape-prone epitopes, RM9 (first row) and SL8 (second row), was fit to the experimental data (black dots) from the infected animals. The solid blue line is the best fit of the model to the data using beta regression, and the solid red line is the best fit of the model using nonlinear least-squares regression. When only one line is visible they are overlapping. The ratio of each first escape mutant epitope to the total population of the wild-type plus the mutant virus, m(t)/[w(t) + m(t)], is plotted with days postinfection. The sequences of each first escape variant are shown in individual panels. For instance, a substitution from RPKVPLRTM to RSKVPLRTM in the animal cy0161 is denoted P2S. Here, we limited the fitting range up to day 30, because the proportion of the wild epitope can exceed the proportion of the first escape mutant at later time points. The estimated rates of CD8-TL escape and the timing of the first CD8-TL escape with asymptotic standard errors are listed in Table 1. The mean estimated first escape rate is 0.36 day⁻¹.

FIG. 4.

Comparison of two data-fitting methods for modeling early CT8-TL escape. (A) Timing of the first CD8-TL escape for each animal estimated from the method of beta regression is represented by a blue star, and that from the method of nonlinear least squares is represented by a red bar. Blue and red vertical lines represent estimated asymptotic standard errors from beta regression and from least squares, respectively. (B) Rate of CD8-TL escape for each animal estimated from the methods of beta regression (blue star) and least squares (red cross). Blue and red vertical lines represent the asymptotic standard errors estimated from least squares and beta regression, respectively. As listed in Table 1, while the least-squares method provides smaller values of the sum of squared errors, beta regression estimates generally have smaller standard errors for the parameter estimates. Although data from each animal were fit to the model separately with no shared information, the beta regression estimates are more consistent across the animals than the least-squares estimates.

FIG. 5.

Escape dynamics of the CM9 epitope and viral load kinetics. The kinetics of changes in the Gag_181-189CTPYDINQM (CM9) epitope are presented as a function of time. Viral load kinetics in each macaque also are plotted (second and fourth columns). In animal rh2122, the transmitted (wild-type) CM9 epitope was dominant for the first 200 days and subsequently was replaced by a single mutant, GTPYDINQM (denoted C1G). In animals rh2124 and rh2126, the transmitted (wild-type) CM9 epitope was preserved for the entire 400-day time course of our analysis. In macaque rh2127, the mutant CCPYDINQM (T2C) appeared slowly, and at day 400, the percentage of viruses containing the mutant epitope was comparable to that of the transmitted (wild-type) virus.

FIG. 6.

Association between the estimated timing of the first CD8-TL escape and (A) the magnitude of or (B) timing of the acute peak CD8-TL response. The correlation between the estimated timing of the first CD8-TL escape and the magnitude of the CD8-TL response (A) or the timing of the peak CT8-TL response (B) was measured for eight macaques. Timing of first escape was defined as the time point when the first escape mutant comprised half of the sum of the wild-type (transmitted) and mutant virus populations in a given animal. Magnitude of CTL was defined as the area under the curve of the expansion kinetics of each tetramer-positive CD8⁺ T cell in peripheral blood (see Fig. S1 in the supplemental material). Peak timing of CTL was defined as the time point when the number of tetramer-positive CD8⁺ T cells was greatest; note that PBMC samples for tetramer analysis were collected at days 7, 14, 21, and 28 for the rhesus macaques (rh) and at days 10, 14, 17, 21, 24, and 28 for the cynomolgus macaques (cy). The earlier estimated timing of the first CD8-TL escape was correlated with an earlier peak in the acute CTL response (r = 0.76, P = 0.027).

FIG. 7.

Kinetics of the loss of the transmitted (wild-type) epitope. (A) The percentage of each transmitted (wild-type) epitope (RM9 and SL8) is plotted as a function of time postchallenge for each individual animal. The red solid lines denote the predicted dynamics based on the acute-sequence evolution model, which assumes the neutral evolution of the epitope population without any selection (9, 29, 33). Ninety-five percent confidence intervals were calculated by sampling 1,414 sequences at each time from 1,000 Monte Carlo simulation runs. This sample size was chosen based on the average number of sequence reads from the pyrosequencing. The rate of the decay of the transmitted epitope in each macaque is greater than that predicted by the model, suggesting that there is strong selection pressure on the epitope region. (B) The average decay dynamics of the transmitted (wild-type) epitopes (RM9 and SL8) from all eight macaques is plotted, together with the standard deviations. The lines denote the best fit of the exponential decay curve to the data for two distinct stages of viral escape: (i) the period between 14 and 28 days (red) and (ii) after 28 days (light blue). The rate of loss was 1.4 × 10⁻¹ day⁻¹ between 14 and 28 days, reflecting rapid clearance for transmitted virus during the resolution of acute infection. The rate of loss was 1.0 × 10⁻² day⁻¹ in the period of the resolution of the acute phase of infection and progression toward the establishment of steady-state virus load.