Evolutionary indicators of human immunodeficiency virus type 1 reservoirs and compartments

Abstract

In vivo virologic compartments are cell types or tissues between which there is a restriction of virus flow, while virologic reservoirs are cell types or tissues in which there is a relative restriction of replication. The distinction between reservoirs and compartments is important because therapies that would be effective against a reservoir may not be effective against viruses produced by a given compartment, and vice versa. For example, the use of cytokines to "flush out" long-lived infected cells in patients on highly active antiretroviral therapy (T. W. Chun, D. Engel, M. M. Berrey, T. Shea, L. Corey, and A. S. Fauci, Proc. Natl. Acad. Sci. USA 95:8869-8873, 1998) may be successful for a latent reservoir but may not impact a compartment in which virus continues to replicate because of poor drug penetration. Here, we suggest phylogenetic criteria to illustrate, define, and differentiate between reservoirs and compartments. We then apply these criteria to the analysis of simulated and actual human immunodeficiency virus type 1 sequence data sets. We report that existing statistical methods work quite well at detecting viral compartments, and we learn from simulations that viral divergence from a calculated most recent common ancestor is a strong predictor of viral reservoirs.

FIG. A1.

Representative latent-cell age distribution.

FIG. A2.

V(t) and component logistic curves for representative parameters.

FIG. 1.

Genetic characteristics of virus from persistent reservoirs and nonreservoir compartments. Panel A illustrates the characteristics predicted for a persistent virus reservoir. The phylogenetic tree depicts viral sequences from blood over four time points (thin lines for times 1 to 3, thick lines for time 4) plus a putative virus reservoir sampled at time 4 (thick gray lines). The relationships between the putative reservoir sequences are summarized in panel B, and the relationships between blood sequences (i.e., nonreservoir) sampled at time point 4 are summarized in panel C. The reservoir sequences in panel A have less temporal structure (indicated by the length of the brackets in panels B and C), higher levels of sequence diversity (the length of lines without arrowheads in panels B and C), and a lower average divergence from the MRCA than contemporaneous virus from the blood (compare arrow lengths in panels B and C). Of these three criteria, the reduction in sequence divergence is the most important for detection of a reservoir. Panels D to F illustrate some examples of samples from a putative reservoir (thick gray lines) that do not fit all of our criteria. The sequences in panel D are temporally structured, since they are phylogenetically intermingled with the contemporaneous blood-derived virus. More importantly, they have high mean divergence from the MRCA. The sequences in panel E have higher mean diversity but also have high levels of divergence equivalent to blood-derived sequences. These relationships would be expected from a compartment having high asymmetric migration from the blood to the compartment with turnover as fast as that of virus in the blood. The sequences in panel F have high divergence and high diversity, as well as high temporal structure. These relationships would be expected for a compartment seeded early in infection and evolving separately from the virus in the blood.

FIG. 2.

Diversity and divergence comparisons between reservoir and contemporary virus samples. The median difference between the reservoir sample divergence (black triangles) and diversity (gray squares) and contemporary divergence and diversity is graphed for three latent-cell half-lives (6, 24, and 44 months) and two effective population sizes (1,000 and 2,500). Each simulation represents 250 iterations. The 95th percentiles of the diversity and divergence distributions are indicated by gray and black bars, respectively.

FIG. 3.

Analysis of sequence data obtained prior to and after ∼5 months of effective therapy. (A) A maximum-likelihood tree with a model of evolution determined to be HKY + G. The tree was constructed by using a neighbor-joining start with an SPR algorithm for branch swapping in PAUP*. (B) Diversity and divergence of sequences within this patient. We note that these quantities are neither normally distributed nor independent measures of distances, and thus, it is not appropriate to use parametric statistics. Instead, we chose to use resampling methods to determine significance levels. The confidence intervals and the mean test are based on 1,000 bootstrap replicates. We found that the samples taken after ∼5 months of HAART had the same evolutionary patterns found in the simulation (N_e of 1,000 and latency half-life of 44 months), whereas diversity was higher and divergence was lower for the latent-reservoir pool. OG, outgroup; BS, bootstrap.