TEXT

Top Abstract Text References

A restricted sequence heterogeneity is generally seen during primary human immunodeficiency virus type 1 (HIV-1) infection (5, 6, 21, 25, 26), after which the viral evolution rates vary among patients (12, 18, 22, 24). Nonsynonymous nucleotide substitutions occur less frequently in progressors than in nonprogressors (4, 11, 16, 17, 22, 23), reflecting a lower selective pressure due to a less efficient immune response. However, the initial strains have seldom been included, and sampling intervals have often exceeded 1 year. This approach may underestimate the true mutation rate if substitutions appear rapidly during primary HIV-1 infection (6) or if original viral strains reappear in later stages of infection. Our aims were to obtain information about the kinetics of viral variability and heterogeneity during the first years of HIV-1 infection and to analyze viral sequence changes in connection with clinical complications.

Samples (n = 64) were collected from seven homosexual men (patients A through E, G, and H) and one intravenous drug addict (patient F) with symptomatic primary HIV-1 infections (Table 1)( 6). Three patients developed symptoms (C, D, and F) during follow-up. Plasma HIV-1 RNA levels were determined by the Roche Amplicor HIV-1 Monitor test. The numbers of CD4⁺ cells were determined by flow cytometry. V3 and p17 regions were amplified from extracted RNA by reverse transcription and nested PCR (6). Amplified products from samples with low copy numbers were pooled. The V3 regions were analyzed in duplicate. A single-stranded DNA template was obtained by using streptavidin-coated magnetic beads (Dynabeads; Dynal AS). Direct solid-phase sequencing was performed with an AutoRead Sequencing Kit (Pharmacia) and an automated laser fluorescent sequencer (A.L.F. Express; Pharmacia Biotech) (6, 14, 20). Sequences of 258 nucleotides (nt) (V3) and 390 or 405 nt (p17) were obtained. Sequence heterogeneity was manually determined, blindly, by identifying the number of polymorphic sites (6, 14). A minor peak had to reach 15% of the total peak height to be included. Quasispecies consisting of 10 to 25% of the virus population can be reliably detected by this method (8, 9). Ambiguities were indicated by using the International Union of Pure and Applied Chemistry (IUPAC) code. The data were manually aligned, and the distance matrix was determined by DNADIST by using the maximum-likelihood method. Phylogenetic trees were generated by using the program DNAML with the transition/transversion parameter set to 3.0 (p17) or 1.42 (V3) (10). Bootstrap values were estimated by the neighbor-joining (NJ) method. The dendrograms were performed by NJ, by using the Jin and Nei methods (gamma distribution set to 0.38 [V3] or 0.25 [p17]) with the Kimura 2 parameter algorithm, included in TREECON (19). There was no difference in branching order between trees constructed by DNAML or NJ. Frequencies of synonymous (rd_S) and nonsynonymous (rd_N) nucleotide substitutions were estimated by pairwise comparison of the initial major sequence with each later sequence, by applying the Jukes-Cantor formula (13) included in MEGA (7). Positions containing gaps were excluded.

The HIV-1 sequences of each patient clustered together and were divergent from those of other patients. No positively charged amino acids were found at positions 11 and 25, consistent with a non-syncytium-inducing phenotype. The intrapatient nucleotide divergence per site ranged from 1.29 × 10³ to 54.18 × 10³ (median, 7.72 × 10³) in the V3 region and from 0.42 × 10³ to 37.12 × 10³ (median, 1.85 × 10³) in the p17 region (P < 0.001 by the Wilcoxon sign rank test). In the V3 region, rd_N were more common than rd_S (P < 0.001 by the Wilcoxon sign rank test), but this was not the case in the p17 region (P = 0.62).

Accumulating mutations in the major sequence were seen in patients A, B, E, G, and H, although the rate was very limited in patient H (Fig. 1 and 2). In patient E, almost no sequence changes were seen after the initiation of therapy. None of these patients developed symptoms or advanced immunodeficiency. The CD4⁺ cell declines were low (range, 1.7 to 15.5 cells/liter per month).

View larger version (86K): [in this window] [in a new window]   FIG. 1.   The intersample DNA distance, rdS, and rdN of the V3 and p17 regions are shown in relation to the number of days after the onset of primary HIV-1 infection. Each data point represents pairwise comparison with the first major sequence.

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Unrooted NJ trees, indicating evolutionary patterns of the major V3 and p17 sequences of patients A, C, D, and F. Boldface numbers to the right of the trees consist of the letter designating the patient and the number of days after the onset of primary HIV-1 infection. A subtype B sequence (G006 from patient G) or a subtype E (V3)/A (p17) sequence (isolate CM240) is used as an outgroup. The numbers at the nodes are percentages of support in 500 bootstrap resamplings. The scale indicates the genetic distance in each tree. A longitudinal accumulation of viral divergence is illustrated by patient A. Nucleotide changes back and forth in relation to the initial major sequence were present in patients C, D, and F. The sequence of the p17 region in the sample taken from patient F on day 347 is missing due to an ambiguous sequence. *, consensus sequence derived from 20 clones (1).

In patients C, D, and F, a new major sequence appeared within the first 2 months. Thereafter, further changes in the major sequence due to reverse mutations or reappearance of the original sequence were identified. These changes coincided with the onset of disease progression and increased viral load. The CD4⁺ cell slopes were steep in patients C (35.9) and D (44.4) but not in patient F (+13.5). In patient C, the initial major viral sequence reappeared completely at day 236, after 2 months of diarrhea and/or dermatitis. The mutated major sequence from day 54 reappeared (day 746) when the symptoms disappeared and the CD4⁺ cells had increased after the initiation of combination therapy. In patient D, five reverse mutations toward the original major sequence were seen at the onset of a multidermatomal herpes zoster (day 265). The virus population contained a mixture of original and mutated sequences at several positions (days 152, 280, and 351). The number of CD4⁺ cells declined but increased after the resolution of the symptoms. In patient F, diarrhea (day 324) was followed by pneumonia and septicemia (day 368). The major viral sequence at day 347 was almost identical with the initial major sequence. In contrast, the major viral sequences obtained before and after this occasion differed substantially from the original sequence.

In the p17 region a sequence variability similar to that in the V3 region was detected, although to a lower extent in most cases (Fig. 1). An early outgrowth of a new major sequence was seen in patients C, D, and F. Later, changes toward the original sequence were found.

The sequence heterogeneity was higher in the V3 region than in the p17 region (P < 0.001 by the Wilcoxon sign rank test). During the first 6 months, the highest heterogeneity was detected in samples from patients C, D, and F, who developed symptoms and rapid major sequence changes. In samples from patients C, D, F, and H, a continuously high heterogeneity was detected. For patients A, B, E, and G, the V3 heterogeneity increased slowly during follow-up.

The HIV-1 population adapts to changes in the host environment, resulting in the emergence of new viral variants (3). Different patterns of replacement of early virus variants have been reported (4). In our study three patterns were present. Four patients (A, B, E, and G) exhibited a successive replacement of the original major sequence. In patient H, the major V3 sequence at day 736 differed only slightly from that at day 6, suggesting that this virus variant was optimally adapted to the new host environment (12). A third pattern, with a rapid shift to a new major viral sequence, was recorded for patients C, D, and F, who also had the highest viral heterogeneity. These early viral population kinetics could reflect the transmission of several HIV-1 strains and/or a pronounced antiviral immune response that resulted in a rapid selection of the most fit master sequence (6).

The original major HIV-1 sequence reappeared completely (patients C and F) or to some extent (patient D) in connection with clinical deterioration and increased plasma viral load. The possibility that the reappearance of founder virus sequences was due to a recombination between V3 and/or p17 regions of founder virus and dominating quasispecies cannot be excluded. It is also possible that reverse mutations toward the original sequence caused the reappearance. Thus, the V3 evolution may be constrained due to the interaction with chemokine receptors. Alternatively, a developing immunodeficiency could have resulted in both a reappearance of the original HIV-1 population and a concomitant appearance of opportunistic infections. Perturbations of the immune system induced by clinical complications (2) might also have contributed to the switch back toward the original sequence. Immune activation has been associated with reversible shifts in the composition of plasma viral quasispecies, possibly as a result of the induction of virus from latently infected cells (15). The mutated viral populations dominated again after the resolution of the clinical complications, indicating their higher fitness in a more stable intrahost environment.

In conclusion, our study shows that the temporal variation of the viral population may differ substantially among individuals during the first years following seroconversion. Also, virus sequences present before seroconversion may reappear as the dominating sequences if the intrahost environment is changed in a direction which is favorable for the initial viral quasispecies.

Nucleotide sequence accession numbers. The sequences identified in this study have been submitted to GenBank under accession no. AF068484 through AF068533 and AF062040 through AF062068.