Effect of immune activation on the dynamics of human immunodeficiency virus replication and on the distribution of viral quasispecies

Abstract

Virus replication in a human immunodeficiency virus (HIV)-infected individual, as determined by the steady-state level of plasma viremia, reflects a complex balance of viral and host factors. We have previously demonstrated that immunization of HIV-infected individuals with the common recall antigen, tetanus toxoid, disrupts this steady state, resulting in transient bursts of plasma viremia after immunization. The present study defines the viral genetic basis for the transient bursts in viremia after immune activation. Tetanus immunization was associated with dramatic and generally reversible shifts in the composition of plasma viral quasispecies. The viral bursts in most cases reflected a nonspecific increase in viral replication secondary to an expanded pool of susceptible CD4(+) T cells. An exception to this was in a patient who harbored viruses of differing tropisms (syncytium inducing and non-syncytium inducing [NSI]). In this situation, immunization appeared to select for the replication of NSI viruses. In one of three patients, the data suggested that immune activation resulted in the appearance in plasma of virus induced from latently infected cells. These findings illustrate certain mechanisms whereby antigenic stimulation may influence the dynamics of HIV replication, including the relative expression of different viral variants.

FIG. 1

Kinetics of plasma viremia postimmunization in patients 2 through 4. Patient 1 was mock immunized. Viral RNA copies (10³/ml) of plasma are annotated on the ordinate, and days postimmunization are annotated on the abscissa. Plasma samples were obtained for sequencing of cDNA at baseline, during peak viremia, and at last follow-up and are indicated (↑). PBMC samples obtained for proviral sequencing at day 0 in patients 2 through 4 and at day 30 in patient 2 are indicated (#). LNMC samples obtained for proviral sequencing at days 0 and 30 postimmunization in patient 2 are indicated (∗). All patients were asymptomatic and were not receiving antiretroviral agents.

FIG. 2

(Left) Phylogenetic analysis of the C2 to V5 region of all sequences (cDNA) sampled from patient 1. Viral sequences are shown by colored symbols according to their sampling time and tissue compartment. Three groups are indicated as A, B, and C; unclassified sequences are indicated by ?. Bootstrap probabilities are shown by percentages. ∗, This sequence has an artifactually long branch length due to a deletion of the V3 region. The following prototype HIV env subtype B strains were used as outgroups: M79345 (derived from a primary isolate), M96155 (from isolate 89.6), M89973 (YU-2), M65024 (HIVSFAAA), and M60472 (ADA). (Right) Frequency of detection of sequence variants in patient 1 (mock immunized) in sequential plasma samples. All sequences were derived from RNA.

FIG. 2

(Left) Phylogenetic analysis of the C2 to V5 region of all sequences (cDNA) sampled from patient 1. Viral sequences are shown by colored symbols according to their sampling time and tissue compartment. Three groups are indicated as A, B, and C; unclassified sequences are indicated by ?. Bootstrap probabilities are shown by percentages. ∗, This sequence has an artifactually long branch length due to a deletion of the V3 region. The following prototype HIV env subtype B strains were used as outgroups: M79345 (derived from a primary isolate), M96155 (from isolate 89.6), M89973 (YU-2), M65024 (HIVSFAAA), and M60472 (ADA). (Right) Frequency of detection of sequence variants in patient 1 (mock immunized) in sequential plasma samples. All sequences were derived from RNA.

FIG. 3

(Left) Phylogenetic analysis of sequences (cDNA and proviral) from patient 2. Samples from plasma represent RNA sequences, and those from PBMC and LNMC are proviral. Note two major groups, A and B, which are found in all compartments and time points sampled. Variant A was predominant in plasma, and variant B was predominant in PBMC at baseline. (Right) Frequency distribution of RNA variants in plasma and proviral variants in PBMC before and after immunization. For comparative purposes the proviral distributions of LNMC and PBMC at day 30 are not illustrated but are as follows: day 30 PBMC, group A, 35%, and group B, 65%; day 0 LNMC, group A, 35% and group B, 65%; day 30 LNMC group A, 45% and group B, 55%.

FIG. 3

(Left) Phylogenetic analysis of sequences (cDNA and proviral) from patient 2. Samples from plasma represent RNA sequences, and those from PBMC and LNMC are proviral. Note two major groups, A and B, which are found in all compartments and time points sampled. Variant A was predominant in plasma, and variant B was predominant in PBMC at baseline. (Right) Frequency distribution of RNA variants in plasma and proviral variants in PBMC before and after immunization. For comparative purposes the proviral distributions of LNMC and PBMC at day 30 are not illustrated but are as follows: day 30 PBMC, group A, 35%, and group B, 65%; day 0 LNMC, group A, 35% and group B, 65%; day 30 LNMC group A, 45% and group B, 55%.

FIG. 4

(Left) Phylogenetic analysis of sequences (cDNA and proviral) from patient 3. Major groups are indicated as A through F. (Right) Frequency distribution of viral variants in plasma and PBMC preimmunization and in plasma postimmunization in patient 3.

FIG. 4

(Left) Phylogenetic analysis of sequences (cDNA and proviral) from patient 3. Major groups are indicated as A through F. (Right) Frequency distribution of viral variants in plasma and PBMC preimmunization and in plasma postimmunization in patient 3.

FIG. 5

(a) HTA of viral sequences (C2 through V5) obtained from patient 3. After referring to the phylogenetic tree, a representative sequence from group B or D was selected as a probe against other variants. Probes derived from group B or D sequences were able to easily distinguish sequences of the same group (B or D) from a panel of sequences obtained from different groups in the same patient or from an unrelated sample (lane U). The experiment was repeated three times, using different representative sequences for probe and panel, with similar results. ∗, probe lane; arrow, position of homoduplex migration. (b) The effect of tetanus immunization on viral quasispecies turnover in patient 3, as displayed by HTA. In order to determine the turnover of groups B and D, an HTA was performed with the probes described above. For plasma samples, RT PCR products amplified from approximately 100 copies of HIV-1 RNA templates were used as the driver. For PBMC samples, PCR products amplified from about 40 copies fo HIV-1 genomic DNA were used as the driver. Arrow, position of homoduplex migration.

FIG. 6

(Left) Phylogenetic analysis of sequences from patient 4. Two major groups, A and B, are identified. ∗, a recombinant of groups A and B. (Right) Frequency distribution of groups A and B in plasma and PBMC preimmunization and in plasma postimmunization.

FIG. 6

(Left) Phylogenetic analysis of sequences from patient 4. Two major groups, A and B, are identified. ∗, a recombinant of groups A and B. (Right) Frequency distribution of groups A and B in plasma and PBMC preimmunization and in plasma postimmunization.

FIG. 7

Mean diversity of plasma viral sequences over time in patients 1 through 4. Error bars represent standard error of the mean.

FIG. 8

Diversity along env over time in patient 4. Entropy (ordinate) is plotted against nucleotide position.

FIG. 9

Effect of tetanus immunization on synonymous (d_s) and nonsynonymous (d_n) substitution rates. Intrasample d_s and d_n values for patients 1 through 4. Significant d_n/d_s ratios away from neutrality, i.e., d_n/d_s either <1 or >1, are defined as those plots with error bars which have significantly deviated either below or above the x = y line, respectively.