Longitudinal dynamics of the HIV-specific B cell response during intermittent treatment of primary HIV infection

Abstract

Background: Neutralizing antibodies develop in natural HIV-1 infection. Their development often takes several years and may rely on chronic virus exposure. At the same time recent studies show that treatment early in infection may provide opportunities for immune preservation. However, it is unknown how intermittent treatment in early infection affects development of the humoral immune response over time. We investigate the effect of cART in early HIV infection on the properties of the memory B cell compartment following 6 months of cART or in the absence of treatment. The patients included participated in the Primo-SHM trial where patients with an early HIV-1 infection were randomized to no treatment or treatment for 24 or 60 weeks.

Methods: Primo-SHM trial patients selected for the present study were untreated (n = 23) or treated for 24 weeks (n = 24). Here we investigate memory B cell properties at viral set-point and at a late time point (respectively median 54 and 73 weeks) before (re)-initiation of treatment.

Results: At viral set-point, the memory B cell compartment in treated patients demonstrated significantly lower fractions of antigen-primed, activated, memory B cells (p = 0.006). In contrast to untreated patients, in treated patients the humoral HIV-specific response reached a set point over time. At a transcriptional level, sets of genes that showed enhanced expression in memory B cells at viral setpoint in untreated patients, conversely showed rapid increase of expression of the same genes in treated patients at the late time point.

Conclusion: These data suggest that, although the memory B cell compartment is phenotypically preserved until viral setpoint after treatment interruption, the development of the HIV-specific antibody response may benefit from exposure to HIV. The effect of viral exposure on B cell properties is also reflected by longitudinal changes in transcriptional profile in memory B cells over time in early treated patients.

Fig 1. Plasma viral load after randomization / treatment interruption in untreated and 24 weeks treated patients.

The patients depicted consist of a subgroup of the initial Primo-SHM trial [30]. The graph shows the median viral load over time. The time points that were selected for analyses in the present study are viral setpoint (defined as 36 weeks after randomization / treatment interruption) (*) and the latest available time point before (re)-start of treatment (median weeks for each group) (**) (see also Table 1).

Fig 2

Treated patients show A: a smaller fraction of antigen-primed, CD21^neg, B cells at viral setpoint as well as B: a smaller fraction of plasmablasts at viral setpoint. Horizontal bars represent median values and statistical significance was calculated using the nonparametric Mann-Whitney U test.

Fig 3

Differences in phenotype of the B cell compartment at viral setpoint (A) and the late time point (B) between untreated (empty symbols) and treated (filled symbols) patients. Scatterplots show the frequency of memory B cells divided by subset. Horizontal bars represent median values and statistical significance was calculated using the nonparametric Mann-Whitney U test. Open symbols represent untreated patients and closed symbols represent treated patients. On the y-axis percentages of cells are depicted gated on aqua^negCD19^posIgG^pos cells

Fig 4. The phenotype of HIV gp-140-specific B cells shows a shift towards a predominance of resting memory B cell phenotype in treated patients.

Scatterplots show the distribution of gp140 positive memory within IgG^posCD19^pos B cells (RM: resting memory; IM: intermediate memory; AM: activated memory; TL: tissue like memory) in untreated (left) and treated (right) patients at viral setpoint. Horizontal bars represent median values. On the y-axis percentages of gp140^pos cells expressed as percentage of IgG^posCD19^pos cells are depicted. P values show the significance level calculated by Friedman test. Further testing for differences between B cell subsets in each group was performed using Wilcoxon matched-pairs signed ranks test and shows the following significance levels. In summary for treated patients: RM compared to IM p<0.001; RM compared to AM p = 0.16; RM compared to TLM p = 0.18. In summary for untreated patients: RM compared to IM p = 0.08; RM compared to AM p = 0.4; RM compared to TLM p = 0.8.

Fig 5. Dynamic shifts in gene expression occurring within each group with time was quantified as a ratio between temporal gene expression differences observed in the treated group to that of the untreated group, for each of the 92 genes.

This temporal variation ratio was taken as the x-coordinate. Differences in gene expression observed between the two groups at each time point was quantified as a ratio between differences observed between the two groups at viral setpoint to that observed at the late time point, for each of the 92 genes. This inter-group variation ratio was taken as the y-coordinate. This graphs shows the distribution of the 92 genes based on the two calculated ratios. The polarization towards the double negative quadrant indicates that temporal variation occurring with time within each group occurs in opposite directions; and that the inter-group variation observed between the two groups at the viral setpoint and the late time point are in opposite directions. In addition, for 65 genes temporal variation ratio was <1 indicating that the changes in gene expression accrued with time in the untreated group was greater than changes accrued with time in the treated group. Also, for 59 genes inter-group variation ratio was <1 indicating that differences in gene expression observed between the two groups were greater at the late time point compared to those observed at the early time point.

Fig 6

Scatterplots showing the serum endpoint titers of antibodies directed against gp140 (A), CD4 binding site (CD4bs) (resurfaced stabilized gp120 core (RSC3))(B) and membrane proximal region (MPER) (C) in untreated (left panels) and treated (right panels) patients at viral set point (VS) and at the late time point (late). Statistical significance was calculated with nonparametric Wilcoxon matched-pairs signed-ranks test.