Dynamics of virological and immunological markers of HIV persistence after allogeneic haematopoietic stem-cell transplantation in the IciStem cohort: a prospective observational cohort study

Abstract

Background: Allogeneic haematopoietic stem-cell transplantation (allo-HSCT) markedly reduces HIV reservoirs, but the mechanisms by which this occurs are only partly understood. In this study, we aimed to describe the dynamics of virological and immunological markers of HIV persistence after allo-HSCT.

Methods: In this prospective observational cohort study, we analysed the viral reservoir and serological dynamics in IciStem cohort participants with HIV who had undergone allo-HSCT and were receiving antiretroviral therapy, ten of whom had received cells from donors with the CCR5Δ32 mutation. Participants from Belgium, Canada, Germany, Italy, the Netherlands, Spain, Switzerland, and the UK were included in the cohort both prospectively and retrospectively between June 1, 2014 and April 30, 2019. In the first 6 months after allo-HSCT, participants had monthly assessments, with annual assessments thereafter, with the protocol tailored to accommodate for the individual health status of each participant. HIV reservoirs were measured in blood and tissues and HIV-specific antibodies were measured in plasma. We used the Wilcoxon signed-rank test to compare data collected before and after allo-HSCT in participants for whom longitudinal data were available. When the paired test was not possible, we used the Mann-Whitney U test. We developed a mathematical model to study the factors influencing HIV reservoir reduction in people with HIV after allo-HSCT.

Findings: We included 30 people with HIV with haematological malignancies who received a transplant between Sept 1, 2009 and April 30, 2019 and were enrolled within the IciStem cohort and included in this analysis. HIV reservoirs in peripheral blood were reduced immediately after full donor chimerism was achieved, generally accompanied by undetectable HIV-DNA in bone marrow, ileum, lymph nodes, and cerebrospinal fluid, regardless of donor CCR5 genotype. HIV-specific antibody levels and functionality values declined more slowly than direct HIV reservoir values, decaying significantly only months after full donor chimerism. Mathematical modelling suggests that allogeneic immunity mediated by donor cells is the main viral reservoir depletion mechanism after massive reservoir reduction during conditioning chemotherapy before allo-HSCT (half-life of latently infected replication-competent cells decreased from 44 months to 1·5 months).

Interpretation: Our work provides, for the first time, data on the effects of allo-HSCT in the context of HIV infection. Additionally, we raise the question of which marker can serve as the last reporter of the residual viraemia, postulating that the absence of T-cell immune responses might be a more reliable marker than antibody decline after allo-HSCT.

Funding: amfAR (American Foundation for AIDS Research; ARCHE Program), National Institutes of Health, National Institute of Allergy and Infectious Diseases, and Dutch Aidsfonds.

Figure 1.. HIV reservoir in blood and tissues.

(A) Pairwise comparison of HIV-DNA for participants with samples from before allo-HSCT and after reaching FDC. (B) Proviral HIV-DNA in longitudinal samples; the grey vertical line represents the allo-HSCT date. (C) Decay ratios of HIV-DNA before allo-HSCT and after reaching FDC by donor cell type (CCR5Δ32/Δ32 and CCR5wt/ wt). (D) usVL pairwise comparison for participants with plasma samples before allo-HSCT and after reaching FDC. (E) usVL in longitudinal plasma samples; the grey vertical line represents the allo-HSCT date. (F) Decay ratios of HIV-RNA before allo-HSCT and after reaching FDC by donor cell type (CCR5Δ32/Δ32 and CCR5wt/wt). (G) Non-paired comparison for the replication-competent viruses in a qVOA measured in CD4⁺ T cells before allo-HSCT and after reaching FDC. (H) Non-paired comparison for proviral HIV-DNA measured in CD4⁺ T cells isolated from bone marrow. (I) Non-paired comparison for proviral HIV-DNA measured in CD45⁺ T cells isolated from ileum biopsies. (J) Proviral HIV-DNA measured in CD4⁺ T follicular helper cells isolated from lymph nodes by fine needle biopsy. (K) Non-paired comparison for ultrasensitive HIV-RNA measured in cerebrospinal fluid. Where more than one sample per participant was tested in tissues after allo-HSCT, the sample with the highest number of cells analysed is shown. For the tissue determinations and qVOA analysis, unpaired tests were used for the comparison because sufficient pairwise determinations of the same participants before and after allo-HSCT were unavailable. Non-filled symbols correspond to undetectable determinations for which the upper limit of detection was assigned on the basis of sample input. Allo-HSCT=allogeneic haematopoietic stem-cell transplantation. FDC=full donor chimerism. usVL= ultrasensitive viral load. qVOA=quantitative viral outgrowth assay. wt=wild-type.

Figure 2.. HIV-specific antibody levels and avidity.

The grey vertical line represents the allo-HSCT date. (A) Pairwise comparison of the standard VITROS EIA for participants with samples from before allo-HSCT and the last sample after reaching FDC. (B) Longitudinal standard VITROS EIA levels. (C) Decay ratios of standard VITROS EIA before allo-HSCT and the last sample after reaching FDC by donor status. (D) Pairwise comparison using the LAg avidity technique for participants with samples from before allo-HSCT and the last sample after reaching FDC. (E) Longitudinal levels of LAg avidity. (F) Decay ratios of LAg avidity before allo-HSCT and the last sample after reaching FDC by donor status. (G) Pairwise comparison for VITROS antibody avidity for participants with samples from before allo-HSCT and the last sample after reaching FDC. (H) Longitudinal levels of antibody avidity according to the VITROS avidity competition assay. (I) Decay ratios of VITROS antibody avidity before allo-HSCT and the last sample after reaching FDC by donor status. Each symbol corresponds to an IciStem participant; open symbols represent levels under the so-called primary infection threshold to interpret antibody reduction. EIA=enzyme immunoassay. S/CO=signal to cutoff ratio. Allo-HSCT=allogeneic haematopoietic stem-cell transplantation. FDC=full donor chimerism. LAg=limiting antigen.

Figure 3.. HIV-specific antibody functionalities.

Analyses performed in the 12 IciStem participants indicated in the legend, based on sample availability. A) Antibody-dependent neutrophil phagocytosis (ADNP) follow-up and paired comparison for participants with samples from before allogeneic stem cell transplantation (allo-HSCT) and the last sample after reaching full donor chimera (FDC). B) Monocyte antibody-dependent cellular phagocytosis (ADCP) and paired comparison for participants with samples from before allo-HSCT and the last sample after reaching FDC. C) Induced complement component C3b deposition on gp120-coated target cells (ADCD) follow-up and paired comparison for participants with samples from before allo-HSCT and the last sample after reaching FDC. D) Antibody-dependent cellular cytotoxicity measured by NK cell activation (ADNKa) using CD107a levels and paired comparison for participants with samples from before allo-HSCT and the last sample after reaching FDC. E) Antibody-dependent cellular cytotoxicity measured by NK cell activation (ADNKa) using IFN-γ levels and paired comparison for participants with samples from before allo-HSCT and the last sample after reaching FDC. F) Antibody-dependent cellular cytotoxicity measured by NK cell activation (ADNKa) using MIP-1β levels and paired comparison for participants with samples from before allo-HSCT and the last sample after reaching FDC. The dotted vertical line represents the allo-HSCT date.

Figure 4.. Mathematical modeling of HIV reservoir dynamics after allo-HSCT.

A) Schematics of the mathematical model, with variables as follows: latently infected cells with replication-competent HIV-DNA (L_P), productively infected cells (I_P), non-productively infected cells with defective HIV-DNA (L_D), total CD4⁺ T-cells (T₄), virus (V), and a surrogate for HIV-specific antibody concentration (A). In the model t = 0 is the transplantation date and the conditioning period is t ∈ [−10,0) days. Parameters: κ_C, cell depletion rate due to conditioning; α_T: CD4⁺ T-cell maximum proliferation rate; K, carrying capacity of cell proliferation; δ_L, death rate of latent cells; ξ reactivation rate of latent cells; α_A: antibody proliferation rate; δ_A, antibody loss rate; δ_P, death rate of productively infected cells; π, virus production rate; γ, virus clearance rate; κ_GVH, rate of cell depletion relative to chimerism (C_t). AICc: corrected Akaike information criterion. During fitting we found κ_GVH > 0 to be the most parsimonious model (ΔAICc=AICc_H0 - AICc_H1=51 and p<0.0001 using the loglikelihood ratio test). Β) T-cell chimerism observations (medians in bold/black). Best model predictions (C-H): C) blood CD4⁺ T-cell count; D) HIV-DNA per CD4⁺ T-cells; E) plasma ultra-sensitive HIVRNA-viral load; F) infected units per million cells per the qVOA assay; G) low sensitivity anti-HIV antibody (diluted ELISA observations); and H) model predictions of the HIV-DNA/mL. Black/bold solid line: model predictions using median parameter estimates (see table 1 in Appendix Page 17). Gray region: 10^th and 90^th quantiles from 1,000 simulations by the best model randomly selecting values from the population distributions of the estimated parameters (see table 1 in Appendix Page 17). Filled circles: observations over the limit of detection. Open circles: observations below the limit of detection. Red dashed line: model simulation assuming κ_GVH = 0. I) Contribution of different components to the total reservoir turnover rate and estimated reservoir half-life at different timepoints before and after transplantation as observed for total HIV-DNA and by qVOA. Blue: reservoir half-life equivalent to cell-loss preconditioning using historical values^,. Purple: reservoir half-life equivalent to depletion due to conditioning. Green: cell proliferation. Red: reservoir half-life equivalent to depletion proportional to levels of chimerism. Mpt: months post-transplantation.