Persistent T cell unresponsiveness associated with chronic visceral leishmaniasis in HIV-coinfected patients

Abstract

A large proportion of HIV-coinfected visceral leishmaniasis (VL-HIV) patients exhibit chronic disease with frequent VL recurrence. However, knowledge on immunological determinants underlying the disease course is scarce. We longitudinally profiled the circulatory cellular immunity of an Ethiopian HIV cohort that included VL developers. We show that chronic VL-HIV patients exhibit high and persistent levels of TIGIT and PD-1 on CD8⁺/CD8^- T cells, in addition to a lower frequency of IFN-γ⁺ TIGIT^- CD8⁺/CD8^- T cells, suggestive of impaired T cell functionality. At single T cell transcriptome and clonal resolution, the patients show CD4⁺ T cell anergy, characterised by a lack of T cell activation and lymphoproliferative response. These findings suggest that PD-1 and TIGIT play a pivotal role in VL-HIV chronicity, and may be further explored for patient risk stratification. Our findings provide a strong rationale for adjunctive immunotherapy for the treatment of chronic VL-HIV patients to break the recurrent disease cycle.

Fig. 1. Study design and patient stratification.

a Flow chart and selection of study participants. HIV: human immunodefiency virus, VL-HIV: visceral leishmaniasis-HIV patients, LS+-HIV: Leishmania-seropositive HIV co-infected individuals. b VL-HIV patients follow-up monitoring timelines. The depicted months of follow-up were calculated backwards and onwards from the time of active VL development (D0) to standardise timepoints between individuals. On the left axis, the number of VL episodes prior to study inclusion are shown beside the red (Chronic VL-HIV) and blue (Non-chronic VL-HIV) patient icons. Stripe indices indicate visits while black dots indicate that a blood sample was taken at that visit. VL disease episodes are represented by red diamond symbols. Deaths are indicated by a black cross. Blue lines indicate the treatment periods. Individuals indicated with a plus sign (+) are non-chronic VL-HIV patients with one prior VL episode more than 10 years before study inclusion. Individuals indicated with an asterisk (*) were included in the single-cell immune profiling. D0 Day 0 (active disease development); W Week, EOT End of Treatment, M Month.

Fig. 2. The cellular composition of PBMCs isolated from participants as measured by flow cytometry.

a, d Cross-sectional analysis of the frequencies of CD8⁺ T cells (CD3⁺CD8⁺) and CD8^- T cells (CD3⁺CD8^-), respectively, at active disease development for the VL-HIV group (D0; n = 23) and at study recruitment for the HIV (n = 19), and Leishmania-seropositive HIV-coinfected (LS+-HIV; n = 20) groups. In the latter group, the plus sign (+) indicates those Leishmania-seropositive individuals with a history of VL. b, e Comparison of the frequencies of CD8⁺ T cells (CD3⁺CD8⁺) and CD8⁻ T cells (CD3⁺CD8⁻), respectively, for non-chronic (blue; n = 7) or chronic (red; n = 16) VL-HIV patients only, at active disease development (D0) and at the End-of-Treatment (EOT). c, f Longitudinal profiling of the cellular frequencies from time of VL development, for non-chronic (blue; n = 7) or chronic (red; n = 17) VL-HIV patients only.

Fig. 3. The cellular composition of PBMCs isolated from participants as measured by single-cell RNA sequencing.

a UMAP representation of 17.308 primary PBMC derived from two representative cases of healthy endemic controls, individuals with HIV, individuals with AL-HIV (Leishmania-seropositive individuals with no history of VL), and four VL-HIV patients at active disease development. b Cellular proportions of inferred immune cell types in the PBMC fractions of the different participant groups at D0, n = number of cells per patient group. c UMAP representation of 12.822 primary PBMC derived from two non-chronic (and long-term cured) and two chronic VL-HIV patients, all sampled at D0 and EOT. d Cellular composition of inferred immune cell types in VL-HIV participant PBMC fractions across the different timepoints of the disease course, n = number of cells per participant group and timepoint.

Fig. 4. The PD-1⁺ and TIGIT⁺ CD8⁺ (CD3⁺CD8⁺) T cell fractions of PBMCs isolated from the different participant groups as measured by flow cytometry.

a, d, g, j Cross-sectional profiling of the frequency of PD-1⁺ CD8⁺ T cells, the MFI of PD-1 on CD8⁺ T cells, the frequency of TIGIT⁺ CD8⁺ T cells, and the MFI of TIGIT on CD8⁺ T cells, respectively, between the VL-HIV (PD-1 n = 23, TIGIT n = 17), the Leishmania-seropositive HIV (PD-1 n = 20, TIGIT n = 12) and the HIV-only (PD-1 n = 19, TIGIT n = 14) groups, using a Benjamini-Hochberg corrected pairwise Mann-Whitney U test to test for statistical differences. The plus sign (*) indicates those Leishmania-seropositive individuals with a history of VL. b, e, h, k Comparison between chronic (PD-1 n = 16 at both timepoints, TIGIT n = 13 at D0 and n = 7 at EOT) and non-chronic VL-HIV patients (PD-1 n = 7 at both timepoints, TIGIT n = 4 at D0 and n = 5 at EOT) at the active disease development (D0) and End-of-Treatment (EOT) timepoints using a BH-corrected Mann-Whitney U to test for statistical differences. c, f, i, l Longitudinal characterisation of the frequencies and MFIs of the same cellular subsets, for non-chronic (PD-1 n = 7, TIGIT n = 6) or chronic VL-HIV (PD-1 n = 17, TIGIT n = 16) patients only, using linear mixed-effects models (see Methods).

Fig. 5. Interferon-γ positivity in the TIGIT⁺ and TIGIT^- CD8⁺ (CD3⁺CD8⁺) T cell fractions of PBMCs isolated from the different participant groups as measured by flow cytometry.

a, c, e, g Comparison of the frequencies of IFN-γ⁺TIGIT^- cells out of all CD8⁺ T cells, IFN-γ^-TIGIT⁺ cells out of all CD8⁺ T cells, IFN-γ⁺ cells out of TIGIT⁺CD8⁺ T cells, and IFN-γ⁺ cells out of TIGIT^-CD8⁺ T cells, respectively, between the chronic (n = 13 at D0, n = 7 at EOT) and non-chronic (n = 4 at D0, n = 5 at EOT) VL-HIV patients at the active disease development (D0) and End-of-Treatment (EOT) timepoints using a BH-corrected Mann-Whitney U to test for statistical differences. b, d, f, h Longitudinal characterisation of the frequencies of the same cellular subsets, for non-chronic or chronic VL-HIV patients, using linear mixed-effects models (see Methods).

Fig. 6. Pathway overrepresentation visualised as a heatmap performed on the differentially expressed genes between.

a CD4⁺ T cells of non-chronic VL-HIV at End-of-Treatment (EOT) versus active disease (D0), b CD4⁺ T cells of chronic VL-HIV at EOT versus D0, c CD8⁺ T cells of non-chronic VL-HIV patients at EOT versus D0, d CD8⁺ T cells of chronic VL-HIV patients at EOT versus D0. N = 2 each group and timepoint.

Fig. 7. The proportions of the top 30 T cell clonotypes per group, and their phenotypes.

The proportions of the top 30 T cell clonotypes per group, and their phenotypes of the (a) non-chronic VL-HIV and (b) chronic VL-HIV groups, at the start and end of VL treatment. c Phenotypes of the top 30 clonotypes of the non-chronic VL-HIV patients and d Phenotypes of the top 30 clonotypes of the chronic VL-HIV patients. n = 2 of each condition.