The majority of VL-HIV patients demonstrated disease chronicity under ART
Our study included a total of 63 participants, including 19 (30.2%) HIV patients, 20 (31.7%) asymptomatically Leishmania-infected HIV patients (AL-HIV), and 24 (38.1%) HIV patients that developed active VL during the course of the study (VL-HIV).
Table 1
Patient socio-demographic, clinical, and biochemical characteristics at study inclusion for HIV and AL-HIV patients, or at active disease development for VL-HIV patients (D0).
|
Total (n = 63)
|
HIV (n = 19)
|
AL-HIV (n = 20)
|
VL-HIV (n = 24)
|
P-value
|
Socio-demographic characteristics
|
Age in years, median (IQR)
|
39 (34-44.5)
|
39 (34–43)
|
39 (35–46)
|
39.5 (31–42)
|
0.827
|
Male, n (%)
|
52 (82.5)
|
8 (42.1)
|
20 (100)
|
24 (100)
|
< 0.001
|
BMI in kg/m2, median (IQR)
|
18 (16-19.7)
|
20 (17.6–21)
|
17.9 (16.8–18.8)
|
16.6 (15.4–18.6)
|
0.004
|
Literacy, n (%)
|
36 (57.1)
|
9 (47.4)
|
10 (50)
|
17 (70.8)
|
0.251
|
Occupation, n (%)
|
0.014
|
Daily labourer
|
30 (47.6)
|
5 (26.3)
|
10 (50)
|
15 (62.5)
|
|
Farmer
|
18 (28.6)
|
4 (21.1)
|
8 (40)
|
6 (25)
|
|
Other
|
15 (23.8)
|
10 (52.6)
|
2 (10)
|
3 (12.5)
|
|
Clinical characteristics and disease history
|
VL history, n (%)
|
32 (50.8)
|
0 (0)
|
13 (65)
|
19 (79.2)
|
NA
|
Past VL episodes, median (IQR)
|
1 (0–2)
|
0 (0–0)
|
1 (0–2)
|
2 (1–6)
|
NA
|
Months since previous VL episode, median (IQR)
|
9.5 (4–80)
|
NA
|
73 (9-135)
|
6 (4-15.5)
|
NA
|
On ART, n (%)
|
62 (98.4)
|
19 (100)
|
20 (100)
|
23 (95.8)
|
1
|
Concomitant diseases, n (%)
|
7 (11.1)
|
0 (0)
|
3 (15)
|
4 (16.7)
|
0.177
|
Laboratory markers
|
rK39 RDT positivity, n (%)
|
43 (68.3)
|
0 (0)
|
20 (100)
|
23 (95.8)
|
NA
|
rK39 ELISA positivity, n (%)
|
39 (61.9)
|
0 (0)
|
18 (90)
|
21 (87.5)
|
NA
|
KAtex positivity, n (%)
|
22 (34.9)
|
1 (5.3)
|
2 (10)
|
19 (79.2)
|
NA
|
DAT positivity, n (%)
|
40 (63.5)
|
0 (0)
|
18 (90)
|
22 (91.7)
|
NA
|
Leishmania PCR positivity, n (%)
|
26 (41.3)
|
0 (0)
|
2 (10)
|
24 (100)
|
NA
|
CD4 count, median (IQR); (cells/µl)
|
259 (99.5–421)
|
425 (262–750)
|
342 (221–420)
|
58 (37–126)
|
< 0.001
|
Lymphocytes, median (IQR); (x103/µl)
|
1.02 (0.67–1.53)
|
1.61 (1.06–1.99)
|
1.11 (0.87–1.31)
|
0.51 (0.3–0.79)
|
< 0.001
|
Platelets, median (IQR); (x103/µl)
|
165 (122.2-245.8)
|
245 (199–300)
|
204 (156–278)
|
114 (85–138)
|
< 0.001
|
Hemoglobin, median (IQR); (g/dL)
|
12.05 (9.73–13.78)
|
13.8 (13.1–14.3)
|
12.3 (11.8–14.3)
|
9.4 (8.6–10.1)
|
< 0.001
|
NA = Not Applicable.
Table 2
Patient socio-demographic, clinical, and biochemical characteristics of primary VL-HIV and chronic VL-HIV patients at various timepoints, including at active disease development (D0), at End-of-Treatment (EOT) and at six months post-treatment (Post M6; VL-HIV only).
|
Primary VL-HIV (n = 7)
|
Chronic VL-HIV (n = 17)
|
P-value
|
Socio-demographic characteristics
|
Age in years, median (IQR)
|
45 (41–51)
|
38 (31–41)
|
0.022
|
Male, n (%)
|
7 (100)
|
17 (100)
|
1
|
BMI in kg/m2, median (IQR)
|
15.8 (15.1–16.4)
|
17 (15.8–18.6)
|
0.192
|
Literacy, n (%)
|
4 (57.1)
|
13 (76.5)
|
0.374
|
Occupation, n (%)
|
0.449
|
Daily labourer
|
4
|
11
|
|
Farmer
|
3
|
3
|
Other
|
0
|
3
|
Clinical characteristics and disease history
|
VL history, n (%)
|
2 (28.6)
|
17 (100)
|
NA
|
Past VL episodes, median (IQR)
|
0 (0-0.5)
|
2 (2–8)
|
NA
|
Months since previous VL episode, median (IQR)
|
185 (164–206)
|
5 (4–10)
|
NA
|
Microscopically confirmed, n (%)
|
5 (71.4)
|
17 (100)
|
0.076
|
Parasite grading out of those microscopically confirmed, n (%)
|
0.311
|
+ 1 to + 3
|
4 (80)
|
7 (41.2)
|
|
+ 4 to + 6
|
1 (20)
|
10 (58.8)
|
VL treatment regime, n (%)
|
1
|
AmBisome + Miltefosine
|
7 (100)
|
16 (94.1)
|
|
AmBisome + Sodium Stibogluconate
|
0
|
1 (5.9)
|
On ART, n (%)
|
6 (85.7)
|
17 (100)
|
0.292
|
Concomitant diseases, n (%)
|
1 (14.3)
|
3 (17.6)
|
1
|
Laboratory markers at D0
|
rK39 RDT positivity, n (%)
|
7 (100)
|
16 (94.1)
|
1
|
rK39 ELISA positivity, n (%)
|
6 (85.7)
|
15 (88.2)
|
1
|
KAtex positivity, n (%)
|
2 (28.6)
|
17 (100)
|
< 0.001
|
DAT positivity, n (%)
|
5 (71.4)
|
17 (100)
|
0.076
|
Leishmania PCR positivity, n (%)
|
7 (100)
|
17 (100)
|
1
|
CD4 count, median (IQR); (cells/µl)
|
66 (52–157)
|
54 (36–113)
|
0.547
|
Lymphocytes, median (IQR); (x103/µl)
|
0.46 (0.26-1)
|
0.56 (0.33–0.74)
|
1
|
Platelets, median (IQR); (x103/µl)
|
124 (92–130)
|
112 (77–149)
|
0.973
|
Hemoglobin, median (IQR); (g/dL)
|
9.5 (8.8–9.9)
|
9.4 (7.7–10.6)
|
1
|
Laboratory markers at EOT
|
Lymphocytes, median (IQR); (x103/µl)
|
0.81 (0.44–1.24)
|
0.9 (0.4–1.27)
|
0.891
|
Platelets, median (IQR); (x103/µl)
|
216 (167–262)
|
145 (108–169)
|
0.077
|
Hemoglobin, median (IQR); (g/dL)
|
9.5 (8.8–10.6)
|
10.1 (8.5–11.2)
|
0.616
|
Laboratory markers at Post M6
|
CD4 count, median (IQR); (cells/µl)
|
219 (100–438)
|
82 (61–104)
|
0.091
|
Lymphocytes, median (IQR); (x103/µl)
|
0.89 (0.51–1.22)
|
0.5 (0.34–0.82)
|
0.462
|
Platelets, median (IQR); (x103/µl)
|
184 (94–188)
|
116 (96–129)
|
0.291
|
Hemoglobin, median (IQR); (g/dL)
|
11.1 (9.8–11.4)
|
10.3 (9.8–11.9)
|
0.892
|
NA = Not Applicable.
All Leishmania co-infected HIV patients were male, and almost all of them (88.6%) were working as farmers or daily labourers, occupations shown to have high risk of Leishmania transmission (Table 1) 43. As expected and in contrast, the majority of HIV patients were female (57.9%) and did not work in high-risk occupations (52.6). A similar proportion of participants with a history of VL was observed for AL-HIV and VL-HIV patients. However, the VL-HIV patients with VL history were only recently (< 1 year) cured from a previous VL episode prior to study inclusion, while the asymptomatic Leishmania-infected HIV patients were cured long-term (> 3 years). This suggests that the AL-HIV group consisted of both true asymptomatic cases as well as already long-term cured VL-HIV patients that were able to control the parasite. All participants except one VL-HIV patient were already following an ART regimen at study inclusion, with reported good adherence (only one patient reported a recently missed dose). VL-HIV patients at time of overt disease exhibited lower CD4+ T cell counts (P < 0.001), total lymphocytes (P < 0.001), platelet (P < 0.001), and hemoglobin (P < 0.001) levels than the other patient groups at study inclusion.
Out of the 24 VL-HIV patients, 7 (29.2%) patients were primary VL-HIV cases which had no prior nor recent (within < 10 years of study inclusion) VL history, and 15 (70.8%) patients were chronic cases with VL history (1–12 episodes before study admission) that developed one or multiple additional VL episode(s) during the study period (Table 2). The VL-HIV patients were followed-up for a median time of 19.5 months (IQR 12-26.5), visualised per patient in Fig. 1. Primary VL-HIV patients were significantly older than chronic VL-HIV patients (P = 0.022). While all chronic VL-HIV patients were KAtex-positive at active disease, all but two primary VL-HIV patients were negative for KAtex (P < 0.001). At study inclusion (not shown in table), 10 (58.8%) chronic VL-HIV tested positive for KAtex while none of the primary VL-HIV patients did (P = 0.019). Besides age and KAtex-positivity, no significant differences in patient characteristics could be identified between primary and chronic VL-HIV patients (Table 2). However, when the VL-HIV patients were alternatively stratified into two groups based on recurrent VL disease (Relapsed VL-HIV, n = 21, 71.4%) or episode-free survival (Cured VL-HIV, n = 6, 28.6%) during our study period, the six cured VL-HIV patients that did not experience VL relapse during our study period showed successful (P = 0.028) CD4+ T cell reconstitution (219 (107–438) cells/µl) six months post-treatment, in contrast to the fifteen patients that represented with recurrent episodes (71 (48–104) cells/µl). Similarly, the cured VL-HIV patients showed a successful (P = 0.032) reconstitution of platelets (184 (133–188) 103/µl) six months post-treatment in comparison to the relapsed VL-HIV patients (108 (84–129) 103/µl).
Finally, no significant differences or bias could be observed in patient characteristics between the 8 selected patients for single-cell sequencing (excluding the healthy endemic controls), indicating representative groups (Table S1).
The peripheral blood immune cell composition is not associated with VL-HIV chronicity at time of active disease
To study how the composition of blood immune cell subsets changes over time in VL-HIV co-infection, we first employed multi-colour flow cytometry to investigate the CD4+ and CD8+ T cell, and NK cell, populations known to be affected in HIV patients.
Consistent with the absolute CD4+ T cell counts shown in Table 1, a lower relative CD4+ T cell proportion in VL-HIV patients, as compared to HIV patients, was observed (Fig. 2A). No difference was observed in the CD4+ T cell proportion between primary and chronic VL-HIV patients at active disease (Fig. 2A). However, the proportion of CD4+ T cells was higher in pVL-HIV (P = 0.014) than in cVL-HIV patients across the follow-up duration (Fig. 2B). The proportion of CD4+ T cells was similar between the patient groups in function of the annualised relapse rate (Fig. 2C). The first and latter show that CD4+ T cell proportions between pVL-HIV and cVL-HIV patients are similar during active disease and across the frequency of recurrent events, suggesting a negligible value of CD4+ T cell counts and proportions in predicting disease chronicity. Inversely to the CD4+ T cell population, the proportion of CD8+ T cells was higher in VL-HIV patients as compared to HIV patients (Fig. 2D). Similarly, the proportion of CD8+ T cells was lower in pVL-HIV (P = 0.021) than in cVL-HIV patients across the follow-up duration (Fig. 2E). However, no difference could be observed between the patient groups in the proportion of CD8+ T cells in function of the annualised relapse rate (Fig. 2E). No differences could be observed in the NK cell proportions between the different patient groups at any timepoint or in function of the relapse rate (Fig. 2G-I).
Next, we applied single-cell RNA sequencing on a representative subgroup to enable a broader and unbiased transcriptional characterisation of possible changes in the peripheral immune cell composition. We sequenced a total of 17.308 cells across the different patient groups (n = 10, see methods), and a total of 12.822 cells for the four VL-HIV patients (both D0 and EOT timepoints, see Fig. 1 for selection). Cell type clustering and annotation resulted in 20 and 21 distinct clusters for the data of all patient groups and the data of VL-HIV patients only, respectively (Fig. 3A, C). Although the cellular composition was relatively similar across patient groups, CD4+ T effector-memory (TEM) cells were notably enriched in the Leishmania co-infected patient groups compared to the healthy endemic controls and HIV patients (Fig. 3B). More pronounced compositional changes were observed between D0 and EOT in the VL-HIV patient cohorts (Fig. 3D). Here, in consistence with the flow cytometry findings, the summed proportion of all CD4+ T cell subsets increased after treatment in primary VL-HIV patients, but decreased in chronic VL-HIV patients. As both primary VL-HIV patients reported long-term cure during the study duration, this finding was consistent with the previously observed CD4+ T cell reconstitution in long-term cured VL-HIV patients only. Similar to the findings at the flow cytometry level, the decrease in CD4+ T cell proportion in cVL-HIV patients after treatment was paired with an increase in the proportion of CD8+ T cells, in particular CD8+ TEM cells. Finally, a notable depletion of CD14+ monocytes was observed in cVL-HIV patients, as compared to the pVL-HIV patients, at both timepoints.
Higher proportions of PD1+ and TIGIT+ CD4+ T cells in chronic VL-HIV patients
Since CD4+ T cell exhaustion and senescence are clear hallmarks of a long-term HIV infection that may lead to decreased functionality against Leishmania, we assessed the proportions of CD4+ T cells (CD3+CD8−) that were positive for a range of cell exhaustion (PD-1, LAG3, TIM3, and TIGIT) and senescence (KLRG1 and CD57) markers across the different patient groups. For all exhaustion markers, we observed a significant increase in marker-positive CD4+ T cell proportions in VL-HIV patients as compared to the other patient groups at active disease (Fig. 4A,D,G,J). For the senescence markers, we observed increased proportions of KLRG1+ and CD57+ CD3+CD8− T cells in VL-HIV patients in comparison to HIV patients, but not AL-HIV patients, at D0 (Fig. S4A,B). These findings suggest VL-HIV patients, as compared to the other patient groups, exhibited increased CD4+ T cell exhaustion and senescence. At active disease, cVL-HIV patients demonstrated a higher proportion of TIGIT+ CD4+ T cells than pVL-HIV patients (Fig. 4J, P = 0.026). Furthermore, a more pronounced and persistently higher proportion of both PD-1+ and TIGIT+ CD3+CD8− T cells was observed for cVL-HIV patients across the study duration (Fig. 4B,K). In addition, out of all 7 pVL-HIV patients, those 2 pVL-HIV that relapsed during the study follow-up reported the highest proportions of PD-1+ CD3+CD8− T cells at all timepoints (Fig. 4B). Further confirming this, the pVL-HIV patients with a higher relapse rate exhibited higher proportions of PD-1 positive CD3+CD8− T cells, demonstrating similar levels as the cVL-HIV patients (Fig. 4C, P = 0.026). Together, these data indicate a marked CD4+ T cell exhaustion in chronic VL-HIV patients that may underlie the differentiation to VL disease chronicity.
A lower proportion of IFN-γ+ CD4+ T cells in relapsed VL-HIV patients
Next, since both VL and HIV alone have been described to induce persistent immune dysregulation in the CD4+ T cell population, and because we observed marked exhaustion in this population, we characterised the temporal dynamics of cytokines in the different CD4+ T cell subsets. No difference could be observed in the proportion of IFN-γ-producing CD4+ T cells across the different patient groups (Fig. S5A). Similarly, no difference could be observed in the proportions of IFN-γ-producing CD4+ T cells between pVL-HIV and cVL-HIV patients at any timepoint nor in function of the annualised relapse rate (Fig. S5A-C). However, when the VL-HIV patients were stratified into two groups based on VL episode-free survival during our study follow-up, the six VL-HIV patients that had reported episode-free survival showed a significantly higher proportion of IFN-γ-producing CD4+ T cells at active disease than the patients that developed a relapse episode (Fig. S6A, p = 0.002). This higher proportion of IFN-γ-producing CD4+ T cells in VL-HIV patients that reported cure during study follow-up persisted across the entire follow-up period (Fig. S6B, P = 0.014). Similar results were observed when using a more stringent definition of T helper 1 cells, including CXCR3 membrane expression (Fig. S6C-D, P = 0.009 and P = 0.015). Together, this indicates that the IFN-γ production ability of CD4 + T cells is rather linked with cure of the ongoing VL episode, and is independent of prior VL history and VL disease chronicity. With regard to other cytokines in the CD4+ T cell population, we identified a significantly higher proportion of IL17-producing CD4+ T cells (Th17) at active disease in the VL-HIV patients as compared to the other patient groups (Fig. S5D). The proportion of Th17 did not differ between pVL-HIV and cVL-HIV patients at any timepoint (Fig. 2D-E). Finally, the proportion of FOXP3+ CD4+ T cells (Tregs) did not differ between patient groups at active disease (Fig. S5G). The proportion of Tregs was lower in primary VL-HIV patients across the study follow-up duration (Fig. S5H, P = 0.045), but this did not differ at active disease (Fig. S5G), nor in function of the relapse rate (Fig. S5I).
Higher proportions of PD-1+ CD8+ T cells associated with chronicity in VL-HIV patients
Because T cell exhaustion and senescence are primarily thought to be CD8+ T cell pronounced processes in both HIV and VL, we investigated the level of immunosenescence- and exhaustion-related markers also on CD8+ T cells in our patient groups. Consistent with the CD4+ T cell population, the proportions of PD-1+, LAG3+, and TIM3+ CD8+ T cells were higher in the VL-HIV group than the other groups at D0 (Fig. 5A,D,G). The proportion of TIGIT+ CD8+ T cells were higher in the VL-HIV group than in the HIV group but not the AL-HIV group (Fig. 5J). As opposed to the CD4+ T cell subset, the chronic VL-HIV patients demonstrated higher proportions of PD-1+, but not TIGIT+, CD8+ T cells at active disease (Fig. 5A, P = 0.013). However, similar to the CD4+ T cell subset, the proportions of PD-1+ and TIGIT+ CD8+ T cells were persistently higher across the study follow-up duration in the cVL-HIV patients (Fig. 5B,K). Again, the two pVL-HIV patients that relapsed during the study follow-up reported the highest proportions of PD-1+ CD8+ T cells (Fig. 5B). Further confirming this, the pVL-HIV patients with a higher relapse rate exhibited similar proportions of PD-1+ CD8+ T cells as the cVL-HIV patients (Fig. 5C). For the senescence markers, no differences were observed between any patient group at any timepoint (Fig. S4).
Next, to account for CD8+ T cell IFN-γ production cooperation in a CD4+ T cell-deprived environment, we assessed the proportions of CD8+ T cells that were positive for a range of functionality markers (I IFN-γ, IL17, CD107a). However, no difference across the patient groups could be observed at any timepoint, nor in function of the annualised relapse rate, for any of the assessed functionality markers in the CD8+ T cell population (Fig. S7).
Pronounced and persistent CD4+ and CD8+ T cell treatment unresponsiveness in chronic VL-HIV patients at the single-cell transcriptional level
To better investigate the association of impaired CD4 + and CD8 + T cell functionality with VL chronicity in HIV patients in more detail, we studied their transcriptional activity at the single cell level.
We first compared the transcriptional profiles of CD4+ T cells between primary VL-HIV patients and chronic VL-HIV patients at the start and end of treatment (Fig. 6A,B). The level of expression of genes related to the antigen processing and presentation pathway, including genes of the Human Leukocyte Antigen (HLA) locus, was already higher in pVL-HIV patients than in cVL-HIV patients at the start of treatment (Fig. 6A), but was even more prominent after treatment (Fig. 6B). Accordingly, the genes of the T cell activation pathway had higher expression in pVL-HIV patients than in cVL-HIV patients after treatment. In addition, pVL-HIV patients showed increased expression of IFNG and genes related to the cellular response to IFNG after treatment, as compared to cVL-HIV patients. CCL5, which is associated with long-term viral suppression in HIV controllers, was markedly upregulated in pVL-HIV patients at both timepoints, suggesting that pVL-HIV patients may be able to control HIV better than cVL-HIV patients 44. Alternatively, we compared the transcriptional profiles of CD4+ T cells before and after treatment within the pVL-HIV and cVL-HIV patient groups (Fig. 6C-D). While parasitological treatment induced upregulation of T cell antigen presentation and activation pathway genes, and IFN-γ response pathway genes, in pVL-HIV patients (Fig. 6C), the CD4+ T cells of cVL-HIV patients instead experienced a downregulation of the genes of these pathways (Fig. 6D). This finding indicates a lack of host treatment responsiveness in cVL-HIV patients. Similarly, the CD8+ T cells of pVL-HIV patients featured increased expression of genes related to T cell antigen presentation and activation pathways as compared to cVL-HIV patients both at D0 and after treatment (Fig. S8).
Finally, due to the consistent upregulation of antigen presentation- and recognition-related pathways in pVL-HIV patients, but not cVL-HIV patients, we verified whether similar findings could be observed in the circulatory antigen-presenting cell (APC) subsets (CD14+ and CD16+ monocytes). Here, we showed that treatment in the pVL-HIV patients induces antigen processing- and presentation-related gene upregulation (Fig. S9). In contrast, only three differentially expressed genes between the D0 and EOT timepoints could be identified in the cVL-HIV patients. While this could indicate that treatment has few consistent effects on the expression of genes of these patients, this could also be partly explained by the observed depletion of CD16+ monocytes within this patient group, as shown in Fig. 3.
Lack of lymphoproliferative response at the clonotype level confirms pronounced CD4+ T cell unresponsiveness
Since we observed consistent T cell unresponsiveness in cVL-HIV patients at a transcriptional level, we set out to confirm this finding at the T cell clonotype level using single-cell TCR profiling. When comparing the treatment-induced expansion of the top 30 clonotypes between the pVL-HIV and cVL-HIV patients (Fig. 7A,B), we observed a marked clonotype expansion in pVL-HIV patients, but not in cVL-HIV patients who experienced subsequent VL episodes. When these clonotypes were highlighted on the projected UMAP, we observed that the pronounced T cell expansion in pVL-HIV patients occurred in both the CD4+ and CD8+ subsets, in contrast to a restricted CD8+ repertoire in cVL-HIV patients (Fig. 7C,D). The lack of lymphoproliferative responses after parasitological treatment further strengthens the observation of a persistent CD4+ T cell anergy in chronic VL-HIV patients experiencing VL relapses, as compared to primary patients without VL relapse. This finding implies an adjuvant host-directed immunotherapy regimen could be beneficial to stimulate a correct activation and subsequent expansion of T cells.