HLA‐DR Class II expression on myeloid and lymphoid cells in relation to HLA‐DRB1 as a genetic risk factor for visceral leishmaniasis

Summary

Genetic variation at HLA‐DRB1 is a risk factor for visceral leishmaniasis (VL) caused by Leishmania donovani. The single nucleotide polymorphism rs9271252 upstream of the DRB1 gene provides a perfect tag for protective versus risk HLA‐DRB1 four‐digit alleles. In addition to the traditional role of the membrane‐distal region of HLA class II molecules in antigen presentation and CD4 T‐cell activation, the membrane‐proximal region mediates ‘non‐traditional’ multi‐functional activation, differentiation, or death signals, including in DR‐expressing T cells. To understand how HLA‐DR contributes to disease pathogenesis, we examined expression at the protein level in circulating myeloid (CD14⁺, CD16⁺) and lymphoid (CD4⁺, CD8⁺, CD19⁺) cells of VL patients (pre‐ and post‐treatment) compared with endemic healthy controls (EHC). Although DR expression is reduced in circulating myeloid cells in active disease relative to EHC and post‐treatment groups, expression is enhanced on CD4⁺ DR ⁺ and CD8⁺ DR ⁺ T cells consistent with T‐cell activation. Cells of all myeloid and lymphoid populations from active cases were refractory to stimulation of DR expression with interferon‐γ (IFN‐γ). In contrast, all populations except CD19⁺ B cells from healthy blood bank controls showed enhanced DR expression following IFN‐γ stimulation. The rs9271252 genotype did not impact significantly on IFN‐γ‐activated DR expression in myeloid, B or CD8⁺ T cells, but CD4⁺ T cells from healthy individuals homozygous for the risk allele were particularly refractory to activated DR expression. Further analysis of DR expression on subsets of CD4⁺ T cells regulating VL disease could uncover additional ways in which pleiotropy at HLA DRB1 contributes to disease pathogenesis.

Abbreviations

anova

analysis of variance

CIITA

Class II transactivator

EHC

endemic healthy controls

gMFI

geometric mean fluorescence intensities

HC

healthy control

IFN‐γ

interferon‐γ

IL‐10

interleukin‐10

SNP

single nucleotide polymorphism

Th1

T helper 1

Th2

T helper 2

VL

visceral leishmaniasis

Introduction

Visceral leishmaniasis (VL) is caused by Leishmania donovani or Leishmania infantum chagasi, which are obligate intracellular parasites of myeloid cells. VL is characterized by fever, hepatosplenomegaly and hypergammaglobinemia, and is fatal in susceptible individuals if left untreated. However, only about 10% of people infected with these parasites proceed to clinical VL. Familial clustering1 and high sibling risk ratios2 suggested that host genetic factors were important in determining these differences in the outcome of infection with Leishmania species causing VL disease. In a genome‐wide association study,3 we demonstrated that the class II gene region HLA‐DRB1‐DQA1 carried the major genetic risk factors for VL disease caused by both L. donovani in India and L. infantum chagasi in Brazil. In further fine mapping studies, we confirmed that genetic risk maps to HLA‐DRB1, with HLA‐DRB1*1501 and DRB*1404/DRB1*1301 being the most significant protective versus risk alleles, respectively. Within these alleles, variants encoding specific residues at amino acid positions 11 and 13 in exon 2 were unique to protective alleles.4 The C allele at the biallelic single nucleotide polymorphism (SNP) rs9271252 lies upstream of the DRB1 gene and provides a perfect tag for protective HLA‐DRB1*01/*15/*16 allele groups, whereas the G allele tags intermediate HLA‐DRB1*03/*04/*07/*09/*10/*12 and risk DRB1 HLA‐DRB1*11/*13/*14 allele groups.3, 4 In addition to coding region variants in exon 2 influencing antigen processing and presentation, polymorphic variants upstream of HLA‐DRB1 could contribute to variable expression of DRB1 molecules.

A major role for HLA‐DR class II molecules is in presentation of antigen to CD4 T cells, the polarization of which into antigen‐specific T helper type 1 (Th1) and Th2/regulatory T cells is associated with the outcome of Leishmania infections in murine models5 and in human disease.6 Clinical VL caused by L. donovani, in particular, has been associated with high Th2/regulatory T cytokine responses including interleukin‐10 (IL‐10),6, 7, 8, 9, 10 whereas interferon‐γ (IFN‐γ) from Th1 cells is higher in L. infantum chagasi‐infected individuals that do not progress to clinical VL than in those who do.11 In addition to the role of the membrane‐distal region of HLA class II molecules responsible for peptide binding and T‐cell receptor engagement, the membrane‐proximal region comprising the extracellular connecting peptide, transmembrane domain, and cytoplasmic tail mediates ‘non‐traditional’ multi‐functional activation, differentiation, or death signals (reviewed in refs 12, 13) on class II‐expressing cells including T cells. HLA‐DR expression on T cells is regarded as a marker of T‐cell activation.14, 15 There is evidence too that antigen presentation by HLA‐DR‐expressing T cells16 can lead to CD4⁺ T‐cell unresponsiveness or tolerance.17 In the context of infectious diseases, HIV‐induced immune activation leads to expansion of both CD4⁺ and CD8⁺ T cells expressing HLA‐DR antigens.18, 19 Few studies of human VL have examined in detail the expression of HLA‐DR class II molecules at the protein level on cells of the myeloid (monocytes/macrophages) and lymphoid (B‐cell/T‐cell) lineages,20, 21 and none have studied HLA‐DR expression in relation to polymorphic variants.

In this study, we look at expression of HLA‐DR at the protein level in cells of both the myeloid and lymphoid lineages in whole blood of VL patients (pre‐ and post‐treatment) infected with L. donovani in India compared with endemic healthy controls (EHC), and in the spleen of individuals with VL. We also explore the ability of IFN‐γ to stimulate HLA‐DR expression on these different immune cell populations and determine whether this is influenced by polymorphism at the rs9271252 SNP upstream of the HLA‐DRB1 gene.

Material and methods

Study participants

The study was performed on VL patients attending the Kala‐Azar Medical Research Centre (KAMRC), Muzaffarpur, Bihar, India. EHC used for ex vivo blood flow cytometry profiling were recruited from individuals attending KAMRC to care for patients. Healthy controls (HC) used in lymphocyte stimulation assays were blood bank donors from the hospital at Banaras Hindu University (Varanasi, India). Demographic details (age, sex) were recorded for all participants, and clinical details (routine blood profiles; drug administered) for patients (Table 1). For a subgroup of pre‐ and post‐treated patients, splenic aspirates were taken as part of routine diagnostic procedure at the KAMRC. Splenic aspirate samples were collected into heparinized M199 medium containing 20% heat‐inactivated fetal bovine serum and used for flow cytometry. Whole blood was collected from all participants for flow cytometry analyses, and from HC and active VL cases for cytokine stimulation assays (cf. below).

Table 1.

Demographic and clinical details of participants used in the study

	Active VL single dose Ambisome	Active VL alternate day Amphotericin B	Endemic healthy controls
			EHC ex vivo studies	Blood donor HC in vitro stimulation studies
n	23	12	26	31
Age (years)	18·38 ± 16·37	24·60 ± 16·03	30·68 ± 7·75	32·58 ± 6·39
Sex % (M/F)	78/22	42/58	65/35	68/32
WBC (cells/μl)	4207 ± 1759	3774 ± 1235	ND	ND
Hemoglobin (g/dl)	8·32 ± 2·06	8·39 ± 1·08	ND	ND
Lymphocytes (% of WBC)	47·66 ± 15·35	42·88 ± 14·92	ND	ND
Eosinophils (% of WBC)	3·82 ± 0·73	3·87 ± 0·53	ND	ND
Neutrophils (% of WBC)	42·05 ± 14·89	46·82 ± 14·99	ND	ND
Monocytes (% of WBC)	1·16 ± 0·42	1·25 ± 0·67	ND	ND
SGOT (IU/ml)	62·75 ± 49·14	57·45 ± 29·12	ND	ND
SGPT(IU/ml)	39·14 ± 36·34	30·61 ± 22·02	ND	ND
Creatinine (mg/dl)	0·75 ± 0·26	0·73 ± 0·28	ND	ND

N/A, not applicable; N/D, not done; SGOT, serum glutamic oxaloacetic transaminase; SGPT, serum glutamic pyruvate transaminase; VL, visceral leishmaniasis; WBC, white blood cells.

Mean value ± SD of aggregated data are shown.

Ethical considerations

Ethical approvals for studies on Indian participants were obtained from the ethics committee of the Institute of Medical Sciences, Banaras Hindu University (Varanasi, India). The study was carried out in accordance with the Declaration of Helsinki Principles, and each participant, or the parent/guardian of individuals < 18 years old, signed informed consent forms to participate in the study and provide a blood sample.

Flow cytometry

Heparinized whole blood samples or splenic aspirate samples were used for flow cytometry analyses. Two panels of conjugated antibodies were used to analyze myeloid (BD Biosciences, West Bengal, India catalogue numbers: FITC‐CD16‐555406, PE‐CD14‐555398, PERCP‐HLA‐DR‐347364) versus lymphoid (BD Biosciences India catalogue numbers: FITC‐CD4‐555346, PE‐CD8‐555635, APC‐CD19‐340722, PERCP‐HLA‐DR‐347364) cell lineages, as indicated. The anti‐HLA‐DR, clone L243, antibody is derived from the hybridization of mouse NS‐1/1‐Ag4 myeloma cells with spleen cells from BALB/c mice immunized with the human lymphoblastoid B‐cell line RPMI 8866. It recognizes the HLA‐DR antigen comprising α‐ (encoded by DRA) and β‐ (encoded by DRB1) subunits that have molecular weights of 36 000 and 27 000, respectively. Cocktails of antibodies in staining buffer were added to samples and incubated for 15–20 min at room temperature in the dark, followed by the addition of whole blood lysis BD 10× lysis solution (BD Biosciences, India) and incubation for 30 min at room temperature. Cells were washed at 430 g for 8 min at 20°, supernatants were removed, and the cell pellet was resuspended in 200 μl of staining buffer. Blood samples were kept at 15–18° and splenic samples at 4–8° for transport from KAMRC to the laboratory at Banaras Hindu University in Varanasi, India. All samples were processed for flow cytometry within 24 hr. The BD FACSCalibur™ system with four‐color capability was used for cellular analysis and the BD CellQuest™ software was used for sample acquisition. Data analysis was performed using flowjo ^® version 10 (Ashland, OR).

DNA isolation

Whole blood (200 μl) for DNA extraction was collected into citrate tubes. Genomic DNA was extracted using QIAamp DNA mini kits (Qiagen, Hilden, Germany) in accordance with the manufacturer's instructions and DNA was eluted in 60 μl MilliQ water. The concentration and purity of extracted DNA was assessed by measuring the absorbance at 260 and 280 nm using an ND‐2000 spectrophotometer (Thermo Fisher Scientific India, Mumbai, India).

Stimulation assay

Heparinized whole blood samples from active VL patients and HC were used for IFN‐γ stimulation assays. Whole blood samples were stimulated with 100 U/ml recombinant human IFN‐γ (Cat. No. 285‐ IF; R&D systems, Minneapolis, MN) in 5% CO₂ at 37° for 24 hr. The cells were then retrieved and used for flow cytometry staining, as indicated.

SNP genotyping

Genotyping for the biallelic SNP rs9271252 was undertaken using TaqMan^® assays (Thermo Fisher Scientific, Hyderabad, India). The C allele for this SNP is a perfect tag for the protective DRB1 alleles, the alternative G allele tags intermediate and risk DRB1 alleles.3, 4

Statistical analysis

When comparing data for four study groups, analysis of variance (anova) with multiple comparisons and correction for multiple testing using the Tukey test was performed in graphpad prism 5 (version 5.00 for Windows; Graph Pad Software, San Diego, CA) to determine statistical significance for between‐group differences either in percentages of cells positive for specific markers or for geometric mean fluorescence intensities (gMFI) for DR expression on specific cell types, as indicated. To compare data for splenic aspirates in cases pre‐ and post‐treatment, a Mann–Whitney U‐test was used.

Results

Clinical and demographic data

The study was performed on 35 confirmed active VL cases and 57 healthy individuals. Active cases treated with either single‐dose liposome‐encapsulated Amphotericin B (Ambisome: n = 23) or alternate‐day Amphotericin B for 30 days (n = 12) were well‐matched for age and sex and clinical parameters (Table 1). Endemic healthy controls used in ex vivo profiling (n = 26) and blood bank donor HC used for lymphocyte stimulation assays in vitro (n = 31) were also well‐matched for age and sex ratio. Controls were older on average than cases because healthy donors younger than 18 years of age were not recruited.

Gating strategies for myeloid and lymphoid cells

To determine expression of HLA‐DR on myeloid cells, we used a panel of antibodies that allowed us to gate (Fig. 1a,b) on CD14⁺ monocytes or CD16⁺ neutrophils, in concert with an anti‐HLA‐DR antibody to study expression on each of these populations. For HLA‐DR expression on lymphoid cells, we used the anti‐HLA‐DR antibody in concert with antibodies that allowed us to gate on CD4⁺ T cells, CD8⁺ T cells or CD19⁺ B cells (Fig. 1c,d). These gating strategies were applied throughout, and back‐gating was used to verify that the gated population included the cells of specific interest.

Figure 1.

Examples of the gating strategies used to study DR expression on myeloid cells (a and b) and lymphoid cells (c and d) in representative endemic healthy controls (EHC) (a and c) and active visceral leishmaniasis (VL) cases (b and d). In (a) and (b) cell populations were first examined for forward and side scatter and the total leukocyte population gated to identify monocytes (CD14⁺) or neutrophils (C16⁺), which were further gated to obtain the frequency and geometric mean fluorescence intensities (gMFI) of CD14⁺ DR ⁺ cells or CD16⁺ DR ⁺ cells. The decrease in DR ⁺ cells is clear when comparing (a) EHC with (b) active VL cases. In (c) and (d) the populations of smaller nucleated leukocytes were selected to identify CD8⁺ DR ⁺, CD4⁺ DR ⁺, or CD19⁺ DR ⁺ cell populations. The increase in DR ⁺ cells is clear when comparing CD4⁺ DR ⁺ or CD8⁺ DR ⁺ cell populations in (c) EHC with (d) active VL cases. Back gating demonstrated that the appropriate populations of cells were present in the cell populations gated on the basis of forward and side scatter (not shown).

HLA‐DR expression in myeloid cell populations

We first determined the influence of VL disease on HLA‐DR expression in myeloid cells. A significant increase in the proportion of CD14⁺ monocytes was observed in whole blood in active cases (pre‐treatment cases throughout) compared with EHC (Fig. 2a), which returned to baseline following treatment (designated throughout as ‘discharge’ at ~30 days after commencement of treatment, and ‘follow up’ at 4–8 months after discharge). In contrast, the proportion of CD16⁺ neutrophils (Fig. 2b) showed a significant decrease in active cases compared with EHC consistent with the general neutropenia observed clinically in VL patients, returning to baseline in the post‐treatment follow‐up group. Although the frequency of CD14⁺ DR⁺ cells as a percentage of the gated CD14⁺ monocyte population did not differ significantly between EHC and active cases (data not shown), the gMFI DR expression on CD14⁺ DR⁺ cells was significantly reduced with active disease (Fig. 2c). For neutrophils, as observed previously,22 the frequency of CD16⁺ DR⁺ cells as a percentage of the gated CD16⁺ cell population increased significantly (data not shown), whereas gMFI was significantly reduced (Fig. 2d). A similar difference in gMFI HLA‐DR expression between active (low MFI) and post‐treatment discharge (higher MFI) cases was observed for CD14⁺ monocytes in splenic aspirates (Fig. 3a) but this did not achieve statistical significance in CD16⁺ neutrophils (Fig. 3b). Overall, the results for circulating and splenic myeloid cells suggest that HLA‐DR‐mediated antigen‐presenting cell function may be reduced in active VL cases.

Figure 2.

HLA‐DR expression in CD14⁺ monocytes and CD16⁺ neutrophils in whole blood taken from endemic healthy controls (EHC), active visceral leishmaniasis (VL) cases (pre‐treatment), treated (discharge) VL cases, and VL cases at follow up 4–6 months after completion of treatment, as indicated on the x‐axes. Panels (a) and (b) show the percentages of CD14⁺ and CD16⁺ cells, respectively, in the total leukocyte population. Geometric mean fluorescence intensities (gMFI) for DR expression are shown for the gated (c) CD14⁺ DR ⁺ and (d) CD16⁺ DR ⁺ cell populations. The bars represent the median and interquartile range. Statistical significance determined using non‐parametric analysis of variance (anova) with multiple comparisons is indicated with ****P < 0·0001, **P < 0·01 and *P < 0·05. Statistical between‐group differences occur despite apparent background bimodality for individuals with higher and lower percentages of cells that are CD14 (a; active cases; higher = 4–10%, lower = 1–2%) or CD16 (b; EHC; higher = 20–55%, lower = 0–15%) positive, or that have higher (200–300) and lower (30–175) gMFI for CD14 (c; EHC).

Figure 3.

Geometric mean fluorescence intensities (gMFI) for HLA‐DR expression in (a) CD14⁺ DR ⁺ monocyte, (b) CD16⁺ DR ⁺ neutrophil, (c) CD19⁺ DR ⁺ B‐cell, (d) CD4⁺ DR ⁺ T‐cell, and (e) CD8⁺ DR ⁺ T‐cell populations in splenic aspirates taken from active and discharged visceral leishmaniases cases. The bars represent the median and interquartile range. Statistical significance determined using a non‐parametric Mann–Whitney test is indicated with **P < 0·01 and *P < 0·05.

HLA‐DR expression in lymphoid cell populations

We next determined whether similar or different changes in expression of HLA‐DR on lymphoid cells were observed when comparing active cases with EHC and post‐treatment groups. No significant differences in frequencies of CD19⁺ B cells (Fig. 4a) or gMFI for DR expression in circulating CD19⁺ DR⁺ B cells (Fig. 4d) were observed in whole blood taken from EHC and the pre‐ and post‐treatment groups. As with myeloid cells, gMFI for DR expression in splenic CD19⁺ DR⁺ B cells (Fig. 3c) was lower in active VL cases compared with cases sampled at discharge. No significant differences were observed between groups in the percentages of CD8⁺ T cells in whole blood taken from EHC and the pre‐ and post‐treatment groups (Fig. 4c). However, there was a trend for an increase in the percentage of CD4⁺ T cells in active cases compared with EHC, which was significant at discharge and had returned to baseline at follow up (Fig. 4b). Of interest, both the frequencies of CD4⁺ DR⁺ T cells and CD8⁺ DR⁺ T cells as a percentage of their respective gated CD4⁺ and CD8⁺ T‐cell populations (data not shown), and the gMFI for DR expression on both CD4⁺ DR⁺ T cells (Fig. 4e) and CD8⁺ DR⁺ T cells (Fig. 4f) increased with active disease returning to baseline at discharge and follow up. Increased frequencies of CD4⁺ DR⁺ and CD8⁺ DR⁺ cells as a percentage of their respective T‐cell subsets, together with enhanced gMFI for expression of DR on these circulating T cells, is an early sign of T‐cell activation23 and is reminiscent of the expansion of both CD4⁺ and CD8⁺ T cells expressing HLA‐DR antigens associated with HIV‐induced immune activation.18, 19 No significant difference in gMFI for DR expression was observed in splenic CD4⁺ DR⁺ T cells between active cases and at discharge (Fig. 3d), but in CD8⁺ DR⁺ T cells there was an increase in gMFI on splenic cells at discharge compared with active cases (Fig. 3e). The results suggest differences in the phenotypes of circulating CD4⁺ DR⁺ T cells and CD8⁺ DR⁺ T cells compared with equivalent cells in splenic aspirates.

Figure 4.

HLA‐DR expression in CD19⁺ B cells, as well as CD4⁺ and CD8⁺ T cells, in whole blood from endemic healthy controls (EHC), active visceral leishmaniasis (VL) cases (pre‐treatment), treated (discharge) VL cases, and VL cases at follow up 4–6 months after completion of treatment, as indicated on the x‐axes. (a–c) The percentages of CD19⁺ B cells, CD4⁺ T cells and CD8⁺ T cells, respectively, in the small leukocyte population. Geometric mean fluorescence intensities (gMFI) for DR expression are shown for the gated (e) CD19⁺ DR ⁺ B‐cell, (f) CD4⁺ DR ⁺ T‐cell, and (g) CD8⁺ DR ⁺ T‐cell populations. The bars represent the median and interquartile range. Statistical significance determined using non‐parametric analysis of variance (anova) with multiple comparisons is indicated with ***P < 0·001 and *P < 0·05.

HLA‐DR expression ex vivo stratified by rs9271252 SNP genotype

Given the association between HLA‐DRB1 and VL disease,3 we interrogated our data further to determine whether variation in DR expression in active cases was related to rs9271252 SNP genotype. Figure 5 shows data for gMFI for DR expression in myeloid and lymphoid cell populations in whole blood from active cases stratified by rs9271252 genotype. As expected, there were no active VL cases homozygous for the protective C allele. There were no significant differences in gMFI for DR expression between GC heterozygotes and GG homozygotes across any myeloid or lymphoid cell populations.

Figure 5.

Geometric mean fluorescence intensities (gMFI) for HLA‐DR expression in (a) CD14⁺ DR ⁺ monocytes, (b) CD16⁺ DR ⁺ neutrophils, (c) CD19⁺ DR ⁺ B‐cell, (d) CD4⁺ DR ⁺ T‐cell, and (e) CD8⁺ DR ⁺ T‐cell populations in whole blood from active visceral leishmaniasis (VL) cases stratified by rs9271252 genotype. No individuals homozygous for the C protective allele were observed among active VL cases. No significant differences in gMFI were observed between heterozygous GC and homozygous GG individuals. Apparent bimodality for groups of individuals with high/low gMFI for CD14⁺ DR ⁺ cells (a) also occurs independently of genotype.

HLA‐DR expression in cells stimulated with IFN‐γ in vitro

Ability to observe any effect of rs9271252 genotype on gMFI expression on myeloid or lymphoid cells in active VL cases was compromised by absence of homozygotes for the protective C allele. Expression of HLA Class II molecules is also strongly driven by IFN‐γ,24 it is also possible that differences related to variants upstream of DRB1 would only be observed following stimulation with IFN‐γ. We therefore examined DR expression on the myeloid and lymphoid cells with/without IFN‐γ (Figs 6 and 7). DR⁺ myeloid cells (CD14⁺ DR⁺ monocytes and CD16⁺ DR⁺ neutrophils) and DR⁺ T cells (CD4⁺ DR⁺ T cells and CD8⁺ DR⁺ T cells) in whole blood from HC all showed an increase in gMFI for DR expression in response to IFN‐γ stimulation (Fig. 6a–e). In contrast, no significant differences were observed in parallel experiments comparing stimulated versus unstimulated cell populations in whole blood taken from active VL cases (Fig. 7). Hence, all circulating cell populations from VL cases appear to be at maximum ‘activation’ status (whether suppressed or enhanced for DR expression) and are unable to respond further to IFN‐γ.

Figure 6.

Geometric mean fluorescence intensities (gMFI) for HLA‐DR expression in (a) CD14⁺ DR ⁺ monocyte, (b) CD16⁺ DR ⁺ neutrophil, (c) CD19⁺ DR ⁺ B‐cell, (d) CD4⁺ DR ⁺ T‐cell, and (e) CD8⁺ DR ⁺ T‐cell populations in whole blood from blood bank healthy controls (HC) before (unstimulated) and after interferon‐γ (IFN‐γ) stimulation. All cell populations responded significantly (****P < 0·0001, ***P < 0·001, *P < 0·05) to IFN‐γ stimulation with higher gMFI.

Figure 7.

Geometric mean fluorescence intensities (gMFI) for HLA‐DR expression in (a) CD14⁺ DR ⁺ monocyte, (b) CD16⁺ DR ⁺ neutrophil, (c) CD19⁺ DR ⁺ B‐cell, (d) CD4⁺ DR ⁺ T‐cell, and (e) CD8⁺ DR ⁺ T‐cell populations in whole blood from active visceral leishmaniasis (VL) cases before (unstimulated) and after interferon‐γ (IFN‐γ) stimulation. No cell populations responded significantly to IFN‐γ stimulation with higher gMFI.

To determine the influence that variants at rs9271252, or in strong linkage disequilibrium with this SNP, might have on HLA‐DR expression we therefore focused on IFN‐γ‐stimulated myeloid (Fig. 8) and lymphoid (Fig. 9) cells in whole blood from HC. For CD16⁺ myeloid cells (Fig. 8d–f) and all lymphoid cells (Fig. 9), significantly higher gMFI for DR expression on DR⁺ cells in stimulated compared with unstimulated cells was only observed for the GC heterozygous genotype. For CD14⁺ DR⁺ myeloid cells significantly higher gMFI was also observed in stimulated cells from homozygous GG individuals when compared with unstimulated cells (Fig. 8b, c). One caveat in this analysis was the reduced sample size for the two homozygous groups of individuals compared with the heterozygous group. Nevertheless, when the level of expression (gMFI) was compared across IFN‐γ‐stimulated cells (Fig. 10), differences between genotypes were observed in CD4⁺ DR⁺ T cells (Fig. 10e; anova P = 0·0003; GG genotype significantly lower than GC genotype). CD4⁺ DR⁺ T cells (Figs 9 and 10e) from individuals homozygous (GG) for the risk allele appeared to be particularly refractory to IFN‐γ stimulation.

Figure 8.

Geometric mean fluorescence intensity (gMFI) for HLA‐DR expression stratified by rs9271252 genotype in (a) to (c) CD14⁺ DR ⁺ monocyte and (d) to (f) CD16⁺ DR ⁺ neutrophil cell populations in whole blood from blood donor healthy controls (HC) before (unstimulated) and after interferon‐γ (IFN‐γ) stimulation. Significant increases in gMFI with stimulation are indicated by results of Mann–Whitney non‐parametric t‐tests (****P < 0·0001, ***P < 0·001, **P < 0·01, *P < 0·05).

Figure 9.

Geometric mean fluorescence intensity (gMFI) for HLA‐DR expression stratified by rs9271252 genotype in (a–c) CD19⁺ DR ⁺ B‐cell, (d–f) CD4⁺ DR ⁺ T‐cell, and (g–i) CD8⁺ DR ⁺ T‐cell populations in whole blood from blood donor healthy controls (HC) before (unstimulated) and after interferon‐γ (IFN‐γ) stimulation. Significant increases in gMFI with stimulation are indicated by results of Mann–Whitney non‐parametric t‐tests (****P < 0·0001, ***P < 0·001, **P < 0·01, *P < 0·05).

Figure 10.

Geometric mean fluorescence intensity (gMFI) for HLA‐DR expression stratified by rs9271252 genotype in (a) CD14⁺ DR ⁺ monocyte, (b) CD16⁺ DR ⁺ neutrophil, (c) CD19⁺ DR ⁺ B‐cell, (d) CD4⁺ DR ⁺ T‐cell, and (e) CD8⁺ DR ⁺ T‐cell populations in whole blood from blood bank healthy controls (HC) after interferon‐γ (IFN‐γ) stimulation. The only significant difference between genotypes was observed when comparing gMFI for IFN‐γ stimulated CD4⁺ DR ⁺ T cells (***P < 0·001; analysis of variance with multiple comparisons; GG homozygous risk genotype lower compared with GC heterozygote group) across genotypes.

Discussion

In recent studies, we demonstrated that polymorphism at HLA DRB1 was the major genetic risk factor for clinical VL.3, 4 The top single nucleotide variants lay upstream of the HLA DRB1 gene, but also tagged protective versus risk classical HLA‐DRB1 four‐digit alleles. Within these alleles, variants encoding specific residues at amino acid positions 11 and 13 in exon 2 were unique to protective alleles.4 The latter provided strong evidence that polymorphisms in the functional peptide‐binding groove of DRB1 would determine differences in antigen‐presenting cell function and so influence the nature of the CD4 T‐cell response. Further support was provided by epitope binding studies, which showed greater peptide promiscuity in sequence motifs for 9‐mer core epitopes predicted to bind to risk compared with protective DRB1 alleles.4 In particular, there was a higher frequency of basic amino acids in risk‐specific epitopes, compared with hydrophobic and polar amino acids in protective‐specific epitopes, at anchor residues P4 and P6, which interact with residues at DRB1 positions 11 and 13. Relating this to T‐cell‐mediated immunity, we found that cured VL patients made variable but robust IFN‐γ, tumor necrosis factor and IL‐10 responses to 20‐mer peptides based on epitopes captured from DRB1 molecules, with peptides based on protective epitopes resulting in a higher proportion of patients with IFN‐γ : IL‐10 ratios more than twofold compared with peptides based on risk epitopes. These observations concur with studies25 demonstrating that it is the balance between pro‐inflammatory IFN‐γ and anti‐inflammatory cytokines like IL‐10 that is important in determining disease outcome in leishmaniasis and suggest that characteristics of epitopes binding to risk versus protective alleles might influence antigen‐presenting cell function.

Our observation that epitopes binding to risk versus protective DRB1 alleles have different 9‐mer core motifs does not preclude the additional possibility that variants lying in regulatory regions upstream of the DRB1 gene could contribute to genetic risk. The top single nucleotide variants (of which rs9271252 is one) from the genome‐wide association study3 are in perfect linkage disequilibrium (data not shown) with distal S′‐Y′‐like motifs that bind RFX5 and the Class II transactivator CIITA, and provide enhancer activity.26 Such regulatory variants might not only drive differing levels of DRB1 expression in individuals carrying risk versus protective alleles, but these regulatory elements could act differently in different immune cell types. We were therefore interested to determine the levels of expression of DR molecules on different immune populations during active VL. In relation to antigen‐presenting cell function, it was of interest that active VL cases had reduced levels of DR on circulating cells of the myeloid lineage. This is consistent with earlier studies demonstrating that macrophage infection with L. infantum or L. donovani is associated with reduced expression of MHC Class II molecules.27, 28, 29, 30 Hence, in addition to differences in epitopes binding to risk alleles, there will be an overall depression of antigen‐presenting cell function during active VL. How this relates to polymorphic variation at the top SNP rs9271252 could not be fully evaluated because (as expected) there were no VL cases homozygous for the protective allele. In addition, cells from active VL cases of myeloid or lymphoid lineages could not be activated to increase DR expression following IFN‐γ stimulation. For these reasons, we undertook additional studies of DR expression in the cells from HC, with/without IFN‐γ stimulation, where the full range of rs9271252 genotypes could be studied.

One interesting observation made in studying DR expression ex vivo in blood from VL cases was that, while DR expression on myeloid cells known to be capable of antigen‐presenting cell function was reduced in active VL cases, DR expression on CD4⁺ and CD8⁺ T cells was increased. Increased frequencies of CD4⁺ DR⁺ and CD8⁺ DR⁺ cells as a percentage of their respective T‐cell subsets, together with enhanced gMFI for expression of DR on these circulating T cells, is generally associated with T‐cell immune activation.13, 14, 20, 23 Persistent T‐cell immune activation is seen as deleterious in chronic HIV‐1 disease, but higher levels of T‐cell activation are also positive predictors of disease outcome and/or response to therapy.18, 19 Upstream regulatory variants might act differently in different cell populations; we were interested to determine the effect of rs9271252 genotype on DR expression in both myeloid and lymphoid cells from HC stimulated with IFN‐γ. Although genotype did not appear to impact significantly on DR expression in myeloid, B or CD8⁺ T cells, it was of interest that CD4⁺ T cells from individuals homozygous for the risk allele seemed to be particularly refractory to IFN‐γ stimulation. This might relate to the observation that expression of DR on T cells is controlled exclusively through the CTIIA promoter III.13 An area of interest in leishmaniasis has been the role of regulatory T cells in directing the immune response.31, 32, 33 In humans, HLA‐DR expression has been shown to define functionally distinct CD25⁺ CD4⁺ regulatory T cells that induce contact‐dependent suppression associated with high Foxp3 expression.34 In contrast, CD25⁺ CD4⁺ HLA‐DR negative regulatory T cells induce early IL‐4 and IL‐10 secretion.34 CD25⁺ CD4⁺ FoxP3⁺ cells are not elevated in active VL.33 Rather, it is CD4⁺ CD25⁻ FoxP3⁻ cells that are the major source of IL‐10 in the spleen.32 Given the refractoriness of CD4⁺ T cells to respond to immune activation and increase HLA‐DR expression in individuals homozygous for the upstream risk allele, further analysis of DR expression on subsets of T cells regulating VL disease could uncover additional ways in which pleiotropy at HLA DRB1 contributes to disease pathogenesis.

In summary, we have found differences in expression of HLA‐DR molecules with active VL disease, including decreased expression on cells of the myeloid lineage concurrent with increased expression in T cells. The former will impact on antigen‐presenting cell function, which we have previously demonstrated is impacted by variants in the coding region of HLA‐DRB1 molecules. The latter is a measure of T‐cell activation. Of interest in the present study are differences in DR expression on CD4⁺ T cells in relation to risk genotypes upstream of the HLA DRB1 gene, further analysis of which will contribute to our growing understanding of genetic risk and the pathogenesis of VL disease.

Disclosure

All authors declare that there is no conflict of interests.

Author contributions

BS, MF, JO and JMB conceived and designed the experiments; BS, JO, MS and SSS performed the experiments; BS, MF and JMB analyzed the data and wrote the paper; and MF, SS and JMB supervised the research. All authors read and approved the manuscript.

Funding

This work was supported by the NIH as part of Tropical Medicine Research Centre award 1P50AI074321‐01.