Abstract
Cryptococcus neoformans is an opportunistic fungal infection and a leading cause of HIV-associated meningitis, contributing significantly to AIDS-related mortality, especially in Africa. Despite advancements in treatment, case fatality rates remain high, emphasizing the need for a better understanding of host immune responses. This study investigated the association between variation in the human leukocyte antigen (HLA) genes and the risk of cryptococcal disease in Ugandan individuals with HIV. We first described the allele prevalence in a prospectively enrolled Ugandan population using next-generation sequencing at a four-field resolution for classic HLA class I and II genes. We then conducted a case-control analysis involving 137 participants categorized as: 1) disseminated cryptococcosis but without meningitis, 2) cryptococcal meningitis, and 3) controls with similar degree of HIV disease but without cryptococcal infection. Across 66 unique alleles that met criteria for analysis, we found no significant associations between HLA variation and altered risk of disseminated cryptococcosis or cryptococcal meningitis after adjusting for multiple comparisons. However, before multiple comparisons correction, HLA-DPB1*04:01 demonstrated lower odds ratios for both disseminated disease and cryptococcal meningitis. Using public online databases, we analyzed predicted Cryptococcus peptide binding to HLA-DPB1*04:01 with predicted interferon-gamma responses and identified 121 viable peptides—leading us to hypothesize a role for the leading epitopes. This research underscores the need to further investigate the importance of the HLA system for host-pathogen interactions in HIV-associated cryptococcal disease.
Keywords: Cryptococcus, HIV, HLA, MHC, cryptococcal meningitis, cryptococcosis
1. Introduction
Cryptococcus neoformans is a fungal pathogen that causes meningitis in persons with impaired CD4+ T cell responses, particularly in those with advanced HIV disease. Cryptococcal meningitis accounts for 19% of AIDS-related deaths each year, with the largest global burden in Africa [1]. Despite advances in access to antiretroviral and antifungal therapies in low- and middle-income countries, case fatality rates remain unacceptably high, particularly outside of clinical trial settings [2]. Significant research gaps remain in understanding how the human immune response influences disease outcome.
Disease due to Cryptococcus is a complex interaction between pathogen virulence and host immune response. The precise factors that determine whether a host immune response is protective or detrimental remain poorly understood, but adaptive CD4+ T cell responses are a critical component of host protection. The T cell receptor on host T cells interacts with human leukocyte antigen (HLA) molecules, which are encoded by genes within the major histocompatibility complex (MHC) and are responsible for antigen presentation to initiate the signaling cascade of the adaptive immune response. However, T cells are said to be “MHC-restricted” in that they require a specific combination of HLA molecule and foreign peptide to activate via the T cell receptor (TCR) interaction. The HLA system may play an important yet understudied role in disease, as not all individuals with low CD4 counts reactivate latent pulmonary Cryptococcus and develop disseminated disease or progress to meningitis. We hypothesized that certain HLA alleles are associated with a protective response, likely through catalyzing a beneficial interferon-gamma driven Type 1 T-helper cell (Th1) polarized adaptive immune response [3]. A parabolic model of disease was proposed in which an insufficient immune response results in uncontrolled pathogen dissemination, while an excessive immune response results in host damage due to inappropriate inflammation [4, 5]. This framework may align with the idea that HLA alleles play a key role in determining whether sufficient immunity is generated to prevent the dissemination of Cryptococcus.
Although no published studies have examined the role of HLA alleles in HIV-associated cryptococcal disease to date, we are aware of two prior cryptococcal meningitis-related studies. First, a study of 53 immunocompetent HIV-negative Han Chinese individuals with cryptococcal meningitis due to C. neoformans found that HLA-DQB1*05:02 is associated with disease when compared to the uninfected control group (odds ratio = 2.51; 95%CI, 1.50–4.20) [6]. Second, a study of 25 individuals with Cryptococcus gattii infection in Papua New Guinea did not find any significant associations for disease when compared to uninfected controls after correction for multiple comparisons; however, they noted that HLA-B*56:01 is less common in meningitis patients (16%) compared to controls (41%) [7]. Outside of cryptococcosis, significant associations exist between various HLA alleles and infectious disease outcomes. HLA-DPB1*13:01 is associated with protection from severe disease in SARS-CoV-2 infected Italian patients [8]; while in another large study, HLA-B*15:01 is associated with asymptomatic SARS-CoV-2 infection. HLA-DRB1*03:01 and HLA-DRB1*13:01 alleles are statistically more prevalent in persons with chronic hepatitis C than in those who spontaneously clear the virus in an Egyptian family study [9]. In HIV populations, overall HLA class I heterozygosity is associated with slower progression to AIDS [10], while the individual HLA-B*27 and HLA-B*57 allele groups may also be protective against disease progression [11-13]. The role of HLA alleles in HIV-associated cryptococcal disease remains unknown.
Our study had three primary objectives. First, we aimed to describe the allele frequencies in a Ugandan patient population using modern next-generation sequencing techniques to obtain high-resolution allele genotyping. Second, we performed a series of case-control analyses to determine whether any alleles were associated with either protection against, or progression of, cryptococcal disease when compared with various defined reference groups. Third, we conducted modeling simulations of the identified major allele(s) of interest to generate hypotheses for immune mechanisms that might modify cryptococcal disease risk.
2. Methods
2.1. Study Population
We enrolled, obtained consent, and collected buccal swabs from adult Ugandans during two prospective observational studies conducted from 2018 to 2020. The COAST study enrolled persons with HIV who presented to the hospital with symptoms of suspected meningitis or central nervous system infection [14]. Participants were recruited from either Mulago National Referral Hospital or Kiruddu National Referral Hospital in Kampala, Uganda. Participants were systematically tested for cryptococcal meningitis using a stepwise testing approach [15]. The ORCAS study enrolled asymptomatic persons with HIV who underwent serum cryptococcal antigen (CrAg) screening based on CD4+ T cell count in outpatient clinics in Kampala [16, 17]. Cryptococcal antigenemia diagnosis was made by detection of cryptococcal antigen (CrAg LFA; IMMY, Norman, OK) in plasma. Cryptococcal meningitis was diagnosed by detection of CrAg in cerebrospinal fluid (CSF). Regulatory approval was obtained from Minnesota and Ugandan institutions (STUDY1304M31361; STUDY00003488). All screened participants provided written informed consent for the collection, storage, and analysis of genetic data.
2.2. DNA Purification
DNA samples were collected using the Zymo DNA/RNA Shield buccal swab collection system (Zymo Research; Irvine, CA, USA). All buccal swabs were collected at study entry and then frozen at ≤−80°C until the time of transport. The swabs were transported in batches to the University of Minnesota, where they were processed for DNA extraction. DNA extraction was performed using the Quick DNA MiniPrep Plus Kit (Zymo Research; Irvine, CA, USA), following the manufacturer’s instructions for biological fluids stored in DNA/RNA Shield. The DNA was eluted in 50 μL elution buffer and frozen at −20°C. The extracted DNA was then transported to the University of California San Francisco for sequencing.
2.3. DNA Sequencing
HLA sequencing was performed using a hybrid-capture protocol that includes HLA class I and II loci, as previously described [18]. A similar approach targeting HLA class I genes was previously reported in [19]. For each sample, 100 ng of high-quality DNA was fragmented using the Library Preparation Enzymatic Fragmentation Kit 2.0 (Twist Bioscience) according to the manufacturer's instructions. Briefly, DNA was enzymatically fragmented, end-repaired, and A-tailed, followed by ligation to Illumina-compatible dual-index adapters with unique barcodes. Following ligation, fragments were purified with AMPure XP magnetic beads at a 0.8x bead-to-sample ratio (Beckman Coulter). Dual size selection was performed (0.42× and 0.15× ratios), and libraries of approximately 600-800 bp were selected, at which point libraries were amplified and purified using magnetic beads. After fluorometric quantification, each sample was pooled (30 ng per sample) via ultrasonic acoustic energy. A Twist Target Enrichment Kit (Twist Bioscience) was then used to perform target capture on pooled samples, according to the manufacturer’s instructions. Sample volumes were then reduced using magnetic beads, and DNA libraries were bound to 1,394 biotinylated probes. Probes were designed specifically to target all exons, introns, and regulatory regions of the classical and non-classical HLA loci, including HLA-A, HLA-B, HLA-C, HLA-E, HLA-F, HLA-G, HLA-H, HLA-DPB1, HLA-DPA1, HLA-DRB1, HLA-DRA, HLA-DQB1, and HLA-DQA1. Then, streptavidin magnetic beads were used to capture fragments targeted by the probes. Captured fragments were then amplified and purified. A bioanalyzer (Agilent) was then used to analyze the enriched libraries. After evaluation, enriched libraries were sequenced using a paired-end 150-bp sequencing protocol on the NovaSeq platform (Illumina). After sequencing, HLA genotypes were predicted using HLA Explorer (Omixon).
2.4. Clinical Data and Analysis
Participants living with HIV were classified into three disease groups for the case-control analysis:
Cryptococcal Meningitis (CM): participants with positive blood CrAg and with positive CSF CrAg
Cryptococcal antigenemia (CrAg+): participants with positive blood CrAg without meningitis symptoms, or symptomatic with confirmed negative CSF CrAg
Control: participants with HIV admitted to the same hospital ward over the same study period, screened by our study for possible nervous system infection, and confirmed to have no evidence of cryptococcal disease by blood and CSF testing.
Baseline demographics and laboratory values collected at the time of study screening were summarized as medians, interquartile ranges, counts, or proportions, and were compared between the three groups by Chi-squared or Kruskal-Wallis testing.
Total allele frequencies are reported as the fraction of the total number of allele copies in the sample relative to the total number of chromosomes tested (i.e. twice the number of participants). Genotypes that could not be fully resolved at a locus were not included in the analysis for that locus. For the case-control analysis, we analyzed six classical loci (HLA-A, B, C, DRB1, DQB1/ DQA1, and DPB1). We excluded rare alleles from the case-control analysis (removing alleles from analysis when the allele frequency was less than 3%). We performed logistic regression to estimate the odds ratio for any given allele associated with a defined disease state compared with the defined reference group. We defined our primary analyses as 1) any disseminated cryptococcosis (CM and CrAg+ vs. control) and 2) progression to cryptococcal meningitis only (CM vs. CrAg+ and control). Secondary analyses of CM vs control alone, CrAg+ vs control alone, and CM vs CrAg+ alone can be found in the Supplementary Results. We controlled for multiple comparisons using the Benjamini-Hochberg correction for false discovery rate within each comparison. A p-value of <0.05 after correction was considered statistically significant.
2.5. Predictions for DPB1*04:01 binding and IFN-γ induction
Cryptococcus neoformans H99 protein sequences were retrieved from UniProt based on their cellular component (Gene Ontology), specifically: extracellular region, cell surface, cell wall, and membrane [20-22]. Additionally, proteins associated with differences in patient outcomes were included [23]. From the sequences, NetMHCII 4.3 (https://services.healthtech.dtu.dk/services/NetMHCIIpan-4.3/) was used to predict binding to HLA-DPB1*04:01 [24]. Since the software requires an alpha chain, the common alpha heterodimers of HLA-DPA1*01:03 and HLA-DPA1*02:01 were used. Prediction was performed using the eluted ligand (EL) method, selecting peptides with a Rank% < 5% and IC50 ≤ 200 nM. The eluted ligand method estimates the likelihood of naturally occurring antigen presentation, while IC50 reflects the binding strength between the peptide and HLA molecule. Peptides that met these criteria were further analyzed using the IFNepitope server (https://webs.iiitd.edu.in/raghava/ifnepitope/index.php) to predict IFN-γ responses. Peptides with at least three positive motifs and no negative motifs were selected [25].
3. Results
3.1. Study Participant Characteristics
Overall, 137 Ugandan participants were included in our analyses: 63 participants with CM, 45 participants that were CrAg+ in blood only, and 29 participants as controls. The median age for the entire cohort was 35 (IQR, 30–44) years, and 60% (83/137) were male. The median CD4 count for the cohort was 27 cells/μL (IQR, 13–89), which importantly did not differ by disease grouping (Table 1). CrAg+ persons were more likely to be receiving ART at baseline (73%; 33/45) compared to either CM participants (51%; 32/63) or controls (41%; 12/29). Additional baseline characteristics are included in Table 1.
Table 1. Baseline characteristics by cryptococcosis status grouping.
Data presented as median (IQR) or N (%). P-values calculated by Kruskal-Wallis rank sum test or Pearson’s Chi-square test.
| Characteristic | N | Cryptococcal meningitis N=63 |
N | Blood CrAg+ N=45 |
N | HIV+ Control N=29 |
P- Value |
|---|---|---|---|---|---|---|---|
| Age, years | 63 | 35 (31, 42) | 45 | 35 (30, 44) | 29 | 36 (30, 46) | 0.81 |
| Male Sex | 63 | 44 (70%) | 45 | 22 (49%) | 29 | 16 (55%) | 0.08 |
| CD4, cells/μL | 44 | 22 (11, 72) | 31 | 31 (16, 92) | 12 | 46 (26, 99) | 0.14 |
| On ART | 63 | 45 | 29 | 0.013 | |||
| Receiving | 32 (51%) | 33 (73%) | 12 (41%) | ||||
| Not Receiving | 31 (49%) | 12 (27%) | 17 (59%) | ||||
| CSF WBC, cells/μL | 60 | <5 (<5, 96) | 11 | --- | 25 | <5 (<5, 35) | 0.30 |
| CSF Opening Pressure, cmH2O | 60 | 20 (12, 27) | 8 | --- | 27 | 12 (9.5, 21.5) | 0.07 |
Abbreviations: CrAg+, cryptococcal antigen; ART, antiretroviral therapy; CSF, cerebrospinal fluid; WBC, white blood cells.
3.2. HLA Allele Frequencies
We found a total of 177 unique alleles amongst all participants (Figure 1). Total allele frequencies for the study are displayed in Supplemental Table 1a. For the three classical HLA class I loci, the most frequent alleles in this study population were HLA-A*02:01 at 13.8% (control only=12.1%), HLA-B*58:02 at 13.0% (control only=17.2%), and HLA-C*06:02 at 15.4% (control only=24.1%). Among the three classical HLA class II loci, the most frequent was HLA-DPB1*01:01 at 33.7% (control only=25.9%), followed by HLA-DQB1*05:01 at 23.1% (control only=25.9%) and HLA-DRB1*11:01 at 16.5% (control only=17.9%). Total allele frequencies broken into control, CrAg+, and CM subgroups can be found in (Supplemental Tables 1b-d).
Figure 1. HLA allele schema for case control analysis.
A total of 177 unique alleles were sequenced across our cohort. We then filtered the alleles to include only alleles with a total allele frequency ≥3%.
3.3. Case-Control Analysis
We performed a case-control analysis to determine if any alleles were associated with 1) disseminated cryptococcosis to the bloodstream or CNS (CM and CrAg+ vs. control) or 2) progression to cryptococcal meningitis (CM vs. CrAg+ and control). The odds ratios (OR) and p-values for both comparisons can be found in Supplemental Tables 2a and 2b. We found two alleles were associated with lower odds for the development of disseminated cryptococcosis before adjustment for multiple comparisons: HLA class I allele HLA-C*06:02 (OR=0.36; 95%CI, 0.15–0.84; p=0.018) and HLA class II allele HLA-DPB1*04:01 (OR=0.22; 95%CI, 0.08–0.66; p=0.006). None of the MHC class I or class II alleles were associated with lower odds of cryptococcal meningitis before adjustment for multiple comparisons, although HLA-DPB1*04:01 demonstrated a similar effect size (OR=0.31; 95%CI, 0.08–0.95; p=0.054). After adjusting for multiple comparisons using a false discovery correction, no alleles remained significantly associated with either disease state. Thus, these results should be viewed as hypothesis-generating, requiring future validation. Figure 2 shows a forest plot of alleles with the most significant association before false discovery rate correction is applied. HLA-DPB1*04:01 had similar odds ratios across each of the secondary analyses (CM vs control: OR = 0.17; CrAg+ vs control: OR = 0.30; CM vs CrAg+: OR = 0.54) but none remained significant after false discovery correction (Supplemental Tables 2c, 2d, and 2e).
Figure 2. Odds ratios for the development of disseminated cryptococcal disease by HLA allele (cryptococcal meningitis [CM] and cryptococcal antigenemia [CrAg+] vs control).
Forest plot for odds ratios of the 10 alleles with the most significant P-values before adjustment for multiple comparisons using a false discovery rate. No alleles were significantly associated with disseminated cryptococcal disease after correction for multiple comparisons.
3.4. Epitope modeling
To identify putative Cryptococcus epitopes that might be recognized by the HLA-DPB1*04:01 allele and could induce a protective CD4+ T cell response that might decrease dissemination, we performed a NetMHCII analysis, followed by identification of epitopes, and analysis for the likelihood that the epitope would induce an MHC II-mediated IFN-γ response. IFN-γ is associated with protective immune responses during cryptococcosis [3, 26]. Of the approximately 1900 protein sequences from Cryptococcus retrieved from UniProt, only 121 peptides from 95 proteins met the criteria for potential presentation by HLA-DPB1*04:01, when paired with the two known interacting alpha chains HLA-DPA1*01:03 and/or HLA-DPA1*02:01, and for inducing an IFN-γ response. These two alpha heterodimers were frequent in our cohort, with HLA-DPA1*01:03 at 36% and HLA-DPA1*02:01 at 19%. The 121 peptides comprised 94 membrane peptides from 89 proteins; 5 extracellular region peptides from 4 proteins; 2 cell surface peptides from 1 protein; 1 cell wall peptide from 1 protein; and 4 peptides from 4 proteins that had previously been related to differences in patient outcomes (Supplemental Tables 3a and 3b). Table 2 (HLA-DPA1*01:03) and Table 3 (HLA-DPA1*02:01) list the epitopes from proteins with the highest likelihood of eliciting an HLA-DPB1*04:01 mediated HLA-DPB1*04:01-mediated IFN-response, which are described in more detail in the discussion.
Table 2. Selected peptides modeled for presentation by HLA-DPB1*04:01 and inducing an IFN-γ response, paired with HLA-DPA1*01:03.
Cryptococcus neoformans H99 proteins sequences were retrieved from UniProt. From the sequences, NetMHCII 4.3 (https://services.healthtech.dtu.dk/services/NetMHCIIpan-4.3/) was used to predict binding to HLA-DPB1*04:01. Peptides with Rank% < 5% and IC50 ≤ 200 were selected to predict IFN-γ induction using the IFNepitope server (https://webs.iiitd.edu.in/raghava/ifnepitope/index.php). Peptides that had at least three positive motifs and no negative motifs were selected.
| MHC pair |
Peptide | Core | UniProt entry |
Protein names | Gene Names |
Eluted Ligand (%Rank) |
IC50 Affinity (nM) |
|---|---|---|---|---|---|---|---|
| DPA1*01:03 | GFSYVGETFLDCDLD | YVGETFLDC | J9VSP4 | Uncharacterized protein | CNAG_02176 | 0.03 | 10.67 |
| DPA1*01:03 | KLSFYTELYTETNAK | FYTELYTET | Q9P8P2 | Phospholipase B | PLB1, CNAG_06085 | 0.08 | 9.3 |
| DPA1*01:03 | GRRIKDTLYSEWADQ | IKDTLYSEW | J9VQL2 | Vacuolar transporter chaperone 4 | CNAG_01263 | 0.36 | 102.78 |
| DPA1*01:03 | EGKDVFYEEWMSFIS | VFYEEWMSF | J9VVE4 | Glycoprotein | CNAG_06000 | 0.76 | 13.85 |
| DPA1*01:03 | ASWDYTIYKFLTIGT | YTIYKFLTI | J9VYJ4 | C2H2-type domain-containing protein | CNAG_06256 | 1.14 | 26.57 |
| DPA1*01:03 | FKVLVEKDWLAFGHK | LVEKDWLAF | J9VEQ0 | Myotubularin | CNAG_00390 | 2.33 | 181.94 |
| DPA1*01:03 | QPVFIFDTELSLEWA | FIFDTELSL | J9VP68 | Uncharacterized protein | CNAG_07703 | 2.34 | 123 |
| DPA1*01:03 | FEKFLEYFFAEDDAN | FLEYFFAED | J9VRQ0 | Endoglucanase C | CNAG_05766 | 2.74 | 13.01 |
| DPA1*01:03 | QPLFIFDTELSFEWA | FIFDTELSF | J9VS66 | Uncharacterized protein | CNAG_05937 | 3.03 | 128.03 |
| DPA1*01:03 | NQGFLWRNMLRARTT | FLWRNMLRA | J9W135 | Sterol regulatory element-binding protein cleavage-activating protein | CNAG_01580 | 3.95 | 197.56 |
Table 3. Selected peptides modeled for presentation by HLA-DPB1*04:01 and inducing an IFN-γ response, paired with HLA-DPA1*02:01.
Cryptococcus neoformans H99 proteins sequences were retrieved from UniProt. From the sequences, NetMHCII 4.3 (https://services.healthtech.dtu.dk/services/NetMHCIIpan-4.3/) was used to predict binding to HLA-DPB1*04:01. Peptides with Rank% < 5% and IC50 ≤ 200 were selected to predict IFN-γ induction using the IFNepitope server (https://webs.iiitd.edu.in/raghava/ifnepitope/index.php). Peptides that had at least three positive motifs and no negative motifs were selected.
| MHC pair |
Peptide | Core | UniProt entry |
Protein names | Gene Names |
Eluted Ligand (%Rank) |
IC50 Affinity (nM) |
|---|---|---|---|---|---|---|---|
| DPA1*02:01 | GFSYVGETFLDCDLD | YVGETFLDC | J9VSP4 | Uncharacterized protein | CNAG_02176 | 0.03 | 21.63 |
| DPA1*02:01 | KLSFYTELYTETNAK | FYTELYTET | Q9P8P2 | Phospholipase B | PLB1, CNAG_06085 | 0.06 | 17.9 |
| DPA1*02:01 | QPVFIFDTELSLEWA | FIFDTELSL | J9VP68 | Uncharacterized protein | CNAG_07703 | 1.13 | 166.53 |
| DPA1*02:01 | ASWDYTIYKFLTIGT | YTIYKFLTI | J9VYJ4 | C2H2-type domain-containing protein | CNAG_06256 | 1.16 | 60.24 |
| DPA1*02:01 | QPLFIFDTELSFEWA | FIFDTELSF | J9VS66 | Uncharacterized protein | CNAG_05937 | 1.79 | 145.1 |
| DPA1*02:01 | FFEKFLEYFFAEDDA | KFLEYFFAE | J9VRQ0 | Endoglucanase C | CNAG_05766 | 3.54 | 23.65 |
| DPA1*02:01 | FEKFLEYFFAEDDAN | FLEYFFAED | 3.56 | 27.54 |
4. Discussion
Cryptococcal disease continues to cause significant morbidity and mortality in high-risk persons. While the most significant risk factor for cryptococcosis is a severely depleted CD4+ T cell level (secondary to HIV/AIDS in our clinical context), CD4+ T cell count does not uniformly predict risk. Some individuals with very low CD4 counts never develop cryptococcosis despite essentially ubiquitous exposure. For this study, our hypothesis was that certain HLA alleles might modify disease risk by either protecting against or facilitating the progression of cryptococcal disease in persons with advanced HIV disease. We utilized our access to three simultaneously enrolled cohorts to conduct a case-control analysis for HIV-associated cryptococcal disease, attempting to control for low CD4 counts.
We demonstrated that HLA-A*02:01 (full cohort=13.8%; control only=12.1%), HLA-B*58:02 (full cohort=13.0%; control only=17.2%), and HLA-C*06:02 (full cohort=15.4%; control only=24.1%) were the prevalent alleles in our Ugandan cohort for each major class I locus, while HLA-DPB1*01:01 (full cohort=33.7%; control only=25.9%), HLA-DQB1*05:01 (full cohort=23.1%; control only=25.9%), and HLA-DRB1*11:01 (full cohort=16.5%; control only=17.9%) were the prevalent alleles for each HLA class II locus across the study cohort. These data are consistent with a 2008 population study of 175 Ugandan individuals, where the HLA-A*02:01 (11.4%), HLA-B*45:01 (10.9%), HLA-B*58:02 (10.3%), and HLA-C*06:02 (19.1%) were the most common class I alleles [27]. For class II alleles, a 2001 population study of 43 Ugandan individuals demonstrated that HLA-DPB1*01:01 (27.7%) and HLA-DQB1*05:01 (25.5%) were the most common HLA-DPB1 variants [28], while a 2016 malaria vaccine study of 122 Ugandan individuals demonstrated that HLA-DRB1*11:01 (43.2%) was the most common DRB variant [29]. Thus, the HLA alleles in our advanced HIV population do not appear drastically different from the general Ugandan population, based on limited datasets.
We compared our dataset with HLA alleles previously identified in other cryptococcal meningitis populations. HLA-DQB1*05:02 was associated with cryptococcal meningitis in an HIV-negative Han Chinese cohort, but only a single copy of this allele was present in our cohort, preventing case-control analysis [6]. Similarly, we found zero HLA-B*56:01 alleles, which were identified as an allele of interest in a Papua New Guinea Cryptococcus gattii cohort [7].
In performing the case-control analysis, we did not find any alleles significantly associated with either our defined cases of disseminated cryptococcal disease (CM and CrAg+ vs. control) or progression to cryptococcal meningitis (CM vs. CrAg+ and control) after accounting for multiple comparisons. However, we would like to highlight HLA-DPB1*04:01 as an allele of interest for future study, given that it was most strongly associated with both disease conditions prior to false discovery rate correction. We found that HLA-DPB1*04:01 was associated with protection against disseminated cryptococcosis (OR=0.22; 95%CI, 0.08–0.66; p=0.006) and trended toward protection against progression to cryptococcal meningitis (OR=0.31; 95%CI, 0.08–0.95; p=0.054). HLA-DPB1*04:01 has previously been associated with susceptibility to anti-neutrophil cytoplasmic antibody (ANCA) vasculitis and relapse [30, 31], protection against hepatitis B infection [32], and protection against multiple sclerosis [33]. We did not identify any previously published literature on this allele’s association with disease from fungal pathogens.
The HLA-DPB1*04:01 binding prediction models identified several peptides that may be presented by this allele and induce an IFN-γ response. A Th1 immune response is associated with improved disease outcomes in several studies of cryptococcal disease, which may explain why patients carrying this allele experience less fungal dissemination [34, 35]. Identifying proteins capable of binding to this allele and inducing an IFN-γ immune response is important to guide future studies and development of effective immunotherapies and vaccines [24]. We have highlighted and provided literature on several such peptides below.
Myotubularin is a gene related to membrane stability, as deletion strains are more susceptible to cell membrane disruption, potentially making this protein a good target for immunotherapy studies [36]. On the other hand, vacuolar transporter chaperone 4, which plays a vital role in polyphosphate storage, has been linked to increased fungal loads in mouse lungs when knocked out, suggesting that it may be recognized by the immune system [37]. Another epitope that might be recognized by the immune system is encoded by the gene CNAG_02176, as its depletion decreased survival in mice [23]. Curiously, in the same study, various single-nucleotide polymorphisms (SNPs) in CNAG_02176 were associated with increased survival among infected patients. Moreover, two genes (CNAG_05937, CNAG_07703) with SNPs associated with differential human survival rates did not influence the survival of mice infected with their knockouts [23]. However, as they may induce a Th1 response, based on our modeling analysis, and are associated with attenuated human disease, they may also warrant further research. Interestingly, HLA-DPB1*04:01 showed a strong binding relationship with phospholipase B, one of the well-studied virulence factors of C. neoformans. This enzyme is reported to influence capsular enlargement, intracellular survival, adhesion to epithelial cells, and immunomodulation [38, 39]. Thus, further studies on the role of phospholipase B in persons living with HIV and its relationship with HLA-DPB1*04:01 and IFN-γ induction are of paramount importance, since new immunotherapy strategies could emerge. Another protein for further study is endoglucanase C, an understudied cell-surface protein. Although included in the Madhani deletion collection, it has not shown significant differences in signature-tagged mutagenesis experiments in infected mice or in vitro assays [40].
These findings highlight the need for further exploration of the HLA-DPB1*04:01 allele and its implications in HIV-associated cryptococcal disease, where future studies in vitro or in vivo could confirm these predictions and strengthen the results. Altogether, the binding and IFN-γ response predictions provide new mechanistic insights into possible host responses to cryptococcal infection, and this general line of inquiry may have implications for vaccine, immunotherapy, or drug-target prioritization. In our case, the response predictions characterized here are restricted to HLA-DPB1*04:01 and therefore may not be uniformly represented in populations where this allele is less common. The relevance of these predictions will depend on the frequency of HLA-DPB1*04:01 in target populations and on whether the same or overlapping epitopes are presented by other common HLA class II alleles. Future work assessing MHC-restricted epitope presentation and T-cell recognition across diverse HLA class II backgrounds will be important to define broader population coverage.
A major strength of our study was the use of modern next-generation sequencing techniques to classify alleles at two-field resolution in the significantly understudied sub-Saharan African population, whereas the established literature predominantly used older techniques. However, several limitations exist. Notably, our sample size was small, limiting our ability to draw definitive conclusions. In particular, the low number of participants in the control arm impaired our case-control analysis. In addition, our allele diversity was high, perhaps owing to population diversity, given that our study was conducted at a tertiary referral hospital. Performing a similar analysis with a more geographically restricted cohort of patients with lower allele diversity may yield greater precision in HLA associations with cryptococcal disease.
In conclusion, our data on the overall allele frequencies in this understudied Ugandan population continues to expand our understanding of global HLA allele frequencies. Sample size limitations reduced statistical power to detect associations across our highly variable loci in determining allele associations with cryptococcal disease. However, we identified the class I allele HLA-C*06:02 and the class II allele HLA-DPB1*04:01 as having low odds ratios and 95% confidence intervals (i.e., protective effect) before correction for multiple comparisons, suggesting that these alleles should be studied further for their potential protective effect against cryptococcal meningitis and antigenemia. In particular, the interaction of HLA-DPB1*04:01 with CD4+ T cell receptors and its potential to induce an IFN-γ-mediated Type 1 response provides an intriguing immune mechanism by which the observed protective trend against Cryptococcus neoformans might be explained. Ultimately, cryptococcal meningitis remains a significant source of morbidity and mortality in Ugandans with advanced HIV disease, and further research is needed to identify risk factors associated with the host immune response that could lead to better stratification of disease severity and targets for adjunctive therapies.
Supplementary Material
Acknowledgements
We would like to acknowledge the Ugandan patients and their families who participated in the various clinical studies. We are thankful to the entire Infectious Diseases Institute Meningitis Research group, without whom this study would not have been possible. Nielsen lab personnel who contributed to the successful completion of the study include: Marina Yoder, Katrina Jackson, Madeline Ahern, Jovany Betancourt, and Elizabeth Okafor.
Funding
We are thankful for the support from the National Institutes of Health: National Institute for Neurological Disorders and Stroke grants K23NS137968, R01NS086312, R01NS110519, R01NS118538; National Institute of Allergy and Infectious Diseases grants R21AI150303, U01AI125003, R01AI176922; Fogarty International Center grant D43TW009345; and National Center for Advancing Translational Sciences (KL2TR002492 and UL1TR002494).
Declaration of interests
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
Caleb Skipper reports financial support was provided by National Institutes of Health. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests.
References
- 1.Rajasingham R, Govender NP, Jordan A, et al. The global burden of HIV-associated cryptococcal infection in adults in 2020: a modelling analysis. The Lancet Infectious diseases 2022; 22:1748–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Loyse A, Burry J, Cohn J, et al. Leave no one behind: response to new evidence and guidelines for the management of cryptococcal meningitis in low-income and middle-income countries. The Lancet infectious diseases 2019; 19:e143–e7. [DOI] [PubMed] [Google Scholar]
- 3.Jarvis JN, Casazza JP, Stone HH, et al. The phenotype of the Cryptococcus-specific CD4+ memory T-cell response is associated with disease severity and outcome in HIV-associated cryptococcal meningitis. The Journal of infectious diseases 2013; 207:1817–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Pirofski LA, Casadevall A. The damage-response framework of microbial pathogenesis and infectious diseases. Advances in experimental medicine and biology 2008; 635:135–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Skipper C, Abassi M, Boulware DR. Diagnosis and Management of Central Nervous System Cryptococcal Infections in HIV-Infected Adults. J Fungi (Basel) 2019; 5(3):65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Liu J, Wei H, Liu J, et al. Analysis of the association of HLA subtypes with cryptococcal meningitis in HIV-negative immunocompetent patients. Future microbiology 2022; 17:1231–40. [DOI] [PubMed] [Google Scholar]
- 7.Van Dam MG, Seaton RA, Hamilton AJ. Analysis of HLA association in susceptibility to infection with Cryptococcus neoformans var. gattii in a Papua New Guinean population. Medical mycology 1998; 36:185–8. [PubMed] [Google Scholar]
- 8.Farias TDJ, Brugiapaglia S, Croci S, et al. HLA-DPB1*13:01 associates with enhanced, and KIR2DS4*001 with diminished protection from developing severe COVID-19. Hla 2023; 103(1):e15251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.El-Bendary M, Neamatallah M, Elalfy H, et al. HLA Class II-DRB1 Alleles with Hepatitis C Virus Infection Outcome in Egypt: A Multicentre Family-based Study. Annals of hepatology 2018; 18:68–77. [DOI] [PubMed] [Google Scholar]
- 10.Carrington M, O'Brien SJ. The influence of HLA genotype on AIDS. Annu Rev Med 2003; 54:535–51. [DOI] [PubMed] [Google Scholar]
- 11.Kaslow RA, Carrington M, Apple R, et al. Influence of combinations of human major histocompatibility complex genes on the course of HIV-1 infection. Nature medicine 1996; 2:405–11. [DOI] [PubMed] [Google Scholar]
- 12.Migueles SA, Sabbaghian MS, Shupert WL, et al. HLA B*5701 is highly associated with restriction of virus replication in a subgroup of HIV-infected long term nonprogressors. Proceedings of the National Academy of Sciences of the United States of America 2000; 97:2709–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gao X, Nelson GW, Karacki P, et al. Effect of a single amino acid change in MHC class I molecules on the rate of progression to AIDS. N Engl J Med 2001; 344:1668–75. [DOI] [PubMed] [Google Scholar]
- 14.Mugabi T, Namombwe S, Dai B, et al. Etiology and Outcomes of Meningitis among Adults in Three Ugandan Referral Hospitals, 2018–2023: A Prospective Cohort Study in a High-HIV Endemic Setting. Am J Trop Med Hyg 2025; 112:1273–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Durski KN, Kuntz KM, Yasukawa K, Virnig BA, Meya DB, Boulware DR. Cost-effective diagnostic checklists for meningitis in resource-limited settings. J Acquir Immune Defic Syndr 2013; 63:e101–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Boulware DR, Nalintya E, Rajasingham R, et al. Adjunctive sertraline for asymptomatic cryptococcal antigenemia: A randomized clinical trial. Medical mycology 2020; 58:1037–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Levin AE, Bangdiwala AS, Nalintya E, et al. Outpatient Cryptococcal Antigen Screening Is Associated With Favorable Baseline Characteristics and Improved Survival in Persons With Cryptococcal Meningitis in Uganda. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2023; 76:e759–e65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Augusto DG, Murdolo LD, Chatzileontiadou DSM, et al. A common allele of HLA is associated with asymptomatic SARS-CoV-2 infection. Nature 2023; 620:128–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Norman PJ, Hollenbach JA, Nemat-Gorgani N, et al. Defining KIR and HLA Class I Genotypes at Highest Resolution via High-Throughput Sequencing. American journal of human genetics 2016; 99:375–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ashburner M, Ball CA, Blake JA, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nature genetics 2000; 25:25–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.UniProt: the Universal Protein Knowledgebase in 2023. Nucleic acids research 2023; 51:D523–d31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Aleksander SA, Balhoff J, Carbon S, et al. The Gene Ontology knowledgebase in 2023. Genetics 2023; 224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gerstein AC, Jackson KM, McDonald TR, et al. Identification of Pathogen Genomic Differences That Impact Human Immune Response and Disease during Cryptococcus neoformans Infection. mBio 2019; 10:e01440–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Nilsson JB, Kaabinejadian S, Yari H, et al. Accurate prediction of HLA class II antigen presentation across all loci using tailored data acquisition and refined machine learning. Science advances 2023; 9:eadj6367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Dhanda SK, Vir P, Raghava GP. Designing of interferon-gamma inducing MHC class-II binders. Biology direct 2013; 8:30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Jarvis JN, Meintjes G, Rebe K, et al. Adjunctive interferon-gamma immunotherapy for the treatment of HIV-associated cryptococcal meningitis: a randomized controlled trial. AIDS (London, England) 2012; 26:1105–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kijak GH, Walsh AM, Koehler RN, et al. HLA class I allele and haplotype diversity in Ugandans supports the presence of a major east African genetic cluster. Tissue Antigens 2009; 73:262–9. [DOI] [PubMed] [Google Scholar]
- 28.Begovich AB, Moonsamy PV, Mack SJ, et al. Genetic variability and linkage disequilibrium within the HLA-DP region: analysis of 15 different populations. Tissue Antigens 2001; 57:424–39. [DOI] [PubMed] [Google Scholar]
- 29.Tougan T, Ito K, Palacpac NM, Egwang TG, Horii T. Immunogenicity and protection from malaria infection in BK-SE36 vaccinated volunteers in Uganda is not influenced by HLA-DRB1 alleles. Parasitol Int 2016; 65:455–8. [DOI] [PubMed] [Google Scholar]
- 30.Gibson KM, Drögemöller BI, Foell D, et al. Association Between HLA-DPB1 and Antineutrophil Cytoplasmic Antibody-Associated Vasculitis in Children. Arthritis & rheumatology (Hoboken, NJ) 2023; 75:1048–57. [DOI] [PubMed] [Google Scholar]
- 31.Hilhorst M, Arndt F, Joseph Kemna M, et al. HLA-DPB1 as a Risk Factor for Relapse in Antineutrophil Cytoplasmic Antibody-Associated Vasculitis: A Cohort Study. Arthritis & rheumatology (Hoboken, NJ) 2016; 68:1721–30. [DOI] [PubMed] [Google Scholar]
- 32.Ou G, Liu X, Xu H, Ji X, Liu X, Wang J. Variation and expression of HLA-DPB1 gene in HBV infection. Immunogenetics 2021; 73:253–61. [DOI] [PubMed] [Google Scholar]
- 33.Watanabe M, Nakamura Y, Sato S, et al. HLA genotype-clinical phenotype correlations in multiple sclerosis and neuromyelitis optica spectrum disorders based on Japan MS/NMOSD Biobank data. Scientific reports 2021; 11:607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Jarvis JN, Meintjes G, Rebe K, et al. Adjunctive interferon-γ immunotherapy for the treatment of HIV-associated cryptococcal meningitis: a randomized controlled trial. AIDS (London, England) 2012; 26:1105–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Jarvis JN, Bicanic T, Loyse A, et al. Determinants of mortality in a combined cohort of 501 patients with HIV-associated Cryptococcal meningitis: implications for improving outcomes. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2014; 58:736–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Jin JH, Lee KT, Hong J, et al. Genome-wide functional analysis of phosphatases in the pathogenic fungus Cryptococcus neoformans. Nature communications 2020; 11:4212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kretschmer M, Reiner E, Hu G, et al. Defects in phosphate acquisition and storage influence virulence of Cryptococcus neoformans. Infection and immunity 2014; 82:2697–712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ganendren R, Carter E, Sorrell T, Widmer F, Wright L. Phospholipase B activity enhances adhesion of Cryptococcus neoformans to a human lung epithelial cell line. Microbes and infection 2006; 8:1006–15. [DOI] [PubMed] [Google Scholar]
- 39.Hamed MF, Araújo GRS, Munzen ME, et al. Phospholipase B Is Critical for Cryptococcus neoformans Survival in the Central Nervous System. mBio 2023; 14:e0264022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Liu OW, Chun CD, Chow ED, Chen C, Madhani HD, Noble SM. Systematic genetic analysis of virulence in the human fungal pathogen Cryptococcus neoformans. Cell 2008; 135:174–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.