Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

HLA Allele E*01:01 Is Associated with a Reduced Risk of EBV-Related Classical Hodgkin Lymphoma Independently of HLA-A*01/*02

  • Paloma Martín,

    Affiliation Group of Molecular Pathology, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯
  • Isabel Krsnik,

    Affiliation Department of Hematology, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯
  • Belen Navarro,

    Affiliation Department of Hematology, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯
  • Mariano Provencio,

    Affiliation Department of Oncology, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯
  • Juan F. García,

    Affiliation Department of Pathology, MD Anderson Cancer Center, Madrid, Spain

    ⨯
  • Carmen Bellas,

    Affiliation Group of Molecular Pathology, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯
  • Carlos Vilches ,

    ‡ These authors are shared senior authors on this work.

    Affiliation Group of Immunogenetics and Histocompatibility, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯
  • Natalia Gomez-Lozano

    ngomez@idiphim.org

    ‡ These authors are shared senior authors on this work.

    Affiliation Group of Immunity and lymphoproliferative diseases, Instituto de Investigación Sanitaria Puerta de Hierro (IDIPHIM), Majadahonda, Spain

    ⨯

Abstract

Background

An inefficient immune response against Epstein-Barr virus (EBV) infection is related to the pathogenesis of a subgroup of classical Hodgkin lymphomas (cHL). Some EBV immune-evasion mechanisms target HLA presentation, including the non-classical HLA-E molecule. HLA-E can be recognized by T cells via the TCR, and it also regulates natural killer (NK) cell signaling through the inhibitory CD94/NKG2A receptor. Some evidences indicate that EBV-infected B-cells promote the proliferation of NK subsets bearing CD94/NKG2A, suggesting a relevant function of these cells in EBV control. Variations in CD94/NKG2A-HLA-E interactions could affect NK cell-mediated immunity and, consequently, play a role in EBV-driven transformation and lymphomagenesis. The two most common HLA-E alleles, E*01:01 and E*01:03, differ by a single amino acid change that modifies the molecule function. We hypothesized that the functional differences in these variants might participate in the pathogenicity of EBV.

Aim

We studied two series of cHL patients, both with EBV-positive and-negative cases, and a cohort of unrelated controls, to assess the impact of HLA-E variants on EBV-related cHL susceptibility.

Results

We found that the genotypes with at least one copy of E*01:01 (i.e., E*01:01 homozygous and heterozygous) were underrepresented among cHL patients from both series compared to controls (72.6% and 71.6% vs 83%, p = 0.001). After stratification by EBV status, we found low rates of E*01:01-carriers mainly among EBV-positive cases (67.6%). These reduced frequencies are seen independently of other factors such as age, gender, HLA-A*01 and HLA-A*02, HLA alleles positively and negatively associated with the disease (adjusted OR = 0.4, p = 0.001). Furthermore, alleles from both HLA loci exert a cumulative effect on EBV-associated cHL susceptibility.

Conclusions

These results indicate that E*01:01 is a novel protective genetic factor in EBV-associated cHL and support a role for HLA-E recognition on the control of EBV infection and lymphomagenesis.

Citation: Martín P, Krsnik I, Navarro B, Provencio M, García JF, Bellas C, et al. (2015) HLA Allele E*01:01 Is Associated with a Reduced Risk of EBV-Related Classical Hodgkin Lymphoma Independently of HLA-A*01/*02. PLoS ONE 10(8): e0135512. https://doi.org/10.1371/journal.pone.0135512

Editor: Joseph S. Pagano, The University of North Carolina at Chapel Hill, UNITED STATES

Received: February 12, 2015; Accepted: July 22, 2015; Published: August 11, 2015

Copyright: © 2015 Martín et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: This work was supported by Miguel Servet programs CP09/00182 (NGL) and CP08/00218 (PM) and the Spanish Cancer Network (RTICC RD 06/0020/0047) all from Instituto de Salud Carlos III (FEDER). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Epstein-Barr virus (EBV) is a ɣ-herpesvirus with a high prevalence worldwide (over 90%) which interacts with the host immune system and typically establishes benign lifelong latent infection in the B-cells of healthy immunocompetent individuals [1]. However, in some cases, the normal homeostasis of the EBV-specific response is disrupted and leads to the malignant transformation of the infected cells. EBV-driven oncogenesis is related to the development of a subset of several B-cell lymphomas in immunocompetent individuals (mainly, Burkitt [BL], diffuse large B cell [DLBCL] and classical Hodgkin’s [cHL] lymphomas) and in immunocompromised patients (post-transplant [PTLD] and HIV-associated lymphoproliferative disorders).

The relationship between EBV infection and cHL is well established. The rates of EBV-related cHL vary by ethnic group, age, gender, socio-economic status, and histological subtype, and its development is associated with a previous history of infectious mononucleosis [2]. The epidemiology of EBV suggests a complex interplay between genetic, viral and environmental factors that converge to an inadequate immune control of EBV infection. However, genetic drivers of EBV-specific immune responses and lymphomagenesis in EBV-related cHL remain unclear.

There is increasing evidence that the NK cell repertoire plays a role in EBV-directed immunity. In this way, NK cells exhibit greater cytotoxicity during acute EBV infection, and in a humanized mouse model, NK cells limit the EBV viral load, infectious mononucleosis symptoms, and tumor formation [3–5]. Their significance in the control of EBV infection is supported by the observation that primary immunodeficiencies affecting NK cells or NK cell recognition of EBV-transformed cells are associated with EBV-positive malignancies and fatal EBV infection [6–8].

Recently, an EBV-driven imprint on the NK compartment has been reported and is characterized by the proliferation of an NK cell population expressing the inhibitory receptor CD94/NKG2A (CD56dimNKG2A+KIR-CD57-) in tissues and peripheral blood during acute symptomatic infection [9, 10]. Moreover, a CD56bright NK cell subset with a similar NKG2A+ phenotype proved to be relevant for controlling infection in the tonsils and to have the potential to inhibit in vitro EBV-induced transformation [11]. These findings support a role for CD94/NKG2A-HLA-E signaling in NK-mediated control of EBV and its associated pathologies.

The inhibitory CD94/NKG2A receptor, a member of the heterodimeric CD94/NKG2 family, is expressed on NK and CD8+ T cell subsets. This receptor interacts with a high affinity and in a peptide-specific manner with the non-classical class I molecule HLA-E loaded with signal peptides derived from other HLA-I molecules [12, 13]. The CD94/NKG2A-HLA-E interaction constitutes a sensor that monitors HLA class I expression on healthy cell surfaces and provides an immunomodulatory mechanism of tolerance. HLA‐E can also present stress- or pathogen-derived peptides (including EBV), whether to CD94/NKG2A or TCR bearing cells, to survey cellular stress or infection by NK and T cells [14–16].

Expression of HLA molecules on infected cells varies during the EBV cycle infection and inversely correlates with sensitiveness to NK killing. Latently EBV-infected B-lymphoblastoid cell lines (B-LCLs) express high levels of HLA molecules, and resist NK lysis [3, 17]; however, HLA class I expression is down-modulated upon entry into the lytic cycle, which is associated with an increase of the NK cell response [18, 19]. HLA class I antigens synthesis and assembly pathways are disrupted in a concerted manner by several EBV antigens (e.g., BNLF2a, BGLF5, BILF1, vIL10) during the lytic cycle, which affects and targets a broad range of HLA antigens [20–22]. HLA-E expression is also affected by EBV immune-evasion mechanisms, presumably circumventing the recognition of HLA-E-viral peptide complexes by CD8+ T cells [23].

Compared to the characteristic genetic diversity of classical class I MHC genes, HLA-E is fairly conserved. The low level of HLA-E polymorphism was present before the generation of classical HLA diversity and seems to be driven by a different selective pressure. These features suggest an essential function for this molecule [24, 25]. Fifteen different HLA-E nucleotide sequences have been described to encode two different protein sequences: E*01:01 and E*01:03 (http://www.ebi.ac.uk/ipd/imgt/hla [26]). These two HLA‐E molecules are present in more than 90% of the population with similar frequencies, suggesting that balancing selection is acting on the dimorphism [27]. Both variants differ at only one non-synonymous substitution (1147A>G; rs1264457) that results in an amino acid change (arginine to glycine) at position 107 in the alpha-2 domain of the heavy chain. This substitution confers different levels of surface expression and peptide affinities. E*01:03 is expressed at higher levels than E*01:01, probably due to an increase in peptide binding ability that results in a more stable configuration [28].

The allelic diversity of HLA-E is associated to several clinical situations. In HCV infection, the homozygous E*01:03 genotype seems to correlate with chronification [29], and in hematopoietic stem cell transplantation, it has been shown to confer protection against graft‐versus‐host disease, leukemia relapse and mortality [30–32]. Other investigators have reported that the E *01:01 allele associates with recurrent miscarriage [33] and a reduced risk of Behcet’s disease [34].

We performed a case-control study to evaluate whether the genetic diversity of HLA-E affects susceptibility to lymphomagenesis in cHL and its relationship to EBV infection.

Materials & Methods

Ethics Statement

The project was approved by the Ethics Committee of Hospital Puerta de Hierro (Approval number: Acta n° 304). Written informed consent was provided by all of the patients and controls.

Patients and clinicopathological data

One hundred and seventy-five patients diagnosed of cHL (73 EBV+ and 102 EBV-) and 400 unmatched healthy individuals from Hospital Puerta de Hierro Majadahonda (Madrid, Spain) were enrolled in this study. Another series of 82 patients diagnosed with cHL (35 EBV+ and 47 EBV-) from MD Anderson Cancer Center (Madrid, Spain) were studied to validate the results.

cHL was diagnosed based on a morphological and immunohistochemical examination of biopsy materials using standard techniques. The histopathological classification and cHL subclassification were based on the World Health Organization (WHO) 2008 criteria as follows: nodular sclerosis (NS), mixed cellularity (MC), lymphocyte-rich (LR), lymphocyte-depleted (LD) or not classified (NC) cHL. HIV-positive patients were excluded from the study.

Nuclear staining of HRS by in-situ hybridization for EBERs using the Bond ISH EBV-EBER probe was used to detect EBV-infection in the cells. The main characteristics of the cHL patient are summarized in Table 1, and the other clinico-pathological parameters of the patients are shown in S1 Table.

thumbnail
Download:
Table 1. Characteristics of the cHL patients and control subjects.

https://doi.org/10.1371/journal.pone.0135512.t001

HLA typing

DNA was extracted from either peripheral blood with DNAzol (MRC, Cincinnati, OH) or using a Maxwell 16 Blood DNA purification kit (Promega Corp. Madison, CA); or from the paraffin-embedded tissue of biopsied tumoral lymph nodes using the QIAmp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions.

HLA-E typing discriminating the single nucleotide polymorphism in codon 107 (A>G) was performed using PCR-SSP as described [35] with minor modifications. To validate the genotyping method, reference DNAs from the IHW cell panel were used. Randomly selected samples and problematic DNAs were examined by DNA sequencing in a 3130xl Genetic Analyzer Sequencer (Applied Biosystems, Foster City, CA, USA) using the amplification primers HLA-E_Fcon290 5’-ACCGCACAGATTTTCCGAGT-3’ and HLA-E_Rcon396 5’-AGGCGAACTGTTCATACCCG, with HLA-E_Fi6 5’-TCAGTTTAGGCCAAAATGCCCA-3’ as an internal primer for sequencing. HLA-A*01 and A*02 typing was performed by PCR-SSP with the primer pairs FHc527 5’-TGGGAGGCGGTCCATGC-3’ and Rhcg570 5’-AGGTATCTGCGGAGCCCG-3’ for A*01, and FHa144 5’-GAGCCCCGCTTCATCGCA-3’ and RHagg256 5´-TTCACTTTCCGTGTCTCCCC-3’ for A*02 (details available upon request).

Statistical Analysis

Differences in the distribution of HLA-E genotypes reported in this study were calculated using the Pearson χ2 test and considering a dominant model for the protective allele HLA-E*01:01 (E*01:01 carriers vs negatives). In both the control and patient groups, HLA-E genotypes were evaluated for adequacy to Hardy-Weinberg equilibrium.

Odds ratio (OR) values and 95% confidence intervals (CIs) for relative risks were calculated for the genotypes. Frequencies of combined HLA-A-HLA-E phenotypes were compared with Pearson χ2 test or Fisher’s test when any expected frequency was lower than five individuals. Linkage disequilibrium (LD) between HLA-E alleles and A*01 and A*02 was measured by coefficient D’ (D/Dmax). The confounding effect of gender and age was analyzed by a stratified analysis of the association of HLA-E genotypes with the risk of cHL (male, female, <45 years and ≥ 45 years subgroups). Multiple logistic regression was performed to measure the independent contribution of HLA-E genetics and other associated HLA variables (A*01 and A*02) on the risk of EBV-positive cHL. Data were analyzed using the Statistical Package for the Social Sciences software (SPSS 15, Chicago, IL, USA).

Results

Characteristics of cHL patients

One hundred and seventy-five Spanish patients with cHL from Hospital Puerta de Hierro (Madrid) and 400 unmatched healthy individuals from the same area were included in the study. The characteristics of the cHL cohort (Table 1 and S1 Table) were consistent with its epidemiology in Western Europe [36, 37]. The histological distribution was as follows: 60.6% (n = 106) of the cases were NS subtype, followed by a 30.3% (n = 53) with MC histology, 4.6% (n = 8) corresponded to other less frequent subtypes (LD and LR), and 4.6% (n = 8) were not histologically classified.

Consistently with other series [38, 39], 41.7% of the cHL patients were EBV-positive with a bimodal age-incidence curve (data not shown) and the presence of EBV was more frequent in males (79.5% males vs 20.5% females, p = 2.6x10-5). Additionally, 64.7% and 30.2% of the MC and NS cases were EBV-positive. The EBV-negative lymphomas presented a unimodal age distribution, were predominantly found in young adults aged <45 years, lacked an association with gender, and the majority had a NS histology (75.5%).

As a replicative series of cHL, we studied 82 patients from a different medical Center in Madrid (MD Anderson Cancer Center, Spain). Clinical and biological data were very similar to the results obtained in our exploratory series, including epidemiology, EBV association and histology, as shown in Table 1 and S1 Table.

HLA-E distribution in controls and cHL patients

Analysis of the HLA-E genotype distribution in healthy controls (S2 Table) showed that the majority of the individuals (49.6%) were heterozygous at the dimorphism (E*01:01, 01:03). The major (E*01:01) and minor (E*01:03) homozygous genotypes were present in 33.3% and 17% of the individuals, respectively. HLA-E allelic and genotypic frequencies fell into the range reported for other European ancestry reference populations from The 1000 genomes project (http://www.1000genomes.org: CEU, TSI, GBR and FIN) and from other reported population studies [40].

The distribution of the HLA-E genotypes in the cHL cohort (regardless of the EBV status), as well as in all the EBV-categorized subgroups, was closely replicated in both series (S2 Table): Hospital Puerta de Hierro (HPH) and MD Anderson Cancer Centre (MDA). The consistency of the results obtained in the two cohorts supports the validity of our data and, additionally, it permitted to compile and analyze both series together to strengthen the statistical power of the study. The genotype frequencies of HLA-E in controls and patients were consistent with Hardy-Weinberg equilibrium.

HLA-E*01:01 is a protective factor in cHL and its negative association is mainly derived from EBV-positive cases

In cHL patients (either with a positive or a negative EBV status) E*01:01 was less frequent compared to controls (individuals with the E*01:01 allele, 72.4% vs 83%; Table 2; OR 0.5, p = 0.001). To assess the selective effect of HLA-E on the risk of EBV-driven lymphomagenesis, EBV-positive and EBV-negative tumors were analyzed separately. Only 67.6% of the EBV-positive cases were homozygous or heterozygous for the E*01:01 allele, which indicated that E*01:01-carriers have a significantly decreased risk of EBV-associated cHL (OR = 0.4, p = 0.0004, Table 2). EBV-negative patients had an intermediate frequency of E*01:01 (75.8%), and showed a non-significant tendency toward a decreased frequency in comparison to controls (OR = 0.6, p = 0.057). This reduction of E*01:01 frequency in EBV-negative cases is moderate compared to EBV-positive patients cases and our study lacks sufficient statistical power to determine significant differences between these two subgroups (p = 0.15). These results indicated that the reduction of E*01:01-carriers in the global cohort of cHL patients (irrespective of the EBV status) is attributable to differences in the distribution of HLA-E genotypes in the EBV-positive subgroup.

thumbnail
Download:
Table 2. Analysis of the HLA-E*01:01 allele as a protective factor in cHL stratified by EBV status.

https://doi.org/10.1371/journal.pone.0135512.t002

We analyzed the potential effect of the E*01:01 allele dose on the decreased risk of EBV-positive cHL. Similar OR values were obtained for carriers of one (OR = 0.4, p = 0.002) or two copies (OR = 0.4, p = 0.002) of E*01:01 compared to donors lacking the protective allele, indicating that the risk of suffering from EBV-positive cHL does not depend on the E*01:01 allele dosage.

The association of HLA-E*01:01 with EBV-positive cHL is not influenced by age or gender

We stratified the series by gender and age, variables previously associated with EBV-associated cHL, to investigate their possible confounding effect on the contribution of the HLA-E genotype to the risk of developing EBV-associated cHL. The subgroup analysis showed that E*01:01 was similarly associated in males and females (OR = 0.4, p = 0.0037 and OR = 0.4, p = 0.028; respectively) and in <45 years old (OR = 0.5, p = 0.021) versus ≥45 years old (OR = 0.4, p = 0.009) categories (Table 3). This analysis confirmed that the negative association between the E*01:01 allele and the risk of EBV-positive cHL was not influenced by age or gender. Similarly, stratifications of the EBV-positive cHL series by histology and other clinical parameters (disease stage, presence of bulky tumor and survival) did not reveal any differences in the distribution of the HLA-E genotypes (data not shown).

thumbnail
Download:
Table 3. Stratification of HLA-E*01:01 frequency for age and gender.

https://doi.org/10.1371/journal.pone.0135512.t003

Analysis of HLA-E/HLA-A interaction: Identification of HLA-E/HLA-A phenotypes associated to the EBV-positive cHL susceptibility

HLA alleles A*01 and A*02 have been reported to associate with an increased and a reduced risk of EBV-positive cHL, respectively [41, 42]. In our study, A*01 and A*02 in healthy controls had similar frequencies to those reported for Spanish population (A*01: 16.8% and A*02: 51.9%) [43] and consistent with previous works, in the EBV-positive cHL cohort, A*01 is of risk (43.2%, OR = 3.8, p = 2.3x10-8) whilst A*02 is protective (31.9%, OR = 0.4, p = 0.001).

Given the strong LD in the HLA region and the vicinity of HLA-A and -E loci, we analyzed the genetic association of HLA-E alleles and A*01 and A*02, in order to study their interaction and relative contribution to EBV-positive cHL. The LD analysis within our control samples showed that the predisposing A*01 allele was in strong positive LD with the protective E*01:01 allele (D’ = 0.65, p = 0.0082) and in weaker, non-significant negative LD with the risk E*01:03 allele (D’ = −0.13, p = 0.096). On the contrary, the protective A*02 presented a random association with the protective E*01:01 (D’ = −0.05, p = 0.64) and a weak LD with the risk E*01:03 (D’ = 0.15, p = 0.026).

Strikingly, in patients, HLA-A and HLA-E alleles with opposed effects on the risk of the EBV-related cHL had a lower LD compared to those described for controls. We found a decreased non-significant LD between the protective E*01:01 and the predisposing A*01 (D’ = 0.34, p = 0.064) and no association between the risk E*01:03 and the protective A*02 (D’ = 0.02, p = 0.86). These reduced LD between HLA-A and HLA-E alleles with a confronting influence on EBV-associated cHL risk would indicate a selection of HLA phenotypes composed by predisposing HLA-A and HLA-E alleles in EBV-positive cHL patients.

A subgroup analysis stratified by HLA-A alleles revealed a similar distribution of E*01:01 in A*01-carriers (OR = 0.3, p = 0.029) and in A*01-negative donors (OR = 0.4, p = 0.003); and, similarly, in A*02-positive (OR = 0.4, p = 0.036) or A*02-negative subjects (OR = 0.6, p = 0.088). In A*01 A*02 double negative subjects, E*01:01 also presents comparable frequencies, but sample size is not large enough to detect significant differences (80.5% vs. 68.6%; OR = 0.5, p = 0.17). These results indicate that A*01 and A*02 alleles are not confounding factors influencing the protective association between the E*01:01 allele and the risk of EBV-positive cHL.

A multiple logistic regression analysis was performed to further assess the possible influence of A*01 and A*02 on the association of HLA-E to the risk of EBV-related cHL (Table 4). After adjusting by presence of HLA-A*01 and A*02, E*01:01 remained as a significant protective factor. This analysis confirmed that E*01:01 is negatively associated with the risk of EBV-positive cHL (OR = 0.4, p = 0.001) and this association is independent of other known HLA factors that impact on the risk of EBV-related cHL.

thumbnail
Download:
Table 4. Multivariate logistic regression analysis: E*01:01 contributes independently of other HLA factors to the risk of EBV+ cHL.

https://doi.org/10.1371/journal.pone.0135512.t004

Finally, we examined the cumulative effect of the HLA-A and-E polymorphism on the risk of EBV-positive cHL through the comparison of protective phenotypes (E*01:01+ve, A*01-ve or/and A*02+ve) to all other phenotypes or to its opposite risk phenotype (E*01:01-ve, A*01+ve or/and A*02-ve) (Table 5). Protective HLA-E-HLA-A combinations were dominant in healthy controls but clearly underrepresented in EBV-associated cHL (E*01:01+veA*01-ve: 67.2% vs 35.8%, E*01:01+veA*02+ve: 42.6% vs 20.9% and E*01:01+veA*01-veA*02+ve: 36.1% vs 9.5%). On the contrary, the reciprocal risk phenotypes had very low rates in healthy controls and were increased in EBV-associated cHL (E*01:01-veA*01+ve: 1% vs 8.4%, E*01:01-veA*02-ve: 7.8% vs 16.8% and E*01:01-veA*01+veA*02-ve: 0.3% vs 5.3%). In summary, comparison of the frequencies of the protective HLA-A-HLA-E phenotypes to the frequencies of any other genotypes or their opposite risk phenotypes showed a marked decrease on the risk of EBV-associated cHL, demonstrating an additive effect of HLA-E and HLA-A genetics on protection against EBV-positive cHL (e.g. OR = 0.013 for the E*01:01+ve A*01-ve A*02+ve phenotype in comparison with its reciprocal, p = 2.9x10-5).

thumbnail
Download:
Table 5. Analysis of the combined effect of HLA-E*01:01, A*01 and A*02 on the risk of EBV-associated cHL.

https://doi.org/10.1371/journal.pone.0135512.t005

Discussion

Inherited variations in host immune responses have long been associated with susceptibility or protection against disease, including tumors and infections. We analyzed the contribution of HLA-E genetics to the susceptibility of developing EBV-driven cHL in a case-control study. Our results showed a reduced frequency of E*01:01 in EBV-positive cHL cases compared to controls. This difference indicated a protective effect of this variant─ E*01:01-positive individuals would have more than a 50% reduction in the risk of developing EBV-associated cHL. The validity of this genetic association is strengthened by the replication of the HLA-E distribution in a supplementary series of cHL patients in the same geographic region. We also found that the association of HLA-E with EBV-positive cHL is not linked to other reported demographic factors associated with this disease such as age and gender. The relationship of cHL resistance to the E*01:01 allele did not show a dose effect, likely implying a dominant influence of this variant on protection against the disease.

A potential biological rationale for the genetic association with EBV-associated lymphomas described herein could be the influence of HLA-E polymorphism on CD94/NKG2A signaling in NK cells. In this model, it may be speculated that lower expression of HLA-E*01:01 in EBV-infected cells, as compared with E*01:03, could confer an attenuated NKG2A-mediated inhibition of NK or CD8+ T cell killing. This phenomenon could be translated into an enhanced EBV-directed cytotoxic response in E*01:01 carriers that could contribute to protection against a malignant course of EBV infection. Another hypothesis is that E*01:01 could more efficiently present EBV-derived peptides to CD8+ T lymphocytes than E*01:03, generating an increased cytotoxic T cell response to EBV infection.

In EBV-negative cHL, the frequency of E*01:01 was intermediate between healthy controls and EBV+ cases. Although we did not find significant differences because the reduction of the frequency of the E*01:01 allele was not as pronounced as that in the EBV+ group, the results could also have a potential biological significance. HLA-E is expressed at high levels by some tumors [44–46], including neoplastic and microenvironment cells in some cHL patients [47]. The expression and function of HLA-E in cancer cells is not fully understood, but it may represent a mechanism of tolerance that could be influenced by HLA-E polymorphism. Further studies with larger cohorts should be performed to better estimate the relevance of HLA-E genetics in EBV-negative cHL.

One limitation of our study is that the control series could not be monitored for antibodies against EBV. To assess how EBV-serostatus would impact on our results, we estimated, from age-adjusted EBV-seroprevalence data of our Institution (Dr. Portero, F, Microbiology Department; personal communication), a 4% of EBV IgG seronegative controls in our cohort. With this seronegativity rate, and even in the worst scenario for our conclusions (i.e. subtracting 4% [N = 25] supposed seronegatives only from the E*01:01 carrier group), we estimated an irrelevant influence of EBV-seronegative subjects on our results (data not shown)

The influence of the HLA-E locus on the genetic background of EBV-induced tumors has also been studied in nasopharyngeal carcinoma (NPC), an EBV-associated malignancy with a restricted racial and geographical distribution. In Thai NPC patients, a homozygous E*01:03 genotype was associated with an increased risk of disease, i.e., E*01:01-carriers were, as in this study, protected for EBV-tumorigenesis [48]. However, although a similar trend was found in Tunisian patients with NPC, neither HLA-E allele was found to have major effects on NPC susceptibility or progression [49].

Previous investigations on the association of the HLA class I region with EBV-associated cHL pointed to HLA-A1 and A2 being positively and negatively associated, respectively, with EBV+ cHL [47, 48]. The HLA-E locus is ~540 kb centromeric to HLA-A (http://vega.sanger.ac.uk). The physical proximity of the two genes raised the question of whether the relationship between HLA-E and EBV-positive cHL could be secondary to an association with HLA-A or vice versa. Interestingly, we found that the association of HLA-E with EBV-positive cHL is independent of those of A*01 or A*02, which were replicated in our study. Despite the extensive amount of studies assessing the LD in MHC region, few of them include data on HLA-E polymorphism. These studies describe an essentially random association of their alleles, probably due to the presence of recombination hot spots [24] but also, some conserved HLA-A-HLA-E haplotypes, including the combination of the EBV-positive Chl protective E*01:01 and the predisposing A*01 [50]. Here, we found also this association between A*01 and E*01:01 in healthy individuals. This haplotype, contradictory in terms of influence on EBV-positive cHL risk, presents a lower LD among EBV-positive cHL patients in whom it is underrepresented. An analysis integrating the effect of E*01:01, A*01 and A*02 revealed the existence of protective phenotypes by which E*01:01-positive, A*01-negative and A*02-positive individuals would have more than 70 times less risk of developing EBV-associated cHL than the opposite high-risk phenotype (E*01:01-negative, A*01-positive and A*02-negative). The high frequencies of the protective HLA-A-HLA-E phenotypes and the low rates of the opposite risk phenotypes in healthy individuals could indicate an EBV selective pressure shaping HLA-A and HLA-E polymorphism. Our results suggest a potential use of combined HLA-A and-E typing as a biomarker of increased risk of EBV-positive cHL

EBV-positive cHL presents an intrincate epidemiology, and the overall incidence rates vary greatly in different regions of the world [36]. The origin of this geographic discrepancy is unknown, but even when modifiable factors such as lifestyle or the environment are controlled for epidemiological studies, some variability remains across the ethnic groups [51]. These persistent differences in susceptibility to EBV-associated cHL among human populations might be partially explained by polymorphic genetic systems that show ethnic variation in the rates of allelic prevalence, such as HLA-E and HLA-A.

The association of the HLA-E genetic diversity with the susceptibility of EBV-driven cHL suggests a role for the HLA-E molecule in the etiopathogenesis of the disease. These findings may contribute to the understanding of the influence of host genetics on EBV-related lymphomas.

Supporting Information

S1 Table. Clinical and histological features of classic Hodgkin lymphoma subjects.

https://doi.org/10.1371/journal.pone.0135512.s001

(DOCX)

S2 Table. Distribution of HLA-E genotypes in controls and patients.

https://doi.org/10.1371/journal.pone.0135512.s002

(DOCX)

Acknowledgments

We are indebted to all of the individuals who kindly donated blood or tissue samples to the Biobanks of the Hospital Puerta de Hierro and Fundación MD Anderson Internacional España. We are also grateful to Dr Elvira Ramil from the Sequencing Core facility of the Instituto de Investigación Puerta de Hierro for excellent technical support.

Author Contributions

Conceived and designed the experiments: PM CV NGL. Performed the experiments: PM NGL. Analyzed the data: PM CV NGL. Wrote the paper: NGL PM IK CV. Principal investigator and takes primary responsibility for the paper: NGL. Provided clinical cases to the study: IK BN MP JFG. Reviewed cHL biopsies: CB. Performed the statistics: NGL CV.

References

  1. 1. Cohen JI. Epstein-Barr virus infection. N Engl J Med. 2000;343(7):481–492. pmid:10944566
  2. 2. Glaser SL, Lin RJ, Stewart SL, Ambinder RF, Jarrett RF, Brousset P, et al. Epstein-Barr virus-associated Hodgkin's disease: epidemiologic characteristics in international data. Int J Cancer. 1997;70(4):375–382. pmid:9033642
  3. 3. Williams H, McAulay K, Macsween KF, Gallacher NJ, Higgins CD, Harrison N, et al. The immune response to primary EBV infection: a role for natural killer cells. Br J Haematol. 2005;129(2):266–274. pmid:15813855
  4. 4. Balfour HH Jr., Odumade OA, Schmeling DO, Mullan BD, Ed JA, Knight JA, et al. Behavioral, virologic, and immunologic factors associated with acquisition and severity of primary Epstein-Barr virus infection in university students. J Infect Dis. 2013;207(1):80–88. pmid:23100562
  5. 5. Chijioke O, Muller A, Feederle R, Barros MH, Krieg C, Emmel V, et al. Human natural killer cells prevent infectious mononucleosis features by targeting lytic Epstein-Barr virus infection. Cell Rep. 2013;5(6):1489–1498. pmid:24360958
  6. 6. Eidenschenk C, Dunne J, Jouanguy E, Fourlinnie C, Gineau L, Bacq D, et al. A novel primary immunodeficiency with specific natural-killer cell deficiency maps to the centromeric region of chromosome 8. Am J Hum Genet. 2006;78(4):721–727. pmid:16532402
  7. 7. Parolini S, Bottino C, Falco M, Augugliaro R, Giliani S, Franceschini R, et al. X-linked lymphoproliferative disease. 2B4 molecules displaying inhibitory rather than activating function are responsible for the inability of natural killer cells to kill Epstein-Barr virus-infected cells. J Exp Med. 2000;192(3):337–346. pmid:10934222
  8. 8. Shaw RK, Issekutz AC, Fraser R, Schmit P, Morash B, Monaco-Shawver L, et al. Bilateral adrenal EBV-associated smooth muscle tumors in a child with a natural killer cell deficiency. Blood. 2012;119(17):4009–4012. pmid:22427204
  9. 9. Hendricks DW, Balfour HH Jr., Dunmire SK, Schmeling DO, Hogquist KA, Lanier LL. Cutting edge: NKG2C(hi)CD57+ NK cells respond specifically to acute infection with cytomegalovirus and not Epstein-Barr virus. J Immunol. 2014;192(10):4492–4496. pmid:24740502
  10. 10. Azzi T, Lunemann A, Murer A, Ueda S, Beziat V, Malmberg KJ, et al. Role for early-differentiated natural killer cells in infectious mononucleosis. Blood. 2014;124(16):2533–2543. pmid:25205117
  11. 11. Lunemann A, Vanoaica LD, Azzi T, Nadal D, Munz C. A distinct subpopulation of human NK cells restricts B cell transformation by EBV. J Immunol. 2013;191(10):4989–4995. pmid:24108698
  12. 12. Braud VM, Allan DS, O'Callaghan CA, Soderstrom K, D'Andrea A, Ogg GS, et al. HLA-E binds to natural killer cell receptors CD94/NKG2A, B and C. Nature. 1998;391(6669):795–799. pmid:9486650
  13. 13. Lee N, Llano M, Carretero M, Ishitani A, Navarro F, Lopez-Botet M, et al. HLA-E is a major ligand for the natural killer inhibitory receptor CD94/NKG2A. Proc Natl Acad Sci U S A. 1998;95(9):5199–5204. pmid:9560253
  14. 14. Jorgensen PB, Livbjerg AH, Hansen HJ, Petersen T, Hollsberg P. Epstein-Barr virus peptide presented by HLA-E is predominantly recognized by CD8(bright) cells in multiple sclerosis patients. PLoS One. 2012;7(9):e46120– pmid:23049954
  15. 15. Michaelsson J, Teixeira de MC, Achour A, Lanier LL, Karre K, Soderstrom K. A signal peptide derived from hsp60 binds HLA-E and interferes with CD94/NKG2A recognition. J Exp Med. 2002;196(11):1403–1414. pmid:12461076
  16. 16. Joosten SA, van Meijgaarden KE, van Weeren PC, Kazi F, Geluk A, Savage ND, et al. Mycobacterium tuberculosis peptides presented by HLA-E molecules are targets for human CD8 T-cells with cytotoxic as well as regulatory activity. PLoS Pathog. 2010;6(2):e1000782– pmid:20195504
  17. 17. Sivori S, Pende D, Bottino C, Marcenaro E, Pessino A, Biassoni R, et al. NKp46 is the major triggering receptor involved in the natural cytotoxicity of fresh or cultured human NK cells. Correlation between surface density of NKp46 and natural cytotoxicity against autologous, allogeneic or xenogeneic target cells. Eur J Immunol. 1999;29(5):1656–1666. pmid:10359120
  18. 18. Pappworth IY, Wang EC, Rowe M. The switch from latent to productive infection in epstein-barr virus-infected B cells is associated with sensitization to NK cell killing. J Virol. 2007;81(2):474–482. pmid:17079298
  19. 19. Keating S, Prince S, Jones M, Rowe M. The lytic cycle of Epstein-Barr virus is associated with decreased expression of cell surface major histocompatibility complex class I and class II molecules. J Virol. 2002;76(16):8179–8188. pmid:12134023
  20. 20. Quinn LL, Zuo J, Abbott RJ, Shannon-Lowe C, Tierney RJ, Hislop AD, et al. Cooperation between Epstein-Barr virus immune evasion proteins spreads protection from CD8+ T cell recognition across all three phases of the lytic cycle. PLoS Pathog. 2014;10(8):e1004322– pmid:25144360
  21. 21. Zeidler R, Eissner G, Meissner P, Uebel S, Tampe R, Lazis S, et al. Downregulation of TAP1 in B lymphocytes by cellular and Epstein-Barr virus-encoded interleukin-10. Blood. 1997;90(6):2390–2397. pmid:9310490
  22. 22. Zuo J, Quinn LL, Tamblyn J, Thomas WA, Feederle R, Delecluse HJ, et al. The Epstein-Barr virus-encoded BILF1 protein modulates immune recognition of endogenously processed antigen by targeting major histocompatibility complex class I molecules trafficking on both the exocytic and endocytic pathways. J Virol. 2011;85(4):1604–1614. pmid:21123379
  23. 23. Griffin BD, Gram AM, Mulder A, van LD, Claas FH, Wang F, et al. EBV BILF1 evolved to downregulate cell surface display of a wide range of HLA class I molecules through their cytoplasmic tail. J Immunol. 2013;190(4):1672–1684. pmid:23315076
  24. 24. Geraghty DE, Stockschleader M, Ishitani A, Hansen JA. Polymorphism at the HLA-E locus predates most HLA-A and-B polymorphism. Hum Immunol. 1992;33(3):174–184. pmid:1618657
  25. 25. Felicio LP, Porto IO, Mendes-Junior CT, Veiga-Castelli LC, Santos KE, Vianello-Brondani RP, et al. Worldwide HLA-E nucleotide and haplotype variability reveals a conserved gene for coding and 3' untranslated regions. Tissue Antigens. 2014;83(2):82–93. pmid:24400773
  26. 26. Robinson J, Halliwell JA, Hayhurst JD, Flicek P, Parham P, Marsh SG. The IPD and IMGT/HLA database: allele variant databases. Nucleic Acids Res. 28-1-2015;43(Database issue):D423–D431.
  27. 27. Grimsley C and Ober C. Population genetic studies of HLA-E: evidence for selection. Hum Immunol. 1997;52(1):33–40. pmid:9021407
  28. 28. Strong RK, Holmes MA, Li P, Braun L, Lee N, Geraghty DE. HLA-E allelic variants. Correlating differential expression, peptide affinities, crystal structures, and thermal stabilities. J Biol Chem. 2003;278(7):5082–5090. pmid:12411439
  29. 29. Schulte D, Vogel M, Langhans B, Kramer B, Korner C, Nischalke HD, et al. The HLA-E(R)/HLA-E(R) genotype affects the natural course of hepatitis C virus (HCV) infection and is associated with HLA-E-restricted recognition of an HCV-derived peptide by interferon-gamma-secreting human CD8(+) T cells. J Infect Dis. 2009;200(9):1397–1401. pmid:19780673
  30. 30. Tamouza R, Busson M, Rocha V, Fortier C, Haddad Y, Brun M, et al. Homozygous status for HLA-E*0103 confers protection from acute graft-versus-host disease and transplant-related mortality in HLA-matched sibling hematopoietic stem cell transplantation. Transplantation. 2006;82(11):1436–1440. pmid:17164714
  31. 31. Ludajic K, Rosenmayr A, Fae I, Fischer GF, Balavarca Y, Bickeboller H, et al. Association of HLA-E polymorphism with the outcome of hematopoietic stem-cell transplantation with unrelated donors. Transplantation. 2009;88(10):1227–1228. pmid:19935378
  32. 32. Danzer M, Polin H, Proll J, Haunschmid R, Hofer K, Stabentheiner S, et al. Clinical significance of HLA-E*0103 homozygosity on survival after allogeneic hematopoietic stem-cell transplantation. Transplantation. 2009;88(4):528–532. pmid:19696636
  33. 33. Mosaad YM, Abdel-Dayem Y, El-Deek BS, El-Sherbini SM. Association between HLA-E *0101 homozygosity and recurrent miscarriage in Egyptian women. Scand J Immunol. 2011;74(2):205–209. pmid:21410502
  34. 34. Park KS, Park JS, Nam JH, Bang D, Sohn S, Lee ES. HLA-E*0101 and HLA-G*010101 reduce the risk of Behcet's disease. Tissue Antigens. 2007;69(2):139–144. pmid:17257316
  35. 35. Lauterbach N, Voorter CE, Tilanus MG. Molecular typing of HLA-E. Methods Mol Biol. 2012;882143–158.
  36. 36. Glaser SL, Gulley ML, Clarke CA, Keegan TH, Chang ET, Shema SJ, et al. Racial/ethnic variation in EBV-positive classical Hodgkin lymphoma in California populations. Int J Cancer. 2008;123(7):1499–1507. pmid:18646185
  37. 37. Jarrett RF, Stark GL, White J, Angus B, Alexander FE, Krajewski AS, et al. Impact of tumor Epstein-Barr virus status on presenting features and outcome in age-defined subgroups of patients with classic Hodgkin lymphoma: a population-based study. Blood. 2005;106(7):2444–2451. pmid:15941916
  38. 38. Gulley ML, Eagan PA, Quintanilla-Martinez L, Picado AL, Smir BN, Childs C, et al. Epstein-Barr virus DNA is abundant and monoclonal in the Reed-Sternberg cells of Hodgkin's disease: association with mixed cellularity subtype and Hispanic American ethnicity. Blood. 1994;83(6):1595–1602. pmid:8123850
  39. 39. Huang X, Nolte I, Gao Z, Vos H, Hepkema B, Poppema S, et al. Epidemiology of classical Hodgkin lymphoma and its association with Epstein Barr virus in Northern China. PLoS One. 2011;6(6):e21152– pmid:21695175
  40. 40. Antoun A, Jobson S, Cook M, Moss P, Briggs D. Ethnic variability in human leukocyte antigen-E haplotypes. Tissue Antigens. 2009;73(1):39–45. pmid:19140831
  41. 41. Niens M, Jarrett RF, Hepkema B, Nolte IM, Diepstra A, Platteel M, et al. HLA-A*02 is associated with a reduced risk and HLA-A*01 with an increased risk of developing EBV+ Hodgkin lymphoma. Blood. 2007;110(9):3310–3315. pmid:17630352
  42. 42. Hjalgrim H, Rostgaard K, Johnson PC, Lake A, Shield L, Little AM, et al. HLA-A alleles and infectious mononucleosis suggest a critical role for cytotoxic T-cell response in EBV-related Hodgkin lymphoma. Proc Natl Acad Sci U S A. 2010;107(14):6400–6405. pmid:20308568
  43. 43. Gonzalez-Galarza FF, Takeshita LY, Santos EJ, Kempson F, Maia MH, Silva AL, Silva AL, Ghattaoraya GS, Alfirevic A, Jones AR and Middleton D. Allele frequency net 2015 update: new features for HLA epitopes, KIR and disease and HLA adverse drug reaction associations. Nucleic Acid Research 2015, 28, D784–8.
  44. 44. Marin R, Ruiz-Cabello F, Pedrinaci S, Mendez R, Jimenez P, Geraghty DE, et al. Analysis of HLA-E expression in human tumors. Immunogenetics. 2003;54(11):767–775. pmid:12618909
  45. 45. Levy EM, Bianchini M, von Euw EM, Barrio MM, Bravo AI, Furman D, et al. Human leukocyte antigen-E protein is overexpressed in primary human colorectal cancer. Int J Oncol. 2008;32(3):633–641. pmid:18292941
  46. 46. Gooden M, Lampen M, Jordanova ES, Leffers N, Trimbos JB, van der Burg SH, et al. HLA-E expression by gynecological cancers restrains tumor-infiltrating CD8(+) T lymphocytes. Proc Natl Acad Sci U S A. 2011;108(26):10656–10661. pmid:21670276
  47. 47. Kren L, Fabian P, Slaby O, Janikova A, Soucek O, Sterba J, et al. Multifunctional immune-modulatory protein HLA-E identified in classical Hodgkin lymphoma: possible implications. Pathol Res Pract. 2012;208(1):45–49. pmid:22177730
  48. 48. Hirankarn N, Kimkong I, Mutirangura A. HLA-E polymorphism in patients with nasopharyngeal carcinoma. Tissue Antigens. 2004;64(5):588–592. pmid:15496202
  49. 49. Hassen E, Ghedira R, Ghandri N, Farhat K, Gabbouj S, Bouaouina N, et al. Lack of association between human leukocyte antigen-E alleles and nasopharyngeal carcinoma in Tunisians. DNA Cell Biol. 2011;30(8):603–609. pmid:21332388
  50. 50. Romero V, Larsen CE, Duke-Cohan JS, Fox EA, Romero T, Clavijo OP, et al. Genetic fixity in the human major histocompatibility complex and block size diversity in the class I region including HLA-E. BMC Genet. 2007;814–
  51. 51. Glaser SL, Clarke CA, Chang ET, Yang J, Gomez SL, Keegan TH. Hodgkin lymphoma incidence in California Hispanics: influence of nativity and tumor Epstein-Barr virus. Cancer Causes Control. 2014;25(6):709–725. pmid:24722952