TLR4 896A/G and TLR9 1174G/A polymorphisms are associated with the risk of infectious mononucleosis

Toll-like receptors (TLRs) recognize pathogen-associated molecular patterns and activate innate and adaptive immune responses. Single nucleotide polymorphisms (SNPs) within the TLR genes may influence host–pathogen interactions and can have an impact on the progression of infectious diseases. The present study aimed to investigate the genotype distribution of TLR2 (2029C/T, rs121917864; 2258G/A, rs5743708), TLR4 (896A/G, rs4986790), and TLR9 (− 1237T/C, rs5743836; − 1486T/C, rs187084; 1174G/A, rs352139; and 2848C/T, rs352140) polymorphisms in 149 children and adolescents with infectious mononucleosis (IM) and 140 healthy individuals. The potential association of TLR SNPs with the clinical manifestations of EBV infection was also studied. The presence of TLR2, TLR4, and TLR9 SNPs was identified by polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP). EBV DNA loads were detected by quantitative real-time PCR assay. The TLR4 896 GG and the TLR9 1174 GA genotypes were associated with an increased risk of EBV-related IM in examined patients (p = 0.014 and p = 0.001, respectively). The heterozygous genotype of the TLR4 896A/G SNP was associated with an increased risk of elevated liver enzyme levels and leukocytosis (p < 0.05). Our preliminary study revealed that the TLR4 896A/G and the TLR9 1174G/A polymorphisms seem to be related to the course of acute EBV infection in children and adolescents.


Results
Clinical outcome assessment. At the time of this study 149 participants with acute IM, including 93 children (median age: 4.3 years; range 3 months-9.0 years) and 56 adolescents (median age: 14.2 years; range 10.0 years-17.5 years), and 140 healthy volunteers were examined. In all patients with IM, EBV infection was confirmed by the serologic and/or virologic tests. In brief, EBV-IgM was positive for 109 individuals, viral capsid antigen (VCA)-IgG for 46, EBV-IgG for 71, and EBNA-IgG was positive for 30 participants. Almost all patients (38/40) in whom specific IgM was undetected were children < 7 years of age. The results are presented separately for the two patient groups (children and adolescents) in Table 1. Higher frequency of EBV-specific IgM in adolescents than children (p = 0.008) was found, whereas children presented higher levels of the EBNA IgG antibodies (p = 0.037). The EBV DNA was detected in whole blood samples from 69 (74.2%) children and 35 (62.5%) adolescents (mean 2.13 × 10 4 copies/ml ± 1.23 × 10 5 copies/ml and mean 1.20 × 10 4 copies/ml ± 4.74 × 10 4 copies/ml, respectively). In all patients with IM, the presence of the EBV DNA in mouth saliva and/or throat swabs was found (mouth saliva: mean 8.77 × 10 5 copies/ml ± 1.75 × 10 6 copies/ml; throat swabs: mean 5.55 × 10 4 copies/ml ± 2.23 × 10 5 copies/ml). Patients were classified as having clinically confirmed IM when the following signs or symptoms were observed: atypical-lymphocytes, pharyngitis, lymphadenopathy, fever, fatigue, and enlarged spleen and/or liver. Clinical manifestations of acute IM were found in all examined children and youth patients. However, as expected from other studies, specific signs and symptoms of IM (elevated liver enzyme levels, atypical-lymphocytes, and thrombocytopenia) were detected more frequently in adolescents than in children (Table 1). It was found that elevated liver enzyme levels were more common in adolescents than children (82.1% vs. 61.3%, p = 0.010). There were also more adolescents than children with atypical-lymphocytes (66.1%  (Table 2). For the TLR2 2029C/T SNP and the TLR4 896A/G SNP, the homozygous recessive genotypes were detected more frequently in children and adolescents with IM than in healthy subjects (4.0% vs. 0.0%, p = 0.030 and 6.7% vs. 0.7%, p = 0.011, respectively; Table 2). Consequently, the wild-type C alleles of these SNPs were detected more frequently in uninfected individuals compared with EBV-infected cases (p = 0.039 and p = 0.004, respectively; Table 2). The heterozygous genotype of the TLR9 1174G/A polymorphism was more common in patients with IM than in healthy individuals (47.0% vs. 30.0%, p = 0.004), while no difference in the frequency of the alleles was observed (p > 0.05).
Mutation present in at least one allele of the TLR9 2848C/T SNP occurred more frequently in patients with IM than in healthy subjects (55.4% vs. 38.9%, p = 0.0001; Table 2). No significant differences for the TLR2 2258G/A and the TLR9 − 1237T/C and − 1486C/T SNPs were found. Moreover, no sex differences in the TLR SNPs frequency among patients with IM were observed. The expected genotype frequencies for TLR9 − 1486T/C and 2848C/T SNPs were not in Hardy-Weinberg Equilibrium (HWE; p < 0.001 for both) and were excluded from further analysis.

The wild-type genotype of the TLR2 2029C/T polymorphism is associated with the EBV replication.
To further examine the association between the TLR polymorphisms and viral infection, we correlated the specific TLR SNPs with the EBV DNA levels in the peripheral blood of IM patients. As shown in Fig. 2, the median EBV DNA levels in blood were lower among individuals who had a wild-type genotype for the TLR2 2029C/T (mean 1.09 × 10 4 copies/ml ± 2.66 × 10 4 copies/ml) compared with those who were heterozygous or homozygous recessive (mean 3.40 × 10 4 copies/ml ± 1.58 × 10 5 copies/ml) for this polymorphism (p = 0.014). Age group-specific analyses confirmed that children carrying the wild-type genotype for the TLR2 2029C/T SNP had lower EBV DNAemia than those with heterozygous or homozygous recessive genotypes (mean 1.37 × 10 4 copies/ml ± 3.24 × 10 4 copies/ml vs. mean 3.21 × 10 4 copies/ml ± 1.9 × 10 5 copies/ml; p = 0.015). Such a correlation was observed also for adolescents (mean 7.81 × 10 3 copies/ml ± 1.59 × 10 4 copies/ml vs. mean 2.07 × 10 4 copies/ ml ± 8.05 × 10 4 copies/ml; p = 0.042, respectively). No association was observed between the EBV DNAemia and any other TLR polymorphisms (p > 0.05) in both study groups.
TLR haplotypes in patients with IM. Haplotype analysis of the TLR2 2029C/T, 2258G/A, the TLR4 896A/G, and the TLR9 − 1237T/C, and 1174G/A SNPs showed that the most frequent haplotype was CGATG (53.5%), which was detected in 48.9% of patients with IM and 58.1% of healthy individuals. The CAATA haplotype was detected at a minor frequency in 1.5% of patients with IM and 2.0% of healthy volunteers, whereas haplotype TAATG was determined only in EBV-infected patients (3.8%). We found no evidence for linkage disequilibrium (LD) for the all examined TLR SNPs (p > 0.05, r 2 < 0.2).

Discussion
TLRs have been demonstrated to play a crucial role in modulating innate recognition of viruses not only by serving as pathogen sensors but also by activating signaling pathways that result in the increased production of proinflammatory cytokines and type I IFNs. This preliminary study provides the first evidence that TLR polymorphisms seem to influence the outcome of the infectious mononucleosis in children and adolescents. Patients with the TLR4 896A/G and the TLR9 1174G/A polymorphisms had a higher risk of IM and significantly increased the frequency of specific manifestations when compared with subjects with the wild-type genotype.
No studies were examining the allele frequency distribution and the role of TLR SNPs in patients with EBVrelated IM. We observed that the TLR4 896A/G SNP and an intronic polymorphism 1174G/A in exon 2 of the TLR9 gene were significantly associated with increased risk of EBV infection in our population. Polymorphisms in TLR4 gene, which is located in chromosome 9q33.1, has been implicated so far in the pathogenesis of several infectious diseases, including bacterial [23][24][25] , fungal 26 , parasitic 27 , and some viral infection, such as human immunodeficiency virus 28,29 , Kaposi sarcoma-associated herpesvirus 30 , and respiratory syncytial virus (RSV) 31 . It was previously described that missense mutations and single nucleotide polymorphisms in the TLR4 gene could alter the function of the receptor and impairs recognition of the pathogens 30 www.nature.com/scientificreports/   www.nature.com/scientificreports/  www.nature.com/scientificreports/ substitution at residue 299 polymorphism (896A/G) of the TLR4 was found to be associated with susceptibility to bacterial pneumonia, possibly through impaired first-line defense mechanisms 33 . Individuals carrying the TLR4 896A/G polymorphisms were significantly less responsive to bacterial peptides derived from Escherichia coli and Porphyromonas gingivalis compared with the wild-type subjects 34,35 . Functional studies have demonstrated that this SNP was also associated with impaired NF-κB and IFN regulatory factor 3 activation in response to lipopolysaccharide, F protein from RSV, and chlamydial Hsp60 36 . It was found that TLR9 contributes to the recognition of EBV and is expressed on B cells, a natural target of the virus infection 18,37 . The TLR9 1174 G/A SNP was linked to rapid disease progression in children with malaria and HIV-infected patients 38,39 and was also associated with nonresponse to anti-TNF treatment among patients with inflammatory bowel disease 40 . In our previous studies, no significant differences in the frequencies of these polymorphisms were detected among CMV-infected and healthy infants within the same ethnic group 20,41 . However, a reduced risk of the CMV infection in infants with the 1174 GG genotype was noticed 20 . We suggest that both the TLR4 896A/G and the TLR9 1174 G/A polymorphisms may influence the immune response and impair NF-κB activation in EBV infection. Several studies have implicated that TLR polymorphisms are involved in the development of lymphoid and non-lymphoid EBV-associated malignancies [42][43][44][45][46][47][48][49][50] . It has been previously reported that the − 16933T/A SNP in the TLR2 gene increased the risk of follicular lymphoma and decreased the risk of chronic lymphocytic leukemia 43 . Mutation present in both alleles of the TLR2 2029C/T and 2258G/A SNPs were associated with susceptibility to gastric carcinoma in China 42 . In the TLR4 gene, the 896A/G polymorphism influences the risk of mucosaassociated lymphoid tissue lymphoma and Hodgkin lymphoma 42 , but not with non-Hodgkin lymphoma 45,46 . The combined CT/TT genotype of the 1196C/T SNP and the GC genotype of the 11350G/C polymorphism in the 3′-untranslated region of the TLR4 gene may alter its expression, influence the expression of inflammatory cytokines and chemokines, and increased the risk of nasopharyngeal carcinoma (NPC) 49,50 . The occurrence and progression of NPC were also associated with the TLR9 − 1486T/C SNP 48 . The TLR9 − 1486 CC genotype increased the susceptibility of NPC in the Chinese population and patients with this genotype were inclined to advanced tumor stage and lymph node metastasis. No association between the TLR9 − 1237T/C and 2848C/T SNPs and the risk of NPC was found 48 . Nevertheless, carriers of the − 1237 C and 2848 A allele had been reported to be associated with an increased risk of Hodgkin's lymphoma in the Caucasian population 47 . The − 1237T/C polymorphism was associated with non-Hodgkin lymphoma in Portuguese and Italian, but not in the United States cohort of patients 44 . Analysis in silico of the TLR9 promoter showed that the mutated − 1237T/C and − 1486T/C variants create a putative c-Rel/NF-kB transcription factor binding sites that influence transcriptional regulation of the gene 51 . The NF-kB binding site in the C allele at the position − 1237 enhanced the transcriptional activity of the TLR9 gene and affect activation of proinflammatory cytokines, chemokines, and the adaptive immune response 52 . In contrast, another study revealed that the TT allelic variant of − 1237T/C SNP had higher transcriptional activity 53 .
A few investigations indicated that the risk of developing infectious mononucleosis after primary infection with EBV correlates with the age of patients and adolescents would react more strongly against EBV infection than children [54][55][56] . It is also known that the frequency of IM symptoms depends on age, ethnicity and geographic location [57][58][59] . Most people are exposed to the virus as children, when the disease produces flu-like symptoms such as pharyngitis, whereas a characteristic triad of fever, pharyngitis, and lymphadenopathy is more frequently presented in adolescents and young adults. In this study, we observed that elevated liver enzyme levels, thrombocytopenia, as well as higher atypical-lymphocytes were more frequent in adolescents than children. This result supports the hypothesis that the prevalence of acute IM rises significantly with patient age, and is common in adolescents 54,55 . These findings are coincident with that described by Wang et al., who demonstrated that young patients had significantly increased levels of liver enzymes and atypical-lymphocytes, while no significant www.nature.com/scientificreports/ difference in the prevalence of fever between children and adolescents was observed 58 . The most significant laboratory abnormality found in the present study was atypical-lymphocytes, which occurred more frequently in adolescents than in children (66.1% vs. 48.5%). The low platelet count, in course of IM, was observed in 26.3% adult patients in France (age 16-53 years) 59 , less than 1% young adults in the United States 60 , 2.3-4.5% preschool children and youth patients in China 58,61 , and 7.3% Mexican children (age 0-17.5 years) 62 . The higher incidence of rash in children than in adolescents (25.8% vs. 10.7%) was observed in our study. Almost one-third population of Israeli patients up to 18 years had a rash during the IM, but the age of the patients was not associated with the development of this symptom 57 . Similarly, no statistically significant difference in patient age and the incidence of rash was found among preschool children (21.5%) and youth patients (13.9%) in Beijing, China 58 . Part of the explanation that would elucidate the more frequent risk of IM in adolescents and young adults than in children is that the percentage of CD8+ T cell counts increased and CD4+ T cell counts decreased with age increment 61 . Moreover, significantly higher levels of early-differentiated CD56 dim NKG2A + killer-cell immunoglobulin-like receptors (KIR) − NK cells in the peripheral blood of children than in adolescents and young adults may affect the course of EBV infection 63,64 . It was also found that polymorphisms in the HLA class I locus may predispose patients to the development of IM upon primary EBV infection 65 . McAulay et al. described that the nature of primary EBV infection and the level of viral persistence can be determined by the genetic variation in T cell responses 65 . It is also possible that adolescents receive a larger amount of the virus through deep kissing, while young children probably acquire EBV from parents or guardians, who during reactivation of infection transmit smaller infectious amounts of EBV, or from siblings or other children 54,66,67 . Furthermore, the heterophile antibody test is less sensitive and often unreliable in young children 3,56,68 . It is also known that EBV VCAs cause lifelong persistent IgG titers, while antibodies of the IgM type are produced only transiently but are not necessarily produced in all patients with primary infections 69 . However, to understand the differential impact age on susceptibility to primary EBV infection, further research in age-matched patient groups are needed. The present study is first to examine the distribution and possible association between the TLR gene polymorphisms and the clinical and/or laboratory findings of infectious mononucleosis. Similar to other genetic association studies, this study has some potential limitations. First, the present study may be limited by a relatively small number of participants. However, we enrolled patients from a well-characterized population and all noticeable symptoms were verified by experienced clinicians. Moreover, serological documentation of EBV infection indicates the higher incidence of advanced-stage disease or the delayed recognition of IM in the examined children than adolescents. Further studies in a larger sample group are needed to confirm our findings and to evaluate the function of the disease-associated TLR polymorphisms.

Methods
Study population. A total of 289 blood samples was obtained from 149 patients with IM (median age: The presence of EBNA-IgG, EBV-IgM, and EBV-IgG or VCA-IgG was measured using an immunochemiluminescence assay (CLIA, DiaSorin, Saluggia, Italy) according to the manufacturer's instructions. The presence of EBV DNA in the whole blood and/or the presence of IgM antibodies to VCA in the absence of measurable antibodies to EBNA-1 was considered the evidence for primary EBV infection as described previously 3,70 . AST, ALT, and GGT were determined by dry chemistry method using Johnson & Johnson Vitros 5.1 FS Chemistry Analyzer (Ortho Clinical Diagnostics, Raritan, NJ, USA) or by the kinetic method using Architect 2000 (Abbott Laboratories, Abbott Park, IL, USA). Mild elevations of ALT and AST were commonly discovered in individuals with mild or no symptoms and were defined as increased liver enzyme values higher than 2-5 times the upper limit of normal. Liver disease was diagnosed in patients with elevated liver enzyme levels greater than 5 times the upper limit of the reference range. Platelet counts were tested using optical and/or impedance methods using the Sysmex XE-2000 Analyzer (Sysmex Corp., Kobe, Japan). Thrombocytopenia was recognized when the platelet count fell below 150 × 10 9 /l of blood, whereas leukocytosis was classified as elevated white blood cells count above the age-specific reference values. All the individuals with EBV infection were ethnically classified as European descents and were enrolled from the Central and South-West areas of Poland. Clinical data from EBV infected patients were summarized in Table 1. This study was approved by the appropriate Bioethics Committee of the Medical University of Lodz (RNN/120/09/KE) and the Ethics Committee of the Polish Mother's Memorial Hospital Research Institute (7/2014). Written informed consent was provided from parents of all children who participated in the study. All experiments were performed in accordance with relevant guidelines and regulations.