Association between human genetic variants affecting the host NK cell response and the development of herpes simplex virus type 1 encephalitis

Herpes simplex virus encephalitis (HSE) is a rare complication of herpes simplex virus type 1 (HSV‐1) infection or reactivation. It is so far unclear why only few patients develop HSE. As natural killer (NK) cells provide an important defense against HSV‐1, we investigated whether there is an association between distinct human genetic variants associated with the host NK cell response and HSE. Forty‐nine adult patients with confirmed HSE and 247 matched controls were analyzed for the distribution of the following genotypes: CD16A (FcγRIIIA) V/F and IGHG1 G1m3/17, both influencing antibody‐dependent cellular cytotoxicity; HLA‐E*0101/*0103, associated with NK cell activation; and SLFN13 rs9916629C/T associated with NK cell response. Homozygous HLA‐E*0101:0101 and HLA‐E*0103:0103 variants as well as the rs9916629CC genotype were overrepresented in HSE patients compared to controls (p ≤ 0.001). Notably, cooccurrence of the homozygous HLA‐E*0101 and rs9916629CC genotypes was present in 19% of patients but totally absent in controls (p ≤ 0.0001). Distribution of CD16A and IGHG1 variants did not differ between patients and controls. Our data show that the rare combination of HLA‐E*0101:0101 and rs9916629CC is significantly associated with HSE. Possibly, these genetic variations could be useful as clinical markers predicting HSE prognosis and helping to adapt the treatment of HSE in the individual patient.

response and thereby may contribute to the development of HSE in individual persons. 7 Individual human genetic factors have already been identified that influence the risk to develop HSE. These factors are involved in the production of interferons and thus in the innate immune response to HSV, and especially affect Toll-like receptor 3 related pathways in children and adults. [8][9][10] The natural killer (NK) cell response is crucial in the defense against HSV-1, as NK cells recognize virus-infected cells by different pathways and elicit cytokines and a cytotoxic response against the infected cells. 11 NK cells play, however, a controversial role in HSE. It was shown that the NK cell response may confer some degree of protection against HSE, as patients displaying impaired NK cell functions have an increased risk for this disease. 12 However, an excessive NK cell response may also lead to extensive tissue destruction and deterioration of HSE. 13 The fine balance between NK cell activation and inhibition may be influenced by individual host genetic variations of factors with impact on the antiviral NK cell response. Such genetic variations may lead to individual differences in the antiviral response against HSV-1 and thus in the patients' clinical outcomes. 11 Previous data have shown that human genetic variations influencing the NK cell-mediated antibody-dependent cellular cytotoxicity (ADCC) may affect the severity of general HSV-1 infections. 14,15 ADCC is based on the interaction between virus-specific antibodies and the activating NK cell CD16A (FcγRIIIA) receptor, which further leads to NK cell-driven lysis of the infected cells. A single amino acid substitution in CD16A (phenylalanine [F] to valine [V] at position 158) increases the binding capacity between NK cell receptor and antibodies, and thus the ADCC response in vitro. 16 In addition, a polymorphism in the IGHG1 gene, coding for the IgG1 constant heavy chain, results in different G1m allotypes (3/17), and the G1m3 allotype shows increased binding to CD16A, thereby also enhancing ADCC. 14 The NK cell activation is modulated by the regulation of HLA-E, a nonclassical major histocompatibility complex (MHC) molecule which is expressed on the cell surface. HLA-E is stabilized by specific virus peptides, and binds to the NK cell inhibitory receptor NKG2A/ CD94. 17 In European populations, HLA-E exists mainly as two alleles, HLA-E*0101 and HLA-E*0103. These exhibit different binding properties to virus peptides, which has further impact on their interaction with NK cells. 17 HLA-E is expressed also on neurons under pathological conditions, however, it is still unclear whether the HLA-E/NKG2A pathway has an impact on the development of HSV-1 infection and HSE.
Recently it was described that the rs9916629 C/T polymorphism on chromosome 17q12 close to the SLFN13 gene is associated with alterations in NK cell subpopulations and functions. 18 The C allele was associated with strong degranulation of NK cells, suggesting enhanced cytotoxicity and release of inflammatory markers, and this was especially evident for individuals carrying the rs9916629 CC variant. 18 While it was demonstrated that some of these genetic variations in the NK cell response have an impact on the clinical course of general HSV-1 infections, their association with the development of HSE caused by HSV-1 is unclear yet. 12 In the present study, we therefore investigated whether there is an association between distinct host genetic variations influencing the NK cell response and the development of HSE caused by HSV-1 in adult patients.

| Peptide prediction
Prediction of HSV-1-specific HLA-E binding nonamers was performed via the Epitope Prediction and Analysis Tools by IEDB Analysis Resource, using the MHC I binding prediction algorithm based on NetMHCpan-4.1a. 19 The whole proteome of 15 HSV-1 strains, including the NCBI reference sequence (strain 17) as well as strains of laboratory and clinical origin, were used for prediction (Supporting Information: Table 1) and consecutively compared by NCBI BLAST. 20

| HLA-E stabilization assay
Functional HLA-E binding was assessed for the four peptides predicted to match best with HLA-E, using a predicted binding score of 0.605 as cut-off. 21

| Genotyping
Genomic DNA was isolated from the diagnostic CSF samples of HSE patients and from the serum of healthy controls using the NucliSens EasyMag extractor (BioMérieux) and eluted in 50 µL H 2 O. G1m allotypes 3/17 and CD16A 158V/F variants were determined by Sanger sequencing as described previously. 22,23 HLA-E genotyping was performed using a previously published TaqMan protocol, with primers and probes as described in detail earlier. 24 The rs9916629 C/T variants were determined using an in-house TaqMan assay. 25 The rs9916629 and HLA-E TaqMan assays were performed with the Luna Universal Probe qPCR Master Mix (NEB Biolabs) as recommended by the manufacturers.

| CD16A and G1m variants in HSE patients and controls
First, we investigated the association of HSE with human genetic variants with impact on NK cell-mediated ADCC. We therefore determined the CD16A 158V/F as well as the G1m 3/17 variants in HSE patients and controls. As shown in Figure 1A,B, no significant differences were observed between both cohorts (p > 0.05, respectively; χ 2 test).  Table 3). Both peptide sequences are highly conserved between HSV-1 strains (NCBI BLAST, Supporting Information: Table 1). Thus, no variations of the HSV-1 peptides exist, which may differently impact the NK cell response over HLA-E presentation.

| Distribution of rs9916629 genotypes between patients and controls
We further investigated variations in the rs9916629 polymorphism in HSE patients and controls. The distribution of the rs9916629 alleles differed significantly between patients and controls (p ≤ 0.001, χ 2 test; Figure 1D). The homozygous CC variant occurred significantly more often in HSE patients than in controls and the rs9916629 CT was found more frequently in the control population, while the distribution of the rs9916629 TT genotype did not differ significantly between patients and controls.

| HLA-E and rs9916629 risk allele combination
Finally, as specific HLA-E and rs9916629 variants were overrepresented in patients with HSE, we assessed the association of specific cooccurrence patterns of these genetic markers with HSE. We

| DISCUSSION
In the present study we reveal that variants of specific host factors In other herpesviruses infections, differences in the interaction between HLA-E and NK cells are in part conferred by virus specific variations within the viral peptides binding to HLA-E. 28 30,31 In this study, we also identified a second variant, the rs9916629 CC genotype, which is significantly overrepresented in patients with HSE. The rs9916629 SNP was included in our investigations as the rs9916629 C allele was previously indirectly associated with alterations in NK cell subpopulations and functions. 18 Especially the presence of rs9916629 CC was linked to an enhanced cytotoxic and proinflammatory NK cell response. 18  Notably, in this study, we only included adult HSE patients with a median age of 65 years, which indicates that HSE in our patients mostly resulted from HSV-1 reactivation. 32 It remains to be further investigated whether the risk variants identified here also play a role in HSE developing during HSV-1 primary infections. 33 We found no significant differences in the distribution of the genetic factors influencing ADCC, that is, CD16A and G1m, between HSE patients and controls. This is in contrast to recent data on general HSV-1 infections, however not including HSE, which showed that the presence of CD16A V/V together with the G1m3/3 variants was associated with an asymptomatic course of HSV-1 infection, probably due to an increase in ADCC activity in these patients. 14 The lacking association of these variants with HSE in our patients might reflect a difference between the role of ADCC in HSV-1 infections in the CNS and in other compartments.
In conclusion, we found that variations in HLA-E and rs9916629 genotypes, especially when cooccurring, are significantly associated with the risk to develop HSE. It will be necessary to further assess whether the identification of these genetic variations could serve as a clinically helpful marker with implications in the composition and duration of therapy against HSE.