Human Immunology 75 (2014) 524–530

Contents lists available at ScienceDirect

www.ashi-hla.org

journal homepage: www.elsevier.com/locate/humimm

HLA-A⁄02, gender and tobacco smoking, but not multiple sclerosis, affects the IgG antibody response against human herpesvirus 6 Elin Engdahl a, Rasmus Gustafsson a, Ryan Ramanujam a, Emilie Sundqvist b, Tomas Olsson b, Jan Hillert a, Lars Alfredsson c,d, Ingrid Kockum b, Anna Fogdell-Hahn a,⇑ a

Karolinska Institutet, Department of Clinical Neuroscience, The Multiple Sclerosis Research Group, Center for Molecular Medicine L8:00, 171 76 Stockholm, Sweden Karolinska Institutet, Department of Clinical Neuroscience, Neuroimmunology Unit, Center for Molecular Medicine L8:04, 171 76 Stockholm, Sweden Karolinska Institutet, Institute of Environmental Medicine, 171 77 Stockholm, Sweden d Centre for Occupational and Environmental Medicine, Stockholm County Council, Sweden b c

a r t i c l e

i n f o

Article history: Received 18 March 2013 Accepted 11 March 2014 Available online 22 March 2014 Keywords: Human herpesvirus 6 IgG p41 HLA-A⁄02 Multiple sclerosis

a b s t r a c t Multiple sclerosis (MS) is an inflammatory, demyelinating disease of the central nervous system. Both genetic and environmental factors contribute to disease susceptibility and two viruses associated with MS are human herpesvirus (HHV)-6A and HHV-6B, together referred to as HHV-6. This study characterized the plasma IgG antibody response against HHV-6 in MS patients (n = 446) and healthy controls (n = 487), and the relationship between MS susceptibility factors and the anti-HHV-6 response was investigated. In addition, 134 samples were further investigated for IgG against the early HHV-6 antigen p41. Antibody levels were measured with ELISA. The overall seroprevalence against HHV-6 was 90%, with no significant difference in positivity or levels between MS patients and controls. Interestingly, carriership of HLA-A⁄02 and tobacco smoking was associated with lower anti-HHV-6 IgG levels (p = 0.0017 and p = 0.026 respectively), whereas females sex was associated with higher levels (p = 0.0090). No difference in IgG titers against p41 was observed between MS patients and controls. In conclusion, the IgG response against HHV-6 was associated with several factors that have previously been associated with MS susceptibility, possibly reflecting a relation between autoimmunity and how the immune system handles viral infections. Ó 2014 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.

1. Introduction Primary infection of human herpesvirus (HHV)-6A and/or HHV6B (family Herpesviridae, subfamily Betaherpesvirinae) usually occurs early in life with most children being infected before 2 years of age [1,2]. Both species primarily infect T cells but have a broad cell tropism [3], and can establish life-long latency in the host. HHV-6B is the causative agent of exanthema subitum [4], but no disease has been clearly linked to HHV-6A. The two viruses share 90% of their nucleotide sequence [5], and most serological methods used are not able to distinguish between the two viruses. In the present study, HHV-6 is used when no such discrimination has been made. Multiple sclerosis (MS) is a chronic, inflammatory, demyelinating disease of the central nervous system (CNS). Despite many years of intense research, the etiology of MS is still unknown but ⇑ Corresponding author. Fax: +46 8 51773909. E-mail address: [email protected] (A. Fogdell-Hahn).

both genetic and environmental factors are considered to contribute to disease susceptibility. Genetic risk factors for MS includes primarily immune genes [6], with human leukocyte antigen (HLA)-DRB1⁄15 as the strongest risk allele [7,8] and HLA-A⁄02 as a protective allele [7,9]. Other risk factors include female sex, smoking [10], low vitamin D levels (reviewed in [11]) and infections (reviewed in [12]). Both HHV-6A and HHV-6B have been associated with MS; HHV6B protein expression has been found more frequently in oligodendrocytes in MS plaques compared with normal appearing white matter from MS patients and controls [13,14] and intrathecal production of anti-HHV-6 IgG antibodies has been identified in approximately 25% of MS patients [15,16], suggesting a role for HHV-6 within the CNS. HHV-6A DNA load in serum, and messenger RNA load in peripheral blood mononuclear cells (PBMCs), has been shown to be higher during MS exacerbations compared with remissions [17,18]. Also, increased titers of serum anti-HHV-6 IgG antibodies have been shown to positively associate with relapse risk in MS [19], suggesting a role for HHV-6 in mechanisms

http://dx.doi.org/10.1016/j.humimm.2014.03.001 0198-8859/Ó 2014 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.

525

E. Engdahl et al. / Human Immunology 75 (2014) 524–530

of disease activity. Finally, IgM antibodies against HHV-6 are more common in early stages of MS [20], indicating a possible role in disease onset. An increasing number of studies report associations of various environmental factors with MS. Until now, most studies have focused on one risk factor only, but interactions between these may occur. The aim of this study was therefore to characterize the anti-HHV-6 IgG antibody response in MS patients and in healthy controls, and to investigate if factors influencing MS susceptibility also regulate the serological response to HHV-6. 2. Materials and methods 2.1. Patients All MS patients (n = 446) and healthy controls (n = 487) included in this study were part of the Epidemiological Investigation of Multiple Sclerosis (EIMS), a Swedish population based case-control study consisting of incident cases of MS patients together with randomly selected controls matched for age, gender and geographic region from the general population [10] (Table 1A). All MS cases fulfilled the McDonald Criteria [21,22]. All participants gave a blood sample, answered a questionnaire regarding lifestyle and environmental exposures, and provided written informed consent. Included samples were collected between 2005 and 2009. Treatment data were obtained for 393 MS patients from the national Swedish MS registry. Of these, 139 (35%) were defined as treatment naïve as no MS treatment had been prescribed more than 1 day before the blood sample was collected, 41 (10%) were treated with IFNb and 11 (3%) with Glatiramer acetate at sampling and for at least 1 year prior to the sampling date. The remaining 177 (48%) MS patients that did not fulfill any of these treatment criteria were excluded from the analysis of MS treatment effects on HHV-6 serology. The study was approved by the Regional Ethical Review Board at Karolinska Institutet and was performed according to the ethical standards of the Declaration of Helsinki. 2.2. Measurement of anti-HHV6 IgG To assess antibody responses against HHV-6, anti-HHV-6 IgG antibody ELISA kits with whole virion lysate (Z29 strain) as antigen were used according to the manufacturers’ protocol (Advanced

Biotechnologies, Columbia, MD, USA. Cat. No: 15-401-000). The ELISA sensitivity and specificity were both 100% compared by immunofluorescence assay (IFA) as reported by the manufacturer. All plasma samples (n = 933) were run in at least duplicate wells. The results are given as normalized optical density (nOD) levels, calculated by ODsample divided by 2ODneg control. The results were interpreted as suggested in the protocol; nOD levels 60.75 were regarded as negative and nOD levels P1.00 as positive. Borderline samples (nOD 0.76–0.99, n = 36) were excluded from the antibody prevalence analysis but included when titers were analyzed. Samples with a nOD level P0.75 were accepted only if the sample’s nOD values had a coefficient of variance (CV) of less than 25%. The positive control’s nOD ratio on all plates had an inter-assay CV of 13%. 2.3. Measurement of anti-p41 IgG and total IgG To further investigate the specificity of the anti-HHV-6 antibodies, the serological response in samples from 67 relapsing remitting (RR) MS patients and 67 healthy controls (Table 1B) were further examined. The two groups were selected to have samples with equal spread in anti-HHV-6 IgG levels (nOD levels between 1.0 and 12.0, with median titers of 5.4–5.5, p = 0.98) and were matched for gender, HLA-A⁄02 status and smoking habits. First, a commercial ELISA kit was used to detect IgG against p41 according to the manufacturer’s protocol (Bioworld Consulting Laboratories, Mt. Airy, MD, USA. Cat. No: H6-EA-096). Results were obtained as nOD levels and interpreted as above. This kit was validated by the manufacturer by showing a correlation between ELISA results from seropositive and negative individuals with results from immunoprecipitation. Second, total IgG was measured using Human IgG ELISA kits (ALP) according to the manufacturers protocol (Mabtech, Stockholm, Sweden. Cat. No: 3850-1AD-6). Briefly, microtiter plates were coated with 2 lg/ml of the monoclonal antibody MT145 overnight. An IgG standard curve (0.2–50 ng/ml) and plasma samples diluted 1:106 were added to the plate. Bound IgG was detected by alkaline phosphatase-conjugated monoclonal anti-IgG antibodies after which Alkaline Phosphate Yellow (pNPP) was used as substrate (Sigma–Aldrich, St. Louis, MO, USA). All plasma samples and controls were run in duplicate wells. In addition to the 134 MS and control samples described above, plasma samples from

Table 1 Demographic and clinical data on MS patients and controls. (A) All patients and controls included in this study, (B) Selected population for further investigation of total IgG and anti-p41 IgG. This population is not representative of the total study population but matched for anti-HHV-6 IgG titer, gender, HLA-A⁄02 status and smoking habits. (A) Total study population Healthy controls

Number of subjects Age; mean ± SD % Female sex % Current smokersa % HLA-DRB1⁄15 carriers % HLA-A⁄02 carriers Vitamin D [nmol/L]; median (min–max)b Disease duration; median (min–max)c % RRMSd % Treatment naivee % anti-HHV-6 IgG positive Anti-HHV-6 IgG titer; median (min–max)

(B) Population for further analysis MS patients

Female

Male

Female

Male

373 41 ± 11 77 25 29 56 70 (10–197) – – – 90 4.2 (0–13)

114 41 ± 11

319 38 ± 11 72 31 62 40 63 (18–235) 3 (0–31) 87 34 91 4.3 (0–16)

127 39 ± 11

24 29 64 62 (20–138) – – – 87 3.3 (0–12)

34 57 43 62 (10–132) 3 (0–30) 87 38 89 3.3 (0–13)

Healthy controls

MS patients

67 36 ± 10 52 46 37 52 59 (20–197) – – – 100 5.4 (1–12)

67 35 ± 9 52 46 52 52 65 (18–125) 2 (0–19) 100 33 100 5.5 (1–12)

a Current smokers = use of cigarettes, cigars or pipes within 2 years of the collection of blood sample. Smoking status was obtained for 486 healthy controls and 418 MS patients. b Vitamin D level at sampling was obtained for 472 healthy controls and 429 MS patients. c Disease duration is given as the time (years) between estimated onset for first symptom and inclusion in the study. d MS course data was obtained for 442 MS patients. e Treatment data was obtained for 393 MS patients. Treatment naïve is defined as never treated with IFN beta, glatiramer acetate, natalizumab or copaxone.

526

E. Engdahl et al. / Human Immunology 75 (2014) 524–530

8 MS patients and 8 healthy controls which were negative for antiHHV-6 IgG were also investigated for total IgG.

3. Results 3.1. Seroprevalence of anti-HHV-6 IgG antibodies

2.4. Measurement of vitamin D levels Levels of 25-hydroxyvitamin D2 and D3 were measured in the plasma samples by a chemiluminescent immunoassay from DiaSorin on a LIAISONÒ instrument (DiaSorin AB, Sundbyberg, Sweden) as previously described [23].

The IgG seroprevalence against HHV-6 was 90% with no significant difference in status between MS patients and healthy controls. In the further analysis of total IgG concentration, no significant difference between anti-HHV-6 IgG negative and positive samples was detected.

2.5. HLA genotyping

3.2. The influence of different factors on anti-HHV-6 IgG levels

DNA extraction was performed according to standard methods and subsequently HLA-A and HLA-DRB1 was genotyped using a commercial kit for Low resolution HLA typing based on sequence-specific primers (Olerup SSP™, Stockholm, Sweden). Genotypes for HLA-A and DR were unavailable for 17 and 23 patients, respectively. For these, HLA allele status was imputed using HLA⁄IMP01 [24] from high density genotypes generated in a genomewide scan for MS susceptibility genes [6]. The accuracy of this method was 95–99% for 2 digit HLA resolutions at HLA class I and II, and these imputations were therefore considered to be as accurate as those determined experimentally.

Gender (p = 0.0057), current smoking (p = 0.025) and HLA-A⁄02 carriership (p = 0.000091) were significantly associated to antiHHV-6 antibody nOD levels (Fig. 1), whereas HLA-DR⁄15 carriership was not significantly associated (data not shown). Also, multiple sclerosis affection status was not significantly associated with anti-HHV-6 IgG nOD levels. Furthermore, no significant difference was observed on nOD levels between the treatment naïve (n = 139, median nOD 3.7), IFNb (n = 41, median nOD 2.7) or Glatiramer acetate (n = 11, median nOD 6.1) treated MS patients. Single linear regression models determined that neither age nor vitamin D levels affected the nOD levels significantly.

2.6. Statistics Mann Whitney U or Kruskal–Wallis ANOVA tests were applied in STATISTICA 11 (StatSoft, Uppsala, Sweden) when statistics between two or three groups of numerical values were calculated. Independent variables were gender, affection status (MS patient or healthy control), smoking, HLA-A⁄02 and HLA-DRB1⁄15 carriership, and treatment. Pearson’s chi-squared tests were performed when statistics between two groups of categorical values were calculated. The anti-HHV-6 antibody nOD distribution was right-skewed (skewness 0.66) and was transformed to obtain a more normal distribution and to constrain residual variance for the linear regression analyses. This was done in STATISTICA 11 using the Box-Cox power transformation (k = 0.44). Single linear regression was conducted using all samples (n = 933) to determine if statistically significant relationships existed between anti-HHV-6 IgG nOD levels and age at sampling and vitamin D levels. In order to determine if confounding might influence measured effects, two distinct multiple linear regression models were constructed, using all samples (n = 933) and MS cases only (n = 446). Stepwise backward variable selection was employed, beginning with all variables included (gender, affection status, smoking, age, vitamin D levels, HLAA⁄02 and HLA-DRB1⁄15 carriership). The case only model included the additional covariate disease duration. The Akaike information criterion [25] was used as a measure of model information to determine optimum model parameters. These linear regression models were conducted using R version 2.15.2. For further investigations of positive samples (n = 134; Table 1B), the ratios ‘‘anti-HHV-6 IgG nOD/total IgG concentrations’’, ‘‘anti-p41 IgG nOD/total IgG concentrations’’ and ‘‘antip41 IgG nOD/anti-HHV-6 IgG nOD’’ were compared between the following groups (all of which were matched for anti-HHV-6 IgG titers): MS patients versus healthy controls, presence versus absence of HLA-A⁄02, current smokers versus never smokers and females versus males. In the analyses involving total IgG, one subject was excluded due to abnormally high total IgG levels (34 g/L). In addition, linear regression analyses were made using GraphPad 5 (GraphPad Software, La Jolla, CA, USA) to compare anti-p41 IgG nOD to anti-HHV-6 IgG nOD or total IgG concentrations, and to compare anti-HHV-6 IgG nOD to total IgG concentrations. All graphs were made in GraphPad 5. The threshold for significance was set at p < 0.05.

3.3. Control for confounders In order to test for the effect of potential confounders on antiHHV-6 antibody nOD levels, a multiple linear regression model was conducted where gender, affection status, smoking, age, vitamin D levels, HLA-A⁄02 and HLA-DRB1⁄15 carriership were included as potential covariates. The results indicated that none of these factors affected the detected associations between gender, smoking and HLA-A⁄02 carriership and nOD levels. An optimized model with the inclusion of gender, smoking and HLA-A⁄02 carriership resulted in similar p-values to the first multiple linear regression model (p = 0.0090, p = 0.026 and p = 0.0017, respectively). In the model using MS cases only, disease duration was not significantly associated with anti-HHV-6 IgG nOD levels. Pearson’s chi-squared tests showed no significant difference between males and females regarding their smoking habits or HLA-A⁄02 status (data not shown). 3.4. Further investigation of anti-p41 IgG and total IgG The ‘‘anti-HHV-6 IgG nOD/total IgG concentration’’ ratio was not significantly different depending on MS affection status (Fig. 2A), HLA-A⁄02 carriership, smoking or gender (data not shown). No significant linear correlation was observed between anti-HHV-6 IgG nOD levels and the total IgG concentration (Fig. 2B). Further investigations of the antibody specificity revealed that 100% of the anti-HHV-6 IgG positive samples were positive for IgG antibodies against the HHV-6 early antigen p41. Borderline correlation was observed between anti-p41 IgG and total IgG levels (p = 0.05; Fig. 2D) but not to anti-HHV-6 IgG levels (Fig. 2F). However, neither MS disease (Fig. 2C and E), HLA-A⁄02 carriership, smoking nor gender influenced the ‘‘anti-p41 IgG nOD/total IgG concentration’’ ratio or the ‘‘anti-p41 IgG nOD/anti-HHV-6 IgG titer’’ ratio significantly (data not shown). 4. Discussion In the present study, an anti-HHV-6 IgG seroprevalence of 90% in patients with MS and in healthy controls is reported, which is in line with previous reports (75–100%) [1,2,15,26–28]. The data

E. Engdahl et al. / Human Immunology 75 (2014) 524–530

527

Fig. 1. Anti-HHV-6 IgG antibody responses in MS patients (MS) and in healthy controls (HC), divided by gender (A), smoking habits (B) and HLA-A⁄02 carriership (C). Current smokers are defined as having used cigarettes, cigars or pipes within 2 years of the collection of blood sample. Each dot represents one study participant. Median titers are marked with red lines. p-Values are calculated using Mann Whitney U Test. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. Ratios and correlations of anti-HHV-6 specific antibodies and total IgG concentrations in MS patients (MS) and in healthy controls (HC). Anti-HHV-6 IgG nOD levels normalized for total IgG concentrations (A), relationship between anti-HHV-6 IgG nOD level and total IgG concentration (B), anti-p41 IgG nOD levels normalized for total IgG concentrations (C), relationship between anti-HHV-6 IgG nOD level and total IgG concentration (D), anti-p41 IgG nOD levels normalized for anti-HHV-6 IgG nOD levels (E), relationship between anti-p41 and anti-HHV-6 IgG nOD levels (F). Each dot represents one study participant and red lines are median titers (A, C and E) or regression lines (B, D and F). Regression analyses include both MS patients and controls. p-Values are calculated using Mann Whitney U Test. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

of this study show a wide spread in antibody levels against HHV-6 in the population and underlying mechanisms for these differences in antibody levels were investigated, with focus on factors previously associated with MS susceptibility. No difference in anti-HHV-6 IgG levels was seen between MS patients and healthy controls, supporting the results of some of

the previous reports also comparing IgG responses against HHV6 in serum between MS patients and healthy controls [20,29] or other neurological disease (OND) controls [15,28]. However, other studies proposed significantly higher anti-HHV-6 IgG antibody titers in serum or plasma of MS patients compared with healthy controls [26,27,30]. One of the negative studies [28] and two of

528

E. Engdahl et al. / Human Immunology 75 (2014) 524–530

the positive studies [26,30] used immunofluorescence assay (IFA) with HHV-6A infected HSB-2 cells whereas the others [15,20,27,29] used ELISA with whole virion lysate as antigen, suggesting that the different outcomes are not solely due to the assay used. Interestingly, it has been suggested that serum IgG responses against HHV-6A are more abundant and of higher titers than against HHV-6B in MS patients compared with OND controls [31]. Further investigation of the HHV-6A or -6B species specificity of the IgG responses measured in the present study would be interesting, but will require development of a less time consuming screening method than the immunofluorescence avidity assay used by Virtanen et al. [31]. All negative studies and one of the positive [30] were performed in different European countries whereas the other positive studies were performed in Iran [27] and USA [26], respectively. Hence, the different results of the different studies cannot directly be explained by differences of the populations from which the study groups were sampled. An important remark, however, is that two of the positive studies [26,27] were performed on small samples of around 20 MS patients and 20 controls and therefore might have been underpowered. With 446 MS patients and 487 controls, our present study is the largest and together with the other negative studies favors the notion that an increased plasma/serum IgG response against HHV-6 is not regularly observed in MS. However, our results do not necessarily rule out a role for HHV-6 in MS, but rather suggest that IgG serology against whole virion lysate measured in the peripheral blood might be too blunt to detect a role of common viruses in autoimmune disease causing events. Virus lysate is a mix of many different proteins, but we assume that glycoproteins and structural proteins have a prominent role in complete virions and that such proteins would be overrepresented in ELISA coatings when viral lysate is used as antigen. According to Ablashi et al. [32], 90% of healthy controls have plasma IgG antibodies against the glycoprotein gp110, but only 28% have IgG against the early nuclear antigen p41/38. As the DNA-binding protein p41 is not an exposed antigen on released virus particles, we speculated that a humoral response against this antigen would be relatively uncommon and could be a sign of cell lysis. Therefore, we wanted to investigate the IgG response against p41 in a subgroup of anti-HHV-6 IgG positive subjects. As 100% of these were positive for anti-p41 IgG, our study could not confirm this hypothesis. Our study could not reveal any difference in plasma anti-p41 IgG levels between MS patients and controls. Previous studies investigating plasma or serum IgG against p41/38 have reported contradictory results with increased frequency of positive MS patients [32] as well as no difference between MS patients and controls [33]. However, studies investigating plasma/serum IgM responses against p41/38 seems less controversial and support that anti-p41 IgM is more common in MS patients compared with controls [32,33]. Whereas serum anti-HHV-6 IgM antibodies are a sign of active infection [34], IgG antibodies is rather a measure of history of infection. IgM antibodies against HHV-6 have been shown to be more common in the early stages of MS [20], but as increased reactivations of HHV-6 have been seen during relapses [17,18] and since increased serum IgG titers against HHV-6 have been shown to be positively associate with relapse risk in MS [19], we speculate that an increased frequency of reactivations might influence the MS disease and furthermore might be reflected by a rise in IgG levels in serum. However, we did not note this in the present study. Instead we observed a vast variation in serological levels both in patients with MS and in controls, indicating that the level of antiviral IgG antibodies against a common virus such as HHV-6 is a reflection of a general activity of the immune system that is influenced by several factors such as HLA types, smoking and gender. These are the same factors that have been shown to influence

susceptibility to MS and autoimmunity in general. There might not be any connections between these but if there are, several possible theoretical explanations can be suggested. Starting with the HLA association, this study reveals for the first time that plasma IgG antibody response against HHV-6 is lower in subjects carrying the HLA-A⁄02 allele. This HLA-A⁄02 association is in line with data reported for IgG against Epstein-Barr virus (EBV) where carriers of the HLA-A⁄02 allele had significantly lower plasma IgG titers against Epstein-Barr virus nuclear antigen 1 (EBNA1) [35], indicating that HLA-A⁄02 positive individuals respond differently to herpesvirus infections. One possible feature of HLA-A2 that might influence the immune response against herpesviruses is the partial independency of transporter associated with antigen processing (TAP) loading due to the HLA-A2 molecule’s ability to present hydrophobic peptides [36]. Many viruses have mechanisms to interfere with the MHC class I peptide loading complex (reviewed in [37]). For example, EBV and cytomegalovirus (CMV) encode for proteins (BNLF2a and US6 respectively) that block TAP-mediated peptide loading on MHC class I molecules [38,39]. Cross-presentation, i.e. the presentation of peptides derived from exogenous antigens on MHC class I on dendritic cells, is a crucial step in the activation of cytotoxic T lymphocytes (CTLs) against infected cells. Accordingly, if HLA-A2 is present, the peptide loading and crosspresentation in dendritic cells can occur even though the TAP complex is blocked by viral proteins during viral infection and CTLs might become activated despite the presence of viral encoded TAP blocking peptides [40]. Also, MHC class I peptide presentation is crucial for CTL recognition of infected cells and CD8+ T cells are important in the control of reactivating herpesviruses (reviewed in [41–43]). If HHV-6A and/or -6B encodes for a protein that can block TAP-mediated peptide loading onto MHC class I, then the lower antibody levels in HLA-A⁄02 carriers observed in this study might be explained by a larger frequency of activated CTLs in the primary acute infection, clearing the infection more efficiently and thus leading to less subsequent reactivations of the virus. As B cells need to encounter their specific antigen in order to be activated, free virions/viral proteins released from cells are needed to mount a B cell response. Subsequently, CTL killing of infected cells by apoptosis leads to less released virions, and anti-viral IgG antibodies might not be produced to the same extent as if T cells had failed to control the intracellular infection. Hence, the lower antibody levels against common herpesviruses in HLA-A⁄02 carriers observed by us and Sundqvist et al. [35] might reflect a general protective T cell effect against infections when having this partial TAP-independent presentation of viral peptides in HLA-A⁄02 allele carriers. We hypothesize that the protective effect of HLA-A⁄02 in MS [9] can be due to an ability of HLA-A2 to reduce the intracellular viral burden of common viral infections when these establish as latent infections early in life and subsequently yield weaker reactivations later in life. A higher antibody response against HHV-6 was observed in females compared with males. Females have a generally more active immune system with higher IgG levels [44] and a highly active immune system might be one mechanism behind the overrepresentation of females seen in autoimmune diseases like MS [45]. It is possible that the increased levels of IgG antibodies against HHV-6 among females seen in our material is a reflection of this increased immune activity rather than a link between gender, MS and HHV-6. However, since we did not observe a significant correlation between total IgG and anti-HHV-6 IgG in our material (Fig. 2), we cannot confirm that anti-HHV-6 IgG levels are dependent on general IgG immunity. Another explanation for the higher anti-HHV-6 IgG levels in females could be that the monthly variations in T cell immunity induced by hormones of the menstrual cycle [46,47] result in monthly fluctuations in virus control [48]. Consequently, females might have insufficient control

E. Engdahl et al. / Human Immunology 75 (2014) 524–530

of HHV-6 reactivations more frequently than males, reflected in higher antibody levels. The increased risk of developing MS in females might be due to a combination of frequent reactivations of latent viruses and a generally active immune response trying to clear the reactivations. This study revealed that smoking also lowered the antibody response against HHV-6. A decrease in general serum IgG by tobacco smoking has been seen previously [44]. The results are therefore not unexpected. However, smoking is known to increase the risk for MS [10], hence this result is somewhat difficult to fit into a model hypothesizing that a low IgG level is a consequence of fewer and weaker virus reactivations due to increased CTL activation. However, lowered antibody production induced later in life by starting smoking might lead to less neutralization of reactivated virions. If HHV-6 reactivation is a risk factor for developing MS [12,20], this lowered serological protection might be one reason for the increased risk of developing MS in tobacco smokers. To investigate this hypothesis, one should measure the anti-HHV6A/6B IgG response and the frequency and magnitude of virus reactivations before and after the individual starts smoking. In conclusion, although no difference was seen in the IgG responses against HHV-6 between MS patients and controls, the serological response was associated with factors previously associated with MS susceptibility. This indicates that an IgG response against a common latent virus like HHV-6 might be regulated by similar factors that also regulate susceptibility to autoimmune diseases like MS. The study does not settle the controversy of whether HHV-6 is a risk factor for MS or not, but warrants further investigations of the role of common viral infections in autoimmune diseases. Grants E.S. has received research grant from The Swedish Association of Persons with Neurological Disabilities, T.O. has received grants from the Swedish research council, the Swedish Brain Foundation, Knut and Alice Wallenbergs foundation and the AFA foundation. L.A. has received grants from the Swedish Research Council and the Swedish Council for Working Life and Social Research. J.H. has received grants from the Swedish Research Council, Bibbi and Nils Jensens Foundation and the European Commission. Disclosure E.E., R.G., R.R., E.S., L.A., I.K. and A.F.H. have no conflicts of interests to declare. T.O. has received lecture and/or advisory board honoraria from BiogenIdec, Merck, Novartis and Genzyme. The same companies have provided unrestricted MS research grants. All authors have approved the final article. J.H. received honoraria for serving on advisory boards for BiogenIdec, Merck-Serono and Novartis and for speaker’s fees from BiogenIdec, Merck-Serono, Bayer-Schering, Teva and Sanofi-Aventis. He has served as P.I. for and received projects supported by BiogenIdec, Merck-Serono, and Bayer-Schering. Author contributions E.E. and R.G. designed the study and performed the anti-HHV-6 ELISAs, E.E. performed the anti-p41 and total IgG ELISAs, provided figures and wrote the paper. E.E., R.G., R.R. and A.F.H. interpreted the data, where E.E. and R.R. performed the statistical analyses. E.E., R.G., R.R., E.S., L.A., I.K. and A.F.H. participated in drafting, reviewing, and revising of the paper for intellectual content. L.A., T.O., J.H. are responsible for the EIMS study. All authors contributed to the final paper.

529

Acknowledgments The authors thank Jenny Link for HLA genotyping and Mohsen Khademi for technical assistance on the handling of the samples used in this study. We also thank Nina Nordin, Karin Kai-Larsen and Anna K. Hedström at the EIMS secretariat for help with collecting data, and we are grateful to all the patients and controls for participating and to all clinics involved in the study; Akademiska sjukhuset, Blekingesjukhuset, Bollnäs sjukhus, Centrallasarettet i Växjö, Centralsjukhuset i Karlstad, Centralsjukhuset i Kristianstad, Danderyds sjukhus, Falu lasarett, Frölunda specialistsjukhus, Gävle sjukhus, Hallands sjukhus, Halmstad, Hallands sjukhus Kungsbacka, Hallands sjukhus Varberg, Helsingborgs lasarett, Höglandssjukhuset Eksjö, Karolinska Universitetssjukhuset Solna & Huddinge, Kullbergska sjukhuset Katrineholm, Kärnsjukhuset i Skövde, Lasarettet i Motala, Länssjukhuset i Kalmar, Länssjukhuset Ryhov i Jönköping, Mälarsjukhuset i Eskilstuna, Norra Älvsborgs sjukhus, Nyköpings lasarett, Odenplans Läkarhus, Sahlgrenska universitetssjukuset, Sjukhuset i Arvika, Sjukhuset i Falköping, Sollefteå sjukhus, Sunderby sjukhus i Luleå, Sundsvalls sjukhus, Södra Älvsborgs sjukhus, Uddevalla sjukhus, Univ. sjukhuset i Linköping, Univ. sjukhuset i Lund, Umeå Universitetssjukhus, Örebro Universitetssjukhus, Malm Universitetssjukhus, Visby lasarett, Vrinnevisjukhuset i Norrköping, Västmanlands sjukhus Västerås, Örnsköldsviks sjukhus, Östersunds sjukhus. References [1] Okuno T, Takahashi K, Balachandra K, Shiraki K, Yamanishi K, Takahashi M, et al. Seroepidemiology of human herpesvirus 6 infection in normal children and adults. J Clin Microbiol 1989;27(4):651–3. [2] Zerr DM, Meier AS, Selke SS, Frenkel LM, Huang ML, Wald A, et al. A population-based study of primary human herpesvirus 6 infection. N Engl J Med 2005;352(8):768–76. [3] De Bolle L, Van Loon J, De Clercq E, Naesens L. Quantitative analysis of human herpesvirus 6 cell tropism. J Med Virol 2005;75(1):76–85. [4] Yamanishi K, Okuno T, Shiraki K, Takahashi M, Kondo T, Asano Y, et al. Identification of human herpesvirus-6 as a causal agent for exanthem subitum. Lancet 1988;1(8594):1065–7. [5] Dominguez G, Dambaugh TR, Stamey FR, Dewhurst S, Inoue N, Pellett PE. Human herpesvirus 6B genome sequence: coding content and comparison with human herpesvirus 6A. J Virol 1999;73(10):8040–52. [6] Sawcer S, Hellenthal G, Pirinen M, Spencer CC, Patsopoulos NA, Moutsianas L, et al. Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature 2011;476(7359):214–9. [7] Brynedal B, Duvefelt K, Jonasdottir G, Roos IM, Akesson E, Palmgren J, et al. HLA-A confers an HLA-DRB1 independent influence on the risk of multiple sclerosis. PLoS One 2007;2(7):e664. [8] Hillert J, Olerup O. Multiple sclerosis is associated with genes within or close to the HLA-DR-DQ subregion on a normal DR15, DQ6, Dw2 haplotype. Neurology 1993;43(1):163–8. [9] Fogdell-Hahn A, Ligers A, Gronning M, Hillert J, Olerup O. Multiple sclerosis: a modifying influence of HLA class I genes in an HLA class II associated autoimmune disease. Tissue Antigens 2000;55(2):140–8. [10] Hedstrom AK, Baarnhielm M, Olsson T, Alfredsson L. Tobacco smoking, but not Swedish snuff use, increases the risk of multiple sclerosis. Neurology 2009;73(9):696–701. [11] Ascherio A, Munger KL, Simon KC. Vitamin D and multiple sclerosis. Lancet Neurol 2010;9(6):599–612. [12] Kakalacheva K, Munz C, Lunemann JD. Viral triggers of multiple sclerosis. Biochim Biophys Acta 2011;1812(2):132–40. [13] Challoner PB, Smith KT, Parker JD, MacLeod DL, Coulter SN, Rose TM, et al. Plaque-associated expression of human herpesvirus 6 in multiple sclerosis. Proc Natl Acad Sci USA 1995;92(16):7440–4. [14] Friedman JE, Lyons MJ, Cu G, Ablashl DV, Whitman JE, Edgar M, et al. The association of the human herpesvirus-6 and MS. Mult Scler 1999;5(5):355–62. [15] Derfuss T, Hohlfeld R, Meinl E. Intrathecal antibody (IgG) production against human herpesvirus type 6 occurs in about 20% of multiple sclerosis patients and might be linked to a polyspecific B-cell response. J Neurol 2005;252(8):968–71. [16] Virtanen JO, Pietilainen-Nicklen J, Uotila L, Farkkila M, Vaheri A, Koskiniemi M. Intrathecal human herpesvirus 6 antibodies in multiple sclerosis and other demyelinating diseases presenting as oligoclonal bands in cerebrospinal fluid. J Neuroimmunol 2011;237(1–2):93–7. [17] Alvarez-Lafuente R, De las Heras V, Bartolome M, Picazo JJ, Arroyo R. Relapsing-remitting multiple sclerosis and human herpesvirus 6 active infection. Arch Neurol 2004;61(10):1523–7.

530

E. Engdahl et al. / Human Immunology 75 (2014) 524–530

[18] Alvarez-Lafuente R, Garcia-Montojo M, De las Heras V, Bartolome M, Arroyo R. Clinical parameters and HHV-6 active replication in relapsing-remitting multiple sclerosis patients. J Clin Virol 2006;37(Suppl. 1):S24–6. [19] Simpson Jr S, Taylor B, Dwyer DE, Taylor J, Blizzard L, Ponsonby AL, et al. AntiHHV-6 IgG titer significantly predicts subsequent relapse risk in multiple sclerosis. Mult Scler 2012;18(6):799–806. [20] Villoslada P, Juste C, Tintore M, Llorenc V, Codina G, Pozo-Rosich P, et al. The immune response against herpesvirus is more prominent in the early stages of MS. Neurology 2003;60(12):1944–8. [21] McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, Lublin FD, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001;50(1):121–7. [22] Polman CH, Reingold SC, Edan G, Filippi M, Hartung HP, Kappos L, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the ‘‘McDonald Criteria’’. Ann Neurol 2005;58(6):840–6. [23] Baarnhielm M, Hedstrom AK, Kockum I, Sundqvist E, Gustafsson SA, Hillert J, et al. Sunlight is associated with decreased multiple sclerosis risk: no interaction with human leukocyte antigen-DRB1⁄15. Eur J Neurol 2012;19(7):955–62. [24] Dilthey AT, Moutsianas L, Leslie S, McVean G. HLA⁄IMP–an integrated framework for imputing classical HLA alleles from SNP genotypes. Bioinformatics 2011;27(7):968–72. [25] Sugiura N. Further analysis of data by akaikes information criterion and finite corrections. Commun Stat A-Theor 1978;7(1):13–26. [26] Ablashi DV, Eastman HB, Owen CB, Roman MM, Friedman J, Zabriskie JB, et al. Frequent HHV-6 reactivation in multiple sclerosis (MS) and chronic fatigue syndrome (CFS) patients. J Clin Virol 2000;16(3):179–91. [27] Behzad-Behbahani A, Mikaeili MH, Entezam M, Mojiri A, Pour GY, Arasteh MM, et al. Human herpesvirus-6 viral load and antibody titer in serum samples of patients with multiple sclerosis. J Microbiol Immunol Infect 2011;44(4):247–51. [28] Enbom M, Wang FZ, Fredrikson S, Martin C, Dahl H, Linde A. Similar humoral and cellular immunological reactivities to human herpesvirus 6 in patients with multiple sclerosis and controls. Clin Diagn Lab Immunol 1999;6(4):545–9. [29] Nielsen L, Larsen AM, Munk M, Vestergaard BF. Human herpesvirus-6 immunoglobulin G antibodies in patients with multiple sclerosis. Acta Neurol Scand Suppl 1997;169:76–8. [30] Sola P, Merelli E, Marasca R, Poggi M, Luppi M, Montorsi M, et al. Human herpesvirus 6 and multiple sclerosis: survey of anti-HHV-6 antibodies by immunofluorescence analysis and of viral sequences by polymerase chain reaction. J Neurol Neurosurg Psychiatry 1993;56(8):917–9. [31] Virtanen JO, Farkkila M, Multanen J, Uotila L, Jaaskelainen AJ, Vaheri A, et al. Evidence for human herpesvirus 6 variant A antibodies in multiple sclerosis: diagnostic and therapeutic implications. J Neurovirol 2007;13(4):347–52. [32] Ablashi DV, Lapps W, Kaplan M, Whitman JE, Richert JR, Pearson GR. Human Herpesvirus-6 (HHV-6) infection in multiple sclerosis: a preliminary report. Mult Scler 1998;4(6):490–6.

[33] Soldan SS, Berti R, Salem N, Secchiero P, Flamand L, Calabresi PA, et al. Association of human herpes virus 6 (HHV-6) with multiple sclerosis: increased IgM response to HHV-6 early antigen and detection of serum HHV-6 DNA. Nat Med 1997;3(12):1394–7. [34] Suga S, Yoshikawa T, Asano Y, Nakashima T, Yazaki T, Fukuda M, et al. IgM neutralizing antibody responses to human herpesvirus-6 in patients with exanthem subitum or organ transplantation. Microbiol Immunol 1992;36(5):495–506. [35] Sundqvist E, Sundstrom P, Linden M, Hedstrom AK, Aloisi F, Hillert J, et al. Epstein–Barr virus and multiple sclerosis: interaction with HLA. Genes Immun 2012;13(1):14–20. [36] Weinzierl AO, Rudolf D, Hillen N, Tenzer S, van Endert P, Schild H, et al. Features of TAP-independent MHC class I ligands revealed by quantitative mass spectrometry. Eur J Immunol 2008;38(6):1503–10. [37] Roder G, Geironson L, Bressendorff I, Paulsson K. Viral proteins interfering with antigen presentation target the major histocompatibility complex class I peptide-loading complex. J Virol 2008;82(17):8246–52. [38] Hislop AD, Ressing ME, van Leeuwen D, Pudney VA, Horst D, Koppers-Lalic D, et al. A CD8+ T cell immune evasion protein specific to Epstein–Barr virus and its close relatives in Old World primates. J Exp Med 2007;204(8):1863–73. [39] Lehner PJ, Karttunen JT, Wilkinson GW, Cresswell P. The human cytomegalovirus US6 glycoprotein inhibits transporter associated with antigen processing-dependent peptide translocation. Proc Natl Acad Sci USA 1997;94(13):6904–9. [40] Ackerman AL, Kyritsis C, Tampe R, Cresswell P. Access of soluble antigens to the endoplasmic reticulum can explain cross-presentation by dendritic cells. Nat Immunol 2005;6(1):107–13. [41] Egan KP, Wu S, Wigdahl B, Jennings SR. Immunological control of herpes simplex virus infections. J Neurovirol 2013;19(4):328–45. [42] Harari A, Zimmerli SC, Pantaleo G. Cytomegalovirus (CMV)-specific cellular immune responses. Hum Immunol 2004;65(5):500–6. [43] Hislop AD, Sabbah S. CD8+ T cell immunity to Epstein–Barr virus and Kaposi’s sarcoma-associated herpes virus. Semin Cancer Biol 2008;18(6):416–22. [44] Gonzalez-Quintela A, Alende R, Gude F, Campos J, Rey J, Meijide LM, et al. Serum levels of immunoglobulins (IgG, IgA, IgM) in a general adult population and their relationship with alcohol consumption, smoking and common metabolic abnormalities. Clin Exp Immunol 2008;151(1):42–50. [45] Whitacre CC, Reingold SC, O’Looney PA. A gender gap in autoimmunity. Science 1999;283(5406):1277–8. [46] Oertelt-Prigione S. Immunology and the menstrual cycle. Autoimmun Rev 2012;11(6–7):A486–92. [47] Pehlivanoglu B, Balkanci ZD, Ridvanagaoglu AY, Durmazlar N, Ozturk G, Erbas D, et al. Impact of stress, gender and menstrual cycle on immune system: possible role of nitric oxide. Arch Physiol Biochem 2001;109(4):383–7. [48] Mostad SB, Kreiss JK, Ryncarz A, Chohan B, Mandaliya K, Ndinya-Achola J, et al. Cervical shedding of herpes simplex virus and cytomegalovirus throughout the menstrual cycle in women infected with human immunodeficiency virus type 1. Am J Obstet Gynecol 2000;183(4):948–55.