Prospective investigation of herpesvirus infection and risk of glioma

Abstract

Glioma is an aggressive neoplasm of the brain with poorly understood etiology. A limited number of pathogens have been examined as glioma risk factors, but data from prospective studies with infection status determined before disease are lacking. Herpesviruses comprise a large family of DNA viruses that infect humans and are linked to a range of chronic diseases. We conducted a prospective evaluation of the association between antibody to six human herpesviruses and glioma risk in the Janus Serum Bank (Janus) and the Cancer Prevention Study-II (CPS-II). In Janus and CPS-II, the risk for glioma was not related to seroprevalence of herpes simplex virus-1, varicella zoster virus, or human herpes viruses 6A or 6B. In Janus, seropositivity to either the Epstein Barr virus (EBV) EA[D] or VCAp18 antigen was associated with a lower risk of glioma (ORs: 0.55 [95% CI 0.32–0.94] and 0.57 [95% CI 0.38–0.85]). This inverse association was consistent by histologic subtype and was observed for gliomas diagnosed up to two decades following antibody measurement. In Janus, seropositivity to at least one of three examined cytomegalovirus (CMV) antigens (pp150, pp52, pp28) was associated with an increased risk of nonglioblastoma (OR: 2.08 [95% CI 1.07–4.03]). This association was limited to tumors diagnosed within 12 years of antibody measurement. In summary, we report evidence of an inverse association between exposure to EBV and glioma. We further report that CMV exposure may be related to a higher likelihood of the nonglioblastoma subtype.

1 |. INTRODUCTION

Glioma is an aggressive primary neoplasm of the brain with a poorly understood etiology. Gliomas represent 80% of malignant brain tumors and contribute disproportionately to the cancer mortality burden in the United States, ranking ninth for cancer-related death.¹ Five-year relative survival for glioblastoma (GBM), the most common histologic subtype of glioma, is only 5%. Risk factors for glioma include advancing age, male gender, European ancestry,¹ inherited genetic susceptibility,² taller height,³ and an older age at menarche in women.⁴ Ionizing radiation remains the only known environmental contributor, highlighting the lack of modifiable risk factors that could translate into prevention or treatment opportunities for this deadly tumor.⁵ A number of infectious agents can cross the blood-brain barrier,⁶ contribute to neuroinflammation and neurological disease,⁷ and have been shown to influence brain tumorigenesis in experimental models,⁸ but only a limited number of pathogens have been examined as glioma risk factors.

Herpesviruses comprise a large family of DNA viruses that routinely infect humans and are linked to a range of diseases. The most consistent viral link to glioma risk to date has been observed for varicella zoster virus (VZV), an alpha herpesvirus that is neurotropic and causes chickenpox and shingles. Past exposure has been associated with a ~30% reduced glioma risk.^9–12 One proposed mechanism whereby VZV confers protection against glioma is the ability of virusdirected immune responses to cross react with proteins on glial cells, triggering a protective immune response against emerging tumor cells in the brain. Consistent associations between immunity to herpesviruses such as Epstein-Barr virus (EBV) and neurodegenerative, autoimmune diseases including multiple sclerosis^13–15 have also been observed, supporting the possibility that antiherpesvirus responses might reduce survival of malignant brain cells.¹⁶

Cytomegalovirus (CMV), a beta herpesvirus that is tropic for glial cells,¹⁷ gained attention two decades ago in a study that found CMV gene products in 100% of glioma tissue examined but not in normal brain.¹⁸ Subsequent investigations have had mixed results,¹⁷ although a consensus statement was published summarizing data supportive of CMV protein expression in glioma.¹⁹ However, as chemotherapy for glioma (temozolomide) may be immunosuppressive and cause reactivation of latent virus, findings from studies in which CMV was measured after glioma diagnosis could reflect effects of cancer treatment on the pathogen (ie, reverse causality).

To shed further light on the relationship of common herpesviruses with glioma risk, we conducted a prospective evaluation of the association between immune responses (antibody) to herpesviruses and risk of subsequent glioma. Antibodies were measured at least 3 years before glioma diagnosis to minimize effects of preclinical tumor on serologic measures.

2 |. METHODS

Prospective associations of six different human herpesviruses, including herpes simplex virus 1 (HSV1), VZV, EBV, CMV, and human herpes virus 6A (HHV6A) and 6B (HHV6B), with glioma risk were examined in studies nested within two large cohorts, the Janus Serum Bank (Janus) and the Cancer Prevention Study-II (CPS-II). Details on study design have been published.^20,21 In Janus, 323 glioma cases diagnosed ≥5 years after blood collection were included, with one cancer-free control individually matched to each case on birth date within 2 years, sex, date of blood draw within 12 months, and county of residence. In the American Cancer Society-sponsored CPS-II, 37 glioma cases diagnosed ≥3 years after blood collection were identified in cancer-free participants who donated a blood sample between 1998 and 2001. For each case, two cancer-free controls were individually matched to cases on birth date within 1 year, sex, and date of blood draw within 6 months. Gliomas (International Classification of Diseases [ICD]-9:191, ICD10: C71) were classified as GBM (ICD-O-[Oncology]-3 histology 9440, 9441) and lower grade glioma subtypes combined (non-GBM) (ICD-O-3 histology 9382, 9400, 9401, 9410, 9411, 9420, 9424, 9425, 9450, 9451).

A multiplex serological assay was used to measure antibodies against the six human herpesviruses. Briefly, this assay utilized a glutathione S-transferase (GST) capture assay combined with fluorescent bead (Luminex) technology. Antigens of interest were bacterially expressed recombinant fusion proteins with an N-terminal GST domain. Each fusion protein was bound to a spectrally distinct bead set (SeroMap, Luminex, Austin, Texas). A 96-well plate was pre-incubated with blocking buffer, and the fusion protein-bound bead mix was then incubated in those 96-well plates with participant serum at 1:1000 final dilution. Bound antibodies were detected with biotinylated goat-anti-human IgG-secondary antibody and streptavidin-R phycoerythrin reporter conjugate. A Luminex analyzer identified the internal color of the individual beads and to quantify their median fluorescence intensity (MFI). Autofluorescence of each bead set and background reactions resulting from binding of secondary reagents were determined in one serum-free well per plate. Mean background values were then subtracted from raw antibody output to determine antigen-specific MFI reactivity.

For HSV1, VZV, HHV6A, and HHV6B, one antibody marking exposure to the virus was examined (Gg, gE, IE1A, and IE1B, respectively). For EBV, antibody against the latent EBV nuclear antigen [EBNA]-1, a viral capsid protein (viral capsid antigen [VCA]-p18), the early antigen (EA) expressed during viral replication, and the Zebra protein responsible for switching the virus from the latent to lytic phase were examined. For CMV, three different established antigens were examined: pp150, pp52, and pp28. High concordance between antibody measured using multiplexed antigens and gold-standard reference assays has been reported for herpesviruses 1–5, with kappa values ranging 0.86 to 0.96 (median kappa 0.93).²² Lack of gold standard reference assays and global seroprevalence data for HHV6A/B has limited full validation of these assays. All samples from both cohorts were analyzed in a single laboratory run to avoid batch effects.

Odds ratios (ORs) and 95% confidence intervals (CIs) for associations between a detectable antibody response (seropositivity) and glioma risk were obtained from each cohort separately using conditional logistic regression. All models were inherently adjusted for the matching factors. In both cohorts, we examined associations by histologic subgroups (GBM and non-GBM). In Janus only, we explored associations between glioma risk and increasing MFI using a regression model that parameterized antibody response into tertiles (divided distribution into equivalent thirds) among seropositive individuals and generated OR estimates for each increasing tertile of antibody response, compared to the seronegative referent group. We also conducted a latency analysis that examined antibody seropositivity and glioma risk separately for gliomas diagnosed 5–12, 13–19, and ≥20 years after blood draw/antibody measurement. The limited sample size in CPS-II did not support dose-response or exposure lag investigations. Statistical analyses were conducted using R 3.6.2 or SAS 9.4 software.

3 |. RESULTS

Characteristics of both study populations are summarized in Table 1. All participants from Janus, a Norwegian cohort, were of European decent. Seropositivity among cancer-free controls was high for HSV1 (Janus: 75%; CPS-II: 74%) and VZV (Janus: 67%; CPS-II: 74%), whereas positivity to the IE1A and B antigens for HHV6A and B, respectively, was markedly lower (Janus: 35%–46%; CPS-II: 27%–28%). Cancer-free control seropositivity was high for CMV established antigens pp150 (Janus: 72%; CPS-II: 66%), pp52 (Janus: 72%; CPS-II: 66%), and pp28 (Janus: 77%; CPS-II: 69%). Seroprevalence was ubiquitous for EBV early antigen (EA[D]: Janus: 86%; CPS-II: 84%), nuclear antigen (EBNA1: Janus: 88%; CPS-II: 92%), Zebra (Zta: Janus: 91%; CPS-II: 91%), and viral capsid antigen (VCAp18: Janus: 93%; CPS-II: 95%).

TABLE 1.

Descriptive characteristics of Janus and CPS-II study participants

	Janus		CPS-II
	Gliomas	Controls	Gliomas	Controls
Total samples (N)	323	323	37	74
Age at blood draw, mean (SD)	40.4 (7.0)	40.4 (7.0)	69.6 (4.8)	69.7 (4.8)
Female sex, N (%)	104 (32.2)	104 (32.2)	20 (54.1)	40 (54.1)
Year of blood draw, mean (SD)	1983 (7.0)	1983 (7.0)	2000 (0.72)	2000 (0.68)
Seropositive (N, %) by virus
HSV1 (gG)	241 (74.6)	242 (74.9)	27 (73.0)	55 (74.3)
VZV(gE)	196 (60.7)	215 (66.6)	24 (64.9)	55 (74.3)
HHV6A (IE1A)	96 (29.7)	114 (35.3)	6 (16.2)	21 (28.4)
HHV6B (IE1B)	169 (52.3)	150 (46.4)	13 (35.1)	20 (27.0)
EBV (EBNA1)	284 (87.9)	285 (88.2)	31 (83.8)	68 (91.9)
EBV (VCAp18)	283 (87.6)	300 (92.9)	32 (86.5)	70 (94.6)
EBV (EA[D])	249 (77.1)	277 (85.8)	28 (75.7)	62 (83.8)
EBV (Zebra)	297 (92.0)	295 (91.3)	30 (81.1)	67 (90.5)
CMV (pp150)	246 (76.2)	234 (72.4)	26 (70.3)	49 (66.2)
CMV (pp52)	238 (73.7)	234 (72.4)	26 (70.3)	49 (66.2)
CMV (pp28)	261 (80.8)	248 (76.8)	27 (73.0)	51 (68.9)

In Janus and CPS-II, the risk for glioma overall was not related to seroprevalence of HSV1 (ORs: 0.98 [95% CI 0.68–1.41] and 0.94 [95% CI 0.39–2.25], respectively), VZV (ORs: 0.76 [95% CI 0.54–1.06] and 0.64 [95% CI 0.27–1.50]), HHV6A (ORs: 0.77 [95% CI 0.55–1.08] and 0.54 [95% CI 0.21–1.39]), or HHV6B (ORs: 1.27 [95% CI 0.93–1.73] and 1.46 [95% CI 0.63–3.40], Table 2).

TABLE 2.

Associations between herpesvirus seropositivity and glioma risk, by histologic subgroup and cohort^a

Janus	Gliomas overall (N = 323)			Glioblastomas (GBM; N = 196)			Non-GBM (N = 127)
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
HSV1 (gG)	0.98	0.68–1.41	.93	0.83	0.51–1.35	.46	1.22	0.70–2.11	.49
VZV(gE)	0.76	0.54–1.06	.11	0.77	0.5–1.18	.23	0.74	0.43–1.27	.28
HHV6A (IE1A)	0.77	0.55–1.08	.13	0.72	0.46–1.12	.15	0.84	0.50–1.41	.51
HHV6B (IE1B)	1.27	0.93–1.73	.14	1.33	0.89–1.97	.16	1.18	0.71–1.95	.52
EBV (EBNA1)	0.97	0.6–1.58	.90	0.85	0.44–1.62	.62	1.15	0.55–2.42	.71
EBV (VCAp18)	0.55	0.32–0.94	.03	0.55	0.26–1.15	.11	0.56	0.26–1.20	.14
EBV (EA[D])	0.57	0.38–0.85	.01	0.78	0.46–1.32	.36	0.36	0.19–0.70	<.01
EBV (Zebra)	1.08	0.63–1.84	.79	1.23	0.59–2.56	.58	0.92	0.42–2.02	.84
CMV (pp150)	1.24	0.85–1.79	.26	1.00	0.63–1.58	1.00	1.86	0.97–3.56	.06
CMV (pp52)	1.07	0.74–1.54	.71	0.85	0.54–1.34	.49	1.67	0.88–3.16	.12
CMV (pp28)	1.32	0.88–1.98	.18	1.04	0.62–1.74	.90	1.92	0.98–3.76	.06
	Gliomas overall (N = 37)			GBM (N = 27)			Non-GBM (N = 10)
CPS-II	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
HSV1 (gG)	0.94	0.39–2.25	.88	1.00	0.37–2.74	1.00	0.75	0.13–4.49	.75
VZV(gE)	0.64	0.27–1.50	.31	0.92	0.35–2.44	.87	0.15	0.02–1.40	.10
HHV6A (IE1A)	0.54	0.21–1.39	.20	0.77	0.28–2.13	.62	0.00	NA	NA
HHV6B (IE1B)	1.46	0.63–3.40	.38	1.89	0.68–5.28	.22	0.80	0.16–4.12	.79
EBV (EA[D])	0.60	0.22–1.61	.31	0.69	0.21–2.25	.54	0.42	0.07–2.64	.36
EBV (VCAp18)	0.34	0.08–1.47	.15	0.30	0.05–1.69	.17	0.50	0.03–8.0	.62
EBV (Zebra)	0.43	0.13–1.40	.16	1.00	0.23–4.35	1.00	0.00	NA	1.00
EBV (EBNA1)	0.46	0.14–1.55	.21	0.34	0.08–1.47	.15	1.00	0.09–11.0	1.00
CMV (pp150)	1.18	0.53–2.66	.68	1.00	0.40–2.53	1.00	2.00	0.36–11.2	.43
CMV (pp28)	1.20	0.52–2.79	.67	1.18	0.44–3.20	.74	1.24	0.25–6.06	.79
CMV (pp52)	1.18	0.53–2.66	.68	1.00	0.40–2.53	1.00	2.00	0.36–11.2	.43

^aOdds ratios (ORs) and 95% confidence intervals (CIs) for associations between a detectable antibody response (seropositivity) and glioma risk were obtained from each cohort separately using conditional logistic regression. All models were inherently adjusted for the matching factors, including date of birth, sex, and date of blood draw, as well as country of residence in Janus.

Individuals in Janus with evidence of exposure to EBV, defined as being seropositive to either VCAp18 or EBNA1, were less likely to be diagnosed with glioma (OR: 0.75 [95% CI 0.38–1.48]). When considering individual EBV antigens, seropositivity to EA(D) and seropositivity to VCAp18 were each associated with statistically significant, 43% to 45% reductions in risk of glioma (ORs: 0.55 [95% CI 0.32–0.94] and 0.57 [95% CI 0.38–0.85], respectively). Inverse associations for these two EBV proteins were observed for both GBM (ORs: 0.55 and 0.78) and non-GBM tumors (ORs: 0.56 and 0.36), with the EA(D) association reaching statistical significance (P < .01) for non-GBMs. We observed consistent evidence of an association between glioma and both EA(D) and VCAp18 across groups defined by antibody level (Figure 1 and Table S1). Regarding latency, inverse associations between glioma and the EBV EA(D) and VCAp18 antigens were observed for cases diagnosed up to 20 years after blood draw, but not for gliomas diagnosed ≥20 years after blood draw (ORs: 1.00 and 0.88, Table S2). Inverse associations between glioma and the EBV EA(D) and VCAp18 antigens were also observed in the CPS-II study population (ORs: 0.60 [95% CI 0.22–1.61] and 0.34 [95% CI 0.08–1.47]). No statistically significant associations were observed for the EBV Zta or EBNA1 antigens in either Janus or CPS-II.

FIGURE 1.

In Janus only, we explored associations between glioma risk and anti-EBV antibody across strata of MFI using a regression model that parameterized antibody response into tertiles (divided distribution into equivalent thirds) among seropositive individuals and generated OR estimates for each increasing tertiles of antibody response, compared to the seronegative referent group. Points indicate odds ratio (ORs), and error bars indicate 95% confidence intervals (CIs)

Individuals in Janus with evidence of exposure to at least one of the three CMV established antigens (pp150, pp28, pp52) were statistically significantly more likely to be diagnosed with a non-GBM tumor (OR: 2.08 [95% CI 1.07–4.03], P = .03). Seropositivity to pp150 or pp28 was associated with an approximately 2-fold increased risk of non-GBM tumors (ORs: 1.86 [95% CI 0.97–3.56] and 1.92 [95% CI 0.98–3.76], respectively). A positive association was also observed for pp52 among non-GBMs (ORs: 1.67 [95% CI 0.88–3.16]). We observed no evidence of a dose-response trend for CMV pp150, pp52, or pp28 antigens (P-trends = .41, .35, and .23). Regarding latency, elevated non-GBM tumor risk in individuals seropositive for pp150 or pp28 was most pronounced within 5–12 years of blood draw (ORs: 4.0 [95% CI 1.13–14.2, P = .03] and 4.5 [95% CI 0.97–20.8, P = .05], Table S3). Analysis of the 10 non-GBM tumors in CPS-II also indicated an elevation in risk for positivity to at least one of the three CMV established antigens (OR: 2.0 [95% CI 0.36–11.2]).

4 |. CONCLUSION

In this prospective evaluation of adult glioma risk, individuals seropositive to EBV early antigen or viral capsid antigen had a lower likelihood of developing glioma. This inverse association was consistent by histologic subtype and was observed for gliomas diagnosed up to two decades following antibody measurement. In contrast, seropositivity to CMV was associated with an increased risk of non-GBM tumors, an association that was limited to cases diagnosed within ~12 years of antibody measurement. The negative association observed for EBV, and the positive association of CMV with non-GBM tumors, was not observed in the CPS-II cohort, likely due to limited case counts. We observed no statistically significant findings in either cohort for the EBV Zta or EBNA1 antigens, nor for antibody targeting established antigens in HSV1, VZV, or HH6A and B.

The most consistent viral link to glioma to date has been observed for VZV, which causes chickenpox. Past exposure has been associated with a reduced glioma risk in epidemiological studies, including the only prospective study to date of 197 glioma cases and 394 controls in Denmark and Sweden.¹⁰ In that study, individuals in the highest quartile of VZV IgG antibody response were observed to have a 37% lower risk of glioma (P-trend = .03). Although not statistically significant, we also observed an inverse association between glioma and VZV seroprevalence (Janus OR: 0.76 and CPS-II OR: 0.64). The mechanism through which VZV may confer protection against glioma is not established, but one possibility is that VZV antibodies cross react with antigens on glial cells and are thereby able to trigger a protective immune response against emerging, premalignant brain cells.

The present study also suggested an inverse association of glioma with prior exposure to EBV. The only previous prospective study also observed an inverse association between exposure to the EBV EBNA1 protein and glioma (OR: 0.64 [95% CI 0.39–1.06]).¹⁰ Unlike the prior study, we examined antibodies against multiple EBV proteins from various aspects of the viral life cycle, and results suggest statistically significant and novel inverse associations with exposure to proteins expressed as part of EBV viral replication (VCAp18 and EA). EBV is part of the gamma subfamily of herpesviruses and was the first discovered oncogenic virus in humans.^23,24 Relevant to brain cancer, EBV has been consistently associated with increased risk of autoimmune diseases of the central nervous system (CNS) such as multiple sclerosis (MS).^13–15 A proposed biologic mechanism underlying the association between EBV and MS is the “molecular mimicry theory,” whereby immune cells primed by exposure to EBV antigens crossreact to and attack brain cell proteins. In support of this hypothesis, examination of postmortem brain samples from donors with progressive MS documented an abundance of cytotoxic (cell killing) T-cells in CNS tissue that can cross-react with EBV proteins.¹⁶ Like the consistently demonstrated inverse association of VZV with glioma, this antibrain cell response by EBV could potentially contribute to lower glioma risk.

Unlike EBV, the neurotropic pathogen CMV directly infects CNS cell types and is a member of the herpesvirus beta subfamily. CMV gained attention in an early study that found elevated CMV gene product expression in gliomas relative to normal brain tissue.¹⁸ Despite these early findings, subsequent investigations had mixed results,¹⁷ with a consensus statement published in 2012 summarizing data supporting a role for CMV in glioma but emphasizing differences in detection methods across studies.¹⁹ It is important to note that the only previous prospective study also reported that CMV did not follow the pattern of reduced risk observed for VZV and EBV, with a nonsignificantly positive association observed for CMV exposure and glioma risk (OR = 1.08).¹⁰ The remaining studies primarily examined presence of antibodies against CMV in patients already diagnosed with brain cancer; for whom both the presence of the tumor and standard-of-care chemotherapy for glioma (temozolomide) may be immunosuppressive and cause viral reactivation. Our findings are novel in that they suggest a prospective (prediagnostic) association that may be specific to non-GBMs. Few previous studies had power to examine associations according to glioma histologic subtype. Whereas the EBV associations were observed for gliomas diagnosed up to 20 years after blood measurement, exposure to CMV established antigens appeared to only be associated with glioma risk in the 5 to 12 years postblood collection. It is possible that this finding reflects a temporary influence of viral exposure on glioma risk that wanes with increasing time from blood collection. It is also possible that CMV plays a late role in promoting gliomagenesis for individuals with susceptible or premalignant brain cells, in line with the “oncomodulation” effect suggested by Michaelis and colleagues.²⁵ As recent data have demonstrated treatment efficacy of anti-CMV therapies against glioma,²⁶ examination of CMV-directed immune responses in non-GBM patients is warranted.

Our study is the largest prospective evaluation of herpesviruses as potential risk factors for CNS tumors. The prospective study design allowed for an evaluation of the association between exposure to herpesviruses prior to disease and glioma onset. This avoided the reverse causality bias that can occur when the presence of early, possibly undiagnosed, disease changes biomarker levels (eg, anti-EBV antibody) and can thereby induce false associations. To ensure this bias was avoided, glioma cases diagnosed close to the time of blood draw were excluded (within 3 years after blood collection in CPSII-NC and 5 years in Janus). In Janus, gliomas were diagnosed 5 to 35 years following blood collection (median of 15 years), allowing us to examine the association between viral exposure and glioma according to differential exposure lags. Furthermore, similar direction of effect estimates in both cohorts, despite different age profiles, suggests that virus-glioma associations are relevant across the adult lifespan. Our study is not without limitations. The negative association observed for EBV, and the positive association of CMV with non-GBM tumors, was not observed in the CPS-II cohort, likely due to limited case counts. Examination of associations by latency and antibody titer could not be performed in CPS-II due to small numbers, and we were also unable to examine associations across different racial/ethnic groupings since both study populations were primarily of European ancestry. Finally, it must be noted that the magnitude of associations observed are not as pronounced as those observed for other oncogenic viruses. This does not rule about the possibility that immune responses to the studies viruses plays a role in the development of at least a fraction of gliomas but does warrant larger mechanistic studies to evaluate this possibility.

In summary, we report the first evidence of an association between exposure to EBV replication and lower likelihood of glioma development. We further report that exposure to CMV may be related to a higher likelihood of a non-GBM brain tumor diagnosis. Glioma prevention and treatment efforts could be informed by future evaluations of both protective and disease-promoting antiherpesvirus immune responses. Such efforts are urgently needed for glioma, a deadly tumor without effective prevention measures and limited treatment options.

What’s new?

Although glioma is an aggressive cancer, little is known about environmental risk factors that influence its development and that could be leveraged for prevention or treatment. Such risk factors potentially include the herpesviruses cytomegalovirus (CMV) and Epstein-Barr virus (EBV). Here, associations between antibodies against herpesviruses and glioma were investigated in two cohorts, including the Janus Serum Bank. In the Janus cohort, seropositivity to the EBV early antigen and viral capsid antigen were associated with reduced glioma risk, while seropositivity to CMV was associated with increased risk of the nonglioblastoma subtype. The findings offer insight into new opportunities for glioma prevention and treatment.

Abbreviations:

CIs

confidence intervals

CMV

cytomegalovirus

CNS

central nervous system

CPS-II

Cancer Prevention Study-II

EA

early antigen

EBV

Epstein-Barr virus

GBM

glioblastoma

GST

glutathione S-transferase

HHV6A

human herpes virus 6A

HHV6B

human herpes virus 6B

HSV1

herpes simplex virus 1

ICD

International Classification of Diseases

MFI

median fluorescence intensity

MS

multiple sclerosis

ORs

odds ratios

VCA

viral capsid antigen

VZV

varicella zoster virus

Zta

Zebra

DATA AVAILABILITY STATEMENT

The data that support the findings of our study are available on request from the corresponding author.