Interaction between Cytomegalovirus and Herpes Simplex Virus Type 1 Associated with the Risk of Alzheimer’s Disease Development

Abstract

Background:

Several environmental factors, including infectious agents, have been suggested to cause Alzheimer’s disease (AD). Cytomegalovirus (CMV) has been associated with AD in several recent studies.

Objective:

To investigate whether carriage of CMV, alone or in combination with Herpes simplex virus (HSV), increased the risk of developing AD.

Methods:

Plasma samples from 360 AD cases (75.3% women, mean age 61.2 years), taken an average of 9.6 years before AD diagnosis, and 360 age-, sex-, cohort-, and sampling date matched dementia-free controls were analyzed to detect anti-CMV (immunoglobulin [Ig] G and IgM), group-specific anti-HSV (IgG and IgM), and specific anti-HSV1 and HSV2 IgG antibodies by enzyme-linked immunosorbent assays. AD cases and dementia-free controls were compared using conditional logistic regression analyses.

Results:

The presence of anti-CMV IgG antibodies did not increase the risk of AD (odds ratio [OR], 0.857; p = 0.497). Among AD cases, an association between CMV and HSV1 carriage was detected (OR 7.145, p < 0.001); in a conditional logistic regression model, the interaction between CMV and HSV1 was associated with AD development (OR 5.662; p = 0.007).

Conclusion:

The present findings do not support a direct relationship between CMV infection and the development of AD; however, an interaction between CMV and HSV1 was found to be associated significantly with AD development. These findings suggest that CMV infection facilitates the development of HSV1-associated AD, possibly via its effects on the immune system.

Keywords

Alzheimer’s disease cytomegalovirus dementia Herpes simplex virus nested case-control study

INTRODUCTION

The cause of Alzheimer’s disease (AD), the most common dementia disorder, remains unknown. Several environmental factors, including infectious agents, have been suggested to cause AD, and accumulating evidence points toward herpes simplex virus type 1 (HSV1) in particular [1 –5]. Recent studies have demonstrated that the peptide amyloid-β, the principal constituent of amyloid plaques in AD, is an antimicrobial peptide with effects against a broad range of pathogens, including HSV [6 –9].Amyloid-β accumulates in nerve cells infected by HSV or other pathogens [10, 11], and cerebral infection can trigger the formation of amyloid plaques [6, 12]. This process provides a possible causal link between cerebral infection and AD development.

Cytomegalovirus (CMV), another virus from the Herpesviridae group, has also been associated with AD in several recent studies. Barnes et al. [13] reported that CMV seropositivity was associated significantly with an increased risk of AD development (relative risk, 2.15) in a cohort of 849 participants followed for an average of 5.0 years. In a case-control study, Bu et al. [14] found that CMV seropositivity was associated with AD (adjusted odds ratio [OR], 2.33), and Carbone et al. [15] reported that persons who developed AD over a 5-year period presented with increased anti-CMV immunoglobulin (Ig) G levels at baseline. Three other studies have documented a relationship between CMV, alone or as a component of a composite measure of viral load, and cognitive decline in longitudinal samples [16 –18]. Lurain et al. [19] reported a relationship between higher anti-CMV IgG levels and neurofibrillary tangle density, and Westman et al. [20] reported a decreased proportion of CMV-specific CD8+ cells among CMV-positive subjects with AD compared with controls.

CMV is a T-cell–associated virus; compared with the neurotropic HSV, it seldom produces clinical central nervous system (CNS) manifestations. The direct involvement of CMV in AD pathology in CNS neurons thus seems unlikely [21]. However, CMV infection has profoundly detrimental immunosenescent effects [22], which may facilitate AD development [21]. CMV infection has been associated with increasing HSV antibody levels withage [23].

Taking these findings into account, we investigated the possible relationship between CMV and AD in a large nested case-control study, in which plasma samples were taken an average of 9.6 years before AD diagnosis. As we previously found that HSV carriage was associated with an approximately doubled risk of AD development [5], we also explored interaction effects between HSV and CMV.

METHODS

Participants and procedures

This nested case-control study included 360 individuals with prior AD diagnoses and 360 age-, sex-, cohort-, and sampling date-matched dementia-free controls with stored plasma samples in the Medical Biobank in Umeå (Northern Sweden Health and Disease Study) [24]. Pairs were selected through a computerized method that identified 372 possible pairs. Out of these 372, samples from both the AD case and the control could be analyzed for 360 pairs. The first available sample was used, ranging from October 10, 1986 to March 2, 2006. AD diagnoses were established at the Memory Clinic of the Umeå University Hospital, between August 1, 1995 and March 10, 2010 with a mean interval of 9.6±4.1 years between sampling and diagnosis. AD was diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, criteria [25], supported by typical symptoms of cognitive failure, physical examination, cognitive screening tests, standard blood test, and at least one brain imaging technique. In many cases the diagnoses were further supported by neuropsychological examination and cerebrospinal fluid analysis. An experienced specialist in psychogeriatric medicine thoroughly reviewed and checked the accuracy of medical records of AD cases before final inclusion in the data set. Matched controls were checked against Swedish diagnosis registries, and those with diagnoses of any dementia disorder were excluded. The mean difference in sampling date between AD cases and matched controls was 0.2±53 days and the mean age difference was 0.1±1.0 years. This nested case-control study is part of the Consortium on Health and Ageing: Network of Cohorts in Europe and the United States (CHANCES) project [26]. The study procedures have been described in detail elsewhere [5]. The Regional Ethical Review Board in Umeå, Sweden approved the study (09–190 M).

Plasma analyses

Plasma samples from the 360 AD cases (75.3% women, mean age 61.2 years) and the 360 dementia-free controls were analyzed. Anti-CMV IgG antibodies were detected using an in-house enzyme-linked immunosorbent assay (ELISA), as described previously [27]. Plasma was diluted 1 : 420 and detected with alkaline phosphatase–labeled goat F(ab)2 anti-human IgG (Invitrogen), diluted 1 : 6000. The net absorbance of each sample was compared with that of a validated in-house positive control, and values are presented as arbitrary units (AU; percentage of the positive control). Samples with <5 AU were considered to be negative.

CMV IgM antibodies were essentially analyzed according to the previously published HSV IgM ELISA, utilizing CMV antigen in substitution to HSV [5, 27]. Plasma, diluted 1 : 441, was preincubated with Rf absorbent (Virion\Serion) and detected with peroxidase-labeled goat anti-human F(ab’)2 fragment (Sigma-Aldrich) diluted 1 : 6000. Samples were determined to be positive or negative using thecriterion described above.∥Anti-HSV IgG antibodies were detected by ELISA, as described previously [5]. For positive samples, the presence of specific anti-HSV1 and anti-HSV2 IgG antibodies was subsequently determined using commercial ELISA kits (HerpeSelect 1, HerpeSelect 2; FOCUS Diagnostics), according to the manufacturer’s instructions. All samples were analyzed with respect to anti-HSV IgG; anti-HSV IgG positive samples were then analyzed with respect to HSV2; anti-HSV2 IgG positive samples were then analyzed with respect to HSV1. These analyses enabled determination of whether an individual carried HSV1 and/or HSV2.

Statistical analyses

An overview of statistical tests is presented in Table 1. Conditional logistic regression was used to investigate the impact of CMV infection (anti-CMV IgG positivity, anti-CMV IgG level, and anti-CMV IgM positivity) on AD risk. Analyses were repeated with the data divided into quartiles with regard to follow-up time to investigate the potential influence of this factor on the results. In addition, separate analyses were performed for men and women, and for individuals aged <60 and ≥60 years at the time of plasma sampling. Logistic regression was used to investigate the associations between anti-CMV IgG positivity and anti-HSV IgG positivity among AD cases and controls. Similar analyses were performed for HSV1 and HSV2 separately. The difference between AD cases and controls in these associations was tested by including AD status, anti-CMV IgG positivity and an interaction term in a logistic regression with anti-HSV1 IgG as the independent variable. A post-hoc conditional logistic regression model including the presence of anti-CMV IgG and anti-HSV1 IgG antibodies, and the interaction between them, was used to investigate the combined impact of the viruses on AD risk. Student’s t test and Fisher’s exact test were used to compare anti-HSV IgG levels and anti-HSV IgM presence, respectively, among HSV carriers with and without anti-CMV IgG antibodies (in AD cases and controls, respectively). IBM^® SPSS^® Statistics (version 23 for Macintosh) was used for data handling and calculations. P values <0.05 were regarded as statistically significant.

Table 1

Overview of statistical tests

Test	Predictor	Outcome	Subgroups	Role
Conditional logistic regression	Anti-CMV IgG positivity, anti-CMV IgG level, and anti-CMV IgM positivity, respectively	AD	Quartiles of follow-up time, sex, age < /≥60 years	Hypothesis testing
Logistic regression	Anti-CMV IgG positivity	Anti-HSV IgG positivity, anti-HSV1 IgG positivity, and anti-HSV2 IgG positivity	AD cases/controls	Exploratory
Logistic regression	Interaction term anti-CMV IgG positivity×AD status^a	Anti-HSV1 IgG positivity	–	Exploratory
Conditional logistic regression	Interaction term anti-CMV IgG positivity×anti-HSV1 IgG positivity^b	AD	–	Exploratory
Student’s t test and Fisher’s exact test	Anti-CMV IgG positivity among HSV carriers	Anti-HSV IgG levels and anti-HSV IgM positivity, respectively	AD cases/controls	Exploratory

AD, Alzheimer’s disease; HSV, herpes simplex virus; Ig, immunoglobulin; CMV, cytomegalovirus. ^aAnti-CMV IgG and AD status were also included as predictors in the regression model. ^bAnti-CMV IgG and anti-HSV1 IgG positivity were also included as predictors in the regression model.

RESULTS

Anti-CMV IgG positivity was detected in 312/360 (86.7%) AD cases and 318/360 (88.3%) controls, as shown in Table 2 presenting baseline characteristics according to AD case or control group. The presence of anti-CMV IgG antibodies did not increase the risk of AD (OR, 0.857; 95% confidence interval, 0.549–1.338; p = 0.497). Equal proportions of individuals in both groups (13/360, 3.6%) were anti-CMV IgM positive, and no difference in the mean anti-CMV IgG antibody level was detected. No significant effect of sex or age was observed (data not shown). The interval between plasma sampling and diagnosis did not affect the relationship between the presence of anti-CMV antibodies and AD.

Table 2

Characteristics of Alzheimer’s disease cases and dementia-free controls, at the time of plasma sampling^a

	Alzheimer’s disease cases, n = 360	Dementia-free controls, n = 360
Age, y, mean±SD	61.2±5.6	61.2±5.6
>60 y, n (%)	228 (63)	228 (63)
Male sex, n (%)	89 (25)	89 (25)
Anti-HSV IgG+, n (%)	338 (94)	324 (90)
Anti-HSV IgM+, n (%)	27 (8)	20 (6)
Anti-HSV1 IgG+, n (%)	329 (92)	317 (88)
Anti-HSV2 IgG+, n (%)	52 (14)	46 (13)
Anti-CMV IgG+, n (%)	312 (87)	318 (88)
Anti-CMV IgM+, n (%)	13 (4)	13 (4)
Anti-HSV1 IgG+ and anti-CMV IgG+, n (%)	295 (82)	282 (78)
Anti-HSV1 IgG- and anti-CMV IgG-, n (%)	14 (4)	7 (2)

HSV, herpes simplex virus; Ig, immunoglobulin; CMV, cytomegalovirus. ^aIg-prevalence was analyzed in plasma samples using enzyme-linked immunosorbent assays. Samples with <5 arbitrary units were considered negative and ≥5 positive.

Anti-HSV IgG antibodies were detected in 338/360 (93.9%) AD cases and 324/360 (90.0%) controls (Table 2). Among AD cases, the presence of anti-HSV IgG antibodies was associated with the presence of anti-CMV IgG antibodies (11/22 [50.0%] and 301/338 [89.1%] anti-HSV IgG–negative and –positive persons, respectively, were concomitantly anti-CMV IgG positive; OR 8.135; 95% confidence interval, 3.298–20.065; p < 0.001). No such association was observed among controls (29/36 [80.6%] and 289/324 [89.2%] anti-HSV IgG–negative and –positive persons, respectively, were concomitantly anti-CMV IgG positive; OR 1.993; 95% confidence interval, 0.813–4.887; p = 0.132). A significant association was found between anti-HSV1 IgG and anti-CMV IgG positivity among AD cases (54.8% and 89.7% of anti-HSV1–negative and –positive persons, respectively, were concomitantly anti-CMV positive; OR 7.145; 95% confidence interval, 3.238–15.767; p < 0.001), but not among controls (83.7% and 89.0% of anti-HSV1–negative and –positive persons, respectively, were concomitantly anti-CMV positive; OR 1.567; 95% confidence interval, 0.648–3.787; p = 0.319). The difference between AD cases and controls in the association between anti-HSV1 IgG and anti-CMV IgG positivity was significant when tested using logistic regression, p = 0.012. No significant association betweenpresence of anti-HSV2 and anti-CMV antibodies was found among cases or controls (OR 1.211; 95% confidence interval, 0.487–3.010; p = 0.681 and OR 2.033; 95% confidence interval, 0.602–6.868; p = 0.253, respectively). Anti-HSV and anti-CMV IgG level among those carrying both viruses showed a tendency towards a weak correlation among both AD cases and controls (β 0.112, p = 0.052 and β 0.103, p = 0.082 respectively; p for difference between AD cases and controls 0.890).

The prevalence of anti-HSV1 IgG and anti-CMV IgG positivity and their overlap among AD cases and controls is shown in Fig. 1. Among AD cases, 295/360 (82%) were concomitantly anti-HSV1 IgG and anti-CMV IgG positive, compared with 282/360 (78%) among controls (Table 2). A conditional logistic regression model demonstrated significant interaction between CMV and HSV1 carriage in the risk of AD development (OR, 5.662; p = 0.007), as shown in Table 3. The ORs for anti-HSV1 IgG and anti-CMV IgG positivity in this model were 0.424 (p = 0.111) and 0.195 (p = 0.005), respectively.

Fig.1

Seroprevalence of herpes simplex virus type 1 (HSV1) IgG and cytomegalovirus (CMV) IgG and their overlap among Alzheimer’s disease (AD) cases and controls.

Table 3

Interaction between cytomegalovirus and herpes simplex virus type 1 carriage in the risk of Alzheimer’s disease development^a

	Odds ratio	95% confidence interval	p
Anti-CMV IgG positivity	0.195	0.062–0.619	0.005
Anti-HSV1 IgG positivity	0.424	0.147–1.218	0.111
Anti-CMV IgG positivity×anti-HSV1	5.662	1.606–19.96	0.007
IgG positivity

CMV, cytomegalovirus; Ig, immunoglobulin; HSV1, herpes simplex virus type 1. ^aAnalyzed using conditional logistic regression.

Anti-HSV IgG level and the prevalence of anti-HSV IgM antibodies among HSV carriers in the case and control groups did not differ between those who were and were not also anti-CMV IgG positive (data not shown).

DISCUSSION

CMV carriage was not associated with an increased risk of AD development in the present study, but a strong interaction between CMV and HSV1 was found to be associated with this outcome. The present results suggest that CMV, possibly through immune system–modulating effects, facilitates HSV1-associated AD development. This CMV–HSV1 interaction may have contributed to the results of previous studies indicating an association between CMV and AD or cognitive decline [13 –18], and is supported by previous studies of CMV-associated immune alterations in AD and in relation to the HSV immune response [20, 23].

The strengths of the present study include the examination of data from large numbers of well-defined preclinical AD cases and closely matched dementia-free controls, and the lengthy intervals between plasma sampling and diagnosis. Thus, the effects measured are likely associated with early disease processes. In addition, generalizability of the results to similar populations should be good.

A limitation of the present study is that APOE genotype was not determined, due to unavailability of genetic material. APOE genotype is strongly correlated with AD, and can affect the course of HSV1 and CMV infection [1 , 29]. However, there is no indication that APOE genotype affect the risk of HSV1 or CMV carriage [4], and consequently, the risk of confounding effects should not be exaggerated. Clinical data at the time of plasma sampling and information on lifestyle factors was also unavailable. The conditional logistic regression model indicated a strong and significant interaction effect between HSV1 and CMV on AD risk. However, we could not determine the exact impacts of carriage of the two viruses alone and in combination in this analysis. The ORs for HSV1 and CMV alone were much less than 1 (significantly so, for CMV) indicating a lower AD risk in individuals infected by only one of the two viruses than in non-carriers. This result could be regarded as a statistical artifact caused by the high seroprevalence of the two viruses and the significant correlation between them among those who developed AD, resulting in a very small number of singly positive individuals especially in the AD group.

The interaction between HSV1 and CMV reflects early processes that may contribute to subsequent clinical AD development. The results of this study are in agreement with the accumulated evidence of infectious causes of AD, especially the role of HSV1[1, 2]. Further epidemiological studies are needed to confirm and further elucidate the interplay between CMV and HSV1 in AD development. The accumulating evidence for a pivotal role of HSV1 in AD development indicates a potential target for drug treatment, although clinical studies are needed to confirm apositive effect of anti-viral drug treatment on AD.

Conclusion

The present data do not support a direct relationship between CMV infection and the development of AD, however, an interaction between CMV and HSV1 was found to be associated significantly with AD development. These findings suggest that CMV infection facilitates the development of HSV1-associated AD, possibly via its effects on the immune system.

Footnotes

ACKNOWLEDGMENTS

This study was conducted in the context of the CHANCES project, funded in the FP7 framework program of DG-RESEARCH in the European Commission. The Hellenic Health Foundation, Greece, coordinates the project. The study was furthersupported financially by grants from the Västerbotten County Council, Kempe Foundations, Swedish Medical Association, Swedish Dementia Association, Trolle-Wachtmeister Foundation, Dementia Fund in Västerbotten, Swedish Alzheimer Fund, Stohne Foundation, Bergvall Foundation, and Umeå University Foundation for Medical Research.

Authors’ disclosures available online ().

References

Steel

, Eslick

(2015) Herpes viruses increase the risk of Alzheimer’s disease: A meta-analysis. J Alzheimers Dis 47, 351–364.

Itzhaki

, Lathe

, Balin

, Ball

, Bearer

, Braak

, Bullido

, Carter

, Clerici

, Cosby

, Del Tredici

, Field

, Fulop

, Grassi

, Griffin

, Haas

, Hudson

, Kamer

, Kell

, Licastro

, Letenneur

, Lovheim

, Mancuso

, Miklossy

, Otth

, Palamara

, Perry

, Preston

, Pretorius

, Strandberg

, Tabet

, Taylor-Robinson

, Whittum-Hudson

(2016) Microbes and Alzheimer’s disease. J Alzheimers Dis 51, 979–984.

Letenneur

, Peres

, Fleury

, Garrigue

, Barberger-Gateau

, Helmer

, Orgogozo

, Gauthier

, Dartigues

(2008) Seropositivity to herpes simplex virus antibodies and risk of Alzheimer’s disease: A population-based cohort study. PLoS One 3, e3637.

Lovheim

, Gilthorpe

, Adolfsson

, Nilsson

, Elgh

(2015) Reactivated herpes simplex infection increases the risk of Alzheimer’s disease. Alzheimers Dement 11, 593–599.

Lovheim

, Gilthorpe

, Johansson

, Eriksson

, Hallmans

, Elgh

(2015) Herpes simplex infection and the risk of Alzheimer’s disease: A nested case-control study. Alzheimers Dement 11, 587–592.

Kumar

, Choi

, Washicosky

, Eimer

, Tucker

, Ghofrani

, Lefkowitz

, McColl

, Goldstein

, Tanzi

, Moir

(2016) Amyloid-beta peptide protects against microbial infection in mouse and worm models of Alzheimer’s disease. Sci Transl Med 8, 340ra372.

Bourgade

, Garneau

, Giroux

, Le Page

, Bocti

, Dupuis

, Frost

, Fulop

Jr (2015) beta-Amyloid peptides display protective activity against the human Alzheimer’s disease-associated herpes simplex virus-1. Biogerontology 16, 85–98.

Soscia

, Kirby

, Washicosky

, Tucker

, Ingelsson

, Hyman

, Burton

, Goldstein

, Duong

, Tanzi

, Moir

(2010) The Alzheimer’s disease-associated amyloid beta-protein is an antimicrobial peptide. PLoS One 5, e9505.

White

, Kandel

, Tripathi

, Condon

, Qi

, Taubenberger

, Hartshorn

(2014) Alzheimer’s associated beta-amyloid protein inhibits influenza A virus and modulates viral interactions with phagocytes. PLoS One 9, e101364.

10.

Santana

, Recuero

, Bullido

, Valdivieso

, Aldudo

(2012) Herpes simplex virus type I induces the accumulation of intracellular beta-amyloid in autophagic compartments and the inhibition of the non-amyloidogenic pathway in human neuroblastoma cells. Neurobiol Aging 33, 430e419–433.

11.

Miklossy

, Kis

, Radenovic

, Miller

, Forro

, Martins

, Reiss

, Darbinian

, Darekar

, Mihaly

, Khalili

(2006) Beta-amyloid deposition and Alzheimer’s type changes induced by Borrelia spirochetes. Neurobiol Aging 27, 228–236.

12.

Little

, Joyce

, Hammond

, Matta

, Cahn

, Appelt

, Balin

(2014) Detection of bacterial antigens and Alzheimer’s disease-like pathology in the central nervous system of BALB/c mice following intranasal infection with a laboratory isolate of Chlamydia pneumoniae. Front Aging Neurosci 6, 304.

13.

Barnes

, Capuano

, Aiello

, Turner

, Yolken

, Torrey

, Bennett

(2015) Cytomegalovirus infection and risk of Alzheimer disease in older black and white individuals. J Infect Dis 211, 230–237.

14.

, Yao

, Jiao

, Zeng

, Liu

, Xiang

, Liang

, Wang

, Cao

, Yi

, Deng

, Liu

, Xu

, Zhang

, Gao

, Xu

, Zhang

, Wang

, Tan

, Xu

, Zhou

, Wang

(2015) A study on the association between infectious burden and Alzheimer’s disease. Eur J Neurol 22, 1519–1525.

15.

Carbone

, Lazzarotto

, Ianni

, Porcellini

, Forti

, Masliah

, Gabrielli

, Licastro

(2014) Herpes virus in Alzheimer’s disease: Relation to progression of the disease. Neurobiol Aging 35, 122–129.

16.

Nimgaonkar

, Yolken

, Wang

, Chung-Chou

, McClain

, McDade

, Snitz

, Ganguli

(2016) Temporal cognitive decline associated with exposure to infectious agents in a population-based, aging cohort. Alzheimer Dis Assoc Disord 30, 216–222.

17.

Aiello

, Haan

, Blythe

, Moore

, Gonzalez

, Jagust

(2006) The influence of latent viral infection on rate of cognitive decline over 4 years. J Am Geriatr Soc 54, 1046–1054.

18.

Strandberg

, Pitkala

, Linnavuori

, Tilvis

(2003) Impact of viral and bacterial burden on cognitive impairment in elderly persons with cardiovascular diseases. Stroke 34, 2126–2131.

19.

Lurain

, Hanson

, Martinson

, Leurgans

, Landay

, Bennett

, Schneider

(2013) Virological and immunological characteristics of human cytomegalovirus infection associated with Alzheimer disease. J Infect Dis 208, 564–572.

20.

Westman

, Lidehall

, Magnusson

, Ingelsson

, Kilander

, Lannfelt

, Korsgren

, Eriksson

(2013) Decreased proportion of cytomegalovirus specific CD8 T-cells but no signs of general immunosenescence in Alzheimer’s disease. PLoS One 8, e77921.

21.

Itzhaki

, Klapper

(2014) Cytomegalovirus: An improbable cause of Alzheimer disease. J Infect Dis 209, 972–973.

22.

Soderberg-Naucler

, Fornara

, Rahbar

(2016) Cytomegalovirus driven immunosenescence-An immune phenotype with or without clinical impact? Mech Ageing Dev 158, 3–13.

23.

Stowe

, Peek

, Cutchin

, Goodwin

(2012) Reactivation of herpes simplex virus type 1 is associated with cytomegalovirus and age. J Med Virol 84, 1797–1802.

24.

Hallmans

, Agren

, Johansson

, Stegmayr

, Jansson

, Lindahl

, Rolandsson

, Soderberg

, Nilsson

, Johansson

, Weinehall

(2003) Cardiovascular disease and diabetes in the Northern Sweden Health and Disease Study Cohort - evaluation of risk factors and their interactions. Scand J Public Health Suppl 61, 18–24.

25.

American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision, American Psychiatric Association, Washington, DC.

26.

Boffetta

, Bobak

, Borsch-Supan

, Brenner

, Eriksson

, Grodstein

, Jansen

, Jenab

, Juerges

, Kampman

, Kee

, Kuulasmaa

, Park

, Tjonneland

, van Duijn

, Wilsgaard

, Wolk

, Trichopoulos

, Bamia

, Trichopoulou

(2014) The Consortium on Health and Ageing: Network of Cohorts in Europe and the United States (CHANCES) project–design, population and data harmonization of a large-scale, international study. Eur J Epidemiol 29, 929–936.

27.

Sjostrom

, Hjalmars

, Juto

, Wadell

, Hallmans

, Tjonneland

, Halkjaer

, Manjer

, Almquist

, Melin

(2011) Human immunoglobulin G levels of viruses and associated glioma risk. Cancer Causes Control 22, 1259–1266.

28.

Aiello

, Nguyen

, Haan

(2008) C-reactive protein mediates the effect of apolipoprotein E on cytomegalovirus infection. J Infect Dis 197, 34–41.

29.

Burgos

, Ramirez

, Sastre

, Valdivieso

(2007) Apolipoprotein E genotype influences vertical transmission of herpes simplex virus type 1 in a gender specific manner. Aging Cell 6, 841–842.