Exposure to Epstein Barr Virus and Cognitive Functioning in Individuals with Schizophrenia

Abstract

Cognitive deficits are a central feature of schizophrenia whose etiology is not fully understood. Epstein Barr Virus (EBV) is a potentially neurotropic infectious agent that can generate persistent infections with immunomodulatory effects. Previous studies have found an association between EBV antibodies and cognitive functioning in different populations, but there has been limited investigation in schizophrenia. In this study, 84 individuals with schizophrenia were administered a comprehensive neuropsychological battery, the MATRICS Consensus Cognitive Battery (MCCB). Participants also provided a blood sample, from which antibodies to the EBV whole virion and specific proteins were measured. Multivariate models were constructed to determine the association between these antibodies and cognitive performance on the MCCB overall and domain scores. Using these models, we found a significant association between the MCCB overall percent composite score and level of antibodies to the EBV Nuclear Antigen-1 (EBNA-1) protein, the Viral Capsid Antigen (VCA) protein, and the EBV whole virion. A significant association was also found for the MCCB social cognition domain with the level of antibodies to the EBV Nuclear Antigen-1 (EBNA-1) protein, the Viral Capsid Antigen (VCA) protein, and the EBV whole virion. In all cases, a higher level of antibodies was associated with a lower level cognitive performance. These findings suggest that exposure to EBV may contribute to cognitive deficits in schizophrenia, a finding which may have implications for new methods of prevention and treatment.

1. Introduction

Cognitive deficits are a central feature of schizophrenia and contribute to the functional and social disability associated with the disorder. (Dickerson et al., 1996; Green, 1996). In individuals with schizophrenia, cognitive functioning is generally reduced compared to persons in the general population across a wide range of domains including attention, verbal memory, reasoning, and processing speed (Blanchard and Neale, 1994; Heinrichs and Zakzanis, 1998). As a group, individuals with schizophrenia also show deficits in social cognitive abilities, including difficulties in identifying emotions, feeling connected to others, inferring people’s thoughts, and reacting emotionally to others (Green et al., 2015). Like impairments in cognitive functioning, these social cognitive deficits are strong determinants of the degree of impaired daily functioning in such individuals. The determinants of cognitive and social cognitive deficits in persons with schizophrenia are not known with certainty but are likely to involve both genetic and environmental factors.

Epstein Barr virus (EBV), also known as human herpesvirus 4, is a member of the Herpesviridae family. EBV is a lymphotropic virus that can produce latent infections with immunomodulatory effects (Ambinder, 2003; Ambinder and Lin, 2005). Primary infection with EBV is generally associated with self-limited fever and adenopathy in a syndrome commonly referred to as infectious mononucleosis. Following acute infection, the virus can persist in host B and T lymphocytes, monocytes, and epithelial cells. Salivary viral shedding leads to person-to-person transmission (Niederman et al., 1976). EBV can establish latency in many body sites including the brain, where reactivation can be associated with encephalitis (Martelius et al., 2011) and brain specific immune responses (Tzartos et al., 2012). The immune response to EBV infection can be monitored by the measurement of levels of antibodies directed at antigens derived from virions as well as specific EBV proteins. Commonly measured anti-EBV antibodies include those directed at: anti-early antigen (EA) that arises early in the course of infection and decreases after 3–6 months; anti-viral capsid antigen (VCA) that also arises early in infection but persists for extended periods of time; anti-EB nuclear antigen 1 (EBV NA or EBNA-1) that does not arise until late in infection but does persist for extended periods of time (Andersson et al., 1994) (Middeldorp, 2015). The response to additional EBV proteins can be measured by Western blotting techniques to further define the immune response to infection (Dickerson et al., 2019).

The association between exposure to EBV and cognitive functioning has been the subject of some previous investigations and with inconsistent findings. A study of 569 children, average age 11, the offspring of fathers with and without substance use disorder, found a significant association between EBV seropositivity and IQ (Vanyukov et al., 2018). However, in a study of 1000 adolescents from the general population, average age 16, EBV exposure was not associated with levels of cognitive performance (Jonker et al., 2014). Focusing on Alzheimer’s disease, a meta analysis of 3 studies found an association between EBV seropositivity and the diagnosis of Alzheimer’s disease (Steel and Eslick, 2015). However, a more recent and large population-based study did not find an association between EBV antibodies in adults and a subsequent diagnosis of dementia (Torniainen-Holm et al., 2019). In addition, previous studies from our group regarding the association between EBV and cognition did not detect a significant association between EBV seropositivity and cognitive performance in either 521 non-elderly adults (Dickerson et al., 2014) or 229 persons with schizophrenia (Dickerson et al., 2003) using a brief cognitive battery.

It is of note that previous studies examining the association between EBV and cognitive functioning have generally measured EBV by reactivity to a single EBV protein rather than to EBV virions or a panel of EBV proteins. Previous studies from our group found that antibodies to different EBV proteins such as nuclear antigen and viral capsid antigen as well as whole virion proteins are associated with differential risks for schizophrenia and major depressive disorder (Dickerson et al., 2019; Jones-Brando et al., 2020). We thus hypothesized that antibodies to different EBV proteins might be associated with differential levels and domains of cognitive functioning in individuals with schizophrenia. In order to address this hypothesis, we investigated the association between levels of antibodies to EBV proteins and cognitive functioning in a cohort of individuals with schizophrenia using the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB).

2. Methods

2.1. Patient sample

Participants in this study were individuals with schizophrenia or schizoaffective disorder who were enrolled at the Stanley Research Program at Sheppard Pratt, Baltimore, Maryland, USA in one of three clinical trials of adjunctive medications for adults with schizophrenia: A Double-Blind Placebo-Controlled Trial of a Sulforaphane Nutraceutical to Reduce the Symptoms of Schizophrenia (clinicaltrials.gov NCT02810964); A Double-Blind Trial of Adjunctive Valacyclovir to Improve Cognition in Early Phase Schizophrenia (VISTA) (clinicaltrials.gov NCT02008773) (Breier et al., 2019); D-Cycloserine Augmentation of Cognitive Behavioral Therapy for Delusions (clinicaltrials.gov NCT01981759) (Diminich et al., 2020). Participants in the Sulforaphane trial were assessed in the study period February 2017 – May 2019; in the Valacyclovir trial March 2014 – March 2017; and in the D-cycloserine trial September 2015 – February 2017. Each of the studies was approved by the Sheppard Pratt IRB.

Inclusion criteria across the three trials were:

• Minimum age of 18 (Sulforaphane through age 65; Valacyclovir through age 40; D-cycloserine through age 68)

• Diagnosis as confirmed by Structured Clinical Interview for DSM-IV-TR (SCID-I/P Patient Edition) (First, 1996) (Valacyclovir and D-cycloserine) or DSM-5 (SCID-5) (First et al., 2015)(Sulforaphane).

• Proficient in the English language

• Current stable treatment with an anti-psychotic agent (8 weeks SUL, no changes 21 days; Valacyclovir 4 weeks; D-cycloserine 8 weeks)

• Outpatient at time of trial enrollment

Exclusion criteria across the three trials were:

• Any clinically significant or unstable medical disorder

• Pregnant, planning to become pregnant, or breastfeeding during the study period

• Current or recent substance use disorder (Sulforaphane - moderate or severe in past 3 months; Valacyclovir- DSMIV substance dependence within past 3 months; D-cycloserine - abused substances in the past 6 weeks, abused any substance beside marijuana))

• Intellectual disability or IQ below 70

Additional inclusion criteria for the Sulforaphane trial included

• Residual psychotic symptoms of at least moderate severity as evidenced by a Positive and Negative Syndrome Scale (PANSS)(Kay et al., 1987) total score of 60 or higher

Additional inclusion criteria for the Valacyclovir trial included:

• Onset of schizophreniform disorder, schizophrenia, or schizoaffective disorder within the past eight years as defined by first medical records documentation of these conditions.

• CGI-S score of less than or equal to 4 (moderately ill) at randomization and must not have experienced an exacerbation of their illness within 4 weeks prior to randomization leading to an intensification of psychiatric care

Additional exclusion criteria for the Valacyclovir trial included

• Receipt of any disallowed medication per protocol including herbal medications with CNS activity, interferon, antiviral medication.

• Receiving cognitive remediation at the time of study entry

Additional inclusion criteria for the D-cycloserine trial included

• Persistent delusions as defined by a score of at least 3 (moderate) on the Scale for Assessment of Psychotic Symptoms (SAPS) (Andreasen, 1894) global delusion score at two assessments, four weeks apart.

Additional exclusion criteria for the D-cycloserine trial included

• Receipt of clozapine, an SSRI anti-depressant, or medication for tuberculosis

• Current or previous receipt of cognitive behavioral therapy

In cases where a participant had been enrolled in more than one of the clinical trials, data from the first trial with the earliest assessment was used.

2.2. Cognitive Measure

All participants were assessed [by trained raters] with the cognitive battery, Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB). The MCCB includes ten tests that measure seven cognitive domains: 1) Speed of processing (BACS: Symbol Coding; Category Fluency: Animal Naming; Trail Making Test: Part A); 2) Attention/vigilance (CPT-IP); 3) Working memory (WMS®-III: Spatial Span; Letter-Number Span); 4) Verbal learning (HVLT-RTM): 5) Visual learning (BVMT-RTM); 6) Reasoning and problem solving (NAB®: Mazes); 7) Social cognition (MSCEIT TM: Managing Emotions). Raw scores on each test were standardized to age- and gender-corrected percentile scores as previously described (August et al., 2012; Kern et al., 2008).

2.3. Other clinical measures

Participants were assessed with the Positive and Negative Syndrome Scale (PANSS). Participants were also asked if they currently smoked cigarettes. Body mass index was calculated from measured weight and height.

2.4. Antibody measures

All participants had a blood sample drawn using standard venipuncture methods. Plasma was separated from the blood by centrifugation and was stored at −70 degrees C until testing.

IgG antibodies to EBV antigens derived from intact virions were measured by means of solid phase enzyme immunoassay using methods which have been described previously(Jones-Brando et al., 2020). These assays consisted of the binding of purified virions to the solid phase surface of microtiter plates followed by sequential reaction with enzyme labeled anti-human IgG and enzyme substrate. IgG antibodies to EBV-viral capsid antigen (VCA) and EBV Nuclear Antigen-1 (EBNA-1) were measured by solid phase immunoassay employing commercially available microtiter plates and reagents (IBL America, Minneapolis, MN). The procedures used have been previously described (Jones-Brando et al., 2020). In order to allow for the comparison of levels across and antigens and microtiter plates, the values of all of the assays were standardized the calculation of a standard score (mean=1, sd=1) based on the reactivity of control individuals without a psychiatric disorder analyzed on the same microtiter plate. Antibodies to early antigen were not measured since we did not suspect that the participants had recent infection. Reactivity to Herpes Simplex Virus type 1 was measured by solid phase immunoassay as previously described (Dickerson et al., 2003).

In all of the studies, blood samples and cognitive testing were done after participants were enrolled in the trial but before any study medication was administered. In the Sulforaphane trial, the blood sample collection and the cognitive testing occurred at the same visit. In the Valacyclovir and the D-cycloserine trials, blood samples were drawn at the screening visit and the cognitive testing was performed within 30 days.

2.5. Statistical Analyses

Linear regression models were constructed to determine whether levels of antibodies to the EBV whole virion and specific proteins were associated with each MCCB domain score and the overall composite score, after adjusting for the covariate educational level; models were also run adjusting for both race and educational level; age and gender were already adjusted for in the percentile scores. Additional regression models were generated with the covariates of cigarette smoking, race, and serological evidence of HSV-1. The level of statistical significance used for all hypothesis tests was alpha = 0.05/3 = .0166, based on the 3 EBV measures that were used; an alpha of .0166 - .05 was considered suggestive as previously described.(Dickerson et al., 2019)

3. Results

3.1. Sample characteristics

Characteristics of the 84 individuals in the overall sample and those in each of the 3 clinical trials from which they were drawn are shown in Table 1. The 3 groups did not differ significantly in any demographic or clinical characteristic except for race. Performance on the social cognition domain of the MCCB significantly differed among the groups.

Table 1.

Characteristics of the Study Population¹

	D-cycloserine n=19	Valacyclovir n=23	Sulforaphane n=42	Total sample N=84
Demographic variables2
Age, years (p<.01)	41.0 (±11.5)	24.4 (3.8)	40.7 (±11.5)	36.3 (11.9)
Sex male	16 (84.2%)	17 (73.9%)	32 (76.2%)	65 (77%)
Race
Caucasian	10 (52.6%)	5 (21.7%)	23 (54.8%)	38 (45.2%)
African-American	9 (47.4%)	16 (69.6%)	17 (40.4%)	42 (50%)
Other	0	2 (8.7%)	2 (4.8%)	4 (4.8%)
Education category
< High school	7 (36.8%)	6 (26.1%)	14 (33.3%)	27 (32.1%)
High school	6 (31.6%)	14 (60.9%)	19 (45.2%)	39 (46.4%)
>High school	6 (31.6%)	3 (13.0%)	9 (21.4%)	18 (21.4%)
Clinical variables
Diagnosis schizoaffective disorder	12 (63.1%)	9 (39.1%)	21 (50.0%)	42 (50.0%)
PANSS total score	77.05 (±13.39)	74.17 (±12.35)	77.45 (±13.02)	76.46 (±12.85)
Body Mass Index	32.59 (±6.49)	30.66 (±8.36)	33.03 (±6.92)	32.28 (±7.23)
Reactivity to HSV-13	7 (36.8%)	7 (30.4%)	9 (21.4%)
MCCB4 cognitive battery domain score – mean %tile
Overall	3.8 (±12.2)	4.7 (±12.2)	4.6 (±11.9)	4.4 (±11.9)
Speed of processing	11.1 (±14.6)	10.0 (±15.3)	8.5 (±13.3)	9.5 (±14.0)
Attention/vigilance	13.5 (±21.3)	11.4 (±19.8)	10.7 (±17.7)	11.5 (±18.9)
Working memory	12.4 (±15.8)	9.3 (±12.7)	12.1 (±18.4)	11.4 (±16.3)
Verbal learning	19.7 (±23.0)	14.7(±20.4)	14.7 (±17.5)	15.8 (±19.5)
Visual learning	11.1 (±23.3)	14.2 (±23.2)	11.3 (±18.1)	12.0 (±20.7)
Reasoning and problem solving	6.7 (±5.0)	11.7 (±17.9)	15.5 (±17.2)	12.5 (±15.8)
Social cognition	17.4 (±27.0)	38.5 (±36.6)	17.7 (±27.7)	23.3 (±31.3)

¹Mean scores (± s.d.) or frequency (percentage)

²Differences among groups all p>.05 except for race (p=.029) and MCCB Social cognition (p=.028)

³Serologic evidence of exposure

⁴MCCB: Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB)

Within the overall sample, a total of 72 (85.7%) were receiving an atypical antipsychotic medication: 16 (19.1 %) received olanzapine; 15 (17.9 %) clozapine; 10 (11.9 %) ziprasidone; 17 (20.2%) risperidone. A total of 23 (27.4%) received valproate, 34 (40.5%) an anticholingeric, and 37 (44.1%) an antidepressant. The groups did not differ significantly except for clozapine (X²=7.78, p=.02), consistent with the exclusion criteria of the studies.

3.2. Association between Antibodies and Cognitive Performance on the MCCB

As shown in Table 2, results of the linear regression models adjusting for educational level indicate a significant association between the MCCB overall percentile composite score and level of antibodies to the whole virion (coefficient = −3.87, 95% CI −6.47, −1.28, p=.004), the EBNA-1 protein (coefficient = −5.40, 95% CI −7.76, −3.04, p= <..000019) and the VCA protein (coefficient = −3.62, 95% CI −5.84, −1.39, p=.002). In all cases, a higher level of antibodies was associated with lower (worse) cognitive score.

Table 2.

Coefficients of the Association between Epstein Barr Virus Antibody Levels and MCCB Cognitive Scores (N=84)

	VCA2		EBNA-13		Virion
MCCB domain1	Coefficient	p value	Coefficient	p value	Coefficient	p value
Overall	−3.62	.002	−5.40	.000019	−3.87	.004
Speed of processing	−2.25	.108	−2.43	.120	−3.34	.038
Attention/vigilance	−2.23	.241	−3.51	.097	−2.44	.267
Working memory	−2.91	.059	−4.05	.018	−2.66	.136
Verbal learning	−2.80	.156	−3.92	.075	−2.77	.224
Visual learning	−3.15	.131	−2.99	.201	−1.97	.417
Reasoning and problem solving	−2.73	.075	−2.88	.093	−3.22	.068
Social cognition	−11.95	<.001	−8.61	.014	−15.02	<.001

¹Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB) percentile scores; coefficients adjusted for educational level

²EBV Viral Capsid Antigen (VCA)

³EBV Nuclear Antigen-1 (EBNA-1)

In terms of the MCCB domain scores, a significant association was found for the social cognition domain score with level of antibodies to the whole virion (coefficient = −15.02, 95% CI −21.51,−8.52, .p<.00002), the EBNA-1 protein (coefficient = −8.61, 95% CI −15.41, −2.66, p= .014), and the VCA protein (coefficient = −11.95, 95% CI −17.68, −6.22, p<.00009). For the speed of processing domain, a suggestive association was found between the score and antibodies to the whole virion (coefficient = −3.34, 95% CI −6.49, −.186, p=.038. For the working memory domain, a suggestive association was found between the score and antibodies to the EBNA-1 protein (coefficient = −4.05, 95% CI −7.38, −.715, p=.018). In all cases, a higher level of antibodies was associated with lower (worse) cognitive scores.

Regression models were also generated including the additional covariates of cigarette smoking, race, and serological evidence of exposure to Herpes Simplex Virus Type 1(HSV-1). As depicted in Table S1, the results with covariates generated by these models were similar to those generated by the primary model which included educational category as the covariate. Age and gender were not evaluated as covariates since they are included in the calculation of the percentile scores. There were no significant bivariate associations between any single antipsychotic medication or valproate and the MCCB overall percentile score (Table S2).

Associations between the other MCCB scales and level of EBV antibodies were not significant (all p >.05).

4. Discussion

In this study we found that cognitive functioning in individuals with schizophrenia was associated with specific patterns of reactivity to EBV. We found that a higher level of antibodies to the EBV Nuclear Antigen-1 (EBNA-1) protein, the Viral Capsid Antigen (VCA) protein, and the EBV whole virion was associated with lower scores on the MCCB overall composite percentile. We also found that a higher level of antibodies to the EBNA-1 protein, the VCA protein, and the EBV whole virion was associated with lower scores on the MCCB social cognition domain.

We found a suggestive association between a higher level of antibodies to the EBV whole virion and lower scores on the speed of processing domain and also a suggestive association between a higher level of antibodies to the EBV EBNA-1 protein and lower scores on the domain of working memory.

Our results are consistent with some previous studies (Steel and Eslick, 2015; Vanyukov et al., 2018) but not others (Dickerson et al., 2014; Dickerson et al., 2003; Jonker et al., 2014; Torniainen-Holm et al., 2018) in showing an association between antibodies to EBV and cognitive functioning in different populations. In schizophrenia, a previous study from our group did not find an association between antibodies to EBV EBNA-1 and cognitive functioning as measured by a brief cognitive battery (Dickerson et al., 2003). It is of note that most of the previous studies used either relatively brief assessments or the diagnosis of dementia as the measure of cognitive status. By contrast, the MCCB used in the current study is a comprehensive battery that allows for more reliable and valid assessment of overall cognition and specific cognitive domains relevant to schizophrenia, including social cognition and working memory. This may partially explain the difference in results between the current study and our previous study in schizophrenia. The RBANS assessment used in the previous study does not contain measures of social cognition, a domain found to be associated with EBV in the current study (Zhang et al., 2017). Furthermore, the use of a single EBV protein might also have masked associations between different phases of EBV reactivity and cognitive functioning.

The biological basis for the association between cognitive functioning and differential immune response to EBV proteins is not known with certainty. Possible reasons for the association include differences in the timing of first infection, the genetic composition of the infecting virus, and differences in host response to EBV based on both genetic and environmental factors (Kachuri et al., 2020) (Vistarop et al., 2020). It is of note in this regard that these study results are consistent with our previous investigations in which we found a different pattern of association between the level of antibodies to specific EBV proteins and different psychiatric diagnoses. In schizophrenia, we found increased levels of antibodies to the EBV whole virion and to the EBV VCA protein but not to EBV EBNA-1, compared to a control population (Dickerson et al., 2019). In a parallel study of individuals with major depressive disorder, we found reduced levels of EBV EBNA-1 compared with controls (Jones-Brando et al., 2020). A differential response to EBV has also been documented in other brain disorders, such as multiple sclerosis. (Castellazzi et al., 2010). These studies are consistent with the differential immune response to EBV playing a major role in several aspects of brain functioning.

To date there are not effective medication treatments for the cognitive impairments associated with schizophrenia. A previous and definitive trial of valacyclovir, an anti-herpes virus medication, did not find a benefit on cognition in individuals with early course schizophrenia (Breier et al., 2019). It is of note that EBV is relatively insensitive to valacyclovir and its parent compound, acyclovir, as compared to other herpesviruses such as HSV-1 and HSV-2. On the other hand, there are a number of antiviral medications with enhanced activity against EBV that have been developed or are in later stages of development (Bar-Or et al., 2020). These medications may prove to be important tools for the prevention and treatment of EBV-associated cognitive impairments in susceptible populations.

Limitations of the current study include the relatively small sample size and that results from this sample may not be generalizable to other samples of persons with schizophrenia. Also, the participants received different medications during the course of the study. In addition, while we adjusted for demographic variables and educational level in our analyses, we cannot rule out the effects of other variables that were not measured that may affect the association between virus exposure and cognition such as diet and underlying host genetics. Strengths of our study include the measurement of antibodies to specific EBV proteins and also that the cognitive assessment was comprehensive and performed by the same research staff who were blind to the results of the serological results.