Immune aberrations in schizophrenia and bipolar disorder have led to the hypotheses that infectious agents or corresponding immune responses might contribute to psychiatric etiopathogeneses. We investigated case–control differences in exposure to the opportunistic fungal pathogen, Candida albicans, and examined associations with cognition, medication, lifestyle, and somatic conditions. We quantified C. albicans IgG antibodies in two cohorts totaling 947 individuals and evaluated odds ratios (OR) of exposure with psychiatric disorder using multivariate regressions. The case–control cohort included 261 with schizophrenia, 270 with bipolar disorder, and 277 non-psychiatric controls; the second included 139 with first-episode schizophrenia, 78 of whom were antipsychotic naive. No differences in C. albicans exposures were found until diagnostic groups were stratified by sex. In males, C. albicans seropositivity conferred increased odds for a schizophrenia diagnosis (OR 2.04–9.53, P⩽0.0001). In females, C. albicans seropositivity conferred increased odds for lower cognitive scores on Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) in schizophrenia (OR 1.12, P⩽0.004), with significant decreases on memory modules for both disorders (P⩽0.0007–0.03). C. albicans IgG levels were not impacted by antipsychotic medications. Gastrointestinal (GI) disturbances were associated with elevated C. albicans in males with schizophrenia and females with bipolar disorder (P⩽0.009–0.02). C. albicans exposure was associated with homelessness in bipolar males (P⩽0.0015). In conclusion, sex-specific C. albicans immune responses were evident in psychiatric disorder subsets. Inquiry regarding C. albicans infection or symptoms may expedite amelioration of this treatable comorbid condition. Yeast exposure as a risk factor for schizophrenia and its associated cognitive and GI effects require further investigation including the possible contribution of gut–brain mechanisms.
Psychiatric disorders may arise from a combination of genetic and environmental factors, and increasingly, immune-related hypotheses are proposed to reconcile both aspects of this interface.1,
Fungal pathogens have not been extensively evaluated in studies of psychiatric disorders. We have previously shown that people with schizophrenia and bipolar disorder have elevated antibodies to Saccharomyces cerevisiae,8,9 a yeast species commonly used by bakers and brewers and not typically considered to be virulent in humans. Studies of the human mycobiome regularly report the presence of S. cerevisiae, but Candida albicans and other Candidal species are the clearly dominant commensal fungus in humans.10,11 C. albicans is a diploid, polymorphic yeast residing in mucosal surfaces of the human respiratory, gastrointestinal (GI) and genitourinary tracts. Under a set of predisposing conditions, C. albicans can become pathogenic ranging in seriousness from local infections of the mouth, throat, and reproductive tract to a systemic invasive candidiasis that impacts the circulatory system, bones and brain.12 Although bacterial dysbioses can contribute to C. albicans overgrowth by failing to provide the competition needed to keep the fungus in check, it is likely that the complexities of the fungal–bacterial, fungal–fungal, and fungal–host relationships are currently underappreciated.13
Here we investigated C. albicans as a new candidate infectious disease target for studies of schizophrenia and bipolar disorder. Blood levels of IgG class antibodies directed against a specific pathogen generally reflect the lifetime exposure of an individual to that antigen.14 Thus, we measured and compared IgG antibodies directed against this fungus in two psychiatric cohorts: one composed of 261 people with schizophrenia, 270 with bipolar disorder and 277 individuals without a history of psychiatric disorder; the other cohort was composed of 139 people with first-episode schizophrenia, 78 of whom were antipsychotic naive. We then determined whether exposure to C. albicans was associated with cognitive symptoms, as evident by scores on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) Form A.15 Finally, to provide a preliminary indication of whether or not C. albicans exposure could be considered a candidate risk factor for these psychiatric disorders, we investigated significant disease associations with outside, potentially confounding, influences. Toward this end, we evaluated our antibody measures in conjunction with a variety of extrinsic factors including antipsychotic medication, homelessness, and somatic comorbidities. Among the somatic conditions tested were cardiovascular, endocrine, GI, genitourinary, hepatic, neoplastic and respiratory disturbances.
First, we examined C. albicans IgG levels in cases compared with controls and found no significant differences among the diagnostic groups (analysis of variance (ANOVA) F=0.93, P⩽0.39). However, when we categorized groups according to sex, we found significant elevations of C. albicans IgG in males with schizophrenia and bipolar disorder compared with male controls (ANOVA F=5.55, P⩽0.0043; data distributions, means, and medians are shown in Figure 1). In multinomial logistic regression models that included the independent variables of age, homelessness, race, and socioeconomic status, we found that irrespective of the positivity cutoff designation used as the dependent variable, elevated C. albicans measures conferred increased odds ratios (OR) to males with schizophrenia compared with control males (Table 1). The elevated C. albicans IgG levels that we identified in bipolar disorder males were found to be a function of a positive homelessness history (positive, n=12, mean±s.d., 0.98±0.55; negative, n=65, mean±s.d., 0.53±0.40; t-tests, t=3.30, two-tailed P⩽0.0015). C. albicans antibody levels in females were not distinguishable between groups, but notably were of the same magnitude as the male schizophrenia and male bipolar disorder groups (Figure 1).
We measured C. albicans antibody levels in a second cohort of 139 individuals with first-episode schizophrenia, 78 of whom were antipsychotic naive. We detected no differences in antibody levels between those who were antipsychotic naive compared with those who were receiving antipsychotic medications (AP+, mean±s.d. 0.60±0.39; AP−, mean±s.d. 0.56±0.33; t-tests, t=0.72, two-tail P⩽0.47). There were also no differences between males in these two medication groups or between females in these two groups (AP+ male, n=30, mean±s.d. 0.52±0.29; AP− male, n=44, mean±s.d. 0.51±0.34; t-tests t=0.07, two-tail P⩽0.94; AP+ female, n=31, mean±s.d. 0.68±0.46; AP− female, n=34, mean±s.d. 0.61±0.31; t-tests, t=0.70, two-tail P⩽0.49). There were trends towards increased antibody levels in females compared to males in both medication groups (t-tests, t=1.36–1.62, P⩽0.06–0.08).
We then tested whether C. albicans positivity was associated with cognitive impairment. We found that females with schizophrenia who were C. albicans IgG-seropositive performed more poorly on these tests than did females with schizophrenia who were C. albicans IgG-seronegative or than female controls (Total RBANS score: multinomial logistic regression, OR 1.12, 95th% CI 1.03–1.23, P⩽0.004). Furthermore, performance on the Immediate and Delayed Memory modules of the RBANS test were significantly reduced in both C. albicans IgG-seropositive women with schizophrenia and those with bipolar disorder compared with respective seronegative groups, with some variation depending on the positivity model (Table 2). Male schizophrenia, male bipolar disorder, male control, and female control groups did not show significant differences in scores on the RBANS test that could be attributable to C. albicans IgG levels.
To determine whether a particular somatic or clinical condition affected the observed patterns of C. albicans antibodies, we compared IgG levels between individuals with psychiatric disorders who were positive and negative for current or chronic conditions in seven categories: cardiovascular, endocrine, genitourinary, GI, hepatological, neoplastic, and respiratory disturbances. Of the categories evaluated, GI, genitourinary and neoplastic conditions were associated with C. albicans levels, again in a sex-specific manner (Table 3). Reported GI symptoms included constipation, Crohn’s disease, diarrhea, gastroesophageal reflux disease, irritable bowel syndrome, lactose intolerance, Norwalk virus, pancreatitis, ulcers, any abdominal surgery, and any abdominal pain. Genitourinary disturbances included urinary tract infections, kidney infections, cysts, urinary incontinence, and sexually transmitted diseases. Neoplastic conditions included pituitary adenoma, melanoma, lung cancer, breast cancer, testicular cancer, endometrial cancer, stomach cancer, liver cancer and ovarian cancer. Significant C. albicans-associated GI symptomatology and genitourinary disturbances affected males with schizophrenia and females with bipolar disorder, although the genitourinary disturbances did not survive Bonferroni multiple comparison corrections. We also observed a trend toward association of C. albicans IgG with neoplastic conditions in males with schizophrenia (Table 3). To determine whether GI disturbances were significantly associated with antibody levels independently of a cancer history or genitourinary source, we reanalyzed the GI association with neoplasm- and genitourinary-positive individuals removed. The positive association of GI symptoms and elevated C. albicans IgG in both groups persisted, although these analyses did not survive Bonferroni multiple comparison corrections (males with schizophrenia, GI negative, n=67, mean±s.d. 0.50±0.36; GI positive, n=52, mean±s.d. 0.64±0.46; t-test, t=−1.83, one-tail P⩽0.03; females with bipolar disorder, GI negative, n=45, mean±s.d. 0.46±0.35; GI positive, n=42, mean±s.d. 0.64±0.52, t-test, t=−1.74, one-tail P⩽0.03).
Our results document that antibodies directed against the opportunistic fungal pathogen, C. albicans, were elevated in distinct subsets of individuals with psychiatric disorders. For some of these individuals, females in particular, C. albicans antibodies were associated with reduced cognitive functioning. Elevated C. albicans could be attributable to extraneous lifestyle or somatic variables for some subgroups, but for others, such as males with schizophrenia, no confounders were evident suggesting the possibility that exposure to this pathogen could be a risk factor for this psychiatric disorder. Sex-specific patterns reflected expectations that yeast overgrowth is found in the female reproductive tract, during immunosuppression and in association with homelessness. C. albicans antibody patterns in both sexes preliminarily implicated the GI tract as a source of microbial dysbioses in both schizophrenia and bipolar disorder. Follow-up studies are required to determine how unregulated C. albicans growth might impact the gut–brain axis in psychiatric disorders.
Many of our observations mirrored epidemiological and sex-specific expectations of heightened exposure to C. albicans in women, in those who were immunocompromised and in people who had a history of homelessness. In this regard, these antibody associations with lifestyle and somatic conditions represented conceptual controls. The portion of our analyses with males showed the most direct and strongest linkage of C. albicans antibodies with schizophrenia, and this association was independent of potential confounders such as age, race, antipsychotic medication, homelessness, socioeconomic status, and a history of cancer. The similarly significant antibody elevation in males with bipolar disorder was found to be purely a function of increased exposures to this pathogen due to a past period of homelessness. In women, we were not able to detect significant differences in strict comparisons of antibody levels among the diagnostic groups. This finding seems logical given that the female reproductive anatomy is prone to yeast infections. Thus, women are likely exposed more often to C. albicans overgrowth regardless of psychiatric status and so any disease-related association might be obscured. Intriguingly, however, it was women who exhibited C. albicans associations with decreased cognition, suggesting that elimination of the overgrowth might help to improve cognition-associated symptoms. We can speculate that this significant association of yeast antibodies with cognition, which is not observed in control women in spite of an equivalent rate of exposure, is consistent with a gene by environmental etiology of psychiatric disorders.16,17 In this scenario, the exposure to an immunopathogenic substance or agent in an individual with a genetically encoded immune system or endothelial barrier defect could be prone to CNS compromise by potentially neurotrophic pathogens or by peripherally generated immune factors.18,19 A role for peripherally acting systems in a CNS disorder was also supported by our findings that antibody levels were higher in people who reported GI conditions compared with those who did not. This GI connection was independent of those variables identified as confounders and preliminarily points toward a GI route of C. albicans immune response generation, perhaps signaling the presence of gut dysbiosis. Our findings, therefore, may be consistent with the hypothesis that disturbance to the gut microbiome in psychiatric disorders can lead to translocation of gut-related products, including those that are derived from fungi, into systemic circulation.18,
These data also serve in part as biological validation of our recently published next-generation sequencing of the oropharyngeal microbiome, where among several dysregulated bacterial species, we detected elevated abundances of the fungal Candida spp., C. dubliensis, in people with schizophrenia compared with controls.21 Fungal material has been found in the cerebrospinal fluid and brains of patients with Alzheimer’s disease and amyotrophic lateral sclerosis.22,
The hypothesis that infectious agents contribute to the etiopathophysiology of psychiatric disorders is most often applied in a context of pathogen exposure or immune activation during critical prenatal neurodevelopmental time windows.6,28,
Several methodological, analytical, and conceptual factors limit the extent to which our results can be interpreted. From a technical standpoint, given that there are over 150 species of Candida, it is possible that the commercial immunoassay kit that we used also detected exposures to other similarly immunogenic Candidal species, including the many that are benign and importantly the few that are pathogenic. In our data analyses, it was not possible for us to correct for all of the multiple variables including lifestyle characteristics that might contribute to our findings. For example, information regarding immunosuppression from sources other than cancer treatment was not available for our study analyses. We addressed potentially confounding basic demographic factors by including these variables in our multiple regressions models. In this manner and as described earlier, we were able to find that a history of homelessness significantly accounted for the elevated C. albicans exposures in males with bipolar disorder. Of note, none of the other demographic factors (age, race, and socioeconomic status) were independently or interactively associated with C. albicans IgG levels. We applied multiple comparison testing for the basic interdiagnostic group analyses; however, applying this correction to the smaller sample size sub-category comparisons would have eliminated detection of informative associations. Therefore, such sub-analyses including the GI associations remain preliminary and support the exploratory nature of this paper. Although our results demonstrate an absence of effect of antipsychotic medications on C. albicans IgG levels, we did not have information regarding over-the-counter or prescription treatments of the somatic conditions evaluated in this study.
In conclusion, our initial results suggested the absence of differential C. albicans IgG levels between psychiatric cases and controls. However, when sexes were evaluated separately, numerous significant disease-specific associations of this fungus with lifestyle, somatic conditions, and cognitive measures were evident. It may be premature to list this pathogen as a risk factor for disease causation, but its status as a comorbidity requires clinical attention. During health care evaluation, inquiry regarding the presence of symptoms of C. albicans infections would help to identify individuals in need of treatment. Furthermore, the early identification of active infection might be an indicator of those at risk for cognitive decline; however, future research is necessary to investigate this connection. In people who have or who are prone to C. albicans overgrowth, dietary management geared toward correcting microbial imbalances may be considered. In the long term, more research is required to understand the mechanisms that trigger pathogenicity of fungal commensals and how this might impact brain function in psychiatric disorders.
Materials and Methods
Cohort 1—Sheppard Pratt Health System, Baltimore, MD, USA
This study cohort was composed of 808 individuals: 277 were control individuals with no history of psychiatric disorder; 261 individuals were diagnosed with schizophrenia; and 270 were diagnosed with bipolar disorder. Diagnoses were made according to criteria defined by DSM-IV-TR46 and have been previously described.47,48 For inclusion in the schizophrenia group, individuals received a DSM-IV-TR diagnosis of schizophrenia, schizophreniform disorder, or schizoaffective disorder. Individuals with bipolar disorder had a DSM-IV-TR diagnosis of type 1 or type 2 bipolar disorder or bipolar disorder not otherwise specified. For both groups, inclusion criteria required an age between 18 and 65 years. Individuals without a history of psychiatric disorder were recruited from posted announcements and were screened to rule out current or past psychiatric disorders with the Structured Clinical Interview for DSM-IV Axis I Disorders Non-Patient Edition.49 Control participants were between the ages of 20 and 60 years, inclusive. Exclusion criteria for all groups included: mental retardation; clinically significant medical disorder that would affect cognitive performance; any history of intravenous substance abuse or a primary diagnosis of substance abuse or substance dependence. For controls, any active substance misuse was considered an exclusion criterion.
At the time of interview, cognitive functioning was assessed with the RBANS Form A.15 Clinical information was gathered, and individuals were asked to report current or chronic conditions related to their health. These conditions could include those that occurred within the last month, were treated with medication, or were chronic. Medical records and self-reports were reviewed for the following categories: cardiovascular, endocrine, genitourinary, GI, hepatological, neoplastic, and respiratory disturbances. Specific conditions falling within each category are reported in the Results section for those categories that exhibited significant case–control differences. Information for these categories was not available for all individuals, but relevant sample sizes are reported in the Results section. Basic demographic data for this study population are shown in Table 4. Maternal education was used as a surrogate for socioeconomic status. Diagnostic groups differed significantly in age, homelessness, race, sex and maternal education. Thus, these variables were all included in the multivariate analyses described below.
These studies were approved by the Institutional Review Boards of the Sheppard Pratt Health System and the Johns Hopkins Medical Institution following established guidelines. All participants provided written informed consent after study procedures were explained. This research was performed in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans.
Cohort 2—University of Cologne, Cologne, Germany
The methods for identifying and characterizing the individuals with a first episode of schizophrenia according to criteria defined by DSM-IV have also been previously described.50 Seventy-eight of these patients were antipsychotic naive and 61 patients were currently receiving antipsychotic medication. Individuals were excluded from the study if they had a relevant comorbidity such as heart disease, liver cirrhosis, known immune-mediated disease (such as multiple sclerosis), or a history of substance dependence. Demographic data regarding age and sex are listed in Table 4. The region from which patients were recruited was generally homogenous regarding socioeconomic characteristics. Informed consent was obtained from all study participants. Protocols for sample collection and analyses were approved by the ethics committee at the University of Cologne, Heidelberg University and Johns Hopkins University, in accordance with the Declaration of Helsinki.
Blood samples were obtained by venipuncture, and plasma and serum separated and assessed for antibodies. Anti-C. albicans IgG levels were measured according to the manufacturer’s protocol using a commercially available kit (Abcam, Cambridge, MA, USA). Each 96-well plate tested contains kit standards as well as study sample replicates for use as internal controls of reproducibility.
Following test procedures, investigators were unblinded for the data analyses. Differences in quantitative levels of antibodies between diagnostic groups were compared using ANOVAs with Sidak post hoc tests. C. albicans IgG seropositivity was designated using three different cutoff values to best capture a range of positivities: (1) absorbance value below which 90% of female control IgG values fall; (2) absorbance value below which 90% of male control IgG values fall; and (3) absorbance value designations assigned by standards of the commercial ELISA kit included for each plate. Higher absorbance values utilized for a cutoff are more conservative than the lower absorbance values that are less rigorous estimates of seropositivity (Figure 1). Odds ratios for disease association of IgG quantitative levels and the three differently calculated IgG seropositivities were assigned using multivariate logistic regressions corrected for age, homelessness, sex, race, and maternal education. Paired comparisons between groups of other continuous variables including age and RBANS scores were also evaluated with t-tests. Multiple comparison testing was evaluated and applied to each analysis depending on the number of comparisons. Results that were significant but fell short of Bonferroni P-value limits are so indicated. Differences in categorical demographic data were evaluated with chi square tests. All the data analyses were performed with groups subdivided according to sex so that sex-specific differences could be identified. For multivariate regressions, sex was dropped as a correction variable when all members of the same sex were evaluated. Statistical analyses were performed with STATA version 12 (STATA Corp LP, College Station, TX, USA).
Availability of the data and materials
Data generated from this study will be made available upon request. Samples may be made available upon request provided that conditions relating to Institutional Review Boards and MTAs at cooperating institutions are met.
We thank Ann Cusic for administrative assistance. Grant support was provided by NIMH P50 Silvio O. Conte Center at Johns Hopkins (grant# MH-94268; E.G.S., K.L.G., R.H.Y.) and by the Stanley Medical Research Institute (all authors).
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/