The interleukin-6/interleukin-23/T helper 17-axis as a driver of neuro-immune toxicity in the major neurocognitive psychosis or deficit schizophrenia: A precision nomothetic psychiatry analysis

Abstract

Background

Schizophrenia and especially deficit schizophrenia (DSCZ) are characterized by increased activity of neuroimmunotoxic pathways and a generalized cognitive decline (G-CoDe). There is no data on whether the interleukin (IL)-6/IL-23/T helper 17 (IL-6/IL-23/Th17)-axis is more associated with DSCZ than with non-deficit schizophrenia (NDSCZ) and whether changes in this axis are associated with the G-CoDe and the phenome (a factor extracted from all symptom domains) of schizophrenia.

Methods

This study included 45 DSCZ and 45 NDSCZ patients and 40 controls and delineated whether the IL-6/IL-23/Th17 axis, trace elements (copper, zinc) and ions (magnesium, calcium) are associated with DSCZ, the G-CoDe and the schizophrenia phenome.

Results

Increased plasma IL-23 and IL-6 levels were associated with Th17 upregulation, assessed as a latent vector (LV) extracted from IL-17, IL-21, IL-22, and TNF-α. The IL-6/IL-23/Th17-axis score, as assessed by an LV extracted from IL-23, IL-6, and the Th17 LV, was significantly higher in DSCZ than in NDSCZ and controls. We discovered that 70.7% of the variance in the phenome was explained by the IL-6/IL-23/Th17-axis (positively) and the G-CoDe and IL-10 (both inversely); and that 54.6% of the variance in the G-CoDe was explained by the IL-6/IL-23/Th17 scores (inversely) and magnesium, copper, calcium, and zinc (all positively).

Conclusion

The pathogenic IL-6/IL-23/Th17-axis contributes to the generalized neurocognitive deficit and the phenome of schizophrenia, especially that of DSCZ, due to its key role in peripheral inflammation and neuroinflammation and its consequent immunotoxic effects on neuronal circuits. These clinical impairments are more prominent in subjects with lowered IL-10, magnesium, calcium, and zinc.

Introduction

Schizophrenia is a severe chronic neuropsychiatric disorder with a heterogeneous genetic and neurobiological background that affects about 0.75% of the world population [1]. Recent research has found new phenotypes that are qualitatively different, such as major neurocognitive psychosis (MNP, formerly called deficit schizophrenia) and simple neurocognitive psychosis (SNP), non-remitters to treatment (NRTT) and partial remitters to treatment (PRTT), and first-episode schizophrenia (FES) and multiple-episode schizophrenia (MES), both with and without worsening [2]. About a quarter of people with schizophrenia have MNP, which is marked by more severe cognitive problems and symptoms [3, 4].

MNP patients show more severe neurocognitive impairments than SNP patients in almost all cognitive domains including sustained visual attention, working memory, strategy use, rule acquisition, attention set-switching, emotional recognition, semantic memory, episodic memory, and delayed recall and recognition [5]. Moreover, one common core (or latent construct) may be extracted from all these cognitive dysfunctions, named generalized cognitive decline (G-CoDe) [2, 5]. MNP is characterized by intertwined increases in the severity of various symptom domains, including psychosis, hostility, excitation, mannerism, and negative (PHEMN) symptoms, psychomotor retardation (PMR), and formal thought disorders (FTD) [6–8]. Again, a common core underpins all seven symptom domains, conceptualized as the phenome of schizophrenia [9].

There is now evidence that schizophrenia, MNP, and FEP/FES are neuroimmune disorders. The first comprehensive theory of schizophrenia, known as the macrophage-T-lymphocyte theory, was published in the 1990s [10]. This theory considers that early neurodevelopmental disorders, as a consequence of intra-uterine infections, may predispose individuals to later injuries causing activation of immune-inflammatory and nitro-oxidative stress pathways. Since then, a growing amount of data has shown that immune-inflammatory processes in the peripheral blood and central nervous system (CNS) have a role in the onset of schizophrenia, primarily through systemic inflammation translating into microglial activation [11–19]. Moreover, schizophrenia is not only characterized by activation of the immune-inflammatory response system (IRS), due to activated M1 macrophages, as well as T helper (Th)1 and Th17 cells, but also by activation of the compensatory immune-regulatory system (CIRS) which tends to downregulate the primary IRS thereby preventing hyperinflammation [16]. Key players in the CIRS are negative immunoregulatory cytokines produced by Th2 (IL-4) and T regulatory (Treg) (IL-10) cells [16]. Importantly, relative deficits in the CIRS are associated with a worse outcome of FEP [20] and with increased neurocognitive impairments and worsening in FEP and MEP [21]. Moreover, increased neurotoxicity caused by increased levels of toxic cytokines (e.g., IL-1β, IL-6, IL-8, TNF-α, and IFN-γ) and chemokines (e.g., CCL11, CCL2, CXCL8, and CXCL10) is significantly associated with the G-CoDe and the schizophrenia phenome [5]. MNP is characterized by highly significant increases in neuroimmunotoxic pathways, attenuated CIRS protection, as well as increased phenome and G-CoDe scores [22–24].

Increasing evidence now suggests that Th17 cells are involved in the immunopathogenesis of schizophrenia [25], and a possible role of the IL-23/IL-17 pathway has been described [26–28]. In addition, proinflammatory cytokines produced by pathogenic Th17 cells, such as IL-17 and IL-21 [29], have a significant role in the onset and progression of schizophrenia [25, 30, 31]. IL-23, IL-1β, and IL-6, which are produced by macrophages and dendritic cells, mediate the production of Th1, Th2, and Th17 cytokines [32, 33]. IL-23, a proinflammatory cytokine from the IL-12 family, is essential for the IL-17-mediated immune response as well as the survival and growth of pathogenic Th17 cells [34]. IL-23 stimulates nave T cells to acquire the pathogenic Th17 phenotype (producing IL-17, IL-21, IL-22, and GM-CSF), whereas without IL-23 (stimulated by IL-6 and TGF-β), more homeostatic or non-pathogenic Th17 cells are acquired (producing IL-17, IL-21, and IL-10) [35–37]. Increased production of IL-23 also drives TNF-α, which together with IL-23 and IL-17, shapes the TNF/IL-23/IL-17 axis [38]. The latter plays a key role in inflammation and (auto)immune disorders, including rheumatoid arthritis, inflammatory bowel disease, multiple sclerosis, and type 1 diabetes mellitus [36]. In addition, IL-6 also induces the production of IL-17 [39] and IL-21 in activated T cells [29]. In turn, IL-21 induces Th17 cell differentiation in the presence of TGF-β1, and Th17 cells produce IL-21, which plays an autocrine role in Th17 cell development [29, 40].

There are also some reports that trace elements such as copper and zinc and ions such as calcium and magnesium, as well as serum albumin, may play a role in schizophrenia. Thus, serum copper is elevated in schizophrenia patients [41], while lower zinc is shown in a recent meta-analysis [42]. Patients with schizophrenia often have lowered serum and intracellular magnesium levels, while magnesium treatment may improve psychosis [43]. There is also some evidence that aberrations in calcium signaling are associated with neurocognitive impairments [44] and that lowered albumin may contribute to the severity of the phenome of schizophrenia [45].

Nevertheless, there is no data on whether the IL-6, IL-23, and Th17 (IL-17, TNF-α, IL-21 and IL-22) axis is more associated with MNP than with SNP and whether changes in this axis are associated with the G-CoDe and the phenome of schizophrenia. Hence, this study was conducted to delineate a) whether the IL-6/IL-23/Th17 axis is activated in MNP and is associated with impairments in the G-CoDe and the phenome; b) the associations between IL-6 and IL-23 and Th17 effector cytokines; and c) the role of IL-10, copper, zinc, calcium and magnesium in the neuro-immune pathophysiology of MNP.

Subjects and methods

Participants

Ninety patients with schizophrenia and forty healthy controls were included in this study. Patients were recruited between February and June 2021 at "The Psychiatry Unit," Al-Hakeem General Hospital, Najaf Governorate, Iraq. All patients were stable for at least 12 weeks. Family and friends of personnel or friends of patients served as controls. Patients and controls were excluded if they had ever taken immunosuppressive treatments or glucocorticoids or if they had been diagnosed with a neurodegenerative or neuroinflammatory illness such as Alzheimer’s disease, Parkinson’s disease, multiple sclerosis, or stroke. Additionally, individuals with (auto)immune diseases such as inflammatory bowel disease, rheumatoid arthritis, COPD, psoriasis, or diabetes mellitus were excluded. None of the controls had a present or lifetime DSM-IV-TR axis I diagnosis or a family history of schizophrenia or psychosis. Patients having axis-1 DSM-IV-TR disorders such as bipolar disorder, major depression, schizoaffective disorder, obsessive-compulsive disorder, psycho-organic illnesses, or substance use disorders were excluded. CRP levels were < 6 mg/L in all subjects, indicating that there was no overt inflammation.

The study followed Iraqi and international privacy and ethics laws. Before participating in this study, all participants and first-degree relatives of participants with schizophrenia gave written informed consent (legal representatives are mother, father, brother, spouse, or son). The study was approved by the ethics committee (IRB) of the College of Science, University of Kufa, Iraq (82/2020), which follows the Declaration of Helsinki’s International Guideline for Human Research Protection.

Clinical assessments

To collect patient and control data, a senior psychiatrist with expertise in schizophrenia conducted semi-structured interviews. Using DSM-IV-TR diagnostic criteria, the Mini-International Neuropsychiatric Interview (M.I.N.I.) was used to diagnose schizophrenia. The same psychiatrist evaluated the Schedule for the Deficit Syndrome (SDS) [46] for the diagnosis of primary deficit schizophrenia, as well as the Positive and Negative Syndrome Scale (PANSS), the Scale for the Assessments of Negative Symptoms (SANS) [47], the Brief Psychiatric Rating Scale (BPRS) [48], and the Hamilton Depression (HAM-D) and Anxiety (HAM-A) rating scales [49, 50]. We calculated z-unit weighted composite scores reflecting psychosis, hostility, excitation, mannerism, PMR and FTD symptom domains as previously reported using BPRS, HAM-D, HAM-A, and PANSS items and the total SANS score was used to assess negative symptoms [7, 8, 23, 51]. Based on the 6 computed z composite scores reflecting PHEM, negative, PMR, and FTD scores, we then extracted the first factor from these data, named the “phenome”. On the same day, a research psychologist performed neuropsychological probes using the Brief Assessment of Cognition in Schizophrenia (BACS) [52] while remaining blind to the clinical diagnosis. The latter battery consists of the Digit Sequencing Task (which measures working memory), the List Learning test (which tests verbal episodic memory), Controlled Word Association (which assesses letter fluency), Category Instances (which measures semantic fluency), the Tower of London (to probe executive functions), Symbol Coding (which probes attention), and the token motor task. The total score was computed. Based on our previous findings [5] that one latent vector underpins all cognitive tests, we here extract the first factor from the 7 BACS test, and named this factor the “G-CoDe” (generalized cognitive decline). Tobacco use disorder (TUD) was diagnosed in accordance with DSM-IV-TR criteria. The following formula was used to compute the body mass index (BMI): body weight (kg) / length (m²).

Assays

Each subject’s fasting venous blood was obtained in the early morning hours. After 15 minutes at room temperature, the blood was allowed to coagulate for 10 minutes before being centrifuged at 3000 rpm for 10 minutes. Separated serum was then transferred to Eppendorf tubes and stored at -80°C until analysis. The concentrations of CRP in serum were measured using a kit supplied by Spinreact^® (Barcelona, Spain). The test is based on the notion of latex agglutination. Serum albumin, calcium, and magnesium concentrations were determined using spectrophotometric kits provided by Biolabo^® (Maizy, France). Copper was determined spectrophotometrically by kits supplied by LTA Co., (Milano, Italy). Zinc concentrations were determined using a kit provided by Cenrionic GmbH (Wartenberg, Germany). We measured serum IL-6, IL-10, G-CSF, and IL-1β using commercial ELISA sandwich kits provided by Sunlong Biotech Co., Ltd. (Zhejiang, China), and Melsin Medical Co. (Jilin, China) provided the other ELISA kits (IL-17, IL-21, IL-22, IL-23, and TNF-α). We diluted samples as required in the presence of very concentrated biomarkers. The within-assay coefficient of variation for all assays was <10.0% (precision within-assay). The sensitivities of the kits were 1.0 pg/ml for IL-10, IL-17, IL-21, IL-22, IL-23, TNF-α, G-CSF, and 0.1 pg/ml for IL-1β, and IL-6.

Statistical analysis

To compare scale variables between groups, one-way analysis of variance was employed, while analysis of contingency tables (two tests) was used to analyze category variables. We used multivariate general linear model (GLM) analysis to examine the association between the 14 biomarkers and the diagnosis while entering age, gender, education, BMI, TUD and drug use as covariates. The between-subject effects tests were employed to investigate the impact of significant explanatory factors on each of the observed biomarkers. As a result, we calculated the estimated marginal means (SE) values produced by the model (adjusted for the significant confounders). These multiple associations were subjected to p-correction for false discovery rate (FDR) [53], and the protected Least Significant Difference (LSD) tests were used to assess multiple pair-wise differences. Multiple regression analysis (automated approach with p-to-entry of 0.05 and p-to-remove of 0.06 while assessing the change in R²) was used to determine the important biomarkers that predict the phenome in schizophrenia and all participants combined. Multivariate normality (Cook’s distance and leverage), multicollinearity (using tolerance and VIF), and homoscedasticity (using White and modified Breusch-Pagan tests for homoscedasticity) were checked. These regression analyses’ results were always bootstrapped using 5.000 bootstrap samples, and the latter results are shown if the results were not concordant. We used two-step cluster analysis to derive meaningful clusters from the data set. Feature reduction was performed using principal component (PC) analysis. IBM SPSS Windows version 25, 2017 was used to conduct the statistical analysis.

Partial Least Squares (PLS)-SEM pathway analysis is a statistical approach for predicting complex cause-effect linkages utilizing both discrete indicators (variables) and latent variables (factors based on a collection of highly connected indicators) [54]. Without imposing distributional assumptions on the data, PLS allows the estimation of complex multi-step mediation models with several latent constructs, single indicator variables, and structural pathways (associations between indicators or latent vectors). This strategy has recently been used to develop novel nomothetic models of affective disorders and schizophrenia by combining the many components of a disease into a causal-effect, mediation model [9, 22]. A causative framework is developed utilizing causome, protectome, adverse outcome pathways, and phenome indicators (either single indicators or latent vectors), which is then analyzed and cross-validated using PLS pathway analysis on bootstrapped data (e.g., 5.000 samples). As a result, pathway coefficients (with exact p values), specific indirect (mediated) and total effects are estimated to determine the influence of direct and mediated pathways. According to the power analysis, the predicted sample size for multiple regression analysis (which applies to PLS) should be at least 70 to obtain a power of 0.8, an effect size of 0.2, and an alpha of 0.05 with 5 preset variables. Complete path analysis with 5.000 bootstrap samples was conducted only when the outer and inner models matched the following quality criteria: a) model quality as measured by the SRMR index is less than 0.08; b) outer model loadings on the latent vectors are greater than 0.666 at p < 0.001; and c) the latent vectors exhibit accurate construct validity as measured by average variance extracted > 0.5, Cronbach’s alpha > 0.7, rho A > 0.7, and composite reliability > 0.8. Compositional invariance was investigated using Predicted-Oriented Segmentation analysis, Multi-Group Analysis, and Measurement Invariance Assessment. The predictive performance of the model was evaluated using PLSpredict with a 10-fold cross-validation.

Results

Sociodemographic data

Table 1 shows the sociodemographic and clinical data of MNP versus SNP and controls. There were no significant differences in age, sex, BMI, marital status, TUD, residency, and employment status between MNP and SNP and healthy controls. Both patient groups were somewhat less educated than the control group, while there were no significant differences in education level between both patient groups. MNP patients have a significantly higher (p<0.001) positive family history of schizophrenia than the SNP group. Since some of the patients were treated with olanzapine (n = 62), clozapine (n = 12), fluphenazine (n = 27), haloperidol (n = 7), quetiapine (n = 5), risperdal (n = 16), benzodiazepines (n = 19), and mood stabilizers (n = 39) we have examined the impact of these drugs on the clinical scores. After adjusting for age, sex, BMI, smoking, education, and the drug state of the subjects, the major symptom domain scores (psychosis, hostility, excitement, mannerism, SANS, PMR, and FTD) were significantly different between the three study samples and increased from controls → SNP → MNP. These differences remained significant after FDR p-correction at p<0.001.

[Figure omitted. See PDF.]

Differences in biomarkers between the study groups

Multivariate GLM analysis (Table 2) was used to assess the relationships between biomarkers and diagnosis after controlling for age, BMI, sex, smoking and drug state. There were highly significant differences in the biomarkers across the groups (which remained significant after FDR p-correction at p<0.012), but no significant impact of the covariates.Tests for between-subject effects in Table 3, which shows the estimated marginal means, revealed that IL-23, IL-6, IL-17, and TNF-α were significantly higher in MNP than in SNP and controls and that IL-17 and TNF-α were higher in SNP than in controls. IL-1β, IL-22, G-CSF, IL-21, and IL-10 were significantly higher in schizophrenia than in controls. There were no differences in albumin between the study groups, and zinc levels were lowest in both schizophrenia groups. Magnesium was significantly lower in MNP as compared with the other two groups, whereas calcium was lower in MNP than in SNP. Copper levels were significantly higher in MNP than in SNP. We found no significant effects of the drug state on the single biomarkers (tested with multivariate GLM analysis and tests for between-subjects effects) even without p correction for FDR.

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

Results of PLS analysis

Fig 1 shows the results of PLS path analysis with IL-23 and IL-6 as primary input variables and latent vectors extracted from all symptom domains (reflecting the phenome) and cognitive tests (reflecting G-CoDe) as output variables.

[Figure omitted. See PDF.]

Interleukin (IL)-23 and IL-6 are primary input variables. Latent vectors extracted from T helper (Th)17 cytokines, including tumor necrosis factor (TNF)-α, and the generalized cognitive decline (G-CoDe) mediate the effects of IL-6 and IL-23 on the phenome of schizophrenia (outcome indicator). The latter is conceptualized as a latent vector extracted from all symptom domains (reflecting the phenome). The model also considers effects of Th-17 on granulocyte-colony stimulating factor (G-CSF) and IL-10. Cat_Inst: Category Instances; Cont_oral: Controlled Word Association; Digit_seq: Digit Sequencing Task; List_learn: the List Learning test; Symbol..: Symbol Coding; Token..: token motor task; TOL: Tower of London; FTD: formal thought disorders; SANS: negative symptoms; PMR: psychomotor retardation; Excited: excitation; Manner..: mannerism. White figures in circles: explained variance. Shown are the pathway coefficients (p values) and factor loadings (p values).

In this model, IL-23 and IL-6 were considered to be correlated with a latent vector (LV) extracted from IL-17, IL-21, IL-22, and TNF-α (labeled Th17 LV) while the latter predicted G-CSF and IL-10. After feature selection, multi-group analysis, PLS predict analysis, and prediction-oriented segmentation, Fig 1 shows the results of the PLS path analysis conducted on 5.000 bootstrap samples. The final PLS model did not contain non-significant pathways and variables. With SRMR = 0.039, this model’s overall fit was more than sufficient. Additionally, the reliability of the factor constructs was satisfactory for all three factors, with Cronbach alpha > 0.754, rho A > 0.785, composite reliability > 0.845, and AVE > 0.58. At p <0.0001, all loadings on the outer models were more than 0.620. We discovered that 70.7% of the variance in the phenome could be explained by regression on the cognitome, Th17 LV (both positively) and IL-10 (inversely), and that 68.9% of the variance in the cognitome could be explained by education (positively) and the combination of IL-23 and Th17 LV. IL-23 and IL-6 explained 17.7% of the variance in the Th17 LV. The cognitome (0.558), the Th17 LV (0.093), and the phenome (0.624) all had acceptable construct cross-validated redundancies (as determined by PLS blindfolding). Confirmatory Tetrad Analaysis (CTA) in PLS demonstrated that the latent vectors were not conceptualized incorrectly as reflective models. PLS prediction revealed that the indicators of the endogenous constructs had positive Q² predict values, indicating that they outperformed the naïve benchmark. There were significant specific indirect effects of IL-23 on the phenome through the Th17 LV (t = 2.82, p = 0.002), the cognitome (t = 1.68, p = 0.047), the path from Th17 LV to the cognitome (t = 3.64, p<0.001), and the path from Th17 LV to IL-10 (t = -1.84, p = 0.033). Although the direct pathway from IL-6 to the phenome was insignificant, the overall impact of IL-6 on the phenome was significant (t = 1.74, p = 0.041). Nonetheless, IL-23 (t = 6.76, p<0.001) and the Th17 LV (t = 5.10, p<0.001) had a much greater influence on the phenome.

Construction of pathway endophenotype and endophenotype classes

Based on the PLS findings that IL-6 and IL-23 are both associated with the Th17 LV we constructed an LV extracted from IL-6 and IL-23 (AVE = 58.7%, composite reliability = 0.738, loadings: 0.823 and 0.705, respectively) and then extracted an LV from this construct and the Th17 LV, yielding a new IL-6/IL-23/Th17 LV (AVE = 70.8%, loadings: 0.841 and 0.841, respectively, composite reliability = 0.844). Tables 2 and 3 show that this IL-6/IL-23/Th17 score significantly differed between the three study groups and increased from controls → MNP → SNP. In addition, we were able to extract a latent vector from the IL-6/IL-23/Th17 (loadings: 0.880), G-CoDe (-0.918) and phenome (0.918) scores (AVE = 81.52%, composite reliability = 0.912). Two-step cluster analysis with the IL-6/IL-23/Th17, G-CoDe, and phenome entered as continuous variables and the diagnosis of schizophrenia as a categorical variable showed three clusters of patients with a silhouette measure of cohesion and separation of 0.66. The formed schizophrenia clusters exactly matched with the MNP (deficit) versus SNP (non-deficit) dichotomy. We found no significant effects of these drugs on the IL-6/IL-23/Th17 score (tested with univariate GLM analysis).

Prediction of symptom domains and G-CoDe by the IL-6/IL-23/Th17 axis score

The first multiple regression analysis in Table 4 entered the phenome LV as the dependent variable and the IL-6/IL-23/Th17 LV and G-CoDe scores as the explanatory variables while allowing for the effects of sex, age, smoking, education, the trace elements, and ions. Table 4, regression #1 demonstrates that 68.8% of the variance in the phenome was explained by the regression on G-CoDe and IL-6/IL-23/Th17 scores. Figs 2 and 3 show the partial regression plots of the phenome on both predictors (after considering the effects of the other covariate).

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

In the restricted study sample of schizophrenia patients (regression #2), the same variables explained 28.5% of the variance in the phenome.

Table 4, regression #3 demonstrates that 54.6% of the variance in the G-CoDE was explained by the regression on IL-6/IL-23/Th17 (inversely) and magnesium, copper, and zinc (all positively). Fig 4 shows the partial regression plot of the G-CoDE on the IL-6/IL-23/Th17 score (after considering the effects of the other covariates).

[Figure omitted. See PDF.]

In the restricted study sample of schizophrenia patients (regression #4), we found that the IL-6/IL-23/Th17 score (inversely) and magnesium and calcium (positively) explained 28.5% of the variance in the G-CoDe. Fig 5 shows the partial regression of the G-CoDe on magnesium (after adjusting for the other variables in this regression).

[Figure omitted. See PDF.]

Discussion

Increased cytokine levels in MNP

The first major finding of this study is that IL-23, IL-6, IL-17, and TNF-α were considerably greater in MNP than in SNP, while IL-1β, IL-22, G-CSF, IL-21, IL-17, IL-10, and TNF-α were significantly higher in schizophrenia than in controls. Furthermore, elevated IL-23 and IL-6 were identified in MNP but not SNP as compared to controls, suggesting that these cytokines are characteristics of MNP rather than schizophrenia in general. Previous research has linked MNP to increased TNF-α, IL-6, and IL-1β production, as well as CCL11, CCL2, and the soluble IL-1 receptor antagonist [17, 55, 56], and increased TNF-α levels have been identified as a critical factor in the pathophysiology of MNP [57].

Increments in cytokines such as IL-1β, TNF-α, IL-6, and IL-10 have frequently been described in schizophrenia [58]. Blood levels of IL-23 were found to be high in people with treated schizophrenia, FEP, and repeated episodes of schizophrenia [28, 59–61]. As reviewed by Roomruangwon et al. [58] and reported in some [27, 28, 30, 60, 62], but not all [63], studies, schizophrenia is accompanied by increased IL-17 levels. Increased IL-22 and IL-6, but not IL-17 or IL-1β, were observed in schizophrenia as compared with healthy subjects [64]. Interestingly, antipsychotic treatment for 3 months was found to significantly reduce plasma levels of Th17-pathway-related cytokines like IL-6 and IL-17A [65]. Contradictorily, a four-week treatment with antipsychotic agents did not affect circulating levels of IL-23 or IL-17 [66, 67]. Most likely, the differences mentioned above come from comparing different research groups, like MNP vs. SNP/controls or schizophrenia vs. controls.

All in all, our results show that schizophrenia is accompanied by indicants of activation of the M1 (IL-1β, IL-6, TNF-α) and Th17 (IL-17, IL-21, IL-22, and TNF-α) phenotypes and that MNP is additionally characterized by increased IL-6 and IL-23, while increments in those cytokines contribute to increased G-CSF and IL-10 [64–66].

IL-23 and IL-6 as drivers of Th17 pathway activation

The second major finding is that our PLS analysis revealed that increasing IL-23 and IL-6 levels are linked to Th17 activation, as measured by extracting a latent vector from IL-17, IL-21, IL-22, and TNF-α. As described in the introduction, IL-23 and IL-6 are key factors in the growth and survival of the pathogenic Th17 phenotype [34–37, 67, 68]. Moreover, IL-6 and IL-23 signaling is required to activate STAT3 and IL-17 production, while IL-21 activates STAT3 [69, 70]. IL-23 is a key cytokine involved in a variety of inflammatory and autoimmune disorders by producing a highly pathogenic T cell population [71]. Th17 cells are the main producers of IL-22 [72] and IL-17 coupled with IL-22 marks a particularly pathogenic population of aggressive autoreactive and proinflammatory T cells [73]. Nevertheless, IL-22 also has protective properties [74], including acting as an antimicrobial peptide and collaborating with IL-17 in regulating the immune response [72, 73]. IL-21 is another proinflammatory cytokine that mediates antibody class switching and production, and this cytokine has pathogenic effects in (auto)immune disorders, including Sjogren’s syndrome, systemic lupus erythematosus, and psoriasis [75–77].

As a result, rather than calling this axis the TNF/IL-23/IL-17 axis [78] or the IL-23/IL-17/G-CSF axis [66], it is more appropriate to refer to it as the pathogenic IL-6/IL-23/IL-21/IL-17/IL-22/TNF-axis, or simply the IL-6/IL-23/Th17-axis. In this way, we were able to find a common core between IL-23, IL-6, and the Th17 cytokines, which confirmed the concept.

It is worth noting that in schizophrenia, activation of the pathogenic IL-6/IL-23/Th17 axis is linked to other important pathways. First, brain-derived neurotrophic factor (BDNF) is adversely related to IL-23 and illness severity ratings in schizophrenia [79], and a first protein-protein interaction (PPI) network created using FEP/FES genes shows that BDNF is part of the same immune PPI network as STAT3, IL-6, TNF-α, IL-1β, and IL-10 [80]. Second, complement component factors that have a role in schizophrenia, such as C1q, C3, and C4 [21, 81, 82], may affect the production of IL-23 and IL-17 family members [83]. Third, blood-brain barrier (BBB) endothelial cells express IL-17 and IL-22 receptors, and binding to IL-17 and IL-22 causes BBB permeabilization [84, 85], thereby aggravating the effects of lipopolysaccharides (LPS), other cytokines and tryptophan catabolites on BBB breakdown [86]. Fourth, the IL-6/IL-23/Th17 axis plays an important role in the immune-inflammatory responses directed to gut commensal microbiota (including Gram-negative bacteria). Thus, commensal bacteria may prime IL-23 expression and production by dendritic cells [87], and induce IL-22 production in the gut [88]. IL-23 increases gut permeability by a breakdown of the tight junction barrier [89] and orchestrate the immune-inflammatory response in the gut [90]. IL-22 generally has more regenerative and protective functions [91–93], including the regulation of antimicrobial peptide production and the composition of gut commensal bacteria [88]. IL-17 plays a key role in response to infectious agents by mounting an innate immune response in epithelial cells, which may lead to systemic inflammatory and autoimmune responses [91]. IL-21 may trigger gut inflammation and plays a role in inflammatory bowel disease, while claudin-5 is a downstream gene of this cytokine [94]. Breakdown of the tight junctions and adherents barriers (leaky gut) with increased bacterial translocation is involved in schizophrenia and, especially, in MNP [86, 95].

Effects of the pathogenic IL-6/IL-23/Th17 axis on G-CoDe and the phenome of schizophrenia

The third major finding of this study is that the latent vector score extracted from the IL-6/IL-23/Th17 data was considerably greater in MNP than SNP, with a difference of more than one standard deviation, and predicted cognitive deficits and the severity of the phenotype of schizophrenia. Moreover, we showed that a reliable latent vector could be extracted from the IL-6/IL-23/Th17, G-CoDe and phenome scores, thereby shaping a new pathway phenotype reflecting the link among peripheral adverse outcome pathways and the cognitome and phenome of schizophrenia. Furthermore, clustering analysis revealed two qualitatively distinct and well-defined patient groups using the same three scores, and these groups perfectly matched the deficit vs. non-deficient dichotomy. These findings support the view that MNP and SNP are qualitatively distinct classes [96] and that the pathogenic IL-6/IL-23/Th17 axis is a major player in this regard.

We previously addressed IL-6 and TNF-’s many neurotoxic actions, which, when coupled with other neurotoxic substances like LPS and lipid peroxidation, may produce neurotoxicity and, hence, explain the cognitive deficits and phenome of schizophrenia [57, 97]. As discussed previously, these compounds may induce a plethora of neurotoxic effects, especially on processes like neurogenesis, neuroplasticity, cerebral cortex radial glia guided migration, synapse assembly, axogenesis, axonal spreading and branching, synapse assembly and structure, pre- and post-synaptic neuronal connectivity, regulation of excitatory synaptic functions, and post-synaptic protein assembly [21]. By inference, the strong effects of the IL-6/IL-23/Th17 axis on both the cognitive detrioration and symptomatome of schizophrenia indicate that IL-23, Th17, IL-22 and IL-21 may further contribute to the neurotoxic effects.

A variety of cells in the central nervous system (CNS) produce IL-17 under physiological conditions, while IL-17R is expressed by immune cells and glial cell populations [98]. IL-23 is produced by CNS cells and is a principal mediator of microglial activation, neuroinflammation, and tissue damage [99]. There are two ways in which Th17 cells are involved in neuroinflammation, either directly via the production of cytokines (IL-17, IL-21; IL-22; IFN-γ; and G-CSF) or indirectly via the stimulation of neutrophil infiltration and microglial cells, which secrete cytokines, and by attracting CD8+ and Th1 cells to the CNS [25]. IL-17 has been hypothesized to adversely influence adult hippocampus neurogenesis even in physiological conditions [100]. Moreover, IL-23-activated Th17 cells in the brain are essential for sustaining persistent neuroinflammation during infection and autoimmune responses [34]. Furthermore, following stroke, IL-21 and its receptor are expressed in the CNS via brain infiltrating CD4+ T cells, and CNS IL-21 strongly contributes to CNS tissue damage [101]. IL-22 is expressed in the brain in physiological conditions, and is upregulated during neuroinflammatory conditions [102, 103]. While this cytokine may have protective effects [103, 104] and induces an acute phase response [105], it may also display more detrimental inflammatory effects [88]. Most importantly, during inflammatory responses, IL-22 is upregulated in the brain and may increase the production of TNF-α, IL-6, and prostaglandins and induce STAT3, MAP-kinase, and JAK-STAT pathways [102], which all play a role in schizophrenia [102, 106].

It is also important to note that after considering the detrimental effects of the IL-6/IL-23/Th17 axis on the phenome of schizophrenia, IL-10 showed an inverse association with the phenome. Since IL-10 levels were higher in patients than in controls and since IL-10 is a negative immunoregulatory cytokine [16], our results indicate that a relative shortage in IL-10 CIRS functions may aggravate the effects of the pathogenic IL-6/IL-23/Th17 axis.

Calcium, magnesium and zinc in schizophrenia

This study shows that schizophrenia is accompanied by lower zinc as compared with controls, and that MNP is characterized by lower magnesium and calcium but higher copper as compared with SNP. As described in the Introduction, some [42, 107–113] but not all [114, 115] authors reported changes in these trace elements and ions in plasma, serum or CSF in schizophrenia patients. Although some studies reported inverse associations between calcium and magnesium levels and the severity of schizophrenia, others could not find any association [107, 116]. Still, we found that magnesium, zinc, copper, and calcium were positively linked to the G-CoDe. This means that these trace elements and ions protect cognitive functions from the neurotoxic effects of the IL-6/IL-23/Th17 axis, and that the lower levels in schizophrenia and MNP contribute to the overall decline in cognitive function.

It is important to note that inflammation or the acute phase response is accompanied by lowered levels of zinc, magnesium, and calcium (reviews in: [117, 118]. Moreover, these three elements and copper all modulate NMDA receptor (NMDAR) functions: a) zinc and magnesium are anatagonists of the glutamatergic NMDAR [119, 120]; b) increases in intracellular calcium exert a negative feedback on NMDA channels [121]; c) copper modulates the affinity of hippocampal NMDARs for the co-agonist glycine pocampus [122]. As a result, changes in the balance between these trace elements and ions may affect the NMDAR and NMDAR-mediated neuroplasticity and excitotoxicity, which may be linked to schizophrenia.

Conclusion

All in all, the pathogenic IL-6/IL-23/Th17 axis, which may be induced by microbiota, contributes to neurocognitive deficits and the phenome of schizophrenia, especially MNP, due to its key role in peripheral inflammation, gut and BBB permeability, neuroinflammation and ensuing neurotoxic effects on CNS circuits. Such effects will appear or be more prominent in subjects with CIRS deficits, including lowered IL-10, magnesium, calcium, and zinc.

Acknowledgments

We acknowledge the staff of Al-Hakeem General Hospital, Psychiatric Unit in Najaf Governorate, for their help in collecting samples. We also acknowledge the work of the high-skilled staff of Asia Clinical Laboratory in Najaf city for their help in the ELISA assays.

Citation: Al-Hakeim HK, Al-Musawi AF, Al-Mulla A, Al-Dujaili AH, Debnath M, Maes M (2022) The interleukin-6/interleukin-23/T helper 17-axis as a driver of neuro-immune toxicity in the major neurocognitive psychosis or deficit schizophrenia: A precision nomothetic psychiatry analysis. PLoS ONE 17(10): e0275839. https://doi.org/10.1371/journal.pone.0275839

About the Authors:

Hussein Kadhem Al-Hakeim

Roles: Investigation, Project administration, Software, Writing – original draft, Writing – review & editing

E-mail: [email protected]

Affiliation: Department of Chemistry, College of Science, University of Kufa, Kufa, Iraq

https://orcid.org/0000-0001-6143-5196

Ali Fattah Al-Musawi

Roles: Conceptualization, Data curation, Formal analysis, Investigation, Methodology

Affiliation: Department of Clinical Pharmacy and Laboratory Sciences, College of Pharmacy, University of Al-Kafeel, Kufa, Iraq

Abbas Al-Mulla

Roles: Data curation, Formal analysis, Resources

Affiliation: Medical Laboratory Technology Department, College of Medical Technology, The Islamic University, Najaf, Iraq

Arafat Hussein Al-Dujaili

Roles: Conceptualization, Data curation, Investigation, Project administration, Supervision, Validation, Visualization

Affiliation: Senior Clinical Psychiatrist at the Faculty of Medicine, University of Kufa, Kufa, Iraq

Monojit Debnath

Roles: Investigation, Methodology, Software, Validation, Writing – review & editing

Affiliation: Department of Human Genetics, National Institute of Mental Health and Neurosciences, Bangalore, India

Michael Maes

Roles: Formal analysis, Investigation, Project administration, Supervision, Visualization, Writing – original draft, Writing – review & editing

Affiliations Department of Psychiatry, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand, Department of Psychiatry, Medical University of Plovdiv, Plovdiv, Bulgaria, IMPACT Strategic Research Centre, Deakin University, Geelong, VIC, Australia

References

1. Moreno-Küstner B, Martín C, Pastor L. Prevalence of psychotic disorders and its association with methodological issues. A systematic review and meta-analyses. PloS one. 2018;13(4):e0195687. Epub 2018/04/13. pmid:29649252; PubMed Central PMCID: PMC5896987.

2. Maes M, Anderson G. False Dogmas in Schizophrenia Research: Toward the Reification of Pathway Phenotypes and Pathway Classes. Front Psychiatry. 2021;12:663985. Epub 2021/07/06. pmid:34220578; PubMed Central PMCID: PMC8245788.

3. Kanchanatawan B, Sriswasdi S, Thika S, Sirivichayakul S, Carvalho AF, Geffard M, et al. Deficit schizophrenia is a discrete diagnostic category defined by neuro-immune and neurocognitive features: results of supervised machine learning. Metabolic brain disease. 2018;33(4):1053–67. Epub 2018/03/13. pmid:29527624.

4. Kirkpatrick B, Galderisi S. Deficit schizophrenia: an update. World Psychiatry. 2008;7(3):143–7. Epub 2008/10/07. pmid:18836581; PubMed Central PMCID: PMC2559917.

5. Maes M, Kanchanatawan B. In (deficit) schizophrenia, a general cognitive decline partly mediates the effects of neuro-immune and neuro-oxidative toxicity on the symptomatome and quality of life. CNS Spectrums. 2021:1–10. Epub 2021/04/12. pmid:33843548

6. Maes M, Vojdani A, Geffard M, Moreira EG, Barbosa DS, Michelin AP, et al. Schizophrenia phenomenology comprises a bifactorial general severity and a single-group factor, which are differently associated with neurotoxic immune and immune-regulatory pathways. Biomol concepts. 2019;10(1):209–25. pmid:31734647

7. Maes M, Sirivichayakul S, Kanchanatawan B, Carvalho AF. In schizophrenia, psychomotor retardation is associated with executive and memory impairments, negative and psychotic symptoms, neurotoxic immune products and lower natural IgM to malondialdehyde. World J Biol Psychiat. 2020;21(5):383–401. pmid:32031479

8. Almulla AF, Al-Hakeim HK, Maes M. Schizophrenia phenomenology revisited: positive and negative symptoms are strongly related reflective manifestations of an underlying single trait indicating overall severity of schizophrenia. CNS spectrums. 2021;26(4):368–77. pmid:32431263

9. Maes M, Vojdani A, Galecki P, Kanchanatawan B. How to construct a bottom-up nomothetic network model and disclose novel nosological classes by integrating risk resilience and adverse outcome pathways with the phenome of schizophrenia. Brain Sciences. 2020;10(9):645. pmid:32957709

10. Smith R, Maes M. The macrophage-T-lymphocyte theory of schizophrenia: additional evidence. Medical hypotheses. 1995;45(2):135–41. pmid:8531836

11. Meyer U, Schwarz MJ, Muller N. Inflammatory processes in schizophrenia: a promising neuroimmunological target for the treatment of negative/cognitive symptoms and beyond. Pharmacol Ther. 2011;132(1):96–110. Epub 2011/06/28. pmid:21704074.

12. Müller N, Myint A-M, Schwarz MJ. Inflammation in schizophrenia. Adv Protein Chem Struct Biol. 2012;88:49–68. pmid:22814706

13. Mongan D, Ramesar M, Föcking M, Cannon M, Cotter D. Role of inflammation in the pathogenesis of schizophrenia: A review of the evidence, proposed mechanisms and implications for treatment. Early intervention in psychiatry. 2020;14(4):385–97. Epub 2019/08/02. pmid:31368253.

14. Debnath M. Adaptive Immunity in Schizophrenia: Functional Implications of T Cells in the Etiology, Course and Treatment. J Neuroimmune Pharmacol. 2015;10(4):610–9. Epub 2015/07/15. pmid:26162591.

15. Momtazmanesh S, Zare-Shahabadi A, Rezaei N. Cytokine alterations in schizophrenia: an updated review. Frontiers in psychiatry. 2019;10:892. pmid:31908647

16. Roomruangwong C, Noto C, Kanchanatawan B, Anderson G, Kubera M, Carvalho AF, et al. The role of aberrations in the immune-inflammatory response system (IRS) and the compensatory immune-regulatory reflex system (CIRS) in different phenotypes of schizophrenia: the IRS-CIRS theory of schizophrenia. Molecular neurobiology. 2020;57(2):778–97. Epub 2019/09/02. pmid:31473906.

17. Al-Hakeim HK, Almulla AF, Maes M. The Neuroimmune and Neurotoxic Fingerprint of Major Neurocognitive Psychosis or Deficit Schizophrenia: a Supervised Machine Learning Study. Neurotoxicity research. 2020;37(3):753–71. Epub 2020/01/10. pmid:31916129.

18. Orlovska-Waast S, Köhler-Forsberg O, Brix SW, Nordentoft M, Kondziella D, Krogh J, et al. Cerebrospinal fluid markers of inflammation and infections in schizophrenia and affective disorders: a systematic review and meta-analysis. Molecular psychiatry. 2019;24(6):869–87. Epub 2018/08/18. pmid:30116031; PubMed Central PMCID: PMC6756288.

19. Al-Dujaili AH, Mousa RF, Al-Hakeim HK, Maes M. High Mobility Group Protein 1 and Dickkopf-Related Protein 1 in Schizophrenia and Treatment-Resistant Schizophrenia: Associations With Interleukin-6, Symptom Domains, and Neurocognitive Impairments. Schizophrenia bulletin. 2021;47(2):530–41. Epub 2020/09/25. pmid:32971537; PubMed Central PMCID: PMC7965081.

20. Noto MN, Maes M, Nunes SOV, Ota VK, Rossaneis AC, Verri WA Jr, et al. Activation of the immune-inflammatory response system and the compensatory immune-regulatory system in antipsychotic naive first episode psychosis. Eur Neuropsychopharmacol. 2019;29(3):416–31. pmid:30594344

21. Maes M, Plaimas K, Suratanee A, Noto C, Kanchanatawan B. First Episode Psychosis and Schizophrenia Are Systemic Neuro-Immune Disorders Triggered by a Biotic Stimulus in Individuals with Reduced Immune Regulation and Neuroprotection. Cells. 2021;10(11):2929. Epub 2021/11/28. pmid:34831151; PubMed Central PMCID: PMC8616258.

22. Al-Hakeim HK, Almulla AF, Al-Dujaili AH, Maes M. Construction of a Neuro-Immune-Cognitive Pathway-Phenotype Underpinning the Phenome of Deficit Schizophrenia. Current topics in medicinal chemistry. 2020;20(9):747–58. Epub 2020/01/30. pmid:31994463.

23. Sirivichayakul S, Kanchanatawan B, Thika S, Carvalho AF, Maes M. A New Schizophrenia Model: Immune Activation is Associated with the Induction of Different Neurotoxic Products which Together Determine Memory Impairments and Schizophrenia Symptom Dimensions. CNS & neurological disorders drug targets (Formerly Current Drug Targets-CNS & Neurological Disorders). 2019;18(2):124–40. Epub 2018/11/20. pmid:30451122.

24. Al-Hakeim HK, Mousa RF, Al-Dujaili AH, Maes M. In schizophrenia, non-remitters and partial remitters to treatment with antipsychotics are qualitatively distinct classes with respect to neurocognitive deficits and neuro-immune biomarkers: results of soft independent modeling of class analogy. Metabolic brain disease. 2021;36(5):939–55. Epub 2021/02/14. pmid:33580860.

25. Tahmasebinia F, Pourgholaminejad A. The role of Th17 cells in auto-inflammatory neurological disorders. Prog Neuropsychopharmacol Biol Psychiatry. 2017;79(Pt B):408–16. Epub 2017/08/02. pmid:28760387.

26. Debnath M, Berk M. Th17 pathway-mediated immunopathogenesis of schizophrenia: mechanisms and implications. Schizophrenia bulletin. 2014;40(6):1412–21. Epub 2014/04/09. pmid:24711545; PubMed Central PMCID: PMC4193719.

27. O’Connell KE, Thakore J, Dev KK. Proinflammatory cytokine levels are raised in female schizophrenia patients treated with clozapine. Schizophrenia research. 2014;156(1):1–8. Epub 2014/04/20. pmid:24742875.

28. Borovcanin M, Jovanovic I, Dejanovic SD, Radosavljevic G, Arsenijevic N, Lukic ML. Increase systemic levels of IL-23 as a possible constitutive marker in schizophrenia. Psychoneuroendocrinol. 2015;56:143–7.

29. Nurieva R, Yang XO, Martinez G, Zhang Y, Panopoulos AD, Ma L, et al. Essential autocrine regulation by IL-21 in the generation of inflammatory T cells. Nature. 2007;448(7152):480–3. Epub 2007/06/22. pmid:17581589.

30. El Kissi Y, Samoud S, Mtiraoui A, Letaief L, Hannachi N, Ayachi M, et al. Increased Interleukin-17 and decreased BAFF serum levels in drug-free acute schizophrenia. Psychiatry Res. 2015;225(1–2):58–63. Epub 2014/12/03. pmid:25453636.

31. Ding M, Song X, Zhao J, Gao J, Li X, Yang G, et al. Activation of Th17 cells in drug naïve, first episode schizophrenia. Progress in neuro-psychopharmacology & biological psychiatry. 2014;51:78–82. Epub 2014/01/23. pmid:24447943.

32. McGeachy MJ, Cua DJ. Th17 cell differentiation: the long and winding road. Immunity. 2008;28(4):445–53. pmid:18400187

33. Korn T, Bettelli E, Oukka M, Kuchroo VK. IL-17 and Th17 Cells. Ann Rev Immunol. 2009;27:485–517. pmid:19132915

34. Langrish CL, Chen Y, Blumenschein WM, Mattson J, Basham B, Sedgwick JD, et al. IL-23 drives a pathogenic T cell population that induces autoimmune inflammation. J Exp Med. 2005;201(2):233–40. pmid:15657292

35. Wacleche VS, Landay A, Routy J-P, Ancuta P. The Th17 Lineage: From Barrier Surfaces Homeostasis to Autoimmunity, Cancer, and HIV-1 Pathogenesis. Viruses. 2017;9(10):303. pmid:29048384.

36. Wu R, Zeng J, Yuan J, Deng X, Huang Y, Chen L, et al. MicroRNA-210 overexpression promotes psoriasis-like inflammation by inducing Th1 and Th17 cell differentiation. J Clin Invest. 2018;128(6):2551–68. pmid:29757188

37. Zhao G, Liu Y, Yi X, Wang Y, Qiao S, Li Z, et al. Curcumin inhibiting Th17 cell differentiation by regulating the metabotropic glutamate receptor-4 expression on dendritic cells. International immunopharmacology. 2017;46:80–6. pmid:28273557

38. Ten Bergen LL, Petrovic A, Krogh Aarebrot A, Appel S. The TNF/IL‐23/IL‐17 axis—Head‐to‐head trials comparing different biologics in psoriasis treatment. Scand J Immunol. 2020;92(4):e12946. pmid:32697374

39. Korn T, Bettelli E, Gao W, Awasthi A, Jäger A, Strom TB, et al. IL-21 initiates an alternative pathway to induce proinflammatory T(H)17 cells. Nature. 2007;448(7152):484–7. Epub 2007/06/22. pmid:17581588; PubMed Central PMCID: PMC3805028.

40. Bauquet AT, Jin H, Paterson AM, Mitsdoerffer M, Ho IC, Sharpe AH, et al. The costimulatory molecule ICOS regulates the expression of c-Maf and IL-21 in the development of follicular T helper cells and TH-17 cells. Nature immunology. 2009;10(2):167–75. Epub 2008/12/23. pmid:19098919; PubMed Central PMCID: PMC2742982.

41. Wolf TL, Kotun J, Meador-Woodruff JH. Plasma copper, iron, ceruloplasmin and ferroxidase activity in schizophrenia. Schizophrenia research. 2006;86(1–3):167–71. pmid:16842975

42. Joe P, Petrilli M, Malaspina D, Weissman J. Zinc in schizophrenia: A meta-analysis. General hospital psychiatry. 2018;53:19–24. Epub 2018/05/05. pmid:29727763.

43. Ordak M, Matras J, Muszynska E, Nasierowski T, Bujalska-Zadrozny M. Magnesium in schizophrenia. Pharmacological reports: PR. 2017;69(5):929–34. Epub 2017/06/27. pmid:28651118.

44. Lidow MS. Calcium signaling dysfunction in schizophrenia: a unifying approach. Brain Res Rev. 2003;43(1):70–84. Epub 2003/09/23. pmid:14499463.

45. Chen S, Xia HS, Zhu F, Yin GZ, Qian ZK, Jiang CX, et al. Association between decreased serum albumin levels and depressive symptoms in patients with schizophrenia in a Chinese Han population: A pilot study. Psychiat Res. 2018;270:438–42.

46. Kirkpatrick B, Buchanan RW, McKenny PD, Alphs LD, Carpenter WT Jr. The schedule for the deficit syndrome: an instrument for research in schizophrenia. Psychiatry research. 1989;30(2):119–23. pmid:2616682

47. Andreasen NC. The Scale for the Assessment of Negative Symptoms (SANS): conceptual and theoretical foundations. The British journal of psychiatry. 1989;155(S7):49–52. pmid:2695141

48. Overall JE, Gorham DR. A pattern probability model for the classification of psychiatric patients. Behav Sci. 1963;8(2):108–16.

49. Hamilton M. A rating scale for depression. Journal of neurology, neurosurgery, and psychiatry. 1960;23(1):56. pmid:14399272

50. Hamilton M. The assessment of anxiety states by rating. British journal of medical psychology. 1959. pmid:13638508

51. Sirivichayakul S, Kanchanatawan B, Thika S, Carvalho AF, Maes M. Eotaxin, an endogenous cognitive deteriorating chemokine (ECDC), is a major contributor to cognitive decline in normal people and to executive, memory, and sustained attention deficits, formal thought disorders, and psychopathology in schizophrenia patients. Neurotoxicity research. 2019;35(1):122–38. pmid:30056534

52. Keefe RS, Goldberg TE, Harvey PD, Gold JM, Poe MP, Coughenour L. The Brief Assessment of Cognition in Schizophrenia: reliability, sensitivity, and comparison with a standard neurocognitive battery. Schizophrenia research. 2004;68(2–3):283–97. Epub 2004/04/22. pmid:15099610.

53. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I. Controlling the false discovery rate in behavior genetics research. Behav Brain Res. 2001;125(1–2):279–84. Epub 2001/10/30. pmid:11682119.

54. Ringle C, Da Silva D, Bido D. Structural equation modeling with the SmartPLS. Braz J Market. 2015;13(2).

55. Maes M, Kubera M, Kotańska M. Aberrations in the Cross-Talks Among Redox, Nuclear Factor-κB and Wnt/Catenin Pathway Signaling Underpin Myalgic Encephalomyelitis andChronic Fatigue Syndrome: A Review and New Hypothesis Based on Results of Network, Enrichment and Annotation Analyses. Preprints 2021;2021090201

56. Maes M, Sirivichayakul S, Matsumoto AK, Michelin AP, de Oliveira Semeão L, de Lima Pedrão JV, et al. Lowered antioxidant defenses and increased oxidative toxicity are hallmarks of deficit schizophrenia: a nomothetic network psychiatry approach. Molecular neurobiology. 2020;57(11):4578–97. pmid:32754898

57. Maes M, Sirivichayakul S, Matsumoto AK, Maes A, Michelin AP, de Oliveira Semeão L, et al. Increased levels of plasma tumor necrosis factor-α mediate schizophrenia symptom dimensions and neurocognitive impairments and are inversely associated with natural IgM directed to malondialdehyde and paraoxonase 1 activity. Molecular neurobiology. 2020;57(5):2333–45. Epub 2020/02/11. pmid:32040834.

58. Roomruangwong C, Barbosa DS, de Farias CC, Matsumoto AK, Baltus TH, Morelli NR, et al. Natural regulatory IgM‐mediated autoimmune responses directed against malondialdehyde regulate oxidative and nitrosative pathways and coupled with IgM responses to nitroso adducts attenuate depressive and physiosomatic symptoms at the end of term pregnancy. Psychiatry and Clinical Neurosciences. 2018;72(2):116–30. pmid:29194848

59. Shahraki A, Sarabandi R, Kianpour M, Zakeri Z. Elevated serum interleukin-23 and interleukin-6 levels in schizophrenic patients compared to those in healthy controls. Shiraz E Med J. 2016;17:6.

60. Li H, Zhang Q, Li N, Wang F, Xiang H, Zhang Z, et al. Plasma levels of Th17-related cytokines and complement C3 correlated with aggressive behavior in patients with schizophrenia. Psychiat Res. 2016;246:700–6. pmid:27829509

61. O’Connell KE, Thakore J, Dev KK. Increased interleukin 23 (IL23) levels in schizophrenia patients treated with depot antipsychotic medication. Cytokine. 2015;73(1):196–8. pmid:25769808

62. Leboyer M, Godin O, Terro E, Boukouaci W, Lu CL, Andre M, et al. Immune Signatures of Treatment-Resistant Schizophrenia: A FondaMental Academic Centers of Expertise for Schizophrenia (FACE-SZ) Study. Schizophrenia bulletin open. 2021;2(1):sgab012. Epub 2021/12/14. pmid:34901861; PubMed Central PMCID: PMC8650073.

63. Dimitrov DH, Lee S, Yantis J, Valdez C, Paredes RM, Braida N, et al. Differential correlations between inflammatory cytokines and psychopathology in veterans with schizophrenia: potential role for IL-17 pathway. Schizophrenia research. 2013;151(1–3):29–35. pmid:24210870

64. Bunte K, Beikler T. Th17 Cells and the IL-23/IL-17 Axis in the Pathogenesis of Periodontitis and Immune-Mediated Inflammatory Diseases. Int J Mol Sci. 2019;20(14):3394. Epub 2019/07/13. pmid:31295952; PubMed Central PMCID: PMC6679067.

65. Shin H-M, Ernstoff A, Arnot JA, Wetmore BA, Csiszar SA, Fantke P, et al. Risk-based high-throughput chemical screening and prioritization using exposure models and in vitro bioactivity assays. Environ Sci Technol. 2015;49(11):6760–71. pmid:25932772

66. Karalyan Z, Voskanyan H, Ter-Pogossyan Z, Saroyan D, Karalova E. IL-23/IL-17/G-CSF pathway is associated with granulocyte recruitment to the lung during African swine fever. Veterinary immunology and immunopathology. 2016;179:58–62. Epub 2016/09/04. pmid:27590426.

67. Iwakura Y, Ishigame H. The IL-23/IL-17 axis in inflammation. J Clin Invest. 2006;116(5):1218–22. Epub 2006/05/04. pmid:16670765; PubMed Central PMCID: PMC1451213.

68. McGeachy MJ, Bak-Jensen KS, Chen Y, Tato CM, Blumenschein W, McClanahan T, et al. TGF-β and IL-6 drive the production of IL-17 and IL-10 by T cells and restrain TH-17 cell–mediated pathology. Nature immunology. 2007;8(12):1390–7. pmid:17994024

69. Mathur AN, Chang H-C, Zisoulis DG, Stritesky GL, Yu Q, O’Malley JT, et al. Stat3 and Stat4 direct development of IL-17-secreting Th cells. J Immunol. 2007;178(8):4901–7. pmid:17404271

70. Frydecka D, Misiak B, Pawlak-Adamska E, Karabon L, Tomkiewicz A, Sedlaczek P, et al. Interleukin-6: the missing element of the neurocognitive deterioration in schizophrenia? The focus on genetic underpinnings, cognitive impairment and clinical manifestation. European archives of psychiatry and clinical neuroscience. 2015;265(6):449–59. Epub 2014/09/13. pmid:25214388; PubMed Central PMCID: PMC4540774.

71. Duvallet E, Semerano L, Assier E, Falgarone G, Boissier M-C. Interleukin-23: a key cytokine in inflammatory diseases. Annals of medicine. 2011;43(7):503–11. pmid:21585245

72. Liang SC, Tan XY, Luxenberg DP, Karim R, Dunussi-Joannopoulos K, Collins M, et al. Interleukin (IL)-22 and IL-17 are coexpressed by Th17 cells and cooperatively enhance expression of antimicrobial peptides. J Exp Med. 2006;203(10):2271–9. Epub 2006/09/20. pmid:16982811; PubMed Central PMCID: PMC2118116.

73. Bettelli E, Oukka M, Kuchroo VK. TH-17 cells in the circle of immunity and autoimmunity. Nature immunology. 2007;8(4):345–50. pmid:17375096

74. Xin N, Namaka MP, Dou C, Zhang Y. Exploring the role of interleukin-22 in neurological and autoimmune disorders. International immunopharmacology. 2015;28(2):1076–83. Epub 2015/08/28. pmid:26311525.

75. Kang KY, Kim H-O, Kwok S-K, Ju JH, Park K-S, Sun D-I, et al. Impact of interleukin-21 in the pathogenesis of primary Sjogren’s syndrome: increased serum levels of interleukin-21 and its expression in the labial salivary glands. Arthritis research & therapy. 2011;13(5):1–10. pmid:22030011

76. Spolski R, Leonard WJ. Interleukin-21: a double-edged sword with therapeutic potential. Nat Rev Drug Discov. 2014;13(5):379–95. pmid:24751819

77. Botti E, Spallone G, Caruso R, Monteleone G, Chimenti S, Costanzo A. Psoriasis, from pathogenesis to therapeutic strategies: IL-21 as a novel potential therapeutic target. Curr Pharma Biotechnol. 2012;13(10):1861–7. pmid:22250707

78. ten Bergen LL, Petrovic A, Krogh Aarebrot A, Appel S. The TNF/IL-23/IL-17 axis—Head-to-head trials comparing different biologics in psoriasis treatment. Scand J Immunol. 2020;92(4):e12946. pmid:32697374

79. Allimuthu P, Nandeesha H, Chinniyappan R, Bhardwaz B, Blessed Raj J. Relationship of Brain-Derived Neurotrophic Factor with Interleukin-23, Testosterone and Disease Severity in Schizophrenia. Indian journal of clinical biochemistry: IJCB. 2021;36(3):365–9. Epub 2021/07/06. pmid:34220013; PubMed Central PMCID: PMC8215017.

80. Maes M, Plaimas K, Suratanee A, Noto C, Kanchanatawan B. The Protein-Protein Interaction Network of First Episode Psychosis and Schizophrenia Reveals Possible Trigger Factors and New Drug Targets among Intracellular Signal Transduction Pathways and Neurotoxicity Processes. Preprints. 2021; 2021070596

81. Maes M, Delange J, Ranjan R, Meltzer HY, Desnyder R, Cooremans W, et al. Acute phase proteins in schizophrenia, mania and major depression: modulation by psychotropic drugs. Psychiatry research. 1997;66(1):1–11. pmid:9061799

82. Gallego JA, Blanco EA, Morell C, Lencz T, Malhotra AK. Complement component C4 levels in the cerebrospinal fluid and plasma of patients with schizophrenia. Neuropsychopharmacology. 2021;46(6):1140–4. Epub 2020/09/23. pmid:32961544; PubMed Central PMCID: PMC8115103.

83. Grailer JJ, Bosmann M, Ward PA. Regulatory effects of C5a on IL-17A, IL-17F, and IL-23. Front Immunol. 2012;3:387. Epub 2013/01/15. pmid:23316190; PubMed Central PMCID: PMC3540403.

84. Kebir H, Kreymborg K, Ifergan I, Dodelet-Devillers A, Cayrol R, Bernard M, et al. Human TH 17 lymphocytes promote blood-brain barrier disruption and central nervous system inflammation. Nat Med. 2007;13(10):1173–5. pmid:17828272

85. Huppert J, Closhen D, Croxford A, White R, Kulig P, Pietrowski E, et al. Cellular mechanisms of IL-17-induced blood-brain barrier disruption. FASEB J. 2010;24(4):1023–34. Epub 2009/11/27. pmid:19940258.

86. Maes M, Sirivichayakul S, Kanchanatawan B, Vodjani A. Breakdown of the Paracellular Tight and Adherens Junctions in the Gut and Blood Brain Barrier and Damage to the Vascular Barrier in Patients with Deficit Schizophrenia. Neurotoxicity research. 2019;36(2):306–22. Epub 2019/05/12. pmid:31077000.

87. Smits HH, van Beelen AJ, Hessle C, Westland R, de Jong E, Soeteman E, et al. Commensal Gram-negative bacteria prime human dendritic cells for enhanced IL-23 and IL-27 expression and enhanced Th1 development. European journal of immunology. 2004;34(5):1371–80. Epub 2004/04/29. pmid:15114670.

88. Zenewicz LA. IL-22: there is a gap in our knowledge. Immunohorizons. 2018;2(6):198–207. pmid:31022687

89. Heinzerling NP, Donohoe D, Fredrich K, Gourlay DM, Liedel JL. Interleukin-23 increases intestinal epithelial cell permeability in vitro. Eur J Pediat Surg. 2016;26(03):260–6. pmid:26007691

90. Maloy K, Kullberg M. IL-23 and Th17 cytokines in intestinal homeostasis. Mucosal immunol. 2008;1(5):339–49. pmid:19079198

91. Lee JS, Tato CM, Joyce-Shaikh B, Gulen MF, Cayatte C, Chen Y, et al. Interleukin-23-independent IL-17 production regulates intestinal epithelial permeability. Immunity. 2015;43(4):727–38. Epub 2015/10/04. pmid:26431948; PubMed Central PMCID: PMC6044435.

92. Wang P, Bai F, Zenewicz LA, Dai J, Gate D, Cheng G, et al. IL-22 signaling contributes to West Nile encephalitis pathogenesis. PloS one. 2012;7(8):e44153. pmid:22952908

93. Valeri M, Raffatellu M. Cytokines IL-17 and IL-22 in the host response to infection. Pathog Dis. 2016;74(9):ftw111. pmid:27915228

94. Wang M, Guo J, Zhao YQ, Wang JP. IL-21 mediates microRNA-423-5p /claudin-5 signal pathway and intestinal barrier function in inflammatory bowel disease. Aging. 2020;12(16):16099–110. Epub 2020/08/29. pmid:32855360; PubMed Central PMCID: PMC7485739.

95. Maes M, Kanchanatawan B, Sirivichayakul S, Carvalho AF. In Schizophrenia, Increased Plasma IgM/IgA Responses to Gut Commensal Bacteria Are Associated with Negative Symptoms, Neurocognitive Impairments, and the Deficit Phenotype. Neurotoxicity research. 2019;35(3):684–98. Epub 2018/12/16. pmid:30552634.

96. Kanchanatawan B, Sriswasdi S, Thika S, Stoyanov D, Sirivichayakul S, Carvalho AF, et al. Towards a new classification of stable phase schizophrenia into major and simple neuro-cognitive psychosis: Results of unsupervised machine learning analysis. Journal of evaluation in clinical practice. 2018;24(4):879–91. Epub 2018/05/24. pmid:29790237.

97. Maes M, Anderson G, Kubera M, Berk M. Targeting classical IL-6 signalling or IL-6 trans-signalling in depression? Expert opinion on therapeutic targets. 2014;18(5):495–512. Epub 2014/02/20. pmid:24548241.

98. Kawanokuchi J, Shimizu K, Nitta A, Yamada K, Mizuno T, Takeuchi H, et al. Production and functions of IL-17 in microglia. J Neuroimmunol. 2008;194(1–2):54–61. Epub 2008/01/01. pmid:18164424.

99. Nitsch L, Zimmermann J, Krauthausen M, Hofer MJ, Saggu R, Petzold GC, et al. CNS-specific synthesis of interleukin 23 induces a progressive cerebellar ataxia and the accumulation of both T and B cells in the brain: characterization of a novel transgenic mouse model. Mol Neurobiol. 2019;56(12):7977–93. pmid:31154574

100. Liu Q, Xin W, He P, Turner D, Yin J, Gan Y, et al. Interleukin-17 inhibits adult hippocampal neurogenesis. Sci Rep. 2014;4(1):1–8. pmid:25523081

101. Leavy O. IL-21 stokes brain inflammation. Nat Rev Immunol. 2014;14(4):215–.

102. Lee D, Jo H, Go C, Jang Y, Chu N, Bae S, et al. The Roles of IL-22 and Its Receptor in the Regulation of Inflammatory Responses in the Brain. Int J Mol Sci. 2022;23(2):757. Epub 2022/01/22. pmid:35054942; PubMed Central PMCID: PMC8775345.

103. Perriard G, Mathias A, Enz L, Canales M, Schluep M, Gentner M, et al. Interleukin-22 is increased in multiple sclerosis patients and targets astrocytes. J Neuroinflam. 2015;12(1):1–18. pmid:26077779

104. Caspi RR, Rigden RC, Kielczewski JL, Zarate-Blades CR, Leger AJS, Silver PB, et al. Neuroprotective effects of IL-22 during CNS inflammation. Investig Ophthalmol Vis Sci. 2017;58(8):1236–.

105. Liang SC, Nickerson-Nutter C, Pittman DD, Carrier Y, Goodwin DG, Shields KM, et al. IL-22 induces an acute-phase response. J Immunol. 2010;185(9):5531–8. pmid:20870942

106. Maes M, Nani JV, Noto C, Rizzo L, Hayashi MA, Brietzke E. Impairments in peripheral blood T effector and T regulatory lymphocytes in bipolar disorder are associated with staging of illness and anti-cytomegalovirus IgG levels. Molecular neurobiology. 2021;58(1):229–42. pmid:32914395

107. Ma J, Yan L, Guo T, Yang S, Liu Y, Xie Q, et al. Association between Serum Essential Metal Elements and the Risk of Schizophrenia in China. Sci Rep. 2020;10(1):10875-. pmid:32001736.

108. Bojarski L, Debowska K, Wojda U. In vitro findings of alterations in intracellular calcium homeostasis in schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry. 2010;34(8):1367–74. Epub 2010/09/04. pmid:20813148.

109. Li Z, Liu Y, Li X, Ju W, Wu G, Yang X, et al. Association of Elements with Schizophrenia and Intervention of Selenium Supplements. Biol Trace Elem Res. 2018;183(1):16–21. Epub 2017/08/16. pmid:28812245.

110. Levine J, Rapoport A, Mashiah M, Dolev E. Serum and cerebrospinal levels of calcium and magnesium in acute versus remitted schizophrenic patients. Neuropsychobiology. 1996;33(4):169–72. pmid:8840338

111. Kirov GK, Birch NJ, Steadman P, Ramsey RG. Plasma magnesium levels in a population of psychiatric patients: correlations with symptoms. Neuropsychobiology. 1994;30(2–3):73–8. Epub 1994/01/01. pmid:7800167.

112. Mazhari S, Arjmand S, Eslami Shahrbabaki M, Karimi Ghoughari E. Comparing Copper Serum Level and Cognitive Functioning in Patients With Schizophrenia and Healthy Controls. Basic Clin Neurosci. 2020;11(5):649–57. Epub 2020/09/01. pmid:33643558.

113. Sharma S, Sood S, Sharma A, Gupta I. Estimation of serum zinc and copper levels patients with schizophrenia: a preliminary study. Sri Lanka J Psychiatry. 2014;5(1).

114. Khan A, Tupling D. A brief note on serum calcium levels in schizophrenics and controls. Can J Psychiatry. 1990. pmid:2317738

115. Jamilian H. Vitamin D, parathyroid hormone, serum calcium and phosphorus in patients with schizophrenia and major depression. Eur Psychiatry. 2011;26(S2):1407–.

116. Kirov GK, Tsachev KN. Magnesium, schizophrenia and manic-depressive disease. Neuropsychobiology. 1990;23(2):79–81. Epub 1990/01/01. pmid:2077436.

117. Mousa RF, Smesam HN, Qazmooz HA, Al-Hakeim HK, Maes M. A pathway phenotype linking metabolic, immune, oxidative, and opioid pathways with comorbid depression, atherosclerosis, and unstable angina. CNS Spectr. 2021:1–15. Epub 2021/05/28. pmid:34039448.

118. Maes M, Vandoolaeghe E, Neels H, Demedts P, Wauters A, Meltzer HY, et al. Lower serum zinc in major depression is a sensitive marker of treatment resistance and of the immune/inflammatory response in that illness. Biol Psychiatry. 1997;42(5):349–58. Epub 1997/09/01. pmid:9276075.

119. Sowa-Kućma M, Szewczyk B, Sadlik K, Piekoszewski W, Trela F, Opoka W, et al. Zinc, magnesium and NMDA receptor alterations in the hippocampus of suicide victims. Journal of affective disorders. 2013;151(3):924–31. Epub 2013/09/24. pmid:24055117.

120. Mlyniec K. Zinc in the Glutamatergic Theory of Depression. Current neuropharmacology. 2015;13(4):505–13. Epub 2015/09/29. pmid:26412070; PubMed Central PMCID: PMC4790399.

121. Xin W-K, Kwan CL, Zhao X-H, Xu J, Ellen RP, McCulloch CA, et al. A functional interaction of sodium and calcium in the regulation of NMDA receptor activity by remote NMDA receptors. J Neurosci. 2005;25(1):139–48. pmid:15634775

122. Stys PK, You H, Zamponi GW. Copper‐dependent regulation of NMDA receptors by cellular prion protein: implications for neurodegenerative disorders. J Physiol. 2012;590(6):1357–68. pmid:22310309