The deficit subtype of schizophrenia is associated with a pro-inflammatory phenotype but not with altered levels of zonulin: Findings from a case-control study

There is evidence that subclinical inflammation and increased gut permeability might be involved in the path- ophysiology of schizophrenia. Less is known about these phenomena in patients with the deficit subtype of schizophrenia (D-SCZ) characterized by primary and enduring negative symptoms. Therefore, in the present study we aimed to compare the levels of zonulin (the marker of gut permeability) and immune-inflammatory markers in patients with D-SCZ, those with non-deficit schizophrenia (ND-SCZ) and healthy controls (HCs). A total of 119 outpatients with schizophrenia and 120 HCs were enrolled. The levels of 26 immune-inflammatory markers and zonulin were determined in serum samples. The following between-group differences were sig- nificant after adjustment for multiple testing and the effects of potential confounding factors: 1) higher levels of interleukin(IL) (cid:0) 1 β and C-reactive protein (CRP) in patients with D-SCZ compared to those with ND-SCZ and HCs; 2) higher levels of tumor necrosis factor- α and RANTES in both groups of patients with schizophrenia compared to HCs and 3) higher levels of IL-17 in patients with D-SCZ compared to HCs. No significant between- group differences in zonulin levels were found. Higher levels of IL-1 β and CRP were associated with worse performance of attention after adjustment for age, education and chlorpromazine equivalents. Also, higher levels of IL-1 β were correlated with greater severity of negative symptoms after adjustment for potential confounding factors. In conclusion, individuals with D-SCZ are more likely to show subclinical inflammation. However, findings from the present study do not support the hypothesis that this phenomenon is secondary to increased gut permeability.


Introduction
Subclinical inflammation represents one of the most widely replicated findings in the pathophysiology of schizophrenia (Misiak et al., 2019). There is convincing evidence from epidemiological studies that immune activation during critical periods of the brain development increases the risk of schizophrenia (Comer et al., 2020). Moreover, it has been documented that autoimmune diseases are more likely to occur in individuals with schizophrenia (Cullen et al., 2019). Studies show that a number of immune-inflammatory alterations, especially the pro-inflammatory profile of cytokine levels, can be observed at various stages of psychotic disorders, including the preclinical stage (Capuzzi et al., 2017;Miller et al., 2011). For instance, increased levels of interleukin(IL)− 6 have been demonstrated in subjects at risk of psychosis and this observation is supported by findings from the most recent meta-analyses (Misiak et al., 2021;Park and Miller, 2020). In subjects with established diagnosis of psychosis, dynamic fluctuations of cytokine levels can be detected at various stages of illness. Therefore, some of them can be classified as trait markers (i.e., cytokine alterations observed across all stages of psychosis) and state markers (i.e., cytokine alterations observed during first-episode psychosis and acute relapse) (Miller et al., 2011).
Several studies indicate the necessity to explore aberrant immuneinflammatory responses of schizophrenia taking into consideration clinical heterogeneity. For instance, it has been found that subclinical inflammation might be greater among individuals with treatmentresistant schizophrenia (Jiao et al., 2022) and those with marked cognitive impairments (Bora, 2019;Misiak et al., 2018;Patlola et al., 2023). Also, recent studies clearly indicate that individuals with the deficit subtype of schizophrenia (D-SCZ), i.e., those with primary and enduring negative symptoms (Carpenter et al., 1988), might show higher levels of subclinical inflammation compared to individuals with non-deficit schizophrenia (ND-SCZ). Specifically, Goldsmith et al. (2018) demonstrated increased levels of IL-6 and tumor necrosis factor-α (TNF-α) in patients with D-SCZ compared to those with ND-SCZ and healthy controls. Another study revealed elevated levels of C-reactive protein (CRP), IL-1β, IL-6, IL-8, interferon(IFN)-γ and total pro-inflammatory cytokines in subjects with D-SCZ (Wang et al., 2022). Finally, it has been found that increased immune responses to kynurenine metabolites might further define individuals with D-SCZ. (Kanchanatawan et al., 2018a;Kanchanatawan et al., 2018b).
The origin of greater subclinical inflammation in subjects with D-SCZ remains unclear. There is a growing body of evidence that individuals with schizophrenia show greater gut permeability assessed by means of the lactulose-mannitol loading test (Ishida et al., 2022) and blood-based markers, such as anti-Saccharomyces cerevisiae antibodies, zonulin, lipopolysaccharide-binding protein and intestinal alkaline phosphatase (Barber et al., 2019;Cihakova et al., 2019;Gokulakrishnan et al., 2022;Usta et al., 2021). These observations are in accordance with increasing number of studies showing altered gut microbiota profile among individuals with schizophrenia (Borkent et al., 2022;McGuinness et al., 2022). However, it remains unknown as to whether intestinal permeability is increased among individuals with D-SCZ and whether this phenomenon contributes to the pro-inflammatory phenotype. Taking into consideration this research gap, the present study aimed to explore the levels of zonulin and cytokines with respect to clinical characteristics in patients with D-SCZ, those with ND-SCZ and healthy controls. We hypothesized that zonulin levels appear to be increased in patients with D-SCZ compared to other participants and contribute to immune-inflammatory alterations.

Participants
The present study was based on the convenience sample of 119 outpatients with schizophrenia (46 individuals with D-SCZ and 73 individuals with ND-SCZ) and 120 healthy controls. Individuals with schizophrenia were enrolled at two university hospitals in Wroclaw and Szczecin (Poland). The inclusion criteria in the group of patients were: 1) age 18 -65 years; 2) a diagnosis of schizophrenia according to the DSM-IV criteria validated using the OPCRIT checklist (McGuffin et al., 1991); 3) remission of positive and disorganization symptoms based on clinical assessment using the Positive and Negative Syndrome Scale (PANSS) (Kay et al., 1987) and 4) maintenance treatment with stable psychopharmacotherapy over the preceding 6 months.
Healthy controls were recruited from the local community using advertisements. They reported negative family history of mood and psychotic disorders among first-and second-degree relatives. To confirm negative lifetime history of psychiatric disorders, all of them were screened using the Mini-International Neuropsychiatric Interview (M.I. N.I.) (Sheehan et al., 1998).
For all participants, we used the following exclusion criteria: 1) age < 18 years or > 65 years; 2) comorbid substance dependence (except of nicotine and caffeine); 3) unstable physical health impairments; 4) current clinical symptoms of infectious disease; 5) individuals who had been taking anti-inflammatory drugs regularly over the period of preceding 6 months and 6) individuals who had been taking any anti--inflammatory drugs during the preceding 24 h. The study was approved by the Bioethics Committees at Wroclaw Medical University (Wroclaw, Poland) and Pomeranian Medical University (Szczecin, Poland). All participants signed a written informed consent. Findings on the levels of CRP were previously published by our group in a smaller sample overlapping with the one reported in the present study (Cyran et al., 2022).

Clinical assessments
A diagnosis of D-SCZ was established using the Schedule for the Deficit Schizophrenia (SDS) (Kirkpatrick et al., 1989). The SDS is a gold standard tool designed to diagnose D-SCZ according to the characteristics (severity, primary character and enduring presence) of six negative symptoms, i.e., restricted affect, diminished emotional range, poverty of speech, curbing of interests, diminished sense of purpose and diminished social drive. The clinical cutoff to diagnose D-SCZ is the presence of at least two primary negative symptoms with moderate severity and enduring presence (i.e., the symptoms are present during the preceding 12 months and are always present during clinical stability).
Additionally, clinical manifestation was rated using the following scales: 1) the PANSS; 2) the Calgary Depression Scale for Schizophrenia (CDSS) (Addington et al., 1993) and 3) the Social and Occupational Functioning Assessment Scale (SOFAS) (Smith et al., 2011). All participants underwent examination of cognitive performance with the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (Randolph et al., 1998). The RBANS is a set of 15 tasks that measure performance across five cognitive domains (i.e., immediate memory, visuospatial/constructional abilities, language, attention and delayed memory).

Dietary habits and cigarette smoking
To control for the effects of diet, we used the Food Frequency Questionnaire-6 (FFQ-6) (Niedzwiedzka et al., 2019). It is a 62-item interview that was developed to measure the frequency of consuming various food products over the period of preceding 12 months. In this study, we focused on the adherence to the Mediterranean diet using the aMED score (Hawrysz et al., 2020;Krusinska et al., 2018). The aMED score (range: 0 -8; higher scores are indicative of greater adherence to the Mediterranean diet) is based on the consumption frequency of vegetables, fruits, whole grains, fish, legumes, nuts and seeds, red and processed meet as well as the ratio of vegetable oils to animal fat.

Peripheral blood markers
Fasting blood samples were collected from all participants (between 7 a.m. and 9 a.m.). Next, all samples were centrifuged (at 2500 rpm for 10 min) to obtain serum that was stored in aliquots at − 80 • C (250 µl in each aliquot).
The levels of zonulin were determined using the IDK® Zonulin Table 2 The levels of immune-inflammatory markers and zonulin in patients with schizophrenia and healthy controls.

Data expressed as mean (SD)
All concentrations provided in pg/ml except of CRP levels (mg/l) and zonulin levels (ng/ml) *Significant differences after the Benjamini-Hochberg correction # Significant differences after adjustment for sex, BMI, the FTND score and the aMED score Abbreviations: GM-CSF, granulocyte macrophage colony-stimulating factor; IFN, interferon; IL, interleukin; IL-1RA, IL-1 receptor antagonist; IL-2R, IL-2 receptor; IP-10, IFN-γ induced protein-10; MCP-1, monocyte chemoattractant protein-1; MIG, monokine induced by IFN-γ; MIP, macrophage inflammatory protein; TNF-α, tumour necrosis factor-α (Serum) ELISA kit (Immundiagnostik AG, Bensheim, Germany, batch No. K5601-150513 IVD, CE) according to the manufacturer's recommendations. The assay was produced and distributed according to the IVD guidelines of 98/79/EC. The assay is based on the competitive binding technique. Biotinylated zonulin tracer was added to the samples, standards, and controls as a competitor to the sample's own zonulin. The intensity of the color was inversely correlated with the zonulin concentration in the sample. Samples were read at 450 nm. The fourparameter algorithm was used to form the standard curve and to analyze data. All tests were carried out in duplicates. The limit of blank (LoB), detection (LoD) and quantitation (LoQ) were 0.140 ng/ml, 0.183 ng/ml, and 0.183 ng/ml, respectively. The evaluation was performed according to the CLSI guideline EP17-A2. The specified accuracy goal for the LoQ was 20% CV.

Data analysis
To avoid the confounding effect of clinically relevant inflammation, we excluded participants with CRP levels > 10 mg/l (4 individuals with D-SCZ, 2 individuals with ND-SCZ and 7 healthy controls). Normality of data distribution was tested using the Kolmogorov-Smirnov test.
Bivariate differences were assessed using the following tests: 1) the χ 2 test (categorical variables); 2) the Mann-Whitney U test (continuous variables across two groups, non-normal distribution); 3) the t-tests (continuous variables across two groups, normal distribution); 4) the Kruskal-Wallis test (continuous variables across three groups, nonnormal distribution) and 5) one-way analysis of variance (continuous variables across three groups, normal distribution). In case of the Kruskal-Wallis test and one-way analysis of variance, additional posthoc comparisons were carried out using the Bonferroni test. Bivariate correlations were analyzed with the Spearman rank correlation coefficients. The Benjamini-Hochberg correction with the false discovery rate set at 25% was applied to control for multiple testing. Significant between-group differences in tested biomarkers and their correlations with clinical characteristics (after adjustment for multiple testing) were further analyzed using the analysis of co-variance (ANCOVA) and partial correlation coefficients. In these analyses, the levels of biomarkers were converted to z-scores. Results were interpreted as significant if the p-value was lower than 0.05. Data analyses were performed in the SPSS, version 28.

The general characteristics of the sample
There were no significant between-group differences with respect to age (Table 1). However, individuals with D-SCZ were significantly more likely to be males and had significantly higher BMI compared to other groups of participants. Individuals with D-SCZ and those with ND-SCZ had significantly lower rates of higher education level compared to healthy controls. Also, both groups of patients scored significantly lower across all RBANS domains than healthy controls. Importantly, individuals with D-SCZ outperformed those with ND-SCZ on tasks measuring immediate and delayed memory as well as attention. As expected, individuals with D-SCZ had significantly higher PANSS scores of negative symptoms and significantly lower SOFAS scores compared to those with ND-SCZ. No significant between-group differences with respect to the measures of medication effects were found. (Table 2).

Between-group differences in the levels of zonulin and immuneinflammatory markers
After adjustment for multiple testing, significant between-group differences with respect to the levels of the following immuneinflammatory markers were found: IL-1β, IL-10, IL-13, IL-6, RANTES, IL-17, IFN-γ, IFN-α, TNF-α, IP-10, IL-4 and CRP. Importantly, no significant between-group differences in the levels of zonulin were observed. The analysis of co-variance (ANCOVA) revealed that the following differences in the levels of immune-inflammatory markers were significant after controlling for the effects of sex, BMI, the aMED score and the FTND score: IL-1β, RANTES, IL-17, TNF-α and CRP. Specifically, the levels of IL-1β (F = 4.342, p = 0.014) and CRP (F = 3.309, p = 0.038) were significantly higher in patients with D-SCZ compared to other groups of participants. In turn, the levels of RANTES (F = 5.632, p = 0.004) and TNF-α (F = 5.079, p = 0.007) were significantly higher in both groups of individuals with schizophrenia. Finally, the levels of IL-17 were significantly higher in patients with D-SCZ compared to healthy controls (F = 4.571, p = 0.011).

Correlations of immune-inflammatory markers with clinical characteristics
Correlations were tested for immune-inflammatory markers showing altered levels in patients with D-SCZ (Table 3). In the whole group of participants, the following correlations were significant after adjustment for multiple testing: 1) the correlation of IL-1β levels with the PANSS negative symptoms score and the RBANS score of attention; 2) the correlation of RANTES levels with the RBANS score of visuospatial/ constructional abilities; 3) the correlation of TNF-α levels with the RBANS scores of immediate memory, visuospatial/constructional abilities, attention and delayed memory and 4) the correlation of CRP levels with the RBANS scores of immediate memory and attention. The analysis of partial correlations controlling for the effects of age, the level of education and CPZeq demonstrated significant associations (Fig. 1) of IL-1β with the PANSS score of negative symptoms (r = 0.385, p < 0.001) and the RBANS score of attention (r = -0.247, p = 0.011). Also, the correlation of CRP levels with the RBANS score of attention remained Table 3 Correlations of immune-inflammatory markers with clinical characteristics.

Discussion
To our knowledge this is the largest study investigating the immuneinflammatory profile of D-SCZ. Moreover, this is the first study investigating zonulin levels with respect to the occurrence of primary and persistent negative symptoms of schizophrenia. In this study, we found that subclinical inflammation might appear with greater extent in patients with D-SCZ compared to those with ND-SCZ even when potential confounding factors are taken into consideration. Specifically, we found that individuals with D-SCZ show increased levels of IL-1β and CRP compared to those with ND-SCZ and healthy controls. Importantly, only the levels of IL-1β were significantly and positively correlated with the severity of negative symptoms. Also, higher levels of IL-17 were observed in patients with D-SCZ compared to healthy controls. The levels of two immune-inflammatory markers, i.e., RANTES and TNF-α were elevated to the similar extent in both groups of patients with schizophrenia. However, contrary to our hypothesis, we did not find that zonulin levels are significantly altered in patients with D-SCZ compared to those with ND-SCZ or healthy controls. In other words, our findings do not indicate that increased gut permeability accounts for the greater extent of subclinical inflammation in D-SCZ.
In general, our findings are in line with those obtained by some previous studies. For instance, Dai et al. (2020) found elevated levels of IL-1β in drug-naïve individuals with first-episode schizophrenia and predominant negative symptoms. Another study of stable outpatients with schizophrenia also demonstrated a significant positive correlation between IL-1β levels and the severity of negative symptoms (Gonzalez-Blanco et al., 2019). Finally, elevated levels of CRP in patients with D-SCZ, compared to those with ND-SCZ and healthy controls, have also been reported by previous studies (Pan et al., 2020;Wang et al., 2022).
Nevertheless, certain discrepancies with previous studies should also be noted. Indeed, as opposed to our results, Wang et al. (2022) found elevated levels of IL-6 and IL-8 in patients with D-SCZ, but not in those with ND-SCZ. Other studies also demonstrated elevated levels of TNF-α and IL-6 in patients with D-SCZ compared to those with ND-SCZ and healthy controls (Goldsmith et al., 2018;Maes et al., 2020). These discrepancies might originate from sample size and methodological differences related to adjustment for specific covariates or the use of specific diagnostic instruments. Indeed, the majority of previous studies used the proxy measures to diagnose D-SCZ (Goldsmith et al., 2018;Wang et al., 2022) that might be characterized by low face validity and temporal stability (Galderisi et al., 2021).
It needs to be highlighted that our findings indicate the involvement of immune-inflammatory processes in both subtypes of schizophrenia. However, there might be differences in specific alterations of immuneinflammatory pathways involved in the pathophysiology of D-SCZ and ND-SCZ. According to our findings, both subtypes of schizophrenia might be characterized by elevated levels of RANTES and TNF-α. Importantly, RANTES represents chemokines and promotes migration of leukocytes to the site of inflammation as well as contributes to the activation of T cells (Appay and Rowland-Jones, 2001). It might be produced by a variety of the central nervous system cells and modulates the release of glutamate (Lanfranco et al., 2017;Musante et al., 2008). It has also been found that increased production of RANTES might mediate the effects of maternal immune activation on the brain development in mice (Openshaw et al., 2019). Elevated levels of RANTES were reported by our group in patients with chronic schizophrenia from an independent sample . In turn, elevated levels of TNF-α have been widely reported in schizophrenia. Previous meta-analyses revealed that increased TNF-α levels might be classified as one of trait markers of schizophrenia characterized by elevated levels during acute exacerbations of psychosis and after antipsychotic treatment (Capuzzi et al., 2017;Miller et al., 2011). Importantly, higher levels of CRP and IL-1β were associated with lower performance of attention in the whole group of participants. At this point, it is also important to note significantly lower performance of attention in patients with D-SCZ compared to those with ND-SCZ and healthy controls from our study. Similar findings were obtained by Baek et al. (2022), who found a significant negative correlation between the levels of IL-β and the scores of sustained attention tasks in patients with first-episode schizophrenia. Moreover, the most recent meta-analysis demonstrated that elevated levels of CRP, IL-1β, IL-6 and TNF-α are associated with cognitive impairment in patients with schizophrenia (Patlola et al., 2023). Importantly, this meta-analysis demonstrated the largest pooled effect size estimates for IL-1β. Various mechanisms can underly the associations between subclinical inflammation and cognition that were observed in our study. Importantly, the RBANS attention score is the sum of scores obtained from two tasks, i.e., digit span and coding. Apart from attention, these tasks measure working memory and psychomotor speed. It has been shown that the intrahippocampal injection of IL-1β impairs working memory in rats (Matsumoto et al., 2004;Matsumoto et al., 2001) and there is evidence that IL-1β might suppress neurogenesis (Kaneko et al., 2006). However, it has also been shown that the effects of IL-1β on cognition might be dose-dependent with lower concentrations improving hippocampal-dependent cognition and higher concentrations exerting detrimental effects on memory processes (Goshen et al., 2007).
Certain limitations of the present study need to be discussed. First, although we investigated a variety of immune-inflammatory markers, assessment of gut permeability was limited to the measurement of zonulin levels. Also, none of markers related to dysfunction of the blood-brain barrier was assessed. Therefore, our observation that subclinical inflammation in D-SCZ cannot be attributed to increased gut permeability should be interpreted with caution. Second, although we controlled for the effects of various confounding factors, it cannot be excluded that some lifestyle characteristics, environmental exposures and comorbid physical health impairments had influenced our results. These include, i.e., physical activity, medication effects, comorbid cardiovascular diseases. However, it should be noted that both groups of patients did not differ significantly in terms of receiving specific groups of psychiatric medications and the dosage of antipsychotics. At this point, it is also important to note that assessment of dietary habits over the period of preceding 12 months with self-reports might be characterized by a recall bias. Third, representativeness of our findings might also be limited as we used the convenience sample of outpatients recruited at two clinical sites. Fourth, due to the lack of longitudinal design, conclusions about causal associations are not possible to be established.
In sum, findings from the present study indicate that subclinical inflammation occurs in both subtypes of schizophrenia, i.e., D-SCZ and ND-SCZ, and might be associated with cognitive impairment. However, this phenomenon might appear with greater extent in patients with D-SCZ. Although our findings do not indicate that subclinical inflammation in D-SCZ is secondary to increased gut permeability, additional studies using markers other than zonulin are needed to confirm this observation. Future studies in the field need to elaborate clinical heterogeneity in order to better understand the contribution of aberrant immune-inflammatory processes to the pathophysiology of schizophrenia. This perspective might hold translational importance for the development of novel and personalized treatments targeting subclinical inflammation in schizophrenia.