A total of 400 subjects were enrolled in the present study and divided into two groups: i) 200 patients diagnosed with SZ attending a psychiatric clinic at King Abdullah University Hospital and Princess Basma Teaching Hospital (Irbid, Jordan); ii) 200 healthy controls free of any psychosis-related symptoms attending the National Center for Diabetes Endocrinology and Genetics (Amman, Jordan) for routine health care. The present study was approved by the Institutional Review Board Committee (approval no. 2019/626) of Jordan University of Science and Technology (Irbid, Jordan) and written informed consent was obtained from all participants before enrollment in the present study.

After collecting the consent forms signed by all of the enrolled patients with SZ and controls, two peripheral blood samples were collected in plain (5 ml) and EDTA (4 ml) tubes. The samples were collected between May 1, 2020 and September 30, 2021. Serum was separated after centrifuging blood samples in plain tubes at 10,000 x g for 10 min at 4˚C and was stored at -80˚C for later use to measure VD concentration. Notably, hemolyzed samples were excluded from the present study. Blood samples in EDTA tubes were used for DNA extraction and further analysis.

Patients were diagnosed by a proficient psychiatrist according to the diagnostic criteria for SZ based on the ICD-10 (DSM-V) (35). Experienced psychiatrists conducted these diagnoses, following standard guidelines to ensure accuracy. Patients were treated according to the latest best practices and medical standards as outlined by the guidelines issued by the American Psychiatric Association. The participants who met the following criteria were included: i) Either sex between 18-60 years old; ii) sporadic and familial cases; iii) no history of other mental disorders; iv) no history of blood transfusion within 1 month; v) no physical and nervous system diseases, such as brain trauma; vi) no history of alcohol abuse and drug abuse. The exclusion criteria were: i) Presence of other medical conditions, which may produce psychotic SZ-related symptoms, such as epilepsy, metabolic disturbance, brain lesions, limbic encephalitis, stroke, multiple sclerosis and dementia; ii) presence of other diseases or medications known to affect VD, such as arthritis, osteoporosis, end-stage renal disease, hypothyroidism, rickets, corticosteroid therapy and malabsorption syndromes; and iii) individuals with drug-induced psychosis, acquired brain injuries and intellectual disabilities.

Genomic DNA was extracted using the QIAamp DNA Mini Kit (cat. no. 51304; Qiagen GmbH) according to the manufacturer's instructions. DNA concentration was measured using a Nanodrop (Thermo Fisher Scientific, Inc.) and integrity was verified using 2% agarose gel electrophoresis.

For SNP genotyping, two sets of allele-specific primers for each SNP were designed using the PRIMER1 tool (http://primer1.soton.ac.uk/primer1.html). Tetra-amplification refractory mutation system-polymerase chain reaction (T-ARMS-PCR) was carried out for rs10741657, rs10877012, rs4646536, rs6013897, rs2228570, rs1544410, rs731236 and rs7975232 genotyping. The primer sequences are listed in Table SI. The annealing temperature for each primer set was optimized through gradient PCR that was carried out on a Veriti™ Dx 96-well Thermal Cycler (Thermo Fisher Scientific, Inc.). Each PCR was carried out in a total volume of 20 µl, containing 4 µl HOT FIREPol^® Blend Master Mix (cat. no. 04-27-00115; Solis BioDyne), 0.5 µl each primer, 1 µl DNA template and 13 µl nuclease-free water. The PCR cycling conditions were as follows: Initial denaturation at 95˚C for 10 min, followed by 35 cycles of denaturation at 95˚C for 15 sec, annealing at 61-64˚C for 30 sec, extension at 72˚C for 2 min, and a final extension at 72˚C for 10 min. Subsequently, 10 µl PCR products were loaded onto a 3% agarose gel, and a 100-bp ladder (cat. no. MBT049; HiMedia Laboratories) was used for size comparison between DNA fragments. For gel visualization, a UV transilluminator (Cleaver Scientific Ltd.) was used.

Serum 25(OH)D levels were determined using the commercially available Roche Elecsys Vitamin D total II electrochemiluminescence immunoassay and Cobas E801 auto analyzer (Roche Diagnostics GmbH).

Statistical analysis was carried out using SPSS software (Version 22; IBM Inc.). The genotype frequencies of all SNPs were compared using a χ² test or Fisher's exact test when >20% of expected cell counts are <5. Logistic regression analysis was performed to determine the odds ratio (OR) and 95% confidence interval associated with SZ risk, with the control considered the reference group. Each polymorphism was tested for Hardy-Weinberg equilibrium (HWE) using χ² test or Fisher's exact test. P<0.05 was considered to indicate a statistically significant difference.

Moreover, a total of 370 cases were included for the receiver operating characteristic (ROC) curve analysis after excluding individuals with missing genotype data for any of the investigated SNPs. For the generalized linear model, the analysis was conducted on a subset of 251 cases due to a lack of VD measurement data or genotype data for one of the studied SNPs. ROC curve analysis was performed using the classify module in SPSS software, encompassing all studied SNPs, to predict the presence of SZ. Subsequently, the generalized linear model was employed, utilizing binary logistic regression in the type of model menu. The dependent variable in the response menu was set as the status of the samples, referencing the control samples. The predictor menu included the three studied SNPs as factors, while VD levels and age served as covariates. The wild-type genotype for each SNP was designated as the baseline category, coded as 1. Within the model menu, the main effect for all SNPs and VD was specified. Finally, the likelihood ratio test with profile likelihood was utilized for model evaluation. These comprehensive methods were implemented to investigate the predictive utility of the included SNPs through ROC curve analysis and to explore the association between genetic variants, VD levels, age and SZ status using generalized linear modeling techniques.

Furthermore, SRplot (http://www.bioinformatics.com.cn/en) was used to perform Mann-Whitney U test when comparing VD levels between patients with SZ and controls, which included 272 cases in total. The effect of sex variation between groups on the genotyping frequency was assessed using Pearson's χ² test or Fisher's exact test. Linkage disequilibrium (LD) analysis was conducted using the SNP linkage LD heatmap module available on SRplot (http://www.bioinformatics.com.cn/srplot). This module calculates pairwise LD statistics, measured by R², between SNPs. These statistical data are visually represented in a triangular heatmap, where the extent of LD between SNP pairs is indicated through a color-coded scale. The color key is used to denote R² values, enhancing the visual interpretation of LD strength. Additionally, the heatmap integrates gene models with SNP sites marked by colored asterisks, providing a clear genetic landscape and facilitating the identification of regions with strong LD. This graphical representation allows for an efficient analysis and easy interpretation of the LD patterns across the genomic regions studied.

A summary of the demographic data of the patients with SZ and controls is presented in Table I. The average age of the control group was ~57 years (range, 18-76 years), while it was 42 years in the SZ group (range, 19-78 years). The case and control groups exhibited a significant difference in their mean age. Among the SZ group, there were 36 female and 164 male patients and in the control group, there were 158 female and 42 male healthy individuals.

T-ARMS-PCR was carried out to amplify the DNA fragments of the eight SNPs in VD metabolic pathway genes, including CYP2R1 (rs10741657), CYP27B1 (rs10877012 and rs4646536), CYP24A1 (rs6013897) and VDR (rs2228570, rs1544410, rs731236 and rs7975232). T-ARMS-PCR is a genotyping method designed to detect SNPs by utilizing two pairs of primers in a single PCR reaction: One pair flanking the SNP (outer primers) and another pair that specifically anneals depending on the allele presence (allele-specific or inner primers). This technique generates two products per allele: One common product and one allele-specific product, which allows for the determination of the zygosity of the sample directly by gel electrophoresis (36). While T-ARMS-PCR is efficient for detecting known variants, Sanger sequencing is essential for validation. Due to budget constraints, the cost of sequencing and limited local sequencing facilities, Sanger sequencing was not possible for utilization in the present study. Therefore, the T-ARMS-PCR method was utilized as a practical alternative for variant detection. Fig. S1, Fig. S2, Fig. S3, Fig. S4, Fig. S5, Fig. S6, Fig. S7 and Fig. S8 illustrate the T-ARMS-PCR genotyping results for the studied SNPs.

CYP2R1 (rs10741657; A>G). The genotype and allele frequencies of rs10741657 among all participants are summarized in Tables II and SII, respectively. Statistical analysis of the results revealed a significant difference in genotype and allele frequencies among patients and controls (P<0.0001). The findings also suggested that the AA genotype and the A allele were more prevalent among patients with SZ compared with the controls.

CYP27B1 (rs10877012; G>T). The distribution of rs10877012 genotypes and alleles among patients with SZ and controls are shown in Tables II and SII. A significant difference was shown in both genotype and allele frequencies between the two groups (P<0.0001). The results also indicated a higher prevalence of the TT genotype and the T allele in patients with SZ compared with the controls.

CYP27B1 (rs4646536; A>G). The genotype and allele frequencies among patients with SZ and controls are revealed in Tables II and SII. Statistical analysis of the data revealed no significant differences in both genotype and allele frequencies between the two groups (P=0.18 and 0.165, respectively).

CYP24A1 (rs6013897; T>A). The genotype and allele distributions of the rs6013897 SNP in patients with SZ and controls are shown in Tables II and SII. The AA genotype was significantly more frequent in the patient group (P<0.0001). Regarding the allele frequency, there was no statistically significant association identified; however, the A allele appeared to be slightly more frequent in the patient group.

VDR (rs2228570; A>G). The frequencies of the rs2228570 genotypes and alleles among patients with SZ and controls are revealed in Tables II and SII. The genotype and allele distributions were not significantly different between patients and controls (P=0.81 and 0.26, respectively).

VDR (rs1544410; C>T). The genotype and allele frequencies for the rs1544410 SNP are presented in Tables II and SII. Data analysis revealed no significant differences between patients and controls regarding both genotype and allele distributions (P=0.45 and 0.88, respectively).

VDR (rs731236; A>G). The frequencies of genotypes and alleles for rs731236 among the two groups are summarized in Tables II and SII. Data analysis revealed no significant differences between patients and controls in both genotype and allele distributions (P=0.48 and 0.51, respectively).

VDR (rs7975232; C>A). The genotype and allele frequencies of the rs7975232 SNP are illustrated in Tables II and SII. Statistical analysis of genotype and allele frequencies revealed no significant difference between patients with SZ and controls (P=0.15 and 0.21, respectively).

Based on the findings obtained, three SNPs: CYP2R1 (rs10741657; A>G), CYP27B1 (rs10877012; G>T) and CYP24A1 (rs6013897; T>A) exhibited statistically significant differences between the two groups. To evaluate the impact of sex variation on the results, a Pearson's χ² test or Fisher's exact test was conducted, which revealed no significant differences in the frequency of SNPs among controls (male and female controls): χ²=0.685, P=0.953 for CYP24A1 (rs6013897; T>A); χ²=1.20, P=0.549 for CYP27B1 (rs10877012; G>T); and χ²=2.92, P=0.571 for CYP2R1 (rs10741657; A>G). Similarly, there were no significant differences in the frequency of these SNPs among patients (male and female patients): χ²=0.449 (P=0.799) for CYP24A1 (rs6013897; T>A); χ²=1.90 (P=0.387) for CYP27B1 (rs10877012; G>T); and χ²=1.48 (P=0.477) for CYP2R1 (rs10741657; A>G) (Table III). These results confirmed that male or female sex does not have a significant impact on the genotype frequency between patients and controls.

Differences between observed and expected genotype frequencies for each SNP were determined by χ² test or Fisher's exact test to assess the deviation from HWE and to identify any possible genotyping error that could exist. As demonstrated in Table IV, the genotype and allele frequency of five SNPs (rs10741657, rs10877012, rs1544410, rs731236 and rs7975232) were in HWE. However, two SNPs (rs6013897 and rs2228570) deviated significantly from HWE in both cases and controls.

The Mann-Whitney U test was utilized to compare VD levels between individuals with SZ and controls, as shown in Fig. 1. The results demonstrated that VD levels were significantly lower (P<0.0001) in patients with SZ (13.8 ng/ml) compared with those in the control group (31.3 ng/ml), indicating that the levels of VD varied significantly between the two groups.

The area under the curve (AUC) analysis was conducted to assess the predictive performance of genetic variants in distinguishing between patients with SZ and controls (Fig. 2). The AUC values for each test result variable (rs2228570, rs1544410, rs731236, rs7975232, rs6013897, rs10877012, rs4646536 and rs10741657) ranged from 0.422 to 0.615. Notably, rs10741657 exhibited the highest AUC value of 0.615, indicating improved discriminative ability compared with the other variants, whereas rs6013897 showed the lowest AUC value of 0.422. The statistical significance of the AUC values varied across the tested variants, with rs10877012 and rs10741657 demonstrating significant discriminative abilities (P=0.003 and P<0.001, respectively). These findings suggested varying levels of predictive power among the tested genetic variants in distinguishing schizophrenic states, with rs10741657 revealing the most promising discriminatory performance.

The binary logistic model revealed significant associations between genetic variants, VD levels and the likelihood of having SZ, as shown in Table V. Notably, rs10741657, rs10877012 and rs6013897 exhibited an increased likelihood of developing SZ. For rs10741657, the AA genotype had an OR of 4.911, although this was not significant (P=0.093), while the TA genotype had an OR of 2.497 (P=0.022). Regarding rs10877012, the GT genotype had an OR of 2.369 (P=0.087), however this was not statistically significant. For rs6013897, the AA genotype had an OR of 13.087 (P=0.096) without statistical power, while the TA genotype had an OR of 0.267 (P<0.001) indicating a protective effect and suggesting that having the variant decreased the probability of having SZ. Additionally, increased VD levels were revealed to be associated with a lower probability of developing SZ, with an OR of 0.888 (P<0.001). Finally, the overall model exhibited a strong statistical significance (χ²=77.209, P<0.001) and good fit (Hosmer and Lemeshow goodness-of-fit test: χ²=4.451, P=0.814), underscoring the robustness of the associations identified (data not shown).

LD analysis between all eight SNPs is shown in Fig. S9. The LD heatmap revealed that these SNPs do not exhibit significant associations, suggesting an independent inheritance within the study population. Notably, no SNP pairs demonstrated high LD (R² values close to 1.0), which would indicate a tendency to be co-inherited.

Given the lack of significant LD, the haplotype construction from these SNPs would likely result in arbitrary combinations of alleles, as these do not represent true biological interactions. This undermines the utility of haplotype analysis in this context, as all SNPs do not appear to influence the phenotype collectively. Instead, each SNP contributes independently, which aligns with the observed distribution of allele frequencies and LD patterns among the studied SNPs.