Modern therapy for schizophrenia, based on the use of antipsychotic drugs, is an effective treatment method for many patients. However, according to various studies, 20 to 30% of patients with schizophrenia do not respond to treatment or have an unsatisfactory response [1].

The reason for the ineffectiveness of schizophrenia treatment may be genetic variability, which, along with other factors, determines individual characteristics of response to drug therapy for schizophrenia [2]. It has been established that both common and rare variants of genes associated with the risk of schizophrenia, antipsychotic metabolism, adverse drug reactions, as well as gene-environment interactions depending on ethnic background, contribute to the genetic determination of response to drug therapy [3, 4].

Pharmacogenetic testing aids in making clinically significant decisions when choosing patient treatment strategies in cases where antipsychotic therapy fails to achieve the expected clinical effect and there is a risk of developing chronic and highly disabling pharmacoresistant schizophrenia [5, 6]. To facilitate the implementation of pharmacogenetics into clinical practice, guidelines are being developed with evidence-based recommendations for optimizing pharmacotherapy, describing gene-drug interactions between genes and antipsychotics [7]. Response to drug therapy is a complex human trait, where the magnitude of the therapeutic effect, given intra- and inter-population polygenic variability, depends on the cumulative small effects of gene-gene and gene-environment interactions along the response pathway [3, 4, 8–10]. The neurobiological mechanisms of therapy-resistant schizophrenia remain poorly understood. Attempts are being made to investigate its polygenic architecture, identify reliably associated specific gene variants, and establish the genetic link between the pathophysiology of schizophrenia and the mechanisms of action of antipsychotic drugs [9, 11]. Therefore, there are currently no specific pharmacogenetic tests for pharmacoresistant schizophrenia. Pharmacogenetic markers with proven impact on the efficacy of schizophrenia therapy include polymorphic loci of the CYP2D6 and CYP2C19 genes [12, 13]. Their genetic variants influence the functioning of the cytochrome P450 enzymes they encode. The algorithms for antipsychotic therapy proposed by the Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for psychiatry are based on these variants [6]. Regarding the response to antipsychotic therapy for schizophrenia, the role of many polymorphic variants of genes such as COMT, MAOA/B, CYP1A2, ABCB1, LINC01795, DDHD2, SBNO1, KCNG2, SEMA7A, RUFY1, and others has been discussed [6, 14, 15], including in specific clinical cases [6, 16]. Currently, many researchers agree that for each distinct population and ethnic sample, the search remains relevant for those significant pharmacogenes whose polymorphic loci may serve as reliable biomarkers of complex gene-environment interaction effects in shaping the overall response to antipsychotic therapy.

Objective of the study was to investigate the association of a panel of polymorphic gene loci with pharmacoresistance in patients with schizophrenia among residents of Belarus.

Participants and study design. This prospective comparative clinical study employed a targeted patient selection method and included 161 individuals with schizophrenia (ICD-10 code: F20) undergoing inpatient treatment at the Republican Scientific and Practical Center for Mental Health. Patient selection criteria included: diagnosis of schizophrenia (ICD-10 code: F20), age 18–60 years, signed informed consent to participate in the study, absence of concomitant somatic diseases in the acute phase, and absence of acute infectious diseases.

The main group of pharmacoresistant patients consisted of 104 individuals aged 18 to 60 years (38.0±1.0 years): 62 women, 42 men; the comparison group included 57 patients without pharmacoresistance, aged 18 to 60 years (41.0±1.1 years): 27 women, 30 men. The groups were homogeneous in terms of age, characteristics of the underlying disease, and social status. All patients received treatment with psychotropic medications according to the clinical protocol for providing medical care to patients with mental and behavioral disorders. Pharmacoresistance was defined as the absence of improvement in psychopathological symptoms despite oral administration of adequate medication doses for 6 to 8 weeks, during which two or more antipsychotics were used, with at least one being atypical [17].

Clinical methods. Sociodemographic data (sex, age, age at disease onset), family history of mental illness, number of hospitalizations over the last 10 years, and smoking status were collected for each patient. The assessment of patients with schizophrenia spectrum disorders was based on the following scales: Clinical Global Impression – Global Improvement scale (CGI); psychometric assessment using the Scale for the Assessment of Negative Symptoms (SANS) and the Scale for the Assessment of Positive Symptoms (SAPS); assessment of functional impairment in different social domains; and the Extrapyramidal Symptom Rating Scale (ESRS-A).

Biological material for the study. Saliva was used as the biological material for genomic DNA extraction, collected using the Oragene OG-500 saliva collection system (DNAgenotek, USA). Sample collection was carried out at the clinical diagnostic laboratory and clinical inpatient departments of the Republican Scientific and Practical Center for Mental Health.

Pharmacogenetic analysis. Genotyping was performed for patients included in the study for polymorphic loci of cytochrome P-450 isoenzyme genes: CYP2D63, CYP2D64, CYP2D610, CYP2C92, CYP2C93, CYP2C192, CYP2C1917, CYP1A2F; the C3435T locus of the MDR1 gene (encoding the P-glycoprotein transport protein); as well as polymorphic loci of drug target genes and functionally related proteins: the TaqI polymorphism of ANKK1 (ankyrin kinase 1 gene), the C-1019G polymorphism of the serotonin receptor gene HTR1A, the 1438G>A polymorphism of the serotonin receptor gene HTR2A, the 5-HTTLPR polymorphism of the serotonin transporter gene SLC6A4, the rs1414334 polymorphism of the serotonin receptor gene HTR2C, the Val158Met polymorphism of the COMT gene (catechol-O-methyltransferase), the uVNTR polymorphism of the MAOA gene (monoamine oxidase A), the Val66Met polymorphism of the BDNF gene (brain-derived neurotrophic factor), the C957T polymorphism of the dopamine receptor gene DRD2, and the UGT1A1*28 polymorphism of the uridine diphosphate glucuronosyltransferase UGT1A1 gene. Genomic DNA from saliva was isolated using the "DNA-BK" kit (Institute of Bioorganic Chemistry, National Academy of Sciences of Belarus, Belarus). Pharmacogenetic testing was conducted at the Institute of Bioorganic Chemistry, National Academy of Sciences of Belarus, using the "OLIGO-GENFARM" diagnostic oligonucleotide kit for determining genetic markers of pharmacoresistance to psychotropic drugs (Institute of Bioorganic Chemistry, National Academy of Sciences of Belarus, Belarus). PCR and real-time PCR were performed on a CFX96 instrument (BioRad, USA). Restriction fragment length polymorphism analysis used restriction endonucleases from NEB (England): AvaII (*CYP2C92 rs1799853), KpnI (CYP2C93 rs1057910), SmaI (CYP2C192 rs4244285), KpnI (CYP2C1917 rs12248560), BstNI (CYP2D64 rs3892097), ApaI (CYP1A21F rs762551), MboI (MDR1* rs1045642), BtsCI (HTR1A rs6295), followed by separation of restriction products in agarose/acrylamide gel according to the "OLIGO-GENFARM" kit instructions. Markers TaqI (rs1800497) of the ANKK1 gene, 1438G>A (rs6311) of the HTR2A gene, rs1414334 of the HTR2C gene, Val158Met (rs4680) of the COMT gene, Val66Met (rs6265) of the BDNF gene, C957T (rs6277) of the DRD2 gene, UGT1A1*28 (rs8175347), and CYP2D63 (rs35742686) were detected by real-time PCR according to the "OLIGO-GENFARM" kit instructions. The 5-HTTLPR polymorphism of the SLC6A4 gene and the 30bp uVNTR polymorphism of the MAOA gene were determined by amplified fragment length analysis using capillary electrophoresis on an AB3500 genetic analyzer (Applied Biosystems, USA).

Statistical analysis. SPSS Statistics 20.0 software was used for the preparation and analysis of clinical data and genotyping data. Odds ratios (OR) and 95% confidence intervals (CI) were used as measures of association between genotypes and the risk of developing a specific phenotype.

This study included 161 patients with schizophrenia (89 women and 72 men), of whom 104 patients were resistant to antipsychotic pharmacotherapy. The control group of treatment-responsive patients consisted of 57 individuals. For the treatment of schizophrenia patients, the most frequently used antipsychotic drug was the atypical antipsychotic clozapine (42.1%); a combination of atypical and typical antipsychotics was also commonly used (39.5%) (Table 1).

Table 1. Frequency of prescribing medications to patients with schizophrenia

MedicationPrescription frequency, %Quetiapine3.5Clozapine42.1Olanzapine5.3Risperidone5.3Haloperidol1.8Zuclopenthixol1.8Fluphenazine0.9Multiple antipsychotics39.5TotalN=161 (100.0)

Genotyping was performed for 19 polymorphic loci in 15 genes. Table 2 indicates the putative risk alleles and the genotypes containing them for each polymorphic locus of the studied genes, according to literature data. In the total patient sample, we analyzed allele frequency distribution to check for compliance with the Hardy-Weinberg equilibrium. Hardy-Weinberg equilibrium was observed for all studied loci. To investigate the association of polymorphic gene loci with pharmacoresistance (PR), an association analysis was performed comparing the groups.

*Note: Statistically significant differences (p < 0.05).

Analysis of genotype and allele frequencies revealed an association of PR with three polymorphic loci: CYP2D6 (rs3892097), HTR1A (rs6295), and COMT (rs4680) (Table 2). PR was associated with the A allele (CYP2D64) (rs3892097) of the CYP2D6 gene (χ²=4.124, p=0.042), the G allele of the serotonin receptor gene HTR1A (rs6295) (χ²=5.522, p=0.019), and the G allele of the COMT gene (rs4680) (χ²=9.006, p=0.003) (Table 2).

The risk of developing PR was identified in carriers of the A allele and the AG genotype at the *4* polymorphic locus (rs3892097) of the CYP2D6 gene (OR 2.159; 95% CI 1.018-4.578 and OR 2.975; 95% CI 1.269-6.977, respectively), the CC genotype (rs6295) of the HTR1A gene (OR 2.703; 95% CI 1.151-6.348), and the G allele (rs4680) of the COMT gene (OR 2.786; 95% CI 1.414-5.489). Protective effects against the risk of developing PR were found for the GG genotype (rs3892097) of the CYP2D6 gene (OR 0.463; 95% CI 0.218-0.928), the G allele (rs6295) of the HTR1A gene (OR 0.367; 95% CI 0.156-0.864), and the AA (Met/Met) genotype (rs4680) of the COMT gene (OR 0.359; 95% CI 0.182-0.707).

To further investigate the association of PR risk with combinations of alleles and genotypes of other polymorphic loci, a comparative analysis of the frequency of occurrence of polymorphic loci of the CYP2D6, HTR1A, COMT genes in combination with polymorphic loci of other genes included in the study was performed. Combinations significantly associated with the PR trait are presented in Table 3.

| :--- | :--- | :--- | :--- | :--- || |  (n) | % |  (n) | % | |

It was found that certain combinations of alleles or genotypes at polymorphic gene loci can significantly increase the risk of PR. An increased OR for PR risk was observed with the combination of the CYP2D6 A allele with the CYP1A2 A allele (OR 2.926; 95% CI 1.206-7.102); with the MDR1 T allele (OR 3.249; 95% CI 1.417-7.448); and with the CYP1A2 A and MDR1 T alleles together (OR 4.833; 95% CI 1.753-13.328). For the COMT G (Val) allele, an increased PR risk was found in combination with the *5-HTTLPR* L allele (OR 3.172; 95% CI 1.500-6.709) and the *5-HTTLPR* LL genotype (OR 6.923; 95% CI 1.900-25.227), as well as with the MDR1 T allele and the *5-HTTLPR* LL genotype (OR 11.143; 95% CI 2.415-51.414). The combination of the HTR1A CC genotype with the MDR1 T allele also increased the PR risk (OR 2.564; 95% CI 1.120-5.873).

The aim of this study was to identify pharmacogenetic risk profiles for the development of PR among patients with schizophrenia. The panel of tested potential pharmacogenetic markers included 19 polymorphic loci in 15 genes (CYP2D6, CYP2C9, CYP2C19, CYP1A2; MDR1, ANKK1, HTR1A, HTR2A, SLC6A4, HTR2C, COMT, MAOA, BDNF, DRD2, UGT1A1), which, according to literature data, affect both the pharmacokinetics and pharmacodynamics of antipsychotics.

We failed to detect significant differences in the distribution of allele and genotype frequencies between the group of pharmacoresistant patients and the group of patients responding to antipsychotic pharmacotherapy for the following loci: CYP2C9 (*2(rs1799853), 3(rs1057910)), CYP2C19 (2(rs4244285), 17(rs12248560)), CYP2D6 (3 (rs4986774), *10 (rs1065852)), NKK1 (rs1800497), HTR2A (rs6311), HTR2C (rs1414334), MAOA (uVNTR (30bp-VNTR)), BDNF (rs6265), DRD2 (rs6277), UGT1A1 (*28 (rs3064744)).

However, for the polymorphic loci of the CYP2D6 (rs3892097), HTR1A (rs6295), and COMT (rs4680) genes, we found a significant association with the development of PR in patients with schizophrenia. These polymorphic loci have been previously discussed in the literature in the context of searching for genetic markers of PR during antipsychotic treatment in different population samples [2, 4, 5]. Our research confirmed the significance of these loci as pharmacogenetic markers of PR for the population of Belarus. Table 4 presents data from other authors' studies regarding the association of specific genetic loci with the risk of developing PR (and adverse drug reactions), associations that were confirmed in the present study. As can be seen, both minor and major alleles may be associated with PR risk. The more frequently a pharmacogenetic marker occurs in a population, the more important its use in genetic testing when choosing antipsychotic therapy.

Table 4. Characterization of polymorphic gene loci

Gene / Polymorphic locus, product functionAllele associated with risk, functional changesFrequency (%) in European population (gnomAD Genome*)Association with riskCYP1A2 *1F/rs762551 EnzymeA (increased function)71.18Low antipsychotic (AP) concentration, inefficacy of AP therapy, tardive dyskinesia [14, 18–20]CYP2D6 *4/rs3892097 EnzymeA (no function)19.86High AP concentration, extrapyramidal symptoms, frequent hospitalization, severe outcome [7, 19, 21–26]COMT rs4680 Enzyme (dopamine catabolism)A (decreased function)51.80Pharmacoresistance, extrapyramidal symptoms, tardive dyskinesia [14, 27–29] G (active function)48.20 MDR1 C3435T rs1045642 TransporterT (decreased function)53.12Altered AP concentration, adverse effects, severe outcome [2, 19, 24, 26] C (active function)46.88 SLC6A4 5-HTTLPR TransporterS (decreased function)44.3 [16]Inefficacy of AP therapy [30, 31]HTR1A rs6295 ReceptorC49.97Inefficacy of therapy, altered AP blood concentration [32] G (decreased neurotransmission)50.03 

*Note: Based on gnomAD v2.1.1 (Genome) unless otherwise cited.

For the polymorphic locus rs3892097 of the cytochrome P450 CYP2D6 gene, our data are consistent with previously obtained results. It has been shown that the probability of developing PR is increased in patients carrying this poor metabolizer A allele [5, 12, 24]. The frequency of this minor A allele in individuals of European descent is 18.55% according to the GnomAD database (Table 4). We demonstrated that the probability of developing PR in Belarusian residents is significantly increased when the A allele (CYP2D64) is combined with the common A allele of the CYP1A2 gene (rs762551) and the T allele of the MDR1 gene (rs1045642). The frequency of pharmacoresistant patients with A-/A- (CYP2D6/CYP1A2), A-/T- (CYP2D6/MDR1), or A-/A-/T- (CYP2D6/CYP1A2/MDR1) genotypes reaches proportions of 38.8%, 46.8%, and 42.2%, respectively. Moreover, for carriers of the AG/AA (CYP2D6/CYP1A2) genotypes, the odds ratio reaches its maximum value (OR 9.143) (Table 3), indicating a very high likelihood of PR development in patients with this combination of genotypes.

The frequency of the G (Val) allele of the polymorphic locus rs4680 of the COMT gene (catechol-O-methyltransferase enzyme) among pharmacoresistant schizophrenia patients from Belarus was significantly higher (65%) than in the general European population (48.2%) (Table 4) and among patients who responded to antipsychotic therapy (40%) (Table 2). Literature data indicate that the G (Val) allele is associated with rapid dopamine metabolism and the risk of PR [14, 33]. However, other studies have shown that PR risk is associated with the A allele [34]. In our study, PR risk was associated with the presence of the G allele at the COMT rs4680 locus (OR 2.786) and increased further in combination with the L allele (5-HTTLPR) of the serotonin transporter gene SLC6A4 (OR 6.923). Carriers of the combination of alleles/genotypes G-/LL/T- (COMT / SLC6A4 / MDR1) had the highest PR risk indicator (OR 11.143) (Table 3).

For the polymorphic locus rs6295 of the serotonin receptor gene HTR1A, it has previously been established that reduced serotonergic neurotransmission is associated with the G allele, which also showed an association with the risk of several psychiatric disorders [35]. However, according to data obtained in the present study, the CC genotype (OR 2.722) acts as a genetic risk factor for PR. Among the European population, according to the GnomAD database, the G and C alleles occur with equal frequency (Table 4). In the present study, the proportion of patients carrying the PR risk genotype CC was twice as high as the proportion of patients responding to antipsychotic pharmacotherapy (30.8% and 14%, respectively), corroborating recently published data [32, 36, 37] indicating the differential significance of the alleles of this locus depending on sex and ethnicity.

Thus, this study conducted on schizophrenia patients from Belarus demonstrated a significant association of carrying polymorphic loci of the CYP2D6 (rs3892097), HTR1A (rs6295), and COMT (rs4680) genes with an increased risk of PR, both as individual genetic factors and in combination with polymorphic loci of other studied genes (CYP1A2 (rs762551), MDR1 (rs1045642), SLC6A4 (5-HTTLPR)). Pharmacogenetic risk profiles for the development of PR during antipsychotic treatment in individuals with schizophrenia were described: the highest PR risk indicator in our study was found in carriers of the combination of alleles/genotypes G-/LL/T- of the COMT / SLC6A4 / MDR1 genes (OR 11.143), as well as carriers of the AG/AA (CYP2D6 / CYP1A2) genotypes (OR 9.143). The identified associations of carrying alleles and genotypes of polymorphic gene loci with the response to antipsychotic drug treatment allow them to be used as pharmacogenetic markers in the treatment of schizophrenia patients of Belarusian and European descent with antipsychotic (typical and atypical) medications. However, the obtained data require confirmation in a larger patient cohort.