Genomic DNA was extracted using the QIAsymphony^SP automated extraction system and a QIAsymphony^® DSP DNA Midi Kit according to the manufacturer’s instructions (Qiagen, Hilden, Germany). A Nanodrop spectrophotometer (Thermo Fisher Scientific, Santa Clara, CA, USA) was used to determine the concentration and purity of the extracted DNA.

The Axiom PharmacoFocus Array (Catalog identifier: 952396; Thermo Fisher Scientific) offers comprehensive coverage of more than 2,000 markers (SNPs and insertions and deletions) in 150 genes across diverse populations (Tilleman et al., 2019) and functional variants that influence the ADME of commonly prescribed medications that are curated by the Pharmacogenomics Knowledge Base (PharmGKB) with clinical annotation levels of evidence 1A–2B (Whirl-Carrillo et al., 2021). These variants are commonly termed actionable PGx markers for testing in clinical practice (Whirl-Carrillo et al., 2021). Specifically, the Axiom PharmacoFocus Array facilitates genotyping in regions of high homology of key pharmacogenes (CYP2A6, CYP2D6, GSTM1, GSTT1, UGT2B17, and SULT1A1), which are usually difficult to obtain by complex multistep traditional methods (Tilleman et al., 2019).

According to the manufacturer’s instructions, genomic DNA was amplified by multiplex polymerase chain reaction (PCR) using a QIAGEN Multiplex PCR Kit (Qiagen). These amplified products were then fragmented, pooled, resuspended, and hybridized to the PharmacoFocus Array platform (Thermo Fisher Scientific, Santa Clara, CA, USA). The array was scanned on the automated Applied Biosystems™ GeneTitan™ Multi-Channel (MC) Instrument (Affymetrix Inc., Santa Clara, CA, USA). The genotyping call rates and quality parameters of all available markers and samples were generated using Applied Biosystems Axiom™ Analysis Suite software (version 5.2, Thermo Fisher Scientific, Santa Clara, CA, USA).

To achieve the highest genotyping performance, samples that did not satisfy the Dish QC (DQC) parameters were excluded. Furthermore, individuals with a <90% genotyping call rate were excluded from the analyses. Moreover, all markers with any of the following criteria were not considered for the bioinformatic analyses: genotyping call rate <95%, minor allele frequency (MAF) < 0.05, Hardy–Weinberg equilibrium (HWE) p < 0.001, and located on the X chromosome.

CYP2C8 is a phase I metabolizing enzyme that has been recently described by PharmGKB as a very important pseudogene (VIP) (Whirl-Carrillo et al., 2021). Interest in CYP2C8 emerged for several reasons, such as its central role in the biotransformation of structurally dissimilar compounds and endogenous molecules, attributed to its ability to bind divergent substrates without extensive conformational changes (Lai et al., 2009); its wide expression in body tissues other than hepatocytes (Sjostedt et al., 2020; Karlsson et al., 2021); more updates in identification of its substrates and inhibitors; and advanced knowledge in characterization of its SNPs and star alleles (Gaedigk et al., 2022). Additionally, the CYP2C8 gene is positioned on chromosome 10q24 in the CYP2C gene cluster (centromere-CYP2C18-CYP2C19-CYP2C9-CYP2C8-telomere), and given the proximity of CYP2C8 and CYP2C9, LD was previously reported between these genes (Yasar et al., 2002). Interestingly, the present exploratory study revealed strong LD between 2 CYP2C8 (rs7909236 and rs2275622) and 2 CYP2C9 (rs4918758 and rs1505) SNPs, all of which were shown to have significant individual associations with 9-OH-RIS plasma levels with decreasing (CYP2C8 pair) and increasing impacts (CYP2C9 pair). Only the CYP2C8 pair was associated with a decreased total active moiety, possibly indicating increased metabolism of 9-OH-RIS. This assumption may be supported by evidence from previous studies indicating that rs7909236 (-271C>A SNP designated as CYP2C8*1B) is associated with normal enzyme function compared with wild-type (Bahadur et al., 2002; Yasar et al., 2002; Rodríguez-Antona et al., 2007). While this CYP2C8 SNP was reported to be absent in Africans, its prevalence in our population (10%) was similar to that in Asians but lower than that in Caucasians (23%) (Bahadur et al., 2002). The other CYP2C8 SNP (rs2275622) (18% in our population) is a variant that was less commonly reported in clinical studies with contradictory functional effects (associated with higher or lower enzymatic activity) depending on the substrate (Kirchheiner et al., 2008; Grau et al., 2009).

Collectively, the observed negative impact of CYP2C8 on 9-OH-RIS and the total active moiety support the assumption that this enzyme simultaneously acts on RIS and 9-OH-RIS, with a greater reduction in the RIS level (possibly highlighting the influence of CYP2C8 on a second RIS metabolic pathway to an alternative metabolite).

However, our observed increase in 9-OH-RIS plasma levels associated with the 2 CYP2C9 variants (both designated as a CYP2C9*1 allele with normal function) (Gaedigk et al., 2017) could be linked to increased RIS metabolism to 9-OH-RIS mediated by the CYP2C9 enzyme, presumably as a minor non-CYP2D6 pathway. In contrast, previous studies with smaller sample sizes have failed to reveal any CYP2C9 impacts on RIS PK parameters (Llerena et al., 2004; Cabaleiro et al., 2014).

Additional studies are needed to explain the relative contributions of CYP2C8 and CYP2C9 to RIS metabolism and the power of simultaneously existing CYP2C8/CYP2C9 haplotypes to predict RIS efficacy and safety before implementation in clinical practice.

Two SNPs from two genes encoding different CYP450 enzymes (1 in CYP2C19 (rs4494250) and 1 in CYP2C18 (rs2860840)) were observed to equally contribute to reduced concentrations of RIS among carriers compared to noncarriers. Unfortunately, the functional status of those two SNPs was not defined in the known resources of CYP2C19 or CYP2C18 Allele Definition Tables (Whirl-Carrillo et al., 2021; Gaedigk et al., 2022). However, both genes are within the CYP2C subfamily, which is responsible for the metabolism of various drugs, including warfarin, escitalopram and omeprazole; however, the role of CYP2C18 in drug metabolism in general remains unclear (Goldstein and de Morais, 1994). Interestingly, the two genes are closely located on the same chromosome, and previous evidence in Japanese subjects suggested complete linkages between the mutated alleles of CYP2C18 and CYP2C19 (Kubota et al., 1998). Furthermore, our observation of lower RIS concentrations among autistic children carrying rs2860840 (CYP2C18_c.*31C>T(3′UTR)) reinforces a recent novel finding by Bråten et al. (2021), indicating that increased CYP2C19-dependent escitalopram metabolism leads to decreased concentrations in CYP2C19 NMs (*1/*1), similar to levels obtained in patients classified as CYP2C19 UMs (*17/*17) or RMs (*1/*17). The latter findings were mainly attributed to the liver enhancer effect induced by only one SNP, CYP2C18 (rs2860840), which was simultaneously carried by the CYP2C19 NMs. This novel SNP finding was later examined in Native American cohorts, and due to its high frequency and clinical implications for the treatment of >20 drugs with official annotations for CYP2C19 polymorphisms, it was recommended to add this SNP to the PGx practice panel to reveal the mismatching between CYP2C19-predicted and exposure-substantiated CYP2C19 metabolic phenotypes (Fernandes et al., 2023).

Our study is the first to report on the significant individual and combined impact of two SNPs in CYP2C19 and CYP2C18 on reducing RIS plasma levels, which could lead to infectiveness and require a higher dosage to achieve the RIS target therapeutic range. In contrast, Cabaleiro et al. (2014) reported an increased chance of RIS-induced neurologic manifestations among CYP2C19 NM healthy individuals compared to CYP2C19 PMs. However, in the same study, no significant impact of CYP2C19 polymorphisms was found on RIS or 9-OH-RIS PK parameters. Whether this finding could be attributed to the absence of both SNPs identified in our study as enhancing the metabolism of RIS or the smaller sample size in the Cabaleiro et al. (2014) study remains to be examined in future large-scale studies.

In the classical RIS metabolic pathway (Whirl-Carrillo et al., 2021), CYP2D6 is the primary enzyme involved in RIS metabolism to 9-OH-RIS. The current work reinforced this fact by revealing that 6 mutated CYP2D6 PGx variants have a significant positive impact on the RIS/9-OH-RIS metabolic ratio, which indirectly reflects a substantial increase in the plasma bioavailability of RIS in comparison to its major metabolite. Notably, all these SNPs were in complete LD in our autistic children. However, none of these variants were shown to have an influence on the individual estimates of RIS, RIS metabolite, or total active moiety, which may reflect the presence of other, undiscovered non-CYP2D6 variants (for example, as discussed earlier, CYP2C8/9 haplotypes) that could further modulate global exposure to and the efficacy and safety of RIS therapy in ASD. This finding is consistent with a similar observation in an exploratory study in Thai children with ASD, which revealed that three CYP2D6 variants existing in strong LD significantly influenced the metabolic ratio but not the discrete measurements of RIS, 9-OH-RIS, or total active moiety in plasma (Medhasi et al., 2016).

CYP2E1 is a member of the CYP450 family that metabolizes relatively few prescription drugs but is better known for the metabolism of toxins and procarcinogens (Hayashi et al., 1991). Several CYP2E1 SNPs in the promoter and intronic regions have been identified; however, luciferase promoter studies have shown that polymorphisms in the *7 haplotype, in particular the rs6413420 variant, increase CYP2E1 transcription (Huang et al., 2012). Interestingly, the current study revealed a significant impact of this SNP (15.7% prevalence) in reducing the total active moiety (at the level of p < 0.01) and 9-OH-RIS (p = 0.018) up to 0.6- and 0.44-fold, respectively, compared to noncarriers. Consistent with this finding, two previous studies highlighted the role of CYP2E1 SNPs in the etiology of schizophrenia and RIS treatment outcomes in the Chinese population (Huo et al., 2012; Shi et al., 2017). Of note, 5 CYP2E1 SNPs (including *5, rs3813867 and rs2031920) were associated with increased total active moiety, suggesting lower enzyme activity (Huo et al., 2012). However, rs2515641 in CYP2E1 was found to be significantly related to nonresponse to RIS treatment for schizophrenia in a Chinese cohort (p = 0.007) (Shi et al., 2017). Additionally, molecular genetic studies found that DNA methylation levels of CYP2E1 in the placenta were associated with children later diagnosed with ASD (Zhu et al., 2019; Bahado-Singh et al., 2022). Consistent with this finding, some molecular expression studies in inflammatory-mediated gastrointestinal diseases reported increased levels of CYP2E1 (122%) (Effinger et al., 2019), which strengthens the evidence of the role of the gut–brain axis in ASD etiology (Lombardi et al., 2023; Morton et al., 2023) and its correlation to our observation of reduced total active moiety and 9-OH-RIS levels in the study.

The CYP1A1 gene encodes the CYP1A1 enzyme, which is a member of the CYP1A subfamily and is responsible for the metabolism of diverse substrate molecules, such as sex hormones, caffeine, therapeutic drugs, environmental pollutants, toxins, and carcinogens. This diversity of CYP1A1 functions implies its involvement in numerous biological pathways (Kukal et al., 2023). Individual divergences in CYP1A1 expression and activity not only are attributed to genetic polymorphisms in CYP1A genes but can also be up- or downregulated through the interaction of environmental and endogenous physiological factors (Kukal et al., 2023). According to our association analysis, only one CYP1A1 SNP (rs2606345), which was highly prevalent in our ASD population (30.9%), had a significant positive impact on the metabolic ratio. In PGx studies of other medications, this variant produced controversial findings regarding its influence on enzyme activity (Whirl-Carrillo et al., 2021). Despite the SNP’s position in the intronic region of the CYP1A1 gene, some PK studies reported decreased function (Allegra et al., 2016; Talwar et al., 2016), while others reported a gain of function (Grover et al., 2010; Cusato et al., 2014; Allegra et al., 2017). Therefore, it is quite challenging to interpret from current data the actual contribution of this CYP1A1 polymorphism to the elevated metabolic ratio (increased RIS or decreased 9-OH-RIS levels in plasma), particularly because it is not directly involved in the primary RIS metabolic pathway. Moreover, this enzyme function was reported to be modified by a wide range of downstream modifications (genetic or epigenetic) and environmental factors that function together to alter the expression of the underlying genetic variant, leading to the ultimate biological response (Ye et al., 2019; Xu et al., 2023). Notably, another CYP1A1 SNP (rs1048943) was highly prevalent in Thai children with ASD (30.3%) (Sukasem et al., 2016), indicating that further molecular studies are needed to define the mechanisms connecting the CYP1A1 polymorphism to ASD and to modulate RIS PK in this disease population.

The CDA gene encodes the cytidine deaminase enzyme, which catalyzes the irreversible hydrolytic deamination of cytidine and deoxycytidine to uridine and deoxyuridine, respectively. It is one of numerous deaminases responsible for preserving the cellular pyrimidine pool. It is known that certain drugs can be rapidly metabolized by the CDA enzyme, which can affect their bioavailability and efficacy (Lavelle et al., 2012). In the context of chemotherapy, CDA plays a significant role (Abbaspour et al., 2023), since it metabolizes several chemotherapeutic drugs, including gemcitabine (Pellicer et al., 2017). Gemcitabine interferes with DNA synthesis and replication either by inhibiting enzymes involved in the synthesis of nucleic acid precursors or by misincorporation of nucleic acids into DNA or RNA macromolecules (Ciccolini et al., 2016). Studies have shown that the activity of the CDA enzyme can be a predictive biomarker in gemcitabine-treated cancer patients. Patients with lower CDA activity had significantly longer survival compared to patients with higher CDA activity (Soo et al., 2009). The current data suggested that one CDA variant (CDA_c.435C>T(T145 = ), rs1048977) was a strong predictor for a higher RIS/9-OH-RIS metabolic ratio. However, to date, there is no specific information in the medical literature about the interaction between CDA and RIS (Whirl-Carrillo et al., 2021). In the context of psychiatry and ASD, there is ongoing research into the role of various enzymes and their potential impact on these conditions. Some studies have demonstrated that CDA gene expression in the brain is associated with certain psychiatric disorders by creating DNA mutations via deamination of cytosine bases, which results in uracil (Gavin et al., 2012; Guidotti and Grayson, 2014). The same studies highlighted antipsychotics, including RIS, as potential targets in altering DNA methylation profiles in the brain. One study found that salivary immunoglobulin A (IgA) levels were significantly decreased in patients with ASD, and this correlated with bacteria-induced downregulation of the polymeric immunoglobulin receptor (Pigr) in salivary glands (Gong et al., 2021). However, this study did not specifically mention the CDA enzyme. Another study discussed the role of activation-induced CDA (AID) in the adaptive immune system and its potential implications in various diseases (Rios et al., 2020). However, it did not specifically link AID to autism or psychiatric disorders. While these studies provide valuable insights, more research is needed to fully understand the complex interactions among CDA, AID, and ASD etiology. Subsequently, it is important to investigate how genetic polymorphisms affecting CDA could act as markers for RIS clinical outcome (i.e., toxicity, efficacy) in real clinical practice.

Another novel variant that was found to be associated with increased RIS levels was rs2297480 (by 1.4-fold, p = 0.0012) in the FDPS gene, which encodes farnesyl diphosphate synthase (FDPS), an essential enzyme in the mevalonate pathway (cholesterol biosynthesis) (Göbel et al., 2020). The FDPS gene is implicated in various diseases, including neuropsychiatric disorders such as ASDs (Segatto et al., 2019; Lin et al., 2023). Additionally, as an RNA-binding protein, FDPS is also involved in transcriptional and post-transcriptional regulation of many enzymes (Wang L. et al., 2023). A recent in silico analysis revealed that FDPS is an overlapping gene that is involved in CNS disorders and is simultaneously associated with the encoding of enzymes in the lipid and cholesterol metabolic pathways (Ang and Moon, 2022). Of note, antipsychotic drugs were observed to result in upregulated expression of the genes involved in cholesterol biosynthesis (Le Hellard et al., 2008), which was suggested as a potential causal pathway for their role in the pathogenesis of neuropsychiatric disorders (Zhou et al., 2021) as well as their subsequent induced adverse metabolic effects (Le Hellard et al., 2008). However, according to the present association study, how FDPS polymorphisms and expression modulate RIS exposure and response in ASD patients and vice versa remain unknown. However, our results highlight the need for further investigation of the pathways underlying this gene-disease-drug interaction.

The TPMT gene encodes the thiopurine S-methyltransferase enzyme, which plays a crucial role in the metabolism of thiopurine drugs (Lee et al., 1995). Moreover, it is dependent on the S-adenosylmethionine (SAM) methyl donor substrate in the methionine pathway (Weinshilboum, 2006). The TPMT enzyme is involved with other conjugation enzymes in phase II detoxification, where liver cells add a substance (such as cysteine, glycine, methyl or a sulfur molecule) to a toxic chemical or drug to make it less harmful and easier for the body to excrete (Zhang, 2011). It has also been implicated in the metabolism of other aromatic and heterocyclic sulfhydryl compounds (Woodson and Weinshilboum, 1983).

Several pharmacogenetic studies have demonstrated that certain polymorphism-induced mutations in the TPMT gene result in completely undetectable TPMT enzyme activity, leading to life-threatening adverse events associated with even normal doses of anticancer drugs, such as azathioprine, cyclosporine, and daunorubicin (Wang et al., 2010). However, TPMT is not involved in the direct metabolic pathway (Phase I) of RIS (Whirl-Carrillo et al., 2021). In our study, two TPMT intronic SNPs (rs12529220 and rs2518463; both known to express the normal functional TPMT allele *1), with a prevalence of 46.1% among Saudi autistic children, were likely to be in LD at chromosome 6. Both were associated with a substantial increase in RIS levels (1.3-fold). Consistent with this finding, a previous RIS PK study in normal volunteers demonstrated significantly higher 9-OH-RIS plasma levels in *1/*1 genotype participants in comparison to mutant genotype carriers (*1/*2,*1/*3C,*1/*3A) (Cabaleiro et al., 2014). Another earlier study reported an association of decreased TPMT activity (mutant genotypes) with olanzapine-induced fatigue and dizziness in healthy volunteers with no significant impact on any of its PK parameters (Cabaleiro et al., 2013). However, no other data are available on their association with the PK of RIS or other antipsychotics (Whirl-Carrillo et al., 2021); therefore, it is challenging to provide a satisfactory interpretation. Further studies are needed to explore the impact of TPMT polymorphisms on chronic RIS therapy in ASD.

The NAT2 gene encodes N-acetyltransferase 2 (arylamine N-acetyltransferase), which is a typical xenobiotic metabolizing enzyme (Ackenheil and Weber, 2004) responsible for acetylation as a phase II conjugation reaction. In previous studies, NAT2 was identified to play a role in the metabolism of benzodiazepines (Camargo et al., 2023) and also hypothetically plays a role in the metabolism of some antipsychotics (Ackenheil and Weber, 2004). To date, 88 SNPs have been identified within the NAT2 gene that can affect NAT2 function by resulting in reduced enzyme stability or altered affinity for a substrate. NAT2 genotypes can be divided into three subgroups: “slow acetylator” (two slow alleles), “intermediate acetylator” (1 slow and 1 rapid allele), and “rapid acetylator” (2 rapid alleles, occasionally referred to as “fast”). Out of 38 NAT2 SNPs examined in our exploratory study via the PharmacoFocus array, NAT2_c.-594G>C (5′UTR) (rs4271002) had a significant impact, increasing the RIS plasma level by 1.867-fold compared to noncarriers. Therefore, it is expected that carriers of this mutant allele may require dose modification to avoid RIS-induced adverse effects. Importantly, this is a novel NAT2 variant that is not a part of any named alleles and has been shown in one study to be significantly associated with the risk of aspirin-intolerant asthma (Kim et al., 2010). According to PharmGKB, the functional consequence of this SNP is currently unknown, but it may affect transcription, and its exact role in the RIS metabolic pathway remains unclear (McDonagh et al., 2014). However, our result is consistent with the impact of NAT2 polymorphisms in the study by Cabaleiro et al. (2014) reporting an increased incidence of RIS-induced headache among NAT2 IM and PM healthy individuals in comparison to NAT2 NMs.

The UGT2B17 gene encodes uridine diphosphate glycosyltransferase 2 family, member B17, which is part of the family of UDP-glucuronosyltransferases (UGTs) (Beaulieu et al., 1996). As part of the phase II liver detoxification system, these genes are responsible for maintaining steady-state levels of a variety of substrates, including steroid hormones, by catalyzing the transfer of glucuronic acid moieties to these molecules and rendering them hydrophilic (Burchell et al., 1995). UGT2B17 is expressed not only in the small intestine and liver but also in steroid target tissues such as the breast, uterus, and prostate, where the extent of glucuronidation can be substantial (Karlsson et al., 2021), indicating its potential role in hormonally induced diseases (Wilson et al., 2004). However, molecular studies reported that UGT2B17 isoforms had a 4.4-fold higher abundance in the intestine than in the liver (Zhang et al., 2018). This fact suggested the potential of UGT2B17 to have a greater first-pass effect on its substrates when orally administered than subsequent liver metabolism, particularly in high UGT2B17-expressing individuals (Zhang et al., 2018). Additionally, in proteome studies, the UGT2B17 isoform, unlike other UGTs, was expressed in different brain regions, particularly in the cerebellum (Karlsson et al., 2021). According to the current study data, UGT2B17 rs78998153, which had a prevalence of 26% among our ASD children, exhibited a very significant effect on RIS exposure. Indeed, after Bonferroni correction for multiple testing, this novel variant is the only PGx marker that still demonstrated a positive impact on the RIS/9-OH-RIS metabolic ratio, indicating a substantial increase in the RIS plasma circulating levels in comparison to its major metabolite. Consistently, an exploratory-based study of RIS PK in Thai children with ASD demonstrated that UGT2B4 c.∗448A>G (rs1131878), an isoform that is more highly expressed in liver than brain tissues (Karlsson et al., 2021), was highly associated with the metabolic ratio (Medhasi et al., 2016). Additional evidence for the impact of UGTs on RIS PK can be drawn from a study in Thai children with ASD, where three SNPs indicating UGT1A1 mutation (an isoform that is highly expressed in brain tissues (Sjostedt et al., 2020)) have shown a significant association with the RIS-induced prolactin response (Hongkaew et al., 2018). Collectively, these results highlight the need for further molecular studies to explore the correlation between various UGT genotypes and their induced modifications in UGT enzyme expression in the brain tissues of patients with ASD. This research topic is anticipated to enable a better understanding of altered RIS disposition in the brain and its precise dose individualization requirements under this central nervous system condition (Sheng et al., 2021).

The MTHFR gene encodes the enzyme methylenetetrahydrofolate reductase (MTHER). MTHER is involved in a chemical reaction involving forms of the vitamin folate. Specifically, this enzyme converts 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate. This reaction is part of the multistep process that converts the amino acid homocysteine to another amino acid, methionine. The body uses methionine to make proteins and other important compounds, such as neurotransmitters (dopamine and serotonin) (Jalgaonkar et al., 2022; Majhi et al., 2023). Individuals homozygous for the SNP rs1801133 (MTHFR_c.665C>T(Ala222Val)) have lower MTHFR activity than CC or CT (heterozygous) individuals and therefore are predisposed to hyperhomocysteinemia associated with lower plasma folate levels (Majhi et al., 2023). A meta-analysis conducted on the association between MTHFR SNPs and ASD susceptibilities indicated that MTHFR rs1801133 was associated with ASD in the five genetic models (Li et al., 2020).

Consistent with this, a growing body of evidence suggests that the severity of autistic symptoms as a whole may be associated with increased levels of homocysteine associated with aggravation of dopamine deficiency (Carpita et al., 2023; Dangmann, 2023; Majhi et al., 2023). According to our data, rs1801133 of MTHFR, despite being less prevalent in our sample (12.9%) than in another cohort of Saudi children with ASD (36%) (Arab and Elhawary, 2019), was associated with a significant decrease in 9-OH-RIS and total active moiety levels, yet no evident impact was observed on RIS plasma exposure. The decreased concentration of the active moiety and a more pronounced effect on 9-OH-RIS could be explained by several factors. 9-OH-RIS undergoes minor hepatic metabolism and is primarily excreted unchanged by the kidney (79.6%). One of the known metabolic pathways for 9-OH-RIS is mediated by oxidative N-dealkylation, forming the acid metabolite M1 (Citrome, 2007; Vermeir et al., 2008). Emerging evidence indicates that high levels of homocysteine may enhance several metabolic pathways, such as oxidation (oxidative stress), nitrosylation, acylation, and hypomethylation (Perna et al., 2003a; 2003b). According to drug metabolism theories, these mechanisms (except hypomethylation) are believed to produce more polar metabolites that cannot diffuse across membranes and may, therefore, be actively transported (Li et al., 2019). Therefore, enhanced 9-OH-RIS excretion linked to hyperhomocysteinemia (induced by the MTHER mutation) via oxidation is assumed to be a superior postulated mechanism. In addition, homocysteine is a sulfur-containing amino acid (Lentz, 2005), which could serve as a co-factor in the conjugation of a drug metabolite by sulphation (Chen and Tang, 2022), leading to increased facilitated excretion in urine (Pan et al., 2020). However, these hypotheses remain uncertain. Clearly, more studies are necessary to elucidate the role of homocysteine in enhancing 9-OH-RIS excretion and its clinical consequences in ASD patients.

The ABCC3 gene encodes a protein that is a member of the superfamily of ATP-binding cassette (ABC) transporters. These ABC proteins transport various molecules across extra- and intracellular membranes. The specific function of this transporter has not yet been determined; however, it was reported to mediate biliary and intestinal excretion of organic anions (Banach et al., 2022). The functional activity of some ABCC3 variants has been fully examined (Singh et al., 2020). The current study revealed for the first time that a specific variant of the ABCC3 gene (ABCC3_c.3890G>A(R1297H)) had a prominent negative impact on 9-OH-RIS plasma levels, indicating decreased excretion (efflux) from target cells (hepatocytes or brain) to the bile or peripheral blood. The negative impact of this variant on RIS clinical efficacy and safety in ASD carriers warrants further investigation.

The ABCB11 gene encodes the ATP-binding cassette subfamily B member 11 protein, which is another member of the superfamily of ABC transporters. This membrane-associated protein is also named bile salt export pump (BSEP) or sister of P-glycoprotein (sPgp) (Strautnieks et al., 1998). Consistent with a previous exploratory study in Thai children with ASD (Medhasi et al., 2016), the current work detected the same 4 ABCB11 variants (ABCB11 rs495714, rs496550, rs473351 and rs497692), which displayed a significant reducing effect on the RIS/9-OH-RIS metabolic ratio, although at a lower rank of importance (Supplementary Table S4). Notably, our results in this population (Figure 2B) are compatible with the previous LD finding (Medhasi et al., 2016) that 3 ABCB11 SNPs were strongly linked. This observation hypothetically indicates a more predominant influence of these mutations on reducing RIS efflux than 9-OH-RIS efflux, probably from various tissues’ cells into bile or blood. In contrast, previous candidate gene studies in Caucasian children with ASD (Correia et al., 2009; Rafaniello et al., 2017) and adults with schizophrenia (Xing et al., 2006; Kuzman et al., 2008) have linked reduced RIS efflux with other ABC transporter subtype mutations (ABCG2_c.421C>A in ABCG2; c.3435C>T, c.1199G>A, c.1236C>T and c.2677G>T in ABCB1). These inconsistent findings emphasized the importance of genome-wide exploratory studies to reveal disease-specific PGx markers of certain drugs, with priority according to clinical significance and their ethnogeographic frequencies.

One of the receptors that RIS blocks is the alpha 2A adrenoceptor (α2A-AR), which is an adrenergic receptor that responds to adrenaline and noradrenaline (Shahid et al., 2009). α2A-AR is encoded by the ADRA2A gene and is mainly found in the brain, where it regulates various functions, such as mood, cognition, attention, and sleep (Nyrönen et al., 2001). The role of α2A-AR in the therapeutic effects and side effects of RIS is not fully understood, but some studies have suggested that blocking this receptor may have both positive and negative consequences (Marcus et al., 2009). Additionally, other factors, such as genetic variations, drug interactions, and individual differences, may influence the response to RIS and the α2A-AR antagonism effect (Uys et al., 2017). The current association analysis revealed that ADRA2A SNP rs553668 carriers exhibit significantly increased RIS and total active moiety plasma levels (by 1.468- and 1.678-fold, respectively) in comparison to noncarriers. This finding could be interpreted in light of a previous study involving pheochromocytoma patients reporting that ADRA2A SNPs rs553668/rs521674 were associated with higher dosage requirements of α-adrenergic receptor blockers to control blood pressure (Berends et al., 2022).

Both our results and previous findings suggested that a certain degree of mutation in the α2A-AR receptor (decreased expression and/or density) mediated by these ADRA2A SNPs could lead to decreased drug-receptor occupancy and affinity, which could explain the subsequent higher PK exposure of any drugs targeted to antagonize it, such as that shown in our ASD patients exhibiting higher RIS plasma levels adjusted by the RIS dose. However, this assumption regarding the correlation between receptor affinity and drug plasma level remains speculative, and the functional consequences of the ADRA2A SNP on its receptor need to be examined. To our knowledge, to date, no other antipsychotic PGx studies have addressed the potential clinical consequences of polymorphisms of any of the genes encoding adrenergic receptors (including the ADRA2A gene) on efficacy and safety (Bousman et al., 2023). Therefore, it is important to investigate the current observation of significantly higher RIS plasma levels associated with ADRA2A polymorphisms in terms of clinical impacts in a larger cohort.

Corticotropin-releasing hormone receptor 2 (CRHR2) is a protein that is encoded by the CRHR2 gene, which is highly expressed in the choroid plexus (part of the blood–brain barrier) of the human brain and to a lesser extent in the plasma membranes of hormone-sensitive cells, including those in the gastrointestinal tract and kidney (Pal et al., 2010). CRH is a hormone secreted from the hypothalamus in response to stress, which needs to efficiently bind with the CRHR2 receptor to stimulate its effects (Grammatopoulos et al., 1999). CRF is a key hormone that is involved in the control of various body systems via its mediatory stimulation of the hypothalamic–pituitary–adrenal (HPA) axis. On the other hand, hypothalamic CRF, via its action on the HPA axis, may be partially involved in the reinforcing effects of metabolic enzymes in phases I and II (Mormede et al., 2011; Wójcikowski and Daniel, 2011; Bromek and Daniel, 2021). In addition, recently, increased activation of the HPA axis was suggested to play an important role in ASD-like social behaviors (Jacobson, 2014; Rusch et al., 2023). As CRHR2 is one of the receptors for the hormones involved in the HPA axis, decreased CRHR2 expression levels in the hypothalamus were recently suggested to increase the risk of ASD (Wang X. et al., 2023). Interestingly, the current data revealed that an intronic CRHR2 variant (rs7793837), which mostly causes mutations in the CRHR2 protein, was significantly associated with a positive impact on the total active moiety level in the plasma of children with ASD. These findings suggest that CRH could play a complex role in drug metabolism and possibly in the clinical response to RIS therapy, and further research could clarify its significance as a PGx marker within the context of ASD. A possible explanation for this increased level of RIS and 9-OH-RIS could be attributed to decreased binding of CRH to the mutated CRHR2 variant 3 (rs7793837), leading to its decreased functional impact on various downstream signaling pathways mediating the metabolism and excretion of both molecules. However, this hypothesis needs to be proven in further studies to elucidate its clinical impact on RIS therapy outcomes.