CYP3A4∗22 Genotyping in Clinical Practice: Ready for Implementation?

Cytochrome P450 3A4 (CYP3A4) is the most important drug metabolizing enzyme in the liver, responsible for the oxidative metabolism of ∼50% of clinically prescribed drugs. Therefore, genetic variation in CYP3A4 could potentially affect the pharmacokinetics, toxicity and clinical outcome of drug treatment. Thus far, pharmacogenetics for CYP3A4 has not received much attention. However, the recent discovery of the intron 6 single-nucleotide polymorphism (SNP) rs35599367C > T, encoding the CYP3A4∗22 allele, led to several studies into the pharmacogenetic effect of CYP3A4∗22 on different drugs. This allele has a relatively minor allele frequency of 3-5% and an effect on CYP3A4 enzymatic activity. Thus far, no review summarizing the data published on several drugs is available yet. This article therefore addresses the current knowledge on CYP3A4∗22. This information may help in deciding if, and for which drugs, CYP3A4∗22 genotype-based dosing could be helpful in improving drug therapy. CYP3A4∗22 was shown to significantly influence the pharmacokinetics of several drugs, with currently being most thoroughly investigated tacrolimus, cyclosporine, and statins. Additional studies, focusing on toxicity and clinical outcome, are warranted to demonstrate clinical utility of CYP3A4∗22 genotype-based dosing.


IMMUNOSUPPRESSIVE AGENTS
Cyclosporine and tacrolimus are immunosuppressants used in solid organ transplantation, metabolized by CYP3A4 into less active compounds, and are characterized by highly variable pharmacokinetics and a narrow therapeutic index (Moes et al., 2014). To prevent overexposure (risking drug-related toxicity) or underexposure (risking transplant rejection), therapeutic drug monitoring is applied (Kahan et al., 2002;Lloberas et al., 2017). Pharmacogenetic testing may optimize the starting dose when pharmacokinetic steady states are not yet achieved.

CARDIOLOGY
Ticagrelor, clopidogrel, and prasugrel are active platelet aggregation inhibitors metabolized by CYP3A4, used in cardiology and neurology. For ticagrelor, CYP3A4 * 22 carriers showed an 89% higher area-under-the-plasma-concentrationtime curve (AUC) and more pronounced platelet inhibition (43% vs. 21%) (Holmberg et al., 2019) (Supplementary Table 3). This increases risk of bleedings (Becker et al., 2011). No significant correlation of CYP3A4 * 22 with active metabolites of clopidogrel or prasugrel was found, although the authors stated that differences in clopidogrel and prasugrel pharmacokinetics <50% cannot be ruled out (Holmberg et al., 2019). For sildenafil, metabolized by CYP3A4 in less active compounds, is used in patients with heart failure, CYP3A4 * 22 carriers showed increased dose-adjusted concentrations (de Denus et al., 2018). For these drugs, only one study is currently available. Further studies are needed to confirm these findings.

PSYCHIATRY Antidepressants
CYP2D6 pharmacogenetic based guiding of antidepressant drug therapy is nowadays quite accepted. Dosing advices, based on CYP2D6 and CYP2C19 genotype, are available at PharmGKB, CPIC and the Dutch Pharmacogenetic Working Group (DPWG) (Hicks et al., 2015). Thus far, CYP3A4 genotype-based dosing for psychotropic drugs has not yet been described. However, CYP3A4 does have a contribution in citalopram [CYP2C19, CYP2D6, CYP3A4 (Rochat et al., 1997) (Störmer et al., 2000)] metabolism. The prominent role of CYP2C19 and CYP2D6 in the metabolism may complicate CYP3A4 studies in these drugs. Yet, compound heterozygotes for CYP2D6, CYP2C19, and CYP3A4 were found by us in patients with therapy related side effects, insufficiently explained by CYP2D6 and/or CYP2C19 genotype alone. Recently, combining CYP2C19, CYP2D6, and CYP3A4 genotypes proved to be a better predictor of citalopram/escitalopram blood levels as compared to individual genes (Shelton et al., 2020). We feel it would be beneficial to study the effect of CYP3A4 * 22, in addition to CYP2D6 and CYP2C19, in antidepressants, but thus far, no clear indications for a prominent role for CYP3A4 * 22 are available.

Anti-anxiolytics
Alprazolam is one of the most commonly prescribed psychoactive agent for mood and anxiety disorders (Stahl, 2002). In patients with alcoholism and anxiety disorders, CYP3A4 * 22 carriers had significantly increased active alprazolam concentration/dose ratios and a decreased treatment response, as reflected in HAMA scale scores (4.0 vs. 3.0) (Zastrozhin et al., 2020) (Supplementary Table 4).

Anti-psychotics
Risperidone, a drug with CYP2D6 genotype-based dosing recommendations, showed CYP3A4 * 22 carriers having 30% lower 9-hydroxyrisperidone (active metabolite, generated from risperidone by CYP2D6) clearance (Vandenberghe et al., 2015). Two other studies could not confirm this: the reason for this discrepancy could be sample size [n = 26, (Rafaniello et al., 2018)], although van der Weide et al. included 130 patients (van der Weide and van der Weide, 2015) (Supplementary Table 5). CYP3A4 * 22 genotype was associated with serum levels active pimozide and C/D ratio, however, this association only explained 5% of total variation in a multiple regression analysis (van der Weide and van der Weide, 2015). For aripiprazole and haloperidol, no significant effect of CYP3A4 * 22 was observed (van der Weide and van der Weide, 2015; Rafaniello et al., 2018), possibly because CYP2D6 plays a more prominent role (Fang et al., 1997;Fang et al., 1999;van der Weide and van der Weide, 2015). Multiple regression analysis demonstrated that 4-17% of the variation in concentration of these anti-psychotics was explained by CYP2D6 genotype (van der Weide and van der Weide, 2015). However, quetiapine, a drug metabolized by CYP3A4 (DeVane and Nemeroff, 2001), showed 150% higher serum concentrations and 67% higher dose-corrected quetiapine serum concentrations in CYP3A4 * 22 carriers (van der Weide and van der Weide, 2014). Significantly more CYP3A4 * 22 patients achieved serum levels above the therapeutic range of 500 µg/L (van der Weide and van der Weide, 2014).

ANTICANCER AGENTS
For anticancer drugs, pharmacogenetic testing for CYP450 enzymes is currently limited to CYP2D6 testing for tamoxifen, which needs activation by CYP2D6 [for review, see Mulder et al. (2021)]. A good example of implementation of pharmacogenetic testing in oncology outside the CYP450 field, is DPYD analysis prior to capecitabine treatment, as well as TPMT testing for 6-mercaptopurine for treatment of Acute Lymphatic Leukemia (Henricks et al., 2018;Roden et al., 2019). Yet, CYP3A4 is involved in the metabolism of many anticancer drugs, and since efficacy and toxicity are important aspects in oncology, factors that may contribute to predicting toxicity should be examined for their potential clinical value. As indicated by Elens et al., decreased CYP3A4 metabolism due to CYP3A4 * 22 could be demonstrated in cancer patients, as determined by midazolam and erythromycin metabolism (Elens et al., 2013b).

Microtubule-Stabilizing Agents: Paclitaxel and Docetaxel
Paclitaxel and docetaxel are used for solid tumors. For paclitaxel, CYP2C8 has been the focus for CYP450 studies, but also CYP3A4 is involved in its metabolism (Harris et al., 1994). Neurotoxicity is often observed as side effect (Lee and Swain, 2006). In 261 cancer patients, CYP3A4 * 22 carrier status was an independent predictive factor for development of paclitaxelinduced neurotoxicity, although no significant association with paclitaxel pharmacokinetics was found for reasons unknown to the authors (Supplementary Table 6) (de Graan et al., 2013). Interestingly, another CYP3A4 variant allele, CYP3A4 * 20, was also associated with paclitaxel neuropathy (Apellániz-Ruiz et al., 2015). This supports the potential predictive value of CYP3A4 genotyping for paclitaxel neurotoxicity. For docetaxel, a study in 150 breast cancer patients showed that CYP3A4 * 22 carriers were at increased risk of grade 3-4 toxicity (Sim et al., 2018). Docetaxel is mainly metabolized by CYP3A4 (Shou et al., 1998;Hirth et al., 2000), and large interindividual variability in docetaxel pharmacokinetics has been reported (Goh et al., 2002;Michael et al., 2012). Since the pharmacokinetic association of CYP3A4 * 22 is lacking, more research on CYP3A4 * 22 and docetaxel treatment is warranted.

Tyrosine Kinase Inhibitors: Sunitinib and Pazopanib
Sunitinib and pazopanib are frequently prescribed tyrosine kinase inhibitors (TKIs) with established exposure-response relationships for renal cell carcinoma (Houk et al., 2010;Suttle et al., 2014). Both drugs are predominantly metabolized by CYP3A4 into less active metabolites (Sugiyama et al., 2011;Thorn et al., 2017). CYP3A4 * 22 status was associated with a significantly decreased clearance of pazopanib (35%) (Bins et al., 2019) and a decreased clearance of sunitinib (22.5%) (Diekstra et al., 2014) (Supplementary Table 6). Bins et al. (2019) proposed that pazopanib dose adjustments based on CYP3A4 * 22 status should be considered since their pharmacokinetic-model showed that 600 mg pazopanib in CYP3A4 * 22 carriers would lead to similar pazopanib exposure as wild-type patients using 800 mg. Feasibility of CYP3A4 * 22 genotype-guided dosing of TKIs in cancer patients is currently under investigation in a large prospective clinical trial; results are expected in 2022 (Dutch Trial Registry 1 ).

PAIN MEDICATION: FENTANYL
Synthetic opioids, like alfentanil, sufentanil, remifentanil, and fentanyl, are all metabolized by CYP3A4 into inactive metabolites (Yun et al., 1992;Tateishi et al., 1996). Unfortunately, no studies regarding the association between CYP3A4 * 22 and pharmacokinetics of alfentanil, sufentanil, and remifentanil are published. Regarding fentanyl, a positive association between CYP3A4 * 22 carriers and increased exposure to fentanyl (higher AUC, lower clearance) was found in two studies with healthy individuals (Saiz-Rodríguez et al., 2019;Saiz-Rodríguez et al., 2020) (Supplementary Table 7). An association with fentanyl MR confirmed this influence, although the authors did not find a significant association with serum fentanyl concentration (Barratt et al., 2014).

CONCLUSION
Although high interindividual variability is observed in CYP3A4 expression and activity, this could not be accounted for by genetic variability in CYP3A4, as polymorphisms that result in a significant altered CYP3A4 activity are rare. Therefore, the discovery that the more common observed CYP3A4 * 22 has significant effects on the pharmacokinetics of several drugs was surprising. This argues in our opinion for further investigation of the potential clinical use of CYP3A4 * 22 genotyping (