Regulation of Human Hydrolases and Its Implications in Pharmacokinetics and Pharmacodynamics

Abstract

Hydrolases represent an essential class of enzymes indispensable for the metabolism of various clinically essential medications. Individuals exhibit marked differences in the expression and activation of hydrolases, resulting in significant variability in the pharmacokinetics (PK) and pharmacodynamics (PD) of drugs metabolized by these enzymes. The regulation of hydrolase expression and activity involves both genetic polymorphisms and nongenetic factors. This review examines the current understanding of genetic and nongenetic regulators of six clinically significant hydrolases, including carboxylesterase (CES)-1 CES2, arylacetamide deacetylase (AADAC), paraoxonase (PON)-1 PON3, and cathepsin A (CTSA). We explore genetic variants linked to the expression and activity of the hydrolases and their effects on the PK and PD of their substrate drugs. Regarding nongenetic regulators, we focus on the inhibitors and inducers of these enzymes. Additionally, we examine the developmental expression patterns and gender differences in the hydrolases when pertinent information was available. Many genetic and nongenetic regulators were found to be associated with the expression and activity of the hydrolases and PK and PD. However, hydrolases remain generally understudied compared with other drug-metabolizing enzymes, such as cytochrome P450s. The clinical significance of genetic and nongenetic regulators has not yet been firmly established for the majority of hydrolases. Comprehending the mechanisms that underpin the regulation of these enzymes holds the potential to refine therapeutic regimens, thereby enhancing the efficacy and safety of drugs metabolized by the hydrolases.

SIGNIFICANCE STATEMENT

Hydrolases play a crucial role in the metabolism of numerous clinically important medications. Genetic polymorphisms and nongenetic regulators can affect hydrolases’ expression and activity, consequently influencing the exposure and clinical outcomes of hydrolase substrate drugs. A comprehensive understanding of hydrolase regulation can refine therapeutic regimens, ultimately enhancing the efficacy and safety of drugs metabolized by the enzymes.

Introduction

The majority of clinically used drugs often undergo two phases of drug metabolism. Phase I metabolism involves modifying the parent compound through reactions, such as oxidation, reduction, and hydrolysis, to convert the drug into a more hydrophilic metabolite. Phase II metabolism (conjugation) includes methylation, acetylation, sulphation, glucuronidation, and glycine or glutathione conjugation reactions, which attach the drug molecule with another molecule to render the compound pharmacologically inert and enhance its water solubility for efficient excretion through urine.

The study of drug-metabolizing enzymes (DMEs) has predominantly focused on the cytochrome P450 (P450) enzyme family, largely due to their significant involvement in metabolizing over half of the drugs used in clinical practice. Recently, there has been an increasing interest in investigating non-P450 enzymes, particularly hydrolases, regarding their roles in drug metabolism. Cerny (2016) found that approximately 37.6% of the 125 Food and Drug Administration (FDA)-approved drugs between 2006 and 2015 were metabolized by non-P450 enzymes (Fig. 1). Moreover, Fukami et al. (2022) reported that approximately 26% of the drugs approved by the Pharmaceuticals and Medical Devices Agency in Japan from 2016 to 2020 were metabolized by hydrolases. However, compared with P450s, hydrolases are understudied regarding their role in the pharmacokinetics (PK) and pharmacodynamics (PD) of their substrate drugs despite their significant role in metabolizing numerous clinically essential drugs. Notably, hydrolyze-mediated metabolism could lead to the deactivation of an active compound or the activation of a prodrug. Prodrug design has become an indispensable strategy for drug development. Prodrugs are often formulated to overcome various PK obstacles, such as poor solubility and absorption and extensive first-pass effect, and many PD issues, such as side effects and subtherapeutic efficacy (Najjar et al., 2020). Prodrug activation is usually catalyzed by enzymatic processes, and hydrolases are an essential class of enzymes responsible for activating ester-based prodrugs.

Fig. 1.

Percentages of small-molecule drugs approved by the FDA between 2006 and 2015 metabolized by different drug-metabolizing enzymes. Non-P450 enzymes: UGT, carbonyl reductase, and aldehyde oxidase (AO). Adapted from Cerny (2016).

Conventionally, most medications follow fixed-dose regimens, even though significant interindividual variability in drug exposure and response has been documented for many drugs. In recent years, individualized and patient-specific dosing regimens have been increasingly developed based on the patient’s characteristics, such as renal clearance, liver function, body weight, and surface area, to reduce the variability and enhance the safety and efficacy of pharmacotherapy (DiPiro et al., 2020). The expression and activity of DMEs vary significantly among individuals, which is a pivotal factor contributing to interindividual variability in PK and PD. Both genetic and nongenetic factors regulate the expression and activity of DMEs. Notably, the pharmacogenetics of DMEs demonstrated a significant impact on patient care, with growing utilization in clinical settings to enhance the effectiveness and safety of pharmacotherapy. For instance, azathioprine is metabolized by thiopurine methyltransferase (TPMT) to 6-mercaptopurine (6-MP). If the patient carries the intermediate or poor metabolizer genotypes of TPMT, there would be a higher risk of toxic accrual of 6-MP in the body leading to myelosuppression. Recognizing this potential risk, the FDA has recommended that poor metabolizers should seek alternative therapy and that intermediate metabolizers should receive a substantially reduced dose (Seidman, 2003). Furthermore, in addition to genetic variants, nongenetic regulators, such as inhibitors and inducers, can exert substantial influence on DME function and expression, thereby impacting the PK and PD of substrate drugs.

Although the importance of hydrolases has been increasingly recognized in drug development and clinical practice, our understanding of the regulatory mechanisms governing the expression and activity of hydrolases has lagged behind other DMEs, such as P450s and UDP-glucuronosyltransferases. In this review, we describe the present knowledge of six of the most clinically significant hydrolases, including carboxylesterase (CES)-1, CES2, arylacetamide deacetylase (AADAC), paraoxonase (PON)-1, PON3, and cathepsin A (CTSA), regarding their pharmacogenetics and nongenetic regulators and discuss the implications in the PK and PD of their substrate drugs.

Carboxylesterases

CES1 and CES2 were first characterized as 60-kDa esterase forms 1 and 2, respectively, in rabbit liver microsomes (Korza and Ozols, 1988). Frey et al. (1994) identified the role of low-barrier hydrogen bonds between His and Asp (substituted by Glu in CES) in aiding Ser’s β-OH group’s nucleophilic attack on the acyl carbonyl group of the peptide in chymotrypsin. These bonds stabilize the catalytic triad in the tetrahedral addition intermediate. It was theorized that the low-barrier hydrogen bond between Glu336 and His450 facilitates Ser203’s nucleophilic attack on the substrate’s carbonyl group in CESs (Hosokawa, 2008). Of note, there is a noticeable conservation of sequences essential for the hydrolytic function of the catalytic triad (Glu, His, Ser) in CES, acetylcholinesterase, butyrylcholinesterase, and cholesterol esterase (Hosokawa, 2008). This common structure typifies α/β-hydrolase–fold families, which is crucial for the breakdown of both endogenous and exogenous substrate compounds. Although both CES1 and CES2 can cleave ester bonds, CES1 prefers substrates with a large acyl group and a small alcohol group, whereas CES2 is more efficient in metabolizing compounds with a small acyl group and a large alcohol group (Fukami et al., 2015).

Carboxylesterase 1

CES1 is the most abundant DME in the liver, responsible for 80%–95% of total hepatic hydrolytic activity in humans (Imai et al., 2006). CES1 is also expressed in the lung, but the CES1 level is rather low in human plasma, kidney, and intestine (Fig. 2). Numerous drugs are CES1 substrates, such as angiotensin-converting enzyme inhibitor prodrugs, antiviral agents (e.g., remdesivir, oseltamivir, and sofosbuvir), central nervous system agents (e.g., methylphenidate and cocaine), anticancer agents (e.g., capecitabine and irinotecan), and antiplatelet/anticoagulation drugs (e.g., clopidogrel and dabigatran) (Table 1).

TABLE 1.

Substrate drugs of major drug-metabolizing hydrolases

Hydrolase	Substrate Drugs
CES1	Enalapril, imidapril, benzapril, quinapril, ramipril, trandolapril, oseltamivir, sofosbuvir, tenofovir alafenamide, sacubitril, methylphenidate, cocaine, heroin, meperidine, flumazenil, rufinamide, capecitabine, irinotecan, telotristat etiprate, clopidogrel, dabigatran, mycophenolate mofetil, clofibrate, fenofibrate, ciclesonide, dimethyl fumarate, oxybutynin
CES2	Acebutolol, aspirin, flupirtine, flutamide, heroin, irinotecan, gemcitabine, doxazolidine, molnupravir, capecitabine
AADAC	Eslicarbazepine acetate, ketoconazole, flutamide, phenacetin, rifampicin, rifabutin, rifapentine, indiplon
PON1	Olmesartan medoxomil, pilocarpine, prulifloxacin, dihydrocoumarin
PON3	Lovastatin, mevastatin, simvastatin, spironolactone
CSTA	Tenofovir alfanamide, sofosbuvir, remdesivir

Fig. 2.

Heat map of mRNA expression (parts per billion) of major drug-metabolizing hydrolases in different tissues. Expression data were obtained from the Expression Atlas database.

Pharmacogenetics of CES1

There is considerable variation in the expression and activity of hepatic CES1 among individuals, which may be partially attributed to genetic polymorphisms. Of all the genetic variants that have been studied to date, six CES1 genetic variants have been found to potentially affect the PK and PD of CES1 drug substrates: rs71647871(G143E), rs200707504 (E220G), rs2307240 (S75N), rs3785161 (-816A>C), rs3815583 (-75G>T), and rs2244613 (1168-33C>A).

The genetic variant rs71647871 is the first loss-of-function mutation identified in the CES1 gene. Its discovery originated from a PK study to evaluate the drug-drug interaction (DDI) between methylphenidate and alcohol (Patrick et al., 2007). A study participant’s C_max values of d- and l-methylphenidate were 7-fold and 100-fold higher than the rest of the study participants, respectively. Later analysis demonstrated that the nonsynonymous variant rs71647871 was responsible for the poor metabolizer phenotype (Zhu et al., 2008). Following this discovery, many clinical studies have validated the significant effects of this variant on the metabolism, PK, and PD of other CES1 substrates, including clopidogrel (Lewis et al., 2013; Tarkiainen et al., 2015a; Bozzi et al., 2016; Jiang et al., 2016), angiotensin-converting enzyme inhibitor prodrugs (Tarkiainen et al., 2015b; Wang et al., 2016; Her et al., 2021), oseltamivir (Zhu and Markowitz, 2009; Tarkiainen et al., 2012), dabigatran (Shi et al., 2016b), and sacubitril (Shi et al., 2016a).

In an in silico analysis, the nonsynonymous variant rs200707504 (E220G), also referred to as c.662A>G, was predicted to reduce CES1 enzymatic activity (Oh et al., 2017). Consistent with the prediction, Wang et al. (2017) reported markedly decreased CES1 activity of the variant on metabolizing CES1 substrate drugs (enalapril, clopidogrel, and sacubitril) in an in vitro study using transfected cell lines. A PK study of a single dose of oseltamivir (75 mg) was completed to evaluate the clinical significance of rs200707504 in CES1-mediated activation of the prodrug in 20 healthy Korean volunteers (Oh et al., 2017). The results showed that the variant carriers (n = 8) had a 10% increase in the area under the curve between 0 and 48 hours of the parent drug and a 5% decrease in the area under the curve between 0 and 48 hours of its active metabolite compared with the noncarrier group (n = 12). Nonetheless, the differences were not statistically significant.

The CES1 nonsynonymous single nucleotide polymorphism (SNP) rs2307240 (S75N) is prevalent across diverse populations, exhibiting a minor allele frequency (MAF) between 2% and 7%. A retrospective study was conducted to examine the influence of the variant on the outcome of clopidogrel treatment in 851 patients diagnosed with coronary syndrome (Xiao et al., 2017). The CES1 S75N carriers (n = 375) showed a higher incidence of cerebrovascular events (P < 0.001), acute myocardial infarction (P < 0.001), and unstable angina (P < 0.001) compared with noncarriers. The study further revealed that the S75N SNP was more common in acute coronary syndrome patients (MAF, 22%) than in the general population (MAF, 5%). However, there are conflicting findings with this result, where the S75N carriers did not have any significant difference in outcomes of patients treated with the CES1 substrate methylphenidate, and an in vitro study showed that the S75N variant did not significantly alter the CES1 activity and expression in transfected cells and human livers (Johnson et al., 2013; Wang et al., 2017).

Other CES1 SNPs, such as rs3785161 (-816A>C), rs3815583 (-75G>T), and rs2244613 (1168-33C>A), have been investigated in limited studies. The results were inconsistent, and therefore, they are not discussed in this review. In summary, although substantial efforts have been devoted to the study of CES1 pharmacogenetics in the past decade, the functions of many CES1 genetic variants remain to be determined.

Nongenetic Regulators of CES1

Developmental CES1 Expression

Considerable evidence demonstrates a developmental expression pattern of CES1 in both human and mouse livers, with hepatic CES1 expression progressively increasing with age. (Zhu et al., 2009; Hines et al., 2016; Boberg et al., 2017). An in vitro study divided human liver samples into five age groups: 1–31 days old (group 1), 35–70 days old (group 2), 89–119 days old (group 3), 123–198 days old (group 4), and over 18 years old (group 5). Neonates (group 1) exhibited 10% of the hydrolysis rate and CES1 expression compared with adults (group 5). The rest of the pediatric groups (groups 2–4) had approximately 50% of the CES1 expression and hydrolytic activity compared with that of the adults (Shi et al., 2011). In an in vitro study involving 165 human liver samples across various pediatric age groups ranging from birth to 18 years old, CES1 protein levels were positively correlated to age. (P < 0.001) (Hines et al., 2016). Thus, CES1 expression and activity levels are lower in neonates and pediatric cohorts and gradually increase with age. The results necessitate investigations to assess the potential influence of CES1 maturation on the treatment response of CES1 substrate medications in pediatric patients.

Sex Differences in CES1 Expression

In both in vitro and clinical settings, evidence suggests that CES1 expression is greater in females compared with males (Patrick et al., 2007; Zhu et al., 2009; Shi et al., 2016b). Males demonstrated significantly greater exposure to d-methylphenidate than females (Patrick et al., 2007), suggesting a higher CES1 function in females. An in vitro study reported that there was a significantly higher CES1 activity in female human livers (n = 56) compared with male samples (n = 46) (Shi et al., 2016b). However, another study using human liver samples (n = 32) and mouse liver samples (n = 9) did not show a significant difference in sex regarding CES1 activity (Zhu et al., 2009).

CES1-Mediated Drug-Drug Interactions

CES1 Inhibitors

Alcohol is currently the only clinically significant CES1 inhibitor that has been confirmed in multiple in vivo and in vitro studies (Griffin et al., 2010, 2013; Bell et al., 2011; Zhu et al., 2017). To eliminate the potential confounding effect of ethanol on methylphenidate absorption caused by faster gastric dissolution due to ethanol coadministration, a pulsatile dosing regimen was adopted for methylphenidate and ethanol. Notably, d-methylphenidate exhibits a pharmacological potency 10-fold higher than l-methylphenidate, whereas l-methylphenidate is a more efficient substrate for CES1. D-methylphenidate can come as a single active ingredient (Focalin) or in a racemic mixture with l-methylphenidate (Ritalin). When alcohol (0.6 g/kg) and d-methylphenidate were coadministered, the C_max of d-methylphenidate was increased by 27% (P < 0.01); when alcohol and dl-methylphenidate were coadministered, the C_max of d-methylphenidate in the racemic mixture was increased by 35% (P < 0.01), and area under the curve between 4 and 8 hours was increased by 25% (P < 0.05) (Zhu et al., 2017).

Aside from alcohol, there are many reported CES1 inhibitors, such as cannabis, protease inhibitors, aripiprazole, calcium channel blockers, tacrolimus, and valproate. An in vitro study with CES1-transfected cells reported that the main psychoactive constituents of cannabis (i.e., tetrahydrocannabinol, cannabidiol, and cannabinol) were potential CES1 inhibitors (Qian et al., 2019). The inhibition constant (K_i) values for tetrahydrocannabinol, cannabidiol, and cannabinol were 0.541, 0.974, and 0.263 µM, respectively. Protease inhibitors (i.e., nelfinavir, amprenavir, atazanavir, ritonavir, and saquinavir) were suggested to be potential inhibitors as well based on an in silico analysis and an in vitro study (Rhoades et al., 2012). Nelfinavir demonstrated markedly stronger inhibitory effects relative to other protease inhibitors (K_i, 3.7 ± 0.7 µM) in the in vitro incubation study using p-nitrophenyl acetate (PNPA) as the CES1 substrate. Aripiprazole, perphenazine, thioridazine, and fluoxetine were identified as potent inhibitors of CES1 in an in vitro study, which was further supported by a PK study in mice, demonstrating that the coadministration of aripiprazole and dl-methylphenidate significantly increased the plasma concentration of dl-methylphenidate (P < 0.01) (Zhu et al., 2010). The IC₅₀ values of aripiprazole, perphenazine, thioridazine, and fluoxetine were 5.7 µM, 13.9 µM, 7.0 µM, and 6.1 µM, respectively. Furthermore, 26 known cardiovascular, antiplatelet, and immunosuppressant drugs have been screened for potential CES1 inhibition utilizing human liver microsomes and recombinant CES1 (Thomsen et al., 2014). The results indicated that 100 µM isradipine (a dihydropyridine calcium channel blocker) and 100 µM tacrolimus (an immunosuppressant) decreased CES1 activity by 17.6% and 28.4%, respectively. Lastly, valproate was suggested to be a CES1 inhibitor by an in vitro study in both microsomes and cytosol and significantly affected rufinamide metabolism (K_i, 363 ± 39 µM; IC₅₀, 2.52 ± 1.32 mM; Williams et al., 2011). This potential DDI could be clinically significant as rufinamide and valproate are often prescribed together as antiepileptic therapy when monotherapy is ineffective.

CES1 Inducers

Currently, CES1 inducers are understudied relative to CES1 inhibitors. Staudinger et al. (2010) showed that various nuclear receptors may be involved in CES1 expression regulation. Since then, several synthetic peroxisome proliferator-activated receptor (PPAR) agonists have been found to induce mRNA expression of several CES1 isoforms in mouse livers (Jones et al., 2013). Zhu et al. (2000) observed a modest elevation in CES1 expression in human hepatocytes upon exposure to 10 µM rifampicin, a prototypical inducer of the human pregnane X receptor. Similarly, Wen et al. (2019) found that the environmental contaminant perfluorooctanoic modulated the expression and function of hepatic CES enzymes, in part through PPARα. They found an increase in Ces1d, 1e, 1f, 2c, and 2e mRNAs between 1.5- and 2.5-fold while seeing an up-regulation of Cyp4a14 mRNA, an indicator of PPARα activation.

Additionally, in vivo findings indicated that glucose induces hepatic CES1 expression by enhancing CES1 promoter activity and histone 3 and 4 acetylations within CES1 chromatin, suggesting CES1 involvement in glucose homeostasis (Xu et al., 2014b). Further, phenobarbital was found to increase CES1 expression in mouse liver, where neonatal mice (treated with 20 mg/kg per day) showed higher inducibility than that of adult mice (treated with 80 mg/kg per day) (Xiao et al., 2012).

Carboxylesterase 2

Different from CES1, CES2 is mainly expressed in the small intestine and also expressed to a lesser extent in the kidney, liver, heart, brain, and testis (Fig. 2) (Satoh et al., 2002; Taketani et al., 2007; Sanghani et al., 2009). CES2 substrates include acebutolol, aspirin, flupirtine, flutamide, heroin, irinotecan, gemcitabine (Pratt et al., 2013), doxazolidine (Barthel et al., 2009), molnupravir (Shen et al., 2022), and capecitabine (Quinney et al., 2005) (Table 1). Similar to CES1, CES2 expression and activity can be regulated by both genetic and nongenetic factors.

Pharmacogenetics of CES2

The study of human CES2 activity on hydrolyzing irinotecan revealed remarkable interindividual variabilities (Xu et al., 2002), which has led to further pharmacogenetic investigations. Marsh et al. (2004) found that there was a significant difference in Europeans and Africans regarding CES2 allele and haplotype frequencies. Eleven CES2 SNPs were identified in at least one population. Although reduced CES2 mRNA expression was found in colorectal tumors with the CES2 intronic SNP (IVS10-88), the SNPs were not associated with mRNA expression in normal colonic mucosa (Marsh et al., 2004).

To date, only a few studies have examined the association between CES2 genetic polymorphisms and response to CES2 drug substrates. Aspirin, a known CES2 substrate drug, is frequently prescribed for primary and secondary prevention of cardiovascular disease. However, around 25%–40% of patients exhibit resistance to aspirin therapy. Tang et al. (2006) initially explored whether aspirin and clopidogrel were affected by competitive inhibition. Interestingly, the results from human liver and intestinal microsome incubations suggested that clopidogrel is primarily hydrolyzed by CES1, whereas aspirin is hydrolyzed by CES2. In addition, they identified an SNP (A139T) that was associated with a 40% decrease in CES2 function in aspirin hydrolysis (Tang et al., 2006). However, conflicting results were reported as 293T cells transfected with the A139T variant, and the wild-type CES2 did not show a significant difference in hydrolyzing the CES2 substrate molnupiravir (Shen et al., 2022).

Two missense variants of CES2, R34W (rs72547531) and V142M (rs72547532), were identified in 165 Japanese cancer patients (Kubo et al., 2005). The mean V_max values on hydrolyzing irinotecan, p-nitrophenyl acetate, and 4-methylumberiferyl acetate were 1.458 ± 0.0495 pmol/mg protein/min, <0.2 µmol/mg per min, and <0.2 µmol/mg per min in COS-1 cells transfected with wild type, R34W, and V142M, respectively. The expression levels of the two variants in the liver were found to be 252% and 360% higher compared with the wild type. This result indicates that the two variants have high expression of CES2 but a deficient catalytic CES2 activity in metabolizing irinotecan. Additionally, a follow-up study by Kim et al. (2007) reported that a novel SNP M1L and the previously reported R34W were associated with 40% and 64% reductions, respectively, in the area under the curve ratios of irinotecan to its active metabolite, SN-38, in patients.

One of the newly discovered CES2 substrates is molnupiravir (Jayk Bernal et al., 2022), a COVID-19 oral drug that was recently granted emergency use authorization by the FDA. The study evaluated several CES2 variants for their hydrolytic capacity in hydrolyzing molnupiravir using cell lysates prepared from cells transfected with the variants. The variants A139T and L45I hydrolyzed molnupiravir with efficacies comparable to the wild-type CES2. However, the variants A178V and F485V were found to reduce the hydrolytic activity by 74% and 61%, respectively, compared with the wild-type enzyme. Moreover, the variant R180H showed a 77% increase in hydrolytic activity toward molnupiravir compared with the wild-type CES2 (Shen et al., 2022).

The findings were consistent with an in vitro study from the same research group (Xiao et al., 2013). The study found that the CES2 variants A139T and A178V were resistant to orlistat inhibition, agreeing with the decreased hydrolysis function in the two variants reported in the studies above. Additionally, the R180H variant was found to be sensitive to orlistat inhibition, which aligns with the increased hydrolytic activity of this variant suggested by Shen et al. (2022) (Xiao et al., 2013).

Overall, very few studies have been conducted to explore the impact of CES2 genetic polymorphisms on the metabolism of CES2 substrates, and it is undetermined whether CES2 variants could significantly alter the PK of CES2 substrate drugs and consequently affect the therapeutic responses of these drugs in patients.

Nongenetic Regulators of CES2

CES2-Mediated Drug-Drug Interactions

CES2 Inhibitors

Several CES2 inhibitors have been identified by in vitro studies. An in vitro study demonstrated that simvastatin and fenofibrate significantly inhibited CES2-mediated hydrolysis of irinotecan (CPT-11) (Fukami et al., 2010). The K_i values for simvastatin and fenofibrate were 0.67 ± 0.09 and 0.04 ± 0.01 μmol/L, respectively. In another study, diltiazem and verapamil showed significant inhibitory effects against CES2, with K_i values of 0.25 ± 0.02 and 3.84 ± 0.99 μM, respectively (Yanjiao et al., 2013). Additionally, a study found that 1 nM of orlistat inhibited CES2 activity by 75% but had no inhibitory effect on CES1 activity (Xiao et al., 2013). The same research group also reported that 1μM of remdesivir, the first COVID-19 drug granted full approval by the FDA, inhibited molnupiravir hydrolysis by 90% in intestinal microsomes (Shen et al., 2022). The inhibition was observed in other organs known to express CES2, such as the liver and kidney (approximately 40% and 60%, respectively) (Shen et al., 2022). The same laboratory demonstrated that 5 μM of the anti–human immunodeficiency virus drug sofosbuvir decreased the hydrolysis activity in the kidney by 82% in vitro (Shen and Yan, 2017). Moreover, in vitro experiments were conducted to assess the potential of remdesivir and sofosbuvir in protecting against the irinotecan-mediated suppression of intestinal epithelium regeneration (Eades et al., 2022). The results indicated that remdesivir and sofosbuvir both reduced irinotecan hydrolysis and reversed the irinotecan-reduced formation of organoids. These findings suggest a potential role of CES2 inhibitors in reducing the gastrointestinal (GI) toxicity effects of irinotecan.

Lastly, loperamide was studied for its inhibitory potential for the CES2-catalyzed hydrolysis of capecitabine, a prodrug used for colorectal and breast cancers (Quinney et al., 2005). By using recombinant CES2 prepared from transfected Sf9 insect cells, the study suggested that CES2 is responsible for the hydrolysis of capecitabine into its active metabolite, 5′-DFCR, and loperamide is a strong CES2 inhibitor with a K_i of 1.5 μM (Quinney et al., 2005). Of note, many providers use loperamide to treat chemotherapy-induced GI toxicities. The study indicates that inhibiting CES2 by loperamide could lead to potential DDIs with other chemotherapy drugs.

CES2 Inducer

Compared with the research done on CES2 inhibitors, the studies on CES2 inducers are limited. Recent studies have shown that activating constitutive androstane receptor, aryl hydrocarbon receptor, PPARs, LXR, pregnane X receptor hepatocyte nuclear factor-4α, and nuclear factor erythroid 2–related factor 2 (Nrf2) transcriptional pathways could regulate the expression of mammalian CES1 and CES2 (Staudinger et al., 2010; Jones et al., 2013; Xu et al., 2014a). Rifampicin, a prototypical human pregnane X receptor–activating agent, has been shown to moderately increase CES2 mRNA and protein expressions in mice (Zhang et al., 2012; Jones et al., 2013). This result was consistent with an in vitro study where Shi et al. observed a moderate increase in CES2 expression in human hepatocytes when treated with rifampicin (Shi et al., 2008). Additionally, CES2 expression was found to be potentially induced by NO1886 (ibrolipim), a lipoprotein lipase–promoting agent, in primary cultures of cryopreserved human hepatocytes (Morioka et al., 2006). Lastly, urethane dimethacrylate was found to induce the mRNA expression of CES2 in human dental pulp cells (Chang et al., 2014). Further investigations are warranted to study the impact of CES2 inducers on drug metabolism and the PK and PD of CES2 substrate drugs.

Arylacetamide Deacetylase

Human AADAC, a microsomal serine esterase with a molecular mass of 45 kDa, hydrolyzes many acetyl-containing drugs, with a strong preference for those with small acyl moieties (Fukami et al., 2015). In contrast to CES enzymes, AADAC exhibits a type II membrane protein configuration, where it retains an uncleaved N-terminal signal anchor sequence, ensuring its localization on the lumen side of the endoplasmic reticulum (Frick et al., 2004). Moreover, AADAC is classified as a lipase due to its active site exhibiting strong homology with that of hormone-sensitive lipase. Notably, when expressed in yeast, human AADAC has been shown to effectively hydrolyze cholesterol ester (Probst et al., 1994). Recently, a study reported a lack of significant correlation between the levels of AADAC mRNA and protein expressions in human livers (Sakai et al., 2024), indicating a potential contribution of post-transcriptional regulation to AADAC protein expression.

AADAC substrates include many clinically important medications (Table 1), such as eslicarbazepine acetate, ketoconazole, phenacetin, rifampin, rifabutin, rifapentine, and flutamide (Fukami et al., 2022). The enzyme has also shown some potential involvement in lipid metabolism. N-deacetyl ketoconazole (DAK), a metabolite of ketoconazole, can trigger hepatocellular toxicity (Rodriguez and Acosta, 1997). Interestingly, Fukami et al. (2016) conducted kinetic and inhibition studies in human liver microsomes and identified AADAC as the enzyme that metabolizes ketoconazole to DAK. AADAC is expressed in the liver and small intestine and is also present in the lungs at a lower level (Fig. 2) (Gabriele et al., 2019; Fukami et al., 2022). Overexpression of AADAC in ovarian cancer tissues has been found to promote therapeutic activity of cisplatin and imatinib against ovarian cancer cells (Wang et al., 2022).

Pharmacogenetics of AADAC

Gabriele et al. (2019) not only demonstrated the presence of AADAC protein in the lung through immunoblot analysis but also observed high interindividual variability in both AADAC mRNA levels and the enzymatic activity in hydrolyzing phenacetin, a known AADAC substrate. The impact of SNPs on the AADAC’s activity hydrolyzing eslicarbazepine acetate was studied in small intestine and liver microsomes (Hirosawa et al., 2021), where samples with the SNPs R248S or X400Q showed lower activity than the wild type (5% or 21%, respectively), and those with V172I showed higher activity than the wild type (174%). Similar trends were observed for other known substrates of AADAC, such as p-nitrophenyl acetate, ketoconazole, phenacetin, and rifampicin. This study was the first to demonstrate eslicarbazepine acetate as a substrate of AADAC in both the liver and the small intestine (Hirosawa et al., 2021). Interestingly, a study showed that black participants with one G allele of the AADAC variant rs1803155 (c.931G>A, V281I) were 3 times as likely to have rifapentine exposure below the target bactericidal level than black participants with the A allele (odds ratio, 2.97; 95% confidence interval, 1.16, 7.58), and the odds ratio was even greater for those with two G alleles. However, in non-black participants, rs1803155 was not associated with rifapentine exposure (Weiner et al., 2021). Shimizu et al. (2012) observed that the AADAC protein produced by the AADAC*3 allele (g.13651G>A/g.14008T>C; V281I/X400Q) showed substantially lower intrinsic clearance values compared with wild-type proteins based on the in vitro hydrolysis rates on metabolizing the substrates flutamide, phenacetin, and rifampicin.

Nongenetic Regulators of AADAC

AADAC Inhibitors

In a study that screened natural compounds for their inhibitory effects on AADAC, CES1, and CES2, Yasuda et al. (2021) found that curcumin and quercetin exhibited strong inhibitory effects against all three enzymes, whereas epicatechin, epicatechin gallate (ECg), and epigallocatechin gallate (EGCg) specifically inhibited AADAC. ECg and EGCg strongly inhibited AADAC-mediated rifampicin hydrolysis in human liver microsomes, with IC₅₀ values of 2.2 ± 1.4 µM and 1.7 ± 0.4 µM, respectively. One study investigated the regulation of mRNA expression of AADAC by 15 microsomal enzyme inducers in male mouse livers. The data suggested that Aadac mRNA expression was suppressed by three aryl hydrocarbon receptor ligands, two constitutive androstane receptor activators, two pregnane X receptor (PXR) ligands, and one Nrf2 activator (Zhang et al., 2012). As previously mentioned, Sakai et al. (2024) found that there was no significant correlation between the mRNA and protein expression of AADAC, indicating a potentially significant involvement of AADAC protein expression regulation through post-transcriptional and post-translational modification. Overexpression of miR-222-3p in Huh-1 cells resulted in increased lipid accumulation, which was reversed by AADAC overexpression. Thus, miR-222-3p is a potential regulator of AADAC expression in human livers and could impact both drug metabolism and lipid accumulation.

AADAC Inducers

Similarly to carboxylesterases, AADAC has also been found to be regulated by nuclear receptors at the transcriptional level. AADAC mRNA and protein levels were significantly induced by the PPARα ligands fenofibric acid (1.8-fold) and WY-14643 (3.8-fold) in human hepatoma Huh-7 cells (Morikawa et al., 2022). Consistent with previous studies, the knockdown and overexpression of PPAR-α resulted in decreased and increased expressions of AADAC, respectively (Trickett et al., 2001). The study also showed that AADAC plays a role in suppressing cellular lipid accumulation.

Paraoxonases

Human PON consists of three isoforms: PON1, 2, and 3. The PON genes are located on chromosome 7, forming a cluster, and their proteins share approximately 60% amino acid sequences. Unlike CES and AADAC, PONs do not possess a serine residue in the active site and, thus, do not belong to the serine hydrolase family. PONs are known to efficiently hydrolyze organophosphates like paraoxon, sarin, and soman, detoxifying these compounds. PON substrate drugs are summarized in Table 1. PON1, an enzyme approximately 40 kDa in size, is primarily expressed in the liver and subsequently released into the bloodstream, where it becomes associated with high-density lipoprotein particles (La Du et al., 1999). Although the structure of PON3 is currently unknown, it is postulated that it is similar to that of PON1, with stronger catalytic activity toward lactones. Tripathy et al. (2017) delineated the active site of PON1 by analyzing the effects of mutations on its interaction with the aryl ester substrate phenyl acetate. This investigation revealed that phenyl acetate binds to the carbonyl oxygen directed toward the catalytic calcium that is positioned at a distance of 2.3 Å. The imidazole ring of residue H115 was positioned in a manner that brought phenyl acetate and imidazole into close proximity (with a distance of 2.9 Å between the carbonyl of phenyl acetate and histidine). Additionally, extensive hydrophobic and aromatic stacking interactions were observed between phenyl acetate and residues F222, I291, F292, L240, and V346 within the active site of PON1 (Tripathy et al., 2017). Essential for hydrolytic activity in human PON1 are 16 residues, all of which, except His-243 and Trp-281, are conserved in rabbit PON3. Two PON1 mutants, H243K and W281L, which correspond to the PON3 sequence, demonstrate markedly reduced arylesterase and organophosphate activity (Josse et al., 1999). This finding may explain, at least partially, the observed limited arylesterase and paraoxonase activity in rabbit serum PON3 as noted by Draganov et al. (2000, 2005). PON2, given the lack of information regarding its role in drug metabolism, is excluded from this review.

Paraoxonase 1

PON1 is named for its ability to hydrolyze paraoxon, the active metabolite of the organophosphorus insecticide parathion, representing its primary and extensively studied substrate (Costa et al., 2011). PON1 is highly expressed in the liver and plasma and capable of metabolizing olmesartan, medoxomil, pilocarpine, spironolactone, mevastatin, simvastatin, lovastatin, and prulifloxacin (Table 1). PON1 also is known to protect low-density lipoproteins and high-density lipoproteins (HDLs) from oxidation (Costa et al., 2003). Extensive research has been conducted on PON1’s involvement in regulating susceptibility to organophosphorus insecticides, cardiovascular disease, and other medical conditions (Costa et al., 2003; Ng et al., 2005).

Pharmacogenetics of PON1

Two PON1 polymorphisms, Q192R and L55M, are known to hinder the catalytic efficiency of PON1. Ishizuka et al. (2012) constructed two recombinant allozymes of human PON1 (PON1 192QQ and PON1 192RR) and demonstrated that the catalytic efficacy of PON1 192RR on hydrolyzing olmesartan was slightly greater than that of the PON1 192QQ. A meta-analysis evaluated the association of organophosphate toxicity risk with the Q192R and L55M polymorphisms, which included nine case-control studies with a total of 1042 patients with organophosphate toxicity and 1014 healthy controls (You et al., 2013). The results indicated that there was an increased risk of organophosphate toxicity among the Caucasian population with PON1 192Q and 55L polymorphisms. Interestingly, similar associations were not seen in Asian populations. However, given that there is high interindividual variability of PON1 activity even within the same PON1 genotype, a high-throughput two-dimensional enzyme assay using two PON1 substrates (usually diazoxon and paraoxon at high salt concentrations) was developed (Richter and Furlong, 1999). This approach provides an assessment of PON1 serum levels, which has been referred to as a determination of PON1 “status” for an individual.

Nongenetic Regulators of PON1

Age and Gender

Age appears to play an important role in regulating PON1 expression. PON1 activity is very low before birth and increases over time during the first 2 years of life in humans (Costa et al., 2005). Additionally, it has been found that PON1 activity decreases with aging, potentially due to the development of oxidative stress conditions (Seres et al., 2004). Females have shown higher PON1 activity than males (Costa et al., 2005, 2011).

PON1-Mediated Drug-Drug Interactions

Given the role of PON1 in protecting against toxic pesticide exposure, metabolizing therapeutic agents, and a number of pathologic conditions (Costa et al., 2003; Camps et al., 2009), there has been an effort to understand the modulation of PON1 activity and expression (Durrington et al., 2002; Costa et al., 2005, 2011). A few clinical studies have been conducted to investigate PON1 regulation in humans. There have been reports of small increases (range, 5%–23%) of serum paraoxonase activity in humans following the treatments of simvastatin or atorvastatin (Kassai et al., 2007; Mirdamadi et al., 2008; Harangi et al., 2009). However, further studies indicated contrasting results regarding the effects of atorvastatin and simvastatin on regulating PON1 (Dullaart et al., 2009). Animal and in vitro studies provided additional contrasting results. Two studies by Beltowski et al. (2010) showed that fluvastatin decreased serum and liver paroxonase activity in rats, whereas pravastatin had no effects (Costa et al., 2005). Moreover, PON1 mRNA and activity levels were decreased by pravastatin, simvastatin, and fluvastatin in a human hepatoma cell line (HuH7) (Gouédard et al., 2003).

Nuclear hormone receptors may be involved in regulating PON1 activity (Gouédard et al., 2003; Shih et al., 2006; Beltowski et al., 2010). The effects of statins on PON1 were antagonized by the LXR receptor activator 22(R)-hydroxycholesterol. Additionally, simvastatin upregulated the activity of the PON1 promoter by increasing a nuclear transcription factor, sterol regulatory element binding protein 2 (SREBP-2); these effects were reversed by mevalonate as well. Moreover, the LXR agonist T091317 restored the serum PON1 activity reduced by leptin in rats (Beltowski et al., 2010). The farnesoid X receptor, a target for bile acids, has been found to decrease serum PON1 activity by approximately 65% and liver PON1 expression in mice upon activation (Shih et al., 2006).

Given that moderate use of alcohol could have beneficial cardiovascular effects through its HDL-modulating effect, investigators were interested in evaluating the effects of alcohol on PON1 expression and activity (Rao et al., 2003). Initially, a comparative study examined the impact of moderate alcohol consumption in rats and humans. In rats, moderate alcohol consumption for 8 weeks led to a range of 20% to 25% increase (P < 0.05) of serum and liver PON1 activity, consistent with a 59% increase (P < 0.001) in hepatic PON1 mRNA levels (Rao et al., 2003). Similarly, in humans, moderate alcohol consumption (13–39 g/day for 6 months or longer) resulted in a greater than threefold increase in serum PON1 activity. Further studies suggested that the effect of alcohol on PON1 may be mediated by protein kinase C, which may phosphorylate Sp1 and regulate its binding to the Sp1 binding site in the promoter region of PON1 (Osaki et al., 2004). On the other hand, heavy consumption of alcohol could produce opposite effects on PON1 activity. In rats, 8-week administration of alcohol [blood concentration of 30 mM (140 mg/dL)] decreased serum and hepatic PON1 activity by 25% (P < 0.05) along with the reduction of PON1 mRNA levels (Rao et al., 2003). The results were consistent with the findings from humans as an 80-g/day consumption of alcohol led to a 45% decrease (P < 0.001) in serum PON1 activity (Rao et al., 2003). Moreover, they observed a 53% decrease in serum PON1 activity in chronic alcoholics and a further 72% decrease in alcoholics with liver cirrhosis (Marsillach et al., 2007a). Further, administration of a diet containing 36% alcohol in rats decreased plasma PON1 activity by 23%–58% (Varatharajalu et al., 2010). Of note, these changes were attenuated by the coadministration of betaine (trimethyl glycline).

Other medications reported to increase PON1 activity in animals or humans include fibrates (gemfibrozil, fenofibrate) (Aviram et al., 1998; Paragh et al., 2006), probucol (Hong et al., 2006), ezetimibe (Turfaner et al., 2010), aspirin (Blatter-Garin et al., 2003; Jaichander et al., 2008), rosiglitazone (van Wijk et al., 2006; Carreón-Torres et al., 2009; Ackerman et al., 2010), eplerenone (Noll et al., 2010), sulfonylureas (glimepiride, gibenclamide) (Wójcicka et al., 2010), and erythropoietin β (Marsillach et al., 2007b). There have been extensive studies regarding the influence on PON1 activity by statins, fibrates, and aspirin; however, the results from these studies are often conflicting. Further studies are warranted to fully determine the effects of these medications on PON1 expression and activity.

Paraoxonase 3

PON3, a 40-kDa glycoprotein, is the least characterized member of the paroxonase family. PON3 is mainly synthesized in the liver and expressed at a much lower level in the kidney. Like PON1, PON3 is often found in circulation tightly bound to HDLs. PON3, like PON2, lacks the ability to hydrolyze organophosphates. Nevertheless, it shares the lactonase and N-acyl homoserine lactone activities with PON1. PON3 shows greater catalytic activity for statin lactones than PON1. Thus, statin lactones, like lovastatin, spironolactone, and canrenone, are commonly used to measure PON3 activity (Draganov et al., 2000, 2005).

Pharmacogenetic Factors of PON3

There are a limited number of studies concerning the polymorphisms of PON3. In a study of 1143 healthy blood donors in southern Italy, 250 DNA samples were randomly selected to identify polymorphisms of PON3 (Campo et al., 2004). The investigators identified three silent (G51G, G73G, G99G) and two missense (S311T, G324D) variants in exons 3, 4, and 9 of PON3. Their frequencies were relatively low with respect to known SNPs of PON1 and PON2. The effect of these variants on the metabolic activity of PON3 is yet to be evaluated. However, three PON3 promoter SNPs (PON3-567, rs11764079; PON3-665, rs11770903; PON3-746, rs17882539) in linkage disequilibrium were found to be significantly associated with changes in serum PON3 concentrations in healthy Mediterranean subjects (n = 356) (Aragonès et al., 2011). It is important to note that the same research group conducted studies in human immunodeficiency virus patients (Aragonès et al., 2012) and coronary artery disease and peripheral artery disease patients (Rull et al., 2012) and reported no significant differences between the SNP carriers and control patients, indicating that the PON3 genotype neither influences serum PON3 levels nor the course of these diseases. Riedmaier et al. (2011) investigated PON1 and PON3 polymorphisms in an in vitro study of atorvastatin-lactone hydrolysis. The common variations analyzed in this study did not affect PON3 expression; one variant located approximately 12 kb downstream of the last PON3 exon and approximately 23 kb upstream of PON1 (var55146; MAF, 5%) was found to be associated with not only increased protein expression of PON3 but also marginally increased atorvastatin-lactone hydrolysis. However, whether this association is due to the high linkage to the six intronic SNPs, unidentified polymorphism, or a downstream enhancer element of PON3 remains to be elucidated. Finally, the investigators conducted a multivariate analysis and found that nongenetic factors clearly influenced PON3 expression more than genetic polymorphisms.

Nongenetic Regulators of PON3

Riedmaier et al. (2011) examined nongenetic factors that potentially impact PON3 activity, such as C-reactive protein (CRP), γ-glutamyl transferase (GGT), bilirubin, or treatment with omeprazole or pantoprazole. In a univariate analysis, clinically elevated CRP levels (>8.2 mg/dL) were associated with 2.6-fold lower PON3 expression (P < 0.01) and lower atorvastatin lactone hydrolysis (1.7-fold; P < 0.05). Similarly, elevated levels of GGT (>64 in men, >36 in women) were associated with a decrease in PON3 expression (1.4-fold) and atorvastatin-lactone hydrolysis (P < 0.01). Bilirubin levels higher than 1.2 mg/dL were associated with 1.8-fold decreased PON3 expression (P < 0.01) and 1.3-fold decreased atorvastatin lactone hydrolysis (P < 0.05). Presurgical drug exposure to omeprazole or pantoprazole was associated with a significant decrease of PON3 mRNA (1.3-fold; P < 0.05), protein (1.6-fold; P < 0.01), and activity (1.7-fold; P < 0.01) compared with patients receiving no drugs (Riedmaier et al., 2011).

Multivariate linear model and stepwise model selection were used to understand the contribution of polymorphisms and nongenetic factors to atorvastatin-lactone hydrolysis and PON3 expression variability. Of the atorvastatin-lactone hydrolysis variability, 15.0% were explained by F21F (rs13226149), PON3-1091A>G promoter polymorphism, and the nongenetic factors, including sex, CRP, GGT, and cancer classification. PON3 expression appeared to be influenced by nongenetic factors by 20.0% when smoking status, cancer classification, CRP, bilirubin level, and GGT were included. Note that the mRNA variation could not be explained reasonably throughout the analysis (Riedmaier et al., 2011).

Cathepsin A

Initially synthesized as a zymogen, the serine hydrolase CTSA exists as a single-chain precursor weighing 54 kDa, which subsequently self-associates into a homodimer. The primary distinction between the precursor enzyme and the active CTSA lies in the lack of a 40-residue activation domain (Schreuder et al., 2014). The crystal structure of the precursor CTSA reveals that each monomer comprises two domains, including a core insertion (Kolli and Garman, 2014). This core domain adopts an α/β-hydrolase fold housing the active site, a characteristic of other serine carboxypeptidases (i.e., CES1, CES2, and AADAC). Within this fold lies a catalytic triad residue (Ser150, His429, and Asp372) responsible for hydrolyzing its substrates.

CTSA is highly expressed in the kidney, liver, lung, and GI tract (duodenum, colon, stomach) (Wang et al., 2019). CTSA exhibits both protease and carboxylesterase activity and is involved in various physiologic and pathologic processes, such as being essential for the stability and activation of β-galactosidase and neuraminidase (Galjart et al., 1991). Furthermore, CTSA has emerged as a highly efficient DME, playing a pivotal role in the initial hydrolysis step for activating the antiviral prodrugs tenofovir alfanamide, sofosbuvir, and remdesivir (Birkus et al., 2007, 2008; Murakami et al., 2010; Li et al., 2021).

Pharmacogenetics of CTSA

The sole pharmacogenetics study on CTSA was conducted by Ito et al. (2021), revealing nine genetic polymorphisms in the CTSA gene in healthy Japanese subjects. Among these variants, one polymorphism in exon 2 (85_87CTG>-) resulted in the deletion of a leucine within a 9-leucine repeat (Leu9) in the signaling peptide region of the CTSA protein. The consequential 8-leucine repeat (Leu8) caused by the 85_87CTG>- polymorphism led to an approximately two-fold increase in CTSA expression (P < 0.05) when compared with Leu9 in Flp-In-293 cells stably expressing CTSA (Ito et al., 2021).

Nongenetic Regulators of CTSA

A study by Murakami et al. (2010) reported that telaprevir inhibited CSTA-mediated cleavage of the fluorogenic peptide substrate (7-methoxycoumarin-4-yl)acetyl-Arg-Pro-Pro-Gly-Phe-Ser-Ala-Phe-Lys-(2,4-dinitrophenyl)-OH, with an IC₅₀ value of 0.21 μM. Another study by Birkus et al. (2025) showed that telaprevir and boceprevir inhibited CSTA-mediated tenofovir alfanamide hydrolysis, with IC₅₀ values of 0.27 ± 0.17 μM and 0.16 ± 0.02 μM, respectively, in CTSA-overexpressed HEK293T cells. Notably, as of the present, no inducers of CTSA have been identified.

Conclusions

Hydrolases play an imperative role in metabolizing many clinically important medications and are essential for activating various prodrugs (Table 1). The expressions of hydrolases exhibit significant variations across different organs (Fig. 2), which could lead to drug metabolism in a tissue-specific manner. The differential expression pattern is particularly relevant for prodrug activation. For example, remdesivir is primarily activated in the liver due to the high abundance of hepatic CES1 protein, which has greatly reduced the concentrations of its active metabolite in the target organ—the lung. Therefore, rational prodrug design based on the differential expression profiles of activating hydrolases could enhance prodrug activation in target organs, resulting in improved efficacy and reduced off-target effects.

The regulation of hydrolase expression and activity in humans involves both genetic polymorphisms and nongenetic (Tables 2 and 3). Although many genetic variants have been reported to be associated with the expression and activity of hydrolases, the clinical significance of most hydrolase SNPs has not been fully demonstrated, with the exception of the CES1 variant rs71647871 (G143E). Among nongenetic factors, numerous inhibitors and reducers were shown to potentially affect the expression and catalytic activity of hydrolases (Table 4). Several clinical studies have consistently demonstrated that the CES1 inhibitor alcohol can significantly alter the PK and PD of CES1 substrate drugs. However, other inhibitors and reducers have not been evaluated in clinical studies for their effects on the PK and PD of substrate drugs.

TABLE 2.

Summary of hydrolase pharmacogenetic studies

Genetic Polymorphisms	Representative Studies	Study Design	Major Findings
CES1
G143E (rs71647871)	Patrick et al., 2007; Zhu et al., 2008	PK of methylphenidate, single dose of 0.3 mg/kg in healthy volunteers (n = 20)	The study found one carrier of the G143E variant who exhibited elevated exposure to methylphenidate, with a Cmax of 62 ng/mL for l-methylphenidate, which was 100-fold higher than that of the rest of the participants.
E220G (rs200707504)	Oh et al., 2017; Wang et al., 2017	Single-dose PK study of 75 mg oseltamivir in healthy Korean volunteers (n = 20) In vitro incubation of enalapril, clopidogrel, and sacubitril in cell lysates of transfected cell lines expressing the SNP of interest (E220G, n = 3)	E220G carriers (n = 8) had a 10% increase in the area under the curve between 0 and 48 hours compared with that of the noncarrier group (n=12). The results were not statistically significant (P = 0.334). The CES1 SNP E220G significantly decreased CES1 hydrolytic activity on enalapril, clopidogrel, and sacubitril by approximately 78.8% ± 2.6%, 82.3% ± 3.5%, and 80.3% ± 3.0%, respectively.
S75N (rs2307240)	Johnson et al., 2013; Wang et al., 2017; Xiao et al., 2017	Methylphenidate, weight-based dosing for 6 weeks in children with diagnosed attention deficit hyperactivity disorder (n = 77) In vitro incubation study in transfected cell lines and human liver S9 fractions Retrospective PD analysis on the outcome of clopidogrel therapy in patients with coronary syndrome (N = 851) Outcome: side effects reported by parents through behavioral questionnaires	No statistical significance was found in enalapril activation rate or CES1 protein expression between S75N carriers and noncarriers No statistical significance in side effects of methylphenidate between S75N carriers and noncarriers CES1 S75N carriers (n = 375) had a higher incidence of cerebrovascular events (P < 0.001), acute myocardial infarction (P < 0.001), and unstable angina (P < 0.001) when compared with noncarriers
CES2
A139T	Tang et al., 2006; Xiao et al., 2013; Shen et al., 2022	Exploring potential competitive inhibition between aspirin and clopidogrel using 293T cells transfected with CES2 plasmid construct of A139T or wild type in vitro analysis to evaluate the impact of the antiobesity drug orlistat on the activity of the CES2 polymorphic variant using 293T cells transfected with CES2 genetic variants or wild type in vitro analysis to evaluate the impact of CES2 genetic polymorphisms on molnupiravir hydrolysis using transfected cell lines	Among the HCE2 variants tested, A139T showed a maximal decrease of 40% in aspirin hydrolysis A139T was relatively resistant to CES2 inhibition by orlistat No statistical difference in the hydrolysis of molnupiravir between the A139T variant and the wild type
R34W (rs72547531)	Kubo et al., 2005; Kim et al., 2007	CES2 SNP detection in Japanese cancer patients (n = 165) and in vitro functional analysis of CES2 SNPs using CES2 plasmid–transfected COS-1 cells Retrospective PK analysis of Japanese cancer patients (n = 176)	Significant reduction in hydrolyzing CES2 substrates irinotecan, p-nitrophenyl acetate, and 4-methylumberiferyl acetate in comparison with wild type Observed 60% decrease in area under the curve ratio (SN-38 + SN-38G / CPT-11)
V142M (rs72547532)	Kubo et al., 2005	In vitro analysis of CES2 SNPs in CES2 plasmid–transfected COS-1 cells	Significant reduction in hydrolyzing CES2 substrates irinotecan, p-nitrophenyl acetate, and 4-methylumberiferyl acetate in comparison with wild type
R180H	Xiao et al., 2013; Shen et al., 2022	In vitro analysis of the inhibitory effect of orlistat on the CES2 variant using transfected 293T cells In vitro analysis to evaluate the impact of CES2 genetic polymorphisms on molnupiravir hydrolysis using transfected 293T cells	R180H was found to be relatively sensitive to orlistat inhibition, but at higher concentrations of orlistat (1 nM) The variant R180H showed a 77% increase in hydrolytic activity toward molnupiravir compared with the wild type
AADAC
R248S, X400Q, V172I	Hirosawa et al., 2021	In vitro analysis of the impact of polymorphic variants of AADAC on hydrolyzing eslicarbazepine, p-nitrophenyl acetate, ketoconazole, phenacetin, and rifampicin using human liver and intestinal microsomes (n = 25)	SNPs that showed lower activity than wild type: R248S (5%), X400Q (21%) SNPs that showed higher activity than wild type: V172I (174%)
rs1803155 (c.931G>A, V281I	Weiner et al., 2021	Prospective, placebo-controlled, randomized phase 2 studies conducted during the first 2 months of TB treatment in adults with and without human immunodeficiency virus coinfection in adult participants with tuberculosis (n = 173)	Black participants carrying a single G allele in AADAC rs1803155 were 3 times more likely than wild-type participants to exhibit bactericidal rifapentine exposure below the target level (odds ratio, 2.97; 95% confidence interval, 1.16, 7.58). The odds ratio was greater in black participants with two G alleles.
AADAC*3 allele (g.13651G>A/g.14008T>C; V281I/X400Q	Shimizu et al., 2012	In vitro analysis to evaluate the impact of AADAC variants on hydrolyzing flutamide, phenacetin, and rifampicin using transfected COS-7 kidney cells	AADAC protein produced by the AADAC*3 allele (V281I/X400Q) showed substantially lower intrinsic clearance values compared with wild types
PON1
PON1192QQ and PON1192RR	Ishizuka et al., 2012	In vitro analysis to compare the activity of the PON1 allozymes on hydrolyzing olmesartan medoximil using human embryonic kidney stem cells overexpressing PON1	The catalytic efficacy of PON1192RR on hydrolyzing olmesartan was slight greater than that of PON1192QQ
Q192R and L55M	You et al., 2013	Meta-analysis evaluating the organophosphate toxicity risk associated with PON1 genetic polymorphisms. The analysis included nine case control studies with patients with organophosphate toxicity (n = 1042) and healthy volunteers (n = 1014)	Caucasians with 129Q and 55L polymorphisms had an increased risk of organophosphate toxicity compared with the wild-type population.
PON3
(PON3-567: rs11764079, PON3-665: rs11770903, PON3-746: rs17882539)	Aragonès et al., 2011, 2012; Rull et al., 2012	Measure serum PON3 concentration and evaluate the influence of PON3 promoter polymorphisms in healthy Mediterranean subjects (n = 356) Investigating whether PON3 promoter polymorphisms would impact the course of human immunodeficiency virus disease. The study enrolled patients with human immunodeficiency virus (n = 207) and healthy volunteers (n = 385) Investigating the influence of genotypic variants on serum PON3 concentrations in patients with coronary artery disease (n = 72), peripheral artery disease patients (n = 118), and healthy volunteer (n = 175)	Three PON3 promoter SNPs in linkage disequilibrium were found to be significantly associated with changes in serum PON3 concentrations No significant differences of the PON3 promoter polymorphism between the carriers and control patients No significant differences in PON3 gene promoter polymorphisms and their haplotypes between patients and controls
CTSA
Exon 2 (85_87CTG>-)	Ito et al., 2021	Identified CTSA genetic variants in healthy Japanese volunteers (n = 76) and conducted an in vitro study to evaluate the effect of the identified variants on tenofovir alafenamide hydrolysis	An approximately twofold increase in CTSA expression in 85_87CTG>-transfected cells compared with the wild type (P < 0.05)

TABLE 3.

Pharmaceutical drugs found to be inhibitors or inducers of hydrolases

Enzyme	Inhibitor	Inducer
CES1	Alcohol, protease inhibitors, aripiprazole, isradipine/tacrolimus, valproate, ACEIs	PPAR agonists, rifampicin, glucose, phenobarbital
CES2	Simvastatin, fenofibrate, diltiazem, verapamil, orlistat, remdesivir, sofosbuvir, loperamide	Rifampicin, NO1886 (ibrolipim), urethane dimethacrylate (UDMA)
AADAC	Epicatechin, ECg, EGCg	Fenofibric acid, WY-14643
PON1	Statins (simvastatin, atorvastatin),a fibrates (gemfibrozil, fenofibrate),a aspirina	Statins (simvastatin, atorvastatin),a fibrates (gemfibrozil, fenofibrate),a probucol, ezetimibe, aspirin,a rosiglitazone, eplerenone, sulfonylureas (glimepiride, glibeclamide,b erythropoietin β
PON3	No literature found	No literature found
CSTA	Telepravir, boceprevir	No literature found

^aConflicting results have been reported by different studies.

^bIncrease in the liver, decrease in plasma.

TABLE 4.

Enzyme kinetic characteristics of drug-metabolizing hydrolases

Substrate	Hydrolase	Km	Vmax	Reference
Trandolapril	CES1 (recombinant)	1734 µM	624 nmol/min per mg protein	Thomsen et al., 2014
Ramipril	CES1(recombinant)	901 µM	956 nmol/min per mg protein
Enalapril	CES1(recombinant)	1721 µM	34 nmol/min per mg protein
Oseltamivir	CES1 (recombinant)	2.9 ± 0.8 mM	7.2 ± 1.0 µmol/min per mg	Xu et al., 2015
Sacubitril	CES1 (recombinant)	767.2 ± 56.4 µM	557.5 ± 18.1 nmol/mg per min	Shi et al., 2016a
Methylphenidate	CES1 (purified native)	89.9 ± 6.6 µM	3.2 ± 0.1 nmol/min per mg	Sun et al., 2004
Meperidine	CES1 (purified native)	1.9 ± 0.3 mM	0.0111 ± 0.0007 µmol/min per mg	Zhang et al., 1999
Capecitabine	CES1 (recombinant)	19.2 mM	88.3 nmol/min per mg	Tabata et al., 2004
Irinotecan	CES1 (liver cytosol)	3.4 ± 3 µM	2.7 pmol/mg per min	Hennebelle et al., 2000
Clopidogrel	CES1 (transfected cell S9)	62.7 ± 15.4 µM	3558 ± 371 pmol/min per mg	Zhu et al., 2013
Dabigatran	CES1 (recombinant)	24.9 ± 2.9 µM	676 ± 26 pmol/min per mg	Laizure et al., 2014
Mycophenolate mofetil	CES1 (HLM)	992 ± 138 µM	3073 ± 224 nmol/min per mg protein	Fujiyama et al., 2010
Fenofibrate	CES1(recombinant)	9.27 ± 0.92 µM	2165 ± 94.16 pmol/min per µg protein	Li et al., 2023
Oxybutynin	CES1(recombinant)	17 µM	310 pmol/min per mg protein	Sato et al., 2012
Aspirin	CES2 (recombinant)	0.36 mM	NA	Tang et al., 2006
Heroin	CES (recombinant)	1440 ± 630 µM	8130 ± 910 pmol/min per mg	Hatfield et al., 2010
Flutamide	CES2 (recombinant)	0.3 ± 0.1 mM	1113.1 ± 206.4 pmol/min per mg	Kobayashi et al., 2012
Gemcitabine	CES2 (recombinant)	43 µM	40 nmol/min per mg protein	Pratt et al., 2013
Capecitabine	CES2 (recombinant)	1.0 ± 0.11 mM	0.22 ± 0.01 µg/min per mg	Quinney et al., 2005
Eslicarbazepine acetate	AADAC (recombinant)	0.79 ± 0.11 mM	157.5 ± 3.9 nmol/min per mg	Hirosawa et al., 2021
Ketoconazole	AADAC (recombinant)	6.93 ± 0.105 µM	35.3 ± 0.865 pmol/min per mg protein	Fukami et al., 2016
Flutamide	AADAC (recombinant)	778 ± 122 µM	0.6 ± 0.1 nmol/min per mg	Watanabe et al., 2009
Olmesartan medoxomil	PON1 (192QQ, 192RR) recombinant	157 µM, 102 µM	124, 140 nmol/min per mg protein	Ishizuka et al., 2012
Pilocarpine	PON1 (192QQ, 192RR) recombinant	14.1 ± 3.1, 8.3 ± 1.9 mM	52.1 ± 3.6, 45.0 ± 2.8 nmol/min per mg protein	Hioki et al., 2011
Simvastatin	PON1 (recombinant)	103.01 µM	0.9 nmol/min per mg protein	Li et al., 2019
Simvastatin	PON3 (recombinant)	337.62 µM	1.30 nmol/min per mg protein	Li et al., 2019
Tenofovir Alafenamide	CTSA (recombinant)	610 ± 190 µM	NA	Birkus et al., 2007
Remdesivir	CTSA (recombinant)	166.20 ± 11.17 µM	14.52 ± 0.34 nmol/min per mg protein	Zhang et al., 2022

NA, not available.

In summary, despite considerable efforts dedicated to investigating the regulation of hydrolases, the clinical significance of genetic polymorphisms and nongenetic regulators of most hydrolases is yet to be firmly established due to the fact that the majority of the studies were conducted either in vitro or in nonhuman animals. More in-depth research in the field, particularly in the realm of clinical studies, is urgently needed to advance our understanding of the regulation of hydrolases and to improve the clinical outcomes of numerous medications metabolized by these enzymes.

Abbreviations

AADAC

arylacetamide deacetylase

CES

carboxylesterase

CRP

C-reactive protein

CTSA

cathepsin A

DDI

drug-drug interaction

DME

drug-metabolizing enzyme

FDA

Food and Drug Administration

GGT

γ-glutamyl transferase

GI

gastrointestinal

HDL

high-density lipoprotein

K_i

inhibition constant

MAF

minor allele frequency

P450

cytochrome P450

PD

pharmacodynamic

PK

pharmacokinetic

PON

paraoxonase

PPAR

peroxisome proliferator-activated receptor

SNP

single nucleotide polymorphism

Authorship Contributions

Wrote or contributed to the writing of the manuscript: Jung, Zhu.