Regulation of carboxylesterases and its impact on pharmacokinetics and pharmacodynamics: an up-to-date review

Abstract

Introduction:

Carboxylesterase 1 (CES1) and carboxylesterase 2 (CES2) are among the most abundant hydrolases in humans, catalyzing the metabolism of numerous clinically important medications, such as methylphenidate and clopidogrel. The large interindividual variability in the expression and activity of CES1 and CES2 affects the pharmacokinetics (PK) and pharmacodynamics (PD) of substrate drugs.

Areas covered:

This review provides an up-to-date overview of CES expression and activity regulations and examines their impact on the PK and PD of CES substrate drugs. The literature search was conducted on PubMed from inception to January 2024.

Expert opinion:

Current research revealed modest associations of CES genetic polymorphisms with drug exposure and response. Beyond genomic polymorphisms, transcriptional and posttranslational regulations can also significantly affect CES expression and activity and consequently alter PK and PD. Recent advances in plasma biomarkers of drug-metabolizing enzymes encourage the research of plasma protein and metabolite biomarkers for CES1 and CES2, which could lead to the establishment of precision pharmacotherapy regimens for drugs metabolized by CESs. Moreover, our understanding of tissue-specific expression and substrate selectivity of CES1 and CES2 has shed light on improving the design of CES1- and CES2-activated prodrugs.

1. Introduction

Carboxylesterases (CESs) are essential members of the serine hydrolase superfamily, responsible for the biological transformation of a wide range of structurally diverse endogenous (e.g., lipids and triacylglycerols) and xenobiotic substances (e.g., drugs and pesticides). CESs catalyze the cleavage of ester, thioester, and amide bonds, resulting in the corresponding alcohol/thiol/amine products and carboxylic acid metabolites. In the presence of ethanol, CES1 can also mediate transesterification reactions [1]. In humans, there are primarily six CES isozymes based on the homology of amino acid sequences, of which CES1 and CES2 are the major isozymes. CES1 and CES2 play crucial roles in metabolizing clinically important medications, such as antihypertensive, anticoagulant, antiviral, and anticancer drugs [2].

The primary differences between CES1 and CES2 are their tissue expression patterns and substrate selectivity. CES1 is predominantly expressed in the liver, with much less expression in the lung, kidney, and intestine [3]. Of note, in the human liver, the CES1 expression level is significantly higher than other drug-metabolizing enzymes, such as cytochrome P450s (CYPs) and uridine diphosphate glucuronosyltransferases (UGTs), and other hydrolases [4]. In comparison, CES2 is the most abundant hydrolase in the intestine. Figure 1 shows the protein expression levels of CES1 and CES2 in the human liver, intestine, lung, and kidney [5]. The mRNA expression levels of the CES1 and CES2 genes were reported in the GTEx database, as shown in Figure S1. As for substrate specificity, CES1 substrates usually contain a large acyl group and a small alcohol group, such as clopidogrel, enalapril, capecitabine (CP), and methylphenidate (MPH). CES2 prefers substrates with a small acyl group and a large alcohol part, such as aspirin and irinotecan.

Figure 1.

CES1 and CES2 protein expressions in human liver, intestine, lung, and kidney. Data are adapted from a previous publication [5].

Large variability in CES1 and CES2 expression and activity has been observed among individuals, contributing to the marked variability in the pharmacokinetics (PK) and pharmacodynamics (PD) of CES substrate drugs [2]. Therefore, it is critical to identify factors regulating CES expression and activity and investigate the effects of these regulators on drug exposure and response. This review comprehensively summarized the genetic and environmental regulating factors of CES1 and CES2 and discussed their influence on drug metabolism, exposure, and clinical outcomes. This review highlighted that a better understanding of CES regulation could facilitate personalized treatment with CES substrate drugs.

2. Method

An extensive literature search was conducted in the PubMed database from inception to January 2024. The search terms were categorized into three main topics: (1) carboxylesterase, (2) drug metabolism and pharmacokinetics (PK), and (3) pharmacodynamics (PD). These search terms were used as title/abstract and Mesh terms, if available, combined with the logical words “and” and “or”. Each retrieved literature was manually screened for inclusion by reviewing its title and abstract. Publications written in non-English languages and those inaccessible for full-text review were excluded.

3. Regulation of CES1 and CES2 expression and activity

CES1 and CES2 play a crucial role in hydrolyzing numerous clinically important drugs, affecting their PK and PD. Understanding the regulation of CES expression and activity is vital for optimizing pharmacotherapy. This regulation involves a complex interplay of genetic variations, epigenetic modifications, and transcriptional, post-transcriptional, translational, and posttranslational mechanisms [6]. This section discusses the regulatory events of CES and the impact of genetic, physiological factors, and CES inhibitors/inducers on CES functionality.

3.1. Genetic Polymorphisms

3.1.1. CES1 Gene Structure Variation

Human CES1 is encoded by the CES1 gene, which has a tail-to-tail configuration of CES1 (also termed CES1A1) and the pseudogene CES1P1 (also termed CES1A3) located on the long arm of chromosome 16: 55802851–55833096 (Figure 2) [7]. The sequence of CES1P1 is different from that of CES1A1 in the 5’ untranslated region (5’-UTR), exon1, and intron 1 regions, and the presence of a premature stop codon within exon 3. Consequently, CES1P1 only generates low levels of truncated mRNA without expressing functional CES1 protein. CES1P1VAR (CES1A2) is a functional CES1 isoform resulting from the double crossover between CES1A1 and CES1P1 and accounts for 2% of CES1 hepatic expression [8,9]. CES1P1VAR has two haplotypes with frequencies of 22% and 74%, respectively. The CES1P1VAR minor haplotype has two binding sites for the Sp1 transcription factor (TF) and exhibits a higher transcription activity than the major haplotype without a Sp1 binding site [10]. However, Rasmussen et al. reported that the minor allele frequency (MAF) of CES1P1VAR is less than 0.022 in Asians, and the pharmacogenetic relevance of CES1A2 is limited in this population [11]. Moreover, genomic translocation between CES1A1 and CES1P1 produces two CES1 variants, i.e., CES1SVAR (identical or closely related to CES1A1b, with a frequency < 0.01%) and CES1A1c (also termed CES1VAR, with a frequency about 17%). CES1VAR reduces hepatic CES1 mRNA expression by ~30% with no apparent alteration in hepatic protein expression and activity of CES1 [11]. Collectively, these genetic variants form 4 major CES1 haplotypes and generate various diplotypes with 2, 3, and 4 functional CES1 copies (Figure 2) [7,12].

Figure 2.

Gene structures and diplotypes of CES1, adapted from Ref. [7,12]. CES1P1 is a pseudogene with no expression of functional CES1 protein, while all others are functional genes. Reprinted with permission from [8], copyright © 2016 Wolters Kluwer Health, Inc.

Genetic structural variation may alter CES1 expression and functionality. An in vitro study found no impact of CES1/CES1VAR, CES1P1/CES1P1VAR, and other diplotypes on CES1 protein expression and hydrolysis activity in human livers [12]. Similarly, serum CES1 level was not different among the healthy Chinese Han population with different copy numbers [13]. However, CES1 activity on hydrolyzing irinotecan in Japanese cancer patients with three or four CES1 copies was 1.24-fold of those with two copies (P = 0.0134) [14]. These findings suggest that the effect of copy number variation on CES1 activity could be substrate-specific and requires further investigations for other CES1 substrates.

3.1.2. CES1 Single Nucleotide Polymorphisms (SNPs)

In addition to structural variations, SNPs in the CES1 gene can regulate CES1 expression and activity. Zhu et al. discovered the first and most well-studied loss-of-function CES1 SNP, rs71647871, which results from a nonsynonymous variant in exon 4, leading to amino acid change at the coding position of 143 from glycine to glutamic acid (also termed G143E). This mutation substantially impairs CES1 catalytic functions, likely by disrupting the oxyanion hole essential for the enzyme’s active domain stabilization [15]. Given the relatively high minor allele frequency (MAF) of G143E in white (3.7%), black (4.3%), and Hispanic (2.0%) populations [15], G143E has clinical relevance in the PK and PD of CES1 substrate medications [2,3]. Notably, G143E is rare in the Asian population, and an MAF of 1% is reported in the gnomAD database (gnomad.broadinstitute.org). Additionally, the study identified a deletion in CES1 exon 6 at codon 260, termed Asp260fs, that results in a frameshift and a truncated nonfunctional protein. However, the MAF of Asp260fs is extremely low, thereby lacking clinical relevance [13]. In 2017, Wang et al. comprehensively evaluated the impact of 20 CES1 SNPs on CES1 activity and expression in CES1-transfected cell lines and human livers [16]. These CES1 SNPs were selected as they either have an MAF greater than 0.5% in any ethnic population or are located within 5 Å from the CES1 active site. L40Ter (rs151291296), G142E (rs121912777), G147C (rs146456965), Y170D (rs148947808), and R171C (rs201065375) independently caused a complete loss of CES1 activity against clopidogrel, enalapril, and sacubitril, whereas A158V (rs202121317), R199H (rs2307243), E220G (rs200707504), and T290M (rs202001817) decreased CES1 activity to a lesser extent in a substrate-dependent manner. Among these SNPs, only G147C resulted in simultaneous mRNA and protein expression reduction. L40Ter, Y170D, and R171C-transfected cell lines exhibited normal mRNA levels but a loss of CES1 activity and deficient protein expression. Like the G143E SNPs, G142E, E220G, A158V, R199H, and T290M diminished CES1 activity without altering the mRNA or protein expression, indicating that these variants might directly affect the enzyme-substrate interaction [16]. Notably, many of the above SNPs capable of reducing CES1 functionality have relatively high allele frequencies in diverse ethnic groups, such as A158V (6.69% in South Asians), T290M (2.86% in South Asians, 1.87% in Latinos,1.21% in Africans), underscoring their potential clinical significance [14].

CES1–816A>C (rs3785161) is a common SNP located in the promoter region of CES1P1. Mixed results were reported for the associations of -816A>C with the clinical response of CES1 substrate drugs, indicating increased or decreased CES1 functionality [17–19]. However, Zhu and colleagues reported no impact of 816A>C on hepatic CES1 expression and activity in 100 normal human liver samples [20]. More recently, Her et al. identified two common (MAF 0.22) cis-acting regulatory CES1 variants in human liver samples. CES1 rs6499788 A allele and rs35918553 G allele were significantly associated with a ~15% decrease in CES1 protein expression and ~27% reduction in CES1 catalytic activity on enalapril in human liver microsomes (HLM) [21]. Besides direct in vitro functional assessments, clinical studies implied that several SNPs (e.g., S75N (rs2307240)) could potentially modulate CES1 function, and their effects on PK and PD appeared to be substrate-dependent [14,22–24]. Further molecular genetics studies are required to clarify the impact of these SNPs on CES1 functionality against different substrate drugs.

3.1.3. CES2 Genetic Polymorphisms

The human CES2 gene, coding for the CES2 enzyme, is located on chromosome 16q13–q22.1 downstream of the CES1 gene. Wu et al. identified three transcriptionally active promoters (P1, P2, P3) in CES2, with P3 contributing to low basal expression in various tissues, whereas P1 and P2 are implicated in tissue-specific transcription regulation [25].

Several studies have focused on the impact of SNPs and haplotypes on CES2 expression and functionality. Compared with the CES1 gene, the CES2 gene has fewer genetic variants, predominantly located in introns [26]. Wu et al. identified 15 SNPs and 12 possible haplotypes of CES2 in 78 healthy subjects. 5’-UTR-363 was found in linkage disequilibrium (LD) with intron1 + 947, +1361, and +1643. The haplotype with the rare homozygous alleles (GTGA/GTGA) was associated with lower mRNA expression (P < 0.01), and the haplotype of the heterozygous rare alleles seemingly correlated with reduced mRNA levels (P = 0.03). However, none of these haplotypes impacted the CES2 activity of metabolizing irinotecan and procaine [26]. Marsh et al. discovered 11 CES2 SNPs and six haplotypes in Europeans, Africans, and Americans (n=120). The CES2 SNPs showed no correlation with mRNA expression in normal colonic mucosa; however, colon cancer patients with intronic IVS10–88 (rs3893757) TT genotype had significantly reduced mRNA expression [26]. Furthermore, two nonsynonymous SNPs, R34W and V142M, increased CES2 protein expression but abolished CES2 activity against irinotecan, p-nitrophenol acetate (PNPA), and 4-methylumbelliferyl acetate. IVS8–2A>G at the splice acceptor site in intron 8 results in aberrantly spliced transcripts, thus truncated proteins [27]. Based on 21 SNPs, including R34W, V142M, and IVS8–2A>G, Kim et al. identified 20 haplotypes in 262 Japanese subjects. The majority (over 94%) belong to the *1 group with normal function and without an impact on CES2 in vivo activity. In contrast, the SNP Met1Leu (A>T) substantially reduced CES2 in vivo activity and in vitro protein expression without altering mRNA expression, likely due to the impaired translation efficiency caused by alterations in the translational start codon [28]. The SNP 830C>G, near the translation start codon, was found to have no significant impact on CES2 functionality [29].

3.2. Epigenetic regulation

Epigenetics broadly refers to stable alterations in gene activity and expression not caused by DNA sequence changes [30]. Typical epigenetic processes such as DNA methylation, histone modification, and various RNA-related regulations (e.g., microRNA) can regulate gene expression transcriptionally and translationally [6,30]. DNA methylation primarily occurs on cytosine in the CpG islands in the promoter and exon regions of a gene, leading to decreased gene expression. The silenced CES1A1 gene expression in human embryonic kidney cells (HEK293) and the low renal CES1 expression were closely related to the highly methylated region near the transcription start site. Treating HEK293 cells with DNA methylation inhibitor dramatically increased CES1A1 expression. In contrast, the genomes from both the human hepatoma cells (HepG2) and the human liver exhibited extremely low levels of DNA methylation in comparable regions [31]. Furthermore, histone acetylation often positively regulates gene transcription [32]. In mice, hepatic CES1 expression increased with glucose treatment and in animals with diabetes but decreased during fasting, likely due to glucose-induced acetylation of histones 3 and 4, which enhances the transcription activity [32]. To our knowledge, no study has explored epigenetic regulations of CES2 expression.

3.3. Transcriptional regulation: alternative splicing, CES inducers, and disease mediators

In addition to genetic polymorphisms and epigenetic regulators, alternative splicing (AS) plays a role in CES expression regulation. Two AS sites within the CES1 gene yield four associated transcripts or isoforms. Wang et al. reported that isoforms 1 and 2 were the major transcripts in human liver tissues, accounting for 73%-90% of CES1 protein expression. Moreover, the ratios of proteins expressed from isoforms 1 and 2 to 3 and 4 exhibited a positive association with hepatic CES1 activity [33].

Transcription factors (TFs) are pivotal in regulating CES gene expression [34,35]. In general, TFs are complex with proteins in the cytoplasm. Upon activation, TFs translocate to the nucleus to bind regulatory elements in target genes, modulating their expression [34]. Notably, various TFs, including glucocorticoid receptor (GR), pregnane X receptor (PXR), constitutive androstane receptor (CAR), nuclear factor erythroid 2-related factor 2 (Nrf2), and nuclear receptor subfamily 1 group H member 3 (NR1H3), have been identified as important players in regulating CES gene expression [34,36,37]. This process can be influenced by several medications, inducers, and disease conditions, exhibiting concentration-, isoform-, and species-dependent regulatory manners [38,39].

Dexamethasone, a widely used immunosuppressive drug, notably decreased CES1 expression at nanomolar concentrations in rats, potentially due to dramatically reduced GR expression. However, only micromolar dexamethasone increased CES2 mRNA expression and irinotecan hydrochloride hydrolase activity, likely via enhancing PXR expression [38,39]. In human primary hepatocytes, dexamethasone slightly induced CES1 and CES2 expressions with drug concentrations above ten micromolar levels, suggesting concentration-dependent effects, possibly related to its varying affinity for GR and PXR. Phenobarbital and rifampicin, widely used for epilepsy and tuberculosis, respectively, modestly induced CES1 and CES2 expression in human primary hepatocytes [38]. The CES-inducing effect of phenobarbital was attributed to CAR and PXR activation [40]. The anticancer drug 5-fluorouracil (5-FU) activated p-53 in HepG2 cells, elevating CES2 mRNA expression through transactivation and increasing mRNA stability [41,42].

Fluoxetine, an antidepressant, repressed CES1 and CES2 expression and activity in HepG2 cells by decreasing PXR and increasing differentiated embryonic chondrocyte-expressed gene 1 (DEC1) expression. This repression occurs as increased DEC1 inhibits the PXR-RXR heterodimer, a crucial component for initiating target gene transcription [43]. Conversely, cis-diamminedichloroplatinum increased CES1 and CES2 expression and function by activating PXR and downregulating DEC1 in HepG2, mouse hepatocytes, liver, and intestinal tissues [44]. Insulin markedly repressed CES expression and hydrolytic activity in HepG2 cells, but PXR overexpression reversed this effect by directly binding to the promoters of CES1 and CES2, enhancing their transcriptional activity [45]. Nrf2 activators, including tert-butylhydroquinone and sulforaphane, substantially increased CES1A1 transcription, protein expression, and activity of hydrolyzing imidapril in HepG2 cells [46].

Disease mediators also influence CES expression by TF-mediated mechanisms. For instance, acetaminophen-induced acute liver injury reduced CES2 protein levels in mice [47]. Similarly, lipopolysaccharide (LPS)-induced immunological hepatic damage and DSS-induced ulcerative colitis suppressed the expression and functionality of CES1 and CES2 in rodents [48,49]. The activation of substrate prodrugs, such as imidapril and irinotecan, was significantly decreased by LPS-induced liver injury [48]. A mechanistic study revealed that IL-6 was significantly associated with the downregulation of CES1, CES2, PXR, CAR, and NF-κB, and the trend was reversed by NF-κB inhibitors [49]. Recently, the IL-33 level in patients’ livers was inversely associated with CES1 expression, likely due to Nrf2 inhibition [50]. Notably, most studies on TF-mediated CES gene regulation were performed on rodents and human hepatoma cells. Given potential species differences and variations in gene expression between cultured tumor cells, normal cells, and in vivo conditions, the regulation direction, magnitude, and mechanisms of CES expression by TFs in humans might differ from the results obtained from in vitro and animal studies.

3.4. Age and sex

CES expression and functionality are age-dependent [51–55], potentially affecting PK and PD in pediatric patients. Significantly lower hepatic CES1 expression and activity were found in neonates (< 21–28 days), infants (< 1 year), children during early and middle childhood (1 – 12 years), and adolescence (12–18 years) compared to that of adults (>18 years) independent of sex and race [52,55]. Hepatic CES2 expression and activity in neonates were significantly lower relative to adults. Hepatic CES2 expression in children from infancy to adolescence was also lower than in adults; however, the difference was not statistically significant [55]. Additionally, CES1 expression and activity were approximately 20% and 30% higher, respectively, in female livers than in males [56]. Age-related CES expression was found to be independent of the growth hormone levels [52]. The mechanisms underlying the age- and sex-related CES expression remain unclear.

3.5. CES inhibitors

CES inhibitors generally bind with the CES regulatory domain to repress its activity without impacting CES expression. There are several review articles focusing on different aspects of CES inhibitors. In 2018, Zou et al. summarized the capacity and mechanisms of a wide range of synthetic, semi-synthetic, and natural CES inhibitors. They discussed the structural-activity relationships for developing potent and selective CES inhibitors [57]. In 2021, Song et al. reviewed three categories of clinically relevant CES inhibitors, including therapeutic agents, pharmaceutical excipients, and herbal constituents, and discussed their impact on CES-mediated drug metabolism [58].

Our discussion primarily centers on CES inhibitors that may interact with CES substrate drugs under clinically relevant scenarios. Many therapeutic agents across diverse chemical structures and indications have been found to inhibit the enzymatic activity of recombinant human CES1 and CES2 (rCES1 and rCES2), HLM and human liver S9 fractions (HLS9), and human jejunum microsomes (HJM). Tables 1 and 2 list the typical therapeutic agents that act as inhibitors for CES1 and CES2, along with details on their clinical use, tested substrates, and inhibitory properties. Notably, many of these inhibitors are commonly prescribed medicines for chronic metabolic diseases, such as hypertension, hyperlipidemia, obesity, and diabetes. They could be co-prescribed with CES substrate drugs, affecting their PK and PD [59,60]. For instance, among 11 antidiabetic and 12 antihyperlipidemic drugs, lactone ring-containing statins (e.g., simvastatin and lovastatin) and thiazolidinediones (e.g., troglitazone and rosiglitazone) potently inhibited the imidapril hydrolytic activity of rCES1. Fenofibrate and simvastatin significantly inhibited rCES2 activity [59]. Within 17 tested antihypertensive drugs, telmisartan and nitrendipine potently inhibited rCES1- and HLM-mediated hydrolysis of imidapril, whereas diltiazem and verapamil significantly impaired CES2 activity against irinotecan [60]. In contrast to the negligible rCES2 inhibition of telmisartan reported by Xu et al. [60], Shimizu et al. suggested that telmisartan was a potent and selective inhibitor of CES2 [61]. Further studies are required to clarify the discrepancy. Additionally, Zhu et al. screened several psychotropic drugs (i.e., aripiprazole, perphenazine, thioridazine, and fluoxetine) for their CES1 inhibitory effects. Aripiprazole was found as a potent CES1 inhibitor, significantly altering the PK of MPH in mice [62].

Table 1.

Therapeutic drug and drug excipients as CES1 inhibitors.

CES1 inhibitor	Category	Substrate	Inhibitory properties	Ref.
Procainamide	Antiarrhythmic drug	Imidapril	Ki: 29.3 ± 4.8 μM (HLM) Ki: 34.5 ± 2.2 μM (HLC)	[66]
Simvastatin	Antihyperlipidemic drug	Imidapril PNPA	Ki: 0.11 ± 0.01 μM (rCES1) Ki: 0.76 ± 0.06 μM (HLM) IC50: 0.76 μM (rCES1) *	[59]
Troglitazone	Antidiabetic drug	Imidapril PNPA	Ki: 0.62 ± 0.08 μM (rCES1) Ki: 5.64 ± 0.23 μM (HLM) IC50: 3.30 μM (rCES1) *	[59]
Aripiprazole Perphenazine Thioridazine Fluoxetine	Psychotropic drug	d-MPH	IC50: 61.7 μM (cell s9) IC50: 65.0 μM (cell s9) IC50: 58.3 μM (cell s9) IC50: 58.9 μM (cell s9)	[62]
Telmisartan	Antihypertensive drug	Imidapril PNPA	Ki: 0.49 ± 0.09 μM (rCES1) Ki: 1.69 ± 0.17 μM (HLM) IC50: 0.89 μM (rCES1) *	[60]
Nitrendipine	Antihypertensive drug	Imidapril PNPA	Ki: 1.12 ± 0.39 μM (rCES1) Ki: 1.24 ± 0.27 μM (HLM) IC50: 3.7 μM (rCES1) *	[60]
Nelfinavir	Antiretroviral drug	PNPA dl-MPH	Ki: 3.7 ± 0.7 μM (cell s9) * IC50: 6.6 ± 4.0 μM (cell s9)	[67]
Sodium lauryl sulphate	Pharmaceutical excipients	Imidapril PNPA	Ki: 0.04 ± 0.01 μg/mL (rCES1) Ki: 0.12 ± 0.03 μg/mL (HLM) IC50: 0.2 μg/mL (rCES1) *	[63]
Polyoxyl 40 hydrogenated castor oil	Pharmaceutical excipients	Imidapril PNPA	Ki: 0.20 ± 0.09 μg/mL (rCES1) Ki: 0.76 ± 0.33 μg/mL (HLM) IC50: 1.48 μg/mL (rCES1) *	[63]

HLC, human liver cytosol; rCES1/2, recombinant human CES1/2; HLM, human liver microsomes; HJM, human jejunum microsomes. Cell s9 refers to the s9 fraction extracted from CES1-over expressing cells.

^*indicates that the inhibitory features are based on PNPA hydrolysis.

Table 2.

Therapeutic drug and drug excipients as CES2 inhibitors.

CES2 inhibitor	Category	Substrate	Inhibitory properties	Ref.
Carvedilol	Antihypertensive drug	Irinotecan	Ki: 1.6 ± 0.2 μM (HLM) Ki: 4.1± 0.3 μM (HLC)	[66]
Simvastatin	Antihyperlipidemic drug	Irinotecan PNPA	Ki: 0.67 ± 0.09 μM (rCES2) Ki: 1.85 ± 0.28 μM (HLM) Ki: 3.67 ± 0.49 μM (HJM) IC50: 0.78 μM (rCES2) *	[59]
Fenofibrate	Antihyperlipidemic drug	Irinotecan PNPA	Ki: 0.04 ± 0.01 μM (rCES2) Ki: 87.7 ± 12.0 μM (HLM) Ki: 0.50 ± 0.06 μM (HJM) IC50: 0.22 μM (rCES2) *	[59]
Diltiazem	Antihypertensive Antiarrhythmic	Irinotecan PNPA	Ki: 0.25 ± 0.02 μM (rCES2) Ki: 2.89 ± 0.39 μM (HLM) Ki: 4.67 ± 2.12 μM (HJM) IC50: 3.89 μM (rCES2) *	[60]
Verapamil	Antihypertensive	Irinotecan PNPA	Ki: 3.84 ± 0.99 μM (rCES2) Ki: 11.54 ± 1.20 μM (HLM) Ki: 15.75 ± 2.63 μM (HJM) IC50: 7.94 μM (rCES2) *	[60]
Vinblastine	Anti-cancer	PNPA Irinotecan	IC50: 9.4 μM (rCES2) * IC50: 2.0 ± 0.4 μM (rCES2) IC50: 37.4 ± 16.1 μM (HLM)	[61]
Loperamide	Anti-diarrhea	PNPA	IC50: 0.1 μM (rCES2) *	[61]
Telmisartan	Antihypertensive Antiarrhythmic	PNPA Irinotecan	IC50: 0.5 μM (rCES2) * IC50: 0.4 ± 0.1 μM (rCES2) IC50: 0.5 ± 0.1 μM (HLM)	[61]
Eserine	Anticholinesterase	PNPA	Ki: 0.222 ± 0.045 μM (rCES2) * Ki: 0.358 ± 0.093 μM (HLS9) *	[68]
Reserpine	Antihypertensive	fluorescent diacetate	IC50: 0.94 ± 0.12 μM (HLM) Ki: 1.53 μM (HLM)	[69]
Clotrimazole	Antifungal	Irinotecan	Ki: 29.0 μM (HLM)	[70]
Orlistat	anti-obesity drug	Irinotecan Acebutolol	IC50: 3.31± 0.95 nM(rCES2) IC50: 2.64 0.45 nM (cell homogenates)	[71]
Tween 20	Pharmaceutical excipient	Irinotecan PNPA	Ki: 0.93 ± 0.36 μg/mL (rCES1) Ki: 2.2 ± 0.55 μg/mL (HLM) Ki: 1.2 ± 0.33 μg/mL (HJM) IC50: 3.16 μg/mL (rCES1) *	[63]
Polyoxyl 35 castor oil	Pharmaceutical excipient	Irinotecan PNPA	Ki: 4.4 ± 1.24 μg/mL (rCES1) Ki: 13.2 ± 2.0 μg/mL (HLM) Ki: 20.54 ± 3.82 μg/mL (HJM) IC50: 8.22 μg/mL (rCES1) *	[63]

HLC, human liver cytosol; rCES1/2, recombinant human CES1/2; HLM, human liver microsomes; HJM, human jejunum microsomes. HLS9, human liver S9 fractions.

^*indicates that the inhibitory features are based on PNPA hydrolysis.

Pharmaceutical excipients are also sources of CES inhibitors and might be clinically relevant. Among 25 evaluated excipients, sodium lauryl sulphate and polyoxyl 40 hydrogenated castor oil substantially inhibited the hydrolysis of imidapril by rCES1 and HLM. Tween 20 and polyoxyl 35 castor oil inhibit rCES2 activity toward irinotecan, with weaker effects on HLM and HJM [63]. Furthermore, a wide range of natural compounds, particularly herbal constituents like cannabinoids, flavonoids, tanshinones, lignans, triterpenoids, phenolic chemicals, and ginsenosides, strongly inhibited CES1 and CES2 activity [64]. A detailed discussion of natural CES inhibitors regarding their inhibitory properties and the impact on the PK and PD of CES substrate drugs can be found in other reviews [58,64,65].

4. Implications of CESs in drug PK and PD

This section discusses the impact of CESs on the PK and PD of selected clinically important medications. Table 3 lists the important regulators of CESs studied in humans and their implications for the exposure and clinical outcomes of these drugs.

Table 3.

Genetic polymorphisms of CES1 and CES2 and their impacts on the PK and PD of CES substrate drugs.

Substrate Drug	CES variants	Populations	PK associations	PD associations	Ref.
Clopidogrel	CES1 rs71647871 (G143E) MAF:0.006(E)	Healthy white subjects (n = 566); Patients with coronary heart disease (n=350)	G143E carriers had significantly greater levels of clopidogrel active metabolite (GE: 30.3±6.1 vs. GG:19.0 ± 0.4 ng/mL, P=0.01).	G143E carriers had better clopidogrel responses, as shown by larger decrease in platelet aggregation in normal subjects (GE: 29% vs. GG: 43%, P=0.03) and patients with coronary heart disease (GE: 25% vs. GG: 45%, P=0.03).	[73]
Clopidogrel	CES1 rs71647871 (G143E)	Healthy white volunteers: Carriers (n=10); Non-carriers (n=12)	AUC0–∞ of clopidogrel and its active metabolite was 123% (P=0.004) and 67% (P=0.009) larger in G143E carriers.	Inhibition of platelet aggregation was 19% higher in G143E carriers (P=0.036).	[74]
Clopidogrel	CES1 Intronic rs12443580 MAF:0.35 (C); intronic rs8192935 MAF: 0.31 (A); rs71647871 (G143E)	Healthy volunteer: PK study, n=106 PD study, n=46	rs12443580 and rs8192935 did not impact clopidogrel PK; AUC0–∞ of clopidogrel and its active metabolite was 163% (P=0.00012) and 59.6% (P=0.0054) larger in G143E carriers.	rs12443580 and rs8192935 did not affect clopidogrel PD; G143E enhanced the antiplatelet effects of clopidogrel by 21.4% (P =0.00090).	[75]
Clopidogrel	CES1 rs3785161 (−816A/C) MAF:0.285 (C)	Han Chinese with coronary artery disease and CYP2C19 wild type (n=249)	N/A	Platelet aggregation was significantly different among AA: 31.7%, AC:23.6%; CC:14.6% (P = 0.002), C allele was associated with greater antiplatelet effects; Only decreasing trend in the incidence of 1-year definite stent thrombosis was found, (AA: 1.83%, AC: 0.4%, CC: 0.0%).	[18]
Clopidogrel	CES1 rs3785161 (−816A/C)	162 patients with coronary heart diseases	N/A	Platelet activity was significantly higher in CES1 −816C carriers (AA: 45.93 ±18.17% vs. 53.18 ±18.28%, P = 0.014). CES1 genotype explains 3.7% of platelet activity variability. −816C allele was not significantly associated with the risks of stent thrombosis.	[17]
Clopidogrel	CES1 rs2307240 (S75N) MAF:0.22 (T)	851 Chinese patients with acute coronary syndrome	N/A	CC carriers have significantly higher occurrence ratios than in the CT + TT group for cerebrovascular events (14% vs. 4.8%, OR = 0.31, P < 0.001), acute myocardial infarction (15.1% vs 6.1%, OR = 0.37, P < 0.001) and unstable angina (62.8% vs 37.7%, OR = 0.36, P < 0.001).	[76]
Clopidogrel	CES1 intronic rs2244613 MAF:0.284(C)	81 patients with acute coronary syndrome 136 healthy subjects	N/A	rs2244613 carriers have higher residual platelet reactivity than noncarriers. AC+CC:183.23 PRU ± 37.24 vs. AA: 154.3 PRU ± 60.36 (P=0.01).	[77]
Dabigatran etexilate	CES1 intronic rs2244613 MAF: 0.18 (C) intronic rs8192935 MAF: 0.33 (A)	1694 patients on long-term anticoagulation therapy	rs2244613 and rs8192935 were significantly associated with a 15% decrease in Ctrough (P=1.2×10−8) and 12% decrease in Cmax of active dabigatran per allele (P=3.2×10−8), respectively	Significant association of rs2244613 C allele with decreased risk of any bleeding (OR=0.67, P=7×10−5), minor bleeding (OR=0.7, P=4×10−4). NO association between rs8192935 and bleeding or ischemic events	[79]
Dabigatran etexilate	CES1 intronic rs2244613 MAF: 0.402(T) intronic rs8192935 MAF: 0.228 (G)	86 Chinese patients with atrial fibrillation	There was no significant association between CES1 rs2244613 and the Ctrough and Cmax of dabigatran. CES1 rs8192935 G carriers have significantly increased Ctrough (AA: 42.73 ± 29.04; AG: 58.66 ± 45.46; GG: 112.69 ± 40.63 ng/mL P =0.013) and but not Cmax of dabigatran.	A significant association between rs2244613 (OR=2.30, P = 0.038) and rs8192935 (OR=2.26, P = 0.042) with increased risks of any bleeding, but no significant association with stroke, major bleeding, or minor bleeding	[81]
Dabigatran etexilate	CES1 intronic rs2244613 MAF: 0.389 (A) CES1 intronic rs8192935 MAF: 0.24 (G)	106 healthy Chinese subjects	No significant association between rs2244613 and dabigatran PK parameters. rs8192935 GG carriers have significantly higher Cmax (AA: 50.96 ± 18.93; AG: 56.31 ± 17.17; GG: 86.12 ± 20.61 kg*ng/ mL*mg) but not AUC.	N/A	[82]
Dabigatran etexilate	CES1 intronic rs2244613 MAF: 0.208 (C)	96 patients with atrial fibrillation and stage 3 chronic kidney disease	Patients with rs2244613 CC (n= 5) genotype had a 70% reduction in the mean Ctrough/dose relative to AA genotype (n=62), p = 0.001.	No patients with rs2244613 CC genotype experienced a bleeding event in this study.	[80]
Dabigatran etexilate	CES1 intronic rs2244613 MAF: 0.275 (A)	60 patients undergoing knee replacement surgery	No significant association between rs2244613 and dabigatran Ctrough.	N/A	[115]
Imidapril	CES1 rs2307240 (−816A/C) MAF: 0.248 (C)	105 Japanese hypertensives with no less than 140/90 mmHg	N/A	AC+CC carriers (n=41) had significantly higher responding rate to imidapril (68.3% vs. 46.9%, P = 0.0440), and greater reduction in systolic BP (24.7±11.8 vs. 17.6±16.8 mmHg, P=0.0184) than AA group (n=64)	[19]
Quinapril Enalapril	CES1 rs71647871 (G143E)	Healthy volunteers with G/A (n=10) and G/G (n=12) genotypes.	G143E carriers had a 20% lower AUC0–∞ of active enalaprilat (P = 0.049) and 35% smaller enalaprilat excretion into the urine (P = 0.044) than wild type. No significant effects of G143E on quinapril PK.	G143E showed no significant effects on the PD of enalapril and quinapril.	[86]
Enalapril	CES1 SNPs copy number variation	Healthy Caucasian volunteers (Wild type n=16 4 copies n=5 G143E n=6 3 copies + active promoter n=2 CES1A1c n=4 3 copies + normal promoter n=10)	No significant difference in the median AUC of enalaprilat was found among the six groups. Carriers with 3 copies and normal promoter exhibited a significantly longer half-life of enalaprilat relative to the wildtype (26.7 vs.1 2.4 hour).	N/A	[88]
Enalapril	CES1 intronic rs2244613 MAF:0.24 (C) rs71647871 (G143E) MAF: 0.013(A) CES1A1c (indicated by rs12149368 rs111604615 rs201577108)	286 Caucasian patients with 1–3 grades arterial hypertension	rs2244613 was associated with a 16.4% lower Cmax (P = 0.027) and 20% lower Ctrough (P = 0.022) of enalaprilat per C allele. CES1A1c/CES1A1c carriers (n=6) showed a 75% lower adjusted mean enalaprilat concentration than heterozygotes and wild type (P=0.043)	N/A	[89]
Enalapril	CES1 diplotypes CES1P1/CES1P1, CES1P1/CES1A2, CES1A2/CES1A2	286 Caucasian patients with arterial hypertension	CES1A2/CES1A2 was associated with decreased Cmax (linear regression coefficient: −0.389) and Ctrough (−0.555) of enalaprilat. CES1A1c/CES1A1c was associated with decreased Cmax (−0.541) and Ctrough (−0.692) of enalaprilat.	N/A	[91]
Enalapril	CES1 rs71647871 (G143E) MAF: 0.02	Non-carriers (n = 15) heterozygote carriers (n = 6)	G143E carriers had 30.9% lower Cmax (P = 0.03) and 27.5% lower AUC of enalaprilat (P = 0.02), and 32.3% lower enalaprilat-to-enalapril AUC0–∞ ratio (P = 0.003).	Non-carriers had approximately 12.4% lower systolic BP compared to the baseline (P = 0.001), while G143E carriers had no significant BP reduction.	[87]
Trandolapril	CES1 rs3785161 (−816 A>C) MAF: 0.167 in Caucasians	Patients with hypertensive coronary artery disease (n=486)	N/A	−816 A>C showed no association with systolic BP response to trandolapril.	[20]
Trandolapril	CES1 rs71647871 (G143E)	Healthy volunteers (GE: n=5, GG: n=11)	GE carriers showed a nonsignificant decrease in Cmax (20%) and AUC0–72h (15%) of trandolaprilat.	G/E carriers exhibited nonsignificant reductions in average maximum reductions of systolic BP (22%) and diastolic BP (23%).	[90]
Methylphenidate	CES1 rs71647871 (G143E) MAF: 0.026 intronic rs2307235 MAF:0.041(T) intronic rs2302722 MAF:0.079 (T)	120 healthy subjects	Significant lower metabolite to parent drug ratios were found for G143E (GG:15.7 vs. GA: 5.4, P<0.0015), rs2307235 (GG:16.4 vs. TG:8.2, P=0.04) and rs2302722 (GG:17.0, TG+TT: 11.5, P=0.04).	N/A	[93]
Methylphenidate	CES1 rs71647871 (G143E) MAF: 0.016(E) Copy number variation	healthy Danish Caucasian subjects (Wild type n=16 4 copies n=5 G143E n=6 3 copies + active promoter n=2 CES1A1c n=4 3 copies + normal promoter n=10)	G143E allele carriers had a significantly larger median AUCd-MPH (53.3 ng mL−1 h−1 vs. 21.4 ng mL−1 h−1, P<0.0001). Subjects with 4 CES1 copies had a significantly larger median AUCd-MPH (34.5 ng mL−1 h−1 vs. 23.8 ng mL−1 h−1, P<0.003).	N/A	[94]
Methylphenidate	CES1 rs114119971	99 children and adolescents with Attention-Deficit Hyperactivity Disorder (ADHD)	N/A	rs114119971 carriers exhibited a lower weight-based dose than those without (0.42 mg/Kg vs. 0.88 mg/Kg, p<0.001).	[97]
Methylphenidate	CES1 rs3815583 (−75T>G) MAF: 0.2(G)	213 children with ADHD	N/A	G allele carriers showed a trend with higher appetite reduction baseline scores (P=0.05), appetite reduction over time (P= 0.03), and a higher risk for appetite reduction worsening (OR = 3.47, P=0.01) than TT carriers.	[22]
Methylphenidate	CES1 rs3815583 (−75T>G) MAF: 0.2(G)	74 children with ADHD TG & GG (n=26) TT (n=48)	N/A	G allele carriers had a significantly greater proportion of weight loss and mean weight loss (88%, −0.279Kg) than that of TT carriers (31%, +0.157 Kg)	[98]
Methylphenidate	CES1 intronic SNPs: rs2244613 MAF:0.22 (G) rs2302722 MAF: 0.32(C) rs2307235 MAF:0.21 (A) rs8192950 MAF:0.39 (T)	140 children and adolescents with autistic spectrum disorder	N/A	rs2244613-G (P=0.04, beta = 0.35), rs2307235-A (P= 0.03, beta = 0.37), and rs8192950-T(P= 0.03, beta = 0.29) alleles were associated with higher risks of MPH-induced side effects, whereas rs2302722-C allele (P=0.02, beta=−0.39) was associated with lesser risk of side effects, after accounting for gender, age, and dose. CES1 haplotypes (i.e., TCCCATC and TCCCGGT) were associated with efficacy (P=0.02) and side effects (P= 0.03), respectively.	[99]
Capecitabine	CES1 intronic rs7187684 intronic rs2244613	338 patients with colorectal and gastric cancer	N/A	rs7187684 T carriers had a lower proportion of stage III-IV (OR =0.182, P=0.001), lower relapse rate (OR = 0.541, P=0.036), and longer event-free survival (EFS, 29.6 vs. 13.9 months, P=0.011). rs2244613 TT carriers had a shorter EFS (12.9 vs. 29.0 months, P=0.009), and progression-free survival (P=0.003), relative to G carriers.	[102]
Capecitabine	SNPs of CES1, CES2	144 cancer patients	N/A	rs3217164 (OR=4.08, P < 0.001), rs2244614 (OR=4.78, P= 0.001), rs2244613 (OR=6.40, P= 0.013), rs7187684 (OR=6.48, P=0.012), rs1186118 (OR=6.51, P=0.012), and haplotype A3 containing the minor alleles of these five variants (ORadditive = 2.2, P=0.012; ORrecessive =10.3, P=0.0038) were significantly associated with early on-set overall toxicity.	[100]
Capecitabine	CES1 SNPs CES1 diplotype	37 colorectal cancer patients treated with capecitabine and oxaliplatin	CES1 genetic variants had no significant associations with PK of capecitabine.	CES1 genetic variants had no significant associations with the toxicity of capecitabine.	[103]
Capecitabine	CES1 rs71647871 (G143E)	161 colorectal patients	N/A	A significant association between CES1 rs71647871 and severe hand–foot syndrome (OR = 11.92, P = 0.03).	[105]
Capecitabine	Several SNPs of CES1 and CES2	446 cancer patients	N/A	Significant association between CES1 rs2244613 and higher risk of hand–foot syndrome ≥ grade 2 (OR=1.888, P = 0.027)	[104]
Irinotecan	CES2 haplotypes rs20583738 (Met1Leu) MAF: 0.02 rs20586162 (Arg34Trp) MAF: 0.02	176 Japanese cancer patients receiving irinotecan	CES2 major *1a, *1b diplotypes and relatively rare *1c diplotypes exhibited no significant association with the median AUC ratio of metabolites-to-parent drug; nonsynonymous Arg34Trp (n=1) and Met1Leu (n=1) carriers showed extremely reduced AUC ratios.	N/A	[28]
Irinotecan	CES2 rs11075646 (830C>G) MAF: 0.14(G)	49 metastatic colorectal cancer patients	CES2 830C>G was not associated with any PK parameters, including clearance of irinotecan and AUC of irinotecan and metabolites.	No significant association was observed between CES2 830C>G with neutropenia, diarrhea, response and bilirubin and alanine aminotransferase levels	[107]
Irinotecan	SNPs in CES1, CES2 and other genes	299 cancer patients	N/A	A significantly lower all-grade risk of thrombocytopenia was observed in CES1 rs2244614 (OR = 0.42, P=0.024) and rs7187684 (OR = 0.255, P=0.018) carriers.	[109]
Irinotecan	CES1 diplotypes and SNPs	177 Japanese cancer patients	The median AUC ratio of metabolites-to-parent drug in patients with three or four copies was 1.24-fold of that in patients with two copies (P=0.0134).	Increasing trends of grade 3/4 neutropenia were observed in patients with CES1 3–4 copies and without UGT1A1*6 or 28*, although significance was not achieved.	[14]
Oseltamivir	CES1 rs71647871 (G143E) MAF:0.022	860 healthy Caucasian Finnish volunteers	G143E carriers (n=9) had a 18% larger oseltamivir AUC0–∞ (P=0.025) and 23% smaller active metabolite-to-oseltamivir AUC ratio (P=0.006)	N/A	[110]

MAF, minor allele frequency; AUC, area under the curve; C_trough, trough concentration; C_max, peak concentration; OR, odds ratio; BP, blood pressure; ADHD, attention deficit hyperactivity disorder; EFS, event-free survival; UGT, uridine diphosphate glucuronosyltransferases

4.1. Clopidogrel

Clopidogrel, combined with aspirin, is the standard antiplatelet therapy for coronary heart disease (CHD) and for those undergoing percutaneous coronary interventions [72,73]. As a thienopyridine prodrug, only 15% of clopidogrel is activated by several CYP enzymes (predominantly CYP2C19) to the active 5-thiol metabolite, which irreversibly suppresses platelet aggregation. Notably, 85% of the parent drug, intermediate metabolites, and 48% of the active metabolite undergo CES1-mediated inactivation, forming carboxylic acid-related inactive metabolites (Figure 3a) [72].

Figure 3.

Metabolic pathways of clinically important CES substrate drugs, including (a) clopidogrel, (b) dabigatran etexilate, (c) enalapril, (d) trandolapril, (e) methylphenidate, (f) capecitabine, (g) irinotecan, (h) oseltamivir and (i) remdesivir.

Given the predominant role of CES1 in clopidogrel inactivation, interindividual variability in CES1 expression and activity caused by various regulators might influence clopidogrel PK and PD. Several clinical studies found that the loss-of-function CES1 G143E variant was significantly associated with increased exposure to the active clopidogrel metabolite and enhanced antiplatelet effects in CHD patients and healthy subjects [73–75]. Although the impact of CES1 G143E on clopidogrel clinical outcomes remains unclear, patients with a high bleeding risk might require personalized dosing of clopidogrel to prevent bleeding [2].

The impacts of CES1 -816A/C on clopidogrel PK and PD in humans are not fully elucidated. In a prospective study involving 162 CHD patients, -816C allele carriers had significantly higher platelet aggregation (P=0.014), indicating enhanced CES1 expression/activity and decreased clopidogrel activation in the carriers [17]. Conversely, Zou et al. observed that -816C carriers showed significantly lower platelet aggregation compared to noncarriers (P = 0.001), suggesting decreased CES1 expression/activity and increased clopidogrel activation in these patients [18]. Since the -816C variant was found predominantly in the promoter region of the pseudogene CES1A3 but rarely in functional CES1A2, its impact on CES expression/activity might be ascribed to the unknown functionally important SNPs in CES1A1 that is in high LD with -816A/C [18].

A large prospective study of 851 Chinese patients with acute coronary syndrome revealed that CES1 S75N (rs2307240) was significantly associated with lower occurrences of cerebrovascular events (CT+TT: 4.8% vs. CC: 14% P < 0.001, OR = 0.31), acute myocardial infractions (CT+TT: 6.1% vs. CC: 15.1%, P < 0.001, OR = 0.37) and unstable angina (CT+TT: 37.7% vs. CC: 62.8%, P < 0.001, OR = 0.36) (CC: n =471, CT: n= 372; TT: n = 2) [76]. This study suggested that the T allele was correlated with decreased CES1 expression/activity, thus enhancing clopidogrel efficacy. Moreover, S75N has an MAF of 0.22 in acute coronary syndrome patients, which is over 4-fold higher than that of the general population (MAF = 0.05). Another intronic CES1 SNP, rs2244613, which presumably impacts CES1 phosphorylation, was associated with increased platelet residual activity [77].

4.2. Dabigatran etexilate (DABE)

DABE is an inactive ester prodrug used for preventing and treating deep venous thrombosis, pulmonary embolism, stroke, and systemic embolism in patients with nonvalvular atrial fibrillation. Following oral administration, DABE is sequentially activated by intestinal CES2 and hepatic CES1, forming the active metabolite dabigatran (DAB), which exerts the anticoagulant effect by inhibiting the thrombin pathway (Figure 3b) [78].

Mixed evidence suggests that CES1 rs2244613 and rs8192935 may impact DABE PK and PD. In a subset of 1490 patients from the Randomized Evaluation of Long-term Anticoagulation Therapy trial, CES1 intronic rs2244613 was associated with a 15% decrease in DAB trough concentration (C_trough) per C allele (P=1.2×10⁻⁸), reduced risks of any bleeding (OR = 0.67, P=7×10⁻⁵), minor bleeding (OR = 0.7, P=4×10⁻⁴), a nonsignificant trend in decreasing major bleeding (OR = 0.66, P=0.66) and ischemic stroke or systemic embolism (OR = 0.70, P=0.34) [79]. Similarly, in patients with atrial fibrillation and chronic kidney disease, the CES1 rs2244613 CC genotype was significantly linked with a 70% reduction in the mean C_trough/dose compared to AA carriers [80]. However, studies in healthy and diseased Chinese subjects found no significant correlation of rs2244613 with DAB exposure [81,82]. The CES1 intronic SNP rs8192935 was associated with a 12% decrease in DAB maximum concentration (C_max) per A allele (P = 3.2×10⁻⁸) but only showed a reduced trend of bleeding and ischemic events [79]. In contrast, studies on the Chinese population reported a significant association of CES1 rs8192935 with increased DAB exposure and risk of any bleeding [81,82].

Additionally, an in vitro study on 104 human liver samples showed that G143E carriers were associated with 47% (P =0.018), 57% (P = 0.004), and 63% (P = 0.001) decreases in the mean activation rates of DABE and its two intermediate metabolites, M1 and M2, respectively, relative to non-carriers after the activity was normalized to CES1 expression levels [78]. More recently, a computational analysis suggested that functional nonsynonymous SNPs, including 332G>A, 581C>T, 717C>A, 913C>T, and 977C>T, might reduce CES1 activity against DABE [83]. Moreover, CES2 was suggested to play a role in DABE activation, as indicated by the drastically reduced DAB formation when incubating DABE in HLM, followed by human intestinal microsomes [84]. It remains unclear whether CES1 G143E, 332G>A, 581C>T, 717C>A, 913C>T, 977C>T, and CES2 SNPs can affect the activation and clinical response of DABE in patients.

4.3. Angiotensin-converting enzyme inhibitors (ACEIs)

ACEIs are a family of drugs that elicit antihypertensive effects by inhibiting the formation of angiotensin II and the degeneration of bradykinin. ACEIs are widely used in patients with hypertension, heart failure, diabetes, and chronic kidney disease [85]. Of the 10 FDA-approved ACEIs, eight are ester-containing prodrugs requiring CES1 activation to form active metabolites (e.g., enalapril and trandolapril, Figure 3c–d) [2]. CES1 genetic polymorphisms have shown varying degrees of associations with ACEIs PK and PD.

In a prospective study in healthy volunteers, CES1 G143E carriers had a 20% lower area under the curve (AUC_0–∞) of the active metabolite enalaprilat (P = 0.049) and a 35% reduction in enalaprilat urine excretion (P = 0.044) but showed no difference in antihypertensive effects, compared to non-carriers [86]. The same study found no association between G143E and the PK or PD of quinapril [86]. In another prospective multi-dose study, G143E heterozygotes (n=6) had a 30.9% lower enalaprilat C_max (38.01 vs. 55.01 ng/mL, P = 0.03), 27.5% lower AUC_0–∞ of enalaprilat (374.29 vs. 515.91 ng*h/mL, P = 0.02), and 32.3% lower enalaprilat-to enalapril AUC_0–∞ ratio (21.9 vs 32.0, P = 0.003) compared with the non-carriers (n=15). The average systolic blood pressure (BP) in G143E heterozygote carriers was not reduced, whereas non-carriers had a 12.4% reduction in systolic BP following enalapril treatment (P=0.001) [87]. These two studies suggested that G143E reduced CES1 catalytic activity toward enalapril, but another two studies reported no significant associations between G143E and enalapril PK [88,89].

For trandolapril, the CES1 SNPs -816A>C and G143E showed no significant impact on its exposure and efficacy [20,90]. However, G143E was related to a decreasing trend in trandolaprilat C_max (20%), AUC_0–72h (15%), and average maximum reductions of systolic BP (22%) and diastolic BP (23%) [90]. In 105 Japanese hypertension patients, -816A>C carriers (AC+CC, n = 41) exhibited a significantly higher responding rate to imidapril and a decrease in systolic BP than AA carriers (n=64) [19]. A study in 286 Caucasian hypertension patients found that CES1 rs2244613 carriers had reduced C_max (16.4%, P= 0.027) and C_trough (20%, P=0.022) of enalaprilat per C allele. Additionally, CES1A1c/CES1A1c carriers exhibited a 75% lower C_trough than non-carriers (P=0.043) [89]. Another study using the same cohort suggested a significant association between CES1A2 and lower enalaprilat C_max and C_trough, likely due to the LD of CES1A2 with rs2244613 [91]. Conversely, Stage et al. reported that CES1 copy number, promoter activity, CES1A1c, and G143E did not significantly affect the median AUC of enalaprilat [88]. These findings highlight the need for further research to clarify the influence of CES1 polymorphisms on ACEIs’ exposure and response.

4.4. Methylphenidate (MPH)

MPH is a first-line therapy for attention deficit hyperactivity disorder (ADHD), a psychiatric disorder affecting approximately 1.5–5% of the global population. MPH has two chiral centers, leading to four isomers. Of these, the dl-threo-MPH isomers are used in therapy, with d-MPH being more pharmacologically potent than l-MPH. CES1 is the primary enzyme for converting active dl-MPH to inactive dl-ritalinic acid (RA) (Figure 3e) [92]

Alterations in CES1 expression and function might significantly impact MPH PK and clinical outcomes. CES1 G143E carriers showed significantly lower d-RA/d-MPH ratios [93] and larger median AUCs of d-MPH compared to non-carriers [94,95]. An in vitro HLS9 incubation study has shown that G143E heterozygous carriers had approximately 50% of MPH-metabolizing rate of non-carriers [96]. In a population PK model, integrating the G143E genotype decreased the unexplained interindividual variability of d-MPH apparent clearance by 8.6% [95].

In addition to G143E, the intronic CES1 SNPs rs2244613, rs2307235, and rs2302722 were significantly associated with reduced d-RA/d-MPH ratios, indicating reduced CES1 activity [93]. A study in healthy Caucasian subjects characterized study participants into six groups based on CES1 haplotype structure and SNPs. It revealed that subjects with four CES1 copies had significantly larger median AUC of d-MPH. In contrast, no difference was found between two copies (wild type) and three copies groups regardless of the transcriptional activity of CES1A2 and the presence of CES1A1c [94]. More recently, Xiao et al. studied the collective effects of three CES1 regulators, including the G143E genotype, alcohol (a CES1 inhibitor) consumption and gender, on MPH PK using physiologically based pharmacokinetic (PBPK) modeling. The PBPK model suggested that male G143E heterozygotes who consume alcohol could have 340.1% and 240.6% greater AUC and C_max of d-MPH, respectively, than healthy G143E non-carrier males without consuming alcohol [96].

Furthermore, CES1 genetic polymorphisms demonstrated a significant correlation with the dosing and side effects of MPH. ADHD patients with the SNP rs114119971 required a lower weight-based dose than non-carriers for symptom control (0.42 mg/Kg vs. 0.88 mg/Kg, P<0.001) [97]. The CES1 rs3815583 G allele was significantly associated with appetite reduction over time (P= 0.03) and a higher risk for appetite reduction worsening (OR = 3.47, P=0.01) [22]. A larger proportion of rs3815583 G allele carriers experienced weight loss than non-carriers (88%, −0.279 Kg vs. 31%, +0.157 Kg, P<0.001) [98]. The association of rs3815583 with increased MPH side effects may reflect a reduced MPH metabolism caused by the variant. The possible mechanism might be that (1) rs3815583 in 5’-UTR directly affects CES1 translation initiation, reducing CES1 expression; (2) rs3815583 is in LD with other functional SNPs [98].

In a retrospective study involving 140 children with autistic spectrum disorder, CES1 SNPs rs2244613, rs2307235, and rs8192950 were significantly associated with increased MPH-induced side effects, whereas rs2302722 were significantly correlated with reduced side effects [99]. Interestingly, rs2244613, rs2307235, and rs2302722 carriers had reduced MPH metabolism [93], which might explain the increased side effects of MPH in rs2244613 and rs2307235 carriers, yet the in vivo and in vitro results were contradictive for rs2302722 carriers.

4.5. Capecitabine (CP)

CP is an inactive oral prodrug of 5-FU, widely used for patients with breast and colorectal cancers. CP undergoes a three-step activation process, including (1) the initial activation of CP into 5-deoxy-fluorocytidine (5-dFCR) by CES1 and CES2; (2) the conversion of 5-dFCR into 5-deoxy-fluorourdine (5-dFUR) by cytidine deaminase (CDA); (3) the transformation of 5-dFUR into 5-FU mediated by thymidine phosphorylase (TP) and uridine phosphorylase 1 and 2 (UPP1/2) [100]. Subsequently, 5-FU is either converted to its main pharmacologically active metabolite or catabolized for excretion (Figure 3f) [101]. A meta-analysis on the genetic polymorphisms of genes involved in the PK and PD of CP has been recently published [101].

In a study involving 338 Chinese patients with colorectal or gastric cancer, the CES1 rs7187684 CC genotype was significantly associated with a higher proportion of stage III–IV and relapse rate relative to the CT or TT genotype. Patients with rs7187684 CC or rs2244613 TT genotypes had a significantly reduced event-free survival rate [102]. Another study of 144 CP-treated cancer patients found that five SNPs in the CES1 non-coding region were associated with early-onset overall toxicity: rs3217164 (OR: 4.08, P < 0.001), rs2244614 (OR: 4.78, P = 0.001), rs2244613 (OR: 6.40, P = 0.013), rs7187684 (OR: 6.48, P = 0.012), and rs1186118 (OR: 6.51, P = 0.012), but not specifically with hand-foot syndrome and diarrhea. As these five CES1 SNPs are in high LD, the resulting haplotype A3 containing all minor alleles was more strongly associated with increased CP-induced toxicity (OR _additive = 2.2, P_adjusted =0.012, OR_recesive = 10.3, P_adjusted = 0.0038) [100]. In contrast, Matsumoto et al. found no association of the above five CES1 SNPs, diplotype structure, and copy number variation with the PK and toxicity of CP in 37 colorectal patients treated with CP plus oxaliplatin [103]. Moreover, CES1 rs2244614 carriers had significantly higher risks of hand-foot syndrome ≥ grade 2 (OR=1.888, P=0.027) [104]. A significant association between CES1 G143E and severe hand–foot syndrome (OR = 11.92, P = 0.03) was recently reported [105], although this finding contradicts the results from Hamzic et al. [100]. Further study in a large patient cohort is required to validate the association of CES1 genetic polymorphism with CP exposure and clinical outcomes and to elucidate the mechanisms by which CES1 SNPs impact the PK and PD of CP.

4.6. Irinotecan (CPT-11)

CPT-11 is a widely used DNA topoisomerase I inhibitor for treating colorectal and lung cancers. Only 3–4% of intravenously infused CPT-11 is converted into its active metabolite SN-38 by CES1, CES2, and butyrylcholinesterase (BES). Subsequently, UGT1 efficiently deactivates SN-38 into inactive SN-38G, which can be reconverted into SN-38 by β-glucuronidases in the intestinal microbiome. Moreover, CYP3A4 and CYP3A5 are involved in the inactivation of CPT-11 (Figure 3g) [14,106]. Our discussion focuses on the impact of CES on the PK and PD of CPT-11, and the influence of other enzymes and transporters involved in CPT-11 biotransformation can be found elsewhere [106].

While CES2 has higher catalytic efficiency towards CPT-11 than CES1, CES2 genetic polymorphism generally had limited influence on CPT-11 PK and PD. In a study with 176 Japanese cancer patients, CES2 major diplotypes containing *1a, *1b, *1c did not affect the AUC ratios of (SN-38+SN-38G)/CPT-11. Extremely low AUC ratios were observed in carriers with *2a/*1a (n =1, contains Arg34Trp (MAF: 0.02)) and *5a/*1a (n=1, contains Met1Leu (MAF: 0.02)) [28]. In addition, CES2 promoter SNP 830C>G showed no association with the AUC of CPT-11 or other metabolites, nor with therapeutic response and toxicities like neutropenia and diarrhea [107].

Notably, CES2 is highly expressed in several cancer cell lines, xenograft, and human pancreatic ductal adenocarcinoma tissue (PDAC). In PDAC tissue, a higher CES2 protein expression level was significantly associated with longer overall survival (Hazard ratio = 0.14, P =0.02) in patients receiving a combination therapy of oxaliplatin, irinotecan, and 5-FU [108]. Moreover, a preclinical study revealed that 5-FU induced CES2 expression by upregulating p-53-mediated transactivation, thereby increasing CPT-11 activation and its anticancer efficacy [42].

CES1 polymorphisms, similar to CES2, showed moderate associations with CPT-11 PK and PD. In a study of 177 Japanese cancer patients, carriers with three or four functional CES1 copies (n=35) had a 24% increase in AUC ratios of (SN-38+SN-38G)/CPT-11 compared with those carrying two copies (n=25). However, no significant impacts were observed for the presence of VAR1A1 and major SNPs [14]. In another study on 299 patients, significantly lower risks of all-grade thrombocytopenia (a type of CPT-11-induced toxicity) were observed in patients with the CES1 variants c.1165–41C (rs2244614, OR = 0.42, P=0.024) and n.95346T>C (rs7187684, OR=0.23, P=0.018), necessitating further validation [109].

4.7. Antiviral drugs

CES1 is the key enzyme involved in activating several antiviral prodrugs, such as oseltamivir and remdesivir. Oseltamivir is an ethyl ester prodrug widely used to prevent and treat influenza virus A and B infections. Oseltamivir is efficiently activated into oseltamivir carboxylate by hepatic CES1 (Figure 3h). CES1 polymorphisms, gender, age, and hepatic impairment have been found to alter CES1 expression/activity, potentially impacting oseltamivir activation and effectiveness.

An in vitro incubation study using HLS9 fractions from CES1 143GE carriers (n=5) showed a 60% decrease in oseltamivir hydrolysis relative to samples from 143 GG carriers (n=104, P = 0.005) [56]. In healthy Finnish subjects, an 18% increase of oseltamivir AUC_0-∞ (P=0.025) and a 23% decrease of AUC ratio of active carboxylate-to-oseltamivir (P=0.006) were observed in 143GE carriers (n=9), indicating the G143E variant impaired oseltamivir activation [110]. In healthy Asian subjects, c.662A>G carriers (n=8) exhibited a 10% increase in AUC_{0–48h, Oseltamivir} and a 5% decrease in AUC_{0–48h, Oseltamivir carboxylate}; however, the differences were not statistically significant [111]. Other CES1 polymorphisms, including rs2244613, rs8192935, -816A>C, CES1P1/CES1P1VAR had no impact on in vitro oseltamivir hydrolysis and CES1 hepatic expression [56]. Additionally, HLS9 from female donors showed a 27.8% higher oseltamivir activation rate compared to males (P=0.076), which was consistent with a 17.3% higher CES1 expression in female liver tissues (P=0.039) [56]. Moreover, the oseltamivir activation rate in HLM from pediatric tissues was significantly lower relative to adults (P<0.05). A PBPK modeling study suggested that subjects with Child-Pugh C hepatic cirrhosis could have 2-fold higher C_max and 6-fold higher AUC of oseltamivir, and 30% lower C_max of the active metabolite due to decreased CES1 expression (~ by 70%), potentially resulting in increased risks for toxicity and impaired efficacy [112].

Remdesivir is the first FDA-approved antiviral prodrug for COVID-19 treatment. Following intravenous administration, over 90% of remdesivir undergoes rapid metabolism primarily by CES1 into GS-704277 in the liver, which is further metabolized to the active triphosphate metabolite by intracellular enzymes (Figure 3i) [113,114]. Currently, research on the impact of CES1 on remdesivir PK and PD is limited. A study using in vitro to in vivo scaling analysis indicated that the CES1 G143E variant could potentially reduce remdesivir clearance [113].

5. Drug-drug interactions mediated by CES

Many therapeutic drugs, pharmaceutical excipients, and herbs have been identified as CES inducers or inhibitors by various in vitro methods (Tables 1 and 2). This section delves into CES-mediated drug-drug interaction studies in animal models and humans.

Ethanol is the most well-recognized CES1 inhibitor and can cause clinically significant interactions with various CES1 substrate drugs. Ethanol suppressed the CES1-mediated metabolism of MPH, leading to a 20% to 35% increase in systemic MPH exposure and enhanced stimulant responses in healthy subjects administered with dl-MPH or d-MPH [116,117]. Ethanol also inhibited CES1-mediated activation of oseltamivir without affecting CES2 activity [118]. In healthy subjects, co-administration of ethanol with oseltamivir led to a 27% increase in AUC_0–6h of oseltamivir (range: 11–46%, P=0.011) and a 34% reduction in AUC_0–6h ratio of active metabolite/oseltamivir (range: 25–41%, P<0.001) [118], posing a risk on decreasing antiviral effects. Similarly, ethanol significantly impaired the hydrolysis of anti-HIV prodrug tenofovir alafenamide fumarate (TAF) in HLS9 and human primary hepatocytes, potentially affecting its activation and therapeutic efficacy [119]. More recently, CES1 was found to be responsible for the pre-systemic activation of oral dimethyl fumarate (DMF, the first prodrug for treating multiple sclerosis) into monomethyl fumarate (MMF). In plasma CES1-deficient mice, alcohol co-administration with DMF reduced MMF C_max and AUC in plasma by 69% and 39%, respectively, and that in the brain by 50% and 20%, respectively, indicating that alcohol could decrease the effectiveness of DMF therapy [120]. In a nongenetically modified mouse study, alcohol pretreatment significantly reduced Ces1 expression and Ces1-mediated clopidogrel inactivation but increased Cyp2c expression and clopidogrel activation by Cyp2c, leading to increased antiplatelet effects and bleeding risks [121]. Further clinical research is warranted to investigate the clinical significance of alcohol-drug interactions with TAF, DMF, clopidogrel, and other CES1 substrates.

In addition to alcohol, a few therapeutic drugs were studied regarding CES1-mediated drug-drug interactions. Several potential concurrently used drug pairs, including dexamethasone-oseltamivir for influenza epidemic [122], dabigatran-simvastatin [123], and clopidogrel-simvastatin [124] for cardiovascular diseases, and cannabidiol-MPH [125], were reported to have marginal drug-drug interaction risks with no clinical relevance. The drug-drug interactions between clopidogrel and ACE inhibitors remain inconclusive. A nested case-control study in 45,918 post-myocardial infarction patients treated with clopidogrel found no association between risks of reinfarction, heart failure, or death with the use of ACEI prodrugs (i.e., ramipril or perindopril) relative to ACEI non-prodrugs (i.e., lisinopril), indicating minimal clinical relevance of clopidogrel-ACEI interactions [126]. However, in an epidemiological study of 70,934 patients with myocardial infarction, cotreatment of trandolapril/enalapril with or without clopidogrel had hazard ratios for clinically significant bleeding of 1.10 (95% CI: 0.97–1.25, P = 0.124) and 0.90 (95% CI: 0.81–0.99, P = 0.025), respectively, suggesting ACEIs increased the bleeding risks of clopidogrel [127]. Additionally, a PBPK model, informed by in vitro and animal experiment results, indicated the potential drug-drug interactions of CPT-11 with capecitabine and the herbal component oroxylin A [128].

6. Conclusion

In conclusion, CES1 and CES2 play important roles in drug metabolism, significantly influencing the exposure, effectiveness, and adverse reactions of substrate drugs. Many factors, including genetic polymorphisms, epigenetic modification, age, sex, hepatic dysfunction, inducers, and inhibitors, were reported to regulate CES expression and activity. Despite great efforts to identify genetic regulators, CES1 G143E stands out as the only SNP consistently impacting the PK and PD of multiple CES1 substrate drugs. In contrast, other genetic variations exhibited mixed results, likely due to their small effect size and the large interindividual variability among study subjects. Although in vitro and animal studies identified numerous CES inducers and inhibitors from therapeutic agents, excipients, and herbal constituents, few have been tested for drug-drug interactions in humans, with ethanol being the only one showing clinical significance. Current evidence suggests that known genetic variations moderately impacted substrate drugs’ metabolism and clinical outcomes. The interindividual variability in the PK and PD of drugs metabolized by CES1 and CES2 is likely attributed to genetic and non-genetic regulators. A comprehensive model integrating those contributing factors could better predict in vivo CES function and the consequent effects on the PK and PD of CES substrate medications.

7. Expert opinion

CES1 and CES2 are the most abundant human drug-metabolizing enzymes, pivotal in activating and deactivating a wide range of commonly prescribed drugs. For substrate drugs with a narrow therapeutic index, interindividual variability in CES expression and activity could markedly affect the clinical outcomes of these medications. Understanding the regulation of CES1 and CES2 will lead to personalized treatment strategies that improve therapeutic efficacy and safety.

Pharmacogenomic research has revealed several CES genetic variants explaining some PK and PD variability. However, no CES1 and CES2 pharmacogenomic biomarkers have been integrated into clinical guidelines, such as the Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines and FDA drug labeling for dose individualization. The lack of reliable biomarkers indicative of CES expression and activity, partly due to multifactorial regulation, is a major challenge for personalized CES substrate drug use. Besides genetic polymorphisms, non-genetic factors, including transcription factors, age, gender, disease states, and inducers and inhibitors, contribute to CES variability to varying degrees.

Despite significant progress in CES research that has been made in the past two decades, CES1 and CES2 are understudied compared to other drug-metabolizing enzymes, such as CYPs and UGTs. We envision that three critical knowledge gaps need to be filled to advance the field of research and improve pharmacotherapy involving CES substrate drugs.

1). Elucidate the regulatory mechanisms of CES expression and activity.

Although many genetic variants and non-genetic factors have been identified as regulators of CES function, a significant portion of the interindividual variability in CES1 expression, activity, and the PK of CES substrate drugs remains unexplained. More research is urgently needed to elucidate the genetic polymorphisms and non-genetic elements that regulate CES expression and activity. Moreover, these regulatory factors can be integrated into comprehensive mathematical models to better predict the in vivo catalytic function of CES1 and CES2 and the metabolisms of their substrate drugs. These regulators can also be included as additional covariates to improve the performance of existing PBPK and population PK models.

2). Identify endogenous biomarkers for predicting CES 1 and CES2 function.

Endogenous biomarkers of drug-metabolizing enzymes provide a valuable tool for PK and PD prediction. Plasma CES1 protein is believed to be released from organs with a high abundance of CES1 protein (e.g., the liver). Plasma CES1 is insignificant for drug metabolism because of its extremely low concentration. A recent study was conducted to test the hypothesis that plasma CES1 could serve as a biomarker for hepatic CES1 expression [129]. The results showed a negative correlation between plasma CES1 protein concentrations and the AUCs of d-MPH in human subjects (P=0.014, r=−0.617). After accounting for the interindividual variability in hepatic protein-releasing rate, plasma CES1 alone can explain ~50% variability of d-MPH AUC (P=0.014, r=−0.703) [129], suggesting that plasma CES1 is a potential protein biomarker for in vivo CES1 metabolic function. Further studies with different CES1 substrates are required to establish plasma CES1 as a reliable biomarker for predicting the PK and PD of CES1 substrate drugs. The relationship between plasma CES2 and in vivo CES2 function has not been explored, probably due to the technical challenges associated with quantifying CES2 in plasma. Notably, extracellular vesicles (EVs) have been proposed as a surrogate to predict the expression levels of various drug-metabolizing enzymes in different organs. An ultrasensitive proteomic method is required to quantify CES1 and CES2 because of the extremely low protein concentrations of enzymes in EVs. Additionally, real-time monitoring of CES1 functionality in biological samples using a highly sensitive biosensor was achieved [130], representing another potential solution to quantify CESs functionality. It should be noted that both CES1 and CES2 are involved in the metabolisms of endogenous substances, such as lipids. Therefore, it is possible that the levels of certain endogenous metabolites may reflect in vivo CES function. An untargeted metabolomics study in human subjects with known CES phenotypes could reveal potential metabolite biomarkers for CES1 and CES2. We expect that these protein and metabolite biomarkers will allow for the personalized pharmacotherapy of CES substrate drugs and improve clinical outcomes.

3). Optimize prodrug design based on tissue-specific expression patterns of CES1, CES2, and other activating enzymes.

Prodrug design has been increasingly utilized in drug development to improve a drug’s PK properties and safety. Prodrug activation is catalyzed by various enzymes in vivo to form pharmacologically active metabolites. About half of all marketed prodrugs are estimated to be activated by hydrolases [131]. CES1 and CES2 are the primary hydrolases in humans, and both enzymes are involved in the activation of many clinically important prodrugs, such as remdesivir, oseltamivir, tenofovir disoproxil, tenofovir alafenamide, sofosbuvir, and irinotecan. Because of tissue-specific expression patterns of CES1 and CES2, CES-mediated prodrug activation is also tissue-dependent. For example, CES1 is the primary hydrolase involved in the hydrolytic step of the activation of remdesivir, tenofovir alafenamide, and sofosbuvir; consequently, the majority of these prodrugs are activated in the liver due to the extremely high hepatic CES1 expression (Figure 1). The highly efficient hepatic activation is a favorable feature for prodrugs that target liver diseases (e.g., tenofovir alafenamide and sofosbuvir) but undesirable for prodrugs targeting extrahepatic organs (e.g., remdesivir). In fact, most dosed remdesivir is activated in the liver rather than in the lung due to the extremely high CES1 expression level in the liver [132]. The hydrolytic reaction of remdesivir activation is catalyzed by both CES1 and cathepsin A (CatA), and unlike CES1, the expression levels of CatA are similar between the liver and the lung [133]. Therefore, it is sensible to develop remdesivir analogs that are more resistant to CES1 but labile to CatA. These analogs are expected to be activated much less in the liver than remdesivir while being efficiently activated in the lung. This strategy can also be applied to the design of other prodrugs with targeted organs outside of the liver.

In summary, we envision that continued efforts in the field will lead to the discovery of additional genetic polymorphisms and non-genetic factors involved in the regulation of CES1 and CES2 expression and activity and a more complete understanding of the observed large interindividual variability in the metabolism, PK, and PD of CES substrate drugs. It is essential to identify and validate biomarkers that reflect the in vivo functions of CES1 and CES2 in order to develop a personalized pharmacotherapy regimen to improve the outcomes of the drugs metabolized by the enzymes. Moreover, we need to take advantage of our knowledge regarding tissue-specific expression patterns of CES1, CES2, and other hydrolases and the substrate selectivity of these enzymes to design prodrugs that are selectively activated in the targeted organs.

Table 4.

Drug-drug interactions mediated by CES1 and CES2 and the implications in PK and PD.

Drug	CES substrate drugs	Type of study	Impact on PK and PD	Ref.
Alcohol	dl-MPH d-MPH (CES1)	Crossover study Healthy subject (male n=12, female n=12)	Combined use of dl-MPH or d-MPH with ethanol significantly increased drug effects	[116]
Alcohol	dl-MPH d-MPH (CES1)	Crossover study Healthy subjects (n=14)	The second pulse of Cmax of d-MPH and AUC4–8h following a dose of dl-MPH and d-MPH were increased by ethanol by 35% (P<0.01) and 25% (P<0.05), as well as 27% (P=0.001) and 20% (P<0.001) respectively. Ethanol significantly potentiated stimulant responses to either formulation.	[117]
Alcohol	Oseltamivir (CES1) Aspirin (CES2)	Crossover study Healthy subjects (n=17)	Alcohol use increased the AUC0–6h of oseltamivir by 27% (P=0.011) and reduced the AUC0–6h ratio of metabolite/oseltamivir by 34% (P<0.001). Alcohol has no impact on aspirin PK.	[118]
Alcohol	Clopidogrel (CES1)	Male C57BL/6J mice	Alcohol pretreatment inhibited Ces1-mediated clopidogrel hydrolysis but increased Cyp2c-mediated activation of clopidogrel in mice.	[121]
Alcohol	Dimethyl fumarate (CES1)	plasma CES-deficient mouse strain	Alcohol use decreased the Cmax and AUC of active metabolite by 69% and 39% in the plasma, 50% and 20% in the brain.	[120]
Dexamethasone	Oseltamivir (CES1)	Multiple dose clinical study Healthy subjects (n=19)	Dexamethasone use lowered the metabolic ratio of oseltamivir by 8% in healthy volunteers.	[122]
Simvastatin	Dabigatran (CES1)	Single sequence study Healthy subjects (n=12)	Concurrent simvastatin treatment had a negligible impact on dabigatran etexilate PK and anticoagulant effects.	[123]
Clopidogrel	Angiotensin-converting enzyme inhibitor (ACEI, CES1)	Case-control study in patients treated with clopidogrel following acute myocardial infarction	No association between reinfarction, heart failure, or death and the use of ACEI prodrug (i.e., ramipril or perindopril) relative to active ACEI (i.e., lisinopril)	[126]
Trandolapril Enalapril	Clopidogrel (CES1)	Epidemiological study in 97,52 patients	In 70,934 patients with myocardial infarction, hazard ratios for clinically significant bleeding in ACEI-treated patients cotreated with or without clopidogrel were 1.10 (P = 0.124) and 0.90 (P = 0.025), respectively, as compared with patients who did not receive ACEIs.	[127]
Cannabidiol	Methylphenidate (CES1)	Crossover study on 12 subjects	The cannabidiol group has slightly larger Cmax (13.5±43.7% ng/mL v.s. 12.2 ± 36.4% ng/mL) and AUCinf (70.7 ± 32.5% and 63.6 ± 25.4%) of MPH than the placebo group. The geometric mean ratio of AUCinf and Cmax for MPH only and MPH+CBD is 1.09 and 1.08, respectively.	[125]

ACEI, angiotensin-converting enzyme inhibitor

Article highlights.

• Genetic polymorphisms in CES1 and CES2 have shown a modest impact on drug exposure and response.

• The CES1 rs71647871 variant, the most studied loss-of-function variant, consistently and significantly affects the PK and PD of CES1 substrate drugs.

• The clinical significance of other genetic variants is debatable, presumably due to their small effect size.

• Regulation beyond genetic polymorphisms, such as age, gender, inducers, and inhibitors, could further explain the variability in CES functionality, PK, and PD.

• CES1 and CES2 in plasma and EVs have the potential to be established as biomarkers for precision pharmacotherapy of CES substrate drugs.

• A better understanding of tissue-specific expression and substrate selectivity of CES1 and CES2 could shed light on improving the design of prodrugs.