Insights into the Substrate Specificity, Inhibitors, Regulation, and Polymorphisms and the Clinical Impact of Human Cytochrome P450 1A2

Abstract

Human CYP1A2 is one of the major CYPs in human liver and metabolizes a variety of clinically important drugs (e.g., clozapine, tacrine, tizanidine, and theophylline), a number of procarcinogens (e.g. benzo[a]pyrene and aflatoxin B₁), and several important endogenous compounds (e.g. steroids and arachidonic acids). Like many of other CYPs, CYP1A2 is subject to induction and inhibition by a number of compounds, which may provide an explanation for some drug interactions observed in clinical practice. A large interindividual variability in the expression and activity of CYP1A2 and elimination of drugs that are mainly metabolized by CYP1A2 has been observed, which is largely caused by genetic (e.g., SNPs) and epigenetic (e.g., DNA methylation) and environmental factors (e.g., smoking and comedication). CYP1A2 is primarily regulated by the aromatic hydrocarbon receptor (AhR) and CYP1A2 is induced through AhR-mediated transactivation following ligand binding and nuclear translocation. To date, more than 15 variant alleles and a series of subvariants of the CYP1A2 gene have been identified and some of they have been associated with altered drug clearance and response to drug therapy. For example, lack of response to clozapine therapy due to low plasma drug levels has been reported in smokers harboring the −163A/A genotype; there is an association between CYP1A2*1F (−163C>A) allele and the risk for leflunomide-induced host toxicity. The *1F allele is associated with increased enzyme inducibility whereas *1C causes reduced inducibility. Further studies are warranted to explore the clinical and toxicological significance of altered CYP1A2 expression and activity caused by genetic, epigenetic, and environmental factors.

INTRODUCTION

The CYP1A gene cluster has been mapped to chromosome 15, with a close link between CYP1A1 and 1A2 sharing a common 5′-flanking region (1). The CYP1A1 and 1A2 genes are separated by a 23-kb segment that contains no open-reading frames (1). Between CYP1A2 and CYP1A1, exons 2, 4, 6, and especially 5 are strikingly conserved in both nucleotides and total number of bases (2). Jaiswal et al. (3) were the first to isolate a cDNA corresponding to the CYP1A2 gene located on human chromosome 15. The human CYP1A2 gene spans almost 7.8 kb comprising seven exons and six introns (2), and the first exon is a 55-bp-long non-coding exon. CYP1A2 is a 515-residue protein with a molecular mass of 58,294 Da (2).

CYP1 family comprises three members, namely CYP1A1, 1A2, and 1B1, with CYP1A2 being one of the major CYPs in human liver (∼13–15%) (4). A large interindividual variability in the elimination of drugs that are metabolized by CYP1A2 has been observed, which has been ascribed to both genetic mutations and environmental factors (5–8). This review updates our current knowledge on the substrate specificity, regulation, structural features, and polymorphisms of human CYP1A2. To retrieve relevant data, the authors have searched through computer-based literatures by full text search in Medline (via Pubmed), ScienceDirect, Genetics Abstracts (CSA), SCOPUS, Chemical Abstracts, Current Contents Connect (ISI), Cochrane Library, CINAHL (EBSCO), CrossRef Search, and Embase (all from inception to 12 June 2009). Keyword search terms used included cytochrome P450, CYP1A2, substrate, inhibitor, inducer, structure, mutation, single nucleotide polymorphism (SNP), mutagenesis, genotype, phenotype, and allele, with combination terms pharmacokinetics, response, clearance, adverse reaction, drug interaction, toxicity, response, and human.

STRUCTURAL FEATURES AND FUNCTIONAL RELEVANCE OF CYP1A2

Homology modeling studies have suggested that most of the CYP1A2 substrates are hydrophobic with high LogP values, suggesting that hydrophobic interactions play an important role in their binding to CYP1A2 (9). Our modeling studies have indicated that the common features of ligands for CYP1A2 are of one to two hydrophobic regions, an aromatic ring and a hydrogen bond acceptor. A number of mutagenesis studies (10,11) have demonstrated that a series of residues in the substrate recognition sequence (SRS) regions of CYP1A2 (e.g., Arg108, Thr124, Thr223, Glu225, Phe226, Lys250, Arg251, Lys253, Asn312, Asp313, Glu318, Thr319, Asp320, Thr321, Val322, Leu382, Thr385, and Ile386) play important roles in substrate–enzyme interactions based on mutagenesis and homology modeling studies (Table I). Several residues in the non-SRS regions, including Lys99, Arg137, Gln141, Phe186, Phe205, Val227, Lys453, Arg455, and Thr501, also appear to play a role in the ligand–CYP1A2 interactions. The functional effect of mutants of these residues appears to be dependent on the substrate.

Table I.

Identified Amino Acid Residues of CYP1A2 that may Play a Role in CYP1A2–Ligand Interactions by Mutagenesis and/or Molecular Modeling Studies

Residue	Possible location	Assays	References
Thr78	–	Homology modeling	(79)
Thr87	–	Homology modeling	(79)
Phe88	–	Homology modeling	(79)
Lys94	Helix B	SDM	(10)
Lys99	Helix B	SDM	(80)
Lys105	Helix B	SDM	(10)
Arg108	SRS1	Homology modeling	(81)
		SDM	(82)
Phe114	SRS1	Homology modeling	(9)
Thr115	SRS1 (Helix B′)	Homology modeling	(83)
Ile117	SRS1	SDM Homology modeling	(9)
Asp119	SRS1 (Helix B′–C)	Homology modeling	(83)
Thr124	SRS1 (Helix B′–C)	SDM	(11)
		Homology modeling	(83)
Ser126	SRS1	Random mutagenesis	(84)
Ser129	SRS1	Random mutagenesis	(84)
Arg135	Helix C	SDM	(10)
Arg136	Helix C	SDM	(10)
Arg137	Helix C	SDM	(80)
Gln141	–	Homology modeling	(81)
His163	–	SDM	(85)
Phe181	–	Homology modeling
Thr184	–	Homology modeling
Phe186	–	Random mutagenesis	(84)
Phe205	–	Homology modeling	(9)
Thr223	SRS2 (Helix F)	SDM	(11)
		Homology modeling	(83)
Glu225	SRS2	Random mutagenesis	(84)
Phe226	SRS2	Homology modeling	(9)
		SDM	(84,86)
		Random mutagenesis	(84)
Val227	–	SDM	(11)
Lys250	SRS3	SDM	(87)
Arg251	SRS3	SDM	(87)
Lys253	SRS3	SDM	(87)
Asn259	SRS3	Homology modeling	(79)
Phe266	–	Homology modeling	(79)
Asn312	SRS4	SDM	(11)
Asp313	SRS4 (Helix I)	Homology modeling	(81,83)
		SDM	(82)
Gly316	SRS4 (Helix I)	Homology modeling	(83)
Glu318	SRS4 (Helix I)	SDM	(10,88–92)
Glu319	SRS4 (Helix I)	SDM	(10,88–91)
Asp320	SRS4 (Helix I)	Random mutagenesis	(84)
Thr321	SRS4 (Helix I)	Homology modeling	(81)
		SDM	(82)
Val322	SRS4	Random mutagenesis	(84)
Phe330	–	Homology modeling	(79)
Leu382	SRS5	SDM	(11)
		Homology modeling	(83)
Pro383	SRS5	Homology modeling	(83)
Thr385	SRS5	Homology modeling	(81)
		SDM	(82)
		Random mutagenesis	(84)
Ile386	SRS5	Random mutagenesis	(84)
Lys401	–	SDM	(10)
Tyr437	–	Homology modeling	(79)
Lys440	–	SDM	(10)
Lys453	–	SDM	(10,80)
Arg455	–	SDM	(10,80)
Arg456	Near heme	Homology modeling	(83)
Ile462	Helix L	SDM and homology modeling	(9)
Lys463	Helix L	SDM	(10)
Tyr495	Near strand β4-1	Homology modeling	(83)
Thr501	–	Homology modeling	(9)
		SDM	(84,86)

SDM site-directed mutagenesis, SRS substrate recognition sequence

CYP1A2 substrates generally contain planar ring that can fit the narrow and planar active site of the enzyme. The planar active site architecture in CYP1A2 is well adapted for the oxidation of relatively large aromatic compounds which is conserved among all CYP1 members. Sansen et al. (12) recently reported the crystal structure of CYP1A2 in complex with α-naphthoflavone (ANF) which has been refined to 1.95 Å (PDB ID: 2HI4). In the 2HI4 structure in complex with ANF, the rather compact active site is closed without clear solvent or substrate access channels with a relatively small volume of the cavity of 375 Å³ (Fig. 1) (13) which is 44.2% larger than that of CYP2A6 (260 Å³) (12).

Fig. 1.

The structures of human CYP1A2 in complex with ANF (PDB ID = 2HI4; a and c) or caffeine b. The distance from the metabolic site of caffeine to the heme is 4.34 Å

The 2HI4 structure contains 12 α-helices designated A–L and four β-sheets designated 1–4 (12). As observed with other mammalian CYPs of known structure, the most conserved regions are the core of the protein forming the heme binding site and the proximal surface that is considered to offer binding sites for cytochrome P450 reductase and cytochrome b₅. The most distinct regions between known CYP structures are the portions that constitute the distal surfaces of the substrate binding cavity, the helix B–C and F–G regions, and the C-terminal loop following helix L. In the 2HI4 structure, the α-helical hydrogen-bonding manner is lost at residues Val220 and Lys221, resulting in one helical turn in the middle of helix F to unwind (12). Two water molecules fill the space and thus form water-bridged contacts between Val220 carbonyl oxygen and Thr223 Oγ, and Lys221 carbonyl oxygen and His224 amide nitrogen, respectively.

Sansen et al. (12) have reported that the substrate binding cavity of CYP1A2 is lined by residues on helices F and I that define a relatively planar binding platform on either side. Helix I bends while it crosses the heme prosthetic group, placing its residues in one flat side of the substrate binding cavity. Coplanarity is thus formed through the Ala317 side chain, the Gly316–Ala317 peptide bond, and the Asp320–Thr321 peptide bond. On the other side of the cavity, the side chain of Phe226 of helix F forms a parallel substrate binding surface. The active site cavity of CYP1A2 is stabilized through a strong hydrogen-bonding interaction between the side chain of Thr223 on helix F and the side chain of Asp320 on helix I. Both Thr223 and Asp320 play a role in forming an extensive network of hydrogen-bonded water molecules and side chains, including Tyr189, Val220, Thr498, and Lys500.

Our docking studies using the published crystal structure of CYP1A2 by Sansen et al. (12) have identified Phe226, Ala317, Gly316, Phe125, and Thr124 as the most important residues to influence the inhibitory potency of a series of natural compounds from Chinese herbal medicines [Zhou SF, et al. (2009) unpublished data]. Based on the structure of 2HI4, we have successfully docked two substrates (amitriptyline and naproxen) and two competitive inhibitors (ciprofloxacin and methoxsalen) into its active site (Fig. 2). The distance from the site of metabolism of amitriptyline and naproxen to the heme group is 4.73 and 3.22 Å, respectively. Both ciprofloxacin and methoxsalen are orientated above the heme group.

Fig. 2.

The docking of two substrates (amitriptyline and naproxen) and two competitive inhibitors (ciprofloxacin and methoxsalen) into the active site of CYP1A2 (PDB ID = 2HI4). The distance from the site of metabolism of amitriptyline and naproxen to the heme group is 4.73 and 3.22 Å, respectively

SUBSTRATE SPECIFICITY OF HUMAN CYP1A2

Phenacetin, caffeine, and theophylline have been frequently used as model substrates for evaluating the activity of CYP1A2 in vivo (14). Industry investigators have used phenacetin O-deethylation as the most common marker reaction for CYP1A2 activity in the in vitro study of 45% new drugs (87 out of 194 studies) (15). Caffeine 3-demethylation and ethoxyresorufin O-deethylation are also commonly used marker reactions for CYP1A2. To a less extent, tacrine (1,2,3,4-tetrahydro-9-aminoacridine, 1-hydroxylation), melatonin (6-hydroxylation) and tizanidine (oxidation) are also used as probes for CYP1A2 in vivo. For in vitro studies, phenacetin (O-deethylation), 7-ethoxycoumarin (O-deethylation), and 7-ethoxyresorufin (O-deethylation) are commonly used probes for determining CYP1A2 activity (16).

CYP1A2 metabolizes a variety of clinically important drugs with variable contributions to their overall elimination (Table II) (17). Drugs that are markedly (>30%) metabolized by CYP1A2 include caffeine, clozapine, ropivacaine, olanzapine, tizanidine, theophylline, tacrine, aminopyrine, zolmitriptan, melatonin, propranolol, duloxetine, leflunomide, promazine, verapamil, thioridazine, nabumetone, ropinirole, and riluzole. CYP1A2 can metabolize some drugs to a minor to moderate extent (10–30%), including mexiletine, paracetamol, flutamide, lidocaine, imipramine, paraxanthine, propafenone, terbinafine, bortezomib, aminoflavone, 5,6-dimethylxanthenone-4 acetic acid (a phase II anticancer drug), trazodone, R-warfarin, R-acenocoumarol, propofol, antipyrine, pentoxifylline, alosetron, furafylline, and tegafur. In addition, CYP1A2 plays a minor role (<10%) in the metabolism of fluvoxamine, pranidipine, zolpidem, thalidomide, cyclobenzaprine, naproxen, coumarin, ondansetron, haloperidol, guanabenz, rofecoxib, efavirenz, bufuralol, cinnarizine, flunarizine, azelastine, almotriptan, carvedilol, amitriptyline, nortriptyline, clomipramine, carbamazepine, albendazole, methadone, amiodarone, and diphenhydramine.

Table II.

Clinical Drugs as Substrates of Human CYP1A2

Substrate	Drug class	Metabolic pathway(s)	Estimated contribution by CYP1A2 (%)a
R-Acenocoumarol	Oral anticoagulant	6-Hydroxylation	10–20
Acetaminophen (paracetamol)	Analgesic drug	Oxidation (to form NAPQI)	10–15
Albendazole	A benzimidazole anthelmintic drug	4-Hydroxylation	5–10
Almotriptan	Selective agonist of 5-HT1B/1D receptor (for treatment of migraine)	N-demethylation	5–10
Alosetron	Selective 5-HT3 antagonist (for irritable bowel syndrome)	N-dealkylation and hydroxylation	10–20
Amiodarone	Antiarrhythmic drug	N-deethylation	5–10
Aminoflavone (NSC686288)	Anticancer agent	3-Hydroxylation	10–20
Aminopyrine	Analgesic drug	N-demethylation	40–50
Amitriptyline	Tricyclic antidepressant	10-Hydroxylation	5–10
Antipyrine	Analgesic drug	N-demethylation and 3- and 4-hydroxylation	10–25
Azelastine	Antihistamine	N-demethylation	5–10
Bortezomib	Proteasome inhibitor (for treatment of multiple myeloma)	Deboronation and dehydrogenation	10–25
Bufuralol	β-Blocker	1′-Hydroxylation	5–10
Caffeine	CNS stimulant	N-demethylation	95
Carbamazepine	Anticonvulsant	2- and 3-Hydroxylation	5–10
Carvedilol	β-Blocker	8-Hydroxylation and O-demethylation	5–10
Cinnarizine	Antihistamine drug	N-dealkylation	5–10
Clomipramine	Tricyclic antidepressant	N-demethylation	5–10
Clozapine	Atypical antipsychotic drug	N-demethylation and N-oxidation	40–55
Coumarin	Natural compound as anticancer agent	7-Hydroxylation	5–10
Cyclobenzaprine	Tricyclic antidepressant	N-demethylation	5–10
5,6-Dimethylxanthenone-4 acetic acid	Anticancer drug	6-Methyl hydroxylation	10–20
Diphenhydramine	Antihistamine drug	N-demethylation	5–10
Duloxetine	Antidepressant	4-, 5-, and 6-Hydroxylation	30–40
Efavirenz	Non-nucleoside reverse transcriptase inhibitor	7- and 8-Hydroxylation	5–10
Erlotinib	Epidermal growth factor receptor tyrosine kinase inhibitor (for cancer treatment)	O-demethylation	5–10
Estradiol	Steroid in oral contraceptives	2- and 4-hydroxylation	5–10
Flunarizine	Antihistamine	N-dealkylation	5–10
Flutamide	Non-steroidal antiandrogen for prostate cancer treatment	2-Hydroxylation	10–25
Fluvoxamine	SSRI	O-demethylation	5–10
Furafylline	Antiasthmatic agent		10–20
Guanabenz	α2-Adrenoreceptor agonist (antihypertensive)	N-hydroxylation	5–10
Haloperidol	Antipsychotic drug	N-dealkylation	5–10
Imatinib	Bcr-Abl tyrosine kinase inhibitor (as anticancer agent)	O-demethylation
Indiplon	Hypnotic drug	N-demethylation	5–10
Imipramine	Tricyclic antidepressant	N-demethylation	10–20
KR-62980	Selective peroxisome proliferator-activated receptor-γ modulator	Hydroxylation	5–10
KR-63198	Selective peroxisome proliferator-activated receptor-γ modulator	Hydroxylation	5–15
Leflunomide	Disease-modifying anti-inflammatory agent	N―O bond cleavage	40–55
Lidocaine	Antiarrhythmic and local anesthetic drug	N-deethylation and 3-hydroxylation	10–20
Maprotiline	Antidepressant		5–10
Melatonin	Pineal hormone	6-Hydroxylation and O-demethylation	40–60
Mexiletine	Antiarrhythmic drug	Para- and 2-hydroxylation	10–20
Mianserin	Antidepressant	N-oxidation and N-demethylation	5–10
Mirtazapine	Antidepressant	N-oxidation, 8-hydroxylation, and N-demethylation	5–15
Naproxen	NSAID	O-demethylation	5–10
Nabumetone	NSAID	Aliphatic hydroxylation	30
Nicotine	Agent for treating nicotine dependence	N-hydroxylation and C-oxidation	5–10
Nortriptyline	Tricyclic antidepressant	10-Hydroxylation	5–10
Olanzapine	Atypical antipsychotic drug	N-demethylation, and 4- and 7-hydroxylation	30–40
Ondansetron	Selective serotonin 5-HT3 receptor antagonist (antiemetic)	7- and 8-Hydroxylation	5–10
Paraxanthine	CNS stimulant	8-Hydroxylation	80
Pentoxifylline	A drug used to treat intermittent claudication	Xanthine 7-demethylation	10–20
Phenacetin	Analgesic drug	O-deethylation	86
Pranidipine	Calcium channel blocker	Methyl hydroxylation	5–10
Promazine	Antipsychotic drug	N-demethylation and 5-sulfoxidation	30–45
Propafenone	Antiarrhythmic drug	5-Hydroxylation and N-desalkylation	10–25
Propofol	Sedative agent	4-Hydroxylation	10–25
Propranolol	β-Blocker	N-desisopropylation	30–50
Riluzole	Antiglutamate agent for treatment of ALS	N-hydroxylation, and possibly 4-, 5-hydroxylation and O-dealkylation	75–80
Rofecoxib	Selective COX-2 inhibitor (withdrawn from the market)	5-Hydroxylation	5–10
Ropinirole	Non-ergoline dopamine agonist (for treatment of Parkinson’s disease)	N-deethylation	30–40
Ropivacaine	Local anesthetic drug	3-Hydroxylation	50–65
Selegiline	Selective irreversible MAO-B inhibitor (for treatment of Parkinson’s disease)	N-demethylation	5–15
Tacrine	Centrally acting cholinesterase inhibitor (for treatment of Alzheimer’s disease)	1-, 2-, 4-, and 7-hydroxylation	50–65
Tegafur (tetrahydrofuranyl-5-fluorouracil)	Anticancer drug (a prodrug)	N-dealkylation	10–20
Terbinafine	Antifungal agent	N-demethylation and side chain oxidation	10–25
Thalidomide	Antiangiogenic agent (for treatment of multiple myeloma)	5′-, 5-, and 6-hydroxylation	5–10
Theophylline	Bronchodilator	N-demethylation	90–95
Thioridazine	Antipsychotic drug	5-Sulfoxidation and N-demethylation	35–45
Tizanidine	Muscle relaxant		80–95
Trazodone	Antidepressant	Hydroxylation, dealkylation, and N-oxidation	15–20
Triamterene	Potassium-sparing diuretic	4′-Hydroxylation	5–10
Verapamil	Calcium channel blocker	N-demethylation	20–30
R-Warfarin	Oral anticoagulant	6-, 8-, and 10-Hydroxylation	10–20
WHI-P131 (JANEX-1)	Janus kinase-3 inhibitor (an anti-inflammatory agent)	7-O-demethylation	5–15
Zileuton	Specific inhibitor of 5-lipoxygenase (for treatment of asthma)	Ring hydroxylation	5–15
Zolmitriptan	Selective agonists of 5-HT1B/1D receptor (for treatment of migraine)	N-demethylation and N-oxidation	30–40
Zolpidem	Non-benzodiazepine hypnotic drug	Ring hydroxylation	5–10
Zotepine	Atypical antipsychotic drug	2- and 3-Hydroxylation	5–10
Zoxazolamine	Muscle relaxant	6-Hydroxylation	5–10

Abbreviations: ALS amyotrophic lateral sclerosis, CNS central nervous system, COX cyclooxygenase, MAO monoamine oxidase, NSAID non-steroid anti-inflammatory drug, SSRI selective serotonin reuptake inhibitor

^aThe estimation contribution is based on inhibition studies using chemical inhibitors and/or inhibitory antibodies, and pharmacogenetic studies where poor and extensive metabolizers of CYP1A2 are compared. Data are from Rendic (17) and our own databases of CYPs

CYP1A2 is involved in the metabolism of a number of natural compounds, probably resulting in toxic metabolites (17,18). Estragole is bioactivated to reactive metabolites by CYP1A2, and CYP1A2 and 2C9 are the major enzymes responsible for conversion of methyleugenol to 1′-hydroxymethyleugenol (19). Aristolochic acids (AAs), naturally occurring nephrotoxins, and rodent carcinogens found in some herbal medicines have been associated with the development of urothelial cancer in humans. Both CYP1A1 and 1A2 contribute to their activation (20).

CYP1A2 together with CYP1A1 and 1B1 play important roles in the bioactivation of a variety of carcinogenic polycyclic aromatic hydrocarbons [PAHs, e.g., benzo[a]pyrene (B[a]P)], heterocyclic aromatic amines/amides (e.g., 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine and 2-amino-3,4-dimethylimidazo[4,5-f]quinoline), and mycotoxins found in some grains such as aflatoxin B1 (AFB₁) and the specificity of the substrates often overlaps (21). AFB₁ is a potent hepatotoxin and procarcinogen in a number of animals and is associated epidemiologically with a high incidence of primary hepatocellular carcinoma in humans. The critically reactive metabolite of AFB₁ is the exo 8,9-epoxide formed by a two-electron oxidation mainly catalyzed by CYP3A4, with contribution from CYP1A2 and other CYPs (22). Oxidation of the chemicals by CYP1A1 and 1A2 serves as an initial step in the conversion of the substrates to more polar metabolites, resulting in increased excretion.

Moreover, CYP1A2 metabolizes several important endogenous substrates, such as melatonin, bilirubin, uroporphyrinogen, estrone and estradiol, and arachidonic acid (17). CYP1A1, 1A2, and 1B1 6-hydroxylate melatonin, with minor contribution from CYP2C19 (23). CYPA2 catalyzed uroporphyrinogen oxidation to form uroporphyrin (24). In addition, CYP1A2, 1A1, and 3A4 are involved in the metabolism of estradiol and estrone (25).

INDUCTION OF CYP1A2 THROUGH AROMATIC HYDROCARBON RECEPTOR (AHR)

CYP1A2 and 1A1 are highly inducible at both mRNA and protein levels by a variety of chemicals, smoking, and several dietary factors (26,27). Many potent inducers of CYP1A1 such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 3-methylcholanthrene, and β-naphthoflavone are also potent inducers of CYP1A2. However, rifampicin, a prototypical inducer of CYP3A, is only a weak inducer of CYP1A2 (28).

CYP1A2 is induced through AhR-mediated transactivation following ligand binding and nuclear translocation (Fig. 3), but the extent of induction is generally lower than that of CYP1A1 (29). AhR is a ligand-activated transcription factor and a basic helix-loop-helix (bHLH) protein belonging to the Per-Arnt-Sim (PAS, where Per stands for Drosophila period clock protein, Arnt refers to AhR nuclear translocator, and Sim is Drosophila single-minded protein) family of transcription factors (26,27). The bHLH motif is located in the N-terminal of the protein and is a common entity in a variety of transcription factors. Members of the bHLH superfamily have two functionally distinctive and highly conserved domains (30). The first is the basic region (b) which is involved in the binding of the transcription factor to DNA, while the second is the helix-loop-helix (HLH) region which facilitates protein–protein interactions. AhR contains two PAS domains, PAS-A and PAS-B, which are stretches of 200–350 amino acids that exhibit a high sequence homology to the protein domains that were found in the Drosophila genes period (Per) and single-minded (Sim) and in Arnt. The AhR exists as cytoplasmic aggregates bound to two 90-kDa heat-shock proteins (Hsps), the cochaperone prostaglandin E synthase 3 (p23), and a 43-kDa immunophilin-like protein hepatitis B virus X-associated protein 2 (XAP2, also called AhR interacting protein 1, AIP1; or AhR-associated protein 9, ARA9) (31). These other proteins are involved in the correct folding and stabilization of AhR. Upon binding a ligand, after the replacement of its associated molecule with Arnt to form a heterodimer with release of 90 kDa HSPs, AhR translocates into the nucleus (Fig. 3). This heterodimer interacts with a 5′-GCGTG-3′ DNA sequence, the core binding motif of the xenobiotic response element or dioxin response element of the target genes, located and present in multiple copies in the upstream region of the CYP1A1 gene promoter. The AhR-regulated genes include CYP1A1, 1A2, 1B1, 2S1, UGT1A1, 1A6, and GSTA1 (32). The regulation of expression of CYP1A1/1A2 is complex because gene transcription not only involves the AhR but also a number of transcription factors, and is potentially influenced by the actions of transcriptional coactivators and corepressors. AhR-mediated signaling pathways provide a first line of defense against potentially toxic environmental contaminants. However, induction of metabolic processes by the AhR can also produce highly carcinogenic metabolites, creating a link between AhR activation and chemical carcinogenesis.

Fig. 3.

A schematic illustration of the aromatic hydrocarbon receptor (AhR)-mediated induction of phase I and phase II drug metabolizing enzymes and drug transporters such as human CYP1A1, 1B1, 1A2, and 2S1, UGT1A1 and 1A6, and MDR1/ABCB1. The AhR exists as cytoplasmic aggregates bound to two 90-kDa heat-shock proteins (Hsps), the cochaperone p23 and a 43-kDa immunophilin-like protein hepatitis B virus X-associated protein 2 (XAP2). Upon binding a ligand, after the replacement of its associated molecule with AhR nuclear translocator (Arnt) to form a heterodimer with release of 90-kDa heat-shock proteins, AhR translocates into the nucleus. This heterodimer interacts with a 5′-GCGTG-3′ DNA sequence, the core binding motif of the xenobiotic response element (XRE) or dioxin response element of the target genes encoding CYP1A1, 1B1, 1A2, and 2S1, and uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1) and UGT1A6

Induction of CYP1A1/1A2 is generally a means of maintaining the homeostasis of the chemical environment in cells by increasing the metabolic clearance of substrates. Since CYP1A1/1A2 catalyzes the metabolic activation of PAHs and heterocyclic aromatic amines/amides to ultimate carcinogens, it is expected that induction of the enzyme is detrimental in humans exposed to high levels of PAHs and heterocyclic aromatic amines/amides such as by cigarette smoking. Induction of the enzyme in humans exhibits large variations; high inducibility may impose additional risk for lung cancer to individuals who are smokers. Furthermore, CYP1A2 can metabolize a range of substrates; induction of the enzymes by one substrate may increase the metabolism of other chemicals (for instance, clinical drugs), resulting in unexpected drug–drug interactions.

INHIBITION OF CYP1A2 AND CLINICAL RELEVANCE

Several therapeutic drugs including carbamazepine, dihydralazine, furafylline, isoniazid, rofecoxib, clorgyline, thiabendazole, and zileuton are mechanism-based inhibitors of CYP1A2 (17). In addition, desethylamiodarone also inactivated CYP1A2 but amiodarone inactivated CYP3A4 only (33). The metabolite of thiabendazole, 5-hydroxythiabendazole, is also a mechanism-based inhibitor of CYP1A2 (34). Antofloxacin, an 8-NH₂ derivative of levofloxacin, was found to be a mechanism-based inhibitor of rat CYP1A2 (35). Oltipraz, a chemo-protective agent, is a competitive and mechanism-based inhibitor of CYP1A2 (36). trans-Resveratrol inactivates CYP1A2, but not CYP1A1 (37).

Typical competitive CYP1A2 inhibitors are relatively small molecules often containing methyl, chloro, or fluoro substitutions such as tolfenamic acid, ciprofloxacin, furafylline, and fluvoxamine. These compounds can fit well into the active site of CYP1A2. ANF is commonly used as selective inhibitors for CYP1A2 in reaction-phenotyping studies (38,39). ANF is a potent competitive inhibitor of CYP1A2 with K_i values of 1–50 nM (38,39). Fluvoxamine, a selective serotonin reuptake inhibitor (SSRI), was a very potent and selective CYP1A2 inhibitor with K_i of 0.035–0.24 μM (40). Other SSRIs, including fluoxetine, norfluoxetine, sertraline, and paroxetine, also inhibited CYP1A2-mediated 7-ethoxyresorufin O-deethylase activity (41). The fluoroquinolone antibiotics such as levofloxacin, ciprofloxacin, moxifloxacin, gatifloxacin, caderofloxacin, and antofloxacin are weak to moderate inhibitors of CYP1A2 (42–44).

Some natural compounds can inhibit CYP1A2 and 1A1. Rutaecarpine, evodiamine, and dehydroevodiamine are quinazolinocarboline alkaloids isolated from Evodia rutaecarpa, which has been used in traditional Chinese medicine for the treatment of gastrointestinal disorder, headache, and hypertension. They are all inhibitors of CYP1A1 and 1A2, with rutaecarpine being the most potent (45). Tanshinone I, tanshinone IIA, and cryptotanshinone extracted from the commonly used Chinese herbal medicines Salvia miltiorrhiza (Danshen) are potent competitive inhibitors of CYP1A2, with K_i of 0.48, 1.0, and 0.45 μM, respectively (46). Because of the potential effects on drug disposition and inhibition of toxicological processes, natural products such as flavonoids are the focus of much current interest, from the perspectives of both nutrition and pharmacotherapy. These natural compounds may be involved in the prevention of malignant transformation by reducing the formation of carcinogens through inhibition of enzymes such as CYP1A1 and 1A2, both of which are known to be involved in carcinogen activation.

Overall, CYP1A2 is subject to reversible and/or irreversible inhibition by a number of drugs, natural substances, and other compounds. Competitive inhibitors of CYP1A2 can readily fit into the active site of CYP1A2. The inhibitory potency of compounds for CYP1A2 is determined by their physico-chemical features. It appears that planar molecules with a small volume to surface area are the most potent inhibitors of CYP1A2. Inactivation and reversible inhibition of CYP1A2 by drugs may cause important drug interactions.

PHENOTYPES OF CYP1A2 KNOCKOUT MICE

The Cyp1a2^−/− mice are viable and fertile; histologic examination of 15-day embryos, newborn pups, and 3-week-old mice revealed no abnormalities (47). Cyp1a1 and Cyp1b1 mRNA levels appeared unaffected by loss of the Cyp1a2 gene and inducibility of Cyp1a1 by dioxin was not altered by the absence of Cyp1a2 (48). When the muscle relaxant zoxazolamine (a known substrate for Cyp1a2 (49)) was administered to the Cyp1a2^−/− mice, they exhibited dramatically lengthened paralysis times relative to the Cyp1a2^+/+ wild-type animals, and the Cyp1a2^+/− heterozygous mice showed an intermediate effect. Using Cyp1a2^−/− mice, it was found that 87% of caffeine clearance depended on Cyp1a2 in this species (50). Notably, Cyp1a2^−/− mice exhibit increased toxicity from some compounds that are predominantly Cyp1a2 substrates. By cDNA microarray approach, it has been found that at least 15 genes are up- or down-regulated (51). The gene exhibiting the greatest down-regulation was insulin-like growth factor binding protein-1, showing only 12% expression of that in the wild-type mouse. These altered genes are associated with cell-cycle control, insulin action, lipogenesis, and fatty acid and cholesterol biosynthetic pathways. Histologically, the Cyp1a2^−/− mouse exhibited an approximately 50% decrease in lipid stored in hepatocytes, and 50% increase in lipid present in interstitial fat-storing cells compared with that in the wild-type (52). These data suggest that CYP1A2 might have additional hepatic endogenous functions heretofore not revealed.

POLYMORPHISMS OF CYP1A2 AND CLINICAL IMPACT

There are wide interindividual differences (10- to 200-fold) in CYP1A2 (also called phenacetin O-deethylase) expression and activity (7). Approximately 15- and 40-fold interindividual variations in CYP1A2 mRNA and protein expression levels have been observed in human livers (2). These findings may reflect a genetically determined difference in constitutive and/or inducible CYP1A2 gene expression. Unimodal, bimodal, and trimodal distributions of CYP1A2 activity when measured by caffeine urinary metabolic ratios have been observed in different study populations (10). The frequency of poor metabolizers in non-smokers was 5% in Australians, 14% in Japanese, and 5% in Chinese (8). There is also marked racial difference in CYP1A2 activity. Swedes had a 1.54-fold higher CYP1A2 activity than Koreans (53). A lower CYP1A2 activity has been found in Asian and African populations compared to Caucasians (54). Environmental factors have been thought to influence the interindividual differences in CYP1A2 activity and expression. Cigarette smoking and intake of oral contraceptive steroids are well-established inducers of CYP1A2 activity. However, it has been suggested that approximately 35% to 75% of the interindividual variability in CYP1A2 activity is due to genetic factors (55). There is a large variability in the clearance of drugs that are mainly metabolized by CYP1A2. Identification of genetic, epigenetic, and environmental factors that regulate CYP1A2 expression and activity can assist in the optimization of therapeutic regimens of drugs that are mainly eliminated by CYP1A2.

To date, more than 15 variant alleles (*1B to *16) and a series of subvariants of the CYP1A2 gene have been identified (Fig. 4; http://www.imm.ki.se/CYPalleles, access date 12 June 2009), and 177 SNPs have been found in the CYP1A2 upstream sequence, six introns and seven exons in NCBI dbSNP (http://www.ncbi.nlm.nih.gov/, access date 12 June 2009). CYP1A2*1A is referred to as the wild type (reference). Among the SNPs located in seven exons, there are 22 non-synonymous that change amino acid sequence.

Fig. 4.

Important alleles of CYP1A2 in humans

In addition, there are several SNPs identified in the 5′-flanking region and intron 1 of CYP1A2 (56). Aitchison et al. (56) identified several SNPs in the 5′-flanking region of CYP1A2, including −3594T>G, −3598G>T, and −3605insT. The −3594T>G SNP had a 1.72% frequency in 174 Caucasians. In the 5′-flanking region, reported SNPs in Japanese include −3860G>A, −3598G>T, −3594T>G, −3113G>A, −3053A>G, −2847T>C, 2808A>C, 2667T>G, and −2467delT (57,58). In intron 1, SNPs including −739T>G, −729T>G, −733G>C, −587A>G, −367C>T, −163C>A, and −99C>T have been reported. −163C>A in intron 1 is present with considerable frequencies in most populations examined. −3860G>A, −3113A>G, −2464delT, −739T>G, and −729C>T have been found with different frequencies in various ethnic groups (57).

The most extensively studied polymorphisms are −3860G>A (CYP1A2*1C), −2467delT (*1D), −739T>G (*1E), and −163C>A located in intron 1 (*1F), which were first reported in a Japanese population. Of the polymorphic CYP1A2 alleles showing variability in the promoter region, CYP1A2*1C, *1D, *1F, and *1K have been associated with altered enzyme activity. The CYP1A2*1C allele contains −3860G>A in the 5′-flanking region of CYP1A2, which was first found in Japanese with a frequency of 23.3% (27/116) (59). This allele was reported to cause decreased enzyme activity as measured by the rate of caffeine 3-demethylation in smokers of Japanese (n = 50), probably due to decreased inducibility and expression of the enzyme (59). The CYP1A2*1F allele (i.e., −163C>A in intron 1) is common with high and comparable frequencies in various population studies (58,60–62). Sachse et al. (60) first reported this SNP in a healthy German population comprising 185 non-smokers and 51 smokers (n = 236). The frequency of this SNP in German Caucasians was 67.8%, with 45.8% being homozygotes (108/236) for −163A/A and 44.1% being heterozygotes (104/236) for −163C/A. In Japanese, the frequency of −163C>A is high (62.8%) (57). This SNP had a frequency of 55.9% in Swedes (61). The −163C>A in intron 1 caused increased enzyme activity when measured by molar ratios of 17X/137X in plasma compared with the A/C and C/C genotypes in 51 German-Caucasian smokers (10). Smokers (n = 51) with the −163C/C (n = 5) or −163C/A (n = 24) genotype had a 67.1% and 55.7% lower CYP1A2 activity, respectively, compared with those with the −163A/A genotype (n = 22) (0.82 and 0.88 vs 1.37; p = 0.008).

CYP1A2*1J (−163C>A; −739T>G) and *1K (−163C>A; −739T>G; −729C>T, all located in intron 1) have been first detected in Ethiopian non-smokers (63). The −739T>G had a frequency of 3.2% in Japanese, while the −729C>T SNP was not found in the Japanese population (57). The CYP1A2*1K haplotype was associated with 40% lower inducibility in vitro, and non-smokers heterozygous for *1K had significantly lower CYP1A2 activity compared with the wild type (63). The −729C>T SNP, which was responsible for a lower activity of *1K, abolishes a binding site for an Ets nuclear factor, resulting in highly decreased CYP1A2 expression and caffeine metabolism (63). CYP1A2*1K was found to be rare in Swedes (0.3%) and absent in Koreans (61). Thus, the effect of CYP1A2*1K could not be shown in these populations.

CYP1A2*2 carries a 63C>G mutation that causes a Phe21Leu substitution, which was first detected from one subject out of 157 Chinese subjects with an allele frequency of 0.32% (64). Its functional impact is unknown. The *2 allele was absent in British (n = 114) (65) and Italian populations (n = 500) (66). The CYP1A2*3 contains 2385G>A in exon 4 causing Asp348Asn and the synonymous 5347T>G (Asn516Asn), with a frequency of 1% in French population (67). The *4 allele contains 2499A>T in exon 5 leading to Ile386Phe. The *5 allele entails 3497G>A in exon 6 leading to Cys406Tyr. The *6 contains 5090C>T in exon 7 resulting in Arg431Trp (67). All these alleles (*3–*6) were first detected in a French population with very low frequencies (≤0.5–1.0%). CYP1A2*7 contains a 3533G>A mutation at the splice donor site of intron 6. This mutation was found in a 70-year-old patient who had very high plasma concentrations of clozapine when administered at normal dose. This SNP caused RNA splicing defect and led to loss of CYP1A2 activity.

There are significant ethnic differences in the distribution of common and rare CYP1A2 SNPs and haplotypes (Table III). The −3860G>A (*1C) SNP is less prevalent in Caucasians than in Asians (0.21–0.25) (58). The frequency of *1C was significantly lower in the Turkish population (0.04) (68) than in Japanese (0.21) (58) and Chinese (0.25), while the frequency of *1C was relatively equal in the Turkish and Egyptian populations (0.07) (69). The frequency of the −2467delT (CYP1A2*1D) allele is lower in Caucasians compared with Asians and Africans. The allelic frequency of the *1D allele is 4.1–7.9% in Caucasians (65,70), while the −2467T deletion appeared at much higher frequencies of 42.0–43.8% in Japanese (57,58) and of 40% in Egyptians (69). In Turkish, the frequency of *1D is very high (92%) (68). The −739T>G SNP (CYP1A2*1E and *1G) was frequent in Ethiopians (0.10) (63), Saudi Arabians (0.096) (63), and Japanese (0.082) (58). However, British (65), German-Caucasian (70), Spaniard (63), Turkish (68), and Egyptian (69) populations had a low frequency for this allele (0.0044, 0.016, 0.017, 0.01, and 0.03, respectively).

Table III.

Frequencies of Important SNPs in Human CYP1A2 Gene

Ethnic group	No. of subjects	−3860G>A (*1C)	−3113A>G	−2467delT (*1D)	−739T>G (*1E, *1G, *1J, and *1K)	−729C>T (*1K)	−163C>A (*1F, *1J, and*1K)	5347T>C (*1B, *1H, *1G, *3, *8, *15, and *16)	Reference
Caucasian
British	114	0.009		0.048	0.004		0.333	0.382	(65)
German	236						0.32		(60)
Italian	95	0.04		0.24			0.33		(93)
Spanish	117				0.02	0.005			(63)
Swedish	194	0.008	0.023	0.193	0.023	0.003	0.286		(61)
Swedish	1,170						0.29		(94)
Turkish	110	0.04		0.92	0.01		0.27		(68)
Asian
Chinese	27	0.22	0.09	0.50	0.09		0.30	0.20	(95)
Chinese	168						0.33		(96)
Japanese	159	0.21		0.42	0.08		0.61		(58)
Japanese	250	0.24	0.03	0.44	0.03		0.32	0.19	(57)
Korean	150	0.267	0.027	0.707	0.027	0	0.373		(61)
Saudi Arabian	173				0.10	0.04			(63)
African
Egyptian	212	0.07		0.40	0.03		0.68		(69)
Ethiopian	173				0.10	0.03	0.40		(63)
Tanzanian	71						0.51		(97)
Tunisian	98	0.07		0.075	0.13		0.44		(98)
Zimbabwean	143						0.43		(97)

A number of clinical studies have been conducted to examine the impact of CYP1A2 polymorphisms on drug clearance and drug response. For example, Obase et al. (71) examined the effect of genetic polymorphisms in the 5′-flanking region to intron 1 of the CYP1A2 gene on theophylline metabolism in Japanese asthmatic patients (n = 75) and healthy volunteers (n = 159). Among asthmatic patients, theophylline clearance was significantly lower in patients with the polymorphism at site −2964G>A whose genotype was G/A or A/A than in those whose genotype was G/G (G/A or A/A vs G/G = 0.029 vs 0.034 L/h/kg). Therapeutic drug monitoring may be needed in patients with the A allele at site −2964 in the CYP1A2 gene, because theophylline clearance is lower in these patients, particularly in young asthmatic individuals. Resistance to clozapine therapy due to low plasma drug levels has been reported in smoking schizophrenic patients harboring the −163A/A (CYP1A2*1F) genotype (72,73). Eap et al. (72) reported four smoking patients who did not respond to clozapine therapy at usual dosage and found these individuals were ultra-rapid metabolizers of CYP1A2 carrying the *1F allele. On the other hand, higher plasma concentrations of clozapine and its metabolite N-desmethylclozapine have been observed in patients carrying two CYP1A2 variants associated with reduced enzyme activity (−3860A, −2467del, −163C, −739G, and/or −729T) compared with those with one or none (74).

Both smoking and CYP2D6 genotype, but not CYP2C9 genotype, can affect thioridazine plasma levels (75). The mean steady-state plasma concentration of trazodone was significantly lower in smokers than in non-smokers (76). Subjects with −163A/A or −163C/A genotype had lower plasma melatonin concentrations after administration of 6 mg melatonin compared with −163C/C carriers (77). A recent study in patients with rheumatoid arthritis (n = 105) revealed that the CYP1A2*1F (−163C>A) allele affected the risk for leflunomide-induced organ toxicities (78). Patients with −163C/C genotype had a 9.7-fold higher risk for overall leflunomide-induced toxicity compared with patients with the CYP1A2*1F C/A or A/A genotype.

CONCLUSIONS AND FUTURE PERSPECTIVES

CYP1A2 is one of the major hepatic CYPs in human liver that metabolizes about 15% of clinical drugs such as clozapine, theophylline, tacrine, and zolmitriptan. CYP1A2 is one of the major enzymes that bioactivate a number of procarcinogens. This enzyme also metabolizes several important endogenous compounds such as steroids, retinols, melatonin, uroporphyrinogen, and arachidonic acids, suggesting its potential role in some physiologic processes in addition to xenobiotic metabolism. Because CYP1A2 has a relatively small active center due to several surrounding aromatic residues, its substrates and inhibitors are usually small, lipophilic, and planar molecules. The elucidation of the CYP1A2 crystal structure has provided deep insights into the mechanism for the interaction of ligands with CYP1A2. The structure has offered clear clues on how CYP1A2 exhibits its substrate specificity. This information may help with the design of new drugs with minimal interaction with CYP1A2.

There is an increasing number of clinical drug interactions caused by CYP1A2 induction and/or inhibition and timely identification of these drug interactions is important in clinical practice. The interaction may be substrate and inhibitor dependent and thus caution should be taken when conducting in vitro–in vivo extrapolation. Polymorphisms of CYP1A2 may alter drug clearance and drug response and thus dosage adjustment and therapeutic drug monitoring for drugs with a narrow therapeutic index may be required. There are increased pharmacogenetic studies on the impact of common CYP1A2 polymorphisms on drug clearance and response, but the evidence is still preliminary. The clearance of several drugs may be altered by the genotype of CYP1A2 and thus patients’ response to these drugs may vary considerably. More well-designed studies are warranted to explore the genotype–phenotype relationships of CYP1A2 in terms of drug clearance and response. Personalized pharmacotherapy and individualized dosing of drugs would need incorporation of both genetic and environmental factors.

The pharmaceutical industry is likely to screen drug candidates regarding their CYP1A2-binding properties early in drug development. Such candidates may be dropped if they have poor pharmacokinetic properties and where alternatives are available. This candidate selection might eventually lead to a less relevant role of this enzyme for newer drugs.