Distribution of genetic polymorphisms of genes encoding drug metabolizing enzymes & drug transporters - a review with Indian perspective

Phase I enzymes

Enzymes involved in phase I drug metabolism largely belong to the cytochrome P450 (CYP) super family of drug metabolizing enzymes which catalyze the reactions such as hydrolysis, oxidation and reduction in which the functional groups of a substrate are added or deleted. This CYP450 system is divided into 18 families and 44 subfamilies consisting of 57 genes and 58 pseudo genes. Among them, the oxidative metabolism of 90 per cent drugs has been controlled by CYP1, CYP2 and CYP3 subfamilies18. The allele frequency distribution of phase I enzymes in various Indian populations is described in Table I7891113192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263.

Table I Allele frequencies (%) of genes encoding phase I drug metabolizing enzymes (DME) in various Indian populations

CYP1A (CYP1A1 & CYP1A2): Human CYP1A belongs to the group of phase I DME CYP450 isoforms. It consists of two members, CYP1A1 and CYP1A2. These are the key enzymes in the biotransformation of estrogens, polycyclic aromatic hydrocarbons and aromatic amines. In addition, CYP1A2 is also involved in the metabolism of a number of therapeutic drugs including caffeine, clozapine, olanzapine, amitriptyline, R-warfarin, veramapil, theophylline, propranolol, clomipramine, imipramine, haloperidol and acetaminophen. The gene encoding CYP1A1 is located at 15q22-q24, spanning 5,810 bp with seven exons and six introns. While the gene encoding CYP1A2 is located at 15q22, extending 7.8 kb with six exons and so far, more than 150 variant alleles have been described worldwide646566. Polymorphisms in these genes lead to variability in the enzyme activity and is shown to be associated with various cancers such as colon, ovarian, breast, lung, oral and acute lymphoblastic leukaemia (ALL). The most common being CYP1A1*2A or m1 (3798T>C), also known as Msp1, CYP1A1*2C or m2 (2455A>G), CYP1A1*3 or m3 (3204T>C) and CYP1A1*4 or m4 (2452C>A) polymorphisms for CYP1A1. For CYP1A2, CYP1A2*1F (163C>A) is the most common variant. The frequency of the mutant alleles CYP1A1*2A and CYP1A1*2C were significantly (P <0.05) different among the Indian populations as well as in comparison with other populations (P <0.0001)919202122232425262728293031323334353637386768697071. CYP1A1*3 was absent in Indians, while CYP1A1*2B (9.5%) was determined only in NEI35, which was significantly different from other populations (P <0.0001). The CYP1A1*4 frequency in NEI35 was different from NI, Caucasians and Africans but absent in Asians (P <0.0001). Similarly, CYP1A1*4 frequency in NI was in agreement with Caucasians but different from Africans (P <0.0001). Conversely, the frequency of CYP1A2 polymorphisms is available only in NI population. The distribution of CYP1A2 alleles were 8, 24.5, 9.5, 50.7 and 0 for *1C, *1D, *1E, *1F & *2, respectively. Significant interethnic differences were observed between NI and other major populations, P <0.001 (Table II)6768697071727374.

Table II Summary of normative allele frequency distribution of genes encoding major human Phase I and II drug metabolizing enzymes (DME) and drug transporters in different ethnics

CYP2A6: Human CYP2A6 is the major hepatic CYP2A enzyme and its role in the metabolism of drugs is small (3%) but it is important in the oxidative metabolism of nicotine. In addition, it is involved in the catalytic metabolism of valproic acid, fadrozole, methoxyflurane, artesunate, coumarin, disulfiram, halothane, losigamone, tegafur and letrozole. It also metabolizes environmental toxins, procarcinogens, retinoic acids and steroids75. The gene encoding CYP2A6 spans about 6kb with 9 exons and is mapped on chromosome 19q13.2 with other CYP2A sub-family (CYP2A7 and CYP2A13) members. Importance of CYP2A6 had risen considerably after the finding of a relationship between defective CYP2A6 alleles, smoking behaviour and cigarette consumption, drug clearance as well as tobacco related cancer risk. The CYP2A6 gene is extremely polymorphic and up to now, more than 81 variant alleles have been known7566. Besides the normal type allele designated as CYP2A6*1, three non-functional alleles which result in absence of CYP2A6 activity *2 (1799T>A, Leu160His), *4A (gene deletion), and *5 (1436G>T and 6582G>T, Gly479Val) have been well known. The frequency data of CYP2A6*1B polymorphisms are available only for NI43 and similarly CYP2A6*2 and *5 are available only in SI populations13. Among the CYP2A6 variants, *1B is the most prevalent allele followed by *4, whereas *2 and *5 were found to be rare (Table I). The comparison of CYP2A6*1B between NI (32.7%) and other populations indicates similarity with Caucasians (27.6%) and significant difference with Africans 11.2 per cent and Asians 42.8 per cent (P <0.001). The prevalence of the defective allele CYP2A6*4A was higher in NI (11.3%), SI (8.9%) and Asians (11%) but significantly lower in Africans (0.5%) and Caucasians 3 per cent (P <0.0001). In contrast, the other variant CYP2A6*2 was lower in SI (1%), Africans (0.3%), Caucasians (2.3%) and absent in Asians. Similarly, CYP2A6*5 were lower in SI (0.7%), Asians (0.5%) and virtually absent in Africans and Caucasians (Table II)13767778.

CYP2C8: The enzyme CYP2C8 comprises approximately 7 per cent of the total hepatic cytochrome system, which metabolizes around 5 per cent of drugs in phase I metabolism64. CYP2C8 plays significant role in metabolizing antidiabetics (troglitazone, pioglitazone, rosiglitazone & repaglinide), anticancer (paclitaxel), antihypertensive (veramapil), non-steroidal anti-inflammatory drugs (NSAID) (ibuprofen), 3-hydroxy-3-methylglutaryl-coenzyme (HMG-CoA) reductase inhibitor (cerivastatin), antimalarial (chloroquine, amodiaquine) and antiarrhythmic (amiodarone) drugs. Additionally, it is the principal enzyme responsible for the metabolism of retinoic and arachidonic acid in the kidneys79. The human CYP2C8 gene which encodes CYP2C8 enzyme is localized in the long (q24.1) arm of the chromosome 10, consists of 9 exons and spans about 31 kilo base. CYP2C8 gene exhibits several polymorphisms and so far, more than 20 polymorphisms have been reported in different populations66. Among them, the most commonly studied SNPs that lead to decreased enzyme activity are CYP2C8*2 in exon 5 (805A>T, I269F), CYP2C8*3 in exons 3 and 8 (416G>A/1196A>G, R139K/K399R), CYP2C8*4 in exon 5 (792C>G, I264M) and CYP2C8*5 (475delA). The enzymes encoded by these variant alleles impair the metabolism of several drug substrates of CYP2C8. Subjects who are homozygous (*2/*2 or *3/*3) have lower intrinsic clearance of CYP2C8 substrates than those who are heterozygous (*1/*2 or *1/*3). Further, the polymorphisms of CYP2C8 have been associated with diseases such as myocardial infarction (MI)80. The only study which determined the frequency distribution of CYP2C8 alleles in any of the Indian population was carried out in SI Tamilians11 with a frequency of about 0.8 per cent (95% CI 0.2-2.1) and 2.9 per cent (95% CI 1.6-4.7) for CYP2C8*2 and *3, respectively (Table I). Population studies from around the world in different populations indicate that CYP2C8*2 variant is rare in SI and Caucasians, whereas it is a common variant among Africans. Allele CYP2C8*3 is a common variant in Caucasians while it occurs either at small frequency or not present in SI and Africans. With regard to Asians subjects, the CYP2C8 polymorphism was monomorphic for CYP2C8*1 allele (Table II)8183.

CYP2C9: Drugs undergoing oxidative metabolism such as S-warfarin, rosiglitazone, tolbutamide, phenytoin, glyburide, glibenclamide, glimepiride, glipizide, losartan, irbesartan, torsemide, tamoxifen, fluvastatin, fluoxetine, amitriptyline and other commonly used anti-inflammatory drugs (diclofenac, ibuprofen, naproxen, piroxicam, aceclofenac, celecoxib) have been described to be principally metabolized by CYP2C9848586. It is encoded by the gene CYP2C9, which is located on chr10q24.2. Among the CYP2C isoforms, CYP2C9 is the most abundant one and constitutes about 20 per cent of total CYP450 hepatic content85. The gene CYP2C9 is known to be polymorphic and inter-individual differences in the enzyme activity of CYP2C9 categorize subjects into EMs, IMs and PMs. Currently, 41 different variant alleles are known for CYP2C9 gene66. Among the many variants, CYP2C9*2 (430C>T, Arg144Cys) in exon 3 and CYP2C9*3 (1075A>C, Ile359Leu) in exon 7 are the most characterized alleles and individuals with these variant alleles are reported to have decreased CYP2C9 activity. These variants are reported to be associated with acute MI, hypertension, colorectal cancer and major depressive disorders; also with certain adverse drug reactions including gingival hyperplasia, hypoglycaemia and gastrointestinal bleeding. In addition, a number of other alleles such as CYP2C9*5 (1080C>G, D360E), *6 (818Adel), *9 (752A>G, H251R), *11 (1003C>T, R335W) and *12 (1465C>T, P489S) may cause impaired metabolism, which may give rise to life threatening drug toxicity and may have impact on drug metabolism and disease susceptibility646684. The prevalence of CYP2C9*2 was significantly different between NI and SI (9% vs. 3.6%, P <0.0001)744454647, whereas CYP2C9*3 (9.7% vs. 8%) allele was equally distributed among the Indian populations (Table I). With regard to CYP2C9*2, NI showed similarity with Caucasians but difference with Africans (P <0.0001)444546478788. In contrast, the SI showed similarity with Africans but different from Caucasians (P <0.0001). Indians had higher frequency of CYP2C9*3, as compared to Africans and Asians (P <0.0001) but no difference was observed with Caucasians (Table II)878889.

CYP2C19: CYP2C19 is a polymorphically expressed CYP450 enzyme, constitutes about 16 per cent of the CYP2C family in liver and it is encoded by the gene CYP2C19 located on chr10q24. It is involved in the metabolism of a broad range of clinically important drugs including antimalarial (proguanil), oral anticoagulants R-warfarin, anti-epileptics (S-mephenytoin, diazepam, phenobarbitone), antivirals (nelfinavir), antiplatelets (clopidogrel), chemotherapeutic agents (cyclophosphamide), proton pump inhibitors (omeprazole, pantoprazole, lansoprazole, rabeprazole) as well as several antidepressants (amitriptyline, clomipramine)6490. Upto now, 35 polymorphisms have been identified. The frequencies of these alleles in different ethnic populations were extremely variable. Carriers of the non-functional alleles such as CYP2C19*2 in exon 5 (681G>A) and CYP2C19*3 in exon 4 (636G>A) have diminished ability to metabolize therapeutic agents that are substrates of CYP2C196684. The CYP2C19*2 allele notably occurred at a higher frequency among the Indian populations (NI 33.1% and SI 36.8%) than Africans 16 per cent, Caucasians 13.3 per cent and slightly higher than Asians 28.4 per cent (P <0.001). The CYP2C19*3 alleles in Indians were 1.9 and 1.1 per cent in NI and SI, respectively (Table I). Marked inter- and intra-ethnic variations were observed among the Indian populations (P <0.05) and as compared to other major ethnics such as Asians and Caucasians (P <0.001) individually with regard to the distribution of the polymorphic allele CYP2C19*374449505152919293. On the other hand, a novel variant allele in the regulatory region defined as CYP2C19*17 (-806C>T, -340C>T) increases the activity of CYP2C19 protein resulting in ultrarapid metabolism of CYP2C19 substrates66. The ultrarapid metabolizer CYP2C19*17 was studied only in SI Tamilians14 with a frequency of about 19.2 per cent (Table I). The comparison of SI subjects with Asian individuals indicates significant difference in UM allele, P <0.0001 (Table II)91929394.

CYP2D6: Cytochrome P450 2D6 (CYP2D6) is responsible for the metabolism of clinically important drugs, namely anticancer, antiarrhythmic, antihistamines, antipsychotics, β-blockers, opioids, antihypertensives and antidepressants. Of the total CYP450 content, CYP2D6 constitutes approximately 2-4 per cent and involved in the elimination of 25 per cent currently prescribed drugs9596. The CYP2D6 gene consists of 9 exons and 8 introns and is located on chromosome 22q13.2. The sequence of CYP2D6 is highly polymorphic. Up till now, over 135 allelic variants have been reported and CYP2D6*2 (2850C>T;4180G>C) in exon 2 and 6, *3 (2549delA) in exon 5, *4 (1846G>A in intron 3 - exon 4 junction), *5 (whole gene deletion), *10 in exon 1(188C>T, P34S), *14 (1758G>A), *17 (1023C>T; 2850C>T) in exon 2 and *41 (2988G>A) variants were the most characterized alleles of CYP2D666. The frequencies of the defective alleles in different races vary widely. Apart from PMs, IMs individuals also had been classified as UMs of CYP2D6 substrates depending upon the presence of CYP2D6 allele combinations. PMs are those who carry two defective alleles (inactive enzyme), resulting in increased concentrations of the parent drug in plasma, whereas in the case of UMs, as a result of gene duplication individuals carry more than two copies of the functional gene leading to increased enzyme activity, resulting in decreased parent drug concentration in blood96.

The dosage recommendation is based upon the CYP2D6 genotype for drugs that are substrates of CYP2D6 in order to avoid both treatment failure and adverse drug reaction. Of the five variants which contribute to the loss of CYP2D6 activity (*3,*4, *5, *6 and *14), only *14 was not detected in Indian populations. Similarly, the other variants (*9, *17 and *29) leading to diminished enzyme activity were also not found in Indians. The functional allele CYP2D6*2 was most prevalent in Indians which was found at comparable frequencies between NI (29.3%) and SI (34.8%). Both the Indian populations showed significant difference compared with Asians and Caucasians (P <0.0001, (Table II)84453. Although the NI showed similarity with Asians, SI showed significant difference (P <0.0001). The CYP2D6*3 allele was found only in NI 9.2 per cent (95% CI 6.9-11.8) and it was absent in SI. Its frequency in NI was significantly higher as compared to Africans 0.4 per cent, Asians 1 per cent and Caucasians 2.8 per cent (P <0.0001)84445979899100. Likewise, the polymorphism CYP2D6*5 which leads to gene deletion was determined with a frequency of about 1.9 and 1.8 per cent in NI and SI, respectively. It was significantly lower than those observed in Africans 6.6 per cent, Asians 7 per cent and Caucasians 6.9 per cent (P <0.0001). On the other hand, CYP2D6*4 was evenly distributed in SI (7.3%), NEI (8.7%) and Africans (6.8%), while it was higher in NI (11.5%) and Caucasians (14%) and lower in Asians (4%), (P <0.0001) (Table II)830353744454654. The occurrence of the CYP2D6 allele *10 which confers partially decreased activity, was found at higher frequencies (27.2%) in NI than (10.2%) SI and the difference was significant (P <0.0001). With regard to CYP2D6*10, both the Indian populations showed significant difference when compared with other populations, P <0.001 (Table II)82344465455. The frequency of allele *41 showed high frequency in NI 12.5 per cent, Caucasians 8.5 per cent and lower in Asians 1.9 per cent (P <0.0001). The active gene duplication alleles of CYP2D6 were studied only in NI Gujarati population44. Alleles associated with ultra-rapid metabolism (1xN and 2xN) and loss of enzyme activity (4xN) were found at fewer frequencies in NI and major ethnics. On the contrary, abridged activity alleles (10xN and 41xN) were absent in NI (Table II)44979899100.

CYP2E1: CYP2E1 is toxicologically important enzyme, which is present mostly in liver, and at lower levels in several extrahepatic tissues. Its levels are increased during fasting, diabetes, and exposure to alcohol. It catalyses the bioactivation of several procarcinogens and protoxins including N-nitrosodimethylamine, styrene, benzene and N-alkylformamides and also chlorzoxazone, acetaminophen, and the volatile anaesthetics (enflurane, sevoflurane, halothane, methoxyflurane and isoflurane) drugs101. The gene encoding CYP2E1 protein is located on chr10q26.3 and until now 12 different SNPs have been reported66. The polymorphisms of CYP2E1 have been linked to many chemically induced cancers and to alcoholic liver disease, in particular liver cirrhosis. Any functional polymorphism of this enzyme might be an important factor in determining the relative risk of alcohol-mediated hepatotoxicity, cancer or susceptibility to drug toxicity102. Polymorphic alleles with mutation in intron 7 CYP2E1*1B, in the 5’flanking region CYP2E1*5B (C1/C2, Rsa I) and intron 6 CYP2E1*6 (C/D, Dra I) are the common variants associated with altered gene function and expression. Their prevalence has been related to the occurrence of alcoholic liver disease and lung cancer. The frequency of CYP2E1*1B was equally distributed in NI, SI and Caucasians, while it was significantly higher in Asians, 18.1 per cent (P <0.05) and relatively highest in Africans, 65.9 per cent (P <0.001)61356103104. With regard to CYP2E1*5B, similarity was observed between SI and NEI59 (1.3% vs. 0.8%) but different in NI, 8.4 per cent (P <0.001). Likewise, similarity was observed between NI and SI (17.7% vs. 22.2%) but different in NEI59, 0.8 per cent (P <0.001) with respect to CYP2E1*6 (Table II). The frequency of CYP2E1*5B in NI was different from other ethnic populations whereas *6 was similar to Asians but different from Africans and Caucasians (P <0.0001). The SI were different from other major populations (P <0.0001) for CYP2E1*5B but similar to Asians and different when compared with Africans and Caucasians for CYP2E1*6 (P <0.0001). The comparison of the frequency of CYP2E1*5B and *6 alleles in NEI with other ethnicities indicates significant dissimilarity (P <0.01)6103104.

CYP3A (CYP3A4 & CYP3A5): The CYP3A isoenzymes metabolizes about 50-60 per cent of all currently prescribed drugs. Its subfamily consists of homologous proteins encoded by four different CYP3A genes namely CYP3A4, CYP3A5, CYP3A7 and CYP3A43105. These are located adjacent to each other on chromosome 7q21. Among these, CYP3A4 is the largest portion of CYP3A protein present in the adult liver and its catalytic activity may show up to 90 fold variation. Till now, 41 CYP3A4 alleles have been identified and characterized66, Out of these *1B (4713G>A) is the only defining variant of CYP3A4 reported at present. Subjects carrying the defective alleles of CYP3A4 have been implicated in disease predisposition to prostate cancer, estrogen receptor negative breast cancer and type 2 diabetes mellitus106. The frequency of CYP3A4*1B was available in NI (1.2%) only and the comparison between NI and other populations indicates significant variations (P <0.01)6061. The other CYP3A4 variants such as *2, *4, *5, *6 and *10 were not detected in NI60 (Table I). On the contrary, CYP3A5 is polymorphically expressed in about 10-30 per cent adult livers and along with CYP3A4, metabolizes > 50 per cent of currently used drugs. It shares about 85 per cent of amino acid sequence identity with CYP3A4 but it has different degrees of catalytic activity and regioselectivity towards substrates105. Kuehl et al107 have demonstrated that at least one CYP3A5*1 allele is needed for expressing CYP3A5 protein. They have identified CYP3A5*3 (A to G at 6986) in intron 3 and CYP3A5*6 (G to A at 14690) in exon 7, which led to the absence of CYP3A5 protein 6986A allele (CYP3A5*1) was before correlated with high expression107. Further, two more variants in the coding regions viz.,*2 (27289C>A, Thr398Asn) in exon 12 and *4 (14665A>G, Gln200Arg) in exon 8 were identified. CYP3A5 may represent upto 50 per cent of the total hepatic CYP3A content in people expressing CYP3A5105106. This gene may be an important contributor to individual and inter-racial variation in CYP3A mediated metabolism of drugs including antipsychotics (olanzapine), antiestrogen (tamoxifen), anticancer (irinotecan, docetaxel, vincristine), antimalarial (mefloquine, artemether, lumefantrine), immunomodulators (tacrolimus, cyclosporine), antihistamines (chlorpheniramine, terfenadine, astemizole), antiplatelets (clopidogrel), antihypertensives (nifedipine, amlodipine, felodipine, verapamil), antivirals (indinavir, nelfinavir, ritonavir, saquinavir), HMG-CoA reductase inhibitors (atorvastatin, cerivastatin, lovastatin) antibiotics (clarithromycin) and steroids (testosterone, estradiol, progesterone and androstenedione).

The presence of non-functional polymorphic alleles of CYP3A5 has been associated with blood pressure, MI, breast cancer and acute myeloid leukemia (AML) or ALL106. In Indians, CYP3A5*3 is the only variant allele present in NI and SI. None of the other variants *2, *4 and *6 were identified in Indians (Table I). The frequency of CYP3A5*3 in SI 56 per cent (95% CI 53.5-58.5) was significantly different (P <0.0001), as compared to NI 68.2 per cent (95% CI 64.5-71.9). Similarly, both the Indian populations were statistically different compared with Caucasians 91.6 per cent, Asians 80.6 per cent and Africans 15 per cent (P <0.0001)13616263. Further, a promoter polymorphism -44A>G was observed only in NI45 34.7 per cent which was higher than Caucasians 9.2 per cent (P <0.0001) and similar to Asians 28.2 per cent (Table II)108109110111112113114115116117.

Phase II enzyme

Phase II enzymes are involved in sulphation, acetylation, conjugation and glucuronidation reactions which may lead to the excretion of drugs by increasing the hydrophilicity of the substrate or deactivation of highly reactive substrates. Glutathione S-transferases (GSTs), N-acetayltransferases 1 and 2 (NAT1 and NAT2), thiopurine S-methyltranferase (TPMT), uridine disphosphate glucoronosyl transferases (UGTs), catechol methyl trasferase (COMT) and sulphotransferases (SULT) are the main phase II enzymes118. Table III illustrates the allele frequency distribution of phase II enzymes in various Indian populations5612192027303233343536375356119120121122123124125126127128129130131132133134135.

Table III Distribution of genes encoding phase II drug metabolizing enzymes (DME) allele frequencies in different Indian populations

GST (GSTM1, GSTT1 & GSTP1): The polymorphic cytosolic glutathione S-transferase (GST) isoenzymes GSTM1, GSTT1 and GSTP1 are the predominant class of phase II drug metabolizing enzymes. The chromosome location of the genes encoding GSTM1 (mu), GSTT1 (theta) and GSTP1 (pi) are chr1p13.3, chr22q11.2 and chr11q13.2, respectively. They play a significant role in the biotransformation and detoxification of a wide range of xenobiotics and endogenous substances including carcinogens (halomethanes and methyl bromide). GSTs protect cells against reactive oxygen metabolites by the conjugation of glutathione with electrophilic compounds136137. Both the genetic polymorphisms and expression pattern of GST genes may have a major impact on cancer susceptibility, inter-individual variability in the prognosis, drug effects and toxicity137. Of all the human GSTs, GSTM1 and GSTT1 isoenzymes are highly polymorphic. Those who carry the respective null genotypes termed as GSTM1*0/*0 and GSTT1*0/*0 due to homozygous gene deletions do not have these enzymes. On the other hand, single base pair substitution 313A>G of GSTP1 (Ile105Val) in exon 5 marks reduced enzyme activity. In addition, individually or in combination the null genotypes of GSTM1 and GSTT1 genes increase the risk of gastric, colon, bladder and lung cancers137.

The frequencies of GSTM1 and GSTT1 null alleles were found to be 30.6 and 16.7 per cent in NI; 28.8 and 14.8 per cent in SI and 34.4 and 19.7 per cent in NEI519202729303234353637119120121122123124125. On the whole, a general uniformity was observed between NI vs. SI and NI vs. NEI with regard to GSTM1 and GSTT1 null polymorphisms. Conversely, it was significantly different between SI and NEI (P <0.04)53032343536124125. The frequency distribution of GSTM1 null variant in Indian populations was significantly different from Africans, Asians and Caucasians (P <0.03). Similarly, the frequency of GSTT1 null polymorphism in SI was in line with NI and different from NEI (P <0.01)5192729303234353637119120121123124125. Further, its frequency in Indian populations was significantly less prevalent as compared to Africans and Asians (P <0.001), but showed similarity with Caucasians (Table II). The GSTP1 105Val allele was equally distributed at a frequency of 23.8 and 21.7 per cent in NI and NEI, respectively (Table III). However, it was significantly different in SI 14.6 per cent (P <0.001), as compared to NI, NEI, Africans (36.8%), Asians (20.6%) and Caucasians (25.9%)629303637121124126127. Likewise, a significant difference was observed in Africans (P <0.0001) as compared to NI and NEI. As the distribution of GSTM1*0 (null), GSTT1*0 (null) and GSTP1 (105Val) alleles occurs at high frequencies in different populations, their genotyping becomes necessary (Table II)124138139140141.

UGT1A1: The uridine diphosphate glucuronosyl transferase1A1 (UGT1A1) is a major phase II drug metabolizing enzyme belonging to UGT1A family. It is of major importance in the conjugation and subsequent elimination of potentially toxic xenobiotics, carcinogens and drugs. UGT1A1 catalyzes glucoronidation of a variety of compounds including estrogens, bilirubin and therapeutic drugs142143. Genetic polymorphism has been described for six of the 16 functional human UGT genes characterized to date, namely UGT1A1, 1A6, 1A7, 2B4, 2B7 and 2B15142. Polymorphisms of UGT1A1 gene causes decrease in enzyme activity which ultimately results in interpatient differences in the pharmacokinetics of irinotecan, an anticancer drug143. Human UGT1A1 enzyme is encoded by UGT1A1 gene, which is located on chromosome 2q37 and spans about 160 kb143. To date, more than 50 variants in the promoter and coding regions of UGT1A1 gene are known to disrupt enzyme activity. The most common being UGT1A1 (TA) 6>7 (UGT1A1*28) polymorphism in the TATA element of the 5’promoter region. It is characterized by (TA) 7 repeats instead of more common (TA) 6 repeats, resulting in lower promoter activity. It is the genetic basis for several clinical conditions such as, mild unconjugated hyperbilirubinemia associated with reduced hepatic bilirubin glucuronidation (Gilbert syndrome) and irinotecan mediated toxicity142143. Irinotecan, an inhibitor of intracellular topoisomerase-I is metabolized to form active SN-38, which is further conjugated and detoxified by UGT1A1 enzyme143. The only study which determined the frequency of UGT1A1*28 in native Indians was carried out in SI12 namely, Andhra Pradesh 32.2 per cent, Karnataka 29 per cent, Kerala 46.5 per cent and Tamil Nadu 52.8 per cent (Table III). Significant differences in the frequencies of this variant were observed in SI 39.7 per cent, as compared to Africans 55.3 per cent, Asians 13.1 per cent and Caucasians 29.6 per cent, P <0.0001 (Table II)12.

UGT1A7: The uridine 5’-diphosphate-glucuronosyltransferase 1A7 (UGT1A7) is an extrahepatic phase II DME expressed in pancreas, lung, stomach, oropharynx and oesophagus. It catalyzes the glucoronidation and detoxification of a wide variety of endogenous and exogenous compounds such as bilirubin, drugs, steroid hormones, phenols, coumarin and environmental carcinogens including nitrosamines and benzo[a]pyrene present in tobacco142144. The gene UGT1A7 is highly polymorphic and localized on chromosome 2q37 and 11 missense variants have been characterized till now. The common being UGT1A7*1 (N129 R131 W208), UGT1A7*2 (K129-K131-W208), UGT1A7*3 (K129-K131-R208), and UGT1A7*4 (N129-R131-R208)144,145. Population epidemiological studies have suggested that individuals carrying UGT1A7 genetic variations which confer low detoxification activity are associated with the development of cancer of the liver, colon, oral, gastrointestinal, pancreas and chronic pancreatitis144. The frequencies of UGT1A7 gene polymorphisms were available only in NI128 which have not been studied in other Indian populations (Table III). Overall, the variant of UGT1A7 seems to be more common in NI. The UGT1A7* 2, *3 and *4 alleles shows significant interethnic variability as compared to Africans, Asians and Caucasians (P <0.01). In contrast, no interethnic differences were observed for UGT1A7*12 polymorphism (Table II)145146147148.

TPMT: The gene thiopurine S-methyltransferse (TPMT) encodes the cytosolic enzyme TPMT which catalyzes S-methylation of thiopurine drugs such as 6-mercaptopurine (6-MP), azathioprine (AZA) and 6-thioguanine (6-TG). These drugs are commonly used for the treatment of ALL, autoimmune disorders, dermatological conditions and as an immunosuppressant in graft transplantations149. Polymorphism of TPMT alters the enzymatic activity of TPMT which is a major factor influencing interindividual differences in toxicity and therapeutic efficacy of thiopurine drugs. The TPMT enzyme activity in erythrocytes shows trimodal distribution in Caucasians149. Individuals with TPMT*1/*1 wild type show high enzyme activity (89-94%), while carriers of heterozygotes and homozygotes of TPMT mutant alleles show intermediate (6-11%) and low enzyme activity (33%) respectively. In contrast, Southeast Asians have unimodal distribution, which is perhaps due to the absence or low frequency of TPMT*3A149.

The gene TPMT has 10 exons spreading over 27kb of genomic DNA on chromosome 6p22.3149. Till date, more than 24 polymorphisms have been reported for TPMT gene. Amongst these, TPMT*2 (238G>C),*3A (460G>A and 719A>G),*3B (460G>A) and *3C (719A>G) are the four major variant alleles that cause 80-95 per cent intermediate and low enzyme activity. Carriers of theses alleles have been shown to have clinical implications with respect to metabolism, toxicity and therapeutic efficacy of thiopurine drugs. TPMT*2, *3A and *3B were rare, while TPMT*3C appears to be the most common variant in Indians. Another variant allele, the African specific TPMT*8 (644G>A) in exon 9 which is responsible for intermediate activity, was absent in Indians12. The mutant allele TPMT*3A which causes the largest decrease in TPMT activity was detected only in NI (0.4%) but it was absent in SI, Similarly TPMT*2 and TPMT*3B was present only in SI (0.1% and 0.1%) but not in NI (Table III). With regard to TPMT*3A, the NI were different from Caucasians (P <0.004) and similar to Asians and Africans12. With regard to TPMT*3C, the NI and SI were different from Caucasians, Africans (P <0.01) and similar to Asians12. On the whole, the Indians have relatively low frequency of TPMT variant and it shows that they have higher TPMT activity and are likely to be at a lower risk of developing toxicity when treated with thiopurine drugs compared to Africans, Asians and Caucasians12 (Table II).

SULT1A1: Human sulphotransferase 1A1 (SULT1A1) is the most widely expressed of the SULTs isoform which catalyzes the sulphate conjugation of hormones, neurotransmitters, drugs (tamoxifen) and other xenobiotics. Differences in the ability of the SULT1A1 protein to catalyze the sulphonation reaction lead to altered therapeutic efficacy, toxicity and disease process (carcinogenesis)144150,. The gene encoding SULT1A1 is mapped to the short arm of the chromosome 16p12.1-p11.2. The SULT1A1*2 variant was the most common of the two common non-synonymous SNPs observed namely SULT1A1*2 (638G>A, Arg213His) and SULT1A1*3 (667A>G, Met223Val). Individuals carrying the variant allele SULT1A1*2 will have diminished capacity to sulphate the substrates of SULT1A1 due to shorter protein life and more susceptible to cancer risk as well. Gene duplication and deletion was more common in SULT1A1 gene and a correlation between the enzymatic activity and SULT1A1 gene copy numbers has been observed by Hebbring et al151 in an in vitro study. The frequency of SULT1A1*2 in Indian populations was established in SI13033 (22.6%) and NEI36 (27.2%) but not in NI (Table III). It was significantly higher than those reported in Asians 8.7 per cent and lower than Caucasians 41.5 per cent (P < 0.0001), however, it was similar between SI, NEI and Africans 28.5 per cent (Table II)152153.

COMT: The enzyme catechol-O-methyltrasferase (COMT) is involved in the metabolism of catecholamines (adrenaline and noradrenaline), catecholestrogens, dopamine and their hydroxlated metabolites. The gene encoding COMT is located on chromosome 22q11.2 and produces two diverse proteins i.e. low affinity soluble COMT (S-COMT) and high affinity membrane bound (MB-COMT) by alternative transcription. Polymorphism in the human COMT is an important cause for the inter-individual variation in the enzyme activity of COMT. Further evidence has shown that the existence of COMT variants in individuals contributes to schizophrenia, breast cancer, endometrial cancer, Parkinson's disease, variation in pain sensitivity and therapeutic response (analgesics, levodopa)154155. However, the results were not consistent and to date, approximately more than 30 SNPs have been described for COMT gene154. Among these, the most common being the COMT rs4680 (472G>A) in exon 4, where the nucleotide change A to G results in the replacement of valine with methionine at codon 158 in MB-COMT and 108 in S-COMT leading to 3-4 fold decrease in methylation activity. Except the SNP rs4680, the frequency of the other reported alleles such as rs3788319, rs737865, rs6269, rs4818, rs4633, rs165599 were available only in SI133 (Table III). The frequency of rs4680 was significantly different between NI and SI (49 vs. 41.6%), P < 0.001131133. As compared to NI population, significant difference was obtained with Africans and Asians (P < 0.0001). The difference in the frequencies of COMT rs4680, rs737865 and rs4633 in SI were significant, as compared to Africans, Asians and Caucasians, P < 0.01 (Table II)156157158159.

NAT2: Human genome consists of two acrylamine N-acetyltransferase enzymes 1 (NAT1) and 2 (NAT2), which catalyze the metabolism of N-acetylation of arylamines, arylhydroxylamines and arylhydrazines160161. The genes encoding both the isoforms, NAT1 and NAT2 localized on chromosome 8p21.3-23.1 and exhibit numerous polymorphisms. Based on NAT genotype, individual phenotype (acetylation activity) can be divided into slow, intermediate and rapid acetylator160161,. Genetic variations of NAT1 were not determined in native Indians. The common defective alleles of NAT2 are 191G>A (Arg64Gln), 282C>T (Tyr94Tyr), 481C>T (Leu161Leu), 341T>C (Ile114Thr), 590G>A (Arg197Gln), 803A>G (Arg268Lys) and 857G>A (Gly286Glu)161. Among these, the most studied alleles were NAT2*5, NAT2*6 and NAT2*7 at positions 341, 590 and 857, respectively.

The variant NAT2*6 corresponding to inactive enzyme was detected with a frequency of 37 per cent in SI134 which was similar to 31.5 per cent in NI and different from NEI35 26 per cent, (P <0.003). Likewise, NAT2*7 accounts to 13.3 per cent in NI and it was similar to NEI (11.1%) but different from SI, 25 per cent (P <0.0001). The other variants NAT*5 and *13 were studied only in SI and occurred with a frequency of 30 and 44 per cent, respectively, while *14 was virtually absent134. The NAT gene polymorphic frequency of *11 was different between NI and SI (29.2 vs. 22%), whereas *12 was equally distributed among them (Table III). A comparison of NAT2 allele frequencies of Indians with Africans, Asians and Caucasians reveals significant interethnic difference (Table II)161. Carriers of these alleles will have variation in the metabolism of isoniazid, hydralazine, ribavirin, retigabine, sulphamethoxazole and may influence susceptibility to some cancers160161.

Drug transporters

Drug transporters are those proteins that carry either endogenous compounds or xenobiotics across the biological membranes. These play an important role in the uptake, bioavailability, efficacy, toxicity and clearance of drugs3. Generally, transporters that influence ADME of drugs are classified into (i) adenosine triphosphate (ATP)-binding cassette (ABC) family, and (ii) solute carrier (SLC) family162163. The ABC transporters are efflux transporters, consisting of seven subfamilies and 49 genes, whereas the SLC are influx transporters with 360 genes and 46 subfamilies. The allele frequency of drug transporter genes in different Indian populations is shown in Table IV101263164165166167168169170.

Table IV Frequency distribution of drug transporter genes in diverse Indian populations

MDR1: The multidrug resistance 1 gene MDR1, also known as ABCB1, is localized at chromosome 7q21.1, consisting of 29 exons ranging in size from 49 to 209 bp, encoding an mRNA of 4.5 kb. P-glycoprotein (P-gp), the product of MDR1 gene is a 170 kDa transmembrane protein, belongs to ABC super family of transporter proteins which is well recognized for its role in drug transport and chemoresistance. It protects the tissues from toxic xenobiotics and other endogenous substances by exporting the substrates from intracellular to extracellular space162171. The amount of expression, regulation and activity of P-gp influenced by MDR1 gene polymorphisms can directly affect the pharmacokinetics and pharmacodynamics of drugs that are substrates of P-gp, leading to inter-individual variation in drug response and toxicity171. The substrate specificity of P-gp is broad including clinically relevant agents, i.e. anti-neoplasmics (doxorubicin, actinomycin D, paclitaxel), antibiotics (erythromycin, levofloxacin, sparfloxacin, rifampicin), antihypersentives (losartan), antivirals (nelfinavir, indinavir, efavirenz), analgesics (morphine), antiepileptics (phenytoin, phenobarbital), antidepressants (amitriptyline), immunosuppressants (cyclosporin A, tacrolimus, rapamycin), antiarrthythmics (digoxin, verapamil), antilipidemic (atorvastatin) and steroids (aldosterone, cortisol, dexamethasone)162. In addition, the polymorphism exhibited by MDR1 gene is one of the factors responsible for individual susceptibility to various diseases such as breast cancer, colorectal cancer, Parkinson's disease and ulcerative colitis171.

Human MDR1 gene is highly polymorphic and over 50 SNPs have been identified. Among these, variants 2677G>T/A/C in exon 21, 3435C>T in exon 26 and 1236C>T in exon 12 are the most studied alleles of MDR1 gene171. MDR1 gene is extensively explored in Indian populations except NEI. In general, the Indians have at least one variant allele of MDR1. Among the Indian populations, the lowest 42.4 per cent and the highest 60 per cent frequency of 2677T/A allele were observed in NI and SI, respectively12. Similarly, the frequency distribution of the synonymous SNP 3435C>T was found to be 53.6 per cent in NI and 59.5 per cent in SI (Table IV). A significant inter- and intra-ethnic difference was observed when these two alleles were compared among Indians and with other major populations (Table II)172173174. On the other hand, the frequency of 1236C>T was available only in NI populations (51.9%) and it was different from other populations, P < 0.0001 (Table IV).

SLC22A1: SLC22A1 belongs to the solute carrier, SLC22, super family of transporters, and these are also known as organic cation transporter 1 (OCT1). These translocate a wide variety of endogenous substances, environmental toxins and therapeutic drugs of cationic nature163175. There are three important isoforms of OCTs namely - OCT1, OCT2 and OCT3 with similar membrane topology consisting of 12 transmembrane domains175. Of these, OCT1 is one of the most highly expressed transporters in the hepatocytes and plays a significant role in the hepatic uptake, elimination, distribution and renal transport of several xenobiotics including therapeutic agents (e.g. metformin, levodopa, amantadine, pramipexole and imatinib)176177178. Animal studies have shown that the concentration of metformin in liver was greatly decreased in OCT1 gene knockout mice than in mice with normal OCT1 transporter activity175176. The gene encoding human SLC22A1 is mapped onto chromosome 6q25.3, spanning 37kb with 11 exons. Numerous polymorphisms have been described for SLC22A1 gene in various populations leading to differences in transporter function175. Studies across the globe have evaluated the association between genetic variations of SLC22A1 gene and the pharmacokinetics and clinical consequences of metformin, levodopa, imatinib with inconsistent results10177178.

In Indians, the frequency of the three variants rs2282143 in exon 6 (Pro341Leu, 1022C>T), rs628031 in exon 7 (Met408Val, 1222A>G) and rs622342 (1386C>A) located in an intron between exon 8 and exon 9 were described in Tamilian healthy subjects10. No data are available in any of the other Indian populations. Genetic variants of OCT1 were common in SI Tamilians with a frequency of 8.9 per cent (95% CI 5.6-13.5), 80.3 per cent (95% CI 75.2-85.6) and 24.5 per cent (95% CI 18.9-30.2) for the alleles rs2282143 (T), rs628031 (G) and rs622342 (C), respectively10. It was different to the frequencies observed from those in Caucasians, P < 0.004 (Tables II and IV). However, the frequency was similar to Africans and other Asian populations for rs628031. With regard to rs2282143, similarity was observed for Africans but not with Asians (P < 0.05), and for rs622342, similarity was shown for Asians but not with Africans (P < 0.05)10. The available data indicate significant delineation in the allele frequencies of the OCT1 gene variants between different ethnic groups leading to dissimilarity in the pharmacokinetics and clinical consequences of the substrates of OCT1.

SLCO1B1: Organic anion transporting polypeptide 1B1 (OATP1B1) is a transmembrane protein with 12 domains, and plays important role in the hepatocellular uptake of various exogenous and endogenous substances of anionic nature including therapeutic drugs such as statins (simvastatin, rosuvastatin, pravastatin, and atorvastatin), methotrexate, repaglinide, irinotecan and rifampin179. It is encoded by the gene OATP1B1, also known as SLCO1B1, and consists of 14 coding and one non-coding exons and spans 108.59 kb, located on chromosome 12p12.2. Polymorphisms in the SLCO1B1 gene decreases the transporter function and ultimately influencing the pharmacokinetics, toxicity and efficacy of OATP1B1 substrates. The variants of SLCO1B1 increase the risk of statin-induced myopathy; methotrexate induced gastrointestinal toxicity and also susceptibility to gallstone disease169179180. The most common polymorphisms were SLCO1B1*1B in exon 5 (388A>G), SLCO1B1*4 (463C>A) and SLCO1B1*5 in exon 6 (521T>C). The prevalence of these variants were available only in NI169 population and its frequency was reported to be 45, 2.6 and 1.4 per cent for *1B, *4 and *5, respectively (Table IV). The comparison of *1B allele frequency showed significant difference with being higher in Africans 87 and Asians 64 per cent but lower in Caucasians 37 per cent (P < 0.001)169181. Similarly, the other variants (*4 and *5) were higher in other populations but Asians showed similarity with NI with regard to SLCO1B1*4 (Table II)181.