Influence of Single Nucleotide Polymorphisms on Rifampin Pharmacokinetics in Tuberculosis Patients

Rifampin (RF) is metabolized in the liver into an active metabolite 25-desacetylrifampin and excreted almost equally via biliary and renal routes. Various influx and efflux transporters influence RF disposition during hepatic uptake and biliary excretion. Evidence has also shown that Vitamin D deficiency (VDD) and Vitamin D receptor (VDR) polymorphisms are associated with tuberculosis (TB). Hence, genetic polymorphisms of metabolizing enzymes, drug transporters and/or their transcriptional regulators and VDR and its pathway regulators may affect the pharmacokinetics of RF. In this narrative review, we aim to identify literature that has explored the influence of single nucleotide polymorphisms (SNPs) of genes encoding drug transporters and their transcriptional regulators (SLCO1B1, ABCB1, PXR and CAR), metabolizing enzymes (CES1, CES2 and AADAC) and VDR and its pathway regulators (VDR, CYP27B1 and CYP24A1) on plasma RF concentrations in TB patients on antitubercular therapy. Available reports to date have shown that there is a lack of any association of ABCB1, PXR, CAR, CES1 and AADAC genetic variants with plasma concentrations of RF. Further evidence is required from a more comprehensive exploration of the association of SLCO1B1, CES2 and Vitamin D pathway gene variants with RF pharmacokinetics in distinct ethnic groups and a larger population to reach conclusive information.


Introduction
Rifampin (RF) was introduced as a part of the combinational chemotherapy regimen for tuberculosis (TB) during the 1960s. This has revolutionized TB treatment by reducing the duration of antitubercular therapy (ATT) and improving the cure rates [1,2]. The antimicrobial effect of RF on Mycobacterium tuberculosis and the development of RF resistance is concentration-dependent [3,4]. RF exhibits antimycobacterial action by arresting the DNA-directed RNA synthesis of Mycobacterium tuberculosis through interaction with the β subunit of RNA polymerase (RNAP) [5,6]. The primary mechanism of RF resistance is due to the mutations in the rpoB gene that encode for the β-subunit of RNA polymerase. The most common mutations in the rpoB gene are found in the rpoB 531, rpoB 526 and rpoB 516 codons of the RF resistance determining region (RRDR) [7,8]. Recent evidence has shown that higher doses of RF from the currently recommended dosage regimens resulted in better treatment outcomes in TB patients [9,10]. A recent comprehensive meta-analysis reported a wide range of interstudy heterogeneity in RF pharmacokinetic parameter estimates. Many variables such as

SLCO1B1
The organic anion transporting polypeptide 1B1 (OATP1B1) is a 691 amino acid protein expressed predominantly on the basolateral (sinusoidal) membrane of hepatocytes. It is encoded by the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene (spans 15 exons) located on chromosome 12. OATP1B1 is one of the major membrane influx transporters that regulate the active hepatic uptake of substrates from the bloodstream into the hepatocyte [16,32]. RF is a strong substrate of the OATP1B1 transporter protein [33,34]. Around 190 genomic variants with minor allele frequency higher than 5% were identified with the SLCO1B1 gene. Among these variants, rs4149056 and rs2306283 have been commonly identified and well-characterized [35].
The missense SNP rs4149056 located in exon 5 (also known as c.521T>C; with T allele defined as the wild-type allele and the C allele as a variant) causes a change of amino acid from valine to alanine at residue 174. This variant is reported to have reduced expression and activity of SLCO1B1 in vitro and in vivo. Hence, drugs that are substrates for OATP1B1 with c.521T>C may tend to have elevated plasma concentrations due to reduced uptake/transporter activity [36,37]. Allegra [38]. The frequency of SLCO1B1 rs4149056 genotype was reported to be 28.3%, 5.7%, 14.9% and 14.8% in Amerindian, African descent, Mulatto and Caucasian descent ethnic groups, respectively [39]. Mwinyi et al. reported a frequency of 15% and 12.2% in German and Turkish populations, respectively, whereas 15% prevalence was reported among the UK population for SLCO1B1 rs4149056 genotype [40,41]. rs2306283 (c.388A>G) is a missense SNP located in the exon 4 of the SLCO1B1 gene that causes a change of amino acid from asparagine to aspartic acid at amino acid position 130. The functional consequences of this variant reported by different in vitro and in vivo studies have yielded conflicting results and may be substrate-specific [36,42]. Dompreh 1) The genes whose SNPs were assessed with plasma RF concentrations in the review and (2) RF biotransformation.

SLCO1B1
The organic anion transporting polypeptide 1B1 (OATP1B1) is a 691 amino acid protein expressed predominantly on the basolateral (sinusoidal) membrane of hepatocytes. It is encoded by the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene (spans 15 exons) located on chromosome 12. OATP1B1 is one of the major membrane influx transporters that regulate the active hepatic uptake of substrates from the bloodstream into the hepatocyte [16,32]. RF is a strong substrate of the OATP1B1 transporter protein [33,34]. Around 190 genomic variants with minor allele frequency higher than 5% were identified with the SLCO1B1 gene. Among these variants, rs4149056 and rs2306283 have been commonly identified and well-characterized [35].
The missense SNP rs4149056 located in exon 5 (also known as c.521T>C; with T allele defined as the wild-type allele and the C allele as a variant) causes a change of amino acid from valine to alanine at residue 174. This variant is reported to have reduced expression and activity of SLCO1B1 in vitro and in vivo. Hence, drugs that are substrates for OATP1B1 with c.521T>C may tend to have elevated plasma concentrations due to reduced uptake/transporter activity [36,37]. Allegra et al. have reported higher plasma RF concentrations in TB patients with SLCO1B1 rs4149056 polymorphism. Multivariate linear regression analysis revealed that SLCO1B1 rs4149056 genotype was found to be a positive predictive factor for increased plasma RF trough concentration (C trough , p = 0.048, β = 0.345, 95% CI [6.458-1313.556]) and maximum concentration (C max , p = 0.019, β = 0.432, 95% CI [452.896-4571.730] at second week of ATT [38]. The frequency of SLCO1B1 rs4149056 genotype was reported to be 28.3%, 5.7%, 14.9% and 14.8% in Amerindian, African descent, Mulatto and Caucasian descent ethnic groups, respectively [39]. Mwinyi et al. reported a frequency of 15% and 12.2% in German and Turkish populations, respectively, whereas 15% prevalence was reported among the UK population for SLCO1B1 rs4149056 genotype [40,41]. rs2306283 (c.388A>G) is a missense SNP located in the exon 4 of the SLCO1B1 gene that causes a change of amino acid from asparagine to aspartic acid at amino acid position 130. The functional consequences of this variant reported by different in vitro and in vivo studies have yielded conflicting results and may be substrate-specific [36,42]. Dompreh et al. had reported that the SLCO1B1 rs2306283 polymorphism was associated with lower RF concentration in the pediatric TB population. Two patients (1.8%) with the SLCO1B1 *1b homozygous variant (AA genotype) had significantly lower RF C max (1.81 (0.81-2.80) µg/mL) and area under the time-concentration curve from 0 to 8 h (9.33 (2.35-16.31) µg*h/mL) and higher apparent oral clearance (44.54 (15.38-73.69) L/h) and apparent volume of distribution (109.23 (54.86-163.59) L) than did those with the wild type (GG genotype) in a pairwise analysis [43]. However, other studies have reported higher frequencies of the SLCO1B1 *1b homozygous variant (AA genotype) in Chilean (18.6%), Macedonian (33.1%) and Albanian (30.8%) population [44,45].
Chigutsa et al. and Gengiah et al. reported a high prevalence of SLCO1B1 rs4149032 (g.38664C>T), which is an intron 2 haplotype tagging SNP (tSNP). SLCO1B1 rs4149032 polymorphism was found to be associated with lower RF exposures in the African population suggesting the need for increasing the RF dose [46,47]. The functional consequences of SLCO1B1 rs4149032 on gene expression and on transporter activity are not yet known. Chigutsa et al. reported an allele frequency of 70% for the SLCO1B1 rs4149032 polymorphism in the South African pulmonary TB (PTB) patients. Patients who were heterozygous and homozygous for the rs4149032 polymorphism in this population had reductions in RF bioavailability by 18% and 28%, respectively. Simulations showed that SLCO1B1 rs41490932 carriers had a predicted reduction in C max of < 8 mg/L and an increase in the daily rifampin dose by 150 mg in the PTB patients in these population would help in achieving plasma concentrations similar to those of wild-type individuals [46]. Gengiah et al. reported an allele frequency of 76% for the SLCO1B1 rs4149032 polymorphism in the TB-HIV coinfected patients in South Africa. The median (IQR) RF concentrations at 2.5 h postdose were 3.4 (2.7-4.7) µg/mL, 3.7 (2.8-5.0) µg/mL and 5.3 (3.8-6.7) µg/mL for homozygous variant, heterozygous variant and wild type carriers of SLCO1B1 rs4149032 polymorphism, respectively, which was well below the recommended target range of 8 to 24 µg/mL [47]. Mukonzo et al. reported an allelic frequency of 66% for the SLCO1B1 rs4149032 polymorphism in the Ugandan population [48].
Lower RF exposures were reported with SLCO1B1 rs11045819 polymorphism in a study conducted by Weiner et al. [49]. SLCO1B1 rs11045819 (c.463 C>A) polymorphism is a missense variant, present on the exon 4 of the SLCO1B1 that cause a change of amino acid from proline to threonine at amino acid position 155 [49]. SLCO1B1 rs11045819 polymorphism was found to reduce the systemic exposure of the substrate for OATP1B1 transporter [50]. Weiner et al. reported the prevalence of SLCO1B1 rs11045819 polymorphism as 19% (n = 7) in African TB patients, 11% (n = 4) in TB patients of US and Spain and 25% (n = 4) among the healthy US population (controls). Patients with the SLCO1B1 rs11045819 variant allele (CA) had 42% lower RF exposure (25.6 µg*h/mL), 34% lower peak concentration levels (5 µg/mL) and 63% greater apparent oral clearance (22 L/h) compared to the wild type allele (CC) [49].
However, recent studies from the African population have not found any association with SLCO1B1 polymorphisms and RF exposures among TB patients [48,51,52]. Similarly, studies conducted by Ramesh et al. and Jeremiah et al. in the Indian and Tanzanian population, respectively, did not report any association of SLCO1B1 polymorphisms with plasma RF exposures (Table 1) [53,54]. The association of SLCO1B1 rs4149056, rs2306283, rs4149032 and rs11045819 polymorphisms with RF pharmacokinetics reported in certain studies were not replicated in other studies that can be attributed due to multiple factors such as lower sample population, ethnic variations, variations in the criteria and timings of sample collection, analytical variations and interindividual factors such as variations in body weight and medication adherence. Therefore, additional studies are warranted to characterize the functional consequences of SLCO1B1 rs4149056, rs2306283, rs4149032 and rs11045819 polymorphism on RF pharmacokinetics in other ethnic groups.

ABCB1
ABCB1 (or MDR1) gene is located on chromosome 7 and consists of 29 exons in a genomic region spanning 251.3 kb. It is one of many ABC genes that encode for the 1280 amino acid ABCB1 transporter protein (P-glycoprotein). P-glycoprotein (Pgp) is a multidomain integral membrane protein that utilizes the energy generated from the ATP hydrolysis to translocate solutes or ions from intracellular to extracellular membranes (efflux pump) in eukaryotes [56][57][58]. RF is a substrate of the Pgp efflux pump [59]. rs1128503, rs2032582 and rs1045642 are the most commonly found SNPs in the ABCB1 gene [60]. rs1128503 and rs1045642 are synonymous mutations, whereas rs2032582 is a missense mutation [61]. None of the studies were able to infer any association between ABCB1 polymorphisms and RF pharmacokinetics ( Table 2). These studies have explored the association of only a limited number of ABCB1 polymorphisms with the RF exposures. There are about 8643 single nucleotide variants (SNV) reported for the ABCB1 gene. The functional consequences of rare ABCB1 variants that may have a significant effect on drug pharmacokinetics have not been largely elucidated [58]. Hence, additional studies with other genetic variants are required to establish the impact of ABCB1 polymorphisms with the RF exposure.

PXR and CAR
PXR and the CAR are members of the group I of the subfamily 1 of nuclear receptors (NRs) that are involved in regulating the transcription of a wide range of drug-metabolizing enzymes and drug transporters genes [62,63]. RF is a substrate for SLCO1B1 and ABCB1 protein and the transcription of genes encoding these proteins are regulated by the PXR and CAR. Few studies have explored the possibility of association of the SNPs of these genes with the plasma RF levels. The PXR (or NR1I2) gene located on chromosome 3 and consisting of 9 exons encodes for the PXR [64]. rs2472677 and rs1523130 variants are present in the intron 1 and 5 UTR regions of the PXR gene, respectively. These regions represent the transcription factor binding sites of PXR regulatory regions [65,66]. The CAR (or NR1I3) gene located on chromosome 1 and consisting of 9 exons encodes for the CAR [67,68]. The rs2307424 variant is due to a synonymous substitution (c.540 C>T) in the CAR gene [69]. None of these SNPs in PXR and CAR affected RF exposures (Tables 3 and 4).

CES1 and CES2
RF is primarily metabolized to 25-desacetylrifampin by B-esterase [70]. B-esterases family comprises CES, acetylcholinesterase and butyrylcholinesterase enzymes [14]. Among these enzymes, CES exhibits broad substrate specificity and is involved in the metabolism of a wide range of endobiotic and xenobiotic compounds by hydrolyzing ester, thioester, amide and carbamate linkages. Human CES1 and human CES2 encoded by CES1 and CES2 gene, respectively, represent the two major isoenzymes of CES that are expressed in the liver [71]. Over the past decade, several CES1 and CES2 functional genetic variants associated with significant variations to various drug therapy responses have been reported. Hence, assessing the genetic polymorphisms of these genes with the pharmacokinetics of the substrate drugs becomes relevant [72]. The CES1 and CES2 genes are located on chromosome 16 and consist of 14 and 12 exons, respectively [73]. Sloan et al. reported that the rs12149368 variant present on the exon 1 (5'UTR) region of the CES1 gene does not affect the plasma RF concentration (Table 5)  The allelic frequencies for this variant were reported to be 0.33 in TB patients and 0.31 in controls and plasma RF concentrations were 8.9 ± 2.9 mg/L, 10.5 ± 3.1 mg/L and 13.9 ± 7.4 mg/L in homozygotes carrying major allele, heterozygotes and homozygotes carrying minor allele, respectively. Results of luciferase reporter analysis revealed that the change from A to G in CES2 c.-22263A>G variant was associated with a consistent decrease in luciferase activity, which may result in decreased RF metabolism and increased plasma RF concentration. [74]. However, Dompreh et al. did not find any changes in the RF exposures with the CES2 rs3759994 variant (Table 5) [43].

AADAC
AADAC is an enzyme expressed primarily in the human liver and intestine that causes the hydrolysis of many drugs [75]. Nakajima et al. reported that human AADAC was the enzyme responsible for the deacetylation of RF to 25-deacetylrifampin [15]. The AADAC rs1803155 and rs61733693 variants which are missense mutations did not affect any changes in the plasma RF concentrations (Table 6) [55].

Vitamin D Pathway Gene Polymorphisms
The Caudal-type homeobox protein 2 (Cdx2) gene variant found in the regulatory region, FokI variant in exon 2 and BsmI, TaqI and ApaI variants in the 3 end of the VDR gene were found to be associated with TB [76]. BsmI (rs1544410), FokI (rs10735810), TaqI (rs731236) and ApaI (rs7975232) represent the most commonly occurring SNPs of VDR gene [77]. At the fourth week of ATT, univariate regression analysis revealed that FokI TC/CC genotype had a negative predictor role on the plasma RF C trough (p = 0.694, β = −0.085, 95% CI [-1314.809-891.285]), possibly due to stronger transcription of the RF influx protein [38]. The FokI variant codes for a shorter 424 amino acid VDR protein isoform which shows a comparatively higher transcriptional activity by displaying enhanced interaction with transcription factor IIB [78]. Recently, Shaik et al. reported the frequencies of FokI TT, TC and CC genotypes to be 30.2%, 34.4% and 27.7%, respectively, in the Saudi Arabian population [79]. Reports from the Brazilian population revealed the frequencies of FokI TT, TC and CC genotypes to be 44.6%, 41.4% and 14%, respectively [80]. Calcagno et al. reported that the VDR regulatory region Cdx2 variant was not associated with any significant changes in the plasma RF concentration [52].
CYP27B1 and CYP24A1are two enzymes that are involved in the biotransformation of Vitamin D and play critical roles in governing the 1α,25-dihydroxyvitamin D 3 (1,25-(OH) 2 D 3 ) concentration. CYP27B1 gene encodes for the 1α-hydroxylase enzyme that is involved in the activation of 25-hydroxyvitamin D 3 (25-OH-D 3 ) to 1,25-(OH) 2 D 3 [81]. CYP24A1 is involved in catalyzing the C-23 and C-24 hydroxylation pathways of 25-OH-D 3 and 1,25-(OH) 2 D 3 [82]. Hence, genetic variants of these genes may alter the Vitamin D levels and may thereby render TB susceptibility as well as alter RF concentrations in plasma. Allegra et al. reported that the CYP24A1 rs927650 and CYP27B1 rs4646536 variants increased plasma RF concentrations which may probably be explained by the increased activation of Vitamin D, resulting in reduced RF elimination (Table 7) [38]. Multivariate linear regression analysis revealed that CYP27B1 rs4646536 variant (+ 2838C>T; CC/CT genotype) located at intron 6 was a positive factor for RF C max concentration ((p = 0.024, β = 0.416, 95% CI [469.172-5857.279]) at second week of ATT. Univariate linear regression analysis revealed that for the CYP24A1 rs927650 (22776C>T) variant located at intron 11, the homologous mutant profile (TT) is a positive predictor factor of RF C trough ((p = 0.924, β = −0.021, 95% CI [-1148.256-1055.303]) at fourth week of ATT [38]. The distribution of CYP27B1 rs4646536 TT, TC and CC genotypes were reported to be 45.7%, 40.4% and 13.9%, respectively, in healthy controls of Germany which were in near similar lines with a previously conducted study among 7435 healthy controls of UK [83,84]. The clear functional status of CYP27B1 rs4646536 is unknown. Intronic variants could influence gene expression by affecting the binding of transcription factors and mRNA splicing [85,86]. Hence, an allele variation of rs4646536 from C to T can cause abnormal expression of CYP27B1, resulting in the alteration of Vitamin D levels. CYP27B1 rs4646536 was associated with Vitamin D levels and Vitamin-D-related diseases [84]. The frequencies of CYP24A1 rs927650 TT, CT and CC genotypes were reported to be 26.1%, 49.7% and 24.2%, respectively, in the healthy controls of Germany and 21.3%, 50.8% and 27.9%, respectively, among type 1 diabetes German patients [83,87]. 1α,25(OH) 2 D 3 exhibit genomic actions that are mediated through the ligand-binding to the VDR, which forms a heterodimer with retinoid x receptor alpha (RXRα) and subsequently binds to Vitamin D response elements (VDRE) to either enhance or repress transcription of various genes [88]. The CYP24A1 gene has a significant role in 1,25(OH) 2 D 3 signaling as the promoter region of the CYP24A1 gene contains VDRE [89]. Polymorphisms in a VDRE of the CYP24A1 gene could reduce the receptor protein binding, transactivation and expression of the CYP24A1 gene in vivo [90]. A suggestive relationship between the CYP24A1 SNP rs927650 and concentrations of 25(OH)D was reported by Hibler et al. [91]. Further research investigating the influence of CYP27B1 and CYP24A1 variants on Vitamin D levels and consequently on RF exposures are required to establish conclusive evidence.

Conclusions
Pharmacokinetic heterogeneity in RF levels represents an austere and ubiquitous problem in TB patient care. This can lead to therapeutic inefficacy, resistance, adverse drug events and increased healthcare expenditures. Genetic variants of SLCO1B1, ABCB1 and VDR have attracted scientific attention for their influence on the pharmacokinetics of a wide range of drugs. While there is a vast number of studies that have explored the influence of SNPs with Isoniazid levels in plasma, only a limited number of studies have explored the influence of genetic variants on the RF pharmacokinetics. Evidence available to date reported a lack of any association of ABCB1, PXR, CAR, CES1 and AADAC genetic variants with the RF concentrations in plasma. Some literature has shown an association of certain genetic variants of SLCO1B1, CES2 and Vitamin D pathway genes with significant variations of RF concentration in plasma. A comprehensive exploration of the role of genetic variants of these genes can be initiated to provide a consensus agreement on their influence on RF pharmacokinetics in different populations.
Genotyping offers to be a potential tool of precision medicine for predicting individual drug-metabolizing and drug transport capabilities before initiation of RF treatment. Further studies assessing RF exposure and correlating it with the genetic polymorphisms are required in different ethnic populations. Besides, such research should be based on a representative and appropriate sample size to validate and implement a cost-effective genotyping-based RF dosage optimization in clinical settings and national policy levels.