Polymorphism in the ABCB1 gene encoding P-glycoprotein, a transmembrane drug efflux pump, contributes to drug resistance and has been widely studied. However, their association with rifampicin and ethambutol resistance in tuberculosis (TB) patients is still unclear. Genotype/allele/haplotype frequencies in c.1236C > T (rs1128503), c.2677G > T/A (rs2032582), and c.3435C > T (rs1045642) were obtained from 218 patients. Of these, 80 patients with rifampicin and/or ethambutol resistance were selected as the case group and 138 patients were selected for the control group through the results of their culture and drug-sensitive tests. Patients aged <18 years and HIV-positive serologic tests were excluded. ABCB1 polymorphisms were determined using a PCR direct-sequencing approach, and restriction fragment length polymorphism (RFLP). A nomogram was constructed to simulate a combined prediction of the probability of anti-TB drug resistance, with factors including genotype c.1236C > T (rs1128503) (P=0.02), clinical form (P=0.03), previous treatment (P=0.01), and skin color (P=0.03), contributing up to 90% chance of developing anti-TB drug resistance. Considering genotype analyses, CT (rs1128503) demonstrated an increased chance of anti-TB drug resistance (odds ratio (OR): 2.34, P=0.02), while the analyses for ethambutol resistance revealed an association with a rare A allele (rs2032582) (OR: 12.91, P=0.01), the haplotype TTC (OR: 5.83, P=0.05), and any haplotype containing the rare A allele (OR: 7.17, P=0.04). ABCB1 gene polymorphisms in association with others risk factors contribute to anti-TB drug resistance, mainly ethambutol. The use of the nomogram described in the present study could contribute to clinical decision-making prior to starting TB treatment.

Introduction

Tuberculosis (TB) disease, caused by Mycobacterium tuberculosis (MTB), remains a major cause of deaths worldwide from infectious disease, and 2 million people die annually as a result of this disease [1]. Combination treatment with rifampicin (R), isoniazid (H), ethambutol (E), and pyrazinamide (Z) is the standard TB therapy. Nevertheless, over the few last decades, MTB strains resistant to anti-TB drugs have been emerging [2]. In 2015, an estimated 480000 people worldwide developed resistance to anti-TB drugs and 100000 people were rifampicin resistant [1]. In Brazil, 8.9 thousand-resistant TB cases were notified; of them, 1.4 thousand were new cases and 7.5 thousand were retreatment cases [3].

The emergence of resistant strains cannot always be attributed to factors linked to bacterial genetics as host genetic variations also play a role [4]. Gene polymorphisms encoding drug metabolizing enzymes or membrane transporters, such as an efflux pump, could have an impact on resistance to TB treatment. P-glycoprotein (P-gp) is a transmembrane drug transporter encoded by the multidrug resistance MDR1 (ABCB1 gene), which transports key molecules during drug uptake and efflux [5]. Expression of P-gp occurs in several organs such as the liver, kidney, colon, placenta, and other cells including leukocytes, and have functions in the transport and/or secretion of its substrates as well as the protection of these tissues from physiologically active substances, cytotoxic agents, and xenobiotics [6]. In recent years, a large number of single nucleotide polymorphisms (SNP) of the ABCB1 gene have been described. Amongst them, a nonsynonymous mutation, c.2677G > T/A (rs2032582), which changes an alanine to either serine or threonine, and synonymous mutations, c.3435C > T (rs1045642) and c.1236C > T (rs1128503), that influence mRNA stability [7]. These mutations are associated with variations in expression/activity of P-gp that may affect drug pharmacokinetics/pharmacodynamics reducing or increasing their biodisponibility [8], which has been extensively studied mainly in the resistance to anticancer drugs [6]. The degree of expression and functionality of the ABCB1 gene product can directly affect the therapeutic effectiveness of such drugs [9].

P-gp has been described in mononuclear phagocytes [10], and it is known that macrophages play a central role in TB pathogenesis. Thus, we hypothesized that differences in drug efflux transporters would alter the intracellular concentrations of anti-TB drugs. Amongst the drugs standard to TB treatment, rifampicin and ethambutol are substrates of P-gp [11], thus ABCB1 gene polymorphism that encodes P-gp could be a predictor for the occurrence of TB-resistant cases.

There is little information in the literature regarding the ABCB1 gene in TB patients. One study was conducted in Mexico to evaluate the association of ABCB1 polymorphism and multidrug-resistant TB (MDR-TB) [12]. However, the study used a small casuistic and presented a problem in the selection of the control population. To our knowledge, the present study is the first conducted in a Brazilian population treated for TB aiming to evaluate a possible association of ABCB1 polymorphism with rifampicin and/or ethambutol resistance.

Methods

Study design and patients

This was a case–control study of patients under TB treatment at the National Institute of Infectious Diseases Evandro Chagas (INI), Oswaldo Cruz Foundation (FIOCRUZ), Rio de Janeiro, Brazil, from October 2009 to March 2015. The INI-FIOCRUZ Ethics Committee approved this protocol under the number CAAE 0048.0.009.000-09.

All patients with a positive culture for MTB were selected and grouped according to the drug susceptibility test (DST) results. Patients who were resistant to R and/or E, associated or not with other anti-TB drugs, were selected as cases and those patients lacking drug resistance as controls. Patients aged <18 years and HIV-positive serologic test were excluded.

To characterize the population studied, demographic, laboratorial, and clinical data were obtained from medical records. The alleles/genotype/haplotype profile of patients, in relation to SNP c.1236C > T (rs1128503), c.2677G > T/A (rs2032582), and c.3435C > T (rs1045642) of the ABCB1 gene, were evaluated as possible predictors of anti-TB drug resistance. Other possible predictors of anti-TB drug resistance were also evaluated such as, patient’s age (years), gender, ethnicity (classified as white or nonwhite-skin color according to personal report), TB clinical forms (defined as pulmonary, extrapulmonary, and disseminated), tobacco use (defined as current use reported by the patient), alcoholism (defined by a positive CAGE questionnaire), illicit drug use, previous TB treatments, their default (defined as starting therapy and having a treatment interruption of two or more consecutive months), and use of drug substrates/inhibitors of P-gp.

For risk factor analyses, the control group was compared with the case group for resistance to any anti-TB drugs (n=80). In analyses for genetic factors, the case group was divided into three categories: patients resistant to any anti-TB drugs (n=80), patients resistant to R associated or not with other anti-TB drugs (n =76), and E-resistant (n=16) patients associated or not with other anti-TB drugs. Each of these three categories was compared with the control group.

Genetic molecular analyses

DNA sample and extraction

After signed written consents, 5 ml of blood was collected in tubes containing EDTA, and the genomic DNA was extracted from 0.2 ml of blood samples using the QIAamp DNA Blood Mini Kit (Qiagen, Maryland, U.S.A.) according to the manufacturer’s instructions.

ABCB1 genotyping

For molecular identification of ABCB1 gene polymorphism (GenBank accession number: NT_007933), the SNPs c.1236C > T (rs1128503) and c.3435C > T (rs1045642) were analyzed using PCR-restriction fragment length polymorphism (PCR-RFLP). For limitations on restriction enzyme use in evaluating the SNP c.2677G > T/A (rs2032582), a PCR direct sequencing approach was performed. For the molecular assay, PCR conditions were: 50 µl reaction mixture containing 3 µl of DNA, 0.4 mM each primer (Table 1), 0.2 mM dNTPs, 1.5 mM MgCl2, and 2.0 units of Platinum TaqDNA polymerase (Invitrogen, Life Technologies, Carlsbad, U.S.A.), which was thermocycled in an Eppendorf-Master Cycler-PCR machine under the following conditions: denaturation at 95°C for 5 min, followed by 35 cycles at 95°C for 60 s, 63°C for 60 s, 72°C for 60 s, and extension at 72°C for 10 min. The presence of the fragment size of each amplified region was detected on 2.0% agarose gel followed by GelRed™ staining (Biotium). The DNA image was digitalized using a transilluminator with an image capture system, L-PIX-ST and L-PIX IMAGE 7.1 M Pixel.

Table 1
Primers used in the PCR to amplify the coding region
Gene ABCB1  Primers Fragment size 
rs1128503 Sense 5′-TATCCTGTGTCTGTGAATTGCC-3′ 370 bp 
 Reverse 5′-CCTGACTCACCACACCAATG-3′  
rs2032582 Sense 5′-AGGCTATAGGTTCCAGGCTTGCT-3′ 455 bp 
 Reverse 5′-ACAGTGTGAAGACAATGGCCTGA-3′  
rs1045642 Sense 5′-TGTTTTCAGCTGCTTGATGGC-3′ 250 bp 
 Reverse 5′-TGCTCCCAGGCTGTTTATTTGA-3′  
Gene ABCB1  Primers Fragment size 
rs1128503 Sense 5′-TATCCTGTGTCTGTGAATTGCC-3′ 370 bp 
 Reverse 5′-CCTGACTCACCACACCAATG-3′  
rs2032582 Sense 5′-AGGCTATAGGTTCCAGGCTTGCT-3′ 455 bp 
 Reverse 5′-ACAGTGTGAAGACAATGGCCTGA-3′  
rs1045642 Sense 5′-TGTTTTCAGCTGCTTGATGGC-3′ 250 bp 
 Reverse 5′-TGCTCCCAGGCTGTTTATTTGA-3′  

The analyses of c.1236C > T (rs1128503) was performed by digesting the PCR products of 370 bp in length using the HaeIII enzyme (New England BioLabs® Inc.). If the samples did not contain any mutation, the fragment sizes generated were 270, 65, and 35 bp, the presence of a heterozygous mutation was shown by sizes of 270, 100, 65, and 35 bp, while bands of 270 and 100 bp indicated that samples were homozygous for mutation.

For the direct sequencing of mutations at position c.2677G > T/A (rs2032582), the PCR products were purified using the Wizard SV Gel and PCR Clean-Up System Kit (Promega, U.S.A.) according to the manufacturer’s instructions. The Big Dye Terminator™ Cycle Sequencing Kit (Applied Biosystem, Inc., U.S.A.) was used for further direct DNA sequencing. Base composition analyses were determined on an ABI Prism®3730 DNA Analyzer at an institutional genomic facility (RPT01A-PDTIS/FIOCRUZ) [13]. The GenBank sequence NT_007933 was used as a reference for the ABCB1 gene. The sequences were edited and assigned using Sequencher 4.1.4 software (demo version).

The PCR products for c.3435C > T (rs1045642) that contained the recognition site for the MboI enzyme (New England BioLabs® Inc.) indicated the wild-type sample with fragment lengths of 162 and 88 bp, while mutant samples presented fragment lengths of 250 bp, heterozygous samples had fragment sizes of 250, 162, and 88 bp.

The mutation results performed with RFLP (rs11228503 and rs1045642) were confirmed by direct sequencing assay.

Statistical analysis

Deviations from the Hardy–Weinberg equilibrium (HWE) were assessed by Chi-square tests. The proportion of individuals carrying each ABCB1 gene was computed for cases and controls separately. Differences between cases and controls were expressed as odds ratios (ORs). To evaluate the associations between genetic and other predictors (demographic, clinical, and behavioral) with each outcome (resistance, resistance to ethambutol, and resistance to rifampicin) we used logistic models. Initially, nongenotypic predictors were selected and only significant ones were used as control variables for each genetic predictor in the multivariate models. To evaluate the association of genotypes with nongenetic predictors, the Kruskall–Wallis, Chi-square, and Fisher exact tests were used, as appropriate. To help readers predict anti-TB drug resistance or estimate the probability of anti-TB drug resistance, a nomogram was constructed from the fitted logistic model, which graphically represented a probability score for resistance. The significance level was set at 0.05 and the program used for analyses was R version 3.2.2.

Results

Characteristics of patients

The present study included 858 TB patients, amongst whom 218 fulfilled the inclusion criteria. Of these, 80 were selected in the case group and 138 in the control group through the results of the DST. Resistance characteristics of 80 MTB strains isolated from patients are detailed in Table 2. Multidrug-resistant tuberculosis (MDR-TB) was detected for most (85%) of the selected cases.

Table 2
Distribution of resistance amongst 80 strains of MTB isolated
Resistance n=80 (%) 
6 (7.5) 
E + S 1 (1.3) 
H + E 1 (1.3) 
R + S 2 (2.5) 
R + H* 44 (55) 
H + E + Z 1 (1.3) 
R + H + S* 7 (8.8) 
R + H + Z* 4 (5.0) 
R + H + E* 8 (10) 
H + Z + E + S 1 (1.3) 
R + H + E + S* 3 (3.8) 
R + H + Z + E* 1 (1.3) 
R + H + Z + S* 1 (1.3) 
Resistance n=80 (%) 
6 (7.5) 
E + S 1 (1.3) 
H + E 1 (1.3) 
R + S 2 (2.5) 
R + H* 44 (55) 
H + E + Z 1 (1.3) 
R + H + S* 7 (8.8) 
R + H + Z* 4 (5.0) 
R + H + E* 8 (10) 
H + Z + E + S 1 (1.3) 
R + H + E + S* 3 (3.8) 
R + H + Z + E* 1 (1.3) 
R + H + Z + S* 1 (1.3) 

*Multidrug-resistant strains. Abbreviations: H, isoniazid; E, ethambutol; R, rifampicin; S, streptomycin; Z, pyrazinamide.

Note: Asterisks show multidrug-resistant strains

Patient’s characteristics such as age, gender, drugs, alcohol and tobacco use were similar amongst cases and controls. The TB disseminate clinical form and the white skin color was a protecting factor for having resistance. A higher risk of resistance to anti-TB drugs was observed amongst patients who had a previous history of TB treatment, OR =10.21 (5.08–21.63). Default was not a risk factor for resistance in our study (Table 3).

Table 3
Risk factors associated with resistance to any first-line anti-TB drugs in the studied population
Risk factors Univariate Multivariate 
 Sensitives (n=138) Resistants (n=80) OR (IC95%) P-value OR (IC95%) P-value 
Age       
  Median 35.5 (26 ± 54) 38.5 (25, 50.2) 1 (0.98–1.01) 0.74   
Gender       
  Female 51 (62.2) 31 (37.8)    
  Male 87 (64) 49 (36) 0.93 (0.53–1.64) 0.79   
Skin color       
  Nonwhite 57 (52.8) 51 (47.2)  
  White 80 (73.4) 29 (26.6) 0.41 (0.23–0.71) <0.01 0.46 (0.23–0.91) 0.03 
Alcohol user       
  No 68 (64.2) 38 (35.8)    
  Yes 66 (65.3) 35 (34.7) 0.95 (0.54–1.68) 0.86   
Tobacco user       
  No 92 (64.8) 50 (35.2)   
  Yes 42 (63.6) 24 (36.4) 1.05 (0.57–1.92) 0.87   
Drug user       
  No 118 (63.4) 68 (36.6)   
  Yes 17 (70.8) 7 (29.2) 0.71 (0.26–1.75) 0.48   
TB forms       
  Pulmonary 101 (58) 73 (42) 
  Disseminated 22 (88) 3 (12) 0.19 (0.04–0.57) 0.01 0.2 (0.04–0.74) 0.03 
  Extrapulmonary 15 (78.9) 4 (21.1) 0.37 (0.1–1.07) 0.09 0.92 (0.24–2.84) 0.9 
Previous treatments       
  =1 123 (78.3) 34 (21.7) 
  >1 15 (24.6) 46 (75.4) 11.09 (5.65–22.86) <0.01 10.21 (5.08–21.63) <0.01 
Number of default      
  =1 133 (64.3) 74 (35.7)   
  >1 5 (45.5) 6 (54.5) 2.16 (0.63–7.71) 0.22   
Substrate use/P-gp inhibitors      
  No 104 (61.9) 64 (38.1)   
  Yes 34 (68) 16 (32) 0.76 (0.38–1.48) 0.43   
Risk factors Univariate Multivariate 
 Sensitives (n=138) Resistants (n=80) OR (IC95%) P-value OR (IC95%) P-value 
Age       
  Median 35.5 (26 ± 54) 38.5 (25, 50.2) 1 (0.98–1.01) 0.74   
Gender       
  Female 51 (62.2) 31 (37.8)    
  Male 87 (64) 49 (36) 0.93 (0.53–1.64) 0.79   
Skin color       
  Nonwhite 57 (52.8) 51 (47.2)  
  White 80 (73.4) 29 (26.6) 0.41 (0.23–0.71) <0.01 0.46 (0.23–0.91) 0.03 
Alcohol user       
  No 68 (64.2) 38 (35.8)    
  Yes 66 (65.3) 35 (34.7) 0.95 (0.54–1.68) 0.86   
Tobacco user       
  No 92 (64.8) 50 (35.2)   
  Yes 42 (63.6) 24 (36.4) 1.05 (0.57–1.92) 0.87   
Drug user       
  No 118 (63.4) 68 (36.6)   
  Yes 17 (70.8) 7 (29.2) 0.71 (0.26–1.75) 0.48   
TB forms       
  Pulmonary 101 (58) 73 (42) 
  Disseminated 22 (88) 3 (12) 0.19 (0.04–0.57) 0.01 0.2 (0.04–0.74) 0.03 
  Extrapulmonary 15 (78.9) 4 (21.1) 0.37 (0.1–1.07) 0.09 0.92 (0.24–2.84) 0.9 
Previous treatments       
  =1 123 (78.3) 34 (21.7) 
  >1 15 (24.6) 46 (75.4) 11.09 (5.65–22.86) <0.01 10.21 (5.08–21.63) <0.01 
Number of default      
  =1 133 (64.3) 74 (35.7)   
  >1 5 (45.5) 6 (54.5) 2.16 (0.63–7.71) 0.22   
Substrate use/P-gp inhibitors      
  No 104 (61.9) 64 (38.1)   
  Yes 34 (68) 16 (32) 0.76 (0.38–1.48) 0.43   

*Examples such as ranitin, losartan, omeprazole, digoxin. Abbreviation: IC, confidence interval.

Note: Bold P-values show significant values

Genotypes, alleles, and haplotypes in patients

The HWE test on the genotype and allele of the two groups of patients demonstrated that the population is a HWE (results not shown).

In the case group, resistant to any drugs, the genotype analyses of c.1236C > T (rs1128503) showed that CT genotype, when compared with wild-type homozygous, was associated with resistance to any anti-TB drugs (OR: 2.34, P=0.02). However, no significant differences were found in the frequencies of the other polymorphisms c.2677G > T/A (rs2032582) and c.3435C > T (rs1045642). With regard to the haplotype analyses, TTC demonstrated an association with resistance (OR: 7.22, P=0.03) (Table 4).

Table 4
Genetic characteristics of studied population with resistance to any anti-TB drugs
Genetic predictors Univariate Multivariate 
 Sensitives (n=138) Resistants (n=80) OR (IC95%) P-value OR (IC95%) P-value 
Genotype       
c.1236C > T (rs1128503)       
  C/C 69 (66.3) 35 (33.7) 
  C/T 47 (56) 37 (44) 1.55 (0.86–2.82) 0.15 2.34(1.14–4.95) 0.02 
  T/T 22 (73.3) 8 (26.7) 0.72 (0.28–1.72) 0.47 0.65 (0.2–1.94) 0.45 
c.2677G > T/A (rs2032582)       
  G/G 73 (66.4) 37 (33.6) 
  G/T 51 (60) 34 (40) 1.32 (0.73–2.37) 0.36 1.77 (0.88–3.66) 0.12 
  T/T 12 (66.7) 6 (33.3) 0.99 (0.32–2.76) 0.98 0.82 (0.2–3.08) 0.78 
  G/A 2 (40) 3 (60) 2.96 (0.47–23.21) 0.25 4.11 (0.51–37.13) 0.17 
c.3435C > T (rs1045642)       
  C/C 57 (63.3) 33 (36.7) 
  C/T 62 (60.8) 40 (39.2) 1.11 (0.62–2.01) 0.72 1.85 (0.88–4.02) 0.11 
  T/T 19 (73.1) 7 (26.9) 0.64 (0.23–1.62) 0.36 0.54 (0.15–1.78) 0.33 
Allele frequencies       
c.1236C > T (rs1128503)       
  C 185 (63.4) 107 (36.6) 
  T 91 (63.2) 53 (36.8) 1.01 (0.66–1.52) 0.97 1.11 (0.67–1.81) 0.69 
c.2677G > T/A (rs2032582)       
  G 199 (64.2) 111 (35.8) 
  T 75 (62) 46 (38) 1.1 (0.71–1.69) 0.67 1.19 (0.71–2) 0.51 
  A 2 (40) 3 (60) 2.69 (0.44–20.65) 0.28 3.32 (0.44–28.55) 0.23 
c.3435C > T (rs1045642)       
  C 176 (62.4) 106 (37.6) 
  T 100 (64.9) 54 (35.1) 0.9 (0.59–1.35) 0.6 1.01 (0.62–1.64) 0.98 
Haplotype       
  CGC 155 (64.6) 85 (35.4) 
  CGT 22 (59.5) 15 (40.5) 1.3 (0.52–11.62) 0.49 1.44 (0.72–19.87) 0.42 
  CTC 2 (28.6) 5 (71.4) 2.45 (0.35–2.81) 0.26 3.79 (0.32–3.33) 0.12 
  TGC 15 (68.2) 7 (31.8) 1 (0.22–5.16) 1.03 (0.33–11.03) 0.96 
  TGT 7 (63.6) 4 (36.4) 1.06 (1.29–24.62) 0.94 1.9 (1.22–42.73) 0.47 
  TTC 3 (27.3) 8 (72.7) 5.64 (0.52–1.46) 0.02 7.22 (0.51–1.77) 0.03 
  TTT 66 (67.3) 32 (32.7) 0.87 (0.36–5.99) 0.6 0.95 (0.38–8.71) 0.86 
  Rare* 6 (60) 4 (40) 1.48 (0.62–2.73) 0.58 1.83 (0.04–0.92) 0.45 
Genetic predictors Univariate Multivariate 
 Sensitives (n=138) Resistants (n=80) OR (IC95%) P-value OR (IC95%) P-value 
Genotype       
c.1236C > T (rs1128503)       
  C/C 69 (66.3) 35 (33.7) 
  C/T 47 (56) 37 (44) 1.55 (0.86–2.82) 0.15 2.34(1.14–4.95) 0.02 
  T/T 22 (73.3) 8 (26.7) 0.72 (0.28–1.72) 0.47 0.65 (0.2–1.94) 0.45 
c.2677G > T/A (rs2032582)       
  G/G 73 (66.4) 37 (33.6) 
  G/T 51 (60) 34 (40) 1.32 (0.73–2.37) 0.36 1.77 (0.88–3.66) 0.12 
  T/T 12 (66.7) 6 (33.3) 0.99 (0.32–2.76) 0.98 0.82 (0.2–3.08) 0.78 
  G/A 2 (40) 3 (60) 2.96 (0.47–23.21) 0.25 4.11 (0.51–37.13) 0.17 
c.3435C > T (rs1045642)       
  C/C 57 (63.3) 33 (36.7) 
  C/T 62 (60.8) 40 (39.2) 1.11 (0.62–2.01) 0.72 1.85 (0.88–4.02) 0.11 
  T/T 19 (73.1) 7 (26.9) 0.64 (0.23–1.62) 0.36 0.54 (0.15–1.78) 0.33 
Allele frequencies       
c.1236C > T (rs1128503)       
  C 185 (63.4) 107 (36.6) 
  T 91 (63.2) 53 (36.8) 1.01 (0.66–1.52) 0.97 1.11 (0.67–1.81) 0.69 
c.2677G > T/A (rs2032582)       
  G 199 (64.2) 111 (35.8) 
  T 75 (62) 46 (38) 1.1 (0.71–1.69) 0.67 1.19 (0.71–2) 0.51 
  A 2 (40) 3 (60) 2.69 (0.44–20.65) 0.28 3.32 (0.44–28.55) 0.23 
c.3435C > T (rs1045642)       
  C 176 (62.4) 106 (37.6) 
  T 100 (64.9) 54 (35.1) 0.9 (0.59–1.35) 0.6 1.01 (0.62–1.64) 0.98 
Haplotype       
  CGC 155 (64.6) 85 (35.4) 
  CGT 22 (59.5) 15 (40.5) 1.3 (0.52–11.62) 0.49 1.44 (0.72–19.87) 0.42 
  CTC 2 (28.6) 5 (71.4) 2.45 (0.35–2.81) 0.26 3.79 (0.32–3.33) 0.12 
  TGC 15 (68.2) 7 (31.8) 1 (0.22–5.16) 1.03 (0.33–11.03) 0.96 
  TGT 7 (63.6) 4 (36.4) 1.06 (1.29–24.62) 0.94 1.9 (1.22–42.73) 0.47 
  TTC 3 (27.3) 8 (72.7) 5.64 (0.52–1.46) 0.02 7.22 (0.51–1.77) 0.03 
  TTT 66 (67.3) 32 (32.7) 0.87 (0.36–5.99) 0.6 0.95 (0.38–8.71) 0.86 
  Rare* 6 (60) 4 (40) 1.48 (0.62–2.73) 0.58 1.83 (0.04–0.92) 0.45 

*Rare – all haplotypes containing rare A allele. Abbreviation: IC, confidence interval.

Note: Bold P-values show significant values.

In the group resistant to R, associated or not with other drugs, genotype, allele, and haplotype frequencies showed no statistical differences for any of the polymorphisms studied. Regarding patients resistant to E, associated or not with other drugs, the analysis showed a statistical significance for the rare ‘A’ allele of c.2677G > T/A (rs2032582) (OR: 12.91, P<0.01,), the GA genotype (OR: 16.37, P=0.01), and haplotype analysis TTC (OR: 5.83, P=0.05), and any haplotype that contained the rare A allele (OR: 7.17, P=0.04) (Table 5).

Table 5
Genetic characteristics of studied population with resistance to ethambutol associated or not with other anti-TB drugs
Genetic predictors Univariate Multivariate 
 Sensitive (n=202) Resistant E (n=16) OR (IC95%) P-value OR (IC95%) P-value 
Genotype       
c.1236C > T (rs1128503)       
  C/C 99 (95.2) 5 (4.8) 
  C/T 75 (89.3) 9 (10.7) 2.38 (0.79–8) 0.13 2.6 (0.84–8.95) 0.11 
  T/T 28 (93.3) 2 (6.7) 0.6 (0.2–6.96) 0.69 1.5 (0.2–7.62) 0.64 
c.2677G > T/A (rs2032582)       
  G/G 104 (94.5) 6 (5.5) 
  G/T 79 (92.9) 6 (7.1) 1.32 (0.4–4.36) 0.64 1.46 (0.43–4.94) 0.54 
  T/T 16 (88.9) 2 (11.1) 2.17 (0.3–10.39) 0.37 2.12 (0.28–10.65) 0.39 
  G/A 3 (60) 2 (40) 11.56 (1.34–84.3) 0.01 16.37 (1.73–136.27) 0.01 
c.3435C > T (rs1045642)       
  C/C 82 (91.1) 8 (8.9) 
  C/T 95 (93.1) 7 (6.9) 0.76 (0.25–2.19) 0.6 0.97 (0.32–2.93) 0.95 
  T/T 25 (96.2) 1 (3.8) 0.48 (0.02–2.39) 0.41 0.41 (0.02–2.44) 0.41 
Allele frequencies       
c.1236C > T (rs1128503)       
  C 273 (93.5) 19 (6.5) 
  T 131 (91) 13 (9) 1.43 (0.67–2.95) 0.34 1.5 (0.69–3.15) 0.29 
c.2677G > T/A (rs2032582)       
  G 290 (93.5) 20 (6.5) 
  A 3 (60) 2 (40) 9.67 (1.22–61.62) 0.02 12.91 (1.52–89.2) 0.01 
  T 111 (91.7) 10 (8.3) 1.31 (0.57–2.82) 0.51 1.34 (0.58–2.95) 0.47 
c.3435C > T (rs1045642)       
  C 259 (91.8) 23 (8.2) 
  T 145 (94.2) 9 (5.8) 0.7 (0.3–1.5) 0.38 0.75 (0.32–1.63) 0.48 
Haplotype       
  CGC 225 (93.8) 15 (6.2) 
  CGT 35 (94.6) 2 (5.4) 0.91 (0.19–12.58) 0.91 0.83 (0.36–2.81) 0.83 
  CTC 6 (85.7) 1 (14.3) 1.55 (0.75–14.81) 0.68 1.56 (0.73–15.22) 0.07 
  TGC 19 (86.4) 3 (13.6) 3.34 (0–0) 0.11 3.33 (0–0) 0.12 
  TGT 11 (100) 0 (0) 
  TTC 8 (72.7) 3 (27.3) 7.88 (0.32–2.65) 0.02 5.83 (0.36–2.81) 0.05 
  TTT 92 (93.9) 6 (6.1) 0.92 (0.93–39.74) 0.88 1 (1.07–48.1) 
  Rare* 8 (80) 2 (20) 6.08 (0.18–4.57) 0.06 7.17 (1.16–10.6) 0.04 
Genetic predictors Univariate Multivariate 
 Sensitive (n=202) Resistant E (n=16) OR (IC95%) P-value OR (IC95%) P-value 
Genotype       
c.1236C > T (rs1128503)       
  C/C 99 (95.2) 5 (4.8) 
  C/T 75 (89.3) 9 (10.7) 2.38 (0.79–8) 0.13 2.6 (0.84–8.95) 0.11 
  T/T 28 (93.3) 2 (6.7) 0.6 (0.2–6.96) 0.69 1.5 (0.2–7.62) 0.64 
c.2677G > T/A (rs2032582)       
  G/G 104 (94.5) 6 (5.5) 
  G/T 79 (92.9) 6 (7.1) 1.32 (0.4–4.36) 0.64 1.46 (0.43–4.94) 0.54 
  T/T 16 (88.9) 2 (11.1) 2.17 (0.3–10.39) 0.37 2.12 (0.28–10.65) 0.39 
  G/A 3 (60) 2 (40) 11.56 (1.34–84.3) 0.01 16.37 (1.73–136.27) 0.01 
c.3435C > T (rs1045642)       
  C/C 82 (91.1) 8 (8.9) 
  C/T 95 (93.1) 7 (6.9) 0.76 (0.25–2.19) 0.6 0.97 (0.32–2.93) 0.95 
  T/T 25 (96.2) 1 (3.8) 0.48 (0.02–2.39) 0.41 0.41 (0.02–2.44) 0.41 
Allele frequencies       
c.1236C > T (rs1128503)       
  C 273 (93.5) 19 (6.5) 
  T 131 (91) 13 (9) 1.43 (0.67–2.95) 0.34 1.5 (0.69–3.15) 0.29 
c.2677G > T/A (rs2032582)       
  G 290 (93.5) 20 (6.5) 
  A 3 (60) 2 (40) 9.67 (1.22–61.62) 0.02 12.91 (1.52–89.2) 0.01 
  T 111 (91.7) 10 (8.3) 1.31 (0.57–2.82) 0.51 1.34 (0.58–2.95) 0.47 
c.3435C > T (rs1045642)       
  C 259 (91.8) 23 (8.2) 
  T 145 (94.2) 9 (5.8) 0.7 (0.3–1.5) 0.38 0.75 (0.32–1.63) 0.48 
Haplotype       
  CGC 225 (93.8) 15 (6.2) 
  CGT 35 (94.6) 2 (5.4) 0.91 (0.19–12.58) 0.91 0.83 (0.36–2.81) 0.83 
  CTC 6 (85.7) 1 (14.3) 1.55 (0.75–14.81) 0.68 1.56 (0.73–15.22) 0.07 
  TGC 19 (86.4) 3 (13.6) 3.34 (0–0) 0.11 3.33 (0–0) 0.12 
  TGT 11 (100) 0 (0) 
  TTC 8 (72.7) 3 (27.3) 7.88 (0.32–2.65) 0.02 5.83 (0.36–2.81) 0.05 
  TTT 92 (93.9) 6 (6.1) 0.92 (0.93–39.74) 0.88 1 (1.07–48.1) 
  Rare* 8 (80) 2 (20) 6.08 (0.18–4.57) 0.06 7.17 (1.16–10.6) 0.04 

*Rare – all haplotypes containing rare allele. Abbreviation: IC, confidence interval.

Note: Bold P-values show significant values

Prediction score

In the multivariate analysis, a final model including variables was fitted. The probability scores were developed based on the nomogram. The graph simulates a combined prediction behavior. The first category of the group includes patients resistant to any anti-TB drugs, genotype c.1236C > T (rs1128503) (P=0.02), clinical form (P=0.03), previous treatment (P<0.01), and skin color (P=0.03). In Figure 1, as an example on how to use the nomogram, suppose that a patient is CT genotype. Draw a vertical line from ‘C/T’ toward the upper line of the graph and register the points (52 points). The same patient has pulmonary clinical form (68 points); more than a previous treatment (100 points) and nonwhite-skin color (38 points), therefore, the total points equals approximately 258. To ascertain the probability represented by this score, draw a vertical line from the ‘Total points’ line toward ‘Probability to resistance’, which is almost 90%. In Figure 2, the same nomogram is represented, but with the category of resistance to ethambutol.

Nomogram to estimate risk probabilities of resistance to any anti-TB drugs according to the study findings

Figure 1
Nomogram to estimate risk probabilities of resistance to any anti-TB drugs according to the study findings

Draw a vertical line from your patient characteristics toward the upper line and sum the points. After doing this for all characteristics, sum all the points from all five characteristics and draw a vertical line from the ‘Score’ line toward the ‘Probability to resistance’.

Figure 1
Nomogram to estimate risk probabilities of resistance to any anti-TB drugs according to the study findings

Draw a vertical line from your patient characteristics toward the upper line and sum the points. After doing this for all characteristics, sum all the points from all five characteristics and draw a vertical line from the ‘Score’ line toward the ‘Probability to resistance’.

Nomogram to estimate risk probabilities of resistance to Ethambutol according to the study findings

Figure 2
Nomogram to estimate risk probabilities of resistance to Ethambutol according to the study findings

Draw a vertical line from your patient characteristics toward the upper line and sum the points. After doing this for all characteristics, sum all points from all the characteristics and draw a vertical line from the ‘Score’ line toward the ‘Risk to resistance ethambutol’.

Figure 2
Nomogram to estimate risk probabilities of resistance to Ethambutol according to the study findings

Draw a vertical line from your patient characteristics toward the upper line and sum the points. After doing this for all characteristics, sum all points from all the characteristics and draw a vertical line from the ‘Score’ line toward the ‘Risk to resistance ethambutol’.

Discussion

Drug resistance is a matter of great concern for TB control programs and drug-resistant TB has different regional trends [14]. Better comprehension of the risk factors associated with MDR-TB strains is highly significant in developing a national policy for public health interventions [15].

The lack of TB treatment compliance does not always explain the risk to TB resistance [16], therefore, genetic markers are of great value. Nowadays, immune-related gene polymorphisms, such as SLC11A1, some HLA genes, TLR7, TLR8, and INFG, have been associated with increased susceptibility to pulmonary TB. Nevertheless, only a few reports exist regarding the role of host pharmacogenetics that may lead to the development of TB-resistant forms [12]. Due to this, we hypothesized that polymorphism in the ABCB1 gene would contribute to better understanding the development of anti-TB drug resistance.

There are conflicting findings in the literature regarding polymorphism c.1236C > T (rs1128503) located in exon 12, which characterizes a silent mutation that does not exert an effect on the structure and function of P-gp, but changes the substrate specificity, contributing to altered substrate pharmacokinetics [17]. Some studies showed a better therapeutic response in-group with homozygous CC or heterozygous CT when compared with the TT genotype [18]. In another study with epilepsy patients, homozygous TT was associated with better outcomes [19]. Actually, our study found that heterozygous patients (CT) have a two-fold chance of resistance to anti-TB drugs, suggesting that each drug should be evaluated according to the specificity of P-gp. In addition, the concomitant use of P-gp substrates/inhibitors that compete for the same anti-TB drug-binding site, could affect the bioavailability. Nevertheless, our statistical analyses did not find significant association, corroborating the clinical trial results that tested P-gp inhibitors [20].

Regarding c.2677G > T/A (rs2032582) located in exon 21, genotypic analyses containing the T allele did not find any association with anti-TB drug resistance. On the other hand, the analyses with rare A allele, that has the greatest impact on both activity and P-gp substrate specificity [10], demonstrated an association with ethambutol resistance and A allele, as well as the genotype/haplotype containing this rare allele, corroborating another study [12].

The three SNPs in loci c.1236C > T, c.2677G > T/A, and c.3435C > T of the ABCB1 gene located in exon 12, 21, and 26, respectively are closely connected. The haplotype is the combination of SNPs that are located and interconnected in a specific region of the chromosome [19]. Our study showed that the TTC haplotype increases the chance of ethambutol resistance by five-fold. In fact, the presence of mutant alleles located in exon 12 and 21, alter P-gp affinity and substrate-related activity [12,17]. However, the same haplotypes can confer a favorable response to other substrates, for example phenytoin [19], confirming the affinity substrate dependence. Moreover, other discrepancies in responses to many diseases correlated with ABCB1 polymorphism have been described, and reasons for this discrepancy may be due to population heterogeneity, sample size, incomplete phenotype [21], or genotypic analyses without haplotype analyses. In addition, the inherited component of the response to drugs is typically polygenic, so the impact of a single gene may be confounded by influences from other genes and by environmental factors [22].

Concerning risk factors not related to pharmacogenetics, our study showed that white-skinned patients are less expected to have resistance to anti-TB drugs compared with nonwhite skin. One study showed that nonwhite-skin individual macrophages allow significantly greater bacillary replication than those of white-skinned individuals [13] showing that nonwhite-skinned individuals are more susceptible to TB [23]. Therefore, TB susceptibility in certain populations could be directly related to TB recurrence and could result in a large number of treatments as well as the probability of having anti-TB drug resistance. In fact, our results showed a ten-fold higher frequency of resistance to anti-TB drugs in patients with more than one previous treatment, as showed by other authors [15,24]. A previous treatment, even an adequate one, still represents a major risk factor for MDR-TB [25].

Regarding TB clinical forms, several studies have described a high probability of having resistant TB in a pulmonary clinical form [15,26]. The bacteria multiply inside the lung, an organ with a high level of oxygenation, which favors bacilli development [27]. Besides, drugs are easily distributed in the lung, a highly perfused organ, enabling higher drug concentrations and consequently a greater probability of bacteria developing resistance [27]. Corroborating this information, the findings of our study demonstrated that the disseminated clinical form is a protecting factor to resistant TB.

Conclusion

Having nonwhite-skin color, pulmonary TB, previous TB treatment, and the genotype CT in exon 12, and GA in exon 21, A allele in exon 21, and haplotype TTC or any haplotype containing A allele of the ABCB1 gene, are associated with a higher chance of developing TB resistance. Based on these results, we constructed a nomogram that simulate the chance of having anti-TB drug resistance. A nomogram is a useful tool for clinical decision-making in order to evaluate TB-resistance probability. All risk factors should be kept in mind while choosing drugs to compose a regimen to prevent treatment resistance.

Clinical perspectives

  • Resistant strains cannot always be attributed to factors linked to bacterial genetics but also to host genetic variations.

  • The present study demonstrated that interindividual variability in anti-TB treatment with P-gp substrate drugs is clinically relevant. According to knowledge, it is the first study in Latin America involving combination of ABCB1 gene SNPs in resistant-TB patients.

  • Thus, we would like to suggest further study with larger sample size to correlate with these three SNPs analyzed, corroborating with nomogram application in clinical practice, since the clinical implementation of ABCB1 genotyping in combination with other clinical information as described in the constructed nomogram could contribute to reduce drug-resistant TB.

We thank the medical, technical, and nursing staff of the INI for their support, and L.S. Rosadas for technical assistance. We also thank all the patients for their participation in the present study. We also thank P.C. Monteiro for English language revision.

Funding

his study was supported by Programa de Incentivo à Pesquisa e Desenvolvimento Tecnológico (PIPDT),FIOCRUZ and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).

Competing interests

The authors declare that there are no competing interests associated with the manuscript.

Author contribution

The work presented here was carried out in collaboration amongst all the authors. L.d.C. defined the research theme, designed the methods and experiments. Y.P. carried out the laboratory experiments and collected clinical data. S.P.M. analyzed the molecular data. M.J.M.C. was the physician who treated the patients and obtained their written informed consents. All the authors worked on patient data collection, interpreted the results, and wrote the paper. All authors have contributed to, seen, and approved the manuscript.

Abbreviations

     
  • DST

    drug susceptibility test

  •  
  • FIOCRUZ

    Oswaldo Cruz Foundation

  •  
  • HWE

    Hardy–Weinberg equilibrium

  •  
  • IC

    confidence interval

  •  
  • INI

    National Institute of Infectious Diseases Evandro Chagas

  •  
  • MDR-TB

    multidrug-resistant tuberculosis

  •  
  • MTB

    Mycobacterium tuberculosis

  •  
  • OR

    odds ratio

  •  
  • P-gp

    P-glycoprotein

  •  
  • SNP

    single nucleotide polymorphism

  •  
  • TB

    tuberculosis

References

References
1
World Health Organization
(
2017
)
Global tuberculosis report 2016
.
2
Gandhi
N.R.
,
Nunn
P.
,
Dheda
K.
,
Schaaf
H.S.
,
Zignol
M.
,
van Soolingen
D.
et al. 
(
2010
)
Multidrug-resistant and extensively drug-resistant tuberculosis: a threat to global control of tuberculosis
.
Lancet
375
,
1830
1843
3
World Health Organization
(
2015
)
Global tuberculosis report 2015
.
4
Beceiro
A.
,
Tomás
M.
and
Bou
G.
(
2013
)
Antimicrobial resistance and virulence: a successful or deleterious association in the bacterial world?
Clin. Microbiol. Rev.
26
,
185
230
5
Kimura
Y.
,
Morita
S.
,
Matsuo
M.
and
Ueda
K.
(
2007
)
Mechanism of multidrug recognition by MDR1/ABCB1
.
Cancer Sci.
98
,
1303
1310
6
Katayama
K.
,
Noguchi
K.
and
Sugimoto
Y.
(
2014
)
Regulations of P-glycoprotein/ABCB1/MDR1 in human cancer cells
.
New J. Sci.
2014
,
e476974
7
Tovilla-Zárate
C.A.
,
Vargas
I.
,
Hernández
S.
,
Fresán
A.
,
Aguilar
A.
,
Escamilla
R.
et al. 
(
2014
)
Association study between the MDR1 gene and clinical characteristics in schizophrenia
.
Rev. Bras. Psiquiatr.
36
,
227
232
8
Schinkel
A.H.
(
1998
)
Pharmacological insights from P-glycoprotein knockout mice
.
Int. J. Clin. Pharmacol. Ther.
36
,
9
13
9
Hoffmeyer
S.
,
Burk
O.
,
von Richter
O.
,
Arnold
H.P.
,
Brockmöller
J.
,
Johne
A.
et al. 
(
2000
)
Functional polymorphisms of the human multidrug-resistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo
.
Proc. Natl. Acad. Sci. U.S.A.
97
,
3473
3478
10
Gollapudi
S.
,
Reddy
M.
,
Gangadharam
P.
,
Tsuruo
T.
and
Gupta
S.
(
1994
)
Mycobacterium tuberculosis induces expression of P-glycoprotein in promonocytic U1 cells chronically infected with HIV type 1
.
Biochem. Biophys. Res. Commun.
199
,
1181
1187
11
Hartkoorn
R.C.
,
Chandler
B.
,
Owen
A.
,
Ward
S.A.
,
Bertel Squire
S.
,
Back
D.J.
et al. 
(
2007
)
Differential drug susceptibility of intracellular and extracellular tuberculosis, and the impact of P-glycoprotein
.
Tuberculosis (Edinb.)
87
,
248
255
12
Rodríguez-Castillo
J.A.
,
Arce-Mendoza
A.Y.
,
Quintanilla-Siller
A.
,
Rendon
A.
,
Salinas-Carmona
M.C.
and
Rosas-Taraco
A.G.
(
2015
)
Possible association of rare polymorphism in the ABCB1 gene with rifampin and ethambutol drug-resistant tuberculosis
.
Tuberculosis (Edinb.)
95
,
532
537
13
Crowle
A.J.
and
Elkins
N.
(
1990
)
Relative permissiveness of macrophages from black and white people for virulent tubercle bacilli
.
Infect. Immun.
58
,
632
638
14
Ministério da Saúde, Brasil (2014) Panorama da Tuberculose no Brasil. http://bvsms.saude.gov.br/bvs/publicacoes/panorama%20tuberculose%20brasil_2014.pdf (in Portuguese)
15
Elmi
O.S.
,
Hasan
H.
,
Abdullah
S.
,
Mat Jeab
M.Z.
,
Bin Alwi
Z.
and
Naing
N.N.
(
2015
)
Multidrug-resistant tuberculosis and risk factors associated with its development: a retrospective study
.
J. Infect. Dev. Ctries
9
,
1076
1085
16
Amante
T.D.
and
Ahemed
T.A.
(
2015
)
Risk factors for unsuccessful tuberculosis treatment outcome (failure, default and death) in public health institutions, Eastern Ethiopia
.
Pan. Afr. Med. J.
20
,
247
,
1
8
17
Farhat
K.
,
Waheed
A.
,
Azhar
H.
,
Pasha
A.
,
Ismail
M.
and
Mansoor
Q.
(
2014
)
Genotype frequencies of C1236t polymorphism of the ABCB1 gene in various ethnic groups of Pakistan
.
Ann. Pak. Inst. Med. Sci.
10
,
3
6
18
Onodera
M.
,
Endo
K.
,
Kakuta
Y.
,
Kuroha
M.
,
Kimura
T.
,
Hiramoto
K.
et al. 
(
2017
)
ATP-binding cassette subfamily B member 1 1236C/T polymorphism significantly affects the therapeutic outcome of tacrolimus in patients with refractory ulcerative colitis
.
J. Gastroenterol. Hepatol.
,
19
Sun
F.
,
Cao
B.-Q.
,
Wang
B.
,
Wu
S.-Q.
and
Jiang
D.-H.
(
2016
)
Association between ABCB1 genetic polymorphism and the effect on epilepsy following phenytoin treatment
.
Exp. Ther. Med.
12
,
1780
1784
20
Kosztyu
P.
,
Dolezel
P.
,
Vaclavikova
R.
and
Mlejnek
P.
(
2015
)
Can the assessment of ABCB1 gene expression predict its function in vitro?
Eur. J. Haematol.
95
,
150
159
21
Palmieri
O.
,
Latiano
A.
,
Valvano
R.
,
D’Incà
R.
,
Vecchi
M.
,
Sturniolo
G.C.
et al. 
(
2005
)
Multidrug resistance 1 gene polymorphisms are not associated with inflammatory bowel disease and response to therapy in Italian patients
.
Aliment. Pharmacol. Ther.
22
,
1129
1138
22
Manna
I.
,
Gambardella
A.
,
Labate
A.
,
Mumoli
L.
,
Ferlazzo
E.
,
Pucci
F.
et al. 
(
2015
)
Polymorphism of the multidrug resistance 1 gene MDR1/ABCB1 C3435T and response to antiepileptic drug treatment in temporal lobe epilepsy
.
Seizure
24
,
124
126
23
Campos
H.S.
(
1999
)
Mycobacterium tuberculosis resistente: de onde vem a resistência?
Bol. Pneumol. Sanit.
7
,
51
64
24
Ullah
I.
,
Javaid
A.
,
Tahir
Z.
,
Ullah
O.
,
Shah
A.A.
,
Hasan
F.
et al. 
(
2016
)
Pattern of drug resistance and risk factors associated with development of drug resistant Mycobacterium tuberculosis in Pakistan
.
PLoS ONE
11
,
e0147529
25
Schön
T.
,
Miotto
P.
,
Köser
C.U.
,
Viveiros
M.
,
Böttger
E.
and
Cambau
E.
(
2017
)
Mycobacterium tuberculosis drug-resistance testing: challenges, recent developments and perspectives
.
Clin. Microbiol. Infect.
23
,
154
160
26
Mulu
W.
,
Mekonnen
D.
,
Yimer
M.
,
Admassu
A.
and
Abera
B.
(
2015
)
Risk factors for multidrug resistant tuberculosis patients in Amhara National Regional State
.
Afr. Health Sci.
15
,
368
77
27
Lachi
T.
and
Nakayama
M.
(
2015
)
Radiological findings of pulmonary tuberculosis in indigenous patients in Dourados, MS, Brazil
.
Radiol. Bras.
48
,
275
281