SLC19A1, SLC46A1 and SLCO1B1 Polymorphisms as Predictors of Methotrexate-Related Toxicity in Portuguese Rheumatoid Arthritis Patients

Abstract

Methotrexate (MTX) is used for rheumatoid arthritis (RA) treatment showing a wide toxicity profile. This study aimed to evaluate the influence of single nucleotide polymorphisms (SNPs) in genes encoding for MTX transporters with the occurrence of MTX-related toxicity (overall and gastrointestinal). A total of 233 Portuguese RA patients were genotyped for 23 SNPs. Haplotype analyses were performed and a toxicogenetic risk index (TRI) was created for SNPs that revealed to be statistically significant. Regarding MTX overall toxicity, an increased risk was associated to SLC19A1 rs7499 G carriers (p = 0.017), SLC46A1 rs2239907 GG (p = 0.030) and, SLCO1B1 rs4149056 T carriers (p = 0.040) and TT (p = 0.019). TRI revealed that patients with Index 3 were 18-fold more likely to present an adverse drug reaction when compared to those with Index 1 (p = 0.001). For MTX gastrointestinal toxicity, results demonstrated an increased risk associated with SLC19A1 rs7499 G carriers (p = 0.012) and GG (p = 0.045), SLC19A1 rs1051266 G carriers (p = 0.034), SLC19A1 rs2838956 A carriers (p = 0.049) and, SLCO1B1 rs4149056 T carriers (p = 0.042) and TT (p = 0.025). Haplotype analyses showed association between GGAG haplotype for SLC19A1 rs7499, rs1051266, rs2838956 and rs3788200 with MTX gastrointestinal toxicity (p = 0.029). TRI revealed that patients with Index 4 were 9-fold more likely to present a gastrointestinal disorder when compared to those with Index 1 (p = 0.020). This study demonstrated that SLC19A1, SLC46A1 and SLCO1B1 genotypes may help to identify patients with increased risk of MTX-related overall toxicity and that SLC19A1 and SLCO1B1 genotypes, and SLC19A1 haplotypes may help to identify patients with increased risk of MTX-related gastrointestinal toxicity.

methotrexate, polymorphisms, rheumatoid arthritis, transporters, toxicity, toxicogenetic risk index

Abbreviations

Abbreviations
 
• A

  adenine

 
• aa

  amino acid

 
• Ala

  alanine

 
• Arg

  arginine

 
• ABC

  ATP-binding cassette

 
• ACR

  American College of Rheumatology

 
• ADR

  adverse drug reaction

 
• BCRP

  breast cancer resistance protein

 
• C

  cytosine

 
• Chr

  chromosome

 
• CI

  confidence interval

 
• CTCAE

  Common Terminology Criteria for Adverse Events

 
• DAS28

  Disease Activity Score in 28 joints

 
• DMARDs

  disease-modifying antirheumatic drugs

 
• EDTA

  ethylenediaminetetraacetic acid

 
• eGFR

  estimated glomerular filtration rate

 
• EULAR

  European League Against Rheumatism

 
• FOLT

  folate transporter

 
• G

  guanine

 
• Gln

  glutamine

 
• Gly

  glycine

 
• HCP1

  heme carrier protein 1

 
• His

  histidine

 
• ID

  identification

 
• Ile

  isoleucine

 
• Lys

  lysine

 
• LST1

  liver-specific transporter 1

 
• MCT2

  monocarboxylic acid – transporter 2

 
• MDRD

  Modification of Diet in Renal Disease

 
• MDR

  multidrug resistance protein

 
• MRP

  multidrug resistance-associated protein

 
• MTX

  methotrexate

 
• NSAIDs

  non-steroidal anti-inflammatory drugs

 
• OAT

  organic anion transporter

 
• OR

  odds ratio

 
• P-GP

  P-glycoprotein

 
• PCFT

  proton-coupled folate transporter

 
• PK

  pharmacokinetics

 
• RA

  rheumatoid arthritis

 
• Ref

  reference

 
• RFC1

  reduced folate carrier 1

 
• SCr

  serum creatinine

 
• Ser

  serine

 
• SNP

  single nucleotide polymorphism

 
• SLC

  solute carrier

 
• SOC

  System Organ Class

 
• T

  thymine

 
• Thr

  threonine

 
• TRI

  toxicogenetic risk index

 
• UTR

  untranslated region

 
• Val

  valine

Rheumatoid arthritis (RA) is a complex, systemic autoimmune disease, characterized by a chronic inflammation of multiple peripheral joints (Smith et al., 2011), with a worldwide prevalence of 0.3–1.1% and incidence of 9–60 cases per 100,000 inhabitants (Silman and Pearson, 2002). In Portugal, the prevalence of this disease is 0.36% and the incidence is between 20-40 cases per 100,000 inhabitants (Branco and Canhao, 2011; Dias, 2001).

Methotrexate (MTX) is currently the most widely used disease-modifying antirheumatic drug (DMARD) for RA treatment (Benucci et al., 2011) in doses up to 25 mg per week. Nevertheless, treatment with MTX is not devoid of drawbacks and significant adverse drug reactions (ADRs) can occur due to interpatient variability (Benucci et al., 2011; Kremer, 2004). This variability can be consequence of MTX pharmacokinetics (PK) changes partly due to single nucleotide polymorphisms (SNPs) in genes encoding for MTX membrane transporter proteins—influx and/or efflux (Bohanec Grabar et al., 2008, 2012; Chatzikyriakidou et al., 2007; Moncrieffe et al., 2010; Plaza-Plaza et al., 2012; Trevino et al., 2009). MTX transporters are expressed in several tissues (International Transporter et al., 2010; Mikkelsen et al., 2011; Qiu et al., 2006) which can affect absorption, distribution and/or elimination. In gastrointestinal tract, at apical membrane of enterocytes, MTX is absorbed through active transport mediated by solute carrier (SLC) family 19 member 1 (SLC19A1) and possibly by SLC family 46 member 1 (SLC46A1) (Qiu et al., 2006; Swierkot and Szechinski, 2006; Tian and Cronstein, 2007). As SLC19A1-mediated transport is saturable, MTX may also be transported by folate receptor alpha (FOLR1) (Spinella et al., 1995; Tian and Cronstein, 2007). Moreover, MTX effluxes from enterocytes to intestinal tract lumen can be mediated by ATP-binding cassette (ABCs) transporters, i.e., by ABC subfamily C member 2 (ABCC2), ABC subfamily B member 1 (ABCB1) and ABC subfamily G member 2 (ABCG2), or to bloodstream by ABC subfamily C member 1 (ABCC1) and ABC subfamily C member 3 (ABCC3) (Mikkelsen et al., 2011). MTX hepatic uptake involves SLC19A1, SLC organic anion transporter family member 1B1 (SLCO1B1) and SLCO family member 1B3 (SLCO1B3) (Mikkelsen et al., 2011); and, most of the MTX in hepatocytes reenters in bloodstream by ABCC3 and ABC subfamily C member 4 (ABCC4) and only a small portion is excreted into the bile duct by ABCC2, ABCB1, and ABCG2 (International Transporter et al., 2010; Mikkelsen et al., 2011). MTX clearance is mainly through renal glomerular filtration and active secretion over the proximal tubular cells (Mikkelsen et al., 2011). Several renal transporters have an affinity for MTX, allowing MTX influx into renal cells by SLC family 22 member 6 (SLC22A6) and SLC family 22 member 8 (SLC22A8), in basolateral membrane, and by SLC family 22 member 11 (SLC22A11) and SLCO family member 1A2 (SLCO1A2), in apical membrane (Inoue and Yuasa, 2014; Mikkelsen et al., 2011). Moreover, SLC family 16 member 7 (SLC16A7) has been described as having a low to moderate expression in plasma membrane of tubular cells, but its function on MTX transport remains unclear (Halestrap, 2013). Furthermore, MTX excretion through urinary tract can be mediated by ABCB1, ABCC2, ABCC4 and ABCG2 (Benucci et al., 2011; Swierkot and Szechinski, 2006) (Fig. 1). Since MTX therapeutic outcome can be conditioned by PK changes we aimed to elucidate the influence of SNPs in genes encoding for MTX membrane transporter proteins on the occurrence of MTX-related toxicity in Portuguese RA patients.

FIG. 1.

Methotrexate membrane transporter proteins location and representation of their influx and/or efflux function. ABCB1: ATP-binding cassette subfamily B member 1; ABCC1: ATP-binding cassette subfamily C member 1; ABCC2: ATP-binding cassette subfamily C member 2; ABCC3: ATP-binding cassette subfamily C member 3; ABCC4: ATP-binding cassette subfamily C member 4; ABCG2: ATP-binding cassette subfamily G member 2; FOLR1: folate receptor alpha; SLC16A7: solute carrier family 16 member 7; SLC19A1: solute carrier family 19 member 1; SLC22A6: solute carrier family 22 member 6; SLC22A8: solute carrier family 22 member 8; SLC22A11: solute carrier family 22 member 11; SCL46A1: solute carrier family 46 member 1; SLCO1A2: solute carrier organic anion transporter family member 1A2; SLCO1B1: solute carrier organic anion transporter family member 1B1; SLCO1B3: solute carrier organic anion transporter family member 1B3.

MATERIALS AND METHODS

Study design

A retrospective study was performed between January 2009 and December 2012 at São João Hospital Center (Porto, Portugal) in a cohort of consecutive Portuguese Caucasian patients with RA treated with MTX. All patients (≥18 years) had to meet the 2010 revised classification criteria of American College of Rheumatology (ACR) and European League Against Rheumatism (EULAR) (Aletaha et al., 2010) and received MTX for at least 1 month. Other concomitant drugs, such as corticosteroids, non-steroidal anti-inflammatory drugs (NSAIDs) and other DMARDs were allowed during the study. Patients were excluded from the study if had drug abuse history, recent pregnancy or desire to become pregnant during the study. The study was approved by the local research ethics committee (reference 33/2009) and informed written consent was obtained from all patients according to the standards of Helsinki Declaration.

Clinical assessments

Patient demographics, clinicopathological and treatment characteristics were collected from clinical records. Clinical efficacy was assessed using the Disease Activity Score in 28 joints (DAS28) as described by Prevoo et al. (1995). Estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease (MDRD) equation available from http://egfrcalc.renal.org/. MTX-related toxicity was defined when patients presented any ADR related to MTX. At the time of each visit, physician directly asked the patient whether MTX-related ADRs were currently occurring and ADRs were recorded. The type of ADR was classified in System Organ Class (SOC) disorders, in accordance with the Common Terminology Criteria for Adverse Events (CTCAE) (U.S. Department of Health and Human Services, 2010). Due to the well-known protective effect of folic acid supplementation for the prevention of toxicity occurrence (Lima et al., 2013b; Ortiz et al., 2000), this drug was prescribed to all patients and their regular compliance was registered.

Sample processing

Whole blood samples were obtained with standard venipuncture technique in ethylenediaminetetraacetic acid containing tubes. Genomic deoxyribonucleic acid (DNA) was extracted with QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany), according to the manufacturer's instructions, and quantified using the NanoDrop 1000 Spectrophotometer v3.7 (ThermoScientific, Wilmington, DE).

SNPs selection and genotyping procedures

A total of 23 SNPs in 10 genes encoding for MTX membrane transporter proteins were selected based on their putative effects on MTX transport function and/or MTX-related toxicity (Table 1). Sequenom Assay Design 3.1 software was used to design the primers and genotyping was performed according to standard Sequenom iPLEX protocol (Bradic et al., 2011). Genomic sequence was amplified by multiplex polymerase chain reaction, the amplified product was treated with Shrimp alkaline phosphatase and used for allele specific primer extension (iPLEX) reaction (Sequenom, Sand Diego, CA). Reaction mixture was then spotted onto a SpectroCHIP microarray and subjected to the matrix-assisted laser desorption/ionization time-of-fight mass spectrometry (MALDI-TOF MS). Genotypes were assigned based on the presence of mass peaks by the MassARRAY Typer v4.0 software (Sequenom) (Bradic et al., 2011). Results were manually inspected and verified, using the MassARRAY Typer Analyzer v4.0 software (Sequenom). For quality control, 10% of the samples were randomly selected for a second analysis and results were 100% concordant.

Characteristics and Putative Effects of 23 SNPs in MTX Transport Function and Related Toxicity

TABLE 1.

Characteristics and Putative Effects of 23 SNPs in MTX Transport Function and Related Toxicity

Genes	Protein name	Chr	rs ID	Alleles	Region (aa change)	Putative effects		Reference
						MTX transport function	MTX-related toxicity
Solute carriers
SLC16A7	MCT2	12	rs3763980	A>T	Exon (Thr445Ser)	T allele: ↑ influx	—	(Halestrap, 2013)
			rs10877333	T>G	Intron	—	—	—
SLC19A1	RFC1 FOLT	21	rs7499	G>A	3′UTR	—	n.a.	(Hinken et al., 2011; Owen et al., 2012)
			rs1051266	G>A	Exon (Arg27His)	AA: ↑ influx	A carriers: ↓ toxicity	(Baslund et al., 2008; Bohanec Grabar et al., 2008, 2012; Chatzikyriakidou et al., 2007; Lima et al., 2013b)
			rs2838956	A>G	Intron	—	A carriers: ↑ toxicity	(Bohanec Grabar et al., 2012)
			rs3788200	G>A	Intron	AA: ↓ influx	—	(O'Byrne et al., 2010)
SLC22A6	OAT1	11	rs11568626	C>T	Exon (Arg50His)	TT: ↓ influx	—	(Bleasby et al., 2005; Sun et al., 2001)
SLC22A11	OAT4	11	rs11231809	T>A	Intron	AA: ↓ clearance	—	(Vormfelde et al., 2006)
SLC46A1	HCP1	17	rs2239907	G>A	3′UTR	—	—	—
	PCFT
SLCO1B1	LST1 OATP1B1	12	rs4149056	T>C	Exon (Val174Ala)	CC: ↓ influx and clearance	—	(Konig et al., 2000; Trevino et al., 2009)
ATP-binding cassette transporters
ABCB1	MDR P-GP	7	rs1045642	C>T	Exon (Ile1145Ile)	TT: ↓ efflux	TT: ↑ toxicity	(Bohanec Grabar et al., 2008; Kooloos et al., 2010; Wang et al., 2005)
			rs1128503	C>T	Exon (Gly412Gly)	—	TT: ↑ toxicity	(Ranganathan et al., 2008)
			rs2032582	G>A/T	Exon (Ala899Ser/Thr)	A/T alleles: ↓ efflux	n.a.	(Bohanec Grabar et al., 2008)
ABCC1	MRP1	16	rs35592	T>C	Intron	—	n.a.	(Berggren et al., 2007; Stamp et al., 2010)
			rs246240	A>G	Intron	—	—	—
			rs2074087	G>C	Intron	—	n.a.	(Ranganathan et al., 2008)
			rs2230671	C>G	Exon (Ser1334Ser)	GG: ↑ efflux	n.a.	(Ranganathan et al., 2008)
			rs3784864	G>A	Intron	—	—	—
ABCC2	MRP2	10	rs717620	G>A	5′UTR	AA: ↑ clearance	n.a.	(Ranganathan et al., 2008; Simon et al., 2013)
			rs4148396	C>T	Intron	—	CC: ↑ toxicity	(Ranganathan et al., 2008; Stamp et al., 2010)
ABCG2	BCRP	4	rs2231142	C>A	Exon (Gln141Lys)	AA: ↓ efflux	AA: ↑ toxicity	(Stamp et al., 2010; Zhang et al., 2013)
			rs13120400	T>C	Intron	—	—	—
			rs17731538	G>A	Intron	—	n.a.	(Stamp et al., 2010)

—: Unknown; n.a.: not associated.

A: adenine; aa: amino acid; Ala: alanine; Arg: arginine; ABC: ATP-binding cassette; BCRP: breast cancer resistance protein; C: cytosine; Chr: chromosome; FOLT: folate transporter; G: guanine; Gln: glutamine; Gly: glycine; HCP1: heme carrier protein 1; His: histidine; ID: identification; Ile: isoleucine; Lys: lysine; LST1: liver-specific transporter 1; MCT2: monocarboxylic acid – transporter 2; MDR: multidrug resistance protein; MRP: multidrug resistance-associated protein; MTX: methotrexate; OAT: organic anion transporter; PCFT: proton-coupled folate transporter; P-GP: P-glycoprotein; Ref: reference; RFC1: reduced folate carrier 1; Ser: serine; SNP: single nucleotide polymorphism; SLC: solute carrier; T: thymine; Thr: threonine; UTR: untranslated region; Val: valine.

Statistical analysis

Statistical analyses were performed with either IBM SPSS Statistics for Windows, Version 20.0 (IBM Corp., Armonk, NY) and SNPStats software (Sole et al., 2006). Genotype frequencies were assessed and tested for Hardy-Weinberg equilibrium (HWE). SNPs were excluded from analysis when genotyping call rates were <95% and when minor allele frequency was <10.0%. The power of the analysis for statistically significant cases was performed with the online software tool OpenEpi (available from www.openepi.com). Binary logistic regression analysis was used to identify which genotypes were associated with MTX-related toxicity. Analysis was performed adjusting to potential confounders, i.e., clinicopathological variables possibly influencing MTX-related toxicity, selected based on literature review and clinical significance (Halilova et al., 2012; Lima et al., 2013b; Morel and Combe, 2005). These variables included: (1) patient-related: gender, age, smoking and renal function (eGFR and serum creatinine—SCr); (2) disease-related: diagnosis age and disease duration; and, (3) treatment-related: folic acid, corticosteroids, NSAIDs, other DMARDs and MTX administration characteristics (dose, treatment duration and administration route). Haplotype analysis was performed to assess possible consequences on the phenotype by the co-presence of several variants of the same gene. Linkage disequilibrium (LD) between SNPs in the same gene was estimated and expressed as D′ coefficients. The measure was interpretable as the proportion of the maximum possible level of association between two loci, given the allele frequencies, ranging from 0 (linkage equilibrium) to 1 (complete LD) (Schaid et al., 2002). Haplotype frequencies were estimated by SNPStats software. Possible haplotypes were tested for association with MTX-related toxicity by taking the most frequent haplotype as reference. Rare haplotypes (estimated haplotype frequency <2.0%) were excluded. A toxicogenetic risk index (TRI) was created for each patient by the sum of risk genotypes from the SNPs that revealed to be statistically significant for MTX-related toxicity. Results were expressed as odds ratios (OR) with 95% confidence intervals (CI) and considering a probability (p) value of 5% or less as statistically significant.

RESULTS

Patients’ characteristics

This study included follow-up data from 233 patients, 196 (84.1%) females and 37 (15.9%) males, with a median age of 52.0 (26.0–87.0) years old, of which 32 (13.7%) were smokers. In this population, the median of SCr was 8.20 mg/l (4.00–19.80), the median of eGFR was 82.0 ml/min/1.73 m² (29.00–186.00) and 30 patients (12.9%) presented chronic renal insufficiency (eGFR < 60 ml/min/1.73 m²). Considering disease-related variables, the mean diagnosis age was 40.3 ± 13.2 years old and the median disease duration was 10.0 years (0.3–51.0). Only 136 patients (58.4%) used MTX as unique DMARD, while 97 patients (41.6%) were treated with MTX combined with other synthetic or biological DMARDs. MTX was administered by per os (PO) in 210 (90.1%) and by subcutaneous route (SC) in 23 (9.9%) of the patients, and the median MTX treatment duration was 47.0 months (1.0–240.0) with a median dose of 15.0 mg/week (2.5–25.0). Regarding concomitant drugs, other than MTX, 188 (80.7%) patients were under corticosteroids therapy, 170 patients (73.0%) used NSAIDs and 118 patients (50.6%) complied regularly with folic acid supplementation. MTX-related toxicity was registered in 77 (33.0%) patients. Figure 2 represents the observed ADRs classified in SOCs disorders according to CTCAE.

FIG. 2.

Observed MTX-related ADRs by SOCs disorders. Gastrointestinal disorders included anorexia, abdominal distension, diarrhea, dyspepsia, nauseas, stomach pain, and/or vomiting. Skin and subcutaneous tissue disorders included alopecia, rash maculo-papular and rheumatoid nodulosis exacerbation. Hepatobiliary disorders included transaminases serum elevation. Respiratory, thoracic and mediastinal disorders included hypersensitivity pneumonitis.

MTX Transporters SNPs: Genotypes and Haplotypes characteristics

Genotypes distribution of studied SNPs is represented in Table 2. SLC22A6 rs11568626 and ABCG2 rs2231142 were excluded from analysis since minor allele frequency was <10.0% and ABCC1 rs2230671 was excluded from analysis because genotyping call rates were <95%. Taking this into account, 20 SNPs were considered. For ABCB1 rs2032582, three patients (1.3%) presented AG genotype and three patients (1.3%) presented AT genotype. Given the low frequency, these six patients were excluded from the analysis. Genotypes distribution was in HWE (p > 0.050) except for the SLC19A1 G>A, rs1051266. Figure 3 demonstrates the estimated D′ coefficients of the studied SNPs for the possible haplotypes in the same gene. SLC16A7, SLC19A1, ABCB1, ABCC1, ABCC2, and ABCG2 SNPs were in LD, except for ABCC1 rs246240 and rs2074087.

FIG. 3.

Estimated D′ coefficients and p-values for possible haplotypes. Dark gray boxes represent the single nucleotide polymorphisms that were in linkage disequilibrium. A: adenine; ABCB1: ATP-binding cassette subfamily B member 1; ABCC1: ATP-binding cassette subfamily C member 1; ABCC2: ATP-binding cassette subfamily C member 2; ABCG2: ATP-binding cassette subfamily G member 2; C: cytosine; G: guanine; SLC16A7: solute carrier family 16 member 7; SLC19A1: solute carrier family 19 member 1; T: thymine.

Genotype distribution for the 23 studied SNPs

TABLE 2.

Genotype distribution for the 23 studied SNPs

Genes	rs ID	Alleles	Genotype frequency
			Ancestral allele homozygotes	Heterozygotes	Minor allele homozygotes
Solute carriers
SLC16A7	rs3763980	A>T	119 (51.1)	92 (39.5)	22 (9.4)
	rs10877333	T>G	161 (69.1)	68 (29.2)	4 (1.7)
SLC19A1	rs7499	G>A	91 (39.1)	104 (44.6)	38 (16.3)
	rs1051266	G>A	80 (34.3)	91 (39.1)	62 (26.6)
	rs2838956	A>G	82 (35.2)	109 (46.8)	42 (18.0)
	rs3788200	G>A	82 (35.2)	109 (46.8)	42 (18.0)
SLC22A6	rs11568626a	C>T	232 (99.6)	1 (0.4)	0 (0.0)
SLC22A11	rs11231809	T>A	80 (34.3)	120 (51.5)	33 (14.2)
SLC46A1	rs2239907	G>A	87 (37.3)	99 (42.5)	47 (20.2)
SLCO1B1	rs4149056	T>C	169 (72.5)	16 (6.9)	48 (20.6)
ATP-binding cassette transporters
ABCB1	rs1045642	C>T	63 (27.0)	109 (46.8)	61 (26.2)
	rs1128503	C>T	77 (33.0)	107 (46.0)	49 (21.0)
	rs2032582b	G>A/T	81 (34.8)	99 (42.5)	47 (20.1)
ABCC1	rs35592	T>C	124 (53.2)	86 (36.9)	23 (9.9)
	rs246240	A>G	159 (68.3)	66 (28.3)	8 (3.4)
	rs2074087	G>C	146 (62.7)	76 (32.6)	11 (4.7)
	rs2230671a	C>G	117 (92.9)	9 (7.1)	0 (0.0)
	rs3784864	G>A	67 (28.8)	111 (47.6)	55 (23.6)
ABCC2	rs717620	G>A	135 (57.9)	91 (39.1)	7 (3.0)
	rs4148396	C>T	75 (32.2)	109 (46.8)	49 (21.0)
ABCG2	rs2231142a	C>A	200 (85.8)	33 (14.2)	0 (0.0)
	rs13120400	T>C	116 (49.8)	96 (41.2)	21 (9.0)
	rs17731538	G>A	137 (58.8)	87 (37.3)	9 (3.9)

^aSNPs were excluded from analysis when genotyping call rates were <95% and when minor allele frequency was <10.0%.

^bSix patients were excluded from the analyses because three (1.3%) presented AG genotype and three (1.3%) presented AT genotype.

A: adenine; ABC: ATP-binding cassette; C: cytosine; G: guanine; SLC: solute carrier; SLCO: solute carrier organic anion transporter; SNP: single nucleotide polymorphism; T: thymine.

MTX Transporters SNPs: Genotypes and MTX-related Toxicity

Tables 3 and 4 represent the relation between MTX-related toxicity and SNPs in SLCs and ABCs, respectively. Regarding MTX overall toxicity, our results demonstrated that for SNPs in SLCs, SLC19A1 rs7499 G carriers (p = 0.017, OR = 3.72), SLC46A1 rs2239907 G homozygotes (p = 0.030, OR = 2.32) and, SLCO1B1 rs4149056 T carriers and T homozygotes (p = 0.040, OR = 2.78; and p = 0.019, OR = 2.82; respectively) were associated with an increased risk for MTX-related overall toxicity. Considering the MTX gastrointestinal toxicity, similar results were shown for SNPs in SLCs. SLC19A1 rs7499 G carriers and G homozygotes (p = 0.012, OR = 5.64; and p = 0.045, OR = 2.39; respectively), SLC19A1 rs1051266 G carriers (p = 0.034, OR = 3.07), SLC19A1 rs2838956 (p = 0.049, OR = 3.21) and, SLCO1B1 rs4149056 T carriers and T homozygotes (p = 0.042, OR = 3.09; and p = 0.025, OR = 2.92, respectively) were associated with MTX-related gastrointestinal toxicity. For statistically significant cases, including the 233 enrolled patients, the power of the analysis was of 80% for the majority of the polymorphisms. No statistically significant differences were observed in relation to the SNPs in ABCs for MTX-related overall and/or gastrointestinal toxicity.

Relation between SNPs in Solute Carriers and MTX-related Toxicity

TABLE 3.

Relation between SNPs in Solute Carriers and MTX-related Toxicity

	Model	Alleles	MTX overall toxicity		MTX gastrointestinal toxicity
			p	OR (95% CI)	p	OR (95% CI)
SLC16A7 A>T (rs3763980)	Dominant	A carriers		Reference		Reference
		TT	0.581	0.72 (0.22–2.35)		0.70 (0.19–2.55)
	Recessive	AA		Reference		Reference
		T carriers	0.447	1.34 (0.63–2.85)	0.539	1.29 (0.57–2.91)
SLC16A7 T>G (rs10877333)	Dominant	T carriers		Reference		Reference
		GG	0.636	0.53 (0.04–7.31)	0.843	0.77 (0.06–10.00)
	Recessive	TT		Reference		Reference
		G carriers	0.761	1.23 (0.52–2.42)	0.842	1.09 (0.48–2.43)
SLC19A1 G>A (rs7499)	Dominant	G carriers		Reference		Reference
		AA	0.017	0.27 (0.09–0.79)a	0.012	0.18 (0.05–0.69)b
	Recessive	GG		Reference		Reference
		A carriers	0.109	0.54 (0.25–1.15)	0.045	0.43 (0.19–0.98)c
SLC19A1 G>A (rs1051266)	Dominant	G carriers		Reference		Reference
		AA	0.150	0.53 (0.23–1.26)	0.034	0.33 (0.12–0.92)d
	Recessive	GG		Reference		Reference
		A carriers	0.690	0.86 (0.40–1.82)	0.512	0.77 (0.35–1.70)
SLC19A1 A>G (rs2838956)	Dominant	A carriers		Reference		Reference
		GG	0.055	0.37 (0.14–1.02)	0.049	0.31 (0.10–1.00)e
	Recessive	AA		Reference		Reference
		G carriers	0.170	0.60 (0.28–1.25)	0.126	0.54 (0.24–1.19)
SLC19A1 G>A (rs3788200)	Dominant	G carriers		Reference		Reference
		AA	0.140	0.48 (0.18–1.27)	0.078	0.36 (0.12–1.12)
	Recessive	GG		Reference		Reference
		A carriers	0.158	0.58 (0.28–1.23)	0.113	0.53 (0.24–1.16)
SLC22A11 T>A (rs11231809)	Dominant	T carriers		Reference		Reference
		AA	0.806	0.89 (0.34–2.33)	0.506	0.70 (0.24–2.02)
	Recessive	TT		Reference		Reference
		A carriers	0.906	0.95 (0.43–2.09)	0.716	1.17 (0.50–2.78)
SLC46A1 G>A (rs2239907)	Dominant	G carriers		Reference		Reference
		AA	0.605	0.79 (0.32–1.94)	0.859	0.92 (0.36–2.33)
	Recessive	GG		Reference		Reference
		A carriers	0.030	0.43 (0.20–0.92)f	0.155	0.55 (0.24–1.25)
SLCO1B1 T>C (rs4149056)	Dominant	T carriers		Reference		Reference
		CC	0.040	0.36 (0.14–0.96)g	0.042	0.32 (0.11–0.96)h
	Recessive	TT		Reference		Reference
		C carriers	0.019	0.36 (0.15–0.84)i	0.025	0.34 (0.13–0.88)j

Notes. p-value <0.05 is considered to be of statistically significance (highlighted in bold). p-value, odds ratio (OR) and 95% confidence intervals (CI) correspond to multivariate logistic regression adjusted to patient-related variables (gender, age, smoking, eGFR and SCr), disease-related variables (diagnosis age and disease duration), and treatment-related variables (folic acid, corticosteroids, NSAIDs, other concomitant DMARDs and MTX administration characteristics, such as dose, treatment duration and administration route).

^aWhen reference was AA genotype: OR = 3.72, 95% CI: 1.26–10.93.

^bWhen reference was AA genotype: OR = 5.64, 95% CI: 1.45–21.86.

^cWhen reference was A carriers: OR = 2.39, 95% CI: 1.06–5.38.

^dWhen reference was AA genotype: OR = 3.07, 95% CI: 1.09–8.65.

^eWhen reference was GG genotype: OR = 3.21, 95% CI: 1.00–10.25.

^fWhen reference was A carriers: OR = 2.32, 95% CI: 1.08–4.97.

^gWhen reference was CC genotype: OR = 2.78, 95% CI: 1.05–7.39.

^hWhen reference was CC genotype: OR = 3.09, 95% CI: 1.04–9.17.

ⁱWhen reference was C carriers: OR = 2.82, 95% CI: 1.18–6.72.

^jWhen reference was C carriers: OR = 2.92, 95% CI: 1.14–7.46.

A: adenine; C: cytosine; CI: confidence interval; eGFR: estimated glomerular filtration rate; G: guanine; MTX: methotrexate; OR: odds ratio; SCr: serum creatinine; SLC: solute carrier; SLCO: solute carrier organic anion transporter; SNP: single nucleotide polymorphism; T: thymine.

Relation between SNPs in ATP-Binding Cassette Transporters and MTX-related Toxicity

TABLE 4.

Relation between SNPs in ATP-Binding Cassette Transporters and MTX-related Toxicity

	Model	Alleles	MTX overall toxicity		MTX gastrointestinal toxicity
			p	OR (95% CI)	p	OR (95% CI)
ABCB1 C>T (rs1045642)	Dominant	C carriers		Reference		Reference
		TT	0.515	0.76 (0.33–1.75)	0.475	0.72 (0.30–1.77)
	Recessive	CC		Reference		Reference
		T carriers	0.794	0.90 (0.41–1.98)	0.540	0.77 (0.34–1.77)
ABCB1 C>T (rs1128503)	Dominant	C carriers		Reference		Reference
		TT	0.481	1.36 (0.58–3.15)	0.777	1.14 (0.45–2.89)
	Recessive	CC		Reference		Reference
		T carriers	0.269	1.55 (0.71–3.37)	0.426	1.39 (0.62–3.14)
ABCB1 G>A/T (rs2032582)	Dominant	G carriers		Reference		Reference
		TT	0.833	0.91 (0.38–2.18)	0.652	0.80 (0.30–2.10)
	Recessive	GG		Reference		Reference
		T carriers	0.167	1.71 (0.80–3.66)	0.346	1.47 (0.66–3.25)
ABCC1 T>C (rs35592)	Dominant	T carriers		Reference		Reference
		CC	0.734	1.21 (0.40–3.73)	0.367	1.68 (0.55–5.15)
	Recessive	TT		Reference		Reference
		C carriers	0.349	1.40 (0.69–2.85)	0.141	1.79 (0.82–3.88)
ABCC1 A>G (rs246240)	Dominant	A carriers		Reference		Reference
		GG	0.978	0.97 (0.14–6.64)	0.356	0.28 (0.02–4.13)
	Recessive	AA		Reference		Reference
		G carriers	0.148	1.78 (0.82–3.86)	0.174	1.78 (0.77–4.11)
ABCC1 G>C (rs2074087)	Dominant	G carriers		Reference		Reference
		CC	0.864	1.15 (0.24–5.46)	0.716	1.34 (0.28–6.40)
	Recessive	GG		Reference		Reference
		C carriers	0.210	0.62 (0.29–1.31)	0.103	0.50 (0.22–1.15)
ABCC1 G>A (rs3784864)	Dominant	G carriers		Reference		Reference
		AA	0.733	1.16 (0.50–2.70)	0.975	0.99 (0.40–2.46)
	Recessive	GG		Reference		Reference
		A carriers	0.328	0.67 (0.30–1.49)	0.325	0.66 (0.28–1.52)
ABCC2 G>A (rs717620)	Dominant	G carriers		Reference		Reference
		AA	0.556	1.74 (0.28–10.84)	0.701	1.50 (0.19–11.89)
	Recessive	GG		Reference		Reference
		A carriers	0.722	0.88 (0.42–1.81)	0.672	0.84 (0.38–1.85)
ABCC2 C>T (rs4148396)	Dominant	C carriers		Reference		Reference
		TT	0.953	1.03 (0.43–2.48)	0.894	1.07 (0.41–2.78)
	Recessive	CC		Reference		Reference
		T carriers	0.995	1.00 (0.45–2.21)	0.562	1.29 (0.54–3.05)
ABCG2 T>C (rs13120400)	Dominant	T carriers		Reference		Reference
		CC	0.457	0.62 (0.18–2.18)	0.354	0.51 (0.12–2.12)
	Recessive	TT		Reference		Reference
		C carriers	0.069	1.99 (0.95–4.16)	0.052	2.21 (0.94–4.92)
ABCG2 G>A (rs17731538)	Dominant	G carriers		Reference		Reference
		AA	0.459	0.52 (0.10–2.89)	0.455	0.48 (0.07–3.31)
	Recessive	GG		Reference		Reference
		A carriers	0.510	0.78 (0.37–1.65)	0.557	0.78 (0.35–1.76)

Notes. p-value <0.05 is considered to be of statistically significance (highlighted in bold). p-value, odds ratio (OR) and 95% confidence intervals (CI) corresponds to multivariate logistic regression adjusted to patient-related variables (gender, age, smoking, eGFR and SCr), disease-related variables (diagnosis age and disease duration), and treatment-related variables (folic acid, corticosteroids, NSAIDs, other concomitant DMARDs and MTX administration characteristics, such as dose, treatment duration and administration route).

A: adenine; ABC: ATP-binding cassette; C: cytosine; CI: confidence interval; eGFR: estimated glomerular filtration rate; G: guanine; MTX: methotrexate; OR: odds ratio; SCr: serum creatinine; SNP: single nucleotide polymorphism; T: thymine

MTX Transporters SNPs: Haplotypes and MTX-related Toxicity

Table 5 represents the relation between MTX transporters genes haplotypes and MTX-related toxicity. Since ABCC1 rs246240 and rs2074087 were not in LD, analyses were performed considering the following combinations: (1) ABCC1 rs35592, rs246240 and rs3784864; and (2) ABCC1 rs35592, rs2074087 and rs3784864. Results showed that no statistically significant differences were observed in relation to MTX transporters genes haplotypes and MTX-related overall toxicity. Nevertheless, for MTX gastrointestinal toxicity, our results demonstrated that GGAG haplotype (constituted by ancestral alleles) for SLC19A1 rs7499, rs1051266, rs2838956 and rs3788200, respectively, was associated with MTX-related gastrointestinal toxicity when compared to AAGA haplotype (constituted by minor alleles) (p = 0.029). No statistically significant differences were observed for the remaining haplotypes.

Relation between MTX Transporters Genes Haplotypes and MTX-related Toxicity

TABLE 5.

Relation between MTX Transporters Genes Haplotypes and MTX-related Toxicity

Haplotype				Estimated frequency (%)	MTX overall toxicity		MTX gastrointestinal toxicity
					p	OR (95% CI)	p	OR (95% CI)
SLC16A7 A>T (rs3763980)	SLC16A7 T>G (rs10877333)
A	T			54.5		Reference		Reference
T	T			29.2	0.780	1.10 (0.58–2.08)	0.940	1.03 (0.51–2.05)
A	G			16.3	0.630	1.22 (0.55–2.70)	0.670	1.20 (0.53–2.72)
SLC19A1 G>A (rs7499)	SLC19A1 G>A (rs1051266)	SLC19A1 A>G (rs2838956)	SLC19A1G>A (rs3788200)
G	G	A	G	48.7		Reference		Reference
A	A	G	A	33.6	0.089	0.58 (0.31–1.08)	0.029	0.46 (0.23–0.92)
G	A	A	G	5.8	0.920	1.07 (0.25–4.52)	0.550	0.58 (0.09–3.55)
G	A	G	A	4.4	0.590	1.50 (0.34–6.56)	0.580	1.66 (0.28–10.02)
A	G	A	G	2.0	0.920	1.10 (0.19–6.27)	0.610	0.59 (0.07–4.64)
ABCB1 C>T (rs1045642)	ABCB1 C>T (rs1128503)	ABCB1 G>A/T (rs2032582)
C	C	G		43.7		Reference		Reference
T	T	T		37.5	0.530	1.21 (0.66–2.23)	0.740	1.12 (0.58–2.15)
T	C	G		10.7	0.084	0.33 (0.10–1.15)	0.100	0.35 (0.10–1.22)
C	T	T		3.9	0.670	1.40 (0.30–6.62)	0.720	1.35 (0.27–6.75)
C	T	G		2.7	0.920	0.91 (0.15–5.57)	0.770	0.75 (0.11–5.24)
ABCC1 T>C (rs35592)	ABCC1 G>C (rs2074087)	ABCC1G>A (rs3784864)
T	G	A		42.6		Reference		Reference
C	G	G		18.1	0.640	1.21 (0.54–2.69)	0.300	1.57 (0.68–3.62)
T	G	G		17.7	0.200	1.83 (0.73–4.56)	0.220	1.85 (0.69–4.94)
C	C	G		9.4	0.370	1.63 (0.56–4.74)	0.320	1.78 (0.57–5.48)
T	C	G		7.4	0.140	0.23 (0.03–1.62)	0.120	0.18 (0.02–1.51)
T	C	A		4.0	0.560	0.47 (0.04–6.06)	0.700	0.60 (0.05–7.75)
ABCC1 T>C (rs35592)	ABCC1 A>G (rs246240)	ABCC1G>A (rs3784864)
T	A	A		46.3		Reference		Reference
T	A	G		18.9	0.890	1.06 (0.44–2.54)	0.710	1.20 (0.46–3.09)
C	A	G		16.4	0.710	1.18 (0.52–2.62)	0.510	1.34 (0.56–3.18)
C	G	G		11.1	0.200	1.91 (0.71–5.16)	0.140	2.19 (0.78–6.12)
T	G	G		6.2	0.290	2.00 (0.56–7.11)	0.810	1.21 (0.25–5.96)
ABCC2 G>A (rs717620)	ABCC2C>T (rs4148396)
G	C			55.3		Reference		Reference
A	T			22.2	0.240	1.60 (0.74–3.48)	0.110	1.98 (0.85–4.59)
G	T			22.2	0.700	1.14 (0.60–2.17)	0.800	1.10 (0.55–2.20)
ABCG2 T>C (rs13120400)	ABCG2G>A (rs17731538)
T	G			48.6		Reference		Reference
C	G			28.7	0.900	1.04 (0.52–2.11)	0.840	1.08 (0.52–2.22)
T	A			21.7	0.210	0.58 (0.25–1.35)	0.290	0.65 (0.29–1.44)

Notes. p-value <0.05 is considered to be of statistically significance (highlighted in bold). p-value, odds ratio (OR) and 95% confidence intervals (CI) correspond to multivariate logistic regression adjusted to patient-related variables (gender, age, smoking, eGFR and SCr), disease-related variables (diagnosis age and disease duration), and treatment-related variables (folic acid, corticosteroids, NSAIDs, other concomitant DMARDs and MTX administration characteristics, such as dose, treatment duration and administration route).

A: adenine; ABC: ATP-binding cassette; C: cytosine; CI: confidence interval; DMARDs: disease modifying antirheumatic drugs; eGFR: estimated glomerular filtration rate; G: guanine; MTX: methotrexate; NSAIDs: non-steroidal anti-inflammatory drugs; OR: odds ratio; SCr: serum creatinine; SLC: solute carrier; T: thymine.

MTX Transporters SNPs: Toxicogenetic Risk Index for MTX

A TRI for MTX-related toxicity was created for each patient by the sum of risk genotypes from SNPs that revealed to be statistically significant for MTX-related toxicity (see Tables 3 and 4). The TRI was adjusted for potential confounders as described in the Methods section. Regarding MTX overall toxicity, the risk genotypes were as follow: SLC19A1 rs7499 G carriers, SLC46A1 rs2239907 G homozygotes and SLCO1B1 rs4149056 T carriers. Figure 4 represents the contribution of the TRI to the occurrence of MTX-related overall toxicity in RA patients treated with MTX. The number (%) of patients is given for each incremental unit of the Index. The TRI ranged from 0 to 3. An increased TRI value was associated with an increased incidence of ADRs (p = 0.020). Patients with Index 2 were 3.06× (95% CI: 1.04–8.99) more likely to present an ADR when compared with those with Index 1 (p = 0.042), patients with Index 3 were 3.52× (95% CI: 1.37–9.03) more likely to present an ADR when compared to those with Index 2 (p = 0.009) and patients with Index 3 were 18.79× (95% CI: 3.39–104.12) more likely to present an ADR when compared with those with Index 1 (p = 0.001). Regarding MTX gastrointestinal toxicity, the risk genotypes were as follow: SLC19A1 rs7499 G carriers, SLC19A1 rs1051266 G carriers, SLC19A1 rs2838956 A carriers and SLCO1B1 rs4149056 T carriers. Figure 5 represents the contribution of the TRI to the occurrence of MTX-related gastrointestinal toxicity in RA patients under MTX. The number (%) of patients is given for each incremental unit of the Index. The TRI ranged from 0 to 4 and an increased TRI value was associated with an increased incidence of gastrointestinal disorders (p = 0.010). No statistically significant differences were observed between patients with Index 1 and patients with Index 2 (p = 1.000) and between patients with Index 2 and patients exhibiting Index 3 (p = 0.429). Patients with Index 4 were 5.11× (95% CI: 1.56–16.70) more likely to present gastrointestinal disorders compared to those with Index 3 (p = 0.007) and patients with Index 4 were 9.50× (95% CI: 1.43–62.95) more likely to present gastrointestinal disorders when compared to those with Index 1 (p = 0.020).

FIG. 4.

(A) Contribution of the toxicogenetic risk index in MTX-related overall toxicity occurrence. (B) Percentage of RA patients with MTX-related ADRs in relation to toxicogenetic risk index for overall toxicity. Index 0 included SLC19A1 rs7499 A homozygotes + SLC46A1 rs2239907 T carriers + SLCO1B1 rs4149056 C homozygotes. Index 1 included SLC19A1 rs7499 G carriers, SLC46A1 rs2239907 G homozygotes or SLCO1B1 rs4149056 T carriers. Index 2 included SLC19A1 rs7499 G carriers + SLC46A1 rs2239907 G homozygotes; SLC19A1 rs7499 G carriers + SLCO1B1 rs4149056 T carriers; and, SLC46A1 rs2239907 G homozygotes + SLCO1B1 rs4149056 T carriers. Index 3 included SLC19A1 rs7499 G carriers + SLC46A1 rs2239907 G homozygotes + SLCO1B1 rs4149056 T carriers.

FIG. 5.

(A) Contribution of the toxicogenetic risk index in MTX-related gastrointestinal toxicity occurrence. (B) Percentage of RA patients with MTX-related gastrointestinal disorders in relation to toxicogenetic risk index for gastrointestinal toxicity. Index 0 included SLC19A1 rs7499 A homozygotes + SLC19A1 rs1051266 A homozygotes + SLC19A1 rs2838956 G homozygotes + SLCO1B1 rs4149056 C homozygotes. Index 1 included SLC19A1 rs7499 G carriers, SLC19A1 rs1051266 G carriers, SLC19A1 rs2838956 A carriers or SLCO1B1 rs4149056 T carriers. Index 2 included SLC19A1 rs7499 G carriers + SLC19A1 rs1051266 G carriers; SLC19A1 rs7499 G carriers + SLC19A1 rs2838956 A carriers; SLC19A1 rs7499 G carriers + SLCO1B1 rs4149056 T carriers; SLC19A1 rs1051266 G carriers + SLC19A1 rs2838956 A carriers; SLC19A1 rs1051266 G carriers + SLCO1B1 rs4149056 T carriers; and, SLC19A1 rs2838956 A carriers + SLCO1B1 rs4149056 T carriers. Index 3 included SLC19A1 rs7499 G carriers + SLC19A1 rs1051266 G carriers + SLC19A1 rs2838956 A carriers; SLC19A1 rs7499 G carriers + SLC19A1 rs1051266 G carriers + SLCO1B1 rs4149056 T carriers; SLC19A1 rs7499 G carriers + SLC19A1 rs2838956 A carriers + SLCO1B1 rs4149056 T carriers; and, SLC19A1 rs1051266 G carriers + SLC19A1 rs2838956 A carriers + SLCO1B1 rs4149056 T carriers. Index 4 included SLC19A1 rs7499 G carriers + SLC19A1 rs1051266 G carriers + SLC19A1 rs2838956 A carriers + SLCO1B1 rs4149056 T carriers.

DISCUSSION

MTX therapeutic outcome can be altered by several factors such as SNPs in genes encoding for MTX membrane transporter proteins. This study evaluated the influence of 23 SNPs in SLCs and ABCs MTX transporters as predictors of MTX-related toxicity in Portuguese RA patients. Hence, we have performed genotype and haplotype based approaches, considering a toxicogenetic risk index, with multivariate logistic regression analysis adjusted to potential confounders. Genotypes distribution of studied SNPs were in HWE and were similar to those previously described for Caucasian populations (de Rotte et al., 2012; Moncrieffe et al., 2010; Owen et al., 2013) and in the National Center for Biotechnology Information (NCBI) database, except for the SLC19A1 rs1051266. Regarding the observed ADRs, results were in accordance with literature and associated with weekly administration of MTX in low-dose, since the most frequent ADRs were gastrointestinal disorders (Bohanec Grabar et al., 2012; Kremer, 2004).

Genotypes and MTX-related Toxicity

SLC19A1 is a bidirectional transporter, described as being expressed in the majority of tissues, with relevance in enterocytes and hepatocytes (Hinken et al., 2011; Qiu et al., 2006). Accordingly to genotype analysis, our results demonstrated that SLC19A1 rs7499 G carriers were associated with an increased risk for MTX-related toxicity (overall and gastrointestinal). This SNP occurs in the 3′-untranslated region (UTR) in a region thought to be important for messenger ribonucleic acid (mRNA) stability, localization and translational efficiency (Lynch et al., 2005) and, therefore, important to membrane transporter expression. Nevertheless, its effect in this bidirectional transporter function is unknown. We propose that GG genotype could provide an increased influx capability leading to higher bioavailability, increased MTX tissues exposure and, consequently, to toxicity. In literature, only one study in Caucasian RA patients dealt with the influence of this SNP in MTX-related toxicity, but no association was found (Owen et al., 2013). Regarding SLC19A1 rs1051266, our results showed that G carriers were associated with MTX-related gastrointestinal toxicity but no associations were observed in accordance to MTX overall toxicity. Bohanec Grabar et al. have previously reported that G homozygotes had a higher risk for MTX overall toxicity (Bohanec Grabar et al., 2008, 2012) but did not clearly described the impact on gastrointestinal disorders, and the majority of other studies did not have associations with toxicity (Chatzikyriakidou et al., 2007; Owen et al., 2013; Plaza-Plaza et al., 2012). Thus, these controversial studies regarding to this SNP and MTX-related gastrointestinal toxicity need further clarification. We hypothesize that GG genotype could provide an increased influx capability leading to higher bioavailability, increased MTX tissues exposure, mainly in tissues where SLC19A1 is highly expressed and, consequently, to gastrointestinal toxicity. Furthermore, is important to consider if enterohepatic recirculation contributes to major differences in bioavailability, reinforcing the importance of understanding the role of transporters in this pathway. In addition, A carriers for SLC19A1 rs2838956, had a borderline trend toward significance for MTX-related toxicity (overall and gastrointestinal). This is possibly caused by an increased influx capability of A carriers which, consequently, leads to greater MTX tissues exposure and toxicity. Our results are in accordance with a previously report by Bohanec Grabar et al. that demonstrated a borderline significant trend toward MTX-related overall toxicity for A carriers, particularly for skin and subcutaneous tissue disorders (Bohanec Grabar et al., 2012), yet other study reported no statistically significant association (Owen et al., 2013). The impact of this variant in SLC19A1 is currently unknown and functional studies are essential since intronic SNPs can potentially influence RNA splicing, which may affect transporter structure and function (Wang and Cooper, 2007).

In relation to SLC46A1 rs2239907, G homozygotes revealed association with MTX overall toxicity. This SNP is also located in a 3′UTR region, which is thought to be important in mRNA stability, localization and translational efficiency (Lynch et al., 2005) and then considered as a potential functional SNP. However, the effect of this SNP in the transporter function is currently unknown. SLC46A1 is mostly expressed in apical membrane of enterocytes but also can be found in other cells (Qiu et al., 2006). Thus, we hypothesized that GG genotype could provide an increased influx leading to higher bioavailability and, consequently, to higher MTX tissues exposure and toxicity. Nevertheless, the complexity of these mechanisms and the presence of other factors that could also play a role, such as the putative contribution of enterohepatic recirculation as well as its possible influence in renal function deserves further consideration. To our best knowledge this is the first report to analyze the effect of this SNP with MTX-related toxicity in RA.

Regarding SLCO1B1 rs4149056, T carriers were associated with MTX-related toxicity (overall and gastrointestinal). T carriers have been associated with an increased membrane expression of SLCO1B1 and higher MTX influx and clearance (Trevino et al., 2009). Despite the SLCO1B1 transporter is mainly expressed on basolateral membrane of hepatocytes (Konig et al., 2000), its mRNA also has been detected in other tissues, including enterocytes (Glaeser et al., 2007), which can explain the MTX intracellular retention (gastrointestinal and hepatic) leading to cytotoxicity. The association of T allele with MTX-related toxicity has been previously described for high-dose MTX (Trevino et al., 2009) but this is the first report to analyze the influence of SLCO1B1 rs4149056 with MTX-related toxicity in RA.

Accordingly to ABCG2 rs13120400, C carriers had a borderline trend toward significance for MTX-related toxicity. ABCG2 transporter, located in apical membranes of enterocytes, hepatocytes and kidney tubular cells, is responsible for MTX efflux from the enterocytes to intestinal tract lumen and MTX excretion into bile and urine (Mikkelsen et al., 2011). Then, it is plausible to explain our results as follows: C carriers should cause a reduced efflux capability, which is translated, in less MTX elimination and higher MTX bioavailability, thus leading to toxicity. This is the first report to associate this SNP with MTX-related toxicity in RA patients.

Haplotypes and MTX-related Toxicity

Haplotypes may have a particular significance in regard to functionality or as genetic markers for unknown functional variants, claiming for full haplotypic information to be encompassed into studies in order to better characterize the role of a candidate gene (Hodge et al., 1999; Lima et al., 2013a). In fact, haplotypes constituted by SNPs with both unknown and known impact in transporter functions could provide a putative association of these, yet unknown variants, toward depicting the role of transporters function in toxicity development. Therefore, haplotype analysis was performed, to assess of possible consequences on the phenotype in the co-presence of several variants of the same gene. From haplotypes analysis, our results showed that GGAG haplotype for SLC19A1 rs7499, rs1051266, rs2838956 and rs3788200, was associated with MTX-related gastrointestinal toxicity when compared to AAGA haplotype. The association of GGAG haplotype with gastrointestinal toxicity was expected from the genotype analyses obtained results for those SNPs in SLC19A1. Considering this, we can hypothesize that GGAG haplotype could provide an increased MTX influx capability, leading to higher MTX intracellular levels, mainly in tissues where SLC19A1 is highly expressed (Qiu et al., 2006) and, thus, have an increased risk for gastrointestinal toxicity development.

Toxicogenetic Risk Index for MTX

To an improved characterization of the impact of studied SNPs that were statistically significant associated with MTX-related toxicity, a TRI was created, both for overall and gastrointestinal toxicity. Accordingly to MTX overall toxicity, an increased TRI value was associated with an increased incidence of ADRs. Our results demonstrated that patients with Index 3 were 18-fold times more likely to present an ADR when compared to those with Index 1. Regarding to the occurrence of gastrointestinal disorders, the TRI demonstrated that patients with Index 4, when compared to those with Index 3 and 1, were 5-fold and 9-fold, respectively, more likely to present gastrointestinal disorders. This highlights the importance of genotyping patients and the urgency of developing the field of therapy personalization for the prediction of MTX-related toxicity development.

Besides the potential importance of our results, we are aware of possible study limitations such the sample size and the study design. Despite this, our data are supported by the fact that: (1) our population is relatively homogenous regarding ethnic origin (all Caucasians from the North region of Portugal) with a prevalence of RA similar to other countries; (2) studied patient group characteristics were in accordance with other reported studies in regard to disease gender epidemiology (Gibofsky, 2012) and to diagnosis age range (Rindfleisch and Muller, 2005); (3) statistical analyses were performed attending to potential confounder variables limiting the selection bias; and (4) having studied 23 SNPs in genes encoding for MTX membrane transporter proteins, many of which had never been studied before, in both RA and Caucasian populations. Moreover, and due the low frequency of SOC disorders, other than gastrointestinal disorders in our population, the influence of the studied SNPs in there occurrence could not be performed. Knowing that MTX transporters are expressed in different tissues, this line of investigation could be proved of remarkable relevance since it would enable the prediction of toxicity in each tissue and help to guide therapeutic choices. Moreover, the influence of SNPs in MTX transporters genes in MTX circulating levels should be performed in order to elucidate SNPs impact in transport function and toxicity development. Despite this, we have to consider that MTX retention is also dependent of MTX polyglutamation levels and, thus, genetic polymorphisms in enzymes involved in MTX polyglutamation process should also be evaluated.

Other studies with similar approach demonstrated irrelevant results concerning the associations of discussed SNPs with MTX-related toxicity. Several explanations may be proposed to clarify these discrepancies as follow: differences in genotyping methodologies, ethnic origins, study population size, and possible confounding variables not considered in studies; majority of authors do not define similarly the concept of toxicity; and, most of studies did not follow standard guidelines for ADRs classification. Those reasons concur to some erratic conclusions to be taken and render adequate comparison between studies difficult to pursuit. Therefore, and due to the lack of SNPs combined studies using both functional and/or associated with MTX-related toxicity, further evidence is necessary to support the interpretation of our results and elucidate previous inconsistent results.

Conclusion

From this study, we can conclude that of all studied SNPs in ABCs, only ABCG2 rs13120400 seems to be associated with MTX-related toxicity occurrence. Interestingly, it also reveals that SNPs in SLCs should be helpful to elucidate which patients will benefit from MTX treatment since SLC19A1 rs7499, SLC46A1 rs2239907 and SLCO1B1 rs4149056, appeared to be associated with increased risk for MTX overall toxicity and SLC19A1 rs7499, SLC19A1 rs1051266, SLC19A1 rs2838956 and SLCO1B1 rs4149056, showed association with MTX gastrointestinal toxicity. Furthermore, SLC19A1 haplotypes may help to identify patients with increased risk of developing MTX gastrointestinal toxicity. Additionally, the proposed toxicogenetic risk index highlights the importance of genotyping patients and the urgency of developing the field of therapy personalization for the prediction of MTX-related toxicity development.

FUNDING

Fundação para a Ciência e Tecnologia (FCT) for the Doctoral Grant (SFRH/BD/64441/2009 to A. Lima).

The authors wish to acknowledge the Genomics Unit-Genotyping Service of Instituto Gulbenkian de Ciência, the Nursing Service of Rheumatology Day Hospital of São João Hospital Center, and, the Physicians from Rheumatology Department of São João Hospital Center.