Human Genetics (2019) 138:1359–1377 https://doi.org/10.1007/s00439-019-02081-x

ORIGINAL INVESTIGATION

The genetic landscape of the human solute carrier (SLC) transporter superfamily

Lena Schaller1 · Volker M. Lauschke1

Received: 4 August 2019 / Accepted: 26 October 2019 / Published online: 2 November 2019 © The Author(s) 2019

Abstract The human solute carrier (SLC) superfamily of transporters is comprised of over 400 membrane-bound proteins, and plays essential roles in a multitude of physiological and pharmacological processes. In addition, perturbation of SLC transporter function underlies numerous human diseases, which renders SLC transporters attractive drug targets. Common genetic polymorphisms in SLC genes have been associated with inter-individual diferences in drug efcacy and toxicity. However, despite their tremendous clinical relevance, epidemiological data of these variants are mostly derived from heterogeneous cohorts of small sample size and the genetic SLC landscape beyond these common variants has not been comprehensively assessed. In this study, we analyzed Next-Generation Sequencing data from 141,456 individuals from seven major human populations to evaluate genetic variability, its functional consequences, and ethnogeographic patterns across the entire SLC superfamily of transporters. Importantly, of the 204,287 exonic single-nucleotide variants (SNVs) which we identifed, 99.8% were present in less than 1% of analyzed alleles. Comprehensive computational analyses using 13 partially orthogonal algorithms that predict the functional impact of genetic variations based on sequence information, evolutionary conserva- tion, structural considerations, and functional genomics data revealed that each individual genome harbors 29.7 variants with putative functional efects, of which rare variants account for 18%. Inter-ethnic variability was found to be extensive, and 83% of deleterious SLC variants were only identifed in a single population. Interestingly, population-specifc carrier frequencies of loss-of-function variants in SLC genes associated with recessive Mendelian disease recapitulated the eth- nogeographic variation of the corresponding disorders, including cystinuria in Jewish individuals, type II citrullinemia in East Asians, and lysinuric protein intolerance in Finns, thus providing a powerful resource for clinical geneticists to inform about population-specifc prevalence and allelic composition of Mendelian SLC diseases. In summary, we present the most comprehensive data set of SLC variability published to date, which can provide insights into inter-individual diferences in SLC transporter function and guide the optimization of population-specifc genotyping strategies in the bourgeoning felds of personalized medicine and precision public health.

Introduction genome with over 400 proteins classifed into 65 subfami- lies based on sequence similarity (Fredriksson et al. 2008; The solute carrier (SLC) gene superfamily is one of two Höglund et al. 2011; Schlessinger et al. 2013). Substrate major human gene families encoding transporters of endog- specifcity difers substantially across the various subfami- enous and exogenous compounds. SLCs constitute the lies. While some subfamilies, such as the carbohydrate and second-largest family of membrane proteins in the human long chain fatty acid transporters of the SLC2 and SLC27 subfamilies, transport only few physicochemically homog- Electronic supplementary material The online version of this enous substrates (Anderson and Stahl 2013; Mueckler and article (https​://doi.org/10.1007/s0043​9-019-02081​-x) contains Thorens 2013), transporters of the SLC22 family mediate the supplementary material, which is available to authorized users. translocation of various dissimilar ions, including organic cations, anions, and zwitterions (Koepsell 2013). Most SLC * Volker M. Lauschke [email protected] transporters are equilibrative, making use of electrochemi- cal and concentration gradients to facilitate the uptake of 1 Department of Physiology and Pharmacology, their substrates into cells. Transport mechanisms can difer Section of Pharmacogenetics, Karolinska Institutet, within subfamilies, however, as seen in the secondary active 171 77 Stockholm, Sweden

Vol.:(0123456789)1 3 1360 Human Genetics (2019) 138:1359–1377 symporters and antiporters of the SLC4 bicarbonate trans- obtained data set constitutes the most comprehensive analy- porter family (Romero et al. 2013). sis of genetic SLC variability published to date and provides Due to their essential roles in the transport of a plethora valuable insights into inter-individual and inter-ethnic dif- of essential organic and inorganic substrates and the high ferences in transporter function with important implica- number (> 100) of SLC transporters that have been asso- tions for drug disposition, efcacy, and toxicity, as well as ciated with human genetic disorders, SLC transporters are population-specifc prevalence of Mendelian SLC diseases. being increasingly investigated as potential drug targets. One prominent example is the development of blockbuster SGLT2 (encoded by SLC5A2) inhibitors for the treatment of Materials and methods diabetic hyperglycaemia, which was inspired by associations between SLC5A2 mutations and familial renal glucosuria Data collection and annotation (OMIM identifer 233100). Besides their role as drug targets, SLC transporters play fundamental roles in the disposition Genetic variability data of 401 genes comprising the human of numerous drugs, including various chemotherapeutics, SLC superfamily were collected from the Genome Aggrega- antidiabetics, and diuretics. Given the extensive genomic tion Database (gnomAD) version 2.1 (Lek et al. 2016). The coverage and the critical role that SLC transporters play in use of these data did not require separate ethical approval, as mediating drug pharmacokinetics and -dynamics (PK/PD), the data are released under the Fort Lauderdale Agreement. the genetic variability of SLC genes is of considerable inter- In total, we analyzed sequencing data of 141,456 unrelated est for human genetics, as well as for drug discovery and individuals spanning seven worldwide populations (64,603 development programs. Non-Finnish Europeans, 12,562 Finns, 12,487 Africans, In the last decade, seminal studies have contributed sub- 9977 East Asians, 15,308 South Asians, 17,720 Latinos, stantially to our understanding of the link between SLC vari- 5185 Ashkenazi Jews, and 3614 from other populations). ability and drug response. Prominent examples include the Variants with low confdence calls were removed. Rare association between variants in SLC22A1, encoding OCT1, and common genetic SNVs were defned as variants with and pharmacokinetics and response to metformin (Dujic MAF < 1% and MAF ≥ 1%, respectively. Copy-number vari- et al. 2015; Sundelin et al. 2017) and variations in SLC19A1, ants’ (CNVs) data from 59,451 individuals were obtained encoding the reduced folate transporter RFT, with toxicity from the Exome Aggregation Consortium and analyzed as of antifolate metabolites (Bohanec Grabar et al. 2012; Cor- previously described (Santos et al. 2018). Linkage analy- rigan et al. 2014; Lima et al. 2014). However, these stud- sis was performed using LDLink (Machiela and Chanock ies were only powered to detect associations with common 2015). Disease associations for the relevant SLC genes were variations. Importantly, recent population-scale sequencing obtained from the Online Mendelian Inheritance in Man projects revealed that rare variations with minor allele fre- (OMIM) database. Deleterious variants in disease-associated quencies (MAF) < 1% greatly outnumber common variants genes were fltered for benign variants using ClinVar Miner in genes involved in drug absorption, distribution, metabo- (Henrie et al. 2018). lism, and excretion (ADME) (Bush et al. 2016; Kozyra et al. 2017; Wright et al. 2018; Zhou and Lauschke 2018). Rare Computational functionality predictions variations are enriched in variants with functional conse- quences and commonly have increased efect sizes compared Missense variants were analyzed using an array of partly to common variants when analyzed in relation to disease orthogonal algorithms that predict the functional impact of (Ingelman-Sundberg et al. 2018; Manolio et al. 2009). genetic variations based on sequence information, evolution- While the extent and functional importance of rare variants ary conservation, structural considerations, and functional is becoming increasingly appreciated, SLC transporters are genomics data. Specifcally, we used SIFT (Ng and Henikof understudied (César-Razquin et al. 2015) and their genetic 2001), Polyphen-2 (Adzhubei et al. 2010), Likelihood Ratio landscape remains to be systematically analyzed. Tests (Chun and Fay 2009), MutationAssessor (Reva et al. Here, we systematically mapped the genetic variability 2011), FATHMM (Shihab et al. 2012), PROVEAN (Choi of the human SLC transporter superfamily by analyzing et al. 2012), VEST3 (Carter et al. 2013), CADD (Kircher consolidated whole-exome and whole-genome sequencing et al. 2014), DANN (Quang et al. 2015), FATHMM-mkl data (WES and WGS, respectively) from 141,456 individu- (Shihab et al. 2015), MetaSVM (Dong et al. 2015), MetaLR als across seven major populations. We profled the SLC (Dong et al. 2015), and GERP++ (Davydov et al. 2010). genetic variability, its functional consequences, and ethno- We considered all variants that resulted in the gain of a stop geographic distribution using 13 partly orthogonal compu- codon, the loss of the start codon, that caused frameshifts tational predictors, as well as structural mapping approaches or that disrupted canonical splice sites as loss-of-function using experimental high-resolution crystal structures. The variants.

1 3 Human Genetics (2019) 138:1359–1377 1361

Structural modeling genes being classifed as haploinsufcient, suggesting high evolutionary constraints and reduced functional redun- The secondary structures of GLUT1 (SLC2A1) and ENT1 dancy in these gene families (Fig. 1e). In contrast, none of (SLC29A1) were obtained from UniProt (UniProt IDs 4PYP the genes encoding transporters of cofactors, carboxylates, and 6OB6, respectively). The structure of OCT1 (SLC22A1) fatty acids, protons, oligopeptides, and urea cycle metabo- was predicted using Phyre2 (Kelley et al. 2015) as no high- lites were found to be haploinsufcient. resolution crystal structure for this transporter was avail- able. Confdence and coverage scores were ≥ 100% and 80%, respectively. Structures were modeled using PyMOL version Population‑specifc frequencies of clinically 2.3. important SLC variants and haplotypes

SLC transporters mediate the transport of a plethora of Results drugs, and multiple SLC variants can impact disposition, efcacy, or toxicity of various medications (Table 1 and Overview of the genetic variability of the human Supplementary Table 1). Here, we analyzed the population- SLC superfamily specifc frequencies of 31 SLC variants with clinically rele- vant pharmacogenomic associations, mostly to antidiabetics, Across 141,456 unrelated individuals, we identifed a total of analgesics, anticoagulants, and various chemotherapeutics. 204,287 exonic single-nucleotide variants (SNVs) and indels Importantly, variant prevalence was highly population-spe- (Fig. 1a). In addition, the data set contained 118,597 intronic cifc and 42% of these variants (n = 13/31) difered more than variations; however, as these were not systematically cov- fvefold between populations. Common variants in SLC22A1 ered, they were excluded in our further analyses. The major- have been repeatedly linked to altered drug disposition and ity of exonic variants resulted in amino acid exchanges of the efcacy of metformin (Todd and Florez 2014), imatinib encoded polypeptide (n = 116,300; 57% of all exonic vari- (Watkins et al. 2015), and various opioids (Tzvetkov 2017). ants). The remaining SNVs included synonymous variants Multiple variations with functional consequences, includ- (n = 56,685; 28%) and variants in the untranslated regions ing M420del (SLC22A1*2), R61C (SLC22A1*3), G401S (n = 9507; 5% in the 5’ UTRs and n = 7400; 4% in the 3’ (SLC22A1*4), and G465R (SLC22A1*5), are absent in East UTRs). Furthermore, we identifed a multitude of variants Asians, whereas they can reach frequencies up to 21.9% in that result in putative loss-of-function of the gene product, other populations (Table 1). In contrast, L160F, which is such as frameshifts (n = 5086), stop-gain variants (n = 3384), associated with altered imatinib pharmacokinetics and an and variations in canonical splice sites (n = 3050). In addi- increased risk of resistance to imatinib (Cargnin et al. 2018; tion to SNVs, we found 3688 copy-number variations Di Paolo et al. 2014; Makhtar et al. 2018), is common in (CNVs), comprised of 2532 duplications and 1156 deletions East Asians (MAF = 14.2%) but lowest in Africans (3.8%). (Fig. 1a). Strikingly, of the 204,287 total exonic variants, While pharmacogenetically important polymorphisms in 203,968 (99.8%) were identifed as rare, with MAFs < 1% the nucleoside transporter genes SLC28A1 and SLC28A2 (Fig. 1b). were common worldwide, their frequencies difered drasti- Next, we focused specifcally on variants that afected the cally between populations. Rs2242046 in SLC28A1, as well amino acid sequence of the encoded gene product. Among as the highly linked variants rs1060896 and rs11854484 the subfamilies, variability was highest in the cholesterol in SLC28A2 (R2 = 0.9), were least common in East Asian transporter family SLC65 with a median of 809 variants per and African populations (MAF = 7.6–17.6%), whereas they gene (n = 2 genes), followed by SLC12 chloride cotransport- were consistently more prevalent in all other populations ers (604 variants per gene; n = 9) and the bicarbonate trans- (MAF = 20.8–65.4%). In contrast, Africans and East Asians porter family SLC4 (556 variants per gene; n = 10; Fig. 1c). were among the populations with the highest frequency of By contrast, four subfamilies harbored less than 100 vari- rs56350726, a variant in SLC28A3 implicated in improved ants per gene (SLC31, SLC48, SLC54, and SLC57). When outcomes and reduced toxicity of antiviral hepatitis C virus stratifying SLC genes by substrate, variability was highest in (HCV) therapy (Doehring et al. 2011; Rau et al. 2013), sug- genes coding for inorganic ion transporters (718 ± 228 s.d. gesting potentially important implications for toxicity risk variants per gene) and fatty acids (646 ± 187 variants per of nucleoside analogs used in the treatment of viral infec- gene), whereas pyruvate transporters (230 ± 213 variants tions and various cancers. Pronounced tenfold diferences per gene) and metal transporters (433 ± 150) harbored sub- between populations were furthermore observed for the mis- stantially fewer variants (Fig. 1d). Loss-of-function variants sense variant rs17235409 in SLC11A1 that is implicated in were depleted in transporters of neurotransmitters, hor- treatment failure of patients with pulmonary tuberculosis to mones, choline, inorganic ions, and metals with 19–25% of

1 3 1362 Human Genetics (2019) 138:1359–1377

AB 350000 Deletions 100% CNVs (3688) 1156 Common 300000 Duplications 2532 10% 99.8% 250000 onic 118597 Intr Frameshift 1% Ra re 5' UTR 5086 97.9% 200000 3' UTR 7400 9507 Stop gained 0.1% 3384 Inframe insertion Ve

584 ry Ra 150000 Splice Synonymous Acceptor 0.01% re 56685 1327 Start lost 57.6% onic 204287 100000 541 Ex Inframe Minor allele frequency Number of variants 0.001% Deletion Stop lost Singletons Splice 50000 Missense 1512 238 116300 Donor 1723 0 100% 80% 60% 40% 20% 0% Fracon of exonic variants

C 1000 38A10 12A3

e 800 9C1 24A1 600 7A2 9A4 38A8 39A12

400 amino acid sequenc 200 2A1

Number of variants affecng 9A6 39A9 22A31 22A20 37A4 38A3 0 1 SLC1 SLC2 SLC3 SLC4 SLC5 SLC6 SLC7 SLC8 SLC9 SLC1 SLC1 0 SLC1 2 SLC1 3 SLC1 4 SLC1 5 SLC1 6 SLC1 7 SLC1 8 SLC1 9 SLC2 0 SLC2 2 SLC2 3 SLC2 4 SLC2 5 SLC2 6 SLC2 7 SLC2 8 SLC2 9 SLC3 0 SLC3 1 SLC3 2 SLC3 3 SLC3 4 SLC3 5 SLC3 6 SLC3 7 SLC3 8 SLC3 9 SLC4 0 SLC4 1 SLC4 2 SLC4 3 SLC4 4 SLC4 5 SLC4 6 SLC4 7 SLC4 8 SLC4 9 SLC5 0 SLC5 1 SLC5 2 SLC5 3 SLC5 4 SLC5 5 SLC5 6 SLC5 7 SLC6 2 SLC6 5 SLC6 6

DE 1400 100%

1200

e Null

s 80% Recessive 1000 Haploinsufficient 60% 800

600 40%

400 Fracon of gene 20% amino acid sequenc 200 Number of variants affecng 0 0% s s s s s n ls s s s s n H+ ls cl e ar er rs ta H+ cl e vate er ta mins ga vate ylat es Othe r Heme mins actor Sug ta Me ylat es Othe r Heme actor Choline nic Ions nic Ions Su ta Me ea Cy y Acid s Vi Pyru Choline ylamines nic Ions nic Ions Cof ea Cy Bile Acid Unknow y Acid s ga ga Vi Pyru ylamines Hormones ansmi Ur Cof Fa Bile Acid Unknow os ga ga Hormones ansmi Ur Fa tr os Or Amino Acid Carb ox tr Or Amino Acid Oligopepdes Inor Carb ox ro Oligopepdes Inor Gl yc ro Gl yc Neu Neu

isoniazid, rifampicin, pyrazinamide, and ethambutol com- was common in Europeans (MAF = 3.1–3.6%), but rare in bination therapy (Salinas-Delgado et al. 2015). all other populations studied. Similar population specifcity The variant rs1529927 in the renal sodium and chloride was observed for the reduced function variant rs11568482 reabsorption transporter SLC12A3 (NCC) is associated with in SLC22A8 (OAT3), which was exclusively found in increased efcacy of diuretics (Vormfelde et al. 2007) and East Asians with frequencies of 5.9%, with important

1 3 Human Genetics (2019) 138:1359–1377 1363

◂Fig. 1 The landscape of genetic variability within the human SLC the average diploid human genome contained less than 0.1 gene superfamily. a Overview of genetic variants across 401 SLC variants in transporters of bile acids, oligopeptides, choline, genes based on the Next-Generation Sequencing data of 141,456 individuals from seven major populations. Of the 204,287 identi- protons, heme, pyruvate, and various cofactors. The contri- fed exonic SNVs, the majority resulted in amino acid exchanges. In bution of rare variants difered considerably between sub- addition, we identifed 3688 copy-number variations (CNVs) of SLC strate classes. While rare variants accounted for 6%, 12%, genes. b 99.8% of all exonic SLC variants were rare with minor allele and 12% of the genetically encoded functional variability frequencies < 1% and 57.6% were only found in a single individual. in urea cycle, organic ion, and amino acid transporters, no c Box and whisker plot depicting the number of variants that afect the amino acid sequence of the respective gene product (missense, common variants with functional efects were identifed in frameshift, start-lost, stop-gain, indels, and splicing variants). Note transporters of pyruvate, heme, or various other substrates that the number of such variants difers drastically between genes and and, thus, rare genetic variants were the only cause of SLC subfamilies. The middle line depicts the median and the edges genetically encoded functional efects in these transporters of the boxes depict the 25th and 75th percentiles. d Violin plot of total exonic SLC variants per gene, classifed by endogenous trans- (Fig. 2c). SLC26A11, SLC10A2, and SLC26A10 harbored porter substrate. White dots represent the median number of variants most rare functional variants per individual, whereas least per gene, with the ends of the white boxes indicating the 25th and were found in SLC54A1, SLC54A2, and SLC51B. When inte- 75th percentiles. Polygons represent density estimates of the data grating rare and common variant data, most deleterious vari- and extend to extreme values. e Stacked bar plot showing the frac- tion of genes under high evolutionary constraint, with genes classi- ants were identifed in the putative ammonium transporter fed by protein substrate. Evolutionary constraint was estimated using SLC42A2 (RhBG; 3 deleterious variants per individual), the the pLI score (Lek et al. 2016), with scores < 0.5 defned as little con- poorly understood ion transporter SLC22A10 (OAT5; 2 vari- straint (“Null”), scores 0.5 ≤ x ≤ 0.9 for genes for which homozygous loss-of-function results in a deleterious phenotype (“Recessive”), and ants per individual), and the highly clinically relevant drug scores > 0.9 defned for haploinsufcient genes transporter SLC22A1 (OCT1; 1.3 variants per individual; Fig. 2d). In contrast, less than 1 in 2000 individuals harbored a deleterious variant in SLC54A3 (0), SLC38A3 (0.0001), implications for the renal clearance of the OAT3 substrate SLC54A1 (0.0005), SLC51B (0.0007), SLC25A51 (0.0007), cefotaxime (Yee et al. 2013). Moreover, rs16889462 in and SLC30A1 (0.0009). Strikingly, when aggregating infor- SLC30A8, a variant associated with increased response to mation about genetically encoded functional variability repaglinide (Huang et al. 2010), was common in Africans across the entire SLC superfamily of genes, each individual (MAF = 10.5%) and East Asians (MAF = 8%), but rare in all was found to harbor on average 29.7 variants with putative other populations analyzed (MAF < 1%). functional consequences in SLC transporters of which rare variants accounted for 18% (5.4 rare variants per individual; Genetic variants in human SLC genes are predicted Fig. 2e). to substantially contribute to inter‑individual diferences in transporter function Genetic variability in SLC genes is highly population specifc with important consequences To gauge the functional efects arising from the observed for the predisposition to Mendelian disease genetic variability beyond variants with known pharma- cogenetic associations, we employed an array of 13 partly When we stratifed the identifed SLC variants that were pre- orthogonal computational algorithms. Of the 116,300 identi- dicted to afect transporter function by ancestry, we found fed missense variants, 53,642 (46%) were predicted to alter that the distribution varied drastically between populations, the functionality of the respective gene product (Fig. 2a). with 83% of variants (n = 56,273) restricted to a single pop- Furthermore, we considered all 14,157 variants that caused ulation (Fig. 3a). Most population-specifc variants were frameshifts, the loss of a start or the premature gain of a identifed in Europeans, whereas the lowest numbers were stop codon or variants afecting canonical splice sites as found in Finns and Ashkenazi Jews, at least in part due to putative loss-of-function variants. The highest median num- unequal cohort sizes. Interestingly, after adjusting for cohort bers of functional variants were identifed in transporters size, we found that East Asians had the largest number of of inorganic ions (239 variants), fatty acids (238 variants), population-specifc variants with predicted functional con- and oligopeptides (216 variants; Fig. 2b). Per gene, most sequence, suggesting that this population might beneft most putatively deleterious variants were found in SLC12A4 (577 from population-adjusted genotyping strategies (Fig. 3b). In variants), SLC12A3 (544 variants), and SLC65A2 (531 vari- contrast, overall genetically encoded functional variability ants), whereas less than 25 variants were found in pyruvate difered only moderately between populations with individu- transporters. als of African (34.6 variants/ individual) and European (28.6 Each individual was found to harbor on average 6.2 variants/ individual) ancestry carrying on average the most and 5.8 variants with functional efects in organic ion and and least deleterious variants, respectively (Fig. 3c). amino acid transporters, respectively (Fig. 2c). In contrast,

1 3 1364 Human Genetics (2019) 138:1359–1377 ), Tzvetkov ( 2007 ), Tzvetkov et al. ( 2012 , 2013 ) et al. ( 2012 ) ( 2015 ) Fukuda ( 2013 ), Shu et al. Corrigan ( 2014 ) et al. References Wang et al. ( 2014 ) et al. Wang Lima et al. ( 2014 ) Lima et al. Suthandiram et al. ( 2014 ) Suthandiram et al. et al. Bohanec Grabar Moncriefe ( 2010 ) et al. Huang et al. ( 2013 ) Huang et al. Drögemöller et al. ( 2017 ) et al. Drögemöller Cai et al. ( 2017 ) Cai et al. Lee et al. ( 2015 ) Lee et al. Johnson et al. ( 2011 ) Johnson et al. Vormfelde et al. ( 2007 ) et al. Vormfelde Salinas-Delgado et al. et al. Salinas-Delgado Kwon et al. ( 2009 ) et al. Kwon 0.33) = subjects after OGTT; reduced morphine reduced subjects after OGTT; after AUC and increased clearance increased codeine administration; and efcacy of plasma concentrations and ondansetron tropisetron etrexate etrexate treatment (HR = 1.8) therapeutic threshold of methotrexate therapeutic threshold of methotrexate p = 0.02) 40% TT; (8.6% GG vs. GI toxicity (OR = 0.38) hepatotoxicity hepatotoxicity (OR = 5.3) trexate trexate (OR = 1.9) continuation due to toxicity (HR toxicity continuation due to (OR = 3.6) ototoxicity ototoxicity (OR = 0.06) sorafenib sorafenib therapy (OR = 0.5) neuropathy neuropathy (OR = 0.5–0.6) therapy therapy (OR = 11) obsessive-compulsive symptoms symptoms obsessive-compulsive (OR = 3.9) Reduced metformin response in healthy in healthy response metformin Reduced Decreased survival time following pem - survival time following Decreased Association Lower fraction of patients above the of patients above fraction Lower Decreased risk of methotrexate-induced riskDecreased of methotrexate-induced Increased risk of methotrexate-induced riskIncreased of methotrexate-induced Increased riskIncreased - metho to of poor response Decreased frequency of methotrexate dis - of methotrexate frequency Decreased Lower rapid response rate to irinotecan to rate response rapid Lower Decreased riskDecreased of cisplatin-induced Decreased warfarin dose requirement warfarin Decreased PFS of HCC patients following Increased Decreased riskDecreased of thalidomide-induced Increased risk of failure of tuberculosis of tuberculosis riskIncreased of failure Increased efcacy of diuretics Increased Increased risk of antipsychotics-induced riskIncreased of antipsychotics-induced 0.5 0.4 5.5 1 6.3 0.6 51.2 61.8 21.5 25.6 48.4 45.4 21.6 21.2 AFR 0.3 0.2 11.2 2.9 N/A 40.8 22.6 25.8 29.9 46.0 34.0 5.4 0.8 33.9 SAS 0 0 0.0 54.3 52.9 30.7 65.6 71.5 54.5 34.2 14.1 24.6 < 0.1 < 0.1 EAS 0.9 0.7 2.3 0.6 21.9 45.3 43.7 13.6 38.0 24.3 40.6 18.3 11.3 40.7 LAT 0.7 1.3 9.2 9.5 1.4 0.6 38.3 39.9 23.0 31.9 37.9 38.7 25.1 51.7 ASJ 1.9 0.7 5.6 2.2 3.1 11.8 45.0 45.0 23.4 18.7 40.8 43.5 18.1 33.9 FIN 2.6 2.3 7.7 2.0 3.6 14.5 43.6 43.2 24.7 26.5 45.4 42.3 24.3 41.7 NFE MAF (in %) R509 K; L350F) Missense (G401S) Missense (G465R) Missense (M420del) Missense (R61C) Type 3 ′ UTR ​ Missense (R27H) Missense (T445S) Synonymous (L41L) Synonymous 3 ′ UTR ​ Missense (P409S; Synonymous (L159L) Synonymous Missense (D543 N) Missense (A264G) Synonymous (T138T) Synonymous rs1143672:G>A; rs2257212:C>T rs34130495:G>A rs34059508:G>A rs35167514:ATG>del rs12208357:C>T Variant rs1051298:C>T rs1051266:G>A rs3763980:G>T rs4788863:C>T rs1051296:T>G rs1143671:C>T; rs1143671:C>T; rs7164902:C>T rs17235409:G>A rs1529927:G>C rs2228622:G>A Worldwide allele frequencies of SLC associated with variants phenotypes clinical drug allele frequencies or toxicity response Worldwide SLC22A1 (OCT1) 1 Table Gene (protein) SLC19A1 (RFC) SLC16A7 (MCT2) SLC16A5 (MCT6) SLC15A2 (PEPT2) SLC12A6 (KCC3) SLC11A1 (NRAMP1) SLC12A3 (NCC) SLC1A1 (EAAC1)

1 3 Human Genetics (2019) 138:1359–1377 1365 ( 2015 ) References Cargnin et al. ( 2018 ) et al. Cargnin Makhtar ( 2018 ) et al. Qiu et al. ( 2018 ) Qiu et al. ( 2014 ) et al. Di Paolo Tecza et al. ( 2018 ) et al. Tecza Song et al. ( 2008 ) Song et al. Makhtar ( 2018 ) et al. Bray et al. ( 2010 ) et al. Bray Yee et al. ( 2013 ) et al. Yee Song et al. ( 2008 ) Song et al. Visscher et al. ( 2015 ) et al. Visscher Cargnin et al. ( 2018 ) et al. Cargnin Yoon et al. ( 2013 ) et al. Yoon Pellicer et al. ( 2017 ) et al. Pellicer Qiu et al. ( 2018 ) Qiu et al. Bray et al. ( 2010 ) et al. Bray Tarasova et al. ( 2012 ) et al. Tarasova Lanvers-Kaminsky et al. et al. Lanvers-Kaminsky Urban et al. ( 2008 ) et al. Urban imatinib (OR = 0.6) (OR = 1.9–3.3) junctival junctival hemorrhage (OR = 4.8) (OR = 1.8) (OR = 1.3–2.6) AC therapy AC taxime cardiotoxicity imatinib (OR = 0.6) toxicity (OR = 3.2) AC therapy AC GI efects (OR = 0.39) ototoxicity ototoxicity (OR = 0.12) entin Association Decreased major molecular response to to molecular response major Decreased Increased riskIncreased imatinib to of resistance Increased riskIncreased of imatinib-induced con - Increased risk of FAC-induced nausea riskIncreased of FAC-induced 25% decreased imatinib clearance 25% decreased Increased riskIncreased of imatinib resistance Decreased requirement for dose delay in dose delay for requirement Decreased - of cefo clearance renal 50% decreased Increased riskIncreased of anthracycline-induced Increased AUC of metformin AUC Increased Decreased major molecular response to to molecular response major Decreased Increased AUC of metformin AUC Increased Increased risk of severe capecitabine riskIncreased of severe Increased riskIncreased of imatinib-induced edema Decreased requirement for dose delay in dose delay for requirement Decreased Dec. risk of metformin-induced adverse adverse Dec. risk of metformin-induced Decreased riskDecreased of cisplatin-induced Increased Cmax and AUC of metformin Cmax and AUC Increased - of gapap clearance renal 30% reduced 3.8 0.1 0 7.2 27.5 27.1 34.6 15.2 35.8 11.2 15.2 AFR SAS 15.4 21.7 37.7 0.1 N/A 12.5 0 62.7 7.5 12.5 12 5.9 4.3 0.4 6.9 0.1 14.2 18.1 27.3 12.8 64.4 12.8 EAS 8.7 0.1 5.2 0 5.2 41.7 20.2 20.4 31.6 12.1 21.7 LAT 0.0 0 16.2 30.4 35.1 24.5 14.7 40.1 11.7 14.7 41.5 ASJ 0.0 6 0 6.8 6.0 16.4 22.5 45.4 15 50.9 31 FIN 0.0 0 8.8 22.3 36.9 41.5 15.5 10.1 39.1 10.1 42.2 NFE MAF (in %) Type Missense (L160F) Missense (H49R) Missense (M408 V) Missense (I305F) 3’ UTR ​ Missense (A270S) Missense (T201 M) Synonymous (L269L) Synonymous Missense (V252A) Missense (S270A) Missense (L503F) Variant rs683369:C>G rs714368:A>G rs628031:G>A rs11568482:A>T rs4149178:A>G rs316019:A>C rs145450955:G>A rs274558:A>G rs723685:T>C rs316019:G>T rs1050152:C>T (continued) 1 Table Gene (protein) SLC22A16 (FLIPT2) SLC22A8 (OAT3) SLC22A7 (OAT2) SLC22A2 (OCT2) SLC22A5 (OCTN2) SLC22A4 (OCTN1)

1 3 1366 Human Genetics (2019) 138:1359–1377 - oral glucose toler glucose oral Stocker et al. ( 2013 ) et al. Stocker Choi et al. ( 2011 ) and Choi et al. Stocker et al. ( 2013 ) et al. Stocker Shin et al. ( 2014 ) Shin et al. Huang et al. ( 2010 ) Huang et al. Huang et al. ( 2010 ) Huang et al. Rau et al. ( 2013 ) et al. Rau Doehring ( 2011 ) et al. References Soo et al. ( 2009 ) Soo et al. ( 2013 ) et al. Rau Ampuero et al. ( 2015 ) et al. Ampuero Soo et al. ( 2009 ) Soo et al. Tecza et al. ( 2018 ) et al. Tecza Bray et al. ( 2010 ) et al. Bray Soo et al. ( 2009 ) Soo et al. Bray et al. ( 2010 ) et al. Bray South Asians, AFR Africans, Cmax peak serum concentration, respiratory disease (OR = 0.13) logical response in HCV treatment logical response HCV patient treated withHCV patient treated ribavirin hematologic toxicity but increased OS but increased toxicity hematologic patients of NSCLC treated with telaprevir or boceprevir or boceprevir withtreated telaprevir (OR = 2.3) gemcitabine (OR = 3.2) AC therapy AC hematologic toxicity hematologic AC therapy AC Decreased metformin response metformin Decreased Enhanced metformin response Enhanced metformin Decreased riskDecreased of aspirin-exacerbated Increased repaglinide efcacy repaglinide Increased Increased repaglinide efcacy repaglinide Increased - sustained viro of reaching Higher chance twofold reduced incidence of anemia in reduced twofold Association Increased riskIncreased of gemcitabine-induced Increased ribavirinIncreased serum levels riskIncreased of anemia in HCV patients Increased OS of NSCLC patients on OS of NSCLC Increased Increased hetatological toxicity of FAC of FAC toxicity hetatological Increased Decreased requirement for dose delay in dose delay for requirement Decreased Increased riskIncreased of gemcitabine-induced Increased requirement for dose delay in dose delay for requirement Increased 8.9 8.8 3.8 21.5 35.1 80.2 10.5 15.4 17.1 36.9 17.6 AFR N/A N/A 22.4 23.0 0.3 7.6 45.4 SAS 27.6 47.8 30.3 11.3 4.1 8.0 7.9 8.6 7.6 46.9 21.5 44.0 11.5 41.6 EAS < 0.1 0.6 2.8 East Asians, SAS Latinos, EAS East 34.5 17.6 19.3 25.7 28.4 22.5 30.2 20.8 11.7 LAT 0.2 4.9 34.8 36.0 29.1 26.1 50.9 30.5 58.3 55.3 27.4 ASJ ​ survival; OGTT OS overall combination therapy; and cyclophosphamide anthracycline 0.0 26.7 23.3 20.3 38.0 12.7 58.2 18.2 59.2 45.2 15.8 FIN 0.1 5.9 27.2 31.2 23.8 30.4 62.2 21.8 65.4 50.1 23.5 NFE MAF (in %) 5 ′ UTR ​ 5 ′ UTR ​ Missense (M50 V) Missense (R325 W) Missense (R325Q) Missense (Y513F) Type Missense (P22L) Synonymous (N104 N) Synonymous Missense (S75R) Missense (D521 N) Missense (M409T) rs12943590:G>A rs2252281:T>C rs1047626:G>A rs13266634:C>T rs16889462:G>A rs56350726:A>T Variant rs11854484:C>T rs6907567:T>C rs1060896:C>A rs2242046:G>A rs12210538:T>C fuorouracil, anthracycline and cyclophosphamide combination therapy; AC combination therapy; and cyclophosphamide anthracycline fuorouracil, (continued) FAC - PFS progression-free survival; GI gastrointes ratio; HR hazard OR odds ratio; NSCLC under the carcinoma; area HCC hepatocellular non-small cell lung cancer; time curve; concentration LAT Jews, ASJ Ashkenazi FIN Finns, Europeans, NFE non-Finnish MAF minor allele frequency, ​ AUC tinal; SLC47A2 (MATE2-K) ance test SLC47A1 (MATE1) SLC30A9 (ZnT9) SLC30A8 (ZnT8) SLC28A3 (CNT3) 1 Table Gene (protein) SLC28A2 (CNT2) SLC28A1 (CNT1)

1 3 Human Genetics (2019) 138:1359–1377 1367

Next, we focused specifcally on genes for which loss-of- that the analysis of loss-of-function frequencies in the gen- function mutations are associated with Mendelian disorders. eral population can be a powerful resource to inform about Notably, the cohorts which we analyzed were sampled from disease risk and population-specifc genetic complexity the general population and individuals with severe congeni- underlying recessive Mendelian diseases. tal diseases were excluded. However, we hypothesized that quantifcation of loss-of-function carrier frequencies in the Structural consequences of SLC variability general population could be used as a proxy for Mendelian disease risk with recessive mode of inheritance. To this end, To obtain mechanistic insights into the efects of SLC vari- we aggregated frequencies of frameshifts and stop-gain vari- ability, we mapped the genetic variants to the corresponding ations, as well as variants afecting canonical splice sites 3D structures of the transporter proteins. To this end, we removed variants previously reported to not cause disease focused on transporters with important roles in human physi- (see “Materials and methods”). Overall, 109 out of 401 ology and pharmacology for which high-resolution crystal human SLC genes were found to have known associations structures were either available or could be modeled with with genetic diseases, of which 84 were autosomal recessive high confdence. (Supplementary Table 2). The glucose transporter GLUT1 encoded by SLC2A1 As expected, the frequency of loss-of-function variants facilitates glucose uptake into erythrocytes and is the major was substantially lower in disease-associated SLC genes glucose transporter in the human blood–brain barrier. Varia- compared to SLC genes that were not associated with genetic tions in GLUT1 can cause GLUT1 defciency syndrome with disease (Fig. 3d, e). To evaluate whether this approach was an autosomal dominant inheritance pattern, which presents indeed suitable to identify population-specifc disease risk, as neurological problems, developmental delays, complex we focused on Mendelian disorders with well-established movement disorders, and, occasionally, hemolytic anemia (De ethnogeographic variation. Loss-of-function of the amino Giorgis and Veggiotti 2013). GLUT1 belongs to the major acid transporter SLC3A1 has been associated with cystinuria facilitator superfamily (MFS) of transporters and consists of (OMIM 220100). The disease has a worldwide prevalence two discretely folded domains, termed N- and C-domain, each of around 1 in 7000 neonates and has been reported to be consisting of six transmembrane helices, that are connected most common among Jews with frequencies up to 1:2500 by an intracellular helical bundle (ICH) (Deng et al. 2014). individuals in certain subpopulations (Eggermann et al. To translocate glucose, GLUT1 undergoes structural changes 2012). Interestingly, we found highest loss-of-function fre- and alternates between inwards and outwards facing confr- quencies of SLC3A1 in Ashkenazi Jews (0.9%), in agree- mations and the ICH has been shown to play essential roles ment with previous reports (Pras et al. 1995). Accordingly, in this process (Yan 2013). The ICH interacts with multiple one in 12,345 Ashkenazim individuals can be expected to transmembrane domains (TMDs) of GLUT1, thereby acting be homozygous for an SLC3A1 loss-of-function variant. as a latch that, in the absence of a ligand, stabilizes GLUT1 Carrier rates in other populations were > tenfold lower. in the outward facing confrmation (Deng et al. 2014). Upon Similarly, variability profles of the aspartate transporter ligand binding, interactions between the N and C domains are SLC25A13 recapitulated increased prevalence of type II cit- altered, resulting in a transition towards the inward-open state. rullinemia (OMIM 605814) in East Asians (Lu et al. 2005), In total, 181 variants in GLUT1 were identifed that were with aggregated loss-of-function frequencies of 0.8%, cor- distributed across all domains of the protein, including the responding to 1 in 15,625 homozygous East Asian carriers. ICH (Fig. 4a). Notably, we identifed rare variations in R400, Lysinuric protein intolerance (OMIM 222700) is most which participates in stabilization of the interaction between prevalent in the Finnish population and has been associated the N- and C-terminal domains (Park 2015), as well as in with mutations in SLC7A7 (Torrents et al. 1999). Impor- R92, R93, R232, and E209, which form a tightly connected tantly, loss-of-function frequencies of SLC7A7 in Finns salt bridge network that controls GLUT1 state transitions were more than fvefold higher than in other populations. (Galochkina et al. 2019). In contrast, no variants were Similarly, our data aligned with reported population dif- observed in the glucose entry pocket (N34, V69, R126, and ferences in the genetic basis of Pendred syndrome (OMIM Y292) or in the amino acids lining the central glucose cavity 274600), the most common form of syndromic genetic deaf- (S73, Q279, Q282, Q283, N288, N411, and N415). As the ness. While genetic variation in SLC26A4 is a major cause analyzed cohort was depleted of patients with congenital of these disorders in Asia, mutations in diferent genes have diseases, these fndings suggest that GLUT1 function can be been reported to be the most important factors in Western permissive to variations that modulate salt bridges involved populations (Park et al. 2003). In agreement with these in state transitions, whereas residues directly involved in genetic roots, frequencies of SLC26A4 loss-of-function vari- glucose translocation appear more conserved. ants in East Asian populations were approximately sixfold ENT1 encoded by SLC29A1 is an essential uptake trans- higher than in Europeans. Based on the results, we conclude porter of nucleosides and nucleoside analogs. As such,

1 3 1368 Human Genetics (2019) 138:1359–1377

A B 140000 (2619) 600 12A4 12A3 120000 65A2 500 26A4 l 65A1 100000 ra 60039 3A1 400 5A2 Neut Missense 27A3 80000 18A1 300 28A1 28A2 60000 s per gene 200

Number of variants 40000 53642 Deleterious variants Missense Deleteriou 100 20000 LoF 14157 0 0 s s s s e e H+ 26A11 7A6 Othe r Heme Sugars Metals Cholin Vitamin Pyruvate Cofactor Bile Acid Unknow n Hormones

C Rare variants Urea Cycl Amino Acid Organic Ions

per gene 14A2 Carboxylate s ( = 0.02) Inorganic Ions Glycosylamines 7

6 12% 12A2 5A9

s 12% 5 Rare l 30A5 25A41 Common 6% 16A8 4 19A1 1A6 15A5 27A3 10A2 15A1 44A2 29A3 25A25 49A2 16A11 3 per individua 21% 2 37% Deleterious variant 13% 31% 19% 1 17% 77% 42% 100% 100% 100% 100% 100% 100% 100% 100% 0 s s s s ns Cycle vate Sugars Heme u Metals rmone CholineProto ino Acid Vitamins Bile Acid Cofactors Pyr Urea Ho Organic AmIons cosylamineCarboxylates Inorganic Ions y Gl

Total: 29.7 D SLC42A2 E 30 SLC22A10

SLC22A1 To 25 SLC7A13 p 10 SLC38A10 SLC35B3 SLC3A2 20 SLC6A18 Common SLC22A24 (82%) SLC23A1 15 SLC35A2 SLC37A4 10

SLC22A20 per individual SLC30A1

SLC25A51 Deleterious variants SLC51B 5 SLC54A2 Rare SLC54A1 (18%) SLC38A3 0 SLC54A3 0.0007 0.0014 123 Superfamily Deleterious variants per individual

1 3 Human Genetics (2019) 138:1359–1377 1369

◂Fig. 2 Rare genetic variants contribute considerably to the genetically In contrast to GLUT1 and ENT1, no crystal structure encoded functional variability of SLC transporters. a Of the 116,300 of human OCT1 has yet been presented. However, struc- identifed missense variants, 53,642 were predicted to alter the func- tionality of the transporter protein. Furthermore, 14,157 variants that ture–function relationships have been inferred from crystal caused frameshifts, the loss of a start or the premature gain of a stop structures of homologous fungal transporters (Pedersen et al. codon, or variants that afecting canonical splice sites were expected 2013) and mutagenesis studies using rat OCT1 (Gorbunov to result in a loss of protein function. b Box and whisker plot of all et al. 2008; Popp et al. 2005). Here, we used the computa- these deleterious variants (n = 67,799) per gene demonstrates that the complexity of genetically encoded functional variability difers dras- tional tool Phyre2 (Kelley et al. 2015) to predict the structure tically between SLC substrate classes. c When aggregating variant of human OCT1 based on multiple sequence alignments and numbers per individual, most variants were identifed in transport- homologous experimentally determined models. We could ers of organic ions and amino acids. Common variants (MAF > 1%) derive a high confdence model (100% confdence score) are shown in light red, while rare variants (MAF < 1%) are shown in dark red. Percentage values within or above stacked columns indicate covering 82% of the human OCT1 protein sequence, which the fraction of the genetically encoded functional variability allotted aligned well with the putative structure of rat OCT1 (Sup- to rare variants. Inlet dot plots depict the total rare deleterious SLC plementary Fig. 1). variants per individual per gene, with the median represented by the SLC22A1 harbors eight common variants that affect dark bar. The gene with the highest number of rare deleterious vari- OCT1 amino acid sequence, of which R61C, G401S, and ants per substrate class is indicated above the inlet. d The aggregated frequency of variants that afect transporter function is plotted for the G465R resulted in strongly reduced OCT1 function in vitro, top 10 and bottom 10 SLC genes. Note that diferences between the whereas substrate-specifc results have been reported for most and least variable genes exceed 1000-fold. e Across the entire M420del and P341L (Choi and Song 2012; Shu et al. 2003; SLC superfamily, each individual was found to harbor 29.7 variants that are predicted to afect the functionality of the encoded trans- Tzvetkov et al. 2011, 2012). The function of OCT1 isoforms porter protein. Of this genetically encoded functional variability, carrying V464I, L160F or M408 V was not found to be 18.7% is attributed to rare variants altered in vitro (Shu et al. 2003). Notably, M420del occurs exclusively together with M408 V, whereas M408 V can occur in isolation (D′ = 1, R2 = 0.061) (Tzvetkov et al. 2014). ENT1 is of tremendous pharmacological importance for the The frequencies of these common variants varied greatly disposition of various antiviral and antineoplastic medica- across populations (Fig. 4c) with MAFs of the M240del tions, and is itself the pharmacological target of multiple variant ranging from 0.09% in East Asians to 21.9% in Lati- antiarrhythmic and antihypertensive medications (Boswell- nos and M408 V from 20.2% in Latinos to 45.4% in Finns in Casteel and Hays 2017). The organization in N- and C-ter- accordance with previous reports on smaller cohorts (Seitz minal pseudo-symmetric domains as well as the transport et al. 2015). In total, we identifed nine OCT1 variants with cycle of state transitions are overall similar between ENT1 a global MAF > 0.1%, half of which were localized to TMDs and GLUT1. However, the C-domain of ENT1 only com- (G38D, M420del, M440I, and G465R localized to TM1, prises fve instead of six TMDs (Wright and Lee 2019). In TM9, TM10, and TM11, respectively) (Fig. 4c). In addition total, our data set contained 263 variants that afected the to these well-characterized variants, we found 445 additional ENT1 amino acid sequence. The only common SLC29A1 variants that alter OCT1 amino acid sequence (Fig. 4c, d). variant was rs45573936, with a frequency of 1.8%, result- Rare variations were found to afect the mechanistically ing in an I216T amino acid substitution in TMD6 that alters important residues S358, R439, I446, Q447, and C450, binding afnity of the adenosine analog inhibitor NBMPR which are directly involved in the coordination of cationic and has been implicated in neurological symptoms upon substrates, as well as F485, a residue essential for state tran- alcohol withdrawal (Kim et al. 2011). We identifed rare sition during substrate translocation (Gorbunov et al. 2008; variants in multiple critical residues of the central cavity Pedersen et al. 2013; Volk et al. 2009). Combined, structural of the transporter that likely impact substrate afnity and mapping of the genetic variability in physiologically and binding kinetics. These include a substitution of the hydro- pharmacologically important SLC transporters supports the phobic methionine M33 that lines the narrowest constric- conclusion that rare, as of yet uncharacterized variants, are tion point at the extracellular side (Wright and Lee 2019) likely to have important functional impacts on transporter with a charged lysine, as well as amino acid exchanges structure and function. afecting various residues that defne the intracellular gate, such as R111, I304, and T429 (Fig. 4b). Furthermore, mul- tiple variants afected the residues M89, Q158, and R345 Discussion that have been shown to directly interact with the structur- ally heterogeneous ENT1 inhibitors dilazep and NBMPR SLC transporters play pivotal roles in diverse physi- (Wright and Lee 2019). We conclude that the genetic vari- ological processes, including the uptake and disposi- ability in SLC29A1 is extensive and multiple variants are tion of nutrients, maintenance of acid-base homeosta- highly likely to afect ENT1 pharmacology. sis, neurotransmission, and the elimination of metabolic

1 3 1370 Human Genetics (2019) 138:1359–1377

ABC

100% 2.5 35 l l

ed Finns Ashkenazim 30 18% (n = 12562) (n = 5185) 2 Shar 80% 1548 720 25 1.5 Latinos 20 60% c (n = 17720) 7720 1 15 East Asians (n = 9977) 10 Europeans Total deleterious variants per individua 83% variants per individua 40% (n = 64603) 6429 0.5 28479 5 SLC South Asians SLC opulaon-specifi (n = 15308) Populaon-specific deleterious

P 0 0 s s 9957 s s

Fracon of exonic variants 20% Africans Finns (n = 12487) Finns Lanos Lanos Africans 6762 Africans European European East Asian 0% East Asian Ashkenazim Ashkenazim South Asians South Asians

DESLC genes associated with Non-disease associated Mendelian diseases SLC genes

0.03 17A9 0.7

0.6 0.025 y y

0.5 0.02 0.4 28A1 0.015 0.3 6A20 0.01 3A1 25A13 0.2 25A38 Loss-of-funcon frequenc Loss-of-funcon frequenc 0.05 0.1 36A2 17A9 0 0 s s s s Finns Finns Lanos Lanos Africans Africans European European East Asian East Asian Ashkenazim Ashkenazim South Asians South Asians

Fig. 3 Genetic SLC transporter variability is highly population-spe- cally encoded functional SLC variability per individual were observed cifc. a Of all putatively deleterious variants (n = 67,799), 83% were across populations. d, e Dot plots depicting the cumulative frequency only detected in a single population. The pie chart depicts the total of putative loss-of-function (LoF) variants (frameshifts, start-lost, number of deleterious, population-specifc variants for each popula- and stop-gain variations, as well as variants afecting canonical splice tion. Values in brackets indicate the size of the cohort for the popu- sites) per gene for SLC genes associated with Mendelian diseases (d) lation in question. b The number of population-specifc, deleterious as well as for non-disease-associated genes (e). Note that LoF fre- SLC variants per individual difered considerably across major human quency of disease-associated genes is much lower than of non-dis- populations. c In contrast, only minor diferences in overall geneti- ease-associated genes products. In addition, they are involved in the disposition (inhibition of SLC22A12/URAT1 by lesinurad), schizo- of a multitude of clinically relevant medications, rang- phrenia (inhibition of SLC6A9/GLYT1 by bitopertin), and ing from chemotherapeutics to antidiuretics. Their bio- depression (inhibition of SLC6 transporters by serotonin- logical importance and pharmacological relevance, as selective reuptake inhibitors) (Lin et al. 2015). Our analy- well as their roles in numerous human diseases, render ses revealed that almost half of all SLC alleles associated SLC transporters attractive drug targets. Current clinical with altered drug response or toxicity had frequencies that applications include the targeted treatment of hyperten- difered more than fvefold between populations. These sion (inhibition of SLC12A1/NKCC2 by diuretics), dia- fndings have important implications for the treatment with betes (inhibition of SLC5A2/SGLT2 by glifozins), gout SLC transporter substrates in an ethnogeographic context,

1 3 Human Genetics (2019) 138:1359–1377 1371 and incentivize the adoption of population-adjusted geno- disorders with a recessive mode of inheritance, including typing strategies to optimize patient outcomes. cystinuria in Jewish individuals, type II citrullinemia in In addition to the previously described SLC alleles, we East Asians, and lysinuric protein intolerance in Finns. We identifed a surprising extent of genetic complexity within thus conclude, in agreement with previous studies (Fujikura SLC transporters. Importantly, less than 0.2% of all iden- 2016), that NGS data of the general population can pro- tifed variants were found in more than 1% of alleles and vide a suitable tool for the analysis of the genetic variability more than half of all variants were singletons. To estimate underlying inherited disorders. Furthermore, we argue that the overall contribution of this plethora of rare variants to the presented data can serve as a unique large-scale resource functional SLC variability, we used an array of 13 partly for clinical geneticists to inform about population-specifc orthogonal computational algorithms that leverage sequence prevalence and allelic composition of risk alleles associ- information, evolutionary conservation, structural consid- ated with Mendelian diseases of SLC transporters. Impor- erations, and functional genomics data in the prediction tantly, the approach is likely not suitable for the analyses process, and have been found to perform reasonably well of diseases following a dominant mode of inheritance, as on both disease-associated and pharmacogenomic data sets individuals with severe congenital diseases were excluded (Li et al. 2018; Zhou et al. 2019). These analyses revealed from the analyzed cohorts, resulting in an underestimation of that each individual genome harbors on average 29.7 puta- dominant disease allele frequencies in our data set. Notably, tively functional SLC variants, with rare variants accounting the relatively high frequencies of loss-of-function variants for 18% of this genetically encoded functional variability. in SLC28A1, which are associated with autosomal dominant Notably, nearly half of all putatively deleterious SLC vari- uridine-cytidineuria (OMIM 618477), might be explained by ants in an individual afected transporters of amino acids, its putatively benign nature (Wevers et al. 2019). organic ions, and urea cycle intermediates. Structural map- NGS is already widely and successfully applied in the ping of the portfolio of genetic variants on available crystal diagnosis of rare monogenic diseases (Boycott et al. 2013; structures of the encoded proteins revealed that rare variants Fernandez-Marmiesse et al. 2018). However, while targeted afect multiple residues that have been shown to be essen- sequencing panels that include multiple SLC transporters tial for transporter function, thus further corroborating the have been developed (Gordon et al. 2016; Gulilat et al. 2019; important functional roles of rare genetic variability. Besides Klein et al. 2019), the incorporation of these genetic data variations that are directly involved in substrate coordination into personalized pharmacogenomic recommendations and or translocation, a variety of missense variants in SLC trans- clinical decision-making is lagging behind. In the absence porters are known to afect transporter function by altering of feasible experimental strategies to characterize the func- subcellular trafcking or localization. Prominent examples tional impact of the plethora of rare SLC variants, com- include variants in SLC22A1 (Chen et al. 2010), SLC12A6 putational methods have to be used. While such in silico (Salin-Cantegrel et al. 2011) and SLC30A5 (Thornton et al. interpretations of pharmacogenetic variants do not yet have 2011). Notably, while these efects are difcult to infer by sufcient accuracy to warrant direct clinical implementation structural mapping, variant efect predictors, such as those (Zhou et al. 2018), these tools can already be used to fag used in this study, faithfully predicted localization defects patients with suspicious variants in key pharmacogenes for and even outperformed specialized subcellular localization closer monitoring to anticipate detrimental drug response tools (Orioli and Vihinen 2019). as early as possible. However, whether NGS coupled with Strikingly, we found that 83% of all variants that were computational pharmacogenomic analyses can indeed facili- predicted to afect SLC function were population-specifc. tate informed decision-making and provide a cost-efective This degree of inter-ethnic variability is similar to other measure to improve patient care, remains to be evaluated in highly variable pharmacogene families, such as CYPs prospective trials (Lauschke and Ingelman-Sundberg 2016, (Fujikura et al. 2015) and UGTs (Kaniwa et al. 2005), as 2018). well as to the related SLCO family of transporters (Zhang In summary, by leveraging consolidated NGS data from and Lauschke 2019). While individuals of African ancestry 141,456 individuals, we comprehensively assessed the harbored most functional SLC variants in agreement with genetic variability of the human SLC transporter superfamily previous fndings of greater levels of genetic diversity in on an unprecedented scale. We demonstrate that SLC genes Africans compared to non-African populations (Campbell are highly variable and each individual genome is estimated and Tishkof 2008; Tishkof et al. 2009), the largest number to contain around 30 variants that afect SLC transporter of population-specifc SLC variants was identifed in East function. The vast majority of variants were rare, and com- Asians. Interestingly, when focusing on SLC genes asso- putational analyses based on evolutionary, structural, and ciated with Mendelian disease, we found that population- functional genomics data indicate that these rare variants specifc carrier frequency in the general population recapitu- contribute approximately 20% to the genetically encoded lated the ethnogeographic variation of various Mendelian functional variability of SLC transporters. Thus, these data

1 3 1372 Human Genetics (2019) 138:1359–1377

A

Extracellular Space

C2A1 ) ICH

UT1 ( SL R232H R218H

GL G233R Cytoplasm E220A N Terminus C Terminus N Terminus D236Y A224T V237M K225N

B C Terminus

Extracellular Space

C29A1 ) I216T I216T Intra/Extracellular Gates Recognition Sites Q158X M33K ENT1 ( SL I304T R345Q Cytoplasm R111W T429I N Terminus N Terminus

Substrate

C88R H115P C M408V M420del 50% NFE EAS e 25% FIN SAS 30% 15% AJ AFR R61C LAT 10% 10% 5% R488M 6% G165C 2% Extracellular Spac 43 148 262 367 429 493

G38D R342H Plasma

Membrane V519F 17 234

m L160F 25% I403V 15% M440I

C22A1 ) 554

Cytoplas S14F S189L SL 5% P341L G401S V464IG465R 15% 3% 20% 3% T1 ( 10% 2% 2% 10% OC 5% 1% 1% D

M420del Extracellular L160F Space M408V al L160F

G465R AF Neutr V464I Very Rare Cytoplasm G401S Rare N Terminus N Terminus Common P341L No Variant V464I C Terminus

1 3 Human Genetics (2019) 138:1359–1377 1373

◂Fig. 4 Structural mapping of GLUT1, ENT1, and OCT1 variability. server for predicting damaging missense mutations. Nat Methods Schematic topology models and experimentally derived 3D protein 7:248–249. https​://doi.org/10.1038/nmeth​0410-248 structures viewed from both sides of the membrane plane are shown Ampuero J, Del Campo JA, Rojas L, Calleja JL, Cabezas J, Lens S, for human GLUT1 (a) and ENT1 (b). Detailed 3D structures of key Crespo J, Forns X, Andrade RJ, Fernández I, Buti M, Millán R, protein domains with functionally relevant variants (sticks) and sub- Romero-Gómez M (2015) Role of ITPA and SLC28A2 genes in strates (sticks in dark blue) are shown as insets under the respective the prediction of anaemia associated with protease inhibitor plus topology models. Red segments of the 3D models represent residues ribavirin and peginterferon in hepatitis C treatment. J Clin Virol with putatively deleterious variants, blue segments represent residues 68:56–60. https​://doi.org/10.1016/j.jcv.2015.05.010 with putatively neutral variants, and gray segments represent residues Anderson CM, Stahl A (2013) SLC27 fatty acid transport proteins. for which no associated variants have been identifed in 141,456 indi- Mol Aspects Med 34:516–528. https​://doi.org/10.1016/j. viduals. Deleterious and neutral common variants are depicted as red mam.2012.07.010 and blue spheres, respectively. ICH = intracellular helical bundle; c Bohanec Grabar P, Leandro-García LJ, Inglada-Pérez L, Logar D, schematic representation of the secondary structure of human OCT1, Rodríguez-Antona C, Dolzan V (2012) Genetic variation in the with all deleterious (shades of red) and neutral (shades of blue) vari- SLC19A1 gene and methotrexate toxicity in rheumatoid arthri- ants mapped to their respective residues. Color intensity of each tis patients. Pharmacogenomics 13:1583–1594. https​://doi. residue indicates variant frequencies. Population-specifc frequen- org/10.2217/pgs.12.150 cies of the common variants R61C, L160F, P341L, G401, M408 V, Boswell-Casteel RC, Hays FA (2017) Equilibrative nucleoside trans- M420del, V464I, and G465R are shown for Africans (AFR; yellow), porters—a review. Nucleosides Nucleotides Nucleic Acids 36:7– Ashkenazi Jews (AJ; purple), Non-Finnish Europeans (NFE; dark 30. https​://doi.org/10.1080/15257​770.2016.12108​05 blue), Latinos (LAT; green), East Asians (EAS; light blue), and South Boycott KM, Vanstone MR, Bulman DE, MacKenzie AE (2013) Rare- Asians (SAS; red) in inlets. d The predicted 3D model of OCT1 disease genetics in the era of next-generation sequencing: dis- viewed from both sides of the membrane plane is shown. Red seg- covery to translation. Nat Rev Genet 14:681–691. https​://doi. ments of the 3D models represent residues with putatively deleterious org/10.1038/nrg35​55 variants, blue segments represent residues with putatively neutral var- Bray J, Sludden J, Grifn MJ, Cole M, Verrill M, Jamieson D, Boddy iants, and gray segments represent residues for which no associated AV (2010) Influence of pharmacogenetics on response and variants have been identifed in 141,456 individuals. Deleterious and toxicity in breast cancer patients treated with doxorubicin and neutral common variants are depicted as red and blue spheres, respec- cyclophosphamide. Br J Cancer 102:1003–1009. https​://doi. tively. Note that R61C is not depicted, because the corresponding fold org/10.1038/sj.bjc.66055​87 could not be accurately modeled Bush WS, Crosslin DR, Owusu-Obeng A, Wallace J, Almoguera B, Basford MA, Bielinski SJ, Carrell DS, Connolly JJ, Crawford D, Doheny KF, Gallego CJ, Gordon AS, Keating B, Kirby J, serve as a powerful resource for the worldwide pattern of Kitchner T, Manzi S, Mejia AR, Pan V, Perry CL, Peterson JF, SLC Prows CA, Ralston J, Scott SA, Scrol A, Smith M, Stallings variability and motivate the integration of comprehen- SC, Veldhuizen T, Wolf W, Volpi S, Wiley K, Li R, Manolio sive NGS-based genotyping into personalized predictions T, Bottinger E, Brilliant MH, Carey D, Chisholm RL, Chute of SLC substrate disposition and precision public health. CG, Haines JL, Hakonarson H, Harley JB, Holm IA, Kullo IJ, Jarvik GP, Larson EB, McCarty CA, Williams MS, Denny JC, Acknowledgements Open access funding provided by Karolinska Rasmussen-Torvik LJ, Roden DM, Ritchie MD (2016) Genetic Institute. The work was supported by the Swedish Research Council variation among 82 pharmacogenes: the PGRNseq data from the (Grant agreement numbers: 2016-01153 and 2016-01154), by the Euro- eMERGE network. Clin Pharmacol Ther 100:160–169. https​:// pean Union’s Horizon 2020 research and innovation program U-PGx doi.org/10.1002/cpt.350 (Grant agreement number 668353), and by the Strategic Research Pro- Cai L-L, Huang WQ, Su Z-Y, Ye H-M, Wang L-S, Wu Y, Zhang Z-Y, gramme in Diabetes at Karolinska Institutet. We would like to thank the Zhang W, Tzeng C-M (2017) Identifcation of two novel genes Exome Aggregation Consortium and all contributing groups for sharing SLC15A2 and SLCO1B3 associated with maintenance dose vari- their data, without which this study could not have been performed. ability of warfarin in a Chinese population. Sci Rep 7:17379. https​://doi.org/10.1038/s4159​8-017-17731​-1 Campbell MC, Tishkof SA (2008) African genetic diversity: implica- Compliance with ethical standards tions for human demographic history, modern human origins, and complex disease mapping. Annu Rev Genomics Hum Genet Conflict of interest VML is co-founder and owner of HepaPredict AB. 9:403–433. https​://doi.org/10.1146/annur​ev.genom​.9.08130​ LS has no conficts of interest to declare. 7.16425​8 Cargnin S, Ravegnini G, Soverini S, Angelini S, Terrazzino S (2018) Open Access This article is distributed under the terms of the Crea- Impact of SLC22A1 and CYP3A5 genotypes on imatinib tive Commons Attribution 4.0 International License (http://creat​iveco​ response in chronic myeloid leukemia: a systematic review mmons.org/licen​ ses/by/4.0/​ ), which permits unrestricted use, distribu- and meta-analysis. Pharmacol Res 131:244–254. https​://doi. tion, and reproduction in any medium, provided you give appropriate org/10.1016/j.phrs.2018.02.005 credit to the original author(s) and the source, provide a link to the Carter H, Douville C, Stenson PD, Cooper DN, Karchin R (2013) Creative Commons license, and indicate if changes were made. Identifying Mendelian disease genes with the variant efect scoring tool. BMC Genomics 14(Suppl 3):S3. https​://doi. org/10.1186/1471-2164-14-S3-S3 César-Razquin A, Snijder B, Frappier-Brinton T, Isserlin R, Gyimesi References G, Bai X, Reithmeier RA, Hepworth D, Hediger MA, Edwards AM, Superti-Furga G (2015) A call for systematic research on Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, solute carriers. Cell 162:478–487. https​://doi.org/10.1016/j. Bork P, Kondrashov AS, Sunyaev SR (2010) A method and cell.2015.07.022

1 3 1374 Human Genetics (2019) 138:1359–1377

Chen L, Takizawa M, Chen E, Schlessinger A, Segenthelar J, Choi JH, intolerance to metformin in type 2 diabetes: a GoDARTS study. Sali A, Kubo M, Nakamura S, Iwamoto Y, Iwasaki N, Giaco- Diabetes 64:1786–1793. https​://doi.org/10.2337/db14-1388 mini KM (2010) Genetic polymorphisms in organic cation trans- Eggermann T, Venghaus A, Zerres K (2012) Cystinuria: an inborn porter 1 (OCT1) in Chinese and Japanese populations exhibit cause of urolithiasis. Orphanet J Rare Dis 7:19. https​://doi. altered function. J Pharmacol Exp Ther 335:42–50. https​://doi. org/10.1186/1750-1172-7-19 org/10.1124/jpet.110.17015​9 Fernandez-Marmiesse A, Gouveia S, Couce ML (2018) NGS tech- Choi M-K, Song I-S (2012) Genetic variants of organic cation nologies as a turning point in rare disease research, diagno- transporter 1 (OCT1) and OCT2 signifcantly reduce lamivu- sis and treatment. Curr Med Chem 25:404–432. https​://doi. dine uptake. Biopharm Drug Dispos 33:170–178. https​://doi. org/10.2174/09298​67324​66617​07181​01946​ org/10.1002/bdd.1783 Fredriksson R, Nordström KJV, Stephansson O, Hägglund MGA, Choi J-H, Yee SW, Ramirez AH, Morrissey KM, Jang GH, Joski PJ, Schiöth HB (2008) The solute carrier (SLC) complement of the Meford JA, Hesselson SE, Schlessinger A, Jenkins G, Castro human genome: phylogenetic classifcation reveals four major RA, Johns SJ, Stryke D, Sali A, Ferrin TE, Witte JS, Kwok P-Y, families. FEBS Lett 582:3811–3816. https​://doi.org/10.1016/j. Roden DM, Wilke RA, McCarty CA, Davis RL, Giacomini KM febsl​et.2008.10.016 (2011) A common 5′-UTR variant in MATE2-K is associated Fujikura K (2016) Global carrier rates of rare inherited disorders using with poor response to metformin. Clin Pharmacol Ther 90:674– population exome sequences. PLoS One 11:e0155552. https​:// 684. https​://doi.org/10.1038/clpt.2011.165 doi.org/10.1371/journ​al.pone.01555​52 Choi Y, Sims GE, Murphy S, Miller JR, Chan AP (2012) Predicting Fujikura K, Ingelman-Sundberg M, Lauschke VM (2015) Genetic the functional efect of amino acid substitutions and indels. PLoS variation in the human cytochrome P450 supergene family. One 7:e46688. https​://doi.org/10.1371/journ​al.pone.00466​88 Pharmacogenet Genomics 25:584–594. https://doi.org/10.1097/​ Chun S, Fay JC (2009) Identifcation of deleterious mutations within FPC.00000​00000​00017​2 three human genomes. Genome Res 19:1553–1561. https​://doi. Fukuda T, Chidambaran V, Mizuno T, Venkatasubramanian R, Ngam- org/10.1101/gr.09261​9.109 prasertwong P, Olbrecht V, Esslinger HR, Vinks AA, Sadhasivam Corrigan A, Walker JL, Wickramasinghe S, Hernandez MA, New- S (2013) OCT1 genetic variants infuence the pharmacokinet- house SJ, Folarin AA, Lewis CM, Sanderson JD, Spicer J, Mari- ics of morphine in children. Pharmacogenomics 14:1141–1151. naki AM (2014) Pharmacogenetics of pemetrexed combination https​://doi.org/10.2217/pgs.13.94 therapy in lung cancer: pathway analysis reveals novel toxic- Galochkina T, Ng Fuk Chong M, Challali L, Abbar S, Etchebest C ity associations. Pharmacogenomics J 14:411–417. https​://doi. (2019) New insights into GluT1 mechanics during glucose trans- org/10.1038/tpj.2014.13 fer. Sci Rep 9:998. https://doi.org/10.1038/s4159​ 8-018-37367​ -z​ Davydov EV, Goode DL, Sirota M, Cooper GM, Sidow A, Batzoglou S Gorbunov D, Gorboulev V, Shatskaya N, Mueller T, Bamberg E, (2010) Identifying a High Fraction of the Human Genome to be Friedrich T, Koepsell H (2008) High-afnity cation binding under Selective Constraint Using GERP ++. PLoS Comput Biol to organic cation transporter 1 induces movement of helix 11 6:e1001025-13. https​://doi.org/10.1371/journ​al.pcbi.10010​25 and blocks transport after mutations in a modeled interaction De Giorgis V, Veggiotti P (2013) GLUT1 deficiency syndrome domain between two helices. Mol Pharmacol 73:50–61. https​:// 2013: current state of the art. Seizure 22:803–811. https​://doi. doi.org/10.1124/mol.107.04017​0 org/10.1016/j.seizu​re.2013.07.003 Gordon AS, Fulton RS, Qin X, Mardis ER, Nickerson DA, Scherer Deng D, Xu C, Sun P, Wu J, Yan C, Hu M, Yan N (2014) Crystal S (2016) PGRNseq: a targeted capture sequencing panel for structure of the human glucose transporter GLUT1. Nature pharmacogenetic research and implementation. Pharmacogenet 510:121–125. https​://doi.org/10.1038/natur​e1330​6 Genom. https​://doi.org/10.1097/FPC.00000​00000​00020​2 Di Paolo A, Polillo M, Capecchi M, Cervetti G, Baratè C, Angelini S, Gulilat M, Lamb T, Teft WA, Wang J, Dron JS, Robinson JF, Tirona Guerrini F, Fontanelli G, Arici R, Ciabatti E, Grassi S, Bocci RG, Hegele RA, Kim RB, Schwarz UI (2019) Targeted next gen- G, Hrelia P, Danesi R, Petrini M, Galimberti S (2014) The eration sequencing as a tool for precision medicine. BMC Med c.480C > G polymorphism of hOCT1 infuences imatinib clear- Genomics 12:81. https​://doi.org/10.1186/s1292​0-019-0527-2 ance in patients afected by chronic myeloid leukemia. Pharma- Henrie A, Hemphill SE, Ruiz-Schultz N, Cushman B, DiStefano MT, cogenomics J 14:328–335. https​://doi.org/10.1038/tpj.2014.7 Azzariti D, Harrison SM, Rehm HL, Eilbeck K (2018) ClinVar Doehring A, Hofmann WP, Schlecker C, Zeuzem S, Sarrazin C, Berg Miner: demonstrating utility of a Web-based tool for viewing T, Müller T, Herrmann E, Geisslinger G, Lötsch J (2011) Role and fltering ClinVar data. Hum Mutat 39:1051–1060. https​:// of nucleoside transporters SLC28A2/3 and SLC29A1/2 genetics doi.org/10.1002/humu.23555​ in ribavirin therapy: protection against anemia in patients with Höglund PJ, Nordström KJV, Schiöth HB, Fredriksson R (2011) The chronic hepatitis C. Pharmacogenet Genomics 21:289–296. https​ solute carrier families have a remarkably long evolutionary his- ://doi.org/10.1097/FPC.0b013​e3283​4412e​7 tory with the majority of the human families present before diver- Dong C, Wei P, Jian X, Gibbs R, Boerwinkle E, Wang K, Liu X (2015) gence of Bilaterian species. Mol Biol Evol 28:1531–1541. https​ Comparison and integration of deleteriousness prediction meth- ://doi.org/10.1093/molbe​v/msq35​0 ods for nonsynonymous SNVs in whole exome sequencing stud- Huang Q, Yin J-Y, Dai X-P, Wu J, Chen X, Deng C-S, Yu M, Gong ies. Hum Mol Genet 24:2125–2137. https://doi.org/10.1093/hmg/​ Z-C, Zhou H-H, Liu Z-Q (2010) Association analysis of ddu73​3 SLC30A8 rs13266634 and rs16889462 polymorphisms with type Drögemöller BI, Monzon JG, Bhavsar AP, Borrie AE, Brooks B, 2 diabetes mellitus and repaglinide response in Chinese patients. Wright GEB, Liu G, Renouf DJ, Kollmannsberger CK, Bedard Eur J Clin Pharmacol 66:1207–1215. https​://doi.org/10.1007/ PL, Aminkeng F, Amstutz U, Hildebrand CA, Gunaretnam EP, s0022​8-010-0882-6 Critchley C, Chen Z, Brunham LR, Hayden MR, Ross CJD, Huang L, Zhang T, Xie C, Liao X, Yu Q, Feng J, Ma H, Dai J, Li M, Gelmon KA, Carleton BC (2017) Association between SLC16A5 Chen J, Zang A, Wang Q, Ge S, Qin K, Cai J, Yuan X (2013) genetic variation and cisplatin-induced ototoxic efects in adult SLCO1B1 and SLC19A1 gene variants and irinotecan-induced patients with testicular cancer. JAMA Oncol 3:1558–1562. https​ rapid response and survival: a prospective multicenter phar- ://doi.org/10.1001/jamao​ncol.2017.0502 macogenetics study of metastatic colorectal cancer. PLoS One Dujic T, Zhou K, Donnelly LA, Tavendale R, Palmer CNA, Pearson 8:e77223. https​://doi.org/10.1371/journ​al.pone.00772​23 ER (2015) Association of organic cation transporter 1 with

1 3 Human Genetics (2019) 138:1359–1377 1375

Ingelman-Sundberg M, Mkrtchian S, Zhou Y, Lauschke VM (2018) hepatocellular carcinoma. Oncotarget 6:16449–16460. https​:// Integrating rare genetic variants into pharmacogenetic drug doi.org/10.18632​/oncot​arget​.3758 response predictions. Human Genomics 12:26. https​://doi. Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fen- org/10.1186/s4024​6-018-0157-3 nell T, O’Donnell-Luria AH, Ware JS, Hill AJ, Cummings BB, Johnson DC, Corthals SL, Walker BA, Ross FM, Gregory WM, Dick- Tukiainen T, Birnbaum DP, Kosmicki JA, Duncan LE, Estrada ens NJ, Lokhorst HM, Goldschmidt H, Davies FE, Durie BGM, K, Zhao F, Zou J, Pierce-Hofman E, Berghout J, Cooper DN, Van Ness B, Child JA, Sonneveld P, Morgan GJ (2011) Genetic Defaux N, DePristo M, Do R, Flannick J, Fromer M, Gauthier L, factors underlying the risk of thalidomide-related neuropathy in Goldstein J, Gupta N, Howrigan D, Kiezun A, Kurki MI, Moon- patients with multiple myeloma. J Clin Oncol 29:797–804. https​ shine AL, Natarajan P, Orozco L, Peloso GM, Poplin R, Rivas ://doi.org/10.1200/JCO.2010.28.0792 MA, Ruano-Rubio V, Rose SA, Ruderfer DM, Shakir K, Stenson Kaniwa N, Kurose K, Jinno H, Tanaka-Kagawa T, Saito Y, Saeki M, PD, Stevens C, Thomas BP, Tiao G, Tusie-Luna MT, Weisburd Sawada J-i, Tohkin M, Hasegawa R (2005) Racial variability in B, Won H-H, Yu D, Altshuler DM, Ardissino D, Boehnke M, haplotype frequencies of UGT1A1 and glucuronidation activity Danesh J, Donnelly S, Elosua R, Florez JC, Gabriel SB, Getz G, of a novel single nucleotide polymorphism 686C>T (P229L) Glatt SJ, Hultman CM, Kathiresan S, Laakso M, McCarroll S, found in an African–American. Drug Metab Dispos 33:458–465. McCarthy MI, McGovern D, McPherson R, Neale BM, Palotie https​://doi.org/10.1124/dmd.104.00180​0 A, Purcell SM, Saleheen D, Scharf JM, Sklar P, Sullivan PF, Kelley LA, Mezulis S, Yates CM, Wass MN, Sternberg MJE (2015) Tuomilehto J, Tsuang MT, Watkins HC, Wilson JG, Daly MJ, The Phyre2 web portal for protein modeling, prediction and MacArthur DG, Consortium EA (2016) Analysis of protein-cod- analysis. Nat Protoc 10:845–858. https​://doi.org/10.1038/nprot​ ing genetic variation in 60,706 humans. Nature 536:285–291. .2015.053 https​://doi.org/10.1038/natur​e1905​7 Kim J-H, Karpyak VM, Biernacka JM, Nam HW, Lee MR, Preuss Li J, Zhao T, Zhang Y, Zhang K, Shi L, Chen Y, Wang X, Sun Z (2018) UW, Zill P, Yoon G, Colby C, Mrazek DA, Choi D-S (2011) Performance evaluation of pathogenicity-computation methods Functional role of the polymorphic 647 T/C variant of ENT1 for missense variants. Nucleic Acids Res 46:7793–7804. https://​ (SLC29A1) and its association with alcohol withdrawal seizures. doi.org/10.1093/nar/gky67​8 PLoS One 6:e16331. https://doi.org/10.1371/journ​ al.pone.00163​ ​ Lima A, Bernardes M, Sousa H, Azevedo R, Costa L, Ventura F, Seabra 31 V, Medeiros R (2014) SLC19A1 80G allele as a biomarker of Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, Shendure J methotrexate-related gastrointestinal toxicity in Portuguese rheu- (2014) A general framework for estimating the relative path- matoid arthritis patients. Pharmacogenomics 15:807–820. https​ ogenicity of human genetic variants. Nat Rev Drug Discov ://doi.org/10.2217/pgs.13.244 46:310–315. https​://doi.org/10.1038/ng.2892 Lin L, Yee SW, Kim RB, Giacomini KM (2015) SLC transporters as Klein K, Tremmel R, Winter S, Fehr S, Battke F, Scheurenbrand T, therapeutic targets: emerging opportunities. Nat Rev Drug Dis- Schaefeler E, Biskup S, Schwab M, Zanger UM (2019) A new covery 14:543–560. https​://doi.org/10.1038/nrd46​26 panel-based next-generation sequencing method for ADME Lu YB, Kobayashi K, Ushikai M, Tabata A, Iijima M, Li MX, Lei genes reveals novel associations of common and rare variants L, Kawabe K, Taura S, Yang Y, Liu T-T, Chiang S-H, Hsiao with expression in a human liver cohort. Front Genet 10:56. https​ K-J, Lau Y-L, Tsui L-C, Lee DH, Saheki T (2005) Frequency ://doi.org/10.3389/fgene​.2019.00007​ and distribution in East Asia of 12 mutations identifed in the Koepsell H (2013) The SLC22 family with transporters of organic SLC25A13 gene of Japanese patients with citrin defciency. cations, anions and zwitterions. Mol Aspects Med 34:413–435. J Hum Genet 50:338–346. https​://doi.org/10.1007/s1003​ https​://doi.org/10.1016/j.mam.2012.10.010 8-005-0262-8 Kozyra M, Ingelman-Sundberg M, Lauschke VM (2017) Rare genetic Machiela MJ, Chanock SJ (2015) LDlink: a web-based application variants in cellular transporters, metabolic enzymes, and nuclear for exploring population-specifc haplotype structure and linking receptors can be important determinants of interindividual dif- correlated alleles of possible functional variants. Bioinformatics ferences in drug response. Genet Med 19:20–29. https​://doi. 31:3555–3557. https​://doi.org/10.1093/bioin​forma​tics/btv40​2 org/10.1038/gim.2016.33 Makhtar SM, Husin A, Baba AA, Ankathil R (2018) Genetic varia- Kwon JS, Joo YH, Nam HJ, Lim M, Cho E-Y, Jung MH, Choi J-S, tions in infux transporter gene SLC22A1 are associated with Kim B, Kang D-H, Oh S, Park T, Hong KS (2009) Associa- clinical responses to imatinib mesylate among Malaysian chronic tion of the glutamate transporter gene SLC1A1 with atypical myeloid leukaemia patients. Journal of Genetics 97:835–842. antipsychotics-induced obsessive-compulsive symptoms. Arch https​://doi.org/10.1007/s1204​1-018-0978-9 Gen Psychiatry 66:1233–1241. https​://doi.org/10.1001/archg​ Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorf LA, Hunter enpsy​chiat​ry.2009.155 DJ, McCarthy MI, Ramos EM, Cardon LR, Chakravarti A, Cho Lanvers-Kaminsky C, Sprowl JA, Malath I, Deuster D, Eveslage M, JH, Guttmacher AE, Kong A, Kruglyak L, Mardis E, Rotimi Schlatter E, Mathijssen RH, Boos J, Jürgens H, am Zehnhof- CN, Slatkin M, Valle D, Whittemore AS, Boehnke M, Clark Dinnesen AG, Sparreboom A, Ciarimboli G (2015) Human AG, Eichler EE, Gibson G, Haines JL, Mackay TFC, McCa- OCT2 variant c.808G>T confers protection efect against cispl- rroll SA, Visscher PM (2009) Finding the missing heritability of atin-induced ototoxicity. Pharmacogenomics 16:323–332. https​ complex diseases. Nature 461:747–753. https://doi.org/10.1038/​ ://doi.org/10.2217/pgs.14.182 natur​e0849​4 Lauschke VM, Ingelman-Sundberg M (2016) Precision medicine and Moncriefe H, Hinks A, Ursu S, Kassoumeri L, Etheridge A, Hubank rare genetic variants. Trends Pharmacol Sci 37:85–86. https​:// M, Martin P, Weiler T, Glass DN, Thompson SD, Thomson W, doi.org/10.1016/j.tips.2015.10.006 Wedderburn LR (2010) Generation of novel pharmacogenomic Lauschke VM, Ingelman-Sundberg M (2018) How to consider rare candidates in response to methotrexate in juvenile idiopathic genetic variants in personalized drug therapy. Clin Pharmacol arthritis: correlation between gene expression and genotype. Ther 103:745–748. https​://doi.org/10.1002/cpt.976 Pharmacogenet Genomics 20:665–676. https://doi.org/10.1097/​ Lee Y-S, Kim BH, Kim BC, Shin A, Kim JS, Hong S-H, Hwang J-A, FPC.0b013​e3283​3f2cd​0 Lee JA, Nam S, Lee SH, Bhak J, Park J-W (2015) SLC15A2 Mueckler M, Thorens B (2013) The SLC2 (GLUT) family of mem- genomic variation is associated with the extraordinary response brane transporters. Mol Aspects Med 34:121–138. https​://doi. of sorafenib treatment: whole-genome analysis in patients with org/10.1016/j.mam.2012.07.001

1 3 1376 Human Genetics (2019) 138:1359–1377

Ng PC, Henikof S (2001) Predicting deleterious amino acid substitu- potassium-chloride Co-transporter 3 is a major pathogenic mech- tions. Genome Res 11:863–874. https://doi.org/10.1101/gr.17660​ 1​ anism in hereditary motor and sensory neuropathy with agenesis Orioli T, Vihinen M (2019) Benchmarking subcellular localization of the corpus callosum. J Biol Chem 286:28456–28465. https​:// and variant tolerance predictors on membrane proteins. BMC doi.org/10.1074/jbc.M111.22689​4 Genomics 20:547. https​://doi.org/10.1186/s1286​4-019-5865-0 Santos M, Niemi M, Hiratsuka M, Kumondai M, Ingelman-Sundberg Park M-S (2015) Molecular dynamics simulations of the human glu- M, Lauschke VM, Rodríguez-Antona C (2018) Novel copy-num- cose transporter GLUT1. PLoS One 10:e0125361. https​://doi. ber variations in pharmacogenes contribute to interindividual org/10.1371/journ​al.pone.01253​61 diferences in drug pharmacokinetics. Genet Med 20:622–629. Park HJ, Shaukat S, Liu X-Z, Hahn SH, Naz S, Ghosh M, Kim H-N, https​://doi.org/10.1038/gim.2017.156 Moon S-K, Abe S, Tukamoto K, Riazuddin S, Kabra M, Erdene- Schlessinger A, Yee SW, Sali A, Giacomini KM (2013) SLC classi- tungalag R, Radnaabazar J, Khan S, Pandya A, Usami S-I, Nance fcation: an update. Clin Pharmacol Ther 94:19–23. https​://doi. WE, Wilcox ER, Grifth AJ (2003) Origins and frequencies of org/10.1038/clpt.2013.73 SLC26A4 (PDS) mutations in east and south Asians: global Seitz T, Stalmann R, Dalila N, Chen J, Pojar S, dos Santos Pereira JN, implications for the epidemiology of deafness. J Med Genet Krätzner R, Brockmöller J, Tzvetkov MV (2015) Global genetic 40:242–248. https​://doi.org/10.1136/jmg.40.4.242 analyses reveal strong inter-ethnic variability in the loss of activ- Pedersen BP, Kumar H, Waight AB, Risenmay AJ, Roe-Zurz Z, Chau BH, ity of the organic cation transporter OCT1. Genome Med 7:56. Schlessinger A, Bonomi M, Harries W, Sali A, Johri AK, Stroud https​://doi.org/10.1186/s1307​3-015-0172-0 RM (2013) Crystal structure of a eukaryotic phosphate transporter. Shihab HA, Gough J, Cooper DN, Stenson PD, Barker GLA, Edwards Nature 496:533–536. https​://doi.org/10.1038/natur​e1204​2 KJ, Day INM, Gaunt TR (2012) Predicting the functional, molec- Pellicer M, García-González X, García MI, Robles L, Grávalos C, ular, and phenotypic consequences of amino acid substitutions García-Alfonso P, Pachón V, Longo F, Martínez V, Blanco C, using hidden Markov models. Hum Mutat 34:57–65. https://doi.​ Iglesias I, Sanjurjo M, López-Fernández LA (2017) Identif- org/10.1002/humu.22225​ cation of new SNPs associated with severe toxicity to capecit- Shihab HA, Rogers MF, Gough J, Mort M, Cooper DN, Day INM, abine. Pharmacol Res 120:133–137. https​://doi.org/10.1016/j. Gaunt TR, Campbell C (2015) An integrative approach to pre- phrs.2017.03.021 dicting the functional efects of non-coding and coding sequence Popp C, Gorboulev V, Müller TD, Gorbunov D, Shatskaya N, Koep- variation. Bioinformatics 31:1536–1543. https://doi.org/10.1093/​ sell H (2005) Amino acids critical for substrate afnity of rat bioin​forma​tics/btv00​9 organic cation transporter 1 line the substrate binding region Shin S-W, Park BL, Chang H, Park JS, Bae D-J, Song H-J, Choi IS, in a model derived from the tertiary structure of lactose per- Kim M-K, Park H-S, Kim LH, Namgoong S, Kim JO, Shin HD, mease. Mol Pharmacol 67:1600–1611. https​://doi.org/10.1124/ Park CS (2014) Exonic variants associated with development of mol.104.00883​9 aspirin exacerbated respiratory diseases. PLoS One 9:e111887. Pras E, Raben N, Golomb E, Arber N, Aksentijevich I, Schapiro JM, https​://doi.org/10.1371/journ​al.pone.01118​87 Harel D, Katz G, Liberman U, Pras M (1995) Mutations in Shu Y, Leabman MK, Feng B, Mangravite LM, Huang CC, Stryke the SLC3A1 transporter gene in cystinuria. Am J Hum Genet D, Kawamoto M, Johns SJ, DeYoung J, Carlson E, Ferrin TE, 56:1297–1303 Herskowitz I, Giacomini KM, Investigators POMT (2003) Evo- Qiu H-B, Zhuang W, Wu T, Xin S, Lin C-Z, Ruan H-L, Zhu X, Huang lutionary conservation predicts function of variants of the human M, Li J-L, Hou X-Y, Zhou Z-W, Wang XD (2018) Imatinib- organic cation transporter, OCT1. Proc Natl Acad Sci 100:5902– induced ophthalmological side-effects in GIST patients are 5907. https​://doi.org/10.1073/pnas.07308​58100​ associated with the variations of EGFR, SLC22A1, SLC22A5 Shu Y, Sheardown SA, Brown C, Owen RP, Zhang S, Castro RA, Ian- and ABCB1. Pharmacogenomics J 18:460–466. https​://doi. culescu AG, Yue L, Lo JC, Burchard EG, Brett CM, Giacomini org/10.1038/tpj.2017.40 KM (2007) Efect of genetic variation in the organic cation trans- Quang D, Chen Y, Xie X (2015) DANN: a deep learning approach for porter 1 (OCT1) on metformin action. J Clin Investig 117:1422– annotating the pathogenicity of genetic variants. Bioinformatics 1431. https​://doi.org/10.1172/JCI30​558 31:761–763. https​://doi.org/10.1093/bioin​forma​tics/btu70​3 Song IS, Shin HJ, Shim EJ, Jung IS, Kim WY, Shon JH, Shin JG (2008) Rau M, Stickel F, Russmann S, Manser CN, Becker PP, Weisskopf M, Genetic variants of the organic cation transporter 2 infuence Schmitt J, Dill MT, Dufour J-F, Moradpour D, Semela D, Mül- the disposition of metformin. Clin Pharmacol Ther 84:559–562. lhaupt B, Geier A, (SCCS) SHCCSG (2013) Impact of genetic https​://doi.org/10.1038/clpt.2008.61 SLC28 transporter and ITPA variants on ribavirin serum level, Soo RA, Wang LZ, Ng SS, Chong PY, Yong WP, Lee SC, Liu JJ, hemoglobin drop and therapeutic response in patients with Choo TB, Tham LS, Lee HS, Goh BC, Soong R (2009) Distri- HCV infection. J Hepatol 58:669–675. https://doi.org/10.1016/j.​ bution of gemcitabine pathway genotypes in ethnic Asians and jhep.2012.11.027 their association with outcome in non-small cell lung cancer Reva B, Antipin Y, Sander C (2011) Predicting the functional impact of patients. Lung Cancer 63:121–127. https​://doi.org/10.1016/j. protein mutations: application to cancer genomics. Nucleic Acids lungc​an.2008.04.010 Res 39:e118. https​://doi.org/10.1093/nar/gkr40​7 Stocker SL, Morrissey KM, Yee SW, Castro RA, Xu L, Dahlin A, Ram- Romero MF, Chen A-P, Parker MD, Boron WF (2013) The SLC4 irez AH, Roden DM, Wilke RA, McCarty CA, Davis RL, Brett − family of bicarbonate ­(HCO3 ) transporters. Mol Aspects Med CM, Giacomini KM (2013) The efect of novel promoter variants 34:159–182. https​://doi.org/10.1016/j.mam.2012.10.008 in MATE1 and MATE2 on the pharmacokinetics and pharmaco- Salinas-Delgado Y, Galaviz-Hernández C, Toral RG, Ávila Rejón CA, dynamics of metformin. Clin Pharmacol Ther 93:186–194. https​ Reyes-Lopez MA, Martínez AR, Martínez-Aguilar G, Sosa- ://doi.org/10.1038/clpt.2012.210 Macías M (2015) The D543N polymorphism of the SLC11A1/ Sundelin E, Gormsen LC, Jensen JB, Vendelbo MH, Jakobsen S, NRAMP1 gene is associated with treatment failure in male Munk OL, Christensen M, Brøsen K, Frøkiaer J, Jessen N (2017) patients with pulmonary tuberculosis. Drug Metab Personal Ther Genetic polymorphisms in organic cation transporter 1 attenuates 30:211–214. https​://doi.org/10.1515/dmpt-2015-0019 hepatic metformin exposure in humans. Clin Pharmacol Ther Salin-Cantegrel A, Rivière J-B, Shekarabi M, Rasheed S, Dacal 102:841–848. https​://doi.org/10.1002/cpt.701 S, Laganière J, Gaudet R, Rochefort D, Lesca G, Gaspar C, Suthandiram S, Gan G-G, Zain SM, Bee P-C, Lian L-H, Chang K-M, Dion PA, Lapointe J-Y, Rouleau GA (2011) Transit defect of Ong T-C, Mohamed Z (2014) Efect of polymorphisms within

1 3 Human Genetics (2019) 138:1359–1377 1377

methotrexate pathway genes on methotrexate toxicity and plasma Hayden MR, Ross CJ, consortium C (2015) Genetic variants levels in adults with hematological malignancies. Pharmacog- in SLC22A17 and SLC22A7 are associated with anthracycline- enomics 15:1479–1494. https​://doi.org/10.2217/pgs.14.97 induced cardiotoxicity in children. Pharmacogenomics 16:1065– Tarasova L, Kalnina I, Geldnere K, Bumbure A, Ritenberga R, Niki- 1076. https​://doi.org/10.2217/pgs.15.61 tina-Zake L, Fridmanis D, Vaivade I, Pirags V, Klovins J (2012) Volk C, Gorboulev V, Kotzsch A, Müller TD, Koepsell H (2009) Five Association of genetic variation in the organic cation transporters amino acids in the innermost cavity of the substrate binding cleft OCT1, OCT2 and multidrug and toxin extrusion 1 transporter of organic cation transporter 1 interact with extracellular and protein genes with the gastrointestinal side efects and lower BMI intracellular corticosterone. Mol Pharmacol 76:275–289. https​ in metformin-treated type 2 diabetes patients. Pharmacogenet ://doi.org/10.1124/mol.109.05478​3 Genomics 22:659–666. https://doi.org/10.1097/FPC.0b013​ e3283​ ​ Vormfelde SV, Sehrt D, Toliat MR, Schirmer M, Meineke I, Tzvet- 56166​6 kov M, Nürnberg P, Brockmöller J (2007) Genetic variation in Tecza K, Pamula-Pilat J, Lanuszewska J, Butkiewicz D, Grzybowska the renal sodium transporters NKCC2, NCC, and ENaC in rela- E (2018) Pharmacogenetics of toxicity of 5-fuorouracil, doxo- tion to the efects of loop diuretic drugs. Clin Pharmacol Ther rubicin and cyclophosphamide chemotherapy in breast cancer 82:300–309. https​://doi.org/10.1038/sj.clpt.61001​31 patients. Oncotarget 9:9114–9136. https​://doi.org/10.18632​/ Wang S-M, Sun L-L, Zeng W-X, Wu W-s, Zhang G-l (2014) Efects of oncot​arget​.24148​ a microRNA binding site polymorphism in SLC19A1 on metho- Thornton JK, Taylor KM, Ford D, Valentine RA (2011) Diferential trexate concentrations in Chinese children with acute lympho- subcellular localization of the splice variants of the zinc trans- blastic leukemia. Med Oncol 31:62. https​://doi.org/10.1007/ porter ZnT5 is dictated by the diferent C-terminal regions. PLoS s1203​2-014-0062-0 One 6:e23878. https​://doi.org/10.1371/journ​al.pone.00238​78 Watkins DB, Hughes TP, White DL (2015) OCT1 and imatinib trans- Tishkof SA, Reed FA, Friedlaender FR, Ehret C, Ranciaro A, Froment port in CML: is it clinically relevant? Leukemia 29:1960–1969. A, Hirbo JB, Awomoyi AA, Bodo J-M, Doumbo O, Ibrahim M, https​://doi.org/10.1038/leu.2015.170 Juma AT, Kotze MJ, Lema G, Moore JH, Mortensen H, Nyambo Wevers RA, Christensen M, Engelke UFH, Geuer S, Coene KLM, TB, Omar SA, Powell K, Pretorius GS, Smith MW, Thera MA, Kwast JT, Lund AM, Vissers LELM (2019) Functional disrup- Wambebe C, Weber JL, Williams SM (2009) The genetic struc- tion of pyrimidine nucleoside transporter CNT1 results in a ture and history of Africans and African Americans. Science novel inborn error of metabolism with high excretion of uri- 324:1035–1044. https​://doi.org/10.1126/scien​ce.11722​57 dine and cytidine. J Inherit Metab Dis 42:494–500. https​://doi. Todd JN, Florez JC (2014) An update on the pharmacogenomics of org/10.1002/jimd.12081​ metformin: progress, problems and potential. Pharmacogenomics Wright NJ, Lee S-Y (2019) Structures of human ENT1 in complex with 15:529–539. https​://doi.org/10.2217/pgs.14.21 adenosine reuptake inhibitors. Nat Struct Mol Biol 26:599–606. Torrents D, Mykkänen J, Pineda M, Feliubadaló L, Estévez R, de Cid https​://doi.org/10.1038/s4159​4-019-0245-7 R, Sanjurjo P, Zorzano A, Nunes V, Huoponen K, Reinikainen A, Wright GEB, Carleton B, Hayden MR, Ross CJD (2018) The global Simell O, Savontaus ML, Aula P, Palacín M (1999) Identifcation spectrum of protein-coding pharmacogenomic diversity. Pharma- of SLC7A7, encoding y + LAT-1, as the lysinuric protein intoler- cogenomics J 18:187–195. https​://doi.org/10.1038/tpj.2016.77 ance gene. Nat Genet 21:293–296. https://doi.org/10.1038/6809​ Yan N (2013) Structural advances for the major facilitator superfamily Tzvetkov MV (2017) OCT1 pharmacogenetics in pain management: is (MFS) transporters. Trends Biochem Sci 38:151–159. https://doi.​ a clinical application within reach? Pharmacogenomics 18:1515– org/10.1016/j.tibs.2013.01.003 1523. https​://doi.org/10.2217/pgs-2017-0095 Yee SW, Nguyen AN, Brown C, Savic RM, Zhang Y, Castro RA, Tzvetkov MV, Saadatmand AR, Lotsch J, Tegeder I, Stingl JC, Brock- Cropp CD, Choi JH, Singh D, Tahara H, Stocker SL, Huang möller J (2011) Genetically polymorphic OCT1: another piece Y, Brett CM, Giacomini KM (2013) Reduced renal clearance in the puzzle of the variable pharmacokinetics and pharmacody- of cefotaxime in asians with a low-frequency polymorphism of namics of the opioidergic drug tramadol. Clin Pharmacol Ther OAT3 (SLC22A8). J Pharm Sci 102:3451–3457. https​://doi. 90:143–150. https​://doi.org/10.1038/clpt.2011.56 org/10.1002/jps.23581​ Tzvetkov MV, Saadatmand AR, Bokelmann K, Meineke I, Kaiser Yoon H, Cho H-Y, Yoo H-D, Kim S-M, Lee Y-B (2013) Infuences R, Brockmöller J (2012) Efects of OCT1 polymorphisms on of organic cation transporter polymorphisms on the population the cellular uptake, plasma concentrations and efcacy of the pharmacokinetics of metformin in healthy subjects. AAPS J 5-HT(3) antagonists tropisetron and ondansetron. Pharmacog- 15:571–580. https​://doi.org/10.1208/s1224​8-013-9460-z enomics J 12:22–29. https​://doi.org/10.1038/tpj.2010.75 Zhang B, Lauschke VM (2019) Genetic variability and popula- Tzvetkov MV, dos Santos Pereira JN, Meineke I, Saadatmand AR, tion diversity of the human SLCO (OATP) transporter fam- Stingl JC, Brockmöller J (2013) Morphine is a substrate of the ily. Pharmacol Res 139:550–559. https​://doi.org/10.1016/j. organic cation transporter OCT1 and polymorphisms in OCT1 phrs.2018.10.017 gene afect morphine pharmacokinetics after codeine administra- Zhou Y, Lauschke VM (2018) Comprehensive overview of the pharma- tion. Biochem Pharmacol 86:666–678. https://doi.org/10.1016/j.​ cogenetic diversity in Ashkenazi Jews. J Med Genet 55:617–627. bcp.2013.06.019 https​://doi.org/10.1136/jmedg​enet-2018-10542​9 Tzvetkov MV, Seitz T, Bokelmann K, Mueller T, Brockmöller J, Koep- Zhou Y, Fujikura K, Mkrtchian S, Lauschke VM (2018) Computa- sell H (2014) Does the haplotype Met408-Del420, which was tional methods for the pharmacogenetic interpretation of next apparently predictive for imatinib efcacy, really exist and how generation sequencing data. Front Pharmacol 9:1437. https://doi.​ strongly may it afect OCT1 activity? Blood 123:1427–1429. org/10.3389/fphar​.2018.01437​ https​://doi.org/10.1182/blood​-2013-11-53586​4 Zhou Y, Mkrtchian S, Kumondai M, Hiratsuka M, Lauschke VM Urban TJ, Brown C, Castro RA, Shah N, Mercer R, Huang Y, Brett (2019) An optimized prediction framework to assess the func- CM, Burchard EG, Giacomini KM (2008) Efects of genetic tional impact of pharmacogenetic variants. Pharmacogenom J variation in the novel organic cation transporter, OCTN1, on the 19:115–126. https​://doi.org/10.1038/s4139​7-018-0044-2 renal clearance of gabapentin. Clin Pharmacol Ther 83:416–421. https​://doi.org/10.1038/sj.clpt.61002​71 Publisher’s Note Springer Nature remains neutral with regard to Visscher H, Rassekh SR, Sandor GS, Caron HN, van Dalen EC, Kre- jurisdictional claims in published maps and institutional afliations. mer LC, van der Pal HJ, Rogers PC, Rieder MJ, Carleton BC,

1 3