Highly Multiplexed Genotyping of Thiopurine S-Methyltransferase ...

Clinical Chemistry 54:10 1637–1647 (2008)

Molecular Diagnostics and Genetics

Highly Multiplexed Genotyping of Thiopurine S-Methyltransferase Variants Using MALDI-TOF Mass Spectrometry: Reliable Genotyping in Different Ethnic Groups Elke Schaeffeler,1 Ulrich M. Zanger,1 Michel Eichelbaum,1 Steven Asante-Poku,2 Jae-Gook Shin,3 and Matthias Schwab1,4*

BACKGROUND: To avoid severe hematotoxicity in patients, determination of the TPMT (thiopurine S-methyltransferase) genotype before commencing thiopurine therapy has become accepted. METHODS:

We used MALDI-TOF mass spectrometry (MS) based on Sequenom iPLEX威 technology to develop novel multiplex assays for comprehensive testing of TPMT. Two assays, a 15-plex and a 7-plex assay, consisting of multiplex PCR, shrimp alkaline phosphatase treatment, primer extension, and MALDI-TOF MS analysis, allow detection of all currently known functionally relevant 24 TPMT alleles (TPMT*2 to *18, *20 to *23). Previously identified variant DNA samples and newly constructed synthetic templates were used as quality controls. RESULTS:

Assay evaluation performed on a panel of 586 genomic DNA samples previously genotyped by other methods (denaturing HPLC, sequencing) resulted in 100% agreement. Analyses of the distribution of TPMT alleles in 116 samples from a Ghanaian population revealed a TPMT*8 allele frequency of 3.4%. In a Korean population of 118 unrelated individuals, we found a TPMT*6 allele frequency of 1.3%.

CONCLUSIONS: The newly developed multiplex MALDITOF MS assay allows efficient genotyping for all currently known functional TPMT variants. To achieve the most accurate prediction of TPMT phenotype, molecular diagnosis of TPMT should include all these variants.

© 2008 American Association for Clinical Chemistry

1

Margarete Fischer-Bosch Institute of Clinical Pharmacology, Stuttgart, Germany, and University of Tuebingen, Tuebingen, Germany; 2 University of Ghana Medical School, Accra, Ghana; 3 Department of Pharmacology and Pharmacogenomics Research Center, Inje University College of Medicine, Busan Paik Hospital, Busan, Korea; 4 Department of Clinical Pharmacology, University Hospital Tuebingen, Tuebingen, Germany. * Address correspondence to this author at: Dr Margarete Fischer-Bosch Institute of Clinical Pharmacology, Auerbachstrasse 112, D-70376 Stuttgart, Germany.

A key aspect of research in human genetics is associating sequence variations with heritable phenotypes. The most common variations are single nucleotide polymorphisms (SNP)5 that occur approximately once in every 100 –300 bases. SNP discovery and detection has recently been enhanced by whole-genome-association studies conducted in various human diseases or therapeutic settings (1 ). Genotyping of SNPs has, therefore, many applications in human genetic studies for the dissection of complex diseases and other clinical applications like pharmacogenetic testing. A growing list of polymorphisms found in genes encoding drug-metabolizing enzymes, drug transporters, and targets have been linked to drug effects in humans (2, 3 ). At present, one of the best-recognized examples of genetic polymorphisms that alter drug response in humans is TPMT (thiopurine S-methyltransferase)6 and its influence on the thiopurine drugs such as azathioprine and 6-mercaptopurine. Because these medications are predominantly inactivated by TPMT in hematopoietic cells, patients who inherit 2 nonfunctional TPMT alleles accumulate excessive concentrations of the active metabolites when treated with thiopurines. This can lead to severe and potentially life-threatening hematopoietic toxicity (3–5 ). The most common variant alleles with decreased activity are TPMT*3A, with a frequency of 4.5% in the white population, followed by TPMT*3C with a frequency of 0.4%, whereas all other mutations are rare (e.g., TPMT*2 in 0.17%) or found only in single cases (TPMT*4 to *23) (6 –17 ). Clinical diagnostic tests are available for detection of inactivating variants in the human TPMT gene (e.g., RFLP, Taq-

Fax ⫹49 711 85 92 95; e-mail [email protected]. Received January 16, 2008; accepted May 30, 2008. Previously published online at DOI: 10.1373/clinchem.2008.103457 5 Nonstandard abbreviations: SNP, single nucleotide polymorphism; MS, mass spectrometry. 6 Human genes: TPMT, thiopurine S-methyltransferase; ITPA, ITPase; ABCC4 and ABCC5, ATP-binding cassette, subfamily C [CFTR/MRP], members 4 and 5.

1637

Man, LightCycler, pyrosequencing); however, a timeand cost-saving assay to test simultaneously for all known inactivating TPMT variants is currently unavailable. Current methods for TPMT diagnostics depend on indirect detection of analytes, for instance by using fluorescent dyes, reporter enzymes, or radioactive labels. To meet the requirements of high-throughput genotyping as a highly automated process, MALDI-TOF mass spectrometry (MS) enables highresolution analyses of nucleic acids (18 –20 ). One novel assay featuring high multiplex levels (up to 36-plex), called iPLEX assay, is based on generating PCR products, which are subsequently treated with shrimp alkaline phosphatase, followed by primer extension reactions using mass-modified base terminators that achieve at least 16-Da gaps between the 4 possible bases incorporated into a single base extension. Moreover, it has been suggested that calculated cost of the iPLEX assay is substantially cheaper per genotype for highplex genotyping than the average cost using approaches such as the TaqMan technology (http://www.sequenom. com/Seq_genotyping.html) (21 ). In the present article, we report the novel establishment and application of the iPLEX assay for detection of all functional relevant 22 TPMT allelic variants. Moreover, we determined the frequency distribution of these TPMT alleles in study populations of white, African, and Asian origin. Materials and Methods STUDY POPULATION

We selected 586 individuals of both sexes from our previously published large-scale population of 1214 randomly recruited unrelated blood donors consecutively entering the Department of Transfusion Medicine, University Hospital Tuebingen, Germany (16 ). All individuals were of white (German) origin. As previously described, TPMT erythrocyte activities were available in all these subjects and demonstrated a concordance rate of 98.4% between TPMT phenotype and genotype (16 ). We used blood DNA samples from previously identified carriers of TPMT variants discriminating for the TPMT alleles *2, *3A, *3C, *9, *11, *16 to *18, and *20 to *22 as controls (15–17, 22 ). Additionally, blood DNA samples were obtained from 116 Ghanians and 118 Koreans. Healthy Ghanians were randomly recruited from hospital staff and medical students at Ghana Medical School, Accra, Ghana, and belonged mainly to the Ga tribe (23 ). Korean subjects were also healthy unrelated volunteers. Evaluation and confirmation of healthy status was by medical history, physical examination, and/or routine laboratory tests. 1638 Clinical Chemistry 54:10 (2008)

The study was approved by the respective local ethics committees in Germany, Ghana, and Korea, and all volunteers gave written informed consent. MULTIPLEX PCR AMPLIFICATION

Primers were designed using MassArray Assay design software (version 3.1), in combination with manual adjustment to exclude amplification of TPMT pseudogene using reference sequence AB045146.1 (see Supplemental Table 1 in the Data Supplement that accompanies the online version of this article at http:// www.clinchem.org/content/vol54/issue10). We verified the uniqueness of primer sequences by NCBI BLAST to exclude amplification of TPMT pseudogene. PCR reactions contained, in a volume of 8 ␮L, 1 pmol of the corresponding primers, 10 ng genomic DNA, and 1⫻ Hot Star Reaction Mix (Qiagen) in 384well plates. PCR conditions were as follows: 94 °C for 15 min, followed by 40 cycles of 94 °C (30 s), 56 °C (60 s), 72 °C (60 s), and a final extension of 72 °C for 3 min. SHRIMP ALKALINE PHOSPHATASE DIGEST

To dephosphorylate excess deoxyribonucleotide triphosphates, 0.3 units shrimp alkaline phosphatase (iPLEX Gold Reaction Kit; Sequenom) was added to the PCR reaction followed by incubation at 37 °C for 20 min. Shrimp alkaline phosphatase was denatured for 10 min at 85 °C. All pipetting steps were carried out using the Puredisk robot (CyBio). IPLEX PRIMER EXTENSION

For extension reaction, primers were designed using MassArray design software (see Supplemental Table 2 in the online Data Supplement). The iPLEX reaction mix consisted of 0.222⫻ iPLEX buffer, 1⫻ iPLEX termination mix, 1⫻ iPLEX enzyme (iPLEX Gold Reaction Kit; Sequenom), and either 0.625 ␮mol/L or 1.25 ␮mol/L primer. We divided the extension primers into a lowmass group and a high-mass group and doubled the concentration of the high-mass group. iPLEX PCR amplification was performed using 2-step 200 short-cycle programs. The sample was denatured at 94 °C, and strands were annealed at 52 °C for 5 s and extended at 80 °C for 5 s. The annealing and extension cycle was repeated 4 more times for a total of 5 cycles, looped back to a 94 °C denaturing step for 5 s, and then entered the 5cycle annealing and extension loop again. The 5 annealing and extension steps with the single denaturing step were repeated an additional 39 times for a total of 40. A final extension was done at 72 °C for 3 min. CLEAN RESIN

iPLEX reaction products were desalted by adding 6 mg resin (Clean Resin; Sequenom) and 16 ␮L water. After incubation for 30 min, we centrifuged the reaction

TPMT Genotyping by MALDI-TOF Mass Spectrometry

mixture at 3500g. All pipetting steps were carried out using the puredisk robot (CyBio). MALDI-TOF MS ANALYSIS AND DATA ANALYSIS

We dispensed samples onto a 384 SpectroCHIP威 Array using a Nanodispenser, and introduced SpectroCHIP Arrays into a MassArray Compact mass spectrometer. Automated spectra acquisition was performed using Spectroacquire (Sequenom). We performed data analysis with the MassArray Typer software version 3.4. Laboratory staff were blinded to case status of study participants. ALLELE NOMENCLATURE AND ADDITIONAL METHODS

We describe sequence variations according to published recommendations for description of sequence variants in DNA (24 ). Exons 8 and 10 were amplified from genomic DNA by PCR, purified, and sequenced using primers and conditions previously described (16 ). We compared observed and expected allele and genotype frequencies by means of Hardy-Weinberg equilibrium calculations (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl). CONSTRUCTION OF SYNTHETIC TEMPLATES

For all variants, we constructed synthetic templates bearing the variant allele as described (25 ). Results SELECTION OF TPMT VARIANTS FOR GENOTYPING

To choose relevant TPMT variants for the design of MALDI-TOF MS assays, we considered all previously published SNPs associated with functional consequences on erythrocyte TPMT activity, independent of their frequency distribution in various ethnic groups. The current TPMT nomenclature system includes 25 TPMT alleles (*2, *3A–D, *4 to *23), all of which except TPMT*19 result in severely decreased erythrocyte TPMT activity (13, 17, 26, 27 ). Because in vivo and in vitro experiments indicated that the TPMT*19 allele shows normal activity (26 ), this variant was not included in the present assay. Data for all SNPs finally included in the MALDI-TOF MS analysis are summarized in Table 1. TPMT IPLEX ASSAY

We performed MALDI-TOF MS– based SNP genotyping of all selected TPMT alleles (Table 1) using the iPLEX assay chemistry. Generally, iPLEX genotyping allows for routine multiplexing levels up to 36-plex. In the case of TPMT genotyping, however, this potential could not be finally achieved. One limitation was the requirement for highly specific amplification of TPMT gene fragments to effectively discriminate between the functional TPMT gene and its processed pseudogene.

Moreover, not all extension reactions could be carried out simultaneously, since their compatibility is strongly restricted by their primer sequences. Therefore, our final assay design consisted of 2 multiplex allele-specific primer extension assays, one 15-plex and one 7-plex (see Supplemental Table 2 in the online Data Supplement). We optimized quantities of primers to obtain highest yields of allele-specific products and to equilibrate signal-to-noise ratios. As masses increase, signal-to-noise ratios tend to decrease. In extreme cases, signals become indistinguishable from noise, resulting in calling errors. To adjust extension primers, primers were divided into a low-mass group and a high-mass group. All primers in the high-mass group were doubled in concentration. We evaluated the 2 assays by genotyping various DNA samples from sequenced positive controls with known TPMT genotypes (validation panel; TPMT*2, *3A, *3C, TPMT*9, *11, *16 to *18, *20 to *22). However, positive controls were not available for all selected alleles (TPMT*4 to *8, *10, *12 to *15, *23) and therefore we constructed synthetic templates carrying these variants by annealing appropriate oligonucleotides and filling the resulting partial duplexes. All synthetic alleles can be determined with the same primers used to genotype genomic DNA. Because the amount of DNA is limited, particularly in cases of rare TPMT variants, we constructed synthetic templates for all TPMT variants. Fig. 1 shows example MALDI-TOF MS spectra representing various homozygous and heterozygous genotypes. Unambiguous detection of all tested TPMT variants (Table 1) was possible. All these analyses were performed in duplicate. Notably, the triallelic variants 488G⬎A, C (TPMT*16 and *22, respectively) could be definitively distinguished using our iPLEX assay because the mass-modified base terminators are incorporated in a single base extension. VALIDATION OF IPLEX ASSAY–BASED GENOTYPING BY MALDI-TOF MS

To further validate the accuracy of the MALDI-TOF MS TPMT assay, we analyzed in a blinded fashion DNA samples from 586 previously genotyped healthy individuals (16, 17 ). All genotypes were unambiguously distinguished by MALDI-TOF MS, and genotype calls were in 100% concordance with the results of our previous study (Table 2). As standard procedure, 10% of randomly selected samples of the study population were genotyped in duplicate, resulting in a strict conformity. On average, about 5% were not automatically identified and had to be assigned manually, resulting in an overall automated call rate of ⬎95% for all SNPs. Supplemental Fig. 1 in the online Data Supplement shows an example mass spectrum that could not Clinical Chemistry 54:10 (2008) 1639

Table 1. Summary of TPMT variants selected for MALDI-TOF MS assay. Allele

Variant

Amino acid change 80

Reference

*2

5

238G⬎C

Ala Pro

(6 )

*3A

7

460G⬎A

Ala154Thr

(7 )

10

719A⬎G

Tyr240Cys

—

7

460G⬎A

Ala154Thr

(7 )

10

719A⬎G

Tyr240Cys

(7 )

5

292G⬎T

Glu98X

(8 )

7

460G⬎A

Ala154Thr

—

10

719A⬎G

Tyr240Cys

—

*3B *3C a

*3D

a

Exon

*4

—

G⬎A transition, splice site of intron 9

—

(8 )

*5

4

146T⬎C

Leu49Ser

(8 )

*6

8

539A⬎T

Tyr180Phe

(8 )

*7

10

681T⬎G

His227Gln

(9 )

215

*8

10

644G⬎A

Arg

His

(10 )

*9

5

356A⬎C

Lys119Thr

(16 )

*10

7

430G⬎C

Gly144Arg

(11 )

*11

6

395G⬎A

Cys132Tyr

(15 )

125

(12 )

*12

6

374C⬎T

Ser

*13

3

83A⬎T

Glu28Val

(12 )

*14

3

1A⬎G

Met1Val

(14 )

*15

—

G⬎A transition, splice site of intron 7

—

(14 )

163

Leu

*16

7

488G⬎A

Arg

His

(16 )

*17

3

124C⬎G

Gln42Glu

(16 )

*18

4

211G⬎A

Gly71Arg

(16 )

*20

10

712A⬎G

Lys238Glu

(17 )

69

*21

4

205C⬎G

Leu Val

(17 )

*22

7

488G⬎C

Arg163Pro

(17 )

*23

8

500C⬎G

Ala167Gly

(13 )

For genotyping of TPMT*3D, both key mutations of the *3 (460A⬎G and 719G⬎A) were determined.

be identified automatically because the signal intensity was slightly below the predefined threshold used automatically by the software. In addition, we visualized automatically determined relative allele signal intensities for each variant using a scatter plot. This visualization of peak intensities showed 3 clearly distinguished genotype clusters for each variant and, again, no discrepant genotype calls in comparison to our previous genotyping data. Fig. 2 depicts exemplary cluster plots for the SNPs 460 G⬎A and 719A⬎G. GENOTYPING OF DIFFERENT ETHNIC POPULATIONS

We used the newly established MALDI-TOF MS assay to perform a comprehensive TPMT genotype analysis in Africans (116 Ghanaians) and Asians (118 Koreans). For both populations, all genotypes were shown to be 1640 Clinical Chemistry 54:10 (2008)

in Hardy-Weinberg equilibrium. All analyses were done in duplicate with identical results. Table 3 summarizes the frequencies of detected TPMT alleles in the Korean and Ghanaian population. As expected, in both populations TPMT*3C was the most frequent deficient allele, at 6.5% and 2.5%, respectively. However, we found 8 heterozygous carriers of the TPMT*8 allele in Ghanaians and 3 carriers of the TPMT*6 allele in Koreans. Fig. 3A shows example spectra for both variants, 539A⬎T (*6) and 644G⬎A (*8), and Fig. 3B illustrates the respective cluster plots with a clear separation of heterozygous individuals. Sequencing of exon 8 and 10 in Ghanaians and Koreans carrying these variants confirmed the MALDI-TOF MS results. In the present article, not all previously genotyped 1214 Caucasian subjects were retyped for the presence of the TPMT*6 (exon 8) and TPMT*8 (exon 10) allele.


Fig. 1. MALDI TOF MS– based analysis of TPMT variants. (A), Examples of MALDI-TOF MS spectra of sequenced positive controls for known TPMT genotypes, including rare variants previously identified in our laboratory (TPMT*2, *3, *9, *11, *16 to *18, *20 to *22). Unambiguous detection of all tested TPMT variants including the triallelic variant 488G⬎A, C (TPMT*16 and *22) could be definitively distinguished. The respective peaks for these TPMT variants are marked by arrows. Because for each variant a multiplex MS spectrum is shown, additional peaks resulting from other variants are visible. continued on page 1642 Clinical Chemistry 54:10 (2008) 1641

Fig. 1. Continued. (B), MALDI-TOF MS spectra of synthetic control samples, carrying all 22 TPMT variants included in our multiplex assays. Masses of the respective unextended primers (P) and both extension products (allele 1 and allele 2) are marked by arrows. As expected, in all cases only peaks for allele 2 and not for allele 1 are visible in the spectrum. Mass is given as m/z.

1642 Clinical Chemistry 54:10 (2008)


In our previous work, however, all individuals with intermediate TPMT activity in erythrocytes were sequenced to exclude the presence of additional functional variants (16 ). Moreover, denaturing HPLC analysis of exon 10 in all 1214 study subjects did not reveal altered chromatograms (16 ). Thus, TPMT*6 and TPMT*8 appear to be absent in whites, at least those of German origin. Discussion Thiopurine S-methyltransferase is, at present, one of the best examples of genetic polymorphisms that alter drug response (4, 5, 28 ). During the last 10 years, several rare variants in different ethnic groups have been identified, and the underlying molecular mechanisms leading to deficient or intermediate TPMT activity have been comprehensively elucidated (4, 29 ). Pharmacogenetically TPMT-guided thiopurine therapy has been suggested for patients with acute lymphoblastic leukemia (30 ), inflammatory bowel diseases (31 ), and dermatological entities (32 ). The TPMT polymorphism is currently suggested to be the only meaningful biomarker in clinical practice to avoid severe thiopurine-related hematotoxicity, since evidence is lacking for other polymorphically expressed candidate genes involved in thiopurine metabolism [e.g., ITPA (ITPase)] and/or transport of metabolites [(e.g., ABCC4, ABCC5 (ATP-binding cassette, subfamily C [CFTR/MRP], members 4 and 5)] (5 ). MALDI-TOF MS offers a technique with high potential for valid multiplex genotyping. In contrast to other genotyping technologies, MALDI-TOF MS directly measures the molecular weight of the oligonucleotide and therefore has several principal and methodological advantages, which have been summarized in previous articles (19, 20, 33 ). Further advantages are the option for

Table 2. Genotyping results in the study population of 586 healthy whites using MALDI-TOF MS vs denaturing HPLC technology. TPMT genotype

MALDI-TOF MS

Denaturing HPLCa

*1/*1

526

526

*1/*2

2

2

*1/*3A

52

52

*1/*3C

4

4

*3A/*3A Total

2

2

586

586

a Data previously reported by Schaeffeler et al. (16 ). Results achieved 100% concordance.

highly automated processes and the relative ease of setting up multiplex assays, thus reducing time and cost while increasing sample throughput. To date, only more common variations of TPMT (TPMT*2 and TPMT*3A, *3C) have been included in most studies (5, 28 ), in part because comprehensive genotyping assays have not been available. Here, we established a MALDI-TOF MS TPMT genotyping platform based on the iPLEX assay chemistry for all currently known 24 TPMT alleles that have pronounced effects on expression/function (Table 1). The iPLEX assay was originally developed for plexes up to 36 compared with conventional homogeneous mass extend (hME) assay (18 ). Our aim was to establish an assay that can be applied in the routine testing of patients, and therefore all functional relevant TPMT SNPs should be integrated. By the use of the normal hME assay, up to 5 assays would be necessary to genotype the relevant 22 TPMT variants. Moreover, since it cannot be excluded that novel TPMT SNPs will be detected, the iPLEX assay offers the advantage of incorporating novel SNPs without changing the existing assay. Key elements in the analytical validation of a novel genotyping method should include evaluation of specificity, analytical sensitivity, reproducibility, and accuracy of the method, as recently recommended (34 ). The specificity relates to the ability of the assay to accurately detect genetic variants without reacting with or detecting related DNA sequences (e.g., pseudogenes) that may interfere with the assay. We evaluated the specificity of our PCR primers in silico during the process of PCR primer design by NCBI BLAST. For all TPMT SNPs of interest, suitable PCR primers were designed particularly with respect to whether the PCR primer binding sites are unique to the TPMT gene and not to the TPMT processed pseudogene or other DNA sequences elsewhere in the genome, and whether all primers were specific to amplify the target TPMT SNP. Of note, not all 22 selected TPMT SNPs could be detected in a single multiplex assay, because of crossbinding problems and the existing pseudogene. We established 2 multiplex allele-specific primer extension assays as a 15-plex and a 7-plex, respectively. The combination of various primer extension reactions into multiplex assays is another important critical aspect, and support by suitable software (MassArray design software) is highly recommended to allow the optimization of several different parameters, i.e., GC content, molecular mass range, annealing temperatures, dimers, and hairpin loop formation. The analytical sensitivity (i.e., minimal recommended input amount of DNA for a given method) of our MALDI-TOF MS assay was a less critical issue, since in contrast to forensic applications, microgram amounts of DNA are usually available from patient Clinical Chemistry 54:10 (2008) 1643

Fig. 2. Example genotype cluster plots for TPMT variants 460G>A and 719A>G. Automatically determined relative allele signal intensities for each variant were visualized with a scatter plot. This visualization of peak intensities showed 3 clearly distinguished genotype clusters for each variant.

blood samples. Although we did not determine explicitly the minimal amount of DNA target, 10 ng genomic DNA was required to make genotyping calls with acceptable confidence. It should be mentioned that the extremely high sensitivity of MALDI-TOF MS technology can lead to problems by detecting signals in negative controls due to contamination during sample preparation. Therefore

the use of automated handling systems (e.g., robots) is recommended to minimize the risk of contamination by amplicons or genomic DNA from neighboring samples. Furthermore, to ensure the highest reliability for diagnostic purposes, the samples should be measured at least in duplicate, and appropriate controls for cross-contamination (e.g., water controls) should be included.

Table 3. Frequency distribution of TPMT alleles in different ethnic groups tested by the MALDI-TOF MS assay. Allele frequency, % (95% CI) Ghanaian

Korean

White

116

118

1214a

TPMT*2

0

0

0.2 (0.10–0.54)

TPMT*3A

0

0

4.5 (3.30–4.90)

n

TPMT*3C

6.47 (3.70–11.0)

TPMT*6 TPMT*8

a

0 3.45 (1.50–6.70)

2.54 (0.94–5.50) 1.27 (0.26–3.70) 0

0.4 (0.12–0.59) 0 0

TPMT*9

0

0

0.08 (0.00–0.30)

TPMT*16

0

0

0.04 (0.00–0.23)

TPMT*17

0

0

0.04 (0.00–0.23)

TPMT*18

0

0

0.04 (0.00–0.23)

Other alleles

0

0

Data previously reported by Schaeffeler et al. (16 ).


0


Fig. 3. (A), Examples of MALDI-TOF MS spectra of carriers of TPMT*6 and TPMT*8. Mass is given as m/z. (B), Respective genotype cluster plots for TPMT*6 and TPMT*8.

The reproducibility of the MALDI-TOF MS assay was tested by multiple analytical runs of our validation samples, which yielded perfect overall reproducibility. We evaluated the accuracy of our novel TPMT MALDI-TOF MS assay by comparing TPMT genotype

calls determined by other genotyping methods with the actual genotypes of MALDI-TOF MS. It was suggested for validation of the accuracy of a novel genetic test that an independent target gene PCR amplification strategy should be used to amplify the region of genomic DNA Clinical Chemistry 54:10 (2008) 1645

by a second method (34 ). In line with this recommendation, blinded determination of TPMT variants with MALDI-TOF MS in all 586 cases previously genotyped by denaturing HPLC and/or direct sequencing revealed 100% accuracy (Table 2). Altogether, MALDI-TOF MS– based SNP typing using both established assays has been proven to be a specific and reliable tool for TPMT genotyping. With regard to cost efficiency, the inability of technologies such as TaqMan to multiplex makes them less applicable for comprehensive genotyping of several genetic variants at one time. This disadvantage also includes costs per genotype. Generally, iPLEX assay chemistry with 20- to 40-plex capabilities is calculated to be approximately 4 times cheaper per genotype than the average cost of a genotype achieved by TaqMan technology (http://www.sequenom.com/Seq_genotyping. html) (21 ). Because for many polymorphisms substantial differences in allele frequencies have been observed among different ethnic groups, we determined the frequency of all relevant TPMT variants in a Ghanaian and a Korean population by using our assay. In accordance with previous findings in African and Asian individuals (35, 36 ), TPMT*3C was found to be the most frequent deficient allele (6.5% and 2.5%, Table 3). Allele frequencies of the *6 allele in the Korean population and the *8 allele in Ghanaians were 1.3% and 3.45%, respectively, whereas no carriers were found in our large-scale population of 1214 German whites. Our data are in line with the first description of TPMT*6 in a child of Korean ethnicity (8 ) and TPMT*8 originally identified in an African-American individual (10 ), both with an intermediate activity. Thus both variants seem to be specific for these ethnic populations. Interestingly, the *6 allele was previously screened in 600 subjects of multiracial Asian origin (Chinese, Malaysian, and Indians), but only one Malay sample was a carrier of *6 (37 ), underpinning the assumption that ethnic diversity exists between various Asian populations (38 ). Occurrence of TPMT*6 and *8 extends our knowledge on TPMT genetic variation in the Asian and African population and consequently is of clinical relevance. Several publications have shown that a high prediction (e.g., 95% to 98%) of the correct TPMT

phenotype can be achieved by the use of screening for the most frequent occurring variants in a certain population (16, 39 ). Our present data, however, clearly indicate that genotyping strategy can be recommended only to displace measurement of TPMT activity if a complete genetic analysis of all currently known functional relevant TPMT alleles in a certain ethnic population is performed to prevent misclassification. Thus, our TPMT MALDI-TOF MS assay offers the possibility to test all functional relevant variants in clinical routine with the highest prediction of the correct phenotype. Moreover, since the value of TPMT phenotyping is limited in patients who received blood transfusions (40 ), TPMT genotyping by MALDI-TOF MS is beneficial to individualize thiopurine therapy in a clinical setting.

Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors’ Disclosures of Potential Conflicts of Interest: Upon submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest: Employment or Leadership: None declared. Consultant or Advisory Role: None declared. Stock Ownership: None declared. Honoraria: None declared. Research Funding: The work was supported by the Robert-Bosch Foundation, Stuttgart, Germany. Expert Testimony: None declared. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. Acknowledgments: We gratefully acknowledge the excellent technical assistance of M Elbl and A Zwicker.

References 1. Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature (Lond) 2007;447:661–78. 2. Eichelbaum M, Ingelman-Sundberg M, Evans WE. Pharmacogenomics and individualized drug therapy. Annu Rev Med 2006;57:119 –37. 3. Weinshilboum RM, Wang L. Pharmacogenetics


and pharmacogenomics: development, science, and translation. Annu Rev Genomics Hum Genet 2006;7:223– 45. 4. Krynetski E, Evans WE. Drug methylation in cancer therapy: lessons from the TPMT polymorphism. Oncogene 2003;22:7403–13. 5. Teml A, Schaeffeler E, Herrlinger KR, Klotz U, Schwab M. Thiopurine treatment in inflammatory

bowel disease: clinical pharmacology and implication of pharmacogenetically guided dosing. Clin Pharmacokinet 2007;46:187–208. 6. Krynetski EY, Schuetz JD, Galpin AJ, Pui CH, Relling MV, Evans WE. A single point mutation leading to loss of catalytic activity in human thiopurine S-methyltransferase. Proc Natl Acad Sci U S A 1995;92:949 –53.


7. Tai HL, Krynetski EY, Yates CR, Loennechen T, Fessing MY, Krynetskaia NF, Evans WE. Thiopurine S-methyltransferase deficiency: two nucleotide transitions define the most prevalent mutant allele associated with loss of catalytic activity in Caucasians. Am J Hum Genet 1996;58:694 –702. 8. Otterness D, Szumlanski C, Lennard L, Klemetsdal B, Aarbakke J, Park-Hah JO, et al. Human thiopurine methyltransferase pharmacogenetics: gene sequence polymorphisms. Clin Pharmacol Ther 1997;62:60 –73. 9. Spire-Vayron de la Moureyre C, Debuysere H, Sabbagh N, Marez D, Vinner E, Chevalier ED, et al. Detection of known and new mutations in the thiopurine S-methyltransferase gene by singlestrand conformation polymorphism analysis. Hum Mutat 1998;12:177– 85. 10. Hon YY, Fessing MY, Pui CH, Relling MV, Krynetski EY, Evans WE. Polymorphism of the thiopurine S-methyltransferase gene in AfricanAmericans. Hum Mol Genet 1999;8:371– 6. 11. Colombel JF, Ferrari N, Debuysere H, Marteau P, Gendre JP, Bonaz B, et al. Genotypic analysis of thiopurine S-methyltransferase in patients with Crohn’s disease and severe myelosuppression during azathioprine therapy. Gastroenterology 2000;118:1025–30. 12. Hamdan-Khalil R, Allorge D, Lo-Guidice JM, Cauffiez C, Chevalier D, Spire C, et al. In vitro characterization of four novel non-functional variants of the thiopurine S-methyltransferase. Biochem Biophys Res Commun 2003;309:1005–10. 13. Lindqvist M, Skoglund K, Karlgren A, Soderkvist P, Peterson C, Kidhall I, Almer S. Explaining TPMT genotype/phenotype discrepancy by haplotyping of TPMT*3A and identification of a novel sequence variant, TPMT*23. Pharmacogenet Genomics 2007;17:891–5. 14. Lindqvist M, Haglund S, Almer S, Peterson C, Taipalensu J, Hertervig E, et al. Identification of two novel sequence variants affecting thiopurine methyltransferase enzyme activity. Pharmacogenetics 2004;14:261–5. 15. Schaeffeler E, Stanulla M, Greil J, Schrappe M, Eichelbaum M, Zanger UM, Schwab M. A novel TPMT missense mutation associated with TPMT deficiency in a 5-year-old boy with ALL. Leukemia 2003;17:1422– 4. 16. Schaeffeler E, Fischer C, Brockmeier D, Wernet D, Moerike K, Eichelbaum M, et al. Comprehensive analysis of thiopurine S-methyltransferase phenotype-genotype correlation in a large population of German-Caucasians and identification of novel TPMT variants. Pharmacogenetics 2004;14:

407–17. 17. Schaeffeler E, Eichelbaum M, Reinisch W, Zanger UM, Schwab M. Three novel thiopurine S-methyltransferase allelic variants (TPMT*20, *21, *22): association with decreased enzyme function. Hum Mutat 2006;27:976. 18. Gut IG. DNA analysis by MALDI-TOF mass spectrometry. Hum Mutat 2004;23:437– 41. 19. Tost J, Gut IG. Genotyping single nucleotide polymorphisms by MALDI mass spectrometry in clinical applications. Clin Biochem 2005;38:335–50. 20. Ragoussis J, Elvidge GP, Kaur K, Colella S. Matrixassisted laser desorption/ionisation, time-of-flight mass spectrometry in genomics research. PLoS Genet 2006;2:e100. 21. Tindall EA, Speight G, Petersen DC, Padilla EJ, Hayes VM. Novel Plexor SNP genotyping technology: comparisons with TaqMan and homogenous MassEXTEND MALDI-TOF mass spectrometry. Hum Mutat 2007;28:922–7. 22. Schaeffeler E, Lang T, Zanger UM, Eichelbaum M, Schwab M. High-throughput genotyping of thiopurine S-methyltransferase by denaturing HPLC. Clin Chem 2001;47:548 –55. 23. Schaeffeler E, Eichelbaum M, Brinkmann U, Penger A, sante-Poku S, Zanger UM, Schwab M. Frequency of C3435T polymorphism of MDR1 gene in African people. Lancet 2001;358: 383– 4. 24. den Dunnen JT, Antonarakis SE. Mutation nomenclature extensions and suggestions to describe complex mutations: a discussion. Hum Mutat 2000;15:7–12. 25. Ranade K, Chang MS, Ting CT, Pei D, Hsiao CF, Olivier M, et al. High-throughput genotyping with single nucleotide polymorphisms. Genome Res 2001;11:1262– 8. 26. Hamdan-Khalil R, Gala JL, Allorge D, Lo-Guidice JM, Horsmans Y, Houdret N, Broly F. Identification and functional analysis of two rare allelic variants of the thiopurine S-methyltransferase gene, TPMT*16 and TPMT*19. Biochem Pharmacol 2005;69:525–9. 27. Salavaggione OE, Wang L, Wiepert M, Yee VC, Weinshilboum RM. Thiopurine S-methyltransferase pharmacogenetics: variant allele functional and comparative genomics. Pharmacogenet Genomics 2005;15:801–15. 28. Wang L, Weinshilboum R. Thiopurine S-methyltransferase pharmacogenetics: insights, challenges and future directions. Oncogene 2006;25: 1629 –38. 29. Wang L, Nguyen TV, McLaughlin RW, Sikkink LA, Ramirez-Alvarado M, Weinshilboum RM. Human

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

thiopurine S-methyltransferase pharmacogenetics: variant allozyme misfolding and aggresome formation. Proc Natl Acad Sci U S A 2005;102: 9394 –9. Cheok MH, Lugthart S, Evans WE. Pharmacogenomics of acute leukemia. Annu Rev Pharmacol Toxicol 2006;46:317–53. Kaskas BA, Louis E, Hindorf U, Schaeffeler E, Deflandre J, Graepler F, et al. Safe treatment of thiopurine S-methyltransferase deficient Crohn’s disease patients with azathioprine. Gut 2003;52: 140 –2. Meggitt SJ, Gray JC, Reynolds NJ. Azathioprine dosed by thiopurine methyltransferase activity for moderate-to-severe atopic eczema: a doubleblind, randomised controlled trial. Lancet 2006; 367:839 – 46. Blievernicht JK, Schaeffeler E, Klein K, Eichelbaum M, Schwab M, Zanger UM. MALDI-TOF mass spectrometry for multiplex genotyping of CYP2B6 single-nucleotide polymorphisms. Clin Chem 2007;53:24 –33. Isler JA, Vesterqvist OE, Burczynski ME. Analytical validation of genotyping assays in the biomarker laboratory. Pharmacogenomics 2007;8:353– 68. Collie-Duguid ES, Pritchard SC, Powrie RH, Sludden J, Collier DA, Li T, McLeod HL. The frequency and distribution of thiopurine methyltransferase alleles in Caucasian and Asian populations. Pharmacogenetics 1999;9:37– 42. McLeod HL, Pritchard SC, Githang’a J, Indalo A, Ameyaw MM, Powrie RH, et al. Ethnic differences in thiopurine methyltransferase pharmacogenetics: evidence for allele specificity in Caucasian and Kenyan individuals. Pharmacogenetics 1999; 9:773– 6. Kham SK, Tan PL, Tay AH, Heng CK, Yeoh AE, Quah TC. Thiopurine methyltransferase polymorphisms in a multiracial Asian population and children with acute lymphoblastic leukemia. J Pediatr Hematol Oncol 2002;24:353–9. Nelson SM. The archaeology of Korea. Cambridge: Cambridge University Press; 1993. 307 p. Yates CR, Krynetski EY, Loennechen T, Fessing MY, Tai HL, Pui CH, et al. Molecular diagnosis of thiopurine S-methyltransferase deficiency: genetic basis for azathioprine and mercaptopurine intolerance. Ann Intern Med 1997;126:608 –14. Schwab M, Schaeffeler E, Marx C, Zanger U, Aulitzky W, Eichelbaum M. Shortcoming in the diagnosis of TPMT deficiency in a patient with Crohn’s disease using phenotyping only. Gastroenterology 2001;121:498 –9.

Clinical Chemistry 54:10 (2008) 1647