| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Division of Developmental Pharmacology and Medical Toxicology, The Children’s Mercy Hospital and Clinics, Kansas City, MO 64108
2 Department of Psychiatry and Medicine, Morehouse School of Medicine, Atlanta, GA 30310
aaddress correspondence to this author at: The Children’s Mercy Hospital, Division of Clinical Pharmacology, 2401 Gillham Rd., Kansas City, MO 64108; fax 816-855-1958, e-mail agaedigk{at}cmh.edu
Accurate prediction of CYP2D6 phenotype from genotype data is important for many clinically relevant drugs. Genotyping strategies targeting allelic variants with diminished or no activity to identify poor metabolizers generally works well in Caucasian (1)(2)(3) and Asian (4)(5) populations, but we have had more limited success (i.e., poor concordance) in African Americans, even after extensive testing (6). For example, genotype analysis predicted poor metabolizer status (by Caucasian-derived antimode definition) in only 4 of 14 individuals, and many individuals predicted to be extensive metabolizers presented as "intermediate" metabolizers (6). A single-nucleotide polymorphism in the CYP2D6 promoter region (-1584G) has been reported to confer higher CYP2D6 activity in vivo than -1584C (7), possibly as a consequence of higher expression of CYP2D6 protein (8). This polymorphism was first associated with CYP2D6*2 alleles, and the CYP nomenclature committee subsequently assigned *2[-1584G] as CYP2D6*2A (and noting that -1584G is probably found on all CYP2D6*2 alleles) and *2[-1584C] as CYP2D6*41, respectively (9). However, -1584G also appears to be linked with the functional CYP2D6*35 allele, which has been found in many duplication-negative "ultrarapid" metabolizers (10). For simplicity, we refer to the *2[-1584G] allele cumulatively as CYP2D6*2 because our genotyping procedure does not differentiate among variants CYP2D6*2A through K. For reference, -1584G corresponds to -1496G in Ref. (7).
The goals of this investigation were (a) to explore whether -1584G is exclusively linked to functional allelic variants and hence would allow rapid "positive" identification of extensive metabolizers and reliably rule out poor metabolizer status, and (b) to reevaluate the genotype-to-phenotype correlation data in our Caucasian and African-American populations.
The use of patients’ DNA samples was approved by the Pediatric Institutional Review Board of Children’s Mercy Hospital and the Review Board of Morehouse School of Medicine. CYP2D6 genotyping was carried out as described previously (3)(6). Briefly, genomic DNA was isolated from whole blood with a QIAamp Blood Kit (Qiagen) and a 5.1-kb CYP2D6-specific fragment amplified by long PCR that served as template for subsequent genotyping assays. Genotyping comprised CYP2D6*2-*12, *14, *15, *17, *18, *29, *40, and *42 as well as *1, *2, and *4 gene duplications. Additional tests were developed for -1584C
G, to distinguish between *2 and *41, and for 31GA (CYP2D6*35).
For the -1584CG assay, we amplified a 327-bp fragment from genomic DNA in the presence of 50 mL/L dimethyl sulfoxide with RedJumpStart (Sigma) with primers 5'-GAATTCAAGACCAGCCTGGACAACTTGGAAGggCC-3', which contains a partial ApaI site (mismatches are shown in bold lower case) and was also used to nest a 2-kb fragment for sequencing (described below), and 5'-GTGGCTCCCCTCCATTGTGC-3'. -1584C PCR products were digested into 292- and 35-bp fragments by ApaI, whereas -1584G products remained uncut. To cut -1584G-derived fragments, we generated a 283-bp amplicon with primers 5'-GCAGCTGCCATACAATCCACCTG-3' and 5'-CAATCCCAGCTAATTTTGTATTTTTTGTAGgGgCC-3'; ApaI cut 1584G-containing products into 248- and 35-bp fragments, whereas -1585C remained uncut.
In the 31GA assay, primers 5'-AGCAGGTTCACTCACAGCAGAGGGCAAtGG-3', which contained a MscI site, and 5'-CCACCAGGAGCAGGAAGATtGCCAC-3' (mismatch, in lower case, destroys interfering MscI site) generated a 224-bp product with use of a CYP2D6-specific long PCR template. CYP2D6*35-derived PCR products carrying 31G were cut into 161-, 33-, and 30-bp fragments; other fragments were cut once into 194 and 30 bp.
Digestion products were separated on 3% Synergel/agarose gels (Diversified Biotech).
For sequence analysis, a 4.6-kb 5'-upstream fragment that included part of exon 1 was generated by long PCR. A nested 2-kb product was amplified from this fragment for direct sequencing on a MegaBACE 500 capillary sequencer, using DYEnamic ET dye terminator chemistry (Amersham Biosciences). Primer 5'-ATTACAGCCTTCTGAGTCACTGAC-3' (GenBank accession no. X90926) was used to generate the 4.6-kb fragment, and 5'CAAACCTGCTTCCCCTTCTCAGCC-3' (GenBank accession no. M33388) served as reverse primer for the 4.6- and 2-kb fragments. One of the forward -1584 genotyping primers (see above) nested the 2-kb amplicon.
Regenotyping of 203 Caucasians indicated that -1584G was associated not only with the CYP2D6*2 allele but also with CYP2D6*35, as suggested previously (10). The linkage of -1584G with 31A, which defines CYP2D6*35, was confirmed by pedigree and sequence analyses, including three individuals with CYP2D6*35/*35 (n = 2) and *1/*35 genotypes, respectively. The segregation of -1584G with 2D6*2 and 2D6*35 was unequivocally demonstrated in three two-generation families: A, B and C (see the figure in the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol49/issue6/). In addition, the pedigree of family A revealed a CYP2D6*35x2 gene duplication. In two unrelated individuals, CYP2D6*2x2 duplications were also associated with -1584G. Although linkage of -1584G to either CYP2D6*2 or 2D6*35 could not be established unequivocally in two additional individuals, a "default" genotype of 2D6*35/*41 was assigned because -1584G was linked in all other instances with 31A (i.e., *35).
Regenotyping of 193 African Americans for -1584CG also identified G exclusively in association with CYP2D6*2 and *35. There were no homozygous -1584G/G individuals among this study cohort because of the dramatically lower frequency of -1584G compared with Caucasians (Table 1
).
|
Regardless of ethnicity, we did not find any genotypes (e.g., CYP2D6*1/*17 or *4/*10) that would have allowed an association of -1584G with any allele other than CYP2D6*2 or *35. Genotyping results for the nonfunctional allelic variants CYP2D6*3, *4, *6, *40, and *42, sequencing data for CYP2D*4 (n = 1 allele), and genotyping data for cloned *40 (n = 2 alleles) and *42 (n = 1 allele) revealed that only -1584C was associated with these alleles. Furthermore, the -1584CG genotyping assay did not produce any amplification product from CYP2D6*5 (using a CYP2D6*5/*5 control DNA) because this region comprises the CYP2D6 gene deletion. DNAs heterozygous for CYP2D6*5 amplify only from their second allele and therefore appeared "homozygous" in all genotyping assays performed.
CYP2D6 phenotyping was conducted in all participants and was expressed as the urinary metabolic ratio of the probe drug dextromethorphan (DM) to its metabolite, dextrorphan (DX). Because the term "extensive metabolizer" covers a wide range of activities, we propose subclassifications based on DM/DX ratios to facilitate more precise characterization and comparison of such individuals across populations. For example, extensive metabolizers (DM/DX <0.3) are subdivided into rapid (DM/DX <0.003), extensive (0.003 
DM/DX < 0.03), and intermediate (0.03 DM/DX < 0.3) metabolizers (Fig. 1
). According to these arbitrary categories, 32% of Caucasians and 14% of African Americans are rapid, 51% and 49% are extensive, and 10% and 30% are intermediate, respectively (3)(6). For Caucasian individuals with at least one -1584G allele, the 95% confidence intervals for each category were 26.4–46.4% (rapid), 46.9–66.9% (extensive), and 1.6–12.1% (intermediate). The positive predictive value of a -1584G-containing genotype was 1.0 in Caucasians and African Americans; the negative predictive value was 0.120 and 0.084, respectively; and the sensitivity of the test was 0.47 and 0.15, respectively. Although there was no significant difference in DM/DX ratios between the Caucasian -1584C/G and G/G individuals (P = 0.5), the -1584C/G African-American cohort had a significantly higher median metabolic ratio compared with the Caucasians (0.0105 vs 0.0042; P = 0.001). Similar results were observed between -1584C/C groups attributable, at least in part, to the presence of reduced-activity alleles (CYP2D6*17 and *29) in African-American but not Caucasian individuals.
|
For 43% of Caucasian and 12% of the African-American individuals, respectively, poor metabolizer status was accurately ruled out by the presence of at least one allele carrying -1584G.
Regenotyping of a Caucasian and an African-American population demonstrated that -1584G is indeed linked to functional CYP2D6*2 allelic variants, as described previously (7), and confirmed the proposed CYP2D6*35 haplotype (10). Phenotype-genotype correlation analysis subsequently showed that poor metabolizer status could effectively be ruled out by positive identification of a single functional -1584G-containing allelic variant in 43% of Caucasians but only 12% of African Americans. The -1584G assay can be performed directly on genomic DNA or on a long CYP2D6-specific PCR product encompassing upstream and coding regions of the CYP2D6 gene. The assay is useful for rapid and cost-effective identification of individuals with functional CYP2D6 activity and confirmation of an extensive metabolizer assignment when incorporated in a more intensive genotyping strategy. Although perfect linkage of -1584G to functional alleles was demonstrated in our population samples, it remains to be validated in populations of other ethnic backgrounds.
The significantly lower frequency of -1584G in African Americans compared with Caucasians further supports previous findings that CYP2D6 genotyping in African Americans presents unique challenges (6). The increased frequency of specific reduced-function and rare nonfunctional alleles (CYP2D6*17, *29, *40, and *42), and the potential existence of additional such alleles because of the residual genotype-phenotype discordance observed by us and others (6)(11)(12), make accurate prediction of phenotype from genotype data difficult in this heterogeneous population. Therefore, exclusion of poor metabolizer status through identification of -1584G carriers offers one solution to this problem.
Overrepresentation of CYP2D6*35 among duplication-negative ultrarapid metabolizers may be attributable to -1584G and increased gene transcription rather than the Val11Met substitution (31GA) (10), which had no effect on activity in vitro (13). Although DM phenotype does not appear to differentiate gene duplication-carrying ultrarapid metabolizers from other groups, a correlation between CYP2D6*35 and rapid metabolism (DM/DX <0.003) was evident (data not shown), further supporting the association of the -1584G polymorphism with more rapid metabolism.
In conclusion, the -1584CG polymorphism rules out poor metabolizer status, identifies additional important ethnic differences, and therefore is a valuable addition to CYP2D6 genotyping strategies applied to Caucasian and African-American populations.
Acknowledgments
We thank Roger Gaedigk, PhD, for support with sequencing; Stephen Simon, PhD, for statistical analyses; and Ivy Hurwitz, PhD, for CYP2D6*4 upstream sequence data. This study was supported by Children’s Mercy Hospital Research Vision Core Laboratory Project Grant 01.4888 (to A.G.).
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. E. Kootstra-Ros, M. J. M. Van Weelden, J. W. J. Hinrichs, P. A. G. M. De Smet, and J. van der Weide Therapeutic drug monitoring of antidepressants and cytochrome p450 genotyping in general practice. J. Clin. Pharmacol., November 1, 2006; 46(11): 1320 - 1327. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Gaedigk, L. D. Bradford, S. W. Alander, and J. S. Leeder CYP2D6*36 GENE ARRANGEMENTS WITHIN THE CYP2D6 LOCUS: ASSOCIATION OF CYP2D6*36 WITH POOR METABOLIZER STATUS Drug Metab. Dispos., April 1, 2006; 34(4): 563 - 569. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
