Pai et al. (2004) described a novel CYP2D7 brain-specific splice variant that contained an open reading frame due to a T-deletion in exon 1 (138delT) and encoded a protein with distinct kinetic characteristics toward codeine. The recent comment by Hoskins et al. in Drug Metabolism and Disposition (Hoskins et al., 2005) pointed out a number of concerns regarding the findings published in this paper, one being the lack of specificity of PCR primers. Indeed, nonspecific amplification of related genes within the locus, including gene duplications (if present), may have contributed to some controversial results presented by Pai et al. (2004) and may explain, for example, the number of homozygous 138delT subjects and apparent lack of any heterozygotes, which is inconsistent with the Hardy-Weinberg equilibrium (Pai et al., 2004; Hoskins et al., 2005).
Considering these and additional concerns of our own, we could not endorse the conclusions by Pai et al. (2004) that the novel CYP2D7 splice variant is brain-specific (not expressed in liver and kidney), constitutes a major splice variant, is expressed only in subjects carrying 138delT (linkage of 138delT and transcript formation), and, finally, is relatively frequent (Pai et al., 2004). To resolve these issues, we have thoroughly investigated splicing of CYP2D7 and CYP2D6 in brain and liver. We also performed CYP2D7-specific genotyping for 138delT on 285 subjects, including North American Caucasians, African Americans, American Asians, and 7 ethnic Indians. In our recent publication (Gaedigk et al., 2005), we reported numerous CYP2D6 and CYP2D7 splice variants in brain and liver including the one described by Pai et al. (2004), as well as several novel transcripts. Furthermore, we demonstrated that the alternate partial intron 6-retaining variant is exclusively CYP2D7-derived, is indeed expressed in liver, constitutes only a minor variant in liver and brain, and is formed independently of 138delT. In addition, we unequivocally demonstrated that the PCR product generated with the primers used by Pai et al. (2004) is nonspecific and gives rise to “heterozygous” band patterns as predicted by Hoskins et al. (2005) when genomic DNA is used as template. In contrast, our CYP2D7-specific genotyping approach correctly identified 138delT in positive controls, but did not reveal the deletion in any of the 285 subjects genotyped.
In the course of our analysis, we also uncovered an additional inconsistency between the CYP2D7 cDNA sequence reported by Pai et al. (2004) (AY220845) and available sequences deposited in GenBank. An alignment between AY220845, M33387, X58468, X07619, and our cloned CYP2D7 cDNA sequences revealed that AY220845 contained three SNPs that were not observed in any other sequence. Two are nonsynonymous cSNPs (326G>A, Ser70Asn; 1226C>T, Ala370Val), whereas the third is located within the partially retained intron 6. As shown in Fig. 1, the latter SNP converts a stop codon to Ser (1130G>C, corresponds to g.14408 in M33387). In essence, translation of the novel CYP2D7 transcript reported by Pai et al. would require 138delT and 1130G>C to give rise to a full-length open reading frame. The existence of 1130G>C was not mentioned in their original report, and even though it is shown in the figure provided by Pai and Ravindranath (2005), its potential significance remained unrecognized. Using a CYP2D7-specific genotyping method (Gaedigk et al., 2005), we could not find any evidence for g.14408G>C in 118 individuals, including 7 Indians, suggesting that it is either rare or an artifact. However, g.14408G>C is of little importance to the correctly spliced CYP2D7 transcript (as it contains no intron 6 sequences), which makes up the majority of transcript in liver and brain (Gaedigk et al., 2005).
We do not agree with Hoskins et al. (2005), however, that CYP2D6- and CYP2D7-specific genotyping is hampered by the lack of a validated Human Genome Project-derived reference sequence as evidenced by many published genotyping procedures (including our own) and recent developments in the arena of commercial kits and genotyping services (M. S. Phillips, personal communication; Gaedigk et al., manuscript in preparation). There are sufficient sequence data deposited in GenBank that allow the design of CYP2D6- and CYP2D7-specific primers. We do agree, however, that genotype information (including testing for the presence of gene duplications) of the subjects used by Pai et al. (2004) would have been beneficial.
Without doubt, the potential contribution of a novel CYP2D7 protein to understand the variability in the response to analgesic compounds has stimulated considerable interest (Pai et al., 2004). However, whether a “novel” CYP2D7 protein indeed plays an appreciable role in codeine biotransformation in a subset of individuals remains to be confirmed with CYP2D7-specific procedures before any definitive conclusions can be drawn. Overall, we could not agree more with Hoskins et al. (2005) and emphasize that extreme care is a prerequisite when investigating gene loci consisting of highly homologous genes and pseudogenes.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.105.007799.
-
ABBREVIATIONS: PCR, polymerase chain reaction; SNP, single nucleotide polymorphism.
- Received October 10, 2005.
- Accepted December 14, 2005.
- The American Society for Pharmacology and Experimental Therapeutics