This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Reiter, J. L. Articles by Wong, D. Search for Related Content PubMed Articles by Reiter, J. L. Articles by Wong, D.

Correspondence re: S. Shintani et al., Intragenic Mutation Analysis of the Human Epidermal Growth Factor Receptor (EGFR) Gene in Malignant Human Oral Keratinocytes. Cancer Res., 59: 4142–4147, 1999.

Tumor Biology Program, Mayo Clinic, Rochester, Minnesota 55905

In their report (1) , Shintani et al. describe two consistent alterations: (a) a tumor-specific BsrI polymorphism; and (b) a novel truncated EGFR¹ mRNA. We propose alternative explanations for the data presented in this report.

First, we have also detected the BsrI polymorphism (A/T) at position 2073 in EGFR. We have sequenced five clones isolated from three normal sources. In one case, we cloned and sequenced PCR products from normal genomic DNA and detected an A at this position (two of two clones). However, we have detected a T at this position in PAC clone DJ1091E12 and in two independent clones isolated from a placental cDNA library (Clontech), suggesting that the proposed tumor-specific allele is also present in normal DNA. Because the presence of an A or T at this position does not change the encoded amino acid, the contribution of the proposed silent mutation to tumorigenesis is unlikely. An alternative explanation for these data is that the allelic frequency of this polymorphism simply favors a T at position 2073. The frequency of this polymorphism could have easily been addressed through the analysis of a representational number of normal control DNA samples.

Second, we have analyzed the sequence of the novel truncated EGFR mRNA using the reported deletion breakpoints (1072–3263 or 1075–3266). A deletion of 2.2 kb from a normal 5.8-kb EGFR mRNA would be expected to give rise to a transcript of 3.5 kb rather than the reported 1.5–1.7-kb transcript. Although the origin of this novel truncated EGFR mRNA is unknown, it is unlikely that alternative splicing is involved because the deletion breakpoints do not occur at exon-intron splice junctions. In addition, this transcript is not the result of a tumor-specific mutation because the authors also detect this reverse transcription-PCR product in normal cells.

The calculated molecular weight of the truncated protein arising from the 3.5 kb transcript, assuming cleavage of the NH₂-terminal signal peptide, is 50,450; however, this putative protein product also would contain at least six potential N-glycosylation sites, which would contribute another 15,000 to the apparent molecular weight of this protein by SDS-PAGE analysis. The significance of the M_r 53,000 band detected in keratinocyte cell lysates is therefore difficult to assess without a direct comparison to the translation product of the novel truncated 3.5-kb EGFR transcript. In addition, it is difficult to explain the relative abundance of the M_r 53,000 protein compared to the M_r 170,000 EGFR when the truncated transcript is only detectable after Southern blot analysis of reverse transcription-PCR products. One would predict that the smaller PCR product (i.e., the 1.5-kb versus the 3.7-kb product) would be preferentially amplified, even if it were present at a much lower copy number initially. An alternative explanation for the M_r 53,000 band is that it results from nonspecific binding to either the primary or secondary antibody.

Finally, we would like to clarify that the truncated EGFR proteins encoded by both the A431 2.8-kb transcript and the avian 2.6-kb mRNA contain only the extracellular domain and lack the intracellular domain. Because the A431 and avian truncated receptors are capable of binding to ligand, the authors propose that their newly discovered truncated EGFR may also be capable of binding to ligand; however, this would seem unlikely because the entire subdomain III, which is known to provide crucial contact sites for ligand interaction, is deleted from this putative protein.

FOOTNOTES

Correspondence re: S. Shintani et al., Intragenic Mutation Analysis of the Human Epidermal Growth Factor Receptor (EGFR) Gene in Malignant Human Oral Keratinocytes. Cancer Res., 59: 4142–4147, 1999.

¹ The abbreviation used is: EGFR, epidermal growth factor receptor.

Received 9/ 9/99. Accepted 11/29/99.

REFERENCES

1. Shintani S., Matsuo K., Crohnin C. C., McBride J., Tsuji T., Donoff R. B., Posner M., Todd R., Wong D. T. W. Intragenic mutation analysis of the human epidermal growth factor receptor (EGFR) gene in malignant human oral keratinocytes. Cancer Res., 59: 4142-4147, 1999.[Abstract/Free Full Text]

Reply

Harvard School of Dental Medicine, Laboratory of Molecular Pathology, 188 Longwood Avenue, Boston, Massachusetts 02115

The letter by Reiter et al. (1) suggests alternative explanations for the data presented in our article published in Cancer Research genotyping intragenic mutations in human EGFR¹ in oral cancer keratinocytes (2) . We offer the following response.

In our analysis, we found a T in both normal oral keratinocyte cultures (OKB2 and OKF4). In tumor oral keratinocytes (we have sequenced three lines), position 2073 is T in both alleles for one line and heterozygous for the other two. We are pleased that Reiter et al. also detected the 2073 BsrI polymorphism in EGFR. They object to what they believe is our assertion that this polymorphism is cancer specific because they have sequenced five clones from three normal sources, normal genomic DNA, PAC clone DJ1091E12, and a placental cDNA library. At position 2073, they found an A in one normal source and a T in the other two. At no instance in our study (2) did we suggest this is a tumor-specific polymorphism. At most, we suggest it is tumor associated. Their proposal that in vivo analysis of a larger number of normal oral epithelial tissues will resolve this issue already appears in the text and is currently underway. These studies will resolve questions regarding culture artifact and tissue specificity that data from neither laboratory address. In vitro culture conditions may select keratinocyte subpopulations harboring a 2073/EGFR genotype that may different from in vivo representation. In addition, none the data of Reiter et al. (1) address whether a 2073/EGFR genotype may be a tissue-specific phenomena. Tissue-specific EGFR polymorphisms have been characterized previously (3) .

Reiter et al. (1) also find our observation regarding a truncated form of EGFR to be inconsistent, not tumor specific, and possibly functionally irrelevant. They contend that the 2.2-kb deletion from the normal 5.8-kb transcript should yield a 3.5-kb mRNA rather than the reported 1.5–1.7-kb transcript. Although the argument is valid, it should be realized that our analysis is based on PCR products using a 3.5-kb PCR template amplified using the 1A/1B primers (see Ref. 2, page 4143, first paragraph). We did not make inference that this truncated EGFR is tumor associated because it is clearly present in the two normal keratinocyte cultures examined (see Ref. 2, Fig. 3F). We have ruled out the detection of the M_r 53,000 bands as nonspecific binding of primary or secondary antibodies. This band was only detectable using the Santa Cruz sc-03 preparation, and it was not detectable using Upstate Biotechnology monoclonal antibody 05-104 (see Ref. 2, Fig. 2B). The detection of the M_r 53,000 band by the sc-03 antibody in the oral keratinocytes (OKB2, OKF4, SCC 15, SCC 25, SCC 66) but not in the epidermal carcinoma cell line A431 further demonstrates the specificity of detection and may suggest the truncated protein to be oral keratinocyte associated. Finally, our predicted truncation affects predominantly subdomain IV, the transmembrane, juxtamembrane, and the protein kinase domains. Subdomain III is mostly intact and may still be ligand binding.

Received 11/10/99. Accepted 11/29/99.

REFERENCES

1. Reiter, J. L., Eley, G., James, C. D., and Maihle, N. J. Correspondence re: S. Shintani et al. Intragenic mutation analysis of the human epidermal growth factor receptor (EGFR) gene in malignant human oral keratinocytes. Cancer Res., 59: 4142–4147, 1999. Cancer Res., 60: 766, 2000.
2. Shintani S., Matsuo K., Crohnin C. C., McBride J., Tsuji T., Donoff R. B., Posner M., Todd R., Wong D. T. W. Intragenic mutation analysis of the human EGFR gene in malignant human oral keratinocytes. Cancer Res., 59: 4142-4147, 1999.
3. Carter T. H., Kung H. J. Tissue-specific transformation by oncogenic mutants of epidermal growth factor receptor. Crit. Rev. Oncog., 5: 389-428, 1994.[Medline]

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Reiter, J. L. Articles by Wong, D. Search for Related Content PubMed Articles by Reiter, J. L. Articles by Wong, D.