Absence of common activating mutations of the epidermal growth factor receptor gene in thyroid cancers from American and Japanese patients

Dear Editor, Mutations of EGFR are found in 10–30% of non-small cell lung carcinoma (NSCLC), with a greater prevalence in never-smokers and Asian patients. Short in-frame deletions of exon 19 (such as del E746-T750), and the G2573T point mutation resulting in a leucine to arginine substitution at codon 858 (L858R), account for ~90% of the EGFR mutations seen in NSCLC 1, 2. In glioblastomas, genomic deletion of exons 2–7 (EGFRvIII) and missense mutations in the extracellular domain of EGFR are found in 20–30% and ~14% of cases, respectively 3, 4. Despite the different distribution of EGFR mutations in NSCLC and glioblastomas, both classes of mutants are oncogenic and predict sensitivity to EGFR kinase inhibitors. Previous studies have failed to detect mutations in the kinase domain of EGFR in thyroid cancer. Until recently, the few thyroid cancer cases with EGFR mutations reported in the literature involved residues that do not correspond to known oncogenic mutants, and which were not verified to be gain-of-function substitutions 5–7. Thus, a recent report in this journal describing EGFR mutations in 7 of 23 (30%) Japanese PTC patients was unprecedented, and a surprise to many in the field 8. The genetic alterations included the most common mutations found in NSCLC: i.e. L858R and a short in-frame deletion of exon 19 (del E746-T750). This prompted us to revisit this question, primarily because of its potential therapeutic repercussions. As EGFR mutations in NSCLC are more prevalent in patients from Asia, we thought it conceivable that prior negative studies were in part a reflection of the patient populations that were screened, and hence we included patients from Japan and the US in our survey. Moreover, as EGFR, ERBB2, RAS and BRAF mutations are mutually exclusive in NSCLC 9, we enriched our sample set for cases that were wild-type for RET, RAS and BRAF (i.e. triple negative), which also rarely overlap in thyroid cancer (Table 1). We utilized mass spectrometry to screen for the most common EGFR kinase domain mutations (Table 2). EGFR exon 19 deletions were screened for by PCR and size separation by agarose gel electrophoresis. No alterations in the kinase domain of EGFR were found in any of the 41 thyroid cancers from American patients: 29 triple negative (4 PTC, 19 PDTC, 6 ATC) and 12 PTC negative for BRAF/RAS (RET unknown). Because 3 of the 7 (43%) thyroid cancers with EGFR mutations identified in the Masago study also harbored BRAF mutations, we expanded our cohort by including cases harboring other genetic alterations activating MAPK: 10 PTC with BRAF-T1799A, and 10 RET/PTC positive cases (5 PTC and 5 PDTC). Once again, no alterations were found. Because EGFR mutations are more common in lung cancer patients of Asian ethnicity, we also genotyped 60 PTC from Japanese patients. Ten of these were triple negative, 30 were negative for BRAF (RET and RAS not screened) and 20 were positive for BRAF mutation, and none showed an alteration in the kinase domain of EGFR. Altogether, none of the 121 thyroid cancer samples had a kinase domain mutation of EGFR. Table 1 Histotype and genotype of thyroid tumors analyzed for EGFR mutations Table 2 EGFR kinase domain mutations screened by mass spectrometry None of the 30 thyroid cancer cell lines or 33 thyroid tumors (cases with cDNA available) from US (17 PDTC and 16 ATC) expressed the EGFRvIII mutant transcript, corresponding to an in-frame deletion of exons 2–7. We also sequenced exons 3, 7, 8 and 15 for mutations of the extracellular domain of EGFR in the 60 Japanese PTC, and 19 ATC and 17 PDTC from US patients. A non-synonymous alteration in the extracellular domain of EGFR in 1 Japanese PTC was found: p.G322C. We were unable to determine whether this mutation was somatic or not as normal thyroid tissue or blood sample from the patient was not available. Although we do not have functional data, and this mutation has not been previously described in the COSMIC database 10, EGFR-G322 is highly conserved from fish to humans. Thus, our data show that EGFR mutations are rare in thyroid cancer. EGFR copy number changes, as determined by quantitative PCR, were reported in a significant fraction of follicular and anaplastic thyroid cancers (32 and 46%, respectively) 5. Despite the frequent overlap of EGFR mutation and amplification seen in NSCLC, amplification alone does not predict outcome or therapeutic response 11. Accordingly, a phase II trial of gefitinib showed no reports of partial or complete response in 27 patients with PTC, FTC, ATC or MTC 12. This was an unselected patient population, and does not allow conclusions on the role of genetics in responsiveness to EGFR inhibitors. In this regard, it should be noted that thyroid cancers have a high prevalence of BRAF and RAS mutations, which predict resistance to EGFR inhibitors 13. In conclusion, our results show that EGFR mutations are rare in thyroid cancers, regardless of ethnicity, and thus do not confirm the recently published report.