SOX2-silenced squamous cell carcinoma: a highly malignant form of esophageal cancer with SOX2 promoter hypermethylation

This study originally aimed to investigate whether the overexpression of SOX2 is associated with the poor prognosis of patients with squamous cell carcinoma of the esophagus. However, we unexpectedly found that esophageal squamous cell carcinomas completely lacking SOX2 expression showed distinct pathologic features and highly aggressive clinical courses. The study cohort consisted of 113 consecutive patients with esophageal squamous cell carcinoma who underwent surgical resection without neoadjuvant therapy. Immunostaining on tissue microarrays and whole sections revealed that 8/113 (7%) cases were entirely negative for this transcriptional factor. SOX2-negative cancers were histologically less differentiated ( P =0.002) and showed higher pT and pStages ( P =0.003 and 0.007, respectively) than SOX2-positive cases. A remarkable finding was widespread lymphatic infiltration distant from the primary invasive focus, which was observed in 4 SOX2-negative cancers (50%), but none of the SOX2-positive cases. All separate dysplastic lesions observed in SOX2-negative cases were also SOX2-negative. The negative expression of SOX2 appeared to be an independent poor prognostic factor (OR=7.05, 95% CI=1.27–39.0). No mutations were identified in the coding or non-coding regions of SOX2 . Fluorescent in situ hybridization did not show any copy-number variations in this gene. Since the SOX2 promoter contains an extensive CpG island, SOX2-negative cases underwent methylation-specific PCR, which disclosed promoter hypermethylation in all cases. In conclusion, SOX2 -silenced squamous cell carcinomas of the esophagus appear to be a minor, but distinct form of malignancy characterized by extensive lymphatic invasion, a poor prognosis, and potential association with multiple SOX2-negative neoplastic lesions. The hypermethylation of the promoter region is seemingly a critical epigenetic event leading to SOX2 silencing.