Gene expression changes in oral tissues following black raspberry exposure: Interim Affymetrix array analysis of a Phase 1 clinical trial.

A129

INTRODUCTION: Chemoprevention is "the prevention, inhibition, or reversal of carcinogenesis by administration of one or more chemical entities, either as individual drugs or as naturally occurring constituents of the diet." (GD Stoner, MA Morse, GJ Kelloff. EHP 105 Suppl 4:945,1997). The aim of this Phase 1 Clinical Trial was to evaluate molecular changes in oral cavity tissues after oral exposure to lyophilized black raspberries (LBR). Biopsy-confirmed oral squamous cell carcinoma (OSCC) patients were administered three LBR troches four times each day (4.3g cumulative dose) between pre-surgical enrollment and their normally scheduled surgery. Tissue biopsies were obtained from tumor (T1) and distant normal tissues (N1) at enrollment and during surgical resection (T2,N2), an exposure range of 7-34 days. An interim cohort of 10 patients was characterized by Affymetrix HG-U133 Plus 2.0 Arrays ("Affymetrix Arrays";1,300,000 unique oligonucleotide features, >47,000 transcripts, ~39,000 genes). METHODS: Four Affymetrix Arrays (N1,T1,N2,T2) were analyzed for each patient. Baseline pre-surgical tumor (T1) and distant normal (N1) tissues were collected at enrollment. After LBR troche administration, post-surgical tumor (T2) and distant normal (N2) tissues were obtained. Affymetrix Arrays were processed by The Ohio State University Comprehensive Cancer Center Microarray Shared Resource (OSUCCC MASR). Affymetrix CEL files were processed by DNA-Chip Analyzer (dChip) Software and analyses performed using custom statistical routines. Patient covariates included treatment duration, lesion site, tumor staging, and age. RESULTS: As expected, significant differences were readily detected between normal and tumor tissues, both before (T1/N1) and after (T2/N2) LBR exposure. Significant differences were also detected between normal tissues (N2/N1) and tumors (T2/T1) before and after LBR exposure. Importantly, significant differences were revealed when duration of LBR exposure was examined as a covariate among the 10 patient cohort. The data from this interim Phase 1 analysis provide substantial insight into possible biomarkers of oral cancer, and specifically of LBR-responsive genes in the normal and tumor tissues of OSCC patients. Furthermore, additional validation of these possible biomarkers may reveal detailed mechanistic, functional, prognostic, or diagnostic roles for these statistically significant modulated transcripts. SUPPORT: NIH/NCI R21DE016361, ACS RSGT-06-126-01-CNE