Abstract P1-07-08: Effect of sample preservation method and transportation duration on tumor gene expression profiling in breast cancer.

Abstract

Abstract Background: Translational research studies in the context of international clinical trials may incur prolonged transportation time of tumor specimens. The effect of prolonged transportation time and sample preservation method on RNA quality and on reliability of derived gene expression profiling (GEP) is unknown. Methods: Tumor material from surgical specimens of patients with primary breast cancer (BC) was obtained with cold ischemia time (CIT) recorded for each patient. Each sample was divided into 8 aliquots 4 that were placed in RNAlater and 4 that were frozen. To simulate prolonged transportation duration, samples were kept 1 week in a bucket of dry ice replenished daily, with temperature monitoring. The bucket was stored at room temperature. RNA extraction and GEP using Affymetrix HG-U133 Plus2 were performed at baseline and at day 2, 4 and 7. Linear mixed effect models were used to explore effects of transportation time and preservation method on RNA quality (RNA Integrity Number-RIN). Pre-defined single genes (n = 12), gene modules (n = 25) and molecular subtypes (PAM50) were evaluated. P < 0.01 was considered for significance. Results: Samples were collected from 13 patients with an average CIT of 37 minutes (range: 20–85). During the simulation period, the mean temperature at which samples were stored was −77.8°C (SD+/−1.6°C). RIN values decreased significantly over time (1-day change −0.08, CI = −0.12 to −0.04, p < 0.001) and by RNAlater (−0.29, CI = −0.07 to −0.51, p < .01), but remained always above 6.5. RNA degradation evaluated with actin 3/5 and gapdh 3/5 ratios was significantly higher over time for samples preserved in RNAlater compared with frozen (difference in changes 0.68, CI=0.49 to 0.87, p < 0.001 and 0.33,0.25 to 0.4, p < 0.001, respectively), and for 5 RNAlater (38%) samples these ratios exceeded the Affymetrix thresholds (3 and 1.25, respectively) at least at 1 time-point. Comparing tissue preservation methods, only the expression of PTEN gene and CASP3 module changed significantly more over time in RNAlater compared with frozen samples (interaction p < 0.01). Genes and gene modules that were significantly differently expressed over time, independent of preservation method, were AURKA, PTEN, CASP3 and WOUND (p < 0.01, range of 1-day changes in log2 expression:-0.05 to 0.10), while the expression of ESR1 and ERBB2 was unaffected. Those that were significantly influenced by sample preservation method, adjusted for time, were genes ESR1, PLAU, VEGF, PIK3CA, PTEN and gene modules GENE21 and GENE70 was (P < 0.01, range: −0.61 to 0.16). Using PAM50, 8 (61%) samples were classified at least once as a different subtype over time compared to baseline, and 4 (30%) samples fell in a different Genomic Grade Index (GGI) class risk (high/low) at least once over time. Conclusions: In this study, frozen samples were significantly less degraded over time compared with samples stored in RNAlater. Preservation method had a stronger influence on genes and gene modules values compared with prolonged transport duration. In international clinical trials we recommend freezing samples to assure more stable gene expression analysis and limit transport duration to less than 3 days. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P1-07-08.