Changes in gene expression induced by chemoradiation in advanced cervical carcinoma: a microarray study of RTOG C-0128.

PURPOSE: To evaluate gene expression patterns in patients with advanced cervix cancer before and during chemoradiation in a multi-institutional cooperative group setting. METHODS: RTOG C0128 was designed as a Phase II trial of radiation therapy with concomitant chemotherapy and Celecoxib at 400 mg twice daily for one year. Tumor samples were obtained for microarray gene expression analysis before treatment and at the time of the first implant (paired sample). RNA was extracted, linearly amplified, and purity was assessed by gel electrophoresis. Each sample was hybridized against a universal RNA mixture on a customized spotted array consisting of >10,000 genes. Gene expression pre-treatment was compared with clinical characteristics. Changes in gene expression following radiation were assessed within the paired samples (same patient) and then compared across all paired samples. Data were normalized using the AROMA software, and clustering analysis was performed using Ward's method in Spotfire. Differences in paired samples were calculated with Significance Analysis of Microarrays (SAM). RESULTS: From August 2001 to March 2004, 84 patients were accrued to the trial. Tissue was obtained prior to initiation of therapy from 34 patients (40%). FIGO stages of the patients providing tissue were IB (23%), II (57%), and IIIA-IVA (20%). RNA quality was sufficient in 22 pre-treatment and 14 post-treatment samples. Among pre-treatment samples, no significant differences in gene expression were observed by FIGO stage, age, or race. However, between comparison of histologic subtypes (adenocarcinoma, n=5; squamous cell carcinoma, n=17) demonstrated 45 genes differentially expressed with a false discovery rate of 0.018. Cluster analysis segregated unpaired samples into 2 groups: 18/22 comprising pre-treatment samples and 10/14 in group 2 representing post-treatment samples. In all 13 paired samples, gene expression after chemoradiation was significantly upregulated in 91 genes and downregulated in 251 genes (false discovery rate of 0.0018). Genes significantly upregulated included bax, cdk inhibitor 1, MMP2, and adhesion molecules PECAM1, VCAM1, and ICAM2. Genes significantly downregulated included topoisomerase II alpha, myc, H2AX, MSH2, RAD51, RAD53, PCNA, and cell cycle-regulating molecules chk1, CDK2, cyclinB1, cyclin D3, cdc2, and cdc25. CONCLUSIONS: Microarray analysis was successfully performed in a multi-institutional cooperative group trial. Gene expression significantly correlated with histology, but not stage, age or race. Cluster analysis identified two groups of gene expression profiles correlating with pre or post-treatment acquisition of tissue. Notably, paired samples showed significant changes in gene expression following chemoradiation, including several downregulated radiation response genes. Further analysis comparing gene expression to clinical outcomes, acute and late toxicities awaits maturation of clinical data. Hopefully, this data will lead to the development of molecularly based therapies.

Feasibility of RNA collection for micro-array gene expression analysis in the treatment of cervical carcinoma: a scientific correlate of RTOG C-0128.

PURPOSE: To determine the feasibility of RNA collection in a multi-institutional cooperative group setting to be utilized for micro-array gene expression analysis, and to describe the methodology. METHODS: RTOG C0128, a phase I-II, protocol was designed to look at the safety and efficacy of external beam radiation therapy to 45 Gy with concomitant 5-FU and cisplatin chemotherapy, brachytherapy to deliver 85 Gy to point A, and Celecoxib at 400 mg twice daily for 1 year. Patients had the option of participating in a tissue collection portion of the protocol to be utilized for micro-array gene expression analysis before treatment and at the time of the first implant. RNA quality was determined by two parameters: the absorbance ratio at 260 nm/280 nm, and by the ratio of the integrated peak of 28S RNA to 18S RNA after gel electrophoresis. RESULTS: From August 2001 to March 2004, 84 patients were accrued to the trial, and tissue was obtained prior to initiation of therapy on 34 patients (40%). FIGO stages for the patients who provided tissue were IB (23%), II (57%), and IIIA-IVA (20%). Additionally, biopsies were obtained at the time of the first implant from 22 of the accrued patients making paired samples available on 26% for RNA extraction and micro-array gene expression analysis. The mean +/- SEM amount of tissue obtained pretreatment was 97 +/- 13 mg compared with 51 +/- 8 mg for tissue obtained at the time of the first implant (P = 0.009). The mean total RNA extracted from the samples prior to treatment was 119 +/- 19 microg versus 35 +/- 6 microg at the time of the first procedure (P = 0.001). The RNA quality was assessed via the absorbance ratio at 260 nm divided by 280 nm. The mean values pretreatment and at first implant were 1.87 +/- 0.07 versus 1.66 +/- 0.11, respectively (P = 0.002); however, the integrated peak of 28S RNA to 18S RNA after gel electrophoresis was not significantly different (P = 0.26). CONCLUSIONS: RNA extraction for gene expression analysis can be successfully performed in the multi-institutional cooperative group setting. Fresh tissue samples were obtained on 40% of accrued patients prior to treatment. The amount of biopsy material and the quantity of RNA extracted were greater prior to treatment compared with the first implant. The quality of RNA was superior prior to treatment as measured by the ratio of absorbance at 260/280 nm. These results indicate that gene expression analysis is feasible in the cooperative group setting utilizing amplification techniques for the RNA. Hopefully, this will allow for improvement in prognosis, therapeutic development, and correlation with acute and late toxicities in patients with cancer.