The Transcriptional Landscape of Radiation-Treated Human Prostate Cancer: Analysis of a Prospective Tissue Cohort

Purpose

The resistance of prostate cancer to radiation therapy (RT) is a significant clinical issue and still largely unable to be guided by patient-specific molecular characteristics. The present study describes the gene expression changes induced in response to RT in human prostate tissue obtained from a prospective tissue acquisition study designed for radiobiology research.

Methods and Materials

A prospective cohort of 5 men with intermediate-risk and clinically localized tumors were treated with high-dose-rate brachytherapy with 2 × 10-Gy fractions. Image-guided transperineal biopsy specimens were taken immediately before and 14 days after the first high-dose-rate brachytherapy fraction. Using genome-wide 3′ RNA sequencing on total RNA extracted from 10 biopsy specimens, we obtained quantitative expression data for a median of 13,244 genes. We computed the fold-change information for each gene and extracted high-confidence lists of transcripts with either increased or decreased expression (≥1.5-fold) after radiation in ≥4 of the 5 patients. Several gene ontology analyses were then used to identify functionally enriched pathways.

Results

The predominant change in response to RT was elevation of the transcript levels, including that of DNA damage binding protein 2 and p21, and collagens, laminins, and integrins. We observed strong upregulation of the p53 pathway, without observable dysregulation of p53 itself. Interstitial remodeling, extracellular matrix proteins, and focal adhesion pathways were also strongly upregulated, as was inflammation. Functional network analysis showed clustering of the changes inherent in apoptosis and programmed cell death, extracellular matrix organization, and immune regulation.

Conclusions

In the present prospective study of matched clinical tissues, we successfully recognized known radiation-sensitive transcriptional pathways and identified numerous other novel and significantly altered genes with no current association with RT. These data could be informative in the development of future personalized therapeutic agents.

Introduction

Prostate cancer (PCa) will be diagnosed in >1 million men every year globally and is associated with significant mortality. Early detection is common, leading to the frequent use of radical therapy aimed at cure. In addition to radical prostatectomy, radical radiation therapy (RT) is a standard approach for clinically localized PCa (1). Although RT is efficacious in most cases, cancer control can be limited by the development of radioresistance.

Decades of radiobiology research have characterized the key processes related to radiation-induced cell death when studied using controlled in vitro systems. Putative markers that are centrally involved in the radiation response, such as p53, MDM2, and Bax, have since been associated with clinical outcomes in large cohorts of patients who underwent RT (2). However, to date, these markers have had little effect on clinical practice, such as the identification of those requiring RT intensification or those who could possibly avoid RT completely. Previous studies have been limited by an inability to accurately define the effect of RT locally within the prostate using clinical failure endpoints. Similarly, these studies did not have access to repeated tissue samples for accurate assessment of the natural history of the RT response; thus, these analyses have been limited to associations of one or several markers from a priori hypotheses across large groups of patients. Consequently, few studies have reported on predictive biomarkers that are specific to RT.

We hypothesized that gene expression changes after RT might yield valuable insight into the causes of the heterogeneous treatment outcomes observed clinically. In the present study, we aimed to demonstrate the feasibility of identifying and characterizing the transcriptomic responses of human PCa to radiation in vivo. Matched PCa biopsy specimens before and after RT were obtained from a prospective tissue acquisition study designed to address radiobiology questions in PCa. We report the first differential gene expression data from this novel cohort. We found it to be a powerful method of assessing transcriptional changes and revealing novel insights into the response of human tissue to RT.