Identification of differentially expressed genes to predict radioresistant prostate carcinomas

• Published in: Journal of clinical oncology Vol. 37; no. 7_suppl; p. 57
• Main Authors: Nestler, Tim, Wittersheim, Maike, Hellmich, Martin, Pfister, David J. K. P., Odenthal, Margarete, Büttner, Reinhard, Schäfer, Stephan, Heidenreich, Axel
• Format: Journal Article
• Language: English
• Published: 01.03.2019
• ISSN: 0732-183X; 1527-7755
• DOI: 10.1200/JCO.2019.37.7_suppl.57

Abstract only 57 Background: Although relapses after radiotherapy are common in prostate cancer (PCa) patients, there are no clinical models or markers to identify patients at high risk for radioresistance. So far, only in vitro studies and xenograft models have been performed to identify gene expression patterns associated with radioresistance. However, studies which address the protein pattern to predict radioresistance in humans are completely missing. In order to determine potential biomarkers for radioresistance, we compared protein expression profiles of radioresistant PCa patients with PCa of primary prostatectomized patients. Methods: Two study groups consisting of: I) 30 patients who were treated by salvage prostatectomy and II) 94 patients treated by primary prostatectomy were formed. Tissue microarrays were constructed and immunostained for 15 proteins which are suggested to be associated with radioresistance by in vitro findings. Kruskal-Wallis test was used for multiple group comparison and followed by Dunn-Bonferroni-Test to detect intergroup differences. Cohen’s d was used to calculate the intergroup effect size. Results: Most proteins studied did not show any relevant differences between radioresistant PCa and primary PCa, except for two (AR and AKR1C3). On comparing immunostaining patterns between radioresistant PCa and primary PCa separated by Gleason risk groups, we observed only AR (androgen receptor) to be most expressed in radioresistant PCa (89.7%) and, in 87.8% of primary PCa of the high-risk group ( > 7a) (p = 0.851, Cohen’s d = 0.05), while only 67.3% PCa of the low-risk group (≤7a) (p = 0.017, Cohen’s d = 0.55) were positive. Considering the highest Gleason pattern per patient, only AKR1C3 (Aldo-Keto Reductase Family 1 Member C3) was seen to be similarly expressed in radiation-resistant PCa and patients with Gleason patterns 4 and 5 (p = 0.827, Cohen’s d = 0.05 and p = 0.893, Cohen’s d = 0.10) as compared to Gleason pattern 3 (p = 0.20, Cohen’s d = 0.69) in primary PCa. Conclusions: This is the first study evaluating protein expression profiles to predict radioresistance in PCa, where AR and AKR1C3 were identified to be the most promising protein markers.