The putative oncogene, CRNDE, is a negative prognostic factor in ovarian cancer patients

Abstract

The CRNDE gene seems to play an oncogenic role in cancers, though its exact function remains unknown. Here, we tried to assess its usefulness as a molecular prognostic marker in ovarian cancer. Based on results of our microarray studies, CRNDE transcripts were further analyzed by Real-Time qPCR-based profiling of their expression. The qPCR study was conducted with the use of personally designed TaqMan assays on 135 frozen tissue sections of ovarian carcinomas from patients treated with platinum compounds and either cyclophosphamide (PC, N = 32) or taxanes (TP, N = 103). Elevated levels of two different CRNDE transcripts were a negative prognostic factor; they increased the risk of death and recurrence in the group of patients treated with TP, but not PC (DNA-damaging agents only). Higher associations were found for overexpression of the short CRNDE splice variant (FJ466686): HR 6.072, 95% CI 1.814-20.32, p = 0.003 (the risk of death); HR 15.53, 95% CI 3.812-63.28, p < 0.001 (the risk of recurrence). Additionally, accumulation of the TP53 protein correlated with decreased expression of both CRNDE transcripts in tumor cells. Our results depict CRNDE as a potential marker of poor prognosis in women with ovarian carcinomas, and suggest that its significance depends on the therapeutic regimen used.

Keywords: CRNDE; TP53; gene expression; ovarian cancer; prognostic factor.

Figure 1. CRNDE – selected prognostic results of the multivariate statistical analysis of gene expression

CRNDE expression is shown as a continuous variable. Black lines on bars represent standard deviations of Real-Time qPCR measurements for each tumor. The linear regression lines (green and dotted) are shown to visualize trends of expression.

Figure 2. Other statistical results

A–B. A correlation between the mean CRNDE expression measured with gene expression microarrays for two CRNDE-specific probe sets, 238021_s_at and 238022_at, and the results obtained in Real-Time qPCR studies with the use of our personally designed TaqMan assays for two different splice variants of CRNDE. Fitted regression lines are marked blue, whereas shaded regions represent the 95% confidence interval. C–D. The association between the accumulation of TP53 and the decreased expression of CRNDE. If the expression decreases, the calculated/expected ratio of the Mann-Whitney U test is lower than 1 and vice versa. Black lines on bars represent standard deviations of Real-Time qPCR measurements for each tumor.

Figure 3. Different CRNDE transcripts seem to occur in a tissue-dependent manner

A. A graphical alignment of two CRNDE transcripts investigated in this study (FJ466685, FJ466686) to four reference RNA sequences (NR_034105, NR_034106, NR_110453, NR_110454) available in GenBank. Regions, where the PCR primers hybridize, are marked with red arrows. B. Three various tissues were tested in this PCR experiment: normal endometrium (NE), ovarian cancer (OC) and HeLa cells (He). The bands characteristic for certain tissues only and also inconsistent with in silico predictions were marked with ellipses (see Figure C for details on the primer sets used and the expected length of PCR products). 50 bp DNA Ladder (New England Biolabs, Ipswich, MA, USA) was utilized as a size standard. The annealing temperature was set to 55°C, while the elongation time lasted for 2 minutes to allow amplification of products up to 2 kb. ND stands for a no-DNA sample. NDs for the primer sets no. 1 and 2 were not shown due to the lack of empty lanes in the gel (they were both negative).

Figure 4. A diagram of two CRNDE transcripts, FJ466685 and FJ466686, investigated herein

An alternatively spliced region within exon 5 was depicted with dashed lines. The presence of this region leads to a disruption of 84aa ORF encoding the CRNDEP peptide. Red arrows represent primer sequences.