Disturbed expression of splicing factors in renal cancer affects alternative splicing of apoptosis regulators, oncogenes, and tumor suppressors

Abstract

Background: Clear cell renal cell carcinoma (ccRCC) is the most common type of renal cancer. One of the processes disturbed in this cancer type is alternative splicing, although phenomena underlying these disturbances remain unknown. Alternative splicing consists of selective removal of introns and joining of residual exons of the primary transcript, to produce mRNA molecules of different sequence. Splicing aberrations may lead to tumoral transformation due to synthesis of impaired splice variants with oncogenic potential. In this paper we hypothesized that disturbed alternative splicing in ccRCC may result from improper expression of splicing factors, mediators of splicing reactions.

Methodology/principal findings: Using real-time PCR and Western-blot analysis we analyzed expression of seven splicing factors belonging to SR proteins family (SF2/ASF, SC35, SRp20, SRp75, SRp40, SRp55 and 9G8), and one non-SR factor, hnRNP A1 (heterogeneous nuclear ribonucleoprotein A1) in 38 pairs of tumor-control ccRCC samples. Moreover, we analyzed splicing patterns of five genes involved in carcinogenesis and partially regulated by analyzed splicing factors: RON, CEACAM1, Rac1, Caspase-9, and GLI1.

Conclusions/significance: We found that the mRNA expression of splicing factors was disturbed in tumors when compared to paired controls, similarly as levels of SF2/ASF and hnRNP A1 proteins. The correlation coefficients between expression levels of specific splicing factors were increased in tumor samples. Moreover, alternative splicing of five analyzed genes was also disturbed in ccRCC samples and splicing pattern of two of them, Caspase-9 and CEACAM1 correlated with expression of SF2/ASF in tumors. We conclude that disturbed expression of splicing factors in ccRCC may possibly lead to impaired alternative splicing of genes regulating tumor growth and this way contribute to the process of carcinogenesis.

Figure 1. Expression of splicing factors mRNA in ccRCC.

A. Changes in expression of particular pairs of control and tumor samples. The diagram was performed based on the data from real-time PCR analysis (Figure S2). Colors represent expression ratio between control and tumor samples. Green: downregulation in tumor samples. Red: upregulation in tumor samples. White: no difference in expression levels. Tumor grading (G1, G2, G3) is shown. B. Distribution of changes in mRNA expression of splicing factors. D: group of samples with tumor-specific downregulation of expression; U: group of samples with tumor-specific upregulation of expression; N: group of samples that did not differ in expression between control and tumor samples. The threshold of 30% difference in expression between control and tumor samples was used to classify samples.

Figure 2. Disturbances of expression of splicing factors mRNA.

Expression of each splicing factor is shown in two groups of samples: with tumor-specific downregulation (D) (left) and upregulation (U) (right). The threshold of 30% difference in expression between control and tumor samples was used to classify samples. C: control samples, T: tumor samples. The data are given as mean ± S.E.. Statistical analysis was performed using paired t-test. * p<0.05; ** p<0.01; *** p<0.001.

Figure 3. Changes in ratios between splicing factors in tissue samples.

A. Matrix showing correlation coefficients between mRNA expression of analyzed splicing factors. The plot was generated based on Pearson correlation coefficients between expression values of splicing factors. Patient number 16 was removed from analysis due to deviation from normal distribution. For SC35 (gene: SFRS2) Spearman nonparametric correlation was used as data were not normally distributed in this group. The values of Pearson or Spearman r are given below the dot diagram. Splicing factors' gene names are shown in brackets. P<0.05 was considered statistically significant. B. mRNA expression ratios of splicing factors known to act antagonistically. The data are given as mean ± S.E. (for SF2/ASF: hnRNP A1 and hnRNP A1: SC35) or as median values and 95% CI (for SC35: SRp55 as data were not normally distributed in this group). Statistical analysis was performed using paired t test (for for SF2/ASF: hnRNP A1 and hnRNP A1: SC35) or Wilcoxon paired test (for SC35: SRp55). n = 37 for C, n = 37 for T, ** p<0.01.

Figure 4. Protein expression of splicing factors in twelve representative pairs of control (C) and tumor (T) samples.

Tumor grades of differentiation are shown (G1, G2, G3). Western blots of SF2/ASF (A) and hnRNP A1 (B) were used for semiquantitative analysis of protein bands after normalization to β-actin. Gray bars represent control samples. Black bars represent tumor samples.

Figure 5. Alternative splicing of genes involved in tumor progression.

A. PCR analysis of alternative splicing patterns in twelve pairs of control (C) and tumor (T) tissue samples. 1) RON; 2) CEACAM-1; 3) Rac1; 4) Caspase-9; 5) GLI1. Positions of primers used for PCR are shown relative to exons. Alternatively spliced exons are shaded. Gradings of tumor differentiation are shown (G1, G2, G3). B. Graph showing expression ratios of splice variants as determined by densitometric analysis of electrophoresed PCR products. Note different axis scales. Gray bars represent control samples. Black bars represent tumor samples.

Figure 6. Correlation between protein expression of SF2/ASF and Caspase-9a (A) and CEACAM1-L (B).

Pearson correlation analysis was performed on data from twelve pairs of control and tumor tissue samples. P<0.05 was considered statistically significant.

Figure 7. Analysis of CEACAM1 exon 7 sequence in a search for potential binding motifs of splicing factors.

Prediction of motifs was performed with ESE Finder software using matrices for prediction of sequences required for binding of splicing factors. For the prediction of SF2/ASF binding sites, two matrices were used: “SF2/ASF/IgM-BRCA1“ (white bars) and “SF2/ASF“ (gray bars). These two matrices were derived in different context (different minigenes and size of random sequence libraries in SELEX [27]). Only high-score motifs above thresholds for SF2/ASF (1.956), SF2/ASF/IgM-BRCA1 (1.867), SC35 (2.383), SRP40 (2.670), and SRp55 (2.676) are shown. The nucleotide sequence of CEACAM1 exon 7 is given on x-axis.