New genomic structure for prostate cancer specific gene PCA3 within BMCC1: implications for prostate cancer detection and progression

Abstract

Background: The prostate cancer antigen 3 (PCA3/DD3) gene is a highly specific biomarker upregulated in prostate cancer (PCa). In order to understand the importance of PCA3 in PCa we investigated the organization and evolution of the PCA3 gene locus.

Methods/principal findings: We have employed cDNA synthesis, RTPCR and DNA sequencing to identify 4 new transcription start sites, 4 polyadenylation sites and 2 new differentially spliced exons in an extended form of PCA3. Primers designed from these novel PCA3 exons greatly improve RT-PCR based discrimination between PCa, PCa metastases and BPH specimens. Comparative genomic analyses demonstrated that PCA3 has only recently evolved in an anti-sense orientation within a second gene, BMCC1/PRUNE2. BMCC1 has been shown previously to interact with RhoA and RhoC, determinants of cellular transformation and metastasis, respectively. Using RT-PCR we demonstrated that the longer BMCC1-1 isoform - like PCA3 - is upregulated in PCa tissues and metastases and in PCa cell lines. Furthermore PCA3 and BMCC1-1 levels are responsive to dihydrotestosterone treatment.

Conclusions/significance: Upregulation of two new PCA3 isoforms in PCa tissues improves discrimination between PCa and BPH. The functional relevance of this specificity is now of particular interest given PCA3's overlapping association with a second gene BMCC1, a regulator of Rho signalling. Upregulation of PCA3 and BMCC1 in PCa has potential for improved diagnosis.

Figure 1. Complexity of PCA3 transcripts.

(A) Partial PCA3 gene structure as originally reported by Bussemakers et al. with 4 exons (open boxes ∼ not to scale) with alternate splicing of exon 2 and three alternate transcription termination sites in exon 4. 5′ RACE experiments (Supplementary Fig. S1) identified 4 novel PCA3 transcription initiation sites (isoforms 1–4 marked by vertical arrows pointing down with nucleotide sequence below located 1150 bp, 699 bp, 640 bp and 136 bp upstream of the original initiation site (renamed here isoform 5). 3′ RACE identified four novel polyadenylation sites (7 in total*) located in exon 4. The size of exon 1 is expanded from the original 120 bp to 1270 bp. Isoform 4 (PCA3-4) is the most highly expressed of the four novel isoforms. Four overlapping ORFs initiate from a single ‘ATG’ start site (vertical arrow pointing up) within PCA3-4 and terminate within one of the alternatively spliced exons (2a or 2b or 2c) or within exon 3. (B) RT-PCR amplification of BPH, PCa and PCa metastasis samples using a forward primer from within the novel PCA3-4 transcription start site and a reverse primer from novel exon 2a. (C) Complete structure of the PCA3 gene. Shading identifies the newly identified regions of the PCA3 gene which has 6 exons with alternate splicing of exon 2a (93 bp) and exon 2b (93 bp) and exon 2c (original exon 2, 165 bp). and (D) RT-PCR amplification of PCA3 using the same forward primer and a reverse primer from novel exon 2b and (E) RT-PCR amplification of PCA3 using a forward primer from PCA3-5F (within the original transcription start site) and a reverse primer spanning exons 1 and 3 .

Figure 2. PCA3 is embedded within BMCC1.

(A) The PCA3 gene (above) is embedded within the intron 6 of BMCC1 (PRUNE2) isoform 1 (BMCC1-1). Gray boxes denote exons of BMCC1 and black boxes denote exons of PCA3. The two genes are in the opposite orientation (NCBI Build 36). Three other isoforms of BMCC1 have also been described, none of which include the complete set of exons present in BMCC1-1. (B) VISTA plot of BMCC1 gene. Peak heights indicate degree of conservation between species of exons (blue) and evolutionary conserved regions (ECR within introns - pink) compared with human. Note that gene orientation is different in Fig 2A, upper and lower panels. We estimated the mutation rate at the DNA sequence level by comparing human and chimpanzee sequences and using a divergence time of 6 million years (Myr). The mutation rate in the PCA3 gene was estimated to be 1.26×10⁻⁹ per nucleotide per year, higher than that (1.00×10⁻⁹ per nucleotide per year) in the non-PCA3 portion of the BMCC1 gene, suggesting the PCA3 region might have a moderately higher mutation rate. Three highly conserved ECRs within BMCC1 are arrowed (A, B & C); Arrow ‘A’ (ECR_in1, 277 bp) is an extremely conserved non-coding sequence with 91% similarity between human and opossum that is positioned within intron 1 of PCA3; Arrow ‘B’ (ECR_ex4c, 403 bp) an ECR positioned within PCA3 exon 4 which is conserved in all mammals and appears to have been the focal point for the linear evolution of the PCA3 gene (see Fig. 3); Arrow ‘C’ (361 bp) is the most highly conserved ECR (a conserved non-coding sequence with 99% similarity between human and opossum) within BMCC1 (intron 6) immediately downstream of PCA3.

Figure 3. Linear evolving structure (3′→5′) of the PCA3 gene.

(A) Vista plot displaying conserved structures of the PCA3 gene. Only primates appear to have a complete PCA3 gene. The two evolutionary conserved regions shared by BMCC1 and PCA3 (see Fig. 2B) are arrowed (arrow A ∼ ECR_ex4c and arrow B ∼ ECR_in1). Both ECRs appear conserved in mammals (identity >90%). ECR_in1 is also present at this site in chicken and lizard but not in fish, frog, or invertebrates. (B) ECR_ex4c appears focal to the linear evolving structure of PCA3. (C) Summary of mammalian PCA3 exons sharing high conservation compared to human. The gene structure annotation was based on Bussemakers et al. . The level of conservation of PCA3 exons during the course of mammalian evolution increases 3′→5′ based on the presence of meaningful ECRs in each exon. The exact sequence identity for the complete exon between human and other species is shown in Supplementary Table S2. It is important to note that no ECR from PCA3 exons was found in any non-mammalian species in this analysis.

Figure 4. PCA3 and BMCC1-1 expression patterns in prostate cancer.

cDNA was prepared from patient tissue specimens including eight BPH, PCa or PCa metastases, respectively, for use in the following PCR reactions. (A). RT-PCR carried out on BPH, PCa and metastases (MET) with different sets of primers for PCA3 and BMCC1 Upper row; BMCC1-2 RT-PCR using a forward primer for BMCC1 exon 5 (BMCC1-Ex5F) and a reverse primer specific for the extended form of exon 6 (BMCC1-Ex7R) unique to BMCC1-2 2nd row; BMCC1-1 specific RT-PCR using primers for BMCC1 exon 6 (BMCC1-Ex6F) and exon 7 (BMCC1-Ex7R). 3rd row; BMCC1-BCH region specific RT-PCR using primers BCHF and BCHR. 4th row; PCA3 was amplified (35 cycles) with primers specific to PCA3 isoform 5 exon 1 (PCA3-5F and the Ex1/3R primer). 5^th row; β₂microglobulin control PCR (B). RT-PCR comparative expression analysis of PCA3, BMCC1-1 and the BMCC1-BCH region for ALVA41, DU145, LNCaP and PC3 prostate cancer lines, RPWE1 a normal prostate cell line, JHP a control lymphoblastoid cell line, RM654 a lymphoblastoid cell line from a patient with AOA2 and MCF7 a breast cancer cell line. RT-PCR was carried out with the primer sets specific for PCA3 isoform 5 (PCA3-5F) and Exon 1/3 (Ex1/3R) and for BMCC1-1 and the BMCC1-BCH region as indicated above. (C) Semi-quantitative PCR analysis of PCA3 and BMCC1-1 expression in the LNCaP cell line in response to dihydrotestosterone. Results were normalised relative to levels of β₂microglobulin. Four cell cultures were starved of serum for 2 days prior to incubation with dihydrotestosterone (µg/ml). Results were normalised and expressed as mean fold increase relative to the level of expression before treatment. Error bars are standard deviations and p values were determined from comparison with untreated samples using a Student's t-test.