Elevated transcription of the gene QSOX1 encoding quiescin Q6 sulfhydryl oxidase 1 in breast cancer

Abstract

The q arm of chromosome 1 is frequently amplified at the gene level in breast cancer. Since the significance of this is unclear we investigated whether 1q genes are overexpressed in this disease. The cDNA levels of 1q-located genes were analysed in a search for overexpressed genes. 26 genes mapping to the 1q arm show highly significant (P≤0.01) overexpression of transcripts in breast cancer compared to normal breast tissue. Amongst those showing the highest levels of overexpression in both expressed sequence tag (EST) and serial analysis of gene expression (SAGE) databases was enzyme quiescin Q6 sulfhydryl oxidase 1 (QSOX1). We investigated QSOX1 cDNA derived from T47D breast carcinoma cells by RT-PCR and 3'-RACE PCR and identified a novel extended form of QSOX1 transcript, containing a long 3'UTR, nearly double the size of the previously reported QSOX1 cDNA, and confirmed its 3' end nucleotide sequence using RACE-PCR. We also used quantitative real-time PCR to analyse a panel of cDNAs derived from 50 clinically-graded normal and malignant breast tissue samples for the expression of QSOX1 mRNAs. QSOX1 transcription was elevated in an increasing proportion in the grade 2 and grade 3 tumours (graded according to the Nottingham prognostic index), with 10 of the 15 grade 3 tumours (67%) examined exceeding the normal range. There was a significant correlation between relative transcript level and clinical grade (P≤0.01) for all qPCR primer sets tested. QSOX1 mRNA levels, based on SAGE expression data, did not correlate with either Estrogen Receptor (ER) or Epidermal Growth Factor Receptor 2 (ErbB-2 or HER2/neu) expression. Our data indicate that QSOX1 is a potential new prognostic marker which may prove of use in the staging of breast tumours and the stratification of breast cancer patients.

Figure 1. 3′ Extension of QSOX1 cDNA.

Panel A: Overlapping database sequences are aligned with the fragment of human genomic sequence from chromosome 1 (AL390718). The 3′ end of QSOX1 exon 12 (NM_002826) is shown as an open box. Light grey shaded boxes show selected overlapping ESTs (accession numbers are indicated). Dark grey boxes denote the overlapping cDNAs identified in GenBank (accession numbers are indicated). Black arrows indicate the approximate positions and orientation of the PCR primers used to check the expression of the extended QSOX1 transcript. Primer names are shown next to their positions. Dashed lines show the individual overlapping PCR products obtained, which continuously cover the QSOX1 3′ extension. Blue arrows indicate the approximate positions of the sequence-specific RACE-PCR primers used (see Supplementary Table S2 for all the primers’ sequences). The polyadenylation site (AATAAA) was found approximately 30 bases upstream of the poly(A) tail (positions 150623 and 150651 respectively). All positions are numbered relative to the human genomic sequence (AL390718). Red arrow sets indicate the approximate positions of real time qPCR primers sets 2 and 3. The blue double-headed arrow indicates the experimentally confirmed 3′ extension of QSOX1 cDNA. Panel B: Alignment of the experimentally identified 3′ end of the QSOX1 cDNA with the human genomic sequence (AL390718). The polyadenylation signal (AATAAA) is underlined in both sequences.

Figure 2. RT-PCR and RACE-PCR amplification of the extended form of the QSOX1 cDNA.

Panel A shows amplified cDNA using primers 45-F and 46-R (expected length 550 bp). Panel B shows amplified cDNA using primers 43-F and 44-R (expected length 735 bp). Other PCR amplifications were using primers: 63-F with 64-R (panel C, expected size 836 bp), 47-F with 48-R (panel D, expected size 922 bp), 49-F with 50-R (panel E, 825 bp), 53-F with 36-R (panel F, 883 bp), 23-F with 54-R (panel G, 620 bp). Panel H shows amplified cDNA of the soluble version of QSOX1 (QSOX1b) using primers 17-F and 18-R (expected length 814 bp). Panel I shows the size of the cDNA fragment amplified with the specific sense primer (106-Int-F) and antisense 3′ RACE adapter primer (Adap-1-R). The specificity of amplification (the identity of all amplified PCR products) was further confirmed by DNA sequencing for all amplified products.

Figure 3. Quantitative Real-Time PCR of QSOX1 transcripts in a panel of RNA samples from normal and breast cancer tissue samples, clinically graded using the Nottingham prognostic index.

Data for three different TaqMan probe sets are shown (panels A-C). In all panels the vertical axis is relevant abundance normalised to 18s rRNA levels. A: Primer Set 1 (based on the QSOX1 CDS). B: Primer Set 2 (based on the QSOX1 3′UTR). C: Primer Set 3 (based on the QSOX1 3′UTR). The horizontal line in each dot plots shows mean expression data for each probe set/condition D: Scatter plot for Set 2 vs. Set 1 data across all preparations. E: Scatter plot for Set 3 vs. Set 2 data across all preparations. F: Scatter plot for Set 1 vs. Set 3 data across all preparations. For all scatter plots, all axes show relevant abundance values for the relevant probe set in logarithmic scale. Best fit linear regression is shown as a dotted diagonal line.

Figure 4. Differential overexpression of the CDS and 3′UTR of QSOX1.

Mean expression data (as in Figure 2A–C) for the three different TaqMan probe sets normalised per mean expression data for probe Set 1 are shown. Sets 2 and 3 (3′UTR region of QSOX1) show a lower degree of expression in Grade 1 tumours and a higher degree of expression in Grade 3 tumours relative to Set 1 (QSOX1 CDS).

Figure 5. Alternative splicing of the QSOX1 mRNA.

QSOX1 exons are shown. Alternatively spliced variants QSOX1a and QSOX1b are based on NM_002826 and NM_001004128 respectively. Other alternative splicing variants are based on the EST mapping data (see Table 2) and are named QSOX1c – QSOX1i for consistency with the previously used nomenclature. The identified splicing sites are shown with bold connecting lines. Exon 12b denotes the sequence fragment missing in the shorter alternatively spliced version of QSOX1b.

Figure 6. QSOX1 internal splicing variants and their translation.

QSOX1 protein domains Trx1, Trx2, HRR and ARV/ALR are shown as light grey shaded boxes. Asterisks and vertical yellow lines indicate the position of cysteine-pairs in the product of the QSOX1 transcript. SP (black boxes) denotes a signal peptide, TM indicates a transmembrane domain and C indicates a C terminus. Square dotted horizontal lines indicate the joining of truncated domains. Hatched boxes show amino acid sequences produced where such joining results in reading frame-shifts.