A gastrin transcript expressed in gastrointestinal cancer cells contains an internal ribosome entry site

Abstract

As the hormone gastrin promotes gastrointestinal (GI) cancer progression by triggering survival pathways, regulation of gastrin expression at the translational level was explored. Sequence within the 5' untranslated region of a gastrin transcript expressed in GI cancer cells was investigated, then cloned into a bicistronic vector upstream of firefly luciferase and transfected into a series of GI cancer cell lines. Firefly luciferase activity was measured relative to that of a cap-dependent Renilla luciferase. A gastrin transcript that was different from that described in Ensembl was expressed in GI cancer cells. Its transcription appears to be initiated within the region designated as the gene's first intron. In GI cancer cells transfected with the bicistronic construct, firefly luciferase activity increased 8-15-fold compared with the control vector, and there was a further induction of the signal (up to 25-fold) following exposure of the cells to genotoxic stress or hypoxia, suggesting that the sequence acts as an internal ribosome entry site. These data suggest that the gastrin transcript within GI cancer cells contains an internal ribosome entry site that may allow continued expression of gastrin peptides when normal translational mechanisms are inactive, such as in hypoxia, thereby promoting cancer cell survival.

Figure 1

Alignment of Ensembl gastrin genomic sequence (Genomic), Ensembl gastrin transcript (Ensembl), GenBank sequence X00183 (X00183) and three representative RLM-RACE clones (1, 2 and 7). Clones 1 and 7 occurred once each, and clone 2 occurred six times. The beginning of exon 1, intron 1 and exon 2 are indicated by ^* and the ATG start codon is highlighted in bold. The central 2860 bp of intron 1 have been omitted for clarity and their position is indicated by - - - - - - - -; only sequence of the first 13-bp and final 171-bp of intron 1 are shown. The RLM-RACE clones have high homology with the region defined by Ensembl as intron 1 and with X00183; ^ indicates positions where one of the clones or X00183 did not match the genomic sequence. There were only 10 nucleotides, which matched the Ensembl transcript, indicated by ∼ below the sequence. The UTRScan-predicted IRES within X00183 is highlighted by shading.

Figure 2

Reverse transcriptase-PCR to investigate expression of the two gastrin transcripts in a panel of GI cancer cell lines: (1) PAN1, (2) BXPC3 (pancreatic); (3) HCT116, (4) HT29 (5) C170HM2 (colonic); (6) ST16, (7) MGLVA1 (gastric); (8) OE19 and (9) OE21 (oesophageal). cDNA (+) and a negative control in which the RT was omitted (−) were amplified using a primer specific for each transcript ((A) alternative transcript and (B) Ensembl transcript) and a common primer. The expected position of the bands is indicated by an arrow based on the position of the 100-bp markers (M).

Figure 3

Activity of the putative gastrin IRES and Myc IRES in pancreatic (PAN1) and colon (HCT116) cancer cell lines. Basal activity of (A) the gastrin IRES (pRGasF) or (B) Myc IRES (pRMF) compared with empty vector (pRF). (C) Activity of the gastrin IRES in a promotorless plasmid (pBRGas) compared with the empty vector (pBR) or pRGasF. (D) Significant knockdown of both Renilla and firefly luciferase activity in HCT116 cells transfected with pRGasF and a Renilla siRNA, compared with cells transfected with pRGasF and a control siRNA (P<0.005 for both). (E) Reverse transcriptase-PCR of cells transfected with pRGasF using forward and reverse primers located within the Renilla and firefly luciferase sequences. A single band of ∼300 bp was observed (Lane 1). No bands were observed in the RT-negative, cDNA reagent and RNA isolation controls (Lanes 2–4). Lane 5: markers. Statistical significance is indicated by ^*P<0.05 or ^**P<0.01.

Figure 4

Effect of MMC and hypoxia on IRES activity. Treatment with MMC significantly reduced cap-dependent translation of the Renilla luciferase (A) but resulted in an increase in the ratio of IRES-dependent to cap-dependent translation in both HCT116 (B) and PAN1 cells (C). Similarly, 24 h exposure to a hypoxic environment increased IRES-dependent translation (D). Statistical significance is indicated by ^*P<0.05 or ^**P<0.01.