A 584 bp deletion in CTRB2 inhibits chymotrypsin B2 activity and secretion and confers risk of pancreatic cancer

Abstract

Genome-wide association studies (GWASs) have discovered 20 risk loci in the human genome where germline variants associate with risk of pancreatic ductal adenocarcinoma (PDAC) in populations of European ancestry. Here, we fine-mapped one such locus on chr16q23.1 (rs72802365, p = 2.51 × 10^-17, OR = 1.36, 95% CI = 1.31-1.40) and identified colocalization (PP = 0.87) with aberrant exon 5-7 CTRB2 splicing in pancreatic tissues (p_GTEx = 1.40 × 10^-69, β_GTEx = 1.99; p_LTG = 1.02 × 10^-30, β_LTG = 1.99). Imputation of a 584 bp structural variant overlapping exon 6 of CTRB2 into the GWAS datasets resulted in a highly significant association with pancreatic cancer risk (p = 2.83 × 10^-16, OR = 1.36, 95% CI = 1.31-1.42), indicating that it may underlie this signal. Exon skipping attributable to the deletion (risk) allele introduces a premature stop codon in exon 7 of CTRB2, yielding a truncated chymotrypsinogen B2 protein that lacks chymotrypsin activity, is poorly secreted, and accumulates intracellularly in the endoplasmic reticulum (ER). We propose that intracellular accumulation of a nonfunctional chymotrypsinogen B2 protein leads to ER stress and pancreatic inflammation, which may explain the increased pancreatic cancer risk in carriers of CTRB2 exon 6 deletion alleles.

Keywords: genetics; pancreatic cancer; pancreatic enzymes; polymorphic variation; splicing QTL.

Figure 1

A Genomic map of the chr16q23.1 pancreatic risk locus (A) Association results from a meta-analysis of PanScan I, II, and III, , , and PanC4 with variants colored indicating linkage disequilibrium (LD) to tag SNP rs72802365 (red, r² ≥ 0.8; yellow, 0.8 > r² ≥ 0.6; green, 0.6 > r² ≥ 0.4; light blue, 0.4 > r² ≥ 0.2; dark blue, r² < 0.2 in 1000G EUR). (B) NCBI RefSeq genes within the chr16q23.1 risk locus as visualized with the UCSC genome browser. (C) The fifteen candidate functional variants at the 16q23.1 risk locus (LR > 1:100 in black, LR > 1:1,000 in green, tag SNP rs7280265 in purple) distributed over a 52 kb region ranging from chr16: 75,234,273 to 75,286,484 (NCBI GRCh37/Hg19). Results after conditioning the association analysis on rs13337397, rs72802365, and the 584 bp CTRB2 deletion variant are shown in Figure S1.

Figure 2

CTRB2 splicing QTL at the chr16q23.1 pancreatic-cancer-risk locus (A) Boxplots showing normalized CTRB2 exon junction reads (Leafcutter²⁰) in GTEx normal pancreas samples for CTRB2 exon 5–6 (Ai), exon 6–7 (Aii), and exon 5–7 (Aiii) based on rs72802365 genotype (risk allele = C as indicated in red text). The number of samples per genotype are G/G, n = 148; G/C, n = 24; C/C, n = 2. (B) Genomic map of the 584 bp CTRB2 exon 6 deletion (orange) and 16.6 kb CTRB2-CTRB1 inversion (teal) variants as well as the deletion tag SNP, rs72802342, and GWAS tag SNP, rs72802365. The location of the initial most significant genotyped SNP, rs13337397, is approximately 37 kb downstream of the start of CTRB1 (chr16: 75,295,639) and is not shown here. Splicing QTLs for the LTG dataset are shown in Figure S3.

Figure 3

Effect of the exon 6 deletion on chymotrypsin B2 activity and localization (A) (Ai) Schematic figure of the CTRB2 gene showing the 584 bp deletion that overlaps all of exon 6 as well as parts of introns 5 and 6. (Aii) Full-length chymotrypsin B2 (CTRB2) protein contains a cleaved signal peptide and three peptide chains (A, B, and C) that are held together by disulfide bonds. When the exon 6 deletion is present, a premature stop codon is introduced (at aa 166 of the truncated protein; at aa 212 as compared to the full-length CTRB2 protein) resulting in the loss of chain C (amino acids 167–263) including one of the three amino acids forming the catalytic triad (denoted by ^∗, Ser213 is lost). (B) Chymotrypsin activity (μU/mg) calculated from triplicate experiments (^∗ denotes p < 0.05) with (dark gray) and without (light gray; activity due to other proteases besides chymotrypsin) chymotrypsin activator (trypsin), error bars represent standard error of the mean (SEM). (C) Amount of CTRB2 protein in cell lysates and media assessed by immunoprecipitation-immunoblot analysis in HEK293T cells transfected with plasmids expressing full-length and truncated (lacking exon 6) CTRB2 (FLAG-tagged). (D) The immunoblot image in (C) was taken with a ChemiDoc Touch Imaging System and bands were quantified with the ImageLab software (see material and methods). The amount of secreted proteins is summarized from triplicate experiments (^∗∗ denotes p < 0.001, error bars represent SEM). The full immunoblot is shown in Figure S8. The effects of the exon 6 deletion on chymotrypsin activity and secretion via nontagged expression clones are shown in Figure S9.

Figure 4

Effect of the truncated CTRB2 protein on ER stress (A and B) Human normal pancreatic-derived HPDE cells transiently expressing GFP-tagged (green) full-length CTRB2 (Ai and Bi) and truncated (with exon 6 deleted) CTRB2 proteins (Aii and Bii) were immunostained with antibodies for the (A) endoplasmic reticulum marker calnexin (red) or (B) cis-Golgi marker GM130 (red). Representative single xy images from a Z stack processed with deconvolution software (Slidebook) are shown, taken with a spinning disk confocal microscope with a 63× oil objective and a sCMOS camera. (C) Transiently transfected HPDE cells were immunostained for BiP (HSPA5), a marker of ER stress (red), and imaged by epifluorescence microscopy with a 63× oil objective. Representative images (Ci and Cii) from ten fields are shown. DAPI-stained nuclei shown in blue. Scale bar represents 10 μM. (D) Quantification of elevated BiP (HSPA5) protein levels in GFP-expressing HPDE cells imaged as described in (C) (>40 GFP-expressing cells were analyzed in each of three independent experiments) (^∗∗∗ denotes p < 0.005, error bars represent SEM). (E) Fold change increase in mRNA expression for ER stress pathway genes in cells expressing the truncated as compared to the full-length CTRB2 proteins. Expression was assessed by qRT-PCR analysis following transient overexpression of constructs containing full-length versus truncated (exon 6 deleted) CTRB2 in HEK293T (black), MIA PaCa-2 (gray), and PANC-1 (white) cells. The data are calculated from four replicate experiments (^∗ denotes p < 0.05, ^∗∗ denotes p < 0.01, error bars represent SEM). Note that the fold change increase in protein expression for ER stress genes is shown in Figure S13.