A Novel PiRNA Enhances CA19-9 Sensitivity for Pancreatic Cancer Identification by Liquid Biopsy

Abstract

Pancreatic cancer is one of the deadliest tumours worldwide, and its poor prognosis is due to an inability to detect the disease at the early stages, thereby creating an urgent need to develop non-invasive biomarkers. P-element-induced wimpy testis (PIWI) proteins work together with piwi-interacting RNAs (piRNAs) to perform epigenetic regulation and as such hold great potential as biomarkers for pancreatic cancer. PIWIL2 and PIWIL4 are associated with better prognosis, while PIWIL1 and PIWIL3 involvement appears to be associated with carcinogenesis. We aimed to discover PIWIL3- and PIWIL4-modulated piRNAs and determine their potential mechanisms in pancreatic cancer and the clinical implications. PIWIL3 or PIWIL4 was downregulated in pancreatic cancer-derived cell lines or in a non-tumour cell line. Differentially expressed piRNAs were analysed by next generation sequencing of small RNA. Nine fresh-frozen samples from solid human pancreases (three healthy pancreases, three intraductal papillary mucinous neoplasms, and three early-stage pancreatic cancers) were included in the sequencing analysis. Two piRNAs associated with PIWIL3 (piR-168112 and piR-162725) were identified in the neoplastic cells; in untransformed samples, we identified one piRNA associated with PIWIL4 (pir-366845). After validation in pancreatic cancer-derived cell lines and one untransformed pancreatic cell line, these piRNAs were evaluated in plasma samples from healthy donors (n = 27) or patients with pancreatic cancer (n = 45). Interestingly, piR-162725 expression identified pancreatic cancer patients versus healthy donors in liquid biopsies. Moreover, the potential of the serum carbohydrate antigen 19-9 (CA19-9) biomarker to identify pancreatic cancer patients was greatly enhanced when combined with piR-162725 detection. The enhanced diagnostic potential for the early detection of pancreatic cancer in liquid biopsies of these new small non-coding RNAs will likely improve the prognosis and management of this deadly cancer.

Keywords: CA19-9; MAPK pathway; PIWI proteins; PIWIL3; PIWIL4; liquid biopsy; pancreatic cancer; piRNA; small non-coding RNA.

Figure 1

Small RNA sequencing of modified human pancreatic cancer and normal cell lines and human pancreatic tissues revealed three differentially expressed piRNAs. (A) Schematic representation of the procedure for PIWIL3 or PIWIL4 downregulation. Here, two pancreatic tumour cell lines and one non-tumour cell line were transfected with two independent siRNA sequences to downregulate PIWIL3 or PIWIL4. As a control, all cell lines were transfected with a scramble siRNA. (B) RNA from fresh-frozen human tissues obtained from pancreatic ductal adenocarcinomas (n = 3), intraductal papillary mucinous neoplasms (n = 3), and healthy pancreatic tissues (n = 3) were included in the NGS of small RNA. (C) cDNA was obtained from the total RNA isolated from all samples. After adapter ligation by PCR, the fragments were loaded onto a polyacrylamide gel, and bands from 145–160 bp were excised and purified to continue with small RNA library preparation. Sequencing was performed on a MiniSeq Platform (Illumina). (D) Three different piRNAs were commonly found in three different data integrations: 1: healthy pancreas tissues vs. PL45 and RWP1 scramble; 2: hTERT-HPNE scramble vs. PL45 and RWP1 scramble; 3: hTERT-HPNE downregulated for PIWIL4 vs. hTERT-HPNE scramble; 4: RWP1 and PL45 downregulated for PIWIL3 vs. hTERT-HPNE downregulated for PIWIL3; 5: RWP1 and PL45 downregulated for PIWIL3 vs. PDAC and intraductal papillary mucinous neoplasm (IPMN) pancreatic samples; 6: RWP1 and PL45 downregulated for PIWIL3 vs. RWP1 and PL45 scramble; 7: RWP1 and PL45 scramble vs. IPMN pancreatic samples; 8: RWP1 and PL45 scramble vs. hTERT-HPNE scramble; 9: RWP1 downregulated for PIWIL3 vs. RWP1 downregulated for PIWIL4; NGS: Next Generation Sequencing.

Figure 2

Validation of the three piRNAs obtained from the NGS of small RNA in untransformed and pancreatic-cancer-derived cell lines. Statistically significant piRNAs obtained from data integration and analyses from the NGS of small RNA (pir-366845, piR-168112, and piR-162725) were evaluated in one untransformed human pancreatic cell line (hTERT-HPNE) and in five human pancreatic-cancer-derived cell lines. Graphs show piRNA fold-change expression normalised to U6B small nuclear RNA gene (RNU6B) expression. NGS: Next Generation Sequencing.

Scheme 1

Expression of piR-168112 in human plasma samples. (A) Relative piR-168112 expression in plasma samples from 14 healthy donors and 16 PDAC patients recruited institutionally. (B) Statistical analysis of piR-168112 expression in plasma samples from our institutional cohort grouped by their status, i.e., healthy or tumour. (C) Relative piR-168112 expression in plasma samples from seven healthy donors and 22 PDAC patients from a validation cohort. (D) Statistical analysis of piR-168112 expression in plasma samples from the validation cohort grouped by healthy or tumour status. IC-H: Institutional cohort healthy samples. IC-T: Institutional cohort tumour samples. VC-H: Validation cohort healthy samples. VC-T: Validation cohort tumour samples. p-values < 0.05 were considered statistically significant.

Figure 3

Expression of piR-162725 in human plasma samples. (A) Relative piR-162725 expression in plasma samples from healthy donors and PDAC patients recruited institutionally. (B) Statistical analysis of piR-162725 expression in plasma samples from our institutional cohort grouped by their status, i.e., healthy or tumour. (C) Relative piR-162725 expression in plasma samples from healthy donors and PDAC patients from a validation cohort. (D) Statistical analysis of piR-162725 expression in plasma samples from the validation cohort grouped by healthy or tumour status. IC-H: Institutional cohort healthy samples. IC-T: Institutional cohort tumour samples. VC-H: Validation cohort healthy samples. VC-T: Validation cohort tumour samples. p-values < 0.05 were considered statistically significant.

Figure 4

Expression of piR-162725 in human plasma samples enhances the diagnostic potential of CA19-9. The bar graph represents the percentage of PDAC patients identified by any factor individually (piR-168112 (white), piR-162725 (yellow), or CA19-9 (red)) or in combination with CA19-9 (piR168112+CA19-9 (green) or piR162725+CA19-9 (blue)). p-value < 0.05 was considered statistically significant.

Figure 5

Survival curves according to the expression levels of both tumourigenic piRNAs. (A) Kaplan–Meier curve for progression-free survival according to piR-168112 expression. (B) Kaplan–Meier curve for overall survival according to piR-168112 expression. (C) Kaplan–Meier curve for progression-free survival according to piR-162725 expression. (D) Kaplan–Meier curve for overall survival according to piR-162725 expression. Statistical analyses were performed using the log-rank test. p-values < 0.05 were considered statistically significant.

Figure 6

KEGG pathway analysis of piR-162725 highlights its potential as a modulator of the MAPK signalling pathway. (A) PiR-162725-mediated KEGG pathway network showing putative mRNA interaction with piR-162725. (B) Dot plot of the top 13 KEGG pathways which were enriched in piR-162725 potentially modulated transcripts.