Relationship Between PTBP1 and Pancreatic Cancer Based on microRNA and Behavior During TYMS‐Mediated Carcinogenesis

ABSTRACT

Polypyrimidine tract‐binding protein 1, an alternative splicing factor, drives cancer‐associated metabolic reprogramming by modulating pyruvate kinase M splicing. Although organ‐abundant microRNAs (miRs) regulate this protein, its role in tumor progression remains unclear. Therefore, we investigated the functional relationship between polypyrimidine tract‐binding protein 1‐associated microRNAs and this protein in pancreatic cancer to identify pancreas‐enriched microRNAs and explore polypyrimidine tract‐binding protein 1 function beyond metabolic regulation. The expression of miR‐216b‐5p, which is typically pancreas‐enriched, was significantly downregulated in clinical pancreatic cancer samples, whereas polypyrimidine tract‐binding protein 1 expression was markedly upregulated. Mechanistically, miR‐216b‐5p directly binds to the 3′ UTR of polypyrimidine tract‐binding protein 1 mRNA, thereby suppressing its expression and inhibiting cancer cell proliferation. Proteomic analysis, integrated with publicly available individual‐nucleotide resolution UV crosslinking and immunoprecipitation data, identified thymidylate synthase as a novel downstream target of polypyrimidine tract‐binding protein 1, likely regulated via splicing‐related RNA‐level mechanisms. Thymidylate synthase expression was also elevated in pancreatic cancer tissues and correlated with poor prognosis. Functional assays revealed that polypyrimidine tract‐binding protein 1 and thymidylate synthase silencing induced cell cycle arrest. Our findings uncover a previously unrecognized miR‐216b‐5p/polypyrimidine tract‐binding protein 1/thymidylate synthase regulatory axis that promotes pancreatic cancer progression and establish polypyrimidine tract‐binding protein 1 as a central molecular regulator of tumor biology beyond its metabolic role.

Corresponding Author and colleagues identify a pancreas‐enriched microRNA, miR‐216b‐5p, that suppresses PTBP1 and its downstream effector TYMS. Their findings reveal a novel miR‐216b‐5p/PTBP1/TYMS regulatory axis driving pancreatic cancer progression and establish PTBP1 as a central molecular regulator of tumor biology beyond its metabolic role.

Abbreviations

iCLIP

individual‐nucleotide resolution crosslinking and immunoprecipitation

miRNA, miR

microRNA

NMD

nonsense‐mediated mRNA decay

PC

pancreatic cancer

PKM

pyruvate kinase M

PTBP1

Polypyrimidine tract‐binding protein 1

TYMS

thymidylate synthase

1. Introduction

Pancreatic cancer (PC) is the seventh leading cause of cancer‐related death worldwide. Global trends indicate a steady increase in its incidence, which is expected to become the second leading cause of cancer‐related mortality in Western countries [1]. PC is typically diagnosed at an advanced stage, resulting in one of the lowest 5‐year survival rates (2%–9%) of all cancers [2]. In Japan, PC ranks fourth in terms of cancer‐related mortality, with 45,819 new cases and 40,175 estimated deaths as of 2023 [3]. More than 80% of patients present with unresectable disease; even among those undergoing curative pancreaticoduodenectomy, 70%–80% relapse within 2 years and eventually die from the disease [4, 5, 6]. Thus, the overall survival remains poor. The molecular basis of PC remains understudied despite its high mortality rate and resistance to standard therapies, consequently hindering advances in treatment. These findings highlight the urgent need to elucidate the underlying molecular mechanisms and develop novel therapeutic strategies.

MicroRNAs (miRNAs, miRs) are single‐stranded noncoding small RNAs that repress gene expression at the translational level by binding to complementary mRNAs [7]. Although miRs often act as fine tuners, some exhibit tissue‐enriched expression patterns and exert strong organ‐enriched regulatory effects [8, 9, 10]. miR dysregulation has been implicated in various cancers [11]. The expression of several miRs is downregulated at the adenoma stage in the colorectal adenoma–carcinoma sequence, suggesting that deregulation may occur early in tumorigenesis [12, 13, 14, 15]. This paradigm may also apply to PC, which is thought to arise from precursor lesions such as pancreatic intraepithelial neoplasia [16].

We recently investigated the miRs involved in cancer energy metabolism (Warburg effect), particularly those that regulate polypyrimidine tract‐binding protein 1 (PTBP1) [17]. PTBP1 is an RNA‐binding protein that controls alternative splicing, mRNA stability, and translation and promotes glycolysis, proliferation, and tumorigenesis [18]. It is highly overexpressed in many primary tumors [14, 18, 19, 20, 21]. miRs that regulate this protein are referred to as “PTBP1‐associated miRs” and often show high organ specificity, such as the brain‐enriched miR‐124 and muscle‐enriched miR‐133b [17].

TargetScan (https://www.targetscan.org/vert_80/) predicted several PTBP1‐targeting miRs that were most enriched in the brain and muscle. However, miRNATissueAtlas (https://ccb‐compute2.cs.uni‐saarland.de/mirnatissueatlas_2025) data revealed high miR‐216b‐5p expression in the pancreas, suggesting that it may be a previously unrecognized pancreas‐enriched PTBP1‐associated miR. The downregulation of miR‐216‐5p expression in PC is associated with aggressive features and shorter disease‐free survival [22]. These findings implicate miR‐216b‐5p in pancreatic tumor development. However, its precise function and underlying mechanisms, particularly its relationship with PTBP1, remain unclear.

In this study, we aimed to elucidate the relationship between miR‐216b‐5p and PTBP1 expression in PC tissues. Furthermore, we explored novel functions of PTBP1 in cancer energy metabolism.

2. Materials and Methods

2.1. Patients and Samples

Human pancreatic tissue samples were obtained from patients who underwent pancreatectomy at Osaka Medical and Pharmaceutical University Hospital (Takatsuki, Osaka, Japan). The study was approved by the Institutional Review Board (Osaka Medical and Pharmaceutical University, approval nos. 2305, 2344, and 2808) and conducted in accordance with the Declaration of Helsinki. We also analyzed publicly available total RNA samples (BioChain, Newark, CA, USA) obtained from normalhuman tissues to assess miR‐216b‐5p expression. Normal tissue miRNA expression profiles were obtained from the TissueAtlas database (https://ccb‐web.cs.uni‐saarland.de/tissueatlas/) [23]. Expression levels across human organs were downloaded and visualized for miR‐216b‐5p.

2.2. Cell Culture and Viability

Human PC cell lines (AsPC‐1 and Capan‐2) were provided by Emeritus Professor Nobuhiko Tanigawa (Osaka Medical and Pharmaceutical University). AsPC‐1 and Capan‐2 cells were cultured in RPMI‐1640 and McCoy's 5A media, respectively, supplemented with 10% (v/v) heat‐inactivated FBS (Sigma‐Aldrich, St. Louis, MO, USA). The cells were maintained at 37°C in a humidified incubator with 5% CO₂. Cell viability was assessed using the trypan blue exclusion assay with a TC20 automated cell counter (Bio‐Rad, Hercules, CA, USA).

2.3. Transfection Experiments

PC cells were seeded in 6‐well plates at a density of 4 × 10⁴ cells/well (10%–30% confluence) on the day of transfection. Transfections were performed using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The following nucleic acids were used: the miR‐216b‐5p mimic (mirVana; Ambion, Foster City, CA, USA), antagomir‐216b‐5p (mirVana; Ambion), PTBP1 siRNA (siR‐PTBP1; Invitrogen), and TYMS siRNA (Invitrogen). A nonspecific miRNA (5′‐GUAGGAGUAGUGAAAGGCC‐3′; HSS, Hokkaido, Japan) was used as the control [6]. The nucleotide sequences were as follows: miR‐216b‐5p, 5′‐AAAUCUCUGCAGGCAAAUGUGA‐3′; siR‐PTBP1 for the 3′UTR (siR‐PTBP1‐1), 5′‐AUCUCUGGUCUGCUAAGGUCACUUC‐3′; siR‐PTBP1 for the ORF (siR‐PTBP1‐2), 5′‐UGUCAUUUCCGUUUGCUGCAGAAGC‐3′; siR‐TYMS for 3′UTR #1 (siR‐TYMS‐1), 5′‐GGACAGGGAGUUGACCAACUGCAAA‐3′; and siR‐TYMS for 3′UTR #2 (siR‐TYMS‐2), 5′‐CAUGUGCGCUUGGAAUCCAAGAGAU‐3′. The effects of introducing miRNAs and siRNAs into cells were assessed 48 h after transfection.

2.4. Western Blotting

Whole‐cell and tissue lysates were prepared by homogenizing the samples in RIPA lysis and extraction buffer (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with a 1% protease inhibitor cocktail (Sigma‐Aldrich), followed by incubation on ice for 20 min. Lysates were centrifuged at 12000 × g for 20 min at 4°C, and the supernatants were collected. Protein concentrations were determined using a DC Protein Assay Kit (Bio‐Rad). Equal amounts of protein (7 μg per lane) were separated by SDS‐PAGE (Cosmo Bio, Tokyo, Japan) on 10.0 or 12.5% polyacrylamide gels (Fujifilm Wako, Osaka, Japan) and transferred to PVDF membranes (Bio‐Rad). Membranes were blocked for 1 h at 25°C in 5% nonfat milk in PBS supplemented with 0.1% Tween 20 (PBS‐T) and incubated overnight at 4°C with primary antibodies diluted in Can Get Signal 1 solution (Toyobo, Osaka, Japan). Primary antibodies against human PTBP1 (1:1000; catalog no. 72669), TYMS (1:1000; catalog no. 9045), and GAPDH (1:1000; catalog no. 2118) were purchased from Cell Signaling Technology (Danvers, MA, USA). The following day, the membranes were washed thrice with PBS‐T and incubated for 1 h at 25°C with a HRP‐conjugated anti‐rabbit IgG secondary antibody (1:10000; catalog no. 7074S; Cell Signaling Technology) in Can Get Signal 2 solution (Toyobo), followed by three washes with PBS‐T. Immunoreactive bands were visualized using the Immobilon Forte Western Horseradish Peroxidase Substrate (Merck, Darmstadt, Hessen, Germany) and detected using the FUSION‐FX7 imaging system (Vilber Lourmat, Marne‐la‐Vallée, France). GAPDH was used as a loading control.

2.5. Real‐Time RT‐PCR

Total RNA was extracted from tumor tissues using a NucleoSpin microRNA isolation kit (TaKaRa Bio, Kusatsu, Shiga, Japan) according to the manufacturer's instructions. RNA concentration and purity were assessed using UV spectrophotometry. Quantitative RT‐PCR was performed to measure miR‐216b‐5p expression using TaqMan MicroRNA Assays (Applied Biosystems, Carlsbad, CA, USA) and THUNDERBIRD Probe qPCR Mix (Toyobo). Reactions were performed using a Thermal Cycler Dice Real‐Time System II Model TP870 (TaKaRa Bio). Relative expression levels were calculated using the ΔΔC_q method. RNU48 was used as an internal control. For the analysis of TYMS splice variants, total RNA was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA was synthesized using SuperScript IV VILO Master Mix (Invitrogen), according to the manufacturer's protocol. Isoform‐specific RT‐PCR was performed using junction‐spanning primers designed to selectively amplify the full‐length, Δ4, and Δ2–3 TYMS transcripts (Thermo Fisher Scientific). Quantification was performed by real‐time PCR using TaqPath qPCR Master Mix (Applied Biosystems) on a QuantStudio 5 Real‐Time PCR System (Applied Biosystems), and expression levels were normalized to those of ACTB. The primer sequences are provided in Table S1.

2.6. Luciferase Reporter Assay

A putative miR‐216b‐5p binding site at positions 96–102 within the 3′UTR of human PTBP1 mRNA was identified using the TargetScan Human 8.0 database (https://www.targetscan.org/vert_80/). The sequence containing the predicted binding site was cloned into the pMIR‐REPORT luciferase miRNA Expression Reporter Vector (Applied Biosystems) according to the manufacturer's instructions. A mutated version of the seed sequence (wild‐type, AGAGAUU; mutant, AACAAUU) was generated using the GeneArt Service (Thermo Fisher Scientific), and sequence integrity was confirmed via direct sequencing. Vector mutations were confirmed using sequence analysis. The pRL‐TK Renilla Luciferase Reporter vector (Promega, Madison, WI, USA) was used as internal control. Capan‐2 cells were seeded in 24‐well plates at 1 × 10⁴ cells/well 1 day before transfection. Cells were cotransfected with 0.01 μg of either wild type or mutant reporter plasmid and 20 nM of either miR‐216b‐5p or nonspecific noncoding siRNA (Dhamacon, Tokyo, Japan). Luciferase activity was measured 24 h after transfection using a Dual‐Glo Luciferase Assay System (Promega) according to the manufacturer's protocol. The results are expressed as the ratio of firefly to Renilla luciferase activity.

2.7. Proteomic Analysis

Tumor tissues were processed in RIPA buffer using a bead homogenizer (TOMY, Tokyo, Japan, MS‐100), reacted with 10 mM TCEP (Thermo Fisher Scientific, 77720) at 55°C for 1 h, treated with 18 mM iodoacetamide (Thermo Fisher Scientific, A39271) for 30 min, and subjected to methanol–chloroform precipitation. The pellet was digested overnight with trypsin (Thermo Fisher Scientific, 90058). Online liquid chromatography‐mass spectrometry analysis was performed using an Easy‐nLC 1200 system coupled to an Orbitrap Eclipse Tribrid mass spectrometer (Thermo Fisher Scientific). Samples were trapped on a C18 guard‐desalting column (Thermo Fisher Scientific; Acclaim PepMap 100, 75 μm × 2 cm, nanoViper, C18, 5 μm, 100 Å) and separated on a C18 analytical column (Nikkyo Technos, Tokyo, Japan; C18, 3 μm, 75 μm × 12 cm). The eluted samples were analyzed using a data‐independent acquisition (DIA) method. The raw data were processed in library‐free mode using DIA‐NN (version 1.8.1).

2.8. Flow Cytometry

Cells were transfected with 5 nM siR‐PTBP1, siR‐TYMS, or control siRNA, and seeded in 6‐well plates at 4 × 10⁴ cells/well. After 48 h, the cells were harvested using 0.05% trypsin, centrifuged at 300 × g for 5 min at 25°C, and washed twice with PBS. Cells were resuspended in 1× Wash Buffer (EXBIO, Vestec, Czech Republic) at a final concentration of 2.5–10 × 10⁶ cells/mL. A 200 μL aliquot of the cell suspension was added dropwise to 2 mL of 70% ethanol on ice and gently mixed. After fixation, the cells were centrifuged at 300 × g for 5 min, and the supernatant was removed. The cells were stained with 0.5 mL of propidium iodide solution (EXBIO) in the dark for 30 min at 25°C. Samples were vortexed and analyzed within 4 h using a MACSQuant Analyzer16 flow cytometer (Miltenyi Biotec, Bergisch Gladbach, Nordrhein‐Westfalen, Germany).

2.9. Data Acquisition and Processing for Transcriptomic Analysis of PC Cohorts

Transcriptomic data were obtained from patients with PC from the Clinical Proteomic Tumor Analysis Consortium (n = 137) [24], TCGA PanCancer Atlas (n = 185) [25], and Gene Expression Omnibus (GEO) datasets, including GSE71533 (n = 88) [26] and GSE183795 (n = 244) [27]. All data were sourced from public repositories on March 4, 2025. Gene expression data were obtained from the GEO database in accession and GPL formats (https://www.ncbi.nlm.nih.gov/geo/). For RNA sequencing data, normalized counts provided by the original authors were used for downstream analyses. Institutional Review Board (IRB) approval was waived as all cohorts were deidentified and obtained from public sources. All patients included in the analysis had histologically confirmed PC. No patients were excluded after initial data acquisition.

The following R packages were used for data acquisition, processing, and visualization: GEOquery for retrieving datasets from the GEO; readr and readxl for data import; tidyverse (including dplyr, ggplot2, stringr, tibble, forcats, lubridate, purrr, and tidyr) for data handling and plotting; and ggpubr for creating scatter plots with linear regression lines and correlation statistics. To investigate the differences in clinicopathological and biological features between the low and high PTBP1 expression groups, patients were divided into low and high expression groups based on the median value within each cohort.

2.10. Statistical Analysis

Each experiment was performed in triplicate. RNA and protein expression levels in PC cell lines and clinical samples were compared using the Mann–Whitney U‐test to assess statistical differences. Statistical analyses were performed using GraphPad Prism (ver. 8.4.3; GraphPad Software, La Jolla, CA, USA). Statistical analyses of the transcriptomic data were conducted using R version 4.3.1 (RRID:SCR_001905). The low and high PTBP1 expression groups were compared using the Mann–Whitney U test for continuous data to determine the significance of differences in proportions. All data are presented as the mean ± SD. Statistical significance was set at p < 0.05.

3. Results

3.1. miR‐216b‐5p Is Preferentially Expressed in the Pancreas, and Its Expression Is Significantly Downregulated in PC Tissues

Based on data from the TissueAtlas (https://ccb‐web.cs.uni‐saarland.de/tissueatlas/), miR‐216b‐5p was preferentially expressed in the pancreas compared with other normal human organs (Figure 1A). RT‐PCR validation using commercially available normal human tissue samples (BioChain) showed high miR‐216b‐5p expression in the pancreas (Figure 1B). To assess its relevance in PC, we analyzed miR‐216b‐ 5p expression in the GSE71533 dataset. miR‐216b expression was significantly downregulated in PC samples compared with that in adjacent nontumor tissues (Figure 1C) Clinical PC samples from patients who underwent pancreatic resection at our institution (Table 1) exhibited similar downregulation (Figure 1D). These findings suggest that miR‐216b‐5p is enriched in the pancreas, and that its dysregulation may be closely associated with PC development.

FIGURE 1.

Pancreas‐enrichment and downregulation of miR‐216b‐5p expression in pancreatic cancer (PC). miR‐216b‐5p expression in (A) various normal human organs based on the TissueAtlas database, (B) commercially available normal tissues (BioChain), and in (C) 36 PC and 16 adjacent normal tissues from the GEO dataset GSE71533. (D) Relative miR‐216b‐5p expression in 10 paired PC and adjacent normal tissues. Data are presented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001.

TABLE 1.

Clinical characteristics of patients with pancreatic cancer.

	Sex	Age	T	N	pStage	Histopathology	Differentiation
Case 1	F	74	T3	N1a	IIB	Invasive ductal carcinoma	Well
Case 2	M	72	T3	N1b	IIB	Invasive ductal carcinoma	Poor > moderately
Case 3	F	82	T1c	N0	IA	Invasive ductal carcinoma	Well
Case 4	M	78	T3	N0	IIA	Invasive ductal carcinoma	Well > moderately
Case 5	M	66	T4	N1b	III	Invasive ductal carcinoma	Poor > moderately
Case 6	M	65	T3	N0	IIA	Invasive ductal carcinoma	Well
Case 7	M	76	T3	N1a	IIB	Invasive ductal carcinoma	Well > moderately
Case 8	F	79	T3	N1a	IIB	Intraductal papillary mucinous carcinoma	Well
Case 9	M	80	T3	N0	IIA	Invasive ductal carcinoma	Moderate
Case 10	F	62	T3	N0	IIA	Invasive ductal carcinoma	Well > moderately

Note: Moderately, moderately differentiated adenocarcinoma; Well, well differentiated adenocarcinoma; Poor, poorly differentiated adenocarcinoma.

Abbreviations: F, female; M, male; N, Lymph node metastasis; pStage, pathological stage; T, tumor class.

3.2. PTBP1 Is Overexpressed in Clinical PC Tissues and Directly Targeted by miR‐216b‐5p

To assess the role of PTBP1 in PC, we analyzed its expression using the GSE183795 dataset. PTBP1 expression was significantly upregulated in PC tissues (Figure 2A). This finding was validated via western blot analysis of tumors and adjacent normal tissues from 10 patients with PC in whom PTBP1 was consistently overexpressed (Figure 2B). Additionally, PTBP1 and miR‐216b‐5p were not correlated with each expression (p = 0.64) in the Clinical Proteomic Tumor Analysis Consortium study [24] (Figure S1).

FIGURE 2.

Upregulation of PTBP1 expression and its direct repression by miR‐216b‐5p in PC. (A) PTBP1 expression in 139 PC and 105 adjacent normal tissues from the GEO dataset GSE183795. (B) Western blot analysis of PTBP1 expression in 10 paired PC samples. GAPDH was used as a loading control. (C) Luciferase reporter assay of WT and mutant PTBP1 3′UTR constructs cotransfected with control or miR‐216b‐5p in Capan‐2 cells. The upper panel shows the predicted miR‐216b‐5p binding site. (D) PTBP1 levels in Capan‐2 and AsPC‐1 cells 48 h after transfection with the control, miR‐216b‐5p (20 nM), or miR‐216b‐5p (10 nM) + miR‐216b‐5p inhibitor (10 nM). (E) Cell viability following the same transfections as in (D). (F) Western blot analysis of PTBP1 in cells transfected with the control, siR‐PTBP1‐1, or siR‐PTBP1‐2 (each 5 nM). (G) Cell growth inhibition after PTBP1 knockdown with siRNAs. Data are expressed as the mean ± SD. **p < 0.01; ***p < 0.001; N.S., not significant.

We performed a luciferase reporter assay to determine whether miR‐216b‐5p directly binds to PTBP1. The miR‐216b‐5p–PTBP1 miRNA–target interactions are described in Table 2. miR‐216b‐5p significantly suppressed the activity of reporters containing WT PTBP1 3′‐UTR in Capan‐2 cells, whereas mutation of the binding site abolished this effect (Figure 2C). Consistently, miR‐216b‐5p transfection into the two PC cell lines reduced PTBP1 expression levels, whereas miR‐216b‐5p inhibition restored PTBP1 expression (Figure 2D). Functionally, miR‐216b‐5p reduced PC cell viability, which was reversed by cotransfection with its inhibitor (Figure 2E). To confirm that these effects were mediated via PTBP1, we silenced PTBP1 using siRNAs targeting the 3′UTR (siR‐PTBP1‐1) or coding region (siR‐PTBP1‐2). PTBP1 expression in PC cells was inhibited (Figure 2F), with siR‐PTBP1 significantly inhibiting PC cell proliferation (Figure 2G). These findings confirmed that PTBP1 functions as an oncogene in PC and is directly repressed by miR‐216b‐5p.

TABLE 2.

Validation of microRNAs (miRNAs) used in this study.

miRNA gene name (ID)	MIR216B (100126319)
Target gene name (ID)	PTBP1 (5725)
Species name (ID)	Homo sapiens (9606)
Genomic location of MTI Nucleotide sequence	5′AGAGATT
Location	19:810922–810,928
Location within a part of a gene	96–102 (location within 3′UTR)
Existence of previous report (searching with miRTarBase)	No
Methods for experimental validation in this study	Luciferase reporter assay Western blotting
Experimental materials used in this study	AsPC‐1 and Capan‐2 cell lines

Abbreviations: AsPC‐1 and Capan‐2, human pancreatic cancer cell lines; MIR, microRNA (gene symbol); MTI, microRNA‐target interaction; PTBP1, polypyrimidine tract binding protein 1.

3.3. TYMS Was Identified as a Downstream Effector of the miR‐216b‐5p/PTBP1 Axis

To explore the downstream effectors of the miR‐216b‐5p/PTBP1 axis, we conducted proteomic analyses of PC cell lines after PTBP1 knockdown. A schematic of the workflow and its integration with the iCLIP data is shown (Figure 3A). The expression of 1914 and 1884 proteins was downregulated in Capan‐2 and AsPC‐1 cells treated with siR‐PTBP1, respectively, whereas that of 945 proteins was downregulated (Figure 3B). These proteins were significantly involved in cell cycle‐related processes (Figure 3C). Proteomic analysis of PC cells transfected with miR‐216b‐5p revealed 234 proteins with downregulated expression in both siR‐PTBP1‐ and miR‐216b‐5p‐treated cells (Figure S2), suggesting shared downstream targets. To identify PTBP1‐regulated targets with potential direct interactions, we used previously published iCLIP data for PTBP1, which provided transcriptome‐wide maps of PTBP1‐binding sites [28]. Cross‐referencing of these data with our list of proteins with downregulated expression revealed 135 proteins that were encoded by transcripts with PTBP1‐binding sites (Figure 3D). Enrichment analysis of these proteins identified cell cycle regulation as the predominant pathway, consistent with our earlier findings (Figure 3E). TYMS, a key enzyme in nucleotide biosynthesis and cell cycle progression, emerged as a central node (Figure 3F), and its position is highlighted in the volcano plot (Figure 3G). These results suggest that TYMS is a direct downstream effector of the miR‐216b‐5p/PTBP1 axis, potentially mediating tumor progression in PC.

FIGURE 3.

Identification of thymidylate synthase as a downstream effector of PTBP1. (A) Schematic of the proteomic analysis workflow and integration with iCLIP data. (B) Proteomic analysis of Capan‐2 and AsPC‐1 cells following PTBP1 knockdown (siR‐PTBP1‐1). (C) Enrichment analysis of 945 proteins whose expression is commonly downregulated in both PC cell lines. Top enriched pathways identified using multiple databases, including Reactome 2022, BioPlanet 2019, KEGG 2021 Human, and ARCHS4 Kinases Coexp. Predominant pathways include cell cycle regulation, nucleotide metabolism, DNA replication, and p53 signaling. (D) Cross‐referencing of downregulated proteins with PTBP1‐binding transcripts based on public iCLIP data. (E) Enrichment analysis of 135 proteins with both PTBP1‐binding sites and reduced expression. Pathway enrichment was performed using multiple databases (Reactome 2022, BioPlanet 2019, KEGG 2021 Human, and ARCHS4 Kinases Coexp). Cell cycle regulation, p53 signaling, and DNA replication were the dominant pathways among the PTBP1‐bound proteins with downregulated expression. (F) Network analysis of TYMS as the central downstream effector. Protein–protein interaction network of 135 PTBP1‐bound proteins with downregulated expression. TYMS (orange) is a central hub linked to multiple cell cycle‐related proteins (green) and pathways (blue), indicating its functional relevance. (G) A volcano plot of proteomic changes following PTBP1 knockdown in PC cells. Differential protein expression in AsPC‐1 (red) and Capan‐2 (blue) cells after PTBP1 silencing. TYMS is highlighted in each dataset, indicating significant downregulation. Downstream effectors of PTBP1.

3.4. TYMS Is Highly Expressed in PC Tissues and Regulated by the miR‐216b‐5p/PTBP1 Axis

To investigate the downstream effects of the miR‐216b‐5p/PTBP1 axis, we assessed TYMS expression in PC cell lines transfected with either miR‐216b‐5p or siR‐PTBP1. TYMS expression was downregulated in both cell types (Figure 4A). In addition, isoform‐specific RT‐PCR analyses using junction‐spanning primers demonstrated that PTBP1 knockdown resulted in a coordinated reduction of multiple protein‐coding TYMS splice variants, including the full‐length, Δ4, and Δ2–3 transcripts, without a marked change in their relative ratios (Figure 4B,C). In contrast, TYMS knockdown did not affect PTBP1 expression, suggesting a unidirectional regulatory relationship (Figure 4D). TYMS knockdown using two independent siRNAs targeting its 3′UTR (siR‐TYMS‐1/2) significantly suppressed cell proliferation in both cell lines (Figure 4E). Additionally, PTBP1 or TYMS knockdown induced G0/G1 arrest and S‐phase prolongation. Furthermore, PTBP1 knockdown increased the proportion of cells in the G2‐phase (Figure 4F). These findings suggest that TYMS is a downstream effector of the miR‐216b‐5p/PTBP1 axis and contributes to PC cell proliferation via cell cycle regulation, partly via PTBP1‐dependent RNA‐level regulation.

FIGURE 4.

TYMS expression is upregulated in PC and suppressed via the miR‐216b‐5p/PTBP1 axis. (A) TYMS expression in Capan‐2 and AsPC‐1 cells 48 h after transfection with miR‐216b‐5p (20 nM) or siR‐PTBP1‐1 (5 nM). (B) Isoform‐specific RT‐PCR analysis of TYMS transcripts in Capan‐2 and AsPC‐1 cells 48 h after siR‐PTBP1‐1 transfection (5 nM), using junction‐spanning primers for the full‐length, Δ4, and Δ2–3 variants. (C) Relative ratios of TYMS splice variants following PTBP1 knockdown. Ratios of Δ4 and Δ2–3 transcripts were calculated relative to the full‐length transcript in each cell line (Δ4/FL and Δ2–3/FL). (D) TYMS and PTBP1 expression in PC cells at 48 h after siR‐TYMS transfection (5 nM). (E) Effects of TYMS knockdown on cell proliferation. (F) Cell cycle distribution, as assessed by flow cytometry after PTBP1 or TYMS knockdown (5 nM). Data are presented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001.

3.5. TYMS And PTBP1 Are Associated With PC Prognosis

To investigate the clinical effect of TYMS, we assessed its expression using the GSE183795 dataset and tumor and adjacent normal tissues from 10 patients with PC. TYMS expression was elevated in PC tissues in both the GSE183795 dataset (Figure 5A) and clinical samples (Figure 5B). Low PTBP1 and TYMS expression was significantly associated with prolonged progression‐free survival (p < 0.05) and tended to correlate with improved overall survival (p = 0.03) in the TCGA PanCancer Atlas data (Figure 5C). PTBP1 and TYMS expression were slightly positively correlated (R = 0.2329, p < 0.01) (Figure 5D). These findings indicate that elevated TYMS expression is associated with PC progression and may be partially regulated by PTBP1‐mediated mechanisms. We summarized the function of the mir‐216b‐5p/PTBP1/TYMS axis in a schematic diagram (Figure 6).

FIGURE 5.

TYMS is associated with PC prognosis. (A) TYMS expression in 139 PC and 105 normal pancreatic tissues from the GEO dataset GSE183795. (B) TYMS expression in 10 paired PC and adjacent normal tissues, as determined via western blotting. (C) Kaplan–Meier curves showing associations between PTBP1 or TYMS expression and overall/progression‐free survival (TCGA Pan‐Cancer Atlas). Patients were classified into two groups based on the median expression levels of TYMS and PTBP1: Those with low (low–low) and high (high–high) expression of both genes. (D) Correlation between PTBP1 and TYMS expression in 185 PC samples (TCGA Pan‐Cancer Atlas). Data are presented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 6.

Schematic illustration of the proposed miR‐216b‐5p/PTBP1/TYMS axis in pancreatic tumorigenesis. Reduced miR‐216b‐5p expression in pancreatic cancer leads to upregulated PTBP1, which promotes PTBP1‐dependent RNA level regulation of TYMS, potentially by limiting the accumulation of non‐productive, nonsense‐mediated mRNA decay (NMD)‐targeted transcripts, thereby increasing the abundance of productive TYMS transcripts and enhancing cancer cell cycle progression.

4. Discussion

In this study, we identified a novel tumor‐suppressive axis involving miR‐216b‐5p, PTBP1, and TYMS in PC. miR‐216b‐5p was preferentially expressed in normal pancreatic tissues, whereas its expression was significantly downregulated in PC specimens. Furthermore, restoration of miR‐216b‐5p expression suppressed the proliferation of PC cells. Consistent with previous reports, our findings confirm that miR‐216b‐5p functions as a pancreas‐enriched anti‐oncomiR in PC [22, 29]. Previous studies have shown that organ‐enriched miRs suppress PTBP1 to maintain cell‐type–specific metabolic programs, mainly through the regulation of pyruvate kinase M (PKM), which encodes PKM1 and PKM2 isoforms, and glycolytic metabolism [17]. In this context, the present study identified miR‐216b‐5p as a pancreas‐enriched miR and demonstrates that miR‐216b‐5p–mediated regulation of PTBP1 is lost in PC. Notably, our findings revealed that TYMS is a novel downstream effector of PTBP1, suggesting that PTBP1 contributes to PC progression by regulating nucleotide metabolism rather than solely glycolysis.

PTBP1 expression is tightly regulated by organ‐enriched miRs promoting normal differentiation. For example, miR‐124 and miR‐133 suppress PTBP1 expression during neuronal and muscle differentiation, respectively [30, 31]. Their expression is often downregulated during malignant transformation, leading to PTBP1 overexpression in glioblastoma and rhabdomyosarcoma [20, 32, 33].

miR‐216b‐5p suppresses oncogenic targets in PC [34, 35, 36]. Germline deletion of the miR‐216/‐217 cluster resulted in embryonic lethality [37], and individual knockouts of miR‐216 members impaired pancreatic regeneration and enhanced acinar‐to‐ductal metaplasia, suggesting that the loss of the miR‐216 family may contribute to PC initiation [38]. In this study, a database‐based analysis of PTBP1‐targeting miRs revealed that miR‐216b‐5p was the only miR predicted to target PTBP1 and was highly enriched in the pancreas. This contrasts with other organs, such as the brain and skeletal muscle, in which multiple organ‐enriched miRs contribute to PTBP1 regulation, suggesting a pancreas‐specific regulatory architecture. In addition, unlike the brain and skeletal muscle, the pancreas is not a predominantly PKM1‐expressing organ under physiological conditions. This difference in metabolic background suggests that the regulation of PTBP1 by organ‐enriched miRs in the pancreas may not be limited to glycolytic control. Consistent with this notion, our data suggest that the loss of miR‐216b‐5p–mediated regulation of PTBP1 in PC is also associated with altered nucleotide metabolism and DNA synthesis via TYMS. These findings suggest that the miR‐216 family maintains acinar cell identity and epithelial homeostasis, and that its loss may contribute to tumor initiation.

Most studies have investigated the role of PTBP1 in regulating PKM isoforms, in which PTBP1‐mediated mutually exclusive exon usage induces a shift toward PKM2, thereby promoting cancer‐specific metabolism and drug resistance [39, 40, 41]. However, previous studies have reported that PTBP1 inhibition does not necessarily result in decreased PKM2 expression across all tumor contexts [42]. This indicates that PTBP1‐mediated metabolic regulation and the associated metabolic vulnerabilities upon PTBP1 suppression are highly context‐dependent.

Consistent with this concept, PTBP1‐dependent regulation of TYMS differs fundamentally from that of the canonical PKM splicing switch. Although the alternative splicing of PKM represents a binary transition between two productive isoforms [17], the alternative splicing of TYMS does not appear to involve such a clear isoform switch. Previous studies have shown that loss of PTBP1 is associated with complex splicing alterations in TYMS, including exon inclusion and intron retention, which generate non‐productive transcripts predicted to undergo nonsense‐mediated mRNA decay (NMD) [28].

Consistent with this framework, our isoform‐specific RT‐PCR analyses demonstrated that PTBP1 knockdown resulted in a coordinated reduction of multiple protein‐coding TYMS splice variants, including the full‐length, Δ4, and Δ2–3 transcripts, without a marked change in their relative ratios (Figure 4B,C). Although aberrant splicing events and NMD‐targeted transcripts were not directly assessed in this study, these findings support a model in which PTBP1 contributes to the maintenance of productive TYMS transcripts at the RNA level, rather than functioning as an isoform switching factor (Figure 6).

Among the multiple PTBP1‐regulated candidates identified by integrated proteomic and iCLIP analyses [28], TYMS was prioritized based on enrichment analyses that highlighted nucleotide metabolism, DNA replication, and cell cycle–related pathways, as well as protein–protein interaction network analyses identifying TYMS as a central hub (Figure 3). Together with its established roles in DNA synthesis, tumor growth, and therapeutic resistance in PC, these findings suggest that TYMS is a PTBP1‐bound, biologically relevant downstream effector that is regulated at the post‐transcriptional level.

TYMS, a key enzyme in de novo nucleotide synthesis and a therapeutic target of 5‐fluorouracil, contributes to tumor cell proliferation through regulation of the cell cycle [43, 44, 45]. Two alternative TYMS splicing variants (Δ2 + 3 and Δ4), which escape by retaining the open reading frame, have been reported [46]. Previous studies have shown that PTBP1 directly controls TYMS alternative splicing, suppressing non–protein‐coding transcripts and promoting protein‐coding TYMS mRNA [28]. Collectively, these findings support a model in which PTBP1‐driven splicing regulation promotes the accumulation of catalytically competent TYMS isoforms to sustain tumor cell proliferation. Further studies are needed to clarify how PTBP1 influences the balance among TYMS splice variants and their enzymatic activities.

PTBP1 and TYMS are overexpressed in PC tissues, and their expression levels are positively correlated. Patients with low expression of both genes showed prolonged progression‐free survival, further supporting the oncogenic role of the miR‐216b‐5p/PTBP1/TYMS axis in PC. A correlation between miR‐216b‐5p and PTBP1 expression was not clearly observed in the bulk RNA sequencing analyses of the clinical specimens (Figure S1). PC tissues are characterized by marked intratumoral heterogeneity and abundant stromal and immune cell components, which may dilute tumor cell–specific miR–mRNA regulatory relationships in bulk RNA analyses. Additionally, miR‐mediated gene regulation primarily occurs at the post‐transcriptional level; therefore, analyses based solely on mRNA expression levels may not fully capture these regulatory interactions. These factors may partly explain the weak correlation between miR‐216b‐5p and PTBP1 expression in clinical samples.

The implications of this study are twofold. First, PTBP1 functions not only as a regulator of alternative splicing and glycolysis but also as a key mediator of an oncogenic pathway linking RNA processing to nucleotide metabolism. Second, this axis represents potential therapeutic vulnerability to PC. Restoring miR‐216b‐5p expression may represent a pancreas‐targeted strategy with minimal systemic toxicity given its tissue‐enriched expression profile. In contrast, the direct inhibition of PTBP1 may pose challenges because of its role in RNA regulation. In addition to potential therapeutic applications, the present study elucidates the molecular pathogenesis of PC.

This study had several limitations. First, the number of clinical specimens available for analysis was limited in the present study for several reasons. Initially, the number of PC patients undergoing surgical resection was limited. Additionally, this was a retrospective study, and only cases with sufficient tumor tissue and the ability to obtain high‐quality RNA and protein simultaneously were included in the analysis. However, the primary aim of this study was not to perform large‐scale clinical association analyses, but to identify a novel regulatory axis involving miR‐216b‐5p, PTBP1, and TYMS and to clarify its biological significance. Therefore, this study can be considered a hypothesis‐generating study. In addition, functional analyses were performed using two pancreatic cancer cell lines, Capan‐2 and AsPC‐1, which are widely used as representative models in PC research. However, the interpretation of the present results should consider certain limitations, including potential cell line–dependent effects.

Second, our functional analyses were based on two PC cell lines, which do not fully capture the complexity of tumor biology. Animal studies are required to establish the causal role of the miR‐216b‐5p/PTBP1/TYMS axis in pancreatic tumorigenesis and evaluate the therapeutic response. Third, proteomic and related protein expression analyses in this study were performed using PC cell lines, and therefore mainly reflected tumor cell–derived protein expression. However, such in vitro analyses may not fully capture the tumor microenvironment present in vivo. In addition, immunohistochemical and immunofluorescence analyses were not performed, and the cell‐type–specific expression and subcellular localization of PTBP1 remain to be clarified. Although the present study demonstrated the biological significance of the miR‐216b‐5p/PTBP1/TYMS axis in PC cells, further in vivo studies are required to fully evaluate its therapeutic potential. Validation using animal models is necessary to assess the feasibility and efficacy of targeting this regulatory axis in the physiological tumor microenvironment.

In conclusion, we identified the pancreatic‐enriched PTBP1‐associated miR‐216b‐5p/PTBP1/TYMS axis as an uncharacterized regulatory pathway in PC. These findings provide novel insights into the molecular pathology of PC and may inform future studies exploring the biological significance of this axis.

Author Contributions

Shigenori Suzuki: conceptualization, validation, investigation, data curation, formal analysis, visualization, methodology, writing – original draft. Hiroaki Hase: methodology, formal analysis. Yosuke Inomata: methodology, project administration, resources, writing – review and editing. Jun Arima: data curation, writing – review and editing. Kentaro Jingushi: data curation, formal analysis. Atsushi Tomioka: visualization. Nao Kawaguchi: software. Hiroki Hamamoto: formal analysis. Ryo Tanaka: formal analysis. Yoshiro Imai: methodology, data curation. Mitsuhiro Asakuma: funding acquisition, supervision. Tomohito Tanaka: funding acquisition, project administration. Kazutake Tsujikawa: formal analysis, funding acquisition, methodology. Kohei Taniguchi: project administration, resources, software, writing – review and editing, supervision. Sang‐Woong Lee: resources.

Funding

This study was supported in part by JSPS KAKENHI (grant numbers 16K10615, 21K08761, and 25K02711). This research was supported by the Research Support Project for Life Science and Drug Discovery (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED (grant numbers JP25ama121052 and JP25ama121054).

Ethics Statement

This study was approved by the Research Ethics Committee of Osaka Medical and Pharmaceutical University.

Consent

Informed consent was obtained from all participants.

Conflicts of Interest

Kohei Taniguchi is a council member of the Japanese Cancer Association. The other authors declare no conflicts of interest.

Supporting information

Figure S1: Correlation between the expression of PTBP1 and miR‐216b‐5p in pancreatic cancer.

Figure S2: A Venn diagram of proteins exhibiting downregulated expression in proteomic analysis.

Table S1: Primer sequences used for isoform‐specific RT‐PCR analysis of TYMS.

Data Availability Statement

The data that support the findings of this studys are available in the Gene Expression Omnibus (GEO), accession numbers GSE71533, GSE183795, and GSE107094. Clinical association and survival data were analyzed using TCGA Pan‐Cancer Atlas datasets. The raw proteomic data generated in this study are available from the corresponding author upon reasonable request.