Low expression of tRF‐Pro‐CGG predicts poor prognosis in pancreatic ductal adenocarcinoma

ABSTRACT Background & Aims tRFs (tRNA‐derived RNA fragments) have been reported to facilitate cancer progression in multiple cancers. However, their role in pancreatic ductal adenocarcinoma (PDAC) remains to be determined. In this study, we mainly investigated the expression of tRF‐Pro‐CGG in pancreatic ductal adenocarcinoma and evaluated its relationship with the clinicopathology and survival time of patients. Methods 37 cases of pancreatic ductal adenocarcinoma, and 15 cases of normal pancreatic tissues were collected which were resected by surgery from January 2017 to June 2020 from the Department of Hepatobiliary and Pancreatic surgery of Changzhou second people's Hospital. The expression of tRF‐Pro‐CGG in paraffin‐embedded tissues was detected by fluorescence in situ hybridization (FISH). The clinical data including age, sex, tumor location, tumor diameter, tumor clinical stage (TNM stage), depth of invasion, regional lymph node metastasis, serum CA199, and serum CEA were collected and analyzed retrospectively, whether the expression tRF‐Pro‐CGG was correlation with the pathological parameters and clinical outcomes of patients. Results The expression level of tRF‐Pro‐CGG was significantly downregulated in PDAC and associated with an advanced TNM stage (P=0.000) and the N stage (P=0.000) of patients. More importantly, low tRF‐Pro‐CGG expression predicted poor survival in PDAC patients (P=0.003). Conclusions TRF‐Pro‐CGG is under‐expressed in PDAC and is associated with short clinical survival and poor prognosis. tRF‐Pro‐CGG is an independent prognostic factor, which highlights its role as a potential biomarker for PDAC progression and therapy.

surgical resection and late chemotherapy. Due to early metastasis and chemotherapy tolerance, the effect of surgical resection and chemotherapy is not ideal, and most patients are not eligible for surgery after diagnosis. 3 Thus, at present, the understanding of the pathogenesis, diagnosis, and treatment of pancreatic cancer needs to be further studied.
tRFs are small RNAs approximately 14-30 nt in length 4,5 that match the ends of mature and initial tRNA transcripts. 6 tRFs are a type of tsRNAs, which share the specificity of tRNA in specific cells or tissues. 7 tRFs are produced by specific cleavage under certain conditions by specific nucleases (dicer, ANG, etc.) or under circumstances of hypoxia or stress. 8 According to the matched tRNA regions, tRFs can be divided into three categories, namely tRF-5, tRF-3, and tRF-1. 6,9 tRF-5 and tRF-3 are derived from the 5′ and 3′ ends of mature tRNA, respectively, while tRF-1 is derived from the 3′ end of the original tRNA. The length of tRF-5 is approximately 1 to 30 nt, and it is formed by cutting the stem region and anticodon loop of the D loop or t loop of tRNA.
The biological functions of tRFs are as follows: (1) tRFs can participate in the regulation of the expression of various genes. 10 (2) tRFs can change the epigenetic status of genes. 11,12 (3) tRFs have a certain correlation with neuron degeneration. 13 (4) tRFs can regulate chromatin. 14 Since this type of tRF has a 5'-phosphate and 3'-OH and is similar in length to miRNA, it has miRNA-like functions in cancer. For example, the biogenesis of dicer relies on tRFs to form RISC complexes with the argonaute protein and results from RNA silencing. 15 Some miRNAs can be directly matched with tRFs, which inhibit the proliferation of tumor cells by binding to the YBX-1 site. 16,17 In addition, under stress conditions, tRFs can promote the assembly of SGs, thereby inhibiting translation. According to relevant literature reports, 18 tRFs play a role in the occurrence and metastasis of breast cancer, 16 ovarian cancer, 19 and prostate cancer, 20 but no studies have been published regarding pancreatic ductal adenocarcinoma. Therefore, the expression levels and underlying mechanism of tRFs in PDAC remain to be elucidated.
Moreover, previously reported results showed that tRF-Pro-CGG (TRF-001391), whose sequence is GAAGCGAGAA TCATACCCC T AGACCAACGA GCC, was differentially expressed in pancreatic cancer and normal pancreatic tissue based on tRF and tiRNA sequencing, which was verified by RT-PCR in tissue. 21 tRF-Pro-CGG was mostly enriched in "neuromuscular processes" (biological process), "neurons" (cellular component), and "PDZ domain binding" (molecular function) and was mostly enriched in "the PI3 K/protein kinase-B signaling pathway." However, the differential expression in pancreatic tissues and the clinical significance of tRF-Pro-CGG are still unclear.
The main purpose of this study was to verify the differential expression of tRF-Pro-CGG between pancreatic ductal adenocarcinoma tissue and normal patient pancreatic tissue and to examine the relationship between tRF-Pro-CGG expression and clinicopathological parameters. This may suggest that tRF-Pro-CGG is a potential tumor marker of PDAC.

| Patients
A total of 37 samples of paraffin-embedded human pancreatic ductal adenocarcinoma tissue, 15 samples of normal pancreatic tissue, and clinical-pathological parameters of Changzhou Second People's Hospital from January 2017 to June 2020 were collected. All specimens were confirmed to be pancreatic ductal adenocarcinoma by definitive histopathological diagnosis. Inclusion criteria were as follows: (1) ten percent formalin-fixed and paraffin-embedded (FFPE) pancreatic cancer tissue specimens; (2) pathologically confirmed pancreatic ductal adenocarcinoma; (3) complete clinical and pathological patient data; and (4) no radiotherapy or chemotherapy was administered before surgery. Exclusion criteria were as follows: (1) other primary tumors and (2) incomplete clinical and pathological data. The pathological diagnosis and TNM staging of pancreatic cancer tissue was based on the UICC/AJCC TNM staging system (8th edition, 2017). 22 The thickness of each paraffin section was approximately 3 mm.
Normal pancreatic tissue was obtained at least 2 cm from the edge of the pancreatic lesion and served as a control group.

| Evaluation of tRF-Pro-CGG expression
Two pathologists performed independent scoring without clinical data. The FISH results were based on the staining intensity score (0 points (negative), 1 point (weak positive) and 2 points (strong positive)) and according to the proportion of positive cells (0 points (0%), 1 point (scattered or 0%-10% positive cells), 2 points (local distribution or 10%-50% positive cells), and 3 points (diffuse distribution or more than 50% positive cells). The staining intensity score and the positive cell proportion score were added to obtain the FISH score, and PDAC tissues were divided into low-expression and highexpression groups. According to the FISH score, high expression in cells was defined as a score ≥2.

| Statistical analysis
The data were analyzed with the software package SPSS, version 23.0. Fisher exact tests were used to assess associations between tRF-Pro-CGG expression and other parameters. Univariate survival analysis was performed with the Kaplan-Meier method, and differences in survival curves were assessed with the log-rank test.
Multivariate survival analysis was performed using the Cox regression model. A p value <0.05 was considered statistically significant.

| tRF-Pro-CGG was downregulated in PDAC
To explore tRF-Pro-CGG expression in PDAC, we performed FISH to detect tRF-Pro-CGG in FFPE pancreatic cancer tissue specimens.
The results demonstrated that tRF-Pro-CGG was mainly expressed and localized in the cytoplasm of PDAC cells ( Figure 1). Compared with the negative control sample ( Figure 1A), certain normal pancreatic tissues had high tRF-Pro-CGG expression ( Figure 1B). However, in cancerous tissues, tRF-Pro-CGG had low expression ( Figure 1C,D).
Compared with normal pancreatic samples, tRF-Pro-CGG was significantly downregulated in PDAC samples ( Figure 2).

| The expression level of tRF-Pro-CGG was associated with an advanced TNM stage
To explore the clinical function of tRF-Pro-CGG, we extracted all the pathological factors and conducted univariate analysis. The results are shown in Table 1. Among these pathological parameters, the expression difference in tRF-Pro-CGG was closely related to the TNM stage (p < 0.01). Further study revealed that lymphatic metastasis was a crucial indicator associated with high tRF-Pro-CGG expression in PDAC patients. According to the pathological stage, matched-pair analysis demonstrated that tRF-Pro-CGG was significantly downregulated in advanced PDAC ( Figure 3).

| Low expression of tRF-pro-CGG predicts poor prognosis in PDAC patients
To further explore the clinical significance of tRF-Pro-CGG for pre-

| ROC analysis of the diagnostic value of tRF-Pro-CGG
Though tRF-Pro-CGG could act as an independent factor for predicting prognosis, its function in clinicopathologic diagnosis is still unknown. ROC analysis was used to evaluate the diagnostic value of tRF-Pro-CGG. The results shown in Figure 5 show that the AUC of tRF-Pro-CGG of the diagnostic curve was 0.92 (95% CI 0.8466-0.9948, p < 0.0001). The sensitivity and specificity were 75.7% and 93.3%, respectively.

| DISCUSS ION
Pancreatic cancer remains one of the malignancies with the worst prognosis despite recent advances in cancer treatment. Surgery is the only curative therapy. Combining surgery with adjuvant gemcitabine, which represents the current gold-standard therapy for resected pancreatic cancer, improves outcomes, although with limited benefits. 23 The characteristics of PDAC are local aggressiveness, early lymphatic and hematogenous dissemination, and chemotherapeutic resistance. 24 The most common tissue diagnosis method is the fine-needle aspiration (EUS-FNA) sampling of pancreatic masses or other suspicious lesions guided by endoscopic ultrasonography (EUS). 25  Previously reported results showed that tRF-Pro-CGG (TRF-001391), whose sequence is GAAGCGAGAA TCATACCCC T AGACCAACGA GCC, was differentially expressed in pancreatic cancer and normal pancreatic tissue by using tRF and tiRNA sequencing, which was verified by RT-PCR in the tissue. 21 To bet- This result is the same as the study that the expression of tRFs prognosis. 38 This fully shows that the low expression of tRFs is related to poor prognosis. Univariate analysis showed that the low expression of tRF-Pro-CGG and clinicopathological factors such as TNM stage (Ⅰ/Ⅱ-Ⅲ) and regional lymph node metastasis (N0/N1-N2) were related to the prognosis of PDAC patients. Multivariate analysis showed that the low expression of tRF-Pro-CGG in tumor tissues was an independent risk factor for poor prognosis of patients with PDAC. The results of univariate analysis and multivariate analysis showed that the low expression of tRF-Pro-CGG may be an independent prognostic factor for patients with PDAC. that the expression of tRF-Pro-CGG is related to the diagnosis of PDAC patients, and it is very likely to become a tumor biomarker for early diagnosis of PDAC patients. It is reported that more tRFs can be found in the exocrine than miRNA, which indicates that tRFs can be stably expressed in body fluids and rich in content. This makes them potential markers for early tumor diagnosis. 39 In liver cancer, Zhu et al found that the level of plasma exocrine tRF-5GluCTC in patients with liver cancer was significantly higher than that in healthy controls, suggesting that tRFs in plasma exosome can be used as a biomarker for early diagnosis of liver cancer. 40 In gastric cancer, the receiver working curve of Zhang et al showed that tRF-3019a could distinguish tumor tissue from non-tumor tissue, and its AUC was 0.689. Although the diagnostic potential of tRF-3019a is not ideal, it is still better than the previously reported values of CEA (AUC = 0.583) and CA199 (AUC = 0.585). TRF-3019a can be used as a suitable biomarker for (GC) diagnosis of gastric cancer. 41 26,43 The results show for the first time that tRF-Pro-CGG expression in FFPE samples can be used to delineate a subgroup of patients who will truly benefit from subsequent surgery.
Patients whose cancers express high levels of tRF-Pro-CGG are predicted to have a greater than 50% survival rate after two years.
In contrast, low levels of tRF-Pro-CGG predict poor outcomes and short OS after surgical resection.
In recent years, research on tumor growth and the migratory and invasive abilities of cancers, including PDAC, has increased.  This study had some limitations that should be considered.
First, this study was retrospective and non-random. We need to perform many prospective studies to further verify the reliability of the results. Second, the small number of patients with PDAC and the lack of adequate normal pancreatic tissue as a control group made the study less reproducible. We need large numbers of patients and multi-center collaborations. Finally, we were unable to obtain sufficient complete patient clinical data for subgroup analysis.

| CON CLUS ION
In summary, this study showed that tRF-Pro-CGG is underexpressed in PDAC and is associated with poor prognosis. tRF-Pro-CGG was an independent prognostic factor, which highlights its role as a potential biomarker for PDAC progression and therapy.

CO N FLI C T O F I NTE R E S T
The authors declare that this research is not related to any commercial or financial interests. The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.