Pyrroline-5-Carboxylate Reductase-2 as a Potential Prognostic Biomarker of Colorectal Cancer: A Study Based on Bioinformatics Approaches

Background: Colorectal cancer (CRC) is associated with high mortality rates. Some studies have attempted to elucidate the molecular mechanisms underlying CRC, and few CRC biomarkers are now used in clinical practice. Methods: Herein our aim was to identify a new potential prognostic biomarker of CRC, while providing further theoretical basis for studying the pathogenesis of CRC. Results: PYCR2 expression was to be much higher in CRC tissues than in corresponding non-cancerous tissues, and the prognosis of patients with high PYCR2 expression was signicantly poorer than that of those with low PYCR2 expression. Further, PYCR2 showed a high correlation with TRMT61B and CDC23 expression. We found that gene mutations had a signicant inuence on the prognosis of CRC. PYCR2 siRNA transfection decreased and inhibited the ability of migration and invasion in CRC cells. Conclusions: To summarize, our ndings provide novel insights to facilitate better diagnosis as well as treatment of CRC.


Background
Colorectal cancer (CRC) is the third most common cancer worldwide and accounts for 9.2% of new cancer deaths, ranking second among all cancers. The incidence and death rates of CRC have increased in the last decade (including in the Baltic countries, Russia, China, and Brazil). [1] Further, treatment methods, such as surgery, chemotherapy, radiation, and immunotherapy, have rapidly evolved over the past few decades, substantially improving the survival rate of patients with CRC. [2] Approximately 15-25% patients with CRC also show liver metastasis, and the overall survival of such patients remains poor. [3][4][5] Early diagnosis and prompt treatment are pivotal to improve the survival rate of patients with CRC.
[6] Despite CRC being a critical condition, to date, prognostic ability remains unsatisfactory across the globe.
Pyrroline-5-carboxylate reductase-2 (PYCR2) is one of three human PYCRs that can catalyze the reduction of D1-pyrroline-5-carboxylic acid to proline; further, it is a non-enzymatic antioxidant with the ability to inhibit apoptosis. [7] Under oxidative stress, apoptosis levels have been reported to increase in the absence of PYCR2; moreover, in malignant melanoma, PYCR2 silencing was found to decrease A375 cell proliferation and induce autophagy via the AMPK/mTOR pathway. [8] In addition, Gao et al. identi ed PYCR2 to be a prognostic biomarker of hepatocellular carcinoma; they found that PYCR2 was related to proteomic subgrouping and involved in the metabolic reprogramming of hepatocellular carcinoma. [9] Based on the aforementioned data, we hypothesized that PYCR2 is closely related to CRC and that it plays a crucial role in CRC development. Herein we analyzed several databases to investigate the expression levels of PYCR2 and related genes in CRC. Our aim was to elucidate the role of PYCR2 in CRC development and to determine whether PYCR2 can be used as a prognostic biomarker of CRC.

Page 3/19
Methods Exploring PYCR2 prognostic ability and expression pattern in CRC ENCORI Pan-Cancer Analysis Platform [10] (http://starbase.sysu.edu.cn/panCancer.php) is designed for decoding pan-cancer networks of lncRNAs, miRNAs, pseudogenes, snoRNAs, RNA-binding proteins, and protein-coding genes by assessing their expression pro les across 32 cancer types integrated from The Cancer Genome Atlas (TCGA) project. In this study, we used this platform to study PYCR2 expression pattern and prognostic ability in CRC.

Functional analysis of PYCR2 overlapping target genes
To further elucidate the biological function of PYCR2 overlapping target genes in CRC, we performed gene ontology (GO) enrichment analysis using Metascape [15] (http://metascape.org/gp/index.html).
Determining the prognostic signi cance of PYCR2 overlapping target genes ENCORI and GEPIA[16] (http://gepia.cancer-pku.cn/) were applied to determine whether the expression levels of PYCR2 overlapping target genes varied between CRC and non-cancerous tissues and whether they affected the prognosis of CRC.
Assessing the co-expression of PYCR2 and related genes and determining the mutational status of target genes cBioPortal [17,18] (http://www.cbioportal.org/), a web-based tool to explore, visualize, and analyze multidimensional cancer genomics data, was used to analyze the co-expression of PYCR2 and related genes. P < 0.05 was the cut-off criterion. Further, the mutational status of the target genes of PYCR2 in CRC was investigated using cBioPortal.
Comparing prognosis between patients with and without gene mutations The International Cancer Genome Consortium (ICGC) [19] (https://dcc.icgc.org/) serves to catalog largescale cancer genome studies in tumors from 50 distinct cancer types/subtypes, containing clinical data as well as data pertaining to abnormal gene expression, somatic mutations, and epigenetic modi cations. We used this portal to compare prognosis between patients with and without gene mutations.
Cell culture HCT116, HT29, and LoVo CRC cell lines were obtained from the American Type Culture Collection Cell transfection HT29, LoVo, and HCT116 cells were seeded into a 6well plate and cultivated at 37°C and 5% CO 2 . Their growth was closely monitored until they reached a fusion rate of 80%, and they were then transfected with siRNA ( nal concentration = 100 nM; Santa Cruz, Texas, USA) in the presence of liposome 2000 (Invitrogen, California, USA). Brie y, Liposome diluted to 5 µl/well, and then siRNA was diluted at a calculated concentration. We incubated them for 5 minutes separately, and incubated the mixture proportioned for 15 minutes. The mixture was added to the cells and the 6-well plate placed into the incubator. After 8 h, the growth medium was changed to fresh media, and the cells were further incubated at 37°C for 24 h (5% CO 2 ). The cells were subsequently collected for further analyses.

Wound healing
We inoculated approximately 3 × 10 5 cells/well into a 6-well plate, with the exact number varying on a cell-to-cell basis. The cells were intervened with siRNA separately. The next day, the plates were scratched vertically using a pipette tip. Scratches were ensured to be straight using a ruler. The cells were washed with PBS. The scratched cells were then removed and a serum-free medium was added into the well. The plates were subsequently incubated at 37°C and 5% CO 2 . Samples were collected at 0, 24, and 48 h, and images were then captured.
Transwell assay Matrigel(8-12 mg/ml) was placed, which was stored at -20 ˚C, onto the ice to melt (keeping the temperature at 2 ˚C 8 ˚C). Matrigel (100 µl) was aspirated with a pipette tip precooled on ice and expelled into precooled 3000 µl serum-free medium (diluted at 1:30-1:40, total volume diluted according to the number of wells), and mixed thoroughly. The diluted 100 µl Matrigel above was added into the upper chamber of the Transwell plate to cover the whole polycarbonyl ester membrane. The Matrigel was polymerized into gelatin at 37°C for 60 minutes. Any remaining liquid medium on the top was discarded carefully.
The pretreated groups of cells were prepared into single-cell suspension (5 ×10^5 cells/ml) with serumfree medium in the regular way. This suspension (200 µl) was added to the upper chamber of the Transwell culture plate, and 700 µl 10% FBS was added to the lower chamber. The plate was then incubated at 37℃ and 5% CO 2 for 48 h. Subsequently, the cells were xed in 4% paraformaldehyde for 30 min and then stained with 0.1% crystal violet for 3 min.
To measure the number of cells that were able to penetrate the lm in the upper chamber, the unpenetrated cells on the upper surface of the lm and the Matrigel in the chamber were gently wiped off with a wet cotton swab. Finally, images were captured, and the cells were counted.

Statistical analysis
GraphPad Prism v9.00 was used for statistical analyses, and differences were determined with Student's t-tests. P < 0.05 indicated statistical signi cance.

PYCR2 expression pattern in CRC
As evident from Fig. 1, starBase v3.0 project(https://starbase.sysu.edu.cn/index.php) showed that the expression level of PYCR2 in CRC tissues was distinctly higher than that in non-cancerous tissues (p= 6.0e−13), and the prognosis of patients with CRC with high expression levels of PYCR2 was worse than that of those with low expression levels of PYCR2 (p < 0.05). These results indicated that PYCR2 plays an important role in CRC and may be an oncogene that promotes CRC progression.

Target gene prediction and data screening
The analysis involving STRING, BioGRID, IntAct, and GeneMANIA led to the identi cation of 10, 64, 44, and 19 genes, respectively. In total, 137 PYCR2-related genes were identi ed (Table 1). To increase the accuracy of target gene prediction, we selected genes that overlapped in at least two of the four databases (Fig. 2), and ultimately, 36 potentially related genes were identi ed ( Table 2).

Functional analysis of the overlapping target genes
Through GO enrichment analysis, we studied the biological roles of the 36 aforementioned genes in CRC ( Table 3). The target genes of PYCR2 were mainly enriched in biological processes such as positive regulation of mitotic metaphase/anaphase transition, mitochondrial respiratory chain complex assembly, apoptotic signaling pathway, response to hydrogen peroxide, regulation of oxidoreductase activity, and mitochondrial gene expression. The apoptotic signaling pathway was one of the most enriched pathways. Prognostic signi cance of the overlapping target genes Of 36, the expression levels of four genes-AHNAK2, CDC23, TRMT61B, and DLD-were signi cantly altered in CRC tissues as compared with those in non-cancerous tissues (p < 0.05) (Fig. 3A). Further, the prognosis of these four genes was also signi cantly altered in CRC and non-cancerous tissues through ENCORI and Gepia (p < 0.05) (Fig. 3B). These data suggested that AHNAK2, CDC23, TRMT61B, and DLD have a certain effect on CRC and that PYCR2 may have an effect on CRC via them, though the speci c mechanism remains unclear.

Co-expression of PYCR2 and related genes
Our co-expression analysis (Table 4) revealed the existence of a close association between PYCR2 and TRMT61B as well as CDC23 in CRC.

Mutation of PYCR2 target genes in CRC
Due to the high frequency of some genes ampli cation in CRC, the role played by gene mutation may be greatly changed. We used cBioPortal to explore gene-speci c changes in TCGA CRC data (Fig. 4). The genes such as MYO1F, AHNAK2, TNFRSF10A, and CIC showed a high mutation frequency in CRC.

Comparison of prognosis between patients with and without gene mutations
Using the ICGC database, we determined the prognosis of patients with and without target gene mutations in CRC and found that CIC and MYO1F had a signi cant impact on the prognosis (p < 0.05).
Moreover, the overall survival curves of CDC16, DLD, IRF2, and FASTKD showed a signi cant change. However, p value was not statistically signi cant, which could be due to the insu cient samples size (Fig. 5). We believe that the mutation of these target genes is of critical signi cance for CRC patients.

PYCR2 promotes both proliferation and invasion in CRC
To explore the biological role of PYCR2 in CRC, we knocked down PYCR2 expression in HCT116, LoVo, and HT29 cells in vitro (p < 0.05) (Fig. 6A, B). Wound healing assay results indicated that the migration of HCT116, LoVo, and HT29 cells decreased when PYCR2 was knocked down (Fig. 6C). Further, transwell assay results validated that in comparison with the negative control group, cell invasion ability was inhibited after PYCR2 expression was knocked down in HCT116, LoVo, and HT29 cells (Fig. 6D).
Statistical analyses revealed that the number of migratory cells showed an obvious decrease after knocking down PYCR2 in HCT116, LoVo, and HT29 cells (Fig. 6E). Collectively, our data indicated that PYCR2 has a momentous role in the development process of CRC.

Discussion
CRC majorly contributes to cancer-related deaths globally, and its incidence rates are approximately 3fold higher in transitioned versus transitioning countries. However, low human development index countries have a higher mean case fatality rate while the difference in mortality rate is smaller. [1] Considering that traditional methods to treat CRC, including chemotherapy and radiotherapy, remain unsatisfactory, new methods, such as tumor-speci c targeted therapy, have emerged. Identifying novel targets has become an indisputable part of tumor-speci c targeted therapy. [20] These insights indicate that it is vital to nd an independent predictive factor to evaluate the prognosis of CRC. PYCR2 is closely related to oxidative stress and cell apoptosis; [8,21] it also plays a key role in tumor energy metabolism, and thus, PYCR2 appears to represent a new group of potential target molecules for tumor treatment. To comprehensively understand the relationship between PYCR2 and CRC prognosis, herein we performed bioinformatics analyses based on TCGA database, and veri ed our results in vitro using a wound healing assay and Transwell test.
To the best of our knowledge, we for the rst time report the abnormal expression of PYCR2 in CRC, which PYCR2 expression levels were notably higher in CRC tissues than in corresponding non-cancerous tissues, based on TCGA dataset validation. We herein predicted PYCR2 target genes (n = 36); overlapping target genes were detected by GO enrichment analysis, which revealed that the apoptotic signaling pathway was signi cantly enriched. Apoptotic signaling is essential to maintain a healthy balance between cell death and survival and to maintain genomic integrity. The imbalance between pro-and antiapoptotic proteins promotes tumorigenesis by reducing the apoptosis of malignant tumor cells. [22] The function of these target genes of PYCR2 in the biological process was investigated, and we found that the expression status of AHNAK2, CDC23, TRMT61B, and DLD in CRC was remarkably different from that in corresponding non-cancerous tissues; furthermore, a notable difference was present in overall survival rate between the high and low expression groups. AHNAK2 has been reported to promote migration, invasion, and epithelial-mesenchymal transformation of lung adenocarcinoma cells via the transforming growth factor-β/Smad3 pathway. [23] Moreover, TRMT61B and CDC23 were co-expressed in CRC (p < 0.05). TRMT61B is a type of methyltransferase. RNA methyltransferases are closely associated with oncogene transcription and expression. TRMT61B also participates in cancer cell proliferation and tumor initiation, progression, and metastasis. [24] CDC23 is a key regulator of mitotic process. [25] We accordingly analyzed the status of these genes in CRC. The mutation frequency of some genes, such as MYO1F, AHNAK2, TNFRSF10A, LRRK2, and CIC, was found to be high. We also analyzed gene mutations and prognosis in patients with CRC, and the results indicated that the prognosis of patients with MYO1F and CIC gene mutations was signi cantly different from that of those without mutations. In addition, the prognosis of patients with CDC16, DLD, IRF2, and FASTKD gene mutations was signi cantly different from that of those without mutations, which shows high importance of these genes in tumorigenesis. TNFRSF10A reportedly mediates endoplasmic reticulum stress-induced apoptosis in human lung cancer cells in a DDIT3 activation-dependent manner.
[26] As for MYO1F, mutant MYO1F has been found to alter the mitochondrial network and induce tumor proliferation in thyroid cancer. [27] To further determine the impact of PYCR2 on CRC prognosis, we knocked down PYCR2 expression. Wound healing and Transwell assay results showed that in comparison with the negative control group, cell invasion ability was signi cantly inhibited after PYCR2 expression was knocked down in HCT116, LoVo, and HT29 cells. We thus report that in vitro, PYCR2 plays a salient role participating in the invasion of CRC cells.

Conclusions
In this study, we used several online bioinformatics platforms and online tools to systematically analyze the expression of PYCR2 and related genes in CRC; further, we assessed their prognostic signi cance. For the rst time, we report that PYCR2 is a potential prognostic biomarker of CRC. Our ndings also shed light on the importance of the expression of PYCR2 and related genes, as well as the potential role of PYCR2-related pathways in the progression of various human cancers. Our analyses provide valuable insights into PYCR2 as a novel biomarker, presenting it as a potential therapeutic target for several human cancers and contributing to the translation of genomic knowledge into clinical practice. Further studies are warranted to elucidate the speci c role, detailed molecular mechanisms, and clinical signi cance of PYCR2 in tumor progression and prognosis.

Declarations
Ethics approval and consent to participate  Venn diagram of target genes predicted from STRING, BioGRID, IntAct, and GeneMANIA. Different colors represent different datasets.

Figure 3
Expression levels and prognosis-related information of selected genes in CRC and non-cancerous tissues.
(A) Four of the 36 identi ed genes are signi cantly altered both in terms of their expression and prognostic status (p < 0.05). Orange and purple represent CRC and non-cancerous tissues, respectively.
(B) Online Kaplan-Meier plotting tools were used to gather prognostic information. TRMT61B, CDC23, AHNAK2, and DLD were closely associated with survival rates (p < 0.05). Mutation frequency of genes. The mutation status of the 36 identi ed genes is shown. The genes such as AHNAK2, AHNAK, LRRK2, MYO1F, and CIC showed a high mutation frequency.