Prognostic value and potential molecular mechanism of the like-Sm gene family in early-stage pancreatic ductal adenocarcinoma.

BACKGROUND: The purpose of this study was to investigate the prognostic significance of like-Sm (LSM) genes in early pancreatic ductal adenocarcinoma (PDAC) and explore the potential molecular mechanism. The protein product of the LSM1 gene is also known as CASM and YJL124C, while that of the LSM4 gene is known as GRP and YER112W.
METHODS: Data from 112 patients attached to the Whipple surgery were collected from the TCGA database of clinical characteristics and survival data. The Kaplan-Meier method and the multivariate Cox proportional risk regression model were used to analyze the impact of LSM genes on outcomes in these 112 patients. We performed gene-gene interaction (GGI) and protein-protein interaction (PPI) analysis to probe interactions between LSM family genes. Bioinformatics techniques were applied to study the potential early-stage molecular mechanisms of LSM genes. Previously, only a few studies have explored the role and potential mechanisms of LSM1 in pancreatic tumor transformation, revealing possible links to transforming growth factor-β, altering the expression of MMP1, uPAR, and SerpinB5 to enhance invasion and metastasis in pancreatic cancer, and facilitating mRNA decapping and degradation. Gene set enrichment analysis (GSEA) also proved that LSM genes are associated with RNA splicing, RNA synthesis, and RNA decomposition, and they may indirectly cause carcinogenesis through other genes such as myc.
RESULTS: The results showed that LSM1 (adjusted P=0.004) and LSM4 (adjusted P=0.034) were associated with the prognosis of patients with PDAC, and patients with high expression levels of LSM1 (adjusted HR =2.338) or LSM4 (adjusted HR =1.803) tended to experience bad outcomes.
CONCLUSIONS: Our study revealed that LSM1 and LSM4 might be used as prognostic biomarkers in early PDAC. 2021 Translational Cancer Research. All rights reserved.

Introduction

About 90% of pancreatic cancers originate from the epithelium of pancreatic ducts, and pancreatic ductal adenocarcinoma (PDAC) is one of the most deadly pancreatic cancers (1). The mortality rate of PDAC is astonishingly high, with 333,000 new PDAC cases diagnosed worldwide every year, and although only accounting for 3.1% of all cancers, more than 300,000 new cases of PDAC die every year. Pancreatic cancer is now the fourth leading cause of cancer-related deaths in the United States, and the annual survival rate is only 7.7% (2,3). A total of 46,420 patients were diagnosed with PDAC in the United States in 2014; it is expected that by the end of 2020, the number of PDAC patients in the United States will double, and by 2030 PDAC is expected to be the second leading cause of cancer deaths worldwide, after lung cancer, indicating poor prognosis (1,4,5).

About 70% of PDAC deaths are caused by cancer metastasis (6), and this coupled with a lack of early warning symptoms and a highly malignant phenotype are the leading causes of high mortality in PDAC patients. Many patients are already at an advanced stage of cancer when diagnosed; thus, early detection and early treatment are crucial for defeating PDAC (1,6,7). Despite tremendous progress in recent years in surgery, radiotherapy, and chemotherapy, the prognosis for pancreatic cancer remains poor. Although systemic chemotherapy is an essential treatment for liver and peritoneal metastasis, it has little effect on advanced pancreatic cancer (8,9), and molecular targeted therapy has no significant impact on the survival of PDAC patients (5). Developing a new PDAC biomarker, verifying its role in the early detection of PDAC, and predicting the prognosis of PDAC patients are therefore urgent for enhancing survival (3).

There are 14 genes in the like-sm (LSM) gene family; in animals, LSMs generally exist in the form of heptameric complexes, including the LSM1-7 and LSM2-8 complexes, and they are closely related to splicing and cytoplasmic mRNA degradation (10). LSM genes were significantly overexpressed in pancreatic cancer compared with surrounding normal tissues, and they act as oncogenes promoting cancer transformation, proliferation, resistance, and metastasis (11). The mechanism by which LSM genes promote PDAC transformation remains unclear. Studies have shown that up-regulation of LSM-1 can promote the growth of pancreatic tumors in mice, while down-regulation of LSM-1 can have the opposite effect (12), and antisense RNA induction significantly inhibits the ability of tumor cells to form anchoring-independent colonies on soft agar (13).

Herein, we used the TCGA database and found that in patients expressing high levels of LSM1 and LSM4 in pancreatic cancer, overall survival time (OS) was significantly decreased. This could provide a new direction for the treatment of pancreatic cancer. LSM genes are not only related to pancreatic cancer, but also closely related to other cancers such as prostate cancer and breast cancer. Cox multivariate proportional risk regression analysis showed that LSM1 and LSM4 may be early pancreas predictors, and are the primary genes involved in the proliferation and metastasis of pancreatic cancer. However, at present, most studies on the LSM gene family have only explored the role of LSM1 in pancreatic tumor transformation and its possible mechanism. In the present work, we comprehensively examined the roles of all members of the LSM gene family in early-stage PDAC, as well as the potential molecular mechanisms. We present the following article in accordance with the MDAR reporting checklist (available at http://dx.doi.org/10.21037/tcr-20-3056).

Methods

Data mining and processing

PDAC patient data (therapeutic function and mRNA profile) were downloaded from the TCGA website (https:/cancergenome.nih.gov/, accessed April 20, 2017). Clinical features included gender, smoking history, histological grade, presence or absence of Whipple surgery and radiotherapy, etc. After downloading, the DESeq package in R (version 3.5.2; www.r-project.org) was used to normalize expression data. All data used in this study were downloaded from a public database, and no humans or animals were involved in this research. Hence, no approval from an ethics committee was required. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Bioinformatics analysis of LSM genes

Gene Ontology (GO) was used for functional annotation of LSM genes. Additionally, the Kyoto Encyclopedia of Genes & Genomes (KEGG) database was accessed to define related signaling pathways, using the DAVID program (https://david.ncifcrf.gov/home.jsp, version 6.8, accessed January 19, 2019) (14). The Corrplot package (version 3.6.1) was applied (www.r-project.org) to analyze the correlation coefficients of LSM genes. We also used a search tool to construct a protein-protein interaction (PPI) map by retrieving gene/protein interactions (https://string-db.org/, version 11, accessed 30 October 2019). All LSM genes were uploaded to the Homo sapiens database, and an interaction score greater than 0.6 was deemed significant. Finally, we used GeneMANIA to construct the gene-gene interaction (GGI) network (http://genemania.org/, accessed October 30, 2019) (15).

Survival analysis

To review the partnership between specific therapeutic variables and survival prognoses in patients with early PDAC, Kaplan-Meier log-rank tests were applied. The multivariate Cox proportional risk regression system integrates related factors to determine whether LSM genes are strongly correlated with patient OS. Survival analysis of a single gene, comprehensive effect survival analysis based on the results, clinical variables, and expression levels of LSM genes were analyzed in R. A nomogram was generated using an RMS kit. To evaluate the PDAC prognosis of LSM genes, the Kaplan-Meier log-rank test was employed, along with the multivariate Cox proportional probability regression model.

Prognostic signature construction

After performing survival analysis, we found that LSM1 and LSM4 were significantly correlated with the prognosis of PDAC. Based on this, a predictive model of gene expression levels was constructed using the following risk scoring formula: Risk score = expression of gene 1×β1+ expression of gene 2×β2+ ... expression of Gene n×βn (16,17), where βn is the coefficient of regression of the related gene Cox threat analysis. Depending on the median risk assessment, high-risk individuals and low-risk patients are categorized into different groups. The ROC survival package in R (18) was used to develop a time-dependent receiver operator characteristic curve (ROC) for survival assessment to evaluate the prognosis of high-risk to low-risk groups.

Gene set enrichment analysis (GSEA)

Whole-genome expression profile datasets were uploaded and corresponding grouping files expressed by LSM genes were subjected to GSEA (19). The molecular feature database MSigDB (20) was used to expand c2 and c5 databases to determine in which pathways LSM genes are enriched. The upper or lower parts of the phenotype-related sorted gene list comprises LSM genes on each chromosome. A sequence of genes was considered statistically meaningful if both false discovery score (FDR) <0.25 and P<0.05 criteria were met.

Statistical analysis

For statistical analysis, we used SPSS 22.0 (IBM) and R version 3.6.1 (https:/www.r-project.org/). LSM gene relative risk between high- and low-expression categories was calculated using hazard-ratios (HRs) and a 95% confidence interval (CI). When evaluating the association of LSM genes, P<0.05 was regarded as statistically significant, and the Pearson correlation coefficient was employed (21-23).

Results

Data acquisition

Clinical characteristics and transcription profiles of PDAC patients were obtained from the TCGA database. After excluding advanced-stage PDAC without Whipple surgery, data for 112 patients undergoing Whipple surgery were retained for further analysis. Gene information was entered into a gene expression profile (GEPIA, http://gepia.pku.cn/), default parameters were employed, and patients were divided into two groups based on the median expression values of genes. Expression of LSM1 and LSM4 in tumor tissues of patients with PDAC was significantly higher than in normal adjacent tissues (P<0.05), which indicates that LSM1 and LSM4 might be involved in tumorigenesis and development in PDAC ().

Figure 1

Expression of LSM in cancer tissues and pericarcinomatous tissue. (A) LSM1, (B) LSM2, (C) LSM3, (D) LSM4, (E) LSM5, (F) LSM6, (G) LSM7, (H) LSM8, (I) LSM10, (J) LSM11, (K) LSM12, (L) LSM12P1, (M) LSM14A and (N) LSM14B.

GO and KEGG enrichment analysis.

As shown in , GO terms and KEGG pathway analysis revealed that the LSM gene family was mainly enriched in systemic lupus erythematosus, spliceosome, histone mRNA metabolic process, via transesterification reactions, histone mRNA metabolic process, mRNA binding, and ribonucleoprotein complex categories.

Figure 2

KEGG pathway and GO term analysis of LSM genes. KEGG, Kyoto Encyclopedia of Genes and Genomes; GO, Gene Ontology; MF, Molecular function; BP, Biological process; CC, Cellular component.

Analysis of LSM family interactions at the gene and protein levels

A complex gene-gene connection network was constructed between LSM family genes, and genes were clustered together (). Meanwhile, to explore the biological functions of LSM genes, we also performed PPI analysis to identify proteins that interact with LSMs. We identified SNRPD, SNRPF, EDC3, ATXN2, ZNF473, DDX6 and SLBP as potential interacting partners (). The Pearson’s correlation analysis matrix diagram () shows that LSM1, LSM2, and LSM5 are closely related (correlation coefficient >0.5). LSM2, LSM3, LSM5, and LSM7 are also closely related (correlation coefficient >0.5). LSM4 and LSM7 is the most closely related pair, with a correlation coefficient of 0.73. The number in each grid of the correlation graph represents the correlation between corresponding genes.

Figure 3

GeneMANIA and STRING gene-gene interaction (GGI) and protein-protein interaction (PPI) association networks of LSM genes. KEGG, Kyoto Encyclopedia of Genes and Genomes; GO, gene Ontology; sTRING, search tool for the retrieval of interacting genes/proteins; MF, molecular function; BP, biological process; CC, cellular component.

Figure 4

Matrix graphs of Pearson’s correlation analysis of LSM genes.

Survival analysis of the LSM gene family

As shown in Table S1, histological ratings, guided molecular therapy, radiation therapy, and residual resection are all substantially related to OS, based on a summary of clinical data for each patient. We found that high expression levels of LSM1 (adjusted P=0.004, adjusted HR =2.338, 95% CI: 1.320–4.141), and LSM4 (adjusted P=0.034, adjusted HR =1.803, 95% CI: 1.046–3.109) is associated with bad outcomes ( and ).

Table 1

Prognostic value of LSM gene expression in PDAC OS based on the TCGA database

Gene expression	Events/total (n=112)	MST (days)	Crude HR (95% CI)	Crude P value	Adjusted HR (95% CI)	Adjusted P-value
LSM1
Low	30/56	593	1		1
High	39/56	498	1.681 (1.032–2.738)	0.035	2.338 (1.320–4.141)	0.004
LSM2
Low	37/56	486	1		1
High	32/56	603	0.918 (0.568–1.484)	0.728	0.707 (0.415–1.203)	0.201
LSM3
Low	30/56	614	1		1
High	39/56	473	1.638 (1.007–2.664)	0.044	1.298 (0.746–2.258)	0.356
LSM4
Low	32/56	603	1		1
High	37/56	473	1.752 (1.078–2.846)	0.022	1.803 (1.046–3.109)	0.034
LSM5
Low	31/56	614	1		1
High	38/56	473	1.997 (1.218–3.275)	0.005	1.370 (0.750–2.501)	0.306
LSM6
Low	35/56	607	1		1
High	34/56	470	1.475 (0.903–2.411)	0.119	0.951 (0.551–1.640)	0.855
LSM7
Low	36/56	518	1		1
High	33/56	511	1.239 (0.766–2.003)	0.382	0.892 (0.520–1.529)	0.677
LSM8
Low	35/56	518	1		1
High	34/56	517	0.954 (0.592–1.536)	0.845	0.800 (0.462–1.386)	0.427
LSM10
Low	38/56	476	1		1
High	31/56	592	0.757 (0.469–1.221)	0.252	0.767 (0.459–1.283)	0.313
LSM11
Low	33/56	511	1		1
High	36/56	592	0.940 (0.583–1.517)	0.8	0.954 (0.564–1.615)	0.861
LSM12
Low	28/56	607	1		1
High	41/56	481	0.762 (0.468–1.241)	0.273	0.981 (0.575–1.674)	0.943
LSM12P1
Low	29/56	627	1		1
High	40/56	458	1.472 (0.909–2.386)	0.114	1.411 (0.381–2.395)	0.202
LSM14A
Low	32/56	568	1		1
High	37/56	517	1.067 (0.660–1.726)	0.792	0.814 (0.479–1.385)	0.449
LSM14B
Low	35/56	517	1		1
High	34/56	592	0.994 (0.614–1.607)	0.979	1.322 (0.772–2.263)	0.309

Adjusted for histological grade, radiation therapy, radical resection, and targeted molecular therapy. MST, median survival time; OS, overall survival; PDAC, pancreatic ductal adenocarcinoma; TCGA, The Cancer Genome Atlas.

Figure 5

Kaplan-Meier survival curve analysis of the association between high and low expression levels of LSM genes and overall survival in patients with early-stage PDAC, generated using The Cancer Genome Atlas (TGCA) database. Overall survival curves for (A) LSM1, (B) LSM2, (C) LSM3, (D) LSM4, (E) LSM5, (F) LSM6, (G) LSM7, (H) LSM8, (I) LSM10, (J) LSM11, (K) LSM12, (L) LSM12P1, (M) LSM14A and (N) LSM14B. PDAC, pancreatic ductal adenocarcinoma.

Combined effect survival analysis of LSM genes.

Patients were classified into high and low expression classes based on the expression of multiple LSM genes, and Table the expression frequency is shown in . Group C was shown to have poor OS compared with Groups A, B and D (adjusted P=0.001; adjusted HR =4.205, 95% CI: 1.778–9.948; ). Using the Cox multivariable hazard regression model, regression coefficients of LSM1 and LSM4 were 0.719 and 0.346, respectively, and risk score = expression of (LSM1×0.719) + (expression of LSM4×0.346). The results were separated into different variables and classified as high-risk and low-risk groups by median cut off. The results showed that the high-risk level for patients with early-stage PDAC was positively correlated with a disappointing outcome (adjusted P=0.011; adjusted HR =2.038, 95% CI: 1.176–3.531; , and ). Time-based ROC analysis showed that prognostic signatures successfully predicted the consequence of PDAC in early-stage patients (1-year area under the curve [AUC] =0.679; 2-year AUC =0.657; 5-year AUC =0.644; ). The nomogram from this study also shows that LSM1 and LSM4 have a significant impact on the clinical outcomes of patients (). The proportions on nomogram rows represent the importance of each vector array, and the duration of row sections indicates how this element relates to the case.

Table 2

Joint survival analysis of LSM1 and LSM4 genes and OS for patients with early-stage PDAC

Group	LSM1	LSM4	Events/total (n=112)	MST (days)	Crude HR (95% CI)	Crude P-value	Adjusted HR (95% CI)	Adjusted P-value
A	Low	Low	17/35	596	1		1
B	Low	High	13/21	470	2.390 (1.124–5.082)	0.024	2.945 (1.270–6,826)	0.012
C	High	Low	15/21	603	2.169 (1.051–4.474)	0.036	4.205 (1.778–9.948)	0.001
D	High	High	24/35	481	2.418 (1.263–4.629)	0.008	3.327 (1.566–7.068)	0.002

Adjusted for clinical variables, neoplasm histological grade, targeted molecular therapy, radiation therapy and residual resection. HR, hazard ratio; MST, median survival time; PLCD, phospholipase C delta; OS, overall survival; PDAC, pancreatic ductal adenocarcinoma; CI, confidence interval.

Figure 6

Combined effects of LSM1 and LSM4 on the overall survival of patients with early-stage PDAC. (A) Overall survival curves for the combined effect of LSM1 and LSM4. Group A, low LSM1+ low LSM4; Group B, low LSM1+ high LSM4; Group C, high LSM1+ low LSM4; Group D, high LSM1+ high LSM1. (B) Overall survival curves of risk scores for LSM1 and LSM4. Low-risk group=low risk scores for LSM1+ LSM4; High-risk group=high risk scores for LSM1+ LSM4.

Table 3

Prognostic values of risk scores for LSM1 and LSM4 gene expression in PDAC OS based on the TCGA database

Risk scores	Events/total (n=112)	MST (days)	Crude HR (95% CI)	Crude P-value	Adjusted HR (95% CI)	Adjusted P-value
Low	32/56	607	1		1
High	37/56	473	1.932 (1.183–3.154)	0.007	2.038 (1.176–3.531)	0.011

Adjusted for histological grade, radiation therapy, radical resection, and targeted molecular therapy. MST, median survival time; OS, overall survival; PDAC, pancreatic ductal adenocarcinoma; TCGA, The Cancer Genome Atlas.

Figure 7

Combined effect of LSM1 and LSM4 on the overall survival of patients with early-stage PDAC. (A) From top to bottom; risk score plot, survival status scatter plot and heat map of the expression levels of LSM1 and LSM4 in low- and high-risk groups. (B) ROC for predicting 1-, 2-, and 5-year events (mortalities) combining clinical data with LSM1 and LSM4 expression. PDAC, pancreatic ductal adenocarcinoma. (C) Nomogram predicting the impact of various clinical variables on the prognosis of patients. The proportions on nomogram rows represent the importance of each vector array, and the duration of row sections indicate how this element relates to the case.

GSEA

Our study proved that LSM1 and LSM4 are significantly correlated with OS for patients with PDAC. Risk scores for LSM1 and LSM4 were combined and stratified according to the median. The GSEA results showed that the oncogenic mechanisms of LSM genes might include facilitating mRNA decapping, mRNA degradation, and mRNA splicing ().

Figure 8

GSEA results for c2 and c5 reference genes for the group with elevated LSM1 and LSM4 expression levels. GSEA, gene set enrichment analysis; ES, enrichment score; FDR, false discovery rate.

Discussion

The pancreatic tumor microenvironment comprises immune cells and cancer-associated fibroblasts (CAFs), which cause fibrosis of the interstitial connective tissue, and combined with the lack of angiogenesis in the pancreatic tumor, this results in a persistent and severely hypoxic tumor microenvironment (24,25). PDAC is a highly malignant tumor of the digestive system; the prognosis of patients with PDAC is extremely poor, and many patients are already at an advanced stage of cancer when diagnosed, thus early detection and early treatment are crucial for defeating PDAC (26). The standard diagnostic methods for PDAC included ultrasonography, computed tomography, positron emission tomography, magnetic resonance imaging, and endoscopic ultrasonography. However, because of the deep position of the pancreas, factors such as obesity and intestinal gas, and even the operator’s experience, can affect the accuracy of the diagnosis. Therefore, new biomarkers are urgently needed for detection of pancreatic cancer in the early stages (27). Existing markers with potentially important diagnostic value include carbohydrate antigen 19-9 (CA19-9), circulating tumor cells (CTCs), cell-free circulating tumor DNA (ctDNA), epigenetic biomarkers such as DNA methylation, histone modification, and chromatin structure, complement component 4 binding-protein a-chain (C4BPA), neutrophil gelatinase-associated lipocalin (NGAL) in urine, and others (28). However, CA19-9 is insufficient as an independent diagnostic tool; for instance, serum CA19-9 elevation can occur in patients with benign diseases such as chronic or acute pancreatitis, cholangitis, obstructive jaundice, liver cirrhosis, or other malignancies such as gastrointestinal cancers (29). CTCs are derived from primary tumors or metastatic sites and circulate in the bloodstream, and portal vein blood CTC counts but not peripheral blood CTC counts are correlated with OS (30). Neutrophil gelatinase-associated lipocalin requires numerous samples and multi-center clinical trials to identify and verify urinary diagnostic biomarkers of PDAC (28). PDAC is related to the oncogene K-ras, and downregulation of p53, p16/CDK2, and DPC4 (9).

LSM genes are significantly more highly expressed in pancreatic cancer than in surrounding normal tissues (11).LSM proteins are ubiquitous in eukaryotes, where they are present in both the nucleus and cytoplasm, and characterized by the Sm domain. LSMs are usually present as hexa- or heptamethyl RNA-binding complexes, forming two stable heptamer rings in LSM1-7 and LSM2-8 complexes, and their nuclear and cytoplasmic localization is tightly regulated. The two sturdy heptamer rings perform different functions; U6 snRNP contains an LSM2-8 heptamer, consisting of LSM2, LSM3, LSM4, LSM5, LSM6, LSM7, and LSM8, that resides in the nucleus and binds to the 3’ end of U6 snRNA and maintains its stability; meanwhile, the LSM1-7 heptamer is present in the cytoplasm and is involved in the degradation of mRNA (9,10,31-38). LSM proteins are involved in the occurrence and transformation of pancreatic cancer, and they are also related to changes in circadian rhythms in plants and mammals (9,39).

Using the TCGA database, we found that LSM genes are highly expressed in patients with pancreatic cancer. High expression levels of LSM1 and LSM4 were closely related to the prognosis of patients with PDAC. In pancreatic cancer patients with increased expression of LSM1 and LSM4, OS was significantly decreased. The multivariate Cox proportional risk regression model found that LSM1 and LSM4 may be early predictors of pancreatic cancer, and are essential target genes involved in the occurrence, proliferation, and metastasis of PDAC. Some previous studies found that the LSM1 gene is not only closely related to pancreatic cancer, but it could also provide a breakthrough in the treatment of this disease. However, only a few studies have explored the role of LSM1 in pancreatic tumor transformation and its possible mechanism, and not all members of the LSM gene family have been investigated. In the present work, we comprehensively explored the roles of all members of the LSM gene family in early-stage PDAC, and probed their potential molecular mechanisms.

Research has shown that LSM1 is essential for the transformation phenotype and anchorage-independent growth of pancreatic cancer cells; it can facilitate mRNA decapping and degradation, and thereby accelerate the cell cycle (40). In vitro experiments showed that LSM1 overexpression could induce the growth and cloning of pancreatic tumor cells. The possible molecular mechanism may involve LSM1 overexpression affecting the expression of various cancer-related pathways related to apoptosis and invasion, including transforming growth factor-β, tumor necrosis factor, and vascular endothelial growth factor, resulting in altered expression of Slug, MMP1, uPAR, and SerpinB5, all of which are known to be associated with invasion and metastasis in pancreatic cancer (11). Our GSEA results showed that LSM genes were related to RNA splicing, RNA synthesis, and RNA decomposition, and they may also indirectly cause carcinogenesis through other genes such as myc. Numerous studies have shown that upregulating MYC is a significant promoter of tumorigenesis, whether caused by other oncogenes (such as RAS mutations) or by MYC itself, and cell growth is correlated with MYC expression levels (41-43).

Progesterone can inhibit cancer cell cycle progression in PDAC and pancreatic neuroendocrine tumors (44-48). Additionally, patients with progesterone receptor-negative pancreatic solid pseudopapillary (SPNP) tumors have poor prognosis (49). Our current GSEA results confirmed that LSM3, LSM4, LSM7, and LSM8 might function by inhibiting progesterone. EZH2 expression is also upregulated in pancreatic cancer tissues, which can promote the proliferation and metastasis of pancreatic cancer cells (50-52), and our GSEA results indicate that LSM7 may play its role through EZH2.

The most important mechanism by which LSM proteins promote the development of cancer is through telomerase. A reverse transcriptase enzyme responsible for telomere elongation in cells is synthesized using a portion of the RNA subunit as a template, which can offset the loss of DNA sequences at the ends of chromosomes, since shortened telomeres have limited ability to replicate, and this enhances the proliferation capacity of cells in vitro (53-56). Studies have shown that many cancer cells display high telomerase activity in most eukaryotes, and the LSM2-8 complex can bind to the telomerase RNA subunit (TER1) of schizotypal yeast, and thereby protect mature TER1 and 3’ termini from external degradation, promoting the proliferation and malignant transformation of tumor cells (57).

Many patients are already at an advanced stage of cancer when diagnosed, and existing treatment methods such as chemotherapy, surgery, and radiation do not significantly improve the OS of patients with PDAC (58). The traditional standard treatment for PDAC is gemcitabine, and the most commonly used first-line treatment options are FOLFIRINOX (oxaliplatin + irinotecan +5-fluorouracil + leucovorin) or gemcitabine combined with nanoparticle albumin-bound (nab)-paclitaxel, all of which have limited efficacy in patients with advanced pancreatic cancer (59). Immunotherapy is useful for the treatment of leukemia and melanoma; it can activate host immune responses to identify and eliminate tumor cells. However, PDAC has a rich fibrotic stroma, and it lack lacks antigen-experienced T effector cells, hence immunotherapy is not effective for pancreatic cancer (60). The latest research shows that heat shock protein-90 can inhibit PSC/CAF in vitro and enhance the anti-PD-1 blocking effect in vivo (61). Additionally, in mouse experiments, combining anti-PD-1 and vaccine can have good anti-tumor activity; anti-CD137 agonist antibody combined with anti-PD-1 and vaccine significantly improved the survival time of the mouse PDAC model, highlighting a potential treatment option (62). Patients who received neoadjuvant and allogeneic PDAC vaccine tended to exhibit more prolonged median OS (63). Encouragingly, LSM1 antisense gene therapy can disrupt the tumor cell cycle, and this is very effective for pancreatic cancer. The growth of Panc02 tumors was significantly inhibited by injecting adenovirus-based LSM1 antisense RNA (Ad-Alsm1) into a subcutaneous tumor model (9,11). Similarly, the systematic application of Ad-ALSM1 in metastatic tumor models can substantially reduce tumor progression and prolong patient survival (11,40,64). LSM1 is also highly expressed in lung cancer, prostate cancer, breast cancer, and pancreatic endocrine tumors at both mRNA and protein levels. Antisense LSM1 transfection may reduce tumor cell proliferation via cyclin B1 and CDK1 proteins (40,65-69). Additionally, GSEA results in the present work indicate that LSM14B might help promote breast cancer progression through the ZNF217 gene.

The present study has some limitations. Firstly, the sample size was too small, and some patients are still being followed up. Thus, the probability of false-negative results is high, and more specimens are needed for further verification. Secondly, the clinical information is not comprehensive enough, and neither are the results. Thirdly, performed analysis only at the transcriptional level. Nonetheless, we discovered new ties between LSM genes and predictions of early PDAC, but further investigation is needed to confirm these results and the fundamental molecular mechanisms.

Despite these limitations, an association between LSM genes and early PDAC patient prognosis was demonstrated for the first time. Additionally, GSEA was used to assess the possible molecular mechanisms by which LSM genes may influence the prognostics of patients with early PDC. Moreover, LSM genes are likely to become new targets for PDAC therapy if future studies confirm these observations.

Conclusions

In conclusion, high expression levels of LSM1 and LSM4 is related to poor prognosis in early-stage PDAC patients, and these genes may serve as potential prognostic biomarkers and therapeutic targets in early-stage PDAC.