A Systematic Study On The Involvement Of Myosin Family In The Development Of Pancreatic Cancer Using Artificial Intelligence Algorithm

BackgroundPancreatic cancer has a 5-year survival rate of only about 11%.Due to the insidious onset and the lack of early diagnosis methods,the prognosis of patients is inferior.Over the past few decades,with the development of science and technology,the mortality rate of many types of cancer has shown a downward trend.However,the mortality rate of pancreatic cancer patients has not improved significantly,mainly because most patients are diagnosed in an advanced stage.The latest research shows that the median survival of patients diagnosed early with pancreatic cancer is nearly ten years.Therefore,early diagnosis and treatment of pancreatic cancer are particularly critical.In the application of early diagnosis,medical scientists try to quickly locate high-risk groups based on biomarkers,pathological examinations,medical images,and other data through artificial intelligence technology and then screen pancreatic cancer lesions early.However,the application of single-center data and only one single algorithm limits the accuracy and robustness of the model.Therefore,it is urgent to train accurate,efficient,and robust early diagnosis models in multi-center databases through integrated algorithms and find core markers for early diagnosis.Myosin is a superfamily of molecular motors,and myosin acts as an energy source to hydrolyze ATP to ADP,creating mechanical forces that travel along the actin filament.Myosin plays an essential role in the development of various cancers,and different myosin families exist in the form of cancer-promoting genes or tumor suppressor genes.They are involved in tumor invasion,metastasis,cell polarity,and immune microenvironment formation.In addition,they are prognostic biomarkers for various solid tumors such as prostate cancer,melanoma,and lung cancer.Myosin Light Chain 12B(MYL12B),as a new member of the myosin family,has not yet been studied to clarify whether it is related to the development of pancreatic cancer.Our previous studies found an enrichment of MYL12B in patients at high risk of pancreatic cancer.Based on this phenomenon,this study aims to construct an early diagnosis model of pancreatic cancer based on the myosin family through artificial intelligence technology,clarify the critical role of the myosin family in the occurrence and development of pancreatic cancer,explore the early diagnosis,tumor microenvironment formation,invasion and metastasis characteristics and prognosis evaluation value of myosin family and model core gene MYL12B in pancreatic cancer patients,provide potential targets and immune checkpoints for the diagnosis and treatment of pancreatic cancer,and provide more accurate individualized treatment for patients with different genotypes.PurposeThis study aims to use artificial intelligence algorithms to construct an early diagnosis model of pancreatic ductal adenocarcinoma based on myosin family expression,explore the role of myosin family core genes in the formation of the pancreatic cancer tumor microenvironment,and then explore the biological function and prognostic significance of myosin in pancreatic cancer.Methods1.Transcriptomics data of PD AC patients were extracted from TCGA,GTEx,GEO,and ICGC databases.Three dichotomous machine learning algorithms,namely support vector machine,random forest,and artificial neural network,were used to construct an ensemble learning pancreatic cancer early diagnosis model.The ROC curve was used to verify its efficacy.Then,based on the unsupervised consensus clustering algorithm,pancreatic cancer patients were divided into two subgroups,and high-and low-risk groups were defined according to the prognosis of the subgroups.The ESTIMATE algorithm evaluated the relationship between core gene expression and tumor microenvironment.Then CIBERSORT algorithm was used to analyze the immune infiltration between the pancreatic cancer group,the normal group,and the pancreatic cancer subgroups to find out the potentially highly correlated immune cell species of the core genes of the model.The expression of different immune checkpoints in the high-and low-risk subgroups was evaluated to determine the applicable immune checkpoints for high-risk groups and observe the immunotherapy response of different immune checkpoints.Use functional enrichment analysis to elucidate potential biological processes,cellular components,molecular functions,and pathways in high-and low-risk subgroups.2.Observe the expression of the core gene MYL12B in pancreatic cancer cell lines.Observe its expression by transient knockdown,the types of co-expression gene changes,and the changed biological processes,cell components,molecular functions,and action pathways.To assess whether MYL12B affects the characteristics of pancreatic cancer cell invasion and migration.The stable cell line with knockout MYL12B was constructed,pancreatic transplantation tumors were implanted in situ,and the size of transplanted tumors in situ and the number of liver and spleen metastases were calculated.We used immunohistochemical staining to evaluate the expression of MYL12B in human pancreatic cancer tissues and adjacent pancreatic tissues.The prognostic significance of MYL12B expression and pancreatic cancer patients was discussed.The survival of patients with different expressions of MYL12B was observed,and the clinicopathological features highly related to prognosis and MYL12B expression were evaluated.Results1.The artificial intelligence algorithm was used to construct an early diagnosis model of pancreatic cancer based on myosin family expression and to elucidate its immune microenvironment characteristicsEighty human-related genes of the myosin family were obtained from the GeneCards database.Four core myosin family genes were obtained through three dichotomous machine learning algorithms of support vector machine,random forest,and artificial neural network to construct a Pancreatic Ductal Adenocarcinoma Early Detection(PDAC-ED)model and the area under the model curve was verified as in multiple databases between 83.8-99.2%.The core genes MYL12B and PPP1CB were significantly higher in pancreatic cancer tissues than in normal samples,while MPRIP and PPP1R12C were expressed down-regulated in pancreatic cancer tissues.In the tumor microenvironment analysis,the stromal fraction of patients in the highexpression group of MPRIP and PPP1CB was higher than that of the low-expression group,and the purity of tumors was lower.In contrast,patients with high expression of MYL12B had significantly lower stromal fraction and immune fraction,and their tumors had higher tumor purity.Immune infiltration analysis showed that memory resting CD4+ T cells and resting dendritic cells were significantly elevated in PDAC patients and high-risk groups,while CD8+T cells and activated NK cells were significantly downregulated.There were differential expressions of CD47,JAK1,and PDCD1 in the high-and low-risk subgroups,and the three checkpoints were highly expressed in pancreatic cancer tissues.Pancreatic cancer patients with high expression of JAK1 or PDCD1 had better responses in immunotherapy with anti-CTLA-4 combined with anti-PD1.DO analysis showed that high-and low-risk subgroup differential expression genes were mainly enriched in malignant tumor diseases.In GO analysis,biological processes were enriched in endosomal transport,leukocyte migration,leukocyte proliferation,etc.Cell components are enriched at the cell leading edge,cell-substrate junction,coated vesicle membrane,etc.The molecular function is enriched in cadherin binding,protease binding,ubiquitin-binding,etc.KEGG analysis showed that 13 malignant features were enriched in the high-risk subgroup,such as focal adhesion,ECM-receptor interaction,PI3K-Akt signaling pathway,cell adhesion molecules,actin cytoskeleton regulation,etc.2.Myosin MYL12B is involved in pancreatic cancer invasion and metastasis and mediates the poor prognosis of pancreatic cancer patientsThe expression of MYL12B in pancreatic cancer cell lines was higher than that of normal pancreatic epithelial cell lines.After knocking down MYL12B,the migration and invasion ability of pancreatic cancer SW1990 and AsPC-1 cell lines decreased significantly,and the wound healing time increased.The GO and KEGG results were consistent with the abovementioned characteristics of mediating cell motility.After knockout MYL12B,the weight of pancreatic transplant tumors was reduced,and the number of hepatic and splenic metastases decreased.Immunohistochemical staining showed that the expression of MYL12B in human pancreatic cancer tissues was significantly higher than in adjacent tissues.The prognosis of patients with high expression of MYL12B was worse.Univariate and multivariate regression analysis showed that MYL12B expression was significantly correlated with the prognosis of pancreatic cancer and was an independent prognostic risk factor.Correlation analysis of clinicopathological features revealed that high expression of MYL12B was associated with non-elderly patients,higher T stage,and lymph node metastasis.Conclusion1.The PDAC-ED model constructed based on an artificial intelligence algorithm can be used as an effective tool for early diagnosis of pancreatic cancer.The expression of the core genes MYL12B and PPP1CB in pancreatic cancer tissues was up-regulated,while MPRIP and PPP1R12C were down-regulated.MPRIP and PPP1CB highexpression groups had higher matrix fractions and lower tumor purity.The matrix fraction and immune fraction of the MYL12B high-expression group were significantly reduced,and the tumor purity was higher.The regulation of the tumor immune microenvironment by the core gene of the model was related to downregulating CD8+T cells and activated NK cells and was highly correlated with the three immune checkpoints of CD47,JAK1,and PDCD1.2.The expression of MYL12B in pancreatic cancer cell lines and tissues was significantly higher than that of normal control.Knockdown of MYL12B can affect the invasion,migration,and wound healing ability of pancreatic cancer cells in vitro.Knockout of MYL12B can reduce tumorigenesis weight in pancreatic cancer and inhibit hepatosplenic metastasis in vivo.High expression of MYL12B is closely related to poor prognosis in patients with pancreatic cancer and is an independent risk factor,and is associated with non-elderly patients,higher T stages,and lymph node metastasis.