KR20230168575A - Biomarker for prognosis of squamous pancreatic cancer and uses thereof - Google Patents

Abstract

The present invention relates to a composition and kit for diagnosing squamous pancreatic cancer using a pancreatic cancer biomarker, and a method of providing information for diagnosing squamous pancreatic cancer using the same. The present invention provides a biomarker that can predict squamous pancreatic cancer. By providing a composition and kit for diagnosing squamous pancreatic cancer, patients with squamous pancreatic cancer can be diagnosed quickly and accurately, and appropriate treatment can be provided to patients with squamous pancreatic cancer.

Description

Biomarker for prognosis of squamous pancreatic cancer and uses thereof}

The present invention relates to a diagnostic composition and kit for the diagnosis and treatment of squamous pancreatic cancer, and a method of providing information for the diagnosis of squamous pancreatic cancer using the same.

Pancreatic cancer is a disease in which a tumor develops in the pancreas, which is responsible for the secretion of digestive enzymes and insulin, located in the upper abdomen between the tip of the solar plexus and the navel. The tumor occurs in the endocrine cells that secrete hormones such as insulin and is related to the secretion of digestive enzymes. It can be divided into tumors originating from exocrine cells.

Causes of pancreatic cancer include smoking, high-calorie diet, chronic pancreatitis, and genetic factors. Since pancreatic cancer has no early symptoms, the early detection rate is very low, less than 10%. As the disease progresses, symptoms such as abdominal pain, jaundice, decreased appetite, and weight loss may appear, as well as gray stool, pain after eating, vomiting, and nausea.

Pancreatic cancer is diagnosed through MRI and CT, and if it is suspected but not found, it is diagnosed through endoscopy and endoscopic ultrasound. When treated after detection, chemotherapy, surgical resection, drug therapy, and radiation therapy are performed.

If pancreatic cancer is a malignant tumor, the average survival time is several months and the 5-year survival rate is less than 5%. There are several types of pancreatic cancer, but more than 90% occur in glandular cells of the pancreatic duct, so pancreatic cancer generally refers to pancreatic ductal adenocarcinoma (PDA).

The initiation and progression of pancreatic cancer require repetitive genomic changes and involve a wide range of intra- and inter-tumor heterogeneous PDA transcripts. This transcriptional diversity can be stratified into two or more major molecular subtypes (Bailey et al., 2016; Chan-Seng-Yue et al., 2020; Collisson et al., 2011; Moffitt et al. , 2015). The subtypes of pancreatic cancer are largely basal-like/squamous/quasi-mesenchymal (hereinafter referred to as ‘basal-like subtype’) and classic. It can be divided into classical/progenitor (hereinafter referred to as 'classical subtype'), and morphological characteristics, prognosis, and response to chemotherapy differ depending on the subtype (Aung et al ., 2018; Bailey et al., 2016; Brunton et al., 2020; Moffitt et al., 2015). It is known that these pancreatic cancer subtypes are determined by functional gain or loss of transcriptional regulators. For example, increased expression of transcription factors including the ΔN isoform of TP63 (DNP63), MYC, and GLI2 is important for maintaining basal-like subtype identity in pancreatic cancer cells (Adams et al., 2019; Bailey et al., 2019). al., 2016; Milles et al., 1999; Somerville et al., 2018), and reduced expression of HNF4A and GATA4/6, characteristic of the classical subtype, is associated with the basal-like subtype (Brunton et al., 2020 ; Morrisy et al., 1998). Additionally, loss-of-function mutations in chromatin-modifying enzymes, including KDM6A, KMT2C, and KMD2D, are believed to contribute to determining squamous pancreatic cancer, and mutations in TP53 are also seen in squamous pancreatic cancer (Andricovich et al., 2018; Bailey et al., 2018). al., 2016).

The gene expression characteristics of these pancreatic cancer subtypes are used as biomarkers or treatment targets to classify or diagnose the subtype. However, because these expressed genes or proteins are overlappingly expressed in multiple subtypes, there are limitations in specifically diagnosing subtypes through them, and the reality is that there are no biomarkers that can complement or replace them. Therefore, in order to specify the subtype of pancreatic cancer through molecular biology and select the optimal treatment method to maximize the patient's survival rate and survival period, research and development of biomarkers that can objectively and quantitatively predict the subtype of pancreatic cancer are necessary. It is being demanded.

Biomarker diagnosis is a diagnostic method that uses indicators to detect changes in the body using cells, blood vessels, proteins, and DNA in the body. It is a diagnostic method that uses proteins, vestibules, cells, DNA, and viruses. Diagnosis is made through bacteria, etc. Biomarkers are attracting attention as a method for personalized medicine and preventive medicine.

Korean Patent No. 10-2356676 Korean Patent No. 10-2289278

The purpose of the present invention is to provide a composition for diagnosis and prevention of squamous pancreatic cancer using biomarkers.

The purpose of the present invention is to provide a kit for diagnosis and prevention of squamous pancreatic cancer using biomarkers.

Additionally, the purpose of the present invention is to provide a method of providing information for diagnosis and prevention of squamous pancreatic cancer using biomarkers.

One aspect of the present invention provides a composition for diagnosing or predicting prognosis of squamous pancreatic cancer, including an agent for measuring the expression level of the CD109 gene or protein.

“Pancreatic cancer” as used in the present invention is divided into endocrine tumors (occurring in endocrine cells that secrete hormones such as insulin) and exocrine tumors (occurring in exocrine cells related to the production and secretion of digestive enzymes) depending on the location and origin. 90% of pancreatic cancer is pancreatic ductal adenocarcinoma, which occurs in the exocrine cells of the pancreatic duct and commonly occurs in men in their 60s to 80s.

These pancreatic cancers or pancreatic adenocarcinomas have various molecular subtypes, among which are enriched TP53 and KDM6A mutations, upregulated TP63ΔN, and genes that determine endodermal identity in the pancreas (e.g. PDX1, MNX1, GATA6). The subtype characterized by hypermethylation and downregulation of HNF1B) is classified as the squamous subtype. The prognosis for patients with squamous pancreatic cancer is poor compared to other subtypes.

In one aspect of the present invention, the pancreatic cancer is exocrine pancreatic cancer. It may be adenocarcinoma, squamous cell carcinoma, adenosquamous cell carcinoma, collagenous carcinoma, neuroendocrine pancreatic cancer, pancreatic neuroendocrine tumor, gastrinoma, insulinoma, glucagonoma, benign cell carcinoma, or cyst, and specifically, squamous pancreatic cancer.

As used herein, “diagnosis” means confirming the presence or characteristics of a pathological condition in an individual who has not yet been diagnosed or has been diagnosed. Diagnosis in the present invention may be to confirm whether pancreatic cancer is squamous using a biomarker.

Methods for diagnosing pancreatic cancer of the present invention may include serum tumor markers, ultrasound, computed tomography (CT), magnetic resonance imaging, endoscopic retrograde cholangiopancreatography, endoscopic ultrasound, and positron emission tomography. It is a diagnostic method using biomarkers.

“Prognosis” as used in the present invention refers to determining whether an individual has recurrence, metastasis, drug responsiveness, resistance, etc. before/after treatment for an individual who has not yet been diagnosed or has been diagnosed. In the present invention, the overall survival rate can be predicted by using biomarkers to predict whether patients with squamous pancreatic cancer will have a good future survival prognosis.

Here, “biomarker” is a substance that is generally detectable in biological samples and includes all organic biomolecules such as polypeptides, proteins, nucleic acids, genes, lipids, glycolipids, glycoproteins, and sugars that can detect biological changes. In the present invention, the CD109 gene or protein can be used as a biomarker for diagnosing or predicting prognosis of squamous pancreatic cancer.

The present invention provides a method for producing and producing biomarkers by obtaining mouse pancreatic cancer cells.

In one embodiment of the present invention, mouse pancreatic cancer cells were grown on ultra-low attachment plates, and single-layer and multi-layer cancer cells were cultured.

In one embodiment of the present invention, the mouse pancreatic cancer cell line may be FC1199, NB390, UN-KC-6141, UN-KPC-960, or UN-KPC-961, but specifically, it is FC1199.

In one embodiment of the present invention, the human pancreatic cancer cell line may be QGP-1, SUIT-2, HUP-T4, DAN-G, or T3M-4, but specifically SUIT-2.

As an example of the present invention, FC1199-Par and -Sph were subcutaneously injected into mice and tumor growth and metastasis were confirmed.

In one embodiment of the present invention, the tumor metastasizes to the liver, lungs, lymph nodes, peritoneum, and bone, thereby creating cancer.

In one example of the present invention, the reduction of pancreatic cancer was confirmed through down-regulation of TEAD2 and compared with the control group.

In one embodiment of the present invention, the control group for reduction of pancreatic cancer confirmed the reduction when drugs were used, and the drugs included 5-fluorouracil, irinotecan, leucovorin, oxaliplatin, gemcitabine, Abraxane, fluoropyrimidine, There are capecitabine, exatecan, and vertepolpine, and specifically gemcitabine.

CD109 is a glycosylphosphatidylinositol (GPI)-anchored cell surface protein involved in regulating tumor growth factor beta (TGF-β) signaling, mainly used in CD34+ acute myeloid leukemia cell lines, T cell lines, activated T lymphoblasts, and endothelium. It is known to be expressed in cells and activated platelets, but nothing is known about its relationship with pancreatic cancer or squamous pancreatic cancer.

In order to detect the CD109 gene or protein, an agent that measures the expression level of each gene or protein is required.

According to one embodiment of the present invention, the agent is one selected from the group consisting of a primer pair, probe, antisense nucleotide, antibody, oligopeptide, ligand, PNA, and aptamer that is complementary to or specifically binds to the CD109 gene or protein. It may be more than that.

As used in the present invention, “primer” is a nucleic acid sequence with a short free 3-terminal hydroxyl group, a single-stranded oligomer that can form base pairs with a complementary template and serves as a starting point for copying the template strand. It means nucleotide (single strand oligonucleotide). The primers can initiate DNA synthesis in the presence of four different nucleoside triphosphates and a reagent for polymerization (i.e., DNA polymerase or reverse transcriptase) at an appropriate buffer solution and temperature. The primer pair consists of sense and antisense oligonucleotides with a sequence of 7 to 50 nucleotides, and contains 15 to 30 nucleotides at a level that does not change the basic properties of the primer that serves as the starting point for DNA synthesis. It may have a nucleotide sequence.

As used herein, “probe” refers to a linear oligomer of natural or modified monomers or linkages, includes deoxyribonucleotides and ribonucleotides, and specifically hybridizes to the target nucleotide sequence. It can exist naturally or be artificially synthesized.

These primers, probes and antisense nucleotides may, if desired, contain labels detectable directly or indirectly by spectroscopic, photochemical, biochemical, immunochemical or chemical means. The detectable label is a label material capable of generating a detectable signal, and may include a fluorescent substance such as Cy3, Cy5, etc. The detectable label can confirm the hybridization results of nucleic acids.

“Antibody” as used in the present invention refers to a specific protein molecule directed to an antigenic site. In the present invention, it refers to an antibody that specifically binds to each protein, and includes all monoclonal antibodies, polyclonal antibodies, and recombinant antibodies. Here, “specifically binding” means that the binding ability to the target substance is superior to that of other substances to the extent that the presence of the target substance can be detected by binding. Additionally, the antibodies include intact forms with two full-length light chains and two full-length heavy chains, as well as functional fragments of the antibody molecule. A functional fragment of an antibody molecule refers to a fragment that possesses at least an antigen-binding function and may be Fab, F(ab'), F(ab')2, Fv, etc.

The antibody can be easily produced through techniques known in the art. For example, monoclonal antibodies can be prepared using the hybridoma method (Kohler and Milstein (1976) European Jounral of Immunology 6:511-519), which is well known in the art, or a phage antibody library (Clackson et al, Nature, 352:624-628, 1991; Marks et al, J. Mol. Biol., 222:58, 1-597, 1991). Polyclonal antibodies can be produced by injecting a target protein antigen into an animal and collecting blood from the animal to obtain serum containing the antibody. These polyclonal antibodies can be manufactured from animals such as goats, rabbits, sheep, monkeys, horses, pigs, cows, and dogs.

Antibodies prepared by the above method can be separated and purified using methods such as gel electrophoresis, dialysis, salt precipitation, ion exchange chromatography, and affinity chromatography.

Agents for measuring the expression level of the CD109 gene or protein have different detection efficiencies depending on the detection method, so they can be selected considering diagnostic accuracy and specificity.

In addition, one aspect of the present invention provides a kit for diagnosing or predicting prognosis of squamous pancreatic cancer comprising the composition.

The “kit for diagnosing or predicting prognosis of squamous pancreatic cancer” used in the present invention is a substance that can diagnose or predict prognosis through biological samples collected from test subjects, more specifically, individuals who have not been diagnosed with pancreatic cancer or whose pancreatic cancer subtype has not been confirmed. This means that the pancreatic cancer subtype of the test subject can be determined quickly, accurately, and easily.

The kit may include, without limitation, diagnostic kits based on conventional mRNA expression and protein quantitative analysis.

According to one embodiment of the present invention, the kit is a group consisting of a polymerase chain reaction (PCR) kit, a reverse transcription PCR (RT-PCR) kit, a DNA chip kit, a next generation sequencing (NGS) kit, a protein kit, and an array kit. It may be one or more types selected from.

For example, when the kit is applied to a PCR amplification process, the kit of the present invention may optionally include reagents necessary for PCR amplification, such as buffer, DNA polymerase, DNA polymerase cofactor, and dNTPs, When the kit is applied to an immunoassay, the kit of the present invention may optionally include a secondary antibody and a labeled substrate. In addition, the kit according to the present invention may be manufactured in a number of separate packaging or compartments containing the above-described reagent components, and the kit according to the present invention may be a diagnostic kit containing the essential elements required to perform DNA chipping. there is. A DNA chip kit may include a substrate to which a cDNA corresponding to a gene or a fragment thereof is attached as a probe, and reagents, agents, enzymes, etc. for producing a fluorescent label probe. Additionally, the substrate may include cDNA corresponding to a quantitative control gene or a fragment thereof.

Another aspect of the invention includes a) collecting a biological sample from an individual in need of diagnosis; b) measuring the expression level of CD109 gene or protein in the biological sample; and c) comparing the measured expression level of the CD109 gene or protein with a control group.

The above steps a) to c) will be discussed in detail below, and description of content in common with the above-described content will be omitted to avoid excessive complexity.

According to one embodiment of the present invention, the pancreatic cancer may be pancreatic adenocarcinoma.

Step a) is a process of collecting a biological sample from an individual or subject who needs a test to determine the pancreatic cancer subtype or predict the prognosis.

According to one embodiment of the present invention, the subject in step a) may be an individual who has not been diagnosed with pancreatic cancer or whose pancreatic cancer subtype has not been confirmed.

According to one embodiment of the present invention, the biological sample in step a) may be one or more types selected from the group consisting of blood, plasma, serum, lymph, saliva, urine, and tissue.

Step b) is a process of measuring the expression level of the CD109 gene or protein in the biological sample.

An example of step b) may be CD109 as well as proteins and genes belonging to the CD109 signal transduction system.

In one specific example of step b), proteins and genes belonging to the signal transduction system of CD109 may be Janus kinase protein (Jak, janus kinase), p-AKT protein, P-STAT3 protein, DNA, and mRNA. .

The expression level of a gene or protein in the present invention can be measured without limitation based on conventional mRNA expression and protein quantitative analysis methods.

According to one embodiment of the present invention, the expression level of the CD109 gene in step b) was determined by fluorescence in situ hybridization, chromatin immunoprecipitation, and next generation sequencing. Polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), competitive RT-PCR, real-time PCR, real-time RT-PCR, nuclease protection assay, in situ hybridization, DNA microarray. And it may be measured by one or more methods selected from the group consisting of Northern blot.

In addition, according to one embodiment of the present invention, the expression level of CD109 protein in step b) is determined by enzyme-linked immunosorbent assay, western blot, radioimmunoassay, and radioimmunodiffusion. ), immunoprecipitation, flow cytometry, immunohistochemistry, immunofluorescein, and protein microarray.

According to one embodiment of the present invention, the expression level of STAT3 phosphorylation (Phosphorylated STAT3, pSTAT3) in step b) was determined by fluorescence in situ hybridization, chromatin immunoprecipitation, and next-generation sequencing ( next generation sequencing). Polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), competitive RT-PCR, real-time PCR, real-time RT-PCR, nuclease protection assay, in situ hybridization, DNA microarray. And it may be measured by one or more methods selected from the group consisting of Northern blot.

In addition, according to one embodiment of the present invention, the expression level of pSTAT3 protein in step b) was determined by enzyme-linked immunosorbent assay, western blot, radioimmunoassay, and radioimmunodiffusion. ), immunoprecipitation, flow cytometry, immunohistochemistry, immunofluorescein, and protein microarray.

Step c) is a process of determining the pancreatic cancer subtype of a subject or individual or predicting the prognosis of a pancreatic cancer patient based on the expression level of the CD109 gene or protein measured in the biological sample.

According to one embodiment of the present invention, step c) may be to diagnose squamous pancreatic cancer or determine that the prognosis of squamous pancreatic cancer is poor when the measured expression level of the CD109 gene or protein is higher than that of the control group.

According to one embodiment of the present invention, the control group in step c) may be normal tissue without pancreatic cancer lesions, tissue of non-squamous pancreatic cancer, or normal tissue of other organs.

According to one embodiment of the present invention, the higher levels compared to the control group in step c) are proteins and genes belonging to the signal transduction system of CD109, such as Janus kinase protein (Jak, janus kinase), p-AKT protein, and P -This is a case where the level of expression of STAT3 protein, DNA, and mRNA increased numerically and quantitatively compared to the control group.

According to one embodiment of the present invention, methods for comparing genes belonging to the signaling system of CD109, which are higher than those in the control group in step c), include fluorescence in situ hybridization and chromatin immunoprecipitation. precipitation), next generation sequencing. Polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), competitive RT-PCR, real-time PCR, real-time RT-PCR, nuclease protection assay, in situ hybridization, DNA microarray. and Northern blot may be used for measurement and comparative analysis using one or more methods selected from the group consisting of.

According to one embodiment of the present invention, methods for comparing proteins belonging to the signaling system of CD109, which are higher than those in the control group in step c), include enzyme-linked immunosorbent assay and western blot. , one type selected from the group consisting of radioimmunoassay, radioimmunodiffusion, immunoprecipitation, flow cytometry, immunohistochemistry, immunofluorescence, and protein microarray. It may be measured and comparatively analyzed using the above methods.

The present invention provides a diagnostic composition and kit for squamous pancreatic cancer by presenting a biomarker that can predict squamous pancreatic cancer, so that patients with squamous pancreatic cancer can be quickly and accurately diagnosed, and appropriate treatment can be provided to patients with squamous pancreatic cancer.

Figure 1 shows a schematic diagram of the anchorage model, the parental and spheroid morphology of mouse cell lines, and the graph comparing the proliferation levels of Par and sph mouse cells.
Figure 2 shows the results of adhesion and culture on Matrigel and agarose using mouse cell lines.
Figure 3 shows the results of analysis based on RNA-seq of par and sph cells of FC1199 among mouse cell lines.
Figure 4 shows the results of confirming the molecular subtype of PDA, signaling of KRAS, and degree of blood vessel development using GESE.
Figure 5 shows the tumor growth curves of FC1199-Par and -Sph cells subcutaneously injected into C57BL/6J mice, respectively, and the results of lung H&E staining and quantification of the percentage of tumor area in the lung. FC1199-Par and -Sph cells were injected subcutaneously. This is the Kaplan-Meier survival curve result of the mouse.
Figure 6 shows the results of GSEA of RNA-seq of mouse tumors derived from Par and Sph using the characteristics of the PDA subtype.
Figure 7 shows the results of tumor weight quantification, images and quantification of metastasis to the liver, and images and quantification of metastasis to the lungs 4 weeks after injection of FC1199-Par and Sph cells into C57BL/6J mice, respectively.
Figure 8 shows the quantification results of tumor weight, lung metastasis image results, and quantification results when SUIT2-Par and Sph were injected into nude mice, respectively.
Figure 9 shows the signal density regions of H3K27ac, H3K4me1, and ATAC using ChIP-seq in a 10 kb region around the center of the GAIN-S region of FC1199-Par and Sph cells, respectively, using H3K27ac, H3K4me1, and H3K4me3 ChIP-seq. The results show the locations of CxCL12, CD109, and CCL2 in the GAIN-S region in FC1199-Par and Sph.
Figure 10 shows the gene ontology of each GAIN-S region, and as a result of confirmation using GAIN-s for RNA seq of FC1199-Par and Sph, the expression levels of CXCL12, CD109, and CCL2 in the PDA cell subtype. This is the result of confirming RNA-seq of PDA cells and PDA organs using GAIN-s.
Figure 11 shows representative images and quantification results of tube formation analysis in conditioned media with FC1199-Par and Sph using HUVEC cells, respectively, and representative images and quantification of tube formation analysis in conditioned media with FC1199-Par and Sph in Matrigel. It is a result.
Figure 12 shows tube formation analysis images and quantification results of FC1199-Par and Sph dispensed on Matrigel in a serum-free medium, respectively, and tube formation analysis images and quantification results in the presence of Par-CM and Sph-CM.
Figure 13 shows the results of the top 10 de nove motifs in the GAIN-s region, the RNA seq comparison results of FC1199-Par and Sph, and the results of quantifying the expression of the TEAD2 gene in PDA cell lines and PDA patients.
Figure 14 shows Metagene results based on ChIP-Seq of Flag-TEAD2 enrichment for the GAIN-S or LOSS-S region of Flag-TEAD2 expressing FC1199, respectively, and FC1199-Par/DMSO vs. Sph/DMSO vs. Sph/vertiporfin. Results showing Metagene based on ChIP-seq for H3K27ac enrichment in the GAIN-S region of (VP) (left) and FC1199-Sph/shRen vs. Sph/shTead2 (right).
Figure 15 shows the ChIP-seq results of H3K27ac and Flag-Tead2 in the indicated cells in the CD109 and CCL2 genes, respectively. GSEA of FC1199-Sph/DMSO, VP and Sph/shRen, and shTead2 RNA-seq genes were identified by GAIN-S. As a result, this is a heatmap confirming the association of TEAD2 central target genes.
Figure 16 shows dot plot results of TEAD2 core target gene expression in conventional and subtyped PDAs using the PanCuRx dataset, respectively, showing high (n = 88) or low expression (n = 78) of 70 TEAD2 core target genes in tumors. ) This is the survival curve result of PDA patients stratified according to ).
Figure 17 is a heatmap result of TEAD2 core target genes in the PanCuRx dataset.
Figure 18 is the result of visualizing cells from scRNAseq of PDA in UMAP.
Figure 19 shows the results of the Transwell migration and matrigel invasion analysis results of FC1199-Sph/shRen or Sph/shTead2 cells, the results of HUVEC recruitment analysis using CM obtained from FC1199-Sph/shRen or Sph/shTead2 cells, and FC1199-Sph/shRen, respectively. Alternatively, this is the result of tube formation analysis and data collection of Sph/shTead2 cells.
Figure 20 shows the tumor growth curve (left) and tumor volume (right) of FC1199-Sph/shRen or Sph/shTead2 cells subcutaneously injected in C57BL/6J mice (n = 8 and 10, respectively), respectively. H&E staining of lung sections from mice bearing subcutaneous tumors formed by Sph/shRen or Sph/shTead2 cells (left). The bar graph shows the results expressed as percentage of tumor area in the lung (right).
Figure 21 shows the quantification of tumor weight 4 weeks after injection of SUIT2-Sph cells expressing the indicated shRNA in NSG mice, respectively. Quantification of the number of metastatic colonies in the lungs and livers of mice orthogonally injected with shRNA-expressing SUIT2-Sph cells, results shown as representative images of H&E staining of the lungs.
Figure 22 shows the results showing RNA subtypes and chymotherapy response due to cancer development in basal and classical types of PDA cells, control group, gemcitabine, verteporfin, or gemcitabine/verteporfin. Tumor volume curve results of subcutaneous tumors by FC1199-Sph cells treated with (n = 7 to 8, respectively).
Figure 23 shows a schematic diagram of the TEAD2-mediated JAK-STAT activation model, the results of molecular signature analysis based on RNA-seq of the basic type or subtype of human PDA, and the basic type or subtype of human PDA RNA-seq data using the characteristics of the STAT3 target gene, respectively. Western blot analysis results of phospho-STAT3 (pSTAT3) and total STAT3 from FC1199-Par and -Sph cells treated with GSEA, DMSO, or pyridone 6.
Figure 24 is a graphical representation of the expression level of upregulated CD109 in PDA confirmed in normal tissues and tissues with pancreatic cancer.
Figure 25 shows the scatter plots of FC1199-Sph/shCd109 and FC1199-Sph/shSTAT3 RNA-seq data, respectively. GSEA of FC1199-Sph/shRen and Sph/shCd109 RNA-seq data shows the characteristics of shSTAT3 down-regulated (DN) genes. As a result of confirmation, the heatmap is a result showing the expression level of the shSTAT3 top DN gene in FC1199-Sph/shRen, Sph/shCd109, and Sph/shSTAT3 cells.
Figure 26 shows tube formation analysis of FC1199-Sph/shRen or Sph/shCd109 cells and Sph/shSTAT3 cells, respectively, and graphing the results.
Figure 27 shows the results showing the effect of pyridone 6 injection on tumor growth after injection of FC1199-Par or Sph cells in C57BL/6J mice, respectively. When mice were injected intraperitoneally with 0.5 mg/kg pyridone 6 (P6) daily for 16 days, the number of pSTAT3-positive cells in the tumor was shown by immunohistochemical (IHC) staining of pSTAT3, and the graph shows the number of pSTAT3-positive cells in the tumor. This result shows the weight, number of micrometastatic colonies in the liver and lungs.
Figure 28 shows the results of tumor volume quantification, results showing an abnormally large number of metastatic nodules to the lung and liver in C57BL/6J mice, and survival curves of patients stratified according to high or low expression of CD109 in PDA, respectively.
Figure 29 shows the results of Western blot analysis of various proteins associated with KRAS signal, subtype marker and EMT marker from CD109, pSTAT3 and 10 human PDA cell lines, respectively, tube formation analysis in CD109 ^high and CD109 ^low human PDA cells (left) and This is a graph of the results of the matrigel invasion analysis (right).
Figure 30 shows the results showing Matrigel invasion and Vascular mimicry in human PDA cell lines.
Figure 31 shows the results of metagenes based on ChIP-seq for H3K27ac enrichment in the GAIN-S region of CD109 ^high and CD109 ^low human PDA cells, respectively, CD109 ^high and CD109 ^low RNA-seq data using characteristics of human PDA subtypes GSEA results, KP4/verteporphin, KP4/DMSO RNA-seq data results using the characteristics of GSEA's TEAD2 core target gene (left) and PDA flat subtype (right).
Figure 32 shows the results of microscopic confirmation of the expression level of CD109 in 92 pathologically confirmed human PDAs.
Figure 33 is a graph comparing the proportion of microparticles between the group with weak CD109 expression and the group with strong CD109 expression, respectively. As a result, the proportion of patients with poor prognosis who relapsed within 6 months is shown in PDA compared to patients with weak CD109 expression in PDA. As a result of comparing the patient groups with strong CD109 expression, the overall survival time of the strong CD109 expression group and the CD109 weak and intermediate expression group is compared.
Figure 34 shows the results showing the location and expression level of POU5F1B (prostate cancer marker), MYC (oncogenic marker), TMEM75 (colon cancer marker), and HPRT1 (stomach cancer marker) using H3K27ac ChIP-seq of PDA cell lines.

Hereinafter, one or more specific examples will be described in more detail through examples. However, these examples are intended to illustrate one or more embodiments and the scope of the present invention is not limited to these examples.

Example 1. Experimental method of the invention

1-1. Preparation of cells for the present invention

Mouse pancreatic cancer cell lines FC1199 and NB490 were cultured in medium containing 10% (v/v) fetal bovine serum (FBS, Welvine serum, Fellgene, Korea) and antibiotics (Wel-01) (antifungal agent, Cat# 15240062, Gibco, Massachusetts, USA). (DMEM, Cat# LM001-05, Korea). Human pancreatic cancer cell lines SUIT2 and HPAF2 were cultured in minimal essential medium (MEM, Cat#11095080, Gibco), and KP2, KP4, KLM1, SNU-2469/2571, T3M4, and AsPC1 were cultured in RPMI1640 (Cat#LM011) containing 10% FBS. -01, WellGene). PANC04.03 was cultured in RPMI containing 15% FBS and 0.1% insulin (Sigma-Aldrich, MO, USA, Cat# I9278). AsPC1, SNU-2469, and SNU-2571 were purchased from Korea Celline Bank (KCLB, Seoul, Republic of Korea), and HPAF2 and PANC04.03 were purchased from ATCC (Virginia, United States). SUIT2 and KP2 were purchased from JCRB Cellbank (Osaka, Japan), and KP4, KLM1, and T3M4 were purchased from RIKEN BRC Cellbank (Saitama, Japan). Human umbilical vein endothelial cells (HUVEC) were purchased from Lonza and cultured in EGM-2 bullet kit medium (Cat# CC-3156 & CC-4176, Lonza, Basel, Switzerland). 293LTV and Plat-A cells were purchased from Cell Biolab and maintained in DMEM with 5% FBS. All cells were grown at 37°C with 5% CO2. All cell lines were routinely tested for mycoplasma using a PCR detection kit (REF # 25235, LiliF Diagnostics, Gyeonggi-do, Korea). Cells were treated with verteporphine (1 μM, Cat#SML0534, Sigma-Aldrich) for 24 h and pyridone 6 (2 μM, Cat#6577, Torqueless, Bristol, UK) for 48 h.

1-2. Construction of the Anchorage independent growth adaptation model

Cells 1.5 × 10 ^per well, in serum-free DMEM/F-12 (Cat#11320033, Gibco) with B-27 supplement (1x final, Cat#17504044, Gibco) and heparin (Cat#H3149, Sigma-Aldrich, New York) and dispensed into ultra-low attachment 24-wells (REF#3473, Corning, USA). Cells were collected in 15 mL tubes every 14 days and centrifuged for 2 minutes. Cells were resuspended in 1 mL of Accutase cell dissociation reagent (Accutase, Cat# A1110501, Gibco), incubated for 10 minutes at 37°C. Dissociated cells were stabilized in DMEM containing 10% FBS and then seeded in 6-well plates.

1-3. Experiments related to proteins, DNA, and RNA

1-3-1. Cloning of sh RNA and preparation of viral vectors

For overexpression experiments, mouse Tead2 DN (TEAD2 wild-type cDNA, Origin, Maryland, USA, Cat# MR206438) was subcloned into a retroviral vector (MSCV-neo, Addgene #105505, Massachusetts, USA). For knockdown experiments, shRNA targeting TEAD2 was cloned into retroviral and lentiviral vectors (LEPG, Addgene #111160; SGEP, Addgene #1170).

1-3-2. production of viruses

Retrovirus was generated using OPTI-MEM (Cat# 31985070, Gibco) and polyethyleneimine (PEI, Cat# 19850, Warrington, UK) in Plat-A cells (Cat# RV-102, Cell Biolabs, San Diego, United States). , Polyscience) was used to prepare plasmid DNA and packaging plasmids (VSVG and GPE). Lentivirus plasmid DNA and packaging plasmids (pMD2G and PAX2) were prepared using OPTI-MEM and PEI in 293LTV cells (Cat# LTV-100, Cell Biolabs). The supernatant containing the virus was collected for 3 days and filtered through a 0.45 μm filter. Infection of pancreatic cancer cells was performed using 1 ml of filtered virus, 1 ml of culture medium, and 20 μg/ml of polybrene (Cat# H2968, Sigma-Aldrich) in a 6-well plate. After 24 hours of incubation, the medium was replaced with antibiotic-containing medium (puromycin 1 μg/ml or G4181 mg/ml).

1-3-3. western blot

Cultured cells were washed with 1ml of phosphate buffered saline (PBS) and collected into a 1.5ml tube. Cells were incubated in 200 μl of RIPA buffer [150 mM NaCl, 50 mM Tris-Cl (pH 8.0), 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), protease inhibitor cocktail. ] and then incubated on ice for 10 minutes. These cell lysates were centrifuged for 10 minutes at 4°C. The supernatant was mixed with 50 μl of 5x sample buffer and boiled at 95°C for 10 min. Each sample was loaded and separated by 8–15% SDS-PAGE gel electrophoresis and transferred to a nitrocellulose membrane for 3 hours at 4°C. Membranes blocked with 5% skim milk were incubated with the corresponding primary antibodies overnight at 4°C. The membrane was washed with TBST and incubated with secondary antibody for 40 minutes at room temperature. Membranes were washed three times with TBST and visualized using ECL solution (Cat#1705061, Bio-Rad, CA, USA). Primary antibodies were vinculin (sc-73614, Santa Cruz, Texas, USA), CD109 (Cat#24765, CST, CST, Massachusetts, USA), STAT3 (Cat#4904, CST), and phospho-STAT3 (Tyr705, Cat#9145). , CST), AKT (Cat#4685, CST), CATK (CAT#47, CAT), CAT3 (Cat#47, CAT), Rosemont, USA), TP63 (Cat#39692, CST), KRT5/6A ( Cat#MAB1620, Sigma-Aldrich), S100A2 (Cat#109494, Cambridge, USA, ABCAM, CST), HNF4A† (Cat#313, CST), GATA6 (Cat#5851, CST, CST), CAT5 (CAT, CAT) , CAT1, CAT5, CAT), CAT5(CAT), CAT5(CAT), CAT), CAT3K27ac(Cat#4729,Abcam), H3K4me1(Cat#8895,Abcam) and H3K4me3(Cat#8580,Abcam).

1-4. In vitro cell tumor analysis

1-4-1. Growth assay of spherical cells

For cell proliferation analysis, 5 x 10 ⁴ cells were stabilized in 2 ml of growth medium and then transferred to a 6-well plate. After 48 hours of culture, the number of cells in each well was quantified daily. For sphere growth assay, 1.5 × 10 cells were re-stabilized in serum-free DMEM/F-12 with ^B -27 supplement and heparin and seeded in ultra-low attachment 24-well plates. Images of cells were captured under the microscope on day 3, and the size of the spheres was measured using the ImageJ tool. All quantified data were plotted and analyzed by the Prism application.

1-4-2. CellTiter-Glo assay for proliferation of human PDA cells

Cell proliferation of human PDA cells was measured by CellTiter-Glo (Promega, WI, USA, Cat# G7570). Cells were seeded for 24 h and CellTiter-Glo assay was performed according to the manufacturer's instructions. Briefly, the same amount of CellTiter-Glo reagent was added to the medium in each well, mixed by pipetting, and incubated at room temperature for 15 minutes. Then, 100 μl of the mixture was transferred to a 96-well plate with opaque walls and luminescence was measured.

1-4-3. Growth assay on agarose

For basic agarose, preheat 2X DMEM (Cat# LM201-50, Welgene) containing 1.2% agar (Cat# 214010, BD, New Jersey, United States), 20% FBS and 2X antibiotics. Mixed in equal volumes. Then, 1.5ml of mixed base agar was added to the 6-well plate and cultured in a culture hood. For the agarose in the upper part, the cells were stabilized in 1ml of 2X DMEM (containing 20% FBS and 2x antibiotics) and then mixed with 1ml of warmed 0.6% agarose. After the base agarose hardened, 1.5 ml of top agarose containing cells was poured onto the base agarose. Then, after the agarose above hardened, medium was added thereon. Cells were cultured in a humidified incubator at 37°C for 3 to 5 weeks and stained with 0.1% crystal violet in 10% ethanol for 30 minutes. Images of cells were taken under a microscope and the size of colonies was measured using the ImageJ tool. All quantified data were plotted and analyzed by the Prism application.

1-4-4. Metrigel penetration and cell migration (transwell migration) analysis

For cell migration analysis, 100 μl of serum-free medium containing 0.5-2 x 10 ⁵ cells was dispensed into the upper chamber (Falcon, USA, Cat# 353097). The 24-well plate containing the upper chamber was incubated for 10 minutes, and medium containing 10% FBS was added to the lower chamber. Incubation time is as shown. Cells within the permeable membrane were fixed with 3.7% formaldehyde, permeabilized with 100% methanol, and stained with crystal violet. All migrated cells were visualized under a microscope and counted with the ImageJ tool. For invasion analysis, 1-4 x 10 ⁵ cells in 100 μl serum-free medium were seeded into the Matrigel-coated upper chamber. The subsequent process is the same as the cell migration analysis.

1-5. In vitro angiogenesis test

1-5-1. HUVEC recruitment analysis

After starving cancer cells and HUVECs for 20 hours in EGM-2 medium containing 0.2% FBS, HUVECs were labeled with Cell Tracker Green CMFDA dye (Cat# C7025, Invitrogen, United States, Massachassetts) for 30 minutes. Cell migration inserts (Cat# 351152, Falcon) were coated with 0.1% gelatin solution (Cat# LS023-01, WelGene) and placed in each well. Labeled HUVECs (4x10 ⁴ ) were seeded into each cell migration insert. After 12-14 hours, the insert was washed twice with PBS and fixed with 4% paraformaldehyde for 15 minutes at 37°C, and all recruited cells were visualized using a microscope and counted with ImageJ tool.

1-5-2. HUVEC tube morphology analysis

Cancer cells were starved in serum-free EBM-2 (Cat# CC-3156, Lonza) medium for 48 hours, then conditioned medium (CM) was obtained, and tube formation analysis was performed in the presence of CM. Briefly, 50 μl of growth factor-reduced (GFR) Matrigel (Cat# 356231, Corning) was added to a 96-well plate and incubated at 37°C for 30 minutes. 1.5 x 10 ⁴ HUVEC stabilized in CM were seeded on top of Matrigel. After 2.5 hours of incubation, the tube-forming structures were visualized under a microscope and analyzed with Angiogenesis Analyzer in ImageJ software.

1-5-3. Vascular mimic test

260 μl of Matrigel was dispensed into a 24-well plate and incubated for 10 minutes at room temperature and 20 minutes in a 37°C incubator, and quantified at 1.5 x 10 ⁵ in 500 μl of serum-free DMEM. Cells were dispensed and added to Matrigel. 16 hours later, the cells were stained with syto 13, and the images of the cells were confirmed under a microscope. For experiments using conditioned medium, 5 x 10 ⁵ cells were distributed in a 6-well plate and cultured overnight and then changed to serum-free medium. After 2 days of culture, the medium was collected in a 15ml conical tube and used for analysis.

1-6. animal research

1-6-1. Subcutaneous and vertical injection of cancer cells

All experiments using animals were conducted in accordance with the approval procedures of the Korea Advanced Institute of Science and Technology Animal Care and Use Committee (KAIST IACUC). The mice used were 6-8 week old female C57BL/6N (B6) mice, male BALB/C nude mice (Orient Bio, Korea), and male NOD, Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice (JA BIO, Korea). ) was injected using. All mice weighed between 15g and 20g, and for subcutaneous injection, FC1199 (5 x 10 ⁵ ) and SUIT2 (1 x 10 ⁶ ) cells were stabilized in cold PBS. 100 μl of cells were injected into the lower flanks of 6- to 8-week-old B6 (FC1199) and naked mice (SUIT2), and 10 days after injection, tumor volume was measured as [mm ³ ] = (length [mm]) x (width [mm ] ² ) Calculated as x 0.5. In the survival analysis, FC1199 (1 x 10 ⁴ ) cells were injected as above and the overall survival time was evaluated. For orthopedic injection, B6 mice and NSG mice were injected using FC1199 and SUIT2 cells. Add 25% matrigel to cold PBS and prepare 100 μl suspension of 2 x 10 ⁴ cells. After anesthetizing the mouse, a 10 mm vertical skin incision was made on the left upper abdomen of the mouse. The peritoneum was then carefully opened and the spleen was lifted to expose the pancreas. 100 μl of cells were injected into the pancreas for 3 seconds and then paused for 10 seconds to prevent reflux. After confirming that cells were not leaking from the pancreas, the peritoneum was sutured (Cat#SK422, ALIE, Busan, Leica), and the skin was closed with 9mm stainless steel suture clips (Cat#39465011, Leica). For pyridone 6 experiments, 1 week after orthopedic injection, mice were injected intraperitoneally with 100 μl of medium (DMSO 10% + corn oil 90% (Cat# C8267, Sigma)) or pyridone 6 (0.5 mg/kg). Treatment was administered daily for 16 days by injection. Four weeks after orthopedic tumor injection, the weight of tumors produced in mice was measured using an electronic balance (Cat# WTB 200, RADWAG, Polska, Poland).

1-6-2. combination of treatments

FC1199-Sph cells ⁵ Tumor volume was calculated as 0.5 x (long axis) x (short axis) ² . Gemcitabine (Cat# G6423, Sigma-Aldrich) was dissolved in saline solution at 5 mg/ml and then intraperitoneally administered to mice at a dose of 25 mg/kg twice a week. Vertiforfin was dissolved in DMSO at 100 mg/ml, diluted in PBS at 5 mg/ml, and administered intraperitoneally to mice twice a week at a dose of 25 mg/kg. Treatment began on day 14 after transplantation, and mice were analyzed on day 25.

1-6-3. RNA-seq isolation of tumor cells

Tumors were surgically disaggregated and incubated in DMEM/F12 (Cat# LM002-04, WelGene) with digestion buffer (1.25 mg/ml), collagenase type 3 (Collagenase, Cat#LS0041822, Lakewood Washington, CO), 100 U/ml. ml Hyaluronidase (Cat#LS005475, Washington) and 100U/ml Penicillin-Streptomycin (Cat#LS202, WellGene) were added and stored at 37°C, and the dissociated tumor was incubated for 2 hours. Washed twice with PBS, trypsinized with 0.25% trypsin-EDTA (Cat#25200072, Gibco) for 5 minutes, and then neutralized with DMEM containing 10% FBS medium. Cells were completely dissociated and 8 x 10 ⁵ cells were seeded into T75 flasks. After the next day, tumor cells were treated with 2 μg/ml puromycin for 2 days. After the processing procedure, total RNA was extracted with QIAzol (Hilden, Hilden, Germany).

1-7. Construction of the next generation of sequence libraries

1-7-1. Chromatin immunoprecipitation (ChIP) assay

5 x 10 ⁶ single trypsinized FC1199 and 10 human PDA cells were used. Trypsinized cells were cross-linked with 1% formaldehyde for 15 minutes, incubated with 0.125 M glycine at room temperature for 10 minutes, and then washed with PBS. The cell pellet was incubated with 200 μl of lysis buffer (10 mM Tris-Cl pH 8.0, 10 mM NaCl, 0.2 % NP-40), protease inhibitor (complete™Protease Inhibitor Cocket, Cat#11697498001; Roche, Vassell, Switzerland), 1mM. It was dissolved with DTT and incubated on ice for 10 minutes. Chromatin was separated by centrifugation at 7,400 rpm for 30 seconds. The pellet was gently dissolved in 200 μl of nucleolysis buffer [50mM Tris-Cl pH 8.0, 10mM EDTA, 1% SDS] using protease inhibitors and 1mM DTT. The chromatinate was subjected to sonication for 10 cycles (30 seconds on/30 seconds off). The sonicated chromatin mixture was incubated for 1 hour and pre-washed with the sonicated chromatin mixture using 10 μg of rabbit IgG and 10 μl of Protein A magnetic beads (Cat# 10001, Invitrogen) for pre-clearing. . The immunoassay was performed with 1 ml of pre-washed chromatin, 1 μg of the above antibody, and 10 μl of protein A magnetic beads overnight at 4°C in a rotator. Flag antibody was purchased from Sigma (Cat# F1804). H3K27ac, H3K4me3, and H3K4me1 antibodies were purchased from ABCam (Cat# 4729, 8580, and 8895, respectively). The next day, the immune complex was washed with IP Wash I Buffer, the second wash was twice with high salt buffer, the third wash was with IP Wash II buffer, and the final wash was washed twice with TE (pH 8.0). The washed immune complexes were eluted with 200 μl of elution buffer [1% SDS and 0.1 M NaHCO3] at 45°C in a thermomixer at 1,000 rpm for 30 minutes. The solute was decrosslinked with RNase A (1 μg/μl) and 0.25 M NaCl and incubated overnight in a 65°C water bath. The next day, after incubation with proteinase K (Cat# P8107S, NEB) for 2 hours, the immunoprecipitated DNA was purified using the QIA Quick PCR purification kit (Cat# 28106, QIAGEN) in 50 μl of elution buffer.

1-7-2. ChIP library construction

The ChIP-seq library was constructed using 40 μl of purified ChIP DNA and the NEXTflex ^TM ChIP-seq kit (Cat# NOVA-5143-02, PerkinElmer) according to the manufacturer's instructions. Briefly, ChIP DNA was final repaired and size selected (250-300 bp) using AMPure XP beads (Beckman, CA, USA, Cat# A63881). Different procedures from adenylation to PCR amplification were performed according to the ChIP-seq library construction steps. The quality of the ChIP-seq libraries was determined by a Bioanalyzer using a High Sensitivity chip (Agilent), and the average size of the ChIP-seq libraries ranged from 250 to 350 bp. For multiplexing, equal molar quantities of libraries were combined considering the sequencing depth per sample (20 to 40 million reads per library). ChIP-seq libraries were sequenced using 76 base single-end reads using the Illumina NextSec platform.

1-7-3. ATAC-seq library configuration

5 x 10 ⁴ cells were prepared, gently dissolved in stabilization buffer, and then centrifuged. For the lysis step, cells were stabilized and incubated on ice for 3 minutes with 50 μl of lysis buffer (stabilization buffer 48.5 μl, 0.5 μl 10% NP-40, 0.5 μl 10% Twin-20, 1% Digitonin 0.5 μl), then washed. The buffer solution (990 μl, 10 μl of 10% Twin-20) was gently shaken three times. Lysate was centrifuged at 500 xg at 4°C for 10 minutes, and only the pellet (nuclei) was stored. Transposition mixture (25 μl of 2X TD Buffer, 16.5 μl of 1X PBS, 0.5 μl of 10% Twen-20, 0.5 μl of Digitonin 1%, 2.5 μl of Tn5 Transposase, 5 μl of nuclease-free HO) was added to the pellet. It was added and gently dissolved. The released nuclei were incubated in a heat mixer at 1000 rpm for 30 minutes at 37°C. DNA purified using the MinElute PCR purification kit (Cat#28004, QIAGEN) was used to generate an ATAC-seq Library and sequenced using the Illumina NextSeq platform with single-end reads of 76 bases.

1-8. Bioinformatics analysis of ChIP seq and ATAC-seq

1-8-1. Alignment of ChIP-seq and ATAC-seq

ChIP-seq and ATAC-seq analyzes were performed as described previously ( Roe et al., 2017 ). Briefly, raw data were mapped to reference mouse (mm9) and human (hg19) genome organizers using Bowtie2, and duplicate data were removed using SAMtools.

1-8-2. 700 GAIN-S Area Identification and Annotation

To define the GAIN-S region, the HOMER suite was used to identify peaks in each H3K27ac ChIP-seq data set and combined to identify total associated H3K27ac peaks (n = 42,788). If the fold change of FC1199-Sph versus Par was greater than 4 (log2 fold change >2), it was assigned to the GAIN-S region (n = 700). To annotate GAIN-S regions, use the annotatePeaks tool to identify peaks as TSS (transcription start region), TTS (transcription end region), exon (coding), 5' UTR exon, 3' UTR exon, intron, or gene. Determined whether it was inside the argument or not. In this study, the 5' UTR and 3' UTR exons were unified into exons.

1-8-3.Visualization of ChIP-seq and ATAC-seq data

BigWig files were generated using the makeBigWig tool (HOMER suite) for visualization with the UCSC Genome Browser. Metagene plots were generated by centering the GAIN-S region extending +/- 5,000 bp (H3K27ac, H3K4me1 ChIP-seq and ATAC-seq) into 25 bp bins using the annotatePeaks tool (HOMER suite). .

1-8-4. Motif enrichment in GAIN-S and LOST-S regions

We search for defined ATAC peaks in FC1199 cells to find enriched motifs in the GAIN-S and LOST-S regions. These genomic sequences were further analyzed with the TRAP algorithm to identify enriched motifs compared to JASPAR vertebrate and Benjamini-Hochberg modified with rat promoter as control.

1-9. Biological information analysis by RNA-seq

1-9-1. Identification of differentiation expressed genes

Raw reads from RNA-seq were aligned to the mouse gene collection mm9 using the STAR mapping tool. Differentially expressed genes were analyzed using the Couplinks tool.

Total RNA was extracted with QIAzol reagent (Cat#79306; QIAGEN) according to the manufacturer's instructions. RNA-seq libraries were constructed using 10 μg of purified RNA and the NEXTFlexTM Rapid Directional mRNA-seq kit (Cat# NOVA-5138-11; PerkinElmer, MA, USA). Briefly, purified RNA was selected for poly-A’ and fragmented with fragmentation enzyme. After first and second strand cDNA synthesis from the template of fragmented RNA, adenylation with PCR amplification was performed following the RNA-seq library construction steps.

1-9-2. GSEA analysis

GSEA was performed according to the guidelines. To generate a custom gene set for the GAIN-S region, genes with the closest TSS to the GAIN-S region were assigned as peak-related genes.

1-9-3. Gene ontology (GO) analysis of genes related to the GAIN-S region.

A total of 555 genes related to 700 GAIN-S regions were determined by proximity. Then, the defined gene list was input for GO analysis using the AmiGO tool.

1-9-4.Biometric analysis of PanCuRx data

The genes of tumors collected from PanCuRx patients were analyzed. Hierarchical clustering was performed using the Pearson correlation coefficient between each patient or TEAD2 core target gene signature (70 genes), and patients (n=88) with high or low TEAD2 scores were classified. Survival analysis of PanCuRx data was performed using the Log-rank (Mantel-Cox) test in GraphPad Prism software.

1-9-5. Bioinformatic analysis of scRNA-seq

Tumors or core biopsies were finely minced and dissociated overnight at 4°C. Tumor cells were enriched through depletion of the CD45+/CD90+/GlyA+ population using MS columns (Miltenyi Biotec). Enriched tumor cells were loaded and dissociated into dropplet emulsion using the Chromium Single Cell 3' v.2 kit (10x Genomics) and subsequent libraries were sequenced on an Illumina HiSeq 2500 platform. Single-cell sequencing data were processed using CellRanger v.3.1 (10x Genomics) and aligned to the official CellRanger human reference genome GRCh38 3.0.0. Further processing of the data was performed in python 3.7.9 and R3.6.2 using the scanpy (v1.8.1) and Sureat (v3.2.3) packages, respectively. Cells with fewer than 200 unique genes and a mitochondrial content of more than 15% were removed. Samples were run through the SAM algorithm pipeline (v0.8.1, (Tarashansky et al., 2019) for normalization, dimensionality reduction and Leiden clustering. Cell types were evaluated for expression of marker genes based on (Puram et al., 2017). Malignant cell identity was confirmed through inferred CNV analysis based on immune cells, fibroblasts, and endothelial cells. Cells were identified using AddModuleScore (Seurat) using two Moffit subtype gene sets (Moffitt et al. , 2015) and were assigned to higher-scoring subtypes. Cells were also scored for expression of the TEAD2 core set of target genes using AddModuleScore.

1-10. histology

At the end of each animal experiment, the mice were sacrificed and perfused with cold PBS. Primary tumors and organs were then surgically resected and cultured overnight in 4% paraformaldehyde (primary tumors) or 10% formalin (organs) at 4°C. Fixed primary tumors and organs were embedded in paraffin and sliced at 4 μm thickness. Sections were stained with hematoxylin and eosin (H&E), and the metastatic area and number of metastatic foci were determined by ImageJ software (NIH). Transition foci were classified as micrometafocal if their diameter was greater than 100 μm but smaller than or equal to 500 μm, and macrometafocal if their diameter was greater than 500 μm.

1-11. clinical sample

A total of 92 patients who underwent surgical resection with a pathological diagnosis of PDA at Yonsei University Severance Hospital were evaluated (IRB number 4-2015-0297). Time to recurrence was defined as the time from the date of curative surgery to the time of tumor recurrence. Recurrence after surgical resection of pancreatic cancer within 6 months was defined as early recurrence. Overall survival was defined as the time from curative surgery until death.

1-12. Immunohistochemical staining

Immunohistochemistry (IHC) was performed on paraffin-embedded tissues using rabbit-derived monoclonal antibodies against CD109 (1:400, E4l2V, Cat#24765, CST) and p-STAT3 (1:100, Tyr705, Cat#9145, CST). It was performed using standard immunohistochemical techniques. For IHC staining, slides in 10 mM sodium citrate buffered distilled water (pH 6.0) were boiled in a water bath at 97°C for 20 minutes, cooled for 30 minutes, and then directly stained after antigen retrieval. Primary antibodies were incubated for 18 h at 4°C. Dako REAL Peroxidase Detection System Kit was used according to the manufacturer's specifications and counterstained with ready-to-use anti-rabbit/mouse secondary antibody (Cat# K5007) and hematoxylin solution (Cat# 03971). For quantification of p-STAT3 positive cells, stained tissue images were acquired using an Axioscan Z1 (Zeiss) at was counted by the IHC Toolbox Plugin. Light brown nuclear staining for CD109 was considered "weak," visible only at high magnification (at least 100x), heterogeneous nuclear staining of varying shades of medium to dark brown, "moderate," and uniform dark brown, "strong." All slides were classified as gland type or non-gland type depending on their shape.

1-13. statistical analysis

All statistical analyzes were performed in Prism (GraphPad Software) and presented as mean ± standard error of the mean (SEM) and unpaired t-test was used to assess significance with p-value < 0.05.

Example 2. Establishment of an in vitro subtype switching model that reduces the basic transcriptional phenotype of PDA.

We hypothesized that the process by which cells survive detachment and then reattach and grow would mimic molecular subtype transitions and “dissemination-seeding” during PDA progression. For this purpose, we generally culture cancer cells growing as a monolayer (Parental; hereinafter referred to as Par) on ultra-low attachment plates, allow anchorage-independent growth for about 2 weeks, and then readapt the surviving cells to monolayer culture. (from Spheroid; hereinafter referred to as Sph) (Figure 1A). Previous studies have shown that most naturally occurring pancreatic tumors growing in Kras ^G12D /+;Trp53 ^R172H /+;Pdx1-Cre (KPC) mice show morphological characteristics similar to existing molecular subtypes and appear well differentiated. (Hingorani et al., 2005). Using the above in vitro adaptation protocol, we identified basal-like lineages of existing subtype PDA cells derived from tumors developed in KPC mice and human PDA.

For this purpose, we used the mouse KPC cell line (FC1199-Par/FC1199-Sph and NB490-Par/NB490-Sph, respectively) and the human PDA cell line SUIT2 (SUIT2-Par/SUIT2-Sph), a well-known subtype for a long time. -Sph pair was set. Because the use of FC1199 cells in PDA studies was well known, we prioritized to further characterize the FC1199-derived Par/Sph pair. For example, mouse KPC cell lines, including FC1199, displayed a flat appearance upon DNp63 overexpression.

When Par-Sph pairs propagated under monolayer conditions, we detected no noticeable differences in their morphology or diffusion rates (Figure 1B, 1C). In contrast, under ultra-low adhesion conditions, Sph cells formed much larger spheres than Par cells (Figure 2A). In support of these observations, results from colony formation grown on soft-agarose showed that anchorage-independent colony formation of Sph cells was significantly increased compared to Par cells (Figure 2B). . The above findings suggested that Par-to-Sph switching in a homogeneous background can enhance the aggressiveness of PDA cells in vitro, which leads to transwell migration and Matrigel invasion of Sph cells. This was further supported by the ability of matrigel to invade (Figure 2A, 2B).

Next, we sought to distinguish transcriptional changes associated with the phenotypic outcomes described above. After performing RNA-seq analysis of FC1199-Par and FC1199-Sph cells, unbiased Gene Set Enrichment Analysis (GSEA) was used to compare transcriptome profiles of individual cell states. Among a total of 7,647 gene sets accumulated in the Molecular Signatures Database (MSIGDB, v7.0), the top genes enriched in Sph cells were mainly categorized into extracellular matrix, KRAS signaling, and angiogenic development ( Figure 3 ). In particular, epithelial to mesenchymal (EMT) transition and hyperactivation of mutant KRAS signaling are known molecular features of basal-type PDA, but the importance of angiogenic development in the basal-type subtype is still unknown.

To further substantiate our observations, we utilized publicly available PDA subtypes and determined whether the Par-to-Sph transition resembles a rudimentary, fundamental molecular subtype switch. In GSEA, we found that pro/classical gene features were negatively enriched in Sph cells, whereas squamous/basal-type gene features were positively enriched (Figure 4A), suggesting that the Par-to-Sph switch causes classical/basal-type molecular changes. indicated that it could be indicated. GSEA of KRAS downstream targets showed negligible changes in KRAS expression or ERK phosphorylation, but confirmed an increase in KRAS ON and a decrease in KRAS OFF signaling in Sph cells (Figure 4B). Finally, we confirmed that a set of genes related to vascular development was significantly enriched in Sph cells (Figure 4C).

To rule out the possibility that anchorage-independent culture conditions selectively enriched cells of the basal-like subtype present in monolayer cultures, RNA-seq analysis was performed on cells in spherical cultures, an intermediate stage between Par and Sph. Increased and decreased expression of basal-like and classical genes, respectively, were absent in spherically cultured cells, indicating that the anoikis induction-adaptation process actively induced reprogramming of molecular subtypes rather than exerting strong selective pressure. . Of note, although aberrant expression of DNp63 is known to drive flat subtype reprogramming in PDA cells (Somerville et al., 2018), we did not identify increased expression of DNp63 or its downstream targets in Sph cells. (Figure 4).

Collectively, our results suggest that transcriptional changes imposed by changes in tissue culture conditions, primarily through manipulation of anchorage-independent growth, reprogram PDA cells into a basal-like subtype in a DNp63-independent manner.

Example 3. Confirmation of promotion of primary tumor formation and metastasis of basal subtype

The research results of Example 2 led us to examine whether phenotypic results showing the characteristics of the basal-like subtype are preserved in vivo. To this end, we injected FC1199-Par and -Sph cells subcutaneously into the flanks of syngenic mice and monitored tumor growth.

We found that subcutaneous tumors derived from Sph cells grew faster than tumors derived from Par cells (Figure 5A). Accordingly, mice carrying Sph-tumors had significantly increased lung and liver metastases compared to mice carrying Par-tumors, indicating that switching to the basal subtype actually enhanced the metastatic spread of PDA cells (Figure 5B). . Additionally, mice with subcutaneous par tumors survived an average of 123 days, whereas mice with Sph tumors survived an average of only 65 days, showing a statistically significant difference (log-rank test, p < 0.05).

To confirm the effectiveness of subtype switching in the human PDA setting, SUIT2-Par and -Sph cells were injected subcutaneously into immunocompromised mice. Similar to what was observed with mouse KPC cells, SUIT2-Sph cells displayed faster primary tumor growth and more metastatic lung and colon cancers when injected subcutaneously (Figure 5C).

It is important to ensure that the molecular changes to the basal-like subtype are maintained in vivo. Otherwise, it is difficult to conclude the causal role of molecular subtype reprogramming. For this purpose, RNA-seq analysis was performed with cancer cells freshly isolated from FC1199-Par/Sph-derived tumors. Consistent with the results of RNA-seq analysis obtained from cultured cell lines, GSEA analysis verified the conservation of transcriptional features of base-like PDA identity in vivo. Previously observed transcriptional phenotypes, including KRAS pathway activation and angiogenic development, were also more active in Sph-derived tumors than in Par-derived tumors ( Fig. 6 ). These data allow us to conclude that the diverse phenotypes observed in vivo indicative of aggressive PDA are accompanied by the acquisition of a basal subtype molecular program.

Next, an orthopedic implantation model of PDA was used to better reflect the patient's PDA progress. Consistent with the results from subcutaneous injections, we observed that transplanted FC1199-Sph cells formed larger primary tumors in the pancreas compared to FC1199-Par cells (Figure 7A). The degree of PDA cell colonization in the liver and lung, the two main organs to which PDA cells metastasize, was also significantly higher in mice transplanted with FC1199-Sph cells than with FC1199-Par (Figure 7B, C). Similar results were observed in orthotopic xenograft experiments using SUIT2-Par and SUIT2-Sph cells (Figures 8A,B). These findings indicate that our approach faithfully recapitulates the fundamental properties of PDA cells and leads us to explore the molecular mechanisms of subtype switching.

Example 4. Confirmation of endothelial transcriptional promoter activity of basal subtype

Recent studies have shown that establishing unique transcriptional signatures for each PDA subtype, indicated by distinct chromatin states of active enhancer elements, may explain lineage switching in PDA and establish distinct epigenetic landscapes. It can be helpful. To explore the association between subtype differences and chromatin changes, we used chromatin immunoprecipitation and sequencing (ChIP-seq) and transposase accessible sequencing (ATAC-seq), respectively. We mapped the distribution of distinct histone modifications and chromatin accessibility. In particular, we prioritized analyzing the genomic distribution of H3K27ac, a reliable marker of enhancer activity. Although there were no apparent differences in global levels of histone modifications based on Western blot results of total histone extractions, bioinformatics analysis revealed hundreds of genomes showing more than a 4-fold difference in H3K27ac between FC1199-Par and FC1199-Sph cells. The area was identified. Afterwards, we used the regions that gain or lose H3K27ac signal in FC1199-Sph cells as GAIN-S (n = 700) and LOST-S (n = 334), respectively. Chromatin accessibility showed no obvious difference between FC1199-Par cells and FC1199-Sph cells, but the H3K4me1 accumulation pattern in FC1199-Sph cells was similar to that of H3K27ac in both GAIN-S and LOST-S regions (Figure 9 A , B). They were listed in "stable" condition.

Next, we performed a GREAT tool-based unbiased ontology analysis of genes located around GAIN-S sites (GAIN-S genes) and LOST-S sites (LOSS-S genes) (McLean et al., 2010). . Unexpectedly, we found that 7 of the top 10 enriched GO terms in the GAIN-S gene were most closely related to blood vessel development, with the term “neovascular” (Figure 10A). Indeed, angiogenic development was one of the top enriched signatures in Sph cells (Figure 3). On the other hand, the LOST-S gene was only associated with the GO term “magnesium ion binding”. Additionally, we found a significant association of enhancer activity and gene expression in GSEA analysis of RNA-seq data performed on the FC1199-Par/Sph pair (Figure 10B). In support of this, we found increased expression of genes associated with angiogenesis, such as CXCL12, CD109, and CCL2, located in close proximity to the GAIN-S enhancer, in human and mouse basal-type subtype PDA cells (Figure 10C). Additionally, GSEA showed increased expression of the GAIN-S gene in the basal-like subtype compared to the classical subtype in patient PDA and patient-derived PDA organoids. Conversely, expression of the LOSS-S gene was downregulated in both the patient's basal-like PDA cells and basal-like subtype PDA (Figure 10D). Together, these results implicate the acquisition of gene expression programs relevant to endothelial cell biology as a readout of enhancer-driven epigenetic changes enriched in basal-like PDA.

The above analysis provided a basis for examining the phenotypic consequences of basal subtype conversion. Under the assumption that reprogramming of basal-like subtypes allows PDA cells to acquire behavioral patterns of endothelial cells that may help increase their aggressiveness, we sought to determine the extracellular or intrinsic angiogenic propensities of PDA cells in vitro. I decided to investigate. We first determined whether Sph cells could promote experimental angiogenesis. For this purpose, tube formation and trans well migration assays were performed using human umbilical vein endothelial cells (HUVEC) in the presence of conditioned medium (CM) of FC1199-Par and -Sph cells. Compared to the CM of FC1199-Par, the CM of FC1199-Sph increased the angiogenic ability of HUVEC, represented by forming tubes in Matrigel or passing through the trans well membrane (Figure 11). Of note, multiple GAIN-S genes, including CXCL12, CD109, and CCL2, encode secretory proteins, raising the possibility that basal-type subtype PDA cells acquire a pro-angiogenic secretory phenotype through an activated GAIN-S enhancer program. It suggests that there is. Our observations are an extension of recent studies suggesting inflammatory cytokines secreted by flat subtype PDA cells as a key to recruiting neutrophils.

It has been suggested that some cancer cells create their own microvascular channels independent of typical blood vessels, termed “vascular mimicry,” which may allow subtypes of breast cancer cells to metastasize to distant sites. Surprisingly, compared to FC1199-Par cells, FC1199-Sph cells formed robust tubular structures when grown on Matrigel, a well-established phenotypic consequence of cell-autonomous vascular mimicry (Figure 12). Additionally, application of CM harvested from FC1199-Sph cells to FC1199-Par cells was sufficient to facilitate the formation of vascular-like tubes, suggesting a cell-non-autonomous mechanism of vascular mimicry. These findings indicate that basal-type subtype PDA cells use enhancer-mediated transcriptional changes to adopt characteristics of non-epithelial lineages, in our case endothelial cells, to support high levels of aggressiveness during cancer progression.

Example 5. Confirmation of use of TEAD2 in tumor endothelial enhancer activation in basal subtype PDA

Previous studies have shown that upregulated DNp63 plays a role in reprogramming the enhancer topography of the flat subtype. Upregulation of the Sonic Hedgehog pathway TF GLI2 is another way to change the classical molecular subtype of PDA to a basal molecular subtype change. However, Tp63 and Gli2 mRNA expression was not detected in FC1199-Par or -Sph cells. These data allowed us to identify the TFs that activate the GAIN-S enhancer and determine whether these sites are functionally related. To evaluate TF motif enrichment, a genomic region extending 500 bp above and 500 bp below the center of each GAIN-S enhancer peak was analyzed using the HOMER de novo motif discovery tool (Figure 13A). The reason we decided to focus on Hippothway TF TEAD2 among the top abundant motifs in the GAIN-S region is that in RNA-seq analysis, TEAD2 showed the greatest increase in expression in FC1199-Sph compared to FC1199-Par (Figure 13B), and the expression gene This is because the amount was the second highest. When analyzing the GAIN-S enhancer, the TEAD2 motif was not found, indicating selective regulation of the GAIN-S enhancer. Consistent with the results of subtype switching in FC1199 basal cells, TEAD2 mRNA expression was significantly upregulated in the patient's PDA and in subtypes of basal PDA cells, including the subtype of basal PDA cells without TP63 mRNA expression (Figure 13C, D).

The above results led us to investigate whether TEAD2 can bind to the GAIN-S enhancer site. Because of the known high homology shared by TEADs 1–4, we constructed a genome-wide map of exogenously expressed Flag-Tagged TEAD2 in FC1199 cells. Through analysis using the HOMER tool, it was confirmed that TEAD-related motifs were enriched in the Flag-TEAD2 peak. Moreover, we observed that the occupancy of Flag-TEAD2 was stronger in the center of the GAIN-S region than in the LOSS-S region, demonstrating that selective regulation of TEAD2 is possible (Figure 14A). These results indicate that TEAD2 is a key factor in the activation of basal-type subtype-related GAIN-S enhancer.

Next, we confirmed whether the increased expression of TEAD2 is essential for the activity of the GAIN-S enhancer and the transcription process in the endothelial cell lineage. To assess the necessity of TEAD2 to maintain the transcriptional activity of endothelial cell enhancers, we used short hairpin RNA (shRNA) (knockdown or vertporfin), an inhibitor of the interaction of the TEAD family with YAP, to impair the function of TEAD2. Through ChIP-seq analysis, we show that two independent TEAD2 shRNAs and verteporfin reduce H3K27ac occupancy to a similar extent in the GAIN-S region of FC1199-Sph cells while having no effect on H3K27ac signaling at randomly selected sites. was found (Figure 14B, Figure 15A). Accordingly, RNA-seq analysis revealed downregulation of the GAIN-S gene in TEAD2-deficient FC1199-Sph cells (Figure 15B). The most downregulated genes by TEAD2 shRNA were associated with several GO periods related to endothelial cell development, indicating that an important function of TEAD2 is to maintain the endothelial lineage program in basal PDA cells (Figure 15C).

To define the downstream targets of TEAD2, we compared H3K27ac ChIP-Seq and RNA-seq data sets of the FC1199-Par/Sph pair. A total of 70 genes, including Ccl2, Cxcl12, and Cd109, were identified as core targets of TEAD2 on the basis that H3K27ac occupancy increased and downregulated expression upon introduction of TEAD2 shRNA. To determine whether the differential expression of these genes was associated with molecular subtypes of human PDAs, we calculated random expression scores with PDAs from the PanCuRx dataset using single-sample GSEA (ssGSEA). According to ssGSEA analysis, PDA classified as a basic subtype showed a higher level of expression of the TEAD2 target gene compared to PDA classified as a classical subtype (Figure 16A). Conversely, as a result of decomposition of molecular subtypes based on TEAD2 target gene expression, it was confirmed that higher expression of these genes was closely related to the basal subtype and was associated with poor prognosis (Figure 16B, Figure 17). To support our observations, when we analyzed publicly available scRNA-seq data from 15 PDAs, expression of TEAD2 core targets was found to be higher in clusters of the basal subtype than in clusters of the classical subtype. (Figure 18).

As described above, mutant KRAS copy number abnormalities are a novel characteristic of the basal molecular subtype PDA. Therefore, we investigated whether TEAD2 expression level was related to the mutation status of KRAS gene in PDA, and found that allelic imbalance of mutant KRAS was closely associated with TEAD2 expression in PDA. Interestingly, TEAD2 deficiency led to transcriptional repression of KRAS targets, suggesting that convergence of mutant KRAS and TEAD2 pathways may play a role in determining basal-type molecular subtype. Together, our results suggest that the TEAD2-dependent endothelial gene expression program likely represents a basal-like feature of aggressive PDA cells.

Example 6. Confirmation of attenuation of tumor progression and increased chemotherapy response in PDA through inhibition of TEAD2

After verifying the casual relationship between TEAD2 upregulation and characteristics of endothelial cells and associated basal subtypes, we next investigated the functional consequences of TEAD2 depletion in vitro and in vivo. Expression of TEAD2 shRNA did not affect cell proliferation and tumorigenesis of PDA cells, while the trans well migration and matrigel invasion abilities of Sph cells were attenuated (Figure 19A, B). Additionally, conditioned medium from TEAD2-depleted cells had a significantly reduced ability to recruit HUVECs (Figure 19C). TEAD2 inhibition by shRNA or verteporfin also reduced the ability of PDA Sph cells to form vasculature-like networks (Figure 19D).

To extend our findings to a physiologically relevant in vivo setting, we depleted TEAD2 or implanted FC1199-Sph cells subcutaneously and monitored primary tumor and metastasis formation. As a result, we showed that FC1199-Sph cells expressing TEAD2 shRNA showed slower tumor growth and reduced colonization in lung parenchyma compared to control cells (Figure 20A,B). Similar results were observed in orthotopic xenograft experiments using TEAD2 shRNA-expressing SUIT2-Sph cells (Figures 21A to C). To complement shRNA-mediated loss-of-function experiments, we also co-overexpressed the well-known TEAD2 dominant negative (DN) mutant cDNA and empty vector in FC1199-Sph cells ( Liu-Chittenden et al., 2012 ). Similar to the results of TEAD2 inhibition through shRNA, primary tumor growth and lung metastatic lesions were observed to decrease when TEAD2-DN cDNA was expressed.

The response of PDA to chemotherapy varies depending on the molecular subtype, with the basal-like subtype being the least sensitive (Aung et al., 2018). Because TEAD2 deficiency impaired the basal-like identity of PDA cells, we speculated that TEAD2 inhibition alleviated the poor response to cytotoxic chemotherapy in vivo. To this end, we monitored the growth of subcutaneous tumors derived from Sph cells through therapeutic intervention. We chose gemcitabine, a chemotherapy agent used to treat pancreatic cancer in hospitals. To evaluate the synergistic effect of spinal lipids, low-dose gemcitabine (25 mg/kg body weight twice a week) was administered instead of the standard dose (50-100 mg/kg body weight twice a week). The combination of gemcitabine and verteporfin treatment significantly inhibited tumor growth, but not gemcitabine or verteporfin alone (Figure 22A,B), suggesting that TEAD2 inhibition increased response to chemotherapy. . Taken together, these findings suggest that TEAD2 is important for maintaining the basal identity and aggressiveness of PDA cells in vivo.

Example 7. Confirmation of CD109 regulation of STAT3 activation and tumor progression.

CD109 is a glycosylphosphatidylinositol (GPI) anchored cell surface protein that is involved in regulating tumor growth factor beta (TGF-β) signaling. CD109 has also been shown to regulate Janus kinase (JAK)-dependent signal transducer and activator of transcription 3 (STAT3) activation and metastatic ability of lung adenocarcinoma cells. The gene encoding CD109 was one of the 70 TEAD core targets, so we sought to determine whether CD109 is an important effector of the TEAD2 pathway that enhances the JAK-STAT3 pathway in basal PDA cells (Figure 23A). We confirmed the abundance of STAT3 features in human PDA and FC1199-Sph cells. (Figure 23B, C). Compared to FC1199-Par, FC1199-Sph cells showed higher levels of phosphorylated STAT3 (pSTAT3) upon treatment with pyridone 6 (Figure 23D), a potent Pan-Jak-Kinase inhibitor. was reduced by . JAK kinase activity requires the interaction of interleukin-6 (IL-6) family cytokines and their receptor glycoprotein 130 (GP130). Except for CD109, other IL-6-type cytokines were not upregulated in FC1199-Sph cells, supporting our hypothesis that CD109 is involved in STAT3 activation in basal-type subtype PDA.

Therefore, we sought to determine whether CD109 is important for maintaining STAT3 signaling. We compared RNA-seq data from FC199-Sph cells expressing shRNA against CD109 with shRNA against STAT3 and found that the gene expression profile of CD109-depleted cells was very similar to that of STAT3-depleted cells. (R2 = 0.6004, Figure 25A). Accordingly, we found that the top 200 down-regulated genes upon STAT3 depletion were down-regulated upon CD109 depletion (Figure 25B,C). The STAT3 signature obtained from STAT3 knockdown in human PDA or FC1199-Sph cells was positively enriched in FC1199-Sph cells, which were impaired by verteporfin treatment or TEAD2 knockdown. In addition, inhibition of CD109 and STAT3 by shRNA abolished the ability of FC1199-Sph cells to form a blood vessel-like network (Figure 26A, B). These data supported our model in which CD109 upregulation drives activation of the JAK-STAT3 pathway and led us to investigate the physiological relevance of the CD109-JAK-STAT3 axis in basal-type PDA subtype cells.

They then determined whether the kinase activity of JAKs was required for tumor growth because several kinases, including receptor tyrosine kinases and Src, can phosphorylate STAT3. Pyridone 6 treatment of mice transplanted with FC1199-Par or FC1199-Sph cells reduced the total number of pSTAT3-positive cells in the tumor without causing weight loss (Figure 27 A,B). In particular, pyridone 6 treatment significantly reduced primary tumor growth and colon metastasis to the liver and lung in FC1199-Sph-derived tumors, but had no effect on FC1199-Par-derived tumors (Figures 27C-E), and the basal-like subtype. It was found that tumors require JAK kinase activity.

Next, we investigated the functional significance of CD109 in PDA in vivo. Analysis of publicly available data sets showed that CD109 expression was not detected in normal human pancreas, but aberrant CD109 expression was detected in several human cancer types, including PDA (Figure 24). Additionally, we found that FC1199-Sph cells expressing CD109 shRNA had slower primary tumor growth than control tumors upon orthotopic injection (Figure 28A). Surprisingly, CD109 depletion significantly delayed metastatic colonization in the lung and liver (Figure 28B, C). Accordingly, high expression of CD109 in PDA correlated with a short median overall survival of PDA patients (Figure 28D). Our data highlight CD109 as a potential therapeutic target and biomarker for PDA with basaloid subtype.

Example 8. Confirmation of basal subtype characteristics in human PDA through CD109 upregulation

To evaluate the relevance of our results to humans, we determined whether the level of CD109 expression had similar characteristics to those of basal human PDA cells. We separated 10 human PDA cell lines into two groups based on Western blot analysis of their endogenous CD109, 5 expressing high levels (CD109 ^high cells) and 5 expressing low levels (CD109 ^low cells). (Figure 29A). Consistent with FC1199-Sph and -Par, mRNA expression of TEAD2 was significantly higher in CD109 ^high cells than in CD109 ^low cells. Consistent with the effect of CD109 deficiency on STAT3 transactivation (Figure 24A-C), CD109 ^high cells showed higher pSTAT3 levels than CD109 ^low cells (Figure 29A). CD109 promotes epidermal growth factor receptor (EGFR)-mediated phosphorylation of protein kinase B (also known as AKT). Accordingly, 4 out of 5 CD109 ^high cell lines showed high phosphorylation of AKT. There was no difference in both ERK phosphorylation and KRAS expression between CD109 ^high cells and CD109 ^low cells (Figure 29A).

Expression of known classical markers (HNF4A or GATA6) was enriched to varying degrees in CD109 ^low cells, whereas expression of known markers of the basal subtime (TP63, KRT5/6A, S100A2) was restricted to only a subset of CD109 ^high cells. There was no correlation between the expression pattern of CD109 and well-established EMT markers (ZEB1, VIM, N-Cad, E-Cad) or cell proliferation (Figure 29A,B). Nevertheless, the aggressive characteristics of CD109 ^high cells influenced their high trans well migration ability and matrigel invasion ability (Figure 30).

Next, the genomic distribution of H3K27ac was investigated through ChIP-seq analysis to determine the transcriptional activity of STAT3 in CD109 ^high cells and CD109 ^low cells. We found more than twice the H3K27ac occupancy at known STAT3 binding sites in CD109 ^high cells than in CD109 ^low cells (Figure 31A), but differences in H3K27ac signal at randomly selected sites were minimal. The enrichment of STAT3 target gene signatures in CD109 ^high cells further supported STAT3-mediated transcriptional activation. Together, our findings suggested that upregulation of CD109 expression is a basal subtype-specific marker of PDA. In support of this possibility, GSEA showed enrichment of basic gene signatures in CD109 ^high cells, whereas classical gene signatures were enriched in CD109 ^low cells (Figure 31B). The top 200 downregulated genes upon TEAD2 or CD109 depletion in FC1199-Sph were associated with CD109 ^high cells. The functional role of TEAD2 was supported by GSEA depletion of the gene signature of TEAD2 core targets or squamous identity in verteporfin-treated KP4 cells (Figure 31C). These findings revealed transcriptional commonalities between base-switched and CD109 ^high cells and human PDA cells.

Lastly, we examined whether CD109 expression correlated with the molecular subtype of PDA and clinical outcomes of PDA patients. Recent studies have identified associations between morphological patterns and molecular subtypes (S et al., 2020). For example, PDAs with a ‘gland forming’ component of more than 50% were associated with the classical subtype, whereas PDAs with a ‘non-gland forming’ component of more than 50% were associated with the basal subtype. In this study, we examined whether the level of CD109 expression was related to the non-gland forming pattern in pathologically confirmed human PDA. A total of 92 surgically resected PDA specimens were evaluated by immunohistochemistry (IHC), and based on CD109 staining intensity, these PDA were classified into a weak group (CD109 ^weak group, n = 36) or a moderate group (CD109 ^moderate group, n = 37). , divided into strong groups (CD109 ^strong group, n = 19) (Figure 32). We sought to determine ^whether CD109 staining intensity was related to morphological subtype and found that the proportion of particulate components was significantly higher in CD109 strong PDA than in CD109 ^weak PDA and CD109 ^moderate PDA (p < 0.0001, Fisher's exact test) (Figure 33A). According to known characteristics of the basal-like subtype, patients with CD109 ^strong PDA had a higher probability of tumor recurrence (p < 0.0001, Fisher's exact test) and shorter overall survival than patients with CD109 ^weak or CD109 ^moderate PDA. (p = 0.0314, log-rank test). (Figure 33B, C). These findings demonstrate the clinical relevance of CD109 upregulation in PDA, which is closely related to the basal subtype, and may provide improved diagnostic and treatment options for PDA patients, and may also serve as a biomarker in the development of other cancers. It was confirmed that it could play a role (Figure 34).

So far, the present invention has been examined focusing on its preferred embodiments. A person skilled in the art to which the present invention pertains will understand that the present invention may be implemented in a modified form without departing from the essential characteristics of the present invention. Therefore, the disclosed embodiments should be considered from an illustrative rather than a restrictive perspective. The scope of the present invention is indicated in the claims rather than the foregoing description, and all differences within the equivalent scope should be construed as being included in the present invention.

Claims (9)

A composition for diagnosing or predicting squamous pancreatic cancer, comprising an agent for measuring the expression level of CD109.
In claim 1,
The agent for confirming the expression level of CD109 is a composition for diagnosing or predicting squamous pancreatic cancer consisting of genes and proteins.
In claim 1,
The agent for measuring the expression level of CD109 is a composition for diagnosing or predicting squamous pancreatic cancer, which is selected from the group consisting of a specifically binding primer pair, probe, antisense nucleotide, antibody, oligopeptide, ligand, PNA, and aptamer.
A kit for diagnosing or predicting squamous pancreatic cancer, comprising an agent for measuring the expression level of CD109.
In claim 4,
The kit for diagnosing or predicting squamous pancreatic cancer is a kit for diagnosing or predicting squamous pancreatic cancer consisting of a kit selected from the group consisting of a PCR kit, RT-PCR kit, protein kit, and array kit.
As a method of providing information on a kit for diagnosing or predicting squamous pancreatic cancer comprising an agent for measuring the expression level of CD109;
a) collecting a biological sample from an individual in need of diagnosis;
b) measuring the expression level of CD109 gene or protein in the biological sample;
c) a method for providing information for diagnosis or prognosis of squamous pancreatic cancer, comprising comparing the measured expression level of the CD109 gene or protein with a control group.
In claim 6,
A method of providing information, wherein the biological sample in step a) is one or more selected from the group consisting of blood, plasma, serum, lymph, saliva, urine, and tissue.
In claim 6,
The expression level of the CD109 gene in step b) was determined by fluorescence in situ hybridization, chromatin immunoprecipitation, and next generation sequencing. Polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), competitive RT-PCR, real-time PCR, real-time RT-PCR, nuclease protection assay, in situ hybridization, DNA microarray. and a method of providing information, which is measured by one or more methods selected from the group consisting of Northern blot.
In claim 6,
In step c), if the measured expression level of the CD109 gene or protein is higher than the control group, squamous pancreatic cancer is diagnosed or the prognosis of squamous pancreatic cancer is judged to be poor.