Research On Exploration For Molecular Classification And Clinical Treatment Of Pancreatic Cancer Based On Establishment Of Pancreatic Organoid Bank

BackgroundPancreatic cancer is well known for its aggressiveness and poor prognosis.In recent years,despite advances in various aspects of research,survival of pancreatic cancer has not been improved significantly.Late diagnosis,rapid progress,and treatment resistance are main reasons for the morbid outcome.The limited understanding of underlying biological mechanism has further restricted the development of diagnosis and treatment in pancreatic cancer.The advent of next-generation sequencing technology has prompted us to have a better understanding of the intrinsic biological characteristics of cancers.After overcoming the problem of excessively abundant stroma by methods such as macrodissection and microdissection,pancreatic cancer has also accumulated many high-quality genomics and transcriptomics sequencing analysis studies.However,bioinformatics analysis and traditional experimental models could barely reflect the actual state of tumors.Firstly,it is difficult to functionalize specific gene mutation or signal pathway change under the regulation of complex signal networks within tumor cells.Secondly,although molecular classifications based on tumor tissue samples can help to acquire comprehensive understanding of heterogeneity in pancreatic cancer,it could hardly facilitate further recognition of potential functional difference and exploration of targeted therapeutic strategy.Finally,whether genomics or transcriptomics sequencing could fully reflect the individualized difference of pancreatic cancer patients and guide personalized treatment in clinical practice remains little known.With the development of modern medicine,patient-derived preclinical research models have received widespread attention.Among those,organoid has been an emerging hotspot and confirmed by many studies in terms of recapitulating tumor characteristics of individualized patient.As a basic research model,organoid can effectively replace traditional 2D cell lines for functional characterization of specific gene or pathway.Moreover,with high purity of tumor cell,organoid have unique advantage for investigation of molecular classification in pancreatic cancer and subsequent identification of functional difference as well as high-throughput chemical screening.In addition,organoid model can reflect chemotherapy sensitivity of patients to clinical treatment.In this study,we tried to establish a comprehensive pancreatic cancer organoid research platform and verify the consistency between pancreatic cancer organoid and original tumor with histopathology and genomics.Using the organoid bank,we further attempted to explore molecular classifications based on transcriptomics and chromatin accessibility following by exploration of potentially therapeutic chemicals.Finally,we tried to demonstrate the accuracy of organoid in reflecting individualized response of pancreatic cancer patients to clinical chemotherapy.Part Ⅰ Establishment and characterization of pancreatic organoid bankObjective To optimize the culture medium of pancreatic organoid,generate pancreatic organoid bank containing normal pancreas and various pathological subtypes of pancreatic cancer,and confirm the consistency between pancreatic cancer organoid and original tumor with histopathology and genomics.Methods 1.Surgical resection and fine needle biopsy samples of pancreatic tissue were used to test the culture medium protocol of organoid and generate pancreatic organoid bank.2.Hematoxylin-eosin staining and immunohistochemical staining were used to identify histopathology of pancreatic cancer organoids,in vivo transplantation tumors and original tumors.3.Genomic sequencing was used to identify the genomics characteristics of pancreatic cancer organoids,and comparison with genomics data of corresponding tumors and public database was performed.Results 1.The most suitable medium for pancreatic cancer organoid and normal pancreas organoid was determined by culture tests of pancreatic cancer tissues and normal pancreas tissue with 24 designed medium protocols.2.6 normal pancreas organoids and 83 pancreatic cancer organoids including 74 cases of ductal adenocarcinoma,1 case of adenosquamous carcinoma,3 cases of neuroendocrine tumor,1 case of acinar cell carcinoma,and 4 cases of intraductal papillary mucinous tumor were successfully generated with surgical resection and fine needle biopsy samples.The generation efficiency of pancreatic cancer organoids was 77.6%.3.The morphology of pancreatic cancer organoids was highly similar to in vivo transplantation tumors and original tumors,and immunohistochemical staining further showed the features of original tumors were also maintained in pancreatic cancer organoids and in vivo transplantation tumors.In particular,pancreatic neuroendocrine neoplasm organoids were identical to original tumor on the neuroendocrine markers SYN and CGA.4.High-quality genomics data was obtained by whole genome sequencing in 34 pancreatic cancer organoids and whole exome sequencing in 10 corresponding tumors.5.Comparison with genomics data of corresponding tumors and public database showed that pancreatic cancer organoids highly conserved the genomic characteristics of pancreatic cancer and were consistent with original tumors.Conclusion 1.Normal pancreas organoids and pancreatic cancer organoids including five pathological subtypes of ductal adenocarcinoma,adenosquamous carcinoma,neuroendocrine tumor,acinar cell carcinoma,and intraductal papillary mucinous tumor were successfully generated by optimizing the culture medium protocol.2.Pancreatic cancer organoids recapitulated the feature of original tumors on histopathology and were further conserved in transplanted tumors.3.Pancreatic cancer organoids maintained the genomic characteristics of pancreatic cancer and were highly consistent with corresponding tumors.Part Ⅱ Exploration for molecular classification of pancreatic cancer in transcriptomics and chromatin accessibility with pancreatic organoid bankObjective To explore molecular classification of pancreatic cancer in transcriptomics and chromatin accessibility with established pancreatic organoid bank,and subsequently screen for effective epigenetic chemicals.Methods 1.Transcriptomics sequencing data was obtained from pancreatic organoid bank and used for bioinformatics analysis.2.ATAC-seq was used to evaluate chromatin accessibility and performed with pancreatic cancer organoids and normal pancreas organoids.ATAC-seq data were used for bioinformatics analysis.3.Pancreatic cancer organoids were used for drug screening with epigenetic chemical library and further analyzed with chromatin accessibility subtypes.Results 1.By analyzing high-quality transcriptomics sequencing data of 43 pancreatic cancer organoids and 3 normal pancreas organoids,we identified 4 subtypes of pancreatic cancer organoids: classical,basal-like,classical/progenitor and newly defined glycolytic active.2.With TCGA public database,verification of the new pancreatic cancer transcriptomics classification demonstrated that 184 pancreatic cancer samples could be well classified into 4 subtypes using the differentially expressed genes of the new classification,among which patients with classical/progenitor subtype had the best prognosis,while all of the other three subtypes,including the newly defined glycolytic subtype,showed poor prognosis.3.Analysis of insert size,correlation of the accessibility in transcription start sites,quality score and chromatin status by the public database has confirmed the high-quality of ATAC-seq data from 41 pancreatic cancer organoids and 3 normal pancreas organoids.4.Further analysis of ATAC-seq data showed that pancreatic cancer organoids can be divided into 6 subtypes based on chromatin accessibility,and each subtype was enriched with different transcription factors.5.After analyzing the primary screening results of 283 epigenetic chemicals in 35 pancreatic cancer organoids,we selected 59 epigenetic chemicals which might had good inhibitory effect for secondary screen with 39 pancreatic cancer organoids.We found that many transcription factors enriched in chromatin accessibility subtypes could serve as biomarkers for the sensitivity of epigenetic chemicalsConclusions 1.Pancreatic cancer organoids could be divided into four subtypes at the transcriptomics level: classical type,basal-like type,classical/progenitor type,and glycolytic active type.2.The new transcriptomics classification of pancreatic cancer organoids could be verified in public database,and the prognosis of patients with classical/progenitor type was markedly better than the other three.3.Pancreatic cancer organoids could be divided into 6 subtypes at chromatin accessibility level,and each subtype is feature of enrichment of different transcription factors.4.Chromatin accessibility classification of pancreatic cancer organoids could be used to guide the individualized therapy by targeting specific transcription factor.Part Ⅲ Evaluation of individualized chemotherapy sensitivity in pancreatic cancer patients with organoid modelObjective To evaluate the individualized chemotherapy sensitivity of pancreatic cancer patients with organoid model and confirm with in vivo transplantation model.Methods 1.Pancreatic cancer organoids were used for chemotherapy sensitivity test.2.Clinical treatment and follow-up data of corresponding patients was collected.3.Combined the clinical data with chemotherapy sensitivity of corresponding organoids,patients were analyzed for survival benefit and drug effectiveness.4.In vivo transplantation model was used to verify the chemotherapy sensitivity of in vitro organoid.Results 1.After analyzing the primary screening results of 35 pancreatic cancer organoids in six chemotherapeutic drugs,we determined chemotherapeutic drugs and concentration gradients for secondary screening.2.According to the area under the curve of the dose response curve from the secondary screening results using 5 first-line chemotherapeutic drugs and 39 pancreatic cancer organoids,the organoid chemotherapy sensitivity was classified into three levels: sensitive,intermediate and resistant.3.The corresponding 39 pancreatic cancer patients were divided into 3 conditions based on clinical treatment: upfront radical surgery with adjuvant chemotherapy(31 cases),neoadjuvant therapy(2 cases),and upfront radical surgery without adjuvant chemotherapy(6 cases).Patients in the last group did not receive chemotherapy and were not included for further analysis.4.In condition of upfront radical surgery with adjuvant chemotherapy,according to the chemotherapy sensitivity of matched organoid and the actual chemotherapy regime of patients,we divided patients into sensitive group,intermediate group and resistant group.Survival analysis demonstrated that patients in sensitive group had significantly better recurrence-free survival than patients in intermediate group and resistant group,which was further confirmed in radiological examination data.5.In condition of neoadjuvant treatment,organoid DAC-36 generated from metastasis after failure of neoadjuvant therapy showed resistance to all used chemotherapeutic drugs in the chemotherapy sensitivity test;Chemotherapy sensitivity of organoid DAC-38,which was generated from fine needle biopsy sample,was highly consistent with the response of corresponding patient during neoadjuvant treatment.6.Compared with in vitro chemotherapy sensitivity,pancreatic cancer organoids DAC-18,DAC-20 and DAC-36 showed completely consistent results in in vivo transplantation model.Conclusions 1.Organoid model could be used to accurately evaluate the individualized chemotherapy sensitivity of pancreatic cancer patients and guide personalized adjuvant chemotherapy.2.Organoid model could be used to guide the neoadjuvant chemotherapy regime of pancreatic cancer patients and explore further treatment after neoadjuvant therapy failure.Summary1.In this study,by optimizing culture medium protocol,we generated a comprehensive pancreatic organoid bank including normal pancreas and 5 pathological subtypes of pancreatic cancer: ductal adenocarcinoma,adenosquamous carcinoma,neuroendocrine tumor,acinar cell carcinoma,and intraductal papillary mucinous tumor.2.Pancreatic cancer organoids were highly consistent with original tumors in histopathology as well as genomics and maintained the characteristics of pancreatic cancer confirmed by public database.3.Pancreatic cancer organoids were divided into 4 subtypes in transcriptomics: classical type,basal-like type,classical/progenitor type,and glycolytic active type,which could be further verified in public database.4.Pancreatic cancer organoids were divided into 6 subtypes in chromatin accessibility characterized by enrichment of distinct transcription factors,which could guide the individualized treatment targeting specific transcription factor.5.Pancreatic cancer organoids can be used to evaluate the individualized chemotherapy sensitivity of patients and guide postoperative adjuvant as well as neoadjuvant chemotherapy regimen.