5hmC修飾及腫瘤旁纖維母細胞調控胰腺癌之致瘤性

Abstract

胰腺癌之5年存活率僅8%，原因為缺乏有效的治療方法。目前的胰腺癌治療策略為針對局部早期腫瘤之手術切除，以及針對局部晚期或遠端轉移腫瘤之全身性化學療法。對於手術治療的病患而言，儘管癌症分期相似，病患之預後卻有顯著差異。因此，需發展術後之預後標記，以協助術後治療策略，針對預後差之病患給予積極治療。針對化學療法的部分，由於化學療法受限於藥物抗性及嚴重副作用，因此，需合併使用標靶治療，以延長病患存活率。根據以上改善胰腺癌治療之策略，本研究之目的包含兩部分。第一部分為開發預後標記，以協助手術後治療策略。第二部分為探索致瘤性之分子機制，有助於發展治療策略。 第一部分的研究，利用人源性腫瘤細胞異種移殖(patient-derived xenografts，簡稱PDX)反映癌細胞本身腫瘤生成的特性，我們將腫瘤區分為移植成功及移植失敗兩組。PDX之觀察發現，腫瘤細胞於NOD/SCID/IL2Rγ小鼠移植成功之患者預後，較移植失敗之患者預後差。但PDX形成時間過久，不適於作為臨床應用之預後標記。為了尋找生物標記以替代PDX，我比較11個免疫組織化學染色法適用之生物標記，其中5-hydroxymethylcytosine(5hmC)的高表現量與移植成功之腫瘤生成特性呈正相關，且高表現量也與較差的預後顯著相關。癌細胞5hmC異常分佈於CDX2及FOXA1等調控細胞譜系轉錄因子的啟動子及基因內區域，造成CDX2及FOXA1在腫瘤細胞的高表現量，導致胰腺癌細胞之異常分化及增加譜系可塑性，因而增加致腫瘤性。總結而言，癌細胞5hmC之高表現，為免疫組織化學染色法適用之預後標記，其分子機制為增加胰腺癌細胞之譜系可塑性及致腫瘤性。 胰腺癌周圍環繞纖維化基質，其中纖維母細胞是纖維化基質之主要細胞，且具有促進腫瘤細胞生長之特性。許多研究顯示，纖維母細胞分泌之細胞激素及生長因子，能促進腫瘤細胞生長及轉移，然而，纖維母細胞如何藉由直接接觸的方式影響腫瘤細胞仍未知。因此，第二部分的研究為探討胰腺癌細胞與纖維母細胞直接接觸以增加腫瘤惡性程度之分子機制。我觀察胰腺癌病患之組織，證實若腫瘤旁有較多活化之纖維母細胞，則病患預後差。此結果暗示，與纖維母細胞直接接觸能增加腫瘤惡性程度。我利用胰腺癌基因轉殖鼠證實纖維母細胞與腫瘤細胞能相互黏附。此外，以3D細胞培養的方式，與纖維母細胞直接接觸能促進腫瘤細胞生長。進而利用cytokine array發現直接接觸使纖維母細胞的Activin A表現量上升，以協助腫瘤生長。為了探討其分子機制，將癌細胞膜蛋白與纖維母細胞進行交叉鏈接(cross-linking)，並利用質譜分析，找到直接接觸之細胞膜蛋白。總結而言，本研究發現腫瘤細胞之細胞膜蛋白與纖維母細胞之KIAA0090接觸後，會誘發纖維母細胞分泌Activin A，進而促進腫瘤細胞之惡性程度。研究結果顯示，阻斷腫瘤細胞與纖維母細胞之細胞膜蛋白接觸，為具潛力之治療方法。

Pancreatic ductal adenocarcinoma (PDAC) is a devastating disease with lowest five-year survival rate (8%) among all common cancers due to lack of effective treatment. Current strategies for PDAC treatment included surgery for resectable tumors, and chemotherapy for locally advanced or metastatic disease. For patients underwent surgery, although resectable PDACs were around the same stage, a great variation in disease progression existed among patients after surgery. Therefore, a prognostic marker was needed to aid in post-operative treatment decision to provide intensive therapy for patients predicted with poor prognosis. For chemotherapy, its effect was limited by chemo-resistance and unacceptable side effects. Thus, discovery of therapeutic targets for developing targeted therapies could combine with chemotherapy to improve patient survival. According to the strategies to improve pancreatic cancer treatment, the specific aims of my study included two parts. Part I was to identify prognostic markers for post-operative treatment decision and part II was to discover the molecular mechanisms for tumorigenesis as potential therapeutic targets. In the first part of my study, I stratified PDAC based on the tumorigenic properties of PDACs determined by engraftment of patient-derived xenografts. The success of tumor engraftment in NOD/SCID/IL2Rγ null mice was significantly correlated to poor survival. However, xeno-engraftment as a prognostic marker was impractical due to time constraint to generate each xenograft. To search immunohistochemistry (IHC)-based biomarkers as surrogate for high tumorigenicity with prognostic values, 11 candidates of potentially clinical useful prognostic markers were selected. Among them, 5hmC content of the cancer cells was validated. Elevated 5hmC content positively correlated with in vivo tumorigenicity and poor prognosis in both primary and validation cohorts. Enrichment of cancer-associated 5hmC in promoters and gene bodies of CDX2 and FOXA1 was correlated to active transcription of these lineage-specific transcriptional factor genes. The result pointed out the potential role of 5hmC in modulating cellular differentiation to enhance tumor malignancy during PDAC progression. In summary, high tumor-associated 5hmC content, which defined a subpopulation of PDAC with high lineage plasticity and tumorigenic potential, was a prognostic IHC marker that provided a clinical basis for future management of PDAC. PDAC is characterized by desmoplastic stroma with large amount of fibroblasts which promote tumorigenesis. Although the tumor-promoting roles of paracrine factors from fibroblasts are well-studied, direct tumor-fibroblast contact mediated tumorigenesis is a mystery in PDAC. Therefore, the second part of my study focused on the role of tumor-fibroblast contact in tumorigenesis. Here, I found that pancreatic cancer patients with high tumor-adjacent myofibroblasts had poor prognosis. This result suggested that direct tumor-fibroblast contact leading to tumor malignancy. In addition, I proved that juxta-tumoral fibroblasts adhered to tumor cells in a transgenic mouse model of PDAC. Functionally, direct contact with fibroblasts was crucial for tumorigenesis in 3D culture system. To discover potential juxtacrine signaling for direct contact-mediated tumorigenesis, secreted proteins were compared between tumor-fibroblast indirect versus direct co-cultured conditions by cytokine array. Activin A was highly secreted by fibroblasts in the direct co-cultured system to enhance contact-mediated tumorigenesis. Crosslinking of tumor and fibroblast membrane proteins followed by mass spectrometry identified KIAA0090, which bound with tumor membrane protein leading to Activin A production. In summary, contact with KIAA0090 on fibroblasts induced juxtacrine Activin A secretion to promote tumorigenesis. The results suggested that blockage of membrane protein interactions between tumor cells and fibroblasts for Activin A secretion might serve as a potential therapeutic strategy.