研究背景：混合型肝细胞癌－胆管细胞癌（CHC）是一种罕见的原发性肝癌，其同时具有肝细胞肝癌(HCC)和肝内胆管细胞癌(iCCA)的临床病理特点。然而，关于独立肿瘤中同时存在肝细胞肝癌和胆管细胞癌成份的分子机制尚未清楚，因此，我们试图研究混合型肝细胞癌－胆管细胞癌中两种肿瘤成份之间的关系。

实验方法：我们的研究共纳入了75例肝癌患者，其中包括15例混合型肝细胞癌－胆管细胞癌，32例肝细胞肝癌和28例肝内胆管细胞癌，我们对所有的患者进行了免疫组化，包括肝细胞样标志物（Hep和GPC3），胆管细胞样标志物（CK7和CK19）。其中，我们选择了7例肝混合细胞癌样本分别进行肝细胞肝癌，胆管细胞癌成份及正常肝组织的显微切割，然后提取DNA进行全外显子测序。我们对测序结果进行了错义突变，拷贝数变异，驱动基因，高频突变基因，易感基因及乙肝病毒整合的分析，进行了克隆分析以进一步明确两种肿瘤成份之间的关系。为了进一步研究混合型肝细胞癌－胆管细胞癌的细胞起源，我们对所有样本进行了免疫组化，标志物为（EpCAM和c-kit），然后我们根据EpCAM的表达情况进行了相关生存分析。

实验结果：我们通过形态学及免疫组化进一步确认了混合型肝细胞癌－胆管细胞癌的诊断，根据DNA降解程度最终入选了7例CHC患者进行了外显子测序。对于全外显子测序结果，我们将肝细胞肝癌和胆管细胞癌成份共有的体细胞突变定义为共有突变，而各自独有的突变则定义为独有突变。我们发现CHC样本中存在大量独有的体细胞单核苷酸变异（SNV）（33.1%－86.4%）及独有拷贝数变异（79.3%－97.3），这些结果表明CHC具有明显的肿瘤异质性。然而，在CHC的肝细胞肝癌和胆管细胞癌中存在大量共有错义SNV及体细胞拷贝数变异，同样两种成份之间也发现了大量共同的错义突变（包括SNV,插入缺失），这些表明肝混合细胞癌中两种不同肿瘤成份为同一起源，克隆分析也进一步支持了该结论。我们发现一些突变的基因，如VCAN, ACVR2A及FCGBP等与干细胞状态的维持及分化相关，参与了干细胞多能性调节通路及WNT, Notch通路，这些突变基因可能导致了混合癌中肝细胞肝癌和胆管细胞癌表型的分化。再者，不同的肝癌类型EpCAM表达情况差异明显，其中肝混合细胞癌80%的阳性表达，CK19(+)及CK19(-)的肝细胞肝癌阳性表达分别为66.7%和17.2%，肝内胆管细胞癌为71.4%。生存分析表明EpCAM与肝癌患者的预后呈负相关关系。

实验结论：全外显子测序分析表明混合型肝细胞癌－胆管细胞癌为单克隆起源，其可促进原发性肝癌基于肿瘤细胞起源的分子分型。再者，混合型肝细胞癌－胆管细胞癌明显的肿瘤内部异质性鼓励我们进行多区域取样测序从而发现肿瘤发生过程中共有的驱动突变基因，从而让靶向药物的筛选变得更加精准而高效。

BACKGROUND: Hepato-Cholangiocarcinoma (CHC) is a rare subtype of primary liver cancer with clinicopathological features of both hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA). However, molecular mechanisms underlying the co-existence of HCC and iCCA components in a single tumor remain elusive. Herein, we aim to investigate the relationship of these two tumor components in CHC.

METHODS: Seventy-five liver cancer patients including 15 with CHC, 32 with HCC, and 28 with iCCA were enrolled in this study. Immunohistochemical staining (IHC) was performed for all samples using hepatocytic markers (Hep and GPC3) and biliary epithelial markers (CK7 and CK19). Seven CHC samples were selected for laser microdissection isolating HCC, iCCA, and the adjacent noncancerous liver tissues. DNA extracted from HCC and iCCA were subjected for whole exome sequencing (WES). Nonsynonymous mutations, somatic copy number variations (CNVs), driver mutations, significantly mutated genes, as well as hepatitis B virus (HBV) integrations were analyzed. To further study the progenitor cell features of CHC, IHC staining with stem cell markers (EpCAM and c-kit) was conducted. In addition, we performed a Kaplan-Meier survival analysis to evaluate the correlation between protein levels of such stem cell markers and mortality.

RESULTS: CHC was further confirmed by IHC staining. As to WES results, we defined somatic mutations shared by HCC and iCCA as ubiquitous mutations, whereas mutations distinct in each component as private mutations. We found that CHC samples contain substantial private somatic single nucleotide polymorphism （SNV）(ranging from 33.1% to 86.4%) as well as private somatic copy number variants (CNVs) (ranging from 79.3% to 97.3%), indicative of intratumor heterogeneity (ITH) of CHC. However, large amount of ubiquitous nonsynonymous mutations and CNVs were overlapped in HCC and iCCA samples, suggesting the monoclonal origin of CHC. Clonal analyses also supported the similar conclusion. Mutated genes identified herein e.g. VCAN, ACVR2A and FCGBP that related to the stem cell status and cell differentiation are speculated to contribute to distinct differentiation of HCC and IHC within CHC. Moreover, with regard to the progenitor cell feature analysis, EpCAM was positive in 80% of all CHC samples. Among HCC patients, EpCAM positivity in CK19 (+) vs. CK19 (-) samples is 66.7% and 17.2%, respectively. And 71.4% of iCCA samples had positive EpCAM staining. Kaplan-Meier curve showed that EpCAM expression was associated with poor prognosis of patients with liver cancer.

CONCLUSIONS:

  Exome sequencing analyses are suggestive of the monoclonal origin of CHC, which may promote the molecule classification of primary liver cancer on the basis of cell origin. In addition, the substantial intratumor heterogeneity noted in CHC urges multiregional sequencing analysis to find the common driver mutations that playing more important role in carcinogenesis, thus make target drugs selection more precision and effective.