Single-cell RNA-seq highlights intra-tumoral heterogeneity
and malignant progression in pancreatic ductal
adenocarcinoma
ABSTRACT
Pancreatic ductal adenocarcinoma(PDCA) is the most common type of pancreatic cancer featured with high intra-tumoral heterogeneity and poor prognosis.
胰腺导管癌是高肿瘤内异质性和低预后胰腺癌最普遍的一种癌症。
To comprehensively delineate the PDAC intra-tumoral heterogeneity and the underlying mechanism for PDAC progression, we employed single-cell RNA-seq(scRNA-seq) to acquire the transcriptomic atlas of 57,530 individual pancreatic cells from primary PDAC tumors and control pancreases, and identified diverse malignant and stromal cell types, including tow ductal subtypes with abnormal and malignant gene expression profiles respectively, in PDAC.
为了全面描绘胰腺导管癌肿瘤内的异质性和胰腺导管癌进展的潜在调控机制，作者利用单细胞技术从胰腺癌病人和对照组获取了57,530个胰腺细胞，鉴定了多元化的恶性和基质细胞，包括二种异常导管亚型细胞和在胰腺癌中各自的表达基因图谱。
We found that the heterogenous malignant subtype was composed of several subpopulations with differential proliferative and migratory potentials.
我们发现异质的恶性细胞亚群由几种有着不同增值和迁移潜性的亚群组成。
Cell trajectory analysis revealed that the components of multiple tumor-related pathways and transcription factors(TFs) were differentially expressed along PDAC progression.
细胞轨迹分析揭示：在胰腺导管癌的进展中，多种肿瘤相关的通路和转录因子在表达。
Furthermore, we found a subset of ductal cells with unique proliferative features were associated with an inactive state in tumor-infiltrating T cells, providing novel markers for the prediction of antitumor immune response.
我们还发现了，一种导管细胞的亚群有着独特的增值能力其和肿瘤浸润中T细胞的失活紧密联系。
Together, our findings provide a valuable resource for deciphering
我们提供有价值的资源来证明这一些列发现。
Introduciton
Although surgical resection remains the main option for PDAC treatment, only 10%-15% of newly diagnosed patients are eligible.
胰腺癌预后差，动手术后生存低.
Unfortunately, none of these findings have been translated into clinical use, mainly due to the very limited knowledge about their potential role during PDAC progression, whereas most patients already at advanced stages at the time of diagnosis.
基因组、转录组、表观组在胰腺癌方面做的贡献，对临床影响小，对胰腺癌进展的研究知识太少，病人在诊断时往往是晚期。
群测序手段被使用并且有一些进展，但是面对分析肿瘤细胞内的异质性仍有局限性。
The overall progress in identifying actionable diagnostic markers and therapeutic targets is till largely hindered due to the limitation of bulk profiling technologies in capturing inra-tumoral heterogeneity.
单细胞技术使用的进展，肿瘤微环境的发现。
Apart form the malignant cells, tumor mass also contains macrophages, T cells and fibroblasts, etc., forming tumor microenvironment(TME) supporting tumor progression.
巨噬细胞，T细胞， 成纤维细胞等。
The transcription profiles of a total of 57,530 cells from 24 primary PDCA tumors and 11 control pancreases were acquire.
24个胰腺癌病人和11个对照组胰腺细胞共 57,530个细胞的转录图谱被获取。
We found that PDCA tumor mass is highly heterogeneous and composed of diverse malignant and stromal cell types.
胰腺癌是高异质性的，并且由多种恶性和基质细胞类型组成。
In addition, malignant ductal subtype could be distinguished by featured gene expression profile and was observed to contain highly proliferative and migratory subpopulations.
恶性导管细胞亚型可以由特征基因辨别，被观察到属于高增值性和转移性的亚群。
We further identified a list of novel gene expression changes that affected several know cancerous pathways, and suppressed tumor-related T cell activation that is associated with clinical pathological features.
我们找出了一些表达基因，它们与已知的一个肿瘤通路相关并且抑制T细胞的活性，这也和临床的病理特征关联。
Thus, our findings will improve our current understanding about the mechanism for PDAC progression, and are potentially valuable in providing novel prognosis markers for PDAC.
所以，我们当前的发现可以提高我们对胰腺癌进展机制的理解和潜在地为胰腺癌预后提供有价值的生物标志物。
RESULTS
Single-cell expression atlas and cell typing in PDAC tumors and control samples
To explore the cellular diversity in PDCA, we generated single-cell RNA-seq profiles from 24 PDCA tumor samples and 11 control pancreases without any treatment.
24个病人，11个对照
To explore the cellular composition of tumors, we applied principle component analysis on variably expressed genes across all cells and identified 10 main variably expressed genes across all cells and identified 10 main clusters including type 1 ductal, type 2 ductal, acinar, endocrine, endothelial, fibroblast, stellate, macrophage, T and B cells.
为了探索肿瘤细胞的分子组成，在所有的表达细胞上，我们应用主成分分析，辨认出了10种主要的表达基因，和辨认了10个细胞群，包括，类型1导管细胞，类型2导管细胞，胰腺泡细胞，内分泌细胞，上皮细胞，纤维细胞，星状细胞，巨噬细胞，T 细胞和 B 细胞。
We then generated cluster-specific marker genes by performing differential gene expression analysis of define the identity of each cell cluster.
通过差异基因表达分析产生了每个细胞群的特异表达marker
In most cases, well-know cell type makers were used to identify cell clusters, such as KRT19 for ductal cells, PRSS1 for acinar cells, etc.
大多数情况下，细胞的maker 被用来鉴别细胞群，如 KRT19是导管细胞，PRSS1是腺泡细胞。
Conversely, type 2 ductal cells showed 4.3-fold more than type 1 (11,315/2,646) in PDAC samples, indicating that type 2 ductal cells are significantly expanded in PDACs.
导管类型2 细胞比导管类型1 细胞高出 4.3倍，说明导管类型2细胞在胰腺癌中显著扩增。
CNV landscape distinguishing malignant ductal cells in PDACs.
To define malignant cells, we calculated large-scale chromosomal copy number variation(CNV) in each cell type based on averaged expression patterns across intervals of the genome. We found that type 2 ductal cells exhibited remarkably higher CNV levels than type 1 ductal cells and other types of cells.
为了定义恶性细胞，基于基因组间隔在平均表达表达模式下的我们计算了大规模下染色体拷贝数目变化。发现，类型2导管细胞比类型1导管细胞和其它细胞表现为更高水平的拷贝数变化。
[图片上传失败...(image-bca376-1577371954773)]
[图片上传失败...(image-d549de-1577371954773)]
We next examined gene expression patterns in two types of ductal cells in PDAC and identified 3107 differentially expressed genes.
我们在胰腺癌病人中对二种导管细胞进行了表达基因的差异表达基因分析，找出了3107种差异表达基因。
The functional enrichment analyses showed that genes up-regulated in type 2 cells were mainly enriched for cancer-related functions, such as cell proliferation, migration and hypoxia, further supporting the malignant state of this subtype.
功能富集分析显示， 在类型2导细胞中上调的基因主要和癌症功能相关，如细胞增殖，迁移，缺氧，进一步显示了这种亚型的恶性程度。
[图片上传失败...(image-210711-1577371954773)]
In contrast, genes expressed at higher level in type 1 than in type 2 cells were related to normal pancreatic functions, including digestion, pancreatic secretion and bicarbonate transport, supporting that type 1 cells still possess a certain degree of ductal cell functions compared to malignant type cells.
相比之下，在类型1 而不是在类型2 中高表达的基因和正常的胰腺功能相关，包括消化，肾上腺素分泌，碳酸氢盐传输。说明和恶性类型2 相比，类型1细胞还拥有着一定程度的导管细胞功能。
We then performed immunohistochemistry(IHC) staining of markers( MUC1, FXYD3 for type 2 ductal cells and AMBP for type 1 ductal cells) in PDAC and control samples to validate these 2 types of ductal cell.
我们之后进行免疫组化对胰腺癌和对照样本markers进行染色标记去验证这二种细胞类型。
We observed the AMBP-positive cells are mainly present in the ductal structure with normal cell features. In contrast, MUC1 or FXYD3-positive cells showed typical neoplastic cell features and were absent in control samples.
观察到 AMBP阳性细胞主要在正常的导管细胞中呈现， 相比之下，MUC1 或FXYD3 阳性细胞主要在肿瘤细胞中而在对照样本中缺失。
Gene expression pattern in ductal cells during PDAC progression.
Pseudo-time of ductal cells with abnormal( type 1 ) and malignant(type 2 ) gene expression profiles was reconstituted.
F3.a
异常细胞（类型1）和恶性细胞（类型2）表达基因的伪时序分析。
[图片上传失败...(image-f13105-1577371954773)]
We further analyzed the gene expression patterns of all genes along the trajectory of PDAC progression and identified 2299 genes with dynamic expression changes.
进一步分析所有胰腺癌进展中的所有表达细胞，鉴别出了共2299个动态表达的基因。
We further clustered these genes into 4 abnormal expression patterns(P1-P4) and 4 malignant patterns(P5-P8) with specific expression patterns.
我们将这些基因分为异常表达群P1-P4,恶性表达群P5-P8
F3.b
[图片上传失败...(image-793f98-1577371954773)]
In particular, a large portion of genes in responsible for cell proliferation and migration were remarkably activated at the late stage of tumor progression.
在癌症进展中的晚期大部分细胞增殖和迁移的功能被激活。
F3.b
We also detected the activation of multiple key regulators and TFs that participate in the tumorigenesis of PDAS.
F3.c
在癌症进展期中检测到了许多关键调控因子和转录因子被激活参与胰腺癌肿瘤的进展。
[图片上传失败...(image-6ec17d-1577371954773)]
Distinct subgroups in malignant ductal cells
The type 2 ductal cells were further divided into 7 subgroups based on t-SNE analysis.
基于t-SNE分析，类型二导管细胞被分为7个亚群
F3.e
[图片上传失败...(image-571e27-1577371954773)]
Since proliferation was found to be the main feature of PDAC, we then performed functional enrichment analysis for each subgroup and found that the unique functions of subgroup 7 were related to cell cycle and cell proliferation, which is further supported by the specific expression of proliferation markers MKI67, TOP2A and cell cycle markers CCNB1 and CCNB2.
由于增殖是胰腺导管癌主要的特征，我们为每一个亚群进行了功能富集分析，发现这一特征属于亚群7， 它和细胞周期以及细胞增殖相关，进一步被增殖的特异表达marker MKI67, TOP2A, 以及细胞周期markers CCNB1 和 CCNB2 所支持。
F3.f
[图片上传失败...(image-b58533-1577371954773)]
[图片上传失败...(image-2f7c2d-1577371954773)]
[图片上传失败...(image-57755c-1577371954773)]
Gene expression pattern analyses in TCGA PAAD samples based on malignant ductal markers
The malignant ductal cell markers were used to cluster the PAAD tumor samples.
在恶性导管细胞中的marker 被用来给GCTA中的胰腺癌样本的群体细胞分群。
F4.a，b
[图片上传失败...(image-b76538-1577371954773)]
[图片上传失败...(image-cf7ee9-1577371954773)]
被分为四个群，增殖的导管细胞markers 在群2和三中高表达，群4当中表达少。群4是其他类型的胰腺癌，群1-3是胰腺导管癌。恶性导管markers 能够明显区别胰腺导管癌和其它胰腺癌。
We used unsupervised non-negative matrix factorization(NNMF) clustering and divided PDAC patients into 3 subgroups according to the expression level of proliferative makers.
根据增殖marker的表达水平，用无监督非抑制性矩阵分解聚类把 PDAC病人划分为3个子群。
F4.c
We found that patients in group 3 showed a high abundance of proliferative ductal markers,
We used unsupervised non-negative matrix factorization(NNMF) clustering and divided PDAC patients into 3 subgroups according to the expression level of proliferative makers.
根据增殖marker的表达水平，用无监督非抑制性矩阵分解聚类把 PDAC病人划分为3个子群。
F4.c
[图片上传失败...(image-639f79-1577371954773)]
We found that patients in group 3 showed a high abundance of proliferative ductal markers, together a significantly lower survival rate compared to other groups.
我们发现群3显示了高程度的增殖性的导管细胞marker, 和其它群比有着更低的生存率。
F4.d
[图片上传失败...(image-6131da-1577371954773)]
Next, we sought to identify drugs that can pharmacological target the proliferative ductal cells and found that CDK1, PLK1 and AURKA can be served as therapeutic targets of PDAC by using their specific drugs.
接下来，我们寻找确定的药物可以在药理学上靶向增殖导管细胞的，发现了靶向CDK1, PLK1, 和 AURKA 的药物可以用他们作为胰腺癌治疗。
F4.e
[图片上传失败...(image-acb78c-1577371954773)]
We next compared the TCGA transcriptome data between group 3 and other two groups to determine the differential expression genes, 759 up-regulated and 946 down-regulated gens were detected in group 3 patients.
我们接下来比较了TCGA在群3和其他二组中的转录数据去找差异表达基因，发现了759个上调的基因和946个下调的基因。
F5.a
[图片上传失败...(image-aac68f-1577371954773)]
Gene ontology(GO) functional enrichment analysis revealed that genes enriched for terms such as cell cycle, DNA replication were highly up-regulated, whereas the downregulated genes were mainly enriched for T cell selection and lymphocyte activation.
基因本体功能富集揭示， 基因富集的术语为细胞周期，DNA复制是上调基因，而下调基因主要是，T 细胞选择和淋巴细胞激活术语。
F5.b
[图片上传失败...(image-546678-1577371954773)]
We then analyzed the expression pattern of the above differentially expressed genes in our single-cell data.
我们之后在我们的单细胞数据中分析了以上的基因表达模式，
We found that the majority of genes were highly expressed in certain specific cell typ. Specifically, ,many up-regulated genes were expressed in type 2 ductal cells, while down-regulated genes were mainly present in macrophage and T cells, suggesting that dysregulated ductal proliferation and immune response occur concurrently in the TME.
我们发现在特定细胞类型中，大部分细胞基因高表达，
特别的是更多的上调基因在类型2导管细胞中表达，下调基因更多的在巨噬细胞和T细胞中表达,说明了在TME阶段中，导管增殖的失调和免疫应答的共发生。
F5.c
[图片上传失败...(image-22e620-1577371954773)]
Inactivation of T cells in PDAC patients with high abundance of proliferative ductal markers.
We observed that samples with high level of proliferative ductal markers conversely have low expression level of T cell markers.
我们观察到增殖性导管细胞markers 在cluster3 中高表达，但是有着低表达的T 细胞markers.
F5.d
[图片上传失败...(image-40127b-1577371954773)]
We then calculated the CD8 T cell activation score and found that the group with high proliferative ductal cell exhibit significantly lower T cell activation score.
我们计算CD8 T 细胞活性分数，发现，T细胞活性和增殖性导管细胞的表达显著相关。
F5.e
[图片上传失败...(image-6c4e35-1577371954773)]
In support, we found an inverse spatial relationship: very limited Ki67+ ductal cells with high T cell infiltration vs high portion of ki67+ ductal cells with rare T cell infiltration.
我们发现了有差异的空间关系，有限的ki67阳性导管细胞有着高 T 细胞渗透和高的ki67导管细胞有着低T细胞渗透。
F5.f
[图片上传失败...(image-f6785b-1577371954773)]
We then compared these subtype signature scores in the three groups of patients with high, medium and low abundance of proliferative markers and found that the squamous subtype score was highest in group 3 patients and immunogenic subtype score was highest in group 1 patients.
我们之后比较了这三个组病人（高，中，低增殖性markers）的子类型细胞的 标记分数，发现麟型的在类型分数是最高的在组三的病人中，免疫型的在组一中是最高的。
F5.g
[图片上传失败...(image-6dc5d5-1577371954772)]