癌相关成纤维细胞（CAF）是肿瘤微环境中的关键参与者。在这里，我们描述了乳腺癌中具有不同特性和激活水平的四个CAF亚群。两种肌纤维母细胞亚组（CAF-S1，CAF-S4）在三阴性乳腺癌（TNBC）中差异性积聚。 CAF-S1成纤维细胞通过多步机制促进免疫抑制环境。通过分泌CXCL12，CAF-S1吸引CD4 + CD25 + T淋巴细胞并通过OX40L，PD-L2和JAM2保留它们。此外，CAF-S1通过B7H3，CD73和DPP4增加T淋巴细胞存活并促进其分化成CD25 ^ HighFOXP3 ^ High。最后，与CAF-S4相比，CAF-S1增强调节性T细胞抑制T效应子增殖的能力。这些数据与富含CAF-S1的TNBC中的FOXP3 + T淋巴细胞积累一致，并显示CAF子集如何促成免疫抑制。
Carcinoma-associated fibroblasts (CAF) are key players in the tumor microenvironment. Here, we characterize four CAF subsets in breast cancer with distinct properties and levels of activation. Two myofibroblastic subsets (CAF-S1, CAF-S4) accumulate differentially in triple-negative breast cancers (TNBC). CAF-S1 fibroblasts promote an immunosuppressive environment through a multi-step mechanism. By secreting CXCL12, CAF-S1 attracts CD4+CD25+ T lymphocytes and retains them by OX40L, PD-L2, and JAM2. Moreover, CAF-S1 increases T lymphocyte survival and promotes their differentiation into CD25^HighFOXP3^High, through B7H3, CD73, and DPP4. Finally, in contrast to CAF-S4, CAF-S1 enhances the regulatory T cell capacity to inhibit T effector proliferation. These data are consistent with FOXP3+ T lymphocyte accumulation in CAF-S1-enriched TNBC and show how a CAF subset contributes to immunosuppression.
引言：乳腺癌（BC）是女性常见的癌症，尽管试图提供有效的治疗方法，但仍是西方国家癌症相关死亡的主要原因。尽管随着早期检测和治疗方法的改进，BC的死亡率总体上缓慢下降，但在晚期癌症的情况下仅取得有限的成功。基于组织病理学分析，BC被定义为分类为三种主要亚型：鲁米那（Lum），HER2和三阴性（TN），这些基因表达谱已被补充（Perou等，2000; Sorlie等，2001）。肿瘤是复杂的生态环境，受到许多影响或增强遗传上皮改变影响的基质因子的影响。尽管正常成纤维细胞抑制肿瘤形成（Dotto等，1988），但癌相关成纤维细胞（CAF）增强肿瘤表型，特别是癌细胞增殖和侵袭，新血管发生，炎症和细胞外基质（ECM）重塑（Costa等。，2014; Gascard和Tlsty，2016; Gentric等，2016）。虽然CAFs是否阻止或驱动癌细胞侵袭仍存在争议（Ozdemir等，2014），但CAFs的肿瘤促进活性已被广泛描述（Olumi et al。，1999; Allinen et al。，2004; Orimo et al。 ，2005; Kalluri和Zeisberg，2006; Erez等，2010; Toullec等，2010; Hammer等，2017）。在人类乳腺肿瘤中，基质肌成纤维细胞（即α-平滑肌肌动蛋白[αSMA]阳性成纤维细胞）的丰度与侵袭性腺癌相关并预测人类疾病复发（Toullec等，2010; Benyahia等，2017） 。此外，已显示CAF有助于抗药性（Straussman等，2012; Paulsson等，2017）并降低抗肿瘤免疫性（Kraman等，2010; Tan等，2011; Feig 2013; Denton等，2014; Ruhland等，2016; Yang等，2016）。在免疫活性小鼠中的几项研究表明FAP（成纤维细胞活化蛋白α1）阳性CAF驱动免疫抑制和对抗PD-L1免疫治疗的抗性。然而，尚不清楚CAF介导的免疫抑制功能是否与人类肿瘤相关，如果是的话，涉及的机制是什么。
INTRODUCTION: Breast cancer (BC) is a frequent cancer in women and remains a major cause of cancer-associated death in western countries, despite attempts to provide effective therapies. Even though the mortality rate for BC is overall slowly declining with the improvement of both early detection and therapies, only limited success has been achieved in case of advanced cancers.Based on histopathological analysis, BC has been defined as a heterogeneous disease classified into three main subtypes: luminal (Lum), HER2, and triple-negative (TN), which have been complemented by gene expression profiling (Perou et al., 2000; Sorlie et al., 2001). Tumors are complex ecologies that are affected by numerous stromal factors that dampen or enhance the effects of genetic epithelial alterations. While normal fibroblasts suppress tumor formation (Dotto et al., 1988), cancer-associated fibroblasts (CAFs) enhance tumor phenotypes, notably cancer cell proliferation and invasion, neo-angiogenesis, inflammation, and extracellular matrix (ECM) remodeling (Costa et al., 2014; Gascard and Tlsty,2016; Gentric et al., 2016). Although it remains controversial whether CAFs prevent or drive cancer cell invasion (Ozdemir et al., 2014), tumor-promoting activities of CAFs have been widely described (Olumi et al., 1999; Allinen et al., 2004; Orimo et al., 2005; Kalluri and Zeisberg, 2006; Erez et al., 2010; Toullec et al., 2010; Hammer et al., 2017). In human breast tumors, the abundance of stromal myofibroblasts (i.e., α-smooth muscle actin [αSMA]-positive fibroblasts) is associated with aggressive adenocarcinomas and predicts human disease recurrence (Toullec et al., 2010; Benyahia et al., 2017). In addition, CAFs have been shown to contribute to drug resistance (Straussman et al., 2012; Paulsson et al., 2017) and to reduce anti-tumor immunity (Kraman et al., 2010; Tan et al., 2011; Feig et al., 2013; Denton et al., 2014; Ruhland et al., 2016; Yang et al., 2016). Several studies in immunocompetent mice showed that FAP (fibroblast activation protein α1)-positive CAFs drive immunosuppression and resistance to anti-PD-L1 immunotherapy. Yet it remains unclear whether this CAF-mediated immunosuppressive function is relevant in human tumors and if so, what are the mechanisms involved.
尽管CAF是最突出的基质成分，但它们在人类癌症中的异质性的特征还远未完成。已经分别研究了几种标记，如αSMA，FAP，整合素β1/ CD29，S100-A4 / FSP1（成纤维细胞特异性蛋白1），PDGFRβ（血小板衍生生长因子受体-β）和CAV1（小窝蛋白1）过去的几年。 事实上，许多研究使用αSMA来染色人类肿瘤中的肌成纤维细胞，并显示它们在不良预后BC中积累（Toullec等，2010）。此外，最近，在胰腺癌中已经鉴定了具有不同αSMA水平的两种CAF亚群，其中一种是肌纤维母细胞性的，另一种是促炎性的（Ohlund等，2017）。除了αSMA之外，高基质PDGFRβ表达与更短的BC患者存活相关（Paulsson等，2014）。此外，FAP显示在BC的基质中大量表达。这种表达与临床病理因素无关（Tchou et al。，2013），或者与长期存活相关（Ariga et al。，2001）。最后，虽然在患者存活方面存在一些矛盾的信息（Rudland等，2000; Lee等，2004; Goetz等，2011; Simpkins等），但在基质中证实了CAV1或FSP1在间质中的表达的临床意义2012）。第一项研究分析了αSMA，PDGFRβ和S100A4 / FSP1一起在小鼠胰腺和BC中进行，并显示它们在CAF中表现出差异表达（Sugimoto等，2006）。在这里，我们分析了人类BC中的CAF异质性并研究了这种异质性与免疫抑制的联系。
Although CAFs are the most prominent stromal components, characterizing their heterogeneity in human cancers is far from complete. Several markers, such as αSMA, FAP, integrin β1/CD29, S100-A4/FSP1 (fibroblast-specific protein 1), PDGFRβ (platelet-derived growth factor receptor-β), and CAV1 (caveolin 1) have been studied individually in the past years. Indeed, numerous studies used αSMA to stain myofibroblasts in human tumors and showed that they accumulate in BC of poor prognosis (Toullec et al., 2010). Moreover, recently, two CAF subpopulations with different levels of αSMA have been identified in pancreatic cancers, with one being myofibroblastic and the other one pro-inflammatory (Ohlund et al., 2017). In addition to αSMA, high stromal PDGFRβ expression was associated with shorter BC patient survival (Paulsson et al., 2014). Furthermore, FAP was shown to be abundantly expressed in the stroma of BC. Such expression either showed no link with clinicopathological factors (Tchou et al., 2013) or, in contrast, has been associated with longer survival (Ariga et al., 2001). Finally, the clinical significance of either CAV1 or FSP1 expression in stroma has been demonstrated in BC, although with some conflicting information on patient survival (Rudland et al., 2000; Lee et al., 2004; Goetz et al., 2011; Simpkins et al., 2012). A first study analyzing αSMA, PDGFRβ, and S100A4/FSP1 together was performed in mouse pancreatic and BCs and showed that they exhibit a differential expression in CAFs (Sugimoto et al., 2006). Here, we analyzed CAF heterogeneity in human BC and investigated the link of this heterogeneity with immunosuppression.
结果——
人类BC中四个CAF亚群的鉴定
为了定义人BC中的CAF异质性，我们首先使用多色流式细胞术（荧光激活细胞分选[FACS]）对CAF进行了详细的表征。为此，我们分别使用CD45，EPCAM和CD31标记来排除造血细胞，上皮细胞和内皮细胞（图1A），并对6种成纤维细胞标志物（FAP，整合素β1/ CD29，αSMA，S100- A4 / FSP1，PDGFRβ和CAV1）（图1B和1C）。所研究的新鲜BC样品（FACS前瞻性队列）包括在任何治疗前手术时的BC患者，有利于Lum BC患者纳入的条件（表S1）。我们观察到CAF构成异质细胞群（图1B）。根据CD29，FAP，αSMA，PDGFRβ，FSP1和CAV1的表达水平（图1B和1C），我们建立了门控策略，使我们能够区分BC中的四种不同的CAF亚群。这四个CAF亚群被称为CAF-S1（红色），CAF-S2（橙色），CAF-S3（绿色）和CAF-S4（蓝色）。接下来，我们通过无偏方法CytoSPADE（Qiu et al。，2011）分析了使用6种CAF标记获得的FACS数据，该方法将细胞组织成相关表型的等级。通过将该算法应用于FACS数据而构建的树证实了BC中存在四个CAF子集（图1D）。我们通过重复进行CytoSPADE分析来评估不同标记之间的冗余性（图S1A）。输入数据的变化（缺少一个基质标记）显着影响了树的全局结构（图S1A），表明这些标记不是多余的并且带来了附加信息。除了CAV1外，CAF-S1亚组的特征在于六种标志物的高表达，而CAF-S2成纤维细胞表现出所有这些标志物的低表达（图1E）。与CAF-S2和CAF-S3亚组相比，CAF-S1和CAF-S4成纤维细胞均表达αSMA，可以认为是肌成纤维细胞（图1C-1E）。此外，与其他CAF亚型相比，CAF-S1亚型是唯一对FAP呈阳性的细胞，并且CAF-S4细胞表现出CD29的最高表达。此外，CAF-S3和CAF-S4亚型均对PDGFRb和FSP1呈阳性。通过FACS在所有CAF亚群中CAV1显示非常低的染色（图1E）。总之，可将这些CAF亚群定义如下：CAF-S1：CD29 ^ Med FAP ^ Hi FSP1 ^ Low-HiαSMA^ HiPDGFRβ^ Med-Hi CAV1 ^ Low; CAF-S2：CD29 ^低FAP ^ Neg FSP1 ^ Neg-LowαSMA^ NegPDGFRβ^ Neg CAV1 ^ Neg; CAF-S3：CD29 ^ Med FAP ^ Neg FSP1 ^ Med-HiαSMA^ Neg-LowPDGFRβ^ Med CAV1 ^ Neg-Low; CAF-S4：CD29 ^ Hi FAP ^ Neg FSP1 ^ Low-MedαSMA^ HiPDGFRβ^ Low-Med CAV1 ^ Neg-Low。由于CAF-S2子集对于通过FACS检测的6个CAF标记物低或者阴性，我们不能排除CAF-S2细胞可能是另一种细胞类型。然而，对系列肿瘤切片进行免疫组织化学（IHC）分析（见下文）证实缺乏所分析的六种基质标记物的CAF-S2的存在。我们没有观察到CAF亚群与FACS前瞻性队列的临床特征之间的任何关联（表S2），很可能是因为该队列主要由Lum BC患者组成。相比之下，我们观察到四个CAF亚群在肿瘤中的差异性积累与其相应的并列型肿瘤相比较，由病理学家定义为健康组织（图S1B，1F和1G）。事实上，CAF-S1和CAF-S4亚群优先在肿瘤中检测到，而CAF-S3亚型与并列型肿瘤和CAF-S2在两个区室中均匀分布显着相关（图S1B，1F和1G）。因此，基于通过FACS检测几种成纤维细胞标记物，我们鉴定了人BC中四种不同的CAF亚群，其在肿瘤和近旁肿瘤中差异地累积。
RESULTS-
Identification of Four CAF Subsets in Human BC
To define CAF heterogeneity in human BC, we first performed a detailed characterization of CAFs using multicolor flow cytometry (fluorescence-activated cell sorting [FACS]).To do so, we used CD45, EPCAM, and CD31 markers to exclude hematopoietic, epithelial, and endothelial cells, respectively (Figure 1A), and performed the concomitant analysis of six fibroblast markers (FAP, integrin β1/CD29, αSMA, S100- A4/ FSP1, PDGFRβ, and CAV1) (Figures 1B and 1C).The fresh BC samples studied (FACS prospective cohort) included BC patients at time of surgery prior to any treatment, conditions in favor of Lum BC patient inclusion (Table S1).We observed CAFs constitute a heterogeneous cellular population (Figure 1B).We established a gating strategy that enabled us to distinguish four different CAF subpopulations in BC, according to the expression levels of CD29, FAP, αSMA, PDGFRβ, FSP1, and CAV1 (Figures 1B and 1C).These four CAF subpopulations were referred to as CAF-S1 (red), CAF-S2 (orange), CAF-S3 (green), and CAF-S4 (blue).We next analyzed FACS data obtained with the six CAF markers through an unbiased method, CytoSPADE (Qiu et al., 2011) that organizes cells into hierarchies of related phenotypes.The tree built by applying this algorithm to FACS data confirmed the existence of the four CAF subsets in BC (Figure 1D).We evaluated the redundancy between the different markers by repeating the CytoSPADE analysis in absence of each of the six markers (Figure S1A).Changes in the input data (lack of one stromal marker) significantly affected the global structure of the tree (Figure S1A), suggesting that these markers are not redundant and bring additive information.The CAF-S1 subset was characterized by high expression of the six markers except CAV1, while CAF-S2 fibroblasts exhibited low expression of all these markers (Figure 1E).In contrast to CAF-S2 and CAF-S3 subsets, both CAF-S1 and CAF-S4 fibroblasts expressed αSMA and could be considered as myofibroblasts (Figures 1C-1E).In addition, CAF-S1 subset was the only one to be positive for FAP and CAF-S4 cells exhibited the highest expression of CD29 compared with the other CAF subsets.Furthermore, both CAF-S3 and CAF-S4 subsets were positive for PDGFRb and FSP1.CAV1 exhibited very low staining by FACS in all CAF subsets (Figure 1E).In conclusion, these CAF subsets can be defined as follows: CAF-S1: CD29^Med FAP^Hi FSP1^Low-Hi αSMA^Hi PDGFRβ^Med-Hi CAV1^Low;CAF-S2: CD29^Low FAP^Neg FSP1^Neg-Low αSMA^Neg PDGFRβ^Neg CAV1^Neg;CAF-S3: CD29^Med FAP^Neg FSP1^Med-Hi αSMA^Neg-Low PDGFRβ^Med CAV1^Neg-Low;CAF-S4: CD29^Hi FAP^Neg FSP1^Low- Med αSMA^Hi PDGFRβ^Low-Med CAV1^Neg-Low.As the CAF-S2 subset was low or negative for the six CAF markers tested by FACS, we could not rule out that CAF-S2 cells might be another cell type.Still, immunohistochemistry (IHC) analyses on serial tumor sections (see below) confirmed the existence of CAF-S2 devoid of the six stromal markers analyzed.We did not observe any association between CAF subsets and clinical features of the FACS prospective cohort (Table S2), most probably because this cohort was mainly composed by Lum BC patients.In contrast, we observed that the four CAF subsets accumulated differentially in tumors compared with their corresponding juxta-tumors, defined as healthy tissues by pathologists (Figures S1B, 1F, and 1G).Indeed, both CAF-S1 and CAF-S4 subsets were preferentially detected in tumors, while CAF-S3 subset was significantly associated with juxta-tumors and CAF-S2 equally distributed in the two compartments (Figures S1B, 1F, and 1G).Thus, based on the detection of several fibroblastic markers by FACS, we identified four distinct CAF subsets in human BC that accumulate differentially in tumors and juxta-tumors.
联系客服
