Chapter Four - Immune Contexture, Immunoscore, and Malignant Cell Molecular Subgroups for Prognostic and Theranostic Classifications of Cancers

https://doi.org/10.1016/bs.ai.2015.12.002Get rights and content

Abstract

The outcome of tumors results from genetic and epigenetic modifications of the transformed cells and also from the interactions of the malignant cells with their tumor microenvironment (TME), which includes immune and inflammatory cells. For a given cancer type, the composition of the immunological TME is not homogeneous. Heterogeneity is found between different cancer types and also between tumors from patients with the same type of cancer. Some tumors exhibit a poor infiltration by immune cells, and others are highly infiltrated by lymphocytes. Among the latter, the architecture of the TME, with the localization of immune cells in the invasive front and the center of the tumor, the presence of tumor-adjacent organized lymphoid aggregates, and the type of inflammatory context, determines the prognostic impact of the infiltrating cells. The description and the understanding of the immune and inflammatory landscape in human tumors are of paramount importance at different levels of patient's care. It completes the mutational, transcriptional, and epigenetic patterns of the malignant cells and open paths to understand how tumor cells shape their immune microenvironment and are shaped by the immune reaction. It provides prognostic and theranostic markers, as well as novel targets for immunotherapies.

Introduction

Tumors start with early dysplasic lesions, followed by malignant transformation into a locally invasive tumor and subsequent distant metastases. In this review, we will provide evidence for immune reactions within tumors and present the cellular composition and organization of the tumor microenvironment and the molecular mediators involved. We will summarize the genetic and epigenetic modifications of malignant cells and how tumor cells may influence their microenvironment. We will underline the novel prognostic markers and therapeutic targets provided by studies of the cross talk between tumors and their microenvironment. We will particularly focus on colorectal cancers (CRCs) and renal cell cancers (RCCs).

A fundamental aspect of cancer immunology is that the immune system, mainly through its adaptive arm, is able to identify and eliminate host cells undergoing malignant transformation and control tumor growth. Although the idea of a cancer-protecting function of the immune system was proposed more than a century ago (Ehrlich, 1909), it remained controversial during the 20th century (Stutman, 1974, Stutman, 1979). This idea is now widely accepted and considered a hallmark of cancer (Hanahan & Weinberg, 2011).

Schreiber and colleagues proposed the 3E theory (Dunn, Old, & Schreiber, 2004), standing for Elimination, Equilibrium, and Escape. The idea is that most cells undergoing malignant transformation are recognized by immune cells and eliminated. Sometimes, some tumor cells proliferate fast enough to compensate for immune-mediated elimination, and the overall tumor size remains at equilibrium. In the third stage, tumor cells escape elimination from the immune system and grow uncontrolled. The challenge in studying how immune cells are able to eliminate cancer cells is that clinically detectable tumors have, by definition, managed to grow from a single mutated cell to a mass of derived malignant cells.

Epidemiological observations in immunodeficient patients indirectly hinted at a protective role of the immune system against cancer occurrence. In particular, it was observed in the 1980–90 decade that patients suffering from HIV/AIDS have higher chances of developing Kaposi's sarcoma (Haverkos and Drotman, 1985, Pape et al., 1983, Schreiber and Podack, 2009), a virus-associated cancer. Transplant recipients, who receive immunosuppressive therapies to avoid transplant rejection, have a consistently higher risk of developing this malignancy (Farge, 1993). Strikingly, transplant recipients are also at higher risk of developing other solid tumors, such as brain (Curtis et al., 1997) and other central nervous system malignancies (Curtis et al., 1997), thyroid (Curtis et al., 1997), bone (Curtis et al., 1997), colon (Curtis et al., 1997), lung (Kasiske, Snyder, Gilbertson, & Wang, 2004), prostate (Kasiske et al., 2004), stomach (Kasiske et al., 2004), esophagus (Kasiske et al., 2004), pancreas (Kasiske et al., 2004), ovary (Kasiske et al., 2004), breast (Kasiske et al., 2004), melanoma (Curtis et al., 1997, Kasiske et al., 2004), leukemia (Kasiske et al., 2004), hepatobiliary tumors (Curtis et al., 1997, Kasiske et al., 2004), cervical and vulvovaginal cancers (Kasiske et al., 2004), testicular (Kasiske et al., 2004), bladder (Kasiske et al., 2004), kidney (Kasiske et al., 2004), nonmelanoma skin cancers (Kasiske et al., 2004), oral cavity cancers (Curtis et al., 1997), and non-Hodgkin lymphoma (Kasiske et al., 2004).

Following these observations, several studies showed an increase in carcinogen-induced tumors numbers in immunodeficient mice compared to wild-type mice (Schreiber and Podack, 2009, Shankaran et al., 2001). Importantly, it was observed that a knockout of the Rag2 gene, which encodes the Recombination-activating gene 2, is sufficient to induce this effect. Rag2 is necessary for the V(D)J recombination and therefore for the survival of B and T cells in the periphery, supporting the idea that the adaptive arm of the immune system is responsible for the control of tumor occurrence. Direct evidence was later obtained using in vivo two-photon imaging: C57BL/6 mice were injected with the syngenic EL4 cell line expressing the exogenous OVA protein and latter injected with OT-I T cells specific for OVA antigens. In situ and in vivo imaging of the CD8 protein and the activity of CASP3, an apoptosis marker, showed that the cytotoxic lymphocytes CD8+ T cells were actively killing tumor cells (Breart, Lemaître, Celli, & Bousso, 2008). The authors also noted that the elimination was a slow process, taking on average 6 h, and proposed that the amount of tumor-targeting CD8+ T cells could be a limiting factor in the control of tumor growth.

The adaptive immune system is able to discriminate between self and nonself. The distinction is clear in the case of invasive pathogens, but cancer cells are host's transformed cells and could therefore be thought as belonging to the self. The malignant transformation of normal cells involves mutations (Vogelstein et al., 2013) in pathways controlling cell proliferation, resistance to apoptosis, and other fundamental characteristics of cancers (Hanahan & Weinberg, 2011). These mutations encode epitopes that are not expressed by the host's normal cells and could therefore be considered as nonself, marking the malignant cell for elimination by the adaptive immune system (Fig. 1). Mutations encoding mutated peptides capable of eliciting an adaptive immune response are known as neoantigens. CD8+ T cells clones specific for tumor antigens were found against the cyclin-dependent kinase 4 gene CDK4 (Wölfel et al., 1995) or mutated β-catenin (Robbins et al., 1996) in melanoma, and the tumor-expressed MHC Class II HLA-A2 gene in RCC (Brändle, Brasseur, Weynants, Boon, & Van den Eynde, 1996).

Interestingly, host's nonmutated proteins were also found to be targeted by the adaptive immune system. In particular, proteins that are expressed in MHC Class I-negative cells, such as sperm cells or trophoblasts, and aberrantly expressed by cancer cells, are able to elicit immune responses. The first example in human was an epitope of the MAGEA1 (van der Bruggen et al., 1991) testis-restricted protein aberrantly expressed in melanoma. Antigens were later found in proteins encoded by the MAGE family of antigens and other germline-specific genes, such as BAGE1, GAGE1, XAGE1B, CTAG2, CTAG1, and SSX2, in melanoma, and also lung, colorectal, breast, and prostate carcinomas (Coulie et al., 1994). More surprisingly, some proteins constitutively expressed by nonmalignant cells were found to also elicit immune responses. These proteins are usually overexpressed in tumor cells, leading to a TCR-mediated activation of the corresponding specific lymphocytes (Coulie et al., 1994), while their expression is too low in normal cells to reach the threshold leading to T-cell activation. Examples of such antigens include the “prostate-specific antigen” protein in prostate cancer (Correale et al., 1997, Olson et al., 2010), the HER2/neu antigen (Fisk, Blevins, Wharton, & Ioannides, 1995) encoded by the amplified ERBB2 gene in breast and ovarian cancers, as well as the Melan-A protein in melanoma (Coulie et al., 2014, Kawakami, Eliyahu, et al., 1994a, Kawakami, Eliyahu, et al., 1994b).

A more straightforward class of tumor-associated antigens are the peptides associated with carcinogenic viruses. These peptides remain expressed in the transformed cells and can elicit immune responses, as shown with Human Papilloma Virus infection in head and neck squamous cell carcinoma (van der Burg & Melief, 2011), Epstein–Barr virus in Hodgkin's lymphoma, nasopharyngeal carcinoma, NKT lymphoma, and Burkitt's lymphoma (Long, Parsonage, Fox, & Lee, 2010). Altogether, these examples show that although tumor cells are derived from normal cells, they express antigens that can be recognized as nonself.

Tumor control resulting from genetic and epigenetic modifications of the transformed cells takes place in the TME which contains immune cells. Thus, in several situations in which a precancerous state can be individualized and studied, a shift from an immunological pattern with a Th1 orientation to a proinflammatory TME correlates with tumor invasiveness and aggressiveness. It is illustrated in cervical carcinoma in which high expression of genes encoding Th1 cytokines is evidenced in cervical in situ neoplasia, whereas IFN-γ expression is lost, and the expression of proinflammatory cytokines IL-6 is high in invasive and aggressive cervical carcinoma (Tartour et al., 1998). A change toward a Th2-type cytokine pattern has also been reported in the evolution from intraepithelial neoplasia to invasive carcinoma (Bais et al., 2005). A similar shift has been described in pancreatic cancer, where there is a decrease in the density of CD8+ T cells and mature dendritic cells (DCs) from low-grade premalignant lesions into invasive ductal adenocarcinoma (Hiraoka et al., 2006, Hiraoka et al., 2011).

It is well established that chronic inflammation in a tumor favors the outgrowth of malignant cells. Inflammation is associated with many carcinogenic events. For instance, mutation of the RET oncogene in thyrocytes is sufficient to induce papillary thyroid carcinoma and is accompanied with up-regulation of proinflammatory genes (Borrello et al., 2005). Mutation of RAS oncogenes in an ovarian cancer cell line xenograft in athymic mice is associated to the production of IL-8 (CXCL8) by the tumor cells, resulting in increased angiogenesis (Sparmann & Bar-Sagi, 2004). Activating mutation of the IL-6-receptor signal transducer gp130 (IL-6ST) triggers an inflammatory program in hepatocytes, favoring adenoma formation which can later develop into hepatocarcinoma (Rebouissou et al., 2009). Exogenous conditions can also trigger inflammatory signals. Helicobacter pylori is associated with gastric cancer (Forman et al., 1991, Nomura et al., 1991, Parsonnet et al., 1991) and mucosa-associated lymphoid tissue (MALT) lymphoma (Wotherspoon et al., 1993, Wotherspoon et al., 1991), viral infection with Hepatitis B or C viruses is associated with the development of hepatocellular carcinoma (Grivennikov, Greten, & Karin, 2010), and tobacco smoke exposure triggers chronic lung inflammation which favors lung carcinoma development (Takahashi, Ogata, Nishigaki, Broide, & Karin, 2010). It has been proposed that 15% of all diagnosed cancers are caused by infection (Coussens & Werb, 2002). Inflammation is a very broad concept that refers to immune response-promoting conditions and can designate almost every immune cell population. In the context of tumor immunology, it usually refers to immune cell populations or cytokines that promote cancer growth. It is paradoxical as inflammation is associated with the promotion of immune responses, including adaptive immune responses. Some authors described adaptive immune responses as “good inflammation” and the protumor signals as “bad inflammation” (Mantovani, Allavena, Sica, & Balkwill, 2008) (Fig. 1). In this chapter, we use “inflammation” to refer to the latter. In the following section, we detail some of tumor-promoting effects associated with inflammation through direct effect on cancer cells.

The role of inflammation in cancer is particularly illustrated in inflamed tissues secondary to persisting infectious agents, whatever their types. Thus, as stated earlier, bacteria such as H. pylori support the emergence of gastric cancer (Salama, Hartung, & Müller, 2013) and MALT lymphoma (Parsonnet et al., 1994), without directly transforming the malignant cell. Viruses, such as Hepatitis B and C, create liver inflammation that may result in some cases in malignant hepatocellular carcinoma (Arzumanyan, Reis, & Feitelson, 2013). A similar phenomenon is observed in Kaposi's sarcoma, induced by Herpes Virus type 8, often in the context of immunodepressed HIV-positive patients (Haverkos and Drotman, 1985, Pape et al., 1983). The case of papilloma viruses and Epstein–Barr virus is more complex since they are oncogenic for the infected cells and also create an inflammatory TME (Castle et al., 2001). The association of parasites, such as Schistosoma haematobium and Opisthorchis viverrini/Clonorchis sinensis with higher incidence of bladder cancer and cholangiocarcinoma, respectively, reveals similar inflammation-favored carcinogenesis (Sripa et al., 2012, Vennervald and Polman, 2009). In addition to chronic infection, cancer-favoring inflammation can also be the consequence of external stimuli, such as smoke, asbestos, silica, alcohol, aflatoxin, or chronic inflammatory diseases, such as gastritis, pancreatitis, inflammatory bowel disease, thyroiditis, and osteomyelitis, which favor cancer emergence in the corresponding organ (reviewed in Fridman et al., 2014).

Preexisting inflammation can directly promote carcinogenesis through mutational effects. In particular, reactive oxygen species (ROS) have been shown to directly modify DNA sequences (Colotta, Allavena, Sica, Garlanda, & Mantovani, 2009). Inflammatory reactions can also lead to epigenetic modifications of the DNA, albeit no causal effect on carcinogenesis was observed (Hahn et al., 2008).

Inflammatory factors, such as IL-6, IL-1β, or IL-22, induce inflammatory responses leading to the activation of transcription factors NF-κB and STAT3 in cancer cells. These pathways can lead to the expression of antiapoptotic molecules (Al Zaid Siddiquee and Turkson, 2008, Fan et al., 2002), such as BCL2 (Mantovani et al., 2008) and BCL-X (BCL2L1) (Elinav et al., 2013) which promotes survival of malignant B cells in follicular lymphoma (Tsujimoto, Finger, Yunis, Nowell, & Croce, 1984). Mutation of STAT3 in nontumorigenic immortalized fibroblasts is able to induce successful xenograft in nude mice (Bromberg et al., 1999), and activation of the IL-6-STAT3 pathway has been shown to have procarcinogenesis activities in many malignancies (Elinav et al., 2013), including pancreatic ductal adenocarcinoma (Fukuda et al., 2011) and intraepithelial carcinoma (Lesina et al., 2011), and lung (Gao et al., 2007) and gastric (Bronte-Tinkew et al., 2009) adenocarcinomas. NF-κB activation promotes cancer cells’ survival, for instance by inhibiting TNF-mediated apoptosis (Micheau & Tschopp, 2003).

Inflammation also triggers proliferation of malignant cells (Adachi et al., 2006), notably by increasing the expression of the cyclins B, D1, and D2 (Bollrath et al., 2009, Elinav et al., 2013, Grivennikov et al., 2009).

When the tumor reaches a certain size, oxygen supply becomes too limited for it to diffuse in all the areas of the tumor and hypoxic conditions arise. Inflammation is one of the mechanisms subverted by tumors to sustain neoangiogenesis and increase blood supply. Inflammatory mediators released by either malignant, hematopoietic or other stromal cells can increase local angiogenesis. For instance, IL-1β released by cancer cells simultaneously triggers angiogenesis and the recruitment of inflammatory cells to the tumor bed in an MYC-dependent pancreatic β-cell cancer mouse model (Shchors et al., 2006). Mast cells in turn can promote angiogenesis (Mantovani et al., 2008, Soucek et al., 2007). While tumor cells have been known for a long time to induce local angiogenesis, the contribution of stromal cells, and in particular innate immune cells, to neoangiogenesis is now established (Bingle et al., 2006, Du et al., 2008, Giraudo et al., 2004, Lin et al., 2006, Rivera and Bergers, 2015, Shojaei et al., 2007), notably through the release of vascular endothelial growth factors (VEGFs), epidermal growth factor, fibroblast growth factor 2, TNF-α, TGF-β, platelet-derived growth factors (PDGFs), placental growth factor, neuropilin-1, and IL-8 (CXCL8) (Rivera & Bergers, 2015). Among innate immune cells, macrophages are the most abundant in tumors (Bindea et al., 2013) and were shown to control angiogenesis in a mouse model of breast cancer (Lin et al., 2006) through their ability to secrete VEGF-A (Lin et al., 2007) (Fig. 1). Inhibition of the colony-stimulating factor 1-receptor (CSF1R), which is required for macrophage differentiation and survival, was shown to inhibit neovascularization in a glioma model (Pyonteck et al., 2013).

Modulating inflammation is a potential treatment modality for the treatment of some cancers. First, prophylactic use of antiinflammatory agents such as aspirin has been shown to be associated with a reduction in CRC incidence (Algra and Rothwell, 2012, Chan et al., 2007, Flossmann and Rothwell, 2007), a finding that has then been extended to breast (Gierach et al., 2008), esophageal, gastric, prostate, and lung cancers (Thorat & Cuzick, 2013). Second, suppressing inflammatory signals in highly inflammatory clinically detectable cancers can lead to a halt of tumor growth (Balkwill, 2009), potentially synergizing with cytotoxic agents (Green, Ferguson, Zitvogel, & Kroemer, 2009).

DNA instability in tumors, associated to selective pressure from immune-mediated elimination of tumor cells, leads to the emergence of escape mechanisms in tumor cells. These can either be directly induced by the tumor cells or through other cell populations of the TME.

CD8+ T cells are the main effector of antitumor immune responses, but their activity requires the presentation of peptides on target cells by Class I MHC molecules. These molecules are heterodimers consisting of two subunits, one of them being encoded by a single gene, B2M. Some tumor cells harbor inactivating mutations in the B2M gene, abrogating expression of any functional Class I MHC molecule and inhibiting the activity of CD8+ T cells (Bernal, Ruiz-Cabello, Concha, Paschen, & Garrido, 2012). NK cells can sense the loss of Class I MHC expression and exert contact-dependent cytotoxicity. However, tumor-infiltrating NK cells have been reported to display inhibited phenotypes compared to NK cells populating nonmalignant tissue distant from the tumor. Indeed, NK cells infiltrating non-small cell lung cancer (NSCLC) tumors were shown to downregulate the expression of the activating receptors NKp30, NKp80, CD16, NKG2D, and DNAM-1 and consistently to have lower degranulation and cytotoxic capacities ex vivo (Platonova et al., 2011), possibly owing to TGF-β signaling (Donatelli et al., 2014). Similar results have been observed in melanoma (Pietra et al., 2012) and breast cancer (Mamessier et al., 2011).

Loss of Class I MHC expression is a striking illustration of tumor adaptation to immune pressure (Angell, Lechner, Jang, LoPresti, & Epstein, 2014). However, in most cases, immune escape occurs in a more subtle and slow way, through the selection of peptides with low immunogenicity. Mouse models have successfully illustrated this phenomenon, as 3-methylcholanthrene-induced tumors from immunocompetent mice have higher xenograft success rates in syngenic fully immunocompetent mice compared to those grown in Rag−/− mice lacking T and B lymphocytes (Shankaran et al., 2001). In human melanoma, vaccination based on the tumor-expressed gp100 peptide induced a reduction in tumor gp100 expression compared to prevaccination samples (Riker et al., 1999).

As other immune cells, CD8+ T cells express, either constitutively or after activation, inhibitory receptors (also known as immune checkpoints) that regulate their activity. Tumors are able to subvert this mechanism and avoid CD8+ T cell-mediated elimination. A notable example includes the expression of PD-1 (PDCD1) ligands by tumor cells. PD-1 is an inhibitory receptor expressed by a variety of immune cells, including T cells. It can bind to either PD-L1 (CD274) or PD-L2 (PDCD1LG2), which results in reduced cytotoxic capacities (Blank et al., 2004), proliferative capacities (Blank et al., 2004, Carter et al., 2002), and response to TCR stimulation (Freeman et al., 2000). In physiological conditions, IFN-γ the major cytokine of the Th1 axis, produced by activated Th1 and CD8+ cells, has been shown to induce PD-L1 and PD-L2 expression by surrounding cells (Lee et al., 2005, Mazanet and Hughes, 2002). Immune checkpoints include other molecules such as LAG-3, CTLA-4, and TIM-3. Contact-dependent mechanisms can even mediate T-cell elimination, such as Fas-ligand expressed by tumor cells which can bind to Fas expressed by surrounding lymphocytes, inducing their apoptosis (Chappell and Restifo, 1998, O'Connell et al., 1996), but the importance of this effect is still debated (Igney et al., 2000, Igney and Krammer, 2005).

Tumor cells can also release soluble factors that result in suppression of T-cell responses in the microenvironment. TGF-β and VEGF-A orientate the functionality of surrounding hematopoietic cells toward a suppressive phenotype. Other factors, such as the antiinflammatory interleukin IL-10, Galectin-1 (LGALS1) (Gabrilovich, Nadaf, Corak, Berzofsky, & Carbone, 1996), gangliosides (Rabinovich, Gabrilovich, & Sotomayor, 2007), and prostaglandin E2 (PGE2) (Chahlavi et al., 2005, Hahne et al., 1996), are implicated in the direct inhibition or elimination of infiltrating T cells.

Many suppressive pathways involve multiple cell populations from the TME. A critical step in the TCR-mediated activation of T lymphocytes is its interaction with an antigen-presenting cell (APC), mostly DCs and macrophages, but also some B-cell subsets. In addition to presenting Class I and Class II MHC-bound peptides to CD4+ and CD8+ T cells, respectively (primary activation signal), APCs deliver co-stimulatory signals. These signals depend on which ligand/receptor couples are engaged between the T cell and the APC. The type of signal depends mostly on the activation status of the APC. Notably after sensing of danger signals, DCs mature and consequently convey co-stimulatory signals to the T cell. On the other hand, immature DCs presenting a specific antigen to a T cell will transduce tolerogenic signals, by repressing the T cell's ability to respond to future TCR stimulation, apoptosis, or differentiation to a Treg phenotype.

Several studies reported a defective presentation by DCs infiltrating tumors. First, the maturation of monocytes to DCs is dampened in favor of a macrophage differentiation through the action of IL-6 and macrophage colony-stimulating factor 1 (CSF1) (Chomarat et al., 2000, Menetrier-Caux et al., 1998). Second, the maturation process of DCs is inhibited (Alcalay and Kripke, 1991, Chaux et al., 1996, Engelhardt et al., 2012, Gabrilovich et al., 1996, Giraldo et al., 2015, Tas et al., 1993) by several mechanisms. Molecular mediators of the response to hypoxia pathway such as VEGF-A have been implicated in impairment of DC maturation (Fig. 1) through the inhibition of the inflammatory transcription factor NF-κB (Oyama et al., 1998). The theory of immunogenic cell death proposes that activation of DC mostly depends on the type of cell death tumor cells underwent before their uptake by phagocytes (Green et al., 2009). Markers of immunogenic cell death notably include the translocation of the chaperone calreticulin from the cytosol to the plasma membrane and the release of adenosine triphosphate and of high-mobility group box 1 (HMGB1) protein in the extracellular milieu. Mice models have shown that after cancer cell line injection in an immunocompetent host and clearance due to cytotoxic chemotherapy, subsequent rechallenge using the same cell line will lead to rejection only in the case of immunogenic cell death (Casares et al., 2005). The fact that myeloid cells in the TME exert tolerogenic roles led to the functional definition of myeloid-derived suppressor cells (MDSCs) (Gabrilovich & Nagaraj, 2009). Lack of consensual markers in humans hamper comprehensive analyses of the MDSC populations in human tumors, and a wide variety of factors have been implicated in their expansion and polarization, including VEGF, GM-CSF, G-CSF, M-CSF, gangliosides, prostaglandins, IFN-γ, complement C5a, TGF-β, interleukins IL-1b, IL-6, IL-10, IL-12, and IL-13, and chemokines CCL2, CXCL5, and CXCL12 (Gabrilovich & Nagaraj, 2009). However, the mechanisms by which they exert their suppressive functions have been well studied, notably the depletion of l-arginine by the enzyme Arginase 1 which leads to impaired T-cell proliferation (Ochoa et al., 2007, Rodriguez et al., 2005).

Lymphoid cells and notably regulatory T cells (Treg) are involved in antigen-specific suppressive function. Upon TCR activation, Treg release the immunosuppressive cytokine IL-10 which downregulates Th1 cytokines and co-stimulatory molecules on APCs. Treg are differentiated under the influence of TGF-β and IL-2 stimulation through the IL-2 high-affinity receptor CD122-CD25 heterodimer. Other cytokines, notably VEGF-A, have been shown to induce regulatory polarization of CD4+ T cells (Terme et al., 2013).

Hypoxia is in general linked to increased immunosuppression. It directly inhibits T-cell responses, as hypoxic conditions inhibit IL-2 and IFN-γ release after TCR-mediated T-cell activation (Becker, Andersen, Schrama, & Thor Straten, 2013). In ovarian cancer, tumor cells response to hypoxia was shown to induce the expression of the chemokine CCL28 which attracts Treg (Facciabene et al., 2011). In mice, hypoxic area has been shown to favor the M2 polarization of macrophages (Movahedi et al., 2010). In mice, VEGF-A signaling was also shown to directly induce T-cell expression of PD-1 and other immune checkpoints, notably Tim-3 and CTLA-4 (Voron et al., 2015).

Fibroblasts can also modify T-cell responses, through several mechanisms. As major producers of the extracellular matrix, they control the trafficking of T cells from the invasive margin (IM) to the tumor stroma (Salmon et al., 2012). Fibroblasts can directly inhibit TNF-α and IFN-γ-mediated antitumor immunity (Kraman et al., 2010), hamper DC maturation (Séguier et al., 2013), and inhibit T-cell proliferation (Bocelli-Tyndall et al., 2006, Haniffa et al., 2007, Jones et al., 2007). They have been shown to constitutively express the immune checkpoint ligand PD-L1 (Pinchuk et al., 2008), and this expression is upregulated upon IFN-γ stimulation (Pinchuk et al., 2008).

Section snippets

The Immune Microenvironment of Different Cancer Subtypes

Once malignant cell transformation has occurred and tumors grow, invade locally, and spread in distant organs, they still interact with their TME, and particularly with immune and inflammatory cells in the organ where the primary tumors develop, but also in the organs invaded by metastatic cells.

The composition of the immunological TME is not homogeneous. Heterogeneity is found between different cancer types, and also between tumors from patients with the same type of cancers. Some tumors

Classification of Tumors in the Era of Omic Techniques

Tumor classifications have been established and continuously refined by clinicians and researchers. The goals are to integrate knowledge of the biology of different malignancies to improve patient's management. The accurate prediction of the evolution of a cancer is of crucial importance since the advent of cytotoxic chemotherapies, to avoid potentially damaging unnecessary treatments. More recently, the development of targeted therapies, with drugs that interfere with a particular feature of

The Tumor Microenvironment and the Molecularly Defined Subgroups of Human Cancers

The immunoscore is a technology usable in routine practice for identifying patients with potential favorable prognosis and responding to selected immunotherapies on the basis of T-cell densities. However, since it quantifies the ultimate result of the very complex tumor–stroma interaction network happening during cancer evolution, the immunoscore does not take into account the myriad of cellular partners that compose the tumor immune contexture. An immunohistochemical approach of such

Conclusions: From the Immune Contexture of Tumor Subgroups to Precision Medicine

The concept of immune contexture which takes into account the organization, location, density, and functional orientation of immune cells in the TME has changed the paradigm of cancer/host interactions. It provides a conceptual tool to understand the physiopathological mechanisms that support the clinical impact of various cells of the immune reaction. Moreover, it allowed the definition of a routine, robust, and quantitative “immunoscore” that is a major prognostic factor in the vast majority

Acknowledgments

We thank the members of teams 10, 13, and 15 of the UMR-S1138 Cordeliers Research Center, of the UMR-1162, of the UMR-S1147, and of the CIT program for their fruitful discussions and performed most of the work cited in this review.

Financial support: This work was supported by the “Institut National de la Santé et de la Recherche Médicale,” the University Paris-Descartes, the University Pierre et Marie Curie, the Institut National du Cancer (2009-1-PLBIO-07-INSERM 6-1, 2010-1-PLBIO-03-INSERM

References (583)

  • A. Busse et al.

    Immunomodulatory effects of sorafenib on peripheral immune effector cells in metastatic renal cell carcinoma

    European Journal of Cancer

    (2011)
  • Y. Adachi et al.

    Interleukin-6 induces both cell growth and VEGF production in malignant mesotheliomas

    International Journal of Cancer

    (2006)
  • T. Ahmadi et al.

    CD40 ligand-activated, antigen-specific B cells are comparable to mature dendritic cells in presenting protein antigens and major histocompatibility complex class I- and class II-binding peptides

    Immunology

    (2008)
  • K. Al Zaid Siddiquee et al.

    STAT3 as a target for inducing apoptosis in solid and hematological tumors

    Cell Research

    (2008)
  • J. Alcalay et al.

    Antigen-presenting activity of draining lymph node cells from mice painted with a contact allergen during ultraviolet carcinogenesis

    Journal of Immunology

    (1991)
  • R.B. Alexander et al.

    Helper T cells infiltrating human renal cell carcinomas have the phenotype of activated memory-like T lymphocytes

    Journal of Immunotherapy with Emphasis on Tumor Immunology

    (1995)
  • A. Algars et al.

    Type and location of tumor-infiltrating macrophages and lymphatic vessels predict survival of colorectal cancer patients

    International Journal of Cancer

    (2012)
  • A.A. Alizadeh et al.

    Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling

    Nature

    (2000)
  • F. Aloisi et al.

    Lymphoid neogenesis in chronic inflammatory diseases

    Nature Reviews Immunology

    (2006)
  • K. Al-Shibli et al.

    The prognostic value of intraepithelial and stromal innate immune system cells in non-small cell lung carcinoma

    Histopathology

    (2009)
  • K.I. Al-Shibli et al.

    Prognostic effect of epithelial and stromal lymphocyte infiltration in non-small cell lung cancer

    Clinical Cancer Research

    (2008)
  • M. Ammirante et al.

    B-cell-derived lymphotoxin promotes castration-resistant prostate cancer

    Nature

    (2010)
  • J. Ananiev et al.

    Prognostic significance of CD83 positive tumor-infiltrating dendritic cells and expression of TGF-beta 1 in human gastric cancer

    Hepato-Gastroenterology

    (2011)
  • T. Ando et al.

    Multiple fuzzy neural network system for outcome prediction and classification of 220 lymphoma patients on the basis of molecular profiling

    Cancer Science

    (2003)
  • T.E. Angell et al.

    MHC class I loss is a frequent mechanism of immune escape in papillary thyroid cancer that is reversed by interferon and selumetinib treatment in vitro

    Clinical Cancer Research

    (2014)
  • E. Angevin et al.

    Analysis of T-cell immune response in renal cell carcinoma: Polarization to type 1-like differentiation pattern, clonal T-cell expansion and tumor-specific cytotoxicity

    International Journal of Cancer

    (1997)
  • M. Anitei et al.

    Prognostic and predictive values of the immunoscore in patients with rectal cancer

    Clinical Cancer Research

    (2014)
  • T. Armaghany et al.

    Genetic alterations in colorectal cancer

    Gastrointestinal Cancer Research

    (2012)
  • A. Arzumanyan et al.

    Pathogenic mechanisms in HBV- and HCV-associated hepatocellular carcinoma

    Nature Reviews Cancer

    (2013)
  • G. Assié et al.

    Integrated genomic characterization of adrenocortical carcinoma

    Nature Genetics

    (2014)
  • S. Attig et al.

    Simultaneous infiltration of polyfunctional effector and suppressor T cells into renal cell carcinomas

    Cancer Research

    (2009)
  • M. Ayyoub et al.

    The frequent expression of cancer/testis antigens provides opportunities for immunotherapeutic targeting of sarcoma

    Cancer Immunity

    (2004)
  • A. Bachmayr-Heyda et al.

    Prognostic impact of tumor infiltrating CD8 + T cells in association with cell proliferation in ovarian cancer patients—A study of the OVCAD consortium

    BMC Cancer

    (2013)
  • C. Badoual et al.

    PD-1-expressing tumor-infiltrating T cells are a favorable prognostic biomarker in HPV-associated head and neck cancer

    Cancer Research

    (2013)
  • A.G. Bais et al.

    A shift to a peripheral Th2-type cytokine pattern during the carcinogenesis of cervical cancer becomes manifest in CIN III lesions

    Journal of Clinical Pathology

    (2005)
  • P. Balermpas et al.

    Tumor-infiltrating lymphocytes favor the response to chemoradiotherapy of head and neck cancer

    Oncoimmunology

    (2014)
  • F. Balkwill

    Tumour necrosis factor and cancer

    Nature Reviews Cancer

    (2009)
  • E. Barbera-Guillem et al.

    B lymphocyte pathology in human colorectal cancer. Experimental and clinical therapeutic effects of partial B cell depletion

    Cancer Immunology, Immunotherapy

    (2000)
  • I.B. Barsoum et al.

    A mechanism of hypoxia-mediated escape from adaptive immunity in cancer cells

    Cancer Research

    (2014)
  • S. Barton et al.

    Are we ready to stratify treatment for diffuse large B-cell lymphoma using molecular hallmarks?

    The Oncologist

    (2012)
  • E. Becht et al.

    Prognostic and theranostic impact of molecular subtypes and immune classifications in Renal Cell Cancer (RCC) and Colorectal Cancer (CRC)

    Oncoimmunology

    (2015)
  • E. Becht et al.

    Shaping of an effective immune microenvironment to and by cancer cells

    Cancer Immunology, Immunotherapy

    (2014)
  • J.C. Becker et al.

    Immune-suppressive properties of the tumor microenvironment

    Cancer Immunology, Immunotherapy

    (2013)
  • C.L. Behnes et al.

    Tumor-associated macrophages are involved in tumor progression in papillary renal cell carcinoma

    Virchows Archiv

    (2014)
  • M. Bernal et al.

    Implication of the β2-microglobulin gene in the generation of tumor escape phenotypes

    Cancer Immunology, Immunotherapy

    (2012)
  • B. Beuselinck et al.

    Molecular subtypes of clear cell renal cell carcinoma are associated with sunitinib response in the metastatic setting

    Clinical Cancer Research

    (2015)
  • L. Bingle et al.

    Macrophages promote angiogenesis in human breast tumour spheroids in vivo

    British Journal of Cancer

    (2006)
  • S.K. Biswas et al.

    Macrophage plasticity and interaction with lymphocyte subsets: Cancer as a paradigm

    Nature Immunology

    (2010)
  • K. Biswas et al.

    GM2 expression in renal cell carcinoma: Potential role in tumor-induced T-cell dysfunction

    Cancer Research

    (2006)
  • C. Blank et al.

    PD-L1/B7H-1 inhibits the effector phase of tumor rejection by T cell receptor (TCR) transgenic CD8 + T cells

    Cancer Research

    (2004)
  • Cited by (161)

    View all citing articles on Scopus
    View full text