PERSPECTIVES

OPINION

The immune contexture in human tumours: impact on clinical outcome

Wolf Herman Fridman, Franck Pags, Catherine Sauts-Fridman and Jrme Galon

(which we have termed the immune contexture (REF.8)) in large annotated collections of human tumours has allowed the identification of components of the immune contexture that are beneficial, as well as those that are deleterious, to patients. In addition, bioinformatics tools9,10 have permitted the identification of chemokines and cytokines that are involved in shaping the immune contexture11.
The immune contexture: characteristics Histopathological analyses of human tumours have provided evidence that variable numbers of infiltrating immune cells are found in different tumours of the same type, and are found in different locations within and around a tumour (FIG.1). Macrophages, mast cells, granulocytes and myeloidderived suppressor cells (MDSCs) are found in most cases infiltrating or surrounding tumour beds both in the core and at the invasive front of the tumour. It is well established that chronic inflammation and the presence of M2 macrophages favour tumour growth and spreading 12. Lymphocytes are not randomly distributed but are located in specific areas. Therefore, NK cells are found in the stroma and are not in contact with tumour cells. Bcells are mostly found in the invasive margin of growing tumours and in TLS that are adjacent to tumour beds13. Tcells, particularly CD8+ Tcells, may be located in the invasive margin but can also be in the tumour core. Few CD8+ Tcells are seen in TLS, which are similar to secondary follicles in lymph nodes that contain naive Tcells and memory Tcells, Bcells and mature dendritic cells. Indeed, immature dendritic

Abstract | Tumours grow within an intricate network of epithelial cells, vascular and lymphatic vessels, cytokines and chemokines, and infiltrating immune cells. Different types of infiltrating immune cells have different effects on tumour progression, which can vary according to cancer type. In this Opinion article we discuss how the context-specific nature of infiltrating immune cells can affect the prognosis of patients.
During the past decade, work that began with the mouse studies by R. D. Schreiber and colleagues14, and that culminated with the demonstration that infiltrating Tcells have a major effect on the clinical attributes of human cancer 5,6, has changed the field of tumour immunology. Indeed, a recent review by Hanahan and Weinberg included avoiding immune destruction (REF.7) as an emerging hallmark of cancer, in addition to the previously established hallmark of tumour-promoting inflammation. The potential effect of a patients immune system on clinical outcome is not only of academic interest but it also has important implications for the identification of prognostic markers, and markers that predict responses to chemotherapy and radiotherapy. A demonstration of the effect of the host immune response on tumour invasion,
Glossary
Cytotoxic Tcells
CD3+CD8+ effector Tcells with cytotoxic granules that contain perforin and granzymes, which are released on interaction with target cells expressing cognate antigen. This leads to the death of target cells by apoptosis. pro-inflammatory cytokines (IL1, IL6 and TNF), release reactive oxygen and reactive nitrate species and have a pro-inflammatory role. M2 macrophages secrete IL4, IL-10, IL13 and TGF and have an anti-inflammatory role, promote angiogenesis and favour tumour progression.

recurrence and metastasis has come from analyses of the insitu immune components and how these are organized within human tumours. Indeed, immune infiltrates are heterogeneous between tumour types, and are very diverse from patient to patient. All immune cell types may be found in a tumour, including macrophages, dendritic cells, mast cells, natural killer (NK) cells, naive and memory lymphocytes, Bcells and effector Tcells (including various subsets of Tcell: T helper cells, T helper 1 (TH1) cells, TH2 cells, TH17 cells, regulatory T (TReg) cells, T follicular helper (TFH) cells and cytotoxic Tcells). These immune cells can be located in the core (the centre) of the tumour, in the invasive margin or in the adjacent tertiary lymphoid structures (TLS). The analysis of the location, density and functional orientation of the different immune cell populations

Regulatory T (TReg) cells

Population of CD3+CD4+ Tcells that inhibit effector B and Tcells. Some produce cytokines with immunosuppressive activities; for example, IL10 and TGF. They have a central role in suppressing anti-self immune responses to prevent autoimmune diseases.

Dendritic cells
Cells that capture microorganisms or dead tumour cells and that process them to present antigen to Tcells in secondary or tertiary lymphoid organs. They express high levels of co-stimulatory molecules, which allows them to activate naive Tcells.

Memory Tcells
CD3+CD4+CD45RO+ and CD3+CD8+CD45RO+ cells that have encountered antigen and that respond faster and with increased intensity on antigenic stimulation compared with naive Tcells.

Tertiary lymphoid structures

(TLS). Ectopic lymphoid aggregates that are generated during the process of chronic immune stimulation and that exhibit the structural characteristics of secondary lymphoid organs.

Myeloid-derived suppressor cells High endothelial venules

Specialized venules that occur in secondary lymphoid organs, except the spleen. They allow continuous transmigration of lymphocytes as a consequence of the constitutive expression of adhesion molecules and chemokines at their luminal surface. (MDSCs). Heterogeneous population of polymorphonuclear and monocytic CD11b+GR1+ cells that inhibit Tcell activation.

T helper cells
Populations of CD3+CD4+ effector Tcells that secrete cytokines with differential activities. TH1 cells produce IL2 and IFN and favour cellular immunity (acting on CD8+ cytotoxic Tcells, NK cells and macrophages). TH2 cells produce IL4, IL5 and IL13 and favour humoral immunity (acting on Bcells). TH17 cells produce IL17A, IL17F, IL21 and IL22 and favour anti-microbial tissue inflammation (acting on epithelial and endothelial cells, fibroblasts and immune cells).

Naive Tcells
CD3+CD4+ and CD3+CD8+ cells that differentiate into effector Tcells (CD4+ T helper cells or CD8+ cytotoxic Tcells) in secondary lymphoid organs or TLS after stimulation with three signals: antigen, co-stimulatory molecules and cytokines.

Macrophages
Can be subdivided according to their cellular properties and cytokine secretion profiles. M1 macrophages secrete

298 | APRIL 2012 | VOLUME 12 2012 Macmillan Publishers Limited. All rights reserved

www.nature.com/reviews/cancer

F O C U S O N tu M our i M M uno L o G Y & i M M H er Punot ERSPE C T IA Vp EY S

a
DC Mast cell Macrophage

b Immune Parameters: positive association with survival contexture

Type Location
NK cell Tumour core MDSC Immature DC Tumour bed Invasive margin

CTLs (CD3+CD8+) Memory T cells (CD45RO+) Core of the tumour Invasive margin
Number of cells per mm2 10 100 1,000 10,000

Density
1 CD3+CT CD3+IM CD8+CT CD8+IM CD45RO+CT CD45RO+IM

Functional orientation
FDC TFH cell CTL Stroma TLS B cell

TH1 cell-associated factors (IFN, IL-12, T-bet and IRF1) Cytotoxic factors (granzymes, perforin and granulysin) Chemokines (CX3CL1, CXCL9, CXCL10, CCL5 and CCL2) TH17 cells, TReg cells and TH2 cells have a variable eect on survival, depending on tumour type

TLS

Presence and quality

Figure 1 | The immune contexture. a | Tumour anatomy showing the features of the immune contexture, including the tumour core, the invasive margin, tertiary lymphoid structures (TLS) and the tumour microenvironment. The distribution of different immune cells is also shown. b | Table depicting the parameters of the immune contexture that predict a good

prognosis. CT, core of the tumour; CTL, cytotoxic T lymphocyte; DC, dendritic cell; FDC, follicular dendritic cell; IFN , interferon ; IL12, interleukin12; IM, invasive margin; IRF1, interferon regulatory factor 1; Nature Reviews Cancer MDSC, myeloid-derived suppressor cell; NK cell, natural killer |cell; T H, T helper; TReg cell, regulatory Tcell.

cells are distributed in the tumour core, in contact with tumour cells or in the surrounding stroma. Mature dendritic cells concentrate in TLS, in close contact with naive Tcells13. TLS may be sites in which tumour-controlling primary and/or secondary immune responses are generated. The distribution of immune cells also varies between tumour types. All subsets of Tcells are present at the core and at the invasive margin of the tumour in colorectal cancer, non-small-cell lung cancer, melanoma, and head and neck cancers. In colorectal cancer, the proportion of tumours with high densities of CD4+ memory Tcells and CD8+ memory Tcells decreases with local tumour invasion, as assessed by the Tstage of the TNM classification (that is, the density is lower in T4stage tumours than in T1stage tumours). Conversely, the proportion of primary tumours with high infiltrates of CD4+ memory Tcells and CD8+ memory Tcells, particularly in the core, is lower in patients with tumours that recur. It has also been reported that Tcells are found only in the invasive margin in liver metastases of colon cancer 14. The fact that functional populations of immune cells are located in different areas of a tumour and that this varies between cancer types suggests that different immune cell populations may have different roles
NATURE REVIEWS | CANCER

in tumour control. Moreover, the variable density and location of these immune cells between tumours in different individuals with the same cancer type prompted the investigation of whether the immune contexture might affect clinicaloutcome.
Clinical impact of the immune contexture The effect on disease-free survival and overall survival. Correlations between the levels of immune cell infiltration of tumours and clinical outcome have been investigated in many cancers. TABLE1 and FIG.2 summarize the effect of Tcells on clinical outcome from 124 published articles. A strong lymphocytic infiltration has been reported to be associated with good clinical outcome in many different tumour types, including melanoma, and head and neck, breast, bladder, urothelial, ovarian, colorectal, renal, prostatic and lung cancer. Therefore, high densities of CD3+ Tcells, CD8+ cytotoxic Tcells and CD45RO+ memory Tcells were clearly associated with a longer disease-free survival (after surgical resection of the primary tumour) and/or overall survival (FIG.2). The fact that the density of CD8+ Tcells seems to correlate with poor prognosis in renal cell cancer, except when these Tcells are proliferating 15, is an intriguing result that deserves further analysis.

In contrast to the effects of cytotoxic Tcells and memory Tcells, analysis of the effect of CD4+ Tcell populations on clinical outcome has resulted in apparent contradictory results, and so their effects have been a matter of debate for the past decade. The case of TReg cells is a striking example of conflicting data that lead to difficult interpretation. There are different subpopulations of TReg cells (including, natural TReg cells, induced TReg cells and so on) but in most studies they are detected as a population of CD4+ Tcells that express phenotypic markers such as high levels of CD25 (also known as interleukin2 receptor subunit (IL2R), which is a subunit of the receptor for the Tcell-stimulating cytokine IL2) and the transcription factor forkhead box protein P3 (FOXP3). In fact, none of these markers is fully restricted to TReg cells; CD25 and FOXP3 are also expressed by activated effector Tcells, and there are also FOXP3 suppressor cells. Nevertheless, the pioneering report by Curiel etal.16, which demonstrated a correlation of intratumoural TReg cells and poor survival in ovarian cancer, was intuitively adopted as proving the deleterious effect of suppressor Tcells on clinical outcome. Indeed, several reports support this concept, and the high infiltration of TReg cells has been correlated with poor overall survival in breast cancer 17,18 and in hepatocellular
VOLUME 12 | APRIL 2012 | 299

2012 Macmillan Publishers Limited. All rights reserved

PERSPECTIVES
Table 1 | The association of immune cell infiltrates with prognosis in cancer
Cells
Melanoma Head and neck cancers Breast cancer Bladder cancer Ovarian cancer Oesophageal cancer Colorectal cancer Renal cell carcinoma Prostatic adenocarcinoma Lung carcinoma Pancreatic cancer Cervical cancer Anal squamous cell carcinoma Brain cancer Hepatocellular carcinoma Gastric cancer Medulloblastoma Merkel cell carcinoma Urothelial cell carcinoma Follicular lymphoma and Hodgkins lymphoma
TH, T helper; TReg cell, regulatory Tcell.

CD8+CD45RO+ Tcells
Good
103106

TH1 cells

TH2 cells

TH17 cells

TReg cells
None23,25 Poor107,108

Good30,109,110 Good
111114

None29 Good115,116 None117 Good123,124 Good128 Good

5,36,79

Good29,30 None26 Poor17,18 Good38

Good42 None117 Poor123 Good125 Good129 None

36

Good118,119 Good120122 Good126,127 Good

5,6,28,35,36,63,79,130148

Good33,61 Poor16 Good32,3436 None28 Poor150

Poor36,149

Good15 Poor15 Good151153 Good13,154157 None158 Good163

Good71

Good13 Poor164,165 Good

166

Poor159

Poor160162 Poor155 None21 None22,24

Good Poor20

167,168

Good

169

Poor Poor171

170

Poor19,20

Good171 Good173 Good

174

Good172

Good119 Good43 Good31,37 None27 Poor43

carcinoma19,20. However, several analyses of other cancer types have found no impact of TReg cell infiltration on survival2128. By contrast, other reports have demonstrated an inverse clinical correlation between the density of intratumoural TReg cells and the local immune control of tumours from patients with head and neck cancers29, and between the density of intratumoural TReg cells and overall survival in Hodgkins lymphoma, colorectal, breast, bladder, ovarian, and head and neck cancers3038. The reasons for these discrepancies are not evident; they may be due to the imperfect markers used to phenotype suppressive cells or due to technical differences, although they may also reflect the fact that tumours have different phenotypes, grow in different organs and are associated with unique host factors, and, therefore, tumours each have unique microenvironments. Because the antigen specificity
300 | APRIL 2012 | VOLUME 12

of tumour-infiltrating TReg cells has not been established in humans, these cells may have opposing effects depending on the tumour microenvironment: they may be deleterious when they block anti-tumour effector Tcells or beneficial when they decrease chronic inflammation. The analysis of other Tcell populations has also yielded apparently contradictory results, TH17 cells have been reported to be associated with poor prognosis in colorectal, lung and hepatocellular carcinoma or have been reported to predict better survival in some oesophagal and gastric cancers (FIG.2; TABLE1). These data are reminiscent of those from mouse models in which IL17 (a cytokine produced by TH17 cells) favours tumour growth in immunodeficient animals by supporting chronic inflammation, whereas it favours tumour rejection in immunocompetent animals by increasing an anti-tumour

immune response3941. TH2 cells, through the activation of Bcells or through the production of the immunosuppressive cytokine IL10, seem to be associated with aggressive tumours. However, this is not a general phenomenon, as TH2 cells are also associated with favourable outcome in Hodgkins lymphoma and breast cancer 42,43, which suggests a protective effect of antibodies in these diseases. In contrast to other T helpercell populations, TH1 cells and the cytokines that they produce (such as interferon (IFN)) are strongly associated with good clinical outcome for all cancer types. Thus, general characteristics emerge in which cytotoxic Tcells, memory Tcells and TH1 cells are associated with prolonged survival; whereas, the differential effect of other T helper cell populations may be due to their plasticity, depending on the tumour microenvironment and on the tumour type (TABLE1).
www.nature.com/reviews/cancer

2012 Macmillan Publishers Limited. All rights reserved

F O C U S O N tu M our i M M uno L o G Y & i M M H er Punot ERSPE C T IA Vp EY S

In many studies, NK cells have been detected using CD57 as a phenotypic marker. Tumour infiltration by CD57+ cells predicts a good prognosis in colorectal44, gastric45, lung 46, renal47 and liver 48 cancers. However, the expression of CD57 is not restricted to NK cells, and it is also expressed by activated cytotoxic Tcells. Using the NKp46 (also known as NCR1) marker it was shown that a high density of NK cells does not correlate with clinical outcome in non-small-cell lung cancer 49, although it seems to be associated with clinical outcome in the early stages of breast cancer 50. In addition, recent reports have underlined that intratumoural NK cells have an anergic pheno type in lung cancers, which means that they are unable to secrete IFN and to kill tumour cells, even after activation with IL2. The anergic NK cells exhibited downregulated expression of the cell surface molecules NKp30 (also known as NCR3), NKp80 (also known as KLRF1), DNAX accessory molecule 1 (DNAM1; also known as CD226), CD16 and immunoglobulin-like transcript 2 (ILT2; also known as LIR1) in lung cancer 49; downregulated expression of NKp30 and NKp46 in cervical cancer 51; downregulated expression of DNAM1, 2B4 (also known as CD244) and CD16 expression in ovarian carcinoma52; and upregulated expression of co-inhibitory receptors in tongue cancer 53 and renal cell carcinoma54. Tumour cells release suppressor factors, such as transforming growth factor (TGF), which maintain infiltrating NK cells in an anergic state. Therefore, it is likely that, if NK cells have a protective role in the very early stages of tumour development, then they will have little effect once a cancer is clinically detectable. The effect of intratumoural Bcells in cancer is far from clear. Mouse models of spontaneous cancers suggest a deleterious role of Bcells, possibly through the production of IL10 (REF. 55) or through the production of IgG, forming antigenIgG antibody complexes56. This may activate an M2 protumour phenotype in macrophages and might promote the early stages of carcinogenesis57. Bcells may also promote metastasis by converting resting CD4+ Tcells into TReg cells58. Histopathological analyses provide evidence for the presence of intratumoural Bcells in human cancers. Infiltrating Bcells are the dominant component of inflammation in some cancers, such as ductal carcinoma insitu and 20% of invasive breast tumours. They represent clonally expanded populations, express somatic hypermutated antibodies and recognize tumour-associated antigens such as ganglioside D3 (REF. 59). The
NATURE REVIEWS | CANCER 2012 Macmillan Publishers Limited. All rights reserved

associated prognosis of infiltrating Bcells is good in some histological subtypes of breast and epithelial ovarian cancers60,61, but correlations remain to be determined in many cancers. Indeed, Bcells can act as antigenpresenting cells and, therefore, may be important for inducing CD4+ Tcell-dependent CD8+ memory Tcells that help to control tumour invasion and metastasis62. An important clinical translation of this body of observations is the establishment of an immune score to determine the prognosis of clinical outcome in patients when there is no cancer-associated prognostic marker 63. The immune score is based on the numeration of two lymphocyte populations: CD45RO+ memory cells and cytotoxic memory (CD8+) Tcells, both in the core and in the invasive margin of tumours. For example, patients with colorectal cancer who have a local primary tumour and no detectable lymph node or distant metastases are usually treated with surgery alone. However, 2025% of these patients will have recurrence of their disease, indicating that occult metastases are already present at the time of surgery. No tumour-associated marker predicts disease recurrence in these patients. The immune score was applied to two large independent cohorts (n=602). In patients with a high immune score (IS4; which means that high densities of these two cell populations are found in both regions), only 4.8% had relapsed and 86.2% were alive after 5years; whereas, 72% of patients with a low immune score (IS0 and IS1) had disease recurrence, and only 27.5% were alive at 5years. These patients could possibly benefit from adjuvant therapy, and no criterion other than the immune score could have predicted which patients relapsed63. The effect on response to cancer therapies. Whether the immune contexture of the primary tumour predicts therapeutic responses is of paramount importance in patient clinical management. Data based on immune signatures have established that a strong immune cell component is predictive of a good response to chemotherapy in breast cancer 6466, a cancer type in which a high lymphocyte infiltrate is associated with a higher response rate to neoadjuvant therapy 67,68. In hepatic metastases of colorectal cancer, high infiltration of CD8+ Tcells in the invasive margin predicts a better response to chemotherapy and prolonged survival14. In melanoma, the expression of an immune signature (that is, high expression of TH1 cells and cytotoxicity-associated genes) correlates with good clinical

100

58/60
90 80 70

14/15

Eect on prognosis Good None Poor

% articles published

60 50 40 30 20 10 0

4/8

4/8

14/33

CD8+ CD45RO+

TH1 cell

T H2 cell

TH17 cell

TReg cell

Figure 2 | The association of immune cell Nature Reviews | Cancer infiltrates with prognosis in various types of cancer. The analysis of 124 published articles studying the impact of cytotoxic Tcells, memory Tcells, regulatory T (TReg) cells and T helper (TH) cell subpopulations with regard to prognosis of cancer patients (20 different cancer types were analysed) is represented. Good means that the cell type is associated with a good prognosis, none means that there was no correlation and poor means that the cells are associated with a poor prognosis. Please also refer to TABLE1 for references.

response to a therapeutic vaccine using the melanoma-associated antigen 3 (MAGEA3) antigen69. However, the high TH1 cell and cytotoxic immune response that is associated with prolonged survival in patients receiving adjuvant therapies might not be a prediction of response to the therapy, but might rather reflect the fact that the host immune response within the tumour protects the patient and thereby prolongslife.
The cytokine and chemokine milieu Understanding the mechanisms by which a favourable immune contexture might be created and maintained is essential for guiding innovative therapies. Fundamental issues to address include the determination of which factors are associated with tumour cells, what factors are dependent on the host, and how the immune contexture evolves during disease progression and therapy. These issues are far from being completely answered, but analyses of the cytokine and chemokine milieu that is associated with an anti-tumour immune contexture are accumulating (FIG.3; TABLE2).
VOLUME 12 | APRIL 2012 | 301

PERSPECTIVES
CD138+ plasma B cell M2 macrophage MDSC Tumour cells TFH cell CD21+ FDC Tmemory cell Endothelial cell M1 macrophage

CCL17

CCL22

CCL19

CCL21

CCL13

IL-16

CCL5

CXCL9

CXCL10

CX3CL1

TReg cell CCR4

TH cell CD4

T lymphocyte CCR7 CXCR5

MDSC

B lymphocyte

TFH cell

Macrophage CX3CR1

TH1 cell

Tmemory cell

NK cell

CTL

CCR1, CCR3 or CCR5

CXCR3

Figure 3 | Cells and chemokines that coordinate the tumour microenvironment. The immune contexture is determined by the recruitment of various immune cells through the secretion of cytokines and chemokines by tumour cells and other infiltrating immune cells in the

Nature Reviews | Cancer tumour microenvironment. CTL, cytotoxic T lymphocyte; FDC, follicular dendritic cell; IL16, interleukin16; MDSC, myeloid-derived suppressor cell; NK cell, natural killer cell; TFH cell, T follicular helper cell; TH, T helper; TReg cell, regulatory Tcell.

In large cohorts of colorectal tumours, a genegene correlation network that is based on experimental data was integrated with insilico analyses to predict which genes may be associated with an anti-tumour immune response (defined as high densities of memory Tcells in the core and in the invasive margin of the tumours)11. The top 65 genes predicted from the analysis were functionally segregated: the major categories included Tcell activation and differentiation, Bcell activation, innate and inflammatory immune responses, negative regulation of immune responses, adhesion and migrationassociated molecules and chemokines (which had the highest score). High expression of the genes encoding the chemokines CX3CL1 (also known as fractalkine), CXCL9 and CXCL10 was associated with the infiltration of memory Tcells and effector Tcells, particularly TH1 cells, and was associated with prolonged disease-free survival and overall survival11. Indeed, various chemokines are associated with generating an anticancer immune contexture in many human tumours,
302 | APRIL 2012 | VOLUME 12

including melanoma, renal cell carcinoma, hepatocellular carcinoma, cervical cancer and colorectal cancer 11,7074. In fact, a fine balance is generated in the tumour microenvironment between the chemokines that are involved in attracting the relevant immune cells into the core and the invasive margin of the tumour and those chemokines that are involved in generating immune responses in secondary lymphoid structures in draining lymph nodes or in TLS that are adjacent to the tumour. The active recruitment of naive and memory Tcells, including TReg cells, from the blood into TLS via high endothelial venules75 is due to the local production of CCL19, CCL17, CCL22, CXCL13 and IL16 (REF.76). We hypothesize that, as in secondary lymphoid organs, the interaction of Tcells with mature dendritic cells generates central memory and effector Tcells in TLS. Bcells interact with follicular dendritic cells to generate affinity maturation of immunoglobulins, and Bcells locally produce antibodies, some of which react with tumour-associated antigens. The CD4+ and CD8+ memory Tcells that are

generated migrate out through lymphatic vessels expressing CCL21 and migrate into the tumour or to the periphery where memory Tcells may patrol for long periods of time to eventually target circulating malignant cells or nascent metastases77,78. Therefore, a complex interplay of immune cells that express receptors for different chemokines produced at selected locations of the tumour environment builds the architecture of the immune contexture (FIG.3), the coordination of which is an essential trait for effective control of a tumour. Subtle modifications of this architecture that are provoked by changes in the tumour cells (such as antigen modulation), in the host immune system owing to infections, or in the chemokine milieu of the local microenvironment, will result in a loss of coordination and inefficiency of immune control of the tumour, even if high densities of memory Tcells arepresent. Prognostic and therapeutic effects of VEGFA. The dynamic interactions between tumour cells and immune cells are dependent on the
www.nature.com/reviews/cancer

2012 Macmillan Publishers Limited. All rights reserved

F O C U S O N tu M our i M M uno L o G Y & i M M H er Punot ERSPE C T IA Vp EY S

Table 2 | Chemokines and tumour-infiltrating immune cells
Cells
CD138 plasma cells, M2 macrophages and tumour cells
+

Chemoattractants
CCL17 and CCL22

Main receptors
CCR4

Attracted cells
TH cells and TReg cells Tcells in SLO and dendritic cells Bcells in SLO and TFH cells Dendritic cells, macrophages and TH cells TH1 cells, macrophages and memory Tcells Memory Tcells and macrophages TH1 cells, NK cells and cytotoxic Tcells

Mature dendritic cells and tumour cells TFH cells, CD21+ FDCs and tumour cells Lymphocytes and tumour cells Endothelial cells, macrophages, memory Tcells and tumour cells Endothelial cells, M1 macrophages and tumour cells Endothelial cells, macrophages and tumour cells

CCL19 and CCL21 CXCL13 IL16

CCR7 CXCR5 CD4

CCL5

CCR5, CCR1 and CCR3 CXCR3

CXCL9 and CXCL10

CX3CL1

CX3CR1

FDCs, follicular dendritic cells; IL16, interleukin16; NK cells, natural killer cells; SLO, secondary lymphoid organ; TFH cells, T follicular helper cells; TH, T helper; TReg cells, regulatory Tcells.

local network of blood and lymphatic vessels. A dense vascular network may allow potential metastatic cells to escape the primary tumour but, conversely, can also favour the infiltration by immune cells provided that the relevant chemokines are produced. Vascular endothelial growth factor A (VEGFA), a cytokine that is produced by tumour cells, is a major inducer of tumour-associated neovascularization. Furthermore, high expression of VEGFA counteracts the beneficial effects of TH1 cells or cytotoxic Tcells by suppressing the expression of interferon regulatory factor 1 (IRF1) and granulysin (GNLY), respectively 79. In addition, VEGFA is known to inhibit dendritic cell maturation, which may result in the induction of suppressor rather than of effector cells, which hampers the necessary balance of immune cells at the tumour site. Indeed, treatment with anti-VEGFA chemotherapy drugs, such as sunitinib, sorafenib or bevacizumab, results in a decrease of TReg cells and MDSCs both in the periphery and in the tumours of some patients80. The analysis of a cohort of patients with metastatic renal cell cancer who were treated with sunitinib or bevacizumab revealed that patients with a decrease in circulating TReg cells after two or three treatment cycles had a significantly longer overall survival than patients with no decrease in circulating TReg cells81. A precise knowledge of the various factors that are involved in generating the immune contexture should provide other targets in order to tune the balance in favour of immune control of thetumour.
NATURE REVIEWS | CANCER

Conclusions and perspectives The complexity of the immune cell populations infiltrating human tumours with their synergistic or opposing effects may influence tumours differently depending on their histological and molecular type, their stage, the microenvironment of the organ in which they grow, or the nature of the primary tumour or its metastases. In addition, the effect of each component may be buffered by the presence of other immune cell populations at different densities. Therefore, a comprehensive analysis of the interactions between three highly complex systems the tumour cells, the immune response and the tissue micro environment requires the use of systems biology approaches82 to integrate tumour

cell-associated, immune cell-associated and microenvironment-associated parameters with clinical data (including, information on histology, staging evolution and treatment history) from large patient cohorts. Such approaches have already provided a clearer picture of the immune contexture of human tumours and have also provided an integrated scheme of the dynamic interactions of tumour cells with the immune system. The importance of this is illustrated by the fact that, although the densities of immune cell infiltrates are variable from tumour to tumour, all immune cell components (Bcells, NK cells, CD8+ Tcells, TH1 cells, TH2 cells, TReg cells, TH17 cells, TFH cells, macrophages and dendritic cells) are found at higher densities in the tumour than in noncancerous tissues. However, the immune cells may be located in different parts of the tumour. A high TH1 and cytotoxic memory Tcell density, both in the core and at the invasive margin of a tumour is a strong prognostic factor that predicts recurrence and longer overall survival. A complex cytokine and chemokine milieu is also present in the tumour microenvironment, allowing immune cell infiltration and potential activation in tumours adjacent toTLS. These data not only shed new light on the relationship between immunity and cancer but they also provide insight into the management of malignant diseases. Indeed, the pattern of immune cell infiltrates remained the only significant criterion used to predict disease-free survival and overall survival in addition to the TNM classification6,83. Together with mouse studies, the analysis of the human immune microenvironment offers a novel paradigm by which the immune response

Box 1 | Immunomodulatory treatments that modify the immune contexture

Immunogenic chemotherapies, such as oxaliplatin, which induce calreticulin expression on the tumour cell membrane, result in an increased uptake of the tumour cell by phagocytic cells and improved antigen presentation, which helps to clear cancer cells89,90. Anti-angiogenic drugs decrease the levels of myeloid-derived suppressor cells (MDSCs), as well as circulating and intratumoural regulatory T (TReg) cells, which modulate anti-tumour immunity81,91,92. Immune checkpoint-blocking antibodies (which target cytotoxic T lymphocyte-associated antigen 4 (CTLA4), programmed cell death protein 1 (PD1), PD1 ligand 1 (PDL1) and CD137, for example) and specific depletion of TReg cells93 increase the infiltration of the tumour microenvironment by Tcells, particularly CD8+ Tcells, which increases immune responses to tumour cells8588,94. Tumour vaccines increase Tcell infiltration within tumours, which increases cytotoxic immune responses to tumour cells9599. Tumour-specific monoclonal antibodies can induce an adaptive immune response to tumour cells94,100102.

VOLUME 12 | APRIL 2012 | 303 2012 Macmillan Publishers Limited. All rights reserved

PERSPECTIVES
at least as much as the histopathological TNM classification affects clinical outcome and the ability to guide therapeutics. The next questions to resolve concern the interactions between tumour cells and immune cells. Colorectal tumours with microsatellite instability often have high levels of frameshift mutations that could potentially generate antigenic peptides, and these tumours have high infiltrations of Tcells, particularly CD8+ Tcells, as well as a good prognosis84. This suggests that genomic instability in the tumour may create antigens that are recognized by the immune system. Deep sequencing of human tumours will generate large amounts of data that could identify potential tumourassociated antigens. As knowledge of the evolution of the immune contexture in the primary tumour, in the periphery and in metastatic sites as diseases progress is established for various tumours, therapeutic tools could be designed to positively influence the immune contexture. The approval of ipilimumab a monoclonal antibody that antagonises cytotoxic T lymphocyteassociated antigen 4 (CTLA4), resulting in the enhanced recruitment of CD4+ memory Tcells and CD8+ memory Tcells into tumours for the treatment of metastatic melanoma is an example of a targeted immunomodulatory agent85. Others, such as anti-programmed cell death protein1 (PD1) or antiCD137 (also known as TNFRSF9) are in clinical trials8688. The modulation of the immunomodulatory activities of VEGFA may fall into the same category. Finally, classic treatments, such as chemotherapy and radiotherapy, are being re-analysed for their immunomodulatory activities and how these affect therapeutic efficacy (BOX1). The issue of what is recognized by the immune response is far from trivial. Successful immunotherapies result in prolonged overall survival, but they do not always result in direct tumourreducing effects85. The improved survival may be due to the destruction of tumour cells. It may also result from improved homeostasis of the immune system, which is illustrated by the capacity of the tumour cells to build an efficient immune contexture with indirect anti-tumour effects such as the control of infections and inflammation. Considering the immune contexture at the core of the problem could accelerate the development of the most efficient treatments that aim to establish a long diseasefree period and, ultimately, good overall survival.
304 | APRIL 2012 | VOLUME 12 2012 Macmillan Publishers Limited. All rights reserved
Wolf Herman Fridman and Catherine Sauts-Fridman are at the INSERM UMRS872, Laboratory of Immune microenvironment and tumours, Paris F75006, France. Wolf Herman Fridman, Franck Pags, Catherine Sauts-Fridman and Jrme Galon are at the Cordeliers Research Centre, Universit Pierre et Marie Curie, Paris F-75006, France. Wolf Herman Fridman, Franck Pags, Catherine Sauts-Fridman and Jrme Galon are at the Cordeliers Research Centre, Universit Paris Descartes, Paris F-75006, France. Wolf Herman Fridman, Franck Pags and Jrme Galon are at the Assistance Publique-Hopitaux de Paris, APHP, Georges Pompidou European Hospital, Paris F-75015, France. Franck Pags and Jrme Galon are at the INSERM UMRS872, Laboratory of Integrative Cancer Immunology, Paris F75006, France. Correspondence to J.G. e-mail: jerome.galon@crc.jussieu.fr doi:10.1038/nrc3245 Published online 15 March 2012
1. 2. 3. Dunn, G.P., Old, L.J. & Schreiber, R.D. The three Es of cancer immunoediting. Annu. Rev. Immunol. 22, 329360 (2004). Koebel, C.M. etal. Adaptive immunity maintains occult cancer in an equilibrium state. Nature 450, 903907 (2007). Schreiber, R.D., Old, L.J. & Smyth, M.J. Cancer immunoediting: integrating immunitys roles in cancer suppression and promotion. Science 331, 15651570 (2011). Shankaran, V. etal. IFN and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature 410, 11071111 (2001). Galon, J. etal. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 313, 19601964 (2006). Mlecnik, B. etal. Histopathologic-based prognostic factors of colorectal cancers are associated with the state of the local immune reaction. J.Clin. Oncol. 29, 610618 (2011). Hanahan, D. & Weinberg, R.A. Hallmarks of cancer: the next generation. Cell 144, 646674 (2011). Galon, J., Fridman, W.H. & Pages, F. The adaptive immunologic microenvironment in colorectal cancer: a novel perspective. Cancer Res. 67, 18831886 (2007). Shannon, P. etal. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 24982504 (2003). Bindea, G. etal. ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics 25, 10911093 (2009). Mlecnik, B. etal. Biomolecular network reconstruction identifies Tcell homing factors associated with survival in colorectal cancer. Gastroenterology 138, 14291440 (2010). Mantovani, A., Allavena, P., Sica, A. & Balkwill, F. Cancer-related inflammation. Nature 454, 436444 (2008). Dieu-Nosjean, M.C. etal. Long-term survival for patients with nonsmallcell lung cancer with intratumoral lymphoid structures. J.Clin. Oncol. 26, 44104417 (2008). Halama, N. etal. Localization and density of immune cells in the invasive margin of human colorectal cancer liver metastases are prognostic for response to chemotherapy. Cancer Res. 71, 56705677 (2011). Nakano, O. etal. Proliferative activity of intratumoral CD8+ Tlymphocytes as a prognostic factor in human renal cell carcinoma: clinicopathologic demonstration of antitumor immunity. Cancer Res. 61, 51325136 (2001). Curiel, T.J. etal. Specific recruitment of regulatory Tcells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nature Med. 10, 942949 (2004). 17. Bates, G.J. etal. Quantification of regulatory Tcells enables the identification of high-risk breast cancer patients and those at risk of late relapse. J.Clin. Oncol. 24, 53735380 (2006). 18. Gobert, M. etal. Regulatory Tcells recruited through CCL22/CCR4 are selectively activated in lymphoid infiltrates surrounding primary breast tumors and lead to an adverse clinical outcome. Cancer Res. 69, 20002009 (2009). 19. Fu, J. etal. Increased regulatory Tcells correlate with CD8 Tcell impairment and poor survival in hepatocellular carcinoma patients. Gastroenterology 132, 23282339 (2007). 20. Gao, Q. etal. Intratumoral balance of regulatory and cytotoxic Tcells is associated with prognosis of hepatocellular carcinoma after resection. J.Clin. Oncol. 25, 25862593 (2007). 21. Grabenbauer, G.G., Lahmer, G., Distel, L. & Niedobitek, G. Tumor-infiltrating cytotoxic Tcells but not regulatory Tcells predict outcome in anal squamous cell carcinoma. Clin. Cancer Res. 12, 33553360 (2006). 22. Heimberger, A.B. etal. Incidence and prognostic impact of FoxP3+ regulatory Tcells in human gliomas. Clin. Cancer Res. 14, 51665172 (2008). 23. Hillen, F. etal. Leukocyte infiltration and tumor cell plasticity are parameters of aggressiveness in primary cutaneous melanoma. Cancer Immunol. Immunother. 57, 97106 (2008). 24. Jacobs, J.F. etal. Prognostic significance and mechanism of Treg infiltration in human brain tumors. J.Neuroimmunol 225, 195199 (2010). 25. Ladanyi, A. etal. FOXP3+ cell density in primary tumor has no prognostic impact in patients with cutaneous malignant melanoma. Pathol. Oncol. Res. 16, 303309 (2010). 26. Mahmoud, S.M. etal. An evaluation of the clinical significance of FOXP3+ infiltrating cells in human breast cancer. Breast Cancer Res. Treat. 127, 99108 (2011). 27. Mizukami, Y. etal. Localisation pattern of Foxp3+ regulatory Tcells is associated with clinical behaviour in gastric cancer. Br. J.Cancer 98, 148153 (2008). 28. Sinicrope, F.A. etal. Intraepithelial effector (CD3+)/ regulatory (FoxP3+) Tcell ratio predicts a clinical outcome of human colon carcinoma. Gastroenterology 137, 12701279 (2009). 29. Zhang, Y.L. etal. Different subsets of tumor infiltrating lymphocytes correlate with NPC progression in different ways. Mol. Cancer 9, 4 (2010). 30. Badoual, C. etal. Prognostic value of tumor-infiltrating CD4+ Tcell subpopulations in head and neck cancers. Clin. Cancer Res. 12, 465472 (2006). 31. Carreras, J. etal. High numbers of tumor-infiltrating FOXP3positive regulatory Tcells are associated with improved overall survival in follicular lymphoma. Blood 108, 29572964 (2006). 32. Frey, D.M. etal. High frequency of tumor-infiltrating FOXP3+ regulatory Tcells predicts improved survival in mismatch repair-proficient colorectal cancer patients. Int. J.Cancer 126, 26352643 (2010). 33. Leffers, N. etal. Prognostic significance of tumorinfiltrating Tlymphocytes in primary and metastatic lesions of advanced stage ovarian cancer. Cancer Immunol. Immunother. 58, 449459 (2009). 34. Michel, S. etal. High density of FOXP3positive Tcells infiltrating colorectal cancers with microsatellite instability. Br. J.Cancer 99, 18671873 (2008). 35. Salama, P. etal. Tumor-infiltrating FOXP3+ T regulatory cells show strong prognostic significance in colorectal cancer. J.Clin. Oncol. 27, 186192 (2009). 36. Tosolini, M. etal. Clinical impact of different classes of infiltrating T cytotoxic and helper cells (Th1, th2, treg, th17) in patients with colorectal cancer. Cancer Res. 71, 12631271 (2011). 37. Tzankov, A. etal. Correlation of high numbers of intratumoral FOXP3+ regulatory Tcells with improved survival in germinal center-like diffuse large Bcell lymphoma, follicular lymphoma and classical Hodgkins lymphoma. Haematologica 93, 193200 (2008). 38. Winerdal, M.E. etal. FOXP3 and survival in urinary bladder cancer. BJU Int. 108, 16721678 (2011). 39. Benchetrit, F. etal. Interleukin17 inhibits tumor cell growth by means of a Tcelldependent mechanism. Blood 99, 21142121 (2002). 40. Tartour, E. etal. Interleukin 17, a Tcellderived cytokine, promotes tumorigenicity of human cervical tumors in nude mice. Cancer Res. 59, 36983704 (1999). 41. Wilke, C.M. etal. Th17 cells in cancer: help or hindrance? Carcinogenesis 32, 643649 (2011). 42. Yoon, N.K. etal. Higher levels of GATA3 predict better survival in women with breast cancer. Hum. Pathol. 41, 17941801 (2010).

4. 5. 6.

7. 8.

9.

10.

11.

12. 13.

14.

15.

16.

www.nature.com/reviews/cancer

F O C U S O N tu M our i M M uno L o G Y & i M M H er Punot ERSPE C T IA Vp EY S

43. Schreck, S. etal. Prognostic impact of tumourinfiltrating Th2 and regulatory Tcells in classical Hodgkin lymphoma. Hematol. Oncol. 27, 3139 (2009). 44. Coca, S. etal. The prognostic significance of intratumoral natural killer cells in patients with colorectal carcinoma. Cancer 79, 23202328 (1997). 45. Ishigami, S. etal. Prognostic value of intratumoral natural killer cells in gastric carcinoma. Cancer 88, 577583 (2000). 46. Villegas, F.R. etal. Prognostic significance of tumor infiltrating natural killer cells subset CD57 in patients with squamous cell lung cancer. Lung Cancer 35, 2328 (2002). 47. Donskov, F. & von der Maase, H. Impact of immune parameters on long-term survival in metastatic renal cell carcinoma. J.Clin. Oncol. 24, 19972005 (2006). 48. Zhu, L.Y., Zhou, J., Liu, Y.Z. & Pan, W.D. Prognostic significance of natural killer cell infiltration in hepatocellular carcinoma. Ai Zheng 28, 11981202 (2009). 49. Platonova, S. etal. Profound coordinated alterations of intratumoral NK cell phenotype and function in lung carcinoma. Cancer Res. 71, 54125422 (2011). 50. Mamessier, E. etal. Human breast cancer cells enhance self tolerance by promoting evasion from NK cell antitumor immunity. J.Clin. Invest. 121, 36093622 (2011). 51. Garcia-Iglesias, T. etal. Low NKp30, NKp46 and NKG2D expression and reduced cytotoxic activity on NK cells in cervical cancer and precursor lesions. BMC Cancer 9, 186 (2009). 52. Carlsten, M. etal. Primary human tumor cells expressing CD155 impair tumor targeting by downregulating DNAM1 on NK cells. J.Immunol. 183, 49214930 (2009). 53. Katou, F. etal. Differing phenotypes between intraepithelial and stromal lymphocytes in early-stage tongue cancer. Cancer Res. 67, 1119511201 (2007). 54. Schleypen, J.S. etal. Renal cell carcinoma-infiltrating natural killer cells express differential repertoires of activating and inhibitory receptors and are inhibited by specific HLA class I allotypes. Int. J.Cancer 106, 905912 (2003). 55. Wong, S.C. etal. Macrophage polarization to a unique phenotype driven by Bcells. Eur. J.Immunol. 40, 22962307 (2010). 56. Andreu, P. etal. FcR activation regulates inflammation-associated squamous carcinogenesis. Cancer Cell 17, 121134 (2010). 57. Mantovani, A. Bcells and macrophages in cancer: yin and yang. Nature Med. 17, 285286 (2011). 58. Olkhanud, P.B. etal. Tumor-evoked regulatory Bcells promote breast cancer metastasis by converting resting CD4 Tcells to Tregulatory cells. Cancer Res. 71, 35053515 (2011). 59. Coronella-Wood, J.A. & Hersh, E.M. Naturally occurring Bcell responses to breast cancer. Cancer Immunol. Immunother. 52, 715738 (2003). 60. Coronella, J.A. etal. Evidence for an antigen-driven humoral immune response in medullary ductal breast cancer. Cancer Res. 61, 78897899 (2001). 61. Milne, K. etal. Systematic analysis of immune infiltrates in high-grade serous ovarian cancer reveals CD20, FoxP3 and TIA1 as positive prognostic factors. PLoS ONE 4, e6412 (2009). 62. DiLillo, D.J., Yanaba, K. & Tedder, T.F. Bcells are required for optimal CD4+ and CD8+ Tcell tumor immunity: therapeutic Bcell depletion enhances B16 melanoma growth in mice. J.Immunol. 184, 40064016 (2010). 63. Pages, F. etal. Insitu cytotoxic and memory Tcells predict outcome in patients with early-stage colorectal cancer. J.Clin. Oncol. 27, 59445951 (2009). 64. Desmedt, C. etal. Biological processes associated with breast cancer clinical outcome depend on the molecular subtypes. Clin. Cancer Res. 14, 51585165 (2008). 65. Iwamoto, T. etal. Gene pathways associated with prognosis and chemotherapy sensitivity in molecular subtypes of breast cancer. J.Natl Cancer Inst. 103, 264272 (2011). 66. Sotiriou, C. & Pusztai, L. Gene-expression signatures in breast cancer. N.Engl. J.Med. 360, 790800 (2009). 67. Andre, F., Berrada, N. & Desmedt, C. Implication of tumor microenvironment in the resistance to chemotherapy in breast cancer patients. Curr. Opin. Oncol. 22, 547551 (2010). 68. Denkert, C. etal. Tumor-associated lymphocytes as an independent predictor of response to neoadjuvant chemotherapy in breast cancer. J.Clin. Oncol. 28, 105113 (2010). 69. Gajewski, T.F., Louahed, J. & Brichard, V.G. Gene signature in melanoma associated with clinical activity: a potential clue to unlock cancer immunotherapy. Cancer J. 16, 399403 (2010). 70. Altenburg, A., Baldus, S.E., Smola, H., Pfister, H. & Hess, S. CD40 ligandCD40 interaction induces chemokines in cervical carcinoma cells in synergism with IFN-. J.Immunol. 162, 41404147 (1999). 71. Kondo, T. etal. Favorable prognosis of renal cell carcinoma with increased expression of chemokines associated with a Th1type immune response. Cancer Sci. 97, 780786 (2006). 72. Hirano, S. etal. Increased mRNA expression of chemokines in hepatocellular carcinoma with tumorinfiltrating lymphocytes. J.Gastroenterol. Hepatol 22, 690696 (2007). 73. Chew, V. etal. Chemokine-driven lymphocyte infiltration: an early intratumoural event determining long-term survival in resectable hepatocellular carcinoma. Gut 61, 427438 (2011). 74. Kunz, M. etal. Strong expression of the lymphoattractant CXC chemokine Mig is associated with heavy infiltration of Tcells in human malignant melanoma. J.Pathol. 189, 552558 (1999). 75. Martinet, L. etal. Human solid tumors contain high endothelial venules: association with T and Blymphocyte infiltration and favorable prognosis in breast cancer. Cancer Res. 71, 56785687 (2011). 76. de Chaisemartin, L. etal. Characterization of chemokines and adhesion molecules associated with Tcell presence in tertiary lymphoid structures in human lung cancer. Cancer Res. 71, 63916399 (2011). 77. Mlecnik, B., Bindea, G., Pages, F. & Galon, J. Tumor immunosurveillance in human cancers. Cancer Metastasis Rev. 30, 512 (2011). 78. Sautes-Fridman, C. etal. Tumor microenvironment is multifaceted. Cancer Metastasis Rev. 30, 1325 (2011). 79. Camus, M. etal. Coordination of intratumoral immune reaction and human colorectal cancer recurrence. Cancer Res. 69, 26852693 (2009). 80. Gabrilovich, D.I., Ishida, T., Nadaf, S., Ohm, J.E. & Carbone, D.P. Antibodies to vascular endothelial growth factor enhance the efficacy of cancer immunotherapy by improving endogenous dendritic cell function. Clin. Cancer Res. 5, 29632970 (1999). 81. Adotevi, O. etal. A decrease of regulatory Tcells correlates with overall survival after sunitinib-based antiangiogenic therapy in metastatic renal cancer patients. J.Immunother. 33, 991998 (2010). 82. Hood, L., Heath, J.R., Phelps, M.E. & Lin, B. Systems biology and new technologies enable predictive and preventative medicine. Science 306, 640643 (2004). 83. Broussard, E.K. & Disis, M.L. TNM staging in colorectal cancer: T is for T cell and M is for memory. J.Clin. Oncol. 29, 601603 (2011). 84. Ogino, S., Galon, J., Fuchs, C.S. & Dranoff, G. Cancer immunology-analysis of host and tumor factors for personalized medicine. Nature Rev. Clin. Oncol. 8, 711719 (2011). 85. Hodi, F.S. etal. Improved survival with ipilimumab in patients with metastatic melanoma. N.Engl. J.Med. 363, 711723 (2010). 86. Kline, J. & Gajewski, T.F. Clinical development of mAbs to block the PD1 pathway as an immunotherapy for cancer. Curr. Opin. Investig Drugs 11, 13541359 (2011). 87. Rosenblatt, J. etal. PD1 blockade by CT011, antiPD1 antibody, enhances exvivo Tcell responses to autologous dendritic cell/myeloma fusion vaccine. J.Immunother. 34, 409418 (2010). 88. Waldmann, T.A. Effective cancer therapy through immunomodulation. Annu. Rev. Med. 57, 6581 (2006). 89. Nardin, A. etal. Dacarbazine promotes stromal remodeling and lymphocyte infiltration in cutaneous melanoma lesions. J.Invest. Dermatol. 131, 18961905 (2011). 90. Zitvogel, L., Apetoh, L., Ghiringhelli, F. & Kroemer, G. Immunological aspects of cancer chemotherapy. Nature Rev. Immunol. 8, 5973 (2008). 91. Ko, J.S. etal. Sunitinib mediates reversal of myeloidderived suppressor cell accumulation in renal cell carcinoma patients. Clin. Cancer Res. 15, 21482157 (2009). 92. Yang, J.C. etal. A randomized trial of bevacizumab, an anti-vascular endothelial growth factor antibody, for metastatic renal cancer. N.Engl. J.Med. 349, 427434 (2003). 93. Teng, M.W. etal. Conditional regulatory Tcell depletion releases adaptive immunity preventing carcinogenesis and suppressing established tumor growth. Cancer Res. 70, 78007809 (2010). 94. Stagg, J. etal. AntiErbB2 mAb therapy requires typeI and II interferons and synergizes with antiPD1 or antiCD137 mAb therapy. Proc. Natl Acad. Sci. USA 108, 71427147 (2011). 95. Finn, O.J. Cancer immunology. N.Engl. J.Med. 358, 27042715 (2008). 96. Phan, G.Q. etal. Immunization of patients with metastatic melanoma using both class I and class IIrestricted peptides from melanoma-associated antigens. J.Immunother. 26, 349356 (2003). 97. Rosenberg, S.A. etal. Durable complete responses in heavily pretreated patients with metastatic melanoma using Tcell transfer immunotherapy. Clin. Cancer Res. 17, 45504557 (2011). 98. Beer, T.M. etal. Randomized trial of autologous cellular immunotherapy with sipuleucelT in androgendependent prostate cancer. Clin. Cancer Res. 17, 45584567 (2011). 99. Kantoff, P.W. etal. Overall survival analysis of a phaseII randomized controlled trial of a Poxviralbased PSA-targeted immunotherapy in metastatic castration-resistant prostate cancer. J.Clin. Oncol. 28, 10991105 (2010). 100. Abes, R., Gelize, E., Fridman, W.H. & Teillaud, J.L. Long-lasting antitumor protection by antiCD20 antibody through cellular immune response. Blood 116, 926934 (2010). 101. Galluzzi, L. etal. Trial Watch, monoclonal antibodies in cancer therapy. Oncoimmunology 1, 2837 (2012). 102. Park, S. etal. The therapeutic effect of antiHER2/neu antibody depends on both innate and adaptive immunity. Cancer Cell 18, 160170 (2010). 103. Clark, W.H. Jr. etal. Model predicting survival in stage I melanoma based on tumor progression. J.Natl Cancer Inst. 81, 18931904 (1989). 104. Clemente, C.G. etal. Prognostic value of tumor infiltrating lymphocytes in the vertical growth phase of primary cutaneous melanoma. Cancer 77, 13031310 (1996). 105. Mackensen, A. etal. Evidence for insitu amplification of cytotoxic Tlymphocytes with antitumor activity in a human regressive melanoma. Cancer Res. 53, 35693573 (1993). 106. Tefany, F.J., Barnetson, R.S., Halliday, G.M., McCarthy, S.W. & McCarthy, W.H. Immunocytochemical analysis of the cellular infiltrate in primary regressing and non-regressing malignant melanoma. J.Invest. Dermatol. 97, 197202 (1991). 107. Miracco, C. etal. Utility of tumour-infiltrating CD25+FOXP3+ regulatory Tcell evaluation in predicting local recurrence in vertical growth phase cutaneous melanoma. Oncol. Rep. 18, 11151122 (2007). 108. Mougiakakos, D. etal. Intratumoral forkhead box P3positive regulatory Tcells predict poor survival in cyclooxygenase-2positive uveal melanoma. Cancer 116, 22242233 (2010). 109. Reichert, T.E., Scheuer, C., Day, R., Wagner, W. & Whiteside, T.L. The number of intratumoral dendritic cells and zeta-chain expression in Tcells as prognostic and survival biomarkers in patients with oral carcinoma. Cancer 91, 21362147 (2001). 110. Shibuya, T.Y. etal. Clinical significance of poor CD3 response in head and neck cancer. Clin. Cancer Res. 8, 745751 (2002). 111. Alexe, G. etal. High expression of lymphocyteassociated genes in node-negative HER2+ breast cancers correlates with lower recurrence rates. Cancer Res. 67, 1066910676 (2007). 112. Mahmoud, S.M. etal. Tumor-infiltrating CD8+ lymphocytes predict clinical outcome in breast cancer. J.Clin. Oncol. 29, 19491955 (2011). 113. Marrogi, A.J. etal. Study of tumor infiltrating lymphocytes and transforming growth factor- as prognostic factors in breast carcinoma. Int. J.Cancer 74, 492501 (1997). 114. Menegaz, R.A., Michelin, M.A., Etchebehere, R.M., Fernandes, P.C. & Murta, E.F. Peri- and intratumoral T and B lymphocytic infiltration in breast cancer. Eur. J.Gynaecol. Oncol. 29, 321326 (2008). 115. Oldford, S.A. etal. Tumor cell expression of HLA-DM associates with a Th1 profile and predicts improved survival in breast carcinoma patients. Int. Immunol. 18, 15911602 (2006).

NATURE REVIEWS | CANCER 2012 Macmillan Publishers Limited. All rights reserved

VOLUME 12 | APRIL 2012 | 305

PERSPECTIVES
116. Teschendorff, A.E. etal. Improved prognostic classification of breast cancer defined by antagonistic activation patterns of immune response pathway modules. BMC Cancer 10, 604 (2010). 117. Camp, B.J., Dyhrman, S.T., Memoli, V.A., Mott, L.A. & Barth, R.J. Jr. Insitu cytokine production by breast cancer tumor-infiltrating lymphocytes. Ann. Surg. Oncol. 3, 176184 (1996). 118. Nakakubo, Y. etal. Clinical significance of immune cell infiltration within gallbladder cancer. Br. J.Cancer 89, 17361742 (2003). 119. Sharma, P. etal. CD8 tumor-infiltrating lymphocytes are predictive of survival in muscle-invasive urothelial carcinoma. Proc. Natl Acad. Sci. USA 104, 39673972 (2007). 120. Zhang, L. etal. Intratumoral Tcells, recurrence, and survival in epithelial ovarian cancer. N.Engl. J.Med. 348, 203213 (2003). 121. Sato, E. etal. Intraepithelial CD8+ tumor-infiltrating lymphocytes and a high CD8+/regulatory Tcell ratio are associated with favorable prognosis in ovarian cancer. Proc. Natl Acad. Sci. USA 102, 1853818543 (2005). 122. Hamanishi, J. etal. Programmed cell death 1 ligand 1 and tumor-infiltrating CD8+ T lymphocytes are prognostic factors of human ovarian cancer. Proc. Natl Acad. Sci. USA 104, 33603365 (2007). 123. Kusuda, T. etal. Relative expression levels of Th1 and Th2 cytokine mRNA are independent prognostic factors in patients with ovarian cancer. Oncol. Rep. 13, 11531158 (2005). 124. Marth, C. etal. Interferon- expression is an independent prognostic factor in ovarian cancer. Am. J.Obstet. Gynecol. 191, 15981605 (2004). 125. Kryczek, I. etal. Phenotype, distribution, generation, and functional and clinical relevance of Th17 cells in the human tumor environments. Blood 114, 11411149 (2009). 126. Cho, Y. etal. CD4+ and CD8+ Tcells cooperate to improve prognosis of patients with esophageal squamous cell carcinoma. Cancer Res. 63, 15551559 (2003). 127. Schumacher, K., Haensch, W., Roefzaad, C. & Schlag, P.M. Prognostic significance of activated CD8+ Tcell infiltrations within esophageal carcinomas. Cancer Res. 61, 39323936 (2001). 128. van Sandick, J.W. etal. Lymphocyte subsets and T h1/T h2 immune responses in patients with adenocarcinoma of the oesophagus or oesophagogastric junction: relation to pTNM stage and clinical outcome. Cancer Immunol. Immunother. 52, 617624 (2003). 129. Lv, L. etal. The accumulation and prognosis value of tumor infiltrating IL17 producing cells in esophageal squamous cell carcinoma. PLoS ONE 6, e18219 (2011). 130. Baier, P.K. etal. Analysis of the Tcell receptor variability of tumor-infiltrating lymphocytes in colorectal carcinomas. Tumour Biol. 19, 205212 (1998). 131. Baker, K. etal. Differential significance of tumour infiltrating lymphocytes in sporadic mismatch repair deficient versus proficient colorectal cancers: a potential role for dysregulation of the transforming growth factor- pathway. Eur. J.Cancer 43, 624631 (2007). 132. Dalerba, P., Maccalli, C., Casati, C., Castelli, C. & Parmiani, G. Immunology and immunotherapy of colorectal cancer. Crit. Rev. Oncol. Hematol. 46, 3357 (2003). 133. Diederichsen, A.C., Hjelmborg, J.B., Christensen, P.B., Zeuthen, J. & Fenger, C. Prognostic value of the CD4+/CD8+ ratio of tumour infiltrating lymphocytes in colorectal cancer and HLA-DR expression on tumour cells. Cancer Immunol. Immunother. 52, 423428 (2003). 134. Graham, D.M. & Appelman, H.D. Crohns-like lymphoid reaction and colorectal carcinoma: a potential histologic prognosticator. Mod. Pathol. 3, 332335 (1990). 135. Halama, N. etal. The localization and density of immune cells in primary tumors of human metastatic colorectal cancer shows an association with response to chemotherapy. Cancer Immun. 9, 1 (2009). 136. Harrison, J.C., Dean, P.J., el-Zeky, F. & Vander Zwaag, R. From Dukes through Jass: pathological prognostic indicators in rectal cancer. Hum. Pathol. 25, 498505 (1994). 137. Jass, J.R. Lymphocytic infiltration and survival in rectal cancer. J.Clin. Pathol. 39, 585589 (1986). 138. Lee, W.S., Park, S., Lee, W.Y., Yun, S.H. & Chun, H.K. Clinical impact of tumor-infiltrating lymphocytes for survival in stage II colon cancer. Cancer 116, 51885199 (2010). 139. Lugli, A. etal. CD8+ lymphocytes/ tumour-budding index: an independent prognostic factor representing a pro-/anti-tumour approach to tumour host interaction in colorectal cancer. Br. J.Cancer 101, 13821392 (2009). 140. Menon, A.G. etal. Immune system and prognosis in colorectal cancer: a detailed immunohistochemical analysis. Lab. Invest. 84, 493501 (2004). 141. Naito, Y. etal. CD8+ Tcells infiltrated within cancer cell nests as a prognostic factor in human colorectal cancer. Cancer Res. 58, 34913494 (1998). 142. Nosho, K. etal. Tumour-infiltrating Tcell subsets, molecular changes in colorectal cancer, and prognosis: cohort study and literature review. J.Pathol. 222, 350366 (2010). 143. Pages, F. etal. Effector memory Tcells, early metastasis, and survival in colorectal cancer. N.Engl. J.Med. 353, 26542666 (2005). 144. Prall, F. etal. Prognostic role of CD8+ tumorinfiltrating lymphocytes in stage III colorectal cancer with and without microsatellite instability. Hum. Pathol. 35, 808816 (2004). 145. Ropponen, K.M., Eskelinen, M.J., Lipponen, P.K., Alhava, E. & Kosma, V.M. Prognostic value of tumourinfiltrating lymphocytes (TILs) in colorectal cancer. J.Pathol. 182, 318324 (1997). 146. Dahlin, A.M. etal. Colorectal cancer prognosis depends on Tcell infiltration and molecular characteristics of the tumor. Mod. Pathol. 24, 671682 (2011). 147. Nagtegaal, I.D. etal. Local and distant recurrences in rectal cancer patients are predicted by the nonspecific immune response; specific immune response has only a systemic effecta histopathological and immunohistochemical study. BMC Cancer 1, 7 (2001). 148. Ogino, S. etal. Lymphocytic reaction to colorectal cancer is associated with longer survival, independent of lymph node count, microsatellite instability, and CpG island methylator phenotype. Clin. Cancer Res. 15, 64126420 (2009). 149. Liu, J. etal. IL17 is associated with poor prognosis and promotes angiogenesis via stimulating VEGF production of cancer cells in colorectal carcinoma. Biochem. Biophys. Res. Commun. 407, 348354 (2011). 150. Jensen, H.K., Donskov, F., Nordsmark, M., Marcussen, N. & von der Maase, H. Increased intratumoral FOXP3positive regulatory immune cells during interleukin2 treatment in metastatic renal cell carcinoma. Clin. Cancer Res. 15, 10521058 (2009). 151. Karja, V. etal. Tumour-infiltrating lymphocytes: a prognostic factor of PSA-free survival in patients with local prostate carcinoma treated by radical prostatectomy. Anticancer Res. 25, 44354438 (2005). 152. Richardsen, E., Uglehus, R.D., Due, J., Busch, C. & Busund, L.T. The prognostic impact of MCSF, CSF1 receptor, CD68 and CD3 in prostatic carcinoma. Histopathology 53, 3038 (2008). 153. Vesalainen, S., Lipponen, P., Talja, M. & Syrjanen, K. Histological grade, perineural infiltration, tumourinfiltrating lymphocytes and apoptosis as determinants of long-term prognosis in prostatic adenocarcinoma. Eur. J.Cancer 30A, 17971803 (1994). 154. Al-Shibli, K.I. etal. Prognostic effect of epithelial and stromal lymphocyte infiltration in non-small cell lung cancer. Clin. Cancer Res. 14, 52205227 (2008). 155. Hiraoka, N., Onozato, K., Kosuge, T. & Hirohashi, S. Prevalence of FOXP3+ regulatory Tcells increases during the progression of pancreatic ductal adenocarcinoma and its premalignant lesions. Clin. Cancer Res. 12, 54235434 (2006). 156. Ito, N. etal. Prognostic significance of T helper 1 and 2 and T cytotoxic 1 and 2 cells in patients with nonsmall cell lung cancer. Anticancer Res. 25, 20272031 (2005). 157. Kawai, O. etal. Predominant infiltration of macrophages and CD8+ T Cells in cancer nests is a significant predictor of survival in stage IV nonsmall cell lung cancer. Cancer 113, 13871395 (2008). 158. Wakabayashi, O. etal. CD4+ Tcells in cancer stroma, not CD8+ Tcells in cancer cell nests, are associated with favorable prognosis in human non-small cell lung cancers. Cancer Sci. 94, 10031009 (2003). 159. Chen, X. etal. Increased IL17producing cells correlate with poor survival and lymphangiogenesis in NSCLC patients. Lung Cancer 69, 348354 (2010). 160. Petersen, R.P. etal. Tumor infiltrating Foxp3+ regulatory Tcells are associated with recurrence in pathologic stage I NSCLC patients. Cancer 107, 28662872 (2006). 161. Shimizu, K. etal. Tumor-infiltrating Foxp3+ regulatory Tcells are correlated with cyclooxygenase2 expression and are associated with recurrence in resected non-small cell lung cancer. J.Thorac. Oncol. 5, 585590 (2010). 162. Tao, H. etal. Prognostic potential of FOXP3 expression in non-small cell lung cancer cells combined with tumor-infiltrating regulatory Tcells. Lung Cancer 75, 95101 (2011). 163. Fukunaga, A. etal. CD8+ tumor-infiltrating lymphocytes together with CD4+ tumor-infiltrating lymphocytes and dendritic cells improve the prognosis of patients with pancreatic adenocarcinoma. Pancreas 28, e26e31 (2004). 164. De Monte, L. etal. Intratumor T helper type2 cell infiltrate correlates with cancer-associated fibroblast thymic stromal lymphopoietin production and reduced survival in pancreatic cancer. J.Exp. Med. 208, 469478 (2011). 165. Tassi, E. etal. Carcinoembryonic antigen-specific but not antiviral CD4+ Tcell immunity is impaired in pancreatic carcinoma patients. J.Immunol. 181, 65956603 (2008). 166. Seresini, S. etal. IFN- produced by human papilloma virus18 E6specific CD4+ Tcells predicts the clinical outcome after surgery in patients with high-grade cervical lesions. J.Immunol. 179, 71767183 (2007). 167. Cai, X.Y. etal. Dendritic cell infiltration and prognosis of human hepatocellular carcinoma. J.Cancer Res. Clin. Oncol. 132, 293301 (2006). 168. Wada, Y., Nakashima, O., Kutami, R., Yamamoto, O. & Kojiro, M. Clinicopathological study on hepatocellular carcinoma with lymphocytic infiltration. Hepatology 27, 407414 (1998). 169. Gao, Q. etal. Tumor stroma reaction-related gene signature predicts clinical outcome in human hepatocellular carcinoma. Cancer Sci. 102, 15221531 (2011). 170. Zhang, J.P. etal. Increased intratumoral IL17producing cells correlate with poor survival in hepatocellular carcinoma patients. J.Hepatol. 50, 980989 (2009). 171. Ubukata, H., Motohashi, G., Tabuchi, T., Nagata, H. & Konishi, S. Evaluations of interferon-/interleukin4 ratio and neutrophil/lymphocyte ratio as prognostic indicators in gastric cancer patients. J.Surg. Oncol. 102, 742747 (2010). 172. Chen, J.G. etal. Intratumoral expression of IL17 and its prognostic role in gastric adenocarcinoma patients. Int. J.Biol. Sci. 7, 5360 (2011). 173. Wiegering, V. etal. TH1 predominance is associated with improved survival in pediatric medulloblastoma patients. Cancer Immunol. Immunother. 60, 693703 (2011). 174. Paulson, K.G. etal. Transcriptome-wide studies of merkel cell carcinoma and validation of intratumoral CD8+ lymphocyte invasion as an independent predictor of survival. J.Clin. Oncol. 29, 15391546 (2011).

Acknowledgements

This work was supported by grants from the National Cancer Institute (INCa) (grants 07/3D1616, PLBio200907 and PLBio2010), the Canceropole Ile de France (grant R09194DD), Ville de Paris, the Association pour la Recherche sur le Cancer (ARC) (grants PEMT, 07/3D1616 and 3,185), ARC, Fondation de France, INSERM, Universit Pierre et Marie Curie, Universit Pa r i s D e s c a r t e s, Q a ta r - Fo u n d a t i o n N P R P ( g ra n t 0911743291), the European Commission (7FP, Geninca Consortium, grant 202,230) and the LabEx Immuno-Oncology. The authors thank M. C. Dieu-Nosjean, I. Cremer, D. Damotte, E. Tartour, J. L. Teillaud, as well as the other members of Immune Microenvironment and Tumours and Integrative Cancer Immunology teams of the Cordeliers Reseach Centre for their invaluable contribution. The authors thank T. Fredriksen for helping with the drawing of the figures.

Competing interests statement

The authors declare no competing financial interests.

DATABASES
National Cancer Institute Drug Dictionary: http://www.cancer.gov/drugdictionary bevacizumab | ipilimumab | sorafenib | sunitinib

FURTHER INFORMATION
Jrme Galons homepage: http://www.ici.upmc.fr/
ALL LINKS ARE ACTIVE IN THE ONLINE PDF

306 | APRIL 2012 | VOLUME 12 2012 Macmillan Publishers Limited. All rights reserved

www.nature.com/reviews/cancer