48 Coller, B. S., Arterioscler. Thromb. Vasc. Biol. 2005. 25: 658–670.
Correspondence:Dr. Filip K. Swirski, Center for Systems Biology, Massachusetts General Hospital and Harvard Medical School, Simches Research Building, 185 Cambridge St., Boston, MA 02114, USA
Fax: 11-617-643-6133
e-mail: fswirski@mgh.harvard.edu Additional Correspondence:Dr. Mikael J. Pittet, Center for Systems Biology,
Massachusetts General Hospital and Harvard
Medical School, Simches Research Building, 185 Cambridge St., Boston, MA 02114, USA Fax: 11-617-643-6133
e-mail: mpittet@mgh.harvard.edu Received: 5/5/2011
Revised: 20/6/2011 Accepted: 1/8/2011
Key words: Atherosclerosis Cancer Monocyte
Abbreviations: LXRs: liver X receptors MDSCs: myeloid-derived suppressor cells
PPARs: peroxisome proliferator-activated receptors TAMs: tumor-associated macro-phages
See accompanying Viewpoints:
http://dx.doi.org/10.1002/eji.201141719 http://dx.doi.org/10.1002/eji.201141894 The complete Macrophage Viewpoint series is available at:
http://onlinelibrary.wiley.com doi/10.1002/eji.v41.9/issuetoc
Cancer-promoting tumor-associated macrophages: New vistas and open questions
Alberto Mantovani1, Giovanni Germano1, Federica Marchesi1, Marco Locatelli2 and Subhra K. Biswas3
1Istituto Clinico Humanitas IRCCS and Dept. Translational Medicine, University of Milan, Rozzano, Italy 2Department of Neurosurgery, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, University of
Milan, Milano, Italy
3Singapore Immunology Network, Agency for Science, Technology and Research, Singapore DOI 10.1002/eji.201141894
Tumor-associated macrophages
(TAMs) are key components of the
tumor macroenvironment.
Cancer-and host cell-derived signals gener-ally drive the functions of TAMs towards an M2-like polarized,
tumor-propelling mode; however, when
appropriately re-educated. TAMs also have the potential to elicit tumor
destructive reactions. Here, we
discuss recent advances regarding the immunobiology of TAMs and high-light open questions including the mechanisms of their accumulation
(recruitment versus proliferation),
their diversity and how to best ther-apeutically target these cells.
Introduction
It is estimated that about 25% of cancers are linked to chronic inflammation sustained by chronic infections (e.g. inflammatory bowel disease, IBD) or inflammatory conditions of diverse origin (e.g. prostatitis) [1]. Moreover, inflam-matory cells and mediators are also present in the microenvironment of
virtually all tumors that are not epide-miologically related to inflammation [1]. Two pathways link inflammation and cancer. The first, the extrinsic pathway, is driven by exogenous conditions which cause non-resolving smouldering inflam-matory responses; the second, the intrin-sic pathway, is triggered by mutation of either oncogenes or tumor suppressor genes that activate the expression of inflammation-related programmes [1]. Macrophages are a key component of cancer-related inflammation (CRI) [1–4]. In many tumors, a correlation between increased numbers and/or density of macrophages and poor prognosis has been observed [5–7]. The molecular
mechanisms underlying the
cancer-promoting activities of tumor-associated macrophages (TAMs) include the promo-tion of proliferapromo-tion and survival of malignant cells, the subversion of adap-tive immune responses, and the promo-tion of angiogenesis, stroma remodelling and metastasis formation [1, 2]. More-over, there is evidence strongly suggest-ing that TAMs are critical determinants of the tumor response to hormonal therapy and chemotherapy [8–10].
Here, we will discuss recent advan-ces that have shed new light onto the immunobiology of TAMs and their
viability as therapeutic targets. In
particular, we will emphasize the open questions and stumbling blocks to diagnostic and therapeutic exploitation
building on previous reviews that
provide the framework for the present contribution, which by virtue of being a Viewpoint is restricted in length [1–4].
TAM accumulation
It has long been held that TAMs originate from circulating monocytes and that tumor-derived chemoattrac-tants play a key role in monocyte recruitment [11] (Fig. 1). Chemokines (e.g. CCL2) have been known to be associated with macrophage infiltration in experimental and human tumors for over 25 years (e.g. [12–16]). For instance, gliomas are characterized by the long recognized production of a wide spectrum of chemokines, by the expression of chemokine receptors and by the correlations between chemokine Eur. J. Immunol. 2011. 41: 2470–2525 Macrophage Viewpoints
2522
production/receptor expression and clinical outcome; the chemokine reper-toire of gliomas includes CCL2, CXCL8, CXCL12, CXCL10, CCL3L1, CX3CL1 (e.g. [17–19]).
It has been assumed that chemokine-driven entry of monocytes accounts for the maintenance of TAM levels in a growing tumor; however, physiologically, mouse microglia cells relevant to gliomas are self-sustained and do not depend on continuous replenishment from the circu-lation [20]. Moreover, in a mouse model of type II inflammation using nematode infection, local IL-4-dependent macro-phage proliferation is shown to be a key determinant of the numbers of M2-polar-ized macrophages in the lung [21]. In the same vein, a recent paper in this Journal by Taylor and colleagues demonstrates that macrophage accumulation in the
inflamed peritoneum depends on
macrophage proliferation [22]. Early studies also showed in situ proliferation of TAMs (e.g. [23–25]) (Fig. 1). For instance, in a murine sarcoma, a para-crine circuit involving M-CSF produced by tumor cells and high levels of c-fms production by TAMs was identified as a mechanism of macrophage proliferation and survival in situ [25]; however, M-CSF is a relatively poor M-CSF in humans and therefore it remains unclear whether proliferation is a major determinant sustaining human macrophage numbers in tissues [26, 27], although proliferation [28] has been detected in human monocytes. Furthermore, macrophage mitosis has rarely been observed in specimens from human tumors, with the exception of Kaposi’s sarcoma [11].
Given the old and new findings discussed here, the issue of in situ macrophage proliferation as a mechan-ism contributing to TAM levels needs to be re-examined using the current tech-nology and focusing on human tumors, particularly as this issue may have important bearings on the development of therapeutic strategies.
TAM diversity: Both location and cancer-type matter
Plasticity and diversity are hallmarks of the mononuclear phagocyte system
[1–4, 29, 30] and there is now evidence that TAMs in murine tumors consist of cell populations with substantial differ-ences [31]. In a transplanted mammary carcinoma in normoxic and hypoxic areas TAMs were reported to have an M1-like and M2-like phenotype, respec-tively [30]. The microanatomical
distri-butions of cells with different
phenotypes and their relation to mono-cyte subsets [30] remain to be deter-mined, however.
A further layer of diversity, namely the diversity of mechanisms responsible for TAM generation, occurs at the level of the tumors and organs involved. Here are some telling examples. In a model of human papilloma virus-driven
squa-mous epithelium carcinogenesis, a
remote control pathway involving
CD41 T cells, B cells, antibodies and
Fcg receptors are responsible for the M2-like phenotype of tumor-promoting TAMs [32]. In contrast, in a mammary carcinoma, Th2-derived IL-4 promotes M2 polarization and metastasis [33]. In a transplanted mammary carcinoma, complement was shown to be involved in myelomonocytic cell recruitment
[34], although the mechanisms
responsible for complement activation in tumors remain to be defined. It should be noted that B cells can use tools other than antibodies to skew
macrophage function and promote
tumor progression, including IL-10 and
lymphotoxin (LT) [35–38], and B
regulatory cells may be well suited for this [38]. Thus, the mechanisms of co-opting the function of myelomono-cytic cells in tumors can differ consid-erably, although M2-like skewing is a recurrent common denominator.
Treg cells are characterized by
expression of the Foxp3 transcription factor. It has recently been observed that a subset of mouse macrophages
express Foxp3 [39]. Foxp31
macro-phages appear late in the B16 mela-noma and, upon adoptive transfer, promote tumor growth. It will be important to confirm and extend these observations to human TAMs.
The diversity of TAMs impacts on the design of therapeutic strategies given that, in different tissues and tumors, the pathways orchestrating TAM develop-ment and cancer-related inflammation
can differ considerably, and that
myelomonocytic cells come in different
Figure 1. Pathways responsible for macrophage commutation and skewing of macrophage function in the tumor microenvironment. The picture is a compound of data obtained in different tumors. Lymphoid cell-derived signals, including Th2 cells, Treg cells and B cells, influence the TAM phenotype. Moreover, tumor cells and stromal cells are a source of mediators influencing recruitment, proliferation and functional orientation of TAM.
Eur. J. Immunol. 2011. 41: 2470–2525 Macrophage Viewpoints FORUM 2523
flavours in different tumor contexts. Nonetheless, mononuclear phagocytes remain as essential common constitu-ents of cancer-related inflammation and determining their diversity in different human cancers is a prerequisite to translate recent progress in this field into clinical benefit.
Therapeutic targeting
In most established malignant tumors TAMs express an M2-like phenotype, although there are variations and excep-tions to this common theme [40]. The TAM phenotype is reversible and these cells can be re-educated to exert anti-tumor activity [41–43]. In a recent study [44], anti-CD40 antibodies with agonist activity have been shown to have anti-tumor potential in both murine and human carcinoma of the pancreas. Circumstantial evidence suggests that the anti-tumor activity of CD40 agonist antibodies is mediated by macrophage
activation. These recent [44] and
previous [4] results, including data from patients suggest that it is indeed possible to tip the macrophage balance in an anti-tumor direction. [10].
Tumor-targeted monoclonal
anti-bodies are part of the cancer ther-apeutic armamentarium. Macrophages can mediate antibody-dependent cellu-lar cytotoxicity and there is evidence that M2-polarized macrophages phago-cytose antibody-sensitized lymphoma cells more efficiently than non-polarized macrophages [45]. This observation may explain the apparently divergent prognostic significance of TAMs in
follicular lymphoma treated with
chemotherapy alone (unfavourable)
and with anti-CD20-containing regi-mens (favourable) [46]. Thus, the interaction of macrophages in different states of activation with antibodies and more generally with B cells is context-dependent and can result in promotion of carcinogenesis (e.g. [33, 35, 38, 47]) or anti-tumor activity [3].
Reducing the numbers or eliminat-ing TAMs are alternatives to their re-education. Strategies include inter-ference with the M-CSF axis [13] or
chemokine expression/function (in
particular CCL2, e.g. [13, 15]), or
depletion of angiogenic monocyte
subsets [48, 49]. Moreover, the
influ-ence of selected chemotherapeutic
agents on the viability and function of myelomonocytic cells may contribute to or be a critical component of the agents’ anti-tumor activity [9, 50, 51]. Thus, the current understanding of TAM biology opens up different and alter-native strategies to target these cells, although the careful definition of the
immunobiology of the context in
different tumors, as noted in the previous section, may be required for successful therapeutic exploitation.
Concluding remarks
Macrophages can act as a double-edged sword in cancer [4, 11]. It has long been known that appropriately
acti-vated mononuclear phagocytes can
express anti-tumor activity in vitro and in vivo by direct killing of tumor cells and/or by eliciting vascular damage and tissue destruction; however, in the Darwinian evolution of tumors, TAMs are co-opted as an integral component of the neoplastic microenvironment. Identification of the pathways respon-sible for the skewing of macrophage function raises the possibility of macro-phage-targeted therapies, complemen-tary to cytoreductive approaches, and of exploiting TAM as prognostic indicators [7]; however, a number of open ques-tions remain as discussed, including the role and mechanism of recruitment versus in situ proliferation, and the diversity both within the tumor micro-environment and between different tissues and tumors. A better under-standing of these issues may help the better exploitation of the diagnostic and therapeutic potential of TAMs.
Acknowledgement: Alberto Mantovani is supported by Associazione Italiana per la Ricerca sul Cancro, Special Project 5x1000. Conflict of interest:The authors declare no financial or commercial conflict of interest.
1 Mantovani, A. et al., Nature 2008. 454: 436–444.
2 Qian, B. Z. et al., Cell 2010. 141: 39–51. 3 Biswas, S. K. et al., Nat. Immunol. 2010.
11: 889–896.
4 Mantovani, A. et al., Curr. Opin. Immu-nol. 2010. 22: 231–237.
5 Bingle, L. et al., J. Pathol. 2002. 196: 254–265.
6 Kurahara, H. et al., J Surg Res. 2011. 167: e211–e219.
7 Steidl, C. et al., N. Engl. J. Med. 2010. 362: 875–885.
8 DeNardo, D. G. et al., Cancer Discov. 2011. 1: 54–67.
9 Germano, G. et al., Cancer Res. 2010. 70: 2235–2244.
10 Balkwill, F. et al., Clin. Pharmacol. Ther. 2010. 87: 401–406.
11 Mantovani, A. et al., Immunol. Today 1992. 13: 265–270.
12 Bottazzi, B. et al., Science 1983. 220: 210–212.
13 Qian, B. Z. et al., Nature 2011. 475: 222–225.
14 Mantovani, A. et al., Cytokine Growth Factor Rev. 2010. 21: 27–39.
15 Zhang, J. et al., Cytokine Growth Factor Rev. 2010. 21: 41–48.
16 Lazennec, G. et al., Trends Mol. Med. 2010. 16: 133–144.
17 Sciume, G. et al., Blood 2011. 117: 4467–4475.
18 Locatelli, M. et al., Eur. Cytokine Netw. 2010. 21: 27–33.
19 Marchesi, F. et al., J. Neuroimmunol. 2010. 224: 39–44.
20 Ajami, B. et al., Nat. Neurosci. 2007. 10: 1538–1543.
21 Jenkins, S. J. et al., Science 2011. 332: 1284–1288.
22 Davies, L. C. et al., Eur. J. Immunol. 41: 2155–2164.
23 Stewart, C. C. et al., Int. J. Cancer 1978. 22: 152–159.
24 Mahoney, K. H. et al., J. Leukoc. Biol. 1987. 41: 205–211.
25 Bottazzi, B. et al., J. Immunol. 1990. 144: 2409–2412.
26 Martinez, F. O. et al., J. Immunol. 2006. 177: 7303–7311.
27 Solinas, G. et al., J. Immunol. 2010. 185: 642–652.
28 Hamilton, J. A., Nat. Rev. Immunol. 2008. 8: 533–544.
29 Mosser, D. M. et al., Nat. Rev. Immunol. 2008. 8: 958–969.
30 Geissmann, F. et al., Science 2011. 327: 656–661.
31 Movahedi, K. et al., Cancer Res. 2010. 70: 5728–5739.
32 Andreu, P. et al., Cancer Cell 2010 17: 121–134.
33 DeNardo, D. G. et al., Cancer Cell 2009. 16: 91–102.
34 Markiewski, M. M. et al., Nat. Immunol. 2008. 9: 1225–1235.
Eur. J. Immunol. 2011. 41: 2470–2525 Macrophage Viewpoints
2524
35 Wong, S. C. et al., Eur. J. Immunol. 2010. 40: 2296–2307.
36 Ammirante, M. et al., Nature 2010. 464: 302–305.
37 Olkhanud, P. B. et al., Cancer Res. 2011. 71: 3505–3515.
38 Schioppa, T. et al., Proc. Natl. Acad. Sci. USA 2011. 108: 10662–10667.
39 Zorro Manrique, S. et al., J. Exp. Med. 2011. 208: 1485–1499.
40 Torroella-Kouri, M. et al., Cancer Res. 2009. 69: 4800–4809.
41 Duluc, D. et al., Int. J. Cancer 2009. 125: 367–373.
42 Watkins, S. K. et al., Eur. J. Immunol. 2009. 39: 2126–2135.
43 Goubau, D. et al., Eur. J. Immunol. 2009. 39: 527–540.
44 Beatty, G. L. et al., Science 2011. 331: 1612–1616.
45 Leidi,M. et al., J. Immunol. 2009. 182: 4415–4422.
46 Taskinen, M. et al., Clin. Cancer Res. 2007. 13: 5784–5789.
47 Sica, A. et al., Eur. J. Immunol. 2010. 40: 2131–2133.
48 De Palma, M. et al., Trends Immunol. 2007. 28: 519–524.
49 Huang, H. et al., Clin. Cancer Res. 2011. 17: 1001–1011.
50 Apetoh, L. et al., Nat. Med. 2007. 13: 1050–1059.
51 Kodumudi, K. N. et al., Clin. Cancer Res. 2010. 16: 4583–4594.
Correspondence:Prof. Alberto Mantovani Istituto Clinico Humanitas IRCCS, University of Milan, Via Manzoni 113, 20089 Rozzano, Italy Fax: 139-0-02-8224-5101 e-mail: alberto.mantovani@humanitasresearch.it Received: 24/6/2011 Revised: 18/7/2011 Accepted: 20/7/2011
Key words: Cancer Chemokines Inflam-mation Macrophages Tumor-associated macrophages
Abbreviation: TAM: tumor-associated macrophage
See accompanying Viewpoints:
http://dx.doi.org/10.1002/eji.201141719 http://dx.doi.org/10.1002/eji.201141727 The complete Macrophage Viewpoint series is available at:
http://onlinelibrary.wiley.com doi/10.1002/eji.v41.9/issuetoc
Eur. J. Immunol. 2011. 41: 2470–2525 Macrophage Viewpoints FORUM 2525