Macrophages are derived from the mononuclear phagocyte system (5). Progenitor cells in the bone marrow can give rise to two types of cells: neutrophils and macrophages. The macrophage pathway is committed to produce monoblasts through serial divisions of promonoblasts from the colony forming unit, granulocyte-macrophage (CFU-GM). Peripheral blood monocytes, derived from bone marrow progenitor cells that give rise to monoblasts and promonocytes, eventually leave the blood stream and enter the tissue where they differentiate into macrophages (6). In the tissues, they have different roles based on the tissue in which they reside. Leek and Harris (5) described macrophages as the ‘Swiss army knife’ of the immune system. However, when they are activated appropriately, macrophages have specific and distinct functions. Many signals influencing macrophages have been identified and include hypoxia, lipids, and cytokines (e.g., CCL2, CCL3 and CSF-1), TNF-α, INF-γ and MIF (7). Macrophage migration into tissues is not haphazard; rather, this process is specifically orchestrated by various chemoattractants. Among the chemoattractants, CCL2 (formally referred to as monocyte chemoattractant prorein-1, MCP-1) is suggested to be important in tumor progression. Fujimoto et al showed recently that stromal CCL2 is associated with recruitment of macrophages to the breast cancer stroma and plays a role in relapse-free survival (8). Macrophage migration inhibitory factor (MIF) is another chemoattractant involved in tumor progression. MIF, an upstream regulator of host immunity, constitutes an important link between chronic inflammation and cancer. By interacting with CXCR2 and CXCR4, MIF recruits leukocytes and activates cellular inflammatory responses (9). Its levels were found to be elevated in different cancers, particularly prostate and breast cancer. Of particular interest, intratumoral MIF levels were inversely related to nodal status, and this determined an interaction between MIF and intratumoral interleukins (10). A series of experiments with breast cancer cell lines and macrophage cell lines showed that MIF was a major gene product that was upregulated when breast cancer cells were cocultured with macrophages. MIF secreted by tumor cells increased the macrophage metalloproteinase production rates and facilitated tumor cell invasion (11).

After entering peripheral tissues, the pattern of macrophage activation depends on the surrounding microenvironment, which accounts for the heterogeneity among macrophage populations and their plasticity. Typically, macrophages are divided into two main groups, the M1 and M2 macrophages.

M1 macrophages (also known as ‘classically’ activated macrophages) are an integral cellular component of the immune system. These cells play an important role in protection against intracellular pathogens and cancer cells by virtue of the type I immune response they activate. They are ‘primed’ by the cytokine INF-γ for activation either by tumor necrosis factor-α (TNF-α) or (and more importantly) by activation of toll-like receptors (TLRs) via exposure to microbes or microbial products such as bacterial LPS (12). Once activated, M1 macrophages start secreting pro-inflammatory cytokines such as interferons and interleukins. M1 macrophages particularly secrete high levels of IL-12 and in humans they also secrete high levels of IL-23 (13). M1 macrophages also function as antigen-presenting cells; a function that is regarded as a potentially antitumor function (14). These cells have an amplified capacity to kill intracellular pathogens by generating toxic oxygen species and by activating the inducible NO synthase (iNOS) gene to produce nitric oxide (NO). Therefore, M1 macrophages can defend the host from tumor cells and infections (14–16) (Fig. 1).

The M2 group of macrophages is referred to as tumor-associated macrophages as they generally promote tumor growth and metastasis (17). They are also referred to as ‘alternatively activated’ macrophages (M2). Unlike M1 macrophages that are activated by INF-γ, M2 macrophages are activated by other cytokines or by immune complexes. This activation induces a T-helper 2 type of response and as such, they are appropriately called M2 macrophages (reflecting this Th2-type response). However, it has recently been suggested that this M1/M2 dichotomous assignment of macrophages might be too simple and does not convey the different functions and activation patterns of M2 macrophages. An alternative classification was proposed by Mantovani et al whereby M2 macrophages are divided into three sub-populations depending on the particular stimulus that activates them (17). M2a (alternative) macrophages are induced by exposure to IL-4 and IL-13 while M2b macrophages are induced by toll-like receptor ligands (e.g., LPS) and immune complexes. Both the M2a and M2b groups promote the Th2 immune response, by secreting Th2 type cytokines and by recruiting Th2 cells (17). The M2c macrophage subgroup represents macrophages that are activated by IL-10 and glucocorticoid hormones. In general, an M2c response is responsible for suppressing inflammation which is an immunologically a Th-1 type of response (Fig. 2).

Macrophages seem to be directed by the tumor microenvironment and in turn depending on the surrounding milieu will acquire a trophic property as opposed to the original immune role they had (18). It is clear that different stimuli polarize macrophages into different phenotypes. In vitro models have shown that macrophages can change their polarization pattern according to the different stimuli (19). This suggests their dynamic interaction with the tumor microenvironment changing it constantly favoring tumor cell invasion and subsequent metastasis.

Lower oxygen concentrations in tumors have been verified and reproduced in many studies. In solid tumors the chaotic disorganized array of vascularization with blind ended vessels and leaky endothelial lining leads to a low oxygen tension (hypoxic condition) in such tumors (20). Breast cancer, like other solid malignancies, is hypoxic (21), making the oxygen supply insufficient for the rapidly growing tumor. In addition to the areas of low oxygen supply, necrotic areas in a tumor bed develop an acute drop in oxygen with ensuing cell death. This hypoxic environment within the tumor is one factor that influences macrophage function as debris in this area attracts macrophages (22). The research done by Turner et al showed that once attracted, macrophages are entrapped in necrotic tumor areas (23). One explanation for this phenomenon is that upregulation of mitogen-activated protein kinase phosphatase (MKP-1) dephosphorylates chemoattractant receptors for VEGF and CCL2 (VEGFR and CCR2, respectively), thus aborting their chemotactic response in TAMs (24,25). This accumulation of TAMs in hypoxic areas has been associated with aggressive tumor behavior in breast cancer (26).

Hypoxia is known to regulate gene expression in both cancer and inflammatory cells within the tumor microenvironment. Hypoxia-induced gene expression in macrophages is in part due to the upregulation of the transcription factors HIF-1α and HIF-2α (27). HIF-1α is a marker for hypoxia and in breast cancer it has been linked to aggressive malignant behavior (28). Indeed, HIF-1α-dependent genes are associated with an increased patient mortality in some cancers (29). So far VEGF remains the most notable gene upregulated by HIF-1α and HIF-2α secondary to hypoxic stress in tumor microenvironment. Lewis et al showed that TAMs express VEGF almost exclusively in perinecrotic and poorly vascularized areas in breast cancer (30). In breast cancer, the overexpression of HIF-2α in TAMs was related to increased tumor vascularity from the upregulation of VEGF expression by the effect of HIF-2α (31).

Hypoxia-induced HIF-1α also promotes tumor progression by promoting immune evasion. Under hypoxic conditions TAMs secrete more immunosuppressive cytokines such as IL-10, which inhibits immune effector T-cells (32). In fact, using a murine model of breast cancer Doedens et al recently reported that tumor growth is decreased with deletion of HIF1α in macrophages, despite normal levels of VEGF and tumor vascularization. In this model, TAMs suppressed the antitumor activity of tumor-infiltrating T cells (33). The loss of HIF-1α in myeloid cells reversed the hypoxia-induced suppression of T-cell activation resulting in tumor progression. Overall, this study demonstrated that T cells of the adaptive immune response are suppressed by the myeloid derived cells of innate immune system in a HIF-1α-dependent and hypoxia-induced fashion.

Promoting tumor metastatic behavior is another mechanism whereby macrophages in hypoxic breast cancers increase tumor aggressiveness. Macrophage migratory inhibitory factor (MIF) is one of the gene products increased by HIF-1α upregulation in hypoxic conditions. It was found that MIF is released by hypoxic macrophages (34). Recently, Oda et al showed that the effect of MIF on HIF-1α is p53-dependent process and that MIF stabilizes HIF-1α protein under hypoxic conditions (35). MIF's stimulatory effect on the release of matrix metalloproteinases [e.g., MIF-induced increased expression of MMP-9 (36)], which degrade the basement membranes in tumors, opens the gates for tumor cells to escape and initiate a nidus of metastatic foci.

Together the above studies reveal that hypoxic conditions drive accumulation of TAMs in necrotic and poorly vascularized areas of breast cancer. Once in the tumor, the microenvironment influences macrophages to enhance the aggressive behavior of tumors. These hypoxia-mobilized pro-tumoral activities of TAMs may serve as potential targets for breast cancer therapy. By interfering at various levels prior to metastatic and invasive tumor activity the progression induced by macrophages may be reversed. For example, since radiation therapy is known to induce hypoxia in tumor beds, blocking VEGF or CCL2 could potentially stop the recruitment of macrophages to these hypoxic tumor areas and slow or stop tumor progression.

Focal TAM infiltration was shown to have an effect on prognosis in breast cancer. It is well known now that TAMs can affect tumors in different ways, and one of the major pathways is by stimulating angiogenesis (26). TAMs in breast carcinomas express numerous tumor-promoting factors, the most prominent of which is the proangiogenic vascular endothelial growth factor (VEGF) (30). In a series of breast cancer cases, Leek et al found that there was a positive correlation between high mean vessel density and increased macrophage index in the tumor, as well as an inverse relationship between macrophage counts and relapse-free and overall survival (26). Subsequently, they showed that there is an association between vascular endothelial growth factor (VEGF) and macrophage infiltration (37) increasing the evidence for a proangiogenic role for TAMs in breast cancer. Therefore, it became reasonable to think that the effect on prognosis which macrophages have is in part related to their proangiogenic effect. Indeed, breast cancer spheroids containing macrophages induce overexpression of VEGF and an increase in the number of vessels when injected in dorsal skin folds of nude mice (38). Further evidence comes from a study that showed the level of post-surgery circulating serum VEGF levels were directly dependent on macrophage infiltration of the primary tumor as well as the presence of somatic mutation of p53 gene (39). However, this relationship needs to be further investigated to validate its effect on the metastatic activity of macrophage rich tumors.

Using an elegant transgenic mouse model, Lin et al demonstrated that inhibition of macrophage maturation and infiltration into tumors delayed the formation of a high-density vessel network. In this model, genetically deleting colony stimulating factor-1 (CSF-1), a key chemoattractant of macrophages, makes PyMT mice susceptible to mammary cancer. However, restoring CSF-1 showed that tumor-associated macrophages play a key role in fostering tumor angiogenesis which is an essential step in the tumor progression to malignancy (40). They later elaborated that genetically restoring VEGF to the tumor microenvironment in these macrophage-depleted animals restores tumor progression through macrophage-produced VEGF stimulating tumor angiogenesis (41).

Gene expression profiling has been utilized to understand the different macrophage populations in breast cancer. In an experimental mouse model, Ojalvo et al were able to use a flow cytometry technique to isolate a specific TAM population from late stage mammary tumors that was previously shown to be associated with an invasive mammary carcinoma. For comparison, they isolated a similar macrophage population from the spleens of mice that did not have mammary tumors. Gene expression signatures for 460 genes from both populations were studied. They found that the transcript abundance was differentially regulated between the two populations. Macrophage functions such as suppression of immune activation, matrix remodeling and secretion of tumor angiogenesis mediators were higher in the TAM population from the mammary tumor (42). This indicates that a TAM gene expression signature in mouse tumors could be used to assess expression of TAMs in human breast cancer. These data suggest TAMs may regulate tumor angiogenesis and therefore support further investigation into the role of other transcriptional mediators of TAM function within the tumor microenvironment (42). In addition, gene expression profiling may be a promising methodology to explain the different populations of TAMs and their potential functions. It also may be helpful in differentiating tumors by their aggressive potential.

From the highlighted studies, it is clear that macrophages contribute to tumor growth, invasion, and metastasis through an angiogenic pathway. Clinically, macrophages could be a potential target of anticancer therapy, where possibly a treatment would limit metastatic spread and thus increase overall survival. In the angiogenesis pathway, clinical trials have already been studying the targeting of proangiogenic factors such as VEGF. Therefore, an indirect targeted therapy against macrophages or their products is already underway, yet they are not selectively aimed at macrophages which could provide excellent opportunities for more targeted therapies (43). Moreover, detailed gene expression profiling of tumor-associated macrophages will eventually be used to determine what pathway(s) to target for each particular tumor.

TAMs produce several enzymes which can degrade the extracellular matrix (ECM). Such enzymes include several metalloproteinases (e.g., MMP-2 and MMP-9) as well as urokinase-type plasminogen activator (uPA) that degrade the ECM (44). Dissolution of the ECM leads to cleavages through which tumor cells can evade and metastasize. Clinically, uPA has been under investigation in breast cancer. In one study, it was shown that those patients whose uPA activity was elevated in breast cancer tissue had a significantly shorter disease-free interval compared to those with low levels of activity (45). Recently uPA and plasminogen activator inhibitor type 1 (PAI-1) have received greater interest in breast cancer research and treatment and are possible novel tumor biological markers. In fact, uPA is possibly a potent independent prognostic factor in breast cancer (46). Currently, the NNBC-3 trial is investigating treatment of node-negative early stage breast cancer classified as high or low risk based on uPA and its inhibitor PAI-1 (47). Therefore, this draws attention to the importance of the ECM and macrophages in breast cancer progression and even targeted therapy.

Fibrillary collagen is an important substrate of the extracellular matrix. Using multi-photon microscopy Wyckoff et al revealed these structures and found that surrounding the tumor there is a dense ring of fibrillary collagen (48). Moreover, tumor cells were seen moving alone the collagen fibers ten times faster than they do through the stroma. Tumor cells move directly towards vessels along the collagen fibers. Ingman et al, found that depleting macrophages during the process of mammary tumor development in a mouse model, reduced the extent of collagen I synthesis and restoration of macrophages resulted in correcting this defect (49). Further studies are needed to understand how exactly macrophage relate to this collagen fibrillogenesis and tumor invasiveness.