Macrophages are among the most abundant cells in the tumor stroma and can contribute to neoplastic growth, invasion and metastatic diffusion by translating instructive signals delivered by transformed cells. These signals comprise soluble factors such as chemokines and cytokines.1 In many cancers, tumor-associated macrophages (TAMs) are constantly recruited to the tumor environment by the CCL2 chemokine that attracts CCR2+ monocytes circulating in the blood.2 It is generally accepted that the tumor environment polarizes TAMs to express a set of genes common to M2-type macrophages, a specialized subset intervening in inflammation resolution, tissue remodeling and control of parasitic infections.1 These genes include the neoangiogenesis-promoter vascular endothelial growth factor (VEGF) and the L-arginine-metabolizing enzyme arginase (ARG). A recent paper by Colegio et al.3 in Nature opens a new scenario, showing that TAMs can ‘sense’ metabolic changes typical of the malignant state.
The Nobel laureate Otto Heinrich Warburg postulated that glucose cell metabolism was fundamental for tumor progression. The ‘Warburg effect’ defines the prevalent energy production in many cancers by a high rate of glycolysis, resulting in lactic acid secretion even in the presence of oxygen (aerobic glycolysis). This marks a straightforward difference with normal cells, where the oxidative breakdown of pyruvate within the mitochondria is the prevalent source of energy.
The presence of hypoxic areas represents another peculiar feature of the anarchic neoplastic growth. The hypoxia-inducible factor (HIF) is the central mediator of transcriptional responses to hypoxia but it can also be activated by O2-independent pathways.4 HIF proteins form heterodimeric complexes comprising an O2-labile α-subunit (HIF1α, HIF2α or HIF3α) and a stable β-subunit (HIF1β). These complexes recognize and bind hypoxia-responsive elements with a shared RCGTG sequence in target genes. Under normoxic conditions, HIF-specific prolyl-hydroxylases modify HIFα subunits and promote their proteasomal degradation by the von Hippel–Lindau tumor suppressor protein. When cells become hypoxic, posttranslational modification and stabilization of HIF1α and HIF2α subunits increase the transcriptional activity.4
Colegio et al.3 demonstrate that tumor-derived mediators stabilize HIF1α under normoxic conditions, leading to the transcription of the VEGF and ARG1 genes in macrophages. A heat-stable factor present in the low-molecular weight (<3 kDa) fraction of tumor-conditioned medium was able to activate HIF1α. Unexpectedly, this factor was lactic acid, a byproduct of tumor glycolysis (Figure 1).
Among the enzymes involved in the glycolytic cascade, the M2 isoform of pyruvate kinase (PKM2) is predominant in tumor cell lines and its expression levels correlate directly with lactate production in the tumor environment. Upon release by cancer cells, uptake of lactic acid by macrophages requires its active transport by the monocarboxylate transporter on the cell membrane, a process facilitated by low pH. Once inside TAMs, lactic acid induces an HIF-1α-dependent, M2-like transcriptional profile in TAMs (Figure 1).3
One gene considered as an emblem of M2 macrophage orientation is ARG1, as TH2-type cytokines IL-4 and IL-13 are potent inducers of its transcription and activity.1,5 However, the role of these cytokines in regulating ARG1 within the tumor is only partially elucidated.2 In gliomas, for example, GM-CSF released by neoplastic cells can upregulate the IL-4 receptor in tumor-associated myeloid cells promoting ARG1 induction by IL-13.6 Hypoxia, on the other hand, regulates both ARG1 and ARG2 in macrophages, fibroblasts and endothelial cells.7 HIF1α can control ARG1 and another L-arginine-metabolizing enzyme, the inducible isoform of nitric oxide synthase, in TAMs, thus enhancing their immunosuppressive activity on T lymphocytes.8
Colegio et al. demonstrate that lack of IL-4R did not alter ARG1 expression in TAMs, at least in a lung cancer model, suggesting that the tumor environment can alternatively use lactate to influence M2 polarization. On the other hand, HIF1α was required for the regulation of some M2 macrophage-associated genes by IL-4.3 It is thus conceivable that metabolic signals and cytokines can cooperate to shape TAMs in different tumor types (Figure 1).
Two sets of data in this manuscript point to an in vivo role for lactate in macrophage polarization and ARG1 expression. Tumor cell lines lacking PKM2 grew slower and had a reduced amount of ARG1 mRNA; conversely, co-injection of cancer cells with macrophages derived from in vitro cultures of bone marrow cells stimulated with lactate grew more rapidly in mice.3
Whereas the role of VEGF in cancer development is well established, positioning arginase intervention requires further studies. Colegio et al.3 show that mice lacking ARG1 in myeloid cells by LysM promoter targeted deletion had a reduced growth of an implanted, subcutaneous tumor. Arginases are metabolic enzymes present in two isoforms that hydrolyze L-arginine to L-ornithine and urea.5 ARG1 in myeloid cells, including TAMs, could act as a tumor-landscaping gene through two main pathways: supporting tumor growth and suppressing antitumor immune responses. Various tumors, both in humans and mice, express ARG isoforms at certain stages of their development, either in tumor-infiltrating stroma or in the very same neoplastic cells.7 ARG1 activation is able to induce immune suppression by depleting L-arginine in the microenvironment. Reduction of this semi-essential amino acid can inhibit T-cell proliferation through downregulation of CD3ζ chain expression in T lymphocytes.2,5 ARG1 could also have a role without the intervention of adaptive immunity, by promoting tumor cell growth and survival. L-Ornithine, produced downstream of ARG1 activity, is the precursor of polyamines, that is, putresceine, spermidine and spermine, which can act as proliferative signals for mammalian cells. However, only the growth of some transplanted tumors is affected in ARG1 knockout mice (unpublished results), suggesting other potential pathways can bypass the need for ARG1 in the myeloid compartment.
ARG1 might also control tissue remodeling as L-ornitine can be converted into L-proline, which is necessary for collagen synthesis.5 However, whereas the absence of ARG1 in the myeloid compartment resulted in prolonged inflammation and a negative effect on matrix deposition during the wound healing process, liver fibrosis was exacerbated in conditional ARG1 knockout mice exposed to Schistosoma mansoni, a pathology dependent on a TH2 lymphocyte response.9 It is conceivable that ARG1-expressing macrophages might represent a subcategory of M2-like macrophages, operating as suppressors rather than inducers of Th2-dependent inflammation and fibrosis.9
Considering that many tumors cannot influence species evolution, as they arise after the peak of the reproductive age, lactate sensing must have a role in other macrophage responses. Is cytosolic lactate sensing a mechanism to program macrophages toward inflammation resolution and regulation of adaptive immunity? This will certainly be of interest for future researches but we can speculate about at least two conditions.
First, lactic acid-producing bacteria constantly interact with our body and are part of the normal microbiota in the gut and other mucosal surfaces. It is thus conceivable that their fermentation products influence the local macrophage response. Bacterial vaginosis is a common clinical syndrome arising when anaerobic bacteria replace the protective lactic acid-producing bacteria (mainly species of the Lactobacillus genus). Although the specific role of lactic acid remains to be proven, some Lactobacilli strains can exert an anti-inflammatory activity, helping to control colitis severity, by regulating M2 orientation and ARG1 activity in macrophages.10
Second, exposure of bone marrow cells to granulocyte-macrophage colony-stimulating factor (GM-CSF) and IL-6 generates immunosuppressive myeloid cells including macrophages and results in a fast (within 24 h) activation of L-arginine-metabolizing enzymes and increased uptake of glucose, which is mainly metabolized by anaerobic glycolysis.11,12 Thus, the main metabolic changes in myeloid cells exposed to cytokines produced by several tumors, such as GM-CSF and IL-6, can also lead to the accumulation of endogenous lactate. Lactic acid might thus represent a converging crossroad integrating external and internal milieu to regulate L-arginine metabolism and polarization in macrophages (Figure 1). However, this altered metabolic state might promote macrophage death. In fact, although macrophages survive in a hypoxic environment, exposure to lactate levels produced by tumors can cause their dismissal and possibly contribute to their continuous replenishment by circulating precursors as well as their spatial distribution within specific areas of tumors.13