DICERing macrophages for reprogramming TAMs

Macrophages are cells of the innate immune system that respond to external stimuli by rapidly adapting their transcriptional programs and behavior. Upon exposure to inflammatory signals or pathogen-associated molecules, macrophages activate a pro-inflammatory program (called “classical” or M1 activation) that enhances their direct microbicidal functions and, arguably, their ability to stimulate the immune system. Vigorous M1 activation of macrophages, as seen during acute inflammation, also promotes cell death and tissue damage. However, macrophages can likewise acquire other activation states. For example, “alternatively” (or M2) activated macrophages tune down acute inflammation and promote tissue remodeling and repair.¹

In the context of cancer, macrophages are traditionally regarded as M2-polarized and, indeed, are immunosuppressive and ostensibly pro-tumoral.¹ However, increasing evidence indicates that, in both human and murine tumors, tumor-associated macrophages (TAMs) display more nuanced M1- and M2-like traits, which are modulated by the highly variegated repertoire of signals present in distinct – and rapidly evolving – tumor microenvironments. Accordingly, TAMs can both limit and facilitate tumor growth.² There is also evidence for the existence of molecularly and functionally distinct TAM subsets, some of which have more marked pro-tumoral functions. For example, perivascular M2-like TAMs (also known as TIE2-expressing macrophages) facilitate angiogenesis and cancer cell dissemination by increasing vascular permeability.³ Tumors also contain macrophages that express an attenuated M1 (or M1-like) phenotype.⁴ However, their potential anti-tumoral and immunostimulatory functions may be blunted by the strongly immunosuppressive signals that emanate from other tumor-associated cells, including M2 TAMs. Thus, interventions aimed to broadly (re)programming TAMs toward an M1 phenotype may curb cancer-associated immunosuppression and help energize anti-tumor immunity.²

We recently showed that myeloid-specific inactivation of the microRNA (miRNA)-processing enzyme, DICER, promotes the functional programming of TAMs to an M1-like phenotype that can delay, or even prevent, tumor growth in mice.⁵ miRNAs are small non-coding RNAs that finely tune gene expression post-transcriptionally. Conditional deletion of Dicer significantly abated miRNA levels in TAMs, but did not obviously influence endogenous miRNA levels in other tumor-infiltrating myeloid cells. When challenged with tumors, mice lacking DICER in macrophages displayed varying degrees of tumor growth retardation, depending on the tumor type.⁵ Intriguingly, in each tumor model analyzed, the DICER-deficient TAMs had acquired a strongly M1-like phenotype, characterized by enhanced expression of pro-inflammatory cytokines and T-cell–recruiting chemokines (Fig. 1). Likely as a consequence of this phenotypic switch in the TAMs, the tumors attracted more cytotoxic T lymphocytes (CTLs), which are key effectors of anti-tumor immunity.¹ Accordingly, the elimination of CTLs rescued tumor growth in the DICER-deficient mice.⁵

Figure 1.

Suppressing DICER activity reprograms TAMs. The genetic inactivation of Dicer induces M2-like TAMs (left, light blue cell) to acquire features of M1-like TAMs (right, light green cell). This is associated with the upregulation of genes denoting a pro-inflammatory signature, including the T cell chemoattractants Cxcl9 and Cxcl10. The production of CXCL9 and 10 by the reprogrammed TAMs enhances CTL infiltration in the tumors; in turn, CTL-derived IFNγ exacerbates M1 programming of the TAMs. Restoring the expression of Let-7 in the Dicer-deficient TAMs using DICER-independent lentiviral vectors and a haematopoietic -cell transplantation strategy (bottom, light blue cell) attenuates the pro-inflammatory gene signature and partly rescues M2-like activation of TAMs. Conversely, blocking Let-7 activity with a LIN28A-expressing lentiviral vector (top, light green cell) prompts cultured macrophages to acquire M1-like features, including immunostimulatory properties.

Although DICER inactivation was sufficient to initiate M1 programming of macrophages in vitro, the elimination of CTLs attenuated M1 TAM activation, suggesting that a cross talk between the DICER-deficient TAMs and the recruited CTLs had reinforced and sustained M1 macrophage programming in the tumors.⁵ Furthermore, the pharmacological neutralization of CTL-derived interferon-γ (IFNγ) impeded M1 programming of DICER-deficient TAMs. Together, these results imply that DICER and miRNA activity in TAMs may function to limit M1 macrophage activation in response to CTL-derived IFNγ. Of note, IFNγ is a prototypical T helper-1 (Th1) and M1-activating cytokine, as opposed to interleukin-4 (IL4), which is a Th2 cytokine that instead promotes M2 macrophage activation.¹ Suppressing DICER activity may, therefore, make the TAMs more sensitive to IFNγ, and possibly other M1-polarizing stimuli. These may include various Th1 cytokines and pro-inflammatory mediators expressed endogenously in tumors, but also exogenously (therapeutically) administered agents that may serve to activate macrophages and stimulate anti-tumor immunity. Importantly, DICER inactivation in TAMs greatly enhanced the efficacy of cancer immunotherapies, namely PD1 checkpoint blockade and CD40 agonistic antibodies, leading to complete tumor regressions in mice.⁵

By using bioinformatics approaches to interrogate the transcriptomes of DICER-deficient and proficient TAMs, we identified the Let-7 miRNAs as potential regulators of the M2/M1 activation state of the TAMs. Twelve miRNAs with similar sequence and conserved seed (target recognition) region compose the Let-7 family.⁶ We employed several strategies to experimentally validate the prediction that Let-7 miRNAs oppose M1 activation of TAMs. Genetically rescuing the expression of a representative Let-7 miRNA in the DICER-deficient TAMs was sufficient to partially restore their M2 phenotype and decrease CTL infiltration in the tumors. Moreover, blocking the maturation of endogenous Let-7 in wild-type macrophages phenocopied some of the features of DICER-deficient macrophages (Fig. 1). These findings suggest that attenuation of Let-7 activity in the DICER-deficient TAMs contributed to increase their responsiveness to IFNγ, enabling full-blown M1 activation. Therefore, inhibiting Let-7 activity may provide a strategy for reprogramming immunosuppressive TAMs into cells capable of stimulating anti-tumor immunity.⁵

In analogy with other therapeutics that target the stromal cell constituents of tumors (such as angiogenesis inhibitors), macrophage-targeting agents do not typically regress tumors in mice,^2,5 and are unlikely to provide benefits to cancer patients when applied as monotherapies.⁷ Nevertheless, reprogramming TAMs to an immunostimulatory mode has the potential to synergize with various anti-cancer therapies that, either directly or indirectly, function to enhance the endogenous immune response against the tumor.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.