Molecular Medicine Israel

NF-κB and TET2 promote macrophage reprogramming in hypoxia that overrides the immunosuppressive effects of the tumor microenvironment

Abstract

Macrophages orchestrate tissue homeostasis and immunity. In the tumor microenvironment (TME), macrophage presence is largely associated with poor prognosis because of their reprogramming into immunosuppressive cells. We investigated the effects of hypoxia, a TME-associated feature, on the functional, epigenetic, and transcriptional reprogramming of macrophages and found that hypoxia boosts their immunogenicity. Hypoxic inflammatory macrophages are characterized by a cluster of proinflammatory genes undergoing ten-eleven translocation–mediated DNA demethylation and overexpression. These genes are regulated by NF-κB, while HIF1α dominates the transcriptional reprogramming, demonstrated through ChIP-seq and pharmacological inhibition. In bladder and ovarian carcinomas, hypoxic inflammatory macrophages are enriched in immune-infiltrated tumors, correlating with better patient prognoses. Coculture assays and cell-cell communication analyses support that hypoxic-activated macrophages enhance T cell–mediated responses. The NF-κB–associated hypomethylation signature is displayed by a subset of hypoxic inflammatory macrophages, isolated from ovarian tumors. Our results challenge paradigms regarding the effects of hypoxia on macrophages and highlight actionable target cells to modulate anticancer immune responses.

INTRODUCTION

Macrophages (MACs) are sentinels of the innate immune system whose fundamental functions encompass phagocytosis, antigen presentation, and modulation of neighboring cells (12). Tissue MACs can originate from precursors established during embryogenesis or, alternatively, differentiate from monocytes (MOs) extravasating from the peripheral blood (3). Before undergoing differentiation within tissues, MOs migrate into niches characterized by diverse physical and biochemical features (4). Within this context, a wide range of interactions between MOs and MACs and their environment shape their phenotype and functions (5). These interactions are mediated by a variety of stimuli occurring either sequentially or simultaneously, highlighting the complexity of signaling pathways involved (6). Pathological states further influence these interactions, leading to altered MAC functions (7). All of these microenvironmental cues, which are specific to tissue and context, have the capacity to modulate MAC phenotype by regulating epigenetic and transcriptional programs (89). Consequently, such signals contribute to the emergence of a heterogeneous range of MACs that coexist in the tissues, with fluctuating proportions in health and disease (10).

Oxygen availability influences MAC function in physiological and pathological situations. Hypoxia is a hallmark of multiple diseased contexts, such as solid tumors (11), arthritic joints (12), or ischemic tissues (13), where appropriate oxygen influx is impaired. Exposure to hypoxic conditions induces the stabilization of hypoxia-inducible factors (HIFs), which are responsible for the cellular adaptation to oxygen deprivation. Among others, HIF transcription factors (TFs) regulate genes related to metabolism, nutrient transport, angiogenesis, and cell migration (14). HIFs have also been described as regulators of inflammatory processes (15). However, the underlying mechanism for this is still controversial, as there is evidence associating HIF activity with both proinflammatory and anti-inflammatory features, depending on cellular and physiological context (1618).

Hypoxia has the ability to rewire the epigenetic landscape of cells (1920), including MACs (21). We and others have demonstrated that DNA methylation, a major epigenetic modification, determines the acquisition of immune features by MACs and other MO-derived cells (2223). Such changes in DNA methylation are dependent on context-specific TFs and occur in orchestration with additional epigenetic modifications such as histone marks (2223). Under conditions of low oxygen levels, the activity of ten-eleven translocation (TET) methylcytosine dioxygenase enzymes, required for active DNA demethylation (24), is hindered (25). In the tumor microenvironment (TME), which is commonly characterized by hypoxia (11), TET inhibition promotes hypermethylation of tumor suppressor genes in cancer cells, affecting their expression (26). Nevertheless, genomic regulation and transcriptional responses induced by hypoxia differ substantially between distinct cell types (1927).

Although there is evidence linking DNA methylation levels with the immunogenic status of the hypoxic TME (28), the direct effect of these alterations on the MAC functions is still unclear. Given the reduced levels of oxygen in the TME and the pivotal role of TET-mediated demethylation for MAC biology, it is likely that hypoxia affects their immunological properties. Moreover, although, in several studies, hypoxia is considered an immunosuppression-associated factor in the TME (162930), the specific effect of oxygen restriction on MACs in this context remains uncertain.

Thus, to study the effect of hypoxia and its role underlying the acquisition of unique phenotypic and molecular features by MACs, we used an in vitro differentiation model from human MOs to MACs in normoxic (21% O2) and hypoxic (1% O2) conditions and additionally stimulated them with lipopolysaccharide (LPS) to evaluate their immune activation potential. We then characterized the functional, epigenomic, and transcriptomic profiles of all those conditions. Our results revealed that MACs differentiated and activated under hypoxic conditions (herein named “mMAC1”) acquire an enhanced proinflammatory program through a DNA methylation–mediated mechanism. The molecular signatures of mMAC1, characterized by nuclear factor κB (NF-κB)–mediated DNA demethylation and overexpression of proinflammatory activation genes, were similarly found in an in vivo MAC subpopulation, isolated from human tumors, suggesting that mMAC1 potentially corresponds to a subset of bona fide tumor-associated MACs. Notably, tumors with high estimated loads of mMAC1 signature–bearing cells were associated with generally better patient survival. Our results highlight the potential of hypoxia as an enhancer of immunogenic properties in MACs in the TME and identify previously unreported mechanisms mediating this phenomenon, posing a refined understanding of the effect of hypoxia upon MACs in vitro and in vivo.

RESULTS

Hypoxia elicits inflammatory features and NF-κB–associated DNA demethylation in activated MACs

To investigate the impact of hypoxia on the immunological properties of MACs and their associated epigenomic reprogramming, we differentiated in vitro human peripheral blood MOs in the presence of MAC colony-stimulating factor (M-CSF) for 5 days in normoxic (21% O2) or hypoxic (1% O2) conditions. In addition, the resulting MACs were treated with LPS for 48 hours to induce their activation/maturation or treated with a vehicle for the same time as a control (Fig. 1A).

Under these conditions, mature hypoxic MACs (mMAC1) produced higher levels of the inflammatory cytokines interleukin-6 (IL-6) and tumor necrosis factor–α (TNF-α) and lower levels of the anti-inflammatory cytokine IL-10 than mature normoxic MACs (mMAC21Fig. 1B). At the cell surface level, mMAC1 expressed higher levels of the major histocompatibility complex (MHC) class II human leukocyte antigen–DR (HLA-DR) and costimulatory proteins CD86 and CD80 than their normoxic counterpart (mMAC21), as determined by flow cytometry (Fig. 1C), which is consistent with an enhanced antigen presentation capacity. On the other hand, resting/immunoregulatory MAC surface proteins CD14, CD206, and CD163 were decreased in mMAC1 versus mMAC21, suggesting a phenotypic switch of these cells to a less anti-inflammatory phenotype. Hypoxic cells, both at steady state and after activation, displayed a decreased capacity to suppress CD8+ T cell proliferation than normoxic cells in a coculture assay (Fig. 1D and fig. S1A). All these results suggest that hypoxia increases the proimmunogenic functions of MACs.

To study the effect of hypoxia on the epigenomic landscape of MACs, we then performed DNA methylation profiling using Illumina Infinium MethylationEPIC arrays. Comparison among conditions revealed a substantial number of differentially methylated positions [DMPs; false discovery rate (FDR) < 0.05 and absolute Δβ > 0.2; see Supplementary Methods] among the different conditions (Fig. 1E, fig. S1B, and table S1A). Specifically, we identified DMPs grouped into three different clusters: cluster C1 (2782 CpGs), C2 (403 CpGs), and C3 (903 CpGs; table S1B). Clusters C1 and C3 corresponded with hypomethylated and hypermethylated CpG sites, respectively, in normoxic MACs when compared to MOs. In these clusters, the methylation tendency was partially inhibited in hypoxia. On the other hand, cluster C2 displayed a marked hypomethylation specifically in mature hypoxic MACs (mMAC1Fig. 1E and fig. S1B) in comparison with mature normoxic MACs (mMAC21).

DMPs in cluster C1, the largest cluster, displayed a tendency consistent with an inhibition of DNA demethylation under hypoxic conditions, as observed in (26), and were enriched in motifs of the activator protein 1 (AP-1) complex, canonically associated with MAC differentiation (Fig. 1F). Similarly, DMPs in cluster C3, which were also enriched in motifs associated with myeloid cell differentiation, such as those in those in the Runt-related (RUNX) and E26 transformation-specific (ETS) transcription factor families (31), displayed hypoxia-associated inhibition of the hypermethylation observed for MACs in normoxia. In contrast, DMPs in cluster C2 displayed hypoxia-associated demethylation specific to activated MACs and were highly enriched in motifs of the NF-κB family, classically associated with Toll-like receptor signaling (32), among others (Fig. 1F). Specific examples of demethylated CpGs in cluster C2 included those in the loci of genes such as IL6 and TNF (fig. S1C), which is consistent with the increased levels of their products in the supernatant of mMAC1 (Fig. 1B).

To characterize the DMP clusters from a genomic standpoint, we annotated them in genomic and CpG context categories. DMPs in clusters C1 to C3, particularly those in cluster C2, were enriched in intergenic and open sea regions (fig. S1D). In addition, reanalysis of public MAC histone mark chromatin immunoprecipitation followed by sequencing (ChIP-seq) data (see Supplementary Methods) revealed that cluster C2 regions gain canonical enhancer (H3K4me1) and enhancer activation (H3K27ac) histone marks after activation in normoxic conditions (fig. S1E), suggesting that C2 regions consist of LPS-dependent de novo enhancers (33). Cluster C2 DMPs also displayed the highest enrichment for human MO enhancer chromatin state, defined by combinations of different histone marks (see Supplementary Methods) (34). These results are consistent with the usual observation that DNA methylation–dynamic genomic regions are associated with distal regulatory elements (35).

All these results indicate that hypoxic MACs display a more proinflammatory phenotype compared with their normoxic counterparts. In addition, hypoxia partially blocks the DNA methylation changes associated with MAC differentiation in normoxic conditions. Last, LPS induces specific NF-κB–associated demethylation in inflammatory genes and enhancers in hypoxic conditions that might contribute to the more proinflammatory phenotype observed in hypoxic MACs.

Hypoxia-induced transcriptomic reprogramming to proinflammatory MACs mainly depends on HIF1α and NF-κB

Next, we analyzed the transcriptome of MACs in all the aforementioned conditions using bulk RNA sequencing (RNA-seq). We identified four different clusters (E1 to E4) comprising a total of 3737 differentially expressed genes (DEGs; FDR < 0.05 and absolute log2 fold change >1; see Supplementary Methods and Fig. 2A and table S2, A and B). E1 (233 DEGs) corresponds to genes up-regulated in hypoxic MACs (iMAC1 and mMAC1). E2 (1452 DEGs) is composed of genes that are up-regulated in activated MACs (mMAC21 and mMAC1) with a more marked increase in mMAC1. E3 (732 DEGs) includes genes that become down-regulated in hypoxia (iMAC1 and mMAC1), which further down-regulate in mMAC1. Last, E4 (1330 DEGs) corresponds to genes that are down-regulated upon activation. Gene ontology (GO) analysis of E1 to E4 clusters revealed an enrichment in functional categories associated with a variety of immune functions (Fig. 2B). In particular, the E2 cluster (up-regulated after activation) was enriched in categories related to response to interferons (IFNs), LPS, and TNF-α, as well as positive regulation of NF-κB (Fig. 2B), consistent with the association of this factor in genes that are demethylated specifically in mMAC1 (Fig. 1F).

Sign up for our Newsletter