TET3 is a common epigenetic immunomodulator of pathogenic macrophages

Abstract

Through a combination of single-cell/single-nucleus RNA-Seq (sc/snRNA-Seq) data analysis, immunohistochemistry, and primary macrophage studies, we have identified pathogenic macrophages characterized by Tet methylcytosine dioxygenase 3 (TET3) overexpression (Toe-Macs) in 3 major human diseases associated with chronic inflammation: metabolic dysfunction-associated steatohepatitis (MASH), non-small cell lung cancer (NSCLC), and endometriosis. These macrophages are induced by common factors present in the disease microenvironment (DME). Crucially, the universal reliance on TET3 overexpression among these macrophages enabled their selective elimination as a single population, irrespective of heterogeneity in other molecular markers. In mice, depleting these macrophages via myeloid-specific Tet3 KO markedly mitigated disease progression, and the therapeutic effects were recapitulated pharmacologically using a TET3-specific small-molecule degrader. Through an unexpected mode of action, TET3 epigenetically regulated the expression of multiple genes key to the generation and maintenance of an inflammatory/immunosuppressive DME. We propose that Toe-Macs are a unifying feature of pathogenic macrophages that could be therapeutically targeted to treat MASH, NSCLC, endometriosis, and potentially other chronic inflammatory diseases.

Keywords: Epigenetics; Hepatology; Immunology; Immunotherapy; Inflammation; Macrophages.

Figure 1. TET3- and NLRP3-overexpressing macrophages in human MASH liver.

(A) Dimension reduction plots showing all myeloid cells from healthy and MASH livers. (B) Dimension reduction plot showing 9 clusters of total myeloid cells and a dot plot indicating that cluster 0 expresses classical KC markers. pct.exp, percentage of expression; avg.exp, average expression. (C) Dot plots showing expression levels of TET3, TET2, TET1 (left panel), and NLRP3 (right panel) in KCs. (D) IHC images of TET3 (red, top panels) and NLRP3 (red, bottom panels) costaining with CD163 (green) and DAPI (nuclei, blue) in liver tissue sections from MASH. The top panels are serial sections of the bottom panels. (E) IHC images of TET3 and NLRP3 costaining with CD163 and DAPI (nuclei) of liver tissue sections from healthy controls. The top panels are serial sections of the bottom panels. Scale bars: 10 μm. tSNE, t-distributed stochastic neighbor embedding. Refer to Supplemental Table 1 for detailed patient sample information.

Figure 2. Toe-Macs induced by DME factors exhibit elevated NLRP3 activity.

(A) MDMs were treated with Veh, TGF-β1 (10 ng/mL), or CCL2 (200 ng/mL). Proteins were isolated at 24 hours (Veh/TGFβ1) or 48 hours (Veh/CCL2) for Western blot analysis. Protein sizes are in indicated in kDa. (B) MDMs were treated with Veh, TGF-β1 (10 ng/mL), CCL2 (200 ng/mL), or TGF-β1 (10 ng/mL) plus CCL2 (200 ng/mL). Proteins were isolated at 48 hours. (C) MDMs were treated with Veh, TGF-β1 (10 ng/mL), or CCL2 (200 ng/mL), with or without SIS3 (a SMAD3 inhibitor) (5 mM). Proteins were isolated at 48 hours. (D) Western blot analysis of MDMs infected with Ad-GFP or Ad-TET3 for 24 hours. (E) CCL2 expression in MDMs assessed by qRT-PCR (RNA harvested at 24 hours) and ELISA (supernatants harvested at 48 hours). (F) TGFB1 expression in human MDMs assessed by qRT-PCR (RNA harvested at 24 hours) and Western blotting (protein harvested at 48 hours). (G) NLRP3 expression in MDMs assessed by qRT-PCR (RNA harvested at 12 hours) and Western blotting (protein harvested at 24 hours). (H and I) MDMs seeded in 96-well plates were infected with Ad-GFP or Ad-TET3 for 24 hours. Cells were primed with or without LPS at 250 ng/mL for 4 hours. Nigericin was added at 20 mM for 1 hour, followed by measurement of caspase-1 activity (H) and IL-1β protein levels (I). (J) MDMs were infected with Ad-GFP or Ad-TET3 for 24 hours prior to stimulation with 10 ng/mL LPS plus 20 ng/mL IFN-γ for 4 hours, followed by qRT-PCR of IL1B mRNA. (K) ELISA analysis (after 8 hours of LPS/IFN-γ stimulation) of IL-1β of MDMs treated as in J. All data represent the mean ± SEM. *P < 0.05, **P < 0.01, and *** P < 0.001, by 2-tailed Student’s t test. Western blot data are representative of 2–3 biological repeats.

Figure 3. TET3 epigenetically regulates TGFB1, NLRP3, IL1B, and CCL2 expression through interaction with phosphorylated STAT3.

(A) Schematic of the human TGFB1 promoter. Numbers depict nucleotide positions relative to the transcription start site labeled +1. The PCR-amplified region is marked in red, with the zoomed-in sequence shown underneath. The PCR primer sequences are underlined. (B) MDMs were infected with Ad-GFP or Ad-TET3 for 16 hours. Chromatins were prepared for ChIP-qPCR analysis to detect enrichment of the specific TGFB1 promoter region outlined in A. (C) MDMs were infected with Ad-GFP (lane 2) or Ad-TET3 (lane 3) for 24 hours, followed by co-IP using anti-Flag antibody. Western blot analysis was carried out using anti-TET3 or anti-STAT3 (Y705) antibodies. Lane 1 shows 5% of input from Ad-TET3–infected cells. IB, immunoblot. (D and E) MDMs were treated as in B. Genomic DNA was collected and subjected to hMeDIP-qPCR (D) and MeDIP-qPCR (E). (F) Schematic of the human NLRP3 promoter. (G) MDMs were treated as in B, followed by ChIP-qPCR. (H) MDMs were treated as in B, followed by hMeDIP-qPCR. (I) Schematic of the human IL1B promoter. The STAT3-binding site is highlighted blue. (J) MDMs were treated as in B, followed by ChIP-qPCR. (K) MDMs were treated as in B, followed by hMeDIP-qPCR. (L) Basal IL1B expression in MDMs infected with Ad-GFP or Ad-TET3 was assessed by qRT-PCR. (M) Schematic of the human CCL2 promoter. The STAT3-binding site is highlighted blue. (N) MDMs were treated as in B, followed by ChIP-qPCR. In B, G, J, and N, 5 × 10⁵ cells per ChIP were used. All data represent the mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001, by 2-tailed Student’s t test. Western blot data are representative of 2 biological repeats. chr, chromosome.

Figure 4. Bc destabilizes TET3 protein through enhancing a ternary complex formation with VHL.

(A) Co-IP of Flag-TET3 and endogenous VHL in MDMs infected with Ad-TET3 (which expresses a Flag-tagged human TET3), with or without the presence of Bc, at 50 μM for 2 hours. (B) AlphaScreen of dose-response curves of Bc at different concentrations (conc) for binding of TET3, TET2, and TET1 to VHL. (C) Western blots of TET3, TET2, and TET1 in MDMs incubated for 24 hours or 48 hours with Veh or Bc at a final concentration of 10 μM. (D) MDMs were incubated for 2 hours with Veh or Bc at a final concentration of 10 μM, followed by time-course analysis of TET3, TET2, and TET1 in the presence of cycloheximide (CHX) at a final concentration of 50 μg/mL. Cells were harvested at 0, 20, 40, and 60 minutes after addition of CHX. Quantifications are displayed on the right. Western blot data are representative of 2–3 biological repeats.

Figure 5. Toe-Macs are mechanistically linked to diet-induced MASH.

(A) Schematic diagram of the experiments. (B–D) Plasma ALT (B), liver tissue triglycerides (C), and liver tissue hydroxyproline content (D) from mice treated as indicated. (E) Representative immunostaining for TET3 (red) in F4/80⁺ (green) macrophages with quantification of macrophage TET3 median fluorescence intensity (MFI) in liver tissue sections. (F) Immunostaining for NLRP3 (red) and F4/80 (green) in macrophages and quantification of macrophage NLRP3 MFI in liver tissue sections. (G) Immunostaining for IL-1β (green) and quantification of IL-1β MFI in liver tissue sections. (H) Photomicrographs and corresponding statistical analysis of TUNEL⁺ (green) cells. (I) Immunostaining for Ly6G (green) and quantification of Ly6G⁺ cells in liver tissue sections. All data represent the mean ± SEM. n = 6–8 mice per group. Each dot represents a mouse. *P < 0.05, **P < 0.01, and ***P < 0.001, by 1-way ANOVA with Tukey’s post test for the statistical differences versus the WT+Veh group. Scale bars: 25 μm.

Figure 6. Toe-Macs present in human NSCLC epigenetically promote PD-L1 expression.

(A) Dimension reduction plots showing all myeloid cells from adjacent and tumor lung samples. pDCs, plasmacytoid DCs. (B) Bar plot showing macrophage percentages of RTMs and mo-Macs in adjacent and tumor tissue. (C) Dot plot showing expression TET3 and NLRP3 in adjacent RTMs and tumor mo-Macs. (D) Immunostaining for TET3 (red) and CD163 (green) in human NSCLC tumor tissue and adjacent normal tissue. Scale bars: Scale bar: 50 μm. (E and F) CD274/PD-L1 expression in MDMs infected with Ad-GFP or Ad-TET3, assessed by qRT-PCR (E, RNA harvested at 12 hours) and Western blotting (F, protein harvested at 24 hours). (G) PCR-amplified fragment of the CD274 promoter is marked in red, with the zoomed-in sequence shown underneath. The PCR primer sequences are underlined with the STAT3-binding site colored blue. (H) MDMs were treated as in Figure 3B. ChIP-qPCR was performed to detect enrichment of the specific region in the CD274 promoter outlined in G. (I) MDMs were treated as in Figure 3B, followed by hMeDIP-qPCR to detect 5hmC levels in the specific region in the CD274 promoter outlined in G. All data represent the mean ± SEM. *P < 0.05 and **P < 0.01, by 2-tailed Student’s t test. Western blot data are representative of 2 biological repeats. Refer to Supplemental Table 1 for detailed patient sample information.

Figure 7. Toe-Macs contribute to an immunosuppressive TME in NSCLC.

(A) Survival rates of WT and KO mice after LLC inoculation. (B) Lung weights of WT and KO mice injected with LLC cells. (C) Macroscopic images and H&E stains (original magnification, ×10) of lungs from WT and KO mice injected with LLC cells. (D) Experimental design. (E and F) Survival rates and lung weights of mice treated as in D. (G) Macroscopic images and H&E stains (original magnification, ×10) of lungs from mice treated as in D. (H) Immunostaining for TET3 (red) in CD163⁺ macrophages (green) in tumor areas and TET3 (red) in Mac2⁺ macrophages (green) in nontumor areas, and quantification of macrophage TET3 MFI in tumor and nontumor areas. (I) Immunostaining for NLRP3 (red) in CD163⁺ macrophages (green) in tumor areas and NLRP3 (red) in Mac2⁺ macrophages (green) in nontumor areas, and quantification of macrophage NLRP3 MFI in tumor and nontumor areas. (J) Costaining for CD8a (red) and GrB (green) in tumor areas in WT and KO mice, and quantification of CD8⁺ T cell percentages in each group. (K) Costaining for CD8a (red) and GrB (green) in tumor areas of Veh- and Bc-treated mice, and quantification of CD8⁺ T cell percentages in each group. All data represent the mean ± SEM. n = 5 per group. *P < 0.05 and ***P < 0.001, by 2-tailed Student’s t test. Scale bars: 50 μm.

Figure 8. Proposed model.

TET3 overexpression is induced by CCL2 and TGF-β1, which are commonly present in the DME. TET3 downregulates let-7 expression, leading to increased production of IL-6. TET3 is recruited to specific gene promoters through interaction with activated STAT3. By binding to the promoters of NLRP3, IL1B, TGFB1, and CD274, TET3 induces 5hmC DNA modification and demethylation and creates an open chromatin state facilitating transcription. In the case of CCL2, a mechanism independent of DNA modification is involved. This results in an increase in protein production of NLRP3, pro–IL-1β, TGF-β1, CCL2, and PD-L1, as well as enhanced NLRP3 inflammasome activity, which in turn promotes the maturation and release of IL-1β. Furthermore, TGF-β1 and CCL2 released from the Toe-Macs can act as autocrine factors to increase TET3 expression, creating a positive feedback loop with TET3.