IL1β Promotes Immune Suppression in the Tumor Microenvironment Independent of the Inflammasome and Gasdermin D

Chronic inflammation can promote tumor development and progression through various means, such as providing survival signals, suppressing T-cell function, inducing angiogenesis, and enabling invasion and metastasis via tissue remodeling (1). In addition, immune cells recruited to the tumor as part of the inflammatory response often antagonize anticancer therapies (2). Hence, counteracting tumor-promoting inflammation appears to be key to improving disease outcomes in many cancer types (3). This, however, requires a more complete understanding of the mechanisms driving tumor-associated inflammation.

IL1β is a proinflammatory cytokine whose role in cancer is increasingly recognized (4, 5). In a recent study, long-term treatment with a neutralizing IL1β-specific antibody led to a dose-dependent reduction of lung cancer incidence and mortality in a large cohort of patients with atherosclerosis with a history of myocardial infarction (6). The importance of chronic IL1β-driven inflammation in cancer is further supported by the identification of an IL1β-induced transcriptional signature in the peripheral immune cells of patients with renal cell cancer and breast cancer (7, 8). Moreover, p53 mutations, late-stage disease, and the basal-like subtype in breast cancer are all associated with significantly increased IL1B expression and may further augment systemic inflammation (8, 9). IL1β also promotes tumor angiogenesis and the recruitment of myeloid cells (5, 10, 11). In contrast, it supports antitumor T-cell responses and suppresses metastatic outgrowth in mice, suggesting its role to be context dependent (5, 12, 13).

IL1β is produced as a biologically inactive precursor (pro-IL1β). The cleavage of this inactive precursor to generate the active form is typically mediated by caspase-1 (14). Activation of caspase-1 is triggered by the inflammasome, a multiprotein complex that assembles upon activation of intracellular receptors such as NLRP3, AIM2, and NLRC4. These receptors are activated by distinct danger- or pathogen-associated molecular patterns, such as extracellular ATP, double-stranded DNA, and bacterial flagellin (14). Once cleaved, IL1β follows an unconventional secretory pathway that typically requires membrane pores composed of gasdermin D (GSDMD), activation of which is also induced by the inflammasome (5, 14).

Despite emerging interest in IL1β as an oncology target, several questions regarding IL1β signaling in the context of cancer remain unanswered. First, although IL1β secretion is elevated in breast and lung tumors compared with adjacent noninvolved tissues (8, 15), the exact cellular source of increased IL1β production within these tumors remains poorly characterized. In fact, malignant cells, fibroblasts, and immune cells are all considered potential sources of IL1β release in various tumor types (16–19). Second, the inflammasome is dispensable for IL1β release in several types of sterile inflammation (20–22), raising the question whether caspase-1 is critically required for the proteolytic maturation and release of active IL1β in the tumor microenvironment (TME). GSDMD is essential for in vivo IL1β release in several mouse models of inflammation, but it has not yet been determined whether it plays a similar role in tumors (23–26). In addition, we still have limited understanding about how IL1β release affects the complex TME and to what extent its effect is conserved across different tumor types. To address these knowledge gaps, we set out to examine the source of IL1β in tumors, the requirement of inflammasomes and GSDMD for its release, and its impact on the TME in mouse models of non–small cell lung cancer (NSCLC) and triple-negative breast cancer (TNBC).

All experiments were performed with age-matched female mice. C57BL/6 mice were from Janvier, Il1b^−/− mice were provided by François Huaux, UBI-GFP mice were from Jackson Laboratories. Nlrp3^−/−, Nlrc4^−/−, and Gsdmd^−/− mice were provided by Vishva M. Dixit. Ripk3^−/−Casp8^−/−, Casp1/11^−/−, and Mlkl^−/− mice were from Thirumala-Devi Kanneganti (St Jude Children's Research Hospital, USA). Casp1/11^−/−Ripk3^−/−Casp8^−/− mice were generated by crossing the Casp1/11^−/− and Ripk3^−/−Casp8^−/− strains. Gsdmd^−/−Mlkl^−/− mice were generated by crossing the Gsdmd^−/− and Mlkl^−/− strains. Both Casp1/11^−/−Ripk3^−/−Casp8^−/− and Gsdmd^−/−Mlkl^−/− mice were generated in the VIB Center for Inflammation Research, Ghent, Belgium. In all experiments involving knock-out mice, wild-type (+/+) or heterozygote (+/−) littermate mice were used as controls as specified in the figures and their legends.

All procedures followed the guidelines of the Belgian Council for Laboratory Animal Science and were approved by the Ethical Committee for Animal Experiments of the Vrije Universiteit Brussel (licenses 14–220–26, 16–220–3, 19–220–35).

Lewis lung carcinoma (LLC) cells (ATCC, cat# CRL-1642) were purchased in 2017. E0771 cells (CH3 Biosystems, cat# 94A001) were received from Professor Massimiliano Mazzone in 2016. B16F10 cells (ATCC, cat# CRL-6475) were purchased in 2012. EG7 cells were received from professor Karine Breckpot in 2007. LLC, E0771, and B16F10 cells were maintained in DMEM (Gibco, cat# 41965–039) supplemented with 10% (v/v) heat-inactivated FCS (Capricorn Scientific, cat# FBS-12A), 300 μg/mL l-glutamine (Sigma, cat# G8540), 100 units/mL penicillin, and 100 μg/mL streptomycin (Gibco, cat# 15140122). For EG7 cells, the medium was RPMI (Gibco, cat# 52400–025), and the supplements were the same. HEK293-IL1R1 cells (Invivogen, cat# hkb-il1r, purchased in 2019) were cultured in DMEM with 10% (v/v) heat-inactivated FCS, 100 units/mL penicillin, 100 μg/mL streptomycin, 100 μg/mL Normocin (Invivogen, cat# ant-nr-1), and HEK-Blue Selection antibiotics (Invivogen, cat# hb-sel). All cell lines used in experiments were cultured for 7 to 14 days and for 3 to 6 passages. Testing for Mycoplasma infection was performed in 2019, and all cell lines were negative. The cell lines were not authenticated in the past year.

For tumor implantation, 3 × 10⁶ LLC cells, 1 × 10⁶ B16F10 cells, or 3 × 10⁶ EG7 cells were injected s.c. into the right flank of mice in 200 μL of HBSS. For orthotopic breast tumor implantation, 5 × 10⁵ E0771 cells were injected into the left fourth mammary fat pad in 50 μL of HBSS mixed with Growth Factor Reduced Matrigel (Corning, cat# 356230) in a 1:1 ratio.

Tumor volumes were determined by caliper measurements and calculated using the formula: V = π × (d² × D)/6, where d is the shortest diameter and D is the longest diameter.

Anti-VEGFR2 (clone DC101, BioXCell, cat# BE0060) or isotype control antibody (clone HRPN, BioXCell, cat# BE0088) was administered i.p. every 3 days starting from day 4 of tumor growth at a dose of 40 mg/kg body weight.

For macrophage depletion, CSF1R inhibitor PLX5622 was administered via rodent chow (1,200 mg PLX5622/kg chow) starting from day 6 of tumor growth. PLX5622 was provided by Plexxikon. Control and PLX5622-containing AIN-76A rodent chow was prepared by Research Diets.

For CD8⁺ T-cell depletion, 200 μg CD8-specific antibody (clone YTS169, provided by Polpharma Biologics) was injected i.p. every 2 to 3 days starting 1 day prior to tumor inoculation.

For neutrophil depletion, mice received daily i.p. injections of 200 μg CXCR2 inhibitor (SB225002, Selleckchem, cat# S7651) dissolved in saline with 5% DMSO and 8% Tween-80 starting from the day of tumor implantation. In addition, starting from day 12 of tumor progression, mice received 75 μg Ly6G-specific antibody (clone 1A8, BioXCell, cat# BE0075–1) i.p. every second day, followed by 150 μg anti-rat immunoglobulin (clone MAR 18.5, BioXCell, cat# BE0122) i.p. 24 hours later. The anti-Ly6G treatment was only maintained for 6 days due to the development of anti-rat antibodies in treated mice after 1 week, which limits its efficacy (27).

Blood was collected from mice in 1 mL syringes containing 0.5 mol/L EDTA (Duchefa, cat# E0511). Tumors were excised; cut in small pieces; incubated with 10 U/mL collagenase I (Worthington, cat# CLSS-1), 400 U/mL collagenase IV (Worthington, cat# CLSS-4), and 30 U/mL DNase I (Worthington, cat# DCLS) in RPMI for 30 minutes at 37°C; squashed; and filtered. Spleens were mashed through a cell strainer; bone marrow was flushed out from the femurs into RPMI. All single-cell suspensions were treated with Ammonium–Chloride–Potassium erythrocyte lysis buffer.

Single-cell suspensions were resuspended in HBSS, and samples for flow cytometry analysis were incubated with Fixable Viability Dye eFluor 506 (1:1,000, eBioscience, cat# 65–0866–14) for 30 minutes at 4°C. Next, cell suspensions were washed with HBSS and resuspended in HBSS with 2 mmol/L EDTA and 0.5% (v/v) FCS. To prevent nonspecific antibody binding to Fcγ receptors, cells were preincubated with CD16/CD32-specific antibody (clone 2.4G2, BD Biosciences, cat# 553142). Cell suspensions were then incubated with fluorescently labeled antibodies diluted in HBSS with 2 mmol/L EDTA and 0.5% (v/v) FCS for 20 minutes at 4°C and then washed with the same buffer.

For intracellular staining of IFNγ, tumor single-cell suspensions were incubated in ex vivo culture medium (see Ex vivo cell culture below) containing 50 ng/mL PMA, 500 ng/mL ionomycin, and Golgiplug (1:1,000, BD Biosciences, cat# 555029) for 4 hours at 37°C. After washing the samples with HBSS, surface proteins were stained first as described above, followed by fixation and permeabilization using the Cytofix/Cytoperm kit (BD Biosciences, cat# 554714), then staining of intracellular IFNγ.

The following fluorochrome-conjugated antibodies were used in the study: CD45 (clone 30-F11, Biolegend, cat# 103116), CD11b (clone M1/70, Biolegend, cat# 101216 and 101228), Ly6G (clone 1A8, Biolegend, cat# 127616 and 127608), SiglecF (clone E50–2440, BD Biosciences, cat# 552126), MHC-II (clone M5/114.15.2, Biolegend, cat# 107632 and 107626), Ly6C (clone HK1.4, Biolegend, cat# 128010), F4/80 (clone CI:A3–1, Bio-Rad, cat# MCA497A488), CD11c (clone HL3, BD Biosciences, cat# 553802), CD24 (clone M1/69, Biolegend, cat# 101822), NK1.1 (clone PK136, Biolegend, cat# 108728), CD19 (clone 1D3, BD Biosciences, cat# 553786 and clone B4, Biolegend, cat# 115538), TCRβ (clone H57–597, Biolegend, cat# 109212 and 109229), CD4 (clone GK1.5, Biolegend, cat# 100434), CD8 (clone 53–6.7, Biolegend cat# 100712 and 100738), FoxP3 (clone FJK-16s, eBioscience, cat# 45–5773–82), CD44 (clone IM7, Biolegend, cat# 103006), CD62L (clone MEL-14, eBioscience, cat# 12–0621–83), CD69 (clone H1.2F3, Biolegend, cat# 104505), GZMB (clone GB11, Biolegend, cat# 515400), CD31 (clone 390, eBioscience, cat# 46–0311–82), CD64 (clone X54–5/7.1, Biolegend, cat# 139306), CD40 (clone 3/23, Biolegend, cat# 124609), IFNγ (clone XMG1.2, BD Biosciences, cat# 554412), and IL1R1 (clone 35F5, BD Biosciences, cat# 557489).

For neutrophil cell death analysis, tumor and spleen cell suspensions were resuspended in HBSS with 2 mmol/L EDTA and 0.5% (v/v) FCS, stained for surface proteins, washed with the same buffer, and resuspended in Annexin V–binding buffer (Biolegend, cat# 422201) containing Pacific Blue Annexin V (1:20, Biolegend, cat# 640918). After 15-minute incubation at room temperature in the dark, cells were washed with Annexin V–binding buffer and resuspended in the same buffer. 7-Amino-actinomycin D (7-AAD; Biolegend, cat# 420404) was added 10 minutes before analysis.

Flow cytometry data were acquired using a BD FACSCanto II (BD Biosciences) and analyzed using FlowJo software. Samples with less than 10% viable cells and tumor samples with cell contamination from the tumor-draining lymph node (identified as outliers in B-cell and naïve T-cell abundance) were excluded from further analyses.

Surface protein expression was quantified by calculating the Δ median fluorescence intensity (MFI) value: ΔMFI = MFI surface protein –MFI isotype control.

For fluorescence-activated cell sorting of myeloid-cell populations, tumor single-cell suspensions were enriched for CD11b⁺ cells using magnetic cell separation according to the manufacturer's protocol (Miltenyi, cat# 130–049–601). 7-AAD staining was used to exclude dead cells. Cell subsets were then sorted into RPMI with 10% (v/v) FCS, 300 μg/mL l-glutamine, 100 units/mL penicillin, 100 μg/mL streptomycin, 1% (v/v) MEM nonessential amino acids (Gibco, cat# 11140050), 1 mmol/L sodium pyruvate (Gibco, cat# 11360070), and 0.02 mmol/L 2-mercaptoethanol (Sigma, cat# M6250). Fluorescence-activated cell sorting was performed using a BD FACSAria II (BD Biosciences).

The viability of cells after cell sorting was confirmed using trypan blue staining. For ex vivo culture, 3 × 10⁵ cells/well were cultured at 37°C for 24 hours in flat-bottom 96-well plates in 200 μL/well RPMI containing 10% (v/v) FCS, 300 μg/mL l-glutamine, 100 units/mL penicillin, 100 μg/mL streptomycin, 1% (v/v) MEM nonessential amino acids, 1 mmol/L sodium pyruvate, and 0.02 mmol/L 2-mercaptoethanol.

Neutrophils from the spleen of LLC tumor–bearing UBI-GFP mice were isolated by magnetic cell separation using anti-Ly6G microbeads according to the manufacturer's protocol (Miltenyi). 5 × 10⁶ GFP-expressing neutrophils in 100 μL HBSS were injected through the tail vein into recipient LLC tumor–bearing mice, which were sacrificed 24 hours later.

2 × 10⁵ neutrophils or monocytes sorted from tumors were added to 2 × 10⁵ naïve C57BL/6 splenocytes stimulated with anti-CD3 (1 μg/mL, clone 145–2C11, BD Biosciences, cat# 550275) and anti-CD28 (2 μg/mL, clone 37.51, eBioscience, cat# 16–0281–85) and cultured in flat-bottom 96-well plates in culture medium for ex vivo cell culture described above. After 24 hours of culture at 37°C, 1 μCi (0.037 MBq) ³H-thymidine (PerkinElmer, cat# NET027A005MC) was added, and after another 18 hours of culture, T-cell proliferation was measured as count per minute in a liquid scintillation counter.

For measuring T-cell proliferation via flow cytometry, splenocytes were labeled with CellTrace Violet dye (Thermo Fisher, cat# C34571) following the manufacturer's instructions and cocultured with tumor-derived neutrophils as described above. To inhibit potential T-cell–suppressive pathways, 0.5 mmol/L Nω-Nitro-l-arginine methyl ester (L-NAME, Sigma, cat# N5751), 0.5 mmol/L Nω-Hydroxy-nor-l-arginine (Nor-NOHA, Sigma, cat# 399275), or 200 U/mL superoxide dismutase (Sigma, cat# S5395) was added to the cocultures. After 42 hours of culture at 37°C, the frequency of CellTrace^low proliferating T cells was determined via flow cytometry.

Tumor tissue was snap-frozen in liquid nitrogen and homogenized in 1 mL TRIzol (Invitrogen, cat# 15596026) in gentleMACS M tubes [Miltenyi, cat# 130–093–236 using the gentleMACS Dissociator (Miltenyi, cat# 130–093–235)]. RNA was extracted using TRIzol, and 1 μg RNA was reverse-transcribed with oligo(dT) and SuperScript II RT (Invitrogen, cat# 18064022) following the manufacturers' protocols. Quantitative real-time PCR was performed in triplicates for each sample in the CFX Connect Real-Time System (Bio-Rad) using the SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, cat# 1725274) and the following primers: Rps12-F: GGAAGGCATAGCTGCTGGAGGTGT, Rps12-R: CCTCGATGACATCCTTGGCCTGAG; Il1b-F: GTGTGGATCCCAAGCAATAC, Il1b-R: GTCTGCTCATTCACGAAAAG; Cxcl1-F: GCTTGAAGGTGTTGCCCTCAG, Cxcl1-R: AAGCCTCGCGACCATTCTTG; Cxcl2-F: TGGAAGGAGTGTGCATGTTC, Cxcl2-R: CACGAAAAGGCATGACAAAA; Cxcl3-F: CACCCAGACAGAAGTCATAGCCAC, Cxcl3-R: TGGTGAGGGGCTTCCTCCTTT; Cxcl5-F: CTCGCCATTCATGCGGAT, Cxcl5-R: CTTCAGCTAGATGCTGCGGC; Cxcl7-F: CTCAGACCTACATCGTCCTGC, Cxcl7-R: GTGGCTATCACTTCCACATCAG; Cxcl12-F: TCATCCCCATTCTCCTCATC, Cxcl12-R: ATAAAGGAGCCTCCCTCTGC; Cxcl9-F: CCTCCTTGCTTGCTTACCAC, Cxcl9-R: TTTTCACCCTGTCTGGCTCT; Cxcl10-F: AATTGCCCTTGGTCTTCTGA, Cxcl10-R: CCTTGGGAAGATGGTGGTTA; Cxcl16-F: GTCTCCTGCCTCCACTTTCT, Cxcl16-R: CTAAGGGCAGAGGGGCTATT; Ccl5-F: GTGCCCACGTCAAGGAGTAT, Ccl5-R: CGAGTGGGAGTAGGGGATTA; Il12b-F: TCAGGGACATCATCAAACCA, Il12b-R: CTACGAGGAACGCACCTTTC; Cd40-F: GCTGTGAGGATAAGAACTTGGAGG, Cd40-R: GCATCCGGGACTTTAAACCACA; Cd86-F: CCTCCAAACCTCTCAATTTCA, Cd86-R: TCGGCTTCTTGTGACATACAAT. The following program was used for real-time PCR: 95°C 3 minutes, 40 × (94°C 30 seconds, 54°C 30 seconds, 72°C 45 seconds). Expression values were calculated using the ΔCt method as follows: 2^−(CtA-CtB), where CtA is the Ct value of the gene of interest and CtB is the Ct value of the house-keeping gene Rps12.

IL1β, IFNγ, and CXCL9 were measured using ELISA (IL1β from cell culture supernatants: R&D Systems, cat# MLB00C; IL1β from serum: R&D Systems, cat# MHSLB00; IFNγ: R&D Systems, cat# DY485; CXCL9: R&D Systems, cat# DY492). ELISA measurements were obtained using a VersaMax microplate reader (Molecular Devices) set to 450 nm. Measurements at 540 nm were used for background correction. G-CSF, CXCL10, CXCL16, and CCL5 were measured using multiplex immunoassay (Bio-Rad, cat# 171G5015M and 12009159) according to the manufacturers' protocols.

Nitrite levels were determined by adding 50 μL Griess reagent (5% phosphoric acid containing 0.2% naphthylethylenediamine dihydrochloride and 2% sulfanilamide) to 50 μL culture supernatant in a 96-well plate and measuring the optical density at 548 nm using a VersaMax microplate reader.

Cell lysates for immunoblots were prepared by resuspending cells in a lysis buffer containing 20 mmol/L, pH 7.4, Tris HCl (Sigma, cat# T3253), 200 mmol/L NaCl, and 1% (v/v) NP-40 (Sigma, cat# I8896). Cell lysates and cell culture supernatants were denatured in 4× Laemmli buffer [250 mmol/L Tris HCl, 8% (w/v) SDS (Sigma, cat# L3771), 40% (v/v) Glycerol (Sigma, cat# G5516), 0.02% (w/v) Bromophenol blue, and 5% (v/v) 2-mercaptoethanol] at 95°C for 10 minutes. Proteins were separated using SDS-PAGE gels [for IL1β: 16% gel (Invitrogen, cat# P00160BOX) 200 V; for caspase-1/8: 12% gel (Invitrogen, cat# P00120BOX) 164 V] and were transferred to PVDF membranes (Bio-Rad, cat# 1704273). Blocking, incubation with antibody, and washing of the membrane were done in PBS supplemented with 0.05% (v/v) Tween 20 and 3% (v/v) nonfat dry milk. Immunoblots were incubated overnight with primary antibodies against caspase-1 (Adipogen, cat# AG-20B-0042-C100), IL1β (Genetex, cat# GTX74034), caspase-8 (full length: Enzo Life Sciences, cat# ALX-804–447-C100; cleaved: Cell Signaling Technology, cat# 8592S), and β-actin (Santa Cruz Biotechnology, cat# sc-47778-HRP). Horseradish peroxidase–conjugated goat anti-mouse (Jackson ImmunoResearch Laboratories, cat# 115–035–146), anti-rabbit (Jackson ImmunoResearch Laboratories, cat# 111–035–144), or anti-rat (Jackson ImmunoResearch Laboratories, cat# 112–035–143) secondary antibody was used to detect proteins by enhanced chemiluminescence (Thermo Scientific, cat# 34578). Mouse bone marrow–derived macrophages (BMDM) treated with 0.5 μg/mL LPS (Invivogen, cat# tlrl-smlps) for 3 hours followed by 5 mmol/L ATP (Roche, cat# 10519987001) for 45 minutes were used as positive controls for IL1β blots, and Casp1/11^−/− BMDMs treated with 1 μg/mL anthrax protective antigen (Quadratech, cat# 171E) and 500 ng/mL anthrax lethal factor (Quadratech, cat# 172B) for 2 hours were used as positive controls for caspase-8 blots. BMDMs were generated by culturing bone marrow cells in IMDM (Lonza, cat# 12–722F) containing 10% (v/v) FCS, 30% (v/v) L929 cell-conditioned medium, 1% (v/v) MEM nonessential amino acids (Lonza, cat# BE13–114E), 100 U/mL penicillin, and 100 μg/mL streptomycin at 37°C in a humidified incubator containing 5% CO₂ for 6 days.

For the assessment of tumor blood vessel perfusion, mice were injected i.v. with 0.05 mg FITC-conjugated lectin (Vector Laboratories, cat#/FL-1171). After 10 minutes, mice were sacrificed, and tumors were harvested.

Tumor hypoxia was detected via i.p. injection of 60 mg/kg body weight pimonidazole hydrochloride (Hypoxyprobe, cat# HP3–100Kit) into tumor-bearing mice. After 1 hour, mice were sacrificed, and tumors were harvested.

Tumor samples were fixed in 2% PFA overnight at 4°C, dehydrated, and then embedded in paraffin. Serial sections of 7 μm thickness were made. Slides were first rehydrated to further proceed with antigen retrieval in citrate solution (DAKO, cat# S1699) at 100°C for 20 minutes. Slides were then incubated in 0.3% hydrogen peroxide in methanol for 20 minutes to block endogenous peroxidases. The sections were blocked with donkey serum (Sigma, cat# D9663) for 45 minutes and incubated overnight at room temperature with the following antibodies: anti-CD31 (BD Biosciences, cat# 550274), anti-FITC (Serotec, cat# 4510-7604), anti–αSMA-Cy3 (Sigma, cat# C6198), and anti-pimonidazole (Hypoxyprobe, cat# HP3-100Kit). Next, appropriate secondary Alexa Fluor 488/647-conjugated antibodies (Invitrogen, cat# A-21206, A-21110, A-31573) or biotin-labeled antibodies (Jackson ImmunoResearch Laboratories, cat# 712-065-153) were applied. After biotin-labeled antibodies, TSA Cyanine 3 or Cyanine 5 amplification kits (PerkinElmer, cat# NEL704A001KT and NEL705A001KT) were used according to the manufacturer's instructions. Hoechst solution was used to stain nuclei. Mounting of slides was done with ProLong Gold mounting medium without DAPI (Invitrogen, cat# P36931). Imaging and microscopic analysis was performed with an Olympus BX41 microscope and CellSense imaging software. Slides were scanned using Zeiss AxioScan Z.1 slide scanner. CD31⁺ blood vessel density and the proportion of FITC-lectin⁺ (perfused) and αSMA⁺ (pericyte-covered) blood vessels were determined by manual counting in six representative microscopic images/tumor. The proportion of pimonidazole⁺ hypoxic areas was determined in whole tumor cross-sections using ImageJ.

The droplet-based single-cell RNA sequencing (scRNA-seq) data of 8 untreated patients with lung cancer (ref. 28; 10× Genomics 3′ RNA library kit, ArrayExpress:E-MTAB-6149 and E-MTAB-6653) were processed and clustered using Seurat (v2.3.4) package. Cell matrix was filtered (nUMI > 400, 200 < nGene < 6,000, mitochondrial RNA < 25%), normalized, regressed for confounding factors (nUMI, patient, mitochondrial RNA, and cell cycle), and scaled. The variable genes (normalized expression between 0.125 and 3, quantile-normalized variance > 0.5) were used to construct principal components, followed by graph-based clustering (tSNE and Louvain algorithm). Cell type annotation was based on the expression of established marker genes. Plasmacytoid dendritic cells (pDC) were initially coclustered with B cells, and then annotated back to myeloid population, where most other dendritic cells (DC) were coclustered with. Then the myeloid cells were subclustered to identify monocytes (SELL, CDKN1C, and MTSS1), macrophages (CD68, CD163, and MCR1), DCs (CLEC9A, XCR1, CD1C, CD1A, and LILRA4), and neutrophils (FCGR3B). Similar analysis was performed for 5′-scRNA-seq data from 14 treatment-naïve breast cancers (ref. 29; ArrayExpress: E-MTAB-8107), and the myeloid cells were further subclustered and annotated.

Statistical analyses were performed in GraphPad Prism software. For relevant pairwise comparisons, unpaired two-tailed t test was used to calculate the P value. Tumor growth curves were compared by two-way ANOVA with Holm–Sidak multiple comparisons test. To assess correlation, Pearson correlation coefficient was calculated. A P value < 0.05 was considered statistically significant. For statistically significant differences, the P value is indicated in graphs as the following: *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001. Comparisons found to be nonsignificant are not shown.

To determine the cellular sources of IL1β in lung and breast tumors with an unbiased approach, we analyzed single-cell RNA-seq datasets from human NSCLC and breast cancer. Unsupervised clustering of the data followed by identification of known cell lineages based on marker gene expression revealed 13 and 10 major cell types in lung and breast tumors, respectively (Fig. 1A; Supplementary Fig. S1A). We found that the cell populations with the highest average expression levels of IL1B in both tumor types were myeloid cells—neutrophils, monocytes, DCs, and macrophages—whereas other cell populations showed considerably (>10-fold) lower or no expression (Fig. 1B and C). Of note, only a small number of neutrophils could be detected in these datasets, presumably due to their low transcript counts, whereas these cells are known to be well represented in both tumor types based on flow cytometry (30, 31).

To investigate IL1β production in more detail, we turned to mouse models of NSCLC and breast cancer: Subcutaneous Lewis lung carcinoma (LLC), a p53-mutant lung adenocarcinoma, and orthotopic E0771, a p53-mutant TNBC model with basal-like characteristics (32, 33). Mice with LLC or E0771 tumors showed significantly elevated IL1β levels in the serum compared with naïve mice, indicating the presence of tumor-induced IL1β-driven inflammation in these models (Fig. 1D). To assess the contribution of myeloid cells to intratumoral IL1β release, we separated the CD11b⁺ and CD11b⁻ fractions of tumors and measured the expression of Il1b mRNA in freshly isolated cells as well as the secretion of the cytokine following 24 hours of in vitro culture. We found that both mRNA expression and protein secretion of IL1β were almost exclusively restricted to the CD11b⁺ fraction of tumors (Fig. 1E and F). Importantly, the LLC and E0771 cell lines did not produce bioactive IL1β (Supplementary Fig. S1B and S1C). The majority of the CD11b⁺ fraction in LLC and E0771 tumors consisted of neutrophils, monocytes, and tumor-associated macrophages (TAM; Fig. 1G; for gating strategies, see Supplementary Fig. S2). Consistent with published reports, the TAM population included MHC-II^high and MHC-II^low subsets, which possess immunostimulatory and anti-inflammatory gene expression profiles, respectively (34, 35). We then isolated these cell populations from tumors and assessed their IL1β secretion in vitro. In LLC tumors, monocytes and neutrophils showed the highest secretion levels, followed by MHC-II^high TAMs and MHC-II^low TAMs (Fig. 1H). In contrast, IL1β secretion was comparable across the different myeloid cell types isolated from E0771 tumors (Fig. 1H).

Altogether, these results demonstrate that myeloid cells are the primary source of IL1β in human and mouse lung and breast tumors.

To investigate the impact of IL1β release on tumor progression, we implanted LLC or E0771 tumors in IL1β-deficient (Il1b^−/−) mice and their wild-type (Il1b^+/+) littermates. Loss of IL1β delayed tumor growth in both tumor models with a more pronounced effect in E0771 breast tumors, where IL1β-deficiency was, in some cases, associated with regression or durable tumor control (Fig. 2A; Supplementary Fig. S3A). Notably, subcutaneous implantation of E0771 tumors in Il1b^−/− mice did not result in significant growth inhibition, suggesting that the effect of IL1β may depend on the tissue microenvironment (Supplementary Fig. S3B).

IL1β release induces neutrophilia during systemic inflammation, and this may have an influence on tumor progression due to the wide range of tumor-promoting activities linked to neutrophils (36). For this reason, we analyzed the frequency of circulating CD11b⁺Ly6G⁺ neutrophils in naïve and tumor-bearing Il1b^+/+ and Il1b^−/− mice. Both LLC and E0771 tumors induced expansion of circulating neutrophils, and this was abrogated in the absence of IL1β (Fig. 2B). These changes were mirrored by G-CSF levels in the blood, suggesting that tumor-induced, IL1β-dependent systemic neutrophil expansion is driven by G-CSF (Fig. 2C). Loss of IL1β prevented the LLC-induced expansion of bone marrow neutrophils, whereas tumor-induced accumulation of splenic neutrophils was prevented by IL1β-deletion in both tumor models (Fig. 2D). Next, we assessed whether IL1β deficiency has an influence on neutrophils infiltrating primary tumors. We found that loss of IL1β strongly reduced the abundance of neutrophils in both LLC and E0771 tumors (Fig. 2E). The reduced abundance of neutrophils observed in tumor-bearing IL1β-deficient mice was likely not due to increased neutrophil cell death, as the proportions of necrotic/apoptotic neutrophils were not elevated in the spleens and tumors of Il1b^−/− mice compared with wild-type controls (Supplementary Fig. S3C and S3D). To test whether the reduced abundance of tumor-infiltrating neutrophils is solely due to their decreased levels in the circulation or also due to altered recruitment, we adoptively transferred equal numbers of GFP-expressing splenic neutrophils into LLC tumor–bearing Il1b^+/+ and Il1b^−/− mice and assessed their frequency in the tumor after 24 hours. As shown in Fig. 2F, recruitment of GFP⁺ neutrophils to the tumor was strongly reduced in Il1b^−/− mice, even though their frequency in the circulation remained comparable with wild-type controls. Because CXCR2 ligands, particularly CXCL1 and CXCL2, have been shown to be critical for neutrophil extravasation (37), we investigated whether these chemokines are affected by IL1β release in the tumor. We found that all CXCR2 ligands, including Cxcl1, Cxcl2, Cxcl3, Cxcl5, and Cxcl7, but not the CXCR4 ligand Cxcl12, showed strongly reduced expression in the absence of IL1β (Fig. 2G).

Of note, the effect of IL1β on the expansion and recruitment of neutrophils was not restricted to the LLC and E0771 tumor models. We also observed a significant reduction of circulating and tumor-infiltrating neutrophils in IL1β-deficient mice with EG7 lymphoma and B16-F10 melanoma tumors, which show greatly differing levels of neutrophil abundance (Supplementary Fig. S3E and S3F).

Neutrophil recruitment to the tumor has been shown to drive therapy resistance and immunosuppression during treatment with antiangiogenic agents targeting VEGF signaling (38, 39). Hence, we analyzed whether IL1β is required for neutrophil infiltration during antiangiogenic therapy and examined the effect of VEGFR2-specifc antibody treatment in Il1b^+/+ and Il1b^−/− mice in the LLC tumor model. Although anti-VEGFR2 treatment did not affect the levels of circulating neutrophils (Supplementary Fig. S3G), we observed a 103% increase in the abundance of tumor-infiltrating neutrophils in treated wild-type mice, and this therapy-induced neutrophil recruitment was completely abrogated in IL1β-deficient animals (Fig. 2H). This was associated with a significantly reduced tumor burden in anti–VEGFR2-treated Il1b^−/− mice compared with Il1b^+/+ mice with the same treatment (Fig. 2I).

Collectively, these results indicate that loss of IL1β delays tumor progression in mouse models of NSCLC and TNBC, and this is accompanied by reduced systemic expansion and tumor infiltration of neutrophils. In addition, IL1β deletion prevents accumulation of neutrophils in the tumor triggered by antiangiogenic therapy.

Next, we determined whether or not the delayed tumor progression and strong reduction of neutrophil recruitment to tumors in Il1b^−/− mice can be recapitulated in mice lacking various inflammasome components, which would indicate their requirement for bioactive IL1β production in tumors. Deficiency of NLRP3 and NLRC4, two caspase-1–activating NOD-like receptors, did not affect in vitro IL1β release of tumor-derived myeloid cells (Fig. 3A; Supplementary Fig. S4A). Consistent with this, deletion of these inflammasome components did not alter tumor progression or neutrophil recruitment in mice with LLC and E0771 tumors as opposed to IL1β deficiency (Fig. 3B and C; Supplementary Fig. S4B and S4C). To more directly assess the potential role of canonical and noncanonical inflammasome pathways, we analyzed tumors in mice with combined deletion of inflammatory caspases 1 and 11 (Casp1/11^−/−). Deletion of caspase-1/11 led to a partial reduction in IL1β secretion levels by LLC tumor–derived myeloid cells (Fig. 3D). However, this was not sufficient to alter tumor progression or neutrophil recruitment in LLC tumors (Fig. 3E and F). IL1β release by E0771 tumor–derived myeloid cells was not reduced in Casp1/11^−/− mice and, correspondingly, tumor growth and neutrophil infiltration remained unaltered in these tumors (Fig. 3D–F). An IL1β immunoblot on the culture supernatants of tumor-derived Casp1/11^−/− myeloid cells confirmed the inflammasome-independent production of mature IL1β in both tumor models (Fig. 3G).

Caspase-8 has shown redundancy with caspase-1 in producing active IL1β in some cases, cleaving pro-IL1β at the same site (40, 41). Active caspase-8 could be detected by immunoblot in sorted tumor-infiltrating myeloid cells but not in their circulating precursors (Supplementary Fig. S4D). Hence, we assessed the contribution of caspase-8 to IL1β release and neutrophil recruitment in tumors by using Ripk3^−/−Casp8^−/− mice, in which Ripk3 deletion rescues embryonic lethality caused by caspase-8 deficiency (42). We also generated Casp1/11^−/−Ripk3^−/−Casp8^−/− mice to evaluate the potential redundant roles of caspase-1/11 and -8. Caspase-8 deletion in both the Ripk3^−/− and Casp1/11^−/−Ripk3^−/− backgrounds led to partial blockade of in vitro IL1β release in myeloid cells derived from LLC tumors but not from E0771 tumors (Supplementary Fig. S4E and S4H). However, this was not sufficient to affect tumor progression and neutrophil recruitment (Supplementary Fig. S4F, S4G, S4I, and S4J). Together, these data suggest slightly different mechanisms of IL1β production by myeloid cells in LLC and E0771 tumors, but an overall independence of tumor growth and neutrophil recruitment from inflammasomes and caspase-8.

Membrane pore formation by GSDMD is critical for IL1β release in mouse models of autoinflammation, steatohepatitis, disseminated intravascular coagulation, and sepsis (23–26). To investigate a potential role for GSDMD in IL1β release, LLC and E0771 tumors were implanted in Gsdmd^−/− mice and Gsdmd^+/+ littermates. However, GSDMD deficiency did not reduce IL1β release of tumor-derived CD11b⁺ myeloid cells, tumor growth, and neutrophil recruitment (Fig. 3H–J). Alternatively, necroptosis induced by membrane pores composed of mixed lineage kinase domain–like protein (MLKL) has been suggested to mediate IL1β release independently of GSDMD-dependent pyroptosis in vitro (43). However, neither MLKL deficiency nor GSDMD/MLKL double deficiency had a significant effect on myeloid cell IL1β release, tumor progression, and neutrophil recruitment (Supplementary Fig. S5A–S5F).

Overall, these data from two distinct mouse tumor models demonstrate that activation of the inflammasome and caspase-8 as well as the formation of membrane pores by GSDMD and MLKL are dispensable for the release of bioactive IL1β by tumor-associated myeloid cells and consequential neutrophil recruitment.

To investigate the mechanism of reduced tumor growth in Il1b^−/− mice, we studied tumor angiogenesis, which has been described as being potentially IL1β-regulated (10, 11). However, we did not find any major differences in tumor blood vessel density, pericyte coverage, and vessel perfusion between Il1b^−/− and Il1b^+/+ mice (Supplementary Fig. S6A–S6G). Slightly less hypoxic areas were observed in the tumors of Il1b^−/− mice; however, this was likely due to the smaller average tumor size because volume-matched tumors did not show such difference (Supplementary Fig. S6H). These observations suggested that IL1β is not essential for tumor angiogenesis and reduced tumor growth in IL1β-deficient animals may be explained by immune-mediated mechanisms.

Next, we examined whether neutrophils recruited by IL1β to LLC or E0771 tumors were able to suppress T-cell proliferation. CD11b⁺Ly6G⁺ neutrophils isolated from primary tumors inhibited proliferation of splenocytes stimulated with anti-CD3 and anti-CD28 (Fig. 4A). Tumor-infiltrating neutrophils from Il1b^−/− mice showed similar T-cell–suppressive activity to neutrophils from Il1b^+/+ controls, suggesting that IL1β dominantly affects their recruitment rather than their immunosuppressive activity (Fig. 4B).

To identify the mechanism by which neutrophils recruited to the tumor are able to suppress T-cell activation, we inhibited three key T-cell–suppressive mechanisms described previously: nitric oxide (NO) synthesis, arginase activity, and superoxide production (44). Neutrophils isolated from LLC tumors were cocultured with activated splenocytes in the presence of L-NAME (NO synthase inhibitor), Nor-NOHA (arginase inhibitor), or superoxide dismutase. Only L-NAME restored T-cell proliferation and IFNγ production in the presence of tumor-derived neutrophils (Fig. 4C and D). Measurement of nitrite, a stable breakdown product of NO, in the supernatants of cocultures confirmed NO production by tumor-derived neutrophils and its reduction upon inhibition of NO synthase (Fig. 4E).

Overall, these results indicate that neutrophils recruited to the tumor can suppress T-cell proliferation and activation via NO production.

Consistent with the immunosuppressive phenotype of tumor-infiltrating neutrophils, impaired neutrophil recruitment in Il1b^−/− mice (Fig. 2E) was accompanied by an elevated abundance of cytotoxic CD8⁺ T cells, whereas the infiltration of CD4⁺ T cells and FoxP3⁺ regulatory T cells remained unaltered (Fig. 5A–C). In addition, a higher proportion of tumor-infiltrating CD4⁺ and CD8⁺ T cells showed an effector T-cell phenotype in Il1b^−/− mice, as indicated by the increased CD44⁺CD62L⁻ effector versus CD44⁻CD62L⁺ naïve T-cell ratio (Fig. 5D and E; Supplementary Fig. S7A). Upregulation of the activation marker CD44 and expansion of IFNγ⁺ cells among CD8⁺ T cells further indicated an enhanced cytotoxic T-cell response in the absence of IL1β (Fig. 5F). Correspondingly, enhanced infiltration of neutrophils upon anti-VEGFR2 therapy in wild-type mice (Fig. 2H) was associated with impaired effector T-cell differentiation, and this was counteracted by IL1β deletion (Fig. 5G; Supplementary Fig. S7B and S7C). Among the immune cells possessing T-cell stimulatory potential, conventional DCs did not show altered infiltration in tumors of Il1b^−/− mice (Supplementary Fig. S7D). In LLC tumors of IL1β-deficient mice, we observed reduced infiltration of monocytes that possessed potent T-cell–suppressive capacity (Fig. 5H and I). However, TAM abundance was not reduced in LLC and E0771 tumors (Fig. 5J). Among TAMs, we found higher abundance of MHC-II^high TAMs (Fig. 5K), a phenotype that has been shown to be driven by effector T cells (45, 46), and, in turn, possesses the capacity to stimulate T-cell responses (34, 35).

These changes in the TME were likely indirect effects of IL1β deficiency rather than the lack of direct IL1β effects on tumor-infiltrating immune cells, as we could not detect IL1R1 expression on T cells, monocytes, macrophages, and neutrophils in either model (Supplementary Fig. S8).

Reducing the abundance of immunosuppressive neutrophils in the TME administering both a CXCR2 inhibitor and a Ly6G-specific antibody did not fully mirror the changes observed in Il1b^−/− mice (Supplementary Fig. S9). Similar to IL1β deficiency, neutrophil depletion led to a reduction in LLC tumor growth and upregulation of the activation marker CD44 on CD8⁺ T cells (Supplementary Fig. S9A–S9C). In contrast to IL1β deletion, however, neutrophil depletion caused an 89% increase in immunosuppressive monocyte infiltration into tumors and did not significantly enhance the abundance of tumor-infiltrating effector CD8⁺ T cells and immunostimulatory TAMs (Supplementary Fig. S9D–S9H).

In summary, these data indicate that IL1β promotes the establishment of an immunosuppressive TME characterized by impaired accumulation of effector T cells and suppression of TAM activation. This is partly driven by IL1β-dependent recruitment of immunosuppressive neutrophils, but likely requires additional, yet unknown, IL1β-activated immunosuppressive pathways in the TME.

Next, we set out to determine whether the increased abundance and activation state of tumor-infiltrating cytotoxic CD8⁺ T cells is responsible for the inhibition of tumor progression in IL1β-deficient mice. Systemic depletion of CD8⁺ T cells using a CD8-specific antibody fully restored LLC tumor growth in Il1b^−/− mice to wild-type levels, whereas it had no effect in wild-type mice (Fig. 6A and B). In addition, CD8⁺ T-cell depletion reduced the activation of tumor-infiltrating CD4⁺ T cells in Il1b^−/− mice, indicated by the decreased frequency of cells showing the effector phenotype and IFNγ production (Fig. 6C and D; Supplementary Fig. S10A). Analogously, acquisition of an immunostimulatory MHC-II^high phenotype by TAMs in Il1b^−/− mice required the presence of CD8⁺ T cells (Fig. 6E). Furthermore, elevated expression of CD40 on MHC-II^high TAMs in Il1b^−/− mice was reduced to wild-type levels upon CD8⁺ T-cell depletion (Fig. 6F). Consistent with these observations, presence of MHC-II^high TAMs but not MHC-II^low TAMs showed a strong positive correlation with effector CD8⁺ T-cell infiltration in both LLC and E0771 tumors (Fig. 6G; Supplementary Fig. S10B).

Overall, these results show that CD8⁺ cytotoxic T cells are required for the inhibition of tumor progression and are key drivers of CD4⁺ T-cell and TAM activation in IL1β-deficient mice.

Last, we wanted to assess whether macrophages participate in amplifying the antitumor T-cell response in the absence of IL1β. To test this hypothesis, we depleted macrophages in tumor-bearing mice using the CSF1R inhibitor PLX5622. This small-molecule inhibitor is highly specific for CSF1R and has been successfully used before to deplete TAMs (47). Administration of PLX5622 reduced TAM infiltration in LLC tumors by 89%, whereas it only caused a 28% reduction in E0771 tumors (Supplementary Fig. S11A). Based on these results, we examined the impact of macrophage depletion in the LLC tumor model. PLX5622 treatment in Il1b^−/− mice restored tumor growth to wild-type levels, whereas it did not have an effect in wild-type mice (Fig. 7A). Analysis of the tumor immune cell composition confirmed that CSF1R inhibition efficiently eliminated TAMs in both Il1b^+/+ and Il1b^−/− mice, whereas it did not deplete neutrophils, monocytes, and DCs in tumors (Fig. 7B and C; Supplementary Fig. S11B). Analysis of tumor-infiltrating T cells revealed that macrophage depletion in Il1b^−/− mice reduced CD8⁺ T-cell abundance to the wild-type level (Fig. 7D). Similarly, macrophage depletion in Il1b^−/− mice restored the ratio of tumor-infiltrating effector versus naïve CD8⁺ T cells to levels similar to Il1b^+/+ mice (Fig. 7E; Supplementary Fig. S11C). In addition, the proportions of CD8⁺ T cells expressing the activation markers CD69 and granzyme B were reduced by macrophage depletion in both Il1b^+/+ and Il1b^−/− mice (Fig. 7F and G). Consistent with these results, TAM depletion led to lower intratumoral expression of chemokines commonly associated with T-cell trafficking, including CXCL9, CXCL10, CXCL16, and CCL5 (Fig. 7H; ref. 48). Depletion of TAMs in Il1b^−/− mice also reduced the intratumoral expression of the costimulatory molecules Cd40 and Cd86 as well as the Th1 stimulatory cytokine Il12b, further supporting that TAMs in LLC tumors are an important source of T-cell stimulatory signals (Fig. 7I).

Together, these results indicate that macrophages support the accumulation and activation of CD8⁺ T cells in tumors and play a critical role in tumor control in the absence of IL1β.

In this study, we demonstrate in two distinct mouse models that IL1β, released mainly by neutrophils, monocytes, and macrophages, plays a key role in systemic neutrophil expansion during tumor progression and in promoting neutrophil infiltration into tumors. Our results are consistent with previous reports demonstrating a role for IL1β in systemic neutrophil expansion in breast cancer; however, these studies primarily focused on its consequences on the metastatic environment and not the primary tumor (13, 17). Earlier studies utilizing IL1β blockade or IL1β-overexpressing cancer cells have demonstrated that this cytokine promotes infiltration of myeloid cells into tumors, but the exact identity of these cells remained ill-defined (10, 49, 50). More recently, IL1β was reported to induce CCL2 and promote the recruitment of monocytes and subsequent accumulation of macrophages in the 4T1 mouse model of breast carcinoma (51). We observed a similar impairment of monocyte recruitment in the absence of IL1β in LLC tumors but not in E0771 breast tumors, suggesting that the link between IL1β release and monocyte recruitment may not be a general phenomenon.

We also show in this study that the canonical and noncanonical inflammasomes are dispensable for the production of bioactive IL1β in LLC and E0771 tumors, and combined deletion of caspase-1/11 was not sufficient to recapitulate the in vivo phenotype observed in Il1b^−/− mice. Although caspase-8 was activated in tumor-infiltrating myeloid cells, deletion of this enzyme was not sufficient to completely block bioactive IL1β release and neutrophil infiltration. A diverse range of additional enzymes have been shown to cleave pro-IL1β, including proteinase 3, neutrophil elastase, cathepsin G, granzyme A, chymase, matrix metalloproteinases, and meprins (41, 52, 53). Several of these enzymes may be active and play redundant roles in the TME; therefore, it might not be possible to pinpoint a single enzyme that is responsible for IL1β production in tumors. Because several reports have demonstrated the beneficial effect of genetic or pharmacologic inhibition of NLRP3 and caspase-1 on tumor progression in mice, it is likely that requirement of the inflammasome for IL1β release depends on the availability of alternative cleavage pathways determined by the immune microenvironment (5). Alternatively, decreased tumor progression observed in some of these studies may also be explained by IL1β-independent effects such as inhibition of inflammasome-mediated IL18 release (14).

To our knowledge, the contribution of GSDMD and MLKL to IL1β release in tumors had not been evaluated before. In LLC and E0771 tumors, these pore-forming proteins were not required for IL1β release and neutrophil recruitment. This suggests the existence of alternative release mechanisms that may include passive release through myeloid cell necrosis, which has been linked to IL1β release in vitro and is likely to occur in the TME (54).

Inflammasome-independent IL1β release promoted tumor progression and immunosuppression in both LLC and E0771 tumor models. Interestingly, tumor control upon IL1β deletion was much more pronounced in the E0771 model compared with LLC. The mechanisms underlying the different tumor growth inhibition in the two models remain to be identified and may relate to differences in implantation sites and immunogenicity.

This study adds further support to previous observations linking IL1β release to immune suppression in the TME (19, 51, 55). Our results provide novel insights primarily about the impact of IL1β release on the phenotype of tumor-infiltrating T cells and macrophages in two distinct mouse models. Importantly, our results indicate that the immunosuppressive activity of IL1β is only partly mediated by neutrophil recruitment and likely requires additional IL1β-activated immunosuppressive pathways. IL1β-dependent recruitment of immunosuppressive monocytes observed in LLC tumors is possibly one of these. Nevertheless, IL1β may have immunosuppressive effects on additional cell types in the TME and/or systemically that were not examined in the current study, such as cancer cells, endothelial cells, and fibroblasts. Further research is needed to better understand how these cells are affected by IL1β release in lung and breast cancer.

A notable observation in this study is that macrophages were required for LLC tumor control in Il1b^−/− mice. These results are consistent with previous reports showing the requirement of macrophages for an effective antitumor cytotoxic T-cell response following therapies that reduce immune suppression in the TME (56, 57). There are several potential mechanisms by which macrophages can participate in tumor control in IL1β-deficient mice, and these are mutually nonexclusive. First, macrophages may be important for T-cell recruitment. Although intratumoral expression of T-cell chemoattractants was not elevated in Il1b^−/− mice, macrophage-depletion strongly reduced their levels. It is conceivable that signals maintaining T-cell exclusion are suppressed in Il1b^−/− mice, allowing macrophage-derived T-cell chemoattractants to exert their function. Second, immunostimulatory TAMs may promote the intratumoral priming and expansion of antitumor CD8⁺ T cells and/or survival of tumor-infiltrating CD8⁺ T cells primed in the lymph node. Third, TAMs activated by CD8⁺ effector T cells may have a direct cytotoxic effect on cancer cells. Fourth, extratumoral macrophages, such as the ones residing in the tumor-draining lymph node, may also contribute to the priming of antitumor CD8⁺ T cells (58). Further studies are needed to elucidate the relative contribution of these mechanisms to the macrophage-dependent tumor control in Il1b^−/− mice.

We found that immunostimulatory MHC-II^high TAMs released large amounts of IL1β. It will be important to consider the detrimental effects of IL1β on antitumor immunity in the context of therapies that aim to reprogram TAMs toward a proinflammatory state that often involves the upregulation of IL1β. In light of our results, it is possible that addition of IL1β blockade would improve the efficacy of such TAM-reprogramming therapies.

In conclusion, this study provides support for the role of IL1β as a tumor-promoting factor whose inactivation results in an immune permissive TME. We suggest that the existence of inflammasome-independent IL1β release and neutrophil recruitment demonstrated here will have to be taken into consideration when applying the growing range of inflammasome inhibitors for cancer therapy (5).

L. Vande Walle reports grants from Kom op tegen Kanker during the conduct of the study. L. Boon is an employee of Polpharma Biologics Utrecht. G. Raes reports grants from FWO (Research Foundation - Flanders) during the conduct of the study, as well as other from Precirix (shareholder and consultancy fee paid to institute) and Abscint (shareholder) outside the submitted work. M. Lamkanfi reports grants from Kom op tegen Kanker during the conduct of the study. J.A. Van Ginderachter reports grants from Kom op tegen Kanker during the conduct of the study; grants from Oncurious NV, Montis NV, Roche Diagnostics, Ablynx NV, Johnson & Johnson, Camel-IDS, and eTheRNA outside the submitted work; and a patent for anti-macrophage mannose receptor single variable domains for targeting and in vivo imaging of tumor-associated macrophages issued and licensed to Oncurious, tumor-associated dendritic cell preparations and uses thereof pending, human PD-L1–binding immunoglobulins pending, and CCR8 nonblocking binders pending. D. Laoui reports grants from FWO (Research Foundation - Flanders) and Kom op tegen Kanker during the conduct of the study. No disclosures were reported by the other authors.

M. Kiss: Conceptualization, formal analysis, investigation, visualization, writing–original draft, writing–review and editing. L. Vande Walle: Formal analysis, investigation. P.H.V. Saavedra: Formal analysis, investigation. E. Lebegge: Formal analysis, investigation. H. Van Damme: Investigation. A. Murgaski: Investigation. J. Qian: Formal analysis. M. Ehling: Investigation. S. Pretto: Investigation. E. Bolli: Investigation. J. Keirsse: Investigation. P.M.R. Bardet: Investigation. S.M. Arnouk: Investigation. Y. Elkrim: Investigation. M. Schmoetten: Investigation. J. Brughmans: Investigation. A. Debraekeleer: Investigation. A. Fossoul: Investigation. L. Boon: Resources. G. Raes: Funding acquisition. G. van Loo: Funding acquisition. D. Lambrechts: Supervision. M. Mazzone: Supervision. A. Beschin: Resources, supervision. A. Wullaert: Supervision. M. Lamkanfi: Supervision, funding acquisition, writing–review and editing. J.A. Van Ginderachter: Conceptualization, supervision, funding acquisition, writing–review and editing. D. Laoui: Conceptualization, supervision, funding acquisition, writing–review and editing.

The authors thank Maria Solange Martins, Lea Brys, Ella Omasta, Marie-Therese Detobel, and Nadia Abou for technical and administrative assistance. They thank the VIB Bioimaging Core for training, support, and access to the instrument park and Amanda Gonçalves for help with slide scanning. The authors thank Maarten Verdonckt for help with mouse genotyping and Ulrika Frising, Tomoko Asaoka, and Maria Giulia Doglio for help with Western blots. They thank Lars Vereecke and Mozes Sze for help with the multiplex immunoassays. The authors thank Zsolt Czimmerer and Ana Rita Pombo Antunes for critically reading the article.

M. Kiss is supported by doctoral grants from Research Foundation - Flanders (FWO, 1S23316N) and Kom op tegen Kanker (Stand up to Cancer). E. Lebegge is supported by a doctoral grant from FWO (1S67419N). A. Murgaski is supported by a doctoral grant from FWO (1S16718N). H. Van Damme is supported by a doctoral grant from FWO (1S24117N). S. Pretto is supported by a doctoral grant from FWO (1S68420N). P.M.R. Bardet is supported by a doctoral grant from FWO (1154720N). S.M. Arnouk is supported by a doctoral grant from FWO (1S78120N). D. Laoui is supported by grants from FWO (12Z1820N), Kom op tegen Kanker, and Vrije Universiteit Brussel. P.H.V. Saavedra, P.M.R. Bardet, S.M. Arnouk, M. Schmoetten, J. Brughmans, A. Debraekeleer, A. Beschin, A. Wullaert, M. Lamkanfi, and J.A. Van Ginderachter are supported by Kom op tegen Kanker (STIVLK2017000401). A. Wullaert is supported by FWO (3G.0447.18). G. Raes is supported by a grant from FWO-NAFOSTED (G0F3616N).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.