All animal work was approved by the National University of Singapore Institutional Animal Care and Use Committee (IACUC), an AALAAC accredited institution and was in accordance with the National Advisory Committee for laboratory Animal Research (NACLAR) Guidelines. TRAMP-C2 cells were injected subcutaneously into C57BL6 mice or AIM2-/- mice on the same background. Tumors were homogenized or treated with collagenase before isolating CD11b+ cells through a column and incubating overnight, after which supernatants were collected for ELISA analysis.

Human THP-1, U937, HeLa cells, A549, HCT-116, and mouse TRAMP-C2 cells were purchased from ATCC (USA). A549, HeLa and TRAMP-C2 cells were grown in DMEM medium (Invitrogen, USA), while HCT116 cells were cultured in McCoy’s 5A medium (Invitrogen, USA), and both mediums are supplemented with 10% heat-inactivated FCS (Hyclone, USA) and 1% pen/strep (Invitrogen). THP-1 and U937 cells were cultured in RPMI medium (Hyclone, USA) supplemented with 10% heat-inactivated FCS (Hyclone, USA). All cells were grown at 37°C in a humidified 5% CO2 incubator (Thermo Scientific, USA). THP1 cells were differerentiated in PMA (#tlrl-pma, InvivoGen, USA) and stimulated with 1 μg/ml LPS (#tlrl-peklps, InvivoGen, USA) overnight before fresh cell medium was replaced and the cells were either treated with 50% cancer CM or transfected with 1 μg/ml poly (dA:dT) (#tlrl-patc, InvivoGen, USA) with lyovec transfection reagent. Inhibitors THP1 Cells were primed with 100 ng/ml PMA, followed by LPS (1 μg/ml) for 3 hours prior to treatment 1 hour with 1 μg/ml Poly(dA:dT) complexes (AIM2 inflammasome inducer) or A549 tumor CM in the presence of ODN TTAGGG (A151; AIM2 inhibitor, 0.5 μM) or MCC950 (NLRP3 inhibitor, 0.5 μM). Cell supernatants were detected using ELISA.

Bone marrow derived macrophages (BMDMs) from mice were obtained by flushing the femurs and tibias of mice with DMEM media. Briefly, red blood cells were removed through osmotic lysis and the bone marrow cell suspension was washed twice with PBS and cultured with BMDM media (DMEM medium containing 10% FBS (v/v), 100 U/mL penicillin, 100 µg/mL streptomycin, 2 mM L-glutamine and 20% L-929-conditioned DMEM (v/v) as a source of M-CSF). After 3 days of culture, the cells were supplemented with fresh BMDM media. At day 7, culture media and non-adherent cells were removed, and the remaining adherent cells were replenished with fresh BMDM media before experiments.

A549, Hela, MCF7 cells are cultured in T-75 flasks until 80% confluency before the existing media was replaced by new DMEM medium. The cells were grown for another 24 h in the new DMEM media before the conditioned media was collected and filtered using 0.45 µm filter (Millipore, USA) prior to use.

Cells were stained with Fixable Viability Dye eFluor 506 (1:1000) in 1x PBS for 30 min at 4°C (Thermo Fisher Scientific, USA) before incubated with Fc block (BioLegend, USA) for 15 min at 4°C. Cells were stained with fluorescence-conjugated antibodies (BioLegend, USA) diluted in FACS buffer and incubated on ice for 30 min in the dark. Macrophages were identified as CD45+ Ly6G- Gr1- F4/80+ CD11b+ cells. Cells were analyzed on a BD LSRFortessa (BD Biosciences, USA) or Attune NxT Flow Cytometer (Thermo Fisher Scientific, USA). Data analysis was performed using FlowJo software (FlowJo, USA).

Immunohistochemistry (IHC) was performed using the Leica Bond max autostainer (Leica Biosystems, Nussloch GmbH) and bond polymer refine detection kit (DS9800, Leica Biosystems) following the manufacturer’s instructions. Briefly, tissues were fixed in formalin, dehydrated and cleared before adding paraffin (pre-heated to 60oC) and left overnight. Sectioning was performed with a microtome. Formalin-fixed tissue sections were de-waxed, rehydrated and antigen retrieval was achieved using citrate buffer at pH 6 for 20 minutes. Endogenous peroxidase activity was quenched in 0.3% hydrogen peroxide for 10 minutes and non-specific sites were blocked with 3% bovine serum albumin for 20 minutes. Further, single immunostaining with primary antibodies for CD68 (1:50, Bio-Rad), inducible nitric oxide synthase (iNOS, 1:50, Abcam), mannose receptor CD206 (1:200, Abcam) followed by incubation with respective HRP-linked secondary antibodies (for CD68: goat anti-rat antibody Bio-Rad; for iNOS and CD206: goat anti-rabbit antibody, DAKO) was performed. 3,3’-diaminobenzidine (DAB) was used as the chromogen to obtain the brown colour. Counterstaining was with hematoxylin for 10 min. The mean score of five high-power fields (HPF) (×400) was counted for each slide under a light microscope (Olympus BX43) and digital images were acquired using Olympus cellSens Entry imaging software. In the IHC stained sections, macrophages observed with their cytoplasm stained tan or brown were considered positive.

Cytosolic dsDNA and DNA-RNA staining were performed as described previously (16). Briefly, 50% formamide (VWR International, USA) diluted in PBS was added for 10 mins at room temperature, followed by incubation for 15 mins at 75°C to denature dsDNA. Cells were washed with PBS and incubated with blocking buffer (1% BSA (Sigma Aldrich), 2% goat serum (Hyclone) in PBS) for 1 h to prevent non-specific binding of antibodies. Cells were then stained with dsDNA (1:200, Sigma-Aldrich, USA) or S9.6 DNA-RNA hybrids (1:100, Kerafast, Boston, USA) antibodies overnight in 4°C. Cells were subsequently washed with PBST (0.1% Tween) thrice followed by anti-mouse IgG coupled with Cy3 (Millipore, Singapore) or anti-rabbit IgG coupled to Alexa Fluor 488 (Invitrogen, Singapore). Finally, DNA fluorophore DAPI (0.5 μg/ml in PBS, KPL Inc., USA) was added for 10 min. Slides were washed once in PBS before mounted with Da-/- fluorescent mounting medium (Da-/-, UK). Confocal images of staining were captured using a Zeiss Axio Imager Z1 fluorescent microscope equipped with AxioVision 4.8 software (Carl Zeiss MicroImaging, USA) or confocal TCS SP5 (Leica, Singapore). Images were analyzed using Photoshop CS4 (Adobe, USA) or ImageJ. Colocalisation of AIM2 with DNA or other proteins were quantified using Metamorph (Metamorph NX, version 8.12, Molecular Devices, USA).

Cells were lysed with RIPA (Radio-Immunoprecipitation Assay) buffer containing protease inhibitors (Calbiochem, USA). 30 μg of total protein was loaded onto a SDSPAGE gel and electroblotted onto Hybond ECL membrane (GE Healthcare, UK). The blots were stained with primary antibodies (AIM2 (1:1000, Cell Signaling Technology, USA), NLRP1 (1:200, Santa Cruz Biotechnology, USA), NLRP2 (1:1000, Proteintech, USA), NLRP3 (1:1000, Adipogen, USA), NAIP (1:1000, Adipogen, USA), ASC (1:1000, Cell Signaling Technology, USA), Caspase 1 (1:200, Santa Cruz Biotechnology, USA), IL1β (1:1000, R & D, USA), Tubulin (!;1000, Cell Signaling Technology, USA), Actin (1:1000, Cell Signaling Technology, USA) overnight at 4°C after blocked with 5% low fat milk (Bio-Rad, USA). The blots were washed with TBST and incubated with a secondary HRP-linked antibody (1:5000, Santa Cruz Biotechnology, USA). Blots were washed TBST before visualization using ECL reagents according to manufacturer’s instructions (Perkin Elmer, USA). Each protein bands’ intensities were normalized against its loading control band intensity for quantification using ImageJ. Briefly, an area of interest rectangle was created around each band and the integrated density was measured, background subtracted, and divided by its specific loading control band.

Cytokines in culture supernatant and tissue lysates were measured by ELISA using commercially available kits. Mouse IL-1β (88-7013, eBioscience, USA) and human IL-1β (88-7261-86, eBioscience, USA) were used in accordance with the manufacturer’s protocol. Absorbance was read with a Tecan Spark 10M microplate reader (Tecan, USA) at absorbance values as indicated in the manufacturer’s protocol.

Tumors were mechanically dissociated and strained through a 40 µm nylon mesh before digestion into single cells with collagenase type II (0.5 mg mL−1), collagenase type IV (0.5 mg mL−1), hyaluronidase (10 U/mL) and DNase I (0.01 mg mL−1) for 2 h at 37 °C. The dissociated cells were collected, lysed by RBC lysis buffer before CD11b+ tumor associated macrophages were isolated out of these suspensions by the MACS technology according to the manufacturer’s instructions using 10 μl CD11b microbeads (130049601; Miltenyi, USA) per 1x10⁷ cells. The isolated CD11b+ cells were grown overnight in complete RPMI media overnight before the cells and cell supernatants were collected for analysis.

AIM2-specific siRNA and scrambled siRNA (Thermo Fisher) were used according to manufacturer’s instructions. THP-1 cells were seeded at a concentration of 5x105/well in a 12 well plate (Greiner Bio-science) before silencing with a final concentration 10 nM of siRNA and grown overnight before treatment with or without 100ng/ml PMA, 1 μg/ml LPS, and 50% A549 tumor conditioned media for another 24 h.

A549 cells were seeded in 24 well plates and treated with IL-1β (#29-8018-65, Thermo Scientific, USA) at the indicated concentrations for 72 h, after which cell viability was measured using an MTT Cell Proliferation and Cytotoxicity Assay Kit (Boster Bio), according to the manufacturer’s instructions. Absorbance was measured using a TECAN spectrophotometer (Tecan Trading AG, Switzerland) at 490 nm.

A549 cells were seeded into cell migration chambers (SPL, Korea) of PET membrane with 8.0 μm pore size. IL-1β was added to the cells and the chambers were incubated for 24 h in a humidified chamber at 37°C, 5% CO2. For the invasion assay, inserts were coated with 10% Matrigel prior to cell seeding. For co-culture, the bottom chamber was seeded with mouse BMDM cells stimulated with LPS (1 µg/ml) one day prior to TRAMPC2 cell seeding before the chambers were incubated for 24 h. At the end of the incubation period, the inserts were washed using 1xPBS. Inserts were fixed with 100% methanol for 15 minutes before washing and air drying, followed by cell staining using 0.5% crystal violet (filtered) for 30 minutes. The upper membrane of the inserts was rinsed and cleaned to remove excess stain and non-migrated cells. Five microscopic fields of each insert were captured using Nikon Stereoscopic Microscope SMZ1500 and Digital Camera DXM1200F (Nikon, Melvile, NY, USA).

Unpaired Student’s T-Test or One-way/Two-way analysis of variance (ANOVA) with Tukey’s post-test was performed when appropriate. Results are presented as mean values ± standard error (SE).

A p-value of < 0.05 or less was deemed significant.

The presence of nucleic acids have been reported in cancer tissues and cell lines (3) which could activate DAMPs in tumor associated macrophages (TAMs). We first confirmed the expression of dsDNA and in a spontaneous breast cancer mouse model which develops tumors at ~6-9 months of age- MMTV-Wnt+ mice, compared to non-tumor fat pads. Indeed, tumors expressed more dsDNA compared to normal mammary tissue ( Figure S1A ). We next investigated the tumor microenvironment in MMTV-Wnt+ fatpads and tumor (which developed at 9 months) and MMTV-Wnt- fat pads. Many more CD11b+F480+ cells were present in tumor tissue ( Figure S1B ). Interestingly macrophage infiltration occurs early before tumor development, confirmed by immunohistochemistry of MMTV-Wnt+ and MMTV-Wnt- fat pads stained with CD68 and counterstained with Haemotoxylin (at 2,4,6 and 9 months of age). Figure 1A shows mouse mammary tissue composed of fat pads (black arrow), ducts (red arrow) and acinar structures (yellow arrow). MMTV-Wnt+ pads exhibit ductal and acinar proliferation with branching and budding, progressing from 2 months to 6 months in comparison to their respective age matched control littermates. Formation of malignant tumor is observed at 9 months which shows atypical epithelial cells forming enlarged irregular glands with intraluminal papillary projections. At places, there is also formation of cysts filled with secretory material and many solid areas of tumor cells can be seen. Macrophages were stained with CD68 (400x) and a significant increase in CD68+ cells was seen in fat pads from MMTV-WNT+ mice progressing from 2-6 months and tumors at 9 months ( Figures 1B, C ). In addition, staining of sections with an inflammatory macrophage marker, iNOS or an alternatively activated macrophage marker CD206 demonstrated a time-dependent and significant increase in the number of CD206 expressing cells at 9 months of age, when the tumor had formed ( Figure 1D ). Next to assess if tumor-derived nucleic acids can activate the inflammasome in leukocytes in the tumor microenvironment, CD11b+ cells isolated and cultured overnight from MMTV-Wnt+ tumors secreted more IL-1β and expressed higher levels of inflammasome components AIM2 and NLRP3 as compared to CD11b+ cells derived from normal fat pads ( Figures 1E, F ).

To determine if tumors secrete mediators which can enhance the IL-1β and inflammasome activation in macrophages, we collected conditioned media (CM) from TRAMPC2 mouse prostate cancer cells overnight. TRAMPC2 CM induced a high level of IL-1β production in LPS-primed bone marrow derived macrophages (BMDM) ( Figure 2A ). This was confirmed with human monocytic THP-1 cells either resting (monocytic) or differentiated with PMA (macrophage-like) and stimulated with CM from several human cancer cells ( Figure 2B ). Monocytic THP1 cells did not secrete IL-1β in response to LPS or LPS + CM, while macrophage-like THP1 cells secreted more IL-1β basally, when stimulated with LPS and significantly more when stimulated with both LPS+CM ( Figure S2A ). In addition, A549 tumor CM significantly enhanced the cleavage and maturation of inflammasome components caspase-1 and IL-1β and upregulated AIM2 ( Figures 2C, D ). Surprisingly, despite the presence of cytosolic DNA or exogenous DNA transfection, epithelial cancer cells do not produce appreciable amounts of IL-1β compared to myeloid THP-1 cells, implying that the IL-1β in the tumor microenvironment does not originate from the tumor cells themselves ( Figure S2B ).

Next, AIM2 was either silenced with siRNA (( Figure 3A ) or treated with A151, an inhibitor of AIM2 and TLR9 ( Figure 3B ) which resulted in the inhibition of IL-1β and inflammatory components after A549 CM in human THP1 cells, with quantified densitometry ( Figure 3C ). AIM2-/- and caspase 1-/- BMDM, treated with LPS and either dsDNA or tumor CM, similarly produced significantly less IL-1β compared to WT BMDM ( Figures 3D, E ). This shows that macrophage AIM2 is important for IL-1β secretion in response to tumor CM. However, as Caspase-1 plays a role here, we examined if other inflammasomes could also be involved. Inhibition of NLRP3 with MCC950 partially inhibited the release and cleavage of IL-1β and induced by LPS+ATP as well as A549 CM, indicating that the NLRP3 inflammasome activation is partially responsible ( Figure S3 ).

Exogenous IL-1β stimulated proliferation of TRAMPC2 cancer cells, and addition of exogenous IL-1β stimulated proliferation of TRAMPC2 cancer cells similarly in the presence of WT and AIM2-/- or Caspase1/11-/- macrophages ( Figure 4A ). However, coculture of TRAMPC2 cancer cells with BMDM increased their proliferation which was AIM2 and Caspase 1/11 dependent as this was not observed with AIM2-/- or Caspase1/11-/- BMDM ( Figures 4B, C ). In addition, IL-1β enhanced the migration and invasion of TRAMPC2 cancer cells ( Figure 4D ) which implies a pro-tumorigenic role for IL-1β.

The presence of dsDNA and DNA-RNA hybrids has been reported in cancer tissues and cell lines (3, 17–19). We hypothesized that these cytosolic nucleic acids may act as intracellular ligands to stimulate AIM2 inflammasome activation. Therefore, we next determined if cytosolic DNA or DNA-RNA hybrids were present in the spontaneous mammary tumors in MMTV-Wnt+mice by staining MMTV-Wnt+ tumors and non-tumor fat pads from MMTV-Wnt- for both DNA-RNA hybrids and dsDNA using immunofluorescence staining. Both cytosolic dsDNA and DNA-RNA hybrids are constitutively present in tumor tissues and are not found in non-tumor mammary fat pads ( Figure 5A ). In addition, AIM2 colocalized with cytosolic DNA and DNA-RNA hybrids in A549 cells ( Figures 5B, C ) as well as other tested cell lines (U937, THP-1, Hela, and HCT-116, Figure S4 ).

To determine which nucleic acid species in the tumor CM could be responsible for inflammasome activation, DNAase, RNAase H, and RNAase A which degrade dsDNA, DNA-RNA hybrids, and dsRNA, respectively, was added to A549 CM prior to addition to THP-1 cells. Tumor CM stimulated the production of IL-1β which was significantly inhibited with RNAase H, and not RNAase A. Interestingly, only a slight reduction was observed with DNAase treatment ( Figure 5D ). Furthermore, treatment with RNAase H significantly reduced the maturation of IL-1β, and inhibited gasdermin D-N expression ( Figure 5E ). Treatment with DNAase, to a smaller extent, reduced the maturation of IL-1β, but did not inhibit the IL1β release, suggesting that DNA-RNA hybrids present in tumor CM, could be responsible for the IL-1β secretion and inflammasome activation in THP1 macrophages, rather than dsDNA.

Finally, to validate that the AIM2 inflammasome can sense dsDNA and DNA-RNA hybrids, BMDM from WT and AIM2-/- mice were generated and transfected with different concentrations of DNA-RNA hybrids or dsDNA poly (dA: dT). The lowest concentration of DNA-RNA hybrid transfection (0.1μg/ml) increased IL-1β secretion which was AIM2 dependent. Interestingly, transfection of DNA-RNA hybrids at higher concentrations (1µg/ml and 5 µg/ml) did not increase IL-1β secretion compared to the transfected control. ( Figure 5F ). However, only higher concentrations of poly (dA: dT) stimulated IL-1β production. This data suggests that AIM2 can respond to both DNA-RNA hybrids and dsDNA, but is more sensitive to DNA-RNA hybrids as compared to dsDNA

In summary, this study shows that macrophage AIM2 is activated by nucleic acids secreted from the cancer cells, which leads to inflammasome activation, IL-1β cleavage and release from macrophages. IL-1β and coculture of tumor cells with macrophages enhances the proliferation of tumor cells which is AIM2 and caspase 1/11 dependent (Gigure 6). AIM2 is more sensitive to stimulation with DNA-RNA hybrids compared to dsDNA, suggesting that both dsDNA and DNA-RNA hybrids can activate AIM2 inflammasome in tumor associated macrophages in the tumor microenvironment.