Although ApoBDs can maintain membrane integrity shortly after their formation and the onset of PMR of ApoBDs has profound implications in disease settings (2, 10), the regulation of ApoBD lysis has not been defined. Given the role of NINJ1 in PMR (11–13), we sought to determine if NINJ1 undergoes oligomerisation on ApoBDs and regulates their PMR. To this end, we first generated NINJ1-deficient immortalised bone marrow-derived macrophages (iBMDMs) using CRISPR/Cas9 gene editing ( Figure 1a ). Compared to Cas9 control, NINJ1 disruption did not significantly alter apoptosis progression as shown by Caspase 3/7 Glo^® assay detecting apoptotic caspase activity upon BH3 mimetics treatment ( Figure 1b ). ApoBDs were isolated from the apoptotic samples using a differential centrifugation-based approach ( Figure 1c ). We confirmed that the isolated ApoBDs are of the expected size (~1-5 μm), of high purity (~90%) and exhibit typical apoptotic markers, including phosphatidylserine exposure (indicated by annexin V staining), cleaved caspase 3 and its cleavage substrate, pannexin 1 membrane channels ( Figures 1d, e , Supplementary Figure 1a ). Using Blue Native-PAGE ( Figure 1f ) and bis(sulfosuccinimidyl) suberate (BS3) crosslinking followed by SDS-PAGE ( Figure 1g ), we detected a high level of NINJ1 oligomerisation on ApoBDs, compared to untreated and apoptotic cell-enriched sample (ACES). This observation not only reveals that NINJ1 could be functional on ApoBDs but also suggests that apoptotic cells complete ApoBD formation prior to transitioning to secondary necrosis, to ensure non-inflammatory cell clearance.

Consistently, by monitoring the apoptotic cell disassembly process using time-lapse microscopy, we found that both NINJ1^-/- and Cas9 control iBMDMs undergoing apoptosis can readily generate ApoBD through the formation of membrane blebs and membrane protrusions ( Figure 1h ). Time-course analysis of NINJ1 oligomerisation following apoptosis induction also showed limited NINJ1 oligomerisation at 4 hours post BH3 mimetics treatment (i.e. the time-point when ApoBDs were collected) ( Supplementary Figure 1b ). Collectively, these data suggest that NINJ1 does not impact ApoBD biogenesis, an important event to aid efficient removal of cell corpses via efferocytosis.

As NINJ1 oligomerises on ApoBDs, we next sought to detect PMR in ApoBDs by using two experimental approaches. First, using the lactate dehydrogenase (LDH) release assay, a standard measure of late-stage PMR (12, 13), we found that NINJ1 deficiency markedly rescued PMR, not only in whole apoptotic iBMDMs but also in ApoBD samples ( Figure 2a ). In addition, we also performed FITC-dextran exclusion assays to visualise PMR at the single vesicle level. NINJ1^-/- iBMDM-derived ApoBDs showed twice as much FITC-dextran exclusion as compared to Cas9 control iBMDM-derived ApoBDs, further indicating reduced PMR of NINJ1 deficient ApoBDs ( Figures 2b, c ).

NINJ1 controls important downstream consequences of cell death by facilitating PMR and the non-selective release of cellular contents, such as DAMPs and viruses from dying cells (11–14). Notably, although ApoBDs can harbour a variety of cellular contents, whether DAMPs are released from ApoBDs and the mechanism of release has not been defined. To determine whether NINJ1 is involved in DAMP release upon ApoBD lysis, we sought to detect HMGB1, a prominent DAMP, in supernatants of Cas9 and NINJ1^-/- ApoBDs at 4 h post-isolation using the Lumit^® HMGB1 bioluminescence assay. First, we found that HMGB1 are released by ApoBDs ( Figure 2d ), under conditions whereby ~80% of ApoBDs had undergone PMR ( Figure 2a , right panel). Furthermore, the release of HMGB1 by ApoBDs was significantly impaired upon NINJ1 deficiency. In addition, viral infections that induce apoptosis can hijack resultant ApoBDs to disseminate to other bystander cells (14), with ApoBD lysis as a potential mechanism for viral release. Using mouse norovirus (MNV) as an apoptosis-inducing viral model for iBMDMs (15), we were able to isolate ApoBDs from the infected cells ( Figure 2e ) and detect NINJ1 oligomerisation on ApoBDs by Blue Native-PAGE ( Figure 2f ). We also observed a marked decrease in LDH release ( Figure 2g ) and a partial reduction (~2.5 fold decrease) of viral release from ApoBDs derived from MNV-infected NINJ1^-/- iBMDMs as compared to MNV-infected Cas9 control iBMDMs ( Figure 2h ).

Together, these data suggest that NINJ1 further oligomerises on ApoBDs post-biogenesis and regulates their stability and thus the release of DAMPs and viral particles ( Figure 2i ). To our knowledge, NINJ1 represents the first regulator of EV stability.

BH3 mimetics ABT-737 (targets Bcl-2, Bcl-xl and Bcl-w) and S63845 (targets Mcl-1) were purchased from MedChemExpress (Monmouth Junction, NJ, USA). Fluorescently-labelled Annexin V (AV) and AV binding buffer were from BD Biosciences. TO-PRO-3 was purchased from Thermo-Fisher Scientific (Carlsbad, CA, USA).

Cas9 control and NINJ1^−/− murine immortalised bone marrow-derived macrophages (iBMDMs) were generated as previously reported (24) and cultured in complete Gibco^® Dulbecco’s Modified Eagle Medium media, 1 g/L glucose (Thermo-Fisher Scientific), supplemented with 10% (v/v) foetal bovine serum (FCS; Scientifix, Australia), 50 IU/ml penicillin, 50 µg/ml streptomycin (Thermo-Fisher Scientific), and 0.2% (v/v) MycoZap (Lonza, Switzerland) at 37°C in a humidified atmosphere with 5% CO₂.

Cas9 or NINJ1^-/- iBMDMs were seeded in T175 flasks at a density of 5 × 10⁶ cells per flask. On the next day, cells were treated with 2 µM of ABT-737 and 10 µM of S63845 in serum-free DMEM media supplemented with 1% bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA) for 4 h.

Caspase 3/7 activity was measured using the Caspase-Glo 3/7 Glo^® assay (Promega, WI, US) according to the manufacturer’s instructions.

ApoBDs were isolated from BH3 mimetic-treated or MNV-infected iBMDMs using differential centrifugation as previously described (25). Briefly, cells were centrifuged at 500 x g for 10 min to remove whole apoptotic cells. The supernatant was collected and centrifuged at 3,000 × g for 20 min to pellet ApoBDs. After three washes with 1× PBS (Thermo-Fisher), isolated ApoBDs were quality controlled by flow cytometry with AV (1:100) as well as TO-PRO-3 (0.5 µM) staining and counted with AccuCount blank flow cytometry particles (ProSciTech) before proceeding to further assays. Flow cytometry was performed using a BD FACSCanto II flow cytometer (BD Bioscience). Data were analysed using FlowJo software v10.10.0, with the gating strategy described by Jiang et al. (2016) (26).

Apoptotic cell-enriched sample (ACES) and ApoBDs were harvested and stained with A5-BV421 (1:100) prior to seeding on an 8-well chambered NuncTM Lab-TekTM II cover glass (Nunc, Denmark). Images were captured using a Zeiss 800/900 confocal laser scanning microscope (Zeiss, Germany) equipped with a 63x magnification oil immersion lens, at 37°C and 5% CO₂. Image analysis was performed using Zen software v2.6.

ACES and ApoBDs were prepared from BH3 mimetic-treated iBMDMs. Untreated cells, included as a viable control, were detached using an Accutase solution (Sigma-Aldrich). All samples were lysed using a native-PAGE lysis buffer containing 150 mM NaCl, 1% Digitonin, 50 mM Tris (pH 7.5), and a protease inhibitor cocktail (Roche) at 4°C for 30 min. NativePAGE sample buffer and Coomassie G-250 (Thermo-Fisher) were then added to the lysates before electrophoresis using a Native-PAGE 3-12% gel (Thermo-Fisher). Immunoblotting was performed subsequently.

Viable, ACES, ApoBD samples were resuspended in 1× PBS containing 3 mM bis(sulfosuccinimidyl) suberate (BS3) crosslinker (Thermo-Fisher) and incubated for 5 min at room temperature. The reaction was stopped by adding 20 mM Tris (pH 7.5), incubated for 15 min at room temperature and then membrane washed with 1× PBS at 3,000 x g for 10 minutes. ApoBDs/cell lysates were obtained by adding chilled lysis buffer containing 1% IGEPAL^® CA-630, 20 mM HEPES (pH 7.4), 10% glycerol, 1% Triton X-100, 150 mM NaCl, 50 mM NaF and Roche protease inhibitor cocktail. Proteins were resolved using NuPAGE 4-12% Bis-Tris gel (Invitrogen). Immunoblotting was performed afterwards.

After gel electrophoresis, proteins were transferred onto a PVDF membrane. Immunoblotting was performed with the following primary antibodies: NINJ1 Polyclonal Antibody (1:1000, Thermo-Fisher, PA5-95755), anti-pannexin 1 (1:1000; Cell Signalling), anti-caspase 3 (1:1000; Cell Signalling), and anti-β-actin (1:4000, clone AC-74, Sigma-Aldrich, A2228) in 3% BSA in PBS-T (i.e., 1× PBS with 0.1% Tween-20). Following the primary antibody incubation, the blots were incubated with HRP-conjugated secondary antibodies: sheep anti-mouse (1:5000, Millennium Science) or donkey anti-rabbit (1:5000, Millennium Science) in 5% milk in PBST (0.1% Tween-20). Chemiluminescence detection of target proteins were done using ECL Prime reagent (Bio-strategy) with a Syngene G:BOX gel documentation system (Syngene).

Cas9 and NINJ1^-/- iBMDMs were seeded in an 8-well Nunc^® Lab-Tek^® II chamber slide overnight. On the following day, BH3 mimetics were added and cells were imaged using a Zeiss spinning disk confocal microscope (Zeiss, Germany) equipped with a 63× oil objective, at 37°C with 5% CO₂. Images were captured every 2 minutes over a 12-hour period and analysed using Zen Blue imaging software (Zeiss, Germany).

Cas9 or NINJ1^-/- iBMDMs were treated with BH3 mimetics in a 96-well plate containing serum-free DMEM media supplemented with 1% BSA in a humidified atmosphere at 37°C, 5% CO₂ for 8 hours. For ApoBD samples, 6×10⁵ ApoBDs (isolated at 4 h post-apoptosis induction) were seeded per well in a 96-well plate for 4 h. Subsequently, supernatants were then harvested and centrifuged at 500 × g for 20 min to remove cells and ApoBDs. LDH was then detected using a LDH Cytotoxicity Assay Kit (Abcam) as per the manufacturer’s instructions. Absorbance was measured at 450 nm using a SpectraMax M5e Plate Reader (Molecular Devices, CA), and data were analysed using SoftMax Pro 5.2 software (Molecular Devices).

A total of 3×10⁵ ApoBDs were seeded in an 8-well Nunc^® Lab-Tek^® II chamber slide pre-treated with poly-L-lysine and incubated at 37°C with 5% CO₂ in a humidified atmosphere. After 1 h, 100 μg/ml FITC-dextran (4 kDa, Sigma-Aldrich) and A5-BV421 (1:100 dilution) were added. Images were captured using confocal microscopy with a Zeiss 800/900 Confocal Laser Scanning Microscope (Zeiss, Germany) and analysed using Zen software version 2.6. To quantify the percentage of FITC-negative ApoBDs, the number of intact vesicles (determined by the absence of FITC-dextran staining) in a 93.8 μM² area was divided by the total number of ApoBDs in the same area. In total, 922 and 1,269 ApoBDs from Cas9 and NINJ1^-/- iBMDMs, respectively, were quantified from 12 tile regions for both cell lines across three independent experiments.

ApoBD supernatants were harvested as aforementioned for the LDH release assay. HMGB1 was then detected using a Lumit^® HMGB1 Human/Mouse Immunoassay kit (Promega) as per the manufacturer’s instructions.

A total of 3.5 × 10⁶ Cas9 iBMDMs or NINJ1^-/- iBMDMs were seeded in a T175 flask for 24 h. Cells were infected with the MNV CW1 strain at an MOI of 5, as previously described (15). Briefly, the spent media was removed and replaced with 6 mL of serum-free DMEM containing MNV, allowing for a 1 h absorption at 37°C. Subsequently, 20 mL of infection media (DMEM supplemented with 2% FCS, 110 mg/L sodium pyruvate, 2 mM GlutaMAX) was added and cells incubated for 17 h at 37°C with 5% CO₂.

ApoBDs from MNV-infected iBMDMs and resultant ApoBD supernatants were prepared and collected as aforementioned. The ApoBD supernatants were stored at −80°C prior to the plaque assay. One day before the assay, RAW264.7 cells were seeded to approximately 70% confluency in a 12-well plate. Six ten-fold serial dilutions of ApoBD supernatants (10⁻² to 10⁻⁷) in DMEM were added onto the RAW264.7 cells for 1 h. After infection, the inoculum was removed and an overlayer containing 70% DMEM, 2.4% FCS, 13.3 mM NaHCO3, 24.4 mM HEPES, 200 mM GlutaMAX, and 0.35% low melt-point agarose was added. Samples were incubated at 4°C for 30 min to allow the agarose to solidify before transferring to a humidified, 37°C, 5% CO₂ incubator for 48 h. Cells were then fixed with 10% formalin for 1 h at room temperature. Plaques were visualised and counted after staining with toluidine blue.

Data are presented as means ± SEM. Statistical analyses were performed using unpaired, two-tailed Student’s t-test.