8-12 weeks old male and female C57BL/6J and C57BL/6NRj mice were purchased from Janvier Labs. Rag2^-/- mice (C57BL/6N-Rag2Tm1/CipheRj) were bred in specific pathogen-free (SPF) conditions at the Animal Facility of the Rega Institute (KU Leuven). C3^-/- mice (B6.129S4-C3tm1Crr/J) and Itgam^-/- mice (B6.129S4-Itgamtm1Myd/J) were purchased from The Jackson Laboratory. Mice were housed in acrylic filtertop cages with an enriched environment (bedding, toys and small houses) and kept under a controlled light/dark cycle (12/12h) at 21°C with water and food provided ad libitum. All experiments were approved and performed following the guidelines of the Animal Ethics Committee from KU Leuven (P125/2019).

Mice were starved for 15h and given a single oral gavage of 600 mg/kg APAP (Sigma-Aldrich) dissolved in warm PBS. Administration via oral gavage reflects the typical route of APAP-induced liver injury in patients. After 24, 48 or 72h, mice were sacrificed under anesthesia containing 80 mg/kg ketamine and 4 mg/kg xylazine, whereafter liver and blood were harvested. ALT in serum was determined with a kinetic enzymatic kit (Infinity, Thermo Fisher Scientific) according to the manufacturer’s instructions.

Serum levels of mouse C3 were determined by a commercially available C3 ELISA kit (ab157711, Abcam) according to the manufacturer’s instructions.

Human neutrophils were purified from whole blood of healthy volunteers by immunomagnetic negative selection (EasySep™ Direct Human Neutrophil Isolation Kit, StemCell Technologies) according to the manufacturer’s instructions. Ethical permission for use of human blood-derived leukocytes was obtained with the ethical committee from the University Hospital Leuven (UZ Leuven, S58418). Mouse bone marrow neutrophils were extracted from femurs and tibias of C57BL/6J mice by flushing the bones with 5 ml cold RPMI-1640 medium using a 26-gauge needle. Cells were filtered through a 70 µm nylon strainer and further purified with the EasySep™ mouse neutrophil enrichment kit (StemCell Technologies), following the manufacturer’s instructions.

Liver sections were stained with hematoxylin and eosin (H&E) and used to estimate hepatic necrosis via measurement of the necrotic area in the images. The livers were washed with 0.9% NaCl and fixed in 4% buffered formalin. Subsequently, the samples were dehydrated in ethanol solutions, bathed in xylol and included in histological paraffin blocks. Tissue sections of 5 μm were obtained using a microtome and stained with H&E. Sections were visualized using a BX41 optical microscope (Olympus) and images were obtained using the Moticam 2500 camera (Motic) and Motic Image Plus 2.0ML software.

The left liver lobes of mice were harvested, embedded in Tissue-Tek O.C.T. Compound (Sakura Finetek Europe) and snap frozen in liquid nitrogen. 10 µm sections were cut using a Cryostat Microm CryoStar and subsequently fixed, permeabilized and blocked. Sections were incubated overnight at 4°C with 10 µg/ml polyclonal rabbit anti-human/mouse fibrin(ogen) (Dako), 5 µg/ml rat anti-mouse C3b/iC3b (clone 3/26, Hycult Biotec) and 5 µg/ml rabbit anti-mouse C1q (clone 4.8, Abcam). Secondary antibodies were added for 3h at RT: Alexa Fluor 647 donkey anti-rabbit, Rhodamine RED-X (RRX) donkey anti-mouse IgM, Alexa Fluor 488 donkey anti-rat, Alexa Fluor 560 donkey anti-rabbit (all at 10 µg/mL, Jackson ImmunoResearch). 10 µg/ml of Hoechst was added for 30 min at RT to stain nuclei. Finally, slides were mounted with ProLong Diamond Antifade Mountant. Images were captured using the Andor Dragonfly High-Speed Confocal Microscope (Oxford Instruments) or a Zeiss Axiovert 200M fluorescence microscope, and analyzed with FIJI. 8 images were acquired per liver with a 25X objective. Stained areas were selected using the threshold tool in FIJI, from which the percentage area of staining was determined. Pearson’s coefficient was calculated in FIJI using the JACoP plugin. Comparisons between WT and Rag2^-/- mice were normalized to the degree of injury [% of fibrin(ogen) labeling]. All images can be provided in different colors upon request.

Mice were anaesthetized with a subcutaneous injection of 80 mg/kg ketamine and 4 mg/kg xylazine. For the experiments with APAP-induced liver injury, fluorescent antibodies (4 µg/mouse) and dyes (2 µl of a 10 mM Sytox Green solution; Thermo Fisher Scientific) were dissolved in 100 µl sterile PBS and injected intravenously 10 minutes before the surgery. The surgical procedure is described in detail in Marques et al. (24). For the FTI experiments, 1 mm³ burns were made with a cauterizer and the injury site was then stained with 10 µl of pHrodo Red succinimidyl ester (SE) (4 µM; Thermo Fisher Scientific). The incision was stitched, and after 6h, mice were again anaesthetized with ketamine and xylazine to image the burn injury site. Images were taken every 30 sec for at least 30 min with the Dragonfly Spinning-Disk Confocal Microscope (Oxford Instruments) using the 25X objective. Quantification of phagocytosis was done in a blind manner by two individuals and counted manually. The percentage of Sytox Green labeling was determined from 2 mosaic images, each composed of 16 images per mouse, with FIJI software using thresholding. The % of CD11b⁺ cells containing DNA was determined using Imaris software. Surfaces overlaying live cells and DNA debris were generated from 3D images and counted manually.

Necrotic debris was generated from HepG2 cells by inducing mechanical disruption with a pellet mixer for 5 min in 0.1 M sodium bicarbonate (pH 8.5). The necrotic debris was labeled by adding 2 µl of 10 mM pHrodo Red SE (Thermo Fisher Scientific) solution per 10x10⁶ cells. The debris was opsonized with 20% normal human serum, C1q-depleted serum (Complement Technology) or C3-depleted serum (Complement Technology) in PBS for 1h at 37°C. Opsonized debris was added to purified neutrophils in a 1:10 cell/debris ratio. Neutrophils were stimulated with 10^-7 M N-formyl-Met-Leu-Phe (fMLF; Sigma-Aldrich) for human neutrophils or 1 µM WKYMVM (Phoenix Pharmaceuticals) for mouse neutrophils, and labeled with 1 µM calcein AM viability dye (Invitrogen). Two 3D mosaics are captured per well, each comprising 9 overlapping images taken after 3h incubation at 37°C with the 25X objective of the Dragonfly Confocal Microscope (Oxford Instruments). Each condition was plated in duplicate and replicated at least 3 times. 3D reconstructions were generated using Imaris software. Additionally, with Imaris, surfaces were overlaid onto live cells and necrotic debris through thresholding, after which the volume of overlap was calculated in µm³.

Liver lobes were surgically removed, put in MACS tubes with RPMI-1640 (Biowest Riverside) and minced with a gentleMACS Dissociator (Miltenyi Biotec). To this suspension, 2.5 mg collagenase D (Roche) and 1 mg DNAse I were added per liver for 1h at 37°C. The cell suspension was washed with PBS (300 g, 5 min, 4°C). Non-parenchymal cells were separated by density gradient centrifugation at 60 g for 3 min at 4°C. Supernatant was collected and filtered through a 70 µm nylon cell strainer. After centrifugation (300 g, 5 min, 4°C), ACK Lysing buffer (Gibco) was added for 10 min to the pellet to lyse red blood cells. 1x10⁶ cells were collected in FACS tubes and washed with PBS (300 g, 5 min, 4°C). Zombie Aqua Fixable Viability dye (Biolegend) together with mouse FcR blocking Reagent (Miltenyi Biotec) were incubated for 15 min in the dark. Then, cells were washed with PBS supplemented with 0.5% bovine serum albumin (BSA) and 2 mM EDTA and the fluorescently labeled antibodies were incubated for 25 min at 4°C in the dark. After a final washing step, cells were read in a Fortessa X20 (BD Biosciences). Data was analyzed using FlowJo 10.8.1 software.

HepG2 cells were mechanically lysed using a 22G syringe in PBS with 20 µg/ml RNAse A (Sigma-Aldrich) to generate RNA-free necrotic debris. The debris was incubated for 30 min at 37°C to allow enzyme activity. Then, the debris was washed twice with PBS containing 2 mM EDTA and 0.1% BSA, and centrifuged at 60g for 3 min to remove intact cells. The debris was opsonized with 20% serum or C3-depleted serum (Complement Technology) and then incubated for 1h at 37°C. Neutrophils from healthy donors were purified by immunomagnetic negative selection with an EasySep kit and stimulated with 10^-7 M fMLF. Cells and debris were co-incubated in a 6 well plate at a 1:10 cell/debris ratio and centrifuged at 300g for 5 min before being incubated for 3h at 37°C. Cells were harvested and total RNA was extracted by lysing the cells with β-mercaptoethanol and a Rneasy Plus Mini Kit (Qiagen) following the manufacturer’s instructions. After extraction, total RNA quality and quantity were determined using a Nanodrop.

cDNA was obtained by reverse transcription using the high-capacity cDNA Reverse Transcriptase kit (Applied Biosystems). mRNA levels were analyzed by quantitative PCR using a TaqMan Gene Expression Master Mix (Applied Biosystems) and a 7500 Real-Time PCR System apparatus. Expression levels of genes of interest were normalized for the average RNA expression of three housekeeping genes (CDKN1A, 18S and GAPDH) using the 2^−ΔΔCT method (25).

Data were analyzed using GraphPad Prism v9.3.1. All data are expressed as mean ± standard error of the mean (SEM). A Shapiro-Wilkinson test was performed to check for normality. Normally distributed data were analyzed with a Student’s t test or One-way ANOVA. Non-parametric data were analyzed with a Mann-Whitney test or Kruskal-Wallis test. Grubb’s test (extreme studentized deviate) was applied to determine significant outliers, which are identified as red dots in the graphs and removed from statistical analysis. A p-value equal or lower than 0.05 was considered significant.

To assess complement activation at necrotic injury sites, a mouse model of paracetamol/acetaminophen (APAP)-induced liver injury was used, as it is characterized by extensive death of hepatocytes through necrosis (16). A sublethal dose of 600 mg/kg APAP was administered via oral gavage, causing significant liver damage as early as 12h after administration. In this acute model, hepatocellular necrosis was observed by elevated levels of serum alanine aminotransferase (ALT) ( Supplementary Figure S1A ), an enzyme primarily found in hepatocytes that serves as a biomarker for liver damage. Using histopathology, necrotic lesions were detected around the centrilobular veins, a typical pattern for APAP-induced injury, with the highest severity 12h after APAP administration ( Supplementary Figures S1B, C ). At 48h, the injury decreased significantly, as depicted by lower serum ALT levels and reduced necrotic areas ( Supplementary Figures S1A, B ). Fibrin deposition, known for its specific accumulation at necrotic sites after the activation of the coagulation cascade (26), was evaluated over time on liver cryosections to estimate the area of necrosis. Significant fibrin(ogen) staining was observed after 24h, whereafter it gradually decreased ( Figures 1A, C ; Supplementary Figure S2A ). A similar pattern was observed for IgM, with the highest deposition occurring after 24h and diminishing at later timepoints ( Figures 1B, D ; Supplementary Figure S2B ). Remarkably, C1q and (i)C3b deposition remained high up to 48h ( Figures 1E, F ), indicating that the deposition of antibodies preceded complement activation through C1q binding and C3 cleavage. The staining in necrotic regions was not because of autofluorescence or unspecific labeling as confirmed in cryosections stained with secondary antibodies only ( Supplementary Figure S1D ). Pearson’s correlation coefficient between (i)C3b and fibrin was significantly higher in comparison to (i)C3b and intact cell nuclei in the liver ( Figure 1G ). In addition, the other components of the classical complement pathway, IgM and C1q, also had high colocalization with (i)C3b ( Figure 1H ). All this shows that complement proteins are deposited specifically at sites of necrotic injury in the liver. These data were supported by significantly lower C3 levels in the serum of APAP-treated mice ( Figure 1I ), which confirm complement activation in response to injury.

To investigate the contribution of C3 to the resolution of necrotic liver damage, C3^-/- mice were subjected to APAP overdose and evaluated at 2 timepoints: a) after 24h, to assess the peak of injury and b) after 48h, to observe the degree of tissue repair. No differences in serum ALT levels were observed after 24h, while significantly higher levels of ALT were found after 48h in C3^-/- mice compared to WT mice ( Figure 2A ). In addition, fibrin staining of liver cryosections revealed no differences at the peak of injury, whereas more fibrin staining was found in C3^-/- mice at the later timepoint ( Figure 2B ), indicating that C3 deficiency leads to larger, unresolving necrotic areas in the liver. During the liver regeneration phase, cellular proliferation can be estimated by the expression of Ki67 in liver cryosections. Using this approach, we observed a significant decrease in cell proliferation in C3^-/- mice 48h after APAP overdose ( Figure 2C ), confirming that the absence of C3 impairs liver regeneration and recovery from injury.

To directly assess whether C3 deficiency affects the amount of necrotic debris in injured tissues, we performed confocal IVM of mouse livers. Considering that DNA is abundantly released by necrotic hepatocytes (16) and based on the observed differences during the resolution phase ( Figures 2A-C ), we chose to measure the amount of DNA exposed in the liver 48h after the APAP challenge using the membrane-impermeable DNA dye Sytox Green. Interestingly, WT mice presented minimal extracellular DNA in the liver at the 48h timepoint, which is consistent with the removal of necrotic debris and tissue recovery at that phase ( Figures 2D, E ). Moreover, the vast majority of the fluorescent signal observed in the images of WT mice consisted of background fluorescence from healthy hepatocyte nuclei ( Figure 2D ). In contrast, C3^-/- mice had significantly more extracellular DNA debris, demonstrating that these mice have a clear defect in the removal of necrotic debris from the liver ( Figures 2D, E ). These results link poor recovery from liver injury in C3^-/- mice to impaired clearance of necrotic cell debris.

To explore how debris persisted in injury sites, an analysis of the recruited leukocytes and their ability to take up debris was performed. The inflammatory response triggered by liver injury led to the recruitment of CD11b⁺ leukocytes to necrotic areas identified by the extracellular DNA staining ( Figure 3A , upper panels). These cells consisted primarily of inflammatory monocytes (CCR2⁺), neutrophils (Ly6G⁺) and macrophages (F4/80⁺) ( Figure 3A , lower panels; Supplementary Figures S3A-C ). Of interest, CD11b, the α_M subunit of the complement receptor CR3 (CD11b/CD18), which is known for its involvement in complement-mediated phagocytosis, was increased in neutrophils and monocytes during liver injury ( Supplementary Figures S3D-F ). Besides CD11b+ leukocytes, other immune cells (DCs, T and B cells) are present in lower percentages in the injured liver ( Supplementary Figure S3G ). However, their role in debris phagocytosis is less anticipated and therefore not examined in this study. We then inquired whether CD11b⁺ cells were able to internalize extracellular DNA debris using IVM. Numerous CD11b⁺ leukocytes were visualized deep within necrotic areas using Z-stacks, and multiple cells contained Sytox Green⁺ particles ( Figure 3B , Supplementary Video S1 ). In total, 12% of the CD11b⁺ cells contained DNA particles, with an average of 2 particles per cell ( Figures 3C, D ). Internalization of DNA debris was confirmed upon intravenous administration of DNAse to remove the bulk of extracellular DNA in necrotic areas. DNA-positive vesicles in CD11b⁺ leukocytes remained after DNAse injection, indicating that the DNA particles were located intracellularly, likely in phagosomes, which are not accessible to the circulating DNAse treatment ( Supplementary Figure S3H , Supplementary Video S2 ).

After observing impaired necrotic DNA removal in C3^-/- mice ( Figures 2D, E ), flow cytometry was performed to quantify DNA uptake by leukocytes in WT and C3^-/- mice (see Supplementary Figure S7 for gating strategy). Again, the membrane-impermeable DNA dye Sytox Green was injected intravenously 2h before sacrificing mice that received an APAP overdose 24h prior. This allowed sufficient time for the fluorescently-labeled necrotic DNA to be phagocytosed. We observed that a significantly lower percentage of neutrophils and macrophages were able to internalize DNA debris in the absence of C3, demonstrating that debris clearance depends at least partially on complement opsonization ( Figures 3E, F ). Importantly, no differences were observed in the number of neutrophils (Ly6G⁺) and macrophages (Ly6G^-/Ly6C^-/F4/80⁺) present in the injured liver between WT and C3^-/- mice ( Figures 3I, J ), suggesting that impaired DNA removal observed in C3^-/- mice may be due to the reduced phagocytic capacity of these cells. In contrast, classical and non-classical monocytes did not require C3 to take up DNA debris in the injured liver ( Figures 3G, H ), suggesting that different leukocyte populations utilize distinct mechanisms for debris clearance. However, the number of classical (Ly6G^-/Ly6C⁺) and non-classical monocytes (Ly6C^-/CX₃CR1⁺) recruited to the injured liver in C3^-/- mice was significantly reduced ( Figures 3K, L ), showing that less monocytes reached the injured liver to perform debris phagocytosis. Overall, these data show that necrotic DNA debris is cleared by leukocytes in the liver and that neutrophils and macrophages require complement opsonization of debris for its uptake.

Reduced clearance of DNA debris in the absence of C3 was observed in a model of acute liver injury induced by APAP overdose. To validate whether this finding applies to other types of injuries, we investigated debris clearance in a model of focal thermal injury (FTI) of the liver. In this model, necrotic lesions are induced locally with a hot needle, facilitating the observation of phagocytosis in vivo. This was challenging in the APAP model due to widespread necrosis throughout the liver causing an abundance of debris. The localized nature of the thermal injury allowed us to label necrotic debris by applying a droplet of the pH-sensitive dye pHRodo Red succinimidyl ester on top of the lesion. This dye binds covalently to proteins and exhibits increased fluorescence when the ingested material is processed in the acidic environment of a phagolysosome.

Our previous work demonstrated that neutrophils predominated in the injured area 6h after FTI, with monocytes being attracted after 12h (23). The entire image of the FTI shows distinct burn injury zones, with neutrophils mostly accumulating around the injury core ( Supplementary Figure S4A ). Using IVM, neutrophils carrying pHRodo-labeled debris were observed crawling from the injury border into the necrotic core ( Figure 4A , Supplementary Video S3 ). Approximately 75% of neutrophils at the injury site had pHRodo-containing phagosomes, whereas less than 5% of neutrophils in healthy areas of the liver phagocytosed debris ( Figure 4B ). This finding was confirmed by flow cytometry, which showed a significantly increased MFI of pHRodo in neutrophils and monocytes at the burn injury site compared to leukocytes in healthy areas ( Supplementary Figures S4B, C ; see Supplementary Figure S7 for gating strategy). Similarly to the APAP model, complement proteins C1q and (i)C3b were specifically deposited on necrotic lesions 6h after FTI ( Figure 4C ). Both components colocalized with each other and fibrin ( Supplementary Figures S4D, E ). Quantification of necrotic debris clearance by flow cytometry demonstrated a significant decrease in the percentage of neutrophils and non-classical monocytes phagocytosing debris in C3^-/- mice compared to WT ( Figures 4D, E ). Interestingly, this phenomenon was not observed in classical monocytes nor macrophages ( Figures 4F, G ). To be noted, the number of neutrophils migrating in the burn injury site ( Supplementary Figure S4F ), and the percentage of neutrophils, macrophages and non-classical monocytes attracted to the injured liver did not differ between WT and C3^-/- mice, while the percentage of classical monocytes was significantly reduced in C3^-/- mice ( Supplementary Figures S4G-J ).

Due to the increase in CD11b⁺ cells in response to liver injury ( Supplementary Figures S3D-F ) and its known role in complement-mediated phagocytosis, the role of CR3 on debris uptake and liver resolution was investigated. In the FTI, using CD11b^-/- mice, we found a significant decrease in debris clearance by neutrophils and non-classical monocytes ( Figures 4D, E ), while no difference was observed in classical monocytes and macrophages ( Figures 4F, G ). Moreover, the attracted phagocyte populations at the injury site were not affected by the deficiency in CD11b^-/-, indicating that the defective clearance is not connected to inhibition of leukocyte recruitment ( Supplementary Figures S4F-J ). Conversely, no significant effect was observed on the progression of liver injury in CD11b^-/- mice or in WT mice that received a CD11b blocking antibody, as evidenced by similar ALT values and fibrin staining ( Supplementary Figures S5A-E ) after APAP overdose. These data show that in the FTI, neutrophils and monocytes migrate to the necrotic lesion to phagocytose necrotic debris, a process which depends on C3 and CD11b/CD18 for neutrophils and non-classical CX₃CR1⁺ monocytes, whereas classical CCR2⁺ monocytes rely on other unidentified receptors.

Investigating the factors driving debris phagocytosis in vivo is challenging due to the necessity of multiple knock-out strains and the technical limitations associated with observing cellular events in living mice. To overcome this, we developed an in vitro phagocytosis assay, enabling us to study the role of specific complement proteins in necrotic debris clearance. This approach also allowed us to verify our findings in human neutrophils and human necrotic debris. In this assay, necrosis was induced by mechanically disrupting HepG2 cells, whereafter sera lacking specific complement components were used to opsonize the debris. Images were taken with a confocal microscope 3h after combining the opsonized necrotic debris with human or mouse neutrophils. Importantly, the HepG2 cell debris itself did not contain detectable levels of C3/(i)C3b, as demonstrated by immunostaining HepG2 debris in vitro ( Supplementary Figures S6A, B ). However, when the debris came in contact with whole serum, it became clearly opsonized by C3/(i)C3b confirming the capacity of debris to induce complement activation.

In vitro engulfment of pHRodo-labeled necrotic debris by live neutrophils was visualized by 3D reconstruction ( Figure 5A ). Interestingly, the percentage of bone-marrow derived mouse neutrophils that performed phagocytosis did not differ when presented with debris opsonized with serum, serum lacking C3 (from C3^-/- mice) or lacking antibodies (from Rag2^-/- mice) ( Figure 5B ). However, the volume of the necrotic debris internalized by neutrophils was significantly reduced when the debris was opsonized with serum lacking C3 ( Figure 5C ). Likewise, freshly isolated neutrophils from healthy donors showed no difference in phagocytosis rates when debris was opsonized with serum, C1q-depleted serum or C3-depleted serum ( Figure 5D ). Nevertheless, a significant decrease in the volume of debris uptake was observed when it was opsonized with serum lacking C3 ( Figure 5E ). Latrunculin served as a positive control, as it inhibits phagocytosis globally by disrupting actin polymerization. These findings underscore the ability of both human and mouse neutrophils to internalize necrotic debris, while highlighting the role of complement on the amount of debris taken up through phagocytosis.

To investigate the impact of debris phagocytosis on gene expression, qPCR was performed on human neutrophils incubated with necrotic debris from HepG2 cells. Neutrophils were exposed to pure unopsonized debris, debris opsonized with normal serum or with C3-depleted serum for 3 hours. The data were normalized to unopsonized debris in order to account for stimulation by DAMPs present in necrotic cell debris. Interestingly, uptake of serum-opsonized debris induced the upregulation of PTGS2 (encoding COX2) in neutrophils ( Figure 5F ). COX2 is an enzyme with dual role in inflammation, catalyzing the production of pro-inflammatory prostaglandins from arachidonic acid, such as PGE2, but also participating in the synthesis of numerous pro-resolving lipid mediators (27). PTGS2 upregulation was reversed when the debris was opsonized with serum lacking C3, indicating a direct effect of complement on the upregulation of COX2, which was already observed for monocytes but not neutrophils (28). CXCR2, coding for a major chemokine receptor in neutrophils that promotes both chemotaxis and reverse migration (29, 30) was also upregulated by incubation with opsonized debris. Moreover, in the absence of C3, CXCR2 expression levels returned to baseline ( Figure 5G ). In addition, incubation with opsonized debris led to the expression of other immunoregulatory and pro-resolving genes in neutrophils, namely, CXCR4, encoding a chemokine receptor associated with homing of neutrophils to the bone marrow for apoptosis and removal (31, 32); the immunoregulatory cytokine IL10, and ANXA1, encoding the protein annexin A1 that dampens leukocyte chemotaxis, respiratory burst and phagocytosis ( Figures 5H-J ) (33). The absence of C3 in the serum significantly reduced the expression of CXCR4, IL10 or ANXA1, indicating the essential role of C3 opsonization in the induction of pro-resolving genes in neutrophils. Moreover, alterations in gene expression in neutrophils are specific, since multiple genes were unaffected by stimulation with opsonized debris, including ALOX5, CYBB, CASP3, ARG1, FPR1 and FPR2 ( Supplementary Figures S6C-H ). Overall, the gene expression induced by the clearance of opsonized necrotic debris reflects a pro-resolving response in neutrophils, which is C3-dependent and plays a central in promoting tissue repair.

Lastly, we investigated whether the classical complement pathway contributed to (i)C3b opsonization of necrotic debris. For this, Rag2^-/- mice, which lack mature T and B cells and therefore also antibodies, were subjected to the APAP overdose. Activation of the classical complement pathway requires the association of C1q with target-bound IgM or multiple IgGs, thus, this complement pathway cannot be activated in Rag2^-/- mice. First, the absence of IgM was confirmed by immunostaining, which revealed that IgM labeling in the injured Rag2^-/- liver was essentially absent ( Figures 6A, D, E ). Interestingly, 24h after APAP overdose, liver cryosections showed significantly increased fibrin deposition in Rag2^-/- mice compared to WT mice ( Figures 6B, C ). In our previous work, the absence of antibodies was shown to be responsible for a delayed liver resolution due to impaired necrotic debris clearance (23). Because the degree of necrotic injury also affects the degree of C1q and (i)C3b deposition, complement immunostaining was normalized to the area of fibrin staining. With this approach, we observed no difference in C1q binding to necrotic areas in Rag2^-/- mice compared to WT ( Figure 6F ). Similarly, the mean fluorescence intensity (MFI) of C1q was not different between WT and Rag2^-/- mice, indicating that C1q binding does not rely on antibody opsonization of necrotic debris ( Figure 6G ). However, the area of (i)C3b deposition was significantly decreased in Rag2^-/- mice, with also significantly lower MFI in (i)C3b-stained areas ( Figures 6H, I ), even though it has been shown that Rag2^-/- mice have elevated C3 levels in blood (34). These data indicate that loss of IgM and IgG antibodies diminishes the level of complement activation on necrotic sites, even though C1q binding remains unaffected. These data also demonstrate that the classical complement pathway is activated in response to necrotic debris, leading to (i)C3b deposition in injury sites.