An unexpected role of neutrophils in clearing apoptotic hepatocytes in vivo

Billions of apoptotic cells are removed daily in a human adult by professional phagocytes (e.g. macrophages) and neighboring nonprofessional phagocytes (e.g. stromal cells). Despite being a type of professional phagocyte, neutrophils are thought to be excluded from apoptotic sites to avoid tissue inflammation. Here, we report a fundamental and unexpected role of neutrophils as the predominant phagocyte responsible for the clearance of apoptotic hepatic cells in the steady state. In contrast to the engulfment of dead cells by macrophages, neutrophils burrowed directly into apoptotic hepatocytes, a process we term perforocytosis, and ingested the effete cells from the inside. The depletion of neutrophils caused defective removal of apoptotic bodies, induced tissue injury in the mouse liver, and led to the generation of autoantibodies. Human autoimmune liver disease showed similar defects in the neutrophil-mediated clearance of apoptotic hepatic cells. Hence, neutrophils possess a specialized immunologically silent mechanism for the clearance of apoptotic hepatocytes through perforocytosis, and defects in this key housekeeping function of neutrophils contribute to the genesis of autoimmune liver disease.


Introduction
Apoptosis is a process of programmed cell death that clears aged or damaged cells to maintain internal tissue homeostasis (Hochreiter-Hufford and Ravichandran, 2013). An estimated hundred billion cells undergo apoptosis daily in a human adult (Fond and Ravichandran, 2016). These apoptotic cells must be disposed of promptly and efficiently without littering cytoplasm or causing inflammation (Franz et al., 2006). Defects in the clearance of apoptotic bodies are often linked to various inflammatory and autoimmune diseases (Nagata et al., 2010;Poon et al., 2014). It is widely believed that apoptotic bodies are cleared by professional phagocytes, such as macrophages and immature dendritic cells, or by local nonprofessional phagocytes, such as epithelial cells, endothelial cells, and fibroblasts (Lauber et al., 2004). Although the morphology of apoptotic cells is easily distinguished in histological sections (Kerr et al., 1972), these cells are infrequently observed in normal human samples due to efficient removal by phagocytes (Lauber et al., 2004;Poon et al., 2014). Hence, the phagocytes responsible for the removal of apoptotic cells in the homeostatic state remain uncertain and it is unknown whether they are tissue specific.
Neutrophils, a type of terminally differentiated and short-living phagocytic cell, represent 50-70% of the total white blood cell population in humans (Amulic et al., 2012;Kolaczkowska and Kubes, 2013). However, neutrophils are not considered key players in apoptotic cell clearance (Nagata et al., 2010;Poon et al., 2014). They instead function as inflammatory cells responsible for killing bacteria and fighting infection. Blood neutrophils often swarm to the site of infection, where they release toxic mediators (e.g., reactive oxygen species and proteases) that not only kill microorganisms but also can damage tissues (Amulic et al., 2012;Kolaczkowska and Kubes, 2013). To prevent inflammation, apoptotic cells have been shown to release 'keep-out' signals, such as lactoferrin, that prevent the recruitment of neutrophils to the apoptotic site (Bournazou et al., 2009). Nevertheless, neutrophils are related to multiple autoimmune diseases (Amulic et al., 2012;Kolaczkowska and Kubes, 2013). For example, mild neutropenia was observed to precede and accompany the onset of type 1 diabetes, a typical autoimmune disease (Harsunen et al., 2013;Valle et al., 2013).
The role of neutrophils in the clearance of apoptotic bodies and contribute to autoimmunity remain to be solved (Jorch and Kubes, 2017).
Here we report an unexpected role of neutrophils in clearing apoptotic hepatocytes under physiological conditions. We found that neutrophils, not macrophages or other cells, are responsible for apoptotic hepatocyte clearance in an immunologically silent manner. We noted that neutrophils penetrate apoptotic hepatocytes and clear them from the inside, thereby avoiding spillage 6 of cytoplasmic content; this mechanism differs sharply from the classical engulfment of dead cells by macrophages. Defects in neutrophil-mediated apoptotic clearance have been observed in human autoimmune liver (AIL) disease. Hence, in addition to their well-known role in combating infection, neutrophils can function as housekeepers for apoptotic clearance and thus maintain tissue homeostasis.

Neutrophils burrow into apoptotic hepatocytes in human livers
We observed a large number of neutrophils in hepatocytes in human noncancerous liver tissue obtained from patients with hepatocellular carcinoma (Fig. 1A) or from patients with hepatic hemangioma (Fig. 1B). We discovered that the hepatocytes occupied by neutrophils were apoptotic, as evidenced by the condensed chromatin signature (Fig. 1A, panels i-vi, black arrowheads indicate condensed chromatin, white arrowheads point to neutrophils with a characteristic multilobed nucleus). Importantly, apoptotic hepatocytes are rarely observed in the human liver, possibly due to the rapid removal of apoptotic cells by phagocytes. In a total of 281 apoptotic hepatocytes from 32 livers, we observed that each apoptotic hepatocyte was engorged by up to 22 neutrophils ( Fig. 1C, Table S1). We also confirmed apoptotic hepatocytes by TUNEL staining or Caspase-3 immunostaining (Fig. 1D). Neutrophils burrowed inside apoptotic hepatocytes were either stained with an antibody against neutrophil elastase (NE), myeloperoxidase (MPO) or recognized by their multilobed nucleus signature (Fig. 1D, S1A). The distances from burrowed neutrophils to the apoptotic hepatocyte border were analyzed by IMARIS software and recorded in Table S2. Other immune cells were rarely associated with apoptotic hepatocytes.
Based on the above observations that neutrophils are the predominant phagocyte associated with apoptotic hepatocytes, we hypothesized that the neutrophil-mediated clearance of apoptotic cells consists of the following three sequential steps. In the initial invading or burrowing stage, activated neutrophils identified and targeted the apoptotic hepatocytes (cells with condensed chromatin, indicated by black arrowheads) and attached to their cell membrane (Fig. 1A, panel i, outlined by black rectangle). Then, the neutrophils invaded apoptotic hepatocytes (Fig. 1A, panels ii and iii). We observed an average of 7 neutrophils entering each apoptotic hepatocyte, and we termed this process perforocytosis (from Latin perfero to bore) (Fig. 1A, white arrowheads point to neutrophils). The second step consisted of phagocytosis and detachment.
The neutrophils within hepatocytes appeared to clear apoptotic bodies from the inside without destroying the cellular membrane or extruding the cytoplasm (two burrowed neutrophils phagocytosed apoptotic debris as shown in Fig. S1C).
Following digestion by neutrophils, the apoptotic hepatocytes decreased in size, 9 and detached from nearby hepatocytes (Fig. 1A, panels iv-vi). The third step involved the complete digestion of apoptotic hepatocytes (Fig. 1A, panels vii-ix).
After the clearance of apoptotic hepatocytes, neutrophils seemed to migrate away from the cleared space, possibly to make room for new hepatocytes generated by rapid division. This 'eating inside' phagocytosis process differed sharply from the well-known engulfment of apoptotic cells or fragmented apoptotic bodies by other phagocytic cells, such as macrophages (Poon et al., 2014).

Visualization of neutrophil perforocytosis in mouse livers
Neutrophils within apoptotic hepatocytes were confirmed in mouse livers by intravital microscopy ( Fig. 2A, B) and electron microscopy (Fig. 2C). Similar to the observations in human samples, apoptotic hepatocytes in WT mouse livers were occupied by neutrophils but not associated with macrophages ( Fig. 2A, neutrophils were labeled with an i.v. injection of anti-Ly6G antibody, macrophages were labeled with anti-F4/80 antibody and apoptotic cells were labeled with Annexin V). A total of 24 apoptotic cells were observed in 8 WT livers with an average of 2 burrowed neutrophils in each mouse apoptotic hepatocyte. The burrowed neutrophils were projected and analyzed by IMARIS software and results were compared side by side with livers from MRP8cre/DTR 10 mice ( Fig. 2A, with more details in the neutrophil depletion section below). The distances of burrowed neutrophils to the border of apoptotic hepatocytes were recorded in Table S6.
By using the Cellvizio System (Confocal Miniprobes), we managed to visualize the entire process of neutrophil perforocytosis in mouse livers ( Fig. 2B and Video S1, S2, apoptotic hepatocytes were labeled with Annexin V and neutrophils were labeled with anti-Ly6G antibody). Consistent with observations in human samples, Annexin V positive apoptotic hepatocytes were burrowed and cleared by Ly6G-labeled neutrophils (Fig. 2B). This eating inside phagocytic process is fast and rigorous in which neutrophils were able to completely digest apoptotic hepatocytes around 4-7 minutes (Fig. 2B, and Video S1, S2, a total of 13 apoptotic hepatocytes in 12 WT mouse livers were observed).
At the end of the apoptotic clearance process, neutrophils simply left the apoptotic sites and were not labeled by Annexin V, indicating neutrophils were not apoptotic.

The selectivity of neutrophil perforocytosis of effete hepatocytes
To study neutrophil-mediated phagocytosis in live cells in vitro, we induced apoptosis of isolated human primary hepatocytes or mouse liver NCTC cells with puromycin ( Fig. S2A) and then applied human neutrophils or neutrophil-like, differentiated HL60 cells (Fig. 3). Isolated human primary hepatocytes formed a hepatic plate-like structure around 7 days in the culture dishes and were further confirmed with anti-albumin antibody (Fig. 3A).
Human blood neutrophils (labeled with a red membrane dye, PKH-26) were able to penetrate and burrow inside the human apoptotic hepatocytes induced by puromycin (Fig. 3B, white arrowheads point to burrowing neutrophils). These apoptotic hepatocytes had markedly decrease in size after neutrophil burrowing as compared with cells without neutrophil burrowing (Fig. 3C, D, Video S3 shows a human neutrophil first burrowed and then started to phagocytose an apoptotic hepatocyte from inside in real time). Similar results were observed with NCTC cells (labeled with a red membrane dye, PKH-26) and HL60 cells (labeled with a green membrane dye, PKH-67).  showed little response to normal NCTC cells (Fig. 3E, top row). In the presence of apoptotic NCTC cells, however, green-labeled HL60 cells polarized and penetrated dead cells (Fig. 3E, white arrowheads point to burrowing neutrophils, bottom row). Next, we quantified phagocytosis by flow cytometry using a pH sensitive dye, PHrodo (with weak fluorescent at neutral pH but high fluorescent with a drop in pH to measure engulfment, see Methods). The phagocytosis of apoptotic hepatocytes by neutrophils was markedly greater than that of normal nonapoptotic hepatocytes (Fig. 3F, G). There was little cytoplasm or DNA leakage during neutrophils phagocytosing apoptotic NCTC cells as assessed by the extracellular levels of DNA, SOD and ROS ( Fig. S1D-F). To address whether neutrophils also phagocytose other apoptotic cells, we determined the ability of neutrophils to phagocytose apoptotic endothelial HUVECs or epithelial HEK293 cells. We did not observe phagocytosis of these cells compared with their nonapoptotic controls (Fig. 3G), indicating the selectivity of neutrophils in phagocytosing apoptotic hepatocytes. Upon comparing the ability of HL60 cells and macrophage U937 cells to phagocytose apoptotic hepatocytes, we found that U937 cells exhibited much lower phagocytosis of apoptotic NCTC cells than did HL60 cells (Fig. S2C). Similar results were observed with human and mouse macrophages ( Fig. S2B, C), consistent with the central role of neutrophils in mediating apoptotic hepatocyte clearance.

The signals that attract neutrophils to apoptotic hepatocytes
To identify the signals that attract neutrophils to apoptotic hepatocytes, we screened for cytokine and chemokine secreted by NCTC cells before and after apoptosis. Apoptotic NCTC cells showed markedly increased secretion of the cytokines IL-1β, IL-6, IL-8, and IL-12 compared with nonapoptotic controls (Fig.   4A-D). In contrast, GM-CSF, IFN-γ, TNF-α, IL-2 and IL-10 were not significantly changed (Fig. S3A). To address the role of the upregulated cytokines, we 13 knocked down the receptors of IL-1β, IL-6, IL-8, or IL-12 in HL60 cells and then examined phagocytosis ability. Knockdown of the IL-1β receptor abolished the phagocytosis of apoptotic NCTC cells (Fig. 4E), whereas knockdown of the IL-8 receptor showed a 50% reduction in phagocytosis (Fig. 4E). Thus, neutrophil chemoattractants, IL-1β and IL-8 secreted by apoptotic NCTC cells are key signals for attracting neutrophils and inducing subsequent phagocytosis.
Consistent with above observations, we found endogenous IL-1β associated with apoptotic hepatocytes but not with normal hepatocytes in human liver samples ( Fig. 4F).
Next, we examined the cell surface molecules in hepatocytes and found that apoptotic NCTC cells had markedly increased P-selectin, E-selection, L-selectin and PECAM as compared with non-apoptotic controls (Fig. 4G). Inhibition of selectins (with a pan-selectin antagonist, bimosiamose) but not PECAM (with a function blocking antibody) abrogated the neutrophil-mediated clearance of apoptotic hepatocytes (Fig. 4E), indicating selectins play a critical role during this process.

Neutrophil depletion impairs the clearance of apoptotic hepatocytes
We next examined whether neutrophil depletion with either an antibody or genetic methods (see Methods) influences the clearance of apoptotic hepatocytes. Antibody or genetic depletion yielded about 90% or 70% reduction in mouse peripheral blood neutrophils, respectively (Fig. S4A, B). We analyzed liver samples from neutrophil-depleted and control mice by intravital microscopy ( Fig.   2A, genetic depletion), or immunostaining (Fig. 5, antibody depletion). Similar to human liver samples, apoptotic hepatic cells in control WT mice were occupied and phagocytosed by neutrophils ( Fig. 2A, 5A, B). After neutrophil depletion, the apoptotic bodies were no longer associated with neutrophils ( Fig.   2A, 5A, B). However, we observed that macrophages were associated with apoptotic hepatocytes in neutrophil-depleted samples ( Fig. 2A, 5A, C), suggesting a compensatory role of macrophages in phagocytosing dead hepatocytes in the absence of neutrophils. The percentage of apoptotic hepatocytes in neutrophil-depleted samples was significantly increased compared to that in controls (0.92% VS 0.2%, p < 0.001, Fig. 5D). Hence, neutrophil depletion impaired the prompt clearance of apoptotic cells in the mouse liver. To rule out the effects of environmental microbes after neutrophil depletion, we also treated the neutrophil-depleted mice with antibiotic (20 mg/kg ampicillin, i.p. injected daily). We obtained similar results in neutrophildepleted mice with or without antibiotic treatment ( Fig. 5A-D). Meanwhile, we observed impaired liver function in neutrophil-depleted livers (with increased aspartate aminotransferase and alanine aminotransferase activity and increased 15 total or direct bilirubin level) compared with nontreated controls (Fig. 6A-D, antibody depletion). Depletion of neutrophils had little effect on other tissues (e.g. kidney, Fig. S4C, D).

Defective neutrophil perforocytosis in autoimmune liver disease
Autoimmune diseases are often linked to defective clearance of apoptotic cells (Nagata et al., 2010;Poon et al., 2014). We first determined whether defective neutrophil-mediated apoptotic clearance contributes to AIL disease. In the present study, we detected an increase in autoantibodies (i.e., against antinuclear antigen, smooth muscle actin, liver-kidney microsome and total IgG antibodies) in neutrophil-depleted mice compared with controls ( Fig. 6E-H, antibody depletion). This increase was not affected by the treatment of antibiotic and was not observed in macrophage-depleted mice ( Fig. 6E-H, macrophages were depleted with clodronate-liposome, see Methods). Hence, neutrophil depletion not only impaired apoptotic hepatocyte clearance but also led to the generation of autoantibodies, suggesting a role of defective neutrophil-mediated removal of apoptotic bodies in AIL disease.
Next, to address whether neutrophil-mediated apoptotic clearance is impaired in AIL disease, we analyzed biopsy samples from patients diagnosed with AIL disease. In contrast to the normal human controls, a total of 22 AIL disease patient samples contained apoptotic hepatocytes that were not phagocytosed or invaded by neutrophils (Fig. 7A). More apoptotic bodies were associated with macrophages, as observed in neutrophil-depleted mouse livers ( Fig. 7A, Table S7). Since the blood neutrophil count in AIL disease patients is within the normal range, we surmised potential defects in the phagocytosis ability of neutrophils from AIL disease patients. We observed markedly decreased phagocytosis of apoptotic NCTC cells by neutrophils from AIL disease patients compared to normal controls (Fig. 7B). Normal human neutrophils burrowed into apoptotic NCTC cells and demonstrated perforocytosis, while AIL disease neutrophils exhibited little response towards apoptotic NCTC cells (Fig.   7C). We further screened differential gene expression between normal human neutrophils and AIL neutrophils. We noted that the expression of IL-1β receptor, IL1R1 and selectin binding protein, P-selectin glycoprotein ligand 1 (PSGL-1) was markedly decreased in AIL neutrophils as compared with normal human neutrophils ( Fig. 7D), which further confirmed the critical roles of IL-1β and selectins in neutrophil-mediated apoptotic clearance. The above data prove defective neutrophil-mediated apoptotic clearance in human AIL disease samples.

Discussion
Our finding of neutrophil burrowing into and clearing of apoptotic hepatocytes under physiological conditions reveals a fundamental mechanism for the removal of effete hepatocytes, helps to solve some mysteries surrounding the apoptotic clearance process and raises several important questions.

Neutrophils as the new scavengers for apoptotic clearance
Although billions of cells undergo apoptosis daily in our bodies, apoptotic cells are rarely observed in tissues under steady state due to the high efficiency of apoptotic removal by both professional and neighboring nonprofessional phagocytes (Poon et al., 2014). In general, professional phagocytes have much higher phagocytic efficiency and capacity than nonprofessional phagocytes, but they are greatly outnumbered by other cell types in tissues (Elliott and Ravichandran, 2016). Therefore, whether there are unidentified scavengers and/or mechanisms for the prompt removal of dead cells in distinct organs remains unclear. Compared to macrophages or their precursors, monocytes, which make up less than 10% of the total white blood cell population, neutrophils are the most abundant white blood cells (up to 70%) in humans, and thus ideally suited for the swift clearance of apoptotic hepatic cells. As macrophages were induced to phagocytose apoptotic hepatocytes following the depletion of circulating neutrophils, our data show compensatory activation of other phagocytes under these conditions. The basis for this switch, however, is not clear.
Another important question is whether other tissues utilize a similar mechanism. As neutrophils do not phagocytose apoptotic epithelial and endothelial cells, it is possible that distinct signals are specifically released by different types of apoptotic cells. It remains largely unknown whether one or more signals are dominant in each specific tissue and how these signals are orchestrated into a complicated network to swiftly clear dying cells while avoiding tissue inflammation.

Neutrophil influx without causing tissue damage
Although apoptotic clearance has been considered an immunologically silent process that does not lead to the influx of inflammatory cells or exposure of self-antigens (Poon et al., 2014;Savill et al., 2002), it has also been suggested that this process is not completely immunologically silent (Green et al., 2009).
Because neutrophils are proinflammatory cells, they have long been thought to be excluded from apoptotic sites. We observed neutrophil influx into apoptotic hepatic cells (up to 22 neutrophils in one apoptotic hepatocyte), and there was little detectable tissue injury or other inflammatory cells. One reason could be that the neutrophil attraction signals released by apoptotic hepatocytes (e.g., IL-1β and IL-8) are sufficient for attracting neutrophils towards apoptotic cells and inducing subsequent perforocytosis, but do not elicit neutrophil inflammatory functions such as oxidative burst. Therefore, neutrophil recruitment into tissues does not always induce inflammation or tissue damage. In consistent with our observations, neutrophils can express both pro-and anti-inflammatory cytokines (Gideon et al., 2019;Mortaz et al., 2018). In contrast, neutrophil depletion caused defective removal of apoptotic bodies and induced autoantibody generation; thus neutrophils play a vital role in the genesis of AIL disease. Considering that billions of neutrophils patrol tissues under physiological circumstances without causing inflammation(Nicolas-Avila et al., 2017), we concluded that these cells not only provide immune surveillance against infection but also contribute to internal tissue homeostasis, as we report in this study.

Neutrophil burrowing to maintain tissue integrity
One intriguing question regarding apoptotic clearance is how the tissue maintains integrity while dead cells are being continually removed (Poon et al., 2014). Apoptotic cell detachment from the extracellular matrix by caspasemediated cleavage or extrusion into the organ lumen has been proposed as the mechanism for maintaining the epithelial barrier during the clearance of dying 20 epithelial cells (Brancolini et al., 1997;Rosenblatt et al., 2001). Our finding that burrowed neutrophils phagocytose apoptotic hepatocytes from the intracellular space provides a novel mechanism for maintaining liver integrity. The neutrophils that phagocytose cells from the inside are efficient at disposing of apoptotic bodies without extruding the cytoplasm and may help to prevent the leakage of toxic bile acids or the release of danger signals that cause tissue damage.
Neutrophil burrowing into other types of cells has also been reported previously in processes not related to apoptotic clearance (Overholtzer and Brugge, 2008). For example, neutrophils can bore into endothelial cells or megakaryocytes to temporarily form so-called cell-in-cell (or emperipolesis, entosis) structures (Overholtzer and Brugge, 2008). The apparent difference is that both the endothelial cells and megakaryocytes entered by neutrophils are viable and nonapoptotic. The purpose of forming cell-in-cell structures with neutrophils and endothelial cells or megakaryocytes is to obtain a passage out of blood vessels or bone marrow, respectively (Overholtzer and Brugge, 2008). It will be interesting to examine whether neutrophils utilize similar invasion mechanisms during these processes.
The important apoptotic clearance function of neutrophils described in this study adds to the repertoire of other known functions of neutrophils (Amulic et 21 al., 2012;Kolaczkowska and Kubes, 2013;Wang et al., 2017). Since the failure to clear dead cells is linked to inflammatory and autoimmune diseases, the critical role of neutrophil-mediated apoptotic clearance may have implications for the pathogenesis and treatment of these diseases.

Cell culture, transfection, and isolation of human and mouse neutrophils.
NCTC, HEK-293, U937 cells and HUVECs were cultured in DMEM supplemented with 10% fetal bovine serum and HL60 cells were cultured in RPMI 1640 with 10% fetal bovine serum. For experiments, HL60 cells were differentiated by adding 1.3% DMSO into the medium for 7 days (Xu et al., 2003).
To establish the stable RNAi cell lines, we transfected shRNA into HEK293T cells. After virus packaging in HEK293T cells, HL60 cells were infected, screened by puromycin and sorted by flow cytometry (Liu et al., 2012). For the isolation of human neutrophils (Liu et al., 2015), blood was collected from healthy human donors or those with AIL disease. Erythrocytes were removed using dextran sedimentation (4.5% dextran) followed by hypotonic lysis using distilled water. Neutrophils were isolated from the resulting cell suspension using discontinuous Percoll gradient centrifugation. This procedure yielded >95% neutrophil purity and >95% viability as assessed by flow cytometry.
For the isolation of human primary hepatocytes, liver tissues from a fresh liver hemangioma surgery were suspended in DPBS and washed for 5-10 times.
Surrounding connective tissues and adipose tissues were removed with a surgery knife and remaining liver tissues were further sliced into small fragments (around 1 mm 2 ), then washed with DPBS for three times and centrifuged at 800 rpm for 3 min. Tissue samples were resuspended in 5 times volume of collagenase IV medium and incubated at 37℃ for 30 min. After collagenase digestion, samples were centrifuged at 1000 rpm for 5min, and  (Chow et al., 1995). A total of 1 × 10 6 NCTC cells were plated in a 35 mm glass-bottom dish (MatTek) for 10 h, and then treated with puromycin (2.5 μg/ml) for 12 h.

Apoptosis was assayed with an Alexa Fluor 488 Annexin V/Dead Cell Apoptosis
Kit (Thermo Fisher) and analyzed by flow cytometry to measure the fluorescence emission at 530 nm and 575 nm (or equivalent) using 488 nm excitation. Representative data are shown in the paper. An independent experiments twotailed Student's t-test was used as the statistical assay for comparisons.
Significant differences between samples were indicated by P < 0.05.            Figure 1D) are remodeled and analyzed with IMARIS software: anti-E-Cadherin blue staining is set as the source channel to detect hepatic cell boundary, and anti-NE red staining is set as the source channel to detect neutrophils inside blue cells. Only neutrophils inside hepatic cells can be detected. The cell positions of both neutrophils and hepatocytes, and the distances from neutrophils to the hepatocyte border are recorded below (a total of 8 apoptotic hepatocytes are analyzed).

Apoptotic hepatocytes
Hepatocyte position X,Y,Z     Figure 2A) are remodeled and analyzed with IMARIS software: Annexin V blue staining is set as the source channel to detect hepatic cell boundary, and Ly-6G green staining is set as the source channel to detect neutrophils inside blue cells. Only neutrophils inside hepatic cells are detected and calculated. The cell positions of both neutrophils and hepatocytes, and the distances from neutrophils to the hepatocyte border are recorded below (a total of 10 apoptotic hepatocytes are analyzed). Video S1, S2: Mouse neutrophils burrowed and digested apoptotic hepatocytes in the mouse liver. Neutrophils were labeled with an i.v. injection of anti-Ly6G antibody (green) and apoptotic cells were labeled with Annexin V (red). Time-lapse images were acquired by the Cellvizio system. The cytoplasm of apoptotic hepatocyte with neutrophil burrowing quickly dwindled and finally disappeared. 2 out of 13 apoptotic hepatocytes are shown in video S1 and S2.
Video S3: Human neutrophils burrowed inside apoptotic hepatocytes. A human neutrophil (green, pointed by white arrow) was burrowing into an apoptotic hepatocyte (red). Time-lapse video was recorded for 60 minutes and followed by reconstructed 3D images at 20 and 60 min. At 20 min, the neutrophil partially entered the hepatic cell and did not phagocytose (neutrophil was extracted and analyzed by IMARIS software and no red dye inside the neutrophil).
At 60 min, the neutrophil completely burrowed inside the hepatocyte and started to phagocytose hepatocyte (with uptake of red dye).