Neutrophil infiltration regulates clock-gene expression to organize daily hepatic metabolism

Virtually all cell types have an internal clock that controls their rhythmicity through the periodic expression of clock genes (Robles et al., 2014; Tahara and Shibata, 2016). However, it is unknown how these multiple cell rhythms are integrated. The liver is an essential metabolic organ that controls body glucose and lipid homeostasis (Manieri and Sabio, 2015), and neutrophil infiltration alters its function (González-Terán et al., 2016). We hypothesized that the metabolic cycles in the liver might be entrained by rhythmic neutrophil infiltration. To test this, we harvested liver, BM, and blood from C57BL6J mice at 4 hr intervals over a 24 hr period. Liver neutrophil infiltration showed a clear diurnal pattern, with a peak at ZT2, coinciding with liver-driven lipogenesis in mice (Zhou et al., 2015), and a nadir during the night, at ZT14 (Figure 1A), correlating with lipolysis (Zhou et al., 2015). These oscillations corresponded directly to changes in neutrophil numbers in blood (Figure 1—figure supplement 1A), suggesting that liver infiltration might result from higher neutrophil migration to the liver. We first confirmed that neutrophils were infiltrated in the liver using 3D microscopy. According to published data (Casanova-Acebes et al., 2018), infiltrated neutrophils presented an intrasinusoidal distribution in the liver, different to that observed in the Kupffer cells population (Figure 1B and Figure 1—figure supplement 1B). Then we evaluated whether myeloid chemokines could be involved in circadian neutrophil recruitment into the liver. Analysis of liver lysates indicated that the expression of the hepatocyte-derived neutrophil chemoattractant Cxcl1 (Su et al., 2018) was higher at ZT2 than a ZT14. Moreover, mRNA of Cxcl1 in liver samples showed the same oscillation pattern than infiltrated neutrophils, suggesting that this chemokine may be important in the regulation of the neutrophil diurnal cycle (Figure 1—figure supplement 1C).

The infiltration pattern correlated with liver expression levels of the clock-gene Bmal1, peaking at ZT2 and bottoming at ZT14 (Figure 1C). Infiltration also correlated inversely with the expression of Nr1d2 (encoding Rev-erb β), Per2, and Cry2 (Figure 1C), which are important proteins in the control of circadian rhythms (Reppert and Weaver, 2002), consistent with the feedback loop that controls their expression. Bmal1 is thought to induce lipogenesis (Zhang et al., 2014), whereas Nr1d2 controls lipid metabolism and its reduced expression promotes lipogenesis and steatosis (Delezie et al., 2012; Solt et al., 2012). In agreement with these studies, liver triglycerides were higher at ZT2 than at ZT14 (Figure 1D).

Our results show a correlation between neutrophil infiltration, hepatocyte Bmal1 expression, and lipid metabolism regulation, raising the possibility that neutrophils signal to hepatocytes to modulate the expression of circadian genes. Exposure of mouse hepatocytes in vitro to freshly isolated neutrophils increased hepatocyte expression of the clock genes Bmal1 and Clock. In contrast, no effect was observed upon exposure to T or B lymphocytes, or macrophages, suggesting the existence of a neutrophil-to-hepatocyte communication that controls hepatocyte clock-gene expression (Figure 1E and Figure 1—figure supplement 1D).

We then investigated whether neutrophil elastase (NE), a proteolytic enzyme reported to regulate liver metabolism, could regulate hepatocyte clock genes (Mansuy-Aubert et al., 2013; Talukdar et al., 2012). Exposure to elastase reproduced the same increase in hepatocyte Bmal1 and Clock expression in contrast with another protease that did not affect Bmal1 expression (Figure 1F and Figure 1—figure supplement 1D).

Next, neutrophil-mediated regulation of liver clock-gene expression in vivo was investigated using a previously characterized genetic model of neutrophil deficiency (Dzhagalov et al., 2007; Steimer et al., 2009; Figure 1—figure supplement 1E,F and Figure 1—figure supplement 2A–C). Low hepatic neutrophil infiltration in neutropenic mice correlated with reduced expression of Bmal1 and Clock (Figure 1G) and increased expression of Cry2 and Per2 at ZT2 (Figure 1G). These changes in clock-gene expression were accompanied by lower liver triglyceride levels (Figure 1H). Furthermore, lack of neutrophils perturbed the diurnal rhythmicity in Bmal1, Clock, and Per2 expression in the liver without affecting clock genes in other organs such as the lung, in which there is no correlation between the peak of neutrophil infiltration and Bmal1 expression (Figure 1—figure supplement 2D,E). Our results thus indicate that neutrophils might specifically control the expression of hepatocyte circadian clock genes in steady state.

Chronic jet lag alters liver circadian genes and disrupts liver metabolism (Kettner et al., 2016). Analysis of a mouse model of jet lag revealed complete disruption of the circadian liver neutrophil infiltration with increased hepatic neutrophil infiltration even at ZT14 (Figure 2A). Abolition of rhythmic neutrophil hepatic infiltration under jet lag correlated with increased steatosis and high levels of liver triglycerides (Figure 2B). To evaluate whether the metabolic effect of circadian perturbation was caused by the increased neutrophil infiltration, we exposed neutropenic and control mice to the jet lag protocol (Figure 2—figure supplement 1A,B). Jet lag-induced steatosis was less severe in neutropenic mice (Figure 2C), and disruption of diurnal liver expression of Bmal1 detected in control jet-lagged mice was partially ablated in neutropenic mice (Figure 2D). Similar results were also observed in mice with impaired neutrophil migration such as Cxcr2^MRP8-KO BM transplanted mice (Eash et al., 2010; Mei et al., 2012) and p38γ/δ^Lyzs-KO mice (González-Terán et al., 2016). In both models, the reduction of neutrophil infiltration correlated with decreased levels of liver Bmal1 expression and protection from jet lag-induced steatosis (Figure 2—figure supplement 1C–G). These results are consistent with the role of neutrophils in the control of liver clock genes.

Inflammation plays a key role in the pathogenesis of non-alcoholic fatty liver disease (Tiniakos et al., 2010) and the development of hepatic steatosis is associated with increased liver infiltration by myeloid cells, particularly neutrophils (González-Terán et al., 2016; Mansuy-Aubert et al., 2013; Talukdar et al., 2012; Tiniakos et al., 2010). Two widely used mouse models of hepatic steatosis, high-fat diet (HFD) and methionine-choline-deficient (MCD) diet, increased liver neutrophil infiltration in WT mice at ZT2, ZT14, and ZT18 (Figure 2E,F). Consistent with a neutrophil-to-hepatocyte communication in the regulation of hepatocyte clock genes, the MCD diet enhanced Bmal1 expression and inhibited Cry2 and Per2 expression in control mice, but not in neutropenic mice at ZT2 (Figure 2G). Altered liver clock-gene regulation in neutropenic mice was associated with protection against steatosis and lower liver triglycerides (Figure 2H). To confirm the role of neutrophils in modulating liver clock genes, we depleted neutrophils by injecting anti-Ly6G antibody into MCD diet-fed mice (González-Terán et al., 2016). Anti-Ly6G administration for 7 days reduced circulating neutrophil levels without affecting monocytes (Figure 2—figure supplement 2A,B), and treatment for 21 days markedly decreased hepatic diurnal Bmal1 and Clock expression, increased expression of Cry2, and Per2 (Figure 2—figure supplement 2C) and consequently reduced steatosis (González-Terán et al., 2016).

To further support the role of neutrophil liver infiltration in the regulation of liver clock genes and hepatic lipogenesis during diet-induced steatosis, we leveraged a mouse model (p38γ/δ^Lyzs-KO) that exhibits deficient neutrophil migration and subsequently, reduced liver neutrophil infiltration after MCD diet (González-Terán et al., 2016). Compared with diet-matched control (Lyzs-Cre) mice, MCD-diet-fed p38γ/δ^Lyzs-KO mice showed hepatic down-regulation of Bmal1, which was associated with higher expression of Cry2, and Per2 (Figure 2I). These results suggest that the reduced neutrophil infiltration in mice lacking myeloid p38γ/δ expression is responsible for the altered expression of circadian clock genes. Overall, these findings strongly support that neutrophil infiltration modulates clock-gene expression in the liver, with downstream effects on liver metabolism.

It has been suggested that JNK activation in the liver may be regulated in a circadian manner with a peak at noon (Robles et al., 2014). To evaluate whether neutrophils might mediate this diurnal regulation of JNK, we analyzed JNK activation in neutropenic mice. Lack of neutrophils was associated with lower liver expression and activation of JNK, lower activation of the JNK downstream effector c-Jun, and lower expression of acetyl-CoA carboxylase (Acaca), a key enzyme in metabolic regulation (acetyl-CoA carboxylase; ACC) that mediates inhibition of beta-oxidation and activation of lipid biosynthesis (Figure 3A and Figure 3—figure supplement 1A). Similar results were found in p38γ/δ^Lyzs-KO mice, in which reduced liver neutrophil infiltration was associated with decreased JNK phosphorylation and ACC protein levels (Figure 3B and Figure 3—figure supplement 1B). Moreover, neutrophil-treated hepatocytes showed increased JNK activation together with increased levels of ACC expression (Figure 3—figure supplement 1C). NE represents a potential mediator of this neutrophil function because elastase-treated hepatocytes also showed higher JNK activation, suggesting that this protease modulates the expression of the clock genes through the JNK signaling pathways (Figure 3C and Figure 3—figure supplement 1D). This JNK activation was accompanied by increased Bmal1 expression (Figure 3D), indicating that neutrophils altered liver clock-gene expression through the elastase-JNK pathway.

Our results suggest that neutrophil-mediated JNK activation might modulate hepatocyte clock genes and metabolism through the regulation of ACC. Supporting this hypothesis, specific JNK depletion in hepatocytes downregulated Bmal1, Clock, and Acaca compared to Alb-Cre (Figure 3E and Figure 3—figure supplement 1E). According to these results, JNK inhibition reduced the expression of Bmal1, Clock and Acaca in WT liver but not in neutropenic mice (Figure 3—figure supplement 1F,G). These data strongly suggest that JNK activation caused by neutrophil infiltration modulates clock genes and daily metabolism in hepatocytes.

JNK is an important modulator of the expression of the hepatokine circadian regulator FGF21 (Vernia et al., 2014), which controls glucose and lipid metabolism (Fisher and Maratos-Flier, 2013; Li et al., 2013; Potthoff et al., 2012). Mice lacking JNK in hepatocytes had higher FGF21 mRNA expression (Figure 3E). In concordance with high JNK activation, FGF21 expression was reduced in neutrophil-exposed hepatocytes (Figure 3—figure supplement 1H). Moreover, neutropenic and p38γ/δ^Lyzs-KO mice showed increased FGF21 expression (Figure 3F and Figure 3—figure supplement 1I,J), which was consistent with the reduced hepatocyte JNK activation in these mice.

To further define the role of FGF21 in the neutrophil-mediated regulation of liver metabolism, we suppressed FGF21 expression using two independent lentiviral shRNA vectors (Figure 3G and Figure 3—figure supplement 1K). The protection of p38γ/δ^Lyzs-KO mice against MCD-diet-induced alterations was abrogated by shFGF21 and these mice developed steatosis with an elevated hepatic triglyceride content (Figure 3H,I). These data further supported the idea that neutrophil infiltration controls liver metabolism through the regulation of FGF21 expression.

To formally confirm the involvement of NE in circadian clock alteration, we first evaluated the diurnal oscillation of NE levels in liver from WT mice fed a normal diet (ND). According to infiltration pattern of neutrophils in the liver (Figure 1A), we found higher NE levels at ZT2 than at ZT14. (Figure 4A). Next, circadian clock-gene expression in NE^-/- mice revealed lower Bmal1 and elevated Per2 and Cry2 expression, compared to control mice (Figure 4B), which mimicked the behavior of neutropenic mice. In addition, NE^-/- mice presented lower respiratory quotient during the lights-on period than WT mice, indicating that these mice have increased fat utilization as a source of energy (Figure 4C), supporting the data that reduced liver-neutrophil infiltration results in higher lipid oxidation. Interestingly, when fed MCD or HFD diet, NE^-/- mice were protected against steatosis (Figure 4D,E and Figure 4—figure supplement 1A,B), presented lower JNK activation, and expressed less ACC than control mice (Figure 4F,G and Figure 4—figure supplement 1D). Besides, NE^-/- mice were protected against alterations in clock-gene expression induced by MCD diet, presenting lower expression of Bmal1 and higher of Cry2 and Per2 comparing to control mice at ZT2 (Figure 4H). Furthermore, under HFD, NE^-/- mice were also refractory to these changes as these mice maintained a pattern of clock-gene expression similar to control mice in ND (Figure 4—figure supplement 1E).

To formally test a direct contribution of NE in the regulation of hepatic clock-gene expression and liver metabolism, we infused WT or NE^-/- neutrophils into neutropenic mice under the jet lag protocol (Figure 5A). The infusion of WT neutrophils was able to increase Bmal1 expression in the liver after jet lag, while neutropenic mice infused with NE^-/- neutrophils presented the same levels of Bmal1 than non-infused neutropenic mice (Figure 5B). In addition, while infusion of neutropenic mice with WT neutrophils increased steatosis, neutropenic mice infused with NE^-/- neutrophils presented the same levels of steatosis than control neutropenic mice (Figure 5C,D). All these data indicate that diet or jet-lag -induced hepatic infiltration of neutrophils results in dysregulation of the liver clock, and the lack of NE is enough to protect mice against these alterations.

Finally, to evaluate the translational relevance of these findings for human physiology we quantified in human livers the expression levels for the genes encoding NE, JUN (as an indicator of JNK activation) and Bmal. Our results suggest that the levels of ELANE expression directly correlate with BMAL1 and JUN mRNA in livers from a human cohort (Figure 5E). These correlations reinforce the idea that a rhythmic neutrophil infiltration in the liver controls the expression of clock genes through the JNK pathway activation and could be a target for therapeutic intervention during non-alcoholic fatty liver disease.

For the analysis of human liver mRNA levels, individuals were recruited among patients who underwent laparoscopic cholecystectomy for gallstone disease. The study was approved by the Ethics Committee of the University Hospital of Salamanca (Spain), and all subjects provided written informed consent to participate. Patients were excluded if they had a history of alcohol use disorders or excessive alcohol consumption, chronic hepatitis C or B, or body mass index ≥35. Baseline characteristics of these groups are listed in. Figure 5—source data 1

Neutropenic mice were generated with MCL1 (B6.129-Mcl1tm3Sjk/J) crossed with B6.Cg-Tg(S100A8-Cre,-EGFP)1Ilw/J mice or B6.129P2-Lyz2tm1(cre)Ifo/J mice. Mice deficient in NE, with compound JNK1/2 deficiency in hepatocytes, with Cxcr2 deficiency in neutrophils or with p38γ/δ deficiency in myeloid compartment have been described (Belaaouaj et al., 1998; Das et al., 2011; Das et al., 2009; González-Terán et al., 2016) All mice were backcrossed for 10 generations to the C57BL/6J background (Jackson Laboratory). Genotypes were confirmed by PCR analysis of genomic DNA.

Mice were housed under a 12 hr light:12 hr dark cycle (Light is on at Zeitgeber Time ZT0 and off at ZT12). For jet lag experiments, the 12 hr:12 hr dark/light cycle was disrupted by extending the dark cycle 12 hr every 5 days over 3 weeks (Kettner et al., 2016). Cxcr2^MRP8-KO chimeras were generated by exposing WT recipient mice to 2 doses of ionizing radiation (625 Gy) and reconstituting them with 5 × 10⁶ donor BM (Cxcr2^MRP8-KO) cells injected into the tail vein.

Mice were fed a methionine-choline-deficient (MCD) diet for 3 weeks or a high-fat diet (HFD) for 8 weeks (Research Diets Inc). For neutrophil depletion, mice mini-osmotic pumps (Alzet) were implanted with anti-Ly6G antibody or saline (0.4 mg/kg per day, 21 days). For JNK inhibition experiments, mice were intraperitoneally injected with SP600125 (15 mg/kg) (Santa Cruz Biotechnology) at ZT0. For neutrophil infusion experiments, mice were intravenously injected with 3 × 10⁶ WT or NE^-/- purified neutrophils each 3–4 days. Neutrophils were isolated from BM using biotinylated anti-Ly6G antibody (Clone:1A8) and streptavidin-labeled magnetic microbeads (Miltenyi Biotec).

All animal procedures conformed to EU Directive 86/609/EEC and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes, enacted under Spanish law 1201/2005.

Hepatocytes were isolated from adult females by collagenase liver perfusion and cells were filtered through a 70 μm strainer. Hepatocytes pelleted from centrifuged Percoll gradients were plated at 4 × 10⁵ cells/well on 6-well plates coated with collagen type one and incubated at 37°C. After 24 hr, cells were treated with 0.5 mM palmitate (Sigma-Aldrich) for 6 hr and then exposed for 1 hr to freshly neutrophils (2 × 10⁶ cells/well) in the presence of 1 µM FMLP (Sigma-Aldrich). Neutrophils were isolated from BM as described above. For some experiments, neutrophils were sorted purified form the BM using an anti-Ly6G antibody (Clone: 1A8). T and B lymphocytes were sorted purified from spleens using anti-CD3 (Clone: 145–2 C11) and anti-B220 (Clone: RA3-6B2), and bone marrow macrophages (BMDM) were differentiated as previously described (González-Terán et al., 2013). All antibodies were purchased from BD Pharmingen. Alternatively, hepatocytes were exposed 2 hr to 5 nM NE (R and D Systems) or 0.5 mg/mL of collagenase A (Roche) after palmitate treatment.

Mice were perfused with 20 mL of PBS and livers were collected and dissociated. Cell suspension was passed through a 70 μm strainer and centrifuged twice at 50 xg for 2 min to discard the liver parenchyma. For some experiments, livers were incubated for 15 min with 1 mg/mL Collagenase A (Roche) and 2 U/mL DNase (Sigma) at 37°C, and lungs were incubated for 25 min with 0,25 mg/ml Liberase TL (Sigma) and 5 U/mL DNase (Sigma) at 37°C Leukocyte fraction was collected and stained with anti-CD45 (Clone: 30-F11), from Invitrogen, anti-CD11b (Clone: M1/70), anti-Ly6G (Clone: 1A8) or anti-Ly6C/G (Clone: RB6-8C5), from BD Pharmingen, and alternatively, with anti-F4/80 (Clone: BM8), from Invitrogen, and Goat anti-Clec4F from R and D Systems and conjugated with anti-goat Alexa 647. Cells were sorted on a FACSAria to >95% purity. Flow cytometry experiments were performed with a FACScan cytofluorometer (FACS Canto BD), and data were analyzed with FlowJo software.

Transient calcium phosphate transfection of HEK-293 cells (#CRL-1573, ATCC) was performed with the pGIPZ empty or pGIPZ.shFGF21 vector (V3LMM_430499 and V3LMM_430501, from Dharmacon) together with pΔ8.9 and pVSV-G. The supernatants were collected, centrifuged (700 xg, 4°C, 10 min) and concentrated (165x) by ultracentrifugation for 2 hr at 121,986 xg at 4°C (Ultraclear Tubes, SW28 rotor and Optima L-100 XP Ultracentrifuge; Beckman). Mice received tail-vein injections of 200 μl of lentiviral particles.

Expression of mRNA was examined by qRT-PCR using a 7900 Fast Real Time thermocycler and Fast Sybr Green assays (Applied Biosystems). Relative mRNA expression was normalized to Gapdh and Actb mRNA. The primers used were as follows: Actb (F: GGCTGTATTCCCCTCCATCG; R: CCAGTTGGTAACAATGCCATGT); Gapdh (F: TGAAGCAGGCATCTGAGGG; R: CGAAGGTGGAAGAGTGGGA); Clock (F: AGAACTTGGCATTGAAGAGTCTC; R: GTCAGACCCAGAATCTTGGCT); Bmal1 (F: TGACCCTCATGGAAGGTTAGAA; R: GGACATTGCATTGCATGTTGG); Nr1d2 (F: CAGACACTTCTTAAAGCGGCACTG; R: GGAGTTCATGCTTGTGAAGGCTGT); Cry2 (F: CACTGGTTCCGCAAAGGACTA; R: CCACGGGTCGAGGATGTAG); Per2 (F: GAAAGCTGTCACCACCATAGAA; R: AACTCGCACTTCCTTTTCAGG); Acaca (F: GATGAACCATCTCCGTTGGC; R: GACCCAATTATGAATCGGGAGTG); Fgf21 (F: CTGCTGGGGGTCTACCAAG; R: CTGCGCCTACCACTGTTCC); Mip1a (F: TTCTCTGTACCATGACACTCTGC; R: CGTGGAATCTTCCGGCTGTAG); Mip2 (F: CCAACCACCAGGCTACAGG; R: GCGTCACACTCAAGCTCTG); KC (F: CTGGGATTCACCTCAAGAACATC; R: CAGGGTCAAGGCAAGCCTC); Sdf-1 (F: GCTCTGCATCAGTGACGGTA; R: ATCTGAAGGGCACAGTTTGG); Elane (F: ATTTCCGGTCAGTGCAGGTAGT; R: GGTCAAAGCCATTCTCGAAGAT); GAPDH (F: CCATGAGAAGTATGACAACAGCC; R: GGGTGCTAAGCAGTTGGTG); ELANE (F: TCCACGGAATTGCCTCCTTC; R: CCTCGGAGCGTTGGATGATA); BMAL1 (F: GCCGAATGATTGCTGAGG; R: CACTGGAAGGAATGTCTGG); JUN (F: GGATCAAGGCGGAGAGGAAG; R: GCGTTAGCATGAGTTGGCAC).

Lipids were extracted from 25 mg of liver in isopropanol (50 mg/mL) and centrifuged (15 min 9500 xg 4°C). Triglycerides were detected in the supernatant (Sigma-Aldrich).

Tissue samples were fixed in 10% formalin for 48 hr, dehydrated, and embedded in paraffin. Sections (5 μm) were cut and stained with hematoxylin and eosin (Sigma-Aldrich and Thermo Scientific). Sections (8 µm) from frozen tissue and embedded in OCT compound (Tissue-Tek) were stained with Oil Red O (American Master Tech Scientific). Sections were examined in Leica DM2500 microscope using 20x objective.

Tissue extracts were prepared in Triton lysis buffer [20 mM Tris (pH 7.4), 1% Triton X-100, 10% glycerol, 137 mM NaCl, 2 mM EDTA, 25 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 10 µg/mL aprotinin and leupeptin]. Extracts (20–50 µg protein) were examined by immunoblot. The antibodies employed were anti-FGF21 (1/1000, #RD281108100, BioVendor), anti-phospho JNK (1/1000, #4668S, Cell Signaling), anti-JNK (1/1000, #9252S, Cell Signaling), anti-phospho c-Jun (1/1000, #9164L, Cell Signaling), anti-c-Jun (1/1000, #9165S, Cell Signaling), anti-ACC (1/1000, #3676S, Cell Signaling), and anti-vinculin (1/5000, #V9131, Sigma). Anti-phospho JNK and anti-JNK antibodies recognize the two different JNK isoform (JNK1 and JNK2) and their two spliced variants (JNK1 (46 kDa), JNK1 (54 kDa) and JNK2 (46 kDa) and JNK2 (54 kDa)). Immunocomplexes were detected by enhanced chemiluminescence (Amersham).

For 3-D imaging, livers were fixed in a solution of paraformaldehyde 4% in PBS at 4°C. After washing in PBS, tissues were stored overnight in 30% sucrose (Sigma) with PBS. Then, livers were embedded in OCT compound (Tissue-Tek) and frozen at −80°C. Cryosections of organs (70 µm) were washed in PBS and blocked/permeabilized in PBS with 10% donkey serum (Millipore) and 1% Triton. Primary antibodies diluted in blocking/permeabilization buffer were incubated overnight at 4°C, followed by three washes in PBS and 2 hr incubation with secondary antibodies and DAPI at room temperature. After three washes in PBS, cells were mounted with Fluoromount-G (SouthernBiotech). The following primary and secondary antibodies were used: rat anti-CD31 (1:200, #553370 BD Pharmingen,), rabbit anti-S100A9 (mrp14) (1:100, #AB242945↗, Abcam,), goat anti-Clec4f (1:100, #AF2784, RD System), Alexa 488 donkey anti rat IgG (1:200, #A-21208, ThermoFisher), Cy3 AffiniPure Fab Fragment Donkey Anti-Rabbit IgG (1:200, #711-167-003, Jackson Laboratories), Alexa Fluor 633 donkey anti goat IgG (H+L) (1:200, #A21082, ThermoFisher). Immunostaining were imaged with a SP8 confocal microscope using 40x objectives. Individual fields or tiles of large areas were acquired every 2.5 µm for a total of 30 µm in depth. 3D images were obtained with Fiji/ImageJ 3D Viewer plugging.

For 2-D imaging, liver sections (12 µm) prepared from frozen tissue and embedded in OCT compound were fixed with 2% paraformaldehyde and permeabilized with PBS 0.1% Triton. After blocking with PBS 5% BSA 0.1% Triton and washing, tissues were incubated overnight at 4°C with primary antibody. Then, sections were washed and incubated with conjugated secondary antibodies for 1 hr at room temperature and nuclei were stained with Sytox Green (Invitrogen) after washing. The following primary and secondary antibodies were used: rat anti-mouse S100A9 (Mrp-14) antibody (1:200, #AB105472↗, Abcam), rabbit anti-Neutrophil Elastase antibody (1:200, #AB68672, Abcam), goat Alexa Fluor 405 anti-rabbit (1:200) and goat Alexa Fluor 568 anti-rat IgG (1:500). Sections were mounted in Vectashield mounting medium (Vector, H-1000) and examined using a Leica SP5 multi-line inverted confocal microscope and 20x objectives.

20 mL of PBS prefunded livers were crushed with a syringe plunger, resuspended in 4 mL of PBS/EDTA 5 mM/0.5% FBS and filtered (70 µm). Cell suspension was centrifuged at 1800 rpm 5 min and the supernatant was filtered (22 µm). Supernatants were concentrated using Amicon Ultra centrifugal filters (Sigma-Aldrich). NE levels were determined with Mouse Neutrophil Elastase ELISA kit (R and D system).

All data are expressed as means ± SEM. For comparisons between two groups, the Student’s t-test was applied. For data with more than two data sets, we used one-way ANOVA coupled with Turkey’s multigroup test. When variances were unequal, Welch’s test or Kruskal-Wallis test coupled with Dunn’s multiple comparison test were applied, respectively. Multiple group comparisons in the rhythmicity of neutrophil infiltration were analyzed with two-way ANOVA followed by Fisher’s post hoc test. Significance was determined as a 2-sided p < 0.05. All statistical analyses were conducted in GraphPad Prism software. Statistical details were indicated in the figure legends.