Melatonin Suppresses Macrophage M1 Polarization and ROS-Mediated Pyroptosis via Activating ApoE/LDLR Pathway in Influenza A-Induced Acute Lung Injury

Influenza virus infection is one of the strongest pathogenic factors for the development of acute lung injury (ALI)/ acute respiratory distress syndrome (ARDS). However, the underlying cellular and molecular mechanisms have not been clarified. In this study, we aim to investigate whether melatonin modulates macrophage polarization, oxidative stress, and pyroptosis via activating Apolipoprotein E/low-density lipoprotein receptor (ApoE/LDLR) pathway in influenza A-induced ALI. Here, wild-type (WT) and ApoE-/- mice were instilled intratracheally with influenza A (H3N2) and injected intraperitoneally with melatonin for 7 consecutive days. In vitro, WT and ApoE-/- murine bone marrow-derived macrophages (BMDMs) were pretreated with melatonin before H3N2 stimulation. The results showed that melatonin administration significantly attenuated H3N2-induced pulmonary damage, leukocyte infiltration, and edema; decreased the expression of proinflammatory M1 markers; enhanced anti-inflammatory M2 markers; and switched the polarization of alveolar macrophages (AMs) from M1 to M2 phenotype. Additionally, melatonin inhibited reactive oxygen species- (ROS-) mediated pyroptosis shown by downregulation of malonaldehyde (MDA) and ROS levels as well as inhibition of the NLRP3/GSDMD pathway and lactate dehydrogenase (LDH) release. Strikingly, the ApoE/LDLR pathway was activated when melatonin was applied in H3N2-infected macrophages and mice. ApoE knockout mostly abrogated the protective impacts of melatonin on H3N2-induced ALI and its regulatory ability on macrophage polarization, oxidative stress, and pyroptosis. Furthermore, recombinant ApoE3 (re-ApoE3) inhibited H3N2-induced M1 polarization of BMDMs with upregulation of MT1 and MT2 expression, but re-ApoE2 and re-ApoE4 failed to do this. Melatonin combined with re-ApoE3 played more beneficial protective effects on modulating macrophage polarization, oxidative stress, and pyroptosis in H3N2-infected ApoE-/- BMDMs. Our study indicated that melatonin attenuated influenza A- (H3N2-) induced ALI by inhibiting the M1 polarization of pulmonary macrophages and ROS-mediated pyroptosis via activating the ApoE/LDLR pathway. This study suggested that melatonin-ApoE/LDLR axis may serve as a novel therapeutic strategy for influenza virus-induced ALI.


Background
Acute lung injury (ALI)/acute respiratory distress syndrome (ARDS) is a rapidly progressing and refractory disease. Particularly, seasonal influenza-induced ALI/ARDS increasingly contributes to annual global mortality. Recent studies indicated that seasonal influenza epidemics annually affected 10-30% of the human population and cause 3-5 million severe cases and approximately 290,000-650,000 deaths worldwide especially among the elderly with chronic diseases [1,2]. And it was estimated that an annual average of 88,100 influenza-associated respiratory deaths occurred in China [3]. Studies found that influenza viruses A-(H3N2-) associated hospitalizations and mortality were the highest among other circulating viruses [4][5][6], because it was easier to progress to severe pneumonia with bilateral pulmonary infiltrates, even consolidation (i.e., "white lung" in imageology), and cause ALI/ARDS [1,3]. Therefore, the further study on cellular and molecular mechanisms of influenzaassociated ALI/ARDS remains urgently needed for more efficient agents and therapeutic strategies.
Persistent influenza virus infection leads to robust oxidative stress and cytokine storms (i.e., hypercytokinemia) with extensive pulmonary leukocyte infiltration, edema, and alveolar haemorrhages [7], which are commonly caused by pulmonary immune cells. Therein, macrophages are the most important innate immune cells and constitute the first line of defense against virus and bacteria in ALI/ARDS [8]. As highly plastic cells, macrophages can be polarized into different functional phenotypes, mainly including classically activated (M1) macrophages and alternatively activated (M2) macrophages, which perform different biological functions [9]. In the acute stage of influenza infection, macrophages are polarized into M1 phenotype and release abundant proinflammatory proteins and reactive oxygen species (ROS) [10]. Studies indicated that ROS released by M1 macrophages were prone to causing pyroptosis via activating the NLR family pyrin domain containing 3 (NLRP3) /gasdermin D (GSDMD) pathway [11][12][13]. Additionally, macrophage polarization may be regulated by Apolipoprotein E (ApoE) signaling [14], and the ApoE/LDLR pathway was suggested to be a potential signaling pathway involved in antiinflammation and antioxidation [15]. ApoE is an argininerich glycoprotein primarily synthesized in the liver, brain, lung epithelia, and macrophages [15,16]. Low-density lipoprotein receptor (LDLR), the primary binding receptor of ApoE, can assist ApoE to mediate lipid metabolism and the development of pulmonary diseases [15,17]. Previous studies demonstrated that ApoE knockout caused more severe airway inflammation and oxidative stress in asthmatic mice, whereas administration of an ApoE mimetic peptide suppressed the negative effects in ApoE-/-asthmatic mice, but no effects in LDLR-/-asthmatic mice [18,19], indicating that ApoE exerted the anti-inflammation and antioxidation probably in a LDLR-dependent manner.
Melatonin (N-acetyl-5-methoxytryptamine, C 13 H 16 N 2 O 2 ), a kind of neuroendocrine hormone, is mainly synthesized and secreted by the pineal gland of brains and exerts effects via its membrane receptors (MT1 and MT2) [20,21]. Melatonin is also synthesized in respiratory epithelia and bone marrow cells as well as macrophages [22,23]. It is widely recognized that melatonin is in charge of regulating sleep and circadian rhythm and also plays important roles in anti-inflammation and antioxidation [20]. Recent evidence indicated that melatonin had potential abilities of antivirus infection, including respiratory syncytial virus (RSV), influenza virus, and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [24,25]. Melatonin may inhibit viral replication, improve mitochondrial metabolism via modulating the circadian rhythm, and restrain the "influenza virus-cytokine-trypsin" cycle via inhibiting NLRP3 inflammasome [25]. Moreover, melatonin is also capable of inhibiting the M1 polarization of macrophages to favor antioxidation and anti-inflammation [26]. However, the protective mechanisms of melatonin in influenza virusinduced ALI/ARDS remain unknown.
Taken together, this study was aimed at investigating whether melatonin inhibited influenza virus-induced ALI/ ARDS and its underlying molecular mechanisms, which may provide potential therapeutic approaches and efficient agents for influenza-associated ALI/ARDS. 2.2. IAV/H3N2 Amplification and Plaque Assay. Influenza A/Anhui/1/2017 (H3N2) virus was obtained from Prof. Yan Liu (Department of Microbiology, Anhui Medical University, China) and isolated from the patient in 2017 and used in laboratory studies under approved standard biosafety procedures. The influenza A (H3N2) virus samples were amplified in Madin-Darby canine kidney (MDCK) cells and specific pathogen-free embryonated chicken eggs, and virus titers were assayed by standard plaque assay on MDCK cells according to previous description [27]. MDCK cells were infected with diluted virus samples for 2 h at 37°C. After being washed with PBS, the cultivation was proceeded with 50% 2× DMEM, 50% avecil (2.35%), and N-acetyl trypsin (1.5 μg/ml) for 72 h. Then, cells were stained with naphthalene blue-black, and plaques were counted for the calculation of virus titers. All experiments involved in influenza virus were performed according to the biosafety level two requirements with well-equipped personal protection for all the researchers. Oxidative Medicine and Cellular Longevity    Then, smeared BAL cells were stained with Wright-Giemsa stain solution (Baso Diagnostics Inc, Zhuhai, China) for differential leukocyte counting mainly including neutrophils and macrophages in a double-blind manner as previously described [29]. Blood was drawn from the left heart through a 1 ml syringe and centrifuged at 4°C, 4000 rpm for 10 min, to obtain the serum. Serum and BALF of mice were stored in a -80°C freezer for further cytokine analysis.  9 Oxidative Medicine and Cellular Longevity according to the following formula [30]: lung injury scores = ½ð20 × aÞ + ð14 × bÞ + ð7 × cÞ + ð7 × dÞ + ð2 × eÞ/ðnumber of fields × 100Þ.

Materials and Methods
The degree of lung edema was estimated according to the wet/dry ratio of lung. After anesthesia with 1% sodium pentobarbital (100 mg/kg, intraperitoneal), whole lung tissues were isolated and weighed as the wet weight of the lung. After oven drying at 60°C for 48 h, lung tissues were secondly weighed as the dry weight of the lung. The weight ratio of the wet and dry (W/D) lung was then calculated.       For further verifying the effects of melatonin on macrophage polarization, BMDMs were infected with H3N2 (MOI = 2) for 12 h to stimulate the M1 polarization. In the melatonin intervention group, BMDMs were pretreated with melatonin (400 μM) for 3 h before H3N2 infection. Additionally, BMDMs were also pretreated with recombinant ApoE (re-ApoE2, re-ApoE3, and re-ApoE4) (10 μg/ml, PeproTech) before H3N2 infection.
2.11. Detection of Cell Viability. Cell viability of Raw264.7 cells stimulated by influenza A (H3N2) with different multiplicities of infection (MOI) was measured by Cell Counting Kit-8 (C0038, Beyotime Technology). Briefly, cells seeded in 96-well plates were incubated with CCK8 working solution for 2 h at 37°C. The absorbance value at 450 nm (OD 450 ) was measured by a microplate reader (BioTek Instruments, Vermont, USA).

Detection of Cell Reactive Oxygen Species (ROS) Levels.
Cell ROS levels were detected using 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA, D6883, Sigma-Aldrich) according to the manufacturer's instructions. Raw264.7 cells were seeded in 96-well plates with a density of 5 × 10 4 /ml with 6 parallel wells, and BMDMs were seeded in 96-well plates with a density of 20 × 10 4 /ml with 6 parallel wells, incubated overnight at 37C, and then treated with corresponding prevention. After 12 h incubation, cells were stained with 10 μM DCFH-DA at 37°C avoiding light for 30 min. ROS levels were determined by Fluorescence Microscopy (Leica, Germany) or Varioskan Flash (Thermo

20
Oxidative Medicine and Cellular Longevity error of the mean (SEM) from at least three independent samples or biological replicates (n ≥ 3). Statistical analysis was performed using GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA). Student's t test was performed for comparisons between two different groups. One-way ANOVA with Bonferroni's post hoc test (for equal variance) or Dunnett's T3 post hoc test (for unequal variance) was performed for comparisons among multiple groups. * p < 0:05 was considered statistically significant.

Melatonin Reversed Influenza A-(H3N2-) Induced
Decreased Expression of Melatonin Receptors. The detection of CCK8 showed that influenza A (H3N2) stimulation with MOI of 2 had no significant effects on the viability of Raw264.7 cells ( Figure S1(a)). And H3N2 infection inhibited the mRNA expression of melatonin receptors (MT1 and MT2), which showed a marked decrease after infection at 12 h and recovered at 24 h in Raw264.7 cells (Figures 1(a) and 1(b)). Oppositely, H3N2 infection upregulated IL-1β mRNA expression, which reached the peak level at 12 h, then reduced at 18 h and 24 h in Raw264.7 cells (Figure 1(c)). Additionally, in H3N2infected mouse lung tissues, the mRNA expression of MT1 and MT2 showed decreases compared to the PBS group and PBS+Mel group (Figure 1(e)). However, melatonin intervention significantly increased the mRNA expression of MT1 and MT2 in H3N2-infected Raw264.7 cells and mice (Figures 1(d) and 1(e)). Moreover, the protein expression of MT-1/2 also decreased after H3N2 infection in Raw264.7 cells and mice (Figures 1(d) and 1(e)), suggesting that H3N2 infection may induce a reduced secretion of melatonin. After melatonin treatment, MT-1/2 expression significantly increased in H3N2-infected

Melatonin Inhibited Influenza A-(H3N2-) Induced ALI
and Pyroptosis. HE staining showed that H3N2 infection induced significant pulmonary destruction and leukocyte infiltration on day 7, whereas severe lung injury and fibrous changes happened on day 14 in lung tissues ( Figure S1(b)). Therefore, we chose H3N2 infection for 7 consecutive days to establish the ALI mouse model ( Figure S1(c)). Administration of melatonin significantly attenuated H3N2-induced ALI with the decreases of lung injury scores (Figure 2(a)) and also reduced the wet/dry ratio of the lung (Figure 3(b)), indicating inhibiting lung edema. Meanwhile, melatonin significantly decreased total leukocyte counting, especially neutrophils and macrophages ( Figure S2 (d-f)). Moreover, compared to the H3N2 infection group, melatonin significantly decreased the levels of IL-1β and TNF-α in BALF (Figure 2(b)) and reversed H3N2-induced increases of MDA contents in the serum of mice (Figure 2(c)). Subsequently, melatonin significantly reduced H3N2-induced increases in the expression of NLRP3, Caspase1, and GSDMD-N protein (Figure 2(d)), indicating suppressing oxidative stress and pyroptosis in mice.   25 Oxidative Medicine and Cellular Longevity increased around the areas of leukocyte infiltration of bronchial epithelia and alveolar walls in H3N2-infected mice (Figure 2(f)). After melatonin treatment, iNOS expressed intensity showed an obvious decrease, whereas Arg1 (M2 marker) expression showed a marked enhancement in H3N2-infected lung tissues (Figure 2(f)), as indicated in western blot analysis of iNOS and Arg1 of lung tissues (Figure 2(e)). Similarly, PCR analysis also showed that melatonin reversed H3N2-induced increases of M1 markers (IL-1β, TNF-α, and MCP1) and upregulated the mRNA expression of M2 markers (Arg1 and Fizz1) in H3N2-infected mice (Figures S2(a) and S2(b)). In order to further investigate the effects of melatonin on the polarization of pulmonary macrophages, alveolar macrophages (AMs) in BALF were defined by flow cytometry analysis based on the specific markers of M1 AMs (CD45+Siglec-F+CD11c+CD86+ population) and M2 AMs (CD45+Siglec-F+CD11c+CD206+ population). Flow cytometry analysis of BAL cells revealed that H3N2 infection significantly decreased the percentage of AMs, particularly CD206+ AMs, and increased CD86+ AMs (Figures 4(g)-4(j)). However, administration of melatonin increased the percentage of CD206+ AMs and decreased the percentage of CD86+ AMs in H3N2-infected wildtype (WT) mice (Figures 4(g)-4(j)). These results suggested that melatonin exerted anti-inflammatory effects by inhibiting the M1 polarization of pulmonary macrophages in H3N2-induced ALI.

ApoE/LDLR Pathway Was Involved in the Protective
Impacts of Melatonin. Recent evidence demonstrated that melatonin exerted anti-inflammation probably involved in the activation of ApoE signaling [31]. Immunohistochemical staining showed that ApoE expressed intensity was obviously enhanced around bronchial epithelia and alveolar walls after melatonin intervention in H3N2-infected mice (Figure 6(a)). And ApoE protein expression also showed an obvious increase after melatonin intervention accompanied by an increase of LDLR expression in H3N2-infected mice (Figure 6(b)). ELISA also demonstrated that serum ApoE levels significantly decreased after H3N2 infection, whereas they were upregulated after melatonin intervention (Figure 6(c)). Additionally, in H3N2-infected Raw264.7 cells, melatonin intervention also significantly upregulated ApoE mRNA expression (Figure 6(d)) and promoted the protein expression of ApoE and LDLR (Figure 6(e)). These results suggested that the ApoE/LDLR pathway was positively related with the protective impacts of melatonin.

ApoE Knockout Mostly Abrogated Anti-Inflammatory Impacts of Melatonin on Influenza A-(H3N2-) Infected
Mice. To investigate whether melatonin inhibited H3N2induced ALI in an ApoE-dependent manner, ApoE-/-mice were infected with H3N2 with melatonin intervention. The PCR analysis showed that ApoE mRNA expression was entirely lost in ApoE-/-mice ( Figure S2(c)). H&E staining showed that H3N2-induced ALI was further aggravated in ApoE-/-mice compared to that of WT mice, mainly manifested in more severe leukocyte infiltration and hyaline membrane formation as well as alveolar septal thickening (Figure 3(a)). And lung morphology showed that H3N2 infection induced more obvious lung hyperaemia and edema in ApoE-/-mice compared to that of WT mice, as indicated in analysis of the wet/dry ratio of the lung (Figure 3(b)). However, melatonin administration failed to improve H3N2-induced ALI and edema of ApoE-/-mice (Figures 3(a) and 3(b)). BAL cell smearing and counting also showed that H3N2 infection further increased total leukocyte counting, especially neutrophils and macrophages in ApoE-/-mice, but no obvious reduction after melatonin intervention (Figure 3(c) and Figure S2(d-f)).

ApoE Knockout Suppressed the Regulation of Melatonin on Macrophage Polarization in Mice.
In H3N2-infected ApoE-/-mice, melatonin failed to increase ApoE and LDLR expression (Figure 4(a) and Figure S4(b)). There was a significant increase of iNOS expression in H3N2-infected ApoE-/-mice compared to that of WT mice. However, melatonin failed to inhibit iNOS expression and promote Arg1 expression in H3N2-infected ApoE-/-mice ( Figure   4(a) and Figure  S4(c)). Similarly, immunohistochemical staining showed that iNOS expressed intensity in areas of infiltration of leukocytes was obviously enhanced, and Arg1 expressed intensity was attenuated, whereas there were no obvious changes in expressed intensities of iNOS and Arg1 after melatonin intervention in H3N2-infected ApoE-/-mice (Figure 4(b)). Additionally, the PCR results indicated that melatonin inhibited H3N2-induced increases of mRNA expression of TNF-α and MCP1 and promoted mRNA expression of Arg1 and Fizz1 in H3N2-infected WT mice (Figures 4(c)-4(f)). Oppositely, in ApoE-/-mice, melatonin failed to do this (Figures 4(c)-4(f)). Similarly, flow cytometry analysis of BAL cells revealed that administration of melatonin had no significant effects on the percentage of AMs, especially CD86+ AMs and CD206+ AMs in H3N2-infected ApoE-/mice (Figures 4(g)-4(j)). These results indicated that ApoE knockout mostly abrogated the regulatory impacts of melatonin on the polarization of pulmonary macrophages.

re-ApoE3 Promoted Influenza A-(H3N2-) Induced M1
BMDMs to M2 Polarization. To further explore the role of ApoE on macrophage polarization regulated by melatonin, recombinant ApoE proteins (re-ApoE2, re-ApoE3, and re-ApoE4) were added in H3N2-infected bone marrow-derived macrophages (BMDMs). Flow cytometry analysis showed that the purity of matured BMDMs reached over 95% ( Figure S5). The PCR analysis showed that re-ApoE3 pretreatment significantly reversed H3N2-induced decreases in the mRNA expression of MT1 and MT2 as well as the protein expression of MT-1/2, but no effects after re-ApoE2 and re-ApoE4 pretreatment in H3N2-infected BMDMs (Figures 7(a) and 7(b) and Figure S6(b)), suggesting that ApoE3 may affect the secretion of melatonin via regulating the expression of melatonin receptors.
The morphology of BMDMs showed marked differentiation from original spindle shapes toward circular shapes after H3N2 stimulation, whereas it was reversed by re-ApoE3 intervention ( Figure S6(a)). Western blot analysis indicated that ApoE expression significantly increased in H3N2-infected BMDMs with re-ApoE2 and re-ApoE3 pretreatments, but no significant increase on re-ApoE4 (Figure 7(b) and Figure S6(c)). And re-ApoE3 significantly reversed H3N2-induced increase of iNOS expression and promoted Arg1 expression (Figure 7(b) and Figure S6(d)). Likewise, immunofluorescence staining also showed that only re-ApoE3 significantly alleviated the fluorescence intensity of iNOS and enhanced that of Arg1 in H3N2infected BMDMs (Figure 7(c)), indicating the M2 polarization of BMDMs by ApoE3. Additionally, re-ApoE3 significantly inhibited the mRNA expression of TNF-α, MCP1, and CD86 and increased the mRNA expression of Arg1, Fizz1, and CD206 in H3N2-infected BMDMs (Figures 7(d) and 7(e)). These results indicated that re-ApoE3 exerts potential anti-inflammatory impacts by modulating macrophage polarization.
Next, in H3N2-infected ApoE-/-BMDMs, melatonin failed to downregulate H3N2-induced increases of ROS levels ( Figure 8(i)) and also did not inhibit the protein expression of NLRP3, Caspase1, and GSDMD-N (Figure 8(k) and Figure S7(d)). There were also no significant changes in IL-1β mRNA expression and LDH release after melatonin intervention in H3N2-infected ApoE-/-BMDMs (Figures 8(j) and 8(l)). Further assessing, we found that melatonin combined with re-ApoE3 significantly decreased ROS levels (Figure 8(i)) and inhibited the protein expression of NLRP3, Caspase1, and GSDMD-N as well as the IL-1β mRNA expression and LDH release in H3N2-infected ApoE-/-BMDMs (Figures 8(j)-8(l) and Figure S7(d)). These results demonstrated that exogenous re-ApoE3 enhanced the beneficial effects of melatonin, and the activation of the ApoE/LDLR pathway improved the modulation of melatonin on macrophage polarization, oxidative stress, and pyroptosis.

Discussion
In the present study, we put forward novel insights into the regulatory role of melatonin-ApoE/LDLR axis in macrophage polarization, oxidative stress, and pyroptosis and confirmed melatonin as a potential therapeutic agent in influenza virus-induced ALI, as indicated in Figure 9. Specifically, we proved that melatonin significantly attenuated influenza A-(H3N2-) induced ALI with the activation of the ApoE/LDLR pathway. ApoE knockout almost abrogated the protective impacts of melatonin on H3N2-induced ALI; re-ApoE3 inhibited H3N2-induced M1 polarization of BMDMs and inflammatory responses. Furthermore, melatonin and re-ApoE3 cotreatment reversed damaged effects induced by ApoE knockout, effectively switched macrophage polarization from M1 to M2 phenotype, and inhibited ROS production and pyroptosis. Taken together, melatonin suppressed macrophage M1 polarization and ROS-mediated pyroptosis via activating the ApoE/LDLR pathway in ALI. The conclusion provided new direct evidence that melatonin 27 Oxidative Medicine and Cellular Longevity exerted the anti-inflammation and antioxidation in an ApoE-dependent manner.
In view of increasing global influenza and COVID-19 pandemics, more studies are devoted to elucidating the pathophysiology of ALI induced by virus infection for further developing the effective therapeutic agents [32,33]. Particularly, relative studies indicated that the clinical and pathogenic features of SARS-CoV-2 infection had many parallels with influenza [34]. Virus-damaged epithelial cells can recruit a series of immune cells, especially macrophages, which induce the cascading amplification of inflammatory responses and the damage of lung structures [35]. Specially, at least three types of macrophages exist in lung tissues: bronchial macrophages, interstitial macrophages (IMs), and alveolar macrophages (AMs). Therein, AMs in the alveolar lumen form 90-95% of the cellular contents at homeostasis [36]. In ALI, monocytes recruited into the lung can differentiate into AMs. Under mild influenza A virus (IAV) infection, AMs can exert protective impacts through phagocytizing apoptotic epithelial cells [37]. With the exacerbation of IAV infection, the phagocytic capacity of AMs is decreased; oppositely, AMs may phagocytize IAV and assist the replication of progeny virus so as to infect surrounding cells [38]. Therefore, AMs also are regarded as a vehicle for virus dissemination. Moreover, IAV infection also induced a conversion of AMs toward M1 phenotype with an increase of iNOS expression [39]. And IAVinfected M2 BMDMs tended to polarize into M1 phenotype with excessive expression of iNOS and TNF-α [40]. Our results also showed that the M1 pulmonary macrophages were predominant in H3N2-induced ALI.
Studies have clarified that melatonin influenced multiple physiological functions of macrophages from host defenses to immune disorders, modulating inflammation and oxidation-antioxidant system [26,41]. In a stress-induced inflammation model, melatonin switched macrophage polarization from M1 to M2 phenotype with increases of the M2 marker Arg1 and MRC1 expression and inhibited inflammatory injuries [42]. And in PM 2.5 -induced atherosclerosis, melatonin effectively alleviated PM 2.5 -induced oxidative damage of the aorta and atherosclerotic plaque formation via inhibiting macrophage M1 polarization and NOX2-mediated oxidative stress [43]. These studies demonstrate that melatonin can inhibit the M1 polarization of macrophages and oxidative stress. Accordingly, we firstly confirmed that melatonin mainly targeted pulmonary macrophages to exert the protective impacts in H3N2-infected mice. Mechanistically, melatonin inhibited H3N2-induced M1 polarization of pulmonary macrophages and oxidative stress.
With unbridled hyperinflammatory reactions, influenza virus infection will also lead to pyroptosis, a kind of programmed necrotic cell death [33]. Pyroptosis commonly relies on the gasdermin family members to cause cell rupturing and the formation of membrane pores. And GSDMD is considered as the real executioner of pyroptosis which is commonly cleaved to expose the N-terminal domains in a caspase1-dependent manner following NLRP3 inflammasome assembly [44]. Recent studies pointed that melatonin attenuated LPS-induced pyroptosis by inhibiting the NLRP3/GSDMD pathway which was primarily activated by ROS [29,45]. Generally speaking, ROS are also regarded as biomarkers of M1 macrophage polarization [9]. Therefore, we can consider that the M1 polarization of macrophages promotes the occurrence of pyroptosis via activating the ROS-mediated NLRP3/GSDMD pathway. Consistently, a recent study proved that the M2 polarization of AMs alleviated LPS-induced lung pyroptosis via downregulating the Caspase1/GSDMD pathway [46]. As indicated in our results, melatonin inhibited ROS-driven activation of the NLRP3/ GSDMD pathway via switching macrophage polarization from M1 to M2 phenotype.
Emerging evidence increasingly recognized that ApoE protein played a protective role in the development of lung diseases based on their ability to regulate inflammation and oxidative stress [15]. Previous studies indicated that ApoE-/mice showed more severe pulmonary toxicity and neutrophil infiltration in the ALI model [47,48], whereas administration of COG1410, an ApoE mimetic peptide, inhibited LPS-induced increases of alveolar neutrophils and macrophages [49]. And ApoE-/-mice transplanted with ApoEexpressed bone marrow showed increased plasma levels of IL-1RA (M2 Marker), and peritoneal macrophages of transplanted mice were also polarized into the M2 phenotype with increases of IL-1RA and CD206 levels [14]. Moreover, ApoE also had positive antioxidant effects, in a spinal cord injury mouse model; exogenous ApoE administration significantly improved oxidative stress and neural function via Nrf2/HO-1 signaling [50]. These studies indicated that ApoE can be considered as an anti-inflammatory and antioxidant protein with an ability of regulating macrophage polarization. Additionally, a recent study demonstrated that melatonin decreased ROS levels and caspase activity by upregulating ApoE expression in oxygen and glucose deprivation-reoxygenation-(OGD-R-) stimulated endothelial cells [31]. Particularly, melatonin and ApoE are all highly expressed in brains so that potential interaction may be existing between them. In our results, ApoE knockout almost abrogated the positive effects of melatonin on ALI, and exogenous ApoE3 remedied the protective impacts of melatonin. Therefore, we considered that melatonin exerted anti-inflammation and antioxidation in an ApoE-dependent manner. It is a completely new application area for the modulation of melatonin on macrophage polarization, oxidative stress, and pyroptosis via the ApoE/LDLR pathway.
Specifically, human ApoE is polymorphic with three variants that encodes different amino acids in codons 112 and 158: cysteine at both sites for ApoE2, arginine at both sites for ApoE4, and respectively, cysteine and arginine for ApoE3 [16]. This variation greatly modifies ApoE protein functions and causes minimal binding activity with LDLR in ApoE2. Accumulated ApoE2 may induce dysbetalipoproteinaemia and accelerate aging [51]. And the variation in ApoE4 may cause the loss of antioxidant ability and become the strongest risk factor of Alzheimer's disease (AD) [52]. ApoE3 is the one of the highest frequencies and accounts for 65-70% of total ApoE [53]. ApoE3 possesses the requisite lipid-binding ability and higher affinity with LDLR [54]. 28 Oxidative Medicine and Cellular Longevity Moreover, the cysteine residues of ApoE3 drive the covalent binding with 4-hydroxynonenal (HNE) so as to inhibit lipid peroxidation and inflammation [44]. Studies indicated that human ApoE3-knockin mice decreased the levels of TNF-α and IL-1β and elevated the survival compared to those of ApoE4-knockin mice in the caecal ligation and puncture model [55]. Interestingly, in our results, re-ApoE3 switched H3N2-induced M1 BMDMs toward the M2 polarization and also increased the expression of melatonin receptors; however, re-ApoE2 and re-ApoE4 failed to do this. These results indicated a potential synergetic protective impact between melatonin and ApoE3. And re-ApoE3 also enhanced the regulated ability of melatonin on macrophage polarization, oxidative stress, and pyroptosis, which further proved that melatonin attenuated H3N2-induced ALI in an ApoE-dependent manner.

Conclusion
In this study, we have provided strong direct evidence for the first time that melatonin attenuated influenza A-(H3N2-) induced ALI by inhibiting macrophage M1 polarization and ROS-mediated pyroptosis via activating the ApoE/LDLR pathway. We found that re-ApoE3 exerted the positive protective impacts by promoting H3N2-induced M1 BMDMs toward M2 polarization and also enhanced the anti-inflammatory and antioxidant abilities of melatonin, indicating that melatonin-ApoE/LDLR axis may serve as a novel intervention signal for treating influenza Ainduced ALI. Meanwhile, this study also provided a potential clue for the therapy of ARDS induced by the novel coronavirus SARS-CoV-2.

Data Availability
The data that support the findings of this study are available from the corresponding authors upon reasonable request.