Influenza A virus selectively elevates prostaglandin E2 formation in pro-resolving macrophages

Summary Respiratory influenza A virus (IAV) infections are major health concerns worldwide, where bacterial superinfections substantially increase morbidity and mortality. The underlying mechanisms of how IAV impairs host defense remain elusive. Macrophages are pivotal for the innate immune response and crucially regulate the entire inflammatory process, occurring as inflammatory M1- or pro-resolving M2-like phenotypes. Lipid mediators (LM), produced from polyunsaturated fatty acids by macrophages, are potent immune regulators and impact all stages of inflammation. Using LM metabololipidomics, we show that human pro-resolving M2-macrophages respond to IAV infections with specific and robust production of prostaglandin (PG)E2 along with upregulation of cyclooxygenase-2 (COX-2), which persists after co-infection with Staphylococcus aureus. In contrast, cytokine/interferon production in macrophages was essentially unaffected by IAV infection, and the functionality of M1-macrophages was not influenced. Conclusively, IAV infection of M2-macrophages selectively elevates PGE2 formation, suggesting inhibition of the COX-2/PGE2 axis as strategy to limit IAV exacerbation.


INTRODUCTION
Infections with influenza A virus (IAV) are a leading cause of morbidity and mortality of pneumonia in both children and adults, accompanied by suppressed and dysregulated host immune functions by multiple mechanisms. 1 However, autopsy case reports revealed that >90% of deaths during the 1918 influenza pandemic proceeded from secondary pneumonia due to bacterial superinfections. 2,3Staphylococcus aureus (S. aureus) is one of the most prominent colonizing pathogens in this respect, which can either be persistent or non-persistent, 4 ranging from skin and soft tissue infections to life-threatening disease states, such as bacteremia, septicemia, or pneumonia. 5,6The simultaneous occurrence of various pathogens can exuberate pathological effects in affected organs.Thus, superinfection of IAV-infected patients with S. aureus leads to increased inflammatory lung damage. 7The underlying mechanism of how IAV substantially impairs host defense against bacteria (e.g., S. aureus) and thus facilitates bacterial infections is still elusive. 8he acute inflammatory response is a major protective mechanism of the host to eliminate invading pathogens.Pathogens are detected mainly by pattern-recognition receptors (PRRs) of tissue-resident innate immune cells, followed by specific release of pro-inflammatory cytokines (e.g., TNF-a, IL-1b, and IL-6) and type I-III interferons (IFN). 9The PRRs for IAV infections recognize viral RNA, while the main pathogen-associated marker pattern (PAMP), are Toll-like receptors (TLRs), retinoic acid-inducible gene-I (RIG-I), and the nucleotide-binding oligomerization domain (NOD)-like receptor family pyrin domain containing 3 (NLRP3) domain. 10,11Besides cytokines, also the production of polyunsaturated fatty acid (PUFA)-derived lipid mediators (LMs) that play crucial roles in host defense by tightly regulating inflammatory and immune responses, are formed in mammals during acute infections with various pathogens. 12Prominent LMs in the initiation phase of inflammation are eicosanoids derived from arachidonic acid (AA) such as prostaglandins (PG) and leukotrienes (LT), formed by the cyclooxygenase (COX) and 5-lipoxygenase (LOX) pathways, respectively. 13In the resolution phase of infectious diseases, specialized pro-resolving mediators (SPM) are produced by various LOXs as key enzymes, which are LM that terminate acute inflammation and promote the resolution of inflammation and the return to tissue homeostatic states. 14,15t was shown that mice infected with highly pathogenic 1918 pandemic H1N1 and H5N1 IAV strains exhibited increased numbers of macrophages and neutrophils in the lungs compared to mice infected with low-pathogenic viruses. 16Detection of replicating viruses in lung

Impact of H1N1 or S. aureus on secretion of cytokines, chemokines, and interferons from human M1-and M2-MDMs
In order to study innate host cell responses to IAV H1N1 or S. aureus infections, especially inflammation-related protein and LM signatures, primary M1-and M2-MDMs were infected according to general standard infection models with IAV. 16,17,28For this purpose, the well characterized and suitable A/Puerto Rico/8/34 H1N1 strain, isolated from humans and propagated on Madin Darby canine kidney (MDCK) cells, was employed that induces reasonable immune responses in murine and human cell cultures.We assessed M1-/M2-MDM responses due to H1N1 or S. aureus infection after short-term incubation periods (<5 h), in order to exclude detrimental effects on the MDM due to excessive viral or bacterial replication.Thus, H1N1 were added to MDMs at a multiplicity of infection (MOI) of 5, and after 30 min of internalization, MDMs were washed and incubated for another 4 h.In parallel, S. aureus at an MOI of 10 was added to MDMs for 4 h to evoke host cell responses; this rather low MOI is less cytotoxic than an MOI of 50 that was used in previous studies as stimulus for induction of LM formation within 1.5-3 h. 22Production of cytokines, chemokines, and IFNs as inflammation-related protein mediators was measured in the supernatant by using the human antivirus response panel LEGENDplex.H1N1 did not significantly induce cytokine, chemokine, or IFN release from M1-or M2-MDMs (Figure 1).In contrast, upon exposure to S. aureus, M1-MDMs strongly released the pro-inflammatory cytokines IL-1b, IL-12, TNF-a, and granulocyte-macrophage colony-stimulating factor (GM-CSF) (but not IL-6) as well as the anti-inflammatory IL-10, while the chemokines IL-8 and IP-10 were not or hardly elevated (Figure 1).In M2-MDM, S. aureus caused overall less pronounced effects, markedly inducing only the release of IL-10, IL-12, TNF-a, and GM-CSF, with minor impact on IL-1b and IL-6 as well as on the chemokines IL-8 and IP-10.Note that the latter chemokines were secreted in high amounts already from resting M1-MDM.Also the absolute amounts of some cytokines and of IFN-g that inhibits IAV attachment and replication, 29 also differed between the MDM phenotypes.Moreover, after S. aureus (but not H1N1) exposure of both M1-and M2-MDM, a strong IFN response was observed, with substantial production of type I IFN-a2 and IFN-b, and type III interferon IFN-l but no significant elevation of IFN-g.Possibly, longer incubation of IAV-exposed MDM is required to evoke type I IFN-a2 and -b release, as observed for epithelial cells where IFN-b was secreted 8 to 12 h upon IAV exposure. 30Together, exposure of M1/M2-MDM to H1N1 failed to elevate cytokine, chemokine, and IFN production, while S. aureus strongly increased all analyzed cytokines and type I/III IFNs in M1-MDM, and, except IL-1b and IL-6, also in M2-MDM.

H1N1 specifically evokes prostaglandin E 2 production in human M2-MDMs and induces COX-2 expression
By employing ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS-MS)-based targeted LM metabololipidomics, we investigated the differential LM signature profiles (25 LMs and three PUFAs) generated and released by M1-and M2-MDMs after exposure to H1N1 or S. aureus.As previously reported, S. aureus evokes differential LM profiles in human M1-and M2-MDMs, where M1 produced mainly COX-derived PGs and 5-LOX-mediated LTs, while M2-MDMs formed 15-LOX-derived SPM and their precursors, 22 confirmed in the present study (Figure 2A).For comprehensive LM analysis, we utilized meaningful LM radar plots (Figure 2A) and heat maps (Figure 2B), revealing that compared to S. aureus, H1N1 induces LM signatures (i) to an overall much lower extent in both macrophage phenotypes and (ii) with a different pattern of individual LM (Figures 2A and 2B).Thus, in contrast to S. aureus, H1N1 neither elevated formation of 5-LOX-mediated LTs nor of 12/15-LOX-derived SPMs (resolvin (Rv)D5 or lipoxin (LX)A 4 ) or SPM precursors (17-hydroxydocosahexaenoic acid (HDHA), 15-hydroxyeicosapentaenoic acid (HEPE) or 15-hydroxyeicosatetraenoic acid (HETE)), indicating that H1N1 did not activate LOX signaling pathways (Figures 2A and 2B).In M1-MDM, H1N1 did not markedly increase any of the 25 LM and the three PUFAs versus mock infection (Figure 2B).However, In M2-MDMs, H1N1 specifically elevated COX-derived PGE 2 formation (approx.40-fold), and also the production of other COX-derived PGs, i.e., PGD 2 and PGF 2a , was significantly increased (approx.6-to 8-fold) in M2-MDMs, and to a minor extent in M1-MDMs, while TXB 2 was the overall most abundant COX product in M1-and M2-MDM (Figure 2C) in line with previous MDM studies 12,20,21 but not enhanced by H1N1 in either phenotype (Figure 2C).By analysis of LM signatures of human macrophages infected with the IAV H3N2 subtype (A/Wisconsin/67/2005), we found similar effects as observed with H1N1 (Figure S1).We next investigated mRNA expression levels of LM-biosynthetic enzymes and pro-inflammatory cytokines (Figure 2D).We found that H1N1 significantly increased mRNA levels of COX-2 in M2-MDMs (about 15-fold).Of interest, expression of IL-6 mRNA but not of IL-1b and TNFa was significantly increased in M2-MDMs, whereas M1-MDMs did not respond in this respect (Figure 2D).In contrast to the increased COX-2 mRNA levels, the absolute amounts of COX-1, COX-2, and mPEGS-1 proteins were not altered within 4 h post H1N1 infection of M1-or M2-MDMs (Figure 2E and Data S1).Furthermore, we found that COX activity and COX-1/2 protein levels were not further increased within 6 h post H1N1 infection compared to 4 h, indicating that formation of COX products, especially of PGE 2 , takes place shortly after H1N1 exposure, as initial response of the M2-MDMs (Figure S2 and Data S1).Absolute data for mock-, single H1N1-or S. aureus-infected M1-and M2-MDMs, given in pg/2 3 10 6 cells, are shown in Figures 4B and S3.

Co-infection of MDMs with H1N1 and S. aureus has a minor impact on bacterial or viral titers
Influenza-associated bacterial co-infections contribute to immune disorders, including failed antibacterial immune response to clear microbial infection, 8 therefore, we studied the effect of co-infection on immune functions of human MDMs.We investigated whether co-infection of MDMs with both H1N1 and S. aureus impacts the load of each pathogen versus single infection.M1-and M2-MDMs were infected with H1N1 (MOI 5) or kept untreated for 30 min, each, cells were washed and then infected with S. aureus (MOI 10) or kept untreated for 4 h, each (Figure 3A).We first analyzed the extracellular numbers of plaque-forming units (PFU) and colony-forming units (CFU) to determine viral and bacterial loads in the extracellular environment (supernatants) of MDMs, respectively (Figures 3B and 3C).Neither viral (Figure 3B) nor bacterial titers (Figure 3C) were significantly affected upon co-infection compared to single-infected MDMs.To investigate intracellular bacterial loads, we used gene-modified S. aureus expressing GFP to analyze fluorescent contents in the MDMs after infection, and we found that the mean fluorescent intensity in M2-MDMs was slightly increased upon co-infection compared to single-infection, whereas the MFI in M1-MDMs was unaffected, suggesting increased bacterial phagocytosis of H1N1-predisposed M2-MDMs (Figure 3D); whether or not this elevation is due to PGE 2 remains to be investigated.

Impact of co-infection of S. aureus-treated MDMs with H1N1 on the release of inflammation-related mediators
To examine the inflammatory status of H1N1-predisposed MDMs after infection with S. aureus, we determined cytokine, chemokine, and IFN levels in the supernatants compared to single S. aureus infection.We found that especially in M2-MDMs, pro-inflammatory cytokines such as IL-1b, IL-6, and TNF-a but also the type II interferon IFN-g are elevated by tendency, without changes in M1-MDMs (Figure 4A).We also analyzed the impact of H1N1 on the LM signatures of MDMs co-infected with S. aureus.Note that S. aureus causes massive formation of various COX-and LOX-derived LM in MDMs, due to robust LM pathway induction by released exotoxins. 21,22In M1-MDMs, co-infection with H1N1 led to only moderate elevation of S. aureus-induced LM release for all determined LM classes (around 10-30% higher versus single-infection with S. aureus), reaching significance solely for 11-HETE and 13-HDHA, while the PUFA levels were not altered.However, by comparison of the H1N1-infected cells versus cells that were coinfected with S. aureus, increased LM production was observed due to the impact of the bacteria (Figures 4C and S3).In contrast, in M2-MDM, H1N1-predisposal caused strong increases of certain LMs, especially PGE 2 (up to 402% versus S. aureus single-infected MDMs) and other COX-derived PGs.While 5-LOX products were hardly elevated, 15-LOX-mediated LM such as the SPM precursor 17-HDHA, 15-HEPE, and 15-HETE, and the SPM RvD5 were significantly increased in M2-MDMs due to additional H1N1, although to a minor degree (%164%) as compared to PGE 2 (Figures 4B and 4C).Furthermore, compared to single H1N1 infection, the co-infection with H1N1 and S. aureus led to increased LM formation overall, except for PGE 2 (and LXA 4 ), indicating that H1N1 is the predominant elicitor for PGE 2 formation in M2-MDMs (Figures 4B and 4C).

DISCUSSION
Here, we demonstrate that infection of human MDMs with an IAV H1N1 subtype in vitro selectively and strongly elevates formation of PGE 2 in M2-MDM but not in M1-MDM.In contrast, 5-LOX-derived pro-inflammatory LTs or pro-resolving 15-LOX-derived LM and also cytokine/interferon release was not altered in either MDM phenotype.Such selective elevation of PGE 2 production by IAV is in contrast to the broad COXand LOX-activating effects caused by human pathogenic bacteria (i.e., E. coli or S. aureus) in MDM, implying that different mechanisms of LM induction by IAV are operative. 21,31M1-MDMs, with superior COX-2/mPGES-1 expression over M2-MDM, possess higher capacities to produce PGE 2 upon exposure to bacterial pathogenicity, 21 but upon infection with H1N1, the PGE 2 levels of IAV-stimulated M2-MDM surpass those of the M1 counterpart.This effect might be of particular relevance for in vivo situations where M2 macrophages dominate over the M1 phenotype.As PGE 2 was shown to induce the switch from pro-inflammatory toward inflammation-resolving LM, 32,33 it is tempting to speculate that also the IAV-mediated induction of PGE 2 formation in pro-resolving M2 macrophages might contribute to such LM class switching.Our data show that the selective induction of PGE 2 in M2-MDMs by H1N1 is still evident after co-infection with S. aureus that utilizes exotoxins to elicit the formation of broad LM profiles in these cells. 22The antigens and molecular mechanisms of how IAV causes such selective PGE 2 induction, even on top of exotoxin-stimulation, are unknown and require further investigations.PGE 2 is a crucial mediator that induces fever in virus-infected organisms and regulates viral replication, 34,35 activating several G-proteincoupled receptors (GPCR), namely EP1-4 which mediate pro-and anti-inflammatory effects of PGE 2 .Our finding that PGE 2 is specifically produced after IAV infection is consistent with observations from IAV-infected mice in vivo, 36 especially in macrophages, 37 but also with results from other airway-related virus infections, such as respiratory syncytial virus (RSV) in epithelial cells 38 or SARS-CoV-2 in humans. 39It was shown before that IAV rapidly elevates the COX-2 mRNA levels in human lung epithelial cells peaking at 2-4 h but later on decline, 34 which fits to the elevated mRNA levels of COX-2 in MDM at 4 h in our study.Moreover, a shift toward pro-inflammatory COX-2 expression was also seen in human alveolar epithelial cells and PBMCs of IAV-infected patients. 40Lee et al. found differences in the induction of COX-2 in macrophages by different IAV strains, where the avian influenza H5N1 was more robust as compared to H1N1 subtype. 40We tested IAV H1N1 and H3N2 and found that these evoke similar responses in MDM, indicating specific mechanisms to induce PGE 2 , which suggests an IAV class effect.
It was shown before that secretory factors in the supernatants of IAV-infected macrophages induce mRNA expression of COX-2 and of other pro-inflammatory cytokines such as TNF-a, IL-6, and IL-1b in A549 cells. 40We found similar up-regulatory effects of IAV on COX-2 and IL-6 mRNA levels in M2-MDMs, while M1-MDMs were unaffected in this respect.Possible reasons for these differences between the phenotypes could be: (i) activation of TLRs and/or PRRs that are present on M2-but not on M1-MDMs, (ii) pre-activation of PRRs in M1-MDMs due to lipopolysaccharide (LPS) treatment during polarization and, thus, lack of further responsiveness as compared to M2-MDM, (iii) higher susceptibility of M2-MDM for IAV invasion, or (iv) production of M2-MDM-specific mediators which cause COX-2 and IL-6 induction. 41,42Based on our results, at least a higher susceptibility of M2-MDMs for IAV invasion can be excluded as reason, since the H1N1 viral burden of M1-and M2-MDMs was about the same.It is intriguing that the substantial and selective increase of PGE 2 was not connected to elevated mPGES-1 mRNA or protein levels, neither in M1-nor in M2-MDMs.Among the three PGES isoforms, the mPGES-1 is an inducible isoform that is frequently coupled to the induction of COX-2. 43Like for COX-2, M1-MDMs express also high levels of mPGES-1, in contrast to the M2-phenotype, 44 but in M2-MDMs only COX-2 expression (but not mPGES-1) was elevated, at least on the mRNA level.Interestingly, while COX-2 mRNA was significantly increased in M2-MDMs by H1N1, elevated translation toward COX-2 protein expression was not readily evident.We reported earlier that IAV could destabilize COX-2 mRNA levels in human epithelial cells after 2 to 4 h and thus only transiently elevated COX-2 protein. 34It seems that IAV-induced elevation of COX-2 protein in M2-MDM is transient as well, which might explain why the protein levels of COX-2 appeared not altered in MDMs at 4 and 6 h versus uninfected cells.Finally, IAV infections may promote additional events that specifically favor the coupling of COX-2 with PGES isoforms to stimulate PGE 2 formation on the enzyme activation level; increased AA substrate supply can be excluded as levels of free AA and of other AA-derived LOX products were not elevated.Our data show that infection of MDMs with H1N1 does not increase the secretion of inflammation-related cytokines/IFN (including IL-6) regardless of the phenotype, yet the mRNA of IL-6 was selectively increased in M2-MDMs.This indicates that the translation into IL-6 protein was comparatively limited at least within the short time frame of only 4 h.Such translation can be stimulated by low amounts of LPS or bacterial infection, which potentiates cytokine release. 45,46Along these lines, our data from the co-infection experiments using S. aureus and H1N1 showed that in the presence of bacteria, at least some pro-inflammatory cytokines (i.e., IL-1b, IL-6, and TNF-a) and type II interferon IFN-g are elevated on the protein level by H1N1 in M2-MDMs by tendency.A similar impact of S. aureus was observed for LM formation.Thus, compared to the strong increase of PGE 2 by H1N1 in M2-MDMs, only PGD 2 but no other LM was elevated in cells infected with IAV alone.However, when co-infected with S. aureus, besides elevating PGE 2 and PGD 2 , IAV caused increased formation of RvD5 and of 15-LOX-derived 17-HDHA, 15-HEPE, and 15-HETE in M2-MDMs, albeit only 1.5-to 1.6-fold as compared to PGE 2 that was elevated by 4-fold.
Bacterial superinfections are characterized by the loss of epithelial barrier function and altered innate immune functions leading to an overall disturbed host immune response. 47Another detrimental effect of IAV is the predisposition of the lung endothelium, causing leakage after exposure to S. aureus.PGE 2 is a likely candidate as respective mediator, resulting in acute respiratory distress syndrome (ARDS), the predominant cause of death in patients with bacterial superinfections. 48,49Since alveolar macrophages in the airways feature an M2-like phenotype, the strong elevation of PGE 2 that we observed in M2-MDMs after IAV, with or without S. aureus, might be of pathophysiological relevance for ARDS.
PGE 2 can affect cytokine production during the inflammatory process, acting as a highly potent immunomodulatory mediator. 50In view of the robust PGE 2 elevation and the potent pyrogenic features of PGE 2 , the COX-2 is a potential therapeutic target in order to reduce PGE 2 levels in the infected tissue.For example, treatment of IAV-infected mice with NSAIDs to lower PGE 2 levels by COX-2 inhibition reduced pathology, 51,52 while the administration of PGE 2 reversed this phenotype. 37Moreover, IAV induced less severe illness in COX-2 À/À compared to wild-type and to COX-1 À/À mice, suggesting that COX-2 deficiency is beneficial, whereas COX-1 deficiency is detrimental to the host during influenza virus infection. 53This supports specific inhibition of the COX-2 pathway as pharmacological strategy.However, COX inhibition to lower PGE 2 levels in IAV-infected patients, as commonly practiced with NSAIDs like ibuprofen or aspirin to reduce fever, is still under debate.For example, PGE 2 may have anti-viral potential visualized by diminished viral replication rates in A549 cells. 34ogether, our results demonstrate that the H1N1 IAV evokes robust and selective biosynthesis of PGE 2 in human M2-MDMs as a rapid response mechanism while cytokine/IFN release and formation of other LM was essentially unaltered.Interestingly, upregulation of PGE 2 in M2-MDM persists after co-infection with S. aureus, although other LM, such as the SPM RvD5 and its precursor 17-HDHA, were increased as well.Notably, M1-MDMs, as primarily anti-microbial macrophage phenotype, did not respond to IAV infection, regardless of co-infection with S. aureus, neither in terms of cytokine/IFN nor LM formation.

Limitations of the study
A limitation of this work is the investigation of isolated macrophages in monocultures infected with pathogens, while the aspect of the epithelial layer is missing but may provide a more comprehensive analysis of the host response to infection.Also, the impact of the robustly generated PGE 2 by M2-MDMs on the biological outcome of IAV or superinfected organisms in vivo remains to be investigated, especially responses of primary human alveolar macrophages or other primary M2-like cells.Furthermore, we investigated the first/immediate response of M1-and M2-MDMs on viral infections within only 4 or 6 h, neglecting replication effects of IAV themselves at later time points due to technical/experimental issues.It also remains challenging to study different primary macrophage populations from tissues with mixed M1-and M2 phenotype.Comparative analysis of M2-MDMs with human alveolar macrophages, the first responders in the lung airway upon IAV infection, may further support the pathophysiological relevance of our findings.But also studies with alveolar macrophages from rodents, for example, may help to confirm the selective PGE 2 induction by IAV in such cells.Also, more detailed analysis is needed to specify this effect on Orthomyxoviridae overall by comparing influenza A, B, and C viruses, and the influence of hemagglutinin and neuraminidase and their antigenic variations due to mutation, specific for the antigenic drift, remain to be investigated.Future studies aiming at evaluating the involvement of TLRs (i.e., TLR-3 and -7, specific for IAV infections), inflammasome activation, or NF-kB, together with translation toward in vivo IAV infection models, are planned.

Analysis of bacterial phagocytosis
To determine intracellular bacterial titers S. aureus/USA300/GFP were used.After the indicated time points, supernatant was removed and cells were washed with PBS and incubated with PBS plus 5 mM EDTA to detach macrophages for 20 min at 37 C. Cells were centrifuged (400 g, 5 min, RT) and then fixed with 4% paraformaldehyde for 20 min.Intracellular GFP fluorescence was measured by using the BD FACSLyric flow cytometer (BD, Heidelberg, Germany).Data analysis was performed with FLOWJO (BD).

Measurement of bacterial colony-forming units (CFUs)
To determine extracellular bacterial titers, supernatants of MDM incubations (100 mL) were plated on brain-heart infusion (BHI) agar plates and incubated overnight at 37 C. CFUs were counted and calculated.Each experiment was performed in technical duplicates.

Measurement of extracellular virus
At the indicated time points, supernatants of MDM incubations were collected to assess the number of infectious particles by standard virus plaque assay.Briefly, the Madin-Darby canine kidney (MDCK II) cell line was used for this purpose and grown in MEM to 90% confluence in sixwell dishes.Cells were washed and infected with serial dilutions of the supernatants in PBS/BA (PBS containing 0.2% BSA, 1 mM MgCl 2 , 0.9 mM CaCl 2 , 100 U/mL penicillin and 0.1 mg/mL streptomycin) for 30 min at 37 C and 5% CO 2 .The inoculum was aspirated and cells were overlaid with 2 mL MEM containing 0.2% BSA, 1 mM MgCl 2 , 0.9 mM CaCl 2 , 100 U/mL penicillin and 0.1 mg/mL streptomycin supplemented with 0.6% agar (Oxoid, Hampshire, United Kingdom), 0.3% DEAE-dextran (Amersham Pharmacia Biotech, Freiburg, Germany), and 1.5% NaHCO 3 (Gibco Invitrogen, Karlsruhe, Germany).After incubation at 37 C and 5% CO 2 for 2 to 3 days, virus plaques were visualized by staining with neutral red (Sigma-Aldrich, Munich, Germany).

QUANTIFICATION AND STATISTICAL ANALYSIS
Results are expressed as means G standard error of the mean (SEM) of n observations, where n represents the number of experiments with separate donors, performed on different days, as indicated.Analyses of data were conducted using GraphPad Prism 8 software (San Diego, CA).Two-tailed t test was used for comparison of two groups.For multiple comparison, one-way analysis of variance (ANOVA) with Dunnett's or Tukey's post hoc tests were applied as indicated.The criterion for statistical significance is p < 0.05.

Figure 3 .
Figure 3. Impact of co-infection of MDMs with H1N1 and S. aureus on bacterial or viral titers (A) Scheme depicting the experimental settings of the in vitro co-infection model.(B and C) Human M1-or M2-MDMs (2 3 10 6 cells) were infected with H1N1 (PR8; MOI = 5) or mock for 30 min, then washed, and coinfected with S. aureus (MOI = 10) or mock for 4 h at 37 C. Supernatants were collected to determine extracellular viral and extracellular bacterial titers.(B) Extracellular H1N1 are represented by plaque-forming units (PFU), data are shown in bar charts as mean + SEM with single values, ns = not significant, ratio-paired t-test; H1N1-infected MDMs versus co-infected MDMs, n = 3. (C) Extracellular S. aureus are represented by colony-forming units (CFU), data are shown in bar charts as mean + SEM with single values, ns = not significant, ratio-paired t-test; S. aureus-infected MDMs versus co-infected MDMs, n = 3. (D) Human M1-or M2-MDM (2 3 10 6 cells) were infected with H1N1 (PR8; MOI = 5) or mock for 30 min, washed, and then co-infected with GFP-expressing S. aureus (MOI = 10) or mock for 4 h at 37 C. Cells were detached and GFP-expressing S. aureus in the MDMs were measured by flow cytometry.Data are presented as mean fluorescence intensity (MFI) of M1-and M2-MDMs.Results are given as means + SEM with single values in bar charts; n = 3; **p < 0.01, ratio-paired t-test; S. aureus infected MDMs versus co-infected MDMs.

Figure 4 .
Figure 4. Impact of co-infection of S. aureus-treated MDMs with H1N1 on the release of inflammation-related mediators (A and B) Human M1-or M2-MDMs (2 3 10 6 cells) were infected with H1N1 (PR8; MOI = 5) or mock for 30 min, washed, and then treated with S. aureus (MOI = 10) or mock for 4 h at 37 C, as indicated.(A) Supernatants were collected to assess secreted cytokines via flow cytometry.Results of co-infection with H1N1 are shown as -fold change against infection with S. aureus alone, given as means + SEM with single values.Raw data were log-transformed for statistical analysis; ns = not significant, ratio-paired t-test; S. aureus-infected MDMs versus coinfected MDMs, n = 3. (B and C) Released LMs from M2-MDMs in the supernatants were assessed by UPLC-MS-MS.LM formed by MDM upon infection with mock, single H1N1, single S. aureus, and co-infection with H1N1 and S. aureus are shown in pg/2 3 10 6 cells as mean G SEM, and the fold change of H1N1/S.aureus-treated samples versus H1N1 single or S. aureus single infection are given and visualized in a heatmap (B), data for mock-, H1N1-and S. aureus-treated cells are taken from experiments shown in Figures 2A-2C; COX product formation is visualized in bar charts (C).Data were log-transformed for statistical analysis; *p < 0.05; **p < 0.01; ratio-paired t-test; coinfected MDMs versus H1N1-infected MDMs or S. aureus-infected MDMs, n = 3 for M1-MDMs, n = 6 for M2-MDMs.See also Figure S3.