Influenza Causes MLKL-Driven Cardiac Proteome Remodeling During Convalescence

Abstract

Rationale:

Patients with and without cardiovascular diseases have been shown to be at risk of influenza-mediated cardiac complications. Recent clinical reports support the notion of a direct link between laboratory-confirmed influenza virus infections and adverse cardiac events.

Objective:

Define the molecular mechanisms underlying influenza virus-induced cardiac pathogenesis after resolution of pulmonary infection and the role of necroptosis in this process.

Methods and Results:

Hearts from wild-type and necroptosis deficient (MLKL-KO) mice were dissected twelve days after initial Influenza A virus (IAV) infection when viral titers were undetectable in the lungs. Immunofluorescence microscopy and plaque assays showed presence of viable IAV particles in the myocardium without generation of interferon responses. Global proteome and phosphoproteome analyses using high resolution accurate mass based LC-MS/MS and label-free quantitation showed that the global proteome as well as the phosphoproteome profiles were significantly altered in IAV-infected mouse hearts in a strain independent manner. Necroptosis deficient mice had increased survival and reduced weight loss post-IAV infection, as well as increased antioxidant and mitochondrial function, indicating partial protection to IAV infection. These findings were confirmed in vitro by pre-treatment of human and rat myocytes with antioxidants or necroptosis inhibitors, which blunted oxidative stress and mitochondrial damage after IAV infection.

Conclusions:

This study provides the first evidence that the cardiac proteome and phosphoproteome are significantly altered post pulmonary influenza infection. Moreover, viral particles can persist in the heart after lung clearance, altering mitochondrial function and promoting cell death without active replication and interferon responses. Finally, our findings show inhibition of necroptosis or prevention of mitochondrial damage as possible therapeutic interventions to reduce cardiac damage during influenza infections.

Graphical Abstract

INTRODUCTION

The top three global causes of death worldwide are heart disease, cancer and respiratory tract (RT) infections^{1, 2}. RT infections are commonly caused by influenza virus, a class of enveloped RNA viruses that trigger an acute respiratory illness occurring in local outbreaks, seasonal epidemics and/or pandemics^{3, 4}. In the United States alone, influenza infections lead to over 500,000 hospitalizations, and over 36,000 deaths each year^{5, 6}. A key factor contributing to the high mortality rate are major adverse cardiac events (MACE) that occur during hospitalization or convalescence, which are more likely to arise in patients with pre-existing cardiac conditions⁷. Clinical studies have shown that elderly patients admitted to the hospital for severe viral and/or bacterial pneumonia experience 10–25% mortality rates; of these, up to 40% perish within one year⁸, and up to 20% experience MACE that can last for up to 10 years after discharge⁹. Accumulating evidence has tightened the association between MACE and infectious diseases including influenza, pneumococcal pneumonia, or bacteremia^9–12. The majority of the clinical and basic research regarding MACE and pneumonia have studied community acquired pneumonia caused by the Gram-positive bacterium Streptococcus pneumoniae¹³, but not respiratory infections from seasonal, epidemic, or pandemic influenza viruses¹⁰. Therefore, it is essential to untangle the potential molecular links between viral pulmonary infections and alterations to cardiac biology, as illustrated by the two most recent pandemics caused by H1N1 influenza (2009) and SARS-CoV-2 (2019), when many patients experienced adverse cardiac events such as cardiac infarct, myocarditis, ischemic injury and stroke^{10, 14}.

While the molecular and epidemiological effects of influenza infection in pulmonary tissue are well-documented, the molecular effects of influenza in other organs remains unclear^{15, 16}. Influenza virus has been shown to be cardiotropic in multiple models of disease^17–25, with virus disseminating from the lungs to cardiac tissue, especially during severe infections, to cause myocarditis, ischemia, and cardiac dysfunction^{18–21, 24}. Fatalities from typical seasonal influenza have shown myocarditis in up to 48% of all autopsies^22–25, and studies from the 1918 H1N1 influenza pandemic showed severe cardiac damage in 90% of fatalities¹⁷. The 1918 influenza strain was later used to infect macaques, and viral particles disseminated from the lungs to the heart after acute respiratory distress and persisted in cardiac tissue up to 8 days post infection¹⁹, the last time point collected. This suggests that pandemic influenza strains may lead to more exacerbated, prolonged cardiac pathology^{17, 19, 24, 25}. In rodent models, both lab-adapted influenza H1N1 strains like A/Puerto Rico/8/1934 (PR8) and pandemic strains A/California/7/2009 and A/England/195/2009 induce clinical signs (weight loss) as well as histological and molecular markers of pulmonary disease (inflammation and cell death)^26–28. In this study, we utilize the mouse adapted A/California/7/2009 strain as a prototype pandemic strain and PR8 as a point of comparison. While cardiac tropism by influenza viruses has been observed for over a century, the underlying molecular mechanisms that promote cardiac pathologies remain undefined.

Cell death is a key step in host response to viral and bacterial pathogens²⁹. The ability of cells to undergo programmed cell death has extensive influence on recovery. One major programmed cell death pathway with direct implications in host responses to viral pathogens is programmed necrosis, i.e. necroptosis³⁰. Necroptosis is a highly inflammatory form of cell death regulated by the receptor-interacting serine/threonine-protein kinases (RIPK) 1 and 3, and the effector molecule mixed lineage kinase domain-like protein (MLKL). Importantly, necroptosis (especially the activity of RIPK3) evolutionarily promotes host survival during viral infections, in contrast to its detrimental role during severe bacterial infection^{31, 32}. During influenza A virus (IAV) infections, the virus utilizes self-encoded proteins to block canonical apoptosis. Host cells then attempt to halt viral replication by expressing RIPK3 to initiate necroptosis and/or re-initiate apoptosis¹⁵. This suggests that inhibition of the necroptosis effector molecule MLKL may promote apoptotic activity, which may in turn enhance host antiviral responses. To date, necroptosis has not been implicated in cardiac responses to influenza infection.

Here, to test if influenza virus infection has a prolonged effect on cardiac health during convalescence, we employed a mouse model of IAV infection. We combined proteomic, phosphoproteomic, and molecular techniques to describe the molecular changes in the heart after clearance of pulmonary influenza infection. Finally, we used mice deficient in necroptosis to investigate the role of this pathway in cardiac tissue alterations after pulmonary resolution of influenza infection.

METHODS

Data Availability.

The data that support the findings of this study are available from the corresponding author upon reasonable request. All data from the proteomic analyses have been made publicly available at ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifiers PXD017936, PXD016412, and PXD015992.

Supplemental Methods and data.

Supplemental figures and detailed experimental procedures are available online.

RESULTS

Influenza infection induces cardiomyocyte necroptosis, oxidative stress and cellular metabolic breakdown without active viral replication.

To determine if myocytes are susceptible to influenza virus infection, we challenged rat H9c2 myocytes and human AC16 cardiomyocytes with pandemic influenza A H1N1 A/California/7/2009 (IAV) or A/Puerto Rico/8/1934 (PR8) and assessed cellular cytotoxicity 48h after infection. We observed increased cellular cytotoxicity (Fig. 1A, Online Figure I. A, B) without active viral replication (Online Figure I. C), suggesting that myocytes could die upon IAV infection while promoting abortive replication^{33, 34}. Necroptosis has been shown to have a crucial role in suppressing viral replication in vitro and in vivo^{15, 35}. 48h after the IAV challenge of rat and human cardiomyocytes, we observed increased immunofluorescent staining of phosphorylated MLKL, the necroptosis effector molecule (Fig. 1B–C, Online Figure I. D), indicating enhanced necroptosis activity in IAV-infected myocytes. Pre-treatment of myocytes with necroptosis inhibitors ponatinib, pazopanib³⁶ or necrosulfonamide³⁷ led to a significant decrease in cell cytotoxicity (Fig. 1D–F). Taken together, necroptosis inhibition can partially rescue IAV-induced myocyte cell death.

Figure 1: Influenza A virus infection leads to programmed necrosis, oxidative stress and metabolic breakdown in cardiomyocytes.

(A) LDH release of H9c2 rat myocytes were infected with an MOI of 0.1 of IAV strain Influenza A virus/California/7/2009 (IAV) for 48-h. (B) Immunofluorescent staining for pMLKL (green) in H9c2 rat myocytes infected with IAV. Cell nucleus was stained in blue. White bar = 50 μm. (C) Mean fluorescence intensity (MFI) of pMLKL. (D) LDH cytotoxicity assay of H9c2 infected with IAV at an MOI of 0.1 for 48-h. Cells were treated with two necroptosis inhibitors: ponatinib and pazopanib (10 μM each) for 1h before infection. (E) Propidium iodide and (F) LDH cytotoxicity assay staining of human AC16 myocytes infected with IAV at an MOI of 0.1 for 48-h. Cells were treated with MLKL inhibitor necrosulfonamide (10 μM each) for 1h before infection. (G) 8-oxoguanine (8-OXO, green) was stained in H9c2 rat myocytes infected with IAV. Cell nucleus was stained in blue. White bar = 50 μm. (H) Mean fluorescence intensity (MFI) of 8-OXO staining. (I) LDH cytotoxicity assay of H9c2 infected with IAV at an MOI of 0.1 for 48-h. Cells were treated with Catalase (10 μM), Tempol (10 μM), Rotenone + Thallium trifluoroacetate (Rot tha, 10 nM/mL/10 nM/mL), Allopurinol (Allu, 10 nM/mL), Apocynin (Apo, 1 μM/mL), Mefenamic acid (Mefe, 20 nM/mL). (J) ATP levels and (K) ROS/superoxide levels of IAV-infected H9c2 myocytes treated with ponatinib (10 μM), compared to mock treated and uninfected H9c2. RLU: relative light unit of luminescence. Mann-Whitney U tests were applied for two-group comparisons. Kruskal-Wallis test with Dunn’s multiple-comparison post-test. Asterisks denote the level of significance observed: * = p ≤ 0.05; ** = p ≤ 0.01; *** = p ≤ 0.001, **** = p ≤ 0.0001.

During pulmonary infection, IAV promotes oxidative stress that leads to tissue damage, enhanced inflammatory responses, and cell death³⁸. To determine if IAV infection in cardiac tissue leads to oxidative stress induced damage, we stained H9c2 myocytes 48h post-IAV infection against 8-Oxoguanine (8-OXO), a marker of DNA damage resulting from reactive oxygen species (ROS)³⁹. We saw a significant increase of 8-OXO staining, indicating elevated ROS-mediated damage (Fig. 1G–H). We then attempted to rescue this by pre-treating myocytes with general antioxidants (catalase or tempol), mitochondria-dependent [rotenone + thallium trifluoroacetate (mitochondria) or apocynin (NADPH)], or mitochondria-independent [allopurinol (xanthine oxidase-dependent) or mefenamic acid (cyclooxygenase-dependent)] antioxidants prior to IAV infection. Pre-treatment with general antioxidants or mitochondria-dependent ROS inhibitors protected myocytes from IAV-induced cell death, but pre-treatment with mitochondria-independent ROS inhibitors had no effect (Fig. 1I), indicating that toxic cellular ROS originated from the mitochondria. IAV-infected H9c2 and AC16 myocytes showed decreased ATP production (Fig. 1J, Online Figure I. E) that could be rescued by necroptosis inhibition (Fig. 1J). We also observed a stark decrease in ROS production in IAV-infected myocytes after necroptosis inhibition (Fig. 1K). Taken together, our data suggest that influenza infection induces necroptosis and promotes mitochondria-dependent oxidative stress.

MLKL deficiency improves antioxidant and mitochondrial activity.

It has been shown that RIPK3 KO mice or MLKL/FADD double-KO mice are more susceptible to IAV infection than wild-type, whereas MLKL KO mice were not susceptible to lethal disease¹⁵. Recently, our group showed that inhibition of MLKL can protect against pulmonary injury induced during bacterial infection secondary to IAV³⁸. Together, these reports suggest that RIPK3 is required for an efficient response against influenza and that inhibition of MLKL can be an efficient approach to reduce IAV-induced necrotic cell damage and virus dissemination^{15, 30, 40}. To test if necroptosis inhibition could potentially protect against IAV-induced cardiac damage we analyzed baseline proteomes of uninfected WT (C57BL/6) and MLKL deficient mice (MLKL KO) using High-Resolution Accurate-Mass (HR/AM) LC-MS/MS platform and a label-free quantitation (LFQ)^{41, 42}. 278 proteins were significantly changed between the two groups (Online Figure II. A). Among them, up-regulated proteins in MLKL KO hearts are largely associated with oxidation-reduction, fatty acid/glutathione metabolism, responses to oxidative stress, and general metabolic processes after Gene Ontology (GO) analysis (Online Figure II. B). Global inhibition of MLKL leads to a baseline increase in mitochondrial and antioxidant activity, which suggests that mice lacking MLKL may be protected against the cellular changes induced by IAV infection (Fig. 1).

It has been shown that during S. pneumoniae infection, bacteria can translocate to cardiac tissue and modulate local inflammatory responses¹³. However, it is not known whether IAV can translocate to heart, or if virus persists in in cardiac tissue after its clearance from the lungs. To test the potential persistence of the virus in the heart, we designed a study in which 6-to-8-week-old WT and MLKL KO mice⁴³ were infected with 250 plaque forming units (PFU) of IAV strain A/California/7/2009 and followed for twelve days (Fig. 2A). Ten days after initial infection, IAV was undetectable in the pulmonary tissue (Online Figure III.), and MLKL KO mice showed increased survival (Fig. 2B) and reduced weight loss (Fig. 2C) during the acute phase of the infection. After 10 days, 76% of WT mice (n = 13) and 92% of MLKL KO mice (n = 12) had recovered from the initial infection (Fig. 2B–C). At day twelve, hearts were collected. Hematoxylin and eosin staining of frozen sections showed some structural changes in cardiac tissue of uninfected vs IAV-infected mice (Online Figure IV. A). Immunofluorescence staining (Online Figure IV. B) and plaque assays (Fig. 2D) showed presence of IAV viral particles in both WT and MLKL KO cardiomyocytes twelve days after initial infection. To test if the presence of IAV drives inflammation in the cardiac tissue during convalescence, we measured the level of inflammation-associated cytokines via enzyme-linked immunosorbent assay (ELISA). No significant changes in TNF-α, IFN-ƛ IFN-Ɣ, IFN-α, IL-1β, IL-10 (Fig. 2F) or IFN-β (below level of detection <5 pg/mL, Online Figure V.) were observed. TNF and IFN responses have been shown to promote necroptosis during viral infection⁴⁴. We tested whether cytokine production during acute infection may exacerbate IAV-induced tissue damage by pre-treating human AC16 cardiomyocytes with either TNF-α or IFN-β before challenge with IAV. Inhibition of MLKL with necrosulfonamide protected cardiomyocytes after challenge with either cytokine in combination with IAV (Online Figure VI. A–B). Inhibition of TNF receptor 1 or blocking of TNF-α by pre-treatment of cells with R7050 or SPD304, respectively, reduced influenza-induced cell death in vitro (Online Figure VI. C). During acute infection, cardiac damage is exacerbated by TNF production, but cytokine levels return to baseline during convalescence despite persistence of IAV in cardiac tissue.

Figure 2: Influenza virus persists in mice hearts during convalescence without exacerbated inflammatory response.

(A) Timeline for mice infection and heart collection. Male and female wild-type and MLKL KO 6-to-8-week-old C57Bl/6 mice were intranasally infected with A/California/7/2009 (IAV) at day 0. Mice were euthanized and heart tissue was collected at day 12 (n = 10–12, 5–6 mice of each sex). (B) survival and (C) percent weight changes up to 10 days post IAV infection. (D) Viral titers were measured for WT C57Bl/6 and MLKL KO mice 12 days post initial IAV infection (n = 6, 3 of each sex) by the plaque assay and plotted in Log PFU/mL. (E) Levels of IFNƔ, TNF⍺, IFN⍺, IL-10, IFN-ƛ and IL-1-β measured by ELISA (pg/mL, n = 3–6, representative data for each sex of mice infected with IAV or challenged with vehicle (PBS) control). Log-rank (Mantel-Cox) test for survival. Kruskal-Wallis test with Dunn’s multiple-comparison post-test. Asterisks denote the level of significance observed: * = p ≤ 0.05; ** = p ≤ 0.01; *** = p ≤ 0.001.

Recently, it has been shown that both apoptosis and necroptosis are present in pulmonary tissue after influenza infection¹⁵. Pre-treatment of H9c2 myocytes with pan-caspase inhibitor Z-vad-fmk before IAV infection also protected against cell death (Online Figure VII. A), indicating a requirement for caspase activity in cardiomyocyte death. A stark increase in active caspase-1 and Gasdermin-D was observed in both WT and MLKL KO mice (Online Figure VII. B), suggesting an additional role for caspase activity⁴⁵ in IAV-induced cardiomyocyte death. Caspase-3 activation was notably increased in MLKL KO hearts after influenza infection (Online Figure VII. C) and in IAV-infected AC16 human cardiomyocytes pretreated with necrosulfonamide (Online Figure VII. D–F), suggesting that blocking necroptosis redirects cells to apoptosis.

Influenza virus infection leads to global proteome remodeling during convalescence.

To gain a comprehensive understanding of the effects of influenza infection on cardiac tissue and how necroptosis influences this process, we investigated the proteome-level changes in cardiac tissue after IAV clearance from the lungs. Hearts from C57BL/6 WT and MLKL KO mice (uninfected or 12 days post IAV infection) were analyzed by label-free quantitative proteomics^{41, 42}. In the global proteome (no filters applied, 1,592 total quantified proteins; Online Table I, Online Figure VIII. and IX), biological replicates within groups correlated tightly (Pearson r = 0.969 ± 0.013, n = 12), suggesting excellent experimental reproducibility (Online Figure VIII. A). Unsupervised hierarchical clustering (Online Figure VIII. B) and principal component analysis (PCA) (Online Figure VIII. C) of the unfiltered proteome showed segregation between experimental groups. Further filtering of the global proteomes yielded 295 proteins that were significantly different in at least one of the groups by ANOVA (permutation-based FDR < 0.01) (Fig. 3A). These proteins clustered into seven main groups based on their abundance profiles among the four conditions (Fig. 3A). GO enrichment analysis of proteins in each cluster revealed oxidation-reduction processes enriched in clusters 2, 4, 5, and 7 and collectively contained 50 mitochondrial proteins (Online Table II), consistent with our observation of altered mitochondrial function in vitro.

Figure 3: Influenza infection induces cardiac proteome remodeling.

Male and female 6-to-8-week-old wild-type and MLKL KO C57Bl/6 mice were intranasally infected with A/California/7/2009 (IAV) or mock challenged with vehicle (PBS) and hearts excised at day 12 post-infection. (A) Proteomic changes of mice hearts after IAV infection or mock challenge (n = 3 per condition). Hierarchical clustering of LFQ intensities of 295 significantly changed proteins (ANOVA, FDR < 0.01) revealed seven distinct clusters. Their abundance profiles among the groups were plotted in the middle panel. Enriched GO biological process terms are indicated for each marked cluster on the right panel. (B) Immunoblots for complement C3 (C3), AIAT, Catalase, HSP90, GST, GAPDH and VDAC and histograms of protein level quantification (n = 3 per group). Proteomic data of mice is representative from 2 separate experiments done with 3 mice of each sex; no sex based differences were observed. Kruskal-Wallis test with Dunn’s multiple-comparison post-test. Asterisks denote the level of significance observed: * = p ≤ 0.05; ** = p ≤ 0.01; *** = p ≤ 0.001.

The three main clusters with the most total proteins are 2, 6, and 7, which have 59, 80, and 75 proteins, respectively. Proteins in cluster 2 showed a decreased abundance in IAV-infected hearts. Several proteins in this cluster are associated with cardiac muscle contraction, including Troponin I (Tnn3), Tropomyosin-1, and Cysteine-rich protein 3 (Csrp3) (Online Table II–V). Subunits from ATP synthase and members of the mitochondrial electron transport chain were also found in this cluster, indicating broadly decreased energy production in cardiac tissues during IAV infection, consistent with our in vitro observation (Fig. 1I). Moreover, several enzymes involved in the TCA cycle were seen, including subunits or components from Pyruvate dehydrogenase (PdhA), Malate dehydrogenase (Mdh), and Isocitrate dehydrogenase (Idh). Another cluster enriched with metabolic enzymes is cluster 4, which contains a different subunit of Idh, Glyceraldehyde-3-phosphate dehydrogenase (Gapdh), and 2-oxoglutarate dehydrogenase (Ogdh), all of which have similar regulation trends. Cluster 4 also contains Heat shock protein 90 (Hsp90), an important molecular chaperone that promotes cell maturation, cytoskeletal maintenance, and proper regulation of cell cycle control and signal transduction⁴⁶. Down regulation of Hsp90 in both IAV-infected WT and MLKL KO cardiac tissue suggests altered heart protein homeostasis due to virus infection (Fig. 3A).

Cluster 7 showed an inverse trend of protein abundance change compared to cluster 2, where the IAV-infected hearts have proteins up-regulated, and those in MLKL KO mice are higher than in WT mice (Fig. 3A). This cluster includes proteins that regulate glutathione metabolism, including several glutathione transferases, and proteins with antioxidant activities, such as superoxide dismutase (Sod) and Catalase (Cat). Changes in these proteins may indicate an attempt of cardiac tissue to mitigate oxidative stress caused by influenza infection⁴⁷. Another interesting group of proteins up-regulated in this cluster were inhibitors of endopeptidase activities, including Serpina1b, Serpina1d, and Serpina3k. Protease activity is required for influenza virus activation⁴⁸, and an up-regulation of protease inhibitors may be an alternative strategy cardiac tissues undertake to combat influenza invasion. Cluster 6 contains proteins drastically increased only in IAV-infected MLKL KO hearts. This cluster is enriched in proteins that function in platelet aggregation, such as fibrinogens, and innate immune proteins, such as complement C3 and protein S100-A10. Several significantly changed proteins were validated by immunoblot (Fig. 3B). Importantly, we observed that hearts of mice infected with IAV strain PR8 showed similar proteomic (Online Figure X) and GO (Online Figure XI) changes to those infected with strain A/California/7/2009, with both infected tissues differing from uninfected hearts.

The cardiac proteome of IAV-infected MLKL KO mice is reshaped during convalescence.

To investigate the response of cardiac tissue to IAV infection when necroptosis is blocked, we compared the proteomes of IAV-infected WT (WT+IAV) and IAV-infected MLKL KO (KO+IAV) hearts (Fig. 4A). Among the 1,029 proteins quantified in the two groups, 220 of them were significantly different by at least 1.5-fold, with 158 and 62 proteins up- and down-regulated in the KO+IAV hearts, respectively. The protein with highest increase was NAD(P) transhydrogenase (Nnt), an inner mitochondrial membrane protein that maintains the mitochondrial NADPH level and can mediate the oxidative stress defense. Decreased cardiac Nnt has been closely associated with heart failure⁴⁹. In addition to antioxidants, proteins associated with glutathione metabolism (Glutathione transferase (GST) and Glutathione peroxidase), inflammatory or immune response (Alpha-1-acid glycoprotein 1, Complement C3), and regulation of protease functions (Serpina3k, Serpina1d) were also found to be highly up-regulated in IAV-infected MLKL KO hearts. Interestingly, components in the regulation of proteasomal protein catabolic process and regulation of phagocytosis were also enriched (Fig. 4B), suggesting that blocking necroptosis may re-activate alternative pathways in the heart to fight against IAV.

Figure 4: MLKL deficiency leads to proteome reshape in mice hearts during convalescence.

(A) Quantitative comparison of IAV infected WT and MLKL KO mice (n = 3 per group). Fold change of the proteins (x-axis) and their significance (p value, y-axis) were plotted. Up- and down-regulated proteins (1.5-fold cutoff) were highlighted in red and blue, respectively. (B) GO enrichment analysis of the significantly changed proteins as shown in panel A. Top 8 enriched biological terms of each group were displayed. Proteomic data of control and infected mice is representative from 2 separate experiments done with 3 mice of each sex; no sex based differences were observed.

Influenza virus infection affects phosphorylation in cardiac tissue proteins.

Using immunoblot analysis, we found the phosphorylation of AKT, CaMKII and AMPK⍺, kinases known to be important in regulating cell growth, apoptosis, and cardiac function and development⁵⁰ (Fig. 5A–B), decreased in both infected WT and MLKL KO cardiac tissue compared to uninfected hearts (Fig. 5A–B). To further evaluate the alteration of cardiac protein phosphorylation during influenza infection in an unbiased manner, we used TiO₂-based phosphopeptide enrichment for global phosphoproteome analyses⁵¹. In total, we identified 2,340 unique phosphopeptides (probability score > 0.75) corresponding to 898 phosphoproteins from the four groups of heart samples. Known phosphosites, as well as novel ones, were identified from our study (Online Tables V and VI). Interestingly, nearly 2/3 of them were only identified in the phosphoproteome but not the global proteome (Online Figure XII. A), possibly due to the lower abundance of phosphopeptides that can only be detected after enrichment. The majority of the 591 unique phosphoproteins were found to be associated with mRNA processing and splicing, including splicing factors and RNA binding motif proteins (e.g. Rbm10, Rbm17, Rbm25, Rbm37, Srsf6, Srsf9). The 307 proteins that were detected in both global proteomic and phosphoproteomic methods were strongly associated with cardiac muscle contraction and myofibril assembly process, including myosin light chain kinase 3 (Mylk3), tropomyosin (Tpm1) and myopalladin (Mypn) (Online Figure XII. A). The phosphorylation states of these proteins have also been implicated in heart diseases such as cardiomyopathy⁵².

Figure 5: Influenza infection leads to downregulation of functional proteins in the heart.

Male and female 6-to-8-week-old wild-type and MLKL KO mice were intranasally infected with A/California/7/2009 (IAV) or mock challenged with vehicle (PBS) and hearts excised at day 12 post-infection. (n = 3 per group, representative of 2 experiments for a total n =6, with 3 mice of each sex). Heart tissue was collected and immunoblots were performed against Rap1B, AKT, pAKT, pAMPKa, and pCaMKII, presented with quantification of relative protein level. Kruskal-Wallis test with Dunn’s multiple-comparison post-test. Asterisks denote the level of significance observed: * = p ≤ 0.05; ** = p ≤ 0.01.

Quantitative assessment of the phosphoproteomic data (Permutation-based FDR < 0.05) resulted in 1,691 phosphopeptides that were significantly different among the four groups (Online Figure XII. B). The majority of phosphopeptides (~97%) were decreased in IAV-infected hearts, regardless of genotype, suggesting a general protein dephosphorylation during IAV infection that may result in probable impact on cardiac function. Fig. 6A illustrates selected proteins with significantly different phosphorylation levels between the four groups. Proteins with decreased phosphorylation after infection include kinases such as Akt1 (S124), AMPK (e.g., Prkab1, S108; Prkab2, S183), PKA (e.g., Prkaca, T198), protein kinase C β (Prkcb, T500) and δ (Prkcd, S662), and apoptosis-associated proteins such as Pea-15⁵³, Gja1 (Connexin-43)⁵⁴, and Bcl-2 family proteins (Bclaf1 and Bag3)⁵⁵, suggesting possible dysregulation in intercellular communication and cell proliferation. Several mitochondrial proteins showed varied phosphorylation trends during infection: upregulated phosphoproteins include ADP/ATP translocase 1 (Ant1) and pyruvate dehydrogenase E1 component subunit α (Pdha1), and down-regulated ones included Vdac1, Bckdha, and Cytochrome b-c1 complex subunit 1 (Uqcrc1) (Fig. 6). Finally, cardiomyocytes may not be the only cells affected by IAV infection, as changes in Rap1B (Fig. 5, 6) indicate altered vascular function. Rap1b is involved in modulating basal vascular tone and blood pressure⁵⁶. Of interest, staining endothelial cells (CD144, VE Cadherin)⁵⁷ or cardiac fibroblasts (gp38)⁵⁸ and IAV in frozen cardiac sections suggests IAV may be able to infect both cell types during in vivo infection (Online Figure XIII. A). When NRK-49F fibroblasts and human umbilical vein endothelial cells (HUVECs) were infected with influenza virus, only endothelial cells showed significantly increased cytotoxicity upon infection (Online Figure XIII. B–C).

Figure 6: Influenza virus infection leads to changes in phosphorylation of cardiac proteins.

(A) Quantitative phosphoproteomic profiling of mouse heart. Plotted here are selected proteins and their identified phosphosites (indicated in the parentheses) grouped by protein functions or related pathways. (B) Quantitative comparison of IAV infected or mock infected (PBS challenge) WT and MLKL KO mice (n = 3 per group). Fold change of the phosphosites (x-axis) and their significance (p value, y-axis) were plotted. Up- and down-regulated proteins (1.5-fold cutoff) are highlighted in red and blue, respectively. Phosphoproteomic data is representative from 2 separate experiments done with 3 mice of each sex, no sex based differences were observed.

To better understand how necroptosis affects phosphorylation in cardiac tissue after IAV infection, pair-wise comparison of the KO+IAV group to the WT+IAV group identified 72 and 76 phosphopeptides that were up- and down-regulated, respectively, by at least 1.5-fold (Fig. 6B). The up-regulated phosphoproteins include histone deacetylases (HDAC) 1 and 2. Similar HDAC phosphorylation patterns have been implicated in regulating cardiac morphogenesis and growth⁵⁹. Among proteins with down-regulated phosphorylation in the KO+IAV group, we found several apoptosis-associated proteins, including Bcl-family proteins (Bclaf1 and Bag3), protein PML, and alpha-crystallin (Cryab)⁶⁰. Interestingly, membrane proteins that mediate Ca²⁺ homeostasis in cardiomyocytes, including the complex formed by sarcoplasmic reticulum Ca²⁺ ATPase (SERCA), phospholamban (Pln)⁶¹, and ryanodine receptor 2 (Ryr2), were all found to be down-regulated in phosphorylation. The phosphorylation intensity of identified peptides can be found in Online Tables V–VI. These results provide functional perspective of the cardiac proteins and pathways that are altered within necroptosis-deficient mice in response to influenza infection.

Inhibition of the mitochondrial permeability transition pore (MPTP) and replenishing of nicotinamide adenine dinucleotide (NAD) protects cardiomyocytes from influenza-mediated injury. Several mitochondrial proteins related to stress response or oxidation-reduction processes were identified in our proteome and phosphoproteome analyses. MPTP has been shown to maintain Ca²⁺ homeostasis and mediate both necrotic cell death (e.g. necroptosis⁶² and pyroptosis⁶³) and apoptosis^{62, 64, 65}. MPTP has also been implicated in major cardiac pathologies such as ischemia and reperfusion injuries⁶⁶. We also observed significant changes in the phosphorylation states of proteins regulating MPTP activity, Voltage-dependent anion channel 1 (VDAC1) and ADP/ATP translocase 1 (Ant1) (Fig. 5, 6). VDAC proteins localize at the outer membrane of mitochondria and regulate metabolite, ion, and ATP/ADP translocation across the membrane⁶⁵. Additionally, one initiator of mitochondrial damage and necroptosis is depletion of nicotinamide adenine dinucleotide (NAD⁺)^{64, 65}. Therefore, we hypothesized that therapeutically increasing NAD⁺ levels, antioxidant treatment, or inhibition of MPTP may rescue IAV-induced cardiomyocyte death. Pre-treatment of rat H9c2 myocytes with either nicotinamide to directly increase NAD⁺, TRO 19622⁶⁷ to inhibit VDAC and indirectly decrease leak of NAD⁺ from mitochondria, or the general antioxidant N-acetylcysteine (NAC), significantly lowered cytotoxicity and cellular ROS, and increased cellular ATP after IAV infection (Fig. 7A–C). Necroptosis and oxidative stress inhibition also showed protection against mitochondrial membrane permeabilization induced by IAV infection in AC16 myocytes (Fig. 7D). Together, these data demonstrate that IAV infection leads to necroptosis-associated mitochondrial damage, ROS production, and NAD⁺ depletion in cardiomyocytes.

Figure 7: Inhibition of the MPTP and replenishment of NAD+ leads to cardiomyocyte protection after IAV infection.

(A) Cytotoxicity of H9c2 myocytes treated with TRO 19622 (10 μM) or nicotinamide (1 mM) infected with A/California/7/2009 (IAV). (B) ROS/superoxide levels and (C) ATP levels of IAV-infected H9c2 myocytes treated with TRO 19622, nicotinamide with same concentration in (A) or n-acetylcysteine (10 μM), compared to uninfected or mock treated H9c2. RLU: relative light unit of luminescence. (D) Mitochondrial membrane permeabilization assay as measured via changes in RFU: relative fluorescent units of human AC16 cardiomyocytes infected with A/California/7/2009 (IAV) or PR8 and treated with rotenone, tempol or ponatinib. Carbonyl cyanide m-chlorophenyl hydrazone (CCCP) positive control for mitochondrial membrane permeabilization. Kruskal-Wallis test with Dunn’s multiple-comparison post-test. Asterisks denote the level of significance observed: * = p ≤ 0.05; ** = p ≤ 0.01; *** = p ≤ 0.001, **** = p ≤ 0.0001.

DISCUSSION

Increasingly strong association of MACE with influenza infection⁶⁸ demands a better understanding of the underlying molecular mechanisms driving this phenotype. Here, we provide the first evidence that viral particles can be detected in the myocardium during convalescence, indicating influenza virus persistence in other organs after its clearance from the lungs. Multiple studies have shown that cellular blocks can lead to abortive influenza virus replication in several different cell types^{33, 34}. This is supported by our study that shows IAV abortive replication in cardiomyocytes. How the virus translocates to cardiac tissue remains to be determined, but the pronounced effect it has on the heart merits further investigation.

While there are limitations in the presented study, mainly due to the use of experimental animal models and murine or human cells in vitro, the presented data offers compelling experimental evidence of the effects of influenza infection in cardiac tissue. Differences in infection outcomes between murine models and human hosts occur due to the complexity and differences of the human immune system and its murine counterpart. In humans, severe respiratory infection from viruses can lead to tachycardia, hypotension, vascular inflammation, myocarditis, cardiac arrhythmias and increase risk of death. The processes identified in this report are common to both mice and human allowing for the translation of important observations upon validation^69–72. Here we present a comprehensive report of the underlying molecular effects of influenza infection in cardiac tissue during convalescence using a multi-proteomic approach (Online Figure XIV). The global proteomic comparison between uninfected and IAV-infected WT hearts showed upregulation of pathways involved in metabolism and oxidative stress, indicating a high-stress, low-energy state of the heart during convalescence. Several endopeptidase inhibitor proteins were also up-regulated in IAV-infected hearts, providing possible novel targets⁷³ to treat IAV-driven cardiac dysfunction. This study serves as a valuable resource for future studies to pinpoint molecular intervention in cardiac pathology after IAV infection.

Phosphoproteomic analysis provided further insight into the functional state of cardiac proteins. The down-regulation of overall phosphorylation in IAV-infected hearts was consistent with a lower energy state, as seen in the global proteome analysis. Furthermore, many kinases and their substrate proteins that are important for cell proliferation, apoptosis, and intercellular communication had reduced phosphorylation. These data support the notion of decreased mitochondrial function and potential dysregulation of cardiomyocyte cellular and metabolic homeostasis^{65, 74}. Using cultured cardiomyocytes, we corroborated our proteomic data, showing that IAV can directly infect cardiomyocytes. IAV infection leads to cell death, oxidative stress, and mitochondrial damage, which is sensitive to treatment with either ROS inhibitors directed against mitochondrial function or repressing NAD⁺ depletion. Future studies should test the role of influenza infection in cardiac function and fibrosis by using left ventricular pressure-volume analyses, 2-D echocardiography and pathology assessments.

Finally, we observed that necroptosis modulates cardiac protein expression and functional profiles upon IAV infection. Necroptosis deficiency did not completely abolish the effects of IAV infection in the heart, suggesting that there are other molecular pathways essential for these changes. However, MLKL KO mice had a significantly altered cardiac proteome upon IAV infection compared with IAV-infected WT mice. Further studies using WT mice from intercrossed heterozygous animals may be required to fully define the role of necroptosis in modulation of influenza driven cardiac changes. Necroptosis inhibition led to an increase in NAD(P) transhydrogenase, protease inhibitors, and elevated detoxifying proteins and innate immunity, potential strategies that may protect cardiac tissue. Indeed, changes in mitochondrial integrity and oxidative stress proteins were partially rescued in IAV-infected MLKL KO mice when compared to IAV-infected WT mice. In vitro cardiac infection showed that necroptosis inhibition was beneficial, as cell toxicity and ROS generation were reduced, and production of ATP was promoted. Inhibition of MLKL leads to the activation of apoptosis via RIPK3⁷⁵ and its beneficial during influenza infection in the lungs¹⁵, however, inhibition of RIPK3 as a therapeutic target in this case may detrimental due to its role in activating apoptosis³⁸. Blocking of necroptosis was shown to increase apoptotic activity, suggesting a possible target to reduce necrotic cardiac injury. Our data also shows that pandemic influenza A H1N1 A/California/7/2009 and A/Puerto Rico/8/1934 alter similar pathways in mice hearts, with the pandemic strain showing a more profound effect, supporting our initial hypothesis that pandemic strains may be more cardiotropic. Thus, future studies could further elucidate similarities and differences between epidemic and pandemic influenza strains.

In conclusion, this report provides comprehensive evidence that influenza infection can modify the molecular homeostasis of the heart. Our results provide new insights into the prolonged cardiac effects after viral epidemics and pandemics. These results could also translate to other pathogenic viruses that frequently cause pulmonary infections. For example, clinical evidence from cases of SARS-CoV-2, causative of coronavirus disease (COVID)-19 have shown systemic manifestations by the virus. Recent reports have indicated the possibility of a direct effect of COVID-19 in causing myocarditis, fatal arrhythmias and other adverse cardiac events. This may be due in part to the presence of SARS-CoV-2 viral particles in cardiac tissue, a phenotype observed in patients that succumbed to infection, and could be further exacerbated indirectly via systemic inflammation^{76, 77}. Our results provide possible underlying mechanisms that may be shared by these respiratory viruses. Future studies are required to define similarities and differences between these pandemic viruses.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• The incidence of cardiovascular events increases after respiratory infections.

• The cell death pathway of necroptosis promotes inflammation and tissue injury.

What New Information Does This Article Contribute?

• Influenza virus can be found in mouse hearts after resolution of lung infection.

• Cardiac influenza virus presence during convalescence leads to proteome and phosphoproteome remodeling.

• Necroptosis inhibition promotes mitochondrial function, proteome and phosphoproteome reshaping and reduces cardiomyocyte damage after influenza infection.

Respiratory infections have been linked to adverse cardiac events. However, the role of influenza viral infection in the context of cardiac health during convalescence has not been explored. Influenza infection of cardiomyocytes showed changes to mitochondrial function, increase oxidative stress and cellular toxicity. The cellular changes in myocytes promoted the cell death mechanism of necroptosis. We identify that the proteomic and phosphoproteomic profiles of mouse hearts taken during convalescence showed significant remodeling. Deletion of the necroptosis effector molecule MLKL led to increase mitochondrial function and decrease cardiomyocyte death. Furthermore, necroptosis inhibition led to a re-shaped proteome and phosphoproteome. Our data collectively raise the possibility that therapies targeting oxidative stress and necroptosis may be effective in reducing the long-term effects of influenza infection in the heart.

Nonstandard Abbreviations and Acronyms:

MACE

Major adverse cardiac events

MPTP

Mitochondrial permeability transition pore

IAV

Influenza A virus

MLKL

Mixed Lineage Kinase Domain Like Pseudokinase

CAP

community acquired pneumonia

RIPK

receptor-interacting serine/threonine-protein kinases

8-OXO

8-Oxoguanine

ROS

reactive oxygen species

HR/AM LC-MS/MS

High-Resolution Accurate-Mass

LFQ

Label-free quantitation

GO

Gene Ontology

PFU

Plaque forming units