Differential effects of microRNAs miR‐21, miR‐99 and miR‐145 on lung regeneration and inflammation during recovery from influenza pneumonia

In a mouse model of influenza pneumonia, we previously documented that proliferating alveolar type II (AT2) cells are the major stem cells involved in early lung recovery. Profiling of microRNAs revealed significant dysregulation of specific ones, including miR‐21 and miR‐99a. Moreover, miR‐145 is known to exhibit antagonism to miR‐21. This follow‐up study investigated the roles of microRNAs miR‐21, miR‐99a, and miR‐145 in the murine pulmonary regenerative process and inflammation during influenza pneumonia. Inhibition of miR‐21 resulted in severe morbidity, and in significantly decreased proliferating AT2 cells due to impaired transition from innate to adaptive immune responses. Knockdown of miR‐99a culminated in moderate morbidity, with a significant increase in proliferating AT2 cells that may be linked to PTEN downregulation. In contrast, miR‐145 antagonism did not impact morbidity nor the proliferating AT2 cell population, and was associated with downregulation of TNF‐alpha, IL1‐beta, YM1, and LY6G. Hence, a complex interplay exists between expression of specific miRNAs, lung regeneration, and inflammation during recovery from influenza pneumonia. Inhibition of miR‐21 and miR‐99a (but not miR‐145) can lead to deleterious cellular and molecular effects on pulmonary repair and inflammatory processes during influenza pneumonia.


| INTRODUCTION
Pneumonia, or the inflammation of the lungs, is a serious complication that arises from many common and novel viral and bacterial infections.Of these, influenza viral infection is one of the most easily transmissible infectious diseases.Influenza virus infects about a billion people worldwide annually, and accounts for about 0.5 million deaths each year. 1 Especially during winter flu seasons, certain populations such as children younger than 1 year old, the elderly, pregnant women, and immunocompromised individuals are also at higher risk to suffer from influenza pneumonia, leading to increased morbidity and mortality. 2In addition, influenza pandemics such as the 1918 Spanish flu, 1968 Hong Kong flu, and 2009 swine flu pandemic resulted in millions of deaths.In light of the recent COVID-19 pandemic caused by SARS-CoV-2 where many patients were afflicted by pneumonia, 3 understanding the basic pulmonary pathophysiology in viral pneumonia is critical for developing novel therapeutic interventions.The histopathology of SARS-CoV-2 infection resembles that of a virulent influenza pneumonia, with diffuse alveolar damage, intra-alveolar fibrinous exudates, and hyaline membrane formation. 4,5Therefore, understanding the process of post-influenza repair and alveolar regeneration is important and relevant for current and future pandemics.Influenza pneumonia can be divided into two main phases-that is, the infection phase and the recovery phase.While the focus of influenza virus pneumonia research has mainly been on the host-pathogen relationship, especially on the pathogenicity of influenza virus during the infection phase, molecular changes at the recovery phase post-influenza pneumonia are yet to be completely understood.Moreover, there is evidence suggesting similarities in the regenerative processes of the lungs following pneumonia induced by influenza and other viral infections.In the COVID-19 pandemic 6 and the 2003 SARS-CoV epidemic, 7 alveolar type II (AT2) cell hyperplasia, and squamous metaplasia were also observed in the lungs of infected patients.
We previously studied the global and cellular processes linked to the regeneration of the lungs during the recovery phase of influenza, and reported the shift from the innate immune response towards adaptive immune responses.In addition, we observed the appearance of proliferating alveolar type II progenitor cells but not P63+/KRT5+ distal airway stem cells (DASCs)-playing important roles in lung regeneration during the recovery phase. 8cro-RNAs (miRNAs) are noncoding RNAs of around 22 nucleotides that regulate the expression of other genes either through mRNA decay and/or translation repression. 9As such, miRNAs are believed to be one of the important regulators of the body's physiological process.Through global lung microRNA (miRNA) analysis, we also documented that 126 miRNAs were significantly differentially regulated during the recovery phase of influenza infection. 10Since the regulations of many miRNAs were altered in regenerating lungs, their specific roles in post-influenza repair remain unclear.Out of the 126 significant miRNAs, we decided to study the role of three of these miRNAs on influenza recovery, namely miR-21, miR-99a, and miR-145.miR-21 was the highest upregulated miRNA (5.8-fold) during the recovery period, and has been clearly associated with oncogenesis, 11 proliferation, 12,13 and tissue regeneration. 14,15R-145 was downregulated (2.6-fold) during the recovery period, and it is reported to be a negative regulator of miR-21 16 and to be a tumor suppressor.17 Lastly, miR-99a was also downregulated (1.5fold) during the recovery period; however, it is not well-characterized in the literature although it is thought to have tumor suppressor [18][19][20] and epithelial-mesenchymal transition roles.21 According to the miRbase database, miR-21 and miR-145 are predicted to target 2780 and 3406 mRNAs, respectively, whereas miR-99a targets 517 mRNAs.22 The relatively high number of predicted targets of miR-21 and miR-145, in addition to their reported antagonistic actions, and lack of understanding on the role of miR-99a, led us to focus on these three miRNAs to further explore their roles in post-influenza repair.
In view of the dysregulation of miRNAs during influenza pneumonia recovery, we hypothesize that these miRNAs play important roles in the regeneration of lungs by targeting the progenitor lung cells, and that the modulation of these miRNAs may impact the pulmonary recovery process.Here, we employed a well-established BALB/c mouse model of sublethal influenza pneumonia, and showed that the downregulation of miR-21 and miR-99a, but not miR-145, resulted in enhanced morbidity during the recovery phase.The population of proliferating AT2 progenitor cells in influenza-challenged mice was significantly decreased when miR-21 was downregulated at 13 days postinfection.However, the AT2 cell proliferation was increased when miR-99a was downregulated at both 13 and 17 days postinfection.No change in AT2 cell proliferation was noted when miR-145 was downregulated.Interestingly, the expression of P63+/KRT5+ cells was not altered during the knockdown of these miRNAs.Gene expression analysis suggested that miR-21 downregulation resulted in elevated innate immune response and impaired transition from innate to adaptive immune responses, whereas miR-99a downregulation impacted growthrelated processes.The data here provide insights on the regenerative processes during influenza pneumonia.Given the similarities in the pathophysiology and histopathology of influenza and SARS-CoV-2 pneumonia, the data may also be relevant in the recent COVID-19 pandemic.

| Influenza virus and sub-lethal infection of animals
Animal experiments were approved under the Institutional Animal Care and Use Committee of the National University of Singapore under protocol numbers 2015-00014 and 2019-00019, and carried out in an animal biosafety level 2 facility.The strain of influenza virus used for infection of animals was influenza A/Puerto Rico/8/ 1934(H1N1) or PR8 at a sublethal dose of 20 plaque-forming units (PFU) per 25 μL of phosphate-buffered saline (PBS).Briefly, the animals were anesthetized with intra-peritoneal (IP) injection of a mixture of ketamine (75 mg/kg body weight) and medetomidine (1 mg/ kg body weight).PR8 virus was then administered via intra-tracheal delivery.The control group of mice was inoculated with sterile PBS.
Anesthesia reversal was then achieved by IP administration of atipamezole (1 mg/kg body weight).Animals were euthanized at 13 and 17 days postinfection (dpi) through overdose of anesthesia using three times the normal dose of anesthesia (mixture of 225 mg/kg of ketamine and 3 mg/kg of medetomidine).The left lung was harvested, and kept in 4% paraformaldehyde for histopathologic and immunofluorescence staining.The following tissues were snap-frozen: the post-caval lobe and middle lobe for RNA extraction; the superior lobe for protein extraction; and the inferior lobe for plaque assay to quantify viral load.Frozen lung tissues were kept in -80°C until further use.Animals were monitored daily, and their weight was recorded daily.In addition, the mice were observed for clinical signs of piloerection, labored breathing, hunched posture, reduced and swaying movement-semiquantitative clinical scores were recorded. 23Each parameter was ascribed a score on a scale between 0 and 5 points depending on the severity, with a higher score indicating more severe morbidity.

| Treatment of mice with miRNA modulators and controls
At 10 dpi, the mice were anesthetized by IP injection of ketamine (75 mg/kg) and medetomidine (1 mg/kg).HPLC-purified miRNA inhibitors (Qiagen) were dissolved in sterile filtered PBS, and administered intra-tracheally as described previously to both infected and mock-infected mice at a dose of 200 μg per mouse in 50 μL volume. 24The sequences of the miRNA inhibitors are as follows:
Paraffinized lungs were then embedded in paraffin blocks, cut with RM2255 rotary microtome (Leica Biosystems) at 4 μm per section, and placed on microscope slides coated with Polysine (Thermo Scientific).Consecutive sections per block were selected, and placed on separate Polysine slides to be used for either immunofluorescence or hematoxylin and eosin (H&E) staining.For histopathologic staining, the lung sections on the Polysine slides were placed in xylene (5 min, repeated twice), before being rehydrated in 100% ethanol (30 s, repeated twice), 90% ethanol (30 s), 70% ethanol (30 s), 50% ethanol (30 s), and distilled water (30 s, repeated thrice).The slides were then immersed in filtered Shandon hematoxylin for 8 min and washed in distilled water, before being differentiated in differentiating fluid (250 μL of hydrochloric acid in 100 mL of 70% ethanol, 10 s).Slides were then rinsed in distilled water again, before being washed in tap water to allow "bluing" of the nucleus (repeated thrice, slides would be left in last wash for 5 min).Following bluing, the slides were dehydrated in 95% ethanol (10 s), and stained with alcoholic eosin (10 s).The slides were finally dehydrated in 95% ethanol (10 s), 100% ethanol (10 s, repeated thrice), and Histoclear (10 s, repeated thrice).
Mounting of the coverslip was carried out using Permount mounting medium (Thermo Scientific).

| Immunofluorescence staining of lung tissue
Lung sections were rehydrated in decreasing ethanol concentration.
For primary antibody incubation, antibodies were diluted in 1% BSA and incubated overnight at 4°C.Secondary antibody incubation was then carried out with either anti-goat Alexa Fluor 488 (1:200, A11055, Thermo Scientific), donkey anti-rabbit Alexa Fluor 555 (1:200, A31572, Thermo Scientific), or goat anti-guinea pig Alexa Fluor 488 (1:200, ab150185, Abcam) diluted with 1× TBS at room temperature for 1 h.Following this, the slides were incubated in 0.1% Sudan Black (diluted in 70% ethanol) for 20 min to reduce background autofluorescence, 25 before mounting of the coverslip with ProLong Diamond Antifade mountant (Thermo Scientific).All washing steps were carried out using 1× TBST, and the microscope slides were kept at 4°C till imaging.
Following RT, the cDNAs were diluted 10× with nuclease-free water.
The qPCR assay was carried out in an Applied Biosystems 7500 Real-Time PCR system (Thermo Scientific) using the following parameters: initial hold at 50°C for 2 min, followed by initial melting stage at 95°C for 10 min.Thermocycling was then performed in alternating stages at 95°C for 15 s (denaturation) and 60°C for 1 min (elongation) for 45 cycles.Samples were loaded in technical duplicates, and the fold change (FC) was calculated using 2 C -ΔΔ T .

| Quantification of mRNA expression of selected genes
The extracted RNA was subjected to conventional RT by mixing 0.5 μL of 500 ng/μL random hexamers (Promega), 1 μL of RNA, 0.5 μL of 10 mM dNTP mix (Thermo Scientific), and 4.5 μL of nuclease-free water to make up a total volume of 6.5 μL per reaction.
The RT mixture was then heated to 65°C for 5 min before incubating on ice for 1 min.Two microliters of 5× first-strand buffer (Thermo Scientific), 0.5 μL of 0.1 M DTT (Thermo Scientific), 0.5 μL of RNaseOUT recombinant RNase inhibitor (Thermo Scientific), and 0.5 μL (100 units) of SuperScript reverse transcriptase were then added to each reaction and mixed, before incubation at 50°C for 60 min and enzymatic inactivation at 70°C for 15 min.Following RT, the cDNAs were diluted 10× with nuclease-free water.The qPCR was then carried out using 5 μL of FastStart Essential DNA Green Master (Roche), 3 μL of nuclease-free water, 0.5 μL of target gene forward primer (10 μM), 0.5 μL of target gene reverse primer (10 μM), and 1 μL of cDNA per sample.The target gene primer sequences are listed in Supporting Information: Table 1.The qPCR assay was then carried out using the LightCycler 96 system (Roche) with the following parameters: preincubation at 95°C for 10 min, followed by thermocycling for 45 cycles each consisting of denaturation (95°C for 10 s), annealing (50 or 60°C for 10 s), and elongation (72°C for 10 s).Samples were loaded in technical duplicates, and the fold change was calculated using 2 C −ΔΔ T .

| RNA sequencing (RNA-Seq) of lung samples
Extracted RNAs from three random lung samples of each infected animal group were also subjected to mRNA-seq (Novogene).The quality of the RNAs was validated using a bioanalyzer before the sequencing.The quality of the raw data output was first inspected using FASTQC, 26 and adaptor sequences were trimmed using trimmomatic. 27The reads were aligned to the mouse genome (GRCm38.p6)using STAR aligner. 28Fragments per kilobase of transcript per million (FPKM) were computed using the RSEM software. 29To mitigate the impact of noise in the low-abundance genes, a small value (1) was added to the FPKM values, and the values were log-transformed (base 2) for differential expression analysis (two-sample t-test and multiple testing correction by Q-value).Heatmap was generated using heatmap.2function of gplots library in R for the genes with Q-value <0.1 and absolute log 2 fold change >1.For functional enrichment, the selected genes from the heatmap were subjected to Gene Ontology (GO) analysis using ClueGO 30 and visualized using the Cytoscape software.

| Imaging and quantification of DASCs and proliferating AT2 cells in lungs
H&E and immunofluorescence slides (whole lung) were scanned at ×20 magnification with the TissueFAXS PLUS scanning system (TissueGnostics).The resultant images of the same whole lungs were overlaid together using PhotoShop image editor (Adobe).The damaged area (DA), boundary area (BA), and undamaged area (UA) were demarcated as previously described. 8For quantification of the proportion of DASCs expressing KRT5-positive marker, the total KRT5-positive area was manually demarcated and expressed as a ratio to the total damaged area.Quantification of SPC and SPC-plus-PCNA-stained cells was derived by selecting five random fields (at ×20 magnification) in the DA, BA, and UA from each lung image, and by manually counting the total SPC-plus-PCNA dual-expressing cells per field using the Fiji cell counter tool.

| Western blot analyses
The superior lobes of the lung were mixed with 700 μL of RIPA lysis and extraction buffer (Thermo Scientific) and Pierce protease and phosphatase inhibitor mini tablets (one tablet per 10 mL of RIPA buffer), and homogenized using the gentleMACS dissociator (Miltenyi Biotec) using the default protein settings.Each lung homogenate was centrifuged at 4°C for 10 min to remove residual cellular debris, and the supernatant was harvested.The lung homogenates from three   (Figure 1).Hence, the mice could withstand the procedure of intratracheal administration of miRNA inhibitor at 10 dpi.In addition, our previous study using this mouse model also showed that proliferating AT2 cells in the boundary area began to increase from 7 dpi, and peaked from 13 to 17 dpi. 8Given that a major objective of this study was to investigate the effects of knockdown of miRNAs on the proliferating AT2 cell population, the mice were euthanized and their lung samples harvested on 13 and 17 dpi to evaluate alveolar regeneration.
To evaluate animal morbidity, we weighed the mice daily and employed a scoring system based on clinical signs, including piloerection, hunched posture, labored breathing, reduced, and swaying movement. 23 The miR-21 inhibitor-treated mice exhibited an average weight loss of 27% and a clinical score of 2.9 at 13 dpi (Figure 1B), while miR-99a inhibitor-treated mice lost 20% body weight with a clinical score of 2.1 at 14 dpi (Figure 1C).In contrast, the miR-145 inhibitor-treated mice exhibited body weights and clinical scores similar to the control groups (Figure 1D).Treatment with the miRNA inhibitors did not result in any morbidity in uninfected mice (Figure 1B-D).Expression of miR-21, miR-99a, and miR-145 decreased significantly in the respective miRNA inhibitor treatment groups (Supporting Information: Figure 1), indicating that the observed morbidities were attributed to the downregulation of the corresponding miRNAs in the infected mice.

| Treatment with miR-21 and miR-99a inhibitors modulate the proliferating AT2 cell population of influenza-infected mice
We previously reported that in an influenza-infected mouse lung, the damage was localized to patchy areas in the lung, and that the lung could be segregated into damaged, undamaged, and boundary areas.
We found that proliferating AT2 cells played important roles in lung regeneration during recovery after influenza pneumonia, especially in the boundary area. 8To determine if there was any relationship between the modulation of miRNA expression and the population of  and 17 dpi (Supporting Information: Figure 2), and in uninfected lungs (Supporting Information: Figure 3).Taken together, the data suggest that there was no association between miR-145 inhibition and cellular lung regeneration with respect to proliferating AT2 cells.The impact of miR-21, miR-99a, and miR-145 downregulation on the proliferating AT2 cell population at 13 and 17 dpi is summarized in Tables 1 and 2, respectively.

| Inhibition of miRNAs does not affect the population of distal airway stem cells of infected mice
Another lung progenitor cell population involved in the lung regeneration process are cells expressing P63 and KRT5, otherwise termed as distal airway stem cells or DASCs. 31They are thought to be capable of regenerating alveolar epithelial type I (AT1) and II (AT2) cells in the lungs following influenza-induced damage. 32However, our previous data indicated that the DASCs did not contribute to the regeneration of lungs of influenza-infected mice up to 25 dpi. 8Here, we quantified and found no significant difference in the number of KRT5-positive DASCs between the different animal groups, regardless of the difference in morbidity of the mice treated with the different miRNA inhibitors at 13 dpi (Figure 4A-F) and at 17 dpi (Figure 4G-L).However, from 13 to 17 dpi (Figure 4A-L), there was a noticeable increase in the DASC population at 17 dpi as evident by the bigger KRT5-positive pods for all animal groups, which corroborated previous findings. 8,33We also established a histologic scoring system for lung repair (Supporting Information: Table 2) based on the presence of pods resembling squamous metaplasia, which we have previously shown to be DASC pods.As with the immunofluorescence assay, there was no difference in the lung repair score of infected mice treated with miR-21, miR-99a, and miR-145 inhibitors when compared to the control groups at 13 and 17 dpi-however, the histologic scores of all the infected treatment groups increased from 13 to 17 dpi (Supporting Information: Figure 4).Together, our data suggest that the modulation of miRNA expression did not affect the DASCs, in contrast to the proliferating AT2 cell population.

| Downregulation of miR-21 of influenzainfected mice impairs the shift from innate to adaptive immunity
To determine the downstream mechanisms through which inhibition of the miRNAs resulted in the observed phenotypes, RNA sequencing was performed on samples of infected lungs treated with miRNA inhibitors.We observed a significantly large cluster of genes that were downregulated in the miR-21 inhibitor treatment groups at 13 dpi (Figure 5A).Gene Ontology (GO) analysis of these genes revealed adaptive immune responses, such as T cell selection, B cell differentiation, interferon-gamma production, and antigen processing and presentation (Figure 5B).However, by 17 dpi, these same genes were no longer downregulated (Supporting Information: Figure 5).In the miR-21 inhibitor group, there was upregulation of a cluster of genes associated with cell survival and migration as shown in cluster (i) of Figure 5A,C and Table 3.To further elucidate the mechanisms underpinning the observed phenotype, we then probed the mRNA expression of 27 genes by real-time RT-qPCR.There was significantly enhanced mRNA expression of IL6 pro-inflammatory cytokine in infected lungs especially at 13 dpi and even at 17 dpi (Figure 6), suggesting augmented acute phase and innate immune responses.
Interestingly, IL6 overexpression was also observed in the uninfected mice treated with miR-21 inhibitor (Supplementary Figure 6).
KBTBD7 is a target of miR-21, and can activate the NF-κB pathway to induce IL6 expression. 34Hence, we determined if miR-21 downregulation could increase KBTBD7 expression to promote IL6 expression in miR-21 inhibitor-treated mice.In the latter mice, we observed elevated expression of KBTBD7 mRNA at 17 dpi (Figure 6B and Supplementary Figure 6B), but not at the protein level (Supplementary Figure 7A,B).At 17 dpi, while IL6 was still elevated, there was also significantly increased transcriptional expression of M2 macrophage markers, YM1 and RETNLA, 35 indicative of a delayed M2 macrophage polarization.This is further supported by the elevated expression of IL4, another M2 macrophage polarizer (at 13 dpi).These data may account for the severe morbidity (Figure 1B), and delayed appearance of proliferating AT2 cells in miR-21 inhibitor-treated mice (Figures 2A and 3A).Thus, miR-21 inhibition led to M2 macrophage polarization, exacerbated the innate immune response, and impaired its shift toward adaptive immunity that is crucial for the resolution of inflammation.
3.6 | Downregulation of miR-99a and miR-145 of influenza-infected mice leads to relatively minor immune dysregulation Surprisingly, the inhibition of miR-99a and miR-145 resulted in somewhat similar patterns of gene expression at 13 dpi (Figure 5A), that is, upregulation of a cluster of 9 genes and downregulation of another cluster of 22 genes, as depicted respectively in clusters (ii) and (iii) in Figure 5C.The upregulated genes in cluster (ii) are mainly associated with the immune response, including CD209a, CCL17 (chemokine 17), and CSF2 (granulocyte-macrophage colony-stimulating factor).These genes serve functions in the activation of adaptive immunity (Table 4).In the downregulated gene cluster (iii) in Figure 5C, almost half of the genes are also immune-related, including chemokines CCL2 and CCL7, acute phase reactant SAA3, complement factor B, and antimicrobial factors IFIT3 and CLEC4D.The functions of these genes in cluster (iii) are summarized in Table 5.While the global RNA expression patterns in the lungs were similar with miR-99a and miR-145 inhibitor treatment, there were obvious differences in animal morbidity and proliferating AT2 cell population between both treatment groups.This suggests that global RNA-Seq analysis could not detect expression differences of certain genes that may be differentially regulated in the proliferating AT2 cells between the two miRNA treatment groups.[38] 3.7 | Downregulation of miR-99a of influenzainfected mice impacts growth-related processes Next, we probed the expression of 27 genes in infected lungs treated with miR-99a inhibitor.These genes are associated with cytokines, immune-related markers, keratins, apoptosisrelated markers, migration markers, growth-related markers, or stem cell markers.The downregulation of miR-99a resulted in significantly diminished expression of genes associated with migration and growth pathways.In infected mice with miR-99a inhibition, the expression of PTEN, an inhibitor of the PI3K/AKT/mTOR pathway, was downregulated at both 13 and 17 dpi (Figure 6A,B), which may account for the increased proliferation of AT2 cells (Figures 2A and 3A).Interestingly, F I G U R E 5 RNA-Seq analysis of the lungs of infected mice at 13 dpi treated with miR-21, miR-99a and miR-145 inhibitors versus infected controls.(A) Heatmap of significant genes (Q-value <0.1, absolute log 2 FC >1) in the phosphate-buffered saline (PBS), control inhibitor (C), miR-21 inhibitor (21), miR-99a inhibitor (99), and miR-145 inhibitor (145) treatment groups.Genes demarcated in green boxes are distinct clusters of genes that were: (i) upregulated in the miR-21 inhibitor treatment group; (ii) upregulated in the miR-99a and miR-145 inhibitor treatment groups; and (iii) downregulated in the miR-99a and miR-145 inhibitor treatment groups.The purple demarcated box includes the large cluster of genes that were downregulated in the miR-21 inhibitor treatment group.(B) Gene ontology (GO) analysis of the biological processes in the cluster of downregulated genes of the miR-21 inhibitor group (in the purple box of the heatmap) revealed that adaptive immune system processes such as T-cell selection, interferon-gamma production, leukocyte activation, and differentiation were significantly dysregulated.GO analysis of the gene clusters in the green boxes of the heatmap did not reveal any significant biological processes or GO terms.GO analysis was conducted via ClueGO.The percentage in each GO term refers to the number of genes in the cluster to the total genes in the GO term.(C) Grouping of genes in the individual clusters (i), (ii) and (iii) in (A) according to five main functions: immune response, cell survival, migration, metabolic process, and others.These genes in clusters (i), (ii), and (iii) are summarized in Tables 3, 4, and 5, respectively.
T A B L E 3 Upregulated gene cluster (i) in the heatmap of the group of infected mice with miR-21 inhibitor treatment (Figure 5).T A B L E 4 Upregulated gene cluster (ii) in the heatmap of the groups of infected mice with miR-99a and miR-145 inhibitor treatment (Figure 5).

Unknown
T A B L E 5 Downregulated gene cluster (iii) in the heatmap of the groups of infected mice with miR-99a and miR-145 inhibitor treatment (Figure 5).Similarly, during the early recovery phase from influenza pneumonia, the significant changes in the global lung miRNA expression result in enhanced regenerative processes, and the transition from innate to adaptive immune response. 10e perturbation in expression of specific miRNAs during influenza pneumonia recovery culminated in undesirable phenotypes in mice, particularly when miR-21 and miR-99a were downregulated.
Inhibition of miR-21 resulted in severe clinical illness.We previously reported that during recovery of mice from influenza pneumonia from 10 dpi, there was an alleviation in immune processes which corresponded with the progression of cellular regeneration in the lungs. 8However, miR-21 downregulation exacerbated the innate immune response, and impaired the shift towards adaptive immunity, thus impacting cellular regeneration in the lungs.We previously noted a five-fold increase in miR-21 expression in lungs of mice during post-influenza recovery. 10The miR-21 is known to play active anti-inflammatory roles by downregulating the expression of proinflammatory cytokines by targeting NF-κB. 34,41The miR-21 also promotes M2 polarization which is involved in the switch from inflammation toward regeneration-miR-21 downregulation thus results in impairment of the M2 macrophage signature. 42,43Overexpression of miR-21 suppresses the M1 macrophage phenotype and enhances resolving inflammation following macrophage-mediated injury by targeting genes resulting in elevated IL10, anti-inflammatory cytokine. 44,45In addition, miR-21 is also a well-known oncogene, 12,13 and promotes cellular proliferation which is an essential process during regeneration.Therefore, miR-21 likely plays a pivotal role during influenza pneumonia recovery given that its downregulation causes an adverse phenotype in mice.
The downregulation of miR-99a resulted in moderate morbidity compared to the severe morbidity of infected mice with downregulated miR-21.The miR-99a is a relatively understudied miRNA in infections, with one study on the role of miR-99a on tissue regeneration. 19This study described the anti-regenerative role of miR-99a, where in vitro overexpression of miR-99a inhibited dermal wound healing by targeting the mTOR gene which is an important component of the PI3K/AKT/mTOR pathway.This is corroborated by our observation of the significantly increased proliferating AT2 cell population in lungs with knockdown of miR-99a during recovery from influenza pneumonia.Moreover, miR-99a is reported to be a tumor suppressor, 18,46 as opposed to miR-21 being an oncogene-this difference was evident in the contrast of their proliferating AT2 cell populations, that is, increased with miR-99a inhibitor treatment but diminished with miR-21 inhibitor treatment.However, while the increase in the lung progenitor cells in the miR-99a inhibition group could suggest increased regeneration, these mice still exhibited significant morbidity, which may be attributed to perturbations in inflammatory processes.Recently, Fitzpatrick et al. 46 showed that knockdown of miR-99a in THP-1 monocytes during Staphylococcus aureus infection culminates in significantly elevated mTOR level, decreased IL6, and increased IL10.Another study reported that overexpression of miR-99a in activated macrophages reduces the M1 phenotype and bactericidal activity, whereas miR-99a antagonism significantly attenuates M2 macrophage activation. 47Thus, miR-99a reduces M1 phenotype activation, but miR-99a plays a specific role in M2 phenotype activation.
Interestingly, the downregulation of the miR-145 did not exert any impact on clinical signs and morbidity.It has been reported that miR-145 exhibits an antagonistic relationship with miR-21. 16We also noted this similar phenomenon in that downregulation of miR-21 or miR-145 displayed opposite effects on disease severity, AT2 cell proliferation, and gene expression profile in the lungs.However, miR-21 expression did not increase when miR-145 was downregulated, and vice versa.This may be attributed to the miRNA expression being analyzed 3 days (at 13 dpi) and 7 days (at 17 dpi) following miRNA inhibitor treatment, when expression of the antagonistic miRNA may have recovered.It is noteworthy that miR-145 inhibitor treatment resulted in the lowest severity of influenza pneumonia compared to miR-99a and miR-21 inhibitor treatment.This implies that miR-145 may be involved in processes that promote recovery from influenza pneumonia.Interestingly, eosinophilic inflammation and Th2 cytokine production of allergic airway disease can be inhibited by miR-145 antagonism, whose anti-inflammatory effects are comparable to glucocorticoid treatment. 48Further studies are warranted to focus on the specific genes that differed between the miR-145 and miR-99a inhibitor treatment groups, which may help to elucidate the mechanisms underpinning lower morbidity during influenza pneumonia recovery.
Since the discovery of miRNAs, many studies have explored the feasibility of miRNA modulation as a means of therapeutic ONG ET AL.
| 15 of 18 intervention.For example, certain miRNAs are proposed as potential therapeutic targets for cancer therapy, including miR-21, 49 miR-99a, 50 and miR-145. 51Certain miRNA-based therapeutic studies have progressed to clinical trials, including Miravirsen, a miR-122 inhibitor against hepatitis C virus (HCV) infection. 52Given that miRNA modulation may target many downstream genes indiscriminately, the potential side effects of miRNA modulation should be seriously considered in miRNA therapeutic studies.Much remains unknown on the potential downstream adverse effects of miRNA modulators on overall host physiology and homeostasis.
Hence, candidate miRNA therapeutics are currently in clinical trials at the developmental phase 1 or 2 53 where their potential adverse effects may not yet be manifested.In a clinical trial of a miR-34-based anticancer therapeutic, some patients experienced undesirable immune side effects such as diminished lymphocyte and neutrophil counts. 54Should miR-21 eventually be developed as a viable anticancer therapeutic, the patients undergoing this treatment may incur a heightened risk of severe influenza pneumonia.
Notwithstanding this, miRNAs may serve as potentially useful clinical biomarkers for diagnosis or prediction of disease severity.For example, specific miRNAs are associated with specific cancers, such as miR-141 in prostate cancer, 55 in ovarian cancer, 56 and miR-25 and miR-223 in lung cancer. 57For viral diseases, specific miRNAs may serve as diagnostic and prognostic biomarkers. 58Among patients with chronic HCV infection, miR-21 may be a promising serum biomarker to assess efficacy of antiviral therapy and improvement of fibrosis. 59terestingly, there are also studies on miRNAs as biomarkers for the severity of pneumonia.One clinical study sequenced the bloodcirculating miRNAs of pneumonia patients with varying severity, and identified miRNA signatures specific to normal subjects versus nonsevere and severe pneumonia. 60[63][64] In conclusion, our study illustrated the importance of specific miRNAs in lung recovery after influenza pneumonia.Thus, inhibition of miR-21 and miR-99a, but not miR-145, resulted in animal morbidity that was associated with differential expression of critical gene transcripts.These findings highlight the deleterious effects and molecular mechanisms of perturbing the homeostatic equilibrium of critical microRNAs during pulmonary regeneration following influenza pneumonia.Similar experiments should be conducted in future using pulmonary histologic, regeneration, and molecular assays at more time-points and for longer periods.][67][68] control inhibitor (C): ACGTCTATACGCCCA; miR-21 inhibitor (21): TCAGTCTGATAAGCT; miR-99a inhibitor (99): GATCGGATC-TACGGGT; miR-145 inhibitor (145): CCTGGGAAAACTGGA.Following treatment, anesthesia reversal was then achieved by IP administration of atipamezole (1 mg/kg).
different mice were then pooled together into one pooled sample, and three pooled samples were prepared per group of mock-infected and infected mice.Proteins in the pooled lung homogenate were quantified using the Pierce BCA protein assay kit (Thermo Scientific), and diluted to a concentration of 20 μg/μL.Laemmli sample buffer (6×) was added to each sample (to a final concentration of 1×), before the protein samples were denatured at 95°C for 5 min.Polyacrylamide (10%) gel electrophoresis was performed at 100 V for 90 min.The proteins on the gel were then transferred onto PDVF membrane using the Trans-Blot Turbo transfer system (Bio-Rad) at 1.3 A (up to 25 V) for 20 min, and blocked with either 5% milk or 1% BSA for 2 h.Primary antibodies were then added and incubated overnight at 4°C.Primary antibodies used were: SPC (1:300, sc-13979, Santa Cruz Biotechnology), PCNA (1:300, sc-9857, Santa Cruz Biotechnology), KBTBD7 (1:300, NBP1-92040, Novus Biologicals), and GAPDH (1:5000, #2118, Cell Signaling Technology).Following primary antibody incubation, the PDVF membrane was washed thrice with 1× TBS buffer for 5 min each, before secondary antibody (conjugated with horseradish peroxidase) incubation for 1 h at room temperature.Secondary antibodies included either donkey anti-rabbit (1:2000, ab6802, Abcam) or anti-goat (1:2000, ab97110, Abcam) IgG H&L.The PDVF membrane was further washed thrice with 1× TBS buffer for 5 min each.The membrane was incubated with Clarity Western ECL substrate (Bio-Rad) before imaging with ChemiDoc MP gel imaging system (Bio-Rad).Each band intensity was quantified using the ImageJ software, and normalized to the GAPDH band intensity.2.10 | Statistical analysesStudent's t-test was applied for statistical analysis of the animal body weight change data in Figure1B-D.Chi-square test was used for the statistical analysis of the clinical scoring data in Figure1B-D, with the cut-off for high morbidity determined as a score of >1.5 and the cutoff for low morbidity as ≤1.5 (the value of 1.5 was chosen as it was the score of the control groups at 10 dpi).For the lung repair scoring in Supporting Information: Figure4, the cut-off for high repair score was determined as ≥3 at 13 dpi and ≥5 at 17 dpi, whereas the cut-off for low repair score was designated as <3 at 13 dpi and <5 at 17 dpi.The values of 3 and 5 were chosen by rounding off the means of the PBS treatment groups at 13 and 17 dpi to the nearest 0.5 value.Oneway ANOVA with Bonferroni multiple comparison correction was conducted for the quantification of proliferating AT2 cells and global lung protein expression in Figures2, 3and Supporting Information:

3 | RESULTS 3 . 1 |
Figure 1A).Mice were treated with miRNA inhibitor at 10 dpi since this time-point coincided with most animals recovering from influenza infection based on body weight change and clinical scoring Following infection, the control mouse groups (PBS treatment and control miRNA inhibitor treatment) lost weight from 4 dpi, reaching maximum average weight loss of 18% and 20%, respectively, of the original weight at 9 dpi, before recovering (Figure1B-D).Clinical scores of the control mouse groups corroborated with the weight loss data, with the mice reaching maximum average morbidity score of 1.6 and 1.8, respectively, at 9 dpi, before decreasing (Figure1B-D).Following miRNA inhibitor treatment at 10 dpi, the groups treated with miR-21 and miR-99a inhibitors began to exhibit significant morbidity compared to the PBS and control inhibitor groups.

Figure 3 )
Figure 2A,B), although an increase in the proliferating AT2 cell population was detected in the undamaged area of the lung at 13 and 17 dpi (Supporting Information: Figure 2C,D).Global protein expression of SFTPC and PCNA was unchanged in all treatment

3. 3 |
No impact of miR-145 inhibition on the proliferating AT2 cell population of influenza-infected mice While both miR-21 and miR-99a inhibition affected the total population of proliferating AT2 cells in the boundary area, the knockdown of miR-145 in infected mice (Supporting Information: Figure1C) did not result in any observable change in the proliferating AT2 cell population in the boundary area at 13 dpi (Figure2A,F) and 17 dpi (Figure3A,F) when compared to the mice treated with PBS or control inhibitor.Similarly for miR-145 inhibition, there was no difference in the proliferating AT2 cell population in the damaged and undamaged areas of infected lungs at 13 Proliferating AT2 cells (cells co-expressing nuclear PCNA and cytoplasmic SFTPC) in the lungs were significantly reduced in infected mice with miR-21 inhibitor treatment, but increased in infected mice with miR-99a inhibitor treatment at 13 dpi.(A) Proliferating AT2 cells were quantified as the average number of cells per ×20 field in the boundary area.The lungs of infected mice subjected to miR-21 and miR-99a inhibitor treatment displayed significantly lower and higher proliferating AT2 cells, respectively, when compared to the control groups.(B-F) Representative immunofluorescence images of proliferating AT2 cells in the lungs of infected mice of the phosphate-buffered saline (PBS), control inhibitor (C), miR-21 inhibitor (21), miR-99a inhibitor (99), and miR-145 (145) inhibitor treatment groups at 13 dpi.(G) Western blot analyses of global lung protein expression of PCNA and SFTPC in infected mice of the different treatment groups at 13 dpi.(H) Quantification of fold change (FC) of global lung PCNA protein expression in infected mice in comparison with the PBS-treated control group.No significant difference was observed.(I) Quantification of FC of global lung SFTPC protein expression in infected mice in comparison with the PBS-treated control group.No significant difference was observed.Data are expressed as the mean ± SD (n = 5-16 per group).Quantification of proliferating AT2 cells was analyzed by analysis of variance, with statistical differences indicated by *p < 0.05, **p < 0.01 and °p < 0.1.
an impairment in the shift towards adaptive immunity in miR-21 inhibitor-treated mice when compared to PBS and control inhibitor treatment groups.Many of the GO terms derived from the downregulated genes in miR-21 inhibitor-treated mice are associated with F I G U R E 3 Proliferating AT2 cells (cells co-expressing nuclear PCNA and cytoplasmic SFTPC) were increased in the lungs of infected mice with miR-99a inhibitor treatment at 17 dpi.(A) Proliferating AT2 cells were quantified as the average number of cells per ×20 field in the boundary area.The lungs of infected mice subjected to miR-99a inhibitor treatment exhibited significantly higher proliferating AT2 cells when compared to the control groups.(B-F) Representative immunofluorescence images of proliferating AT2 cells in the lungs of infected mice of the phosphate-buffered saline (PBS), control inhibitor (C), miR-21 inhibitor (21), miR-99a inhibitor (99), and miR-145 (145) inhibitor treatment groups at 17 dpi.(G) Western blot analyses of global lung protein expression of PCNA and SFTPC in infected mice of the different treatment groups at 17 dpi.(H) Quantification of fold change (FC) of global lung PCNA protein expression in infected mice in comparison with the PBS-treated control group.No significant difference was observed.(I) Quantification of fold change (FC) of global lung SFTPC protein expression in infected mice in comparison with the PBS-treated control group.No significant difference was observed.Data are expressed as the mean ± SD (n = 11-18 per group).Quantification of proliferating AT2 cells was analyzed by ANOVA, with statistical differences indicated by **p < 0.01.
No difference in the distribution of KRT5-positive distal airway stem cells in the lungs of infected mice in different treatment groups.Representative immunofluorescence staining of KRT5+ pods in the lungs of infected mice of (A) phosphate-buffered saline (PBS), (B) control inhibitor, (C) miR-21 inhibitor, (D) miR-99a inhibitor, and (E) miR145 inhibitor treatment groups at 13 dpi.KRT5+ pods are indicated by white arrows.(F) Quantification of the percentage of the KRT5+ cell area over the total damaged area in the lungs of infected mice of the PBS, control (C), miR-21(21), miR-99a (99), and miR-145 (145) inhibitor treatment groups at 13 dpi.No significant difference in the KRT5+ cell area was observed between the treatment groups.Representative immunofluorescence staining of KRT5+ pods in the lungs of infected mice of (G) PBS, (H) control inhibitor, (I) miR-21 inhibitor, (J) miR-99a inhibitor, and (K) miR-145 inhibitor treatment groups at 17 dpi.KRT5+ pods are indicated by white arrows.(L) Quantification of the percentage of the KRT5+ cell area over the total damaged area in the lungs of infected mice of the PBS, control (C), miR-21(21), miR-99a (99), and miR-145 (145) inhibitor treatment groups at 17 dpi.No significant difference in the KRT5+ cell area was observed between the treatment groups.However, the KRT+ cell area at 17 dpi was generally greater than that at 13 dpi for all groups.Data are expressed as the mean ± SD.Each datapoint in (F) and (L) represents an individual mouse lung in the treatment group (n = 5-16 per group at 13 dpi; n = 11-18 per group at 17 dpi).T A B L E 1 Changes in the proliferating AT2 cell population in boundary, damaged, and undamaged areas of the lungs of infected mice at 13 dpi in response to inhibition of miR-21, miR-99a, and miR-145.F I G U R E 5 (See caption on next page).ONG ET AL.|11 of 18

FGF7 39
and HGF,40 both of which are ligands of the PI3K/ AKT/mTOR pathway, were also downregulated at both 13 and 17 dpi (Figure6A,B).The downregulation of these two growth factors may not have severely impacted the activation of the pathways, given that there are many other concomitantly expressed growth factors in the host.These findings are indicative of dysregulation in growth-related pathways in infected lungs treated with miR-99a inhibitor during recovery, F I G U R E 6 Expression of selected genes in the lungs of infected mice in different treatment groups.Expression of cytokine-related, immunerelated markers, keratins, apoptosis-related, migration, growth-related and stemness genes of infected mice treated with phosphate-buffered saline (PBS), control inhibitor (C), miR-21 inhibitor (21), miR-99a inhibitor (99), and miR-145 inhibitor (145) at (A) 13 dpi and (B) 17 dpi.Expression of genes in each group was evaluated by RT-qPCR with reference to the PBS treatment group.(A) At 13 dpi, mRNA expression of pro-inflammatory cytokines (especially IL6) was significantly upregulated in the miR-21 inhibitor treatment group.The pro-inflammatory cytokine TNFA, growth-related and migration-associated genes were downregulated in the miR-99a and miR-145 inhibitor treatment groups.The pro-inflammatory cytokine IL1B, neutrophil marker LY6G, and M2 macrophage marker YM1 were downregulated in the miR-145 inhibitor treatment group.(B) At 17 dpi, mRNA expression of pro-inflammatory cytokines and immune-related genes (especially IL6 and YM1) was significantly upregulated in the miR-21 inhibitor treatment group.Growth-related and migration-associated genes were downregulated in the miR-99a inhibitor treatment group.IL1B and TGFB mRNAs were downregulated in the miR-145 inhibitor treatment group.Data are expressed as the mean of log 2 fold change (n = 5-16 per group at 13 dpi; n = 11-18 per group at 17 dpi).Fold change was determined by calculating the 2 C −ΔΔ T values of the respective group, with the PBS treatment group as the comparator and RPL13A as the housekeeping gene.Statistical analysis was conducted using ANOVA analysis on the individual ΔC T values between the treatment groups at each time-point.Statistical differences versus the PBS-treated group are indicated by *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.The solid box indicates that the gene is an experimentally verified target of the corresponding miRNA, whereas the dotted box indicates that the gene is a predicted target of the corresponding miRNA via TargetScan.
motif chemokine 2. Acts as a ligand for C-C chemokine receptor CCR2.−1.70 Inflammation SAA3 Serum amyloid A-3 protein.Major acute phase reactant.−1.63 Inflammation CCL7 C-C motif chemokine 7. Chemotactic factor that attracts monocytes and eosinophils, but not neutrophils.−1.60 Inflammation CFB Complement factor B. Part of the alternate pathway of the complement system.−1.42 Inflammation CCR1 C-C chemokine receptor type 1. Receptor for a C-C type chemokine.−1.25 Inflammation SPP1 Osteopontin.Acts as a cytokine involved in enhancing production of IFN-gamma and IL-12.−1.12 Inflammation IFIT3 Interferon-induced protein with tetratricopeptide repeats 3. IFN-induced antiviral protein which acts as an inhibitor of cellular and viral processes (cell migration, proliferation, signaling, viral replication).−1.12 Inflammation CLEC4D C-type lectin domain family 4 member D. Involved in innate recognition of pathogen-associated molecular patterns (PAMPs).−1.06 Inflammation VSIG4 V-set and immunoglobulin domain-containing 4. Suppresses activation of complement pathways, and induces regulatory T-cell differentiation.−1.04 Inflammation CYP4F18 Cytochrome P450 4F subfamily protein.A cytochrome P450 monooxygenase involved in metabolism of the pro-inflammatory leukotriene B4. −0.78 Inflammation MMP8 Neutrophil collagenase.Can degrade fibrillar type I, II, and III collagens.−1.30Migration MMP9 Matrix metalloproteinase-9.Cleaves type IV and V collagens.−1.21 Migration TGFBI Transforming growth factor-beta-induced protein ig-h3.Plays a role in cell adhesion.−1.03 Migration PLA1A Phospholipase A1 member A. Hydrolyzes the ester bond at the sn-1 position of glycerophospholipids and produces 2-acyl lysophospholipids.types I and II.Membrane glycoproteins.−1.56 Others FOLR2 Folate receptor beta.Binds to folate and reduced folic acid derivatives.−1.09Others RTN4RL2 Reticulon-4 receptor-like 2. Cell surface receptor that plays a functionally redundant role in inhibition of neurite outgrowth mediated by MAG (by similarity).−1.03Others and may account for the higher proliferating AT2 cell populations at 13 and 17 dpi.Some impact of miR-99a downregulation on the immune responses may account for the relatively higher morbidity of the mice when compared to the PBS control group.This may also explain the comparatively lower morbidity of the mice when compared to the miR-21 inhibitor group (where significant immune response dysregulation was evident).

3. 8 |
Downregulation of miR-145 of influenzainfected mice culminates in minor alleviation of innate immune response At 13 dpi, both miR-99a and miR-145 inhibition resulted in generally reduced expression of growth-related and migration-associated genes, indicating dysregulation of growth-related pathways in murine lungs treated with miR-99a and miR-145 inhibitors.Intriguingly, a crucial difference between these two miRNA inhibitor treatment groups was the decreased expression of the IL1B pro-inflammatory cytokine and LY6G neutrophil marker in mouse lungs treated with the miR-145 inhibitor at 13 dpi (Figure6A).This suggests that altered expression of genes involved in the innate immune response may play important roles in alleviating animal morbidity during influenza pneumonia.4| DISCUSSIONHost gene expression is governed by a myriad of mechanisms, such as transcription factors, epigenetics, posttranslational modifications, miRNAs, and many others.These mechanisms work together to modulate the expression patterns of each of the thousands of genes in the body.These in turn control and regulate the countless number of cellular and physiological processes.During infection, the body responds by self-regulating biological processes at the site of insult.
Probable pathogen-recognition receptor, and termed as mouse DC-SIGN ortholog.