Antibody Treatment against Angiopoietin-Like 4 Reduces Pulmonary Edema and Injury in Secondary Pneumococcal Pneumonia.

Despite extensive global efforts, secondary bacterial pneumonia still represents a major cause of death in developing countries and is an important cause of long-term functional disability arising from lung tissue damage. Newer approaches to improving treatment outcomes are needed to reduce the significant morbidity and mortality caused by infectious diseases. Our study, using an experimental mouse model of secondary S. pneumoniae infection, shows that a multimodal treatment that concurrently targets host and pathogen factors improved lung tissue integrity and extended the median survival time of infected mice. The immunoneutralization of host protein cANGPTL4 reduced the severity of pulmonary edema and damage. We show that host-directed therapeutic strategies as well as neutralizing antibodies against pathogen virulence factors are viable adjuncts to standard antimicrobial treatments such as antibiotics. In view of their different modes of action compared to antibiotics, concurrent immunotherapies using antibodies are potentially efficacious against secondary pneumococcal pneumonia caused by antibiotic-resistant pathogens.

IMPORTANCE Despite extensive global efforts, secondary bacterial pneumonia still represents a major cause of death in developing countries and is an important cause of long-term functional disability arising from lung tissue damage. Newer approaches to improving treatment outcomes are needed to reduce the significant morbidity and mortality caused by infectious diseases. Our study, using an experimental mouse model of secondary S. pneumoniae infection, shows that a multimodal treatment that concurrently targets host and pathogen factors improved lung tissue integrity and extended the median survival time of infected mice. The immunoneutralization of host protein cANGPTL4 reduced the severity of pulmonary edema and damage. We show that host-directed therapeutic strategies as well as neutralizing antibodies against pathogen virulence factors are viable adjuncts to standard antimicrobial treatments such as antibiotics. In view of their different modes of action compared to antibiotics, concurrent immunotherapies using antibodies are potentially efficacious against secondary pneumococcal pneumonia caused by antibiotic-resistant pathogens. KEYWORDS ANGPTL4, secondary bacterial pneumonia, antibiotic resistance, hostdirected immunotherapeutics, vascular permeability B acterial pneumonia may present as a primary disease or as the terminal event in individuals who are already debilitated. Secondary bacterial infections resulting from an influenza infection caused over half of the deaths during the 1918 flu pandemic and remain one of the leading causes of mortality due to flu infection (1). Streptococcus pneumoniae (S. pneumoniae) is the most common pathogen responsible for secondary bacterial infections during influenza pandemics (2,3). During the 2009 H1N1 pandemic, although antibiotics were widely used for treatment, secondary bacterial infection was detected in 55% of the fatal cases, with 26% of mortality being attributed to secondary S. pneumoniae infection (4). That value increased to 34% for patients in intensive care units (5). Current vaccines against S. pneumoniae, such as the 13-valent pneumococcal conjugate vaccine, have reduced pneumococcal infection rates (6,7). However, the benefits of vaccination are also significantly diminished by epidemiological shifts (8). Among the six clinical trials that evaluated vaccine effectiveness against all-cause pneumonia in older adults, only one trial demonstrated a risk reduction in the vaccinated group (9). These findings underscore the importance of managing patients with influenza who may also have bacterial pneumonia.
Although antibiotics remain the mainstay of treatment for bacterial pneumonia, their broad usage has limitations and may cause certain complications. While bacteriolytic antibiotic treatment may eradicate S. pneumoniae, it causes the release of cytoplasmic virulence factors such as pneumolysin (10,11), which is a pore-forming toxin produced by all clinical isolates of S. pneumoniae (12). This results in clinical complications even after the eradication of the pathogen (13). During a lung infection, free pneumolysin causes a disruption of the alveolus-capillary barrier, leading to pulmonary leakage and exacerbating disease severity (14). In addition, antibiotic resistance is recognized as a growing global threat. Resistance to new antibiotics is emerging at an ever-increasing pace, and at the same time, the rate of development of new antibiotics has slowed substantially. The increasing incidence of drug-resistant S. pneumoniae (DRSP) has complicated the treatment and management of various pneumococcal disease manifestations (15). Over the years, S. pneumoniae has continued to develop antibiotic resistance and poses a serious challenge to public health, particularly in Asian countries, which have among the highest rates of antibiotic resistance in the world (16,17). In Asia, among the isolates belonging to the most prominent non-PCV7 serotype, 19A, 86.0% and 79.8% showed erythromycin resistance and multidrug resistance, respectively (16). Clearly, novel alternative countermeasures against pneumonia are urgently needed (18)(19)(20).
Many strategies that exploit components of the immune system are being actively pursued, although most are still in their infancy (21,22). Immunotherapy with antibod-ies has taken a leading role in this field. In view of their different modes of action compared to antibiotics, antibodies are effective against infections with antibioticresistant pathogens. The synergistic use of antibodies with antibiotics can be an effective approach to managing difficult-to-treat pneumonia. High-throughput RNA sequencing of lung tissue samples from patients during the 1918 and 2009 influenza pandemics revealed that the angiopoietin-like 4 protein (ANGPTL4) was one of the most significantly upregulated mRNAs (23). We previously showed that upregulation of the C-terminal ANGPTL4 fragment (cANGPTL4) was associated with increased severity of pulmonary leakage and damage in primary influenza pneumonia (24). ANGPTL4 knockout (Ϫ/Ϫ) mice exhibit reduced pulmonary edema and improved lung tissue integrity in primary influenza pneumonia (24). These observations underscore the importance and conserved role of cANGPTL4 in the host response to infection-induced lung injury. In this study, we examined the host response to concurrent treatment of secondary bacterial pneumonia with a neutralizing anti-cANGPTL4 antibody together with antibiotics and/or a bacterial pneumolysin-targeted antibody.

RESULTS
Secondary bacterial infections following primary influenza compromise pulmonary integrity and augment lung edema. Among the clinical cases of influenza infection, the associated secondary bacterial pneumonia is a leading cause of death (1), with S. pneumoniae being the most common causative bacterium (2,3). The most prevalent serotype encountered in the Asian population is 19F, whereas serotype 3 is more virulent (16,25). To address the possible role of cANGPTL4 in secondary pneumococcal pneumonia, we developed a sequential dual-infection mouse model, mimicking the clinical scenario wherein patients with a primary influenza infection are susceptible to secondary pneumococcal pneumonia. Mice were intratracheally infected with the influenza virus on day 0 and subsequently infected intratracheally with either S. pneumoniae serotype 3 (S3) or 19F at 7 days post-influenza infection (dpi) (Fig. 1A).
Next, we examined the host expression profile and the pathological phenotype of the lung tissues in this sequential dual-infection mouse model. The bronchiolar epithelium and alveolar tissues were infiltrated with immune cells and damaged by the influenza virus infection at 7 dpi prior to bacterial challenge (Fig. 1A). From 8 to 11 dpi, lungs were harvested daily from mice infected by the influenza virus (Flu), bacteria (S3 or 19F), or influenza with bacterial superinfection (FluϩS3 or Fluϩ19F). To assess the Antibody Therapy Improves Bacterial Pneumonia Survival ® extent of edema in the infected lungs, the volumes of the harvested lung lobes were measured. Lung lobes with more edema display larger volumes due to fluid infiltration from the circulation (Fig. 1B). The lungs of the primary influenza-infected mice showed a small, but significant, 1.5-fold increase in lung volume, while primary bacterial 19Fand S3-infected lungs exhibited no significant change in lung volume compared with the sham-infected lungs (Fig. 1C). Secondary bacterial infection resulted in significantly greater edema than primary influenza infection. They also showed clear pneumococcal serotype-specific patterns ( Fig. 1B and C). Mice infected with influenza and serotype 3 of S. pneumoniae (FluϩS3) reached the criteria for euthanasia (i.e., at least 30% loss of their body weight) before 11 dpi (Fig. 1B). The lungs of mice infected with FluϩS3 displayed a significant Ն2.0-fold increase in lung lobe volume compared with shaminfected lungs (Fig. 1C). S. pneumoniae cells were detected in the pulmonary epithelium and alveolar spaces, as determined by fluorescence in situ hybridization (FISH). The fluid-filled alveolar spaces may encourage bacterial growth, and S. pneumoniae cells were detected in abundance in those spaces (see Fig. S1A in the supplemental material).
Histological analysis of the lungs revealed significantly greater lung tissue damage in secondary pneumococcal infections compared with primary influenza infection alone (Fig. S1B). Regions of lung tissue from FluϩS3-superinfected mice lost the basic structure of the alveolar walls and were replaced with excessive fluid arising from severe lung edema. Compared to the FluϩS3 mice, the Fluϩ19F-superinfected mice displayed better preserved alveolar structures, although there was still significantly more fluid infiltration in the alveolar spaces compared to mice with influenza infection alone. More intra-alveolar erythrocytes and immune cells infiltrated the lung parenchyma in influenza-infected mice than in primary bacterial pneumonia-infected mice. The lungs of primary bacterial pneumonia-infected mice exhibited pulmonary tissue morphology similar to that of healthy control lungs, suggesting that primary bacterial infection did not cause noticeable tissue damage. The trends in lung tissue damage among the different experimental groups were indicated by the morphology of the alveolar structures and the percentage of the intact alveoli compared to healthy lung lobes. Sham-infected, S3-infected, and 19F-infected mice showed uncompromised alveolar space integrity, whereas secondary pneumococcal pneumonia-infected mice suffered more substantial lung tissue injury than influenza-infected mice. Secondary pneumococcal pneumonia-infected mice revealed distinct serotype specificity, with FluϩS3 mice suffering more severe edema and loss of alveolar structures than Fluϩ19F mice (Fig. S1C). Taken together, secondary pneumococcal infection following primary influenza infection significantly increased the severity of lung edema and pulmonary tissue damage in a serotype-specific manner.
cANGPTL4 expression is upregulated in the lung tissues of mice with secondary pneumococcal pneumonia. In correlation with the severity of lung damage, quantitative PCR (qPCR) analysis showed that ANGPTL4 mRNA levels peaked at 9 dpi in lungs infected with Flu (9-fold), 19F (4-fold), S3 (2-fold), Fluϩ19F (7.5-fold), and FluϩS3 (10-fold) compared to sham-infected lungs ( Fig. 2A). Immunoblot analysis revealed elevated levels of cANGPTL4 protein from 8 to 10 dpi across all of the diseased lungs compared to the healthy lungs (Fig. 2B). Immunofluorescence staining of the lung tissue revealed a widespread distribution of cANGPTL4 in mice with primary influenza and secondary bacterial pneumonia (FluϩS3 and Fluϩ19F) at all examined time points (Fig. 2C). In contrast, primary bacterial pneumonia-infected lungs showed cANGPTL4 immunostaining patterns that were similar to those in healthy control lung sections, in which cANGPTL4 was restricted to the bronchiolar structures (Fig. 2C). Thus, elevated cANGPTL4 expression levels correlate well with pulmonary damage in secondary bacterial pneumonia.
Deficiency in host cANGPTL4 reduces pulmonary edema and protects lung tissue integrity in mice with secondary pneumococcal pneumonia. The abundance of the cANGPTL4 protein in secondary bacterial pneumonia-infected mouse lung tissue suggests an important role for cANGPTL4 in infection-induced pulmonary damage. To this end, we examined its impact on lung edema and tissue integrity during secondary bacterial pneumonia using ANGPTL4 knockout mice (ANGPTL4 Ϫ/Ϫ ) ( Fig. 3A and B) and immunoneutralization of cANGPTL4 ( Fig. 3C and D). ANGPTL4 Ϫ/Ϫ mice with secondary FluϩS3 and Fluϩ19F infections exhibited less pulmonary edema than their ANG-PTL4 ϩ/ϩ counterparts ( Fig. 3A and B). As indicated by the arrows in Fig. 3B, Fluϩ19Finfected ANGPTL4 ϩ/ϩ mouse lung tissue was characterized by extensive edema, bleeding, and immune cell infiltration in the alveolar spaces, whereas ANGPTL4 Ϫ/Ϫ mice with Fluϩ19F infection showed better-preserved integrity of the alveolar spaces. Among the FluϩS3-infected mice, ANGPTL4 ϩ/ϩ mouse lung tissue featured severe edema that destroyed the alveolar structures, leaving diminished alveolar walls, in contrast to ANGPTL4 Ϫ/Ϫ mice, which had preserved alveolar structures and less edema (Fig. 3B). In ANGPTL4 ϩ/Ϫ mice, the improvement in lung tissue integrity was less obvious (  Antibody Therapy Improves Bacterial Pneumonia Survival ® and B). We observed an improvement in lung tissue integrity in ANGPTL4 Ϫ/Ϫ mice compared with their wild-type counterparts for both FluϩS3 and Fluϩ19F superinfections, as indicated by the higher percentage of intact alveolar spaces (see Fig. S2A in the supplemental material).
To confirm the results observed in ANGPTL4 knockout mice and to explore the therapeutic potential of anti-cANGPTL4 treatment, we next studied the effects of an cANGPTL4 deficiency caused by the injection of an anti-cANGPTL4 monoclonal antibody (MAb) into secondary pneumococcal pneumonia-infected mice. Infected ANG-PTL4 ϩ/ϩ mice treated with anti-cANGPTL4 MAb were protected against infectionassociated lung damage, with better lung tissue integrity and reduced edema compared to infected but untreated ANGPTL4 ϩ/ϩ mice, thus confirming the findings of the ANGPTL4 knockout experiments (Fig. 3C). As indicated by the higher percentage of intact alveolar spaces, we observed an improvement in lung tissue integrity in anti-cANGPTL4 MAb-treated mice compared with control IgG-treated mice for both FluϩS3 and Fluϩ19F superinfections (Fig. S2A). Electron micrographs further confirmed that Representative image of lung harvested at 11 dpi from ANGPTL4 knockout (Ϫ/Ϫ), heterozygous (ϩ/Ϫ), and wild-type (ϩ/ϩ) mice infected with influenza virus and different pneumococcal serotypes (S3 and 19F). Scale bar ϭ 5 mm. n ϭ 3 independent experiments with 5 mice per time point for each experimental group. (B) Representative H&E images of the indicated primary and secondary pneumococcal infections of lungs from ANGPTL4 knockout (Ϫ/Ϫ), heterozygous (ϩ/Ϫ), and wild-type (ϩ/ϩ) mice. In Fluϩ19F-infected mice, the ANGPTL4 knockout group showed significant reduction of immune cell infiltration and edema compared with their wild-type counterparts (black versus white arrows); In FluϩS3-infected mice, the wild-type group showed vanished alveolar structures with severe edema, while the ANGPTL4 knockout group preserved the alveolar structures with reduced edema (black versus white arrows). Scale bar ϭ 50 m. (C) Infected wild-type mice treated with the anti-cANGPTL4 antibody compared with untreated mice. In the Fluϩ19F-infected mice, anti-cANGPTL4 treatment induced improved bronchiolar epithelial integrity and reduced infiltration (black versus white arrows). In the FluϩS3-infected mice, anti-cANGPTL4 treatment better preserved the structure of the alveolus (white arrow), compared to the diminished alveolar structures due to severe edema in the lungs of the FluϩS3 group of mice (black arrow). Scale bar ϭ 50 m. (D) Representative transmission electron microscopy images of lungs from mice with a secondary infection of serotype 3 that were administered the anti-cANGPTL4 antibody (i and ii) or control IgG (iii and iv). White arrows and arrowheads indicate cell-cell junctions and collagen fibrils, respectively. Viral budding (v) is visible at the damaged cell boundary, and bacteria are visible in control vehicle-treated mice with a secondary infection of serotype 3 (vi). Red arrowheads indicate viral budding and bacteria. Scale bar ϭ 500 nm.
anti-cANGPTL4 MAb treatment improved the integrity of intercellular tight junctions between lung epithelial cells compared to the IgG control treatment (Fig. 3D, panels i to iii). In contrast, the lungs of control IgG-treated mice showed increased fibrosis and bacterial spread (Fig. 3D, panels iv to vi). RNA sequencing analysis revealed that anti-cANGPTL4 MAb treatment improved immune function against secondary bacterial infection (FluϩS3), as well as coagulation to reduce bleeding and edema in the lungs (Fig. S2B). Hence, an cANGPTL4 deficiency reduces pulmonary edema and protects lung tissue integrity, suggesting that the use of anti-cANGPTL4 therapy during secondary bacterial pneumonia may be beneficial.  Antibody Therapy Improves Bacterial Pneumonia Survival ® efficacy in ameliorating lung edema, diminishing tissue damage, and prolonging survival time (Fig. 4B; Fig. S3B and C). As indicated in Fig. 4B and quantified in Fig. S3C, FluϩS3 superinfection destroyed the alveolar structures and caused severe edema in the lungs of infected wild-type mice. Antibiotic treatment alone protected the alveolar structures but could not clear the edema and infiltration of immune cells. Although the anti-cANGPTL4 or antipneumolysin antibody treatments alone improved lung tissue integrity and reduced edema, these improvements were not as significant as those observed when using a combined treatment with antibiotics and anti-cANGPTL4 MAb or with anti-cANGPTL4 and antipneumolysin antibodies ( Fig. 4B; Fig. S3C). The efficacy of a combined approach was further confirmed by marked improvements in the mouse survival rate and lung tissue integrity in treatment groups of wild-type mice receiving antibiotics, anti-cANGPTL4, and antipneumolysin MAb. Similar trends were observed in ANGPTL4 Ϫ/Ϫ mice receiving antibiotic and antipneumolysin treatment (Fig. S3).
Next, we performed transcriptomic analysis of the lung tissues to elucidate the detailed host response factors to secondary pneumococcal pneumonia during the various treatment modalities (Fig. 4C). Compared with antibiotic treatment alone, our analysis indicated that the anti-cANGPTL4 MAb treatment combined with an antibiotic notably improved the immune responses against bacterial infection as well as coagulation function, thus attenuating intra-alveolar hemorrhage and edema in the lungs of secondary bacterial pneumonia-infected mice. For example, this combination treatment yielded a 3-fold increase in activation of the coagulation system pathway compared to MAb treatment alone (Fig. 4C).
Immunoneutralization of pneumococcal pneumolysin abrogates its poreforming action on alveolar epithelium. Antibiotic resistance among clinical strains of S. pneumoniae underscores the urgency for alternative treatment strategies. Multimodal host-and pathogen-directed immunotherapy is a feasible option. Thus, we employed the experimental dual-infection mouse model to explore concurrent immunotherapy for secondary bacterial pneumonia. The host response protein cANGPTL4 and bacterial virulence factor pneumolysin were targeted using cognate neutralizing antibodies. The antipneumolysin antibody mitigated the pore-forming action of pneumolysin in human alveolar epithelial cells and reduced tissue damage. This effect persisted for 48 h (Fig. 4D). The concurrent antibiotic and antibody treatment significantly improved lung tissue integrity and, importantly, extended the median survival time of mice with secondary bacterial pneumonia compared to treatment with either the antibiotic or a single antibody ( Fig. 4A and B). Taken together, these observations highlight that host-directed therapeutic anti-cANGPTL4 MAb can complement pathogen-directed treatment, such as conventional antibiotics or a novel antipneumolysin antibody, to enhance lung tissue integrity and augment host survival during secondary pneumococcal pneumonia. The improved tissue integrity and considerably longer median survival time suggest that a multimodal treatment approach targeting both host and pathogen factors can be highly efficacious against DRSP infection.
Immunotherapy against cANGPTL4 enhances host immune responses against secondary pneumococcal pneumonia. To gain further insights into the host responses to secondary pneumococcal infection and the various treatments, we analyzed the immune cells and gene expression of the infected mouse lungs. qPCR on fluorescence-activated cell sorter (FACS)-sorted major immune cell groups revealed that macrophages, neutrophils, natural killer (NK) cells, cytotoxic T cells, and B cells expressed ANGPTL4 at various levels during secondary pneumococcal pneumonia (Fig. 5A). In FluϩS3-infected mouse lungs, semidigested bacteria appeared within the lymphocyte cytoplasm, while the antibody-mediated neutralization of cANGPTL4 augmented the ability of macrophages to phagocytose and fully digest the bacteria (Fig. 5B). The improved phagocytosis and bacterial digestion suggest that anti-cANGPTL4 treatment enhances host immune function.
Next, we further interrogated the transcriptome of lung tissues to understand the detailed host response factors to secondary pneumococcal infection and the various treatments. Our analysis showed that the anti-cANGPTL4 antibody treatment, either alone or in combination with antibiotic treatment, increased the expression of host immune defense-related genes (such as those coding for interleukin-1␤ [IL-1␤], Toll-like receptor 4 [TLR4], NF-B, tumor necrosis factor [TNF], and gamma interferon [IFN-␥]), while decreasing the expression of anti-inflammatory genes (such as IL10RA) (Fig. 5C). Interestingly, the combined treatment using anti-cANGPTL4 and an antibiotic generated a greater enhancement of gene expression than the anti-cANGPTL4 treatment alone, whereas antibiotic treatment alone produced the opposite effect in terms of the regulation of these genes (Fig. 5C). Given that an insufficient host immune response is a major contributory factor to the lethal outcome of secondary bacterial pneumonia, the apparently positive effects of anti-cANGPTL4 therapy on the immune response could be considered beneficial to the host's defense against bacterial infection (1) and are consistent with the improved phagocytosis and bacterial digestion (Fig. 5B).
Elevated cANGPTL4 protein levels in the bronchoalveolar lavage fluid of patients with pneumonia. All clinical isolates of S. pneumoniae secrete toxic poreforming pneumolysin. This supports a potential broad application of an antipneumolysin antibody in new treatments against DRSP infection (12). Similarly, to underscore the potential use of anti-cANGPTL4 therapy in pulmonary diseases, we examined the protein level of cANGPTL4 in bronchoalveolar lavage fluid (BALF) from patients with acute respiratory distress syndrome (ARDS; n ϭ 7). Compared with healthy volunteers (control; n ϭ 16), the ARDS BALF samples showed significantly higher cANGPTL4 protein levels (ϳ15-fold), as detected by immunoblotting (Fig. 6A). However, this marked increase in cANGPTL4 was not observed in serum samples from pneumonia patients (n ϭ 17) compared with non-pneumonia patients (control; n ϭ 8) (Fig. 6B). This finding Antibody Therapy Improves Bacterial Pneumonia Survival ® indicates that the cANGPTL4 protein was elevated in the pulmonary fluid of patients with severe pneumonia, whereas its levels in the blood circulation may be influenced by other factors. It also supports the potential utility of cANGPTL4 as a biomarker and therapeutic target for the diagnosis and treatment of infection-associated lung injury.

DISCUSSION
S. pneumoniae is a major cause of acute pneumonia and is the most commonly identified pathogen in secondary bacterial lung infection (4). The increasing prevalence of multidrug-resistant S. pneumoniae also poses a serious public health concern worldwide, particularly in Asian countries (4,16,26). Secondary bacterial pneumonia presents a unique set of host responses and challenges. As this is a rapidly progressing infection, often with a poor prognosis, newer approaches for improving treatment outcomes are needed to reduce its high morbidity and mortality. Few studies have examined the effect of antibody-based immunotherapy on secondary pneumococcal pneumonia. In this study, a multimodal treatment that concurrently targets the host and pathogen factors in an experimental mouse model of secondary S. pneumoniae infection was shown to significantly improve lung tissue integrity and extend the median survival time of infected mice. Immunoneutralization of the host protein cANGPTL4 significantly diminished the severity of pulmonary edema and damage when combined with pathogen-directed therapies, such as antibiotics or antipneumolysin MAb, which attenuated bacterial replication or toxicity, ultimately considerably prolonged the survival of infected mice. Effects of anti-cANGPTL4 antibody treatment in secondary bacterial infections were also confirmed using ANGPTL4 Ϫ/Ϫ mice.
Complications or death arising from influenza and secondary pneumonia are often associated with hyperinduction of proinflammatory cytokines and chemokines, a phenomenon also known as a "cytokine storm" or hypercytokinemia (27)(28)(29)(30). More patients die from ARDS, a common consequence of hypercytokinemia, than from the infection per se (31). Pathogen-induced lung injury can manifest as mild to severe ARDS, as seen with the severe acute respiratory syndrome (SARS) coronavirus, influenza virus, and secondary bacterial pneumococcal infections (1,31,32). Although inflammatory processes constitute important treatment targets (32-34), anti-inflammatory treatments can inhibit the immune functions that mediate pathogen clearance. A compromised  host immune response during influenza infection renders the host more vulnerable to secondary bacterial infection. This is particularly crucial since secondary bacterial infection is usually facilitated by the influenza virus, which impairs host immune responses and alters cytokine production, thus favoring bacterial superinfection and greatly increasing host morbidity and mortality (1,32,(35)(36)(37)(38). For example, treatments targeting proinflammatory cytokines, reactive oxygen species (ROS), and reactive nitrogen species (RNS) come with a risk of enhancing pathogen replication (39,40). Ideally, potential treatment regimens should minimize the tissue damage caused by inflammation and facilitate recovery without interfering with the host's antiviral or antibacterial responses.
Both host and bacterial factors contribute to severe tissue damage in pneumonia, thus offering critical targets for therapy development (41,42). The therapeutic use of passive antibodies has been well established for several primary viral infections (43)(44)(45). Similarly, the passive immunization afforded by antibodies to the pathogenic virulence factor pneumolysin can enhance the survival rate of mice with primary bacterial pneumonia (13). The pneumolysin gene has very limited variability and is therefore a prime pathogen target (46). Our antipneumolysin MAb targets the oligomerization domain, which is conserved across strains-i.e., it is not strain-specific, unlike previously reported antipneumolysin antibodies (13). Recent insights into pathogen-host interactions and host innate immune responses have also led to the identification and development of a wide range of host-directed therapies with different mechanisms of action. One such approach is the targeting of vascular leakage to combat respiratory diseases (24,(47)(48)(49). Recent studies have identified cANGPTL4 as a key player in tissue leakiness (24,50,51). Primary influenza pneumonia and lipopolysaccharide-induced acute lung injury are associated with an increased expression of cANGPTL4 (23,24,48). Influenza-infected cANGPTL4-deficient mice exhibit reduced lung tissue leakiness and damage compared to mice with functioning cANGPTL4 (24). Similarly, ANGPTL4 suppression by small interfering RNA (siRNA) also protects mice against lipopolysaccharideinduced acute lung injury (48). Indeed, compared to individual therapies, our concurrent treatment with anti-cANGPTL4 MAb and moxifloxacin considerably extended the survival of mice with secondary bacterial pneumonia. Due to the different modes of action compared to antibiotics, immunotherapies with antibodies are highly effective against secondary DRSP infections caused by antibiotic-resistant pathogens. The coimmunoneutralization of cANGPTL4 and pneumolysin was effective against secondary bacterial pneumonia, culminating in improved pulmonary function and longer survival times. Taken together, our study highlights that host-directed therapeutic strategies are valuable adjuncts to complement and enhance standard antimicrobial treatments.

MATERIALS AND METHODS
Virus and bacteria. Influenza virus A/Puerto Rico/8/34 H1N1 strain (PR8) was obtained from the American Type Culture Collection and propagated in embryonated eggs. Viral titers were determined by plaque assay. Serotypes 3 and 19F of S. pneumoniae were derived from clinical isolates in Singapore and cultured in brain heart infusion broth (Sigma 53286) supplemented with 5% heat-inactivated fetal bovine serum (FBS) under anaerobic conditions until the mid-logarithmic phase.
Animals and infection. Female BALB/c mice (8 to 12 weeks old) were purchased from InVivos, Singapore, and kept at the biosafety level 2 (BSL-2) animal facility of Nanyang Technological University, Singapore. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at Nanyang Technological University, Singapore (A0321 and A0200AZ), and National University of Singapore (050/11). Mice were anesthetized using 80 mg/kg body weight ketamine and 10 mg/kg body weight xylazine and infected via intratracheal delivery with the PR8 virus at a sublethal dose of 10 PFU in phosphate-buffered saline (PBS). At day 7 post-influenza infection, 100 CFU of serotype 3 or 5,000 CFU of serotype 19F of S. pneumoniae in PBS was delivered intratracheally to the mice. Sham infections were performed using sterile PBS. Immediately after bacterial infection, mice were injected intraperitoneally with moxifloxacin (100 mg/kg body weight), anti-ANGPTL4 antibody (10 mg/kg body weight), antipneumolysin antibody (10 mg/kg body weight), or control IgG (10 mg/kg body weight) daily starting from day 7 post-influenza infection until harvest. The mice were euthanized at specified time points, and lung tissues were harvested. Lung lobes were either fixed in 4% paraformaldehyde for histologic analysis or snap-frozen in liquid nitrogen and kept at Ϫ80°C for subsequent assays.
Antibodies. The anti-mouse cANGPTL4 antibody and anti-human cANGPTL4 antibody were produced in-house as described previously (52,53). The antipneumolysin antibody was prepared by culturing an MG12 hybridoma culture in Dulbecco's modified Eagle's medium (DMEM) with 10% FBS and purified from the culture medium using the IgG purification kit (Merck Millipore, no. LSK2ABG20). The anti-CD68 antibody was purchased from Abcam (no. 31630).
Immunofluorescent staining of lung sections. Fixed mouse lungs were dehydrated and embedded in paraffin. Five-micrometer sections were mounted on slides coated with Superfrost Plus (Thermal Scientific). Sections were dewaxed with xylene and rehydrated in water. Immunofluorescent staining was performed as previously described (24). Sections were mounted with an antifade reagent containing DAPI (4=,6-diamidino-2-phenylindole [Invitrogen]), and images were captured by Observer Z1 fluorescence microscope (Carl Zeiss). Hematoxylin and eosin (H&E) staining was performed for histopathologic analysis of the lung tissues. S. pneumoniae in the lung sections was detected by employing fluorescent in situ hybridization (FISH) as described, with minor modifications (54). Briefly, following the dehydration of fixed lung sections in an ethanol series (3 min each in 50, 80, and 98% ethanol), overnight hybridization to 16S rRNA of S. pneumoniae was performed with an Spn probe tagged with Cy5 in a hybridization buffer containing 30% formamide (55). After the washing steps, the sections were mounted with an antifade reagent and imaged using the LSM780 inverted confocal laser-scanning microscope (Carl Zeiss) fitted with a Plan Apochromat 100ϫ/1.4 oil lens objective, with excitation at 561 nm (autofluorescence) and 633 nm (Cy5). Microscopic images were processed using the Imaris software (version 8.2.0). S. pneumoniae was pseudocolored green.
Sorting of immune cells. Immune cells were harvested from influenza virus-and S. pneumoniae serotype 3-coinfected mice at day 9 post-influenza virus infection by bronchoalveolar lavage. One milliliter of sterile PBS was injected into the mouse lung through the trachea and retrieved into the syringe. The BALF contained immune cells and damaged lung tissues. Cells were pelleted by centrifugation at 500 ϫ g for 10 min, stained for flow cytometry, and sorted using a BD FACSAria. Quantitative real-time RT-PCR and immunoblotting. Total RNA was extracted and reverse transcribed as previously described (56,57). PCR was performed with the Bio-Rad CFX-96 real-time system using KAPA Sybr fast qPCR master mix according to the manufacturer's instructions (KAPA Biosystems). The primers used were mouse ANGPTL4 (forward, 5=-CCCCACGCACCTAGACAATG-3=; reverse, 5=-CCTCC ATCTGAAGTCATCTCA-3=) and housekeeping primer RPL27 (forward, 5=-CGCAAAGCTGTCATCGTG-3=; reverse, 5=-CGCAAAGCTGTCATCGTG-3=). Lung tissues were lysed with M-PER mammalian extraction buffer (Pierce), and the protein concentration was determined using the Bio-Rad protein assay kit. Protein samples were heated in 2ϫ Laemmli sample buffer, resolved by SDS-PAGE, and transferred to lowfluorescence polyvinylidene difluoride (PVDF) membranes (Millipore). Immunoblotting was performed as previously described (58).
RNA sequencing. Total RNA was extracted from frozen mouse lung tissues using TRIzol (Thermo Fisher). The integrity of the total RNA and the level of DNA contamination were assessed with an Agilent 2100 bioanalyzer (Agilent Technologies) and Qubit 2.0 fluorometer (Invitrogen). Library preparation was performed according to the mRNA-seq sample preparation kit (Illumina) and sequenced using the Illumina MiSeq platform with single-end 75-bp sequencing. Sequencing reads were mapped onto the viral, S. pneumoniae, and mouse genomes using CLC Genomics Workbench 9.0. Differential gene expression was determined using the R statistic package DESeq2 (59). The following criteria were adopted to filter the unique sequence reads: maximum number of hits for a read of 1, minimum length fraction of 0.9, minimum similarity fraction of 0.8, and maximum number of two mismatches. A constant of 1 was added to the raw transcript count value to avoid any problems caused by 0. The DESeq2 package from R/Bioconductor was used for the statistical analysis of the transcript count table. The transcript counts were normalized to the effective library size. The differentially expressed genes were identified by performing a negative binomial test. Transcripts were determined as differentially expressed when they showed a Ͼ2-fold change with an adjusted P value of Ͻ0.05. Pathway analysis and data presentation were performed using Ingenuity Pathway Analysis software (Qiagen Bioinformatics).
Measurement of cANGPTL4 in BALF and serum from lung disease patients. BALF samples were obtained from pneumonia patients who fulfilled the ARDS diagnostic criteria of the American European Consensus Conference (60). Controls for BALF samples were obtained from healthy volunteers. BALF was collected with three instillations of sterile physiological saline (50 ml) through a flexible bronchoscope as previously described (61). The collected lavage fluid was passed through two sheets of gauze and centrifuged at 400 ϫ g for 10 min at 4°C, and the supernatant was stored at -20°C until use. This study protocol was approved by the Institutional Review Board of Nagasaki University Hospital, Japan. Serum samples were obtained from patients with diagnosed pneumonia before successful treatment with antibiotics from the Biological Resource Center Ferdinand Cabanne (BB-0033-00044; Dijon) and the Centre Hospitalier Universitaire de Dijon, France. Informed consent was obtained from all subjects, and all experiments conformed to the principles outlined in the WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report.
Statistical analysis. Statistical analysis was performed with the unpaired t test using GraphPad Prism version 5.03 (GraphPad Software). Values are depicted as mean Ϯ standard error of mean (SEM). A P value of Ͻ0.05 was considered statistically significant.
Accession number(s). RNA sequencing data have been submitted to the NCBI BioProject database under BioProject no. PRJNA394932 and SRA accession no. SRP113049.