During pneumococcal infection, interaction of bacteria with the host epithelium initiates recruitment of neutrophils (Yamamoto et al., 2014). Following recruitment, activated neutrophils are capable of inducing programmed cell death in epithelial cells (Jia et al., 2014). Pneumococcus can induce both direct and indirect apoptosis in neutrophils (Zysk et al., 2000; Engelich et al., 2001), epithelial cells (Schmeck et al., 2004), and macrophages (Marriott et al., 2004), observations that have been mechanistically tied to PLY, cell wall components, and H₂O₂ (Marriott and Dockrell, 2006; Rai et al., 2015). In murine pneumococcal pneumonia, macrophage apoptosis has been shown to be crucial for resolution of pulmonary inflammation (Marriott et al., 2006). Induction of apoptosis, depending on cell type, can have detrimental or beneficial results, but coordinated and organized cell death is essential during infection resolution. Apoptosis modulates innate immune signaling, cell fate of responding phagocytes, promotes resolution of inflammation, and limits tissue damage (Haslett, 1999; Kennedy and Deleo, 2009; Esmann et al., 2010; Fox et al., 2010; Bordon et al., 2013; Greenlee-Wacker, 2016). The resultant effects of pneumococcal injury and apoptosis within the tissues strongly correlates with wound healing and repair (Matute-bello and Martin, 2003; Geiser et al., 2004; Wu and Chen, 2014; Domon et al., 2016; Holzmann et al., 2016). Numerous experimental in vitro, in vivo, and clinical studies have reported several pro-inflammatory cytokines up-regulated during the early phase of pneumococcal pneumonia, including IL-6, G-CSF, TNF-α, IL-1β, IL-12, IFN-ϒ, and IP-10 (Van Der Poll et al., 1997; Ethuin et al., 2004; Yamamoto et al., 2004; Palaniappan et al., 2006; Sun et al., 2007; Endeman et al., 2011; Yamada et al., 2011; Gomez et al., 2015; Skovbjerg et al., 2017; Loughran et al., 2018). However, in the last two decades, over-expression of IL-12, IFN-ϒ, and IP-10 has emerged as a highly potent pro-inflammatory, neutrophil-mediated signaling axis in murine models of pneumococcal pneumonia (Yamamoto et al., 2004; Sun et al., 2007; Seyoum and Yano, 2011; Yamada et al., 2011; Gomez et al., 2015).

Neutrophils are the primary effector cells of the host innate response to pneumococcal pneumonia (Craig et al., 2009). In pneumococcal pneumonia patients, the activation of neutrophils after extravasation into the lungs delays the rate of spontaneous apoptosis (Droemann et al., 2000). However, apoptosis of neutrophils is an essential feature of pneumococcal infection, as apoptotic neutrophils must be cleared from the site upon performing their effector functions (Koedel et al., 2009). Pneumococcus can induce apoptosis in neutrophils in an H₂O₂-dependent fashion and accelerate apoptosis by PLY membrane perforation at relatively lower multiplicities of infection (MOI) (Zysk et al., 2000).

PUMA (p53-up-regulated modulator of apoptosis), a pro-apoptotic member of the BH3-only family of the BCL-2 superfamily of eukaryotic proteins, is one of the primary effector proteins that promotes apoptosis (Nakano and Vousden, 2001; Vousden, 2005). In a murine model of pneumococcal infection, neutrophil-mediated protection against lethal dissemination of pneumococci was shown to be Puma-dependent apoptosis (Garrison et al., 2010). However, it is unclear which steps of apoptosis signaling pathways are involved in the early phase interactions between neutrophils and pneumococcus. Subsequently, the effects of Puma-dependent apoptosis on global gene expression and innate immune signaling in the lung during the early phase of pneumococcal pneumonia have not yet been investigated.

In this study, we attempted to characterize Puma-dependent responses in the lung during the early phase of pneumococcal pneumonia. Given the clinical relevance of targeting the cytokine response during the early phase of the innate immune response during acute community acquired pneumonia (Woods and José, 2017), we sought to quantitate differences in inflammatory cytokines in the lungs of Puma^+/+ and Puma^-/- mice. Finally, to further understand what Puma-dependent transcriptional responses may be induced during early pneumococcal infection, we attempted to characterize changes in the lung transcriptome.

Apoptosis is integral in the host defense response against pneumococcal infection. Neutrophils are the primary cell type responsible for direct clearance of pneumococcus from the lung (Garrison et al., 2010). Phagocytosis of pneumococcus induces apoptosis in neutrophils via phagocytosis-induced cell death (PICD) (Kobayashi et al., 2017). As neutrophils become apoptotic, they are rapidly cleared by macrophages to avoid the detrimental effects of secondary necrosis and prolonged inflammation within the lung (Savill, 1997; Henson and Bratton, 2013). Neutrophil apoptosis directly and indirectly signals surrounding cells to aid in resolution (Henson and Bratton, 2013; Jia et al., 2014). Apoptosis of other cell types also contributes to pneumococcal infection resolution, including macrophages (Marriott et al., 2004; Rai et al., 2015), epithelial cells (Schmeck et al., 2004; Marriott and Dockrell, 2006), and T-cells (Kemp et al., 2002). During the early innate immune response in the lung, all of these cell types may contribute to apoptosis-induced signaling. Multiple cell types have been documented undergoing PUMA-induced apoptosis, including epithelial cells (Dang et al., 2020), neutrophils (Garrison et al., 2010), and T-cells (Fischer et al., 2008).

PUMA distinctly contributes to protective infectious disease responses during acute viral infection by herpes simplex virus (HSV-1) in T-cells (Fischer et al., 2008), and degradation of PUMA induced by Chlamydia trachomatis contributes to persistent infection (Dong et al., 2005). However, PUMA-induced apoptosis contributes to Helicobacter pylori pathogenesis during gastritis (Dang et al., 2020). These opposing roles of PUMA during infection may lie in the pathophysiology of the respective pathogen and the cell types involved. Puma-deficient mice have been intensively studied for their resistance to cancer-causing stimuli (Garrison et al., 2010). HSV-1 can induce oral cancer in mice and is a suspected causative cancer agent in humans (Fischer et al., 2008), and H. pylori is now recognized as a causative agent of gastric cancer. C. trachomatis is an obligate intracellular pathogen and epidemiologically correlated with cancers of the female urogenital epithelium, and almost exclusively infects epithelial cells (Zhu et al., 2016). HSV-1 can invade most cell types but primarily infects epithelial cells and resides in neurons as long-term reservoirs within the host. H. pylori and pneumococcus are both generally acknowledged as extracellular pathogens and each induces strong neutrophil recruitment, which exacerbates local tissue damage. The up- and down-stream inducers of puma in the context of the immune response to infection are currently poorly understood. The contradictory observations on the Puma^-/- mouse model in cancer studies versus our pneumococcal pneumonia model continue to make the Puma phenotype and its up- and down-stream regulators molecules of interest. Our findings support the notion that Puma’s role in infectious disease immunobiology is multifactorial and further demonstrates that loss of Puma may enhance risk of negative health outcomes during infection. The precise pathways involved in the protective phenotype of Puma remain to be thoroughly investigated in the context of infection.

While many cytokines have been associated with inflammation modulation during pneumococcal infection, the pro-inflammatory IL-12/IFN-ϒ/IP-10 signaling axis has emerged as a potent arm of the innate immune response. Infection-induced IL-12 secretion was first identified from dendritic cells and neutrophils and can be induced by pneumococcus via TLR3 activation (Rijneveld et al., 2002; Groom and Luster, 2011). The effects of increased IL-12 during pneumococcal pneumonia also include stimulating IFN-ϒ-dependent neutrophil recruitment through increased abundance of MIP-2 and TNF-α (Ethuin et al., 2004; Sun et al., 2007). At 24 hours p.i. Puma^+/+ mice expressed less IL-12 in their lungs than Puma^-/- . This could be due to the anti-inflammatory effects of apoptotic neutrophils, as pro-inflammatory signaling is dampened once apoptosis is initiated in neutrophils (Haslett, 1999). In accordance with this notion, we previously demonstrated that Puma^-/- lungs contained more neutrophils than their Puma^+/+ counterparts (Garrison et al., 2010). However, pro-inflammatory efferocytosis of neutrophils also induces an IL-12^low phenotype in macrophages (Filardy et al., 2010). This likely contributes to the low levels of IL-12 in Puma^+/+ lungs following infection, but the precise mechanisms of temporal IL-12 production and its effect on cell death in specific cellular subsets within the lung during early pneumococcal pneumonia need to be further investigated.

The importance of IFN-ϒ signaling is well demonstrated during the early host response to pneumococcal infection. A previous study showed that pneumococcal infection induced significant increases of IFN-ϒ from greater than 98% of neutrophil in the lungs at 24 hours p.i (Yamada et al., 2011). Neutrophilic production of IFN-ϒ during early pneumococcal pneumonia is likely mediated through NADPH oxidase activity, suggesting IFN-ϒ production in neutrophils may be ROS-dependent, the generation of which is also a result of induction of apoptosis (Gomez et al., 2015). Production of IFN-ϒ in pneumococcal pneumonia models has been linked to strain-dependent effects, and suppresses pneumococcal outgrowth during the early phase by recruiting neutrophils (Rubins and Pomeroy, 1997; Groom and Luster, 2011). As a ligand for the CXCR3 receptor, IP-10 can recruit and activate effector CD4⁺ Th1, CD8⁺ CTL, natural killer (NK), and the hybrid NK-T cells (Christaki et al., 2015). In human neutrophils, the combination of IFN-ϒ and TNF-α is the most potent inducer of IP-10 secretion (Cassatella et al., 1997; Gasperini et al., 2019). In accordance with this observation, IFN-ϒ and TNF-α were significantly higher in Puma^-/- mice. IP-10 expression is also strain-dependent and associated with acute lung injury during early pneumococcal pneumonia (Seyoum and Yano, 2011). Interestingly, secretion of IL-12 and IP-10 are increased by human peripheral blood monocytes upon exposure to intact pneumococci, but inhibited when exposed to autolyzed pneumococci (Skovbjerg et al., 2017). In pneumococcal pneumonia, IFN-ϒ signaling is thought to exert beneficial effects by activating NK-T cells but detrimental effects upon activating NK cells (Christaki et al., 2015). However, a role for Puma in these cell types has not yet been characterized. It should be noted, though, that in an experimental pneumonia murine model at 24 hours p.i., T cells, NK cells, and macrophages did not produce IFN-ϒ (Gomez et al., 2015). The increase of IFN-ϒ and subsequent IP-10 production in Puma^-/- mice 24 hours p.i. underscores the importance of apoptosis in resolving inflammation. While both groups displayed increased interstitial inflammation, overall, Puma^-/- mice exhibited a greater degree of inflammation with multiple foci and vasculitis. The excessive inflammation observed in the Puma^-/- mice is likely a result of accumulation of neutrophils that remain active and continue to propagate the IL-12/IFN-ϒ/IP-10 axis. Previous works of others using murine models and data presented here strongly suggest that PUMA-dependent neutrophil apoptosis suppresses the IL-12/IFN-ϒ/IP-10 signaling axis, thereby limiting inflammation and consequent pneumococcal dissemination. However, it is currently unknown if the reported IL-12/IFN-ϒ/IP-10 signaling in mice recapitulates the acute-phase cytokine signaling in humans. Previously, IFN-ϒ had been ruled out as an acute-phase cytokine in humans (Endeman et al., 2011). Although, another clinical study found significant differences in bronchoalveolar lavage fluid IFN-ϒ between healthy control subjects and those with CAP and in serum between severe and non-severe CAP patients upon admission (Paats et al., 2013). Beyond the intrinsic differences in murine and human physiology, a likely explanation for these differences may be the temporal differences in surveillance and detection of infection in humans, with testing only occurring within about 72 hours of the first onset of symptoms (Mandell et al., 2007). Consequently, the mechanistic details of apoptotic suppression of the IL-12/IFN-ϒ/IP-10 pro-inflammatory axis, temporospatial regulation of the intercellular signaling, and whether the cascade results in feedback in humans during development of bacterial pneumonia remain unknown.

Puma-dependent responses also regulate the pulmonary transcriptome. In the lung, ADAM19 is expressed in a variety of cell and tissue types. ADAM19 is expressed higher in bronchiolar than alveolar epithelial cells but expressed higher in smooth muscle cells than in bronchiolar epithelial cells (Dijkstra et al., 2009). However, ADAM19 is only weakly expressed in vascular endothelial cells. The detection of ADAM19 on the apical surfaces of airway epithelial cells may suggest a role in the positive regulation in the early innate immune response to infection since ADAM19 can induce inflammation via shedding of TNF-α (Becherer et al., 2003). ADAM19 also contributes to angiogenesis, neurogenesis, synaptogenesis, and adhesion of and invasion by leukocytes (Shirakabe et al., 2001; Dreymueller et al., 2012; Fukazawa et al., 2013). Several of the other ADAM family members have reportedly been observed modulating the immune response to multiple infectious agents, including pneumococcus (Aljohmani and Yildiz, 2020). Down-regulation of Adam19 in Puma^-/- lungs suggests an immunomodulatory role in the early innate immune response that is short-lived due to yet unidentified Puma-dependent responses to pneumococcal infection. Nrxn2, a member of the neurexin gene family, encodes a transmembrane protein on pre-synaptic neurons that forms a Ca²⁺-dependent complex with the protein neuroligin on post-synaptic neurons (Bottos et al., 2009). In addition to regulation of synaptic transmission in the nervous system, neurexin-neuroligin complexation has been observed in vascular tissue. Each neurexin gene encodes a long (α) and short (β) isoform. While the sequence mapping could not discern Nrxn2 α or β transcripts in the analysis reported here, β isoforms have been detected in complexes with neuroligin in the vascular system while positively modulating angiogenesis in conjunction with vascular endothelial growth factor and increasing vascular tone (Berk et al., 2005; Graziano et al., 2013). Eln, the gene encoding elastin, was also up-regulated in a Puma-dependent manner. One explanation for this observation could be that the accumulation of neutrophils in the lung and the hypoxic nature of pneumonia suppress elastin fiber deposition and repair (Berk et al., 2005; Benjamin et al., 2021). Together, these data suggest a Puma-dependent integrated systems physiologic host response to pneumococcal pneumonia. However, the details of these intermolecular and intercellular interactions within the lung have not been thoroughly interrogated during infection.

For the first time, we have demonstrated the immunomodulatory effects of Puma during pneumococcal pneumonia. Puma-dependent responses suppressed pro-inflammatory cytokine signaling, including the highly inflammatory, neutrophil-mediated IL-12/IFN-ϒ/IP-10 signaling axis. Dissemination of pneumococcus and erosion of the pulmonary vasculature was attenuated by Puma-dependent responses. The precise, temporal expression and regulation of IFN-inducible genes remains to be elucidated in the context of Puma-dependent resolution of inflammation. Finally, the activity of ADAM19 during pneumococcal and other bacterial infections also requires further, more intensive study, as well as to what degree, if any, NRXN2 is playing in pulmonary blood pressure regulation during the early phase of acute pneumonia.

The sequencing raw reads have been deposited in the NCBI Short Read Archive under Bioproject accession number: PRJNA748408.

The animal study was reviewed and approved by Mississippi State University Institutional Animal Care and Use Committee.

DK, PM, WT, and JR obtained the data. DK, J-FG, AO, JR, KS, and JT, performed the data analysis. DK, JT, and KS wrote the manuscript. JT was responsible for study concept and design. JT was a supervisor of laboratory analysis of this study. JT is the guarantor of this study and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors participated in data interpretation and review of the manuscript and approved the final manuscript.

Research reported in this publication was supported by the Center for Biomedical Research Excellence (COBRE) in Pathogen–Host Interactions, National Institute of General Medical Sciences, National Institutes of Health (P20 472 GM103646-06). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.