(Inter)feron the Reaper: New Insights on an Old Cytokine in Secondary Bacterial Pneumonia

Over the past ∼100 years, there have been four well-documented influenza pandemics. Autopsy and microbiological data indicate that secondary bacterial bronchopneumonia, predominantly due to Streptococcus pneumoniae (Spn), was likely responsible for the majority of deaths in the 1918 influenza A H1N1 (“Spanish flu”) pandemic of the preantibiotic era (1). Although mortality rates were considerably lower in the 1957, 1968, and 2009 pandemics, bacterial superinfection with Spn was again prevalent and was associated with increased morbidity and mortality (1, 2). Reports as early as 1949 had identified that influenza A virus (IAV) infection compromises lung clearance of Spn (3), but the mechanism remained obscure.

In 1957, Isaacs and Lindenmann reported that heat-inactivated IAV induces the release of a soluble factor that interferes with viral replication; they called this factor “interferon” (4). Since that discovery, three major families of IFNs (type I, typified by IFN-β; type II, or IFN-γ; and type III, the IFN-λs) have been identified and characterized for their complex effects on immunity, which are believed to occur through induction of hundreds of so-called IFN-stimulated genes (5). Type I IFN induced during IAV infection has in particular been implicated in causing susceptibility to secondary bacterial pneumonia. Thus, IFNAR (type I IFN receptor)–null mice have improved bacterial clearance and survival during post-IAV Spn pneumonia (6). IAV-induced type I IFN similarly compromises host defense against Staphylococcus aureus and gram-negative bacteria (7). Mechanisms identified include the suppression of lung neutrophilia and IL-17–dependent immunity (6, 8). Type I IFN is also well known to compromise host defense against tuberculosis through multiple mechanisms, including induction of IL-10 and suppression of IL-1 expression, IL-1 signaling, and IFN-γ signaling (1, 9). Complicating matters, however, reports have also identified protective roles for type I IFN in bacterial infections, including Spn, as well as some deleterious roles for type I IFN in viral infection (1, 9, 10). Moreover, as IFNAR is ubiquitously expressed, the primary cellular targets of type I IFN in secondary bacterial pneumonia have remained unclear. Macrophages have been of interest, as they reportedly show greater and differential expression of IFN-stimulated genes compared with B cells, T cells, and fibroblasts, and they undergo wide-ranging chromatin remodeling in the promoter regions of proinflammatory genes in response to type I IFN (5).

Against this backdrop, in this issue of the Journal, Palani and colleagues (pp. 264–274) use an array of approaches to provide evidence that IFNAR signaling specifically in alveolar macrophages (AMs) impairs antibacterial host defense in mice inoculated with Spn six days after IAV infection (11). The authors found that global IFNAR-null mice and myeloid (LysM-Cre)–specific IFNAR-null mice, but not neutrophil (Mr8-Cre)–specific IFNAR-null mice, had improved Spn clearance compared with control animals. A type I IFN reporter allele (Mx1^CremTmG) revealed IFNAR signaling in AMs, monocytes, and neutrophils of post-IAV/Spn-infected mice. Nonetheless, pointing away from monocytes, monocyte-derived AMs, and neutrophils being responsible for the effect of IFN, enhancement of Spn clearance with IFNAR deficiency persisted in monocytopenic Ccr2^−/− mice and neutrophil-depleted mice. By contrast, IFN-γ receptor–null mice also exhibited augmented post-IAV Spn clearance compared with wild-type (WT) mice, but this effect was abolished in the setting of neutrophil depletion, suggesting dependence on neutrophils.

Busulfan conditioning for bone marrow transfer does not deplete AMs and thus does not permit marrow-derived engraftment into the AM compartment unless AMs are otherwise depleted, as is the case in Csf2rb^−/− mice (11). Using this to their advantage, the authors showed that engraftment of IFNAR-null marrow into WT recipient mice (thus engrafting circulating and non-AM lung myeloid cells) does not enhance post-IAV Spn clearance, whereas it does enhance Spn clearance when performed in Csf2rb^−/− recipients (under which conditions the engrafted marrow also populates AMs). By contrast, IFNAR-null recipients (which retain native IFNAR-null AMs during busulfan-conditioned bone marrow transfer) have enhanced post-IAV Spn clearance regardless of whether they receive WT or IFNAR-null marrow. Finally, extending the findings to other viruses, the authors showed that respiratory syncytial virus (RSV) and human metapneumovirus but not human endemic coronavirus induce type I IFN responses in AMs at two days after inoculation. Spn lung burdens were higher in mice infected after RSV and human metapneumovirus but not human endemic coronavirus, and Spn burdens and mortality were both rescued in IFNAR-null mice in the setting of RSV. Taken together, the authors conclude that IFNAR signaling in AMs is responsible for IAV-induced compromise of anti-Spn host defense and that this is not specific to IAV but may also be seen, albeit with different kinetics that correspond to the type I IFN response, after other viruses.

The elegant approaches in this report notwithstanding, the evidence provided for AM-intrinsic IFNAR signaling is largely indirect. In addition to macrophages, monocytes, and neutrophils, LysM-Cre also targets dendritic cells (12). Adoptive intratracheal transfer of WT versus IFNAR-null AMs into Csf2rb^−/− mice might have allowed more specific insight into whether tissue-resident AM IFNAR signaling is sufficient. Conversely, it might have been interesting to test whether adoptive transfer of IAV-experienced WT versus IFNAR-null AMs is sufficient to confer Spn susceptibility. Additional AM-targeting Cre drivers, even if also imperfectly selective (12), might have lent more confidence to the conclusions; future intersectional genetic approaches will hopefully allow more specific AM targeting. Given that recent single-cell RNA sequencing studies have identified putative AM subsets (13), one also wonders which of these subsets are the most important IFNAR responders. Moreover, one wonders what the identity and spatial location of the major type I IFN–producing cells are in the IAV-infected lung and whether there is a post-IAV time point at which depleting these cells or using other host-directed interventions (e.g., baricitinib, IFNAR-blocking biologics) might be therapeutic. Another major issue not addressed in the current study is the identity of the AM function(s) compromised by type I IFN signaling and whether they are intervenable. Finally, as suggested by the authors, the anti-Spn host defense phenomenon is time dependent. One wonders how the findings of the present study relate to recent studies that have found that prior pulmonary viral infection is protective against secondary bacterial pneumonia at late timepoints (e.g., 4 wk after virus) (14).

In closing, the study by Palani and colleagues (11) adds to our knowledge on the complex effects of type I IFN in secondary bacterial pneumonia and generates new excitement about the potential for host-directed therapies but also generates many new questions. Given the wide heterogeneity of findings to date in rodent models, one is reminded that in vivo human models (15), wherever possible, will likely be critical in moving the field forward toward clinically actionable insights.