Interferon with Dogma in Cytokine Release Syndrome and Acute Lung Injury

Cytokine release syndrome (CRS) is characterized by elevated circulating cytokines, symptoms of systemic inflammation, and organ injury attributable to inflammation (1). Causes may be genetic, autoimmune, iatrogenic (e.g., chimeric antigen receptor [CAR] T-cell therapy), malignancy, or infectious (e.g., bacterial/viral sepsis or bacterial superantigens). If occurring with infection, organ injury typically exceeds what would normally be expected for the pathogen alone. Therapies used to treat iatrogenic or autoimmune CRS include corticosteroids, Janus kinase inhibition, and IL-6 inhibition. Because these therapies also improve clinical outcomes in moderate–severe coronavirus disease (COVID-19) (2), CRS and its relationship to COVID-19–induced acute respiratory distress syndrome has gained global attention while remaining controversial (3, 4). The incidence of CRS is also rising because of the use of CAR T-cell therapies in malignant disease, resulting in increased usage of critical care for associated organ injuries (5). As such, additional work to evaluate maladaptive immune responses in preclinical models of CRS is warranted.

It is widely accepted that Th1 (T helper 1) and Th17 (T helper 17) T-cell responses are pathogenic in CRS because of the production of IFN-γ and IL-17, respectively (1). In this issue of the Journal, Sun and colleagues (pp. 75–89) provide new and antidogmatic insights into the roles of IFN-γ and IL-17A in CRS-associated extrapulmonary acute lung injury (eALI) (6). To model CRS, staphylococcal enterotoxin B (SEB) was administered systemically to mice expressing the human major histocompatibility complex (MHC) protein HLA-DR3. Human MHC molecules bind SEB with greater efficiency than murine MHC, leading to more robust activation of T cells and release of prototypical CRS cytokines, IFN-γ, and IL-17A, resulting in significant lung injury (7, 8). To determine the specific effects of IFN-γ and IL-17A in CRS, they crossed the HLA-DR3 transgenic mice to IFN-γ or IL-17A knockout mice. In contrast to prior work in direct lung injury models (9–11) and CRS generally (1), the authors found that DR3.IFN-γ^−/− mice had significantly enhanced lung injury compared with DR3.WT and DR3.IL-17A^−/− mice. This included enhanced neutrophil infiltration, extracellular trap production, and release of S100A8/A9 in the lung after SEB challenge.

Whole lung gene expression at early time points, when few neutrophils have infiltrated the tissues, showed upregulation of matrix metalloproteinases and S100 proteins, indicating a potential role of resident lung cells in granulocyte activation and recruitment. DR3.IFN-γ^−/− mice had the highest concentrations of IL-17A, TNFα, IL-10, and several neutrophil/eosinophil chemokines in BAL fluid at peak injury. Interestingly, the Th2-type cytokines IL-4 and IL-5 were also higher in the BAL of DR3.IFN-γ^−/− mice as compared with DR3.WT and DR3.IL-17A^−/− mice, suggesting a protective role of IFN-γ through suppression of Th2 responses. Circulating IL-17A, IL-5, and CCL4 were also more highly upregulated in DR3.IFN-γ^−/− mice, whereas more traditional CRS cytokines such as TNF-α were reduced. Neutralization of IFN-γ in DR3.WT mice reproduced these findings. Surprisingly, though systemic IL-17A was increased, the enhanced inflammatory response in DR3.IFN-γ^−/− mice was not mediated by increased concentrations of IL-17A. In contrast to DR3.IFN-γ^−/− mice, DR3.IL-17A^−/− mice were protected from SEB-induced eALI and also exhibited the highest concentrations of IFN-γ. With the neutralization of IFN-γ, DR3.IL-17A^−/− developed a more severe lung injury. Taken together, these results suggest a reciprocal role of IFN-γ and IL-17A in modulating eALI mediated by SEB. IFN-γ appears to be protective through suppression of local Th2/Th17 responses and granulocyte activation/recruitment in the lung. In the absence of IFN-γ, IL-17A is dispensable for the enhanced ALI observed, suggesting that the Th2 responses or possibly other Th17 cytokines (IL-21 and IL-22) are pathogenic. Though not specifically evaluated in this study, the pathogenic role of IL-17A in this model seems to occur partly through its suppression of IFN-γ. The authors highlight the paradoxical protective role of IFN-γ in the lung and therefore urge caution regarding IFN-γ modulation in clinical ALI and acute respiratory distress syndrome.

Paradoxical and pleiotropic roles of cytokines have been recognized for decades, though our understanding of the biology behind this remains poor (12). In addition, the inflammatory response is compartmental in that different tissues respond differently to the same inflammatory mediators (13, 14). In fact, previous work by this group demonstrated the opposite roles for IFN-γ and IL-17A in CRS-induced small bowel inflammation and cell death; IFN-γ is pathogenic, and IL-17A is protective in the small bowel (8). Similar paradoxical IFN-γ effects have been reported in models of graft-versus-host disease (GVHD) (15, 16). IFN-γ protects against idiopathic pneumonia syndrome but promotes GVHD in the gut. In the GVHD model, the protective effects are because of IFN-γ signaling in the lung the parenchyma, implying tissue specificity. And yet, in other models of CRS, there are contradictory observations. In experimental macrophage activation syndrome induced by repeated administration of CpG, IFN-γ promotes pathogenic pulmonary and systemic inflammation (17). Besides its controversial role in ALI, IFN-γ has important antimicrobial effects that might influence the contextual net benefit or harm of increased IFN-γ in the setting of infectious CRS. These observations directly challenge the idea of a protective role of IFN-γ and urge a more specific type of caution: the effects of any inflammatory mediator, including IFN-γ, must be considered in the context of trigger, tissue, and systemic effects.

This study is novel and highlights deficiencies in our understanding of the immunopathology of CRS. Given the breadth of literature identifying IFN-γ as pathogenic, this study should lead us to pause and ask: In what other specific situations and through what mechanisms may IFN-γ be protective for the lung parenchyma? Future work will benefit from cell-specific genetic modulation to determine the relative importance of IFN-γ and its signaling molecules in immune and lung parenchymal populations. Complimentary models of infectious CRS will enhance our understanding of how the body balances regulators of host defense and injury, such as IFN-γ. Whether these observations will change clinical practice with IFN-γ immunomodulation remains to be seen. IFN-γ inhibitors are already approved for use in children for CRS owing to hemophagocytic lymphohistiocytosis (18). In addition, there are reports of IFN-γ inhibition improving refractory CRS owing to CAR T-cell therapy (19). In light of these novel and robust data, we should proceed cautiously with IFN-γ inhibition in CRS-associated lung injury. We should also recognize that a more tailored (e.g., compartment, trigger, and/or pathogen-specific) modulation of cytokines might be of most benefit to patients.