Viral infections and type 2 immune responses are thought to be critical for the development of chronic respiratory disease, but the link between these events needs to be better defined. Here, we study a mouse model in which infection with a mouse parainfluenza virus known as Sendai virus (SeV) leads to long-term activation of innate immune cells that drive IL-13–dependent lung disease. We find that chronic postviral disease (signified by formation of excess airway mucus and accumulation of M2-differentiating lung macrophages) requires macrophage expression of triggering receptor expressed on myeloid cells-2 (TREM-2). Analysis of mechanism shows that viral replication increases lung macrophage levels of intracellular and cell surface TREM-2, and this action prevents macrophage apoptosis that would otherwise occur during the acute illness (5–12 d after inoculation). However, the largest increases in TREM-2 levels are found as the soluble form (sTREM-2) long after clearance of infection (49 d after inoculation). At this time, IL-13 and the adapter protein DAP12 promote TREM-2 cleavage to sTREM-2 that is unexpectedly active in preventing macrophage apoptosis. The results thereby define an unprecedented mechanism for a feed-forward expansion of lung macrophages (with IL-13 production and consequent M2 differentiation) that further explains how acute infection leads to chronic inflammatory disease.
A critical step toward improved diagnosis and treatment of chronic inflammatory diseases depends on defining the immune mechanisms for the persistent accumulation of activated immune cells in the target tissue. In the case of the lung, clinical evidence suggests that acute infection with a respiratory virus might lead to chronic lung diseases such as asthma and COPD (Holtzman, 2012). To determine precisely how acute infection causes chronic lung disease, we developed a high-fidelity mouse model of this process. In this model, mouse parainfluenza virus (also known as Sendai virus, SeV) is substituted for the related human pathogen to achieve more efficient viral replication and thereby produce the severe acute illness and subsequent chronic respiratory disease that is typical of the pathology found in humans (Walter et al., 2002). Using this model system, we determined that postviral lung disease depends on airway progenitor epithelial cell (APEC) production of IL-33 to drive invariant NK T cells (iNKT cells) and lung macrophages toward IL-13 production (Kim et al., 2008; Byers et al., 2013). The result is IL-13–dependent inflammation (signified by type 2 activation and accumulation of lung macrophages) and airway mucus production (signified by MUC5AC mucin gene expression). This innate epithelial to immune cell loop also appears relevant to human disease because increased numbers of IL-33–expressing APECs are found in association with an IL-13 gene expression signature (including increased MUC5AC mRNA and protein) in the lungs of humans with severe chronic obstructive pulmonary disease (COPD; Kim et al., 2008; Agapov et al., 2009; Alevy et al., 2012; Byers et al., 2013).
In our previous work, we recognized that the APEC population was capable of self-renewal and inducible release of IL-33 to sustain ongoing activation of the innate immune system (Holtzman et al., 2014). However, the existing data did not explain the selective activation of the lung macrophage population and the special dominance of type 2 (M2) macrophages as a downstream part of the disease process. In the present study, we therefore aimed to better understand how the lung macrophage component of this disease process is triggered by acute infection and then is manifest for months. We reasoned that triggering receptor expressed on myeloid cells 2 (TREM-2) might contribute to this process because M2 polarization is associated with TREM-2 expression in isolated macrophages (Turnbull et al., 2006). In pursuing this possibility, we found that the soluble form of TREM-2 (sTREM-2) was linked to the development of chronic postviral lung disease and was active in promoting macrophage survival. The data stand in contrast to the conventional view that cleavage of cell surface TREM-2 to sTREM-2 results in an inactive end product. The results thereby provide for a previously unrecognized control over macrophage survival and a consequent type 2 immune response that can serve both as a pathogenic mechanism and as a therapeutic target and accompanying biomarker for chronic inflammatory disease.
Macrophage control of postviral disease
To further define the role of macrophages in our postviral mouse model of chronic lung disease (Walter et al., 2002), we first assessed the impact of a new strategy for macrophage deficiency. We previously showed that mice that were treated with clodronate or mice that were homozygous for the op mutation in the Csf1 gene (Csf1op/op) led to marked macrophage depletion and blockade of postviral lung disease, but both pharmacologic and genetic approaches to depletion led to mice that were also more susceptible to viral infection (Tyner et al., 2005; Kim et al., 2008). In the present case, we started with Csf1op/opT (op/opT) mice, which carry an osteocalcin-driven Csf1 transgene (T) that rescues the osteopetrotic, but not the macrophage defect, in op/op mice (Abboud et al., 2002). We then used these mice to generate heterozygous Csf1wt/opT (wt/opT) mice with a subtler defect in macrophage level and perhaps function. Of note, we still found that the increase in lung levels of tissue monocytes (SSClowLy6G–F4/80+CD11b+) and interstitial macrophages (SSChighCD11c–Ly6G–Siglec-F–F4/80+CD11b+) that occurs in WT (wt/wt) mice at 5 d post-inoculation (dpi) with Sendai virus (SeV) was significantly attenuated in wt/opT mice (Fig. 1 A and Fig. S1). We observed no increase (and instead found a significant decrease) in alveolar macrophages (SSChighCD11c+Ly6G–Siglec-F+F4/80+CD11b–) in wt/wt and wt/opT mice at 5 dpi, reflecting a predominant effect of Csf1 deficiency on tissue monocytes and interstitial macrophages during acute infection. Despite these differences in lung monocyte–macrophage levels, we found the same degree of acute illness (0–12 dpi) as signified by essentially identical body weight changes, viral titers, and pattern of tissue inflammation in wt/wt and wt/opT mice (not depicted).
In contrast to the similarities in acute illness in wt/wt and wt/opT mice, we found marked differences in the development of chronic lung disease that develops after clearance of infectious virus and is fully manifest at 49 dpi (Walter et al., 2002; Kim et al., 2008; Byers et al., 2013). Under these chronic conditions, we observed that the increases in each lung monocyte–macrophage population at 49 dpi in wt/wt mice were all markedly attenuated in wt/opT mice (Fig. 1 B). Tissue monocytes, which were the most abundant monocyte–macrophage population in the lung, exhibited no increase in wt/opT mice at 49 dpi. In concert with the decreased lung monocyte–macrophage levels in wt/opT mice, we also found that induction of IL-13 expression and mucous cell metaplasia in wt/wt mice was markedly inhibited in wt/opT mice at 49 dpi (Fig. 1 C). Similarly, the postviral increases in lung levels of Il13 mRNA and downstream Muc5ac mRNA and type 2 macrophage (M2) markers Arg1 and Chitinase 3–like-3 (Chi3l3) mRNA in wt/wt mice were completely blocked in wt/opT mice at 49 dpi (Fig. 1 D). These findings are consistent with our previous work showing that lung macrophages are critical for IL-13 production, and in turn, that IL-13 is required to drive mucus production and M2 differentiation in our postviral mouse model of lung disease (Tyner et al., 2006; Kim et al., 2008; Byers et al., 2013). The results also provide a basis for next determining what factors might regulate the expansion of disease-producing lung macrophages under these conditions.
TREM-2 control of postviral disease
Based on the association of TREM-2 with M2 differentiation (Turnbull et al., 2006), we investigated the role of TREM-2 in macrophage-dependent postviral lung disease. In that context, we found that the development of this type of lung disease was associated with increased lung Trem2 (but not other Trem) mRNA levels based on gene expression microarray and real-time qPCR assay (Fig. 2, A and B; and Table S1). Moreover, the development of postviral lung disease at 49 dpi in WT mice was markedly attenuated in Trem2−/− mice. Thus, Trem2−/− mice showed marked decreases in lung inflammation and airway mucus production based on tissue staining (Fig. 2 C) as well as significantly decreased lung levels of the key cytokine Il13 mRNA, the predominant mucin Muc5ac mRNA, and the M2 markers Arg1 and Chi3l3 mRNA based on real-time qPCR assay (Fig. 2 D). Dap12−/− mice exhibited a similar inhibition of postviral disease (Fig. 2, C and D), consistent with the association of TREM-2 and DAP12 phenotype in other systems (Bouchon et al., 2000; Kaifu et al., 2003; Turnbull et al., 2006). In contrast, Trem1/3−/− mice manifested no significant difference in the development of chronic postviral disease (Fig. 2, C and D), perhaps divergent from the pro-inflammatory role of TREM-1 in other disease models (Bouchon et al., 2001a; Schenk et al., 2007; Klesney-Tait et al., 2013).
The distinct nature of TREM-2 actions in postviral lung disease prompted a further examination of mechanism. We first established that there was no significant difference in acute illness among Trem2−/−, Dap12−/−, and Trem1/3−/− mice as signified by similarities in body weight change, viral titer and tissue distribution, and pattern of tissue inflammation (Fig. 3, A–D). Therefore, we returned to our interest in the role of TREM-2 in modifying long-term macrophage function. We first found that TREM-2 expression at 49 dpi was specific to lung tissue macrophages based on co-localization with macrophage markers F4/80 and Mac-3 in tissue immunostaining (Fig. 4 A). Similarly, TREM-2 expression was found in lung macrophages (CD45+F4/80+) and not stromal cells, e.g., epithelial cells (CD45–EpCAM+), based on flow cytometry staining (unpublished data).
As demonstrated previously (Kim et al., 2008), we also found a marked expansion of IL-13+ lung macrophages in WT mice at 49 dpi (Fig. 4 B). Here, we show that this expansion is markedly down-regulated in Trem2−/− mice based on tissue immunostaining (Fig. 4 B). In addition, FACS-based separation of lung macrophage populations revealed that tissue monocytes and interstitial macrophages exhibited a marked increase in Il13, Arg1, and Chi3l3 mRNA levels with the development of postviral disease at 49 dpi in WT mice, and this postviral induction was markedly attenuated in Trem2−/− mice (Fig. 4 C). In contrast, alveolar macrophages showed no induction of Il13 mRNA but nonetheless still manifested a slight increase in Arg1 mRNA and a marked increase in Chi3l3 mRNA at 49 dpi in WT mice. The observed levels of Il13, Arg1, and Chi3l3 mRNA in alveolar macrophages from WT mice were nearly identical to those in Trem2−/− mice (Fig. 4 C). This subset profile might again reflect a predominant effect of TREM-2 deficiency on tissue monocytes and interstitial macrophages. Given the predominance of the tissue monocyte population, these findings were consistent with the decrease in whole lung levels of Il13, Arg1, and Chi3l3 mRNA found in Trem2−/− mice.
Cell surface TREM-2 expression and function
We next assessed the regulation of TREM-2 expression and function under the current paradigm that TREM-2 expressed on the cell surface works in coordination with its adapter protein DAP12 to signal for macrophage activation (Turnbull et al., 2006). In the present model, cell surface expression of TREM-2 on lung macrophages was detected on tissue monocytes and interstitial macrophages but not alveolar macrophages (Fig. 5 A). Moreover, flow cytometry staining for TREM-2 was found exclusively on CD11b+CD11c– lung cells at 5 dpi (Fig. S2), which is consistent with localization to tissue monocytes and interstitial macrophages and not to alveolar macrophages or dendritic cells. Postviral induction of cell surface TREM-2 was found only at 2–8 dpi (Fig. 5, A and B) and was therefore similar to induction for lung levels of SeV, suggesting that SeV replication might increase cell surface TREM-2 on macrophages. To follow up this possibility, we studied BMDMs infected with GFP-labeled SeV and showed an increase in cell surface staining for TREM-2 that correlated closely with viral titer (Fig. 5, C and D). Moreover, SeV infection caused a dose-dependent increase in cell surface TREM-2 levels on BMDMs that was reproduced using poly(IC) treatment, and this increase was unaffected by DAP12, TLR3, or PKR deficiency (Fig. 5 E). In addition to the increase in surface TREM-2 levels, we also found that SeV infection caused a smaller but sustained increase in intracellular levels of TREM-2 in BM-derived macrophages (BMDMs) unaccompanied by any detectable production of sTREM-2 or any increases in Trem2 mRNA (Fig. 5, F–H). A similar increase in intracellular TREM-2 level was found in vivo at 5 and 49 dpi. Together, the results suggest that SeV promotes an increase in surface TREM-2 levels based on increased TREM-2 formation and either transport to or retention at the cell surface with relatively little if any TREM-2 cleavage from the cell surface.
The relatively transient increase in cell surface TREM-2 levels suggested an acute effect on macrophage function after viral infection. Consistent with this idea, we found that the increase in the number of tissue monocytes (and interstitial macrophages but not neutrophils) at 5 dpi in WT mice was attenuated in Trem2−/− mice (Fig. 6 A and not depicted). This attenuation was not accompanied by decreases in monocyte chemokine production based on lung levels of Ccl2 and Cx3cl1 mRNA at 3–5 dpi (Fig. 6 B) or decreases in monocyte influx into the lung based on tracking CD45.2 transferred monocytes by flow cytometry at 5 dpi (Fig. 6 C). However, we did detect increased macrophage apoptosis in Trem2−/− mice compared with WT mice based on immunostaining for active caspase-3 and Mac-3 at 5–12 dpi (Fig. 6, D and E), suggesting a role for TREM-2 in specifically promoting macrophage survival after viral infection. Cells that were active caspase-3+ were not also SeV+ (Fig. 6 F), suggesting that TREM-2 antiapoptotic actions were independent of viral cytopathic effect. In addition, the timing of apoptosis corresponded to a relative decrease in Csf1 mRNA and CSF-1 protein levels in the lung (Fig. 6, G and H), consistent with the link between CSF-1 and macrophage survival via DAP12 signaling (Otero et al., 2009). Moreover, we detected no defect in CSF-1 production in Trem2−/− mice suggesting that the CSF-1 survival signal required downstream activation of TREM-2 to prevent apoptosis. Indeed, we found a marked increase in apoptosis with CSF-1 withdrawal in BMDMs from Trem2−/− mice compared with WT control mice based on propidium iodide staining, Annexin V staining, and active Caspase-3 immunoblotting (Fig. 6, I–K). We detected a slight increase in apoptosis in BMDMs from Trem2−/− mice even in cultures with CSF-1, but this effect was also attributable to CSF-1 withdrawal because CSF-1 was added only once at the start of the 2-d culture period and was therefore being consumed and depleted from the culture medium. This condition therefore represents a partial degree of CSF-1 withdrawal and a consequent loss of the CSF-1 to TREM-2 survival signal in Trem2−/− cells. Together, our results suggested that cell surface TREM-2 might preserve macrophage survival and promote macrophage accumulation after viral infection and thereby lead to long-term macrophage-dependent disease. However, the relatively transient effect of TREM-2 during acute infection also raised the possibility that TREM-2 had additional effects at later time points that influenced long-term disease outcome.
sTREM-2 expression and function
To pursue the possible chronic actions of TREM-2 in the postviral model, we first determined whether there was long-term production of TREM-2 protein in the lung in this model. Indeed, we found a marked increase in lung levels of TREM-2 at 49 dpi compared with baseline or to the relatively small increase found at 5 dpi (Fig. 7 A). Given the absence of detectable cell surface TREM-2 at 49 dpi and the relatively small increase in intracellular TREM-2 at either 5 or 49 dpi (Fig. 5 A), this late increase in TREM-2 levels at 49 dpi was likely a marker of sTREM-2 formation in the diseased lung. Since IL-13 production was also markedly increased at 49 dpi, we questioned whether IL-13 might promote TREM-2 cleavage at the cell surface and consequent formation of sTREM-2. In support of this possibility, we found that IL-13 administration caused a time-dependent loss of cell surface levels of TREM-2 in BMDMs (Fig. 7 B). The effect of IL-13 was specific for the IL-13 receptor since the same effect was found for IL-4 (that also activates this receptor) but was not detectable for other type 2 cytokines (IL-5 and IL-9; Fig. 7 C). Treatment with IL-13 also resulted in a time-dependent increase in cellular TREM-2 levels in BMDMs, as expected for M2 differentiation, but caused an even greater increase in sTREM-2 levels released from BMDMs (Fig. 7 D). The increases in cellular TREM-2, especially the even greater increases in sTREM-2 production, were attenuated in BMDMs from Dap12−/− mice (Fig. 7 D). Moreover, the expected increase in sTREM-2 levels was not found during postviral disease in Dap12−/− mice (Fig. 7 E). Together, these findings indicated that both IL-13 and DAP12 were contributing to sTREM-2 formation and any consequent function of sTREM-2 in chronic postviral disease.
To pursue a possible role for sTREM-2 in postviral disease, we analyzed the effect of recombinant sTREM-2 (amino acids 19–136) on macrophage function. Immunostaining of BMDMs incubated with sTREM-2 showed that the protein was localized to the cell cytoplasm and not the cell surface and was thereby accessible to intracellular proteins (Fig. 8 A). In that context, we found that sTREM-2 administration to BMDMs caused a significant increase in Arg1 mRNA levels (as a sign of M2 differentiation), but the increase was not preserved for other M2 markers (e.g., Chi3l3 mRNA) and was not nearly as effective as IL-13 for driving these responses (Fig. 8 B). Thus, the direct effect of sTREM-2 on M2 differentiation was relatively small and was therefore unlikely to contribute to biological function or postviral disease. However, in the process of conducting these experiments, we recognized that sTREM-2 treatment of BMDMs resulted in marked protection against the usual progression to cell death in this culture system, and this protective effect resulted in expanded numbers of BMDMs from both WT and Trem2−/− mice (Fig. 8 C). Similarly, sTREM-2 treatment caused a dose-dependent expansion of BMDMs from WT, Trem2−/−, and Dap12−/− mice (Fig. 8 D). Here again, the baseline numbers of BMDMs were increased in WT mice compared with Trem2−/−, and Dap12−/− mice under these culture conditions, but this difference was corrected in response to sTREM-2 treatment (Fig. 8 D). The protective effect was specific for active sTREM-2 because it was lost with heat inactivation (Fig. 8 E). Flow cytometry analysis with propidium iodide staining indicated that sTREM-2 administration resulted in attenuation of apoptosis due to CSF-1 withdrawal in BMDMs from WT mice, as well as Trem2−/− and Dap12−/− mice (Fig. 8 F). As expected, the level of apoptosis with CSF-1 withdrawal in BMDMs from WT mice was markedly increased in Trem2−/− and Dap12−/− mice, but this effect was also fully preventable and dose-dependent with sTREM-2 administration (Fig. 8 F). Treatment with sTREM-2 also reversed the increase in activate caspase-3 levels that develops with CSF-1 withdrawal, confirming an effect on apoptosis (Fig. 8 G). In contrast, sTREM-2 had no effect on cellular proliferation based on BrdU incorporation (Fig. 8 H), as expected for cell cultures deprived of CSF-1 stimulation. This result is also consistent with the absence of significant cell proliferation in the lung in vivo when disease becomes manifest (Tyner et al., 2006). Together, these findings support a role for sTREM-2 in protection against macrophage apoptosis under low CSF-1 conditions, and thereby explained the increased levels of apoptosis under these conditions in Trem2−/− and Dap12−/− macrophages that are unable to generate sTREM-2.
Given the difficulties in monitoring apoptosis in vivo (e.g., we detected no active caspase-3 tissue staining at 49 dpi; unpublished data) we sought additional markers of sTREM-2 action during postviral lung disease. In that regard, we were able to establish that sTREM-2 treatment resulted in a time-dependent and selective increase in molecular pathways connected to cell survival and proliferation, e.g., ERK1/2 and MAPK14 phosphorylation but not AKT phosphorylation or BCL-2 induction (Fig. 9 A). The effect of sTREM-2 on phospho-ERK1/2 and phospho-MAPK formation was more pronounced in BMDMs from Trem2−/− mice compared with WT mice (Fig. 9 A), consistent with the enhanced effect of sTREM-2 on viability of BMDMs from Trem-2−/− compared with WT mice based on fold-change (Fig. 8 D). In translating these findings to the postviral model in vivo, we found that phospho-ERK1/2 immunostaining was markedly increased and co-localized to lung macrophages at 49 dpi compared with 5 dpi or SeV-UV control conditions in WT mice (Fig. 9 B). Moreover, the increase in phospho-ERK1/2 immunostaining was not detected at 49 dpi in Trem2−/− mice (Fig. 9 B). These results were consistent with increases in the numbers of total immune cells and macrophages in the lung at 49 dpi in WT mice, as well as marked inhibition of this cellular response at 49 dpi in Trem2−/− mice (Fig. 9 C). Together, these findings provide for a crucial regulatory role of TREM-2 in a feed-forward pathway that expands the M2 macrophage population and thereby promotes IL-13–dependent chronic lung disease.
This report identifies TREM-2 as a key component of an innate immune pathway for the pathogenesis of chronic lung disease. We use a postviral mouse model of this type of disease to show that: (1) during acute illness, viral replication increases lung macrophage levels of intracellular and cell surface TREM-2, and this action mediates a CSF-1–dependent cell survival signal that prevents macrophage apoptosis; (2) during chronic postviral disease, IL-13 and DAP12 signals promote increased expression and cleavage of TREM-2 from the cell surface to form sTREM-2 at increased levels in the lung; and (3) sTREM-2 is not inactive as previously thought, but is instead capable of promoting macrophage survival via antiapoptotic signals such as ERK1/2 activation. Together, these events allow for TREM-2 to participate in a feed forward mechanism to amplify the accumulation of macrophages in the lung. In the context of our previous work, the increased number of macrophages would be available to interact with iNKT cells and drive further IL-13 production thereby resulting in IL-13–dependent differentiation of the macrophage population toward an M2 pattern of gene expression (including Arg1 and Chi3l3) and differentiation of an airway progenitor epithelial cell (APEC) niche to airway mucous cells (AMCs) as signatures of chronic lung disease (as diagrammed in Fig. 10). The scheme is consistent with our observation that TREM-2–deficient mice are markedly protected from the development of chronic lung disease as a long-term consequence of viral infection.
The role of TREM-2 in chronic disease described in this study provides for at least four major and unexpected insights. First, TREM-2 signaling was often portrayed as antiinflammatory and reparative, whereas increased TREM-1 signals were thought to be proinflammatory and destructive in models of intestinal and CNS disease (Schenk et al., 2007; Takahashi et al., 2007; Seno et al., 2009). This view is consistent with studies showing that engagement of TREM-1 promotes secretion of proinflammatory cytokines (including IL-6 and TNF) and increases the inflammatory responses downstream of TLR and NLR engagement (Bouchon et al., 2000, 2001a; Bleharski et al., 2003; Netea et al., 2006; Fortin et al., 2007; Schenk et al., 2007). Meanwhile, similar approaches showed that TREM-2 activation generally antagonized activation signals for inflammation (Daws et al., 2001; Colonna, 2003; Hamerman et al., 2006; Turnbull et al., 2006). In contrast, the present results provide for a paradigm in which TREM-2 activation promotes the development of chronic inflammatory disease. This model is consistent with reports for up-regulation of M2 gene expression (including TREM2) in macrophages isolated from cigarette smokers with and without COPD (Shaykhiev et al., 2009; Koth et al., 2010). However, these previous studies as well as others that assess association with disease cannot determine cause and effect actions of TREM signaling in the development of disease. Moreover, the present application of new tools to separately study both TREM-1 and TREM-2 deficiency in vivo and in vitro in the same model system was critical to dissect any differences in TREM family member function.
A second unexpected aspect of the present findings is the biological activity of sTREM-2. In fact, we initially approached the issue of TREM-2 activity as being a function of TREM-2 localization to the cell surface, where it could presumably bind to an as yet poorly defined ligand, and in association with DAP12, initiate a signaling pathway for cell survival (Otero et al., 2009). Indeed, we were able to secure evidence for this mechanism in the setting of acute infection, wherein viral replication increases TREM-2 on the surface of infected macrophages. However, this relatively transient effect on macrophage survival did not appear to explain the long-term effect of TREM-2 on postviral lung disease. We therefore pursued evidence that the greatest increases in TREM-2 lung levels were found long after clearance of infectious virus and were manifest as marked increases in sTREM-2 that had been cleaved from the cell surface. These circumstances led to the unexpected function of sTREM-2 as a potent antiapoptotic factor for macrophages. Our data suggests that sTREM-2 acts as an intracellular messenger to trigger the activation of kinases (e.g., ERK1/2) that promote cell survival. However, additional work will need to define the precise molecular pathway downstream of sTREM-2 actions, including those proposed for TREM-2–DAP12 signaling for cell survival (Otero et al., 2009, 2012). Meanwhile, however, our results may also help to explain the proposed benefit of sTREM-2 cleavage in neurodegenerative disease, where decreased sTREM-2 cleavage caused by mutation may be associated with impaired macrophage (i.e., microglial) function in the CNS (Kleinberger et al., 2014).
A third unanticipated result of our study was the potent role of DAP12 and IL-13 in regulating TREM-2 cleavage and sTREM-2 formation. Similar to signaling function of cell surface TREM-2, the formation of sTREM-2 and any consequent downstream action of sTREM-2 was still dependent on DAP12. This new connection between sTREM-2 and DAP12 could help explain how distinct TREM signals are transduced into different downstream effects on cell behavior. Thus, both TREM-1 and TREM-2 on the cell surface are proposed to signal via DAP12, and this adaptor has only a single immunoreceptor tyrosine-based activation motif (ITAM) for subsequent signal transduction. It has been proposed that the avidity of TREM (or other receptor) interaction with DAP12 might dictate a positive versus negative signal from this ITAM to the downstream kinase (Turnbull and Colonna, 2007). In the present experiments, we find that TREM-2 function is linked to DAP12 by a protein maturation process that appears distinct from TREM-1–DAP12 and TREM-2–DAP12 signaling. Similarly, IL-13 also contributed to the TREM-2 cleavage process in what appears to be a novel role for IL-13. In both cases, additional work is needed to define the molecular mechanism for TREM-2 cleavage. This type of insight could provide a useful target for antiinflammatory therapy by inhibiting cleavage and a consequent excess of sTREM-2 formation.
A fourth issue that is distinct for our results relates to the view of M2 macrophages and macrophage subsets in mediating inflammatory disease. For example, other disease models such as allergen challenge concentrate on blocking the macrophage response to IL-13 and conclude that there is no role for M2 macrophages in the disease process (Nieuwenhuizen et al., 2012). In contrast, our model focuses on the role of macrophages in IL-13 production as the fundamental step in producing disease, so it is this step that needs to be targeted for disruption. Notably, however, allergen-challenge models may depend on recruited monocytes but not resident alveolar macrophages for the development of lung disease (Zasłona et al., 2014). Similarly, we found differences in tissue monocyte and interstitial macrophage versus alveolar macrophage characteristics (e.g., level of Il13 mRNA induction and sensitivity to TREM-2 regulation) during the development of postviral disease. These results also support the paradigm of inflammatory monocyte recruitment and subsequent differentiation into disease-producing macrophages and provide a basis for assessing these events in situ with reporter labeling and intravital imaging in the future.
Our study focused on the role of TREM-2 in macrophage function based on evidence that macrophages are key to the development of postviral disease. However, we recognize that myeloid (conventional) DCs may contribute to postviral disease at least in the first few weeks after viral inoculation (Grayson et al., 2007) and may express TREM-2 at least in vitro (Bouchon et al., 2001b; Ito and Hamerman, 2012). We found TREM-2 expression on CD11b+CD11c– cells (i.e., monocytes and macrophages) and not on CD11b+CD11c+ cells (i.e., DCs) at 5 dpi with virus using FACS, but we cannot yet exclude expression on DCs at 49 dpi given the low levels of cellular expression of TREM-2 and the low numbers of DCs at this later time point. Thus, further studies with different technologies will be needed to determine any role for DC-expressed TREM-2 after viral infection.
In sum, we have defined a novel role for TREM-2 in the development of chronic lung disease using a postviral mouse model of this process. Our previous work led to a new consideration of the role of innate immune activation of macrophages in this type of disease. However, there was little subsequent progress toward further understanding and thereby specifically blocking the macrophage component of postviral disease in experimental models or humans. Herein, we describe an unanticipated mechanism whereby TREM-2 allows for macrophages to accumulate in the lung and thereby amplify disease-promoting IL-13 production. In this setting, TREM-2 appears to affect the chronic disease process without influencing acute viral illness, thereby creating an opportunity to inhibit excessive TREM-2 action without compromising host defense. The findings thereby provide an opportunity to down-regulate the type 2 immune response and prevent the development of this type of immune-mediated disease by interrupting sTREM-2 formation or action. Initial studies already suggest an increase in TREM-2 expression in concert with excess levels of airway mucus and M2 macrophages in human lung disease (Byers and Holtzman, 2011; Byers et al. 2013. International Meeting of the American Thoracic Society. Abstr. A141.). The present findings justify further validation of these observations in humans with chronic lung disease and additional work to correct any abnormalities in TREM-2 action in this type of disease and other diseases that implicate a role for TREM-2 in regulating macrophage survival.
MATERIALS AND METHODS
WT C57BL/6J mice were obtained from The Jackson Laboratory. The op/opT mice were generated as described previously (Abboud et al., 2002) and were provided by N. Ghosh-Choudhury (University of Texas, San Antonio). These mice were crossed to WT mice to generate heterozygous wt/opT mice and littermate wt/wt control mice. Dap12−/− mice were provided by T. Takai (Tohoku University, Sendai, Japan). Trem2−/− mice were generated as described previously (Turnbull et al., 2006). Trem1/3−/− mice were generated as described previously and were backcrossed to C57BL/6 mice until >99% of the loci were derived from this strain. Tlr3−/− mice were generated as described previously (Alexopoulou et al., 2001). Pkr−/− mice were generated as described previously (Yang et al., 1995) and rederived from frozen embryos obtained from M. White and H. Virgin IV (Washington University, St. Louis, MO). All mice were maintained on a C57BL/6 genetic background and were used in protocols approved by Animal Studies Committee of Washington University School of Medicine in accordance with National Institutes of Health guidelines.
Viral inoculation and monitoring.
Mice were inoculated with SeV (105 pfu), an equivalent amount of UV-inactivated SeV (SeV-UV), or PBS as described previously (Kim et al., 2008). SeV-UV and PBS inoculations gave identical results, and therefore are treated interchangeably as controls. Viral titers for stock solutions and lung infections were monitored by plaque-forming and real-time PCR assays as described previously (Kim et al., 2008).
Analysis of mRNA.
For gene expression microarrays, RNA was isolated and analyzed using Illumina Mouse-WG6 v2 BeadChips (Illumina) as described previously (Byers et al., 2013). Visualization and plotting was performed using DecisionSite for Functional Genomics (TIBCO Spotfire). Microarray data were deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) under series accession nos. GSE49603 and GSE61437. For real-time qPCR assays, RNA was purified from lung homogenate using TRIzol (Invitrogen) and was converted to cDNA using the High-Capacity cDNA Archive kit (Applied Biosystems). Target mRNA was quantified by real-time PCR assay using specific fluorogenic probes and primer sets and the Fast Universal PCR Master Mix system (Applied Biosystems). Primer-probe sets for mouse Ccl2 (Mcp-1; Mm00441242_m1), Cx3cr1l (Mm00436454_m1), Trem1 (Mm00451738_m1), Trem2 (Mm00451744_m1), Chi3l3 (Mm00657889_mH), Arg1 (Mm00475988_m1), and Dap12 (Mm00449152_m1) were obtained from Applied Biosystems. The forward and reverse primers and probes for mouse Il13 were 5′-GGTGCCAAGATCTGTGTCTC-3′, 5′-CCACACTCCATACCATGCTG-3′, and 5′-AAGACCAGACTCCCCTGTGCAAC-3′; for Muc5ac were 5′-TACCACTCCCTGCTTCTGCAGCGTGTCA-3′, 5′-ATAGTAACAGTGGCCATCAAGGTCTGTCT-3′, and 5′-TATACCCCTTGGGATCCATCATCTACA-3′; for Csf1 were 5′-CTTCATGCCAGATTGCCTTTG-3′, 5′-CGCATGGTCTCATCTATTATGTCT-3′, 5′-CAGCTGGATGATCCTGTTTGCTACCT-3′; and for mouse Gapdh were 5′-AAGGTGAAGGTCGGAGTCA-3′, 5′-CATGTAAACCATGTAGTTGAGGT-3′, and 5′-CGGATTTGGTCGTATTGGGCGC-3′. All probes were designed to span introns and did not react with genomic DNA. Samples were assayed on the 7500 Fast Real-Time PCR System and analyzed using the 7500 Fast System Software (Applied Biosystems). Levels of specific gene expression were normalized to Gapdh mRNA levels.
For TREM-2 expression, BMDMs were seeded in 6-well tissue culture plates and were infected with SeV at MOI 1 and 10 for 48 h. For caspase-3 expression, BMDMs were seeded in 6-well tissue culture plates and were cultured in the presence or absence of L cell-conditioned medium for 16 h. Cells were lysed and boiled in reducing SDS-PAGE sample buffer containing 100 mM Tris, pH 6.8, 4% SDS, 5% β-mercaptoethanol, 20% glycerol and 0.2% bromophenol blue. Lysates was subjected to 4–15% gradient gel electrophoresis (Bio-Rad Laboratories) and transferred to PVDF membrane (Millipore) in 10% methanol, 25 mM Tris and 192 mM glycine, pH 8.3. Anti-Caspase-3 Ab was from Cell Signaling Technology. Rat anti–mouse TREM-2 mAb (clone 178.9.4) was generated as described previously (Turnbull et al., 2006). Antibody binding was detected using the enhanced chemiluminescence (ECL) system (GE Healthcare). Molecular weight standards were run on each blot to validate that each of the indicated bands corresponded to the appropriate molecular weight.
Recombinant protein expression and purification.
The mouse TREM-2 Ig domain (amino acid residues 19–136) tagged with 6X His (sTREM-2-his) was expressed, secreted, and purified from mammalian cell culture as previously described (Kober et al., 2014). Precautions were taken to prevent introduction of endotoxin either by using either new, sterile plasticware or by soaking glassware overnight in 0.5 M NaOH. The final purified protein was exchanged in PBS using a microconcentrator (Sartorius). For heat inactivation, sTREM-2-his was boiled at 100°C for 1 h.
Tissues were fixed with 10% formalin, embedded in paraffin, cut into 5-µm sections, and adhered to charged slides. Sections were deparaffinized in Citrosolv (Fisherbrand), hydrated, and in the case of immunofluorescence microscopy, treated with heat-activated antigen unmasking solution (Vector Laboratories). Immunostaining of mouse tissue was performed using anti-MUC5AC mAb (Thermo Fisher Scientific; clone 45M1), anti–Mac-3 mAb (BD; clone M3/84), anti–Caspase-3 mAb (clone 5A1E; Cell Signaling Technology), anti–mouse TREM-2 (R&D Systems), anti–mouse F4/80 mAb (clone CI:A3-1; Abcam), anti–mouse phospho-ERK1/2 (clone D13.14.4E, Cell Signaling Technology), anti–mouse IL-13 mAb (R&D Systems), and anti-SeV antibody (SPAFAS, Inc.). Antibody binding was visualized using Alexa Fluor 488 or 594 fluorochromes (Invitrogen). Slides were counterstained with DAPI-containing mounting media (Vector Laboratories).
For sTREM-2 immunostaining, BMDMs were seeded into 2-well chamber slide (Nalge Nunc International) at 2 × 105 cell per well. For surface sTREM-2 immunostaining, BMDMs were incubated with sTREM2-his (10 µg/ml) for 30 min on ice and then washed four times with chilled PBS and incubated with mouse anti-6X His tag antibody (Abcam; clone His.H8) followed by Alexa Fluor 488–conjugated anti–mouse IgG (Invitrogen) and fixation with 4% PFA. For intracellular sTREM-2 immunostaining, cells were incubated with sTREM-2-his for 30 min at 37°C, washed with PBS, fixed with 4% PFA, and then permeabilized with 0.2% Saponin, 2% BSA in PBS for 30 min before incubation with anti-6X His tag antibody and Alexa Fluor 488–conjugated anti–mouse IgG. All microscopy was performed using an Olympus BX51 microscope and a Retiga-2000R camera system (QImaging).
Flow cytometry and FACS.
Single-cell lung suspensions were made from minced lung tissue subjected to collagenase (Liberase Blendzyme III; Roche), hyaluronidase (Sigma-Aldrich), and DNase I (grade II; Roche) digestion for 45 min at 37°C and then treated with ACK buffer to remove red blood cells. Cell percentages were unchanged if lungs were perfused before digestion (suggesting that intravascular cells did not contribute significantly to the analysis), so unperfused lungs were used for all experiments. Staining of surface markers was performed using FcR blockade and fluorochrome-conjugated mAb at 4°C. We used the following antibodies: rat anti-mCD11b (clone M1/70; eBioscience), anti-Ly6C (clone al-21; BD), anti-Ly6G (clone 1a8; BD), anti-mouse F4/80 (clone BM8; eBioscience), anti-mouse TREM-1 (clone 174031; R&D Systems), anti-m/hTREM-2 (clone 237920; R&D Systems), anti–mouse CD11c (clone HL3; BD), anti–mouse CD45.2 (clone 104; eBioscience), and anti–mouse Siglec-F (clone E50-2440; BD). We used flow cytometry and FACS to identify and purify specific cell populations in mouse lungs using the following gates: tissue monocytes as SSClowLy6G–F4/80+CD11b+, interstitial macrophages as SSChighCD11c–Ly6G–Siglec-F–F4/80+CD11b+, alveolar macrophages as SSChighCD11c+Ly6G–Siglec-F+F4/80+CD11b–, and neutrophil gate as SSChighLy6G+ (Fig. S1). Analysis was performed using Moflo (DAKO-Cytomation) and FlowJo software (Tree Star). Cells were counted using flow cytometry and fluorescent counting beads (CountBright; Life Technologies).
Bone marrow cells were flushed from femurs and differentiated in 150 × 15 mm Petri dishes in DMEM medium containing 0.29 g/liter l-glutamine, 1 mM sodium pyruvate, 10% FBS, and 30% L929 fibroblast-conditioned medium as a source of CSF-1. At day 6 or indicated times, BMDMs were resuspended by incubation in 1 mM EDTA in PBS for 10 min on ice, washed twice with PBS and plated into tissue culture plates. In some experiments, BMDMs were incubated with poly(IC) (Sigma-Aldrich) or mouse IL-13 (R&D Systems). For cell viability assays, BMDMs were seeded into 96-well tissue culture plates at 5 × 104 cells per well and were assessed for viability using the Cell Proliferation kit (XTT) from Roche. For cell proliferation assays, BMDMs were seeded into 6-well tissue culture plates, labeled for 18 h with BrdU (10 µM), and then assessed using flow cytometry with the FITC BrdU Flow kit (BD).
For propidium iodide (PI) assay, BMDMs were cultured in L cell–conditioned medium for 5 d and deprived of CSF-1 for 1 d to induce apoptosis (Otero et al., 2009). Cells were detached using EDTA (1 mM) in PBS, washed, fixed with 70% ethanol, and incubated with PI (50 µg/ml) and RNase (40 µg/ml) for 1 h. Samples were analyzed for DNA content using a FACScan (BD). For Annexin V assay, BMDMs were differentiated in L cell–conditioned medium for 5 d, detached with EDTA, seeded onto the poly-D-Lysine-coated coverslips (BD) placed in a 24-well plate, and cultured in medium without CSF-1 for 1 d. PI and Annexin V immunostaining was performed according to the manufacture’s protocol (Invitrogen). Annexin V–positive, PI-negative staining cells were quantified using microscopy, and values were expressed as the percentage of apoptotic cells (Annexin V positive, PI negative) relative to the total number of cells in 6 randomly selected fields of the two wells.
For analysis of monocyte influx into the lung, BMMs were purified from CD45.2 WT and Trem2−/− mice using negative selection with the EasySep Mouse Monocyte Enrichment kit (StemCell Technologies). Purified monocytes were then introduced into CD45.1 WT mice (3 × 106 cells per mouse via tail vein injection) at 0 dpi with SeV, and the number of CD45.2+ cells in the lung was determined 3 or 5 d later using FACS and fluorescent counting beads.
For lung tissue samples, lung tissue was minced and then homogenized in PBS containing protease inhibitor cocktail (cOmplete; Roche) using a rotor-type homogenizer (Tissue-Tearor; Biospec Products). For cell supernatant and lysate samples, BMDM (2 × 106 cells in 2 ml of complete DMEM containing 15% L cell–conditioned medium) were seeded into 12-well tissue plates, treated with or without IL-13 (20 ng/ml) or SeV (MOI 1 and 10). Cell supernatants were collected, and cells were washed twice with chilled PBS and then lysed in radio-immunoprecipitation assay (RIPA) buffer (Sigma-Aldrich) containing protease inhibitors. Levels of CSF-1 were determined using a Duoset ELISA kit (R&D Systems) and levels of cellular TREM-2 and sTREM-2 with a TREM-2 ELISA kit (Antibodies-online.com).
For experiments in mice and cell culture, all experiments were repeated a minimum of three times, and representative values are presented. PCR data were presented as mean ± SEM and was analyzed using an unpaired Student’s t test. If variances were unequal, Welch’s correction was applied. For all analyses, P < 0.05 were considered significant. For SeV-GFP versus TREM-2 expression, nonlinear curve fit was performed using a simple hyperbolic function where Y = Bmax × X/(Kd + X).
Online supplemental material.
Fig. S1 shows the flow cytometry scheme for analysis of lung macrophages. Fig. S2 shows a scheme for FACS analysis of cell surface and intracellular TREM-2 immunostaining. Table S1 shows microarray gene expression values for Fig. 2 A, including values for all TREM family members.
We thank the Alvin J. Siteman Cancer Center at Washington University School of Medicine and Barnes-Jewish Hospital for the use of the High Speed Cell Sorter Core. We also appreciate valuable scientific and technical input from Loralyn Benoit, Darlene Stewart, and Theresa Tolley.
This work was supported by grants from the National Institutes of Health (OD and NIAID grants AADCRC U19-AI07048 and NHLBI grants T32-HL-007317, P50-HL-084922, P01-HL-029594, R01-HL-073159, and R01-HL-121791) and the Martin Schaeffer Fund.
The authors declare no competing financial interests.
airway progenitor epithelial cell
chronic obstructive pulmonary disease
DNAX activation protein of 12 kD
epithelial cell adhesion molecule
extracellular signal-regulated kinase
- iNKT cell
invariant NK T cell
protein kinase R
soluble TREM 2
triggering receptor expressed on myeloid cells