Critical role of phospholipase A₂ group IID in age-related susceptibility to severe acute respiratory syndrome–CoV infection

Because chronic inflammation in the lung increases with aging (MacNee, 2011) and AM and rDC (CD11c⁺ cells) are critical for maintaining lung homeostasis and for prostaglandin production, we initially examined CD11c⁺ cells harvested from young and middle-aged mice for global changes in gene expression. Expression levels of 186 genes changed during aging, with many involved in antigen presentation and inflammation (Fig. 1 B and Table S1). Many of these were predicted to contribute to a chronic inflammatory response. However, our previous study showed that levels of a prostaglandin with antiinflammatory properties, PGD₂, were elevated in aged mice (confirmed in Fig. 1 A) and contributed to poor outcomes after SARS-CoV infection (Zhao et al., 2011). Therefore, we next focused on genes involved in the AA pathway, including several phospholipases, cyclooxygenases, and synthases, which could modulate PGD₂ levels in the lungs. Of all the genes in the AA cascade, only levels of secreted phospholipase A₂ group IID (PLA₂G2D) were significantly higher in CD11c⁺ cells in the lungs of middle-aged compared with young mice (using twofold differences as a cutoff; Fig. 1, C and D). We confirmed these results by quantitative real-time PCR (qPCR) analyses of CD11c⁺ cell RNA by separating CD11c⁺ and CD11c⁻ cells from the lungs of naive young and middle-aged mice using magnetic beads and quantifying the expression of PLA₂G2D mRNA (Fig. 2 A). We observed 25–30 fold higher expression of PLA₂G2D mRNA in lung CD11c⁺ compared with CD11c⁻ cells from middle-aged mice. Comparison of CD11c⁺ cells from young and middle-aged mice revealed that PLA₂G2D was expressed at levels >50-fold higher in middle-aged mice (Fig. 2 A). Flow cytometric analysis also showed higher expression of PLA₂G2D in CD11c⁺ compared with CD11c⁻ cells on a per cell basis in both young and middle-aged mice, as measured by mean fluorescent intensity (MFI; Fig. 2 B). In addition, middle-aged mice harbored a higher frequency of CD11c⁺ cells expressing PLA₂G2D compared with young mice (Fig. 2 C).

Together, these results indicate that PLA₂G2D levels are highly increased in the lungs of middle-aged mice and that CD11c⁺ cells are the major source for its production. However, regulatory T cells (T reg cells; CD4⁺FoxP3⁺) have been implicated as a source for PLA₂G2D in previous studies (von Allmen et al., 2009) and lungs of uninfected middle-aged mice harbor higher frequency and number of T reg cells compared with young mice (Fig. 2 D). To address whether T reg cells are an additional source for PLA₂G2D, we depleted CD4 T cells, which includes T reg cells, from middle-aged mice and analyzed lungs for expression of PLA₂G2D. We observed no changes in the mRNA expression of PLA₂G2D after CD4 T cell depletion (Fig. 2 E), suggesting that CD4 T cells, including T reg cells, are not a major source of PLA₂G2D in the lungs.

Because CD11c⁺ cells comprise only a minority of the total cells in the lungs, expression of PLA₂G2D by these cells could make a relatively small contribution to its total production. However, this was not the case because levels of PLA₂G2D were significantly higher in unfractionated lung mRNA from middle-aged compared with young mice. Notably, PLA₂G2D levels were even higher in samples obtained from aged (22 mo) mice (Fig. 3 A). Assessment of other genes in the AA cascade generally confirmed the microarray results shown in Fig. 1 C and showed that there were no differences in the levels of expression of the cyclooxygenase enzymes COX₁ and COX₂ or other PLA₂s (PLA₂G2E, PLA₂G3, PLA₂G4, PLA₂G5, and PLA₂G10) as mice aged (Fig. 3, B and C). One exception is that there were significantly higher mRNA levels of hematopoietic PGD₂ synthase (HPGDS), which is important for PGD₂ production and could contribute to elevated PGD₂ levels in the lungs of middle-aged compared with young mice (Fig. 3 C; Murata et al., 2013).

Because robust expression of PLA₂G2D in lymphoid tissue CD11c⁺ cells was previously reported (Miki et al., 2013), we determined whether its levels increased with age in the MLN. As expected, we observed higher expression of PLA₂G2D in CD11c⁺ compared with CD11c⁻ cells in the MLN (Fig. 2 F). However, unlike in the lungs, there were no age-dependent differences in PLA₂G2D mRNA expression in MLN CD11c⁺ cells between young and middle-aged naive mice. Collectively, these results indicate that PLA₂G2D is nearly exclusively expressed in CD11c⁺ cells in both lymphoid tissue and lungs and that age-related differences in its expression are confined to the lungs.

PLA₂G2D has been previously described as a proresolving phospholipase acting upstream in the AA-cascade (Miki et al., 2013). Hence, we reasoned that higher expression levels of PLA₂G2D in the lungs of middle-aged mice would result in perturbation in lipid metabolism and that lipid mediators with proresolving functions, other than PGD₂, would also be produced at higher levels. To investigate this possibility, we performed electrospray ionization mass spectrometry (ESI-MS) to determine the levels of free polyunsaturated fatty acids (PUFAs), as well as their oxygenated metabolites, including PGD₂, in the lungs of naive young and middle-aged mice (Fig. 4, B–D). Initial analyses revealed altered profiles of lipid mediators with both pro- and antiinflammatory function in the lungs of middle-aged relative to young mice. Levels of free PUFA such as AA (ω6) and docosahexaenoic acid (DHA; ω3) were not significantly different between the groups, whereas the level of eicosapentaenoic acid (EPA) was higher in middle-aged mice relative to young mice (Fig. 4 B). AA metabolites with antiinflammatory (PGE₂ and PGD₂) and proinflammatory properties (PGF2α and Thromboxane [TX] B₂ [a TXA₂ metabolite]) were produced at higher levels in the lungs of middle-aged mice compared with young mice, but levels of 6-keto-PGF1α (a PGI₂ metabolite) did not differ between the groups (Fig. 4, C and D).

PLA₂G2D is one of many PLA₂ expressed in the lungs, but is the only one that was up-regulated as mice aged, as shown above. Therefore, we next assessed whether PLA₂G2D mediated the perturbation in pulmonary lipid profile in middle-aged mice using Pla2g2d^−/− and Pla2g2d^+/+ mice (Miki et al., 2013). Lipid profiling of the lungs of middle-aged Pla2g2d^−/− and Pla2g2d^+/+ mice revealed an important role for PLA₂G2D in the levels of most lipids examined (Fig. 4, H–J). Of note, these changes were not observed when young Pla2g2d^−/− and Pla2g2d^+/+ mice were compared (Fig. 4, E–G). AA was produced at lower levels in Pla2g2d^−/− relative to Pla2g2d^+/+ middle-aged mice (Fig. 4 H), suggesting that phospholipase activity required to release AA was at least partly dependent on PLA₂G2D. Levels of 6ketoPGF1α, TXB_2, PGF2α, PGD₂, and PGE₂ were also higher in the lungs of Pla2g2d^+/+ compared with Pla2g2d^−/− mice (Fig. 4, I and J), suggesting that differences in AA levels resulted in differences in the levels of its downstream metabolites, some of which have antiinflammatory function (PGD₂ and PGE₂; Vancheri et al., 2004; Kalinski, 2012; Murata et al., 2013). The results also identify PLA₂G2D-dependent pools of two other PUFA, DHA, and EPA in the lungs. DHA and EPA are further metabolized largely into proresolving mediators, such as resolvin D1, protectin D1, and resolvin E1 among others (Arita et al., 2005; Rogerio et al., 2012).

Having established that both proinflammatory and antiinflammatory/proresolving lipid mediators are up-regulated in the uninfected lungs of middle-aged mice, we next determined whether levels of these mediators were further perturbed in the context of infection. We initially analyzed mice infected with SARS-CoV because middle-aged but not young C57BL/6 mice succumb to infection with this virus, mimicking the human disease (Fig. 5 A; Frieman et al., 2012; Zhao et al., 2012). Because mortality in middle-aged mice begins at 4 d after SARS-CoV infection, we investigated whether levels of lipid mediators, including PGD₂, changed at this time. Expression levels of PLA₂G2D increased in both young and middle-aged mice after infection (Fig. 5 B). Further, pulmonary lipid profiling on day 4 postinfection (p.i.) revealed significant increases in lung levels of free PUFAs such as AA, EPA, and DHA in middle-aged mice but less so in young mice (Fig. 5 C). Levels of lipid mediators with proinflammatory activity, 6ketoPGF1α, TXB₂, and PGF2α, increased in the lungs of young but not middle-aged mice after SARS-CoV infection (Fig. 5 D). In contrast, levels of PGE₂ and PGD₂, both of which have antiinflammatory activity in the lungs, increased in both young and middle-aged mice after infection (Fig. 5 E).

To determine whether the inducible pool of these lipids is dependent on PLA₂G2D, we determined the lung lipid profile in SARS-CoV–infected middle-aged Pla2g2d^−/− and Pla2g2d^+/+ mice. Whereas levels of free PUFAs, AA, DHA, and EPA increased in the lungs of both groups of mice during infection, Pla2g2d^+/+ mice harbored higher levels of all three fatty acids at all times examined (Fig. 5 F), indicating that the inducible pool of each of these fatty acids was at least partly mediated by PLA₂G2D. We did not observe an infection-dependent increase in the levels of 6-keto-PGF1α, TXB₂, or PGF2α in either group of mice, although levels remained higher in Pla2g2d^+/+ mice relative to Pla2g2d^−/− mice on day 4 p.i. (Figs. 4 I and 5 G). PGE₂ and PGD₂ levels increased only in Pla2g2d^+/+ mice after infection (Fig. 5 H). These results indicate multifactorial regulation of the inducible AA cascade with basal levels of all lipid mediators were maintained at least partly by PLA₂G2D, whereas increases in PGE₂ and PGD₂ levels observed after SARS-CoV infection in the lungs were largely dependent on PLA₂G2D. These data also agree with published studies showing that induction of PLA₂G2D generated pools of proresolving but not proinflammatory mediators in the lung and skin in the context of allergic disease (Nagao et al., 2003; Kunikata et al., 2005; Liu et al., 2012; Miki et al., 2013).

The preferential induction of antiinflammatory lipid mediators in Pla2g2d^−/− mice could contribute to less efficient activation of the innate and adaptive immune responses. To address this possibility, we initially analyzed the activation status and function of lung CD11c⁺ cells harvested from middle-aged and young mice at day 1 after SARS-CoV infection. Cells from middle-aged mice expressed significantly higher levels of maturation markers (CD40, CD80, and MHCII) and expressed higher levels of TNF in the absence of ex vivo stimulation (basal levels) compared with young mice (Fig. 6, A and B [RP10]). However upon TLR stimulation with LPS or poly I:C, similar frequencies of CD11c⁺ cells from both young and middle-aged mice produced TNF, although the total number of TNF-producing cells was higher in middle-aged mice (Fig. 6 B, middle and bottom).

To investigate the role of PLA₂G2D in this enhanced CD11c⁺ cell activation in middle-aged mice, we measured CD40, CD80, and MHCII levels on lung CD11c⁺ cells harvested from infected middle-aged Pla2g2d^+/+ and Pla2g2d^−/− mice and also quantified TNF expression after TLR stimulation. PLA₂G2D was required for MHCII but not CD40 or CD80 expression on lung CD11c⁺ cells in middle-aged mice, and for increased frequency but not number of cells expressing TNF in the absence of any direct ex vivo stimulation with TLR ligands. CD11c⁺ cells from both Pla2g2d^+/+ and Pla2g2d^−/− mouse lungs were able to express equivalent amounts of TNF after direct ex vivo stimulation (Fig. 6, C and D). Thus, these results indicate that CD11c⁺ cells from infected middle-aged Pla2g2d^+/+ mice were more highly activated in vivo compared with young C57BL/6 mice, but that PLA₂G2D has a relatively minor role in age-dependent differences in CD11c⁺ cell maturation and function.

These results show that the major effect of PLA₂G2D is to increase levels of antiinflammatory lipid mediators, but changes in expression may also have indirect or direct effects on levels of cytokines and other inflammatory mediators. However, we observed no differences in expression of several proinflammatory molecules or IL-10, a cytokine with antiinflammatory properties, in the lungs of Pla2g2d^+/+ and Pla2g2d^−/− mice at days 1 or 4 p.i. (Fig. 6 E). Thus, PLA₂G2D has only small effects on general innate immune function in middle-aged mice and its absence largely affects only the expression of lipid mediators.

PGD₂ inhibits the migration of airway DCs from lungs to the MLN in middle-aged naive and SARS-CoV–infected mice (Hammad et al., 2003; Zhao et al., 2011) and while levels of PGD₂ decreased in Pla2g2d^−/− mice, it was not known a priori whether this decrease was sufficient to correct the defect in rDC migration. To determine whether this was the case, we first instilled OVA-FITC intranasally (i.n.) into uninfected middle-aged Pla2g2d^+/+ and Pla2g2d^−/− mice and analyzing MLN 18 h later for OVA-FITC⁺ rDCs, indicative of migration from the lungs. We observed that a higher frequency and number of rDCs migrated to the MLN in the absence of PLA₂G2D (Fig. 6 F). We next investigated whether rDC migration was augmented in mice infected with SARS-CoV by instilling CFSE i.n. into middle-aged Pla2g2d^+/+ and Pla2g2d^−/− mice 6 h before infection. 18 h p.i. MLN were harvested and analyzed for the frequency of rDCs that migrated from lungs (CD11c⁺MHCII⁺CFSE⁺ population). Pla2g2d^−/− mice harbored a higher frequency and number of CFSE⁺ rDCs in MLN compared with Pla2g2d ^+/+ mice (Fig. 6 G). Migration of DCs from the site of infection to the MLN is mediated by chemokine receptor CCR7 binding to CCL19/21 (Hamilton-Easton and Eichelberger, 1995; Fainaru et al., 2005) and DCs differentiated in the presence of PGD₂ express lower levels of CCR7 (Gosset et al., 2005; Zhao et al., 2011). Consistent with higher expression of PLA₂G2D and PGD₂ in the lungs, Pla2g2d^−/− rDCs expressed higher CCR7 surface levels than Pla2g2d^+/+ rDCs, contributing to enhanced rDC migration to the draining lymph nodes in uninfected and SARS-CoV–infected mice (Fig. 6 H).

Having demonstrated superior migration of rDCs to the MLN in Pla2g2d^−/− mice, we next assessed whether this translated into more robust virus-specific CD4 and CD8 T cell responses in infected lungs (Fig. 7). We addressed this possibility by examining the CD4 and CD8 T cell response in SARS-CoV–infected middle-aged Pla2g2d^−/− and Pla2g2d^+/+ mice. A higher frequency and number of virus-specific (epitope S525-specific) CD8 T cell and (epitope N110-specific) CD4 T cells (Fig. 7 A) were present in the lungs of Pla2g2d^−/− mice compared with Pla2g2d^+/+ mice. In addition, we also observed that CD8 T cells from Pla2g2d^−/− compared with Pla2g2d^+/+ mice produced more IFN-γ on a per cell basis (Fig. 7 B), although there were no differences in the fraction of CD8 T cells that produced both IFN-γ and TNF (Fig. 7 C). Further, the absence of PLA₂G2D augmented CD8 T cell cytolytic activity, measured in an in vivo killing assay (Fig. 7 D). To investigate if there was also a T cell–intrinsic defect in Pla2g2d^+/+ mice, we performed a set of adoptive transfer experiments, transferring splenocytes from uninfected Pla2g2d^+/+ or Pla2g2d^−/− mice to CD45 disparate Pla2g2d^−/− or Pla2g2d^+/+ mice, respectively, before SARS-CoV infection. In each case, the frequency of transferred CD8 T cells that expressed IFN-γ after peptide stimulation was similar to that of endogenous cells, indicating that the defects were T cell extrinsic (Fig. 7 E).

We next assessed whether these enhanced SARS-CoV–specific T cell responses decreased mortality after SARS-CoV infection. To this end, we infected middle-aged Pla2g2d^−/− and Pla2g2d^+/+ mice with SARS-CoV and monitored survival (Fig. 8 A). More than 80% of Pla2g2d^−/− mice survived the infection, whereas all Pla2g2d^+/+ mice died. Increased survival reflected enhanced kinetics of viral clearance in Pla2g2d^−/− mice (Fig. 8 B), mediated by the more robust T cell response observed in these mice. In addition, decreased pathological changes were observed in the lungs of Pla2g2d^−/− compared with Pla2g2d^+/+ mice. We detected no differences in the lungs of uninfected mice (Fig. 8 C, naive). Pla2g2d^+/+ mice showed much greater perivascular cuffing and cellular infiltration compared with Pla2g2d^−/− mice by day 4 p.i. By day 6 p.i., multifocal inflammatory lesions around blood vessels and alveolar edema were significantly greater in Pla2g2d^+/+ mice, with few changes observed in Pla2g2d^−/− mice.

To determine whether elevated PLA₂G2D levels in middle-aged mice affected the host immune response to other respiratory pathogens, we infected Pla2g2d^−/− and Pla2g2d^+/+ mice with IAV (PR8 strain) and monitored rDC migration, T cell responses and survival. Both the frequency and number of rDCs were increased in the MLN of Pla2g2d^−/− compared with Pla2g2d^+/+ mice when assessed at 18 h p.i. (Fig. 9 A). Consistent with increased rDC migration, we found that CD8 T cell responses were significantly higher in the lungs of Pla2g2d^−/− mice, when measured at 10 d p.i. (Fig. 9 B). Enhancement of the CD8 T cell response did not increase survival in IAV-infected mice (Fig. 9 C). This may occur because the CD8 T cell response is preserved to a greater extent in IAV compared with SARS-CoV–infected middle-aged Pla2g2d^+/+ mice, so that its augmentation has less of an effect on survival (Zhao et al., 2011).

Aging is accompanied by oxidative stress, which triggers lipid peroxidation (Cannizzo et al., 2012). A consequence of lipid peroxidation is increased expression of various PLA₂s (Brown et al., 2003), raising the possibility that increased oxidative stress in the middle-aged lung is a key factor in PLA₂G2D up-regulation. To evaluate whether oxidative stress is increased in the lungs of middle-aged mice, we first measured the levels of malondialdehyde (MDA), a product of lipid peroxidation, and an indicator of oxidative stress (Gutteridge and Halliwell, 1990). Analyses of middle-aged compared with young mice revealed higher lung concentrations of MDA (Imai et al., 2008), indicative of greater pulmonary oxidative stress (Fig. 10 A).

Oxidative stress can be countered by treatment with N-acetyl cysteine (NAC) and, as shown in Fig. 10 B, treatment of middle-aged mice with NAC for 10 d reduced MDA levels. Notably, NAC treatment also diminished levels of PLA₂G2D mRNA in lung CD11c⁺ cells, suggesting that augmented PLA₂G2D expression in middle-aged mice was related to oxidative stress (Fig. 10 B). Further, to induce oxidative stress in young mice, we treated them for seven days with a low dose of LPS (0.1 µg/mouse; Kratzer et al., 2012). This dose has been shown to result in low level induction of proinflammatory mediators in murine macrophages (Maitra et al., 2012). Chronic LPS treatment resulted in higher concentrations of MDA in the lungs of mice compared with controls (Fig. 10 C). More importantly, inducing oxidative stress with LPS also increased mRNA levels of PLA₂G2D (Fig. 10 C), but not of other PLA2s (PLA₂G4, PLA₂G2E, PLA₂G10; Fig. 10 D) in lung CD11c⁺ cells.

To extend the clinical relevance of these results, we next determined whether chronic oxidative stress resulted in PLA₂G2D up-regulation in human monocyte-derived macrophages (MDMs). For this purpose, PBMCs were treated with macrophage colony stimulating factor (M-CSF) for 4 d, and the cell-adherent population was incubated with M-CSF in the presence or absence of low levels of LPS (0.1 ng/ml) for an additional 1–10 d. PLA₂G2D expression was then assessed. In agreement with previous work (Miki et al., 2013), LPS exposure for 1 d did not result in increased PLA₂G2D levels (unpublished data). However, we observed increased PLA₂G2D expression after chronic LPS treatment (10 d), and this was accompanied by increases in MDA levels, indicative of higher oxidative stress (Fig. 10 E). Thus, chronic oxidative stress, a well-described aspect of aging in the human lung, results in increased PLA₂G2D levels, which could contribute to age-dependent defects in the immune response to SARS-CoV and other respiratory pathogens.

Although many infections are more severe in aged populations, SARS, and possibly MERS show remarkable age-dependent differences in outcomes such that young humans (less than 24 yr) all survived the infection, whereas 15% and 52% of middle-aged and aged individuals, respectively, succumbed (Peiris et al., 2004; Assiri et al., 2013; Saad et al., 2014; Alsolamy, 2015). SARS-CoV–infected mice also show similar steep age-dependent changes in mortality, which we show is due, in large part, to increases in levels of a single secretory PLA₂, PLA₂G2D, as mice age. A hallmark of aging is lipid peroxidation and subsequent up-regulation of PLA₂s (Brown et al., 2003; Sun et al., 2004; Hermann et al., 2014) and our results suggest that, in the lungs, levels of PLA₂G2D singularly increase as mice age. PLA₂G2D is considered an antiinflammatory/proresolving enzyme, suggesting its up-regulation is actually a measure used in the lung to counter the low-grade chronic inflammation that occurs with aging (Franceschi et al., 2000; Miki et al., 2013). In support of this notion, treatment with low doses of LPS, which induces oxidative stress and low-grade inflammation, resulted in PLA₂G2D up-regulation both in CD11c⁺ cells in the mouse lung (Fig. 10 C) and in human monocyte-derived macrophages (Fig. 10 E). Further, interventions that diminish reactive oxygen species decreased PLA₂G2D expression (Fig. 10 B). A recent study demonstrated the proresolving role of PLA₂G2D in the setting of contact dermatitis in mice; in its absence, resolution was prolonged (Miki et al., 2013). PLA₂G2D would be expected to have similar effects in ameliorating disease in the context of autoimmune disease.

Increased PLA₂G2D expression by lung CD11c⁺ cells results in up-regulation of several eicosanoids, with especially prominent effects on PGD₂ production (Figs. 4 D and 5, E and H). Activation of the PGD₂–DP1 signaling axis on rDCs results in decreased CCR7 expression so that DP1 antagonism results in augmented rDC migration to the draining LN in uninfected and SARS-CoV–infected mice (Hammad et al., 2003; Zhao et al., 2011). Of note, inhibition of DC migration by PGD₂/DP1 signaling is not confined to the lung, because, after skin invasion, Schistoma mansonii releases PGD₂ to inhibit Langerhans cell migration to the draining LN, delaying the initial immune response (Angeli et al., 2001). Similar effects are observed in Pla2g2d^−/− mice infected with either SARS-CoV or IAV, resulting in enhanced T cell responses (Fig. 7, A, B, and D, and 9 B) and, in the case of SARS-CoV, improved survival (Fig. 8 A). In addition, PGD₂ can also mediate its antiinflammatory effects through binding of 15d-PGJ₂, a dehydration end product of PGD₂, to a nuclear receptor PPARγ (Hilliard et al., 2010).

Although we primarily focused on the role of PLA₂G2D in increased PGD₂ production, PLA₂G2D is critical for the synthesis of additional antiinflammatory/proresolving molecules. PUFA–oxygenated metabolites with dedicated proresolving activities, including resolvins, protectins, and maresins (also called specialized proresolving mediators [SPMs]), are derived from ω3 fatty acids such as EPA and DHA (Serhan et al., 2002; Buckley et al., 2014; Serhan, 2014). PLA₂G2D is known to release free PUFAs into a substrate pool that can be further mobilized to produce proresolving lipid mediators such as DHA-derived resolvin D1 and EPA-derived resolvin E1 among many others (Serhan et al., 2002; Miki et al., 2013). Notably, EPA levels were higher in the lungs of uninfected middle-aged compared with young mice (Fig. 4 B), and both EPA and DHA production were higher in uninfected and infected Pla2g2d^+/+ mice compared with their Pla2g₂d^−/− counterparts (Figs. 4 H and 5 F). However we failed to detect any resolvins or protectins, the downstream metabolites of EPA or DHA in the lungs of mice, possibly reflecting low levels of these molecules in lung tissue. Notably, although EPA is higher in middle-aged compared with young lungs (DHA is nearing to significance; P = 0.12; Fig. 4 B), SPMs are lower in the uninflamed peritoneum of aged compared with young mice (Arnardottir et al., 2014), suggesting that age-dependent differences in baseline levels of these molecules are organ-specific and could possibly contribute to disease-specific outcomes.

Our results also indicated that age-dependent differences in lung PLA₂G2D levels were confined to CD11c⁺ cells. CD11c⁺ cells in draining LN also express PLA₂G2D, and relative levels are much higher than in the lung (Miki et al., 2013), but these do not change as mice age (Fig. 2 F). This specific effect of aging on pulmonary CD11c⁺ cells may reflect constant exposure to airborne antigens, most of which are innocuous. An immune response to these antigens would be undesirable and would contribute to allergic reactions and asthma. In contrast, PLA₂G2D in lymph nodes might be expected to have a different role, namely to down-regulate proinflammatory responses, and thereby contribute to resolution of the immune response after pathogen control or allergic or autoimmune reactions.

A key question is why SARS-CoV–infected mice are especially affected by age-dependent changes in PLA₂G2D expression. One unique feature of SARS is that maximal virus titers in mouse lungs are detected within 24 h after infection (Fett et al., 2013), nearly coincident with the time of maximal migration of rDCs from the lungs to the DLNs. Furthermore, coronavirus replication results in extensive membrane remodeling with the formation of double membrane vesicles and other membranous structures, the sites of viral RNA synthesis and assembly (Novoa et al., 2005; Knoops et al., 2008; Miller and Krijnse-Locker, 2008). This virus-induced reorganization of the host cell membrane causes the accumulation of cone-shaped lipids, specifically lysophosphatidylcholine (LPC) or phosphatidylethanolamine (PE; Escribá et al., 1997; Lee and Ahlquist, 2003; Diaz et al., 2010; Lorizate and Kräusslich, 2011; Xu and Nagy, 2015). Because PLA₂G2D specifically mobilizes PE to release AA and DHA (Miki et al., 2013), robust replication of SARS-CoV during the first 24 h of infection likely increases the amount of PLA₂G2D substrate. Additionally, LPC and PE may form large lipid aggregates; most secreted phospholipases, including PLA₂G2D, exhibit a tendency to show increased activity when the substrate is presented in this form (Carman et al., 1995; Gelb et al., 1995). Thus, SARS-CoV infection results in enhanced release of AA into the eicosanoid pathway, which increases the concentration of PGD₂ and downstream products of EPA and DHA, further dampening the immune response. Notably, similar processes may occur in patients with MERS-CoV infections, who also exhibit age-dependent increases in susceptibility (Assiri et al., 2013; Saad et al., 2014; Alsolamy, 2015).

In conclusion, it is likely that in the absence of challenge with a pathogen such as SARS-CoV, elevated PLA₂G2D expression benefits the host by increasing the level of various immunosuppressive lipid mediators such as PGD₂, thereby dampening the response to environmental antigens. However in the context of an infection, the higher basal levels of PGD₂ and other proresolving lipid mediators adversely affect the initiation of a protective innate immune response and ultimately adaptive immune response (Fig. 10 E). Thus, therapeutic interventions that increase the levels of proresolving lipid mediators might have undesirable effects in the context of infection. However, these results also suggest that PLA₂G2D would serve as a useful therapeutic target in the lungs of patients after infection with SARS-CoV and perhaps other respiratory viral pathogens, including MERS-CoV. Short-term targeting of a single inducible PLA₂ would avoid undesirable effects that might be observed by generalized inhibition of PLA₂ function. Such drugs have been developed for some PLA₂s (Oslund et al., 2008), although not yet for PLA₂G2D, demonstrating the feasibility of this approach.

6 wk to 22 mo specific pathogen–free C57BL/6 mice were purchased from Charles River Laboratories or the National Cancer Institute (National Institutes of Health). Pla2g2d^−/− and Pla2g2d^+/+ mice (H-2^b) were generated as described (Miki et al., 2013). After initial development, mice were backcrossed greater than 12 times to C57BL/6NCrSLc mice. Pla2g2d^+/+ C57BL/6NCrSLc mice were used as controls for all experiments. Mouse-adapted SARS-CoV (MA15) and IAV A/PR8/34 were obtained from Kanta Subbarao (National Institutes of Health) and Kevin Legge (University of Iowa, Iowa City, IA), respectively. For infections, mice were lightly anesthetized using isoflurane and infected i.n. with 1 × 10⁴ PFU SARS-CoV or 10³ TCID IAV in a 50 µl volume. All animal studies were approved by the University of Iowa Animal Care and Use Committee and meet stipulations of the Guide for the Care and Use of Laboratory Animals.

Lungs were harvested on the indicated days p.i. Tissue was homogenized in PBS using a manual homogenizer and titered on Vero E6 cells. For plaque assays, cells were fixed with 10% formaldehyde and stained with crystal violet 3 d p.i. Viral titers are expressed as PFU/gram.

Lung single-cell suspensions were prepared as described. CD11c⁺ cells were separated using anti–mouse CD11c MicroBeads (Miltenyi Biotec) following the manufacturer’s protocol.

CD11c⁺ cells from the lungs of young and middle-aged mice were separated as described above. RNA was purified using a mirVana kit (Life Technologies) according to the manufacturer’s instructions. RNA samples were assessed for purity and quality using an Agilent 2100 Bioanalyzer. RNA for the microarray was processed using a NuGEN WT-Ovation Pico RNA Amplification System together with a NuGEN WT-Ovation Exon Module. Samples were hybridized and loaded onto MouseRef-8 v2.0 Expression BeadChip (Illumina). Arrays were scanned with a HiScan bead array system, and data were collected using GeneChip Operating Software.

Data from IIlumina MouseRef-8 v2.0 Expression BeadChip were normalized and median polished using Robust Multichip Average background correction with log₂-adjusted values. After obtaining the log₂ expression values for genes, significance testing was performed comparing the two groups (CD11c⁺ cells from lungs of young and middle-aged mice). False discovery rate was applied to all p-values to correct for multiple testing. Significance of expression differences was assessed using an false discovery rate (FDR) cutoff of 0.05 and a twofold change. Analysis and visualization of data and pathway analysis were performed using PartekGS software and Ingenuity Pathway Analysis software. Complete microarray data have been deposited at the Gene Expression Omnibus under accession no. GSE71868.

CFSE and OVA-FITC were purchased from Molecular Probes. CFSE was diluted to a concentration of 8 mM in DMEM, and then administered i.n. (50 µl) 6 h before infection (Legge and Braciale, 2005; Zhao et al., 2011). OVA-FITC was dissolved in PBS, and 200 µg were administered i.n. in 75 µl.

Mice were euthanized at the indicated time points. After perfusion through the right ventricle with 5 ml of PBS, lungs and DLNs were removed. Lungs were diced into small pieces and digested in HBSS buffer containing 2% FCS, 25 mM HEPES, 1 mg/ml Collagenase D (Roche) and 0.1 mg/ml DNase (Roche) for 30 min at 37°C. After pressing through 70 µm nylon filters, lungs homogenates were spun down and treated with red blood cell lysis buffer. Cells were counted using a hemocytometer after dilution in 0.2% Trypan blue. Lymph nodes were dispersed and single-cell suspensions prepared using 70 µm nylon filters and counted.

Animals were anesthetized and transcardially perfused with PBS followed by zinc formalin. Lungs were removed, fixed in zinc formalin, and paraffin embedded. Sections were stained with hematoxylin and eosin and examined by light microscopy. Images were acquired using a BX53 microscope (Olympus), DP73 camera (Olympus), and CellSens software (Olympus). All sections were examined at room temperature using 10× or 40× objective lens. Sections were evaluated by a veterinary lung pathologist using postexamination masking techniques (Gibson-Corley et al., 2013). Sections were scored as follows: Inflammation, 1, normal range; 2, focal small patches of perivascular leukocyte aggregates (0–33% of lung); 3, multifocal patches of perivascular leukocyte aggregates (33–66% of lung); 4, widespread patches of perivascular leukocyte aggregates (>67% of lung). Edema: 1, none; 2, minor (0–33% lung fields); 3, (33–66% lung fields); 4, (>67% lung fields).

The following monoclonal antibodies were used: PE or PercP Cy5.5-conjugated rat anti–mouse CD4 (RM4-5), FITC-conjugated rat anti–mouse CD8 (53–6.7), FITC or e450-conjugated hamster anti–mouse CD11c (HL3), APC or e450-conjugated rat anti–mouse CD11b, and rat anti–mouse CD16/32 (2.4G2) (eBioScience); PercP Cy5.5-conjugated hamster anti–mouse CD3 (145-2C11), PercP Cy5.5-conjugated rat anti–mouse I-A/I-E (M5/114.15.2), PercP Cy5.5 or PE-conjugated rat anti–mouseCD45 (30-F11), Alexa Fluor 647 or APC-conjugated rat anti–mouse IFN-γ (XMG1.2), and APC-conjugated rat anti–mouse TNF (MP6-XT22; BioLegend); PE or FITC-conjugated hamster anti–mouse CD80 (16-10A1), and FITC-conjugated rat anti–mouse CD40 (1C 10; BD). For PLA₂G2D staining, rabbit anti–mouse PLA₂G2D antibody (Miki et al., 2013) was used as primary antibody, followed by staining with Alexa Fluor 647-conjugated goat anti–rabbit (Life Technologies).

For surface staining, 1 × 10⁶ cells were blocked with 1 µg of anti-CD16/32 antibody and stained with the indicated antibodies at 4°C, except for anti-CCR7 antibody, which was used at 37°C. After washing, cells were fixed using Cytofix Solution (BD). To detect antigen-specific T cells after SARS-CoV infection, 1 × 10⁶ cells were cultured in a 96-well round bottom plate in the presence of Brefeldin A and peptides corresponding to a CD8 T cell epitope (S525) or a CD4 T cell epitope (N110; Bio-Synthesis). To assess the functionality of CD11c⁺ cells, lung cells were prepared as described above and stimulated with LPS (1 µM), poly I:C (10 µg/ml), or vehicle (RP10) in the presence of Brefeldin A for 6 h at 37°C. To determine the frequency of CD11c⁺ cells expressing PLA₂G2D, lung cells were incubated for 6 h at 37°C in the presence of Brefeldin A. Cells were then fixed and permeabilized using Cytofix/Cytoperm (BD) and labeled with anti–IFN-γ (T cells), anti-TNF (CD11c⁺ cells) or anti-PLA₂G₂D antibody. To detect antigen-specific T cells after IAV infection, lung cells were prepared as described above and PA224-specific CD8 T cells were identified by staining with H2D^b/PA224 tetramer (National Institute of Allergy and Infectious Diseases Tetramer Facility, Atlanta, GA). Flow cytometric data were acquired using a FACSVerse (BD) and were analyzed using FlowJo software (Tree Star).

Total RNA was extracted from CD11c⁺ and CD11c⁻ lung cells or whole lungs of naive or SARS-CoV–infected mice at the specified time p.i. using TRIzol reagent (Invitrogen). 2 µg of RNA were used for cDNA synthesis. 2 µl of cDNA were added to 23 µl of PCR cocktail containing 2XSYBR Green Master Mix (ABI) and 0.2 µM of forward and reverse primer. Amplification was performed in an ABI Prism 7700 thermocycler. The list of genes and of forward and reverse primers is shown in Table S2. Cycle threshold (Franceschi et al., 2000) values were normalized to those of the housekeeping gene hypoxanthine phosphoribosyltransferase (HPRT) by the following equation: ΔCt = Ct_{(gene of interest)} − Ct_(HPRT). All results are shown as a ratio to HPRT calculated as 2^−(ΔCt).

Lungs were harvested into 2 ml methanol and homogenized using a tissue grinder. 200 µl were transferred to a second tube and diluted to a final volume of 3 ml of 60% methanol (vol/vol). After incubation for 30 min at 37°C, 100 µl (0.1 ng/µl) of internal standard cocktail was added. Samples were diluted to a final concentration of 10% methanol (vol/vol) with water. For AA, PGD₂, PGE₂, PGF2α, TXB₂, and 6-keto-PGF1α, internal standards were AA-d8, PGD₂-d4, PGE₂-d4, PGF2α-d4, TXB₂-d4, and 6-keto-PGF1α-d4, respectively, all purchased from Cayman Chemical and added at the aforementioned concentration. For DHA and EPA, LTB₄-d4 was used as the internal standard. Samples were centrifuged at 3,000 rpm for 10 min and the supernatant was subjected to solid phase extraction using a Strata-X 33 µm Polymeric Reversed Phase columns (60 mg/3 ml). Columns were purchased from Phenomenex, activated with 3 ml methanol, and washed with 3 ml water. Samples were eluted with 1 ml methanol, vacuum dried, and resuspended in 100 µl of mobile phase A. The analyses of PUFAs and their metabolites were performed using a Waters Acquity triple quadrupole mass spectrometer coupled with liquid chromatography (Acquity H class UPLC). Samples were then applied to an Acquity UPLC BEH C18 column (130 Å, 1.7 µm, 2.1 mm × 100 mm) and fractionated using a step gradient with mobile phase A (63:37:0.02% water/acetonitrile/formic acid) and mobile phase B (1:1 acetonitrile/isopropanol) at a flow rate of 300 µl/min for most analytes. For DHA and EPA, the mobile phase A was 65:35:0.01% methanol/water/acetic acid and mobile phase B was 100% methanol with 0.01% acetic acid. Columns were run at a flow rate of 250 µl/min. Primary standards were analyzed in parallel with the samples to enable quantification of the endogenous lipids. Detection of each lipid was performed by multiple reaction monitoring.

Naive CD45.1 (Pla2g2d^+/+) or CD45.2 (Pla2g2d^−/−) splenocytes were adoptively transferred by intravenous inoculation into CD45.2 (Pla2g2d^−/−) or CD45.1 (Pla2g2d^+/+) mice, respectively, at 50 × 10⁶ cells/mouse. 1 d after transfer, mice were infected with SARS-CoV i.n. Lungs were harvested 6 dpi and analyzed for virus-specific T cells as described in the preceding sections.

This assay was performed as described previously (Barber et al., 2003). In brief, splenocytes from CD45.1 mice (Pla2g2d^+/+, WT) were stained with high (1 µM, CFSE^hi) or low (100 nM, CFSE^lo) concentrations of CFSE. CFSE^hi cells were then pulsed with S525 (10 µM) at 37°C for 1 h. 5 × 10⁵ cells from each group (CFSE^hi and CFSE^lo) were mixed together (10⁶ cells total) and transferred i.n. into mice CD45.2 Pla2g2d^+/+ and Pla2g2d^-/- mice at 6 d p.i. Lungs were harvested 12 h later and analyzed for numbers of CFSE^hi and CFSE^lo CD45.1 cells. Percentage killing was calculated using the formula: 100 − ([% peptide pulsed in infected/% unpulsed in infected]/[% peptide pulsed in uninfected/% unpulsed in uninfected]) × 100) (Barber et al., 2003).

Young mice (6–8 wk) were treated i.n. with LPS (E. coli O111:B4) at 0.1 µg/mouse in 50 µl for 7 d.

Middle-aged mice (10–13 mo) were treated with N-Acetyl-l-Cysteine (Sigma-Aldrich; 100 mg/kg, i.p.) for 7 d.

Pulmonary oxidative stress was measured using a TBARS assay kit according to the manufacturer’s protocol (Cayman Chemicals). Data are expressed as the concentration of MDA (Imai et al., 2008) in infected lungs.

Human peripheral blood samples from leukocyte reduction cones were obtained from anonymous volunteers that had consented to blood donation at the DeGowin Blood Center at the University of Iowa. Consent forms were approved by the University of Iowa’s Institutional Review Board. To obtain monocytes, PBMCs were cultured in tissue culture plates at a seeding density of 1 × 10⁶ cells/ml in RP-10 media (RPMI-1640 medium [Invitrogen] with 10% FBS [Atlanta Biologicals] and 2 mM l-glutamine) supplemented with 5 ng/ml M-CSF (eBioscience) at 37°C with 5% CO₂. After 4 d, the plates were washed with Hanks’s balanced salt solution devoid of divalent cations (Invitrogen) to remove nonadherent cells. Adherent cells (predominantly MDMs) were then trypsinized and pelleted. Pelleted MDMs were resuspended at 1 × 10⁶ cells/ml in RP10 supplemented with M-CSF (5 ng/ml) and incubated at 37°C overnight. MDMs were subsequently seeded at a density of 5 × 10⁵/ml and incubated for 10 h at 37°C and 5% CO₂. MDMs were then treated with 0.1 nM LPS (E. coli O111:B4; Sigma-Aldrich) and M-CSF (1 ng/ml) daily for 10 d.

A Student’s t test was used to analyze differences in mean values between groups except in the case of the microarray data. All results are expressed as mean ± SEM. P-values of <0.05 were considered statistically significant. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Differences in mortality were analyzed using Kaplan-Meier log-rank survival tests. Microarray data were subjected to multiple comparisons and FDR is shown, as described above.

Fig. S1 is a schematic representation of the effects of PLA₂G2D in lungs of mice during SARS-CoV infection. Table S1 shows a list of genes in pulmonary CD11c⁺ cells expressed differentially in young and middle-aged mice. Table S2 shows qPCR primer sequences used in this study.

We thank Drs. Anthony Fehr, Rudragouda Channappanavar, Jincun Zhao and Kevin Legge for critical review of the manuscript and Dr. Ramakrishna Sompallae for help with the statistical analyses.

This work was supported by grants from the National Institutes of Health (AI060699, AI091322, S. Perlman; HL36235, M. Gelb) and by grants-in aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (22116005, 24390021) and Core Research for Evolutional Science and Technology from the Japan Science and Technology Agency.

The authors declare no competing financial interests.