Deletion of FADD in Macrophages and Granulocytes Results in RIP3- and MyD88-Dependent Systemic Inflammation

Myeloid cells, which include monocytes, macrophages, and granulocytes, are important innate immune cells, but the mechanism and downstream effect of their cell death on the immune system is not completely clear. Necroptosis is an alternate form of cell death that can be triggered when death receptor-mediated apoptosis is blocked, for example, in stimulated Fas-associated Death Domain (FADD) deficient cells. We report here that mice deficient for FADD in myeloid cells (mFADD-/-) exhibit systemic inflammation with elevated inflammatory cytokines and increased levels of myeloid and B cell populations while their dendritic and T cell numbers are normal. These phenotypes were abolished when RIP3 deficiency was introduced, suggesting that systemic inflammation is caused by RIP3-dependent necroptotic and/or inflammatory activity. We further found that loss of MyD88 can rescue the systemic inflammation observed in these mice. These phenotypes are surprisingly similar to that of dendritic cell (DC)-specific FADD deficient mice with the exception that DC numbers are normal in mFADD-/- mice. Together these data support the notion that innate immune cells are constantly being stimulated through the MyD88-dependent pathway and aberrations in their cell death machinery can result in systemic effects on the immune system.


Introduction
Dendritic cells (DCs), macrophages and monocytes are closely related cells derived from the same common myeloid progenitors [1,2]. They share common functions like antigen presentation, participation in T cell development and maintenance of gut immune system homeostasis. However, each also plays additional distinct roles in the immune system [3,4]. DCs are required for initiation of immunity; DC-less mice exhibit impaired innate immunity and diminished NK and CD8 + T cell responses to infection [5,6]. DCs also play an important role in peripheral T cell tolerance as mice with apoptosis-resistant DCs develop autoimmunity [7,8]. In contrast, loss of macrophages and monocytes has no overt effects on innate immunity but instead results in reduced Th1 adaptive immunity or defective wound healing [9,10]. the numbers of macrophages and neutrophils are not decreased but are instead elevated. Loss of RIP3 rescued the mFADD -/phenotype, indicating that these phenotypes are due to RIP3-dependent necroptosis and/or inflammatory activity. We also found that systemic inflammation was abrogated following deletion of MyD88 in these mice. These data illustrate a dynamic interplay between macrophages and other innate cells while demonstrating the importance of MyD88 in maintaining immune system homeostasis.

Ethics Statement
All animals were handled in strict accordance with good animal practice as defined by the relevant national and/or local animal welfare bodies, and all animal work was approved by the UC Berkeley ACUC Animal Care and Use Committee.

Mice
Mice were sacrificed using carbon dioxide (CO2) followed by cervical dislocation at 6-12 weeks of age unless otherwise noted. Littermates or sex-and age-matched mice were used as controls. mFADD -/mice were generated by crossing LysM-Cre transgenic mice with C57BL/ 6 FADD flox/flox mice generated previously in the lab [33]. MyD88 -/mice (C57BL/6) were provided by Dr. Shizuo Akira [43] through Dr. Greg Barton, and RIP3 -/mice were from Xiaodong Wang [22]. Experimental mice were housed in the animal facility at the University of California, Berkeley.

Macrophage Cell Enrichment
For collagenase digestion, spleens were cut into small pieces and digested with collagenase VIII at 37°C for 35 min, then incubated with 25 mM EDTA for 5 min at room temperature. Cells were dissociated and filtered through a 100 μm strainer. Alternatively, spleens were directly dissociated through a 100 μm strainer in 5ml of PBS. Red blood cell lysis with ACK lysis buffer was performed, and single cell suspensions were generated for downstream application.

Cell Death Induction
BMDMs were plated in 24-well non-tissue culture-treated plates at 10 6 cells/well in 1mL complete RPMI media. Samples were plated in duplicates. The cells were pre-treated with 10 μM zVAD-FMK and 30 μM Necrostatin-1 for 30 min and stimulated with 10 ng/ml LPS. Following 16-18 hr of stimulation, BMDMs were harvested in cold PBS and surface stained with anti-F4/ 80 and labeled with 7AAD. Samples were analyzed by flow cytometry.

ELISA and Cytometric Bead Array
Blood was collected from tail vein or cardiac puncture. Flt3L (R&D Systems) was analyzed by ELISA kit. Inflammatory cytokines were quantitated using the mouse inflammation cytometric bead array kit (BD Biosciences). Samples were collected on an LSRII and analyzed with FCAP Array Software (BD Biosciences).

Statistical Analysis
Statistical significance was calculated using paired Student's t test. Mann-Whitney U test was used to compare survival curves. Statistical analysis was completed with GraphPad Prism. ÃÃÃ p<0.001 ÃÃ p<0.01 Ã p<0.05

mFADD -/-Mice Exhibit Splenomegaly and Systemic Inflammation
To examine the role of FADD in macrophages, we crossed FADD fl/fl mice to LysM-Cre mice to generate LysM-Cre/FADD fl/fl mice (mFADD -/mice). LysM-Cre mice express Cre under the control of the lysozyme promoter. When crossed to loxP-flanked target genes, deletion was reported in~95-100% of macrophages and neutrophils [44]. No deletion was seen in T or B cells but partial deletion (16%) was seen in splenic DCs [44]. To confirm the extent of FADD deletion in macrophages, we generated bone marrow-derived macrophages (BMDM) and performed western blot analysis with FADD-specific antibodies. As seen in Fig 1A, FADD expression was undetectable in mFADD -/-BMDM. We investigated the susceptibility of these macrophages to cell death and found that addition of LPS to mFADD -/-BMDM but not the wild-type BMDM led to increased cell death (Fig 1B). This death was further enhanced by addition of zVAD-FMK, a general caspase inhibitor and could be rescued by Necrostatin-1 (Nec-1), a RIP1 kinase inhibitor [45] (Fig 1B). Interestingly, the rescue by Nec-1 was only partial, suggesting that some of the death was RIP1-independent but RIP3-dependent. In support of this, loss of RIP3 alleles completely abolished the LPS-induced necroptosis of mFADD -/macrophages (S1 Fig).
We analyzed 6-12 week-old mice and compared mFADD -/mice to their littermate controls (FADD fl/fl or LysM-Cre/FADD fl/+ ). Spleen and lymph nodes were analyzed for various immune populations in a manner similar to that which was performed for dcFADD -/mice [34]. Interestingly, similar to dcFADD -/mice, mFADD -/mice suffer from splenomegaly ( Fig 1C) and lymphadenopathy ( Fig 1E) with increased neutrophils (Ly6C lo CD11b + ), inflammatory monocytes (Ly6C hi CD11b + ), macrophages (CD11b + F4/80 + ), B cells (B220 + ) and Ter119 + erythrocytes (Fig 1D, 1F, and 1G). The elevated number of macrophages is surprising given that FADD-deficient macrophages are sensitive to necroptosis. In addition, T cell composition and number appear to be the same between mFADD -/and littermate controls (Fig 1F and data not shown). We also analyzed the splenic DC number. As expected, they are similar to that of littermate controls for both splenic DCs ( Fig 1H) and DCs in the gut-associated lymph nodes (see below). In support of this, we measured Flt3L levels in the serum by ELISA. Flt3L is required for DC differentiation and its levels inversely correlate with the number of DCs in vivo [46][47][48]. Both DC-less mice and dcFADD -/mice, which have no DCs and fewer DCs, respectively, exhibit elevated levels of Flt3L [34,46]. As shown in Fig 1H, the serum Flt3L levels are the same between mFADD -/mice and their littermate controls. Thus, this result confirms that mFADD -/mice contain normal numbers of DCs in their immune organs.
To see if mFADD -/mice exhibit elevated inflammatory cytokines, we performed cytometric bead array on sera of 6-12 week-old mFADD -/mice. Similar to dcFADD -/mice, these mice exhibit a slight elevation of serum TNF (Fig 2A). However, unlike dcFADD -/mice, they also showed statistically significant increases in IL-6, IL-10, and IL-12. Moreover, injection with a low dose of LPS resulted in death of 80% of mFADD -/mice within 30 hrs (Fig 2B). In contrast, LPS did not cause any lethality to the littermate controls. Thus, mFADD -/mice exhibit systemic chronic inflammation and succumb to LPS-induced endotoxic shock similar to that of dcFADD -/mice.

mFADD -/-Systemic Inflammation is RIP3-Dependent
In dcFADD -/mice, the lower number of DCs in GALT can be rescued by RIP3 deficiency. Moreover, systemic inflammation is resolved in these dcFADD -/-/RIP3 -/mice. These data suggest that gut microbiota stimulate DCs to die through necroptosis when apoptosis is blocked [34]. Necroptotic DCs then release inflammatory contents, which can be sensed by other cells, resulting in systemic inflammation. Subsequently, we assessed if FADD-deficient macrophages die through necroptosis after being stimulated by commensal microflora through their TLRs. Examination of F4/80 macrophages in mesenteric lymph nodes however, did not reveal a decrease in the number of macrophages ( Fig 3A). As expected, no changes were observed in the numbers of CD103 + migratory DCs in mesenteric lymph nodes (mLN),which is consistent with the lack of changes in Flt3L serum levels (Figs 3B and 1G). Thus, the systemic inflammation of mFADD -/is unlikely due to any possible effects from their DCs. As seen in the spleen, neutrophil and inflammatory monocyte numbers were still increased in the mesenteric lymph nodes of mFADD -/mice (Fig 3C, 3D, and 3E).
To examine the role of RIP3, we crossed mFADD -/mice to RIP3 -/mice. As shown in Fig 4, analysis of mFADD -/-/RIP3 -/and RIP3 -/littermates in comparison to mFADD -/and wildtype controls indicate that the mFADD -/phenotypes are partially rescued in mFADD -/-/ RIP3 -/mice. Although there is a slight increase in spleen weight, neutrophils, and inflammatory monocytes in mFADD -/-/RIP3 -/mice when compared to RIP3 -/mice, the numbers were reduced as compared to mFADD -/mice (Fig 4A, 4B and 4C). Furthermore, the number of splenic macrophages as well as peripheral lymph node and mesenteric lymph node weights in mFADD -/-/RIP3 -/mice were similar to that found in RIP3 -/mice (Fig 4D, 4E, and 4F). The most well characterized function of RIP3 is its role in necroptosis induction, the subsequent promotion of inflammation is thought to be a secondary event due to the release of damage-associated molecular patterns (DAMPs) by necrotic cells [34,49]. However it has recently been appreciated that RIP3 can also function directly in promoting inflammation through production of inflammatory cytokines [38][39][40]42]. Thus, RIP3-dependent inflammatory activity, whether indirectly through necroptosis or directly through promotion of inflammatory cytokines, is responsible for the systemic inflammatory phenotype found in mFADD -/mice.

MyD88-Dependent Signaling Is Crucial for mFADD -/-Inflammation
We have previously shown that MyD88, an adapter protein essential for most TLR signaling, is required for the systemic inflammatory phenotypes of dcFADD -/mice [34]. To assess the requirement of MyD88 in FADD-deficient macrophage-induced inflammation, we similarly crossed mFADD -/mice to MyD88 -/mice. Analysis was then carried out for mFADD -/-/ MyD88 -/mice for comparison to MyD88 -/littermates. In some cases, we were also able to obtain mFADD -/littermates for our analysis. We found that loss of MyD88 rescued the inflammatory phenotype in mFADD -/mice. A decrease in spleen weight and normal numbers of neutrophils, inflammatory monocytes and macrophages were seen in mFADD -/-/MyD88 -/mice ( Fig 5). Likewise, loss of MyD88 rescued cell numbers seen in the mesenteric lymph nodes (Fig 3). These data suggest that aberration in cell death machinery, whether in macrophages or DCs, results in MyD88-driven inflammation [34].

Discussion
In this paper, we showed that FADD is not required for normal macrophage development or proliferation but that the loss of FADD results in macrophage sensitivity to LPS-induced necroptosis and RIP3-dependent inflammation. This is similar to TLR-stimulated FADD-deficient DCs [34]. TLR3 and TLR4 in macrophages/DCs can presumably activate necroptosis through association of RIP3 with the adapter protein TRIF [35,36]. We have previously shown that FADD-deficient DCs can be stimulated to undergo necroptosis through MyD88 as well, although the molecular mechanism of MyD88-dependent death in DCs is not clear. In bone marrow-derived macrophages, MyD88-dependent necroptosis in vitro was reported to occur through a TNF-dependent mechanism [36]. However, we found that addition of a TNF neutralizing antibody did not rescue LPS induced death of mFADD -/-BMDM (S1 Fig). Nevertheless, it is possible that other innate immune stimuli may result in TNF-dependent necroptosis and therefore contribute to the inflammation observed. Alternatively, MyD88 may directly activate RIP3 in vivo through a novel mechanism without going through the TNF pathway. Here we report that mFADD -/inflammatory phenotypes disappear on a MyD88-deficient background. It is unlikely that the rescue is due to a complete absence of necroptotic macrophages/neutrophils as TRIF can still provide signals to trigger cell death. Consistent with this, DC-specific loss of MyD88 only partially rescues the dcFADD -/phenotypes [34]. Although we didn't generate macrophage-specific MyD88-deficient mice, we expect the results to be similar to that of dcFADD -/-/dcMyD88 -/mice. In dcFADD -/mice, the inflammatory phenotype is also rescued by antibiotics administration. Together, these data are consistent with the notion that innate immune cells are primed continuously through MyD88 signaling, and this is crucial for their ability to respond to danger signals including those released by necrotic cells.
Unlike dcFADD -/mice where a significant reduction of DC number in GALT was detected, the number of macrophages in mFADD -/mice is surprisingly elevated instead of decreased. Since the observed phenotypes in mFADD -/mice are similar to that of dcFADD -/mice, we considered the possibility that leakiness of the lysozyme-driven Cre in DCs could result in a significant number of necroptotic FADD-deficient DCs in mFADD -/mice. Although we did find the number of CD103 + DCs in mesenteric lymph nodes of mFADD -/mice to be mildly decreased, the number was not statistically different from their littermate controls (Fig 3B). Moreover, the serum Flt3L levels were completely normal in mFADD -/mice. Flt3L levels are an excellent indicator of DC homeostasis in mice as they inversely correlate with DC number as shown in our previous analysis of dcFADD -/mice and in DC-less mice [34,46]. Thus, our data indicate that it is unlikely that necroptotic DCs are responsible for the mFADD -/phenotypes. Our observation of increased macrophage cell numbers appears to disagree with necroptosis being the sole or major contributor to mFADD -/systemic inflammation. Recently, it was reported that RIP3 may promote inflammation independent of its role in necroptosis [38][39][40]42]. Our data on increased macrophage cell numbers appears consistent with this direct RIP3-inflammatory role. It has been reported that macrophages and dendritic cells deficient for caspase-8 or cIAP1, cIAP2, and XIAP can promote RIP3-dependent production of IL-1β in response to LPS [41,42]. Furthermore, LPS-induced IL-1β maturation in dendritic cells was shown to consist of two RIP3-dependent pathways involving activation of caspase-1 or caspase-8 [38]. We found that loss of RIP3 from mFADD -/mice rescued the inflammatory phenotype of our mice (Fig 4). Thus, although RIP3-deficient macrophages have been reported to have no defect in NF-κB activation or pro-inflammatory cytokine production in response to TLR or TNF stimulation, it is possible that other innate immune stimuli may activate nonnecroptotic RIP3 inflammatory activity in mFADD -/mice [39,50].
Alternatively, it is plausible that a small population of FADD-deficient macrophages that are stimulated to die release inflammatory DAMPs, which results in proliferation of the rest of the macrophage population. Increased numbers of macrophages in response to necroptotic DCs was one of the phenotypes observed in the dcFADD -/mice [34]. In addition, we observed that LPS stimulation alone caused a small increase in cell death of FADD-deficient macrophages (Fig 1). Consequently, the inflammatory contents released by these dying macrophages in the mFADD -/mice may also stimulate other immune cells, leading to their activation and contributing to chronic inflammation. Given the emerging data on RIP3's role as both a direct and indirect contributor to inflammation, it is likely that rescue of systemic inflammation in mFADD -/-RIP3 -/mice is due not only to loss of macrophage necroptosis but also loss of RIP3-dependent inflammatory activity. This suggests that in vivo, FADD may play an important role in limiting RIP3 driven inflammatory activity, whether it be through necroptosis or other inflammatory pathways. As many of the same proteins identified in necroptosis induction have also been implicated in RIP3's non-necroptotic inflammatory activity, there is a need for additional studies to fully evaluate the contribution and activation of these disparate functions [42,51,52].
In conclusion, the data presented here as well as that from dcFADD -/mice demonstrate the dynamic relationship between immune cells and the microbiota. They support the notion that these innate immune cells are important sentinels of the immune system, poised to respond to aberrations in cell death signaling and DAMPs.