FADD deficiency impairs early hematopoiesis in the bone marrow

Abstract

Signal transduction mediated by FADD represents a paradigm of co-regulation of apoptosis and cellular proliferation. During apoptotic signaling induced by death receptors including Fas, FADD is required for the recruitment and activation of caspase 8. In addition, a death receptor-independent function of FADD is essential for embryogenesis. In previous studies, FADD deficiency in embryonic stem cells resulted in a complete lack of B cells and dramatically reduced T cell numbers, as shown by Rag1^−/− blastocyst complementation assays. However, T-specific FADD-deficient mice contained normal numbers of thymocytes and slightly reduced peripheral T cell numbers, whereas B cell-specific deletion of FADD led to increased peripheral B cell numbers. It remains undetermined what impact a FADD deficiency has on hematopoietic stem cells and progenitors. The current study analyzed the effect of simultaneous deletion of FADD in multiple cell types including bone marrow cells by using the IFN-inducible Mx1-cre transgene. The resulting FADD mutant mice did not develop lymphoproliferation diseases, unlike Fas-deficient mice. Instead, a time-dependent depletion of peripheral FADD-deficient lymphocytes was observed. In the bone marrow, a lack of FADD led to a dramatic decrease in the hematopoietic stem cells and progenitor-enriched population. Furthermore, FADD-deficient bone marrow cells were defective in their ability to generate lymphoid, myeloid and erythroid cells. Thus, the results revealed a temporal requirement for FADD. Whereas dispensable during lymphopoiesis post lineage commitment, FADD plays a critical role in early hematopoietic stages in the bone marrow.

Introduction

Apoptosis plays a critical role in mammalian development and homeostasis (1–2). The intrinsic apoptotic signaling is mediated by the mitochondrion, involving cytochrome C and the Bcl-2 family proteins (3). The extrinsic apoptotic pathways are initiated by ligation of death receptors (DRs) by either their cognate ligands or cross-linking antibodies (4–5). DRs, including TNF-R1, Fas/Apo-1, and TRAIL-Rs (DR4 and DR5), activate a caspase cascade through the adaptor protein, FADD, which recruits pro-caspase 8 to form the death-inducing signaling complex (DISC) (6–11). The assembly of the DISC promotes the activation of caspase 8 by self processing, and the resulting active caspase 8 cleaves downstream caspases and other cellular proteins, leading to cell death.

The importance of apoptosis in development is exemplified by the embryonic defects caused by a lack of Bcl-x or Mcl-1, members of the Bcl-2 family which mediate the intrinsic pathway (12–13). Deficiencies in pro-apoptotic Bcl-2 family members such as Bax and Bak resulted in interdigital webbing due to insufficient death of superfluous cells (14–15). Deletion of another pro-apoptotic Bcl-2 family member, Bim, leads to autoimmune diseases caused by impaired death of autoreactive lymphocytes (16). The extrinsic pathways are essential for maintaining homeostasis in the immune system. In particular, a systemic loss of Fas leads to the development of an age-dependent lymphoproliferative (lpr) and autoimmune disease (17).

Although expressed in a wide range of tissues, death receptors do not appear to play an overt role in mouse development (18–22). Interestingly, mice deficient in FADD or caspase 8 die in utero by day 9.5–10.5 of gestation (23–25). Conditional deletion of FADD or caspase 8 following lineage commitment in double-negative thymocytes or pro-B cells resulted in no significant defects in the maturation of T or B cells within primary lymphoid organs (26–30). When tested in vitro, FADD^−/− or caspase 8^−/− lymphocytes are defective in death receptor-induced apoptosis. Additionally, FADD^−/− and caspase 8^−/− T cells failed to expand efficiently upon stimulation through the TCR, while FADD^−/− and caspase 8^−/− B cells are impaired in TLRinduced proliferation. The effect of FADD deficiency on the development and function of myeloid cells and hemaptopoietic progenitors has not been determined.

The process of hematopoiesis commences with hematopoietic stem cells (HSCs) (31). HSCs are located in the endosteal niche within the bone marrow, and are characterized as Lin⁻ Sca-1⁺c-Kit⁺ (32). HSCs proceed to differentiate into multipotent progenitor cells (MPPs), which subsequently develop into either common lymphoid progenitors (CLPs) or common myeloid progenitors (CMPs) (33–34). The differentiation of MPPs indicates the branching point at which developing hematopoietic cells commit to the lymphoid or myeloid lineage. CLPs develop into precursors for the T, B and natural killer cell populations, while CMPs further differentiate into the granulocytic-macrophage progenitors (GMP) and megakaryocytic-erythroid progenitors (MEP) (33, 35). GMPs differentiate into macrophages and granulocytes, while MEPs are capable of differentiation into megakaryocytes and erythrocytes (36–37). Both CLPs and CMPs lead to dendritic cell differentiation, therefore, dendritic cells can be of either myeloid or lymphoid origin (34, 38–39).

Apoptosis, proliferation, and differentiation of HSCs and progenitors are tightly regulated. For example, a severe impairment of hematopoiesis occurs following the loss of anti-apoptotic functions mediated by proteins of the intrinsic pathway, including Bcl-x and Mcl-1 (12, 40–42). The function of proteins of the extrinsic apoptotic pathways has also been investigated. Whereas there has been no HSC defect reported in mice lacking individual DRs, inducible deletion of caspase 8 resulted in greatly diminished in vitro differentiation of hematopoietic progenitor cells (43). In a previous study, the effect of dominant negative mutants of FADD and caspase 8 on fetal liver progenitor cells was analyzed (44). However, the impact of simultaneous deletion of FADD in early hematopoietic lineages has not been formally tested. In this study, we utilized a Cre recombinase under the control of the Mx1 promoter in order to induce the deletion of FADD in multiple cell types including bone marrow cells. These results help establish that FADD is essential at early stages of hematopoiesis, as its deletion in bone marrow cells impaired the peripheral lymphoid, myeloid and erythroid lineages.

Materials and Methods

Mice and primary cell isolation

FADD:GFP mice have been described (26–27), and were crossed to mice bearing the Mx1-cre transgene purchased from the Jackson Laboratory. Mice were housed in germ-free rooms in Thomas Jefferson University research animal facilities. The procedures were approved by the IACUC. Excision of the FADD:GFP transgene in mice bearing the Mx1-cre transgene was induced via injection of the double-stranded RNA, polyinosine-polycytidylic acid (poly I:C; Sigma-Aldrich; Invivogen). Mice were injected 3 times with 400 μg of poly I:C in 200 μl of H₂O intraperitoneally every other day as described previously (42, 45–47). At the indicated times after the last injection, cells were isolated from the thymus, spleen, lymph nodes, femurs and tibias. The peritoneal cavity was lavaged with PBS. Red blood cells were lysed hypotonically with ACK lysis buffer. Cells were washed 2–3 times with PBS and counted using a hemocytometer.

Flow cytometry and cell sorting

To detect HSCs, bone marrow cells were washed with staining buffer (3% BSA, 1 mM EDTA, 0.05% NaN₃ in PBS), and stained with the following antibodies; Sca-1-PeCy5 (eBioscience), c-Kit-Phycoerythrin (PE) and an Allophycocyanin(APC)-Lineage Cocktail (BD Pharmingen), and 4-color flow cytometric analysis performed on a FACS Calibur (BD Biosciences), due to FADD:GFP expression. Deletion was assessed by determining the GFPpopulation present in the c-Kit⁺Lin⁻ or Sca-1⁺c-Kit⁺Lin⁻ population as previously described (42, 48). For lymphoid progenitor analyses, bone marrow cells were stained with APC-conjugated lineage cocktail, c-Kit-PECy7 (BD Pharmingen), Sca-1-PECy5 and IL7Rα-PE (eBioscience) (5-color analyses including GFP). To analyze myeloid and erythroid progenitor populations, bone marrow cells were stained with APC-lineage Cocktail, c-Kit-PECy7, CD16/CD32/FCγR_II/III-PE, biotinylated-Sca-1 (eBioscience), Streptavidin-PE-Texas Red (Caltag-Invitrogen), and CD34- PECy5 (Biolegend) (6-color analysis including GFP). In some experiments, purified anti-CD34 (BD Pharmingen)/anti-rat-IgG-Texas Red (Jackson Immunoresearch) and Sca-1-PECy5 (eBioscience) were used. These 6-color analyses of progenitors were performed on a MoFlo cell sorter (Dako Cytomation). To analyze lymphocyte populations, cells were washed once with staining buffer, and labeled with the following lineage-specific flourochrome-conjugated antibodies: CD4-PE, CD8-TriColor (TC), B220-TC, streptavidin-R670 (Caltag-Invitrogen), CD5-PE (eBioscience), CD3-PE, CD19-Bio, c-Kit-PE, CD25-PE, IgD-PE (BD Pharmingen), IgM-FITC (Jackson Immunoresearch), IgM-PE, IgD-Bio (Southern Biotech). Cells were analyzed using a Coulter Epics XL cytometer (Beckman Coulter). FlowJo (Tree Star Inc) was used for the generation of histograms and dot plots. For sorting, bone marrow cells were resuspended in sorting medium (1:1 PBS: RPMI). GFP- bone marrow cells were isolated using a MoFlo high speed cell sorter in the Flow Cytometry Facility at Thomas Jefferson University.

L-cell conditioned media

We used a protocol as described previously (49). L929 cells were purchased from American Type Culture Collection and were cultured in DMEM (Mediatech) supplemented with 10% Fetal Bovine Serum (FBS; Mediatech), 100 U/ml penicillin and 100 μg/ml streptomycin (Mediatech). Confluent cells were detached with 3 ml of a solution containing 0.25% Trypsin and 2.21 mM EDTA, washed with DMEM and counted. Cells were resuspended at a density of 2.4 X 10⁵ in 58 ml of DMEM, plated in a 175 cm³ tissue culture flask and were ncubated at 37°C with 10% CO₂ for 7 days. On day 7, DMEM from the previous 7 days was filtered through .45 μm filters (Corning) and stored at −20°C as week 1 L-cell conditioned media (LCCM). Fresh DMEM (58 mL) was added to the flasks and incubated for an additional 7 days. On day 14, the DMEM was filtered and stored at −20°C as week 2 LCCM. This media was added to DMEM to make bone marrow macrophage media (see below).

Preparation of bone marrow-derived macrophages

The protocol was based on those described previously (22, 50). Bone marrow cells were resuspended (10⁶/ml) in DMEM (Mediatech) containing 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin and 30% LCCM (15% Week 1 LCCM, 15% Week 2 LCCM). Cells were cultured for 5 days at 37°C with 5% CO₂ in 10 cm plates. Fresh media was added on day 5 and the media was replaced on day 6, in order to remove non-adherent cells, and on day 7, plates were photographed with a Nikon digital still camera (Model #DXM1200) on a Nikon Eclipse TS100 inverted microscope. Adherent cells were removed by incubating with 10 mM EDTA in PBS as described (51) and counted using a hemocytometer.

Preparation of bone marrow-derived dendritic cells (DCs)

The protocol was modified from those published previously (52–53). Bone marrow cells were diluted to 10⁶/ml in RPMI 1640 (Mediatech) supplemented with 10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, 10 ng/ml of rGM-CSF (R & D Systems) and 10 ng/ml of rIL-4 (R & D Systems), plated for 6 days with 5% CO₂. On day 6, the top 80% of the media was discarded and the bottom 20% used to wash the semi-adherent cells from the bottom of the plate, without disturbing the macrophage monolayer. The resulting cell suspension was added to fresh RPMI 1640 supplemented with 10% FBS, 100 U/mL penicillin, 100 ug/mL streptomycin, 10 ng/ml of rGM-CSF and 10 ng/ml of rIL-4 and plated for an additional 2 days, and on day 8, cells in plates were photographed with a Nikon digital still camera (Model #DXM1200) on a Nikon Eclipse TS100 inverted microscope. DC yield was assessed on day 8 by counting cells by trypan blue exclusion, and staining with anti-CD11b-PE and anti-CD11c- fluorescein isothiocyanate (FITC) (BD Biosciences).

In vitro colony forming unit (CFU) assays

This assay was performed as described (54–55) GFP⁻ bone marrow cells were isolated by sorting and washed 2 times with Isecove’s Modified Dulbecco’s Medium (IMDM) (Gibco) containing 2% FBS. Cells were diluted to 2 x 10⁵/ml in IMDM-2% FBS. For each dish, 2.5 x 10⁴ cells were mixed with 1 ml of methylcellulose medium containing 50 ng/ml recombinant mSCF, 10 ng/ml recombinant mIL-3, 10 ng/ml recombinant hIL-6 and 3 U/ml erythropoietin (M3434, StemCell Technologies) and plated in 35 mm cell culture dishes utilizing a 3 ml syringe and a 16 gauge blunt-end needle. Assay plates were cultured at 37°C with 5% CO₂ for 10–14 days. Plates were then scored for erythroid burst-forming units (BFU-E), granulocytic-erythrocyticmegakaryocytic-macrophagic (GEMM)-CFUs and granulo-macrophagic (GM)-CFUs by microscopic analysis based upon morphological appearance, using a gridded scoring dish (StemCell Technologies)

Splenic colony forming unit (CFU-S) assays

The protocol was based on those described (43, 54). Bone marrow cells were counted and 1 X 10⁵ cells were injected retroorbitally in 200 μl PBS into irradiated recipient C57BL/6 (B6) mice (8.5 Gy, ¹³⁷Cs source). Mice were treated with Sulfatrim over the course of the experiment. 8 Mice were sacrificed 8 days post transplantation. Spleens were photographed and weighed, then fixed in Bouin’s fixative (75 ml picric acid, 25 ml formalin, 5 ml glacial acetic acid) (43, 56). Colonies were counted following fixation.

Mouse hematocrit assays

Following administration of poly I:C, mice were retroorbitally bled using heparinized micro-hematocrit capillary tubes (22–362–566, Fisher). Capillary tubes were sealed with Critoseal (McCormick/Fisher), centrifuged for 5 minutes in a Readacrit centrifuge (Clay Adams), and red blood cell percentage was measured by dividing red blood cell volume from total volume (57)

Rag-1^−/− adoptive transfer experiments

GFP⁻ bone marrow cells (1 X 10⁶) in 200 μl PBS were injected retroorbitally into sublethally irradiated Rag-1^−/− recipient mice (4 Gy, ¹³⁷Cs source). Mice were kept on Sulfatrim and were sacrificed 8 weeks post-injection. Organs were harvested and analyzed by flow cytometry.

Results

Inducible deletion of FADD

We previously described the FADD^−/− FADD:GFP^flox mice in which a floxed FADD:GFP fusion transgene corrected the developmental defect caused by deletion of the endogenous FADD alleles (23, 26). Lineage-specific deletion of FADD:GFP^flox in pro/pre T and B cells was achieved using the Lck-cre and CD19-cre transgenes, respectively (26–27). To analyze the effect of simultaneous deletion of FADD in multiple cells types in adult mice, including bone marrow progenitor cells, the Mx1-cre transgene was crossed into FADD^−/− FADD:GFP^flox mice. Systemic administration of poly I:C was carried out in the resulting FADD^−/− FADD:GFP^flox Mx1-cre mutant mice and FADD^+/− FADD:GFP^flox Mx1-cre control mice which contain one allele of endogenous FADD. The Mx1 promoter can be activated by stimulation with type I interferons (IFNs), or endogenous IFN can be induced by injection of the double stranded viral RNA mimic poly I:C (45). At 14 days post induction of deletion, cells from the bone marrow, thymus, spleen, and lymph nodes were isolated and analyzed by flow cytometry. FADD:GFP is ubiquitously expressed in bone marrow cells, thymocytes, and peripheral cells as detected by flow cytometric analyses (Fig. 1). As reported in previous studies using theses mice (26–27), deletion of FADD:GFP was readily detectable by flow cytometry, as indicated by the presence of a GFPpopulation in control and mutant mice injected with poly I:C, as well as by western blotting (Fig 1A and B). The relative deletion efficiencies across mice (average±standard deviation) range as follows. Bone marrow: control 79.5 to 93.1% (87.1±5.9%) and mutant 67.6–92.1% (82.3±10.4%); thymus: control 60.8–98.5% (84.78±17.2%) and mutant 83.1–94.3% (89.8±4.8%); spleen: control 54.3–62.8% (58.1±4.1%) and mutant 40.7–61.7% (55.8±10.1%); lymph nodes: control 40.8–57.4 (51.9±5.9%) and mutant 40.0–56.0% (47.8±6.8%) (n=5).

Figure 1. Inducible deletion of FADD:GFP using the Mx1-cre system.

A, Single cell suspensions prepared from the indicated organs were subject to flow cytometric analysis, and presence and absence of GFP was illustrated in histograms. FADD^+/− mice were used as GFPcontrols, and FADD^−/− FADD:GFP^flox mice, uniformly expressing FADD:GFP in cells of various hematopoietic organs as indicated by a discrete GFP-positive peak, were used as GFP⁺ controls. FADD^+/− FADD:GFP^flox Mx1-cre control mice expressing an endogenous wild type allele of FADD and FADD^−/− FADD:GFP^flox Mx1-cre mutant mice were injected with poly I:C. Deletion of FADD:GFP in various organs was indicated by the presence of GFP⁻ cells, two weeks after the final injection. Numbers indicate percentages of GFP⁻ cells. The data showed a representative of 5 control and mutant mice. The relative deletion efficiencies across mice (average ± standard deviation) was described the results. B, Western blotting showed that sorted GFP⁻ bone marrow cells and spleen cells contained undetectable FADD:GFP protein. Protein transfer was confirmed by staining the nitrocellulose membrane with Ponceau S (Sigma).

Total cell numbers in various hematopoietic organs were determined. When compared to FADD^+/− FADD:GFP^flox Mx1-cre control mice, a significant decrease in the total cellularity in the bone marrow and thymus was observed in FADD^−/− FADD:GFP^flox Mx1-cre mutant mice (Fig. 2A). FADD deficiency also resulted in somewhat lower numbers of splenocytes. The percentages of double negative (DN), double positive (DP), and single positive populations of thymocytes in FADD^−/− mutants were similar to those in control mice (data not shown). Analysis of splenocytes and lymph node cells showed a mild increase in B220⁺ B cells in FADD^−/− FADD:GFP^flox Mx1-cre mutant mice as compared to FADD^+/− FADD:GFP^flox Mx1-cre control mice (Fig. 2B). Conversely, fewer CD3⁺ T cells were found in mutant mice than in controls.

Figure 2. FADD deletion led to reduced hematopoietic pools.

A, Total cell numbers of the indicated organs from pairs of wild type (WT) and FADD mutant (Mut) were determined at 4 days after the last dose of poly I:C injection. Horizontal bars indicate the mean values. B, Mice of the indicated genotypes were injected with poly I:C, and two weeks after the final injection, flow cytometric analysis of splenic and lymph node cells were performed. Deletion of FADD:GFP in CD3⁺ T cells and B220⁺ B cells was indicated by the production of the GFP⁻population. Numbers indicate the percentages of lymphocyte lineages and the percentages of the GFP⁻ population in each lineage in the indicated organs.

Time-dependent depletion of FADD^−/− cells

Recent studies have shown that conditional deletion of Fas using CD11c-cre led to an lprlike phenotype, suggesting a role for Fas-induced apoptosis in non-lymphoid cells (58). Therefore, FADD mutant mice were aged for three additional months following poly I:C injections, yet they did not appear to develop lpr-like phenotypes akin to that of the Fas-deficient mice. However, flow cytometric analysis revealed an unexpected phenotype. Although the initial deletion efficiencies (5 days post injection) in the control (45.2 ± 9.3%) and mutant mice (43.6 ± 7.8%) were similar (n=5), the percentage of FADD^−/− (GFP⁻) cells in the bone marrow, thymus and periphery of FADD^−/− FADD: GFP^flox Mx1-cre mice, were dramatically decreased compared to control FADD^+/− FADD: GFP^flox Mx1-cre mice as they aged, (Fig. 3A). At three months post poly I:C injection, control mice contained more GFP⁻ T cells (>50%) and GFP⁻ B cells (>74.7%) in the spleen and lymph nodes (Fig. 3B) than control mice at two weeks after induction of deletion (Fig. 2B). In contrast, FADD mutant mice contained greatly reduced numbers of FADD^−/− T cells (GFP⁻, 7.5%) and B cells (GFP⁻, 15%) (Fig. 3B and S1). These results indicate that FADD is required for the maintenance of various hematopietic cells in the bone marrow, thymus and periphery.

Figure 3. Time-dependent depletion of FADD^−/− cells.

Three months after the final injection of poly I:C, cells of the indicated organs were subject to flow cytometric analyses. Percentages of GFP⁻ FADD^−/− mutant cells in the total organs (A) and in the peripheral lymphocyte populations (B and C) in FADD^−/− FADD:GFP^flox Mx1-cre mutant mice were dramatically decreased, when compared to the percentages of GFP⁻ FADD^+/− cells in the FADD^+/− FADD:GFP^flox Mx1-cre control mice. The data shown is a representative of 5 control and mutant mice.

In vitro myeloid and lymphoid cell derivation from FADD^−/− bone marrow cells

Depletion of FADD^−/− cells in the bone marrow and periphery may be due to defects in hematopoietic progenitor cells. To test this hypothesis, in vitro differentiation assays were performed. Bone marrow macrophage progenitor cells can be induced to differentiate into mature macrophages in vitro by culturing them in the presence of macrophage colony–stimulating factor (M-CSF) (49, 59). GFP⁻ bone marrow cells were sorted from FADD^+/− FADD:GFP^flox Mx1-cre control and FADD^−/− FADD:GFP^flox Mx1-cre mutant mice and cultured in conditioned medium containing M-CSF as described (22). As seen in Figures 4A and B, FADD^+/− control bone marrow cells readily differentiated into macrophages, while FADD^−/− bone marrow cells showed a dramatically decreased propensity for macrophagic cell production.

Figure 4. In vitro bone marrow progenitor-derived macrophages and DCs.

In vitro generation of macrophages from bone marrow cells isolated from FADD^+/− FADD:GFP^flox Mx1-cre control and FADD^−/− FADD:GFP^flox Mx1-cre mutant mice following a 7 day culture in MCSF-containing media (A and B). Dendritic cells were generated by culturing in media containing GM-CSF and IL-4 for 8 days (C, D, and E). Cells in the plate were photographed (A and C), cell numbers were enumerated and bar graphs generated (B and D), flow cytometric analysis of dendritic cell populations was undertaken by staining with antibodies specific for CD11b and CD11c (E).

To ascertain whether the defects seen in macrophage generation apply to other hematopoietic lineages, GFP⁻ FADD^−/− and FADD^+/− bone marrow cells were induced to differentiate into dendritic cells (DCs) by culturing with GM-CSF and IL-4 (53, 60). Following 8 days in culture, FADD^−/− bone marrow cells generated a lower yield of bone marrow-derived dendritic cells as determined by cell number, when compared to FADD^+/− bone marrow (Fig. 4C and D). Additionally, a lower percentage of CD11b⁺CD11c⁺ dendritic cells developed in FADD^−/− bone marrow cultures than in FADD^+/− cultures (Fig. 4E). Taken together, these results illustrate impairment of the capability of hematopoietic progenitor cells to generate macrophages and DCs in vitro when FADD deletion is induced in the bone marrow.

FADD^−/− bone marrow cells have decreased in vitro colony forming capability

Multipotent hematopoietic progenitors develop from hematopoietic stem cells following differentiation and loss of self-renewal potential (61). The obstruction in differentiation displayed by FADD^−/− hematopoietic progenitors (Fig. 4) may be caused by a defect in HSCs and/or the further differentiated multipotent progenitors. In order to investigate the differentiation capability of FADD^−/− HSCs and progenitors, in vitro colony forming unit (CFU) assays were performed. Hematopoietic progenitor cells, when cultured in a semisolid methylcellulose-based medium supplemented with suitable growth factors, proliferate and differentiate to produce clonal clusters of maturing cells. FADD^−/− and FADD^+/− bone marrow cells were cultured as a single-cell suspension within a semi-solid methylcellulose matrix containing recombinant mouse (rm)-SCF, rmIL-3, recombinant human (rh)-IL-6 and erythropoietin. This allowed for the suspended isolation of solitary HSCs within the matrix and the measurement of their cytokine-induced differentiation into hematopoietic progenitor colonies. When compared to FADD^+/− controls, mutant FADD^−/− bone marrow cells displayed a significantly reduced capacity to differentiate into multipotent myeloid and erythroid colonies (Fig. 5A–D). This demonstrates that the previously observed defect in macrophage and DC development in FADD^−/− bone marrow progenitors in vitro is caused by a block in the differentiation of HSCs into further committed progenitors.

Figure 5. Defects in in vitro colony formation from FADD^−/− HSCs.

Bone marrow cells from FADD^+/− FADD:GFP^flox Mx1-cre control and FADD^−/− FADD:GFP^flox Mx1-cre mutant mice were cultured in a semi-solid media containing rm-SCF, rm-IL-3, rh-IL-6 and EPO to differentiate hematopoietic stem cells for 10–14 days. Plates were photographed for morphological appearance of colonies (A) and colonies scored for the presence of granulocytic/macrophagic (CFU-GM), granulocytic/erythrocytic/macrophagic/megakaryocytic (CFU-GEMM) and erythrocytic (BFU-E) (B-D).

The in vivo effect of bone marrow FADD deficiency

To determine whether the in vitro defect in bone marrow cells is applicable in vivo, we performed splenic colony forming assays (62). FADD^+/− and FADD^−/− bone marrow cells were injected into lethally-irradiated wild type recipient C57BL/6 mice. Eight days post transfer, mice were sacrificed, spleens were weighed and fixed, and colonies were enumerated. Gross morphology of the spleens in mice injected with FADD^+/− cells differed from FADD^−/− injected mice (Fig. 6A). Spleens repopulated with FADD^−/− cells weighed less and contained fewer hematopoietic colonies than did mice injected with FADD^+/− cells (Fig. 6B–C). These results show that the decreased myelopoiesis of FADD^−/− HSCs observed in vitro is reproducible in vivo.

Figure 6. In vivo colony formation analyses.

(A) Bone marrow cells from FADD^+/− FADD:GFP^flox Mx1-cre and FADD^−/− FADD:GFP^flox Mx1-cre were injected into lethally irradiated B6 mice. Spleens were isolated 8 days later and fixed in Bouin’s solution. (B) Weight of spleen isolated from uninjected control mice, or mice injected with FADD^+/− or FADD^−/− bone marrow cells. (C) Colonies scored from experimental mice following fixation in Bouin’s solution.

In addition to myeloid differentiation, hematopoietic progenitors also develop into cells of the erythroid lineage. We have shown that FADD^−/− bone marrow progenitors are defective in in vitro generation of erythroid precursors (Fig. 5A and D). To study whether FADD functions in in vivo erythrocyte survival and development, we performed hematocrit assays. Experimental mice were bled at multiple time points, both prior to and following deletion, and red blood cell volume was measured. A significant decrease in the volume of red blood cells in collected peripheral blood was seen in FADD^−/− FADD:GFP^flox Mx1-cre mice when compared to FADD^+/− FADD:GFP^flox Mx1-cre mice during the time period leading from the final injection until 6 days post-injection (Fig. 7). These results reveal that FADD is important for the survival and development of erythrocytes in vivo.

Figure 7. Decrease in red blood cells in FADD^−/− mice.

Following administration of poly I:C, mice were retro-orbitally bled. Red blood cell percentage in peripheral blood from FADD^+/− FADD:GFP^flox Mx1-cre and FADD^−/− FADD:GFP^flox Mx1-cre mice was analyzed by centrifuging blood samples and measuring hematocrit volume. * p<0.05.

Although we demonstrated that it is required for the differentiation of non-lymphoid cells from hematopoietic progenitors, no distinct function has been determined for FADD with regard to lymphocyte development. As transplantation of hematopoietic progenitors into lymphopenic hosts initiates their differentiation and repopulation of lymphoid organs, GFP⁻ FADD^+/− and FADD^−/− bone marrow cells were adoptively transferred into sub-lethally irradiated Rag-1^−/− mice to examine whether a lack of FADD leads to defects in lymphopoiesis. Analysis of reconstituted mice revealed a drastic decrease in the cellularity of lymphoid organs in hosts reconstituted with FADD^−/− bone marrow cells (Fig. 8A). Additionally, FADD^−/− bone marrow cells exhibited defective reconstitution of lymphoid progenitors and immature lymphocytes in the primary lymphoid organs, the bone marrow and thymus, as determined by flow cytometric analysis (Fig. 8B and C). This defect extended to the repopulation of peripheral organs such as the spleen and peritoneal cavity with mature T and B lymphocytes when compared to FADD^+/− -injected counterparts (Fig. 8D and E). These results demonstrate that deletion of FADD in bone marrow hematopoietic precursors significantly impairs their ability to differentiate into cells of multiple hematopoietic lineages in vivo, in addition to the previously described in vitro defects.

Figure 8. In vivo lymphoid reconstitution analyses.

(A) Cellularity of lymphoid organs isolated 8–10 weeks following adoptive transfer of bone marrow cells from FADD^+/− FADD:GFP^flox Mx1-cre and FADD^−/− FADD:GFP^flox Mx1-cre into sub-lethally irradiated Rag-1^−/− mice. (B–E) Flow cytometric analysis of lymphocyte populations. Thymocyte populations were analyzed by CD4 and CD8 expression (B), bone marrow by B220 and IgM expression (C), spleens analyzed by immunostaining for CD3 and CD19 (D) and peritoneal cavity cells by staining with antibodies for CD5 and B220 (E). n =3 *, p<0.005.

FADD deficiency results in decreased hematopoietic progenitor pools

Our data thus far have shown that FADD^−/− bone marrow cells are defective in their capability to generate multiple lineages in vitro or in vivo. Yet, no insight was provided as to the mechanism causing this deficiency, as the defect may be due to either a flaw in the perpetuation of hematpoietic progenitor cells or an impairment of their proliferation and differentiation. As indicated in Figure 2A, the absolute number of bone marrow cells was decreased in FADD^−/− FADD:GFP^flox Mx1-cre mutant mice when compared to FADD^+/− FADD:GFP^flox Mx1-cre control mice. FADD^−/− progenitors displayed decreased differentiation and repopulation of the periphery (Fig. 3). To determine whether the deletion of FADD impacted the in vivo maintenance of hematopietic progenitors, we performed multiparameter flow cytometry analyses of bone marrow cells. In FADD^−/− FADD:GFP^flox Mx1-cre mutant mice, the HSC and MPPenriched Lin⁻Sca-1⁺cKit⁺ (LSK) population was significantly reduced, in comparison with FADD^+/− FADD:GFP^flox Mx1-cre control mice (Fig. 9A and B). Additionally, a concomitant decrease was also detected within the Lin⁻c-Kit⁺Sca-1⁻ (LK) population, which contains multipotent progenitor cells (Fig. 9A and B). We further analyzed the multiple lineage progenitor populations present in the bone marrow, including the common lymphoid progenitors (CLP), common myeloid progenitors (CMP), megakaryocyte-erythrocyte progenitors (MEP) and the further differentiated granulocyte-macrophage progenitors (GMP), to determine the effect of FADD deletion on these cells. We found that the bone marrow of FADD^−/− FADD:GFP^flox Mx1- cre mice contained decreased numbers of CLP, CMP, MEP and GMP cells (Fig. 9B). In total, these results demonstrate that FADD is required for bone marrow hematopoietic progenitor cell maintenance.

Figure 9. Analysis of HSCs and hematopoietic progenitors in FADD^−/− mice.

A, four-color flow cytometric analysis of HSCs (Lin⁻Sca-1⁺c-Kit⁺) and progenitors (Lin⁻Sca-1⁻c-Kit⁺) at day 4 after Cre induction. Results are representative of six pairs of control and mutant mice. B, Differentiated hematopoietic progenitors were further analyzed by six-color flow cytometry using sequential gating to study the absolute numbers of LSK, LK, CLP (Lin⁻IL7Rα⁺c-Kit^loSca-1^lo), CMP (Lin⁻ c-Kit⁺Sca-1⁻FcγR^loCD34⁺), MEP (Lin⁻ c-Kit⁺Sca-1⁻FcγR^loCD34⁻), and GMPs (Lin⁻ c-Kit⁺Sca-1⁻FcγR^hiCD34⁺) present in the bone marrow as described (33–34). Four to six pairs of control and mutant mice were analyzed. Horizontal bars indicate the mean values.

Discussion

DRs are initiators of diverse signaling responses, which are involved in the regulation of homeostasis, tumor surveillance, inflammatory responses and adaptive immune responses (5, 63). Previous studies have indicated that Fas, TNF-R1, DR3, and TRAIL-Rs (DR4 and DR5) require FADD for relaying apoptotic signals (6–8, 64–67). Non-apoptotic signals can also be induced by some of the DRs, leading to the activation of nuclear factor kappa B (NF-κB) and mitogen-activated protein kinases (MAPKs). Nonetheless, DRs do not appear to play a significant role in embryos because mice lacking each DR or mice lacking both Fas and TNF-R1 appear to develop normally. Therefore, the embryonic defects in FADD^−/− mice are likely due to the loss of a DR-independent function of FADD (23–24). Although essential in embryos, the function of FADD during hematopoiesis has remained a paradox. Initial analyses showed that FADD^−/− ES cell→Rag-1^−/− blastocyst chimeric mice contained no detectable B cells and few T cells, suggesting a potential function for FADD in lymphopoiesis (23). However, when postlineage commitment deletion of FADD was initiated at the pro-B and pro/pre-T stages using the CD19-cre (68) and Lck-cre (69), respectively, subsequent bone marrow B cell development and thymic T cell development were not affected (26–27). The disparity between these data led us to hypothesize that FADD is required at earlier, pre-lineage commitment stages of hematopoiesis, with a potential function in other hematopoietic developmental pathways as well. In this study, we addressed this issue by inducing the deletion of FADD in bone marrow cells using Mx1-cre. Our data demonstrated that FADD plays a critical role in hematopoietic progenitors.

Previously, Mx1-cre was used to induce the deletion of Mcl-1, an anti-apoptosis protein of the Bcl-2 family, which resulted in severe defects in HSC survival (42). Similarly, efficient deletion of Pten in HSCs was induced using Mx1-cre, revealing a critical role for Pten in HSC maintenance and lineage regulation (48). In our system, injection of poly I:C into mice induced excision of the FADD:GFP^flox transgene in multiple hematopoietic organs in mice containing no or just one endogenous FADD allele (FADD^−/− FADD:GFP^flox Mx1-cre and FADD^+/− FADD:GFP^flox Mx1-cre). Flow cytometric assessment of deletion efficiencies as well as the tracking of the GFP⁻ FADD^−/− mutant cells in mice were facilitated by using the GFP tag fused to FADD (Fig. 1, 2, and 3). Deletion occurs concurrently in both B and T cells in peripheral lymphoid organs (Fig. 2 and 3). Induction of FADD deletion had resulted in dramatic reduction of total cellularities in the bone marrow and thymus (Fig. 2A). The contraction of these primary organs is likely due to defects in hematopoieitic progenitor cells in which FADD deletion was induced by Mx1-cre, as such an impact was not seen when FADD was deleted after lineage commitment using Lck-cre or CD19-cre (26–27). Importantly, GFP⁻ FADD^+/− cells persist in control FADD^+/− FADD:GFP^flox Mx1-cre mice, but the GFP⁻ FADD^−/− population was drastically reduced in FADD^−/− FADD:GFP^flox Mx1-cre mutant mice particularly at 3 months after the last dose of poly IC injection (Fig. 3). This time-dependent depletion of the FADD^−/− population phenotype further implies a deficiency in the replenishment of various hematopoietic compartments by FADD^−/− HSCs and progenitors.

In supporting a role for FADD in early hematopoietic development stages prior to lineage commitment, we demonstrated that bone marrow cells isolated from FADD^−/− mice displayed impaired in vitro generation of macrophages and DCs (Fig. 4). Furthermore, FADD^−/− bone marrow cells had diminished ability to form colonies in vitro of various hematopoietic lineages including CFU-GM, CFU-GEMM, and BFU-E (Fig 5). These observed in vitro deficiencies were recapitulated in vivo, as FADD^−/− bone marrow cells were defective in the generation of hematopoietic cells in vivo, following adoptive transfer into immunodeficient hosts (Fig. 6 and 8). Defects in erythroid progenitor differentiation as detected by in vitro assays may lead to reduced peripheral red blood cell numbers, as indicated by a deminished hematocrit in FADD^−/− FADD:GFP^flox Mx1-cre mutant mice (Fig. 7). Therefore, these data help establish a function for FADD in hematopoietic progenitor cells.

HSCs are characterized by their ability to undergo asymmetric division, with one daughter cell maintaining the characteristics of the HSC and the other committing to differentiation (70). As such, both maintenance and differentiation of the HSC and progenitor populations are required for successful development and maintenance of the hematopoietic system. A 4-color flow cytometric analysis of bone marrow cells revealed a dramatic reduction in the HSC-enriched LSK population in FADD mutant mice (Fig. 9). In addition, FADD deletion also resulted in a greatly diminished early progenitor population (LK). Further multicolor flow cytometric analyses demonstrated that the total numbers of various lineage progenitors, CLPs, CMPs, MEPs and further differentiated GMPs, were also dramatically reduced in FADD^−/− mutant mice (Fig. 9B). The reduction in both the HSC and progenitor populations indicates a requirement for FADD in the maintenance of early hematopoietic precursors.

FADD mediates its function during apoptotic signaling through the recruitment of caspase 8 and c-FLIP to a signaling complex, initiating a caspase cascade (6–8). FADD, caspase 8, and c-FLIP knockout mice may have embryonic heart defects (23–25, 71). However, disruption of hematopoiesis may also contribute to the embryonic lethality in mice lacking FADD, caspase 8 or c-FLIP. Both FADD^−/− ES cell→Rag-1^−/− blastocyst and c-FLIP^−/− ES cell→Rag-1^−/− blastocyst chimeras displayed defective T and B lymphopoeisis, and caspase 8^−/− embryos contained depleted hematopoietic precursor populations, implicating these proteins as integral regulators of hematopoiesis (23, 25, 72). Surprisingly however, deletion of these proteins at the pro-T or pro-B cell stages of development did not inhibit the development of T and B lymphoid precursors within the thymus and bone marrow, respectively, bringing to light that the requisite need for these proteins is found at earlier stages of hematopoiesis prior to lineage commitment (26–30, 73–74). Our results detailed in this study provide the first evidence that FADD deficiency impairs the maintenance of hematopoietic progenitors. Interestingly, inducible deletion of caspase 8 led to defective differentiation of hematopoietic progenitors, without affecting the maintenance of HSCs and progenitors (43). The nature of non-apoptotic signaling pathways mediated by FADD, caspase 8, and c-FLIP has been investigated vigorously. In particular, recent studies have implicated a role for FADD and caspase 8 in the regulation of necrosis and/or autophagy in T cells (75–76). Furthermore, there is evidence supporting a role for these three proteins in the regulation of the NF-κB pathways (77–79). In summary, the DISC proteins constitute a signaling axis with diverse roles mediating a variety of pathways that function in a cell-specific manner during embryogenesis, hematopoiesis, and immune responses.