Dominant negative FADD dissipates the proapoptotic signalosome of the unfolded protein response in diabetic embryopathy

Animals and reagents.

The procedures for animal use were approved by the University of Maryland School of Medicine Institutional Animal Care and Use Committee. Wild-type (WT) C57BL/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME). FADD-DN overexpression in C57BL/6J background mice were described previously (26), in which the human FADD-DN transgene was driven by the β-actin promoter.

Mouse models of diabetic embryopathy.

Our mouse model of diabetic embryopathy has been described previously (16, 29–31, 37, 38). Briefly, 8-wk-old WT female mice were injected intravenously daily with 75 mg/kg streptozotocin (STZ; Sigma, St. Louis, MO), dissolved in sterile 0.1 M citrate buffer (pH 4.5), over 2 days to induce diabetes. Using STZ to induce diabetes is not a complicating factor because STZ is cleared from the bloodstream rapidly (STZ serum half-life is 15 min), and pregnancy is not established until 1–2 wk after STZ injection. Diabetes was defined as a 12-h fasting blood glucose level of ≥14 mM. After diabetes in the mice was established, males and females were paired at 3 PM, and day 0.5 of pregnancy (E0.5) was established by the presence of the vaginal plug the next morning (8 AM). WT female mice were treated with vehicle injections to serve as nondiabetic controls. At 6 PM on E8.75, mice were euthanized and conceptuses dissected out of the uteri, and the embryos, taken from the decidua and without the yolk sacs, were used for analyses.

Overall, the use of spontaneous diabetic mouse models in research of diabetic complications is premature (14). There are two spontaneous diabetic mouse models, the AKITA mice and the NOD mice. AKITA mice possess gene mutations, which may affect pregnancy outcomes and add further genetic complexities in our genetically modified mice (14). The NOD mice are in a rare type of genetic background that is not comparable with the C57BL/6J background used in our study. The onset of diabetes in AKITA mice is very early at 3–4 wk of age and is associated with ER stress (14). For these reasons, the STZ-induced diabetic model used in a relatively short window (∼5 wk) is ideal for research of diabetic embryopathy.

Cell culture, transfection, glucose, and tunicamycin treatment.

C17.2 mouse neural stem cells, originally obtained from the European Collection of Cell Culture (Salisbury, UK), were maintained in DMEM (5 mM glucose) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO₂. TRADD plasmid (Myc-DDK-tagged, MR223646) was obtained from Santa Cruz Biotechnology (Dallas, TX). The pCDNA3.1-GFP-FADD-DN construct was made by cloning the cDNA coding for human FADD-DN (residues 80–208) into the pcDNA3.1/GFP vector (Invitrogen, Grand Island, NY). Lipofectamine 2000 (Invitrogen) was used according to the manufacturer's protocol for transfection of plasmid into the cells. After the cells were seeded overnight, C17.2 mouse neural stem cells were transfected with TRADD and FADD-DN plasmid and then cultured for 24 h. Cells transfected with TRADD plasmid were harvested directly, and cells transfected with FADD-DN plasmid were treated with high glucose for another 48 h before being harvested and used in assays. To assess the effect of FADD-DN on TRADD-induced ER stress, cells were cotransfected with TRADD and FADD-DN plasmid and cultured for 24 h before being assayed. To determine the effect of 4-phenylbutyric acid (4-PBA) on TRADD-induced apoptosis, cells were changed to DMEM medium with 2 mM 4-PBA at 6 h after transfection with TRADD plasmid. After 48 h of treatment with 4-PBA, cells were assayed and analyzed. Tunicamycin (T7765, 4 μg/ml; Sigma) or DMSO (final concentration is 0.04% DMSO volume/total medium volume) as control was added to growth medium for another 6 h, and then cells were harvested.

E14 mouse embryonic stem cells were obtained from the American Type Culture Collection (Manassas, VA) and then seeded onto a culture plate with a layer of STO feeder cells maintained in embryonic stem cell culture medium [DMEM (25 mM glucose) supplemented with 10% FBS, 1% NEAA, 1% β-mercaptoethanol, 2 mM l-glutamine, 1,000 U/ml mLIF, 100 U/ml penicillin, and 100 μg/ml streptomycin] at 37°C in a humidified atmosphere of 5% CO₂. At second passage, E14 cells were transferred to feeder cell-free plates and maintained in embryonic stem cell culture medium supplemented with 15% FBS. After two passages, feeder cell-free E14 cells were changed to and maintained in low glucose embryonic stem cell culture medium (5 mM glucose DMEM) for at least eight passages. Plasmid transfection and high glucose treatment were the same as described previously (31).

Immunoprecipitation and immunoblotting.

Immunoprecipitation (IP) and immunoblotting were performed as described previously (16, 30, 31). Protease inhibitor cocktail (Sigma), lysis buffer (Cell Signaling Technology), protein A magnetic bead slurry (New England BioLabs, Ipswich, MA), and rabbit anti-ASK1 antibody (Santa Cruz Biotechnology) were used for IP. For immunoblotting, equal amounts of protein and the Precision Plus protein standards (Bio-Rad, Hercules, CA) were resolved by sodium dodecyl sulfate-polyacryamide gel electrophoresis and transferred onto Immunobilon-P membranes (Millipore, Billerica, MA). Membranes were incubated in 5% nonfat milk for 45 min and then incubated for 18 h at 4°C with the following primary antibodies at 1:1,000 to 1:2,000 dilutions in 5% nonfat milk: caspase 8, caspase 3 (Millipore), TRADD (Santa Cruz Biotechnology), human FADD (Millipore), phospho (p)-ASK1, p-IRE1α, TRAF2, p-PERK (protein kinase R-like endoplasmic reticulum kinase), p-eIF2α (eukaryotic initiation factor 2), BiP, CCAAT/enhancer-binding protein (C/EBP) homologous protein (CHOP), p-JNK1/2, p-c-Jun, p-ATF2 (activating transcription factor 2), p-Elk1, p-FoxO3a, tbid, p-Bad, Bax, Puma, Bak, Bim, cleaved caspase 3 (Cell Signaling Technology, Beverly, MA), and GFP (Torrey Pines Biolabs, Secaucus, NJ). Membranes were exposed to goat anti-rabbit or anti-mouse secondary antibodies. To ensure that equivalent amounts of protein were loaded among samples, membranes were stripped and probed with a mouse antibody against β-actin (Abcam, Cambridge, MA) and prohibitin (Millipore). Signals were detected using the SuperSignal West Femto Maximum Sensitivity Substrate kit (Thermo Scientific, Waltham, MA). All experiments were repeated three times with the use of independently prepared tissue lysates.

RNA extraction and real-time PCR.

Total RNA was isolated from embryos using an RNeasy Mini Kit (Qiagen, Frederick, MD) and reverse transcribed by using the high-capacity cDNA archive kit (Applied Biosystems, Life Technologies, Grand Island, NY). MicroRNA was extracted by using the mirVana miRNA isolation kit (Ambion, Life Technologies) and reverse transcribed using NCode VILO miRNA cDNA synthesis kit (Invitrogen, Life Technologies). RT-PCR for ER stress chaperone genes (calnexin, BiP, CHOP, IRE1α, PDIA, and GRP94), β-actin, mmu-mir-17, mmu-mir-34a, mmu-mir-96, and mmu-mir-U6 was performed using Maxima SYBR Green/ROX qPCR Master Mix assay (Thermo Scientific). RT-PCR and subsequent calculations were performed by a 7700 ABI PRISM sequence detector system (Applied Biosystem).

TUNEL assay.

The TUNEL assay was performed as described previously by using the In Situ Cell Death Detection Kit (Millipore) (30). Cells were seeded on the eight-well Nunc Lab-Tek Chamber Slide system (Sigma). After transfection and 4-PBA or high glucose treatment, cells were fixed with 4% paraformaldehyde in PBS and incubated with TUNEL reagent, counterstained with DAPI, and mounted with aqueous mounting medium (Sigma). TUNEL-positive cells in each well were counted. The percentage of apoptotic cells was calculated as the number of TUNEL-positive (apoptotic) cells divided by the total number of cells multiplied by 100.

Statistical analysis.

Data are presented as means ± SE. Three embryos from three different dams were used for immunoblotting, and five embryos for five different dams were studied for RT-PCR. One-way ANOVA was performed using SigmaStat 3.5 software, and a Tukey test was used to estimate the significance. Chi-square test was used to estimate the significance of NTD formation among different groups. Statistical significance was accepted when P < 0.05.

FADD-DN ameliorates maternal diabetes-induced caspase activation, apoptosis, and NTD formation.

Maternal diabetes induces caspase 8-dependent apoptosis in the neuroepithelium of the developing embryo (30, 33, 35, 39). Caspase 8 is an initiator caspase whose cleavage or activation leads to the activation of the executive caspase 3. Indeed, the cleavage of caspase 3 and caspase 8 was observed in embryos exposed to maternal diabetes (Fig. 1A). FADD-DN overexpression abolished maternal diabetes-induced caspase 3 and caspase 8 cleavage and TRADD upregulation (Fig. 1A). Consistent with caspase activation, the number of apoptotic cells was significantly increased in the neuroepithelia of embryos exposed to maternal diabetes compared with that in the neuroepithelia of embryos from nondiabetic dams with or without FADD-DN overexpression (Fig. 1B). FADD-DN overexpression suppressed maternal diabetes-induced neuroepithelial cell apoptosis (Fig. 1B) and NTD formation (Table 1 and Fig. 1C). These findings support the hypothesis that FADD-DN blocks maternal diabetes-induced neuroepithelial cell apoptosis, leading to reduction of diabetic embryopathy.

FADD-DN suppresses ER stress by disrupting the interaction of TRADD-FADD with the IRE1α-TRAF2-ASK1 complex.

Maternal diabetes activates IRE1α (17, 30), which forms a complex with TRAF2 and ASK1 (18, 31). Because TRADD interacts with TRAF2 (12), we sought to determine whether the maternal diabetes-induced IRE1α signalosome, the IRE1α-TRAF2-ASK1 complex, interacts with TRADD and FADD and whether FADD-DN disrupts the interaction of IRE1α signalosome with TRADD. ASK1 immunoprecipitation using a rabbit anti-ASK1 antibody pulled down p-IRE1α, p-ASK1,TRAF2, TRADD, and FADD in embryonic lysates from nondiabetic or diabetic dams (Fig. 2A), whereas control (normal rabbit IgG) did not pull down any of these proteins (Fig. 2A), indicating that p-IRE1α forms a complex with TRAF2, p-ASK1, TRADD, and FADD. The abundance of p-IRE1α, TRAF2, p-ASK1, TRADD, and FADD proteins in the ASK1 immunoprecipitates was significantly higher in embryos from diabetic dams than in embryos of nondiabetic dams (Fig. 2A). FADD-DN overexpression diminished the increased presence of p-IRE1α, TRAF2, p-ASK1, TRADD, and FADD in ASK1 immunoprecipitates under diabetic conditions (Fig. 2A).

To test whether the disruption of the IRE1α signalosome by FADD-DN overexpression has any impact on the UPR pathway, whose prolonged activation leads to ER stress, levels of p-IRE1α, p-PERK, and p-eIF2α were measured. FADD-DN overexpression suppressed maternal diabetes-induced phosphorylation of IRE1α, PERK, and eIF2α (Fig. 2, B–D). Because the two UPR sensors IRE1α and PERK were deactivated by FADD-DN, we predicted that ER stress was relieved. The levels of ER stress markers CHOP and BiP and ER chaperone gene expression in embryos of diabetic dams that overexpressed FADD-DN were reduced to levels comparable with those in embryos of nondiabetic dams (Fig. 3, A–C). Another index of ER stress, X-box-binding protein-1 mRNA splicing, was also diminished by FADD-DN overexpression (Fig. 3D). These results demonstrate that FADD-DN disrupts the assembly of the IRE1α signalosome, leading to blockage of ER stress.

FADD-DN represses the proapoptotic ASK1-JNK1/2 pathway in the developing embryo.

One of the functional readouts of the IRE1α signalosome is the activation of the proapoptotic ASK1-JNK1/2 pathway. Diabetes-induced phosphorylation of ASK1 and JNK1/2 was blocked by FADD-DN overexpression (Fig. 4, A and B). Furthermore, the phosphorylation of three transcription factors, c-Jun (Fig. 4C), Elk1 (Fig. 4D), and ATF-2 (Fig. 4E), that are downstream effectors of JNK1/2 activation was blunted by FADD-DN overexpression. FoxO3a is also downstream of JNK1/2, and its activation relies on dephosphorylation. Maternal diabetes induced FoxO3a dephosphorylation, and FADD-DN overexpression restored FoxO3a phosphorylation (or inactivation; Fig. 4F). These findings reveal the inhibitory role of FADD-DN in the ASK1-JNK1/2 pathway in diabetic embryopathy.

FADD-DN rescues mitochondrial function altered by maternal diabetes.

TRADD-FADD interacts with the IRE1α signalosome, which activates the proapoptotic ASK1-JNK1/2 pathway. The ASK1-JNK1/2 pathway induces apoptosis through the intrinsic mitochondrial pathway. Mitochondrial dysfunction was manifested in diabetic embryopathy (34). To reveal whether FADD-DN overexpression affects mitochondrial dysfunction induced by maternal diabetes, a set of mitochondrial dysfunction markers was assessed. The activation and mitochondrial translocation of the proapoptotic members of the Bcl-2 family reflect mitochondrial dysfunction. Maternal diabetes triggered Bad dephosphorylation (activation) and Bid cleavage, whereas FADD-DN overexpression prevented these changes (Fig. 5A). Maternal diabetes-induced Bax, Bak, and Puma, and Bim mitochondrial translocation was also blocked by FADD-DN overexpression (Fig. 5B), indicating that FADD-DN restores mitochondrial function that is suppressed by maternal diabetes.

TRADD mimics high glucose to induce UPR and ER stress.

Because TRADD participates in the IRE1α signalosome, we tested whether TRADD overexpression is sufficient to induce the UPR and ER stress. Ectopic TRADD overexpression in neural stem cells activated the two UPR arms by inducing phosphorylation of IRE1α, PERK, and eIF2α (Fig. 6, A and B). TRADD overexpression also upregulated the ER stress markers CHOP and BiP (Fig. 6B). The activation of key components of the IRE1α signalosome, p-ASK1 and p-JNK1/2, was induced by TRADD overexpression (Fig. 6C). Levels of miR17, miR34a, and miR96, indices of the IRE1α endonuclease activities (31), were significantly downregulated by TRADD overexpression (Fig. 6D). UPR activation and ER stress induction by TRADD were not secondary effects of apoptosis because they occurred before TRADD-induced apoptosis. TRADD-induced UPR and ER stress took place as early as 24 h after TRADD transfections (Fig. 6, B–D), whereas the appearance of apoptotic cells, cleaved caspase 3, and cleaved caspase 8 were observed after 36 h onward (Fig. 6, E–G). The ER stress inhibitor 4-PBA (17, 31) significantly reduced TRADD-induced apoptosis (Fig. 6, E–G), suggesting that ER stress is essential for TRADD-induced apoptosis. The above results suggest that TRADD stimulates UPR and ER stress, leading to apoptosis.

FADD-DN blocks high glucose- or TRADD-induced UPR and ER stress in neural stem cells.

The neuroepithelium during embryonic neurulation contains essentially neural stem cells, and maternal diabetes imposes an adverse effect on the developing neuroepithelium by inducing neural stem cell apoptosis (28, 34–36). The C17.2 neural stem cell line was derived from neonatal cerebellum and immortalized by retroviral transduction with v-myc (25). We have demonstrated that the C17.2 cell line mimics the in vivo diabetic embryopathy model in ER stress and apoptosis induction when exposed to high glucose (REF). Therefore, the C17.2 cell line was chosen for our in vitro experimentations. We used FADD-DN to determine whether the TRADD functional output through FADD mediates the stimulatory effect of high glucose on the UPR and ER stress. FADD-DN suppressed high glucose-induced IRE1α, PERK, and eIF2α phosphorylation (Fig. 7A), diminished the increase of ER stress markers CHOP and BiP (Fig. 7A), terminated the activation of ASK1 and JNK1/2 by high glucose (Fig. 7B), blunted the decrease of miRNAs downstream of IRE1α (Fig. 7C), and blocked caspase activation and apoptosis (Fig. 7, A and D). Likewise, FADD-DN abolished TRADD-induced IRE1α, PERK, and eIF2α phosphorylation (Fig. 7E) and upregulation of CHOP and BiP (Fig. 7E). These results indicate that FADD-DN is able to suppress TRADD-induced ER stress.

We further examined the effect of FADD-DN on tunicamycin-induced UPR and ER stress in C17.2 cells. FADD-DN blocked tunicamycin-induced IRE1α phosphorylation and CHOP upregulation and attenuated PERK phosphorylation and BiP upregulation (Fig. 7F). These findings demonstrate that FADD-DN blocks TRADD-, high glucose-, or tunicamycin-induced ER stress.

Both TRADD and high glucose induce the UPR and ER stress in embryonic stem cells.

Although the findings in the C17.2 cell line were supportive of in vivo results, the C17.2 cell line may not truly reflect the cell biology of neural stem cells in the developing neuroepithelium. The C17.2 cell line is developmentally more differentiated than neural stem cells in the neuroepithelium of the neurulating embryo (25). Additionally, maternal diabetes adversely impacts the entire period of embryogenesis, including the stages of embryonic stem cells. Thus we chose E14 embryonic stem cells to recapitulate the finding that TRADD triggers UPR and ER stress in neural stem cells. We examined the effect of TRADD overexpression in an embryonic stem cell line. TRADD overexpression activated the two UPR sensors IRE1α and PERK and increased the ER stress markers CHOP and BiP (Fig. 8A).

Maternal diabetes adversely impacts the entire period of embryonic development and thus may also affect the function of embryonic stem cells. High glucose induced the phosphorylation of IRE1α and PERK and upregulated ER stress markers in embryonic stem cells. FADD-DN abrogated high glucose-induced UPR and ER stress (Fig. 8B).