Oxidative Stress–Induced JNK1/2 Activation Triggers Proapoptotic Signaling and Apoptosis That Leads to Diabetic Embryopathy

C57BL/6J mice (median body weight 22 g) were purchased from The Jackson Laboratory (Bar Harbor, ME). Sustained-release insulin pellets (Linplant) were purchased from LinShin Canada. Streptozotocin (STZ) from Sigma-Aldrich was dissolved in sterile 0.1 mol/L citrate buffer (pH 4.5).

The procedures for animal use were approved by the institutional animal care and use committee.

For decades, we (20,29,30) and others (31–36) have used an internationally accepted rodent model of STZ-induced diabetes in research of diabetic embryopathy. Ten-week-old (median body weight 22 g) jnk1 heterozygous (jnk1^+/−), jnk2 homozygous (jnk2^−/−), and wild-type (WT) female mice in C57BL/6J background (The Jackson Laboratory) were intravenously injected daily with 75 mg/kg STZ during a 2-day period to induce diabetes. Hemizygous SOD1-Tg mice in C57BL/6J background were revived from frozen embryos by The Jackson Laboratory, and a subsequent colony of SOD1 mice was established in the laboratory. The transgene in the SOD1-Tg mice is the full-length human SOD1 gene (hSOD1) (37). To generate SOD1 embryos, we crossed SOD1-Tg male mice with diabetic WT female mice. Nondiabetic WT and jnk1^+/− female mice with vehicle injection served as nondiabetic controls and were mated with respective genotypes of male mice. Blood glucose levels were monitored daily by tail vein puncture and using the FreeStyle Blood Glucose Monitoring System (TheraSense; Abbott Diabetes Care, Alameda, CA). Diabetes was defined as 12-h fasting blood glucose levels ≥250 mg/dL, which normally occurred at 3–5 days after STZ injections. To successfully establish pregnancies, insulin pellets were subcutaneously implanted in diabetic mice to restore euglycemia (80–150 mg/dL) prior to mating. WT, jnk1^+/−, or jnk2^−/− female mice were mated with male mice of the same respective genotype. Embryos from nondiabetic WT mice and from diabetic WT mice crossed with SOD1-Tg male mice and jnk2^−/− embryos from jnk2^−/− mice were harvested at embryonic day (E)8.75 (2:00 p.m. at E8.5) for biochemical and molecular analysis. jnk1^+/− female diabetic mice mated with jnk1^+/− male mice produced WT (jnk1^+/+), jnk1^+/−, and jnk1^−/− embryos under the same maternal diabetic conditions for malformation assessment. STZ has a very short (30 min) half-life, and pregnancies in our study were established after 1–2 weeks of STZ injections. Our previous study using the STZ/insulin-administered and the non–STZ-administered animals in the mouse model (15) and in the rat model (7,38) did not detect any difference in embryonic development. Therefore, there is no residual toxic effect of STZ in our animal model.

On the basis of a published report (39), male and female mice were paired at 3:00 p.m., and day 0.5 (E0.5) of pregnancy was established by the presence of the vaginal plug the next morning (8:00 a.m.). On day 5.5 of pregnancy (E5.5), insulin pellets were removed to permit frank hyperglycemia (>250 mg/dL glucose level) so the developing embryos would be exposed to a hyperglycemic environment during the critical period of closure of the neural tube (neurulation) (E8–E10.5). Therefore, our mouse model of diabetic embryopathy specifically impacts neurulation and induces NTDs. On the basis of our extensive studies (15,20), insulin treatment from E0.5 to E5.5 is essential for successful implantation establishment and, thus, prevents early embryonic lethality (resorption) caused by hyperglycemic exposure at early embryonic stages (≤E5.5). Control nondiabetic pregnant mice were sham operated on for insulin implantation and removal. E10.5 embryos were examined for malformations. Normal embryos were classified as possessing a completely closed neural tube and no evidence of other malformations. Malformed embryos were classified as showing evidence of failed closure of the anterior or posterior neural tubes resulting in anencephaly, a major type of NTD. Even at E15.5, we still observed NTD embryos with no sign of delayed embryonic development (Fig. 2D and E). At late embryonic stages, anencephaly leads to the absence of skull and, consequently, an uncovered brain (Fig. 2E). Embryos were first examined for the open neural tube structure under a Leica MZ16F stereomicroscope (Bannockburn, IL) to identify NTDs. Images of embryos were captured by a DFC420 5 MPix digital camera with software (Leica). In Fig. 2A–C, the open neural tube structure in embryos with NTDs was verified by series sections through the anterior or posterior neural tubes, and the open neural tube structure at the ventral side was shown. Because our model specifically induces NTDs at E10.5, other major structural malformations, such as cardiovascular defects, which are not formed until late embryonic stages (E15.5), were not examined.

We genotyped embryos from jnk1^+/− female mated with jnk1^+/− male mice, and embryos from diabetic WT mice crossed with SOD1-Tg male mice, by using the yolk sac DNA according to The Jackson Laboratory’s protocol.

Western blotting was performed as previously described (20). The following antibodies were used: phospho (p)-JNK1/2, SOD1, JNK2, JNK1, p-E twenty-six–like transcription factor 1 (p-Elk1) (Ser383), p-c-Jun (Ser63), p-ATF-2 (Thr71), Sox1, and p-Foxo3a (Thr32) (Cell Signaling, Beverly, MA), anti-caspase 3 (Chemicon, Billerica, MA), anti-caspase 8 (Alexis Biochemicals, San Diego, CA), and anti–β-actin (Abcam, Cambridge, MA).

Frozen 10 μm-thick embryonic sections through the anterior neural tube were fixed with 4% paraformaldehyde in PBS and incubated with transferase-mediated dUTP nick-end labeling (TUNEL) reaction agents (In Situ Cell Death Detection kit; Roche, Indianapolis, IN) with or without Sox1 immunostaining. TUNEL⁺ cells in the neural tube area of each section were counted. On the basis of total cell nuclei counts, the sizes of the neural tube areas of each section were relatively constant. Thus, apoptotic cell numbers were expressed as total TUNEL⁺ cells per neural tube area per section. TUNEL assays were performed without the observer knowing the experimental group of the embryo examined.

Densitometric data were presented as means ± SE. One-way ANOVA was performed using SigmaStat 3.5 software. In ANOVA analysis, Tukey test was used to estimate the significance of the results. Significant difference between groups in malformation incidences was analyzed by χ² test. Statistical significance was accepted at P < 0.05.

We hypothesized that oxidative stress is responsible for diabetes-induced JNK1/2 activation and its downstream signaling in embryonic tissues. To test this hypothesis, E8.75 embryos were harvested from nondiabetic WT and diabetic WT mice that were mated with SOD1-Tg male mice to avoid any potential maternal influence by the SOD1 transgene. E8.75 embryos are suitable models to study the impact of diabetes on embryonic neural tube closure. This is based on the fact that the E8.75 is a key stage of neurulation, and the major part of E8.75 embryos is the developing neural tube. p-JNK1/2 levels in WT embryos exposed to maternal hyperglycemia were significantly higher than those in WT embryos from nondiabetic WT mice (Fig. 1A and B). SOD1-overexpressing embryos expressed high levels of hSOD1 protein (Fig. 1A and C). SOD1-overexpressing embryos had significantly lower levels of p-JNK1/2 and their downstream effectors, p-c-Jun, p-ATF-2, and p-Elk1, compared with those in WT embryos from the same group of diabetic WT mice mated with SOD1-Tg male mice (Fig. 1A–D). Levels of p-JNK1/2, p-c-Jun, p-ATF-2, and p-Elk1 in SOD1-overexpressing embryos exposed to maternal hyperglycemia did not differ from those in WT embryos from nondiabetic WT mice (Fig. 1A–D). Maternal hyperglycemia-suppressed p-Foxo3a levels, indices of Foxo3a inactivation, were restored in SOD1-overexpressing embryos (Fig. 1C and D). Under nondiabetic conditions, SOD1 overexpression did not affect the basal levels of p-JNK1/2 (Fig. 1E).

Western blotting detects two bands of p-JNK1/2: p54JNK and p46JNK. Both p54JNK and p46JNK are highly phosphorylated in embryos exposed to maternal diabetes. Since the jnk1 gene contributes to both p54JNK and p46JNK, we proposed that JNK1 activation plays a role in the induction of diabetic embryopathy. jnk1^+/− mice were used to test this hypothesis because jnk1^−/− mice are poor breeders. To generate jnk1^−/− embryos, diabetic jnk1^+/− female mice were mated with jnk1^+/− male mice. The mating of jnk1^+/− male and female mice produced WT, jnk1^+/−, and jnk1^−/− embryos under the same maternal diabetic conditions (Fig. 2). This mating scheme resulted in effective comparison in malformation rates of these three genotype embryos while it minimized the maternal influence. Average blood glucose levels in jnk1^+/− mothers at E8.5–E10.5, a critical period of organogenesis, were 371.8 ± 27.2 (SE) mg/dL. E10.5 embryos from 11 jnk1^+/− mothers were examined for malformations (Table 1). As described above, malformed embryos were classified as showing evidence of failed closure of the anterior or posterior neural tubes, namely, NTDs. Other defects, such as heart defects, were not examined because they develop at late embryonic stages (E15.5). In total, 80 embryos were harvested, and the ratios of WT, jnk1^+/−, and jnk1^−/− embryos were close to the expected ratio, 1:2:1 (Table 1). There was a total of 22 jnk1^−/− embryos, with 4 of them having NTDs. The malformation rate of jnk1^−/− embryos was 18.2% (Table 1). Of 40 jnk1^+/− embryos harvested, 22 of them had NTDs (Table 1). The malformation rate of jnk1^+/− embryos was 55.0% (Table 1). A total of 10 out of 18 WT embryos exhibited NTDs, and the malformation rate of WT embryos was 55.6%, which was comparable with that in jnk1^+/− embryos (Table 1). A representative litter from a diabetic jnk1^+/− female mated with a jnk1^+/− male showed NTDs in WT (Fig. 2C) and jnk1^+/− (Fig. 2B) embryos, whereas none of the four jnk1^−/− embryos in the same litter had NTDs (Fig. 2A). The NTDs observed were typically anencephaly (Fig. 2). In agreement with our previous study (15,20), we still observed NTD embryos at E15.5 (Fig. 2D and E) with no sign of delayed development. Consistent with the previous observation that targeted deletion of jnk1 gene does not affect embryonic development, no NTDs were observed in WT, jnk1^+/−, and jnk1^−/− embryos from nondiabetic jnk1^+/− mice injected with saline (Table 2).

Because both JNK1 (Table 1) and JNK2 deficiency (20) reduce maternal diabetes-induced NTDs, we propose that JNK1 and JNK2 equally contribute to the activation of JNK1/2 downstream effectors in diabetic embryopathy. To this end, E8.75 jnk2^−/− embryos from JNK2KO mice and jnk1^−/− embryos from jnk1^+/− females mated with jnk1^+/− males were used to detect the phosphorylated levels of critical JNK1/2 downstream effectors c-Jun, ATF-2, Elk1, and Foxo3a. The activation of these transcription factors is critical for the induction of apoptotic genes. Levels of p-c-Jun, p-ATF-2, and p-Elk1 in WT embryos exposed to maternal hyperglycemia were significantly higher than those in embryos from nondiabetic WT mice (Fig. 3A and 4A). This finding suggests that hyperglycemia activates c-Jun, ATF-2, and Elk1. Hyperglycemia significantly reduced p-Foxo3a levels in WT embryos exposed to maternal hyperglycemia (Fig. 3B and 4B), demonstrating that Foxo3a activities are induced by hyperglycemia because dephosphorylation of Foxo3a correlates with its transcriptional activities. Levels of p-c-Jun, p-ATF-2, and p-Elk1 in jnk2^−/− embryos exposed to maternal hyperglycemia were significantly lower than those in WT embryos from diabetic mice and were comparable with those in embryos from nondiabetic mice (Fig. 3A and C–E). In a similar manner, jnk1^−/− embryos exposed to hyperglycemia had significantly lower levels of p-c-Jun, p-ATF-2, and p-Elk1 than those in WT embryos from the same group of diabetic jnk1^+/− mothers (Fig. 4A and C–E). Levels of p-FoxO3a in jnk2^−/− embryos exposed to hyperglycemia were significantly higher than those in embryos from diabetic WT mice and were comparable with those in embryos from nondiabetic mice (Fig. 3B and F). jnk1 gene deletion in jnk1^−/− embryos prevented maternal hyperglycemia-induced p-Foxo3a reduction (activation) (Fig. 4B and F).

Caspase activation is manifested in diabetic embryopathy. However, the mechanism whereby hyperglycemia induces caspase activation is not clear. The JNK1/2 pathway is a proapoptotic signaling pathway that mediates caspase activation in other systems. We, therefore, reasoned that JNK1/2 activation leads to caspase activation in diabetic embryopathy. Since our previous studies (15,17) define the initiator caspase, caspase 8, and the effector caspase, caspase 3, in diabetic embryopathy, we determined levels of cleaved (active) caspase 3 and 8 to test this hypothesis. Cleaved caspase 3 and 8 were robustly present in WT embryos exposed to hyperglycemia of diabetes but were not present at all in WT embryos from nondiabetic mice (Fig. 5A and B). Similar to observations in WT embryos from nondiabetic mice, we observed no cleaved product of caspase 3 and 8 in jnk2^−/− embryos exposed to hyperglycemia (Fig. 5A). Neither were cleaved caspase 3 and 8 detected in jnk1^−/− embryos exposed to hyperglycemia (Fig. 5B). In contrast, WT embryos, which were from the same group of diabetic jnk1^+/− mothers as the jnk1^−/− embryos, had robust cleaved caspase 3 and 8 (Fig. 5B). These results support that both JNK1 and JNK2 activation lead to caspase cascade activation.

Hyperglycemia-induced apoptosis in the developing neural tube causes NTDs. Since JNK1/2 activation mediates hyperglycemia-induced caspase cascade activation, which is a biochemical hallmark of apoptosis, it is reasoned that JNK1/2 activation is a proapoptotic signal that is responsible for diabetes-induced apoptosis in the developing neural tube. To determine whether JNK1/2 deficiency blocks maternal diabetes-induced apoptosis in the developing neural tube, we used the TUNEL assay to assess apoptosis. We detected fewer apoptotic cells in WT embryos from nondiabetic mice and jnk2^−/− and jnk1^−/− embryos from diabetic mice (Fig. 6A and B). WT embryos from diabetic mice exhibited significantly higher numbers of apoptotic cells in the developing neural tube compared with those observed in WT embryos from nondiabetic mice and jnk2^−/− and jnk1^−/− embryos from diabetic mice (Fig. 6A and B). Although double TUNEL/Sox1 staining was not as sensitive as the TUNEL assay alone, it detected the apoptotic cells at the dorsal side of the neural tube as Sox1⁺ neural progenitors (Fig. 6C and D).

Using genetically modified mice, our studies provide strong evidence that hyperglycemia-induced oxidative stress activates the proapoptotic JNK1/2 signaling, which triggers apoptosis in the developing neural tube resulting in NTDs. In a separate study (submitted for publication), we demonstrated that SOD1 overexpression inhibits maternal diabetes-induced caspase activation and apoptosis. SOD1-overexpressing (29), jnk2^−/− (20), and jnk1^−/− embryos (Table 1) are all resistant to diabetes-induced embryonic malformations, demonstrating that oxidative stress and JNK1/2 activation are key steps in the induction of diabetic embryopathy. JNK1/2 activation mediates the downstream events of oxidative stress and plays a causative role in the induction of diabetic embryopathy. In this study, we used a mouse diabetic model in C57BL/6J background to induce sustained maternal hyperglycemia (≥250 mg/dL glucose) that mimics the human condition seen in type 1 and type 2 maternal diabetes. In our studies (15,20) and those of others (32,33), this mouse model in C57BL/6J background consistently produces >22% NTDs in embryos exposed to hyperglycemia. In contrast, embryos from nondiabetic control mice display ∼1% NTDs. This rate of spontaneous NTD formation is lower than previously reported (35). The high NTD rate in our study is probably due to the calculation based on the gross NTD rate per genotype. Under the same maternal diabetic environment, it is also possible that reduced susceptibility of jnk1^−/− embryos increases NTD incidences of WT and jnk1^+/− embryos. Nevertheless, another group has reported comparable malformation rates in diabetic mice (32).

Using jnk1^+/− male and female mice, we produced jnk1^−/− embryos, which, when exposed to diabetes, had lower malformation rates than those in jnk1^+/− or jnk1^+/+ (WT) embryos. These findings clearly show that JNK1 mediates the teratogenicity of maternal diabetes. Because double deletion of jnk1 and jnk2 genes is lethal, our initial plan was to generate jnk2^−/−jnk1^−/+ or jnk2^+/−jnk1^−/− mice and, thus, produce jnk2^−/−jnk1^−/+ or jnk2^+/−jnk1^−/− embryos to study the synergistic effects of JNK1 and JNK2 on diabetic embryopathy. Unfortunately, jnk2^−/−jnk1^−/+ or jnk2^+/−jnk1^−/− mice are poor breeders, and no successful pregnancies were established from these matings. Nevertheless, we found that both jnk1^−/− and jnk2^−/− embryos (20) from diabetic dams have significantly lower malformation rates than those in WT embryos exposed to hyperglycemia. Targeted deletions of either jnk1 or jnk2 gene are able to completely block downstream events of JNK1/2 activation, including activation of transcription factors, caspase cascade activation, and apoptosis in the developing neural tube induced by maternal diabetes. The current study provides compelling evidence that hyperglycemia-induced apoptosis in the developing neural tube accounts for NTD formation.

The expression of JNK1 and JNK2 in the neural tube starts at E7, whereas JNK3 is expressed at E11 (26). Therefore, JNK3 appears not to be involved in hyperglycemia-induced NTDs. Under normal conditions, virtually all neural tube cells are Sox1⁺ neural progenitors at E9.5 (40). In addition, we have shown that the apoptotic cells are Sox1⁺ neural progenitors, and no glia cells exist until E12.5 (41). Therefore, it is clear that JNK1 and JNK2 are expressed in Sox1⁺ neural progenitors along with apoptosis in some of these cells in the dorsal side of the neural tube. Double deletion of jnk1 and jnk2 increases apoptosis in the forebrain and decreases apoptosis in the lateral edges of the hindbrain. In contrast, single deletion of either jnk1 or jnk2 does not affect apoptosis in the neural tube. Therefore, JNK1KO and JNK2KO mice are suitable models in studying JNK1/2 activation–induced apoptosis in diabetic embryopathy.

A robust increase in JNK1/2 activities, which is associated with NTDs, is observed in both in vivo maternal diabetes models and in vitro embryo cultures exposed to high glucose (15). Subsequent pharmacological and genetic data (20), including these in the current study, confirm the causative role of JNK1/2 activation in the induction of diabetic embryopathy. Thus, JNK1/2 are the key intracellular kinases mediating the proapoptotic signals in diabetic embryopathy. JNK1 and JNK2 are cytoplasmic kinases and relay their proapoptotic signals to the nucleus through activation of an array of transcriptional factors. The current study demonstrates that the transcription factors c-Jun, ATF-2, Elk1, and Foxo3a, which are known downstream effectors of JNK1/2 activation, are activated in diabetic embryopathy. The activation of these transcription factors are presumably responsible for altered gene expression in diabetic embryopathy (36,42,43). Our study uncovers a proapoptotic signaling cascade emanating from JNK1/2 activation in diabetic embryopathy.

Studies suggest that Pax3 deficiency (13), transforming growth factor (TGF)-β pathway disturbance (44), and diminished Wnt signaling (45) are involved in diabetic teratogenesis. Because homeobox (Hox) genes are essential for Pax3 gene expression (46) and Hox genes suppress JNK1/2 activation (47), Hox genes may mediate the interaction between Pax3 and the JNK1/2 pathway. JNK1/2 inhibit the Wnt/β-catenin pathway (48), suggesting that JNK1/2 may contribute to the altered Wnt signaling in diabetic embryopathy (45). JNK1/2 inhibition enhances TGF-β1/Smad2 signaling in embryonic tissues (49), which suggests that JNK1/2 may be responsible for the TGF-β signaling disturbance (44). Therefore, there is potential crosstalk between the JNK1/2 pathway and other identified factors/pathways in mediating the teratogenicity of maternal diabetes.

In summary, the activation of the JNK1/2 pathway in diabetes-induced NTDs is the result of oxidative stress. Targeted deletion of jnk1 gene ameliorates diabetes-induced NTDs, and both JNK1 and JNK2 contribute to the induction of diabetic embryopathy. JNK1/2 activation induces activation of its downstream transcription factors and is responsible for diabetes-induced caspase activation and apoptosis. Thus, our study shows the causative role of the oxidative stress cascade (JNK1/2-caspase-apoptosis) in the induction of diabetic embryopathy.

This study is supported by National Institutes of Health (NIH) grants R01-DK-083770 (to E.A.R.) and R01-DK-083243 (to P.Y.). P.Y. is a Building Interdisciplinary Research Careers in Women’s Health scholar supported by NIH Grant K12-HD-043489 (principal investigator: Dr. Patricia Langenberg, University of Maryland School of Medicine).

No potential conflicts of interest relevant to this article were reported.

X.L., H.W., and C.X. researched data. E.A.R. designed the experiments and wrote the manuscript. P.Y. researched data, designed the experiments, and wrote the manuscript. All authors approved the final version of the manuscript. P.Y. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in abstract form at the 68th Scientific Sessions of the American Diabetes Association, San Francisco, California, 6–10 June 2008.

The authors are grateful to Hua Li at the University of Maryland School of Medicine for her technical support.