Differential placental ceramide levels during gestational diabetes mellitus (GDM)

Immunohistochemistry (IHC) was performed for ceramide localization in the placenta as previously performed in our laboratory [18]. Briefly, placental slides (n = 6) were deparaffinized, washed in TBS and blocked for 30 min with Background Sniper (Biocare Medical, Concord, Ca). Slides were incubated for 1 h with a mouse monoclonal primary antibody against cytokeratin 7 (for trophoblast localization; Dako, Carpinteria, CA), ceramide (R&D Systems, Minneapolis, MN) or with a universal IgG negative control (Biocare Medical; Concord, CA). Sections were incubated with Mach 2 secondary antibody (Biocare Medical, Concord, CA). Slides were developed with diaminobenzidine (DAB) for cytokeratin 7 or ceramide. Slides were imaged at 20X magnification.

Individual images were analyzed using imageJ software when evaluating the staining intensity of external peripheral tissue for controls (ceramide and isotype) and treatments (GDM-D and GDM-I) [19]. ImageJ images were quantified by first filtering for DAB specific staining and then subsequently images were converted to a grayscale for analysis [20]. A universal threshold was applied to the tissue to eliminate areas of negative space from the analysis. The membrane of each treatment (GDM-D and GDM-I) was measured (n = 10) and subsequently quantified by assessing the mean gray value across each membrane; of note, the lower the gray intensity the darker the staining.

Nuclear and cytosolic proteins were extracted from placental biopsies from GDM-I, GDM-D, and control samples using the NE-PER nuclear protein extraction kit (Pierce, Rockford, IL). Briefly, 100 mg of placental tissues was weighed, placed in 500 μl of cytoplasmic extraction reagent I (CER I) and homogenized; 27.5 μl of CER II were added to the samples, vortexed, and incubated on ice for 1 min. Samples were spun and the pellets were resuspended in 125 μl of ice-cold nuclear extraction reagent (NER). The samples were vortexed and returned to ice and vortexing continued for 15 s every 10 min for a total duration of 40 min. Samples were centrifuged, and the supernatant (nuclear protein) was transferred immediately to a pre-chilled tube and placed on ice. When not used immediately, all extracts were stored at − 80 °C. The quality of the extraction was tested by Western blotting of both the cytoplasmic and the nuclear extracts with antibodies against lamin B (a nuclear housekeeping protein, Santa Cruz Biotechnology, Dallas, TX) or actin (Abcam, Cambridge, MA).

Control, GDM-D, and GDM-I samples were obtained from the Research Centre for Women’s and Infant’s Health Biobank. Immunoblotting was performed as previously done in our laboratory [21]. Whole tissue lysates (50 mg) or cytoplasmic and nuclear extracts lysates were loaded (15 mg of protein) and separated on 4–12% Bis-Tris Midi Gel (Novex by Life Technologies, Carlsbad, CA). Proteins were transferred to nitrocellulose membranes using Invitrogen iBlot (Novex by Life Technologies, Carlsbad, CA). For protein determination, membranes were blocked in 5% milk in TBST for 1 h followed by overnight incubation with primary antibodies against: mouse NFAT5 (Affinity Bioreagents, Golden, CO), mouse SLC5A3 (SMIT; Fisher Scientific, St. Louis, MO), rabbit AR (Santa Cruz Biotechnology, Santa Cruz, CA) serine palmitoyltransferase 1 (SPT1, Sigma-Aldrich, St. Louis, MO), active caspase 3 (Cell Signaling, Danvers, MA), XIAP protein (an inhibitor of caspase activation Abcam, Cambridge, MA) Lamin B1 (Santa Cruz Biotechnology, Dallas, TX) or beta-actin (Abcam, Cambridge). Membranes were incubated with a secondary anti-rabbit horseradish peroxidase (HRP) conjugated antibody (Pierce Biotechnology, Rockford, IL,) for 1 h at room temperature followed by development using ECL substrate (Advansta, Menlo Park, CA). Proteins were detected by exposure of membranes to X-ray film and development. The presence of these proteins was confirmed and quantified. Bands were analyzed digitally with AlphaEaseFC software (Alpha Innotech Corporation, San Leandro, CA).

Human BeWo choriocarcinoma cells (which have a villi syncytiotrophoblastic phenotype) were maintained in F12K media supplemented with 10% fetal bovine serum (FBS) and 1% penicillin and streptomycin. Cells were plated at a density of two hundred thousand cells per well cm in six-well plates. Cells were treated with C2-ceramide (1 μM; Sigma-Aldrich, St. Louis, MO), insulin (50 nM, Sigma-Aldrich, St. Louis, MO) or fresh media for 24 h. Importantly, C2-ceramide is an oft-used agent, due to its solubility. After treatment, BeWo cells were used for mitochondrial respiration determination. Cell lysates were collected and evaluated for active caspase 3 and XIAP immunoblot determination.

High-resolution O₂ consumption was determined at 37 °C in permeabilized BeWo cells using the Oroboros Instruments O2K oxygraph. Before the addition of the samples into the respiration chambers, a baseline respiration rate was determined. After addition of the sample, the chambers were hyperoxygenated to ~ 350 nmol/ml. Following this step, electron flow through complex I was supported by GM (glutamate+malate; 10 and 2 mM respectively). Following stabilization, ADP (2.5 mM) was added to determine oxidative phosphorylation capacity (GMD). The integrity of the outer mitochondrial membrane was then tested by adding cytochrome c (10 μM; not shown). Succinate was added (GMSD) for complex I + II electron flow into the Q-junction. To determine the full ETS (electron-transport system) capacity over oxidative phosphorylation, the chemical uncoupler FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) was added (GMSE; 0.05 μM).

Demographics of human placental sample donors were analyzed for significant differences between control (non GDM normal health pregnancy), GDM-D and GDM-I groups. There were no significant differences in maternal age, BMI, gestational weeks and fetal weight between control and both GDM pregnancies (Table 1).

Ceramide is present in the villi of trophoblast cells [10, 11] so we investigated ceramide levels in control placentas and GDM placentas induced with either diet or insulin. A set of representative images of ceramide staining is shown in Fig. 1. Immunohistochemistry quantification confirm increased ceramide staining in the villous trophoblast of the placenta during GDM-I but not in the GDM-D tissues (Fig. 1).

We next wanted to investigate the degree to which the de novo ceramide biosnynthetic pathway was affected. Thus, we explored SPT1 levels, one isoform of the rate-limiting biosynthetic enzyme [13]. No significant differences were observed for cytosolic SPT1 expression between control and GDM placental tissues (Fig. 2a). In contrast, highly upregulated expression of the nuclear SPT1 enzyme was present only in the GDM-I placenta (3.4-fold; p < 0.05) when compared to controls (Fig. 2b), highlighting the potential relevance of a nuclear source of ceramides.

Studies have shown that an increase in osmolarity leads to the activation of TonEBP/NFAT5 [22]. Activation of TonEBP/NFAT5 leads to increased expression of transmembrane proteins such as sodium-dependent myo-inositol transporter (SMIT) as well as the induction of the aldose reductase enzyme (AR; responsible for sorbitol production), which regulates the production and accumulation of inositol and sorbitol. Collectively, these factors regulate production and transport of organic osmolytes into cells to maintain normal osmolarity and cell volume [22]. Figure 3a shows a characteristic western blot for NFAT5, SMIT and AR of treated trophoblast cells as compared to controls. We first investigated the cytosolic and nuclear expression of NFAT5 in the human placenta of control and GDM patients. We observed increased expression of nuclear NFAT5 in both GDM-D (2.8-fold; p < 0.003) and GDM-I (2.5-fold; p < 0.0001), but cytosolic NAFT5 was not elevated in the GDM placentas when compared to controls (Fig. 3b, c). A significant increase in SMIT was observed in the GDM-D (1.8-fold; p < 0.02) and GDM-I (2-fold; p < 0.005) placenta when compared to controls (Fig. 3d). No expression differences were observed for AR when comparing GDM and control placentas (Fig. 3e).

Decreased apoptosis is present in the GDM placenta when compared to control placentas [3]. Active caspase 3 and the anti-apoptotic inhibitor of caspase XIAP was evaluated in the placenta of control and diet or insulin treated GDM patients. Specifically, there was an upregulation of active caspase 3 (1.2-fold; p < 0.05) in the placentas from both GDM-I and GDM-D when compared to control placental tissue (Fig. 4a). Interestingly, a significant decrease of XIAP expression (1.7-fold; p < 0.05) was only observed in the GDM-I placenta when compared to controls (Fig. 4b).

To provide further evidence of altered cellular function, and to mimic the pregnancy milieu of GDM, we treated human placental trophoblast villi cells (BeWo) with insulin (50 nM) or ceramide (C2-ceramide; 1 μM), as used previously [23], prior to placement into respirometer chambers. Oxygen flux was determined in conditions of multiple substrates (Fig. 5a; see methods or legend for details). Both treatments resulted in a significant reduction in mitochondrial respiration compared with controls, which became apparent upon the addition of succinate (GMSD) and remained with addition of FCCP (GMSE). Despite the difference in respiration rates across treatments, respiratory control ratios (RCR; Fig. 5b), a general indicator of mitochondrial function, revealed no apparent differences in the functionality or overall health of the mitochondria. Lastly, the profound disparity across treatments in response to succinate (GMS) was very apparent when we determined the complex II factor, an indicator of succinate sensitivity (Fig. 5c), wherein C2 and insulin (INS) treatments were significantly lower vs. controls (CON), albeit to varying degrees. Active caspase 3 and the anti-apoptotic inhibitor of caspase XIAP was also evaluated in control and ceramide treated BeWo cells. There was no significant change in active caspase in the ceremide treated trophoblast when compared to control placental tissue (Fig. 6). Interestingly, a significant increase of XIAP expression (1.7-fold; p < 0.03) was observed in the treated trophoblasts when compared to controls (Fig. 6).