INTRODUCTION
Gestational diabetes mellitus (GDM), a subtype of diabetes marked by carbohydrate intolerance and insulin resistance, is a common obstetric disorder and a maternal medical complication of pregnancy typically presenting with various degrees of hyperglycemia[1]. It is estimated by The International Diabetes Federation that by 2045, there will be 700 million GDM cases across the globe[2]. In mainland China, GDM prevalence among pregnant women is approximately 14.8%[3]. In addition to the adverse outcomes on mothers including gestational hypertension, preeclampsia, GDM also imposes great impact on infant development, such as congenital anomalies, metabolic disturbances, macrosomia, respiratory difficulties, shoulder dystocia, stillbirth and so on[4,5]. Beyond that, GDM has historically been regarded as a harbinger of the long-term risk of type 2 diabetes mellitus in both mothers and offspring[6]. The crosstalk between environmental and genetic factors has been evidenced to drive the complex etiology of GDM[7]. So far, pharmacologic therapy and lifestyle modifications are extensively applied to lower blood glucose and counter insulin resistance in the management of GDM[8,9]. However, further research into the concealed etiological mechanism of GDM is integral to the discovery of feasible therapeutic biomarkers.
Endoplasmic reticulum (ER) is a membranous intracellular organelle responsible for protein synthesis and folding, calcium transport and storage, as well as lipid synthesis. ER stress (ERS) occurs when the protein-folding capacity of the ER is overwhelmed, leading to the accumulation of unfolded or misfolded proteins. To restore ER homeostasis, cells activate the unfolded protein response (UPR), an adaptive signaling network that reduces protein load, enhances protein folding capacity, and promotes degradation of misfolded proteins. However, persistent ERS may shift the UPR toward pro-apoptotic signaling, thereby contributing to cellular dysfunction. Accumulating evidence indicates that ERS is involved in placental dysfunction in GDM and contributes to disease progression[10-12].
General control nonderepressible 2 (GCN2) is a serine/threonine-protein kinase that acts as a sensor of amino acid deprivation. Under conditions of amino acid deficiency, uncharged tRNAs accumulate and bind to GCN2, leading to its activation. Activated GCN2 subsequently phosphorylates eukaryotic initiation factor 2α (eIF2α), resulting in global inhibition of protein synthesis while selectively promoting the translation of stress-responsive genes[13]. The current investigations have introduced the crucial roles of GCN2 as a central effector in the integrated stress response across tumors[14], immune-related diseases[15] and pulmonary diseases[16] through mediating a cascade of biological processes. A considerable body of literature has broadly documented GCN2 as a pivotal participator in diabetes and diabetic complications[17-20] and the beneficial impact of inhibiting GCN2 has been well established. Also, the preceding reports have demonstrated that GCN2 is implicated in the placental glucocorticoid barrier function and fetal growth[21-23]. More importantly, GCN2 has been supported as a regulator of ERS in multiple human diseases[17,24]. These findings suggest that GCN2 may serve as a critical link between metabolic stress and ERS signaling; however, whether GCN2 mediates ERS-associated mitochondrial dysfunction in the placental context of GDM remains unclear.
Although previous studies have implicated GCN2 in metabolic disorders and ERS regulation, most of these investigations have focused on systemic metabolic effects, while its role in GDM, particularly in linking ERS to mitochondrial dysfunction and placental injury, has not been fully elucidated. Therefore, this study aimed to investigate whether GCN2 contributes to GDM progression by mediating ERS-associated mitochondrial dysfunction in placental tissues, thereby clarifying its role in linking cellular stress signaling to placental injury.
MATERIALS AND METHODS
Animals
C57BL/6J mice [6-8 weeks old; 24 females and 12 males; 20-25 g; Guangdong Provincial Medical Laboratory Animal Center; SPF grade; SCXK (Yue) 2022-0002] were maintained in a controlled environment (21-23 °C, 40%-60% humidity, 12 hours light/dark cycle) and had unrestricted access to food and water. All experimental procedure were reviewed and approved by the Animal Ethics Committee of the Guangdong Provincial Medical Laboratory Animal Center (Approval No. D202506-14).
Study design of the preclinical animal model
Female and male mice were co-housed overnight for mating. Mating success was verified by the presence of a vaginal plug the following morning. Pregnant mice were randomly allocated to four groups (n = 6/group) using a random number-based grouping method, and sample size determination was informed by earlier comparable studies and standard principles for animal experiments[25]. The groups were as follows: A control group (standard rodent chow), a GDM group (high-fat/high-sugar diet consisting of 65% standard feed, 10% sugar, 10% yolk, and 15% lard), a GDM + sh-NC group (high-fat/high-sugar diet with tail vein administration of 1 × 107 units of sh-NC lentivirus), and a GDM + sh-GCN2 group (high-fat/high-sugar diet with tail vein administration of 1 × 107 units of sh-GCN2 lentivirus)[26]. The sh-GCN2 and sh-NC lentiviruses were produced by Sbo-Bio (Shanghai, China). Mouse body weight was measured on gestational days 10 and 20. The GDM model was assessed using metabolic measurements, comprising blood glucose, insulin concentrations, glucose tolerance, insulin sensitivity, and plasma lipid profiles. During outcome assessment, investigators involved in histological observation and image-based quantitative analysis were blinded to group allocation.
Assessment of blood glucose and insulin
On gestational days 10 and 20, blood specimens were harvested from the tail vein to assess blood glucose using a glucometer (Nova Biomedical, Waltham, United Kingdom). After 15 minutes of centrifugation of blood samples at 2000 × g, the serum insulin concentrations were examined via a Mouse Insulin ELISA kit (Cat. No. 80-INSHUU-E01.1; ALPCO, Salem, NH, United States). A Berthold Technologies microplate reader (Bad Wildbad, Germany) was operated at 450 nm to quantify the absorbance.
Glucose tolerance test and insulin tolerance test
On gestational day 20, mice received glucose (2.0 g/kg) by intragastric gavage or insulin (1.0 U/kg) via intraperitoneal injection following a 6-hours fast. Blood glucose concentrations were documented at 0-, 30-, 60-, and 120-minute via a glucometer.
Examination of lipid biochemistry
On the 20 days of pregnancy, the plasma was isolated from blood samples by centrifugation (1000 × g; 10 minutes) to separately quantify total cholesterol (TC; Cat. No. D799799), triglyceride (TG; Cat. No. D799795), low-density lipoprotein cholesterol (LDL-C; Cat. No. D799859) and high-density lipoprotein cholesterol (HDL-C, Cat. No. D799861) concentrations (all from Sangon Biotech, Shanghai, China). A microplate reader was set to 500 nm, 420 nm, 600 nm and 600 nm to quantify the absorbance.
Hematoxylin-eosin assay
At gestational day 20, following biochemical assessments, euthanasia of pregnant mice was performed using intraperitoneal sodium pentobarbital (200 mg/kg), followed by collection of liver and placental tissues for subsequent analyses. The hepatic and placental tissues underwent fixation in 4% paraformaldehyde before being transferred to 70% ethanol. Next, the tissues underwent paraffin embedding followed by sectioning into 5-μm sections, which were subjected to hematoxylin staining for 5 minutes and eosin (both from Sangon Biotech, Shanghai, China) for 30 seconds. Histological structures were examined using a light microscope.
Oil Red O staining
Fresh hepatic specimens were immersed in 4% paraformaldehyde for fixation, cryosectioned, dipped in 60% isopropanol for 20 seconds, and subsequently processed with Oil Red O staining (Servicebio, Wuhan, Hubei Province, China) for 10 minutes. 60% isopropanol was utilized again for color separation. After being counterstained with hematoxylin, Oil Red O-positive lesion areas were recorded under a light microscope.
Periodic acid-Schiff staining
Mouse hepatic tissues were preserved in 4% paraformaldehyde and then placed in 70% ethanol. Subsequently, the tissues were paraffinized and sectioned to 5 μm. Following oxidation with periodic acid, the sections were rinsed in distilled water and placed in Schiff reagent (Wellbio, Changsha, Hunan Province, China) for 15 minutes. After being counterstained with hematoxylin, alcohol dehydration and neutral gum sealing, the Periodic acid-Schiff (PAS)-positive areas were recorded under a light microscope.
Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling
Paraffin-embedded mouse placental tissues were subjected to deparaffinization in xylene, followed by rehydration through a graded ethanol series and permeabilization using proteinase K (20 μg/mL). The transferase-mediated dUTP nick end labeling (TUNEL) assay reagent (Proteintech, Wuhan, Hubei Province, China), prepared following the manufacturer’s instructions, was applied to the sections and incubated for 1 hour in the dark, followed by DAPI counterstaining (2 μg/mL, 10 minutes). After conventional ethanol dehydration, xylene permeation and mounting with antifluorescence quenching sealing solution, TUNEL-positive signals were visualized under a fluorescence microscope.
Dihydroethidium and dichlorodihydrofluorescein diacetate staining
Dihydroethidium or dichlorodihydrofluorescein diacetate staining (10 μmol/L; ApplyGen, Beijing, China) was used to stain mouse placental specimens or HTR-8/SVneo cells at 37 °C for 30 minutes in the dark, after which fluorescence signals were recorded by fluorescence microscopy after PBS washing.
Transmission electron microscopy
After being treated by 2.5% glutaraldehyde, the placental tissues were subjected to post-fixation in 1% osmium tetroxide (1 hour, 4 °C) prior to dehydration and epoxy resin embedding. After being divided into ultrathin slices (60-80 nm), the samples were contrasted with 2% uranyl acetate. The mitochondrial morphology was monitored using a transmission electron microscope (Hitachi, Japan).
Detection of ATP synthesis
Homogenized mouse placental tissues were centrifuged (10000 × g, 10 minutes), and ATP content was determined using an ATP detection kit (Cat. No. E2031; ApplyGen). After being lysed in lysis buffer, HTR-8/SVneo cells were exposed to ATP-Lite assay system (Cat. No. T007; Vigorous Biotechnology, Beijing, China) for measurement of cellular ATP content.
Evaluation of oxidative stress-related indexes
Placental tissue homogenates from mice were centrifuged (10000 × g, 15 minutes), whereas HTR-8/SVneo cells were centrifuged (2000 × g, 10 minutes), after which malondialdehyde (MDA) levels as well as the enzymatic activities of superoxide dismutase (SOD) and catalase (CAT) were separately quantified using an MDA assay kit (Cat. No. A003-1-1; Jiancheng Corp, Nanjing, Jiangsu Province, China), and SOD and CAT activity assay kits (Cat. No. STA-340 and STA-341; Cell Biolabs, Inc., San Diego, CA, United States). A microplate reader was set to 532 nm, 490 nm and 520 nm to quantify the absorbance.
Cell treatment and shRNA transfection
After culture in RPMI-1640 medium (Life-iLab, Shanghai, China) containing 10% FBS, human villous trophoblasts (HTR-8/SVneo; NCACC, Shanghai, China) were employed for the generation of GDM cell model upon exposure to 25 mmol/L glucose. Cells treated with 5 mmol/L glucose were referred to as the non-glucose group. Additionally, high glucose-challenged cells were pretreated with 5 μg/mL ERS activator tunicamycin (TM; InvivoChem, United States) for 6 hours[27]. Besides, high-glucose-treated cells were also underwent transfection with sh-GCN2 or sh-NC after the addition of HighGene transfection reagent (Abclonal, Wuhan, Hubei Province, China).
Reverse transcription-quantitative PCR
Using RNAVzol reagent, RNA extraction from tissues and cell specimens was carried out, with subsequent reverse transcription into cDNA using the VigoScript First Strand cDNA Synthesis Kit (both from Vigorous Biotechnology). Quantitative PCR was then performed using PCR SuperMix, and mRNA expression levels were subsequently calculated according to 2-ΔΔCt approach.
Following the isolation of total DNA via DNAVzol reagent (Vigorous Biotechnology), the mtDNA copy number was quantified by PCR. The primer sequences used for reverse transcription quantitative PCR are provided in Supplementary Table 1.
Western blot
After RIPA lysis of tissues and cells (Life-iLab), protein concentrations were determined using a BCA kit (Vigorous Biotechnology). PVDF membranes containing equal amounts of SDS-PAGE-separated proteins were subjected to 5% BSA blocking. Primary antibody incubation was next carried out at 4 °C overnight, after which membranes were exposed to corresponding secondary antibodies for 1 hour at room temperature. After treatment with the Western Luminescent Detection Kit (Vigorous Biotechnology), the protein bands were detected, and band intensities were quantified using Quantity One software (v4.6.2; Bio-Rad Laboratories, Hercules, CA, United States).
JC-1 probe
HTR-8/SVneo cells underwent incubation with 1 mL JC-1 dye (Life-iLab) at 37 °C for 15 minutes under dark conditions. After being rinsed with JC-1 washing buffer, the red (λex = 550 nm; λem = 600 nm) and green (λex = 485 nm; λem = 535 nm) emissions were acquired by fluorescence microscopy.
Flow cytometry analysis
Following the manufacturer’s protocol for the Annexin V-FITC/PI Apoptosis Detection Kit (Life-iLab), HTR-8/SVneo cells were collected and reconstituted in 100 μL of 1 × binding buffer. Subsequently, 3 × 105 cells were incubated with Annexin V-FITC (5 μL) plus propidium iodide for 10 minutes at room temperature in darkness. Apoptotic rates were determined using flow cytometry.
Statistical analysis
SPSS 22.0 software was used for all analyses. Results are expressed as mean ± SD. Data normality and homogeneity of variance were evaluated using the Shapiro-Wilk test and Levene’s test, respectively. Differences among multiple groups were analyzed by one-way analysis of variance (ANOVA) with Tukey’s post hoc test. For repeated measurements, including glucose tolerance test and insulin tolerance test, two-way repeated-measures ANOVA was applied. Multiple comparisons were adjusted using appropriate post hoc tests to control for type I error. Statistical significance was defined as P < 0.05.
RESULTS
GCN2 is upregulated in the placentas of GDM mice
After GDM modeling, GCN2 expression was markedly increased in the placentas of GDM mice vs control mice (Figure 1A and B). To further investigate the functional role of GCN2, two shRNA constructs targeting GCN2 (sh-GCN2-1 and sh-GCN2-2) were generated and validated prior to loss-of-function experiments (Figure 1C and D). Therefore, sh-GCN2-1 was employed in the subsequent assays. Consistently, the elevated GCN2 expression in the placentas of GDM mice was markedly reduced at both the transcript and protein levels following sh-GCN2 treatment (Figure 1E and F).
Figure 1 General control nonderepressible 2 is upregulated in placental tissues of gestational diabetes mellitus mice.
A and B: Reverse transcription quantitative PCR (RT-qPCR; A) and Western blotting analysis (B) of general control nonderepressible 2 (GCN2) expression; C and D: RT-qPCR (C) and Western blotting analysis (D) of GCN2 knockdown efficiency; E and F: RT-qPCR (E) and Western blotting (F) analysis of GCN2 expression in placental tissues of gestational diabetes mellitus (GDM) mice. aP < 0.001 vs control; bP < 0.001 vs sh-negative control (NC); cP < 0.001 vs GDM + sh-NC. GCN2: General control nonderepressible 2; GDM: Gestational diabetes mellitus; NC: Negative control.
GCN2 knockdown improved metabolic parameters in GDM mice
Compared with control mice, GDM mice exhibited significantly increased body weight at gestational days 10 and 20. After injection of sh-GCN2, the body weight of GDM mice was lowered (Figure 2A). Simultaneously, blood glucose concentrations were markedly increased, whereas insulin concentrations were decreased in GDM mice; both alterations were significantly reversed following GCN2 knockdown (Figure 2B). Also, Intraperitoneal glucose tolerance testing revealed significantly impaired glucose tolerance in GDM mice, which was markedly improved following GCN2 knockdown (Figure 2C). Similarly, intraperitoneal insulin tolerance testing demonstrated reduced insulin sensitivity in GDM mice, which was partially restored by GCN2 knockdown (Figure 2D). Besides, circulating lipid indices, comprising TC, TG, and LDL-C, all exhibited an upward trend in GDM mice whereas HDL-C displayed a downward trend. GCN2 knockdown significantly reduced plasma TC and TG concentrations and restored HDL-C concentrations, with LDL-C concentrations also showing a decreasing trend (Figure 2E), highlighting the beneficial impact of GCN2 loss on hyperlipidemia in GDM mice.
Figure 2 General control nonderepressible 2 knockdown improved metabolic parameters in gestational diabetes mellitus mice.
A: Body weight; B: Blood glucose and insulin levels; C: Glucose tolerance test; D: Insulin tolerance test; E: Plasma lipid levels. aP < 0.05 vs control; bP < 0.001 vs control; cP < 0.01 vs gestational diabetes mellitus (GDM) + sh-negative control; dP < 0.001 vs GDM + sh-negative control. GCN2: General control nonderepressible 2; GDM: Gestational diabetes mellitus; NC: Negative control; TC: Total cholesterol; TG: Triglyceride; LDL-C: Low-density lipoprotein cholesterol; HDL-C: High-density lipoprotein cholesterol.
GCN2 knockdown reduced hepatic lipid deposition and restored glycogen accumulation in GDM mice
Hematoxylin-eosin (H&E) staining revealed disrupted hepatic architecture and increased cytoplasmic vacuolation in GDM mice, which were alleviated following GCN2 knockdown (Figure 3A). Additionally, Oil Red O staining demonstrated marked lipid deposition within the hepatic tissues derived from GDM mice, which was markedly reduced by GCN2 knockdown (Figure 3B). Reverse transcription-quantitative PCR analysis showed that the transcript level of lipogenic gene SREBP-1c was significantly upregulated in GDM mice, whereas the expression levels of lipid oxidation-associated genes PPAR-α and ACOX1 were reduced. These alterations were reversed following GCN2 knockdown (Figure 3C). As Figure 3D depicted, PAS staining demonstrated that hepatic glycogen levels were strikingly decreased after GDM modeling. PAS staining further demonstrated that hepatic glycogen content was markedly decreased in GDM mice and was restored following sh-GCN2 treatment (Figure 3D).
Figure 3 General control nonderepressible 2 knockdown reduced hepatic lipid deposition and restored glycogen accumulation in gestational diabetes mellitus mice.
A: Hepatic histological changes assessed by hematoxylin and eosin staining; B: Hepatic lipid accumulation assessed by Oil Red O staining; C: Expression of lipid metabolism-related genes; D: Hepatic glycogen content assessed by Periodic acid-Schiff staining. aP < 0.001 vs control; bP < 0.01 vs gestational diabetes mellitus (GDM) + sh-negative control (NC); cP < 0.001 vs GDM + sh-NC. GCN2: General control nonderepressible 2; GDM: Gestational diabetes mellitus; NC: Negative control.
GCN2 knockdown alleviated apoptosis and oxidative damage in placental tissues of GDM mice
H&E staining revealed pronounced histopathological alterations in placental tissues of GDM mice, which were alleviated following GCN2 knockdown (Figure 4A). Specifically, TUNEL staining demonstrated a significant increase in apoptotic cells, which was markedly reduced following GCN2 knockdown (Figure 4B). Evidently, Western blot analysis showed downregulation of the anti-apoptotic protein Bcl-2 along with upregulation of the pro-apoptotic protein Bax in GDM placentas; these changes were reversed following GCN2 knockdown (Figure 4C). Consistently, reactive oxygen species (ROS) staining revealed enhanced oxidative stress in GDM placentas, which was attenuated by GCN2 knockdown (Figure 4D). Further, biochemical analysis showed that MDA levels were increased, whereas SOD and CAT activities were reduced in GDM placentas (Figure 4E). GCN2 knockdown significantly decreased MDA levels and restored SOD and CAT activities in GDM placentas (Figure 4E).
Figure 4 General control nonderepressible 2 knockdown attenuated apoptosis and oxidative stress in placental tissues of gestational diabetes mellitus mice.
A: Histological changes in placental tissues; B: Apoptosis level in placental tissues as detected by terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling staining; C: Expression levels of apoptosis-related proteins determined by Western blotting; D: Reactive oxygen species production assessed by dihydroethidium staining; E: Oxidative stress-related indices including malondialdehyde, superoxide dismutase, and catalase. aP < 0.001 vs control; bP < 0.01 vs gestational diabetes mellitus (GDM) + sh-negative control (NC); cP < 0.001 vs GDM + sh-NC. GCN2: General control nonderepressible 2; GDM: Gestational diabetes mellitus; NC: Negative control; MDA: Malondialdehyde; SOD: Superoxide dismutase; CAT: Catalase.
GCN2 knockdown alleviated ERS-associated mitochondrial dysfunction in placental tissues of GDM mice
Furthermore, the expression levels of ERS-associated proteins GRP78, CHOP, and p-eIF2α exhibited a marked increase in GDM placentas, and this increase was markedly reduced following administration of sh-GCN2 lentivirus (Figure 5A). Transmission electron microscopy revealed that mitochondria exhibited normal morphology in the control group, whereas pronounced mitochondrial structural damage was observed in the GDM group (Figure 5B). However, GCN2 deficiency noticeably eased mitochondrial injury in the placenta of GDM mice (Figure 5B). Moreover, mitochondrial ATP production, which was markedly reduced in the placenta of GDM mice, was significantly restored following GCN2 knockdown (Figure 5C).
Figure 5 General control nonderepressible 2 knockdown alleviated endoplasmic reticulum stress-associated mitochondrial dysfunction in placental tissues of gestational diabetes mellitus mice.
A: Expression levels of endoplasmic reticulum stress-related proteins determined by Western blotting; B: Mitochondrial ultrastructural changes observed by transmission electron microscopy; C: ATP content in placental tissues. aP < 0.001 vs control; bP < 0.05 vs gestational diabetes mellitus (GDM) + sh-negative control (NC); cP < 0.001 vs GDM + sh-NC. eIF2α: Eukaryotic initiation factor 2α; GCN2: General control nonderepressible 2; GDM: Gestational diabetes mellitus; NC: Negative control.
GCN2 absence inactivated ERS to mitigate mitochondrial dysfunction in HG-exposed HTR-8/SVneo cells
To further determine whether GCN2 is involved in the pathogenesis of GDM, an in vitro GDM model was established by exposing HTR-8/SVneo cells to HG. Similarly, GCN2 expression was markedly elevated at both the transcriptional and translational levels in HTR-8/SVneo cells upon HG exposure and was effectively reduced by sh-GCN2 transfection (Figure 6A and B). Particularly, HG-induced upregulation of GRP78, CHOP, and p-eIF2α was markedly attenuated in GCN2-silenced HTR-8/SVneo cells (Figure 6C). JC-1 dye revealed that HG exposure significantly decreased mitochondrial membrane potential in HTR-8/SVneo cells, whereas GCN2 knockdown partially restored mitochondrial membrane potential, as indicated by increased red J-aggregate fluorescence and decreased green J-monomer fluorescence. However, after pretreatment with ERS activator TM, the red fluorescence was weakened and the green fluorescence was boosted again in GCN2-interfering HTR-8/SVneo cells under HG conditions (Figure 6D), indicating a reduction in mitochondrial membrane potential. Besides, ATP levels and mtDNA content were notably reduced in HG-exposed HTR-8/SVneo cells, whereas both were markedly increased following GCN2 knockdown; these effects were attenuated by TM pretreatment (Figure 6E and F).
Figure 6 General control nonderepressible 2 absence inactivated endoplasmic reticulum stress to mitigate mitochondrial dysfunction in high glucose-exposed HTR-8/SVneo cells.
A and B: General control nonderepressible 2 (GCN2) expression at mRNA and protein levels determined by reverse transcription quantitative PCR (RT-qPCR) and Western blotting; C: Expression levels of endoplasmic reticulum stress-related proteins; D: Mitochondrial membrane potential assessed by JC-1 staining; E: ATP content in cells; F: The mtDNA content determined by RT-qPCR. aP < 0.001 vs normal glucose; bP < 0.01 vs high glucose (HG) + sh-negative control (NC); cP < 0.001 vs HG + sh-NC; dP < 0.001 vs HG + sh-GCN2. GCN2: General control nonderepressible 2; GDM: Gestational diabetes mellitus; NC: Negative control; eIF2α: Eukaryotic initiation factor 2α; NG: Normal glucose; HG: High glucose.
GCN2 absence inactivated ERS to retard the apoptosis and oxidative damage in HG-exposed HTR-8/SVneo cells
Under HG conditions, HTR-8/SVneo cells exhibited pronounced elevations in intracellular ROS and MDA, accompanied by reduced SOD and CAT activities. GCN2 knockdown alleviated HG-induced oxidative stress, whereas these protective effects were abolished upon exposure to the ERS agonist TM (Figure 7A and B). Besides, flow cytometric analysis showed that HG exposure significantly increased apoptosis in HTR-8/SVneo cells, while GCN2 silencing substantially attenuated this apoptotic response. Consistently, GCN2 silencing decreased Bax expression and increased Bcl-2 expression, whereas these anti-apoptotic effects were partially abolished by TM (Figure 7C and D).
Figure 7 General control nonderepressible 2 absence inactivated endoplasmic reticulum stress to retard the apoptosis and oxidative damage in high glucose-exposed HTR-8/SVneo cells.
A: Reactive oxygen species production assessed by 2′,7′-dichlorodihydrofluorescein diacetate staining; B: Oxidative stress-related indices including malondialdehyde, superoxide dismutase and catalase; C: Apoptosis level determined by flow cytometry; D: Expression levels of apoptosis-related proteins. aP < 0.001 vs normal glucose; bP < 0.001 vs high glucose (HG) + sh-negative control; cP < 0.05 vs HG + sh-GCN2; dP < 0.01 vs HG + sh-GCN2; eP < 0.001 vs HG + sh-GCN2. NG: Normal glucose; HG: High glucose; MDA: Malondialdehyde; NC: Negative control; GCN2: General control nonderepressible 2; TM: Tunicamycin; SOD: Superoxide dismutase; CAT: Catalase.
DISCUSSION
GDM is generally accepted as a transient form of diabetes first diagnosed during pregnancy and may also be closely linked to an increased risk of developing type 2 diabetes mellitus later in life[28]. As an amino acid deprivation sensor activated under diverse stress conditions, GCN2 is increasingly recognized as an important modulator of metabolic diseases, including obesity and pathological diabetes[18,29,30]. The present findings support a role of GCN2 in GDM-associated metabolic and placental alterations, potentially involving ERS-related mitochondrial dysfunction. Specifically, the concurrent improvement in metabolic parameters and placental injury following GCN2 silencing suggests a potential association between GCN2 and ERS-associated mitochondrial dysfunction, oxidative stress, and apoptosis in GDM. However, whether these effects are direct or secondary to systemic metabolic improvements remains unclear and warrants further investigation. Furthermore, ERS activation reversed the cytoprotective outcomes of GCN2 inhibition against mitochondrial dysfunction, oxidative injury, and apoptosis in vitro. The GCN2-ERS-mitochondrial pathway proposed in this study should be considered a putative mechanism based on current evidence rather than a definitively established causal relationship.
GCN2 is activated in human placental trophoblasts and placentas from pregnant mice exposed to cadmium, a prevalent environmental contaminant[21-23]. As an initial observation, GCN2 levels exhibited a marked elevation in GDM placentas and in HG-treated HTR-8/SVneo trophoblasts across both animal and cell-based GDM models established in this study. Similar to type 2 diabetes, GDM represents a pathological condition characterized by insulin resistance accompanied by defective insulin secretion, ultimately leading to disturbances in both glucose and lipid metabolism[31,32]. Moreover, patients with GDM showed increased concentrations of conventional lipids (TC, TG, and LDL-C) and reduced HDL-C levels relative to pregnant women with normal glucose tolerance[33]. Consistently, GDM mice displayed increased body mass, hyperglycemia, diminished circulating insulin levels, impaired glucose tolerance, as well as elevated TC, TG, and LDL-C concentrations and reduced HDL-C concentrations. GCN2 deficiency can alleviate hyperlipidemia and insulin resistance in diabetic mice[17,18]. Moreover, GCN2 upregulation aggravated lipid accumulation in diabetic models[17]. Also, during our investigation, GCN2 depletion improved glucose and lipid profiles. However, these systemic metabolic improvements may not be entirely placenta-specific and could partly account for the observed placental protection, which warrants further investigation.
The liver plays an important role in systemic glucose and lipid metabolism[34], and hepatic alterations have been associated with metabolic disturbances in GDM[35], which may indirectly influence placental function. SREBP-1c, a transcriptional regulator governing genes involved in fatty acid synthesis, is markedly upregulated in pregnant women with GDM, suggesting its possible implication in driving insulin resistance and disturbances in lipid metabolism[26,36]. PPARα and ACOX1 abundantly expressed in the liver are master regulators for fatty acid oxidation, which are lowly expressed in GDM placentas[37-39]. Glycogen is a glucose-based branched polymer predominantly present in the liver, the content of which is also upregulated in GDM placentas[40,41]. Increasing evidence has well documented that GCN2 absence can diminish hepatic steatosis and repress lipogenesis genes in obese mice[18,42]. In alignment with these publications, downregulation of GCN2 was noted to mitigate hepatic damage, lipid droplets, suppressed SREBP-1c expression whereas enhanced PPAR-α, ACOX1 expressions and glycogen content in the hepatic tissues of GDM mice.
Moreover, interference with GCN2 has emerged to protect against oxidative damage and apoptosis in cardiac tissues of diabetic mice[17] and HG-exposed ARPE-19 human retinal pigment epithelial cells[19] via repressing ROS, MDA levels, boosting Bcl-2 expression and CAT level. Oxidative stress and apoptosis are reckoned as frequent physiological processes during the advancement of GDM[43,44]. ROS is abnormally produced in GDM, bringing about oxidative stress, which may impair placental functions[45]. Besides, pro-oxidant MDA level is ascending and anti-oxidant SOD and CAT levels were descending in GDM rats[2]. Similarly, our findings evidenced that, the apoptotic level was boosted, Bcl-2 expression, SOD and CAT concentrations were eliminated, Bax expression, ROS and MDA concentrations were fortified in the placenta samples of GDM mice and HG-challenged HTR-8/SVneo cells. After GCN2 was deleted, these alterations were reversed, suggesting that GCN2 may be involved in oxidative stress and apoptosis through ERS-mediated mitochondrial dysfunction, which could contribute to metabolic disturbances and placental injury, but systemic metabolic effects should also be considered.
Mitochondria are highly specialized organelles and central regulators of cellular energy homeostasis by generating ATP, the primary cellular energy currency[46]. Mitochondrial membrane potential serves as an indicator of mitochondrial metabolic and functional integrity[47]. The mtDNA is essential for mitochondrial function by encoding key components required for the formation and operation of mitochondrial respiratory chain complexes[48]. Besides working as the cellular powerhouse, mitochondria are also involved in diverse critical cellular functions, including lipid synthesis, ROS release, and cell death[49-51]. The ER and mitochondria are closely connected organelles within cells and ER may act as a relay station between stressors and mitochondria[52]. Notably, ERS signaling is activated and ERS-associated factors are all hyper expressed in GDM[12,39]. Also, mitochondrial function is impaired in the setting of GDM[53,54]. As reported, GCN2 knockdown can also downregulate CHOP and p-eIF2α expressions to resist ERS in cerebral ischemia in mice[24] and GCN2 could contribute to mitochondrial dysfunction in gastric cancer cells[55]. Herein, both in vitro and in vivo, interfering GCN2 could deplete ERS-associated GRP78, CHOP and p-eIF2α expressions. Regarding mitochondrial dysfunction, suppression of ERS markers including GRP78 was accompanied by alleviated mitochondrial damage and restored ATP production in placental tissues of GDM mice, as well as improved mitochondrial membrane potential, increased mtDNA content, and enhanced ATP production in HG-treated HTR-8/SVneo cells. These findings support a potential role of GCN2 in regulating ERS and mitochondrial dysfunction during GDM; however, the causal relationship requires further validation through more specific mechanistic approaches. Furthermore, it turns out that ERS can function as a contributor to mitochondrial dysfunction to impact oxidative stress and apoptosis in a broad range of human diseases[56-58]. In the present study, we also noticed that pretreatment with ERS activator TM also repressed mitochondrial membrane potential, ATP synthesis, reduced mtDNA, SOD, CAT contents, Bcl-2 expression, strengthened ROS, MDA contents, Bax expression, as well as exacerbated the apoptotic capacity of HG-treated GCN2-silenced HTR-8/SVneo cells.
Despite these findings, several limitations should be acknowledged. First, the relatively small sample size in the animal experiments may affect the robustness and reproducibility of the results. Second, the GDM model induced by a high-fat/high-sugar diet was primarily evaluated using metabolic parameters and does not fully meet clinical diagnostic criteria such as oral glucose tolerance test, and the absence of functional validation in human samples further limits the translational relevance of the findings. Third, the in vitro experiments were conducted using a single trophoblast cell line (HTR-8/SVneo), which may restrict the generalizability of the mechanistic conclusions, and these findings should therefore be interpreted with caution.
