Placental Antioxidant Defenses and Autophagy-Related Genes in Maternal Obesity and Gestational Diabetes Mellitus

Maternal obesity and gestational diabetes mellitus (GDM) are increasing worldwide, representing risk factors for both mother and child short/long-term outcomes. Oxidative stress, lipotoxicity and altered autophagy have already been reported in obesity, but few studies have focused on obese pregnant women with GDM. Antioxidant and macro/chaperone-mediated autophagy (CMA)-related gene expressions were evaluated herein in obese and GDM placentas. A total of 47 women with singleton pregnancies delivered by elective cesarean section were enrolled: 16 normal weight (NW), 18 obese with no comorbidities (OB GDM(–)), 13 obese with GDM (OB GDM(+)). Placental gene expression was assessed by real-time PCR. Antioxidant gene expression (CAT, GPX1, GSS) decreased, the pro-autophagic ULK1 gene increased and the chaperone-mediated autophagy regulator PHLPP1 decreased in OB GDM(–) vs. NW. On the other hand, PHLPP1 expression increased in OB GDM(+) vs. OB GDM(–). When analyzing results in relation to fetal sex, we found sexual dimorphism for both antioxidant and CMA-related gene expressions. These preliminary results can pave the way for further analyses aimed at elucidating the placental autophagy role in metabolic pregnancy disorders and its potential targetability for the treatment of diabetes outcomes.


Introduction
Maternal obesity (MO) is expanding exponentially worldwide to almost epidemic proportions, representing a significant risk factor for negative pregnancy outcomes, with short-and long-term consequences for both the child and the mother [1,2]. Gestational diabetes mellitus (GDM) is the most frequent complication of pregnancy, presenting in 2-14% of cases [3]. Up to 50% of obese women (OB) develop GDM during pregnancy, with an increased risk for the mother to later develop type 2 diabetes. Moreover, the adverse intrauterine environment of MO and GDM alters the programming of fetal metabolic Pregnant women were enrolled in the Obstetric Units of the L. Sacco Hospital and the V. Buzzi Children Hospital (ASST Fatebenefratelli-Sacco) in Milan.
The study was conducted in accordance with the Declaration of Helsinki and in compliance with all current Good Clinical Practice guidelines, local laws, regulations and organizations. The protocol was approved by the hospital ethical committee (Prot. N. 17739/2018). All participants gave their informed consent to collect personal data and biological samples.
Forty-seven Caucasian pregnant women with single-term pregnancies were prospectively recruited at elective cesarean section. The study population was divided according to pre-pregnancy body mass index (BMI, kg/m 2 ), following the 2009 Institute of Medicine (IOM) guidelines [22]:
According to the protocols in use, counselling was provided for dietary and lifestyle intervention to reach the appropriate gestational weight gain (GWG, kg) according to pregestational BMI and as recommended by the IOM [22]: NW women: 11.5 ≤ GWG ≤ 16; 2.
Gestational diabetes mellitus (GDM) was diagnosed in a subgroup of OB women (OB GDM(+), n = 13) according to the International Federation of Gynecology and Obstetrics (FIGO) guidelines, by the Oral Glucose Tolerance Test (OGTT-75 g) performed at 24-28 weeks of gestation [23,24]. The OGTT procedure requires a fasting blood glucose test (OGTT I) and then glycemia checks after 60 min (OGTT II) and after 120 min (OGTT III) from the administration of 75 g of anhydrous glucose dissolved in 300 mL of water. Blood glucose quantification at each time point was performed after venous blood withdrawal and standardized clinical biochemistry laboratory dosage by enzymatic spectrophotometric analysis (the hexokinase/glucose-6-phosphate dehydrogenase (g6pd) method). GDM was diagnosed for one or more glycemic curve values higher than 92,180,153 mg/dL [23].

Placental Tissue Sample Collection
Human placentas were collected immediately after elective cesarean section, cleaned of excess blood and then weighed after discarding the membranes and cord from the disc. Biometric measurements were performed as previously described [25,26]. Briefly, placental area was estimated by calculating the area of an ellipse from the diameters (D × d × π/4). Assuming constant density, placental thickness was obtained as weight divided by area. Placental efficiency was calculated as the ratio between fetal weight and placental weight.
Chorionic villi biopsies of 1 cm 3 were collected in different sites of the placental disc (central, median and peripheral) from the maternal side, after discarding the maternal decidua. Placental tissue samples were carefully washed in phosphate-buffered saline to eliminate excessive blood and conserved in RNAlater at −80 • C.

Gene Expression Analysis
Placental tissue samples were mechanically shredded in a Potter homogenizer with TRI Reagent and total RNA was extracted from the tissue homogenate using the columnbased RiboPure Kit (ThermoFisher Scientific Baltics UAB, Vilnius, Lituania) following the manufacturer's instructions. RNA concentration was determined spectrophotometrically by NanoDrop ND-1000 (NanoDrop Technologies, Wilmington, DE, USA), and then 2 µg of RNA were retrotranscribed using the GoScript Reverse Transcription Mix, Random Primers (Promega, Madison, WI, USA), following the manufacturer's instructions. For the real-time assay, 4 ng of cDNA were used for each sample amplification.
Gene expressions were obtained by a normalization strategy that allows using two different housekeeping genes simultaneously and takes into account the amplification efficiency of each assay in the specific investigated tissue. Specifically, E −∆Cq /NF was determined, where E is the amplification efficiency of each assay (calculated by a calibration curve) and NF is the normalization factor. This normalization factor was calculated from the geometric mean of two selected housekeeping genes [27], β-actin (Beta-actin) and YWHAZ (Tyrosine 3-Monooxygenase/Tryptophan 5-Monooxygenase Activation Protein Zeta). In more detail, for each sample and assay, the relative quantity Q was calculated: Q = E −∆Cq , where E is the amplification efficiency of each assay and ∆Cq = sample Cq-reference Cq (the reference is a chosen subject, i.e., the one showing the lowest Cq value).
The normalization factor (NF) was then calculated from the geometric mean (GM) of the relative quantities (Q) of the two endogenous control genes: NF = GM of b − actin and YW H AZ relative quantities (Q) for each sample GM of all the geometric means Lastly, for each sample, the expression of target genes was calculated by Q NF . Only Cq values with standard deviation ≤0.25 across triplicates were included in the statistical analysis.

Statistical Analysis
Maternal characteristics, placental and fetal data and gene expression values were compared among groups by One-way ANOVA or the Kruskal-Wallis test according to data distribution (assessed by the Kolmogorov-Smirnov test). In post hoc analyses, Tukey's HSD test and the Mann-Whitney U test with Bonferroni correction (thus considering statistical significance when p ≤ 0.017, as three groups were analyzed) were used.
Frequencies of placental efficiency, using its median value (6.9) as cut-off, and of fetal sex were evaluated among population groups by performing the chi-square test.
A Two-way ANOVA was conducted to explore the impact of maternal characteristics (pregestational BMI and hyperglycemia) and fetal sex on fetal and placental parameters and gene expression.
Comparisons between the two subgroups of placentas with different efficiency were performed using the independent sample t-test or the Mann-Whitney U test according to data distribution.
Correlations between variables were assessed using the Spearman rank order correlation. Differences and correlations were considered significant when p < 0.05. Analyses were performed using the statistical package SPSS, v.26 (IBM; Armonk, NY, USA).  Table 2 summarizes maternal, fetal and placental data of the three study groups. OB GDM(−) women were significantly younger than the other mothers. Following defined inclusion criteria, both obese groups had significantly higher pregestational BMI than NW, and OB GDM(+) showed significantly higher OGTT values than normoglycemic groups.

Clinical Data
Gestational weight gain was lower in OB women, as suggested by the IOM (11.5-16 kg for NW; 5-9 kg for OB).
No significant differences were observed among groups for gestational age and fetal weight.
Placental efficiency (fetal weight/placental weight) was lower in the OB groups vs. NW, and OB women placentas were heavier and thicker compared to NW, though not significantly.
Interestingly, when using the median value of placental efficiency (6.9) as a cut-off to split placentas into subgroups of different efficiency, we found a significant difference in their distribution among the three study groups (chi-square test: χ2 (1, n = 47) = 6.001, p = 0.049, Φ = 0.36). Indeed, most of the placentas from NW women (68.75%), but only 23.1% of those from OB GDM(+) women, were in the "more efficient" subgroup. Differently, OB GDM(−) placentas were equally distributed in the two subgroups of efficiency (50%) (Figure 1).
A Two-way ANOVA was conducted to explore the impact of maternal characteristics (pregestational BMI and GDM) and fetal sex on these (i.e., fetal weight, placental weight and efficiency) parameters. The interaction effect between maternal BMI/GDM and fetal sex was not statistically significant. In the whole population and within each BMI group, no significant differences were found between male and female fetuses for these parameters. A Two-way ANOVA was conducted to explore the impact of maternal chara (pregestational BMI and GDM) and fetal sex on these (i.e. fetal weight, placental we efficiency) parameters. The interaction effect between maternal BMI/GDM and fetal not statistically significant. In the whole population and within each BMI group, no cant differences were found between male and female fetuses for these parameters Interestingly, when comparing the three BMI groups separately in female fetu male fetuses, placentas of OB GDM(+) women resulted in being significantly hea less efficient than NW in the female fetuses subgroup (Table 3).

Expression of Antioxidant Defense Genes in Placenta
To assess the profile of redox homeostasis-related genes in the placentas of GDM(−) and OB GDM(+) women, we evaluated the gene expression level of t modulators of the detox machinery.
The mRNA levels of CATalase (CAT), Glutathione SynthetaSe (GSS), Glutath ductaSe (GSR) and Glutathione PeroXidase 1 (GPX1) genes are reported in Table   Figure 1. Distribution of placentas according to their efficiency for the three study groups (cut-off = 6.9, median value of placental efficiency). Data shown as a pie chart graph. Chi-square test: Interestingly, when comparing the three BMI groups separately in female fetuses and male fetuses, placentas of OB GDM(+) women resulted in being significantly heavier and less efficient than NW in the female fetuses subgroup (Table 3).

Expression of Antioxidant Defense Genes in Placenta
To assess the profile of redox homeostasis-related genes in the placentas of NW, OB GDM(−) and OB GDM(+) women, we evaluated the gene expression level of the main modulators of the detox machinery.
Placental expression was detected for all the analyzed antioxidant genes, showing lower levels of CAT, GPX1 and GSS in the normoglycemic obese women vs. normal weight, though not significantly (Table 4).
Interestingly, although CAT did not reach statistical significance in the comparison between groups, its placental levels were significantly and positively correlated with GSS gene expression (r = +0.6, p < 0.001).

Autophagy-Related Gene Expression
To assess the autophagic processes, the expression of the genes encoding for the main drivers of the autophagic pathways was analyzed. The mRNA levels of Unc-51-Like  Table 5.  Placental expression was detected for all the analyzed antioxidant genes, showing lower levels of CAT, GPX1 and GSS in the normoglycemic obese women vs. normal weight, though not significantly (Table 4).
Interestingly, although CAT did not reach statistical significance in the comparison between groups, its placental levels were significantly and positively correlated with GSS gene expression (r = +0.6, p < 0.001).

Autophagy-Related Gene Expression
To assess the autophagic processes, the expression of the genes encoding for the main drivers of the autophagic pathways was analyzed.  Table 5.  Significant results are reported in Figure 3.  Significant results are reported in Figure 3.
As regards macroautophagy, ULK1 showed significantly increased placental mRNA levels in the OB groups compared to NW ( Figure 3A).
BECN1 gene expression (Table 5) did not differ among the three groups. BECN1 mRNA levels were significantly and strongly correlated with GSS levels (r = +0.76, p < 0.001) in the whole population ( Figure 3B), and in each of its subgroups.
As regards chaperone-mediated autophagy (CMA), we measured the expression of genes that are involved at different levels in CMA activity or regulation.
HSC70 expression was significantly different among groups. Indeed, both OB GDM(−) and OB GDM(+) presented lower levels compared to NW ( Figure 3C).
LAMP-2A mRNA levels did not differ among groups (Table 5). In the NW subgroup, LAMP-2A placental levels were significantly and positively correlated with those of GSS (r = +0.8, p = 0.002).
PHLPP1 expression was significantly different among groups. Post hoc analysis showed that its levels were significantly increased in OB GDM(+) compared to OB GDM(−) placentas ( Figure 3D).
When considering the NW subgroup, PHLPP1 placental levels were significantly correlated with HSC70 expression (r = +0.53, p = 0.034), while this correlation was not significant in both OB GDM(−) and OB GDM(+) groups.   As regards macroautophagy, ULK1 showed significantly increased placental mRNA levels in the OB groups compared to NW ( Figure 3A).
BECN1 gene expression (Table 5) did not differ among the three groups. BECN1 mRNA levels were significantly and strongly correlated with GSS levels (r = +0.76, p < 0.001) in the whole population ( Figure 3B), and in each of its subgroups.
As regards chaperone-mediated autophagy (CMA), we measured the expression of genes that are involved at different levels in CMA activity or regulation.
HSC70 expression was significantly different among groups. Indeed, both OB GDM(−) and OB GDM(+) presented lower levels compared to NW ( Figure 3C).
LAMP-2A mRNA levels did not differ among groups (Table 5). In the NW subgroup, LAMP-2A placental levels were significantly and positively correlated with those of GSS (r = +0.8, p = 0.002).
PHLPP1 expression was significantly different among groups. Post hoc analysis showed that its levels were significantly increased in OB GDM(+) compared to OB GDM(−) placentas ( Figure 3D).
When considering the NW subgroup, PHLPP1 placental levels were significantly correlated with HSC70 expression (r = +0.53, p = 0.034), while this correlation was not significant in both OB GDM(−) and OB GDM(+) groups.
Finally, the expression of genes encoding proteins involved in the cell response to stressors and related to CMA activity was evaluated ( Table 6): Nuclear factor erythroid 2-Related Factor 2 (NRF2), Hypoxia-Inducible Factor 1 (HIF-1α), Vascular Endothelial Growth Factor (VEGF). NRF2, HIF-1α and VEGF mRNA levels did not differ among the three groups (Table 6). NRF2 placental levels resulted in being significantly correlated with LAMP-2A in the study population (r = +0.73, p < 0.001; Figure 4) and within each study group.
To understand if there was a possible joint effect of fetal sex and maternal metabolic characteristics (BMI and maternal glycemia) on levels of the analyzed genes, we performed additional analyses.
A Two-way ANOVA explored the interaction effect of maternal BMI/hyperglycemia and fetal sex on our molecular variables, but no statistical significance was found.
A comparison among the three BMI groups was then performed in female or male placentas, using non-parametric statistics due to the sample sizes of the fetal sex subgroups. NRF2, HIF-1α and VEGF mRNA levels did not differ among the three groups (T NRF2 placental levels resulted in being significantly correlated with LAMP-2 study population (r = +0.73, p < 0.001; Figure 4) and within each study group.
To understand if there was a possible joint effect of fetal sex and maternal m characteristics (BMI and maternal glycemia) on levels of the analyzed genes, formed additional analyses.
A Two-way ANOVA explored the interaction effect of maternal BMI/hyperg and fetal sex on our molecular variables, but no statistical significance was found A comparison among the three BMI groups was then performed in female or m centas, using non-parametric statistics due to the sample sizes of the fetal sex subgrou In female, but not in male, placentas, a lower expression was recorded for CA and  Macroautophagy genes did not show any difference when dividing by both BMI/GDM and fetal sex.

Discussion
This work focused on the description of placental function and molecular features related to cell response to oxidative stress and autophagy, in the context of metabolic deregulation characterizing maternal obesity and gestational diabetes. Indeed, placentas obtained from normal weight and obese women with or without gestational diabetes were analyzed for the expression of different gene subsets.
We found a decrease, though not significant, of placental antioxidant defense genes (CAT, GSS, GPX1) in OB GDM(−), but not in OB GDM(+), compared to NW. Interestingly, CAT was strongly correlated with placental efficiency and had significantly lower levels in less efficient placentas. Among genes playing an important role in macroautophagy, placental ULK1 was increased in OB groups compared to NW, while BECN1 presented a strong and positive correlation with GSS gene expression. The chaperone-mediated autophagy regulator PHLLP1 was reduced in OB GDM(−) placental tissues compared to NW and significantly increased in OB GDM(+) compared to OB GDM(−). PHLLP1 also correlated with HSC70 in NW placental tissues, but not in OB populations. Other CMA-related genes (NRF2, HIF-1a, VEGF) did not show significant differences in obesity and GDM.
When analyzing placental gene expression in relation to fetal sex, a sexual dimorphism was found for both antioxidant and CMA-related genes.
Importantly, in this study, placentas derived only from elective cesarean section, i.e., in the absence of labor, were included. This allowed avoiding any bias related to labor-induced oxidative stress, inflammation and changes in cell metabolism.
Furthermore, the clinical characteristics of the study population were deeply characterized: patients were carefully selected, and both NW and OB groups did not present any associated pathology, except for GDM in OB GDM(+). We excluded any maternal or fetal infection or autoimmune disease, maternal drug-alcohol abuse, fetal malformations, chromosomal disorders, preeclampsia and intrauterine growth restriction, all of which can affect the oxidative and the inflammatory status [28][29][30][31]. Moreover, only Caucasian women were selected, as different metabolic and oxidative characteristics have been identified and have been suggested to possibly contribute to the genetic component of complex disorders [32,33]. Our cases were also carefully matched to NW controls with similar characteristics except for BMI and GDM presence. Finally, all women included in this study were counseled with nutritional and lifestyle advice and recommendations on weight gain during pregnancy, and obese patients had regular specific checkups in a dedicated antenatal clinic with specific dietary indications. Therefore, we may assume that the reported findings can be related to metabolic dysfunctions related to increased BMI or hyperglycemic status [34].
Placental efficiency, calculated as fetal/placental weight, was lower in both obese groups compared to NW. When using the median value of placental efficiency as a cut-off, the percentage of the most efficient placentas (above the cut-off) decreased progressively, moving from the normal weight population to the obese and obese with GDM women. Since a decrease in placental efficiency has been described in relation to the oxidative stress induced by obesity [25,35], the expression of genes coding for proteins involved in cell redox homeostasis was analyzed in the different groups.
Expression of antioxidant defense genes (CAT, GSS, GSR, GPX1) tended to decrease in OB GDM(−) compared to NW. This trend supports an increase in oxidative stress and the evidence of altered placental metabolites involved in antioxidant defenses that has been previously described in the obese population [6,7,36].
Interestingly, expression levels of these genes in GDM women appeared to be more similar to those of NW. This might be due to different responses in the obese and diabetic impaired environment [37], leading to alternative mechanisms that result in an unchanged expression of the analyzed antioxidant genes. Indeed, previous studies showed that while placentas from obese women without comorbidities presented impaired mitochondrial biogenesis and respiratory chain enzyme activities [31], these were not altered in OB GDM(+) placentas, although mitochondrial morphological abnormalities were shown, accounting for mitochondrial dysfunctionality [7,38]. Moreover, we previously reported a differential alteration in antioxidant metabolites' content in OB placentas, depending on the presence or absence of GDM, showing different profiles in these two groups [36].
The CAT gene aroused particular attention because of its strong correlation with placental efficiency. In fact, in most efficient placentas, a significant increase in CAT expression was found, suggesting the importance of the redox potential in the maintenance of organ functionality [39].
Another cellular mechanism involved in the maintenance of cell homeostasis is autophagy. Importantly, reactive oxygen species are master inducers of autophagy. Thus, the activation of this process is a key part of the cellular response to oxidative stress because it allows scavenging of defective components before further damage [9,40].
Autophagy activity has been documented in placentas, but only a few studies have investigated it in obese and diabetic placentas [20,21,41]. Autophagy includes different degradation systems, mainly divided into unspecific and selective strategies. The former group is mainly represented by macroautophagy, which acts by including cytoplasmic material and organelles within an autophagosome whose fusion with a lysosome induces the degradation of the content [42]. On the other hand, the selective strategies such as chaperone-mediated autophagy degrade only selected targets containing a specific consensus sequence called KFERQ. This motif is usually recognized and bound by the chaperone Heat Shock Cognate 70 (HSC70); the complex is then delivered to the lysosome where the protein, upon its interaction with the LAMP-2A multimer, forming a channel on the lysosomal membrane, is unfolded and enters the lysosome where it will be digested by different proteases [43].
The Unc-51-Like Kinase 1 (ULK1) complex plays a central role in the macroautophagy initiation stage and is also involved in promoting autophagosome-lysosome fusion [44]. Interestingly, our study showed an increase in ULK1 expression in the placentas of OB groups, possibly due to the metabolic stress caused by obesity and activating macroautophagy. This phenomenon correlates with the already mentioned expression reduction in genes involved in the detoxification from oxygen reactive species, inducing macroautophagy.
Another main player of macroautophagy is BECliN 1 (BECN1), being the principal driver of this degradative pathway [45]. Even if no statistically significant changes were observed in the three populations, a strong and positive correlation with GSS gene expression was noted in placentas and was sustained in each subgroup, as a clue of the relation between autophagy and the cell response to oxidative stress [46].
On the other hand, the chaperone-mediated autophagy machinery showed some changes in the expression of its players. In detail, HSC70 expression decreased in obese mothers, and also in the presence of diabetes, compared to normal weight, indicating a potential impairment in CMA activity. LAMP-2A expression did not change among groups. However, LAMP-2A positively correlated with GSS in NW, accounting for the expected oxidative stress-induced CMA activity in the control group. This correlation was not significant in OB GDM(−) and OB GDM(+). Nevertheless, a positive correlation between LAMP-2A and HIF-1α (a well-known CMA target) emerged only in OB placentas, suggesting a different activity of CMA in obesity. However, this hypothesis needs further investigations.
CMA is finely regulated by several proteins, including Pleckstrin homology domain and leucine-rich repeat protein phosphatase 1 (PHLPP1) [47]. This phosphatase is able to induce CMA activity by favoring the stabilization of the LAMP-2A multimers in the lysosomal membrane. Moreover, this protein has been described to be also involved in insulin resistance mechanisms [48,49]. Indeed, it supports insulin resistance by reducing insulin-dependent signal transduction, leading to a decrease in glucose transport and hyperglycemia [50].
In our study, PHLLP1 expression was reduced in OB GDM(−) placental tissues compared to NW, while it resulted in being significantly increased in OB GDM(+) compared to OB GDM(−), supporting a correlative hypothesis between PHLPP1 activity and GDM, as already hypothesized in type 2 diabetes by Andreozzi and colleagues [49]. Moreover, a positive and strong correlation between PHLPP1 and HSC70 was observed in NW placental tissues, but not in OB populations, again supporting the hypothesis of an impaired CMA function in obesity. This interesting result could suggest a role for PHLPP1 in GDM placentas. However, whether it has a causative or a secondary role in relation to insulin resistance still needs to be investigated [48]. In fact, it has been described that PHLPP1 overexpression induces insulin resistance. However, an increase in insulin levels also resulted in the overexpression of PHLPP1 itself, being probably involved in a regulating feedback loop that is lost in diabetes, for reasons that are not yet understood.
The expression of other genes whose transcription is triggered by stressors or modulated by CMA activity (NRF2, HIF-1α, VEGF) was also studied and did not show significant differences in obesity and GDM. Among these, NRF2 plays an important role in driving the transcription of different genes involved in cell detoxification and is also related to CMA activation because its activity is directly related to LAMP-2A expression [51]. Data reported herein confirm the strong positive correlation between NRF2 and LAMP-2A in each group.
We also explored the possibility of significant differences in relation to fetal sex. Indeed, different responses to an adverse intrauterine environment have been previously extensively documented depending on fetal sex [52][53][54][55].
We analyzed fetal weight, placental weight and placental efficiency in males and females across BMI groups. When considering female newborns, the diabetic obese placentas resulted in being heavier, with a significantly decreased efficiency when compared to normal weight. These results are supported by our previous reports in diabetic obese placentas [7,55], thus suggesting interesting evidence of sexual dimorphism in placental biometrical and functional features [21].
An interesting result was also obtained for the expression of PHLPP1, which dropped significantly in females' placentas in the OB GDM(−) group, while it increased in both male and female placentas in OB GDM(+). However, PHLPP1 modulation seemed to have different consequences in the two sexes, with a significant decrease in VEGF expression only in males, which might be due to CMA activation. In fact, HIF-1α is both a CMA target and one of the main transcription factors driving VEGF expression ( Figure 6). Supporting the hypothesis of a CMA activation, an increase in the expression of its main player, LAMP-2A, has been detected in the same samples even if it did not reach statistical significance. This mechanism was not observed in females' tissues, providing clues about a dimorphic regulation and possibly different GDM outcomes depending on fetal sex. feedback loop that is lost in diabetes, for reasons that are not yet understood.
The expression of other genes whose transcription is triggered by stressors or modulated by CMA activity (NRF2, HIF-1α, VEGF) was also studied and did not show significant differences in obesity and GDM. Among these, NRF2 plays an important role in driving the transcription of different genes involved in cell detoxification and is also related to CMA activation because its activity is directly related to LAMP-2A expression [51]. Data reported herein confirm the strong positive correlation between NRF2 and LAMP-2A in each group.
We also explored the possibility of significant differences in relation to fetal sex. Indeed, different responses to an adverse intrauterine environment have been previously extensively documented depending on fetal sex [52][53][54][55].
We analyzed fetal weight, placental weight and placental efficiency in males and females across BMI groups. When considering female newborns, the diabetic obese placentas resulted in being heavier, with a significantly decreased efficiency when compared to normal weight. These results are supported by our previous reports in diabetic obese placentas [7,55], thus suggesting interesting evidence of sexual dimorphism in placental biometrical and functional features [21].
An interesting result was also obtained for the expression of PHLPP1, which dropped significantly in females' placentas in the OB GDM(−) group, while it increased in both male and female placentas in OB GDM(+). However, PHLPP1 modulation seemed to have different consequences in the two sexes, with a significant decrease in VEGF expression only in males, which might be due to CMA activation. In fact, HIF-1α is both a CMA target and one of the main transcription factors driving VEGF expression ( Figure 6). Supporting the hypothesis of a CMA activation, an increase in the expression of its main player, LAMP-2A, has been detected in the same samples even if it did not reach statistical significance. This mechanism was not observed in females' tissues, providing clues about a dimorphic regulation and possibly different GDM outcomes depending on fetal sex.

Conclusions
In this study, we presented preliminary data on the expression of antioxidant and autophagy-related genes in placentas from normal weight and obese mothers, with or without gestational diabetes mellitus. Despite the small population size, which limited the statistical significance of some modulations, our analysis showed intriguing results, reporting differential alterations depending on the presence/absence of GDM and on fetal sex. These results confirm how metabolism, nutrient availability and cellular mechanisms such as OxS responses and autophagy are strictly connected. Further studies will be nec-

Conclusions
In this study, we presented preliminary data on the expression of antioxidant and autophagy-related genes in placentas from normal weight and obese mothers, with or without gestational diabetes mellitus. Despite the small population size, which limited the statistical significance of some modulations, our analysis showed intriguing results, reporting differential alterations depending on the presence/absence of GDM and on fetal sex. These results confirm how metabolism, nutrient availability and cellular mechanisms such as OxS responses and autophagy are strictly connected. Further studies will be necessary to unravel processes below this observation. In detail, the role of PHLPP1 and chaperone-mediated autophagy activation in the etiopathology and also possibly in adverse outcomes of GDM might be suggested, deserving further investigations.
In the context of the current research, to our knowledge, this is the first report describing an increase in PHLPP1 expression in placentas in relation to GDM. New studies are needed to understand the mechanisms underlying this deregulation. Their comprehension could be useful for preventing the development of GDM. The elucidation of PHLPP1 involvement will pave the way for further analyses aimed at explaining its role in placental alterations in the context of metabolic disorders and its potential targetability for the treatment of negative consequences of diabetes.