A Decrease in DKK1, a WNT Inhibitor, Contributes to Placental Lipid Accumulation in an Obesity-Prone Rat Model¹

Abstract

Placenta, as the sole transport mechanism between mother and fetus, links the maternal physical state and the immediate as well as lifelong outcomes of the offspring. The present study examined the consequences of maternal obesity on placental lipid accumulation and metabolism. Pregnant obesity-prone (OP) and obesity-resistant (OR) rat strains were fed a control diet throughout gestation. Placentas were collected on Gestational Day 21 for mRNA and oxidative stress analysis, and frozen placental sections were analyzed for fat accumulation as well as beta-catenin and Dickkopf homolog 1 (Xenopus laevis) (DKK1) localization. JEG3 trophoblast cells were cultured in vitro to determine the relationship between DKK1 and lipid accumulation. Maternal plasma and placental nonesterified fatty acids and triglycerides (TG) were elevated in OP dams. Placental Dkk1 mRNA content was 4-fold lower in OP placentas, and a significant increase was noted in beta-catenin accumulation as well as in mRNA content of fat transport and TG synthesis genes, including Ppard (peroxisome proliferator-activated receptor delta), Slc27a1 (fatty acid transport protein 1; also known as Fatp1), Cd36 (cluster of differentiation 36; also known as fatty acid translocation [Fat]), Lipin1, and Lipin3. Significant lipid accumulation was found within the decidual zones in OP, but not OR, placentas, and thickness of the decidual and junctional zones was significantly smaller in OP than in OR placentas. Overexpression of DKK1 in JEG3 cells decreased lipid accumulation and mRNA content of PPARD, SLC27A1, CD36, LIPIN1, and LIPIN3. Our results demonstrate that DKK1 is regulating certain aspects of placental lipid metabolism through the WNT signaling pathway.

Introduction

In the United States, one-third of women of reproductive age are classified as obese, and since the mid-1990s, about half of women of childbearing age have been classified as overweight [1, 2]. Although the connection between fetal outcomes and maternal obesity is confirmed, the molecular mechanisms connecting maternal weight to fetal health remain undefined. As the principal link between mother and child, the placenta has become an important factor in unraveling these mechanisms. Human placental studies are limited, but several have shown that maternal obesity has a substantial effect on placental inflammation and oxidative stress [3–5]. In animals, maternal obesity in a baboon animal model leads to an increase in placental macrophages (CD68) as well as a decrease in placental system A transporter activity [6]. In a diet-induced obesity pregnant rat model, the placentas of heavier animals did not differ in size when compared to those of lean animals, but the ratio of fetal weight to placental weight was significantly decreased, suggesting that the placentas of the obese dams were less efficient in supporting fetal development than the placentas of the lean animals [7]. Gestational high-fat feeding with corresponding obesity development in a mouse model led to mixed placental inflammation, oxidative stress, and cellular necrosis and vasculopathy [8] as well as to an increase in placental glucose and amino acid transport [9].

Plasma triglyceride (TG) content is elevated during pregnancy in response to the changes in fuel demands by the mother and fetus [10]. Because TG are not transported across the placenta, their hydrolysis to free fatty acids (nonesterified fatty acids [NEFA]) provides a vital source of essential fatty acids for both the placenta and fetus [10]. Additionally, placental uptake of free fatty acids from maternal circulation is not a passive event and occurs primarily via transport proteins, such as fatty acid transport proteins 1–6 (FATP; official symbol SLC27A), the plasma membrane fatty acid-binding protein (FABP-pm), as well as a cluster of differentiation 36 (CD36; also known as fatty acid translocase [FAT]) [11]. Once these free fatty acids are taken up, they may be oxidized as an energy source for the placenta, transferred directly to the fetus, or esterified to TG before being delivered to the fetal circulation. Although the mechanisms behind placental fatty acid metabolism are not thoroughly understood, it has been suggested that nuclear receptors, especially the peroxisome proliferator-activated receptors (PPAR), which regulate lipid metabolism in other tissues, also participate in placental fat metabolism [12]. Whereas an adequate supply of fatty acids is essential for placental development, excessive lipid accumulation may be potentially detrimental, as is the case in other nonadipose tissues.

The wingless-type MMTV integration site family (WNT) signaling pathway has a vital role in the development of the placenta. Of the 19 known WNT ligands, 14 were found in the placenta, and of the 10 frizzled (FZD) receptors, eight were detectable in placental tissues [13]. Additionally, β-catenin, the key mediator of canonical WNT signaling, was implicated in trophoblast adhesion, survival, and differentiation [14–16]. WNT7A and TCF/LEF1 are required for chorioallantoic fusion [17, 18], and WNT2 [19], WNT3A [20], WNT5A, WNT10B [21], and SFRP4 [22, 23] may be essential for proper placental vascularization and growth. Dickkopf homolog 1 (Xenopus laevis) (DKK1), a potent inhibitor of WNT signaling, also is involved in trophoblast invasion and migration [24]. Whereas studies have shown that WNT signaling has a vital role in regulating adipogenesis and lipid accumulation [25], little is known about the metabolic action of the WNT signaling pathway in the placenta. The STARD7 gene encodes a StAR-related lipid transfer protein involved in intracellular transport and metabolism of lipids. STARD7 mRNA is expressed in trophoblastic tissues and tumor cell lines. A study in JEG3 trophoblasts showed that beta-catenin and TCF4 could activate the STARD7 gene [26], documenting a potential connection between WNT signaling and lipid metabolism in placental cells in vitro.

The present study utilized a well-defined, obesity-prone (OP) rat model [27–30], which is a unique and novel approach that allowed us to observe the consequences of excessive body weight on the placenta without the effect of high-fat intake often observed in other models of diet-induced obesity. Although many unknowns remain in this model of obesity development, it appears that these rats have many of the characteristics of humans who are obese, which makes this an appropriate model for the present study. Therefore, studying the events occurring in obesity-resistant (OR) and OP rats during pregnancy will be valuable from the mechanistic perspective and help to further classify the physiology of this unique model. The aim of the present study was to reveal the potential role of DKK1 in placental lipid accumulation by utilizing this OP rat model. We demonstrate that gestational obesity resulted in significant placental lipid accumulation in this model and that this process was potentially regulated, in part, by the WNT inhibitor, Dkk1, resulting in an aberrant activation of placental WNT signaling. Additionally, PPAR delta (PPARD) has been shown to be a WNT target in colon tissue [31], and our cell culture model demonstrated that it may be involved in the WNT-mediated placental lipid accumulation we observed.

Materials and Methods

Animal Model and Dietary Treatment

Timed-pregnant rats were obtained from Charles River Laboratories on Gestational Day 2. These rats originated from a line of Crl:CD (SD) rats (Charles River Laboratories, with two lines being developed from the outbred colony: the OP (CD) rats become obese when fed high-fat diets, and the OR (CD) rats do not become obese when fed high-fat diets. Five pregnant rats from each strain group were fed a control diet (64% CHO, 20% Pro, and 16% fat) based on a standard AIN93G diet ad libitum until Gestational Day 20, when they were fasted overnight and then underwent cesarean delivery to collect placentas. All samples were immediately frozen in liquid nitrogen and stored at −80°C for further analysis. Fetal and placental weights were also recorded. Additionally, trunk blood from each dam was collected and immediately spun down to isolate serum (using noncoated glass tubes) and plasma (using ethylenediaminetetra-acetic acid [EDTA]-coated glass tubes).

Maternal Plasma and Serum Analysis

To determine plasma TG levels, plasma samples were thawed on ice and analyzed via the Infinity Triglycerides Liquid Stable Reagent (Thermo Fisher Scientific) following the manufacturer's protocol and using a commercially available standard reference kit (Matrix Plus Chemistry Reference Kit; Verichem Laboratories). Maternal serum NEFA was analyzed utilizing a commercially available kit (HR-2 Series; Wako Diagnostics). For TG and NEFA analysis, within-assay coefficient of variation (CV) was acceptable when less than 5%. Maternal plasma leptin was analyzed using the Rat Leptin ELISA Kit (catalog no. EZRL-83K; Millipore) according to manufacturer's instructions. For leptin analysis, results were considered to be acceptable with a CV of less than 15%, as recommended by the manufacturer; the results fell within the recommended detection limits.

Lipid and Lipid Peroxidation Analysis

Placental samples (50 mg) were ground using a mortar and pestle with liquid nitrogen and mixed with 0.15 ml of saline (0.9% w/v NaCl) as previously reported [32]. Homogenized samples were quickly frozen in liquid nitrogen and kept at −70°C until analysis. The samples were quickly thawed at 37°C and diluted five times in saline to 0.75 ml. Twenty microliters of the diluted samples were incubated with 20 μl of 1% deoxycholate at 37°C for 5 min, and 10 μl of the samples were used to analyze either TG or NEFA. TG content was analyzed via the Infinity Triglycerides Liquid Stable Reagent following the manufacturer's protocol and using a Matrix Plus Chemistry Reference Kit. Placental NEFA concentration was determined using a commercially available kit (HR-2 Series). Lowry assay was performed to determine the protein concentration of each sample, which was then used to normalize the TG and NEFA concentrations, with the data presented as the amount (mg) of either TG or NEFA per gram of placental tissue.

Thiobarbituric acid-reactive substances (TBARS) assay (catalog no. 10009055; Cayman Chemical) was performed per the manufacturer's protocol to determine the concentration of malondialdehyde (MDA), a product of lipid peroxidation, within placental tissues as well as in JEG3 cells following NEFA treatment. Data are presented as MDA values normalized to protein content of samples as measured by the Lowry assay.

Placental Oil Red O Staining

Frozen placentas from five dams (n = 2 from each dam) in each strain group were embedded in Tissue-Tek O.C.T. compound (VWR) on dry ice, and two sections were cut from each sample at a thickness of 7 μm (Cryostat CM3050S; Leica) and mounted on glass slides. Slides were then briefly washed in water and 60% isopropanol and then stained for 10 min in a 60% working Oil Red O solution (catalog no. 1277; Newcomer Supply). Slides were then washed well in water and 60% isopropanol, counterstained for 20 sec in hematoxylin, washed again in running water, and coverslipped using an aqueous mounting media (Crystal Mount; Sigma-Aldrich, St. Louis, MO). Images were obtained using a NanoZoomer Slide Scanner and NDP View software (Hamamatsu). All images were analyzed by two unbiased observers who were blinded to the identity of the samples. The observers were asked for their input on both the decidual/junctional and labyrinthine zones, and the figures show representative images of samples from both groups.

Placental RNA Isolation and RT-PCR Analysis

Frozen placentas from five dams (n = 2 from each dam) from each strain were randomly chosen for mRNA analysis and matched to the placentas used for NEFA, TG, and TBARS analyses. Total RNA was isolated using the GenElute Mammalian Total RNA Miniprep Kit (Sigma-Aldrich) and quantified using Nano Drop Spectophotometer ND-1000. Two micrograms of cDNA were synthesized in a 20-μl reaction volume using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) with random primers and a thermal cycler (Applied Biosystems 2700) with the following program: 25°C for 10 min, 37°C for 120 min, 85°C for 5 sec, and a 4°C hold. Real-time PCR was performed using 25 ng of cDNA as the template, SYBR Green PCR Master Mix (Applied Biosystems), and 5 μM of each forward and reverse primer (Table 1) in the 7300 Real-Time PCR System (Applied Biosystem) with the following program: 95°C for 10 min, 95°C for 15 sec, 60°C for 1 min, 95°C for 15 sec, 55°C for 1 min, and 95°C for 15 sec, with 40 cycles of steps 2 and 3. A serial dilution was used to create a standard curve for quantification, and a dissociation curve was analyzed following each reaction.

Table 1

Primer information.

Gene	Species	Forward sequence and location	Reverse sequence and location	Ensembl ID
Dkk1	Rat	ATGCCCTCTGACCACAGCCATT (+439)	CACCGTGGTCATTGCCAAGGT (+517)	ENSRNOT00000015771
	Human	GATCATAGCACCTTGGATGGG (+635)	GGCACAGTCTGATGACCGG (+742)	ENST00000373970
Ppard	Rat	GCTCACCGAGTTCGCCAAGAAC (+1024)	CCTCATGCACGCCGTACTTGAG (+1112)	ENSRNOT00000042539
	Human	CCCCACGTCTGTCCTCCTTTCTTAT (+1954)	TGTGCAAAAGCAGAGGTCCTGTTC (+2039)	ENST00000360694
Slc27a1	Rat	CCACCATTCCTACAGCAT (+996)	TGCTGAGTGGTAGAGAGGTA (+1062)	ENSRNOT00000024659
	Human	AACCTCAGAGGAACCCGTGCCT (+2137)	TGAAAAGCAGGGAGAGGAGGC (+2206)	ENST00000252595
Cd36	Rat	AGTGCTCTCCCTTGATTCTGC (+151)	GAGCCCACAGTTCAGATCACA (+213)	ENSRNOT00000061687
	Human	CAAGAAAAATGGGCTGTGACCG (+247)	AACACAGCCAGGACAGCACCAAT (+319)	ENST00000447544
Dgat1	Rat	TCAATCTGTGGTGCCGCCAG (+708)	CCCACTGACCTTCTTCCCTGCA (+775)	ENSRNOT00000039795
	Human	GTGGCTTCAGCAACTACCGT (+507)	CGGGCATTGCTCAAGATC (+573)	ENST00000332324
Lipin1	Rat	AGCCTGGTAGATTGTCAGAG (+738)	GAGGACAAGAGCTAGAGAGAAC (+804)	ENSRNOT00000005863
	Human	GCGTAAAATGTCCCAAGCAGCC (+2761)	CGGGGAGACCTATCCTTTAATGGG (+2836)	ENST00000256720
Lipin2	Rat	CCAGTTACCCACAGACAGTGTGCC (+792)	CTCTCGGATGGCTTCACCTCCA (+855)	ENSRNOT00000020476
	Human	ACCACCTATCCCCAGACAGCGT (+936)	CAGGCTCTCCGCAGGTTTCA (+1004)	ENST00000261596
Lipin3	Rat	CCCTGAAGAGAAGCCAGCACCT (+1107)	TCAGAGTCCAGGGAGGGCAGAT (+1176)	ENSRNOT00000067531
	Human	CGGCACCATCACCAAGTCAGAT (+2029)	TGGTGTGTCCAGTCTTTCCCCA (+2097)	ENST00000373257
Rpl7a	Rat	GAGGCCAAAAAGGTGGTCAATCC (+64)	CCTGCCCAATGCCGAAGTTCT (+127)	ENSRNOT00000006754
	Human	TTTGGCATTGGACAGGACATCC (+145)	AGCGGGGCCATTTCACAAAG (+208)	ENST00000323345

Placental Nuclear Extraction and Western Blot Analysis

Nuclear β-catenin translocation was analyzed as a critical indicator of canonical WNT signaling [33]. To isolate nuclear protein, 200 mg of placental tissue from four placentas in each group (n = 1 from each dam) were homogenized and then rinsed twice in ice-cold PBS. The samples were then incubated for 5 min in a buffer containing 10 mmol/L of Hepes (pH 7.9), 10 mmol/L of KCl, 0.1 mmol/L of EDTA, and 1 mmol/L of dithiothreitol and washed in this buffer twice. The samples were then washed and resuspended in a buffer containing 20 mmol/L of Hepes (pH 7.9), 0.4 mol/L of NaCl, 1 mmol/L of EDTA, and 10% glycerol. The resulting suspension was sonicated three times for 30 sec each time at 10 W and then centrifuged at 20 000 × g for 10 min at 4°C to collect the supernatant. Protein concentration was determined using the Lowry method, and 15 μg of protein were diluted and loaded onto a 7.5% SDS-PAGE gel and resolved by electrophoresis at 80 V for 2 h. Gels were then transferred to a polyvinylidene fluoride membrane for 1 h at 0.3 Amps. The membranes were blocked in a solution of 10% milk/Tris-buffered saline with Tween 20 (TBST) for 1 h, incubated in 1:1000 dilution of primary antibody against β-catenin (Santa Cruz Biotechnology) or LaminA (Santa Cruz Biotechnology) in 10% milk/TBST at room temperature for 3 h, washed five times for 5 min each time in 5% milk/TBST, incubated in 1:10 000 dilution of secondary anti-rabbit IgG in 5% milk/TBST at room temperature for 1 h, and washed again five times for 5 min each in 1% Milk/TBST. Signal was detected using SuperSignal West Dura Extended Duration Substrate (product no. 34076; Thermo Scientific) and quantified with Bio-Rad ChemiDoc imaging system.

Placental Immunofluorescence

Frozen placentas from five dams (n = 1 from each dam) in each strain group were embedded in Tissue-Tek O.C.T. compound on dry ice, and two sections were cut from each sample at a thickness of 10 μm and mounted on glass slides. Samples were allowed to air-dry and then fixed in ice-cold acetone for 10 min, followed by 10 min of washing in PBS. Samples were then blocked in Image-iT FX (catalog no. I36933; Invitrogen) for 45 min, washed three times for 5 min each time with PBS, incubated in primary antibody for 1.5 h (1:150 for β-catenin [catalog no. sc-1496; Santa Cruz] and 1:50 for DKK1 [catalog no. ab61034; AbCam], counterstained with Alexa Fluor 647 (catalog no. A-21245; Invitrogen) secondary antibody at a 1:200 dilution for 1 h at room temperature, and then washed three times for 5 min each time with PBS. The Alexa Fluor 647 antibody was chosen because it eliminates the nonspecific autofluorescence observed with blood-infiltrated tissues. Before beginning the experiment, other placental tissues were first tested to confirm that both antibodies were specific when using this tissue and protocol. Nuclei were stained with Hoechst (catalog no. 33342; Invitrogen) for 15 min, followed by washing twice for 5 min each time with PBS. Slides were then coverslipped using ProLong Gold (catalog no. P36934; Invitrogen), stored at room temperature overnight, and viewed using the Zeiss Axiovert 200M with the Apotome Structured Illumination Optical Sectioning System. Quantification of the fluorescence intensity from the red channel (β-catenin or DKK1 protein) was performed using the ImageJ software (National Institutes of Health; http://rsb.info.nih.gov/ij/) on placentas within both decidual/junctional and labyrinthine zones, with eight areas quantified for each region and each placenta and one placenta being measured from three dams in each strain. The appropriateness of the areas selected for quantification was determined by also observing the signal from the blue channel (nuclear stain) to assure that tissue morphology was similar between samples.

Cell Culture and Treatment

Extravillous trophoblasts function as anchors connecting maternal and fetal interfaces, which is the reason for the frequent use of JEG3 cells in placental functional studies [34–36]. The cell line has also been utilized in numerous studies to model syncytiotrophoblasts, or the direct layer of villous chorionic villi exposed to the maternal blood supply [37]. In the present study, human JEG3 trophoblast cells were obtained from ATCC. Minimum essential medium (MEM) was purchased from SCS Cell Media Facility at the University of Illinois at Urbana-Champaign. Fetal bovine serum (FBS) and other cell culture media supplements were purchased from Mediatech. Cells were maintained at 37°C in a 5% CO₂/95% air incubator and were cultured and maintained in MEM media containing 10% (v/v) FBS, 100 IU/ml of penicillin, 100 μg/ml of streptomycin, and 250 ng/ml of amphotericin B.

Cells at passage 3 were seeded in a six-well plate at 0.2 million cells/well on top of 22- × 22-mm coverslips for Oil Red O staining and immunofluorescence, at 8 million cells/dish in 100-mm dishes for TBARS assay, at 1 million cells/dish in 60-mm dishes for RNA isolation in triplicate, and at 0.05 million cells/well in 24-well plates for Oil Red O quantification. For the initial experiment using wild-type (WT) cells, cells were treated with either 0 or 400 μM Fatty Acid Supplement containing 2 mol of linoleic and 1 mol of oleic acid per mole of albumin (catalog no. F7175; Sigma-Aldrich) in triplicate. This fatty acid mixture and concentration have been utilized in previous studies of trophoblast fat accumulation [38–40], and the single treatment concentration of 400 μM was chosen based on results of concentration gradient studies (data not shown) showing that Dkk1 mRNA was significantly decreased by 50, 100, and 200 μM fatty acids when compared to 0 μM, with the most significant decrease following the 400 μM treatment. Treatment media without fatty acids was supplemented with additional bovine serum albumin to maintain the same osmolarity.

Overexpression of DKK1 in JEG3 Cells

Transfection-ready DKK1 DNA inserted into a pCMV6-XL5 vector was purchased from OriGene (catalog no. SC303946; mRNA National Center for Biotechnology Information Reference Sequence NM_012242.2). OriGene has sequenced this clone, and we checked by restriction enzyme digest to confirm the identity. The empty pCMV vector was cloned and used as a transfection control. Cells were grown as described above and transfected with either the DKK1-containing vector or the empty pCMV. The transfection mixture was prepared with each vector (1.5 and 0.5 μg/well for 6- and 24-well plates, respectively), serum-free media, and Superfect Transfection Reagent (catalog no. 301307; Qiagen). After being transfected with the resulting mixture for 3 h at 37°C in a 5% CO₂/95% air incubator, cells were replenished with fresh complete MEM for 48 h to allow the expression of both DKK1 mRNA and protein. Cells were then treated with either 0 or 400 μM Fatty Acid Supplement containing 2 mol of linoleic and 1 mol of oleic acid per mole of albumin.

Analysis of WNT Components and Lipid Accumulation in Normal and DKK1-Overexpressed JEG3 Cells Following NEFA Treatment

After 24 h of NEFA treatment, cells were collected into TRI Reagent (catalog no. T9424; Sigma-Aldrich) to isolate RNA per the manufacturer's protocol. Methodologies for RNA quantification and mRNA analysis in JEG3 cells are identical to those described for placental tissues.

After 24 h of NEFA treatment, cells were fixed in 4% paraformaldehyde (catalog no. P6148; Sigma-Aldrich) in PBS for 1.5 h for Oil Red O staining. Each well was then washed with PBS, and the same protocol was followed as described for tissue Oil Red O staining. Following staining, coverslips containing the stained cells were mounted on glass slides and visualized as described for tissue samples. For quantification of the Oil Red O staining, cells grown in 24-well plates were also fixed in 4% paraformaldehyde as discussed above, washed three times with water, stained with Oil Red O stain for 20 min, and washed three times very well with water. The dye was extruded with 100% isopropanol for 10 min, and the resulting supernatant was transferred to 96-well plates and analyzed spectrophotometrically at an optical density of 490 nm.

For the initial experiment in JEG3 cells, data are presented using a modified y-axis to account for the background from the visible Oil Red O stain remaining on the walls of all wells. For the transfection experiment, data are presented as the percentage change between the pCMV and DKK1-transfected cells following 400 μM NEFA treatment. The lack of a transfection effect on cell number was confirmed by performing Lowry analysis on adjacent cells exposed to the identical treatment.

The protocol for immunofluorescence staining of β-catenin protein in JEG3 cells was similar to that used for the tissue samples, except that coverslips with cells were first mounted on glass slides before proceeding with the staining protocol. Quantification of the fluorescence intensity from the red channel (β-catenin protein) was performed using the ImageJ software, with eight areas quantified for each of the triplicates in each group. The appropriateness of the areas selected for quantification was determined by also observing the signal from the blue channel (nuclear stain) to assure that cell density was similar between samples.

Statistical Analysis

Results are reported as the mean ± SEM (n = 5) for maternal food intake, weight gain, and plasma analysis. Fetal and placental weight analysis was performed by taking the average of weights by litter (n = 5). The fetal:placental ratio was obtained for every individual offspring and placenta, and the average was taken for all litters (n = 5). For placental mRNA analysis, Oil Red O staining, TBARS, lipid content, and placental thickness measurements, five dams are represented, with two samples from each dam. For determination of decidual/junctional thickness, from five to seven measurements were taken per placenta using NDP View software from samples previously stained and scanned for Oil Red O content. For placental immunofluorescence, n = 5 (n = 3 for fluorescence quantification), and for all cell culture experiments, n = 3.

Maternal weight and food intake were analyzed using the repeated-measures ANOVA with SAS software (SAS Institute, Inc.). For all remaining data, Student t-test was used to test differences between means, and means were considered to be significantly different when P < 0.05.

Statement of Ethics

We certify that all applicable institutional and governmental regulations regarding the ethical use of animals were followed during this research (University of Illinois Institutional Animal Care and Use Committee approval no. 09112).

Results

Maternal Gestational Characteristics

Gestational food intake (Fig. 1A) was not different between OR and OP dams throughout gestation; however, OP dams were significantly (P < 0.01) heavier than OR dams throughout gestation (Fig. 1B) and had a significantly (P < 0.01) lower ratio of food intake to body weight throughout gestation (Fig. 1C). Total gestational weight gain was not different between OR and OP dams, and neither was mass gain (calculated by subtracting fetal and placental weights from total gestational weight gain) (Table 2). At the time of cesarean delivery on Gestational Day 21, OP dams had significantly higher levels of fasting plasma TG (Fig. 2A) (P < 0.01) as well as serum NEFA (Fig. 2B) (P < 0.01). Additionally, fasting plasma leptin was significantly (P < 0.01) higher in OP dams when compared to OR dams (Table 2).

Table 2

Maternal and fetal observations.

Variable	Straina
	OR	OP
Total gestational weight gain (g)	124.20 ± 2.67	128.78 ± 6.36
Total mass gain (g)	74.98 ± 4.74	78.63 ± 6.93
Birth weight (g)	4.14 ± 0.24	3.44 ± 0.08#
Placental weight (g)	0.50 ± 0.006	0.49 ± 0.011
Fetal:placental ratio	8.26 ± 0.41	7.11 ± 0.16#
Litter size	10.6 ± 0.68	12.80 ± 0.58*
Products of conception (g)	49.23 ± 3.17	50.16 ± 1.26
Leptin (ng/ml)	0.91 ± 0.11	3.01 ± 0.19#

^a

Results are reported as means ± SEM; n = 5 litters.

*

P < 0.05 when compared to OR

^#

P < 0.01 when compared to OR.

Fig. 1

Gestational food intake (A), body weight (B), and food intake:body weight ratio (C) curves in OR and OP dams throughout pregnancy. Results are reported as the mean ± SEM (n = 5 litters). *P < 0.01 when compared to OR dams after determining an overall difference throughout gestation by repeated-measures ANOVA.

Fig. 2

Fasting plasma TG (A) and NEFA (B), as well as placental TG (C) and NEFA (D), in OR and OP dams at the time of cesarean delivery on Gestational Day 21. Results are reported as the mean ± SEM (n = 5 litters). *P < 0.05 when compared to OR dams, ^#P < 0.01 when compared to OR dams.

Offspring Observations and Placental Lipid Analysis

At the time of cesarean delivery on Gestational Day 21, offspring of OP dams were significantly (P < 0.01) lighter when compared to offspring of OR dams (Table 2), and the size of OP litters was significantly (P < 0.05) larger than that of OR litters (Table 2), with no difference in the weight of products of conception (calculated as the sum of all placentas and offspring from each dam) (Table 2). Placental weight was not different between the two strains, but the fetal weight:placental weight ratio was significantly (P < 0.01) lower in OP dams when compared to OR dams (Table 2).

Placental TG and NEFA content was analyzed to determine whether an increase of circulating TG and NEFA corresponded to their increased accumulation within the placenta. OP dams had significantly (P < 0.05) higher placental TG (Fig. 2C) and NEFA (Fig. 2D) content when compared to OR dams.

Because OP placentas had increased accumulation of TG and NEFA when compared to OR placentas, we wanted to characterize the localization of lipid within these tissues. Oil Red O staining of OR and OP placentas suggested that OP placentas (Fig. 3B) had visibly more lipid associated with the decidual and junctional zones (arrows in Fig. 3B), whereas OR placentas (Fig. 3A) had less lipid accumulation within all zones. Additionally, the thickness of the decidual and junctional zones combined was significantly (P < 0.01) smaller in OP placentas when compared to OR placentas (Fig. 3C).

Fig. 3

Representative image of Oil Red O staining of placental fat accumulation in placentas of OR (A) and OP (B) dams (Dz, decidual zone; Jz, junctional zone; Lz, labyrinthine zone). Arrows indicate areas of significant red staining. Measurement of the decidual/junctional thickness (C), mRNA analysis (D), and relative TBARS measurement (E) were performed in placentas of five dams (n = 2 from each dam) from each strain, with from five to seven measurements per placenta for thickness measurement. Results are reported as the mean ± SEM. *P < 0.01 when compared to OR dams.

The OP placentas had visible lipid accumulation within all regions, so we wanted to determine whether the mRNA expression of lipogenic genes was also increased in these placentas. The mRNA content of several fatty acid transporters, including Slc27a1 (Fatp1) and Cd36 (Fat), were significantly (P < 0.01) higher in OP dams, as was the mRNA content of genes associated with TG synthesis, including Ppard, Dgat (diglyceride acyltransferase1), Lipin1, and Lipin3, without a significant change in Lipin2 (Fig. 3D).

Because lipid accumulation has been shown to be accompanied by an increase in oxidative stress, we utilized TBARS to determine whether this was the case in OP placentas. Despite the increase in lipid content within the OP placentas, no significant difference was found in the MDA concentration between OR and OP placentas (Fig. 3E), indicating that the changes we observed likely had little to do with oxidative stress.

Placental Dkk1 and β-Catenin Content

Our PCR analysis showed that Dkk1 mRNA content was significantly (P < 0.01) lower (>4-fold) in placentas of OP dams when compared to those of OR dams (Fig. 4A). Because of this, we wanted to confirm that the protein content of DKK1 was also decreased. Additionally, we wanted to determine whether the decrease in DKK1 was also associated with an increase in β-catenin, an indicator of aberrant WNT signaling. The amount of DKK1 protein within the decidual/junctional (Fig. 4, B–D) and labyrinthine (Fig. 4, E–G) zones was significantly (P < 0.01) decreased in OP placentas when compared to OR placentas.

Fig. 4

Whole-tissue Dkk1 mRNA content (A) in placentas of OR and OP dams. Representative images of DKK1 protein localization in the decidual/junctional zones in placentas of OR (B) and OP (C) dams with quantification of red fluorescence intensity relative to OR dams (D), as well as in the labyrinthine zone of placentas of OR (E) and OP (F) dams with quantification of red fluorescence intensity relative to OR dams (G), are shown. Red represents DKK1 protein, and blue represents nuclear staining (n = 5 [with 10 samples total] for mRNA analysis, n = 5 for staining, and n = 3 placentas, with eight images per placenta for fluorescence quantification). Results are reported as the mean ± SEM. *P < 0.01 when compared to OR dams. Original magnification ×20 (B, C, E, and F).

Nuclear β-catenin was significantly (P < 0.01) higher in OP placentas when compared to OR placentas when whole homogenates were analyzed (Fig. 5, A and B). Additionally, immunofluorescence analysis confirmed that OP placentas (Fig. 5, D and G) had a visibly increased amount of β-catenin protein within all regions when compared to OR placentas (Fig. 5, C and F), and this increase was significant (P < 0.01) following quantification of the signal from decidual/junctional (Fig. 5E) and labyrinthine (Fig. 5H) zones.

Fig. 5

Representative blot image of nuclear β-catenin and LaminA protein content in whole placental homogenates from OR and OP dams (A) as well as quantification of normalized nuclear β-catenin protein content (B). Representative image of β-catenin protein localization in the decidual/junctional zones in placentas of OR (C) and OP (D) dams with quantification of red fluorescence intensity relative to OR dams (E), as well as in the labyrinthine zone of placentas of OR (F) and OP (F) dams with quantification of red fluorescence intensity relative to OR dams (H), are also shown. Red represents β-catenin, and blue represents nuclear staining (n = 4 for nuclear protein blotting, n = 5 for staining, and n = 3 placentas, with eight images per placenta for fluorescence quantification). Results are reported as the mean ± SEM. *P < 0.01 when compared to OR dams. Original magnification ×20 (C, D, F, and G).

Lipid Accumulation and DKK1 mRNA Content in NEFA-Treated JEG3 Cells

Treating JEG3 cells with 400 μM NEFA resulted in a significant (P < 0.01) decrease in the content of DKK1 mRNA when compared to the 0 μM treatment (Fig. 6A). Additionally, obvious lipid accumulation was observed in cells treated with 400 μM NEFA (Fig. 6C) when compared to the 0 μM treatment (Fig. 6B), as shown by Oil Red O staining and quantified by extruding the Oil Red O dye (Fig. 6D) (P < 0.01), without a significant difference in the MDA concentration (TBARS) between the two treatments (Fig. 6E).

Fig. 6

DKK1 mRNA content in JEG3 cells following treatment with 400 μmol/L of NEFA (A), with representative images of Oil Red O staining in cells following 0 μM (B) and 400 μM (C) NEFA treatment and quantification of the Oil Red O staining (D), as well as the relative TBARS measurement from the same experiment (E) (n = 3 for mRNA and staining analyses). Results are reported as the mean ± SEM. *P < 0.01 when compared to the 0 μM treatment.

Lipid Accumulation and WNT Signaling in NEFA-Treated JEG3 Cells Overexpressing DKK1

Because we observed that DKK1 mRNA was down-regulated and lipid accumulation increased in NEFA-treated JEG3 cells, we wanted to determine whether an overexpression of DKK1 would prevent the accumulation of fat in these cells following NEFA treatment. The overexpression of DKK1 was confirmed by testing mRNA content of the DKK1 gene following transfection and NEFA treatment (Fig. 7A). Overexpression of DKK1 and NEFA treatment (Fig. 7C) resulted in a decrease of β-catenin protein when compared to the empty vector-transfected WT cells (Fig. 7, B and D). Oil Red O staining showed that cells overexpressing DKK1 (+DKK1 cells) (Fig. 8B) had visibly fewer bright-red lipid droplets than WT-transfected cells (Fig. 8A) following NEFA treatment, suggesting a decrease in lipid accumulation in these cells. This observation was confirmed by quantifying the Oil Red O staining, which showed a significant (P < 0.01) decrease of greater than 10% in fat accumulation in +DKK1 cells following NEFA treatment when compared to WT-transfected cells (Fig. 8C).

Fig. 7

DKK1 mRNA content following DKK1 overexpression in JEG3 cells treated with 400 μM NEFA (A) as well as representative images of β-catenin protein localization in WT (following transfection of empty pCMV; B) and +DKK1 (C) JEG3 cells following treatment with 400 μM NEFA with quantification of red fluorescence intensity relative to WT cells (D). Red represents β-catenin protein, and blue represents nuclear staining (n = 3 each for mRNA analysis, staining, Oil Red O quantification, and fluorescence quantification, with eight images per triplicate). Results are reported as the mean ± SEM. *P < 0.05 when compared to WT cells, ^#P < 0.01 when compared to WT cells.

Fig. 8

Representative image of Oil Red O staining in WT (A) and +DKK1 (B) JEG3 cells following treatment with 400 μM NEFA and quantification of the Oil Red O staining (C) and mRNA analysis of +DKK1 JEG3 cells following NEFA treatment (D) (n = 3 each for mRNA analysis and Oil Red O quantification). Results are reported as the mean ± SEM. *P < 0.05 when compared to WT cells, ^#P < 0.01 when compared to WT cells. Bar = 80 μm.

When compared to WT-transfected cells, +DKK1 cells also had a significant decrease in the mRNA content of PPARD (P < 0.01), SLC27A1 (FATP1) (P < 0.05), DGAT1 (P < 0.01), LIPIN2 (P < 0.05), and LIPIN3 (P < 0.05) without a significant change in CD36 (Fat) and only a trend of decrease of LIPIN1 (P = 0.059) (Fig. 8D).

Discussion

To our knowledge, the present study is the first to provide evidence that placental lipid uptake and accumulation are regulated, in part, through the placental WNT pathway. These novel findings demonstrate that WNT signaling is directly involved in the excessive accumulation of fat within the placenta in obese pregnancy, a pathophysiology that may have dire consequences for placental efficiency, nutrient transport, and fetal development.

Current recommendations for pregnant women focus primarily on controlling weight gain during pregnancy. However, increased evidence indicates that prepregnancy weight may be as important to gestational and fetal health as the total amount of weight gained during pregnancy. The model of OP and OR rats has been thoroughly studied and characterized and is a valuable tool for modeling obesity and metabolic syndrome in the human population. Studies showed that OP rats can be characterized by an increase in plasma TG [41] and have reduced capacity for fatty acid oxidation [42], especially in the liver [43]. Additionally, OP rats have decreased leptin sensitivity [44, 45], are more sensitive to stress-induced weight gain [46], have inappropriate insulin signaling [47, 48] and beta-cell mass with age [49], and defend their body weight in response to exercise [50], potentially because of their altered energy expenditure [51]. In the present study, OP dams had significantly higher leptin levels than OR dams, which confirms the obesity status in the current model and is also frequently observed in pregnant and nonpregnant individuals with obesity [45, 52]. Fasting TG levels were higher in OP than in OR dams, suggesting that the obese animals have a decreased ability to clear TG, even after an overnight fast. This observation has also been shown to be characteristic of the current model in nonpregnant rats, and it has been likened to individuals with severe obesity who also have a dysregulation in TG metabolism [41]. Elevated plasma NEFA levels, as observed in OP dams in the present study, are often seen in individuals with obesity [53], and it has been observed that pregnancies complicated with intrauterine growth restriction (IUGR) are also associated with elevated NEFA [54, 55], which agrees with our present data.

Maternal obesity has most often been associated with fetal overgrowth, but evidence exists that obesity can also result in small-for-gestational-age (SGA) offspring [56–59]. In the present study, OP offspring were smaller than OR offspring; however, because OP dams also had larger litters, the total products of conception were not different between the two groups. Because these results do not translate well to the human condition, it is difficult to make definitive statements regarding birth weight based on the present study. However, conditions that lead to decreased placental size or efficiency (and, therefore, to decreased area for nutrient transport and waste export) lead to IUGR [7, 60, 61]. Additionally, studies have suggested that the ratio of fetal to placental weight is a better marker of placental efficiency, because it represents the amount of fetus produced for each gram of placenta. The ratio has been evaluated by several studies, in which the researchers have concluded that a large placenta, when accompanied by a smaller fetus, is associated with IUGR as well as SGA offspring, gestational diabetes, and possible risks of late-onset diseases [7, 62–64]. In the present study, we found that placental weight was not affected by maternal weight but that the fetal weight:placental weight ratio was decreased in obese dams, indicating that despite being the same size as those of lean dams, the placentas of these obese dams were not able to support proper fetal development. We do admit that these differences in fetal growth may be a direct response to maternal metabolism and altered nutrient supply to the fetus. In the present study, we observed that OP dams ate less per gram of body weight than OR dams, and the expression of the placental glucose transporter, Glut1, was slightly, but significantly, increased in OP placentas when compared to OR placentas (P < 0.05; data not shown). Because the fetus uses glucose as its primary fuel source during development, this adaptation may potentially signal the presence of a nutrient deficit and function as a protective mechanism. Because the lipid accumulation within the placenta mirrors what is occurring within maternal circulation, it also may act as a direct reflection of the changes that occur in cases of obesity during pregnancy. Whether placental lipid accumulation itself results in placental inefficiency will need to be further verified, but our results suggest that the placenta directly responds to maternal lipid overload, which may regulate nutrient transport and, therefore, fetal development.

We showed that placentas of OP dams had a significant decrease in the size of the decidual and junctional zones, with a concurrent increase in lipid accumulation. Although little data are available regarding the consequence of increased lipid within the placenta, studies in other organ systems have suggested that excessive fat accumulation results in substantial damage to cellular components and tissue structure [65]. Appropriate vascularization and angiogenesis within the decidual as well as junctional zones is imperative for placental and fetal development [66], and any disruptions within the development of these zones could be potentially detrimental to fetal growth. A recent study of gestationally high-fat-fed (HF) rats showed that offspring of HF dams were growth restricted and that the junctional zone of HF placenta was decreased [67], which is consistent with what we observed in the present study. Appropriate placentation is characterized by full remodeling of the spiral arteries and their invasion into the decidual/junctional zones [68], so although more studies are needed to determine the exact reason for the decreased thickness of the decidual/junctional zones in response to HF diet or obesity, excessive lipid likely prevents appropriate arterial remodeling, thus interfering with the structure and, probably, function of these zones.

In the present study, we propose that the WNT signaling pathway is directly involved in placental lipid handling and show its dysregulation in obese rats. WNT antagonists are either secreted molecules or function at the intracellular level, acting through various routes. Secreted-type inhibitors include the SFRP family, WIF1, DKK1, DKK2, DKK3, and DKK4 [69], where the DKK family of molecules interact with and cause the endocytosis of LRP5/LRP6 coreceptor to prevent the formation of the WNT-FZD-LRP5/LRP6 complex, thus decreasing the accumulation of β-catenin within the nucleus. In a pregnant mouse model, DKK1 was highly expressed in the maternal decidual compartment, a culture of decidual cells in the presence of antisense DKK1 decreased trophoblast invasiveness, and treatment with antisense β-catenin increased trophoblast attachment and invasiveness, suggesting that an increase in WNT signaling may prevent decidualization and trophoblast cell invasion [70]. The inhibitive role of WNT signaling on vascularization in association with decreased DKK1 has been shown in cases of diabetic retinopathy [71], in human umbilical vein endothelial cells [72], and in human breast tumors [73]. Our data suggest that a decrease in DKK1 leading to aberrant WNT signaling results in “fatty” placentas with less decidual/junctional thickness and, therefore, a decrease in placental efficiency. Although PPARD has been shown to be expressed in the placenta [74], its exact actions have not been thoroughly classified. Because PPARD contains a β-catenin responsive element within its promoter [31], we propose that the aberrant activation of WNT signaling in OP placentas results in the activation of Ppard, which has been shown to be a potent inducer of adipogenesis and lipogenesis [75]. Because PPARG is another member of the PPAR family associated with lipogenesis, we also tested its mRNA amount in our placental tissues as well as in JEG3 cells. Whereas Pparg mRNA was elevated in OP placentas (data not shown), no effect was found on its expression in +DKK1 JEG3 cells following NEFA treatment. This suggests that although PPARG may be potentially important for placental lipid metabolism, PPARD is unique in its regulation by the WNT signaling pathway.

In conclusion, we show, to our knowledge for the first time, the regulation of placental lipid accumulation by a canonical WNT inhibitor, DKK1. Although the placenta and fetus require an extensive supply of essential fatty acids during gestation, we show that an excessive availability of NEFA in the circulation leads to placental fat accumulation, in coordination with an aberrant regulation of the WNT signaling pathway. As the primary mode of communication between mother and fetus, the placenta plays an integral part in the adaptations that occur in a nutrient-restricted or -abundant pregnancy. Any dysregulations within placental structure may be deleterious for both placental and fetal development, and altered placental lipid metabolism in obesity may serve as a signal to the fetus to make changes in preparation for the expected postnatal environment.

Acknowledgment

We would like to acknowledge the staff at the Core Facilities (Institute for Genomic Biology) at the University of Illinois for their knowledgeable advice and instruments.

Author notes

¹

Supported by the USDA Cooperative State Research, Education, and Extension Service, Hatch project ILLU-698-374.