placental antiangiogenic prolactin fragments are increased in human and rat maternal diabetes
TRANSCRIPT
Biochimica et Biophysica Acta 1842 (2014) 1783–1793
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
j ourna l homepage: www.e lsev ie r .com/ locate /bbad is
Placental antiangiogenic prolactin fragments are increased in human andrat maternal diabetes
P. Perimenis a,f,g, T. Bouckenooghe a, J. Delplanque b,g, E. Moitrot a, E. Eury b,g, S. Lobbens b,g, P. Gosset c,L. Devisme d, B. Duvillie e, A. Abderrahmani b,g, L. Storme a, P. Fontaine a,g,h, P. Froguel b,g,h, A. Vambergue a,g,h,⁎a EA 4489, Perinatal environment and Growth, University of Lille 2, 1 place de Verdun, 59045 Lille, Franceb CNRS-UMR8199, Lille 2 University, 1 rue du professeur Calmette, 59000 Lille, Francec Anatomopathological Department, Hospital Group of Catholic Institute of Lille (GHICL), 115 rue du grand but, 59160 Lomme, Franced Anatomopathological Department, CHRU Lille, University Hospital, 59037 Lille, Francee INSERM U845, Research Center Growth and Signaling, 96 rue didot, 75014 Paris, Francef Department of Diabetology, Hospital Group of Catholic Institute of Lille (GHICL), 115 rue du grand but, 59160 Lomme, Franceg European Genomic Institute for Diabetes (EGID), FR 3508, 59000 Lille, Franceh Department of Diabetology, CHRU Lille, University Hospital, Claude Huriez Hospital, rue Polonovski, 59037 Lille, France
Abbreviations: BMP-1, bone morphogenetic protein-gestation; dPRL, decidual prolactin; DPRP, decidual prolacdiabetes; IUGR, intra uterine growth restriction; MD, maproteinase; NCT, nicotinamide; PAS, Periodic Acid Schiff; Pnumber; SGA, small-for-gestational-age; STZ, streptozoto⁎ Corresponding author at: Department of Diabeto
rue Polonovski, 59037 Lille, France. Tel.: +33 3 20 44 68E-mail addresses: [email protected] (P. Perimenis),
[email protected] (T. Bouckenooghe), jer(J. Delplanque), [email protected] (E. Moitrot)(E. Eury), [email protected] (S. Lobbens), [email protected] (L. Devisme), [email protected] (A. Abderrahmani), lau(L. Storme), [email protected] (P. Fontaine), p.f(P. Froguel), [email protected] (A. Vambergue
http://dx.doi.org/10.1016/j.bbadis.2014.06.0260925-4439/© 2014 Elsevier B.V. All rights reserved.
a b s t r a c t
a r t i c l e i n f oArticle history:
Received 22 February 2014Received in revised form 20 June 2014Accepted 23 June 2014Available online 28 June 2014Keywords:DiabetesDecidual prolactin-related proteinPregnancyProlactinVasoinhibin
Introduction/objectives: The role of the placenta in diabetic mothers on fetal development and programming isunknown. Prolactin (PRL) produced by decidual endometrial cells may have an impact. Although full-lengthPRL is angiogenic, the processed form by bone morphogenetic protein-1 (BMP-1) and/or cathepsin D (CTSD) isantiangiogenic.The objectiveswere to investigate the involvement of decidual PRL and its antiangiogenic fragments in placentasfrom type-1 diabetic women (T1D) and from pregnant diabetic rats with lower offspring weights than controls.Methods: PRL, BMP-1, and CTSD gene expressions and PRL protein level were assessed in T1D placentas (n=8) atdelivery and compared to controls (n=5).Wistar rats received, at day 7 of pregnancy, streptozotocin (STZ) (n=5) or nicotinamide (NCT) plus STZ (n=9) or vehicle (n=9). Placental whole-genome gene expression and PRLwestern blots were performed at birth.Results: In human placentas, PRL (p b 0.05) and BMP-1 (p b 0.01) gene expressions were increasedwith a higher
amount of cleaved PRL (p b 0.05) in T1D than controls. In rats, diabetes was more pronounced in STZ than inNCT–STZ group with intra-uterine growth restriction. Decidual prolactin-related protein (Dprp) (p b 0.01) andBmp-1 (p b 0.001) genes were up-regulated in both diabetic groups, with an increased cleaved PRL amount inthe STZ (p b 0.05) and NCT–STZ (p b 0.05) groups compared to controls. No difference in CTSD gene expressionwas observed in rats or women.Conclusions: Alterations in the levels of the PRL family are associatedwithmaternal diabetes in both rats and T1Dwomen suggesting that placental changes in these hormones impact on fetal development.© 2014 Elsevier B.V. All rights reserved.
1; CTSD, cathepsin D; D, day oftin-related protein; DT1, type 1ternal diabetes; MMP, metallo-RL, prolactin; RIN, RNA integritycinlogy, Claude Huriez Hospital,18; fax: +33 3 20 44 58 42.
[email protected], [email protected]@ghicl.net (P. Gosset),[email protected] (B. Duvillie),[email protected]@imperial.ac.uk).
1. Introduction
Maternal diabetes (MD) accounts for a variety of fetal adverseeffects, including spontaneous abortion, intrauterine fetal death,andmaternal and perinatal complications (preeclampsia, prematuri-ty, neonatal respiratory distress syndrome) associated with abnor-mal birth weight [1], and depends on the severity of diabetes [2],especially when vascular complications are present [3]. The perinatalperiod represents a critical window of vulnerability to the environ-ment [4] and can modulate health conditions later in life [5,6]. MDinduces damage to placenta [7,8] but the molecular mechanismsinvolved are still elusive.
The placenta produces numerous hormones including prolactin(PRL). PRL is released fromdecidualized endometrial cells, and is thoughtto regulate the immune system [9,10], amniotic-fluid homeostasis [10,
1784 P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
11], and placental angiogenesis [12,13] at the feto-maternal interface.PRL stimulates cell migration and invasion in human trophoblastin vitro [14]. Angiogenesis is crucial for trophoblast invasion [15] andfeto-placental growth [16]. Findings report that endothelial cells produceand release PRL [17].
In this regard, full-length proangiogenic PRL (23-kDa) proteolysisgenerates N-terminal fragments, products known to be antiangiogenicandproapoptotic, and are also known as vasoinhibins [13,17–20]. Sever-al candidate proteases have been reported in rodents and/or in humans:i.e., certain metalloproteinases (MMPs) [21], bone morphogenetic-protein-1 (BMP-1) in neutral conditions [22], or cathepsin D (CTSD) inacid environment [23,24].
PRL and/or its vasoinhibins are thought to be involved in preeclamp-sia [12,13,25,26].More recently, in a cohort of 501preeclampticwomen,the risk of small-for-gestational-age (SGA) infants was increased whenvasoinhibins were present in the amniotic fluid [13]. Immunemaladap-tation to pregnancy may contribute to predominant 16-kDa PRL decid-ual production [25].
The decidual production coupled to the involvement in placentalangiogenesis and preeclampsia, with an increased risk for SGA infants,suggests that decidual PRL (dPRL) and/or its fragments could play arole in MD at the feto-maternal interface.
The aim of our studywas to investigate the involvement of dPRL andits antiangiogenic fragments in placentas from type-1 diabetic women(T1D) and pregnant diabetic rats with lower offspring weights thancontrols.
2. Methods
2.1. Collection of human placentas
Placentas were obtained from controls (n = 5) and women withT1D (n = 8) at birth, who were recruited from the Hospital RegionalCenter (CHRU)of Lille (France). For each placenta, sampleswere obtain-ed from4 various locations between the decidual and chorionic plates inorder to limit the tissue heterogeneity, near the umbilical cord on thefetal side and pooled for gene expression or protein analyses or used in-dividually for immunohistochemical analyses. The sampling locationwas uniform, performed by the same technical assistant.
All the diabetic women were managed with standardized protocolsregarding the treatment of diabetes, especially insulin therapy andprenatal care. Control women were defined by having normal glucosetolerance and a newbornwith a normal birthweight. This studywas ap-proved by the ethics committee of the CHRU of Lille (DG7 2007-0340/CCP07/43). All women involved received verbal information about thestudy and gave their written informed consent before examination.
2.2. Animal study
Female Wistar rats (200–250 g, Janvier, Le Genest Saint Isle, France)were housed individually with a controlled 12/12 h light/dark cycle, at22 ± 1 °C, and were fed ad libitum with a standard chow diet. On the7th day (D) of gestation, three groups of animals were formed, i.p.injected sequentially after 15 min: a diabetic group with severe hyper-glycemia (the STZ group: n = 5, received vehicle (NaCl 0.9%) andstreptozotocin (STZ) [65 mg/kg, dissolved in citrate buffer 0.1 mmol/l,pH 4.5; Sigma-Aldrich, St Quentin-Fallavier, France]); a diabetic groupwith moderate hyperglycemia (NCT–STZ group: n = 9, received nico-tinamide [NCT, Sigma-Aldrich, 75 mg/kg dissolved in NaCl 0.9%] andSTZ [65 mg/kg]); and a control group (C group: n = 9, received thesame volume of two vehicles). By injecting STZ and NCT at day 7 of ges-tation, we aimed to prevent the early formation of the fetal pancreas. Inaddition, at this stage, STZ induces MD without directly affecting fetalpancreas. The GLUT2 glucose transporter, through which STZ triggersbeta cell destruction, is not expressed in the early pancreas develop-ment [27,28].
Daily capillary blood-glucose (ACCU-CHEK Performa Glucometer,Roche Diagnostics, Mannheim, Germany) and weight were evaluatedfrom D7–18 in dams. At D19, for all groups, an oral glucose-tolerancetest was performed after 16 h of fasting. Animals received 2 g/kg ofglucose, and blood samples were collected from the tail vein at 0, 30,60, 90, and 120min to assess glucose (ACCU-CHEK Performa glucometer)and insulin (insulin Rat kit ELISA, Mercodia, Uppsala, Sweden) levels.
All animal experiments were conducted in accordance with theEuropean Communication Council Directive of November 24, 1986(86/609/EEC).
2.3. Collection of blood and tissue samples from rats
Pregnant rats were sacrificed before delivery at D21, correspondingto the end of pregnancy, andmaternal bloodwas collected to assess glu-cose (ACCU-CHEK Performa glucometer), insulin (insulin rat kit, ELISA),and PRL (rat prolactin ELISA kit, CUSABIO BIOTECH, China) plasmalevels. Fetuses were removed with their respective placentas and sepa-ratelyweighed. Placentaswerewashed several times in Phosphate Buff-ered Saline (PBS) to eliminate red blood cells and transferred inRNAlater (Qiagen, Courtaboeuf, France) for gene expression analysis,or were snap-frozen in liquid nitrogen and stored at−80 °C for proteinanalysis, or were fixed in 4% paraformaldehyde (without previouswashing in PBS) for histological analysis. Fetuses' pancreases were re-moved and fixed in 4% paraformaldehyde. Placentas and pancreases,fixed in 4% paraformaldehyde, were paraffin-embedded (ParaplastPlus, McCormick LLC, St. Louis, MO, USA). Fetuses' blood samples werecollected for the determination of glucose levels (ACCU-CHEK) andplasma samples in each litter were pooled to assess insulin levels(insulin rat kit, ELISA).
2.4. Gene-expression analyses of human and rat placentas
2.4.1. Extraction of RNAPlacental samples from women were transferred in RNAlater
(Qiagen, France). For rats, five placentas were randomly selected fromeach dam and placentas from each group were pooled together for mi-croarray analysis or used individually for qRT-PCR validation.
Total RNAwas isolated using the RNeasy® kit (Qiagen). The quanti-fication and quality of RNA were assessed by spectrophotometric anal-ysis with NanoDrop (Thermo Fischer Scientific, Illkirch, France) and byAgilent Bioanalyzer capillary electrophoresis system (Agilent technolo-gies France, Massy, France), respectively. Measurements were carriedout with a Bioanalyzer using RNA integrity number (RIN): the correlat-ing electropherogram and gel-like image were generated for eachsample. The RIN was obtained prior to array hybridization. Only high-quality RNA (RIN N 8) was used for analysis.
2.4.2. Quantitative RT-PCR (qRT-PCR) gene-expression analysis of humanplacentas
For humans, total RNA from each placenta was reverse-transcribedto cDNA using a cDNA synthesis kit (ThermoScript™ RT-PCR Systemfor First-Strand cDNA Synthesis, Invitrogen, France). qRT-PCR wasrun in duplicate using a LightCycler 480 SYBR Green I Master Mix(Roche Diagnostics, Meylan, France), primers (Table 1), and a real-time PCR system (Thermocycler LightCycler LC 2.0 Roche, France).Gene expression was normalized with the TATA box-binding proteingene.
2.4.3. Rat placenta microarray procedure and qRT-PCR gene-expressionanalysis
Whole-gene expression profiling, using the Illumina Rat Ref-12BeadChip (22523 probes for a total of 21792 genes from NCBI RefSeq;Illumina, San Diego, CA, USA), was carried out according to standardprocedures. Aliquots of 300 ng of total RNA were reverse-transcribedto cDNA, transcribed to cRNA, amplified and biotin-labeled with a
Table 1Human primers sequences used for RT-PCR.Hu: homo sapiens.
ID name Forward primer Reverse primer Gene name
Hu PRL 5′ACCAGGAAAAGGGAAACGAATGCC 3′ 5′CGTTGCAGGAAACACACTTCACCA3′ ProlactinHu BMP-1 5′TTCTCAGACAAGGACGAGTGCTCC3′ 5′TGCCGAACGTGTTGACGCA-GT 3′ Bone morphogenetic protein 1Hu CTSD 5′CAGGGCGAGTACATGATCCC3′ 5′TTGTAGCCTTTGCCTCCCAG3′ Cathepsin DHu TBP 5′ACGCCAGCTTCGGAGAGTT3′ 5′ GCACGAAGTGCAATGGTCTTT3′ TATA box binding protein
1785P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
Target Amp Nano-g Biotin-aRNA labeling kit for the Illumina System(Epicentre Biotechnologies, Madison, WI, USA), and purified withRNeasy MinElute Cleanup Kit (Qiagen). Biotinylated cRNA probes werehybridized on BeadChips, incubated for 16 h at 58 °C, streptavidin-Cy3-labeled and processed on an Illumina BeadArray Reader. Normalizedbead-intensity data (using the cubic spline-normalization method)were analyzed with Illumina GenomeStudio software. Gene-expressiondata were loaded into GeneSpring (Agilent Technologies France, Massy,France) for management, analysis, and interpretation. Pathway analysiswas conducted with Ingenuity Pathway Analysis (IPA) Software(Ingenuity Systems, RedwoodCity, USA). A list of differentially expressedgenes was identified using one-way ANOVA followed by Tukey's post-hoc test. Differences were considered significant at p b 0.05 and amean difference greater than 1.4-fold.
qRT-PCR TaqMan was used to confirm the expression of selectedgenes from transcriptomic analysis. Each assay was run in triplicatefor each RNA sample from the rat placentas. Total RNA was reverse-transcribed using a high-capacity cDNA RT kit (Applied Biosystems,Foster city, CA). Assays were run with Universal Master Mix (AppliedBiosystems) using TaqMan probes for selected genes (Table 2) on anApplied Biosystems 7900HT detection system (cycling program includ-ed 10 min incubation at 95 °C, followed by 40 cycles at 95 °C). Datavalues (Cycle Threshold values) were extracted from each assay withthe SDS software tool (Applied Biosystems). Gene expression wasnormalized with the beta-2 microglobulin (B2M) gene.
2.5. Human and rat placenta western-blot analysis
Equal amounts of proteins (70 μg) from the placentas from thetwo human groups and the three animals groups were loaded into12% SDS-PAGE. For the rat study, three to five placentas were ran-domly selected from each dam. Western-blot analysis was per-formed using rabbit anti-PRL (1:200, Santa Cruz Biotechnology)and β-actin (1:5000, Sigma) for human placentas or a mouse anti-alpha-tubulin (1:5000, Sigma) antibody for rat placentas. ExtractedPRL from rat pituitary gland was used as positive control. Afterbeing transferred on a nitrocellulose membrane (Hybond-ECL,Amersham BioSiences, Chalfont, UK), the proteins were detectedusing horseradish-conjugated secondary antibody (1:5000; SantaCruz Biotechnology) and Chemiluminescence Supersignal (Pierce).Densitometric analysis was performed using ImageJ software(National institutes of Health, Maryland, USA) and data were nor-malized for alpha-tubulin or β-actin.
Table 2TaqMan probes used for RT-PCR in rat placentas.Rat: rattus norvegicus.
ID name Assay ID
Rat Bmp1 Rn01466024_m1Rat Dprp Rn00575655_m1Rat Ctsd Rn00592528_m1Rat B2M Rn00560865_m1
2.6. Histological and immunohistochemical analyses from rats and humans
2.6.1. Rat pancreas immunochemistryFrom each dam, the pancreas fromone to three fetuseswas random-
ly selected for each different immunostaining. Five-μm-thick sectionswere deparaffinized in xylene and rehydrated through a graded seriesof ethanol. Antigen retrieval was performed by warming the sectionsfor 5 min at 96 °C in citrate buffer (pH 6).
After a cooling period of 30 min, sections were incubated with amouse anti-insulin antibody (1:500, AbDSerotec, Colmar, France), rab-bit anti-glucagon (DAKO), and rabbit anti-pancreatic and duodenal ho-meobox 1 (Pdx1) (1:500, Abcam, Paris, France). Sections were thenwashed in Tris-buffered saline and incubated for 30 min at room tem-perature with Alexa 488 coupled goat anti-mouse (for insulin), Alexa488 goat anti-rabbit (for Pdx1), and Alexa 594 anti-rabbit (for glucagon)antibodies (1:500, Life Technologies). The samples were analyzed byinverted fluorescence microscopy (Zeiss, Oberkochen, Germany). Fornon-fluorescent evaluation of islets, slides, following incubation withthe insulin anti-mouse antibody, were washed and incubated for 1 hwith a biotinylated sheep-anti-mouse antibody (1:2000, Chemicon, Te-mecula, CA, USA). They were then washed and incubated for 1 h with astreptavidin–peroxidase antibody (1:2000, Amersham, Buckinghamshire,UK), and revealed with 3,3′diaminobenzidine tetra-hydrochloride-DAB(Sigma-Aldrich).
2.6.2. Rat and human placental histologyPlacental samples from women were paraffin-embedded. For rats,
from each dam, the placenta from one to three term fetuses was ran-domly selected for each different histological or immunohistochemicalanalysis. Five-μm-thick sections were deparaffinized in xylene andrehydrated through a graded series of ethanol. Sections were stainedwith Hematoxylin–Eosin or incubated with a monoclonal mouse anti-CD31 antibody (1:40, DAKO, Glostrup, Denmark) using Ventana kits(Ventana, Tucson, Arizona, USA) according to instructions for both spe-cies placentas or with Periodic Acid Schiff (PAS)-diastase for only ratplacentas.
Vascular areas were estimated using ImageJ software (National In-stitutes of Health). In rat and women placenta, capillary density was es-timated by a combination of CD31 staining for the distinction betweenmaternal and fetal blood and of morphological analyses by mm2 in thelabyrinth zone in rat placenta and in terminal villi in women placentas.Capillary diameter (μm) was estimated by counting with an ocular mi-crometer from 10 fields of view. In women placentas, capillary density
Accession number Gene name
NM_031323.1 Bone morphogenetic protein 1NM_022846.1 Decidual prolactin-related proteinNM_134334.2 Cathepsin DNM_012512.2 Beta-2 microglobulin
Table 3Patient characteristics.BMI: body mass index; NA: not available; NS: not significant; HbA1C: glycated hemoglobin. Data are means ± SEM. *p b 0.05.
Characteristics Controls (n = 5) Diabetics (n = 8) p-Value
Women age, years 28.80 ± 2.78 27.88 ± 1.69 NSDuration of diabetes, years – 14.75 ± 3.44 –
Women's preconception BMI, kg/m2 21.15 ± 0.64 24.95 ± 1.64 NS1st trimester women's HbA1C, % NA 6.85 ± 0.44 –
2nd trimester women's HbA1C, % NA 6.41 ± 0.27 –
3rd trimester women's HbA1C, % NA 6.45 ± 0.25 –
Gestational age, weeks 39.20 ± 0.37 38.38 ± 0.42 NSNewborn birth weight, g 3366 ± 94.05 3026 ± 100.1 *p b 0.05Placental birth weight, g 428.0 ± 16.11 428.0 ± 17.80 NSPlacental efficiency ratio 7.91 ± 0.37 7.10 ± 0.29 NS
1786 P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
was estimated by counting in terminal villi. Capillary structures weredefined as CD31-positive capillary endothelial cells.
2.7. Statistical analyses
Data are presented as their means ± SEMs. All statistical analyseswere performed using GraphPad Prism 5.0 software (San Diego, USA).Comparisons between women groups were performed using the non-parametric Mann–Whitney test. Comparisons between rat groupswere performed using one-way ANOVA followed by Tukey's post-hoctest. A p-value of ≤0.05 was considered statistically significant.
3. Results
3.1. Birth weight was lower in newborns from diabetic mothers
The women's characteristics are summarized in Table 3. Control anddiabetic subjects were similar regarding age, bodymass index, and ges-tational age at delivery. T1D women had no diabetic vascular complica-tions, no hypertension and their HbA1c level was 6.45 ± 0.25%(47 mmol/mol) at delivery with a good glycemic control during the en-tire pregnancy. Birth weight was significantly lower in newborns fromdiabeticmothers (3026± 100.1 g) compared to newborns from controlwomen (3366± 94.05 g) (p b 0.05). No differencewas found in placen-tal birth weight between the two groups studied. The ratio of fetalweight to placental weight used to evaluate the placental efficiencydid not vary between the two groups.
3.2. PRL is increased at mRNA and protein levels in the placentas from DT1women
We evaluated the effects of diabetes on gene expression of PRL andon its protein level in the placenta. A significant increase in PRL wasobserved in the diabetic group compared to control at mRNA level(Fold Change (FC) ≈ 9, p b 0.05) (Fig. 1A) and confirmed at proteinlevel with an increased total PRL amount (p b 0.01) (Figs. 1B and C).
3.3. Expressions of BMP-1 and PRL cleavage are increased in the placentasfrom DT1 women
We evaluated if diabetes affected the PRL cleavage and generated anincreased vasoinhibin amount at protein level and if cleavage proteases,BMP-1 and/or CTSD could be involved in this process. As anticipated,western-blot analysis showed native full-length (23-kDa) and cleavedforms of PRL within the total protein extracts from human placentas(Fig. 1B). The ratio of cleaved PRL to full-length PRL (p b 0.05)(Fig. 1D) was significantly increased in placentas from diabetic patientscompared to controls with an increased processed form amount. BMP-1(FC ≈ 13, p b 0.01) gene expression was effectively increased (Fig. 1E)in diabetic placentas compared to controls. However, CTSD gene expres-sion in both groups was similar (Fig. 1E).
As vasoinhibins play a role in angiogenesis, we performed immuno-histochemical analyses to study vasculature. Hematoxylin and Eosinstained sections showed an up to 10% reduction in vascular surfaces indiabetic placentas compared to the controls (Fig. 1F) and demonstratedstructural differences of terminal villi. Normal terminal villi are of small-er diameter than diabetic villi. CD31 was used as an endothelial markerfor quantifying fetal capillaries. Capillary density was reduced in the di-abetic sections (6.16 ± 0.32, p b 0.05) (Fig. 1G) compared to controls(7.02 ± 0.26) with a smaller diameter of the capillaries (6.03 ±0.36 μm, p b 0.05) (Fig. 1H) compared to controls (7.22 ± 0.44 μm) interminal villi.
3.4. Intra uterine growth restriction (IUGR) in rat diabetic models
Pregnant rats were injectedwith STZ or NCT–STZ to obtain severe ormoderate diabetic conditions in rat models. During gestation, the dams'weight gain in the three groups was not significantly different untilD14. After D14, only the STZ-treated dams lost weight significantly(p b 0.05). By D21, the dams' weight in STZ (p b 0.001) and NCT–STZ(p b 0.05) groups was significantly lower compared to that in the con-trol group (Fig. 2A).
As expected, dams that received STZ (p b 0.001) or NCT–STZ (pb 0.001) developed significant hyperglycemia compared to the controlsby 24 h after STZ injection (Fig. 2B), which was more severe in the STZgroup (p b 0.05) than in the NCT–STZ group for thewhole period of ges-tation. By D19, fasting blood-insulin was significantly reduced in thetwo diabetic groups (0.14 ± 0.02 μg/l in the STZ group, 0.23 ±0.05 μg/l in the NCT–STZ group) compared to the controls (0.93 ±0.214 μg/l, p b 0.001). There was no significant difference in plasma-insulin values between the STZ and NCT–STZ groups (data not shown).
We evaluated the level of plasma PRL in control and diabetic dams.Maternal plasma PRL level at D21 was significantly higher in the STZ(29.19 ± 3.61 ng/ml, p b 0.01) and NCT–STZ (18.81 ± 2.89 ng/ml, pb 0.01) groups compared to the controls (4.47 ± 0.45 ng/ml) (Fig. 2C).
We analyzed fetuses' weight and their placenta at D21. Fetuses' birthweight was significantly reduced in both diabetic groups (STZ: 3.55 ±0.14 g, p b 0.001; NCT–STZ: 4.18 ± 0.09 g, p b 0.001) when comparedto the control group (5.57 ± 0.04 g) (Fig. 3A) and this reduction wasmore severe in STZ group than NCT–STZ group (p b 0.001). Theplacental weight was significantly increased in the STZ (0.73 ±0.02 g, p b 0.001) and NCT–STZ (0.63 ± 0.01 g, p b 0.001) groupscompared to the controls (0.56±0.01 g) (Fig. 3B)with a placental hyper-trophy more pronounced in STZ group than NCT–STZ group (p b 0.001).The ratio of fetal weight to placental weight was significantly reduced inboth diabetic groups (STZ: 5.04± 0.22, p b 0.001; NCT–STZ: 6.83± 0.20,p b 0.001) compared to the controls (10.10 ± 0.14). In line with theplacental hypertrophy, stained PAS-diastase sections of STZ, NCT–STZand control placentas showed glycogen accumulation in the tropho-blastic cells of the hemo-placental barrier in the two diabetic groupscompared to the controls (Fig. 3C). No edema or ischemic zones wereobserved.
Controls Diabetics
42 kDa-
23 kDa-
15 kDa-17 kDa-
ββ-actin
PRL
C
A
DB
E
0
5
10
15
PRL
* ControlsDiabetics
mR
NA
exp
ress
ion
0246
5060708090
100
BMP-1
** DiabeticsControls
CTSD
mR
NA
exp
ress
ion
0.0
0.5
1.0
1.5
2.0
2.5 ** ControlsDiabetics
Tot
al P
RL
/β-a
ctin
Controls Diabetics
HE
CD
31
F
H
G
0
2
4
6
8* Controls
Diabetics
Cap
illar
y de
nsity
0.00.20.40.60.81.020406080
DiabeticsControls*
Cle
aved
PR
L/ P
RL
23 k
Da
0
2
4
6
8
10
*
ControlsDiabetics
Cap
illar
y di
amet
er (µ
m)
Fig. 1. Analysis of PRL gene expression and cleavage in human placentas. Placental mRNA expression of prolactin (PRL) (A). Representative western-blot analysis of prolactin proteolyticproducts in placentas (B). Relativemolecular mass of each band is shown on the left. Quantification of total amounts of PRL fragments (C) and of cleaved PRL to 23-kDa PRL ratio (D) fromwestern-blot data. Cleaved PRL represents the sum of ≈17-kDa PRL and ≈15-kDa PRL fragments amounts. Placental mRNA expressions of bone morphogenetic protein-1 (BMP-1) andcathepsinD CTSD (E). Placental histological sections stained byHematoxylin–Eosin (HE) and placental CD31 immunostaining (F), stereological analyses of fetal capillary network showingcapillary density (G) and themean capillary diameter (H) in terminal villi. Magnification ×200 and in the square ×400. Data are expressed asmeans± SEM and are significantly differentat *p b 0.05; **p b 0.01.
1787P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
We analyzed fetuses' glycemia levels at D21 and their pancreas. AtD21, fetuses from STZ and NCT–STZ groups were hyperglycemic(25.52 ± 1.09 mmol/l, p b 0.001 and 20.02 ± 0.63 mmol/l, p b 0.001,respectively) compared to those of the controls (3.82 ± 0.10 mmol/l)(Fig. 4A). Neonatal plasma-insulin levels were lower in both diabeticgroups (0.33 ± 0.09, p b 0.001 and 0.49 ± 0.10 mmol/l, p b 0.01 forSTZ and NCT–STZ groups, respectively) with greater severity in theSTZ group compared to the controls (0.76 ± 0.07 mmol/l) (Fig. 4B).
Plasma-insulin level was not statistically different but tended to belower in the STZ group compared to the NCT–STZ group. We performedimmunofluorescence staining of pancreases from fetuses (Fig. 4C). Isletultrastructure from control fetuseswas normal. Insulin and Pdx1 immu-nostaining of islets was intense and homogeneous: i.e., in the mantle ofglucagon-positive cells surrounding the insulin-positive cells. In sharpcontrast, the number of islets in NCT–STZ and STZ fetuses was reduced,but the STZ group was more severely affected than the NCT–STZ group.
200
220
240
260
280
300
320
340
360
380
400
D7 D8 D9 D10 D11 D12 D13 D14 D15 D16 D17 D18 D21
Dam
s wei
ght
(g)
Gestation (Days)
C
STZ
NCT-STZ
**
§
**** *
0
5
10
15
20
25
30
35
D7 D8 D9 D10 D11 D12 D13 D14 D15 D16 D17 D18 D21
Blo
od G
luco
se (m
mol
/l)
Gestation (Days)
§
§§§
§§§§§§ §§§ §§§ §§§ §§§ §§§
§§§§§§
§§***
†††*** † ******†***†***†***†***†***†***† ***†***†
A
C
B
0
10
20
30
40**
§§
CSTZNCT-STZ
Plas
ma
prol
actin
(ng/
ml)
C
STZ
NCT-STZ
Fig. 2.Analysis of dams' parameters.Weight (A) and blood-glucose level (B) of damsduring gestation. Plasma prolactin level of dams at the end of gestation (D21) (C). STZ: streptozotocin;NCT: nicotinamide; C: controls. Data are expressed asmeans±SEMs and are significantly different at *p b 0.05; **p b 0.01; ***p b 0.001 STZ vs C, §p b 0.05; §§p b 0.01; §§§p b 0.001NCT–STZvs C and †p b 0.05; †††p b 0.001 STZ vs NCT–STZ.
1788 P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
Insulin and Pdx1 immunostaining was profoundly reduced in theSTZ group compared to controls or the NCT–STZ group. However,there was no difference between the NCT–STZ and control groups.Glucagon-positive cells were preserved in the two diabetic groupscompared to the controls. Lower insulin blood levels in pups are duein part in a reduction in beta cell mass as we demonstrated a reductionin insulin and Pdx 1 immunostaining in the diabetics groups. As expect-ed, this reduction wasmore pronounced in STZ group due to protectiveeffect of the NCT in NCT-STZ group.
3.5. Dprp gene expression increases in placentas from diabetic rat models
A microarray differential transcriptome analysis was performed toexplore global gene-expression profiling and the involvement of the
PRL family members in our diabetic models. We showed modificationin 693 mRNA in the placentas from rats treated with STZ. Amongthe modified mRNA, 325 were up-regulated and 368 were down-regulated. When compared to the controls, 677 mRNA were modifiedin the placentas from the NCT–STZ group: 301 were up-regulated and376were down-regulated. Compared to the control group, gene expres-sion was similar in both diabetic groups for 457 mRNA (243 were up-regulated and 214 were down-regulated), and was different in 98mRNA (59 up-regulated and 39 down-regulated). We analyzed someof the gene functions of the 457 mRNA that were expressed similarlyin both diabetic groups (compared to controls) using IPA software.Most of these genes were involved in “cell growth/proliferation”, and“cellular mobility”, or “lipid metabolism” (Fig. 5A). Function analysisof the up- and down-regulated mRNAs is shown in Fig. 5B and C.
0
2
4
6 CSTZNCT-STZ
***§§§
†††
Fetu
ses b
irth
wei
ght (
g)0.0
0.2
0.4
0.6
0.8 CSTZNCT-STZ
***†††
§§§
Plac
enta
l wei
ght (
g)
A B
CST
ZN
CT
-ST
Z
PAS before diastase staining PAS after diastase stainingC
Fig. 3.Analysis of fetus and placentaweights at day 21 of pregnancy. Fetuses' birthweight (A) and placentalweight (B). Data are expressed asmeans± SEM and are significantly differentat ***p b 0.001 STZ vs C; §§§p b 0.001NCT–STZ vs C; †††p b 0.001 STZ vs NCT–STZ. Placental stainingwith Periodic Acid Schiff (PAS), before and after diastase (C). Magnification ×200. Pres-ence of placental glycogen deposits after PAS staining (red-purple coloration) in STZ and NCT–STZ groups, with disappearance of the red-purple coloration after digesting glycogen withdiastase. No glycogen deposits occur in the control placentas. STZ: streptozotocin; NCT: nicotinamide; C: controls.
1789P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
Regarding the members of the PRL family, only decidual prolactin-related protein (Dprp) (also known as Prl8a2) gene expression, con-firmed by qRT-PCR was approximately 4-fold increased in the STZ(p b 0.01) and NCT–STZ (p b 0.01) groups compared to the controls(Fig. 6); all other members that included Prl gene expression didnot differ between groups.
No difference was also noted at protein level by western blot analy-sis of PRL between all groups (data not shown).
3.6. Bmp-1 gene expression and PRL cleavage are increased in the placentasfrom diabetic rat models
Bmp-1, by qRT-PCR,was up-regulated in STZ (FC≈ 1.5, p b 0.05) andNCT–STZ (FC≈ 2, p b 0.001) groups compared to the controls. Howev-er, therewas nodifference in the Bmp-1 gene expression levels betweenthe diabetic groups. Ctsd gene expression did not vary between any ofthe groups (Fig. 7A).
Bywestern blot, the unprocessed (23-kDa) and cleaved forms of PRLwere also evident in rat placentas (Fig. 7B). The ratio of cleaved PRL tofull-length PRL was significantly higher in the STZ (p b 0.05) and NCT–STZ (p b 0.05) groups compared to the controls (Fig. 7C).
As vasoinhibins play a role in angiogenesis, we performed histo-logical analysis to study vascular surfaces. Hematoxylin and Eosinstained sections showed an up to 20% reduction in vascular surfaces
in STZ and NCT–STZ placentas compared to the controls (Fig. 7D).Using CD31 to differentiate fetal capillaries frommaternal blood, ste-reological analyses of the placental labyrinth zone showed a reduc-tion in capillary density in the STZ (15 ± 1.08, p b 0.01) and NCT–STZ(15 ± 1.00, p b 0.01) sections compared to controls (23 ± 0.57)(Fig. 7E) with a smaller diameter of the capillaries in STZ (5.37 ±0.26 μm, p b 0.001) and NCT–STZ (5.57± 0.23 μm, p b 0.001) comparedto controls (10.37 ± 0.55 μm) (Fig. 7F).
4. Discussion
Our study was designed to uncover the potential contributionof dPRL and its antiangiogenic derivatives to fetal development inwomen with T1D and in rats with MD. In diabetic placentas, we ob-served a significant increase in PRL gene expression in women and ofDprp gene expression in rats. We also found increased expression atthe protein level of two antiangiogenic PRL fragments in the womenand in rats. To our knowledge, this is the first report to describe PRLand its fragments in diabetic placentas.
Masumoto et al. showed a significant increase of placental antian-giogenic PRL fragments in cases of pregnancy-induced hypertension[29]. It is noteworthy that none of our cases with T1D experiencedpregnancy complications, such as preeclampsia or pregnancy-inducedhypertension.
0
10
20
30 CSTZNCT-STZ
***§§§
†††
Fetu
ses b
lood
glu
cose
(mm
ol/l)
0.0
0.2
0.4
0.6
0.8
1.0 CSTZNCT-STZ
***
§§
NS
Fetu
ses p
lasm
a in
sulin
(µg/
l)
A B
STZNCT-STZC
Insulin
Ins/Glu
Pdx1/Ins
C
Insulin InsulinInsulin
Pdx1 /Ins Pdx1 /Ins Pdx1/ Ins
Ins/Glu Ins/GluIns/Glu
Fig. 4. Analysis of fetuses' glycemia levels at day 21 of pregnancy. Fetuses' blood-glucose level (A) and plasma insulin level (B). Data are means± SEM. ***p b 0.001 STZ vs C; §§§p b 0.001NCT–STZ vs C; †††p b 0.001 STZ vs NCT–STZ. Fetuses' pancreatic immunohistochemistry (C) in STZ, NCT-STZ and control groups. Immunostaining against insulin as visualized by the HRP-DABmethod, double immunofluorescence staining of islets against insulin (ins, green) and glucagon (glu, red) or against pancreatic and duodenal homeobox 1 (Pdx1, green) and insulin(ins, red) in the controls, NCT–STZ and STZ groups. Magnification ×200. STZ: streptozotocin; NCT: nicotinamide; C: controls.
Fig. 5.Microarray analysis using Ingenuity Pathway Software. Biological function analysis of STZ and NCT–STZ common genes compared to the control group (A) and biological functionanalysis of STZ (B) and NCT–STZ (C) differential expression genes compared to control group. STZ: streptozotocin; NCT: nicotinamide; C: controls.
1790 P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
0.0
0.2
0.4
0.6
Dprp
CSTZNCT-STZ
**
§§
mR
NA
rel
ativ
e ex
pres
sion
Fig. 6. Placental mRNA expression of decidual prolactin-related protein (Dprp). Data aremeans ± SEM. **p b 0.01 STZ vs C; §§p b 0.01 NCT–STZ vs C.
1791P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
We studied rats in which diabetes was induced by either STZ aloneor in combination with NCT in order to achieve less severe diabetes[30]. STZ-treated rats arewidely used for inducing diabetes butwith dif-ferent results of generated hyperglycemia fluctuating from severe(N300 mg/dl) [31] to mild (120–300 mg/dl) [32] depending on theSTZ dose or on the time of injection. Obtainingmoderate hyperglycemiain rodents treated with low doses of STZ is difficult and may be relatedto the interaction of polygenic and nutritional factors as maternal diet[33] that leads to different responses to diabetogenic agent in beta-cells. To circumvent this question, we chose another strategy byinjecting STZ followed by NCT for ensuring a relative protective effecton beta-cells with a partial destruction [34] and guaranteeing a milddiabetic status in our animals. This model has been recently used todeterminate the effects of mild mother's hyperglycemia on the fetalgrowth and some hypothetical links between placental dysfunctionsand abnormal fetal growth [35]. In contrast with models of diabetesusually described in the literature, the NCT–STZ treatment had no effecton the viability of the fetus, induced no fetalmalformations or IUGR. Thedose of NCT used (250 mg/kg) was higher than the one used in ourstudy.
In these animals, placental Dprp gene expression was increasedwhen compared to controls, but placental gene expression of theother members of the PRL family, especially PrlmRNA level, did not sig-nificantly change. The expression of some members of the PRL familyvaries according to species [36]. Rat PRL has only 64% homology withinhumans [37]. The genome contains 24 genes related to PRL in ratswhereas, humans have only a single constituent: PRL itself [36,38,39].Rodent PRL family members can undergo sub-functions compared tothe whole function of human PRL [39]. In women, the increase of PRLmRNA and of its cleaved forms levels was more pronounced than thatin diabetic pregnant rat. The difference did not rely on the severity ofdiabetes. The expression ofDprp in STZ andNCT–STZ groupswas similarwhereas diabeteswasmore severe in STZ group. Nonetheless, hypergly-cemia and/or other abnormalities that usually are associated withdiabetes such as inflammation, individually or together, could affectthe expression and maturation of PRL.
In our study, the higher rat PRL maternal plasma levels and yet nodifference in dPRL gene expression in diabetic rats suggests higher PRLanterior pituitary-gland production. It has been shown that the deciduaproduces large amounts of PRL, which can enter the amniotic fluid [40],but only a small portion enters thematernal blood [41,42]. In contrast tothe rats, we did not assess maternal PRL blood levels in the T1D or con-trol women.
The exact role of DPRP in the control of gestation is not clear.DPRP is a 29-kDa protein isolated in the decidua; it resides in the de-cidual extracellular matrix and is a member of the PRL family [43].Alam et al. have shown, in Dprp-null mice, that DPRP deficiencyinterfered with pregnancy-dependent adaptations to hypoxiaand resulted in total pregnancy failure at D7.5 [44]. Our datasuggest that placental up-regulation of Dprp could be an adaptive
consequence of diabetes for maintaining pregnancy in this deleteri-ous situation.
In our study, amounts of 15- and 17-kDa PRL fragments were in-creased in human placentas from the diabetic group. In the diabetic an-imal models, two PRL fragments, of 14- and a 16-kDa, were highlighted.Our results are in accordancewith the literature in rat and in human [18,22,24,45]. Human PRL proteolysis generates vasoinhibins with distinctmolecular masses from that obtained with rodent hormone, identifiedby N-terminal sequencing and mass spectrometry as corresponding to15-kDa, 16.5-kDa, [23,24,45,46] and 17-kDa [24]. Studies have reportedthat amounts of 14- or 16-kDa PRL are increased in the circulation,urine, amniotic fluid, or placenta in cases of preeclampsia because ofrenal glomerulus or decidual local production [13,25,29]. No study hasyet demonstrated an increase of these fragments in MD either in theplacenta or fluids. An increase of these fragments in the placenta couldlead to decreased angiogenesis or increased apoptosis, causing a hypox-ic atmosphere as observed in MD [47].
In line with this hypothesis, placental histology showed ahypovascularization with a reduction in vascular surfaces and cap-illary density with small capillary diameter in both species, whichcould be caused by a defect in the angiogenesis mechanisms. T1D influ-ences placental microvascular bed by change size and course of capil-laries, and altered structure of villous stroma [48]. The development ofthe placenta is different between insulin-dependent diabetic pregnan-cies if fetal macrosomia or maternal hypertensive disorders are presentwith partial transformation of the spiral arteries [49]. Little is known onthe placenta of low birth offspring weight in T1D, especially when anymaternal complication as a microvascular one or a hypertension wasfound as in our study. Placental efficiency in women was surprisinglysimilar in all groups despite the placental hypovascularization, but MDinfluences the placental structure in a complex manner. For the rats,our results on placental hypovascularization are in agreementwith pre-vious studies [50]. The reduction in the specific diffusing capacity of thevillous membrane may contribute to the fetal hypoxia and increasedfetal and neonatal morbidity associated with diabetes.
In addition, in women with preeclampsia, a linear dose–responsepattern between values of each of the vasoinhibins and the risk of SGAhas been demonstrated [13]. In this respect, while data support an in-creased cleavage of PRL in a diabetic status, we could not exclude thatincreases in vasoinhibins could account for the increased risk of someperinatal outcomes, such as a lower birth weight in diabetic offspringor IUGR in our animal models as suggested in preeclampsia. At thistime, there is no hard evidence to support this relationship betweenthis and fetal growth.
We also found that placental BMP-1 gene expression was increasedin women and in rats and could lead to increased PRL proteolysis inthe placenta. In contrast, CTSD gene expression was not affected in ourstudy. We do not exclude the possibility that CTSD, and perhaps otherproteases, such as MMPs, Kallikrein, and carboxypeptidases may beinvolved to some extent in processing PRL.
In conclusion, the levels of members of the PRL family are modifiedin the placenta during MD. We may speculate from our data that PRLfragments are associated with the physiopathology of placental dys-function in diabetes mellitus. The placental changes in these hormonescould impact on intrauterine fetal development.
Contribution statement
PP performed the management of the animals, contributed to theperformance of all the experiments, analyzed and collected the data,andwrote themanuscript. AV conceived and designed the experiments,recruited the patients, analyzed the data, and wrote the manuscript.PFR, TB, JD, PFO, BD, AA, and LS reviewed themanuscript. TB performedpancreas immunohistochemical experiments. PG and LDperformedhis-tological experiments. PG performed histological analyses. EM, EE, and
A
0.00
0.05
0.10
0.15
0.20
0.25
Bmp-1 Ctsd
*§§§
CSTZNCT-STZ
mR
NA
rel
ativ
e ex
pres
sion
0
5
10
15
20
25
** §§
CSTZNCT-STZ
Cap
illar
y de
nsity
E
0
5
10
15
*** §§§
CSTZNCT-STZ
Cap
illar
y di
amet
er (µ
m)F
0.0
0.2
0.4
0.6
0.8 CSTZNCT-STZ
*§
Cle
aved
PR
L/2
3kD
a PR
LC STZ NCT-STZ PC
81 kDa-
23 kDa-
14 kDa-16 kDa-
α-Tubulin
PRL
B C
DC STZ
100 µm 100 µm100 µm
NCT-STZ
HE
CD
31
Fig. 7. Analysis of placental PRL cleavage in rat diabetic models. Bone morphogenetic protein-1 (Bmp-1) and cathepsin D (Ctsd) mRNA expressions (A). Representative western-blot ofprolactin (PRL) proteolytic products in placentas (B). PC: positive control. The relative molecular mass of each band is shown on the left. Quantification of cleaved PRL to 23-kDa PRLratio in the placental tissue (C). Cleaved PRL represents the sum of≈16-kDa PRL and≈14-kDa PRL fragments. Placental histological sections stained by Hematoxylin–Eosin and placentalCD31 immunostaining in controls, in STZ and NCT-STZ groups (D). STZ: streptozotocin; NCT: nicotinamide; C: controls. Vascular surfaces are reduced (20%) in the STZ group compared tothe controls. Stereological analysis of labyrinth capillary density (E) and capillary diameter (F). Magnification ×200 and in the square ×400. Data are means ± SEM. STZ: streptozotocin;NCT: nicotinamide; C: controls. *p b 0.05; **p b 0.01 STZ vs C; §p b 0.05; §§p b 0.01; §§§p b 0.001 NCT–STZ vs C.
1792 P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
SL performed biological molecular experiments and analyses. All au-thors reviewed and approved the final version of the manuscript.
Acknowledgements
We thank CNRS (REGION CPER CARDIODIA 28228), and theDiamenord group (DIAMENORD-19-11-2010 and DIAMENORD-02-12-2011) for their financial support for this study.
References
[1] P. Boulot, N. Chabbert-Buffet, C. d'Ercole, M. Floriot, P. Fontaine, A. Fournier, J.Y. Gillet, H.Gin, S. Grandperret-Vauthier, A.M. Geudj, B. Guionnet, S. Hauguel-de-Mouzon, S.Hieronimus, M. Hoffet, D. Jullien, M.F. Lamotte, V. Lejeune, J. Lepercq, F. Lorenzi, P.Mares, A. Miton, A. Penfornis, B. Pfister, E. Renard, M. Rodier, P. Roth, G.A. Sery, J. Timsit,A.S. Valat, A. Vambergue, O. Verier-Mine, French multicentric survey of outcome of preg-nancy in women with pregestational diabetes, Diabetes Care 26 (2003) 2990–2993.
[2] Z. Ergaz, M. Avgil, A. Ornoy, Intrauterine growth restriction-etiology and conse-quences: what do we know about the human situation and experimental animalmodels? Reprod. Toxicol. 20 (2005) 301–322.
1793P. Perimenis et al. / Biochimica et Biophysica Acta 1842 (2014) 1783–1793
[3] U.J. Eriksson, J. Cederberg, P. Wentzel, Congenital malformations in offspring ofdiabetic mothers—animal and human studies, Rev. Endocr. Metab. Disord. 4(2003) 79–93.
[4] H. Lehnen, U. Zechner, T. Haaf, Epigenetics of gestational diabetes mellitus and off-spring health: the time for action is in early stages of life, Mol. Hum. Reprod. 19(2013) 415–422.
[5] B. Brenseke, M.R. Prater, J. Bahamonde, J.C. Gutierrez, Current thoughts on maternalnutrition and fetal programming of the metabolic syndrome, J. Pregnancy 2013(2013) 368461.
[6] S. Galjaard, R. Devlieger, F.A. Van Assche, Fetal growth and developmental program-ming, J. Perinat. Med. 41 (2013) 101–105.
[7] T. Radaelli, J. Lepercq, A. Varastehpour, S. Basu, P.M. Catalano, M.S. Hauguel-De,Differential regulation of genes for fetoplacental lipid pathways in pregnancy withgestational and type 1 diabetes mellitus, Am. J. Obstet. Gynecol. 201 (2009) 209.
[8] G. Sisino, T. Bouckenooghe, S. Aurientis, P. Fontaine, L. Storme, A. Vambergue, Diabe-tes during pregnancy influences Hofbauer cells, a subtype of placental macrophages,to acquire a pro-inflammatory phenotype, Biochim. Biophys. Acta 1832 (2013)1959–1968.
[9] T. Bouckenooghe, G. Sisino, S. Aurientis, G. Chinetti-Gbaguidi, J. Kerr-Conte, B. Staels,P. Fontaine, L. Storme, F. Pattou, A. Vambergue, Adipose Tissue Macrophages (ATM)of obese patients are releasing increased levels of prolactin during an inflammatorychallenge: a role for prolactin in diabesity? Biochim. Biophys. Acta 1842 (2013)584–593.
[10] S. Shelly, M. Boaz, H. Orbach, Prolactin and autoimmunity, Autoimmun. Rev. 11(2012) A465–A470.
[11] N. Demir, M. Celiloglu, P.A. Thomassen, A. Onvural, O. Erten, Prolactin and amnioticfluid electrolytes, Acta Obstet. Gynecol. Scand. 71 (1992) 197–200.
[12] C. Clapp, F.J. Lopez-Gomez, G. Nava, A. Corbacho, L. Torner, Y. Macotela, Z. Duenas, A.Ochoa, G. Noris, E. Acosta, E. Garay, G.M. de la Escalera, Expression of prolactinmRNA and of prolactin-like proteins in endothelial cells: evidence for autocrine ef-fects, J. Endocrinol. 158 (1998) 137–144.
[13] C. Gonzalez, A. Parra, J. Ramirez-Peredo, C. Garcia, J.C. Rivera, Y. Macotela, J. Aranda,M. Lemini, J. Arias, F. Ibarguengoitia, G.M. de la Escalera, C. Clapp, Elevatedvasoinhibins may contribute to endothelial cell dysfunction and low birth weightin preeclampsia, Lab. Investig. 87 (2007) 1009–1017.
[14] I. Stefanoska, K.M. Jovanovic, S. Vasilijic, D. Cujic, L. Vicovac, Prolactin stimulates cellmigration and invasion by human trophoblast in vitro, Placenta 34 (2013) 775–783.
[15] G.E. Lash, J.E. Cartwright, G.S. Whitley, A.J. Trew, P.N. Baker, The effects of angiogenicgrowth factors on extravillous trophoblast invasion andmotility, Placenta 20 (1999)661–667.
[16] J.H. Toft, I.A. Lian, A.L. Tarca, O. Erez, J. Espinoza, I.P. Eide, L. Bjorge, S. Draghici, R.Romero, R. Austgulen, Whole-genome microarray and targeted analysis ofangiogenesis-regulating gene expression (ENG, FLT1, VEGF, PlGF) in placentasfrom pre-eclamptic and small-for-gestational-age pregnancies, J. Matern. FetalNeonatal Med. 21 (2008) 267–273.
[17] A.M. Corbacho, G.M. de la Escalera, C. Clapp, Roles of prolactin and related membersof the prolactin/growth hormone/placental lactogen family in angiogenesis, J.Endocrinol. 173 (2002) 219–238.
[18] C. Clapp, C. Gonzalez, Y. Macotela, J. Aranda, J.C. Rivera, C. Garcia, J. Guzman, M.Zamorano, C. Vega, C. Martin, M.C. Jeziorski, G.M. de la Escalera, Vasoinhibins: afamily of N-terminal prolactin fragments that inhibit angiogenesis and vascularfunction, Front. Horm. Res. 35 (2006) 64–73.
[19] J. Ferraris, D.B. Radl, S. Zarate, G. Jaita, G. Eijo, V. Zaldivar, C. Clapp, A. Seilicovich, D.Pisera, N-terminal prolactin-derived fragments, vasoinhibins, are proapoptotic andantiproliferative in the anterior pituitary, PLoS One 6 (2011) e21806.
[20] I. Struman, F. Bentzien, H. Lee, V. Mainfroid, G. D'Angelo, V. Goffin, R.I. Weiner, J.A.Martial, Opposing actions of intact and N-terminal fragments of the human prolac-tin/growth hormone family members on angiogenesis: an efficient mechanism forthe regulation of angiogenesis, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 1246–1251.
[21] Y. Macotela, M.B. Aguilar, J. Guzman-Morales, J.C. Rivera, C. Zermeno, F. Lopez-Barrera, G. Nava, C. Lavalle, G.M. de la Escalera, C. Clapp, Matrix metalloproteasesfrom chondrocytes generate an antiangiogenic 16 kDa prolactin, J. Cell Sci. 119(2006) 1790–1800.
[22] G. Ge, C.A. Fernandez, M.A. Moses, D.S. Greenspan, Bone morphogenetic protein 1processes prolactin to a 17-kDa antiangiogenic factor, Proc. Natl. Acad. Sci. U. S. A.104 (2007) 10010–10015.
[23] R.A. Baldocchi, L. Tan, D.S. King, C.S. Nicoll, Mass spectrometric analysis of the frag-ments produced by cleavage and reduction of rat prolactin: evidence that the cleav-ing enzyme is cathepsin D, Endocrinology 133 (1993) 935–938.
[24] D. Piwnica, P. Touraine, I. Struman, S. Tabruyn, G. Bolbach, C. Clapp, J.A. Martial, P.A.Kelly, V. Goffin, Cathepsin D processes human prolactin into multiple 16 K-like N-terminal fragments: study of their antiangiogenic properties and physiologicalrelevance, Mol. Endocrinol. 18 (2004) 2522–2542.
[25] A. Leanos-Miranda, I. Campos-Galicia, K.L. Ramirez-Valenzuela, Z.L. Chinolla-Arellano, I. Isordia-Salas, Circulating angiogenic factors and urinary prolactin as
predictors of adverse outcomes in women with preeclampsia, Hypertension 61(2013) 1118–1125.
[26] A. Parra, J. Ramirez-Peredo, The possible role of prolactin in preeclampsia: 2001, ahypothesis revisited a quarter of century later, Med. Hypotheses 59 (2002)378–384.
[27] S. Lenzen, The mechanisms of alloxan- and streptozotocin-induced diabetes,Diabetologia 51 (2008) 216–226.
[28] W.J. Schnedl, S. Ferber, J.H. Johnson, C.B. Newgard, STZ transport and cytotoxicity.Specific enhancement in GLUT2-expressing cells, Diabetes 43 (1994) 1326–1333.
[29] A. Masumoto, H. Masuyama, N. Takamoto, Y. Akahori, Y. Hiramatsu, Expression ofantiangiogenic prolactin fragments in the placentas of women with pregnancy in-duced hypertension, Acta Med. Okayama 64 (2010) 249–255.
[30] P. Masiello, E. Bergamini, Nicotinamide and streptozotocin diabetes in the rat. Fac-tors influencing the effectiveness of the protection, Experientia 33 (1977)1246–1247.
[31] R.C. McEvoy, J. Andersson, S. Sandler, C. Hellerstrom, Multiple low-dosestreptozotocin-induced diabetes in the mouse. Evidence for stimulation of a cyto-toxic cellular immune response against an insulin-producing beta cell line, J. Clin.Invest. 74 (1984) 715–722.
[32] B. Portha, C. Levacher, L. Picon, G. Rosselin, Diabetogenic effect of streptozotocin inthe rat during the perinatal period, Diabetes 23 (1974) 889–895.
[33] C. Kappen, C. Kruger, J. MacGowan, J.M. Salbaum, Maternal diet modulates placentagrowth and gene expression in a mouse model of diabetic pregnancy, PLoS One 7(2012) e38445.
[34] P. Masiello, C. Broca, R. Gross, M. Roye, M. Manteghetti, D. Hillaire-Buys, M. Novelli,G. Ribes, Experimental NIDDM: development of a new model in adult rats adminis-tered streptozotocin and nicotinamide, Diabetes 47 (1998) 224–229.
[35] O. Cisse, I. Fajardy, A. ckes-Coopman, E. Moitrot, V. Montel, S. Deloof, J. Rousseaux, D.Vieau, C. Laborie, Mild gestational hyperglycemia in rat induces fetal overgrowthand modulates placental growth factors and nutrient transporters expression,PLoS One 8 (2013) e64251.
[36] S.M. Alam, R. Ain, T. Konno, J.K. Ho-Chen, M.J. Soares, The rat prolactin gene familylocus: species-specific gene family expansion, Mamm. Genome 17 (2006) 858–877.
[37] Y.N. Sinha, Structural variants of prolactin: occurrence and physiological signifi-cance, Endocr. Rev. 16 (1995) 354–369.
[38] M.J. Soares, T. Konno, S.M. Alam, The prolactin family: effectors of pregnancy-dependent adaptations, Trends Endocrinol. Metab. 18 (2007) 114–121.
[39] M.J. Soares, S.M. Alam, T. Konno, R. Ain, The prolactin family: regulators of uterinebiology, in: J.D. Aplin, A.T. Fazleabas, S.R. Glasser, L.C. Giudice (Eds.), The Endometri-um: Molecular, Cellular and Clinical Perspectives, CRC Press, 2008, pp. 352–363.
[40] S.M. Rosenberg, I.A. Maslar, D.H. Riddick, Decidual production of prolactin in lategestation: further evidence for a decidual source of amniotic fluid prolactin, Am. J.Obstet. Gynecol. 138 (1980) 681–685.
[41] D.H. Riddick, A.A. Luciano, W.F. Kusmik, I.A. Maslar, Evidence for a nonpituitarysource of amniotic fluid prolactin, Fertil. Steril. 31 (1979) 35–39.
[42] A. Kauppila, P. Chatelain, P. Kirkinen, S. Kivinen, A. Ruokonen, Isolated prolactin de-ficiency in a woman with puerperal alactogenesis, J. Clin. Endocrinol. Metab. 64(1987) 309–312.
[43] K.F. Roby, S. Deb, G. Gibori, C. Szpirer, G. Levan, S.C. Kwok, M.J. Soares, Decidualprolactin-related protein. Identification, molecular cloning, and characterization, J.Biol. Chem. 268 (1993) 3136–3142.
[44] S.M. Alam, T. Konno, G. Dai, L. Lu, D.Wang, J.H. Dunmore, A.R. Godwin, M.J. Soares, Auterine decidual cell cytokine ensures pregnancy-dependent adaptations to a phys-iological stressor, Development 134 (2007) 407–415.
[45] D. Piwnica, I. Fernandez, N. Binart, P. Touraine, P.A. Kelly, V. Goffin, A new mecha-nism for prolactin processing into 16 K PRL by secreted cathepsin D, Mol.Endocrinol. 20 (2006) 3263–3278.
[46] M. Lkhider, R. Castino, E. Bouguyon, C. Isidoro, M. Ollivier-Bousquet, Cathepsin D re-leased by lactating rat mammary epithelial cells is involved in prolactin cleavageunder physiological conditions, J. Cell Sci. 117 (2004) 5155–5164.
[47] K. Teramo, M.A. Kari, M. Eronen, H. Markkanen, V. Hiilesmaa, High amniotic fluiderythropoietin levels are associated with an increased frequency of fetal and neona-tal morbidity in type 1 diabetic pregnancies, Diabetologia 47 (2004) 1695–1703.
[48] M. Jirkovska, T. Kucera, J. Kalab, M. Jadrnicek, V. Niedobova, J. Janacek, L. Kubinova,M. Moravcova, Z. Zizka, V. Krejci, The branching pattern of villous capillaries andstructural changes of placental terminal villi in type 1 diabetes mellitus, Placenta33 (2012) 343–351.
[49] E. Jauniaux, G.J. Burton, Villous histomorphometry and placental bed biopsy investi-gation in type I diabetic pregnancies, Placenta 27 (2006) 468–474.
[50] R. Padmanabhan, A.G.H. Al-Zuhair, A.H. Ali, Histopathological changes of the placen-ta in diabetes induced by maternal administration of streptozotocin duringpregnancy in the rat, Congenit. Anom. 28 (2006) 1–15.