at SciVerse ScienceDirect

Placenta 34 (2013) 606e612

Contents lists available

Placenta

journal homepage: www.elsevier .com/locate/placenta

Emerging role of SUMOylation in placental pathology

D. Baczyk a,*, S. Drewlo a,1, J.C.P. Kingdom a,b

aResearch Centre for Women’s and Infants’ Health, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, University of Toronto, 25 Orde Street,Room 6-1020, Toronto, Ontario M5T 3H7, CanadabDepartment of Obstetrics and Gynecology, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada

a r t i c l e i n f o

Article history:Accepted 27 March 2013

Keywords:PlacentaPreeclampsiaSUMOylation

* Corresponding author. Tel.: þ1 416 586 4800x832E-mail address: baczyk@lunenfeld.ca (D. Baczyk).

1 Present address: C.S. Mott Center for Human GrowState University School of Medicine, 275 E. Hancock,

0143-4004/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.placenta.2013.03.012

a b s t r a c t

Introduction: Small ubiquitin-like modifiers (SUMO) conjugate to target proteins in a dynamic, reversiblemanner to function as post-translational modifiers. SUMOylation of target proteins can impinge on theirlocalization, in addition to their activity or stability. Differential expression of deSUMOylating enzymes(SENP 1 and 2) contributes to altered mammalian placental development and function in mice. Severepreeclampsia (sPE) is associated with abnormal placental development and chronic ischemic injury.Extra- and intracellular stimuli/stressors that include hypoxic-activated pathways are known modulatorsof SUMOylation. In this current study we hypothesized that placentas from sPE patients will display upregulation in the SUMO regulatory pathway.Methods: Utilizing qRT-PCR, immuno-blotting and Western techniques, we determined the expressionlevels of SUMO pathway genes in healthy and diseased placentas. We also exposed placental explants tohypoxia to study the effect on the SUMOylation pathway.Results: We observed steady-state expression of SUMO1e3, SUMO-conjugated enzyme-UBC9 anddeSUMOylating enzymes e SENPs, throughout normal gestation. An elevated level of free SUMO1e3 andSUMO-protein conjugates was observed in sPE placentas. Furthermore, placental UBC9 levels werestrikingly increased in the same sPE patients. Hypoxia-induced SUMOylation in first trimester placentalexplants.Discussion: Our data demonstrate an elevated steady-state of SUMOylation in sPE placentas comparedwith gestational aged-matched controls. The observed hyper-SUMOylation in sPE placentas correlateswith elevated expression of UBC9 rather than with reduced expression of SENPs Hypoxia may contributeto alterations in placental SUMOylation pathway.Conclusion: Increased placental SUMOylation may contribute to the pathogenesis of serious placentalpathology that causes extreme preterm birth.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

One of the most common placental insufficiency disorders ispreeclampsia. While the full spectrum of preeclampsia is a multi-factorial disease of pregnancy presenting across the full range ofgestation [1], the severe form of this disease is characterized byischemic injury, inflammation and abnormal development of theplacental villi [2]. The resultant anti-angiogenic state in severepreeclampsia (sPE) created by the abnormal placenta ultimatelypresents as hypertension, proteinuria, and fetal growth restriction.

2; fax: þ1 416 586 8565.

th and Development. WayneDetroit, MI 48201, USA.

All rights reserved.

Proper differentiation and function of the outer placental layer isdependent on a number of transcription factors such as glial cellmissing-1(GCM1) [3]. In sPE placentas, GCM1 expression is reduced[4], due to upstream repression by DREAM [5]. The consequences ofreduced GCM1 expression include defective syncytialization, sinceGCM1 promotes the expression of the fusogenic protein Syncytin[6]. In this context it is of interest that UBC9-mediated SUMOylationof GCM1 inhibits its transcriptional activity [7]. This process can bereversed upon cAMP stimulation, leading to deSUMOylation ofGCM1 [8]. SUMOylation has also been implicated in the regulationof Hif 1-a protein stability and activity [9,10].

SUMOylation is a reversible post-translational modification ofproteins localized mostly within the nucleus. This process isinvolved in a number of cellular mechanisms from DNA repair, cellcycle progression to regulation of protein’s function, localizationand stability. Mammals express four SUMO (small ubiquitin-like

Table 1Primer sequences.

Primer name Primer sequence (50e30)

Sumo 1 For CAT TGG ACA GGA TAG CAG TGSumo 1 Rev TTC ATT GGA ACA CCC TGT CTSumo 2 For CCT GCA CAG TTG GAA ATG GAG GSumo 2 Rev AGT AGA CAC CTC CCG TCG TCT GCTSumo 3 For CCA AGG AGG GTG TGA AGA CASumo 3 Rev TCT CGC AGT AGG CCT TCA TCUBC9 For CCA TTT GGT TTC GTG GCT GTUBC9 Rev TCA TGA GGT TCA TCG TGC CASENP 1 For GTA CAG CAG AAG AGA CAG TSENP1 Rev CAT CTG TAG CAG CTG TCT GTSENP2 For TCT GAA GAG AGT GGC AAG GGSENP2 Rev TGA ACA CCC TCC TCC ACA GTSENP 3 For GCG TAC AGA GCA TCT TGG ACSENP 3 Rev CTA CCT CAT CAG TGC TGA GGSENP 5 For TCT GGA TTC CTA GAT GAG GTSENP 5 Rev CAT CAT CCA CTG GAG AGG CTSENP 6 For AGA TGG GAC AAA TCT GCT CASENP 6 Rev CAC CAG AGC TAT TAG CAA CASENP 7 For CAC CAG AAA GGA TAC CCA GASENP 7 Rev CTC AGT TAC AGG TGG GGT CTop1 For GAT GAA CCT GAA GAT GAT GGCTop1 Rev TCA GCA TCA TCC TCA TCT CGCYC1 For CAG ATA GCC AAG GAT GTG TGCYC1 Rev CAT CAT CAA CAT CTT GAG CCYWHAZ For ACT TTT GGT ACA TTG TGG CTT CAAYWHAZ Rev CCG CCA GGA CAA ACC AGT AT

D. Baczyk et al. / Placenta 34 (2013) 606e612 607

modifiers) isoforms: SUMO 1, SUMO 2, SUMO 3 and SUMO 4(pseudo gene). SUMO 2 and SUMO 3 share over 95% homology andtherefore are referred to as SUMO 2/3. The SUMO proteins arecovalently conjugated to its targets in a series of steps, involvingactivating enzyme E1 and a conjugating enzyme E2 (hereafterreferred to as UBC9) with or without SUMO ligase E3 enzyme[11,12]. It is noteworthy, that UBC9 is the only known conjugatingenzyme although a number of E1 and E3 enzymes have beendescribed [13]. A family of 6 sentrin-specific proteases (SENPs, 1e3,5e7) exhibits isopeptidase activity by cleaving the covalentconjugation between SUMO and its target [14]. Thus, UBC9 conju-gation promotes steady-state SUMOylation and SENPs act todeSUMOylate the proteins. For detail review of the mechanism seeRefs. [12,15].

Disruption of SUMO homeostasis has been linked to cancerdevelopment and progression [16], deregulation of mitosis [17]pathogenic infections [18] and neurodegenerative disorders [19e21].

Knockout studies, in a number of species, have revealed thatsome of the SUMO pathway proteins are redundant (excellent re-view [22]). For instance in zebra fish, single deletions of SUMO 1 orSUMO 2 do not appear to affect its development. Deletion of all 3SUMOs however, leads to serious developmental defects [23].Conversely, UBC9 is a critical, non-redundant component of theSUMOylation pathway. UBC9 is essential for early development andknockout mice die by E7.5 [24]. Altered placental development andfunction affecting embryonic viability has also been described intwo additional mouse models: SENP1 and SENP2 transgenic mice[25,26].

SUMOylation is induced in response to external stimuli/stressorssuch as hypoxia and ROS [27,28], features which are commonlyfound in SPE and intrauterine growth restriction (IUGR) (reviewedin [29]). We thus hypothesized that placentas from sPE patientsshow up regulation in the SUMO pathway expression. The studyobjectives were to examine the expression levels of the SUMOpathway genes in healthy and diseased placentas. Our resultsconfirm that sPE is linked to UBC9-mediated hyper-SUMOylation.

2. Methods

2.1. Tissue collection

Placental villous samples from first and second trimester were obtained from theMorgentaler Clinic, Toronto, Canada, following an elective legal termination ofpregnancy. Mount Sinai Hospital (MSH) Research Ethics Board approval (MSH REB#04-0018-U) was obtained for this study and all patients gave written informedconsent.Gestational age andviabilitywere establishedpre-operatively byultrasound.

Placentas from singleton pathologic pregnancies and age-matched controls oflive born infants delivered between 24þ 0 and 34þ 6 weeks of gestation, to healthywomen, were obtained by the Research Centre for Women’s and Infants’ Health(RCWIH) BioBank program of Mount Sinai Hospital, in accordance with the policiesof theMount Sinai Hospital Research Ethics Board (MSH REB#10-0128-E). Theywereclassified into different groups, with characteristics summarized in SupplementaryTable 1: Severe intrauterine growth restriction (sIUGR): Birth weight �10&,abnormal umbilical artery Doppler (absent end diastolic flow velocity (AEDFV) orreversed end diastolic flow velocity (REDFV)), normotensive women. Severe pre-eclampsia (sPE): Birth weight >10%ile, BP> 140/90, proteinuria >300 mg/day or�1þ on dipstick. Preterm control (PTC): Birth weight >20th centile, no abnormalumbilical artery Doppler measurements, BP< 140/90, no history of gestationaldiabetes, no histologic evidence of acute chorio-amnionitis. Control term placentas(>37þ n weeks gestation) were also collected for gestational age profile studies.Similar patient populationwas previously used by our group [30]. Placental sampleswere para-formaldehyde fixed for immunohistochemistry. Pooled samples fromindividual placentas were generated by combining randomly selected 4 tissue piecesfrom each of the four cores. Samples were then snap frozen and stored at �70 �C forfuture RNA and protein isolation.

2.2. First trimester explant cultures

First trimester placental tissues (8e10 weeks, n¼ 5) were obtained followingelective termination of pregnancy (see Section 2.1 above). Small pieces (20e30 mgwet weight) were micro-dissected under a microscope as previously described [5].

Explants were cultured overnight in 1 ml of DMEM/F12 media (Invitrogen, Bur-lington, ON, Canada) containing 1% liquid media supplement (ITSþ 1 (Sigma)),100 U/ml of penicillin, 100 U/ml streptomycin, 2 mM L-glutamine and 100 mg/mlgentamacin and 2.5 mg/ml fungizone (all from Invitrogen) under physiological ten-sion of 8% O2. The following morning the explants were transferred to fresh platescontaining new media pre-equilibrated with various oxygen tensions overnight.Explants were then cultured for a further 3 h or 24 h under hypoxic (1% and 3%) ornormoxic 8% O2. All of the experiments were performed in technical triplicates.Following the completion of the experiment tissues were collected into QiaZol�

extraction buffer and RNA isolated according to the manufacturer’s specifications(Qiagen, Toronto, ON, Canada).

2.3. Immunohistochemistry

Para-formaldehyde fixed (4%) placental tissues were wax embedded. Immuno-histochemistry was performed on rehydrated sections with the biotin-streptavidinstaining procedure, using Peroxidase Dako LSAB kit (Dako Canada Inc. Mis-sissauga, ON, Canada). Primary antibodies to SUMO1, 1:100 dilution (Abcam, Cam-bridge, MA, USA) and SUMO 2/3, 1:100 dilution (Abcam) and UBC9, 1:400 dilution(Abcam) were applied to sections overnight at 4 �C. Specific biotinylated secondaryantibodies were applied to sections for 1 h at RT.

Negative controls included omission of the primary antibody and use ofnonspecific matched IgG. Slides were counterstained with hematoxylin (Sigma,Oakville, ON, Canada), visualized using a Nikon DMRX light microscope and pho-tographed using a Sony PowerHAD 3CCD color video camera DXC-970ND (Sony,Toronto, ON, Canada). Slides were assessed blind to clinical category.

2.4. Reverse transcription and qRT-PCR

A total of 90 placental RNA samples from 9 different clinical groups were iso-lated. These groups are summarized in Supplementary Table 1 and include; healthyfirst trimester (early and late), healthy second trimester, preterm controls, termcontrols, severe early onset intrauterine growth restriction (sIUGR) and severepreeclampsia (sPE). RNAwas isolated using the QiaZol� extraction kit (Qiagen), thenquantified and quality-controlled using Nanodrop (Thermo-Fisher, Rockford, IL,USA) and Experion RNA chips (Bio-Rad, Mississauga, ON, Canada). One microgram oftotal RNA from each sample was reverse-transcribed simultaneously using theBIORAD i-Script Supermix (Bio-Rad). cDNA samples (10 ng) were pipetted with arobot from Perkin Elmer (total volume of the reactionwas 7 ml and included 3.5 ml ofLuminoCT from Sigma) and run on the CFX Bio-Rad 384 PCR machine. All sampleswere run at the same time per gene to avoid inter-assay variations with thefollowing PCR protocol: initial 95 �C for 5 min followed by 38 cycles of 95 �C for 15 sand 60 �C for 20 s. Housekeeping genes (TOP1, CYC1 and YWHAZ) were chosenbased on an earlier study of this sample pool [30]. Primer sequences are summarizedin Table 1.

D. Baczyk et al. / Placenta 34 (2013) 606e612608

2.5. Western blotting

Snap frozen sPE and control placental samples were pulverized in liquid nitro-gen. The resulting powder was then homogenized in lysis buffer containing: 0.3%(w/v) SDS, 0.3% glycerol, 50 mM Tris, pH 6.8 and 10 mM NEM. Thirty milligrams ofpurified proteins/well were separated on gradient pre-cast gels (cat #4561096, Bio-Rad) in 1�TG-SDS buffer (Wisent, Saint-Jean-Baptiste, QC, Canada). Proteins weretransferred to PVDF membrane using Trans-Blot Turbo� Transfer System (Bio-Rad).Membranes were blocked with 5% blotting-grade blocker (Bio-Rad) in 1�TBSTbuffer for 1 h. Primary antibodies: (SUMO1 (1:300), SUMO 2/3 (1:300) and UBC9(1:2000)) incubation was carried out overnight at 4 �C. Secondary HRP-conjugatedantibodies were incubated for 1 h at room temperature. Western Lumiol reagent(Perkin Elmer, Woodbridge, ON, Canada) was used to start luminescence reactionand subsequently exposed to X-ray film. The blots were stripped using Restore�PlusStripping buffer (Thermo-Fisher) and re-probed with a-tubulin antibody (Sigma;1:3000 dilution) for normalization purposes. Densitometry analysis was performedwith the Bio-Rad ImageQuant software.

2.6. Statistical analysis

All experiments were performed in at least technical triplicates and biologicalquadruplicates. Data are presented as mean with errors expressed in S.E.M. Whereapplicable, unpaired Student’s t-test for two group comparison and one way ANOVAtests for multiple group comparisons were applied with the appropriate post-tests.Statistical calculations were performed using SigmaStat software and p-values�0.05 (*) were considered significant.

3. Results

3.1. SUMO expression in healthy pregnancy and placentalinsufficiency syndromes

RNA from healthy first, second and third trimester placentas aswell as from sPE, and preterm controls (PTC) was subjected to qRT-PCR analysis. We observed a steady-state mRNA expression ofSUMO1, 2 and 3 in healthy placentas over gestation. sPE placentas

Fig. 1. Elevated expression levels of SUMO1e3 in sPE. qRT-PCR was utilized to assess mRNAand in placental insufficiency syndromes. Data were normalized to 3 housekeeping genes aN¼ 9e11/group; early first GA¼ 6e7 weeks, late first¼ 8e12 weeks, 2nd¼ 13e16 weeks,uterine growth restriction without preeclampsia (<34 weeks), sPE¼ severe early onset pre

however, showed up regulation of mRNA expression in all threeSUMO genes as compared to aged-matched controls (Fig. 1AeC).SUMO1 mRNA expression in sPE placentas was significantlyelevated (1.38� 0.22 fold, p¼ 0.013) compared with age-matchedcontrols (Fig. 1A). SUMO2 and SUMO 3 mRNA expression in sPEplacentas were significantly elevated (2.80� 0.88 fold, p¼ 0.05 and2.15� 0.417 fold, p¼ 0.028, respectively) compared with age-matched controls (Fig. 1B and C). Late first trimester placentasshowed a significant reduction in SUMO3 mRNA expression(0.91�0.026 fold, p¼ 0.017) compared to placentas from pretermcontrol patients (Fig. 1C).

Our observations were confirmed with immuno-blotting; wedetected increased immunostaining for SUMO1 and SUMO2/3(SUMO 2 and 3 are indistinguishable on the protein level) in sPEplacentas as compared to age-matched controls (Fig. 2). ElevatedSUMO mRNA levels and SUMO immunostaining suggest elevatedlevels of global SUMOylation in placentas complicated by sPE. Wetherefore undertook quantification of total SUMO-conjugatedproteins, evidenced by the laddering of proteins above the freeSUMO band, in extracts from sPE and PTC.Western analysis showedsignificantly higher levels of SUMO1 (1.43� 0.095 fold, p¼ 0.0058)and especially SUMO2/3 (2.96� 0.92 fold, p¼ 0.0313) conjugatedproteins in sPE placentas as compared to age-matched controls(Fig. 3).

3.2. UBC9-mediated hyper-SUMOylation in sPE

The next set of experiments was designed to explore the po-tential mechanism inducing the hyper-SUMOylation in sPE pla-centas. Since UBC9 is a critical conjugating enzyme in theSUMOylation pathway, we analyzed its mRNA expression level andfound no changes over gestation. However, placenta insufficiency

expression of SUMO1 (A) SUMO2 (B) and SUMO3 (C) in human placenta over gestationnd compared to preterm age-matched control group and is presented as mean� SEM.PTC¼ preterm age-matched control before 34 weeks, IUGR¼ severe early onset intra-eclampsia with the absence of growth restriction (<34 weeks), *p< 0.05.

Fig. 2. SUMO 1 and 2/3 staining in control and sPE placentas. Histological assessment of controls (top panel) and sPE placentas (bottom panel) revealed induced expression ofSUMO1 (A, C) and SUMO2/3 (B, D) as compared to the placental pathology. In sPE placenta, SUMOs are mainly expressed in nuclei of a subset of trophoblast and mesenchymal cells.Lighter cytoplasmic expression of SUMO2/3 was also observed in sPE placenta (D). Each field is a representative of 10 samples examined (original magnification; AeD¼ 400�).

Fig. 3. Hyper-SUMOylation in sPE placentas. Western blot analysis was performed on placental protein isolated from sPE and age-matched controls (AeB). Free SUMO levels areindicated with arrows and SUMO-conjugated proteins are observed above. Total densitometry of SUMO1 (C) and SUMO2/3 (D) conjugated proteins was quantified and normalizedto loading control a-tubulin. Data are presented as mean� SEM of 5 placentas/group. *p< 0.05.

D. Baczyk et al. / Placenta 34 (2013) 606e612 609

D. Baczyk et al. / Placenta 34 (2013) 606e612610

syndromes such as, sIUGR, MIUGR and sPE placentas showed asignificant increase in UBC9 mRNA expression (3.45�1.01,p¼ 0.042; 5.89�1.49, p¼ 0.022 and 8.43�1.87, p¼ 0.013) fold,respectively, as compared to PTC (Fig. 4A). These results werefurther validated with UBC9 Western, where sPE placentasdemonstrated 2.36� 0.472 fold (p¼ 0.035) increase in UBC9 pro-tein expression compared to age-matched controls (Fig. 4B).Moreover, PTC samples showed very low level of UBC9 immuno-staining (Fig. 4C) whereas by contrast we observed intense im-munostaining in sPE placentas (Fig. 4D).

3.3. Hypoxia induces SUMOylation

Since hypoxia has been shown to induce SUMOylation in manyother tissues and models [28,31e33], we tested the hypothesis thatthe exposure of first trimester placental explants to hypoxia willinduce SUMO pathway. Exposure of explants to hypoxia (1% and3%) up regulated the mRNA expression of all of the participatinggenes as compared to normoxia (8%). Specifically, 3-h exposure to1% O2 up regulated SUMO1e3 by 1.42� 0.0.09, p¼ 0.042;1.46� 0.12, p¼ 0.048; 1.36� 0.10, p¼ 0.0093 fold, respectively, andUBC9 by 1.34� 0.10, fold p¼ 0.022. Same time point with milderhypoxia of 3% resulted in up regulation of SUMO 2 and 3 by

Fig. 4. Elevated expression level of UBC9 in sPE placentas. qRT-PCR was utilized to assess Usyndromes (A). Data were normalized to 3 housekeeping genes and compared to pretermweeks, late first¼ 8e12 weeks, 2nd¼ 13e16 weeks, PTC¼ preterm control before 34 we(<34 weeks), sPE¼ severe early onset preeclampsia with absence of growth restriction (<34(B). UBC 9 Western blot analysis on sPE and age-matched control samples was normalized t*p< 0.05. Observations were confirmed with histological examination of placentas (CeD). SsPE placentas (D). Each field is a representative of 10 samples examined (original magnific

1.18� 0.07 fold, p¼ 0.041 and 1.25� 0.11 fold, p¼ 0.05, respec-tively. Furthermore hypoxia (1%) resulted in elevated mRNA levelsof DREAM (1.53� 0.11, p¼ 0.0033). Explants exposed to milderhypoxia (3%) also exhibited elevated levels of DREAM mRNA(1.59� 0.156, p¼ 0.009) and reduced levels of GCM1 mRNA(0.50� 0.060, p¼ 0.0002). Exposure of explants to 24 h of hypoxia,1% and 3%, led to reduced GCM1 mRNA levels to 0.82� 0.059 fold,p¼ 0.021 and 0.59� 0.11 fold, p¼ 0.01, respectively. (Fig. 5).

Hyper-SUMOylation may be due to reduced levels of thedeconjugating SENPs. We analyzed the mRNA expression levels ofSENPs in our placental samples. Consistent with observed SUMOand UBC9 data, healthy placentas from first, second and third tri-mesters showed no differences in mRNA expression level. However,we found no differences in the expression levels of SENP1e3 andSENP6e7 in the sPE as compared to PTC (Supplementary Fig. S1).SENP5mRNA expressionwas below the detection levels of the qRT-PCR reaction.

4. Discussion

This is the first report linking SUMOylation to seriousplacental pathology that causes extreme preterm birth. Our datademonstrate an elevated steady-state SUMOylation in sPE placentas

BC9 mRNA expression in human placenta over gestation and in placental insufficiencycontrol group and is presented as mean� SEM. N¼ 9e11/group; early first GA¼ 6e7eks, IUGR¼ severe early onset intrauterine growth restriction without preeclampsiaweeks). Elevated expression of UBC9 in sPE placenta was also observed on protein levelo loading control a-tubulin. Data were presented as mean� SEM of 5 placentas/group,trong nuclear expression of UBC9 was detected in trophoblast cells and stromal cells ofation; CeD¼ 630�).

Fig. 5. Hypoxia induced SYMOylation. First trimester human placental explants (8e10 weeks, n¼ 5) were cultured under hypoxic (1% and 3%) or normoxic (8%) O2 for 3 h (A) or 24 h(B). qRT-PCR was utilized to assess the mRNA levels of SUMO1e3, UBC9 and transcription factors, DREAM and GCM1. Data were normalized to 3 housekeeping genes and comparedto 8% O2 and is presented as mean� SEM. Three hour exposure to hypoxia, up regulated all of the SUMOs, UBC9 as well as DREAM, while GCM1 mRNA levels were elevated (A). After24 h of culture, GCM1 mRNA levels remained significantly reduced (B). White bars¼ 1% O2 and filled bars¼ 3% O2.

D. Baczyk et al. / Placenta 34 (2013) 606e612 611

as compared to aged-matched controls. This phenomenon appearsto be hypoxia-induced.

The qRT-PCR and immuno-blotting data indicate a steady-stateexpression of SUMO peptides, UBC9 and deconjugating enzymes inthe placenta over human gestation. Significant up regulation inSUMO1e3 and UBC9 was, however, observed in placentas from se-vere early onset pre-eclamptic placentas as compared to age-matched control placentas. More importantly, in addition toincreased expression of free SUMOs, we also observed elevatedSUMO1 and especially SUMO2/3-protein conjugates; taken togetherour data suggest global hyper-SUMOylation of proteins in sPE pla-centas. Furthermore, exposure offirst trimester placental explants tohypoxia resulted in up regulation of SUMO pathway in vitro.

Twomousemodels support thepremise thathyper-SUMOylationis detrimental to placental development and function. The firstmodel takes advantage of a randomretroviralmutation in themouseSENP1 gene, resulting in reduced SENP1 expression [25]. Althoughelevated levels of SUMO1 conjugated proteins were observed in thedeveloping placentas and the fetuses of the SENP1 mutants, em-bryonic death (between E12.5 and E14.5) was entirely due to defectswithin the placental labyrinth [25]. Detailed histologic analyses ofthe placentas are missing but the authors noted reduced fetal andmaternal blood spaces. The second model is the SENP2 knockoutmouse. First, the authors reported that in thewild typemice, SENP2expression appears to be restricted only to the extra embryonictissue [26]. Second, SENP2 knockout mice die by E11.5 due toimpaired placental development. Detailed histological analysisrevealed that all three trophoblast layers: labyrinth, spongio-trophoblast and trophoblast giant cells, were reduced or absent inthe knockout mice (by E10.5) as compared to their wild type litter-mates [26]. Furthermore, the maternal and fetal blood spaces werediminished in the SENP2 knock-outs.

Possible contributors to elevated levels of global SUMOylation insPE placenta might include the known culprits associated with thisplacental defect namely: hypoxia [34,35], ROS [36] and inflamma-tion [37]. High oxidative stress or prolonged hypoxia enhancesSUMOylation particularly SUMO2/3 levels [34]. Our data showelevation in free SUMOs and in SUMOylated protein conjugates insPE placentas, suggesting that sPE placenta up regulate the SUMOpathway in response to stressors. Our observations from placentalexplant experiments, support the hypothesis that hypoxia couldcontribute to the observed global up regulation of the SUMOpathway in sPE placentas.

It has been postulated previously that SUMO2/3-ylation is anadaptive response to elevated ROS production [28,32]. We observedelevated levels of SUMO2and3atboth1%and3%oxygen tension.We

thus speculate that the observed hyper-SUMOylation in sPE placentamight play a protective role providing tissue tolerance to chronicischemic injury. Hibernating animals, such as squirrels, drastically upregulate the levels of Ubc9 during torpor [33]; furthermore Lee et al.,demonstrated that elevated global SUMOylation in Ubc9 overexpressing mice, protects their brains from ischemic brain damage[38]. Future studies inour laboratorywill include themanipulationofUBC9 expression in BeWo cells and placental explant models to testthe hypothesis that SUMOylation regulates the expression of sPEassociated genes such as GCM1 and sFLT1.

Inmost scenarios elevated SUMOylation represses the activity oftranscription factors [39,40] such as GCM1 [7]. SUMOylation of arelated upstream transcription factor, PPARg [41], results inrepression of its translocation [42] and thus functional inactivation.Conversely, SUMOylation of the target proteins can result in in-duction of their stability and activity, e.g. Hif1a [32], or in trans-location and thus induction of stability, e.g. DREAM [5,43]. Theexpression and/or activity of all the above-mentioned transcriptionfactors is altered in sPE placentas. An elevated level of globalSUMOylation observed in sPE placentas contributes to SUMOyla-tion and induction of DREAM activity, a transcriptional repressorwe have recently shown to have increased expression in placentasfrom sPE women [5]. In turn, DREAM reduces the expression of itsdownstream target GCM1 [3]. GCM1 has been reported to havereduced expression in sPE placentas [4]. Furthermore UBC9-mediated SUMOylation of GCM1 represses its transcriptional ac-tivity [7], thus possibly leading to reduced levels of Syncitin 1(known downstream target of GCM1) in sPE placentas [44,45].

Irrespective of whether or not hyper-SUMOylation is the resultof ischemia-reperfusion injury [46] or it plays a protective roleagainst such external stressors, its homeostasis is central to propercell/organ function. sPE placentas are associated with alterations inthe expression and/or activity of a number of transcription factors.Our results point to a potentially novel common mechanism oftheir regulation. Further research to decipher SUMO’s specific rolein normal and abnormal placental development and function couldlead to novel targets for therapeutic interventions.

Acknowledgments

The research was supported by Canadian Institute of HealthResearch (Grant No. 64302) to JK as well as Department of Ob-stetrics & Gynecology and Rose Torno Chair atMount Sinai Hospital.Furthermore, the authors thank the donors and the Research Centrefor Women and Infant’s Health Biobank Program of the CIHR Groupin Development and Fetal Health (CIHR no. MGC-13299), the

D. Baczyk et al. / Placenta 34 (2013) 606e612612

Samuel Lunenfeld Research Institute and the MSH/UHN Depart-ment of Obstetrics and Gynecology for the human specimens usedin this study.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.placenta.2013.03.012.

References

[1] North RA, McCowan LM, Dekker GA, Poston L, Chan EH, Stewart AW, et al.Clinical risk prediction for pre-eclampsia in nulliparous women: developmentof model in international prospective cohort. BMJ 2011;342:d1875.

[2] Walker MG, Fitzgerald B, Keating S, Ray JG, Windrim R, Kingdom JC. Sex-specific basis of severe placental dysfunction leading to extreme pretermdelivery. Placenta 2012;33(7):568e71.

[3] Baczyk D, Drewlo S, Proctor L, Dunk C, Lye S, Kingdom J. Glial cell missing-1transcription factor is required for the differentiation of the human tropho-blast. Cell Death Differ 2009;16(5):719e27.

[4] Chen CP, Chen CY, Yang YC, Su TH, Chen H. Decreased placental GCM1 (glialcells missing) gene expression in pre-eclampsia. Placenta 2004;25(5):413e21.

[5] Baczyk D, Kibschull M, Mellstrom B, Levytska K, Rivas M, Drewlo S, et al.DREAM Mediated Regulation of GCM1 in the Human Placental Trophoblast.PLoS One 2013;8(1):e51837.

[6] Knerr I, Beinder E, Rascher W. Syncytin, a novel human endogenous retroviralgene in human placenta: evidence for its dysregulation in preeclampsia andHELLP syndrome. Am J Obstet Gynecol 2002;186(2):210e3.

[7] Chou CC, Chang C, Liu JH, Chen LF, Hsiao CD, Chen H. Small ubiquitin-likemodifier modification regulates the DNA binding activity of glial cellmissing Drosophila homolog a. J Biol Chem 2007;282(37):27239e49.

[8] Chang CW, Chang GD, Chen H. A novel cyclic AMP/Epac1/CaMKI signalingcascade promotes GCM1 desumoylation and placental cell fusion. Mol CellBiol 2011;31(18):3820e31.

[9] Bae SH, Jeong JW, Park JA, Kim SH, Bae MK, Choi SJ, et al. Sumoylation in-creases HIF-1alpha stability and its transcriptional activity. Biochem BiophysRes Commun 2004;324(1):394e400.

[10] Carbia-Nagashima A, Gerez J, Perez-Castro C, Paez-Pereda M, Silberstein S,Stalla GK, et al. RSUME, a small RWD-containing protein, enhances SUMOconjugation and stabilizes HIF-1alpha during hypoxia. Cell 2007;131(2):309e23.

[11] Creton S, Jentsch S. SnapShot: the SUMO system. Cell 2010;143(5):848e848.e1.[12] Geiss-Friedlander R, Melchior F. Concepts in sumoylation: a decade on. Nat

Rev Mol Cell Biol 2007;8(12):947e56.[13] Hay RT. SUMO: a history of modification. Mol Cell 2005;18(1):1e12.[14] Yeh ET, Gong L, Kamitani T. Ubiquitin-like proteins: new wines in new bottles.

Gene 2000;248(1e2):1e14.[15] Wilkinson KA, Henley JM. Mechanisms, regulation and consequences of pro-

tein SUMOylation. Biochem J 2012;428(2):133e45.[16] Bawa-Khalfe T, Yeh ET. SUMO losing balance: SUMO proteases disrupt SUMO

homeostasis to facilitate cancer development and progression. Genes Cancer2010;1(7):748e52.

[17] Dasso M. Emerging roles of the SUMO pathway in mitosis. Cell Div 2008;3:5.[18] Wilson VG, Rangasamy D. Viral interaction with the host cell sumoylation

system. Virus Res 2001;81(1-2):17e27.[19] Terashima T, Kawai H, Fujitani M, Maeda K, Yasuda H, et al. SUMO-1 co-

localized with mutant atrophin-1 with expanded polyglutamines acceleratesintranuclear aggregation and cell death. Neuroreport 2002;13(17):2359e64.

[20] Pountney DL, Huang Y, Burns RJ, Haan E, Thompson PD, Blumbergs PC, et al.SUMO-1 marks the nuclear inclusions in familial neuronal intranuclear in-clusion disease. Exp Neurol 2003;184(1):436e46.

[21] McMillan LE, Brown JT, Henley JM, Cimarosti H. Profiles of SUMO and ubiq-uitin conjugation in an Alzheimer’s disease model. Neurosci Lett 2012;502(3):201e8.

[22] Lomeli H, Vazquez M. Emerging roles of the SUMO pathway in development.Cell Mol Life Sci 2011;68(24):4045e64.

[23] Yuan H, Zhou J, Deng M, Liu X, Le Bras M, de The H, et al. Small ubiquitin-related modifier paralogs are indispensable but functionally redundant dur-ing early development of zebrafish. Cell Res 2010;20(2):185e96.

[24] Nacerddine K, Lehembre F, Bhaumik M, Artus J, Cohen-Tannoudji M,Babinet C, et al. The SUMO pathway is essential for nuclear integrity andchromosome segregation in mice. Dev Cell 2005;9(6):769e79.

[25] Yamaguchi T, Sharma P, Athanasiou M, Kumar A, Yamada S, Kuehn MR. Mu-tation of SENP1/SuPr-2 reveals an essential role for desumoylation in mousedevelopment. Mol Cell Biol 2005;25(12):5171e82.

[26] Chiu SY, Asai N, Costantini F, Hsu W. SUMO-specific protease 2 is essential formodulating p53-Mdm2 in development of trophoblast stem cell niches andlineages. PLoS Biol 2008;6(12):e310.

[27] TathamMH, Jaffray E, Vaughan OA, Desterro JM, Botting CH, Naismith JH, et al.Polymeric chains of SUMO-2 and SUMO-3 are conjugated to protein sub-strates by SAE1/SAE2 and Ubc9. J Biol Chem 2001;276(38):35368e74.

[28] Manza LL, Codreanu SG, Stamer SL, Smith DL, Wells KS, Roberts RL, et al.Global shifts in protein sumoylation in response to electrophile and oxidativestress. Chem Res Toxicol 2004;17(12):1706e15.

[29] Myatt L. Review: reactive oxygen and nitrogen species and functional adap-tation of the placenta. Placenta 2010;31(Suppl.):S66e9.

[30] Drewlo S, Levytska K, Kingdom J. Revisiting the housekeeping genes of humanplacental development and insufficiency syndromes. Placenta 2012;33(11):952e4.

[31] Leitao BB, Jones MC, Brosens JJ. The SUMO E3-ligase PIAS1 couples reactiveoxygen species-dependent JNK activation to oxidative cell death. Faseb J2011;25(10):3416e25.

[32] Huang C, Han Y, Wang Y, Sun X, Yan S, Yeh ET, et al. SENP3 is responsible forHIF-1 transactivation under mild oxidative stress via p300 de-SUMOylation.EMBO J 2009;28(18):2748e62.

[33] Lee YJ, Miyake S, Wakita H, McMullen DC, Azuma Y, Auh S, et al. ProteinSUMOylation is massively increased in hibernation torpor and is critical forthe cytoprotection provided by ischemic preconditioning and hypothermia inSHSY5Y cells. J Cereb Blood Flow Metab 2007;27(5):950e62.

[34] Saitoh H, Hinchey J. Functional heterogeneity of small ubiquitin-related pro-tein modifiers SUMO-1 versus SUMO-2/3. J Biol Chem 2000;275(9):6252e8.

[35] Bossis G, Melchior F. Regulation of SUMOylation by reversible oxidation ofSUMO conjugating enzymes. Mol Cell 2006;21(3):349e57.

[36] Pandey D, Chen F, Patel A, Wang CY, Dimitropoulou C, Patel VS, et al. SUMO1negatively regulates reactive oxygen species production from NADPH oxi-dases. Arterioscler Thromb Vasc Biol 2011;31(7):1634e42.

[37] Bekes M, Drag M. Trojan horse strategies used by pathogens to influence thesmall ubiquitin-like modifier (SUMO) system of host eukaryotic cells. J InnateImmun 2012;4(2):159e67.

[38] Lee YJ, Mou Y, Maric D, Klimanis D, Auh S, Hallenbeck JM. Elevated globalSUMOylation in Ubc9 transgenic mice protects their brains against focal ce-rebral ischemic damage. PLoS One 2011;6(10):e25852.

[39] Gill G. Post-translational modification by the small ubiquitin-related modifierSUMO has big effects on transcription factor activity. Curr Opin Genet Dev2003;13(2):108e13.

[40] Verger A, Perdomo J, Crossley M. Modification with SUMO. A role in tran-scriptional regulation. EMBO Rep 2003;4(2):137e42.

[41] Parast MM, Yu H, Ciric A, Salata MW, Davis V, Milstone DS. PPARgammaregulates trophoblast proliferation and promotes labyrinthine trilineage dif-ferentiation. PLoS One 2009;4(11):e8055.

[42] Shimizu M, Yamashita D, Yamaguchi T, Hirose F, Osumi T. Aspects of theregulatory mechanisms of PPAR functions: analysis of a bidirectional responseelement and regulation by sumoylation. Mol Cell Biochem 2006;286(1e2):33e42.

[43] Palczewska M, Casafont I, Ghimire K, Rojas AM, Valencia A, Lafarga M, et al.Sumoylation regulates nuclear localization of repressor DREAM. BiochimBiophys Acta 2011;1813(5):1050e8.

[44] Langbein M, Strick R, Strissel PL, Vogt N, Parsch H, Beckmann MW, et al.Impaired cytotrophoblast cell-cell fusion is associated with reduced Syncytinand increased apoptosis in patients with placental dysfunction. Mol ReprodDev 2008;75(1):175e83.

[45] Chen CP, Wang KG, Chen CY, Yu C, Chuang HC, Chen H. Altered placentalsyncytin and its receptor ASCT2 expression in placental development and pre-eclampsia. BJOG 2006;113(2):152e8.

[46] Burton GJ, Yung HW. Endoplasmic reticulum stress in the pathogenesis ofearly-onset pre-eclampsia. Pregnancy Hypertens 2012;1(1e2):72e8.