- Email: [email protected]
Anti-angiogenic collagen fragment arresten is increased from 16 weeks' gestation in pre-eclamptic plasma
Placenta 36 (2015) 1300e1309
Contents lists available at ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Anti-angiogenic collagen fragment arresten is increased from 16 weeks' gestation in pre-eclamptic plasma Hannah E.J. Yong*, Padma Murthi 1, May H. Wong, Bill Kalionis, Shaun P. Brennecke, Rosemary J. Keogh Department of Perinatal Medicine Pregnancy Research Centre and The University of Melbourne Department of Obstetrics and Gynaecology, The Royal Women's Hospital, Parkville 3052, Victoria, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 June 2015 Received in revised form 24 August 2015 Accepted 26 August 2015
Introduction: Arresten and canstatin are endogenous anti-angiogenic factors derived from type IV collagen a-chains COL4A1 and COL4A2 respectively. While their functions are explored in cancer studies, little is known about their role in pregnancy. Pre-eclampsia (PE) is a common, serious hypertensive disorder of pregnancy that is characterised by systemic endothelial dysfunction. COL4A1 and COL4A2 are maternal PE susceptibility genes that have increased mRNA expression in PE decidua. Our study aim was to determine the levels of arresten and canstatin in plasma and decidua from PE and gestational age matched normotensive patients. Methods: Plasma was collected from normotensive (n ¼ 44) and PE (n ¼ 39) women during the second and third trimesters of pregnancy. Third trimester decidua was collected at delivery from normotensive and PE women (n ¼ 4 each). Levels of arresten and canstatin were determined by Western immunoblotting. Results: Arresten levels were significantly increased in second and third trimester PE plasma, and in third trimester PE decidua (p < 0.05). Third trimester PE plasma arresten levels also significantly correlated with the need for MgSO4 treatment, where a 1.7 fold increase was observed in patients requiring MgSO4 treatment (p < 0.05). No significant change in canstatin levels was observed between normotensive and PE patients. Discussion: This is the first study to report significant increases in the levels of collagen fragment arresten in PE plasma and decidua. Given its significant increase before the onset of clinical disease and associations with clinical severity in the third trimester, arresten may be a useful biomarker for predicting PE and monitoring its severity. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Pre-eclampsia Arresten Canstatin Plasma Clinical severity
1. Introduction Pre-eclampsia (PE) is a serious pregnancy complication with no known cause. PE is characterised by new onset hypertension and proteinuria after 20 weeks' gestation in a previously normotensive woman [1,2]. PE affects between 2 and 8% of all pregnancies and is a leading cause of maternal and fetal morbidity and mortality worldwide [3]. The only cure for PE is the removal of the placenta
* Corresponding author. The Royal Women's Hospital, Locked Bag 300, Corner Grattan Street and Flemington Road, Parkville 3052, Victoria, Australia. E-mail address: [email protected] (H.E.J. Yong). 1 Present Address: Department of Medicine, Monash University, Monash Medical Centre, Clayton 3168, Victoria, Australia. http://dx.doi.org/10.1016/j.placenta.2015.08.013 0143-4004/© 2015 Elsevier Ltd. All rights reserved.
[4]. Approximately 5% of women with PE also develop the haemolysis, elevated liver enzymes, and low platelet count (HELLP) syndrome, which can progress rapidly to become life-threatening [5,6]. Another complication that may develop from PE is eclampsia, where the woman suffers from seizures that may lead to death [5]. Depending on the severity of the mother's condition, the fetus may need to be delivered prematurely. Over 40% of medically indicated preterm births are attributed to PE [7]. Preterm births are associated with adverse outcomes including increased risks of neonatal death, poor long term health and delayed mental development outcomes [8]. Our previous genetic linkage studies identified COL4A1 and COL4A2 as maternal PE susceptibility genes [9]. COL4A1 and COL4A2 code for type IV collagen a chain 1 and a chain 2 respectively. The mRNA expression levels of these genes are significantly increased
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
in PE decidua and these increased levels correlate with an earlier onset of PE [10]. The a1 and a2 chains form a triple helix isoform [a1(IV)]2a2(IV), which is expressed in all basement membranes [11]. Type IV collagen forms the scaffolding of basement membranes in tissues, allowing other basement membrane proteins to be incorporated and for cell interaction to occur with the basement membrane [11]. Type IV collagen is detectable in maternal blood and is significantly increased in third trimester PE serum [12]. Serum collagen levels reflect the degradation of the basement membrane and damage to the endothelium [12]. Degradation of type IV collagen generates many protein fragments [13]. A study by Bieglmayer and Hofer [14] showed an increase of type IV collagen non-collagenous (NC1) domain fragments in third trimester PE serum. Circulating NC1 fragment levels are also increased in patients with autoimmune, metastatic tumour, liver and renal disease [15,16]. These studies did not identify which of the six different a chains of type IV collagen was responsible for the observed increases. The six a chains of type IV collagen produce six distinct NC1 fragments, which are derived from the carboxyl terminus of the a chains [17]. NC1 fragments arresten (COL4A1), canstatin (COL4A2) and tumstatin (COL4A3) have anti-angiogenic effects that increase endothelial cell apoptosis and decrease endothelial cell proliferation, migration and tube formation [17e19]. Research on these fragments is currently focused on their potential as anti-cancer therapeutics by targeting tumour angiogenesis [20]. To date, the levels of each NC1 fragment of type IV collagen have not been investigated and their roles in pregnancy are unknown. Our earlier findings showed increased decidual expression of COL4A1 and COL4A2 in PE, which correlated with an earlier onset of PE, a marker of clinical severity [10]. Having shown a link between PE and increased collagen expression, the study aims were to measure and compare the second and third trimesters levels of arresten and canstatin in the plasma of PE women compared with normotensive controls, to correlate these levels with clinical parameters relating to the severity of PE and to determine whether the maternal decidua was a source of these fragments. 2. Methods 2.1. Study criteria PE was defined as pregnancy induced hypertension with a systolic pressure of 140 mmHg and/or a diastolic pressure of 90 mmHg, accompanied by proteinuria of 0.3 g in 24 h or a spot urine dipstick of ‘2þ’ [21,22]. Exclusion criteria were factors known to increase the risk of PE such as systemic lupus erythematosus and pre-existing hypertension. Blood pressures of normotensive patients were less than 140/90 mmHg. A non-treating obstetrician independently verified patient clinical records. Each patient provided written informed consent. Research and ethics approval were obtained from The Royal Women's Hospital Research and Ethics Committees. 2.2. Plasma sample collection Blood samples were collected from n ¼ 44 normotensive and n ¼ 39 PE women, who presented during the second or third trimesters of their pregnancies, at The Royal Women's Hospital, Melbourne, Australia. Second trimester blood samples were collected prospectively from women during their antenatal checks at 16e21 weeks' gestation, while third trimester blood samples were from women who presented from 26 weeks' gestation. Blood samples were collected into EDTA coated tubes and centrifuged at 2000 g for 10 min at 4 C. Platelet-poor plasma was then
1301
removed, aliquoted and stored at 80 C until use. 2.3. Decidual tissue collection Third trimester decidua basalis samples were collected at delivery from n ¼ 4 normotensive and n ¼ 4 PE women during Caesarean section at The Royal Women's Hospital, Melbourne, Australia. Normotensive patients underwent Caesarean section due to breech presentation, maternal request or previous history of Caesarean section. During the Caesarean section, 10 international units of oxytocin were given intravenously to the mother after the delivery of the baby. The placenta was then removed after spontaneous separation from the uterine wall. A piece of decidual tissue (approximately 1 cm3) was collected from the placental bed and snap frozen in liquid nitrogen. Decidual tissues were then stored at e 80 C until use. All tissue samples were verified as decidual by the Royal Women's Hospital pathologists. 2.4. Protein extraction of decidual tissues Up to 500 mg of tissue was added to 500 ml of radio immunoprecipitation assay buffer [158 mM NaCl, 1 mM EDTA, 10 mM Tris.HCl, 1% (v/v) Triton X-100, 1% (w/v) sodium deoxycholate, 0.1% (w/v) sodium dodecyl sulphate, pH 7.2] containing Halt™ protease and phosphatase inhibitors (Thermo Fisher Scientific, Waltham, MA, USA). Each sample was snap frozen in liquid nitrogen and then rapidly thawed in a 37 C water bath. This freeze/thaw process was performed two more times. Samples were then centrifuged at 15,900 g for 15 min at 4 C. The supernatant of each sample was collected, aliquoted and stored at 30 C until use. Protein concentrations of tissue and cell lysates were determined by Coomassie Plus (Bradford) Protein Assay (Thermo Fisher Scientific). 2.5. Western immunoblotting Plasma samples (0.5 ml) or decidual tissue lysates (25 mg) were diluted in Tris-buffered saline (TBS, 20 mM Tris, 150 mM NaCl, pH 7.6) to an equivalent volume and mixed with 2 Laemmli's Sample Buffer (Bio-Rad, Hercules, CA, USA). For SDS-PAGE, samples were heated at 95 C for 5 min and loaded onto Any kD™ Criterion TGX Gels (Bio-Rad). Proteins were then transferred to a polyvinylidene difluoride (PVDF) membrane using the Trans-Blot® SD Semi-Dry Transfer Cell system (Bio-Rad). To minimise batch effects, all samples from the same trimester were electrophoresed concurrently on multiple gels and transferred onto the same PVDF membrane for analysis. The PVDF membrane was briefly washed in TBS, blocked with 5% skim milk powder diluted in TBS (w/v) for 2 h at room temperature and then incubated with primary antibody overnight at 4 C. Rat anti-human alpha1 (IV) NC1 monoclonal IgG 7070 and rat anti-human alpha2 (IV) NC1 monoclonal IgG 7071 antibodies (Chondrex, Redmond, WA, USA) were used to detect arresten and canstatin respectively. Both primary antibodies were diluted to 2 mg/ml in 0.5% normal goat serum in TBS (v/v). GAPDH was used as the housekeeping protein for decidual tissues and detected with rabbit anti-GAPDH polyclonal IgG IMG5143A-050 antibody (Imgenex, San Diego, CA, USA) diluted to 0.5 mg/ml in 5% skim milk powder in TBS (w/v). All subsequent washes were performed in TBS-T [TBS with 0.1% (v/v) Tween 20]. PVDF membranes were subjected to 3 5 min TBS-T washes and then incubated with goat anti-rat horseradish peroxidaseconjugated IgG HAF005 (R&D Systems, Minneapolis, MN, USA, 1:1000 dilution of stock in TBS) or goat anti-rabbit horseradish peroxidase-conjugated IgG sc-2004 (Santa Cruz Biotechnology Inc., Dallas, TX, USA, diluted to 80 ng/ml in TBS) for 1 h at room temperature. Following further washes, the PVDF membrane was
1302
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
incubated with Clarity Western ECL Substrate (Bio-Rad). The ImageQuant™ LAS4000 imager (GE Healthcare, Buckinghamshire, United Kingdom) was used to detect and image immunoreactive, luminescent bands. ImageJ software version 1.44 [23] was used to quantitate the densitometric intensity of immunoreactive bands. As no suitable, validated loading control exists for western immunoblotting of plasma obtained from pregnant women, plasma comparisons were made based on densitometric intensities alone. Decidual levels of arresten and canstatin were normalised to GAPDH to correct for protein loading. 2.6. Peptide competition The specificity of the arresten and canstatin antibodies was tested by peptide competition experiments performed using synthetic peptides (Mimotopes, Victoria, Australia), which were identical in sequence to the immunising peptides used to generate the antibodies. The peptide sequences for arresten and canstatin were KKPTPSTL and DTLKAGLIR respectively. For peptide competition experiments, primary antibodies were pre-incubated with arresten and canstatin peptides at a concentration of 20 mg/ml for 48 h at 4 C before application on the PVDF membrane.
were performed on GraphPad Prism 5.0 (GraphPad Software Inc., La Jolla, CA, USA). A value of p < 0.05 was considered statistically significant. 3. Results 3.1. Patient characteristics Analysis of the n ¼ 30 normotensive and n ¼ 14 PE patients in the second trimester cohort at delivery showed significant differences in gestational age after delivery and infant birth weight, but no significant difference in maternal age, gestational age at sampling, infant birth weight percentiles, gravidity, parity and infant sex (Table 1). Analysis for the n ¼ 14 normotensive and n ¼ 25 PE patients in the third trimester cohort after delivery showed significant differences in gestational age at delivery, infant birth weight and infant birth weight percentiles, but no significant difference in maternal age, gestational age at sampling, gravidity, parity and infant sex (Table 2). No significant difference except for maternal age was observed in the parameters tested between the n ¼ 4 normotensive and n ¼ 4 PE patients involved in the decidual study (Table 3). 3.2. Peptide competition
2.7. Clinical characteristics and severity analyses Correlations between collagen fragment levels and clinical parameters were assessed. The parameters studied were maternal age, gestation at sampling, gestation at delivery of fetus, infant birth weight percentile, gravidity, parity, highest recorded systolic and diastolic blood pressures, highest recorded degree of proteinuria, lowest recorded platelet level, required treatment with magnesium sulphate, total number of antihypertensive medications prescribed and gestation at onset of PE. 2.8. Statistical analyses Unpaired Student's t test with Welch's Correction, One-Way Analysis of Variance (ANOVA) with Bonferonni's post-test and 2 2 contingency table with Fisher's Exact Test were used for analysing patient characteristics and fragment levels where appropriate. Linear regression was used to determine associations of fragment levels with gestation. Pearson's correlation was used to determine associations between data sets. All statistical analyses
The specificity of the arresten and canstatin antibodies was tested by peptide competition. The arresten antibody was immunoreactive with three distinct bands at approximately 48 kDa, 85 kDa and 150 kDa (Fig. 1A, e peptide). Arresten peptide addition competed out the smallest band at 48 kDa (Fig.1A, þ peptide), which corresponds to the dimer form of arresten [24]. The higher molecular weight bands may correspond to other degradation fragments of collagen. For the canstatin antibody, four distinct immunoreactive bands were detected at approximately 45 kDa, 80 kDa, 100 kDa and 150 kDa (Fig. 1B, e peptide). The lowest molecular weight immunoreactive band at 45 kDa was competed out by canstatin peptide addition (Fig.1B, þ peptide), which corresponds to the dimer form of canstatin [24]. The additional higher molecular weight bands may possibly correspond to other collagen degradation fragments. The monomer forms were rarely detectable and often degraded (data not shown). Therefore, the dimer forms of arresten and canstatin were quantitated in subsequent experiments for the plasma samples. Examples of immunoreactive bands of arresten and canstatin in normotensive and PE plasma are presented in Fig. 1C.
Table 1 Summary of second trimester plasma patient characteristics. Patient characteristicsa
Normotensive (n ¼ 30)
Pre-eclamptic (n ¼ 14)
P-valueb
Maternal age (years) Gestation at sampling (weeks) Gestation at delivery (weeks) Infant weight (g) Infant weight percentilesc Gravidity Parity Infant sexd Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria (dipstick)c MgSO4 treatment Anti-hypertensive treatment Onset of pre-eclampsia (weeks)
30.9 ± 0.8 17.9 ± 0.3 (16.0e20.0) 39.8 ± 0.3 (35.0e42.0) 3309.6 ± 75.9 25e50 26 primi, 4 multi 29 primi, 1 multi 14F, 17M <140 <90 NA NA NA NA
29.4 ± 1.1 18.9 ± 0.5 (16.0e21.0) 36.2 ± 0.8 (28.0e39.0) 2503.7 ± 269.9 25e50 8 primi, 6 multi 11 primi, 3 multi 6F, 9M 162.9 ± 4.4 106.8 ± 2.3 2þ 4 3 35.2 ± 0.9
0.29 0.16 <0.001 <0.05 0.25 0.05 0.09 1.00 NA NA NA NA NA NA
NA, not applicable. a Shown is the mean ± SEM with the entire range of values shown in brackets unless stated otherwise. b Student's t test with Welch's Correction for parametric data and 2 2 contingency table with Fisher's Exact Test for categorical data were used where appropriate. c Data shown as median. d Twin pregnancies were recorded for n ¼ 1 normotensive and n ¼ 1 PE patients.
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
1303
Table 2 Summary of third trimester plasma patient characteristics. Patient characteristicsa
Normotensive (n ¼ 14)
Pre-eclamptic (n ¼ 25)
P-valueb
Maternal age (years) Gestation at sampling (weeks) Gestation at delivery (weeks) Infant weight (g) Infant weight percentilesc Gravidity Parity Infant sex Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria (dipstick)c MgSO4 treatment Anti-hypertensive treatment Onset of pre-eclampsia (weeks)
29.9 ± 1.2 34.3 ± 1.2 (26.0e39.0) 38.5 ± 0.8 (31.0e41.6) 3192.0 ± 183.0 25e50 4 primi, 10 multi 9 primi, 5 multi 7F, 7M <140 <90 NA NA NA NA
29.6 ± 1.2 32.4 ± 0.9 (26.6e39.0) 32.6 ± 0.9 (26.6e39.0) 1729.3 ± 163.1 10e25 11 primi, 14 multi 16 primi, 9 multi 16F, 9M 172.2 ± 2.9 106.5 ± 1.2 3þ 16 23 31.8 ± 0.9
0.86 0.21 <0.001 <0.001 <0.01 0.50 1.00 0.50 NA NA NA NA NA NA
NA, not applicable. a Shown is the mean ± SEM with the entire range of values shown in brackets unless stated otherwise. b Student's t test with Welch's Correction for parametric data and 2 2 contingency table with Fisher's Exact Test for categorical data were used where appropriate. c Data shown as median.
Table 3 Summary of third trimester decidua patient characteristics. Patient characteristicsa
Normotensive (n ¼ 4)
Pre-eclamptic (n ¼ 4)
P-valueb
Maternal age (years) Gestation at delivery (weeks) Infant weight (g) Infant weight percentilesc Gravidity Parity Infant sex Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria (dipstick)c MgSO4 treatment Anti-hypertensive treatment Onset of pre-eclampsia (weeks)
33.0 ± 2.7 36.1 ± 1.8 (31.0e39.0) 2171.3 ± 439.7 25e50 0 primi, 4 multi 1 primi, 3 multi 2F, 2M <140 <90 NA NA NA NA
23.3 ± 1.9 34.1 ± 0.8 (32.6e35.9) 1978.5 ± 230.3 10e25 1 primi, 3 multi 4 primi, 0 multi 3F, 1M 168.8 ± 7.7 107.5 ± 2.5 3þ 3 4 33.5 ± 0.9
<0.05 0.37 0.21 0.19 1.00 0.14 1.00 NA NA NA NA NA NA
NA, not applicable. a Shown is the mean ± SEM with the entire range of values shown in brackets unless stated otherwise. b Student's t test with Welch's Correction for parametric data and 2 2 contingency table with Fisher's Exact Test for categorical data were used where appropriate. c Data shown as median.
3.3. Second trimester plasma measurements Arresten and canstatin levels were measured in n ¼ 30 normotensive and n ¼ 14 PE s trimester plasma samples. In the second trimester, arresten levels were significantly increased in PE compared with normotensive plasma by 2.3-fold (Fig. 2A). There was no significant change in arresten levels in plasma across gestation for either normotensive or PE women (Fig. 2B). Arresten levels in the second trimester did not correlate with any clinical parameter. Canstatin levels showed no overall difference in the second trimester (Fig. 2C). Both normotensive and PE patients showed significant declines in canstatin levels as the second trimester progressed with non-zero R2 values of 0.72 and 0.37 respectively (Fig. 2D). There was also no significant difference in canstatin levels detected between normotensive and PE samples when comparisons were made using 20 weeks' gestation as a cut-off, which is the common timepoint by which most women present for antenatal care (One-Way ANOVA with Bonferroni's post tests, p > 0.05). Canstatin levels in the second trimester did not correlate with any clinical parameter (Pearson's correlation, p > 0.05).
normotensive and n ¼ 25 PE third trimester plasma samples. Arresten levels were significantly increased in the third trimester by 1.4 fold (Fig. 3A). There was no significant change in arresten levels across gestation for either normotensive or PE women (Fig. 3B). Increased arresten levels in all PE patients were significantly associated with the requirement of magnesium sulphate treatment of r ¼ 0.47 (Pearson's correlation, p < 0.05). A significant increase of 1.7 fold was observed in PE patients treated with magnesium sulphate compared with PE patients who were not treated with magnesium sulphate (Fig. 3C). No significant difference in canstatin levels was detected between normotensive and PE plasma in the third trimester (Fig. 3D). No trend across gestation was observed for canstatin levels in normotensive or PE women (Fig. 3E). There was also no difference in canstatin levels between normotensive and PE samples when data were stratified using the early/late onset PE cut-off of 34 weeks' gestation (One-Way ANOVA with Bonferroni's post tests, p > 0.05). Canstatin levels did not correlate with any clinical parameter (Pearson's correlation, p > 0.05).
3.5. Decidual tissue pilot study 3.4. Third trimester plasma measurements Arresten and canstatin levels were measured in n ¼ 14
A pilot study was conducted on n ¼ 4 normotensive and n ¼ 4 PE third trimester decidual tissues to determine if the decidua was a
1304
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
Fig. 1. Representative immunoblots of plasma samples. The corresponding positions of the 150, 100, 75, 50 and 37 kDa molecular weight marker bands are shown. The lanes marked ‘peptide e’ had no competitor peptides added to the antibodies prior to incubation with the PVDF membranes, while the lanes marked ‘peptide þ’ had competitor peptides added to the antibodies. The black arrows denote the dimer forms of arresten and canstatin, while the grey arrows denote the other immunoreactive bands detected by the antibodies. A: Peptide competition of arresten antibody. B: Peptide competition of canstatin antibody. C: Representative immunoblots of arresten and canstatin in normotensive (N) and preeclamptic (PE) plasma samples.
potential source of the circulating collagen fragments. Both the dimer and monomer forms (approximately 45 kDa and 25 kDa respectively) were detectable in the decidua (Fig. 4A and B). There was no significant difference of the dimer levels of arresten and canstatin (Fig. 4C and D). The arresten monomer was significantly increased in PE decidua by 4.2 fold compared with the normotensive decidua (Fig. 4E). No change was observed between the normotensive and PE samples for the canstatin monomer (Fig. 4F). 4. Discussion This is the first study to examine the levels of anti-angiogenic type IV collagen NC1 fragments arresten and canstatin in the plasma and decidua of normotensive and PE pregnant women. We demonstrated significant increases of the arresten fragment in plasma during the second trimester, several weeks before the onset of clinical disease and in the third trimester after the onset of clinical disease. The increase in arresten during the third trimester correlated with the requirement for magnesium sulphate treatment. Decidual levels of arresten were also increased in the third trimester. Arresten levels were significantly increased in the plasma of PE women from as early as 16 weeks' gestation, many weeks before the onset of clinical disease, and during the third trimester, after the onset of clinical disease. Increased levels of arresten in the third trimester also correlated with the need for magnesium sulphate
treatment, a marker of clinical severity. Thus, arresten levels may be a potential tool to monitor the severity of PE once the clinical signs of the disease manifest. These increased levels of arresten in PE are consistent with our previous decidual mRNA expression study where we found increased expression of COL4A1 in PE [10]. Therefore, given the increase in levels of arresten before the onset of clinical disease, arresten may potentially serve as a novel predictive biomarker alongside a panel of current PE biomarkers to improve test sensitivity. In this study, semi-quantative Western immunoblotting was used to measure the levels of arresten and canstatin as degradation of type IV collagen produces fragments of different sizes [25]. Current antibodies available are unable to discriminate between the complete type IV collagen a chains and their respective NC1 fragments. The antibodies used in this study also detected bands at approximately 85 kDa and 80 kDa for arresten and canstatin respectively. The intensities of these bands were also reduced when competitive peptides were added (Fig. 1). Further studies would benefit from the development of specific antibodies that can discriminate the collagen fragments from the complete collagen a chains, which could be used for enzyme-linked immunosorbent assays to test fragment levels in larger cohorts of women and determine their clinical usefulness as biomarkers. A major strength of this study is the use of gestational age matched controls. Gestational age, in addition to PE, can also significantly influence the expression of proteins at the
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
1305
Fig. 2. Plasma levels of arresten and canstatin during the second trimester from n ¼ 30 normotensive and n ¼ 14 PE patients. A: Densitometric analysis of plasma arresten levels during the second trimester. B: Plasma arresten levels across gestation at sampling. C: Densitometric analysis of plasma canstatin levels during the second trimester. D: Plasma canstatin levels across gestation at sampling. The dotted and straight lines represent the trendlines for normotensive and PE patients respectively. Data presented are the mean ± SEM. *p < 0.05, **p < 0.01, Student's t test with Welch's Correction.
maternalefetal interface during pregnancy [26]. To match for gestational age at sampling, the study included a few plasma samples (2 in each cohort), which were obtained from women who subsequently had preterm deliveries due to spontaneous preterm labour. As the severity of PE often necessitates premature delivery, normotensive preterm samples are used in similar studies to control for gestational age. Nevertheless, these preterm delivery samples represent less than 15% of each normotensive cohort in this study, and the majority of the normotensive patients sampled before delivery, delivered at term. Additionally, the collagen genes were not reported to change with gestation at the maternalefetal interface [26], and our earlier study of collagen expression in the third trimester PE decidua suggested that increased levels of collagen were due to the PE condition rather than gestational age [10]. Therefore, it is unlikely that preterm delivery significantly affects the expression of these collagen fragments. Type IV collagen is expressed in the extracellular matrix at the maternalefetal interface [10,27,28]. The NC1 domains of type IV collagen a chains 1 and 2 from which arresten and canstatin are derived, are also present in this interface [28]. In our study, we detected the fragments arresten and canstatin in the decidua, and show significantly increased expression of the arresten monomer in PE. The mechanisms of arresten and canstatin generation are yet to be determined [29]. However, a recent study demonstrated that metalloproteinases are involved in the proteolytic process to degrade collagen into its fragments, as shown by a strong reduction
in arresten production by adenocarcinoma tumour cells, when a non-specific metalloproteinase inhibitor is used [30]. Metalloproteinases are involved in the remodelling of the spiral arterioles at the maternalefetal interface, which is completed during the second trimester [31]. The decrease in canstatin levels during the second trimester may be a reflection of this process. Hence, increased pre-eclamptic decidual collagen expression [10] coupled with expression of metalloproteinases at the maternalefetal interface may result in increased production of these fragments. Therefore, the pre-eclamptic decidua may be a source of the increased circulating levels of fragments observed. The increase in circulating levels of both collagen fragments could damage the maternal endothelium and contribute to the endothelial dysfunction observed in PE. As the maternalefetal interface is also dysfunctional in other placental pathologies such as intrauterine growth restriction and placenta accreta [32], it would be informative in future studies to determine if the levels of arresten and canstatin are also altered in these pathologies. The functional roles of arresten and canstatin in pregnancy are currently unknown. The anti-angiogenic effects of both arresten and canstatin are well-characterised in several studies [18,19,33,34]. Functional studies of arresten and canstatin are focused on their potential as anti-cancer therapeutics targeting tumour angiogenesis [20]. Arresten binds to the a1b1 integrins that are present on both macrovascular and microvascular endothelial cells to induce its anti-angiogenic effects [33,35]. Arresten induces
1306
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
Fig. 3. Plasma levels of arresten during the third trimester from n ¼ 14 normotensive and n ¼ 25 PE patients. A: Densitometric analysis of plasma arresten levels during the third trimester. B: Plasma arresten levels across gestation at sampling. C: Densitometric analysis of plasma arresten levels stratified by the requirement of magnesium sulphate treatment. D: Densitometric analysis of plasma canstatin levels during the third trimester. E: Plasma canstatin levels across gestation at sampling. The dotted and straight lines represent the trendlines for normotensive and PE patients respectively. Data presented are the mean ± SEM. *p < 0.05, **p < 0.01, Student's t test with Welch's Correction.
apoptosis of endothelial cells by decreasing levels of anti-apoptotic Bcl-family molecules such as Bcl-2 [33]. Interestingly, Bcl-2 is decreased in the serum of women with PE [36]. Additionally, arresten is able to block the activation of MMP2, a key enzyme at the maternalefetal interface, by forming a stable binding complex with MMP2 and increasing expression of MMP2 inhibitors such as TIMP2 by endothelial cells [37]. Other studies have also explored the function of arresten in non-endothelial and cancer cell types. Arresten is able to bind renal cell types through the avb3 and a3b1 integrins and may affect kidney function [38], which is impaired in PE. Another study examined the effect of arresten on primary
vascular smooth muscle cells (VSMC) by transfecting the cells with the arresten gene, which reduced VSMC proliferation [39]. Arresten may thus be involved in regulating the function of VSMC, which play a major role at the maternalefetal interface during the crucial spiral arteriole remodelling process in pregnancy [31]. Therefore, arresten may potentially play a role in the pathogenesis of PE through these cellular interactions. Although there were no significant overall changes in canstatin levels between normotensive and PE patients observed in this study, several PE patients had excess plasma canstatin levels of two to three-folds above that of the normotensive patients in the third
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
1307
Fig. 4. Expression of collagen fragments in the decidua from n ¼ 4 normotensive (N) and n ¼ 4 pre-eclamptic (PE) patients. The corresponding positions of the 50, 37and 25 kDa molecular weight marker bands are depicted on the immunoblots. The black arrows denote the dimer and monomer forms of arresten and canstatin. A: Representative immunoblot of arresten with GAPDH loading control in N and PE decidua. B: Representative immunoblot of canstatin with GAPDH loading control in N and PE decidua. C: Densitometric analysis of decidua arresten dimers. D: Densitometric analysis of decidua canstatin dimers. E: Densitometric analysis of decidua arresten monomers. F: Densitometric analysis of decidua arresten monomers. Data presented are the mean ± SEM. *p < 0.05, Student's t test with Welch's Correction.
trimester. Additionally, while plasma canstatin levels significantly declined over the second trimester in both normotensive and subsequent PE pregnancies (Fig. 2D), the decline in women who subsequently developed PE was not as steep as that in women who remained normotensive. Therefore, given the anti-angiogenic functions of canstatin, it is too early to dismiss a role for canstatin in the pathogenesis of PE for a subset of PE women. The N-terminus of canstatin primarily induces endothelial cell apoptosis, while the C-terminus primarily inhibits proliferation in vitro [40,41]. Canstatin binds the avb3 and avb5 integrins in endothelial cells and
triggers the mitochondrial caspase-9 apoptotic mechanism [42]. Canstatin also inhibits proliferation, tube formation and migration of stimulated endothelial cells by reducing Tie2 and VEGFR3 expression [43]. In PE, serum levels of soluble Tie2 are reduced, while placental VEGFR3 expression is decreased [44e46]. Additionally, canstatin inhibits Akt activation and induces Fasdependent apoptosis in endothelial cells [34]. Decreased endothelial Akt activity increases the release of soluble endoglin, which is significantly elevated in the circulation of PE women and contributes to the observed endothelial dysfunction through increased
1308
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
endothelial permeability and decreased endothelial tube formation [47,48]. Fas ligand is also significantly increased in the circulation of PE women [49]. Hence, increased circulating levels of canstatin in a subset of PE women may contribute to the endothelial dysfunction of PE through these pathways. In conclusion, the anti-angiogenic type IV collagen fragment arresten, which is derived from the product of maternal PE susceptibility gene COL4A1, was significantly increased in the plasma of PE women in the early second trimester before the onset of PE and persisted into the third trimester. Additionally, increased third trimester arresten levels correlated with greater clinical severity. Arresten may thus be potentially useful in predicting and monitoring the severity of PE. The decidua may be a potential source of these fragments circulating in the maternal blood. Further functional studies of arresten and canstatin are warranted to determine their roles at the maternalefetal interface and how they may contribute to the development of PE. Acknowledgements We thank all the patients and clinical research midwives who were involved with sample donation and collection. This study was supported by funding from The Royal Women's Hospital, Melbourne, Australia. H.E.J. Yong was supported by The University of Melbourne's Melbourne International Fee Remission Scholarship and the Felix Meyer Scholarship in Obstetrics and Gynaecology. References [1] S.A. Lowe, L. Bowyer, K. Lust, L.P. McMahon, M.R. Morton, R.A. North, M.J. Paech, J.M. Said, The SOMANZ guidelines for the management of hypertensive disorders of pregnancy 2014, Aust. N. Z. J. Obstet. Gynaecol. 55 (1) (2015) 11e16. [2] A.L. Tranquilli, G. Dekker, L. Magee, J. Roberts, B.M. Sibai, W. Steyn, G.G. Zeeman, M.A. Brown, The classification, diagnosis and management of the hypertensive disorders of pregnancy: a revised statement from the ISSHP, Pregnancy Hypertens. An Int. J. Women's Cardiovasc. Health 4 (2) (2014) 97e104. [3] L. Duley, The global impact of pre-eclampsia and eclampsia, Semin. Perinatol. 33 (3) (2009) 130e137. [4] E.A. Steegers, P. von Dadelszen, J.J. Duvekot, R. Pijnenborg, Pre-eclampsia, Lancet 376 (9741) (2010) 631e644. [5] M.D. Lindheimer, S.J. Taler, F.G. Cunningham, Hypertension in pregnancy, J. Am. Soc. Hypertens. 2 (6) (2008) 484e494. [6] S.A. Lowe, M.A. Brown, G.A. Dekker, S. Gatt, C.K. McLintock, L.P. McMahon, G. Mangos, M.P. Moore, P. Muller, M. Paech, B. Walters, Guidelines for the management of hypertensive disorders of pregnancy 2008, Aust. N. Z. J. Obstet. Gynaecol. 49 (3) (2009) 242e246. [7] P.J. Meis, R.L. Goldenberg, B.M. Mercer, J.D. Iams, A.H. Moawad, M. Miodovnik, M.K. Menard, S.N. Caritis, G.R. Thurnau, S.F. Bottoms, A. Das, J.M. Roberts, D. McNellis, The preterm prediction study: risk factors for indicated preterm births. Maternal-Fetal Medicine Units Network of the National Institute of Child Health and Human Development, Am. J. Obstet. Gynecol. 178 (3) (1998) 562e567. [8] S. Saigal, L.W. Doyle, An overview of mortality and sequelae of preterm birth from infancy to adulthood, Lancet 371 (9608) (2008) 261e269. [9] M.P. Johnson, E. Fitzpatrick, T.D. Dyer, J.B. Jowett, S.P. Brennecke, J. Blangero, E.K. Moses, Identification of two novel quantitative trait loci for pre-eclampsia susceptibility on chromosomes 5q and 13q using a variance componentsbased linkage approach, Mol. Hum. Reprod. 13 (1) (2007) 61e67. [10] H.E. Yong, P. Murthi, A. Borg, B. Kalionis, E.K. Moses, S.P. Brennecke, R.J. Keogh, Increased decidual mRNA expression levels of candidate maternal preeclampsia susceptibility genes are associated with clinical severity, Placenta 35 (2) (2014) 117e124. [11] K. Kuhn, Basement membrane (type IV) collagen, Matrix Biol. 14 (6) (1995) 439e445. [12] N. Furuhashi, H. Kimura, H. Nagae, A. Yajima, C. Kimura, T. Saito, Serum collagen IV and laminin levels in preeclampsia, Gynecol. Obstet. Invest. 37 (4) (1994) 250e253. [13] S.S. Veidal, M.A. Karsdal, A. Nawrocki, M.R. Larsen, Y. Dai, Q. Zheng, P. Hagglund, B. Vainer, H. Skjot-Arkil, D.J. Leeming, Assessment of proteolytic degradation of the basement membrane: a fragment of type IV collagen as a biochemical marker for liver fibrosis, Fibrogenes. Tissue Repair 4 (2011) 22. [14] C. Bieglmayer, G. Hofer, Radioimmunoassay for immunoreactive noncollagenous domain of type IV collagen (NC1) in serum: normal pregnancy and preeclampsia, J. Clin. Chem. Clin. Biochem. 27 (3) (1989) 163e167.
[15] F. Keller, Y. Lyreal Ser, D. Schuppan, Raised concentrations of the carboxy terminal propeptide of type IV (basement membrane) procollagen (NC1) in serum and urine of patients with glomerulonephritis, Eur. J. Clin. Invest. 22 (3) (1992) 175e181. [16] D. Schuppan, M. Besser, R. Schwarting, E.G. Hahn, Radioimmunoassay for the carboxy-terminal cross-linking domain of type IV (basement membrane) procollagen in body fluids. Characterization and application to collagen type IV metabolism in fibrotic liver disease, J. Clin. Invest. 78 (1) (1986) 241e248. [17] E. Petitclerc, A. Boutaud, A. Prestayko, J. Xu, Y. Sado, Y. Ninomiya, M.P. Sarras Jr., B.G. Hudson, P.C. Brooks, New functions for non-collagenous domains of human collagen type IV. Novel integrin ligands inhibiting angiogenesis and tumor growth in vivo, J. Biol. Chem. 275 (11) (2000) 8051e8061. [18] P.C. Colorado, A. Torre, G. Kamphaus, Y. Maeshima, H. Hopfer, K. Takahashi, R. Volk, E.D. Zamborsky, S. Herman, P.K. Sarkar, M.B. Ericksen, M. Dhanabal, M. Simons, M. Post, D.W. Kufe, R.R. Weichselbaum, V.P. Sukhatme, R. Kalluri, Anti-angiogenic cues from vascular basement membrane collagen, Cancer Res. 60 (9) (2000) 2520e2526. [19] G.D. Kamphaus, P.C. Colorado, D.J. Panka, H. Hopfer, R. Ramchandran, A. Torre, Y. Maeshima, J.W. Mier, V.P. Sukhatme, R. Kalluri, Canstatin, a novel matrixderived inhibitor of angiogenesis and tumor growth, J. Biol. Chem. 275 (2) (2000) 1209e1215. [20] J.C. Monboisse, J.B. Oudart, L. Ramont, S. Brassart-Pasco, F.X. Maquart, Matrikines from basement membrane collagens: a new anti-cancer strategy, Biochim. Biophys. Acta 1840 (8) (2014) 2589e2598. [21] M.A. Brown, E.D.M. Gallery, S.P. Gatt, G. Leslie, J. Robinson, Management of hypertension in pregnancy- Executive summary, Med. J. Aust. 158 (10) (1993) 700e702. [22] M.A. Brown, W.M. Hague, J. Higgins, S. Lowe, L. McCowan, J. Oats, M.J. Peek, J.A. Rowan, B.N. Walters, The detection, investigation and management of hypertension in pregnancy: executive summary, Aust. N. Z. J. Obstet. Gynaecol. 40 (2) (2000) 133e138. [23] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis, Nat. Methods 9 (7) (2012) 671e675. [24] A. Ries, J. Engel, A. Lustig, K. Kuhn, The function of the NC1 domains in type IV collagen, J. Biol. Chem. 270 (40) (1995) 23790e23794. [25] E. Werle, L. Hao, C. Hasslacher, W. Fiehn, Western blotting of NC1 type IV collagen fragments in human plasma, Eur. J. Clin. Invest. 27 (7) (1997) 579e588. [26] I.A. Lian, M. Langaas, E. Moses, A. Johansson, Differential gene expression at the maternal-fetal interface in preeclampsia is influenced by gestational age, PLoS One 8 (7) (2013) e69848. [27] M. Iwahashi, Y. Muragaki, A. Ooshima, M. Yamoto, R. Nakano, Alterations in distribution and composition of the extracellular matrix during decidualization of the human endometrium, J. Reprod. Fertil. 108 (1) (1996) 147e155. [28] C.M. Oefner, A. Sharkey, L. Gardner, H. Critchley, M. Oyen, A. Moffett, Collagen type IV at the fetal-maternal interface, Placenta 36 (1) (2015) 59e68. [29] J.D. Mott, Z. Werb, Regulation of matrix biology by matrix metalloproteinases, Curr. Opin. Cell Biol. 16 (5) (2004) 558e564. [30] S. Assadian, W. El-Assaad, X.Q. Wang, P.O. Gannon, V. Barres, M. Latour, A.M. Mes-Masson, F. Saad, Y. Sado, J. Dostie, J.G. Teodoro, p53 inhibits angiogenesis by inducing the production of Arresten, Cancer Res. 72 (5) (2012) 1270e1279. [31] J.E. Cartwright, R. Fraser, K. Leslie, A.E. Wallace, J.L. James, Remodelling at the maternal-fetal interface: relevance to human pregnancy disorders, Reproduction 140 (6) (2010) 803e813. [32] E.R. Norwitz, Defective implantation and placentation: laying the blueprint for pregnancy complications, Reprod. Biomed. Online 14 (2007) 101e109. Spec No 1. [33] P. Nyberg, L. Xie, H. Sugimoto, P. Colorado, M. Sund, K. Holthaus, A. Sudhakar, T. Salo, R. Kalluri, Characterization of the anti-angiogenic properties of arresten, an alpha1beta1 integrin-dependent collagen-derived tumor suppressor, Exp. Cell Res. 314 (18) (2008) 3292e3305. [34] D.J. Panka, J.W. Mier, Canstatin inhibits Akt activation and induces Fasdependent apoptosis in endothelial cells, J. Biol. Chem. 278 (39) (2003) 37632e37636. [35] A. Sudhakar, P. Nyberg, V.G. Keshamouni, A.P. Mannam, J. Li, H. Sugimoto, D. Cosgrove, R. Kalluri, Human alpha1 type IV collagen NC1 domain exhibits distinct antiangiogenic activity mediated by alpha1beta1 integrin, J. Clin. Invest. 115 (10) (2005) 2801e2810. [36] F. Varol, R. Uzunoglu, H. Erbas, N. Sut, C. Sayin, VEGFR-1, Bcl-2, and HO-1 ratios in pregnant women with hypertension, Clin. Appl. Thromb. Hemost. 21 (3) (2015) 285e288. [37] Y.A. Sudhakar, R.K. Verma, S.C. Pawar, Type IV collagen alpha1-chain noncollagenous domain blocks MMP-2 activation both in-vitro and in-vivo, Sci. Rep. 4 (2014) 4136. [38] A.S. Aggeli, P.V. Kitsiou, A.K. Tzinia, A. Boutaud, B.G. Hudson, E.C. Tsilibary, Selective binding of integrins from different renal cell types to the NC1 domain of alpha3 and alpha1 chains of type IV collagen, J. Nephrol. 22 (1) (2009) 130e136. [39] D. Shang, Q. Zheng, Z. Song, Y. Li, X. Wang, X. Guo, Eukaryotic expression of human arresten gene and its effect on the proliferation of vascular smooth muscle cells, J. Huazhong Univ. Sci. Technol. Med. Sci. 26 (2) (2006) 202e205. [40] G.A. He, J.X. Luo, T.Y. Zhang, Z.S. Hu, F.Y. Wang, The C-terminal domain of canstatin suppresses in vivo tumor growth associated with proliferation of endothelial cells, Biochem. Biophys. Res. Commun. 318 (2) (2004) 354e360.
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309 [41] G.A. He, J.X. Luo, T.Y. Zhang, F.Y. Wang, R.F. Li, Canstatin-N fragment inhibits in vitro endothelial cell proliferation and suppresses in vivo tumor growth, Biochem. Biophys. Res. Commun. 312 (3) (2003) 801e805. [42] C. Magnon, A. Galaup, B. Mullan, V. Rouffiac, C. Bouquet, J.M. Bidart, F. Griscelli, P. Opolon, M. Perricaudet, Canstatin acts on endothelial and tumor cells via mitochondrial damage initiated through interaction with alphavbeta3 and alphavbeta5 integrins, Cancer Res. 65 (10) (2005) 4353e4361. [43] J. Hwang-Bo, K.H. Yoo, J.H. Park, H.S. Jeong, I.S. Chung, Recombinant canstatin inhibits angiopoietin-1-induced angiogenesis and lymphangiogenesis, Int. J. Cancer 131 (2) (2012) 298e309. [44] J.F. Sung, X. Fan, S. Dhal, B.K. Dwyer, A. Jafari, Y.Y. El-Sayed, M.L. Druzin, N.R. Nayak, Decreased circulating soluble Tie2 levels in preeclampsia may result from inhibition of vascular endothelial growth factor (VEGF) signaling, J. Clin. Endocrinol. Metab. 96 (7) (2011) E1148eE1152. [45] M. Marini, D. Vichi, A. Toscano, G.D. Zappoli Thyrion, E. Parretti, G. Mello, G. Gheri, A. Pacini, E. Sgambati, Expression of vascular endothelial growth factor receptor types 1, 2 and 3 in placenta from pregnancies complicated by hypertensive disorders, Reprod. Fertil. Dev. 19 (5) (2007) 641e651.
1309
[46] D.I. Sokolov, A.V. Kolobov, L.V. Pecherina, N.L. Kramareva, E.V. Mozgovaya, I.M. Kvetnoi, S.A. Selkov, Expression of VEGF and VEGF-R3 receptor by placental endothelial cells in health and gestosis, Bull. Exp. Biol. Med. 145 (3) (2008) 348e351. [47] M.J. Cudmore, S. Ahmad, S. Sissaoui, W. Ramma, B. Ma, T. Fujisawa, B. Al-Ani, K. Wang, M. Cai, F. Crispi, P.W. Hewett, E. Gratacos, S. Egginton, A. Ahmed, Loss of Akt activity increases circulating soluble endoglin release in preeclampsia: identification of inter-dependency between Akt-1 and heme oxygenase-1, Eur. Heart J. 33 (9) (2012) 1150e1158. [48] S. Venkatesha, M. Toporsian, C. Lam, J. Hanai, T. Mammoto, Y.M. Kim, Y. Bdolah, K.H. Lim, H.T. Yuan, T.A. Libermann, I.E. Stillman, D. Roberts, P.A. D'Amore, F.H. Epstein, F.W. Sellke, R. Romero, V.P. Sukhatme, M. Letarte, S.A. Karumanchi, Soluble endoglin contributes to the pathogenesis of preeclampsia, Nat. Med. 12 (6) (2006) 642e649. [49] T.B. Kuntz, R.D. Christensen, J. Stegner, P. Duff, J.M. Koenig, Fas and Fas ligand expression in maternal blood and in umbilical cord blood in preeclampsia, Pediatr. Res. 50 (6) (2001) 743e749.
Contents lists available at ScienceDirect
Placenta journal homepage: www.elsevier.com/locate/placenta
Anti-angiogenic collagen fragment arresten is increased from 16 weeks' gestation in pre-eclamptic plasma Hannah E.J. Yong*, Padma Murthi 1, May H. Wong, Bill Kalionis, Shaun P. Brennecke, Rosemary J. Keogh Department of Perinatal Medicine Pregnancy Research Centre and The University of Melbourne Department of Obstetrics and Gynaecology, The Royal Women's Hospital, Parkville 3052, Victoria, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 June 2015 Received in revised form 24 August 2015 Accepted 26 August 2015
Introduction: Arresten and canstatin are endogenous anti-angiogenic factors derived from type IV collagen a-chains COL4A1 and COL4A2 respectively. While their functions are explored in cancer studies, little is known about their role in pregnancy. Pre-eclampsia (PE) is a common, serious hypertensive disorder of pregnancy that is characterised by systemic endothelial dysfunction. COL4A1 and COL4A2 are maternal PE susceptibility genes that have increased mRNA expression in PE decidua. Our study aim was to determine the levels of arresten and canstatin in plasma and decidua from PE and gestational age matched normotensive patients. Methods: Plasma was collected from normotensive (n ¼ 44) and PE (n ¼ 39) women during the second and third trimesters of pregnancy. Third trimester decidua was collected at delivery from normotensive and PE women (n ¼ 4 each). Levels of arresten and canstatin were determined by Western immunoblotting. Results: Arresten levels were significantly increased in second and third trimester PE plasma, and in third trimester PE decidua (p < 0.05). Third trimester PE plasma arresten levels also significantly correlated with the need for MgSO4 treatment, where a 1.7 fold increase was observed in patients requiring MgSO4 treatment (p < 0.05). No significant change in canstatin levels was observed between normotensive and PE patients. Discussion: This is the first study to report significant increases in the levels of collagen fragment arresten in PE plasma and decidua. Given its significant increase before the onset of clinical disease and associations with clinical severity in the third trimester, arresten may be a useful biomarker for predicting PE and monitoring its severity. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Pre-eclampsia Arresten Canstatin Plasma Clinical severity
1. Introduction Pre-eclampsia (PE) is a serious pregnancy complication with no known cause. PE is characterised by new onset hypertension and proteinuria after 20 weeks' gestation in a previously normotensive woman [1,2]. PE affects between 2 and 8% of all pregnancies and is a leading cause of maternal and fetal morbidity and mortality worldwide [3]. The only cure for PE is the removal of the placenta
* Corresponding author. The Royal Women's Hospital, Locked Bag 300, Corner Grattan Street and Flemington Road, Parkville 3052, Victoria, Australia. E-mail address: [email protected] (H.E.J. Yong). 1 Present Address: Department of Medicine, Monash University, Monash Medical Centre, Clayton 3168, Victoria, Australia. http://dx.doi.org/10.1016/j.placenta.2015.08.013 0143-4004/© 2015 Elsevier Ltd. All rights reserved.
[4]. Approximately 5% of women with PE also develop the haemolysis, elevated liver enzymes, and low platelet count (HELLP) syndrome, which can progress rapidly to become life-threatening [5,6]. Another complication that may develop from PE is eclampsia, where the woman suffers from seizures that may lead to death [5]. Depending on the severity of the mother's condition, the fetus may need to be delivered prematurely. Over 40% of medically indicated preterm births are attributed to PE [7]. Preterm births are associated with adverse outcomes including increased risks of neonatal death, poor long term health and delayed mental development outcomes [8]. Our previous genetic linkage studies identified COL4A1 and COL4A2 as maternal PE susceptibility genes [9]. COL4A1 and COL4A2 code for type IV collagen a chain 1 and a chain 2 respectively. The mRNA expression levels of these genes are significantly increased
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
in PE decidua and these increased levels correlate with an earlier onset of PE [10]. The a1 and a2 chains form a triple helix isoform [a1(IV)]2a2(IV), which is expressed in all basement membranes [11]. Type IV collagen forms the scaffolding of basement membranes in tissues, allowing other basement membrane proteins to be incorporated and for cell interaction to occur with the basement membrane [11]. Type IV collagen is detectable in maternal blood and is significantly increased in third trimester PE serum [12]. Serum collagen levels reflect the degradation of the basement membrane and damage to the endothelium [12]. Degradation of type IV collagen generates many protein fragments [13]. A study by Bieglmayer and Hofer [14] showed an increase of type IV collagen non-collagenous (NC1) domain fragments in third trimester PE serum. Circulating NC1 fragment levels are also increased in patients with autoimmune, metastatic tumour, liver and renal disease [15,16]. These studies did not identify which of the six different a chains of type IV collagen was responsible for the observed increases. The six a chains of type IV collagen produce six distinct NC1 fragments, which are derived from the carboxyl terminus of the a chains [17]. NC1 fragments arresten (COL4A1), canstatin (COL4A2) and tumstatin (COL4A3) have anti-angiogenic effects that increase endothelial cell apoptosis and decrease endothelial cell proliferation, migration and tube formation [17e19]. Research on these fragments is currently focused on their potential as anti-cancer therapeutics by targeting tumour angiogenesis [20]. To date, the levels of each NC1 fragment of type IV collagen have not been investigated and their roles in pregnancy are unknown. Our earlier findings showed increased decidual expression of COL4A1 and COL4A2 in PE, which correlated with an earlier onset of PE, a marker of clinical severity [10]. Having shown a link between PE and increased collagen expression, the study aims were to measure and compare the second and third trimesters levels of arresten and canstatin in the plasma of PE women compared with normotensive controls, to correlate these levels with clinical parameters relating to the severity of PE and to determine whether the maternal decidua was a source of these fragments. 2. Methods 2.1. Study criteria PE was defined as pregnancy induced hypertension with a systolic pressure of 140 mmHg and/or a diastolic pressure of 90 mmHg, accompanied by proteinuria of 0.3 g in 24 h or a spot urine dipstick of ‘2þ’ [21,22]. Exclusion criteria were factors known to increase the risk of PE such as systemic lupus erythematosus and pre-existing hypertension. Blood pressures of normotensive patients were less than 140/90 mmHg. A non-treating obstetrician independently verified patient clinical records. Each patient provided written informed consent. Research and ethics approval were obtained from The Royal Women's Hospital Research and Ethics Committees. 2.2. Plasma sample collection Blood samples were collected from n ¼ 44 normotensive and n ¼ 39 PE women, who presented during the second or third trimesters of their pregnancies, at The Royal Women's Hospital, Melbourne, Australia. Second trimester blood samples were collected prospectively from women during their antenatal checks at 16e21 weeks' gestation, while third trimester blood samples were from women who presented from 26 weeks' gestation. Blood samples were collected into EDTA coated tubes and centrifuged at 2000 g for 10 min at 4 C. Platelet-poor plasma was then
1301
removed, aliquoted and stored at 80 C until use. 2.3. Decidual tissue collection Third trimester decidua basalis samples were collected at delivery from n ¼ 4 normotensive and n ¼ 4 PE women during Caesarean section at The Royal Women's Hospital, Melbourne, Australia. Normotensive patients underwent Caesarean section due to breech presentation, maternal request or previous history of Caesarean section. During the Caesarean section, 10 international units of oxytocin were given intravenously to the mother after the delivery of the baby. The placenta was then removed after spontaneous separation from the uterine wall. A piece of decidual tissue (approximately 1 cm3) was collected from the placental bed and snap frozen in liquid nitrogen. Decidual tissues were then stored at e 80 C until use. All tissue samples were verified as decidual by the Royal Women's Hospital pathologists. 2.4. Protein extraction of decidual tissues Up to 500 mg of tissue was added to 500 ml of radio immunoprecipitation assay buffer [158 mM NaCl, 1 mM EDTA, 10 mM Tris.HCl, 1% (v/v) Triton X-100, 1% (w/v) sodium deoxycholate, 0.1% (w/v) sodium dodecyl sulphate, pH 7.2] containing Halt™ protease and phosphatase inhibitors (Thermo Fisher Scientific, Waltham, MA, USA). Each sample was snap frozen in liquid nitrogen and then rapidly thawed in a 37 C water bath. This freeze/thaw process was performed two more times. Samples were then centrifuged at 15,900 g for 15 min at 4 C. The supernatant of each sample was collected, aliquoted and stored at 30 C until use. Protein concentrations of tissue and cell lysates were determined by Coomassie Plus (Bradford) Protein Assay (Thermo Fisher Scientific). 2.5. Western immunoblotting Plasma samples (0.5 ml) or decidual tissue lysates (25 mg) were diluted in Tris-buffered saline (TBS, 20 mM Tris, 150 mM NaCl, pH 7.6) to an equivalent volume and mixed with 2 Laemmli's Sample Buffer (Bio-Rad, Hercules, CA, USA). For SDS-PAGE, samples were heated at 95 C for 5 min and loaded onto Any kD™ Criterion TGX Gels (Bio-Rad). Proteins were then transferred to a polyvinylidene difluoride (PVDF) membrane using the Trans-Blot® SD Semi-Dry Transfer Cell system (Bio-Rad). To minimise batch effects, all samples from the same trimester were electrophoresed concurrently on multiple gels and transferred onto the same PVDF membrane for analysis. The PVDF membrane was briefly washed in TBS, blocked with 5% skim milk powder diluted in TBS (w/v) for 2 h at room temperature and then incubated with primary antibody overnight at 4 C. Rat anti-human alpha1 (IV) NC1 monoclonal IgG 7070 and rat anti-human alpha2 (IV) NC1 monoclonal IgG 7071 antibodies (Chondrex, Redmond, WA, USA) were used to detect arresten and canstatin respectively. Both primary antibodies were diluted to 2 mg/ml in 0.5% normal goat serum in TBS (v/v). GAPDH was used as the housekeeping protein for decidual tissues and detected with rabbit anti-GAPDH polyclonal IgG IMG5143A-050 antibody (Imgenex, San Diego, CA, USA) diluted to 0.5 mg/ml in 5% skim milk powder in TBS (w/v). All subsequent washes were performed in TBS-T [TBS with 0.1% (v/v) Tween 20]. PVDF membranes were subjected to 3 5 min TBS-T washes and then incubated with goat anti-rat horseradish peroxidaseconjugated IgG HAF005 (R&D Systems, Minneapolis, MN, USA, 1:1000 dilution of stock in TBS) or goat anti-rabbit horseradish peroxidase-conjugated IgG sc-2004 (Santa Cruz Biotechnology Inc., Dallas, TX, USA, diluted to 80 ng/ml in TBS) for 1 h at room temperature. Following further washes, the PVDF membrane was
1302
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
incubated with Clarity Western ECL Substrate (Bio-Rad). The ImageQuant™ LAS4000 imager (GE Healthcare, Buckinghamshire, United Kingdom) was used to detect and image immunoreactive, luminescent bands. ImageJ software version 1.44 [23] was used to quantitate the densitometric intensity of immunoreactive bands. As no suitable, validated loading control exists for western immunoblotting of plasma obtained from pregnant women, plasma comparisons were made based on densitometric intensities alone. Decidual levels of arresten and canstatin were normalised to GAPDH to correct for protein loading. 2.6. Peptide competition The specificity of the arresten and canstatin antibodies was tested by peptide competition experiments performed using synthetic peptides (Mimotopes, Victoria, Australia), which were identical in sequence to the immunising peptides used to generate the antibodies. The peptide sequences for arresten and canstatin were KKPTPSTL and DTLKAGLIR respectively. For peptide competition experiments, primary antibodies were pre-incubated with arresten and canstatin peptides at a concentration of 20 mg/ml for 48 h at 4 C before application on the PVDF membrane.
were performed on GraphPad Prism 5.0 (GraphPad Software Inc., La Jolla, CA, USA). A value of p < 0.05 was considered statistically significant. 3. Results 3.1. Patient characteristics Analysis of the n ¼ 30 normotensive and n ¼ 14 PE patients in the second trimester cohort at delivery showed significant differences in gestational age after delivery and infant birth weight, but no significant difference in maternal age, gestational age at sampling, infant birth weight percentiles, gravidity, parity and infant sex (Table 1). Analysis for the n ¼ 14 normotensive and n ¼ 25 PE patients in the third trimester cohort after delivery showed significant differences in gestational age at delivery, infant birth weight and infant birth weight percentiles, but no significant difference in maternal age, gestational age at sampling, gravidity, parity and infant sex (Table 2). No significant difference except for maternal age was observed in the parameters tested between the n ¼ 4 normotensive and n ¼ 4 PE patients involved in the decidual study (Table 3). 3.2. Peptide competition
2.7. Clinical characteristics and severity analyses Correlations between collagen fragment levels and clinical parameters were assessed. The parameters studied were maternal age, gestation at sampling, gestation at delivery of fetus, infant birth weight percentile, gravidity, parity, highest recorded systolic and diastolic blood pressures, highest recorded degree of proteinuria, lowest recorded platelet level, required treatment with magnesium sulphate, total number of antihypertensive medications prescribed and gestation at onset of PE. 2.8. Statistical analyses Unpaired Student's t test with Welch's Correction, One-Way Analysis of Variance (ANOVA) with Bonferonni's post-test and 2 2 contingency table with Fisher's Exact Test were used for analysing patient characteristics and fragment levels where appropriate. Linear regression was used to determine associations of fragment levels with gestation. Pearson's correlation was used to determine associations between data sets. All statistical analyses
The specificity of the arresten and canstatin antibodies was tested by peptide competition. The arresten antibody was immunoreactive with three distinct bands at approximately 48 kDa, 85 kDa and 150 kDa (Fig. 1A, e peptide). Arresten peptide addition competed out the smallest band at 48 kDa (Fig.1A, þ peptide), which corresponds to the dimer form of arresten [24]. The higher molecular weight bands may correspond to other degradation fragments of collagen. For the canstatin antibody, four distinct immunoreactive bands were detected at approximately 45 kDa, 80 kDa, 100 kDa and 150 kDa (Fig. 1B, e peptide). The lowest molecular weight immunoreactive band at 45 kDa was competed out by canstatin peptide addition (Fig.1B, þ peptide), which corresponds to the dimer form of canstatin [24]. The additional higher molecular weight bands may possibly correspond to other collagen degradation fragments. The monomer forms were rarely detectable and often degraded (data not shown). Therefore, the dimer forms of arresten and canstatin were quantitated in subsequent experiments for the plasma samples. Examples of immunoreactive bands of arresten and canstatin in normotensive and PE plasma are presented in Fig. 1C.
Table 1 Summary of second trimester plasma patient characteristics. Patient characteristicsa
Normotensive (n ¼ 30)
Pre-eclamptic (n ¼ 14)
P-valueb
Maternal age (years) Gestation at sampling (weeks) Gestation at delivery (weeks) Infant weight (g) Infant weight percentilesc Gravidity Parity Infant sexd Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria (dipstick)c MgSO4 treatment Anti-hypertensive treatment Onset of pre-eclampsia (weeks)
30.9 ± 0.8 17.9 ± 0.3 (16.0e20.0) 39.8 ± 0.3 (35.0e42.0) 3309.6 ± 75.9 25e50 26 primi, 4 multi 29 primi, 1 multi 14F, 17M <140 <90 NA NA NA NA
29.4 ± 1.1 18.9 ± 0.5 (16.0e21.0) 36.2 ± 0.8 (28.0e39.0) 2503.7 ± 269.9 25e50 8 primi, 6 multi 11 primi, 3 multi 6F, 9M 162.9 ± 4.4 106.8 ± 2.3 2þ 4 3 35.2 ± 0.9
0.29 0.16 <0.001 <0.05 0.25 0.05 0.09 1.00 NA NA NA NA NA NA
NA, not applicable. a Shown is the mean ± SEM with the entire range of values shown in brackets unless stated otherwise. b Student's t test with Welch's Correction for parametric data and 2 2 contingency table with Fisher's Exact Test for categorical data were used where appropriate. c Data shown as median. d Twin pregnancies were recorded for n ¼ 1 normotensive and n ¼ 1 PE patients.
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
1303
Table 2 Summary of third trimester plasma patient characteristics. Patient characteristicsa
Normotensive (n ¼ 14)
Pre-eclamptic (n ¼ 25)
P-valueb
Maternal age (years) Gestation at sampling (weeks) Gestation at delivery (weeks) Infant weight (g) Infant weight percentilesc Gravidity Parity Infant sex Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria (dipstick)c MgSO4 treatment Anti-hypertensive treatment Onset of pre-eclampsia (weeks)
29.9 ± 1.2 34.3 ± 1.2 (26.0e39.0) 38.5 ± 0.8 (31.0e41.6) 3192.0 ± 183.0 25e50 4 primi, 10 multi 9 primi, 5 multi 7F, 7M <140 <90 NA NA NA NA
29.6 ± 1.2 32.4 ± 0.9 (26.6e39.0) 32.6 ± 0.9 (26.6e39.0) 1729.3 ± 163.1 10e25 11 primi, 14 multi 16 primi, 9 multi 16F, 9M 172.2 ± 2.9 106.5 ± 1.2 3þ 16 23 31.8 ± 0.9
0.86 0.21 <0.001 <0.001 <0.01 0.50 1.00 0.50 NA NA NA NA NA NA
NA, not applicable. a Shown is the mean ± SEM with the entire range of values shown in brackets unless stated otherwise. b Student's t test with Welch's Correction for parametric data and 2 2 contingency table with Fisher's Exact Test for categorical data were used where appropriate. c Data shown as median.
Table 3 Summary of third trimester decidua patient characteristics. Patient characteristicsa
Normotensive (n ¼ 4)
Pre-eclamptic (n ¼ 4)
P-valueb
Maternal age (years) Gestation at delivery (weeks) Infant weight (g) Infant weight percentilesc Gravidity Parity Infant sex Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Proteinuria (dipstick)c MgSO4 treatment Anti-hypertensive treatment Onset of pre-eclampsia (weeks)
33.0 ± 2.7 36.1 ± 1.8 (31.0e39.0) 2171.3 ± 439.7 25e50 0 primi, 4 multi 1 primi, 3 multi 2F, 2M <140 <90 NA NA NA NA
23.3 ± 1.9 34.1 ± 0.8 (32.6e35.9) 1978.5 ± 230.3 10e25 1 primi, 3 multi 4 primi, 0 multi 3F, 1M 168.8 ± 7.7 107.5 ± 2.5 3þ 3 4 33.5 ± 0.9
<0.05 0.37 0.21 0.19 1.00 0.14 1.00 NA NA NA NA NA NA
NA, not applicable. a Shown is the mean ± SEM with the entire range of values shown in brackets unless stated otherwise. b Student's t test with Welch's Correction for parametric data and 2 2 contingency table with Fisher's Exact Test for categorical data were used where appropriate. c Data shown as median.
3.3. Second trimester plasma measurements Arresten and canstatin levels were measured in n ¼ 30 normotensive and n ¼ 14 PE s trimester plasma samples. In the second trimester, arresten levels were significantly increased in PE compared with normotensive plasma by 2.3-fold (Fig. 2A). There was no significant change in arresten levels in plasma across gestation for either normotensive or PE women (Fig. 2B). Arresten levels in the second trimester did not correlate with any clinical parameter. Canstatin levels showed no overall difference in the second trimester (Fig. 2C). Both normotensive and PE patients showed significant declines in canstatin levels as the second trimester progressed with non-zero R2 values of 0.72 and 0.37 respectively (Fig. 2D). There was also no significant difference in canstatin levels detected between normotensive and PE samples when comparisons were made using 20 weeks' gestation as a cut-off, which is the common timepoint by which most women present for antenatal care (One-Way ANOVA with Bonferroni's post tests, p > 0.05). Canstatin levels in the second trimester did not correlate with any clinical parameter (Pearson's correlation, p > 0.05).
normotensive and n ¼ 25 PE third trimester plasma samples. Arresten levels were significantly increased in the third trimester by 1.4 fold (Fig. 3A). There was no significant change in arresten levels across gestation for either normotensive or PE women (Fig. 3B). Increased arresten levels in all PE patients were significantly associated with the requirement of magnesium sulphate treatment of r ¼ 0.47 (Pearson's correlation, p < 0.05). A significant increase of 1.7 fold was observed in PE patients treated with magnesium sulphate compared with PE patients who were not treated with magnesium sulphate (Fig. 3C). No significant difference in canstatin levels was detected between normotensive and PE plasma in the third trimester (Fig. 3D). No trend across gestation was observed for canstatin levels in normotensive or PE women (Fig. 3E). There was also no difference in canstatin levels between normotensive and PE samples when data were stratified using the early/late onset PE cut-off of 34 weeks' gestation (One-Way ANOVA with Bonferroni's post tests, p > 0.05). Canstatin levels did not correlate with any clinical parameter (Pearson's correlation, p > 0.05).
3.5. Decidual tissue pilot study 3.4. Third trimester plasma measurements Arresten and canstatin levels were measured in n ¼ 14
A pilot study was conducted on n ¼ 4 normotensive and n ¼ 4 PE third trimester decidual tissues to determine if the decidua was a
1304
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
Fig. 1. Representative immunoblots of plasma samples. The corresponding positions of the 150, 100, 75, 50 and 37 kDa molecular weight marker bands are shown. The lanes marked ‘peptide e’ had no competitor peptides added to the antibodies prior to incubation with the PVDF membranes, while the lanes marked ‘peptide þ’ had competitor peptides added to the antibodies. The black arrows denote the dimer forms of arresten and canstatin, while the grey arrows denote the other immunoreactive bands detected by the antibodies. A: Peptide competition of arresten antibody. B: Peptide competition of canstatin antibody. C: Representative immunoblots of arresten and canstatin in normotensive (N) and preeclamptic (PE) plasma samples.
potential source of the circulating collagen fragments. Both the dimer and monomer forms (approximately 45 kDa and 25 kDa respectively) were detectable in the decidua (Fig. 4A and B). There was no significant difference of the dimer levels of arresten and canstatin (Fig. 4C and D). The arresten monomer was significantly increased in PE decidua by 4.2 fold compared with the normotensive decidua (Fig. 4E). No change was observed between the normotensive and PE samples for the canstatin monomer (Fig. 4F). 4. Discussion This is the first study to examine the levels of anti-angiogenic type IV collagen NC1 fragments arresten and canstatin in the plasma and decidua of normotensive and PE pregnant women. We demonstrated significant increases of the arresten fragment in plasma during the second trimester, several weeks before the onset of clinical disease and in the third trimester after the onset of clinical disease. The increase in arresten during the third trimester correlated with the requirement for magnesium sulphate treatment. Decidual levels of arresten were also increased in the third trimester. Arresten levels were significantly increased in the plasma of PE women from as early as 16 weeks' gestation, many weeks before the onset of clinical disease, and during the third trimester, after the onset of clinical disease. Increased levels of arresten in the third trimester also correlated with the need for magnesium sulphate
treatment, a marker of clinical severity. Thus, arresten levels may be a potential tool to monitor the severity of PE once the clinical signs of the disease manifest. These increased levels of arresten in PE are consistent with our previous decidual mRNA expression study where we found increased expression of COL4A1 in PE [10]. Therefore, given the increase in levels of arresten before the onset of clinical disease, arresten may potentially serve as a novel predictive biomarker alongside a panel of current PE biomarkers to improve test sensitivity. In this study, semi-quantative Western immunoblotting was used to measure the levels of arresten and canstatin as degradation of type IV collagen produces fragments of different sizes [25]. Current antibodies available are unable to discriminate between the complete type IV collagen a chains and their respective NC1 fragments. The antibodies used in this study also detected bands at approximately 85 kDa and 80 kDa for arresten and canstatin respectively. The intensities of these bands were also reduced when competitive peptides were added (Fig. 1). Further studies would benefit from the development of specific antibodies that can discriminate the collagen fragments from the complete collagen a chains, which could be used for enzyme-linked immunosorbent assays to test fragment levels in larger cohorts of women and determine their clinical usefulness as biomarkers. A major strength of this study is the use of gestational age matched controls. Gestational age, in addition to PE, can also significantly influence the expression of proteins at the
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
1305
Fig. 2. Plasma levels of arresten and canstatin during the second trimester from n ¼ 30 normotensive and n ¼ 14 PE patients. A: Densitometric analysis of plasma arresten levels during the second trimester. B: Plasma arresten levels across gestation at sampling. C: Densitometric analysis of plasma canstatin levels during the second trimester. D: Plasma canstatin levels across gestation at sampling. The dotted and straight lines represent the trendlines for normotensive and PE patients respectively. Data presented are the mean ± SEM. *p < 0.05, **p < 0.01, Student's t test with Welch's Correction.
maternalefetal interface during pregnancy [26]. To match for gestational age at sampling, the study included a few plasma samples (2 in each cohort), which were obtained from women who subsequently had preterm deliveries due to spontaneous preterm labour. As the severity of PE often necessitates premature delivery, normotensive preterm samples are used in similar studies to control for gestational age. Nevertheless, these preterm delivery samples represent less than 15% of each normotensive cohort in this study, and the majority of the normotensive patients sampled before delivery, delivered at term. Additionally, the collagen genes were not reported to change with gestation at the maternalefetal interface [26], and our earlier study of collagen expression in the third trimester PE decidua suggested that increased levels of collagen were due to the PE condition rather than gestational age [10]. Therefore, it is unlikely that preterm delivery significantly affects the expression of these collagen fragments. Type IV collagen is expressed in the extracellular matrix at the maternalefetal interface [10,27,28]. The NC1 domains of type IV collagen a chains 1 and 2 from which arresten and canstatin are derived, are also present in this interface [28]. In our study, we detected the fragments arresten and canstatin in the decidua, and show significantly increased expression of the arresten monomer in PE. The mechanisms of arresten and canstatin generation are yet to be determined [29]. However, a recent study demonstrated that metalloproteinases are involved in the proteolytic process to degrade collagen into its fragments, as shown by a strong reduction
in arresten production by adenocarcinoma tumour cells, when a non-specific metalloproteinase inhibitor is used [30]. Metalloproteinases are involved in the remodelling of the spiral arterioles at the maternalefetal interface, which is completed during the second trimester [31]. The decrease in canstatin levels during the second trimester may be a reflection of this process. Hence, increased pre-eclamptic decidual collagen expression [10] coupled with expression of metalloproteinases at the maternalefetal interface may result in increased production of these fragments. Therefore, the pre-eclamptic decidua may be a source of the increased circulating levels of fragments observed. The increase in circulating levels of both collagen fragments could damage the maternal endothelium and contribute to the endothelial dysfunction observed in PE. As the maternalefetal interface is also dysfunctional in other placental pathologies such as intrauterine growth restriction and placenta accreta [32], it would be informative in future studies to determine if the levels of arresten and canstatin are also altered in these pathologies. The functional roles of arresten and canstatin in pregnancy are currently unknown. The anti-angiogenic effects of both arresten and canstatin are well-characterised in several studies [18,19,33,34]. Functional studies of arresten and canstatin are focused on their potential as anti-cancer therapeutics targeting tumour angiogenesis [20]. Arresten binds to the a1b1 integrins that are present on both macrovascular and microvascular endothelial cells to induce its anti-angiogenic effects [33,35]. Arresten induces
1306
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
Fig. 3. Plasma levels of arresten during the third trimester from n ¼ 14 normotensive and n ¼ 25 PE patients. A: Densitometric analysis of plasma arresten levels during the third trimester. B: Plasma arresten levels across gestation at sampling. C: Densitometric analysis of plasma arresten levels stratified by the requirement of magnesium sulphate treatment. D: Densitometric analysis of plasma canstatin levels during the third trimester. E: Plasma canstatin levels across gestation at sampling. The dotted and straight lines represent the trendlines for normotensive and PE patients respectively. Data presented are the mean ± SEM. *p < 0.05, **p < 0.01, Student's t test with Welch's Correction.
apoptosis of endothelial cells by decreasing levels of anti-apoptotic Bcl-family molecules such as Bcl-2 [33]. Interestingly, Bcl-2 is decreased in the serum of women with PE [36]. Additionally, arresten is able to block the activation of MMP2, a key enzyme at the maternalefetal interface, by forming a stable binding complex with MMP2 and increasing expression of MMP2 inhibitors such as TIMP2 by endothelial cells [37]. Other studies have also explored the function of arresten in non-endothelial and cancer cell types. Arresten is able to bind renal cell types through the avb3 and a3b1 integrins and may affect kidney function [38], which is impaired in PE. Another study examined the effect of arresten on primary
vascular smooth muscle cells (VSMC) by transfecting the cells with the arresten gene, which reduced VSMC proliferation [39]. Arresten may thus be involved in regulating the function of VSMC, which play a major role at the maternalefetal interface during the crucial spiral arteriole remodelling process in pregnancy [31]. Therefore, arresten may potentially play a role in the pathogenesis of PE through these cellular interactions. Although there were no significant overall changes in canstatin levels between normotensive and PE patients observed in this study, several PE patients had excess plasma canstatin levels of two to three-folds above that of the normotensive patients in the third
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
1307
Fig. 4. Expression of collagen fragments in the decidua from n ¼ 4 normotensive (N) and n ¼ 4 pre-eclamptic (PE) patients. The corresponding positions of the 50, 37and 25 kDa molecular weight marker bands are depicted on the immunoblots. The black arrows denote the dimer and monomer forms of arresten and canstatin. A: Representative immunoblot of arresten with GAPDH loading control in N and PE decidua. B: Representative immunoblot of canstatin with GAPDH loading control in N and PE decidua. C: Densitometric analysis of decidua arresten dimers. D: Densitometric analysis of decidua canstatin dimers. E: Densitometric analysis of decidua arresten monomers. F: Densitometric analysis of decidua arresten monomers. Data presented are the mean ± SEM. *p < 0.05, Student's t test with Welch's Correction.
trimester. Additionally, while plasma canstatin levels significantly declined over the second trimester in both normotensive and subsequent PE pregnancies (Fig. 2D), the decline in women who subsequently developed PE was not as steep as that in women who remained normotensive. Therefore, given the anti-angiogenic functions of canstatin, it is too early to dismiss a role for canstatin in the pathogenesis of PE for a subset of PE women. The N-terminus of canstatin primarily induces endothelial cell apoptosis, while the C-terminus primarily inhibits proliferation in vitro [40,41]. Canstatin binds the avb3 and avb5 integrins in endothelial cells and
triggers the mitochondrial caspase-9 apoptotic mechanism [42]. Canstatin also inhibits proliferation, tube formation and migration of stimulated endothelial cells by reducing Tie2 and VEGFR3 expression [43]. In PE, serum levels of soluble Tie2 are reduced, while placental VEGFR3 expression is decreased [44e46]. Additionally, canstatin inhibits Akt activation and induces Fasdependent apoptosis in endothelial cells [34]. Decreased endothelial Akt activity increases the release of soluble endoglin, which is significantly elevated in the circulation of PE women and contributes to the observed endothelial dysfunction through increased
1308
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309
endothelial permeability and decreased endothelial tube formation [47,48]. Fas ligand is also significantly increased in the circulation of PE women [49]. Hence, increased circulating levels of canstatin in a subset of PE women may contribute to the endothelial dysfunction of PE through these pathways. In conclusion, the anti-angiogenic type IV collagen fragment arresten, which is derived from the product of maternal PE susceptibility gene COL4A1, was significantly increased in the plasma of PE women in the early second trimester before the onset of PE and persisted into the third trimester. Additionally, increased third trimester arresten levels correlated with greater clinical severity. Arresten may thus be potentially useful in predicting and monitoring the severity of PE. The decidua may be a potential source of these fragments circulating in the maternal blood. Further functional studies of arresten and canstatin are warranted to determine their roles at the maternalefetal interface and how they may contribute to the development of PE. Acknowledgements We thank all the patients and clinical research midwives who were involved with sample donation and collection. This study was supported by funding from The Royal Women's Hospital, Melbourne, Australia. H.E.J. Yong was supported by The University of Melbourne's Melbourne International Fee Remission Scholarship and the Felix Meyer Scholarship in Obstetrics and Gynaecology. References [1] S.A. Lowe, L. Bowyer, K. Lust, L.P. McMahon, M.R. Morton, R.A. North, M.J. Paech, J.M. Said, The SOMANZ guidelines for the management of hypertensive disorders of pregnancy 2014, Aust. N. Z. J. Obstet. Gynaecol. 55 (1) (2015) 11e16. [2] A.L. Tranquilli, G. Dekker, L. Magee, J. Roberts, B.M. Sibai, W. Steyn, G.G. Zeeman, M.A. Brown, The classification, diagnosis and management of the hypertensive disorders of pregnancy: a revised statement from the ISSHP, Pregnancy Hypertens. An Int. J. Women's Cardiovasc. Health 4 (2) (2014) 97e104. [3] L. Duley, The global impact of pre-eclampsia and eclampsia, Semin. Perinatol. 33 (3) (2009) 130e137. [4] E.A. Steegers, P. von Dadelszen, J.J. Duvekot, R. Pijnenborg, Pre-eclampsia, Lancet 376 (9741) (2010) 631e644. [5] M.D. Lindheimer, S.J. Taler, F.G. Cunningham, Hypertension in pregnancy, J. Am. Soc. Hypertens. 2 (6) (2008) 484e494. [6] S.A. Lowe, M.A. Brown, G.A. Dekker, S. Gatt, C.K. McLintock, L.P. McMahon, G. Mangos, M.P. Moore, P. Muller, M. Paech, B. Walters, Guidelines for the management of hypertensive disorders of pregnancy 2008, Aust. N. Z. J. Obstet. Gynaecol. 49 (3) (2009) 242e246. [7] P.J. Meis, R.L. Goldenberg, B.M. Mercer, J.D. Iams, A.H. Moawad, M. Miodovnik, M.K. Menard, S.N. Caritis, G.R. Thurnau, S.F. Bottoms, A. Das, J.M. Roberts, D. McNellis, The preterm prediction study: risk factors for indicated preterm births. Maternal-Fetal Medicine Units Network of the National Institute of Child Health and Human Development, Am. J. Obstet. Gynecol. 178 (3) (1998) 562e567. [8] S. Saigal, L.W. Doyle, An overview of mortality and sequelae of preterm birth from infancy to adulthood, Lancet 371 (9608) (2008) 261e269. [9] M.P. Johnson, E. Fitzpatrick, T.D. Dyer, J.B. Jowett, S.P. Brennecke, J. Blangero, E.K. Moses, Identification of two novel quantitative trait loci for pre-eclampsia susceptibility on chromosomes 5q and 13q using a variance componentsbased linkage approach, Mol. Hum. Reprod. 13 (1) (2007) 61e67. [10] H.E. Yong, P. Murthi, A. Borg, B. Kalionis, E.K. Moses, S.P. Brennecke, R.J. Keogh, Increased decidual mRNA expression levels of candidate maternal preeclampsia susceptibility genes are associated with clinical severity, Placenta 35 (2) (2014) 117e124. [11] K. Kuhn, Basement membrane (type IV) collagen, Matrix Biol. 14 (6) (1995) 439e445. [12] N. Furuhashi, H. Kimura, H. Nagae, A. Yajima, C. Kimura, T. Saito, Serum collagen IV and laminin levels in preeclampsia, Gynecol. Obstet. Invest. 37 (4) (1994) 250e253. [13] S.S. Veidal, M.A. Karsdal, A. Nawrocki, M.R. Larsen, Y. Dai, Q. Zheng, P. Hagglund, B. Vainer, H. Skjot-Arkil, D.J. Leeming, Assessment of proteolytic degradation of the basement membrane: a fragment of type IV collagen as a biochemical marker for liver fibrosis, Fibrogenes. Tissue Repair 4 (2011) 22. [14] C. Bieglmayer, G. Hofer, Radioimmunoassay for immunoreactive noncollagenous domain of type IV collagen (NC1) in serum: normal pregnancy and preeclampsia, J. Clin. Chem. Clin. Biochem. 27 (3) (1989) 163e167.
[15] F. Keller, Y. Lyreal Ser, D. Schuppan, Raised concentrations of the carboxy terminal propeptide of type IV (basement membrane) procollagen (NC1) in serum and urine of patients with glomerulonephritis, Eur. J. Clin. Invest. 22 (3) (1992) 175e181. [16] D. Schuppan, M. Besser, R. Schwarting, E.G. Hahn, Radioimmunoassay for the carboxy-terminal cross-linking domain of type IV (basement membrane) procollagen in body fluids. Characterization and application to collagen type IV metabolism in fibrotic liver disease, J. Clin. Invest. 78 (1) (1986) 241e248. [17] E. Petitclerc, A. Boutaud, A. Prestayko, J. Xu, Y. Sado, Y. Ninomiya, M.P. Sarras Jr., B.G. Hudson, P.C. Brooks, New functions for non-collagenous domains of human collagen type IV. Novel integrin ligands inhibiting angiogenesis and tumor growth in vivo, J. Biol. Chem. 275 (11) (2000) 8051e8061. [18] P.C. Colorado, A. Torre, G. Kamphaus, Y. Maeshima, H. Hopfer, K. Takahashi, R. Volk, E.D. Zamborsky, S. Herman, P.K. Sarkar, M.B. Ericksen, M. Dhanabal, M. Simons, M. Post, D.W. Kufe, R.R. Weichselbaum, V.P. Sukhatme, R. Kalluri, Anti-angiogenic cues from vascular basement membrane collagen, Cancer Res. 60 (9) (2000) 2520e2526. [19] G.D. Kamphaus, P.C. Colorado, D.J. Panka, H. Hopfer, R. Ramchandran, A. Torre, Y. Maeshima, J.W. Mier, V.P. Sukhatme, R. Kalluri, Canstatin, a novel matrixderived inhibitor of angiogenesis and tumor growth, J. Biol. Chem. 275 (2) (2000) 1209e1215. [20] J.C. Monboisse, J.B. Oudart, L. Ramont, S. Brassart-Pasco, F.X. Maquart, Matrikines from basement membrane collagens: a new anti-cancer strategy, Biochim. Biophys. Acta 1840 (8) (2014) 2589e2598. [21] M.A. Brown, E.D.M. Gallery, S.P. Gatt, G. Leslie, J. Robinson, Management of hypertension in pregnancy- Executive summary, Med. J. Aust. 158 (10) (1993) 700e702. [22] M.A. Brown, W.M. Hague, J. Higgins, S. Lowe, L. McCowan, J. Oats, M.J. Peek, J.A. Rowan, B.N. Walters, The detection, investigation and management of hypertension in pregnancy: executive summary, Aust. N. Z. J. Obstet. Gynaecol. 40 (2) (2000) 133e138. [23] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis, Nat. Methods 9 (7) (2012) 671e675. [24] A. Ries, J. Engel, A. Lustig, K. Kuhn, The function of the NC1 domains in type IV collagen, J. Biol. Chem. 270 (40) (1995) 23790e23794. [25] E. Werle, L. Hao, C. Hasslacher, W. Fiehn, Western blotting of NC1 type IV collagen fragments in human plasma, Eur. J. Clin. Invest. 27 (7) (1997) 579e588. [26] I.A. Lian, M. Langaas, E. Moses, A. Johansson, Differential gene expression at the maternal-fetal interface in preeclampsia is influenced by gestational age, PLoS One 8 (7) (2013) e69848. [27] M. Iwahashi, Y. Muragaki, A. Ooshima, M. Yamoto, R. Nakano, Alterations in distribution and composition of the extracellular matrix during decidualization of the human endometrium, J. Reprod. Fertil. 108 (1) (1996) 147e155. [28] C.M. Oefner, A. Sharkey, L. Gardner, H. Critchley, M. Oyen, A. Moffett, Collagen type IV at the fetal-maternal interface, Placenta 36 (1) (2015) 59e68. [29] J.D. Mott, Z. Werb, Regulation of matrix biology by matrix metalloproteinases, Curr. Opin. Cell Biol. 16 (5) (2004) 558e564. [30] S. Assadian, W. El-Assaad, X.Q. Wang, P.O. Gannon, V. Barres, M. Latour, A.M. Mes-Masson, F. Saad, Y. Sado, J. Dostie, J.G. Teodoro, p53 inhibits angiogenesis by inducing the production of Arresten, Cancer Res. 72 (5) (2012) 1270e1279. [31] J.E. Cartwright, R. Fraser, K. Leslie, A.E. Wallace, J.L. James, Remodelling at the maternal-fetal interface: relevance to human pregnancy disorders, Reproduction 140 (6) (2010) 803e813. [32] E.R. Norwitz, Defective implantation and placentation: laying the blueprint for pregnancy complications, Reprod. Biomed. Online 14 (2007) 101e109. Spec No 1. [33] P. Nyberg, L. Xie, H. Sugimoto, P. Colorado, M. Sund, K. Holthaus, A. Sudhakar, T. Salo, R. Kalluri, Characterization of the anti-angiogenic properties of arresten, an alpha1beta1 integrin-dependent collagen-derived tumor suppressor, Exp. Cell Res. 314 (18) (2008) 3292e3305. [34] D.J. Panka, J.W. Mier, Canstatin inhibits Akt activation and induces Fasdependent apoptosis in endothelial cells, J. Biol. Chem. 278 (39) (2003) 37632e37636. [35] A. Sudhakar, P. Nyberg, V.G. Keshamouni, A.P. Mannam, J. Li, H. Sugimoto, D. Cosgrove, R. Kalluri, Human alpha1 type IV collagen NC1 domain exhibits distinct antiangiogenic activity mediated by alpha1beta1 integrin, J. Clin. Invest. 115 (10) (2005) 2801e2810. [36] F. Varol, R. Uzunoglu, H. Erbas, N. Sut, C. Sayin, VEGFR-1, Bcl-2, and HO-1 ratios in pregnant women with hypertension, Clin. Appl. Thromb. Hemost. 21 (3) (2015) 285e288. [37] Y.A. Sudhakar, R.K. Verma, S.C. Pawar, Type IV collagen alpha1-chain noncollagenous domain blocks MMP-2 activation both in-vitro and in-vivo, Sci. Rep. 4 (2014) 4136. [38] A.S. Aggeli, P.V. Kitsiou, A.K. Tzinia, A. Boutaud, B.G. Hudson, E.C. Tsilibary, Selective binding of integrins from different renal cell types to the NC1 domain of alpha3 and alpha1 chains of type IV collagen, J. Nephrol. 22 (1) (2009) 130e136. [39] D. Shang, Q. Zheng, Z. Song, Y. Li, X. Wang, X. Guo, Eukaryotic expression of human arresten gene and its effect on the proliferation of vascular smooth muscle cells, J. Huazhong Univ. Sci. Technol. Med. Sci. 26 (2) (2006) 202e205. [40] G.A. He, J.X. Luo, T.Y. Zhang, Z.S. Hu, F.Y. Wang, The C-terminal domain of canstatin suppresses in vivo tumor growth associated with proliferation of endothelial cells, Biochem. Biophys. Res. Commun. 318 (2) (2004) 354e360.
H.E.J. Yong et al. / Placenta 36 (2015) 1300e1309 [41] G.A. He, J.X. Luo, T.Y. Zhang, F.Y. Wang, R.F. Li, Canstatin-N fragment inhibits in vitro endothelial cell proliferation and suppresses in vivo tumor growth, Biochem. Biophys. Res. Commun. 312 (3) (2003) 801e805. [42] C. Magnon, A. Galaup, B. Mullan, V. Rouffiac, C. Bouquet, J.M. Bidart, F. Griscelli, P. Opolon, M. Perricaudet, Canstatin acts on endothelial and tumor cells via mitochondrial damage initiated through interaction with alphavbeta3 and alphavbeta5 integrins, Cancer Res. 65 (10) (2005) 4353e4361. [43] J. Hwang-Bo, K.H. Yoo, J.H. Park, H.S. Jeong, I.S. Chung, Recombinant canstatin inhibits angiopoietin-1-induced angiogenesis and lymphangiogenesis, Int. J. Cancer 131 (2) (2012) 298e309. [44] J.F. Sung, X. Fan, S. Dhal, B.K. Dwyer, A. Jafari, Y.Y. El-Sayed, M.L. Druzin, N.R. Nayak, Decreased circulating soluble Tie2 levels in preeclampsia may result from inhibition of vascular endothelial growth factor (VEGF) signaling, J. Clin. Endocrinol. Metab. 96 (7) (2011) E1148eE1152. [45] M. Marini, D. Vichi, A. Toscano, G.D. Zappoli Thyrion, E. Parretti, G. Mello, G. Gheri, A. Pacini, E. Sgambati, Expression of vascular endothelial growth factor receptor types 1, 2 and 3 in placenta from pregnancies complicated by hypertensive disorders, Reprod. Fertil. Dev. 19 (5) (2007) 641e651.
1309
[46] D.I. Sokolov, A.V. Kolobov, L.V. Pecherina, N.L. Kramareva, E.V. Mozgovaya, I.M. Kvetnoi, S.A. Selkov, Expression of VEGF and VEGF-R3 receptor by placental endothelial cells in health and gestosis, Bull. Exp. Biol. Med. 145 (3) (2008) 348e351. [47] M.J. Cudmore, S. Ahmad, S. Sissaoui, W. Ramma, B. Ma, T. Fujisawa, B. Al-Ani, K. Wang, M. Cai, F. Crispi, P.W. Hewett, E. Gratacos, S. Egginton, A. Ahmed, Loss of Akt activity increases circulating soluble endoglin release in preeclampsia: identification of inter-dependency between Akt-1 and heme oxygenase-1, Eur. Heart J. 33 (9) (2012) 1150e1158. [48] S. Venkatesha, M. Toporsian, C. Lam, J. Hanai, T. Mammoto, Y.M. Kim, Y. Bdolah, K.H. Lim, H.T. Yuan, T.A. Libermann, I.E. Stillman, D. Roberts, P.A. D'Amore, F.H. Epstein, F.W. Sellke, R. Romero, V.P. Sukhatme, M. Letarte, S.A. Karumanchi, Soluble endoglin contributes to the pathogenesis of preeclampsia, Nat. Med. 12 (6) (2006) 642e649. [49] T.B. Kuntz, R.D. Christensen, J. Stegner, P. Duff, J.M. Koenig, Fas and Fas ligand expression in maternal blood and in umbilical cord blood in preeclampsia, Pediatr. Res. 50 (6) (2001) 743e749.