Fetal growth restriction (FGR) is a failure to fulfill the genetically predetermined growth potential during the course of pregnancy. The causes of FGR can be broadly subdivided into fetal, maternal, and placental etiologies. Identification of the contributing environmental and genetic factors, combined with an understanding of the pathological changes preceding development of FGR, is essential so that future care can be targeted toward developing potential therapies to alleviate adverse perinatal outcomes. Based on a plethora of studies, placental insufficiency has been consistently shown to be involved in the majority of FGR cases (Simonazzi et al., Reference Simonazzi, Curti, Cattani, Rizzo and Pilu2013). However, in previous studies (Cosmi et al., Reference Cosmi, Visentin, Favretto, Tucci, Ragazzi, Viel and Ferrara2013; McCarthy et al., Reference McCarthy, Cotter, McElwaine, Twomey, Mooney, Ryan and Vaughan2007; Sitras et al., Reference Sitras, Paulssen, Leirvik, Vartun and Acharya2009) where a singleton fetus was used as a model to study the role of placental inefficiency in the pathogenesis of FGR, contributions of maternal, genetic, and metabolic factors were still evident.
Growth discordance is one of the most common complications of monochorionic diamniotic (MCDA) twin pregnancies (Bejar et al., Reference Bejar, Vigliocco, Gramajo, Solana, Benirschke and Resnik1990; Valsky, Eixarch, Martinez, Crispi et al., Reference Valsky, Eixarch, Martinez, Crispi and Gratacos2010), with a prevalence of 10–19% (Acosta-Rojas et al., Reference Acosta-Rojas, Becker, Munoz-Abellana, Ruiz, Carreras and Gratacos2007; Lewi et al., Reference Lewi, Devlieger, De Catte and Deprest2014; Valsky, Eixarch, Martinez, & Gratacós, Reference Valsky, Eixarch, Martinez and Gratacós2010). The mechanisms and progressive development of growth discordance largely depend on the relative effects of unequal placental sharing and inter-twin blood transfer through placental anastomoses (Chang et al., Reference Chang, Chang, Chao, Hsieh, Wang and Wang2009; Hack et al., Reference Hack, Nikkels, Koopman-Esseboom, Derks, Elias, van, Gemert and Visser2008; Lewi et al., Reference Lewi, Cannie, Blickstein, Jani, Huber, Hecher and Deprest2007). MCDA twins share the same genetic background and maternal host environment. Thus, the differences in intrauterine growth of MCDA twins may be due to differential regulation of genes at the transcriptional and/or post-transcriptional level that are involved in placental development and function. On this basis, growth-discordant MCDA twins represent ideal models to study the underlying mechanisms of placental insufficiency independent of confounding genetic factors and maternal host environment.
Previous twin studies have demonstrated that phenotypic placental variation is associated with epigenetic factors (Bell & Saffery, Reference Bell and Saffery2012; Gordon et al., Reference Gordon, Joo, Andronikos, Ollikainen, Wallace, Umstad and Craig2011; Kaminsky et al., Reference Kaminsky, Tang, Wang, Ptak, Oh, Wong and Petronis2009; Yu et al., Reference Yu, Furukawa, Kobayashi, Shikishima, Cha, Sese and Toda2012). In a recent study, that specifically examined placental mRNA expression of angiogenesis-related genes, significant up-regulation in the placental share of the growth-discordant MCDA twin for leptin (LEP) and several other angiogenic factors were reported (Schrey et al., Reference Schrey, Kingdom, Baczyk, Fitzgerald, Keating, Ryan and Drewlo2013). In other studies, differential expression of osteopontin (Li et al., Reference Li, Fang, He, Gao and Zhou2013) and the human endogenous family W, Env(C7), member 1 protein (Gao et al., Reference Gao, He, Wang, Luo, Sun, Zhou and Jiang2012) were both found to be strongly associated with growth discordance. While there have been several studies focused on global gene expression in growth-discordant MCDA twin placentas (Gordon et al., Reference Gordon, Joo, Andronikos, Ollikainen, Wallace, Umstad and Craig2011; Nishizawa et al., Reference Nishizawa, Ota, Suzuki, Kato, Sekiya, Kurahashi and Udagawa2011; Sitras et al., Reference Sitras, Paulssen, Leirvik, Vartun and Acharya2009), cross-study analysis of the differentially expressed genes identified found no consensus genes strongly associated with growth discordance (Marr et al., Reference Marr, Boughorbel, Presnell, Quinn, Chaussabel and Kino2016).
In order to shed further light on the key placental genes associated with growth discordance, our study compared the placental mRNA expression profiles between eight sets of growth-discordant MCDA twins using RNA sequencing (RNA-seq), which is now a well-established methodology to detect and quantitate the levels of genome-wide expressed transcripts in two closely related samples (Mortazavi et al., Reference Mortazavi, Williams, McCue, Schaeffer and Wold2008). Thus, the identification of any differentially expressed placental genes between growth-discordant MCDA twin placentas would, therefore, be a proxy measure of epigenetic discordance. In this study, we identified two placental proteins, namely LEP and age-related maculopathy susceptibility 2 (ARMS2), which were strongly associated with growth discordance in MCDA twins.
Material and Methods
Ethics Statement
The research study was approved by the Human Research Ethics Committee of the Shanghai First Maternity and Infant Hospital (KS1677, June 2016–June 2019). Prior to delivery, patients provided written informed consent for the research studies conducted on their placental tissue collected after delivery.
Clinical Characteristics of Placental Samples
The selected study samples (Table 1) consisted of eight growth-discordant MCDA twin placentas and six growth-concordant MCDA twin placentas without complications (Table 1). MCDA twins were selected by chorionicity confirmation from ultrasound during the first trimester and placental pathology after birth. Growth-discordant MCDA twins selected for the study were defined by more than a 20% birth–weight discordance, according to the formulae (A-B)×100/A, where A is the birth weight of the larger twin, and B is the birth weight of the smaller twin (Lewi et al., Reference Lewi, Devlieger, De Catte and Deprest2014). Conversely, growth-concordant MCDA twins were defined as having a birth–weight discordance less than 10% (Lewi et al., Reference Lewi, Devlieger, De Catte and Deprest2014) and normal umbilical artery doppler for both fetuses (Gratacós et al., Reference Gratacós, Lewi, Muñoz, Acosta-Rojas, Hernandez-Andrade, Martinez and Deprest2007; Valsky, Eixarch, Martinez, Crispi et al., Reference Valsky, Eixarch, Martinez, Crispi and Gratacos2010). Other MCDA complications, such as preeclampsia, twin-to-twin transfusion, twin anemia-polycythemia sequence, fetal structural abnormalities, genetic syndromes, or chromosomal abnormalities were excluded. All pregnant women recruited to the study had no maternal or pregnancy complications.
TABLE 1 Clinical Characteristics of MCDA Pregnancies and Twin Placentas
Collection of Placental Tissue and Isolation of RNA
Immediately following delivery, pieces of tissue (approximately 100 mg) were removed from the middle portion of the placenta, which was identified by inter-twin membrane, each cord insertion, and placental vessel distributions. The biopsied tissue was washed repeatedly in phosphate-buffered saline to remove all traces of contaminating blood, snap frozen in liquid nitrogen, and then stored at -80 degree Celsius. Frozen tissue (~100 mg sliced into smaller fragments) was placed directly into 1.5 ml siliconized eppendorf tubes containing 1 ml of Trizol solution (Qiagen). Total RNA was extracted according to the recommended protocol. In brief, tissue was homogenized and then tubes incubated at room temperature for at least 5 min to dissociate nucleoprotein complexes. Following addition of 0.2 ml of chloroform and vigorous vortexing for 15 s, tubes were further incubated at room temperature for 3 min. Tubes were centrifuged at 12,000xg for 15 min at 4 degree Celsius and the aqueous phase transferred to a fresh tube. Glycogen (10 μg) was added and RNA precipitated by addition of 1 ml absolute ethyl alcohol and incubation at -80 degree Celsius for 1 hour. Tubes were centrifuged at 12,000xg for 10 min at 4 degree Celsius, the RNA pellet washed with 75% ethanol, briefly air-dried, and then finally dissolved in RNase-free water. The concentration (260 nm absorbance) and quality of the purified RNA (260 nm/280 nm absorbance ratio) was determined by Nanodrop-c (Thermo Scientific).
RNA Sequencing and Data Analysis
Poly A+ containing mRNA was isolated from total RNA using poly-T oligo-attached magnetic beads (Illumina, US). A total of 100 ng of poly A+ mRNA was fragmented into an average size of 300 bp by incubation in 100 mM MgCl2 at 95 degree Celsius for 3 min. Fragmented mRNA was reverse transcribed into first strand cDNA using random primers and reverse transcriptase, then copied to double stranded DNA using DNA polymerase and RNase H (Enzymatics, US). Following fragment end repair using DNA modifying enzymes T4 DNA polymerase, Klenow and T4 polynucleotide kinase and, addition of a single A base to the 5ʹ ends using Klenow Fragment (3ʹ to 5ʹ exo-), sequencing adaptors (Liang et al., Reference Liang, Lv, Wang, Xu, Liu, Li and Wu2013) were ligated to each end and then fragments finally amplified by PCR to generate the placental cDNA libraries. Libraries were then subjected to massively parallel sequencing on the Illumina HiSeq2500 platform to produce in the order of 20 million 2 × 100 bp paired end reads.
Paired end reads were uniquely mapped to the hg19 reference genome and contiguous sequences aligned using the TopHat2 algorithm (Kim et al., Reference Kim, Pertea, Trapnell, Pimentel, Kelley and Salzberg2013) incorporated in the web-based analysis protocol http://ccb.jhu.edu/software/tophat/index.shtml. The relative abundance of each transcript was normalized by calculating the reads per kilobase per one million reads (RPKM values; Mortazavi et al., Reference Mortazavi, Williams, McCue, Schaeffer and Wold2008) according to the formula RPKM = total exon reads/mapped reads (millions) × exon length (kb). The significance (p value) of the level of individual gene expression was calculated from the ratio of the RPKM values determined for the small and large placental samples, using the edgeR algorithm (Robinson et al., Reference Robinson, McCarthy and Smyth2010). In a Microsoft excel sheet, significant transcripts (p value < .05) showing >2-fold change in gene expression between the eight small and large growth-discordant twin placental samples (columns) were ordered by gene name (rows), allowing identification of differentially expressed genes.
Quantitative Real-Time PCR
Quantitative real-time PCR (qPCR) to quantitate relative gene expression levels between placental RNA samples was performed using the Kapa SYBR qPCR kit (KAPABIOSYSTEMS) according to the recommended protocol. Briefly, 20 μl reactions were prepared in 200 μl PCR tubes comprising 10 μl of Kapa SYBR FAST qPCR master mix ABI Prism, 0.4 μl of 10 μM forward primer, 0.4 μl of 10 μM reverse primer, 2 μl of cDNA (20 ng) and 7.2 μl of molecular grade water. Reactions were heated at 95 degree Celsius for 3 min to activate the heat stable DNA polymerase and then subjected to 40 two-step PCR cycles of 95 degree Celsius for 3 s and 60 degree Celsius for 25 s. Melting curve analysis was performed and fold changes between comparative samples calculated by the delta delta CT method, with data normalized using the expression levels of the house keeping gene β-actin.
Immunohistochemical Analysis
Paraffin-embedded placental tissue biopsies were cut into 5 µm sections, deparaffinized with xylene, and then gradually rehydrated with decreasing concentrations of ethyl alcohol. Boiled 10 mM citrate buffer (pH 6.0) was used for antigen retrieval and endogenous peroxidase activity was blocked in 3% hydrogen peroxidase prepared in PBS. After blocking further with 10% rabbit serum, sections were probed overnight at 4 degree Celsius with either rabbit anti-LEPTIN antibody (1:100, Proteintech, USA) or rabbit anti-ARMS2 antibody (1:200, Novus, USA). Specific protein expression was visualized using the EnVision Peroxidase/DAB detection system (Dako, Denmark) according to the manufacturer's instructions. Negative control sections were treated similarly, except the primary antibody, was excluded.
Results
The overall study design to identify differentially expressed genes associated with growth-discordant twins is shown in Figure 1. We initially, compared gene expression patterns by RNA-seq between the larger and smaller placentas of eight growth-discordant MCDA twins (144, 149, 152, 166, 170, 197, 202, and 228) with an average gestational age of 33.4 ± 1.21 weeks, range 31+3–34+4 weeks (Table 1). As controls, we also analyzed the gene expression patterns in the smaller and larger placentas of three MCDA growth-concordant twins (234, 262, and 272) with a slightly higher average gestational age of 36.8 ± 0.3 weeks, range 36+3 ± 37 weeks. We hypothesized that by comparing these two sets of clinically well-defined placental samples, any genes associated with growth discordance would be differentially expressed in smaller of the two growth-discordant placentas but not differentially expressed in the smaller of the two growth-concordant placentas.
FIGURE 1. Study design to identify placental genes associated with growth discordance.
Poly A+ cDNA libraries were generated from growth-discordant and concordant placental pairs and subjected to deep sequencing. By Pearson correlation, the transcript profiles between the smaller and larger placenta pairs were highly concordant (range R 2 values of 0.832–0.949), indicating that high-quality gene expression data had been obtained from all placental samples for meaningful comparative analyses. Across all placental mRNA profiles, the top 16 expressed gene transcripts were CSH1, CSH2, CGA, CGB, HBA1, HBA2, PSG1, PSG2, PSG3, HBG2, GH2, and CSHL1, specifically associated with pregnancy, and KISS1, H19, TFPI2, and FBLN1, mainly known to be associated with oncogenesis and tumor suppression (Table 2). All these genes have been previously shown to be highly expressed in independent placental samples analyzed by both microarray (Sood et al., Reference Sood, Zehnder, Druzin and Brown2005) and RNA-seq (Kim et al., Reference Kim, Zhao, Jiang, Lu, Wang, Murray and Xing2012) methodologies. In all cases, none of these genes were differentially expressed in the growth-discordant placenta pairs.
TABLE 2 List of Highly and Differentially Expressed Placental Genes
Note: aRPKM values represent the relative transcript abundance.
To identify differentially expressed genes in the smaller of the two growth-discordant placentas, the fold change for each transcript was plotted against its relative abundance (Figure 2, selected plots). The subset of the statistically significant transcripts (p < .05) with a fold change of >2 were compared across all growth-discordant placenta pairs. Two genes, leptin (LEP) and age-related maculopathy susceptibility 2 (ARMS2), were found to be specifically up-regulated to different degrees in all eight growth-discordant placentas (Table 2). Importantly, the expression levels of these two genes were relatively similar between the smaller and larger placentas of the three growth-concordant twins, suggesting that up-regulation of both LEP and ARMS2 mRNA was associated with growth discordance.
FIGURE 2. Changes in mRNA gene expression between growth-discordant MCDA twin placentas. Note: Transcript distributions are plotted as log fold change (differential expression) versus average log CPM (abundance). The upper blue line marks 2-fold up-regulation and the lower blue line marks 2-fold down-regulation. The distribution of transcript expression change is shown by the cloud of black dots (non-significant change, p > .05) and red dots (significant change, p < .05). Representative plots for four of the eight growth-discordant MCDA placentas (152, 197, 144, and 202) are shown.
Independent confirmation of LEP and ARMS2 mRNA differential expression was performed by qPCR on a second placental biopsy taken in close proximity to the first biopsy used for RNA-seq. Up-regulation of LEP and ARMS2 mRNA was also exclusively observed in each of the smaller of the two growth-discordant placentas at levels, generally similar to those measured by RNA-seq (Figure 3). Levels of increased LEP mRNA expression in the smaller placenta varied from as low as 7.2-fold (sample 228) to as high as 1,774-fold (sample 170). In contrast, levels of increased ARMS2 mRNA expression in the smaller placenta were generally lower, ranging from 3.5-fold (sample 228) to 53.4-fold (sample 170).
FIGURE 3. Differential expression of LEP and ARMS2 mRNA. Note: The relative expression levels of LEP mRNA (A) and ARMS2 mRNA (B) were determined by the ratio of transcript abundance between the small (S) and large (L) placenta of the growth-discordant and concordant MCDA placentas. qPCR and RNA sequencing results are shown by black and grey bars, respectively.
Given that the three control growth-concordant placenta pairs used in the RNA-seq and PCR studies were of a slightly higher gestational age compared to the growth-discordant pairs (Table 1), we further analyzed LEP and ARMS2 mRNA expression levels by qPCR in three additional growth-concordant placentas (397, 424, and 480) with similar gestational ages to the eight growth-discordant MCDA twin placentas (Table 1). In all three cases, we confirmed that the levels of LEP and ARMS2 mRNA were similar between the smaller and larger placentas in these independent control samples (Figure 3).
To further determine whether leptin and ARMS2 were also differentially expressed at the protein level, immunohistochemistry (IHC) was conducted on paraffin embedded tissue sections from selected growth-discordant placental samples (Figure 4). Strong leptin staining was consistently observed in approximately 5–10% of blood vessels and the surrounding chorionic villous cells of the smaller placenta. In contrast, all the blood vessels cells in the larger placenta showed relatively weak or no leptin staining. ARMS2 staining was observed in all placental cell types in both twin placentas, although marginally increased staining was also consistently observed in the microvillous membranes of the smaller placenta.
FIGURE 4. Differential leptin and ARMS2 protein expression. Note: IHC using leptin and ARMS2 specific antibodies was performed on selected tissue sections of growth-discordant placentas (magnification either x200 or x400). IHC results for smaller (S) and larger (L) placenta are shown in the top and bottom panels, respectively. H&E blue staining indicates cellular nuclei and brown cellular staining represents positive leptin (panels A, B, C, D, and E) and ARMS2 (panels G and H) staining. There was no staining of the negative control sections (panels F and I). BV = blood vessel (panel B), CV = chorionic villous, and MV = microvillous (panel H).
Discussion
In this pilot study of clinically defined growth-discordant and concordant MCDA placentas (Table 1), we successfully applied next generation RNA-seq to elucidate differentially expressed placental genes involved in growth discordance. This was performed using human models of growth-discordant MCDA twin pregnancies in order to reduce the contribution of maternal factors and focus specifically on identifying novel placental factors. Detailed comparison of the poly A+ transcriptomes derived from eight growth-discordant twin MCDA placentas identified two genes, LEP and ARMS2. The RNA-seq results were independently confirmed at the mRNA and protein level by qPCR and IHC analyses.
From the very large number of genes found to be expressed in placental tissue, remarkably, only ARMS2 and LEP were significantly up-regulated in all of the smaller discordant MCDA twin placentas, leading us to speculate on their possible functional roles. The first gene that was found to be significantly up-regulated in the growth restricted twin placenta was LEP. In a previous study that examined placental angiogenic gene expression in MCDA growth-discordant twin placentas, three genes LEP, Flt (fms-like tyrosine kinase), and Eng (Endoglin) were consistently up-regulated in the smaller of the co-twin placentas, although leptin showed the strongest association (Schrey et al., Reference Schrey, Kingdom, Baczyk, Fitzgerald, Keating, Ryan and Drewlo2013) and agrees with the findings in this study. Further, in other independent studies, higher placental levels of leptin gene expression has also been documented in cases of pre-eclampsia and insulin-dependent gestational diabetes, as well as in singleton pregnancies with small gestational age placentas and placentas with IUGR (Sagawa et al., Reference Sagawa, Yura, Itoh, Mise, Kakui, Korita and Fujii2002). Leptin is known to be expressed in a variety of tissues; in particular, the white and brown adipose tissue of the placenta (Allison & Myers, Reference Allison and Myers2014). The most recognized function of leptin is regulation of the amount of stored fat in the body and thus is well-known as the ‘satiety’ hormone (Allison & Myers, Reference Allison and Myers2014; Maffei et al., Reference Maffei, Halaas, Ravussin, Pratley, Lee, Zhang and Ranganathan1995).
Despite its important homeostatic role in regulating hunger, leptin has also been shown to have important autocrine effects on placental growth and function (Pérez-Pérez et al., Reference Pérez-Pérez, Sánchez-Jiménez, Maymó, Dueñas, Varone and Sánchez-Margalet2015). Of further interest are two previous reports of leptin acting as a pro-angiogenic factor in the cornea of the eye (Sierra-Honigmann et al., Reference Sierra-Honigmann, Nath, Murakami, García-Cardeña, Papapetropoulos, Sessa and Flores-Riveros1998) and in developing embryos (Talavera-Adame et al., Reference Talavera-Adame, Xiong, Zhao, Arias, Sierra-Honigmann and Farkas2008). In this study, IHC analyses revealed that 10–20% of the blood vessels as well as the immediate surrounding chorionic villous cells had increased leptin expression, exclusively in the smaller placenta. This observation would support the notion that leptin may act as one of the key angiogenic factors to increase vascularization in specific regions of the placenta when under metabolic stress. Therefore, increased local levels of leptin expression, in concert with other angiogenic factors such as VEGF (Schrey et al., Reference Schrey, Kingdom, Baczyk, Fitzgerald, Keating, Ryan and Drewlo2013), may be involved in modifying the vascular endothelium of the growth-restricted placenta. Thus, we postulate that during growth restriction, leptin signaling may form part of a critical survival or compensatory mechanism to increase blood supply and nutrient flow.
The second gene found to be up-regulated in growth restricted placenta was the ARMS2 gene associated with age-related maculopathy, which is a degenerative condition of the retina that leads to progressive visual loss (Horie-Inoue & Inoue, Reference Horie-Inoue and Inoue2014; Ratnapriya & Chew, Reference Ratnapriya and Chew2013). To date, the function of ARMS2 has remained elusive due to a lack of a non-human homologue for manipulation in an animal model. ARMS2 mRNA is known to be highly expressed in the both the retina (Wang et al., Reference Wang, Scott, Whitehead, Court, Kovach, Schwartz and Pericak-Vance2011) and in the placenta (Fritsche et al., Reference Fritsche, Loenhardt, Janssen, Fisher, Rivera, Keilhauer and Weber2008), and the mature form of the protein is secreted into the cytosol (Kortvely et al., Reference Kortvely, Hauck, Duetsch, Gloeckner, Kremmer, Alge-Priglinger and Ueffing2010). In the retina, ARMS2 binds to a number of extracellular matrix proteins implicated in macular dystrophies (Kortvely et al., Reference Kortvely, Hauck, Duetsch, Gloeckner, Kremmer, Alge-Priglinger and Ueffing2010). Further, it is localized mainly to regions of choroid pillars in the retina where pathological drusen deposits occur in patients with age-related maculopathy (Ratnapriya & Chew, Reference Ratnapriya and Chew2013). Thus, based on the available evidence by association with age-related maculopathy, we speculate that elevated levels of ARMS2 protein detected in the growth-restricted twin placenta may reflect an underlying pathology originating from dysfunction of the normal extracellular matrix. Furthermore, as ARMS2 mRNA is almost exclusively expressed in the placenta (Fritsche et al., Reference Fritsche, Loenhardt, Janssen, Fisher, Rivera, Keilhauer and Weber2008) and is secreted (Kortvely et al., Reference Kortvely, Hauck, Duetsch, Gloeckner, Kremmer, Alge-Priglinger and Ueffing2010), it may represent a useful measurable biomarker of growth discordance in MCDA twins.
One limitation of the study was that the growth-discordant MCDA placentas consented for comparative analyses were obtained after deliveries during the late third trimester of pregnancy; thus, we were only able to focus on genes exerting an effect at the end stage of growth discordance. Obtaining access to placental mRNA, prospectively during the course of a MCDA twin pregnancy would allow the identification of placental genes more relevant to the initiation and perpetuation of growth discordance. One emerging approach is the analysis of the circulating RNA fraction in the maternal plasma during pregnancy (Koh et al., Reference Koh, Pan, Gawad, Fan, Kerchner, Wyss-Coray and Quake2014). Robust techniques for isolating the total RNA fraction (Spornraft et al., Reference Spornraft, Kirchner, Haase, Benes, Pfaffl and Riedmaier2014) and identifying key elements of the entire transcriptome, including mRNA and miRNA (Huang et al., Reference Huang, Yuan, Tschannen, Sun, Jacob, Du and Wang2013), are already well-advanced. In addition, proof of concept for detection of placental-specific mRNAs such as H19 that was found to be highly expressed in our study, as well as pregnancy-specific RNAs, has now been recently demonstrated (Tsui et al., Reference Tsui, Jiang, Wong, Leung, Chan, Chiu and Lo2014). Therefore, multiple blood sampling to obtain a liquid plasma RNA biopsy would be an ethically acceptable research study for women with MCDA twin pregnancies. This type of minimally non-invasive approach may provide novel insights into the key genes and biological pathways involved in the development of growth discordance in MCDA twins. Furthermore, such an approach would also have potential to reveal valuable measurable biomarkers for identifying and monitoring MCDA twin pregnancies that are at high risk for growth discordance.
In conclusion, we identified two placental proteins, leptin and ARMS2, strongly associated with growth discordance in MCDA twins. We postulate that leptin and ARMS2 may play a key role to drive angiogenesis as a survival mechanism to compensate for the decreased blood and nutrient supply in the growth-restricted MCDA twin. Further, since both leptin and ARMS2 are produced as soluble proteins, they may also have potential as secreted biomarkers for predicting growth discordance in MCDA twins. Continued studies are required to understand the key epigenetic changes in growth-discordant MCDA twin pregnancies, so that medical interventions or potential therapies can be developed to alleviate the adverse perinatal outcomes associated with this pregnancy complication.
Acknowledgments
We thank the staff in the Fetal Medicine Unit and Prenatal Diagnosis Center, Department of Obstetrics in Shanghai First Maternity and Infant Hospital for their clinical support during the course of this research project. This work was supported by the Fund of Shanghai Municipal Science and Technology Commission to Dr Luming Sun (grant number 16411963100).
Conflict of Interest
LS, JZ, and DSC are employees of Berry Genomics Corporation, Beijing.
