Gene expression profiling of compromised umbilical cords (CUCs) derived from somatic cell nuclear transfer (scNT) clones was performed to determine why scNT-derived clones often exhibit malformed umbilical arteries. Umbilical cord samples were obtained from 65 scNT piglets, and of these, nine displayed a CUC. Microscopic analyses of the scNT clones with CUCs (scNT-CUCs) revealed complete occlusive thrombi that were not detected in the arteries of scNT clones with normal umbilical cords (scNT-Ns). Moreover, whereas the allantoic ducts of the scNT-Ns contained columnar epithelium, the scNT-CUCs lacked this epithelial layer. Compared to scNT-Ns, the scNT-CUCs exhibited severe histological damage, including tissue swelling and vein and arterial damage with complete occlusive thrombi. To investigate functional abnormality, gene expression profiles were created in duplicate using the Platinum Pig 13K oligonucleotide microarray, which contains 13 610 probes of 70 bp in length and is capable of interrogating 13 297 targets with up to one probe per target. Probe sets were selected according to a 2-fold or greater increase or decrease of gene expression in scNT-CUCs compared to scNT-Ns. Most genes expressed in scNT-Ns were also expressed by scNT-CUCs. However, most genes involved in transcriptional regulation, such as JUN, JUNB, and FOSL2, showed a significant decrease in expression in the scNT-CUCs, which may produce a ripple effect capable of altering the transcriptomes of many other cellular processes, including angiogenesis, antioxidation, and apoptosis. The scNT-CUCs with thrombosis showed extensive apoptosis leading to placental insufficiency and related pathology. Considering that the umbilical cord plays a role in the transportation of metabolites to the fetus, placental insufficiency in scNT-CUCs may be caused by an increase in apoptotic protein expression from scNT-derived umbilical cords with hypoplastic arteries, and our results provide evidence that porcine oligonucleotide microarray analysis is a useful tool for screening scNT-derived abnormalities in pigs.
INTRODUCTION
Despite the advantages of somatic cell nuclear transfer (scNT), this procedure produces normal offspring with a very low efficiency, in part because of a high incidence of fetal death resulting from a variety of complications during pregnancy, including fetal overgrowth and placental malformations [1–5]. In addition, scNT cloning is associated with particularly high levels of phenotypic variability [2, 6–9], with 93% of mice, 64% of cattle, 60% of pigs, and 40% of sheep exhibiting some form of abnormality, including pulmonary hypertension and respiratory distress. Many other researchers, however, have observed that cloned animals are healthy and normal [10, 11]. Nevertheless, the usefulness of scNT is clearly diminished by its low cloning efficiency and the high levels of phenotypic variability observed in cloned animals that survive to adulthood.
Compromised umbilical cords (CUCs) are defined as malformed arteries that exhibit an artery-artery diameter difference of approximately 50% [12]. Very recently, we reported that scNT clones with CUCs (scNT-CUCs) are closely associated with poor perinatal outcome, including stillbirth and early death after birth [12]. Despite its importance in development and the need for a basic understanding of umbilical cord abnormalities, the underlying mechanisms have been poorly studied. To this end, we set out to compare the expression of multiple genes in scNT clones with and without umbilical cord abnormalities. The microarray-based genomic survey is a high-throughput approach that allows parallel study of the expression patterns of thousands of genes [13]. In the present study, we discovered previously unknown gene expression patterns related to CUCs. In agreement with the sites of phenotypic alterations found in scNT-CUCs, many differentially expressed genes were shared between the gene lists for comparisons between scNT-derived clones with normal umbilical cords (scNT-Ns) and scNT-CUCs, suggesting that common pathways are affected by the technique and culture system used for scNT and that many of these unique genes are good candidates for screening umbilical cords with compromised arteries.
MATERIALS AND METHODS
Ethics Statement
The treatment of the pigs used in this research followed the guidelines of the National Institute of Animal Science's Institutional Animal Care and Use Committee, Suwon, South Korea (approval no. 2009‐004, D-grade).
In Vitro Maturation of Oocytes
Ovaries were collected from prepubertal gilts at a local slaughterhouse and transported to the laboratory at 25–35°C. Antral follicles (diameter, 2–6 mm) were aspirated with an 18-gauge needle, and aspirated oocytes, with an evenly granulated cytoplasm and surrounded by at least three uniform layers of compact cumulus cells, were selected and washed three times in Tyrode lactate-Hepes with 0.1% polyvinyl alcohol (PVA). Oocytes were cultured in four-well plates containing North Carolina State University-23 (NCSU-23) medium (500 μl/well) supplemented with 10% porcine follicular fluid, 0.6 mmol/L of cysteine, 1 mmol/L of dibutyryl cAMP (dbcAMP; Sigma-Aldrich), and 0.1 IU/ml of human menopausal gonadotropin (hMG; Teikokuzoki) for 20 h. Oocytes were further cultured without dbcAMP and hMG for another 18–24 h as reported previously [3–6, 8, 9, 12, 14–17].
scNT and Embryo Transfer
Nuclear transfer was carried out as described in previous reports [3–6, 8, 9, 12, 14–17]. Briefly, mature eggs with the first polar body were cultured in NCSU-23 medium supplemented with 0.4 mg/ml of demecolcine (Sigma-Aldrich) and 0.05 mol/L of sucrose for 1 h. Sucrose was used to enlarge the perivitelline space of the eggs. Treated eggs with a protruding membrane were moved to medium supplemented with 5 mg/ml of cytochalasin B and 0.4 mg/ml of demecolcine, and the protrusion was removed with a beveled pipette. A single donor fetal fibroblast cell derived from a Gestational Day 30-derived Duroc, Berkshire, or three-way hybrid (Landrace × Duroc × Yorkshire) fetus was injected into the perivitelline space of each egg and electrically fused using two direct current pulses of 150 V/mm for 50 msec in 0.28 mol/L of mannitol supplemented with 0.1 mM MgSO4 and 0.01% PVA (Sigma-Aldrich). Fused eggs were cultured in medium with 0.4 mg/ml of colcemid for 1 h before parthenogenetic activation and then cultured in 5 mg/ml of cytochalasin B-supplemented medium for 4 h. The reconstructed oocytes were activated by two direct current pulses of 100 V/mm for 20 msec in 0.28 mol/L of mannitol supplemented with 0.1 mmol/L of MgSO4 and 0.05 mmol/L of CaCl2. Activated eggs were cultured in the NCSU-23 medium for 6 days in an atmosphere of 5% CO2 and 95% air at 39°C.
Sampling of CUCs
Gilts (Duroc × Yorkshire × Landrace; age, ≥8 mo) were used as recipients, and estrous synchronization of recipients was carried out as reported previously [3–6, 8, 9, 12, 14–17]. The scNT embryos were surgically transferred into the oviducts of synchronized recipients, and the pregnancy status of recipients was determined by ultrasound between Days 30 and 35. The recipients produced scNT-derived piglets via vaginal delivery. In total, 65 cloned pigs were obtained from 24 recipients using each donor fibroblast cell derived from a Duroc, Berkshire, or three-way hybrid (Landrace × Duroc × Yorkshire) fetus [12]. Of these, nine piglets were derived from Duroc, 12 piglets from three-way hybrid, and 44 piglets from Berkshire. Three clones with CUCs, based on the morphological changes of umbilical cords, were found in Birkshire-derived piglets (referring to scNT-CUC-205, ‐207, and ‐208). In addition, to minimize breeder variability, normal umbilical cord samples were also collected from three Berkshire-derived scNT-Ns showing the normal growth even 6 mo after birth and no phenotypical abnormalities. However, three scNT-CUCs showing malformed umbilical cords died suddenly within the first week of birth.
Generation of Pig Oligonucleotide 13K Microarray Chip
A total of 13 297 oligonucleotide probes (pig genome Array-Ready Oligo Sets, Version 1.0) were purchased from Operon Biotechnologies, Inc. [18]. These probes represent 10 665 Sus scrofa genes with a similarity to known human and mouse transcripts or 3′ expressed sequence tag. They were originally designed using the Institute of Genome Research Gene Index Database SsGI Release 5.0. To verify chip quality and effective normalization, we also spotted control samples in each block. In this pig oligonucleotide 13K chip, each of 24 blocks consists of 24 columns and 24 rows and contains 568 genes.
Hybridization
The microarray experiments were performed as described previously [19]. Briefly, each RNA (30 μg) from the umbilical cord of three scNT-Ns and three scNT-CUCs was labeled with Cy3- and Cy5-conjugated dCTP (GE Healthcare) during reverse transcription reaction using a reverse transcriptase, SuperScript II (Invitrogen). The labeled cDNAs were mixed and hybridized simultaneously to the pig oligonucleotide 13K chip. To control the gene-specific dye bias, we performed a dye-swap experiment for all samples as described previously [20]. Processed slides were scanned using an Axon 4000B Scanner (Axon Instruments), with excitation at 532 and 635 nm for the Cy3 and Cy5 dyes, respectively. The scanned images for each slide were analyzed using the Gen-Pix Pro 5.1 Software (Axon Instruments).
Microarray Data Processing and Analysis
Microarray data were managed with GeneSpring 7.2 software (Sillicongenetics). The raw intensity data were normalized by intensity-dependent normalization in Lowess method [21] and then again by with-print-tip group normalization method for each print-tip. A total of 48 tips were used for making this pig oligonucleotide 13K chip. S-plus PLUS software (TIBCO Software) was used to determine the means of data from triplicate experiments. The gene expression values for each array were normalized to their respective median values. All clustering analyses were performed using standard correlations as described previously [22, 23]. Fold-change filters included the requirement that the genes should be expressed in umbilical cords of scNT-CUCs at least 2-fold higher or lower than in those of scNT-N controls.
RNA Isolation and Real-Time RT-PCR
Total RNA was extracted from umbilical cord tissue using a Micro-to-Midi Total RNA Purification System (Life Technologies, Inc.). Real-time RT-PCR was conducted using a DNA Engine OPTICON2 system (MJ Research) and SYBR Green as the double-stranded DNA-specific fluorescent dye (SYBR Green qPCR premix; Finnzymes). Target gene expression levels were normalized to GAPDH gene expression, which was unaffected in scNT-derived pigs. The RT-PCR primer sets are shown in Table 1. Real-time RT-PCR was performed independently in triplicate for different samples, and the data are presented as the mean value of gene expression measured in each individual scNT-N and scNT-CUC sample, which is used as an experimental unit.
TABLE 1.
Primer sets used in this study.
Western Blot Analysis
Western blot analysis was performed as described previously [5, 12]. Molecular weight standards were obtained from New England Biolabs. Membranes were probed with primary antibodies recognizing SOD2 (gifts from Dr. Han Geuk Seo, GyeongSang National University, Jinju, South Korea); JUNB, JUN, FOSL2, YWHAE, PRDX2, BAX, BAD, and CYCS antibodies were purchased from Santa Cruz Biotechnology. Active CASP3 was from Merck & Co., and the ACTB antibody was from Chemicon. Thereafter, the membranes were incubated with an appropriate horseradish peroxidase-conjugated secondary antibody (Jackson Immunoresearch) and subjected to enhanced chemiluminescence (ECL) analysis (Amersham). An anti-actin antibody (1:500; Chemicon) was used to verify equal protein loading, and signals were visualized using the ECL kit (Amersham). Band intensities representing the expression of each protein were quantified by Image processing and analyzed using Image J 1.23 (NIH Image).
Statistics
Values are reported as the mean ± standard deviation. Ratios of fetal weight to placental weight in Figure 1B were analyzed using 10 scNT-N and nine scNT-CUC samples. Analysis of relative expression of genes in Figure 3 was performed using three scNT-N and three scNT-CUC umbilical cords. Statistical significances was confirmed using t-test.
RESULTS
Cloned scNT-Derived Piglets Exhibit CUCs
Previously, we reported that 9 of 65 scNT clones showed CUCs and compromised umbilical veins, presenting with gross necrosis and thrombosis, respectively [12]. In the present study, microscopic analyses of the nine scNT-CUCs revealed complete occlusive thrombi in the umbilical arteries (Fig. 1Ab) that were not detected in the umbilical arteries (Fig. 1Aa) of scNT-Ns. Moreover, whereas the allantoic ducts of the scNT-Ns exhibited columnar epithelium (Fig. 1Ac, arrow), the scNT-CUCs lacked this epithelial layer (Fig. 1Ad).
Next, the relationship between postnatal death and survival to adulthood were compared, and the associated factors were analyzed. The fetal weight:placental weight ratios (n = 9; 3.78 ± 0.37) of scNT-CUCs were significantly lower than those of the scNT-Ns (n = 10; 5.69 ± 1.3) (Fig. 1B). In general, scNT clones survived to adulthood when the fetal weight:placental weight ratio exceeded five, whereas a ratio of less than five showed significant association with postnatal early death among scNT clones. This observation may partly explain the high mortality rates of scNT-CUCs soon after birth [12].
To assess differentially expressed genes of the CUC in more detail, gene expression profiles of scNT-CUC-205, ‐207, and ‐208 were analyzed using Pig 13K oligonucleotide microarrays.The results showed that 134 genes were differentially expressed with a more than 2-fold change in expression. Of note, most of the differentially expressed genes in scNT-CUCs were down-regulated (113 genes), with only 21 genes showing up-regulation. These up-regulated genes included cytoskeleton-related gene (B2M), proteolysis-related gene (UCHL5, UBE2V2, and STEFIN A8), transcription-related gene (CDK9 and C14orf166), transport-related gene (IAG2), and others (PDZD3, PCF11, RASGRP1, ARGLU1, C8G, NELL2, LOC68117, KARS, CD163, PCK2, LOC283951, SMPD2, CHDH, and PPP6C) (Table 2).
TABLE 2.
Differential up-regulated genes in SCNT-derived umbilical cord with CUC.
Classification and Characterization of Identified Genes
The 134 unique genes identified with high confidence were classified in terms of molecular function and biological process using the Panther classification system. Panther is a software program freely accessible on the Web (www.pantherdb.org) that provides a platform for assigning families, functional classifications, and pathways to gene products [24, 25]. The results, expressed as pie charts using the Panther classification system, are shown in Figure 2, B and C. In the classification based on biological process (cellular component in Fig. 2B), proteins categorized as “nucleus proteins” were the most abundant and accounted for 32.4% of the total number of genes identified. In the classification based on molecular function, the categories containing most genes were “transcription” (13.6%), “signal transduction” (11.1%), and “angiogenesis” (8.5%) (Fig. 2C).
Differential DNA Expression in scNT-CUCs
Differentially expressed genes were identified using fold-change analysis of pairwise comparisons between groups of abnormal scNT-CUC and scNT-N samples. Of all the differentially expressed genes, 113 genes were down-regulated in all scNT-CUCs when compared to scNT-Ns and their fold-change. GenBank accession numbers and annotations are listed in Table 3. Genes are classified into 15 groups based on function: 1) angiogenesis-related genes (SERPINE1, VEGFA, CYR61, CTGF, TNFRSF12A, RHOH6, TNR12, RHOB, PLVAP, and CRIM1), 2) apoptosis-related genes (PHLDA2, DIABLO, PPP1R15A, and NFKBIA), 3) cell cycle-related genes (CDKN1A, HEATR7A, C13orf15, and PPP1CB), 4) chaperone-related genes (HSPA1B, CCT3, HSPH1, and AHSP), 5) cytoskeleton-related genes (DCTN1, DCTN4, KRT7, MARK4, and INPP5K), 6) hypoxia-related genes (PLAUR, DUSP1, ALAS2, and HIGD2A), 7) protein modification-related genes (MPI, ALG3, GALNT5, FES, and NAA10), 8) proteolysis-related genes (UBQLN3, FBXO9, and CAPN10), 9) redox-related genes (YWHAE, ACOX1, FTL, and CPT1A), 10) stress-related genes ([RAD23B, RAD23A, MPG, HSPA4, GADD45B, and CTGF), 11) signal pathway-related genes (SMAD4, FIBP, ADM, RPTOR, RRAD, PRKAA2, RGS3, RND3, ADORA3, RGS2, RANGAP1, and RAP1GDS1), 12) transcription-related genes (HBP1, NCOR1, BTG1, ZMIZ1, RCAN3, CHAF1B, CHAF1A, FOS, JUNB, JUN, JUND, ZSCAN21, NAB2, ZFP36L2, BUD31, ZFP36, MYC, BTG2, and KLF10), 13) transport-related genes (TPCN2, ABCA9, ATP6V1H, ABCC10, COX10, and PEX5), 14) RNA processing-related genes (PNPT1, PAPD4, DKC1, and SRSF1), and 15) others (LOC396596, ATPIF1, BRD2, SDS, TM4SF4, HPS4, SLC25A39, CRYAB, NFKBIL2, GGCT, GLCCI1, SYNGR1, S100A2, VAMP3, MOBK1B, ESYT1, PDXK, FAM129B, SCHIP1, C10orf58, MPDZ, TEX264, and ZFAND2A).
TABLE 3.
Differential down-regulated genes in SCNT-derived umbilical cord with CUC.
Down-Regulation of JUN, JUNB, and FOSL2 Cause a Ripple Effect that Alters Receptor Tyrosine Kinase Signaling
In our DNA chip study, the most important finding was that most molecules (JUN, JUNB, JUND, FOS, FOSB, FOSL1, and FOSL2) involved in receptor tyrosine kinase (RTK) signaling, including vascular endothelial growth factor (VEGF), showed lower expression in scNT-CUC samples than in scNT-N samples (Table 3). To illustrate the reliability of the microarray data and to validate the findings, we conducted quantitative real-time RT-PCR on umbilical cord total RNA from scNT-N and scNT-CUC samples for seven chosen genes (JUN, JUNB, JUND, FOS, FOSB, FOSL1, and FOSL2) (Fig. 3). In addition, we evaluated protein expression in scNT-N and scNT-CUC samples for three genes, JUN, JUNB, and FOSL2, which are involved in the regulation of transcription (Fig. 4D). All seven genes and three proteins showed the same trend with respect to changes between scNT-N and scNT-CUC samples when analyzed by real-time RT-PCR and microarray. Quality-control measures, including the ratios of the housekeeping genes glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ACTB, scaling factors, background, and Q-values were all within acceptable limits. Taken together, our findings show that the CUC is associated with relatively low expression of RTK signaling molecules.
Low Antioxidant Gene Expression Related to Dysfunction of scNT-Derived Umbilical Cord
Previously, we reported that scNT-derived umbilical cords with thrombosis showed extensive apoptosis [12]. Apoptosis can be triggered by many different cellular stimuli, including various types of stress and damage, and consists of a cascade of events leading to the ordered dismantling of critical cell survival components and pathways. As shown in Tables 2 and 3, we found that apoptosis-related genes (UCHL5 and CDK9) were up-regulated but that antioxidation-related genes (YWHAE, DUSP1, ALAS2, and HIGD2A) and stress-related genes (RAD23A, RAD23B, MPG, HSPA5, GADD45B, and CTGF) were significantly down-regulated.
The 14‐3‐3 protein family members have been implicated in the protection of cells from apoptosis through binding to the pro-apoptotic protein BAD [5, 26]. Therefore, we next examined expression of the YWHAE gene, which was among the 113 genes that were down-regulated in scNT-CUC samples (Table 3). Although YWHAE gene expression was high in all scNT-N samples, most scNT-CUC samples exhibited significantly lower expression of this gene (12.5-fold). To determine the effect of YWHAE protein down-regulation on apoptosis, we performed Western blot analysis of scNT-N and scNT-CUC samples. It is interesting to note that Western blot analysis demonstrated that BAX, CYCS, and CASP3 protein expression in scNT-CUC samples showed a significant increase when compared to scNT-N samples (Fig. 4C), suggesting that apoptosis in scNT-CUCs occurred following the down-regulation of antioxidants, such as SOD2, PRDX2, and YWHAE. Considering that the umbilical cord plays a role in the transportation of metabolites to the fetus, placental insufficiency in scNT-CUCs may be caused by an increase in apoptotic protein expression from scNT-derived umbilical cords with hypoplastic arteries. Taken together, our results provide evidence that porcine oligonucleotide microarray analysis is a useful tool for screening for scNT-derived abnormalities in pigs.
DISCUSSION
The umbilical cord normally contains one vein and two arteries [27]. Many intrapartum complications and adverse perinatal outcomes, including stillbirth and intrauterine growth restriction, have been associated with umbilical cord abnormalities. At present, the classification of these malformations is based on angiographic, intraoperative, and histological findings [28–33]. Therefore, the scNT-CUC piglet is a good model for the study of reduced cloning efficiency.
Signal Transduction
Previously, we reported that 9 of 65 scNT cloned pigs exhibited a CUC [12], and the present study extends our proteomic analysis to create a transcriptional profile for the mechanism underlying umbilical cord abnormalities in scNT clones. As far as we are aware, no study has been conducted with the purpose of elucidating a specific association between CUC and mRNA expression patterns. As shown in Figure 4, B and C, in scNT-CUCs the expression of two signal transduction-related genes, extracellular signal-regulated kinase (ERK) 1/2 and c-Jun N-terminal kinase (JNK), was significantly down-regulated. At nontoxic levels, H2O2 acts as a mediator of signal transduction in various cellular systems via the mitogen-activated protein kinase (MAPK) pathway [34, 35]. MAPK is a family of serine/threonine-specific protein kinases, consisting of three isoforms: MAPK1/3 (ERK 1/2), MAPK14 (p38 MAPK), and MAPK8 (JNK) [36, 37]. On the other hand, high levels of H2O2 are toxic to cells and cause the oxidation of biological macromolecules, resulting in DNA breaks, lipid peroxidation, and oxidative damage of proteins. This is referred to as oxidative stress, and it is associated with the vascular apoptosis mediated by JNK [38–40].
In the present study, scNT-CUC samples showed relatively low expression of JUNB, JUN, and FOSL2 proteins involved in RTK signaling (Figs. 3 and 4D) [4]. In general, the RTK signaling pathway is involved in a wide range of fundamental cellular processes, including the cell cycle, cell migration, proliferation, and differentiation [41]. In a mouse model, JUNB−/− fetuses die between Embryonic Days 8.5 and 10.0 because of multiple defects in extraembryonic tissues, such as the placental labyrinth [42]. In addition, c-Fos participates in cell differentiation, growth, and transport processes that occur in mouse extraembryonic tissues [43]. Taken together, the presence of abnormal umbilical cords in scNT clones suggests that a failure in the fetomaternal exchange of nutrients, resulting from defects in umbilical cord tissues, may be the ultimate cause of the low efficiency in scNT cloning. Moreover, the results reported here provide the first evidence, to our knowledge, that members of the AP-1 transcription factor complex, such as JUNB, JUN, and FOSL2, play a key role in umbilical cord development.
Metabolism
Oxidative stress involves the physiological modulation of cells, tissues, and lipids as a result of their interactions with free radicals. These interactions can increase, decrease, or alter the function of specific proteins, depending upon the degree and type of protein modulation. Free radicals are highly unstable molecules that interact quickly and aggressively with other molecules in our bodies to create abnormal cells [44]. They are capable of penetrating and damaging the DNA of a cell to produce mutations that cause the cell to replicate out of control. In the present study, stress-related genes (RAD23A, RAD23B, MPG, HSPA5, GADD45B, and CTGF) and antioxidant enzymes, such as SOD2 and PRDX2 (Fig. 4A), were down-regulated significantly. SOD2 and PRDX2 are well-known proteins that protect the cell against oxidative stress. This reduction in protein expression could be a result of the inability of the cell to protect itself against oxidative stress, which suggests that down-regulation of these genes can disrupt endothelial cell function in the umbilical cord.
Angiogenesis
CYR61 is an estrogen-regulated gene that promotes cell adhesion, migration, and neovascularization [45–47]. Previous studies have demonstrated that hypoxia-inducible factor-1 alpha (HIF1A) interacts with JUN and may thereby contribute to the transcriptional regulation of CYR61 under hypoxic conditions in human melanoma cells [46]. In the present study, we found that the angiogenesis-related genes (SERPINE1, VEGFA, CYR61, CTGF, TNFRSF12A, RHOH6, TNR12, RHOB, PLVAP, and CRIM1) in scNT umbilical cords with hypoplastic arteries were significantly down-regulated (Table 3). Thus, a reduction in the number of cells expressing CYR61 mRNA may be caused by the down-regulation of other angiogenesis-related genes, resulting in the development of an umbilical cord with compromised arteries. However, the specific mechanisms through which CYR61 mediates the pathogenesis of the umbilical cord remain to be fully elucidated.
A strong inducer of vascular permeability, VEGF is also a stimulator of endothelial cell migration and proliferation and an important factor in the survival of newly formed blood vessels [48–50]. In the present study, most of the angiogenesis-related genes were significantly down-regulated (Table 3). Furthermore, adrenomedullin (AM) and VEGF gene expression was significantly down-regulated in compromised scNT-derived umbilical cords. In cultured endothelial cells, AM and VEGF act together to induce angiogenic effects in vitro [51, 52]. The angiogenic actions of AM, however, appear to be independent of VEGF secretion [53], suggesting that AM is not influenced by the up-regulation of VEGF.
A recent study, however, suggested that AM-mediated signals can interact with VEGF-mediated signaling pathways and that AM plays a key role in development of the vascular system, because AM homozygous knockout mice (AM−/−) died in utero as a result of poor angiogenesis in the placenta [54]. Therefore, the down-regulation of AM and VEGF in endothelial cells from compromised scNT-derived umbilical cords would mean that AM and VEGF could not promote cell proliferation and differentiation into cord-like structures. In fact, our previous data showed that scNT-CUC-derived endothelial cells migrated and formed tubules more slowly than control and scNT-N-derived cells [12]. Furthermore, endothelial cells in scNT-CUC samples could not elicit mitogenic responses and induce tubule formation as a result of the down-regulation of the vascular signaling molecules, VEGF, angiopoietin, and/or their cognate receptors, which play essential roles in vascular development. Collectively, our studies clearly demonstrate that scNT-CUCs show impaired endothelial cell migration and angiogenesis, which could lead to the development of CUC and/or placental insufficiency.
Apoptosis
The 14‐3‐3 proteins block BAD-mediated cell death by promoting the accumulation of BAD (phosphorylated at Ser-155), rendering it incapable of binding to BCL2L1 [5, 55, 56]. Interestingly, our results showed that BAD expression increased in scNT-CUCs, whereas expression of its counterparts, the 14‐3‐3 proteins, decreased. Moreover, we also observed a marked increase in the expression of apoptotic signal-related proteins, such as CYCS and CASP3, in scNT-CUCs. As mentioned above, the causative mechanism for CUC is poorly understood, but these results suggest that it may entail an apoptosis-related signal transduction pathway, involving up-regulated, proapoptotic BAD activity and down‐regulated 14‐3‐3 expression.
Recently, we reported that increased oxygen demand during CUC development leads to an increased rate of production for reactive oxygen species (ROS): Oxidative stress contributes to ROS formation in the CUC, and ROS levels increase gradually during early pregnancy, resulting in apoptotic cell death in CUCs [12]. In the present study, the expression of most detoxification-related proteins and antioxidant enzymes, particularly SOD2, was down-regulated in scNT-CUCs (Fig. 4A). Therefore, our data suggest that disruption of the antioxidant defense system and subsequent apoptotic cell death is another characteristic feature of scNT-CUCs.
In conclusion, in the present study, gene expression profiling of CUCs derived from scNT clones was performed to examine why scNT-derived clones often exhibit malformed umbilical arteries. Compared to scNT-Ns, scNT-CUCs exhibited severe histological damage, including tissue swelling and vein and arterial damage with complete occlusive thrombi. Moreover, scNT-CUCs showed a significant decrease in the expression of genes involved in transcriptional regulation, such as JUN, JUNB, and FOSL2, which may produce a ripple effect capable of altering the transcriptomes of many other cellular processes, including angiogenesis, antioxidation, and apoptosis. In fact, scNT-CUCs with thrombosis showed extensive apoptosis leading to placental insufficiency and its related pathology. Even though conflicting data from Affymetrix chips and quantitative RT-PCR are found and could be a result of cross-hybridization and probe location [57, 58], the results of the present study, like those of other studies, showed a strong correlation between microarray data and quantitative RT-PCR measurements, promoting confidence in the quantitative nature of microarray data. In addition, our findings confirm the correlation between histopathological data and gene expression levels in scNT piglets, providing the evidence that porcine oligonucleotide microarray analysis is a useful tool for screening scNT-derived abnormalities in pigs.
Acknowledgment
We thank Dr. Patrick Hughes (Bioedit) for assistance with preparation of the manuscript.