This research investigated gene expression in the canine placenta during the peripartum period. Previous studies have recognized molecular changes that occur in the placenta around the time of placental release in other species, but no study has looked at gene expression in the late gestation canine placenta. Of particular significance for this thesis work is the groundwork laid for future studies modeling placental abnormalities in dogs (e.g. subinvolution of placental sites) and humans (e.g. preeclampsia, placenta accreta). Despite years of research in multiple species, the exact mechanisms and processes regulating trophoblast invasion and placental release remain unclear. Therefore, the specific objective of this research was to characterize gene expression changes that occur during the peripartum period in the dog using microarray and real-time RT-PCR. Following total RNA isolation, the microarray analysis was performed by hybridizing total RNA to the Canine 2.0 Array (Affymetrix, Santa Clara, CA). Microarray analysis was carried out using the limma and affy packages through the Bioconductor software in the R statistical environment. Differential expression was defined as p ��� 0.05, FDR p ��� 0.10 and a log fold change of ��� 1.2. Following cDNA synthesis, real-time RT-PCR was performed using TaqMan primer and probes that were pre-made and pre-optimized for canine tissues (Applied Biosystems, Carlsbad, CA). Microarray analysis showed differential expression in 18 genes between pre-term and pre-labor sample groups, 38 genes that were differentially expressed between pre-term and parturient samples and no genes that were differentially expressed between pre-labor and parturient samples. Microarray analysis led to the identification of several candidate genes for closer investigation using real-time RT-PCR. These genes included MMP-1, MMP-2, MMP-9, TIMP-2, VEGF-A, Flt-1, CD44, DAG-1, IL-6 and CXCL10. All of these genes have been linked to trophoblast invasion or regression or placental release in a number of species including humans, cattle and rodents. Using real-time RT-PCR, there was a significant difference in MMP-9 mRNA expression in pre-term samples compared to pre- labor and parturient samples (p<0.05). However, there was no significant difference in mRNA expression of MMP-2, TIMP-2, VEGF-A, Flt-1 CD44, DAG-1, IL-6 or CXCL10. Future studies may focus on additional candidate genes identified by microarray that play a role in tissue remodeling at the end of canine gestation such as IL-8, EPHX2, PI3 and SERPINE1. / Graduation date: 2013

Although not obvious at first sight, several parallels can be drawn between pregnancy andcancer. Many proliferative, invasive and immune tolerance mechanisms that supportnormal pregnancy are also exploited by malignancies to establish a nutrient supply andevade or edit the immune response of the host. The human placenta, of crucial importancefor pregnancy success, and its main cells, the trophoblast, share several features withmalignant cells such as high cell proliferation rate, lack of cell-contact inhibition andinvasiveness. Both in cancer and in pregnancy, the immune defense mechanisms,potentially threatening the survival of the tumor or the fetus, are progressively blunted oreven turned into tumor- or pregnancy-promoting players. Amongst immune mechanisms that are meant to protect the host from cancer and can be apotential threat to the fetus, the NKG2D receptor-ligand system stands out as the mostpowerful, stress-inducible “danger detector” system that comprises the activating NK cellreceptor NKG2D and its ligands, the MIC (MHC class I Chain-related proteins A and B)and ULBP (UL-16 Binding Proteins) families. It is the major cytotoxic mechanism in thebody promoting surveillance and homeostasis. In the present thesis we investigate theNKG2D receptor-ligand system in human early normal pregnancy and in theleukemia/lymphoma cell lines Jurkat and Raji and ask the questions “How is the NKG2Dreceptor-ligand system functioning in pregnancy and tumor? How is the danger of cytotoxicattack of the fetus avoided? Why is the immunosurveillance function compromised incancer patients?” We developed a method to isolate and culture villous trophoblast from early human normalplacenta and used it to study the NKG2D receptor-ligand system. We discovered that theNKG2D ligand families of molecules MICA/B and ULBP1-5 are constitutively expressedby the syncytiotrophoblast of the chorionic villi. Using immnunoelectron microscopy, westudied the expression of these molecules at the subcellular level and could show for thefirst time that they are preferably expressed on microvesicles in multivesicular bodies(MVB) of the late endosomal compartment and are secreted as exosomes. Exosomes arenanometer sized microvesicles of endosomal origin, produced and secreted by a great7variety of normal and tumor cells. The exosomes are packages of proteins and ribonucleicacids that function as “mail” or “messengers” between cells conveying different biologicalinformation. We isolated and studied exosomes from placental explant cultures. We foundthat they carry NKG2D ligands on their surface and are able to bind and down-regulate thecognate receptor on NK-, CD8+ and <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Cgamma" /><img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Cdelta" />T cells. The down-regulation selectively causedimpairment of the cytotoxic response of the cells but did not affect their lytic ability asmeasured by perforin content and gene transcription. Thus, the NKG2D ligand-bearingexosomes suppress the cytotoxic activity of the cells in the vicinity of the placenta, leavingtheir cytolytic machinery intact, ready to function when the cognate receptor isrestored/recycled. These findings highlight the role of placental exosomes in the fetalmaternalimmune escape and support the view of placenta as an unique immunomodulatoryorgan. Next, we studied the expression and exosomal release of NKG2D ligands by tumor cellsusing the leukemia cell lines Jurkat and Raji as a tumor model. We found that NKG2Dligand-bearing exosomes with similar immunosuppressive properties as placental exosomesare constitutively secreted by the tumor cells, as a mechanism to blunt the cytotoxicresponse of the immune cells and thus protect themselves from cytotoxic attack by the host.Interestingly, we found that thermal- and oxidative stress up-regulates the exosomesecretion and the amount of exosome-secreted NKG2D ligands. Our results imply thattumor therapies that cause stress-induced damage, such as thermotherapy and stripping ofoxygen supply to the tumor, might have a previously unrecognized side effect causingenhanced exosome production and secretion, which in turn suppresses the natural antitumorimmune response and thus should be taken into account when designing an optimaltherapy of cancer patients. In conclusion, we describe a novel stress-inducible mechanism shared by placenta andtumors as an immune escape strategy. We found that placenta- and tumor-derived NKG2Dligand-bearing exosomes can suppress immune responses to promote the survival and wellbeing of the fetus or the tumor. Our work comprises an important contribution to theelucidation of the NKG2D ligand-receptor system and its mode of operation in the humanbody and opens new perspectives for designing novel therapies for infertility and cancer.

In mammals, the trophoblast lineage of the embryo is specified before implantation. It is restricted to become the fetal portion of the placenta. We have isolated and cultured trophoblast-derived cells from day 10 and day 13 porcine embryos. These cells demonstrate morphological and biological characteristics that make them unique. We have demonstrated that these cells can grow in vitro in a defined, serum-replacement medium for over a year without showing any signs of senescence. Trophoblast-derived cells placed into serum-containing medium, however, rapidly senesce and fail to proliferate. Gene expression analysis by RT-PCR and Fluidigm analysis of cells in culture from 0-30 days confirmed expression of genes involved in trophoblast function (CDX2, TEAD4, CYP17A1, HSD17B1, FGFR2, PLET, HAND1) as well as some genes known to mediate pluripotency (POU5F1, KLF4, CMYC). These experiments revealed changes in gene expression over time and in response to serum-containing medium. We have demonstrated that these trophoblast-derived cells are easily stably transfected with an exogenous transgene (eGFP) by a variety of methods, and show the ability to survive and to be passaged repeatedly after transfection. Also, immunofluorescence analysis results demonstrated that these cells do not only demonstrate epithelial characteristics by the expression of KRT18, but also they show expression of VIMENTIN which is a protein found in mesenchymal cells. These findings contradict studies done by Ramsoondar in 1993 and Flechon in 1995 which reported the negative expression of VIMENTIN in similar cells. In summary, early embryonic porcine trophoblast-derived cells have demonstrated unique characteristics which have taken us to the conclusion that they could be used as valuable tools for laboratory work. Anticipated applications include the study of trophoblast physiology as well as possible solutions for improving efficiency of transgenesis by somatic cell nuclear transfer and for pluripotency reprogramming of cells.

Preeclampsia and molar pregnancy are two devastating placental pathologies characterized by an immature proliferative trophoblast phenotype accompanied by excessive cell death. It is therefore of paramount importance to study the regulation of cell fate in the placenta, to gain a further understanding of the mechanisms that contribute to these diseases. In this dissertation we report that during normal placental development and in preeclampsia, Matador (Mtd), a pro-apoptotic member of the Bcl-2 family, has a dual function in regulating trophoblast cell proliferation and death. Importantly, we reveal a novel role of Mtd-L in promoting cyclin E1 expression and cell cycle progression. Of clinical importance, we also identify that both cyclin E1 and the CDK inhibitor p27, are increased in severe early onset preeclampsia. However, the inhibitory function of p27 in this pathology may be hampered due to its increased phosphorylation at Ser10, resulting in its nuclear export. Of equal importance, data presented demonstrate that placentae from severe early onset preeclampsia display a molecular profile distinct from late onset preeclampsia or intrauterine growth restricted pregnancies. In the final data chapter we demonstrate that Mtd is highly expressed in molar tissue, where it localizes to both apoptotic and proliferative cells. Our data suggests that an abundance of Mtd and cyclin E1 in conjunction with the low level of p27 may contribute to the hyperproliferative nature of the disorder. The body of work in this dissertation uncovers novel insights into the regulation of trophoblast cell fate. Importantly, the impact of Mtd on cyclin E1 to promote G1-S transition is a novel mechanism found to regulate trophoblast cell proliferation in normal and pathological placentation. Equally important is our identification of molecular differences between placental pathologies that may help to differentiate early and late onset preeclampsia, IUGR and molar pregnancy.

Preeclampsia and molar pregnancy are two devastating placental pathologies characterized by an immature proliferative trophoblast phenotype accompanied by excessive cell death. It is therefore of paramount importance to study the regulation of cell fate in the placenta, to gain a further understanding of the mechanisms that contribute to these diseases. In this dissertation we report that during normal placental development and in preeclampsia, Matador (Mtd), a pro-apoptotic member of the Bcl-2 family, has a dual function in regulating trophoblast cell proliferation and death. Importantly, we reveal a novel role of Mtd-L in promoting cyclin E1 expression and cell cycle progression. Of clinical importance, we also identify that both cyclin E1 and the CDK inhibitor p27, are increased in severe early onset preeclampsia. However, the inhibitory function of p27 in this pathology may be hampered due to its increased phosphorylation at Ser10, resulting in its nuclear export. Of equal importance, data presented demonstrate that placentae from severe early onset preeclampsia display a molecular profile distinct from late onset preeclampsia or intrauterine growth restricted pregnancies. In the final data chapter we demonstrate that Mtd is highly expressed in molar tissue, where it localizes to both apoptotic and proliferative cells. Our data suggests that an abundance of Mtd and cyclin E1 in conjunction with the low level of p27 may contribute to the hyperproliferative nature of the disorder. The body of work in this dissertation uncovers novel insights into the regulation of trophoblast cell fate. Importantly, the impact of Mtd on cyclin E1 to promote G1-S transition is a novel mechanism found to regulate trophoblast cell proliferation in normal and pathological placentation. Equally important is our identification of molecular differences between placental pathologies that may help to differentiate early and late onset preeclampsia, IUGR and molar pregnancy.

Galectins are a family of secreted animal lectins with a high affinity to betagalactosides commonly involved in cellular functions such as apoptosis, adhesion and migration. Galectin 15 (LGALS15), a newest member of the galectin superfamily, has a unique C-terminal RGD sequence and participates in integrin-mediated ovine trophectoderm cell attachment and migration. In the ovine uterus, LGALS15 is expressed only by the endometrial luminal (LE) and superficial glandular (sGE) epithelia, induced by progesterone between Days 10 and 12 of the cycle and pregnancy, and then stimulated by interferon tau (IFNT) from the conceptus after Day 14 of pregnancy. During early pregnancy, the canonical janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway is not active in the endometrial LE/sGE. Therefore, IFNT may utilizes a non-canonical signaling pathway to increase transcription of genes, including CST3, CTSL, HIF2A, LGALS15, and WNT7A, specifically in the endometrial LE/sGE. Alternatively, IFNT and progesterone could indirectly affect epithelial gene expression by influencing gene expression in the stroma, which then communicates with the epithelium. Although the LGALS15 gene is present in ovine, caprine and bovine species, it is only expressed in uteri of sheep and goats. Available data shows a tissue- and speciesspecific expression pattern for LGALS15, likely involving multiple layers of transcription regulation in the ruminant endometrium. Further analysis of the LGALS15 5? promoter/enhancer region revealed similar predicted transcription factor binding sites in all three species, including; PU.1, Ets-1, AP1, Sp1, and GRE or PRE sites. Interestingly, the proximal promoter region of the LGALS15 gene in all three species exhibited a conserved Sp1 binding site upstream of an AP1 binding site on both sense and antisense strands, and with similar spacing between binding sites. Sequence analysis revealed key differences in LGALS15 gene structure between ruminant species including the proximity of repetitive DNA sequences to the transcription start site (+1). Bovine LGALS15 has repetitive DNA sequences start at - 145 whereas in ovine or caprine LGALS15 it starts at about -300. The length of the repetitive DNA sequence is similar (~1.2 kb) in the 5' promoter/enhancer region of LGALS15 in all three species. Transient transfection analyses found that repetitive DNA sequences reduced basal promoter activity and responsiveness to treatments. None of the promoter construct showed responsiveness to interferon tau (IFNT). The bovine LGALS15 gene promoter showed no activity under any experimental conditions. The current studies indicate that uterine LGALS15 is expressed in ovine and caprine but not bovine species. Additionally, repetitive DNA sequences found in the promoter region may contribute to modulating the LGALS15 gene expression. Therefore, the ruminant LGALS15 gene, like other galectins, is under tight transcriptional control involving hormones, requisite transcription factors and potentially chromatin remodeling complexes working synergistically for LGALS15 promoter transactivation.

[Truncated abstract] The placenta forms a barrier regulating the transfer of gases, nutrients and wastes between the mother and the developing conceptus, and also produces hormones affecting both the fetus and the mother. This barrier is formed by the differentiation of the outer layer of the blastocyst- the trophoblast- to facilitate implantation and subsequent invasion of the uterus. The trophoblast consists of an underlying proliferative pool of cytotrophoblasts, which differentiate to replenish the overlying continuous, multi-nucleated syncytiotrophoblast that forms the barrier between the mother and fetus. Moreover, the location of the syncytiotrophoblast directly in contact with the maternal circulation suggests an endothelial role for the trophoblast regulating blood flow, thrombosis and immune cell adhesion. Disruption to the function of the human trophoblast may result in preeclampsia, a maternally manifested disorder of pregnancy characterised by hypertension and proteinurea. Blood flow to preeclamptic placentae is reduced and the cytotrophoblast pool is diminished; however the exact cause (or causes) remains elusive. Many potential causes are hypothesised, including endothelial damage, premature remodelling of maternal spiral arteries, increased oxidative stress and impaired trophoblast differentiation and apoptosis. Caspase-14 is an unusual caspase in that it is not involved in apoptosis. Furthermore, it possesses a limited, predominantly epithelial, tissue distribution. In the epidermis, caspase-14 is expressed in the apical differentiating layers. Here it cleaves profilaggrin to stabilise intracellular keratin intermediate filaments, and indirectly provides natural hydration and UV protection to the corneocytes. Thus, caspase-14 is vital to the maintenance of the barrier function of the skin. ... As differentiation-associated genes were elevated in the absence of caspase-14, this implies that caspase-14 suppresses biochemical trophoblast differentiation. The cytoskeletal keratin network was also examined following RNA Interference. The synthesis of cytokeratin 18 was significantly enhanced after caspase-14 suppression during BeWo differentiation, linking caspase-14 with keratin homeostasis. Therefore caspase-14 suppresses trophoblast differentiation, potentially through modulation of the cytoskeletal keratin filament network. The precise mechanism remains to be elucidated, however the identification of pathways regulated by caspase-14 advances our knowledge of trophoblast differentiation and potential causes of disorders of pregnancy. In summary, caspase-14 appears to be involved in the suppression of differentiation in the human trophoblast. As disorders of pregnancy such as preeclampsia often feature disturbed differentiation and a diminished cytotrophoblast pool, a greater understanding of caspase-14 biology in the human placenta could lead potential therapies for various disorders of pregnancy.

This thesis investigates the role of Crim1, a transmembrane protein that is expressed in a number of areas in the vertebrate embryo including the developing kidney, eye, testis and spinal cord, which we believe may be a regulator of vertebrate tissue development. To dissect the function of Crim1 in normal mammalian development, two vertebrate models were used, zebrafish and mice. The results show that in zebrafish, crim1 is expressed early in development from the 16-cell stage through to 30 hours post fertilisation (Chapter 3). At 24 hours post fertilisation crim1 is expressed in the intermediate cell mass (icm), the site of haemangioblast development. Haemangioblasts are precursor cells that contribute to the formation of the blood and endothelial cell lineages. Injection of crim1 antisense oligonucleotides into zebrafish embryos (crim1 morphants) lead to an expansion of the icm and defects in the trunk, tail, somites and vasculature. The injection of crim1 antisense oligonucleotides into transgenic fli:GFP zebrafish revealed defects in the intersegmental, dorsal longitudinal anastomotic and parachordal vessels. Although crim1 is expressed during haemagiogensis the primary defect in the crim1 morphant zebrafish appears to be vascular. Further experiments used a ‘knock-in’ mouse, Crim1KST264, in which a loss of functional Crim1 leads to defects in limb (syndactyly), skeleton, eye, vascular, kidney and placental development. Analysis of the kidney phenotype in the embryonic Crim1KST264 homozygotes showed that a loss of Crim1 affects ERK1/2 and phosphorylated-Smad1/5/8 protein expression, although has no direct effect on BMP or TGFβ protein expression (Chapter 4). Analysis of the adult Crim1 outbred kidneys revealed they have albuminuria and leaky vasculature. The complex phenotype presented by the Crim1KST264 homozygote kidneys suggests Crim1 may be regulating multiple growth factor pathways. As Crim1 was shown to be expressed in the placenta, we characterised the role of Crim1 in placental development using the Crim1KST264 mouse (Chapter 5). Crim1KST264 homozygote placentas and embryos are smaller than their wild-type littermates. Our investigations revealed that Crim1 is expressed in trophoblast giant cells and in spongiotrophoblasts. A loss of Crim1 causes a developmental defect in that the junctional zone (region of the placenta containing spongiotrophoblasts and glycogen cells) is expanded, although this phenotype does not appear to be due to a defect in proliferation or apoptosis. Further analysis of E15.5 Crim1KST264 homozygote placentas revealed there was a reduction in the number of labyrinth trophoblast gaint cells. Thus, by using zebrafish and mouse as two model organisms of vertebrate development, this thesis has showed that Crim1 is clearly important for normal embryonic development. To dissect the complex phenotype presented by the Crim1KST264 mouse, further studies of Crim1 and its interaction with other growth factor pathways is needed to elucidate how and to what extent they interact with Crim1 to determine its biological effect on vertebrate tissue.

Thesis (Ph.D.)--Memorial University of Newfoundland, Faculty of Medicine, 1997. / Typescript. Bibliography: leaves [222]-256.

Thesis (Ph. D.)--Ohio State University, 2007. / Full text release at OhioLINK's ETD Center delayed at author's request