Who Let the Dogs In: A Canine Trophoblast Invasion Model for Pre-eclampsia

Authors


Author's address (for correspondence): M Kutzler, 112 Withycombe Hall, Oregon State University, Corvallis, OR 97331, USA. E-mail: michelle.kutzler@oregonstate.edu

Contents

Pre-eclampsia affects 2–8% of pregnant women worldwide and is the third leading cause of maternal mortality in the United States, accounting for 20% of maternal deaths, for which the only known cure is delivery of the placenta. It is known that pre-eclampsia results from defects within the trophoblast invasion of the endometrium and myometrium. At a morphological level within the pre-eclamptic human placenta, trophoblast invasion is shallow, and this results in hypoperfusion, which is a life-threatening condition for both the mother and the foetus. Pre-eclampsia has been intensively investigated for over 50 years, and yet the causes are largely unknown. Despite a large body of data, it is still unknown exactly which mechanisms regulate trophoblast invasion. An effective animal model may be crucial to understanding the underlying causes of pre-eclampsia. A canine model is a proposed improvement on the current efforts to investigate disorders of shallow trophoblast invasion throughout gestation and to improve understanding of the factors that regulate trophoblast invasion. The objectives of this research were to elucidate and compare cellular and molecular similarities between normal canine trophoblasts with those from recently published reports on pre-eclampsia in women.

Introduction

Pre-eclampsia affects 2–8% of pregnant women worldwide (Young et al. 2010), is the third leading cause of maternal mortality in the United States, and accounts for 20% of maternal deaths, (MacKay et al. 2001). Currently, the only known cure of pre-eclampsia is delivery of the placenta (Conde-Agudelo et al. 2004; Högberg 2005). Pre-eclampsia arises from defective placental development because of shallow trophoblast invasion into the endometrium and myometrium resulting in insufficient remodelling of the spiral arteries. Pre-eclampsia has been intensively investigated for over 50 years (Stedge and Gent 1954); yet the causes are largely unknown and the effectiveness of current models has been limited.

An effective animal model may be crucial to understanding the underlying causes of pre-eclampsia. Although pre-eclampsia can occur in higher apes, it does not occur spontaneously in most non-human primates. Other animal models have failed to be effective in the study of pre-eclampsia for a number of reasons. Rodents have been commonly used to study the induction of clinical signs associated with pre-eclampsia through transgenic and knockout studies (e.g. high blood pressure, proteinuria) (Carter 2007; McCarthy et al. 2011), but these models have proven to be less useful in studying shallow trophoblast invasion. To our knowledge, no existing animal model has yet been able to demonstrate the shallow trophoblast invasion aspect observed in pre-eclampsia.

A larger, non-rodent animal model can be more easily manipulated to demonstrate changes in morphology, histochemistry and gene expression throughout pregnancy. Compared with other domestic animal models (e.g. sheep, pig) used to study human pregnancy-related disorders, the canine placenta is significantly more invasive making it potentially more relevant to the invasive human placenta. It is interesting to note that the morphologic and histologic similarities between normal canine trophoblast invasion and that of the pre-eclamptic human trophoblast invasion are striking. In both types of placentation, trophoblasts invade the endometrium (uterine epithelium) and the endometrial stroma (decidua) but do not completely invade the myometrium (Anderson 1969; Wynn and Corbett 1969; Fox and Agrafojo-Blanco 1974; De Wolf et al. 1980, 1982; Fernández et al. 1998; Grether et al. 1998; Stoffel et al. 1998). A second similarity of both forms of placentation is that transformation of the endometrium (decidualization) may play a role in regulation of trophoblast invasion.

For an animal model of pre-eclampsia to be feasible, it should closely resemble changes that occur in the human form of the disorder. Therefore, it should demonstrate (i) shallow trophoblast invasion in vivo (which the canine model does naturally); (ii) changes in trophoblast invasive properties from early pregnancy to term; (iii) changes in substances known to regulate trophoblast invasion [e.g. matrix metalloproteinases (MMPs)] from early pregnancy to term; and (iv) changes that are reversible using promoters of trophoblast invasion. This hitherto unused canine model may be a useful improvement on the current efforts to investigate disorders of shallow trophoblast invasion throughout gestation and to improve our understanding of the factors that regulate trophoblast invasion.

The central hypothesis is that canine trophoblasts display several cellular and molecular similarities to human pre-eclamptic trophoblasts. The long-range goal of the proposed research is that with a canine model, new treatments (and possibly preventions) for the underlying cause of pre-eclampsia (e.g. shallow trophoblast invasion) can be developed.

Materials and Methods

Six pregnant dogs were used for this research. Pregnant dogs were either housed at the Veterinary Medical Animal Isolation Laboratory at Oregon State University or with their owners. Late gestation (60–65 days past the LH surge) placental tissue was recovered via hysterotomy (n = 3) or at delivery (n = 3). The experiments focused on trophoblast cellular behaviour and gene expression near the end of gestation, which is important for comparison with published human studies where placentas are collected from the third trimester after C-section or delivery.

Trophoblasts were isolated using a series of enzymatic digestion and Percoll density centrifugation as previously described for human placentas (Aboagye-Mathiesen et al. 1996; Bloxam et al. 1997; Tanaka et al. 1998; Blaschitz et al. 2000; Frank et al. 2000; Genbacev and Miller 2000; Frank et al. 2001; Pötgens et al. 2001; Chakraborty et al. 2002; Hunkapiller and Fisher 2008; Guibourdenche et al. 2009).

Cells were plated on 22-mm2 coverslips and cultured in standard cell culture atmospheric conditions (38 °C, 5% CO2) in culture media that were replenished every 48 h until 70–80% confluence at which time coverslips were fixed in 70% methanol for 24 h at 4°C. Cellular morphology was monitored daily. Once fixed, cells were fluorescently immunostained for cytokeratin-7 (no. p103620; DAKO, Carpinteria, CA, USA) at a 1 : 200 dilution using the secondary antibody Alexa Flour 488 (no. A21202; Invitrogen, Carlsbad, CA, USA) at a 1 : 1000 dilution. Hoescht 33342 (no. H1399; Invitrogen) was used to count the number of immunopositive trophoblasts. Cytokeratin 7 (CK7) is a type II cytokeratin that has been shown to specifically label human trophoblasts (Aboagye-Mathiesen et al. 1996; Bloxam et al. 1997; Tanaka et al. 1998; Blaschitz et al. 2000; Frank et al. 2000; Genbacev and Miller 2000; Frank et al. 2001; Pötgens et al. 2001; Chakraborty et al. 2002; Hunkapiller and Fisher 2008; Guibourdenche et al. 2009). Canine trophoblasts were also fluorescently immunostained for matrix metalloproteinase-2 (MMP2) (MS-806-P0; Thermo Scientific, Fremont, CA, USA) and matrix metalloproteinase-9 (MMP9) (RB-1539-P0; Thermo Scientific) at a dilution of 1 : 100, and per cent immunopositive cells were determined. All of the cell culture experiments were run in duplicate.

For molecular analysis, placental tissues were flash frozen in liquid nitrogen and stored at −80°C. Total RNA was extracted from each placenta using the TRIzol Plus RNA Purification Kit (Invitrogen) according to the manufacturer's instructions. RNA quality and quantity were assessed using a NanoPhotometer (IMPLEN, Munich, Germany). RNA integrity was confirmed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) at the Oregon State University Center for Genome Research and Biocomputing. Total RNA was reverse transcribed to cDNA using the SuperScript III System (Invitrogen) according to the manufacturer's instructions. Real-time PCR (RT-PCR) studies were carried out using TaqMan gene expression assay kits (Applied Biosystems) according to the manufacturer's instructions on a 7300 RT-PCR machine (Applied Biosystems) according to the equipment manual for a two-step RT-PCR procedure. The primer and probe sequences for 18S rRNA, MMP2 and MMP9 used for RT-PCR studies were pre-made and pre-optimized for canine tissue by Applied Biosystems. Gene expression was normalized to mRNA expression of 18S rRNA in each sample. The first parturient sample collected was used as a calibrator sample. Relative expression was calculated using the 2−ΔΔCt method. All RT-PCR samples were performed in triplicate.

Statistical analyses

Data are represented as mean ± standard deviation (SD). MMP2 and MMP9 protein and RNA expression were compared by Student's t-test. Statistical computer software (GraphPad Prism version 5.0 software, San Diego, CA, USA) was used for all comparisons. Significance was defined as p < 0.05.

Results

Isolated canine trophoblasts were large polygonal cells that arranged themselves into a cobblestone configuration (Fig. 1). In addition, >97% canine trophoblasts were immunopositive for CK7. Both MMP2 and MMP9 were expressed in late gestation canine placentas. More cells were immunopositive for MMP9 (54.7 ± 3.4%) compared with MMP2 (40.3 ± 1.8%; Fig. 2). In addition, MMP9 gene expression was significantly higher than MMP2 (Fig. 2). 18S rRNA was stable in all of the samples tested (data not shown).

Figure 1.

Isolated canine trophoblasts are large polygonal cells that arranged themselves into a cobblestone configuration

Figure 2.

Both MMP2 and MMP9 were expressed in late gestation canine placentas. Using both immunocytochemistry and RT-PCR methods, mean ± SD MMP9 protein and gene expression was significantly higher than MMP2, respectively (p < 0.05)

Discussion

Isolated canine trophoblasts display similar morphology to that of human trophoblasts (e.g. large polygonal cells that arrange themselves into a cobblestone configuration). More than 97% of canine trophoblasts were positive for CK7. This was consistent with results from previous studies on isolated human trophoblasts using similar methods validating this method for canine trophoblast cell isolation (Aboagye-Mathiesen et al. 1996; Bloxam et al. 1997; Tanaka et al. 1998; Blaschitz et al. 2000; Frank et al. 2000; Genbacev and Miller 2000; Frank et al. 2001; Pötgens et al. 2001; Chakraborty et al. 2002; Hunkapiller and Fisher 2008; Guibourdenche et al. 2009).

The invasion and tissue remodelling associated with placentation is mediated in part by the action of MMPs, a family of proteolytic enzymes which together can breakdown all components of the extracellular matrix (Birkedal-Hansen et al. 1993). Numerous studies have demonstrated that human cytotrophoblast production of various matrix metalloproteinases (MMPs) is essential for the penetrative ability of trophoblasts in vitro and in vivo (Aplin 1991; Librach et al. 1991; Strickland and Richards 1992; Lim et al. 1997; Campbell et al. 2003; Lian et al. 2010). Abnormally low concentrations of circulating MMP2 and MMP9 are reported to correlate with shallow trophoblast invasion during pre-eclampsia (Shimonovitz et al. 1998; Palei et al. 2008). Similar to what is reported in other cases of shallow trophoblast invasion (e.g. pre-eclampsia), (Sood et al. 2006; Lian et al. 2010; Mayor-Lynn et al. 2011; Zhang et al. 2011), we expected that MMP2 and MMP9 would be expressed in isolated canine trophoblasts and canine placental tissues. Using RT-PCR, Lei et al. (2007) reported an upregulation in MMP9 expression that was associated with differentiation of mouse trophoblast stem cells expected to occur at the end of pregnancy. It has also been shown that progesterone downregulates MMP9 in human trophoblasts during early pregnancy (Cockle et al. 2007). Because progesterone concentrations were rapidly declining at the end of gestation when these canine placental tissues were collected, it is likely that the withdrawal of progesterone allowed for an upregulation of MMP9. An increase in both protein and gene expression of MMP9 was seen in comparison with MMP2 in isolated canine trophoblasts and canine placental tissues, respectively.

Conclusion

Through multiple approaches, these data provide evidence for the feasibility and validity of a canine shallow trophoblast invasion model (CSTIM). This model may be an improvement on the current efforts to investigate disorders of shallow trophoblast invasion throughout gestation and to improve understanding of the factors that regulate trophoblast invasion.

Acknowledgements

The authors thank Dr. Timothy Hazzard and Shaundra Epperson for animal care.

Conflicts of interest

None of the authors have any conflicts of interest to declare.

Funding sources

This study was funded by the Department of Animal Sciences at Oregon State University.

Author contributions

Kutzler: provided oversight and arranged funding for all experiments, wrote manuscript; Sahlfeld: performed cell culture studies, contributed to literature review, edited manuscript; Fellows: performed molecular biology studies, contributed to literature review, edited manuscript.

Ancillary