|Home | About | Journals | Submit | Contact Us | Français|
Ex vivo perfusion systems offer a reliable, reproducible method for studying acute physiological responses of an organ to various environmental manipulations. unlike in vitro culture systems, the cellular organization, compartmentalization and three-dimensional structure of ex vivo–perfused organs are maintained. these particular parameters are crucial for the normal physiological function of the placenta, which supports fetal growth through transplacental exchange, nutritional synthesis and metabolism, growth factor promotion and regulation of both maternally and fetally derived molecules. the perfusion system described here, which can be completed in 4–5 h, allows for integrated, physiological studies of de novo synthesis and metabolism and transport of materials across the live mouse placenta, not only throughout a normal gestation period but also following a variety of individual or combined genetic and environmental perturbations compromising placental function.
Organ perfusion systems can simulate physiological conditions, providing a framework for the study of numerous biochemical functions, including metabolism, membrane transfer, drug transport and delivery, toxicological studies and drug efficacy studies1–28. Ex vivo perfusion of isolated organs in particular allows for precise control of the physiological environment and can provide a unique opportunity for studying organ-specific physiological responses to environmental factors with a resolution and reproducibility that may not be possible in intact animals1–3,8–12,18,20,22,24–26,29,30. In this article, we describe the procedure used in our laboratory for the dissection and perfusion of mid-to-late-gestation mouse placentas.
To investigate directly the impact of placental function on fetal brain development9,31,32, we sought to devise a mouse placenta perfusion system allowing for transport and metabolic studies throughout gestation. This model can be used in a wide array of integrated physiological studies, such as assessing changes in placental de novo synthesis, molecular transport and metabolism of various compounds throughout normal gestation and amidst manipulations of genetic and environmental conditions. We recently used this method to demonstrate that a fraction of the maternal delivery of the essential amino acid tryptophan is enzymatically converted to serotonin within the placenta; the neo-synthesized placental serotonin is then released into the fetal blood stream through the umbilical vein, from where it can reach the fetal forebrain9.
The method described here allows for perfusions of live mouse placentas in early (embryonic day (E)12–14) and late pregnancy (E16–18), providing a framework for directly comparing physiology at different stages of pregnancy. Furthermore, the use of the mouse placenta as a model allows for direct investigation of the effect of a variety of perturbations, individually or in combination, such as genetic deletions, inflammation or maternal drug exposure on placental physiology. For instance, the ex vivo placenta perfusion system, combined with the application of transgenic mouse models such as trophoblastic or syncytiotrophoblastic-specific gene deletions33–36, allows for the unique opportunity to explore cellular and biochemical processes underlying the impact of environmental or maternal perturbations on placental function throughout gestation. Another potential application of the ex vivo perfusion system is the real-time, simultaneous visualization of blood flow and transport or metabolism of fluorescently labeled molecules across the live placenta. Our preliminary studies indicate that the ex vivo perfusion apparatus, coupled with a two-photon imaging system, can be used to monitor maternal and fetal blood flows simultaneously by infusing fluorescently labeled dextran molecules through the uterine and/or umbilical arteries (Supplementary Fig. 1 and Supplementary Video 1; N.G., J. Burford, J. Peti-Peterdi and A.B., unpublished data). This approach can also be used for in vivo visualization of cell dynamics in the mouse placenta37. Although this system has been used primarily for placental perfusions in our laboratory, the application of this method to other organs or tissue systems can be readily accomplished, pending that target organs have a simple path of blood irrigation (e.g., simple arterial-venous input/output).
Ever since Panigel et al.4,13 demonstrated the first dual recirculation placental perfusion model in 1967 (ref. 14), human term placental perfusion studies have been preferred for pathophysiology, metabolism and developmental pharmacology studies. There are several benefits to conducting postpartum placental studies in humans. Perhaps most notably, the use of postpartum tissue precludes ethical concerns, as both the isolation and perfusion of placental tissue are noninvasive with regard to both the mother and the newborn. Furthermore, in vivo transplacental transfer data obtained by comparing maternal blood concentrations with fetal cord blood concentrations of a given compound at term make postpartum placental perfusion studies directly comparable to in vivo measures.
However, the drawback of using term placenta is that mid-gestation and term placentas are very different in terms of transport and metabolic capacities, as suggested by reports of differential gene expression38–40. The transfer and regulation of compounds that enter the fetal blood supply are controlled by at least four distinct mechanisms, ranging from passive and active transmembranous transport to enzymatic metabolization and phagocytosis41–46. At the molecular level, changes in expression of transporter genes responsible for active transport of nutrients during pregnancy leads to a dynamic state of maternofetal and feto-maternal transfer throughout gestation47–49. There are also physical differences in the barrier structure that emerge during gestation; for instance, in the human placenta the syncytiotrophoblastic cell layer is ~20 μm thick during the first trimester, and it progressively thins during gestation down to ~5 μm in thickness at term50–53. This change in thickness, coupled with an increase in the number of fetal capillaries, may also contribute to the dynamic state of transplacental transfer that changes continuously throughout gestation54.
Thus, very little is known about the transplacental transfer and metabolism of molecules at earlier stages of human gestation. However, this knowledge is crucial if we are to better understand the mechanisms underlying fetal programming of adult diseases. Indeed, it is now becoming evident that several metabolic, cardiovascular and even psychiatric disorders that manifest in adulthood are linked to alterations of maternal-fetal interactions and fetal development that occur at specific periods of gestation31,55–63. The existence of time-sensitive windows during pregnancy for the development of specific mental disorders in the offspring suggests that alterations of maternal homeostatic conditions during early-to-mid gestation can have profound consequences on fetal brain development.
Another common method for studying placental physiology is in vitro culture of isolated placental cell types (e.g., trophoblasts). Although in vitro explant or dissociated cell culture systems allow for studies of transport and metabolic mechanisms at the molecular and cellular level, they lack the cellular organization, compartmentalization and three-dimensional structure of intact, physiologically active placentas. These parameters are particularly important for the normal function of the placenta39,54,64,65.
Our ex vivo perfusion system can provide a direct validation of the role of cellular and molecular pathways initially characterized in vitro. Therefore, we see the ex vivo perfusion method as a complement to current in vitro methods. The overall principle of the method described is comparable to human placental perfusion systems, adapted for an organ of much smaller size. Because of key interspecies differences66,67, results obtained in the mouse model must always be carefully extrapolated to human placental physiology. Initial studies in our laboratory were devoted to determining the effect of perfusion rate, incubation temperature, incubation medium composition and various surgical techniques on the viability of the perfused organ. Below we present the optimized protocol that we are now using for the perfusion of mid-to-late-gestation mouse placentas. Despite the straightforward appearance of the techniques, some hands-on experience and fine motor control are necessary to obtain accurate and reproducible results. The incubation and infusing media should be optimized for each individual experiment, as some compositions may not be suitable for certain metabolic studies.
In our experience and in that of others, it is essential to begin perfusing with fresh, viable tissue, as ischemic cell death progresses rapidly after mice are killed68–70. This implies that organs should be harvested as quickly as possible from the animal and transferred directly into a thermodynamically controlled incubation chamber that has been prefilled with oxygenated incubation medium. It should be noted that different tissues and organ systems require specific storage and incubation media71, and that each experiment should be optimized for the individual needs of the organ in question. In addition, organ viability should be monitored during and after perfusion to ensure that the specific medium and perfusion parameters meet the organ’s metabolic needs. Although the surgical and perfusion procedures are still being perfected, extra tissue to be used for cannulation can be kept on ice to limit the harmful effects of warm ischemia. However, in our experience, exposure to low temperatures increases vascular wall stiffness, which can impede the cannulation process. For this reason, we recommend the use of fresh tissue whenever possible.
Because of the small diameter of the vasculature of mid-to-late-gestation mouse placentas, the catheters used in this perfusion system are terminated with a polyimide-coated fused silica capillary tube, available in various sizes between 90 and 200 μm in external diameter. The optimal diameter of the catheter depends on the size of the vasculature in question, as well as on the operator preference. Although this perfusion system is optimized for use with E14 placenta, the system can be easily adapted to any mid-to-late-gestation mouse placenta simply by increasing or decreasing catheter size within this 90–200-μm range. Although smaller-diameter tubing may appear more straightforward for the cannulation of micro-vasculature, tubing diameter is directly proportional to rigidity, and overly pliable catheters can be more troublesome than their larger counterparts. It is therefore our opinion that a catheter of just slightly smaller diameter than the vasculature in question should be chosen, as it provides the best compromise between rigidity and size.
Although the choice of medium is not completely evidence based, most researchers use culture or incubation media such as Dulbecco’s modified Eagle’s medium (DMEM) or Krebs-Ringer saturated with 95% O2/5% CO2 (refs. 9, 12, 19 and 27). In our experience, this oxygen saturation is sufficient to maintain placental health for the duration of the experiment (ANTICIPATED RESULTS). Perfusion studies conducted under relative hypoxia should be possible by changing the oxygen saturation of perfusion and incubation media. Specific medium may not be suitable to every experiment, and any medium that is capable of sustaining a viable organ throughout the duration of the experiment is suitable.
A variety of buffering solutions may be used to simulate maternal blood flow for the perfusion system. Freshly prepare buffer solutions and maternal and fetal input solutions on the day of the experiment using ultrapure water, and then store the solutions at 4 °C until use. Before perfusion, warm input solutions to above 40 °C to promote degassing (see PROCEDURE Step 4), and then oxygenate the incubation buffer using 95% O2/5% CO2. The buffers we use are as follows: PBS, composed of 10 mM Na2HPO4, 2.7 mM KCl and 137 mM NaCl; HEPES, composed of 140 mM NaCl, 4 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 10 mM HEPES and 10 mM glucose; and DMEM (see Reagents).
Load the BubbleStop syringe warmer with 60-ml syringes and connect it to the Teflon valve box with 1/16th-inch Teflon tubing (Fig. 1a,b). Connect the valve box to the ValveLink 8.2 software and ensure that it is controlled by the software. Connect the output of the Teflon valve box to 1/16-inch Teflon tubing, stepped down to 0.033-inch Micro-Renathane tubing, and then seal it with cyanoacrylate. Connect this Micro-Renathane tubing to the low-flow peristaltic pump, which delivers the simulated maternal input to the placenta. Bubble the incubation medium with 95% O2/5% CO2 through 3/4-inch PVC tubing connected to a bubbling stone, and then pump through the medium-flow peristaltic pump into the in-line solution heater and finally into the incubation chamber. Use the TC324B temperature controller to control the in-line solution heater and the incubation dish to maintain a steady state of 37 °C. Stimulate the fetal circulation by infusing input medium from the BubbleStop syringe warmer through the umbilical artery via an ultralow-flow pump. Collect the flow through the umbilical vein by a second ultralow-flow peristaltic pump and deliver it to a series of collection tubes. The low- and ultralow-flow peristaltic pumps may be controlled using a computer interface through the use of a USB controller.
Load the low-flow and ultralow- flow peristaltic pumps (Fig. 1c) with C-Flex tubing sets, 0.031-inch i.d. and 0.020-inch i.d., respectively. Connect Micro-Renathane tubing to the C-Flex tubing sets and seal them with cyanoacrylate. Terminate this tubing with one of four sizes of polyamide-coated fused silica capillaries and seal it with cyanoacrylate.
Typically, a mouse placenta at mid-to-late gestation is capable of sustaining flow in this ex vivo system for a minimum of 120 min before vascular collapse leads to reduced output from the umbilical vein. Several parameters are monitored to assay organ viability throughout perfusion; however, as noted earlier for human term placenta16, there is no standard set of parameters, and thus viability measures vary between laboratories. Common parameters include fetal circulation pressure, eluate pH, oxygen consumption, fetal oxygen transfer, glucose consumption, lactate production, the synthesis and secretion of proteins and cell death assays. Some indications point to fetal volume loss being the optimal measure of tissue viability and integrity20. However, there is no generally accepted maximum allowable volume loss and these measures vary across species. In our experience, fetal volume loss and vascular collapse precedes cell death, but additional viability measures should still be performed; therefore, we used a combination of placental lactate dehydrogenase (LDH) activity, fetal eluate loss and cell apoptosis assays as organ viability controls.
Fetal eluate volume should be monitored throughout the perfusion to confirm that no major volume loss is present. In our experience, an input of 5 μl min −1 through the umbilical artery of a healthy E14 placenta elicits a flow rate of 4.5 μl min −1 through the umbilical vein directly after euthanasia, with no more than an average fetal volume loss of 20% over a 90-min perfusion (supplementary Fig. 3a). Consequently, perfusions yielding fetal volume loss in excess of 20% should be discarded.
In normal tissues, hypoxia, tissue damage and cell toxicity have been shown to increase metabolic activities, such as glycolysis76–78. Hypoxic conditions in term human placenta, such as those caused by incomplete perfusion, have been shown to enhance glycolysis and increase the activity of LDH, which converts pyruvate to lactate76. Quantification of LDH activity can therefore be used as a reliable measure of cell toxicity throughout placental perfusions. LDH assays can be performed using a colorimetric assay kit (Sigma, cat. no. MAK066) that quantifies placental production of NADH, an index of glucose conversion to pyruvate and thus of placental LDH activity. Results show that in our conditions placental LDH activity remains low and stable throughout the perfusion (supplementary Fig. 3b), until 90 min after euthanasia. When we adjusted experimental parameters to allow for an increased perfusion time of 120 min, we observed a threefold increase in LDH activity, thus confirming that our experimental parameters are optimized for 90-min perfusions. Increased placental LDH activity indicates that cellular toxicity has commenced and the organ should be discarded.
Caspase-3, a member of the cysteine-aspartic acid protease (caspase) family, which is sequentially activated during the execution phase of cell apoptosis79–82, can be used after perfusion as a marker for cell apoptosis. We stained fresh placentas, as well as perfused and unperfused placentas at 40, 80 and 120 min after euthanasia, for activated caspase-3 (Millipore, cat. no. AB3623) as previously described83. We designated a region of interest within the decidua (0.6 mm2) across the different time points and conditions, and we performed cell counting across three sections for each time point under the criteria of colocalized activated caspase-3 and DAPI staining (supplementary Fig. 3c–l). In accordance with the lactate production assays, activated caspase-3 staining indicates that generalized cell apoptosis does not commence until well after completion of the perfusion.
This work was supported by the National Institute of Child Health and Human Development (NICHD) (grant 5R21HD065287 to A.B.) and a NARSAD (National Alliance for Research on Schizophrenia and Depression; now the Brain and Behavior Research Foundation) Young Investigator award (to A.B.). We thank P. Levitt for his support during the initial development of this protocol. We acknowledge J. Burford and J. Peti-Peterdi for their valuable contributions in two-photon live imaging.
Note: Supplementary information is available in the online version of the paper.
AUTHOR CONTRIBUTIONS N.G. conducted the experiments. N.G. and A.B. conceived the protocol, interpreted the data and wrote the manuscript.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.