Ex vivo lungs are useful for a variety of experiments to collect physiological data while excluding the confounding variables of in vivo experiments. Commercial setups are often expensive and limited in the types of data they can collect. We describe a method for building a fully modular setup, adaptable for various study designs.
Ex vivo lung preparations are a useful model that can be translated to many different fields of research, complementing corresponding in vivo and in vitro models. Laboratories wishing to use isolated lungs need to be aware of important steps and inherent challenges to establish a setup that is affordable, reliable, and that can be easily adapted to fit the topic of interest. This paper describes a DIY (do it yourself) model for ex vivo rat lung ventilation and perfusion to study drug and gas effects on pulmonary vascular tone, independent of changes in cardiac output. Creating this model includes a) the design and construction of the apparatus, and b) the lung isolation procedure. This model results in a setup that is more cost-effective than commercial alternatives and yet modular enough to adapt to changes in specific research questions. Various obstacles had to be resolved to ensure a consistent model that is capable of being used for a variety of different research topics. Once established, this model has proven to be highly adaptable to different questions and can easily be altered for different fields of study.
Ex vivo lung perfusion (EVLP) techniques1 have seen a rise in usage in the past decade as a means of studying lung transplantations2, ischemia/reperfusion3, lung metabolism4, and immune responses5. Isolated, but intact, ventilated and perfused lungs offer the critically important ability to directly assess the response of the lungs, including the pulmonary vasculature, to potential interventions and/or therapeutics without potential confounders, such as neuronal and hormonal input or changing hemodynamics in vivo. At the same time, they maintain the physiological interplay of ventilation and perfusion, in contrast to in vitro conditions. A proposal looking at immune responses in lungs5, for example, needs the same quality of data as a study focused on increasing the donor pool size6 for lung transplantations. EVLP can be used across a variety of species, including mice3, rats7,8,9,10,11,12, pigs13, and humans2. Therefore, it is necessary to establish a model that can produce reliable data from a variety of different experimental parameters. Clinical relevance will be generated in subsequent studies using the EVLP model as a tool.
While commercial setups are available for purchase for most species, they can often be cost-prohibitive and confine researchers to a specific brand of equipment and proprietary software. Any deviation from the out-of-the-box setup (e.g., going from one species to another) requires foresight and working around the provided setup, which may prove to be difficult or impossible. In the following, a DIY (do it yourself) setup for rat isolated lungs that is both modular and cost-effective, as well as the surgical procedure for isolating the lungs, are described.
The in vivo portion of the experiments (from general anesthesia to euthanasia) requires prior approval by the respective Institutional Animal Care and Use Committee (IACUC). All procedures described herein were approved (protocol number M1700168) by the IACUC at Vanderbilt University Medical Center, Nashville, Tennessee, and were performed in compliance with the ARRIVE guidelines14. Prior to experimentation, all the rats were housed in the institute's animal care facility, with free access to water and food. Including different studies outside the purview of this manuscript, we have used a total of 148 male Sprague Dawley rats, 7-10 weeks old, with a weight between 250 g and 400 g so far.
1. Apparatus construction
NOTE: All parts, including respective manufacturers, are listed in the Table of Materials.
2. Procedure
3. Data acquisition
Following 10 min of stabilization and baseline readings, we randomized a first set of 10 male Sprague Dawley rats into five small groups: global no-flow ischemia for 5, 7.5, 8, 9, or 10 min (n = 2 per group) followed by reperfusion; these limited preliminary dose-finding experiments were conducted to identify the longest possible ischemia time to still allow sufficient ventilation and reperfusion before the eventual development of a precipitous and irreversible increase in airway pressure and edema formation. Importantly, during ischemia, we chose not to stop the ventilation, but rather reduce its rate from 60 to 20 breaths/min-1 to prevent deflation and irreversible atelectasis of the ex situ lungs. Applying a constant PEEP likely achieves the same goal, in analogy to the tracheas of to-be-transplanted lungs being clamped after inspiration in clinical practice17. Figure 4 suggests that, under these conditions, the longest ischemia time to still allow 150 min of reperfusion was 8 min.
Figure 1: A custom-built lattice to allow for easy configuration and incorporation of new devices. (A) Photograph of the custom-made lattice used to house all the equipment and devices. Not shown are the pressure overflow, gas mixer, gas supply, DAQ, or computer. (B) Schematic of the closed circuit used to perfuse the lung and each device involved in the setup. 1) The buffer is stored in a volumetric flask to reduce the volume and surface area. 2) Heating plates are used to heat both the buffer and the lung chamber. A digital thermometer is used to monitor the temperature inside the lung chamber. 3) A circulating water bath heats the heating coil and air trap to maintain a constant temperature. 4) In the constant-flow setup described here, the roller pump is used to control lung buffer perfusion. 5) Buffer is passed through a heating coil to maintain its temperature. 6) The air trap ensures no air bubbles reach the lung and helps maintain the temperature. 7) A custom double boiler is used to create a humid environment for the lung and maintain its temperature. 8) A pressure overflow is placed between the ventilator and the lung to prevent over-ventilating the lung. 9) A volume-controlled ventilator is used for reliable tidal volumes. 10) The gas mixer can be used to create different gas compositions. 11) Various gas tanks, such as O2, CO2, and N2, are needed for ventilating different compositions of gas. 12) A simple water column is connected to the ventilator to adjust the PEEP on the lung. 13) The DAQ system is responsible for collecting all the data and sending it to the computer. 14) Data are collected and visualized live on the computer. Abbreviations: DAQ = data acquisition system; PEEP = positive end expiratory pressure; AWP = airway pressure. Please click here to view a larger version of this figure.
Figure 2: Custom-made cannulas. (A) Tracheal cannula (made from an 18 G needle). (B) Pulmonary artery cannula. (C) Pulmonary vein cannula. Please click here to view a larger version of this figure.
Figure 3: Isolation process of the rat lungs. (A) The rat is securely taped down after anesthesia, with the mouth taped as well to ensure minimal movement. (B) Proper isolation and placement of the forceps for performing the tracheostomy.
(C) Spreading the rib cage to open the surgical area and minimize the risk of a broken rib puncturing the lungs. (D) Placement of the pulmonary artery cannula. (E) Placement of the pulmonary vein cannula after being secured to the heart. (F) Removal of the trachea and heart-lung block from the rat. (G) The rat isolated lungs hanging prior to placement in the chamber. Please click here to view a larger version of this figure.
Figure 4: Representative results on lung viability after ischemia and reperfusion. Lung viability was defined as survival time until the lungs exhibited a fulminant increase in airway pressure and became fully edematous. Data of limited experiments per group are displayed as mean ± standard error of the mean. Statistical tests were not conducted because of the low number of experiments per group. Please click here to view a larger version of this figure.
NaCl | 119.0 mM |
NaHCO3 | 24.0 mM |
Glucose | 5.5 mM |
CaCl2 | 1.6 mM |
KCl | 4.7 mM |
MgSO4 | 1.17 mM |
NaH2PO4 | 1.18 mM |
Bovine serum albumin (98%) | 4% solution (4.0816 g / 100 mL) |
Table 1: Krebs buffer composition.
More than 100 experiments have been successfully performed in our lab using this setup. The modular design of this customized setup gave great flexibility to potential changes in experimental requirements. While other setups utilize a deoxygenator18 to mimic constant oxygen consumption and CO2 production by end organs, this simplified model did not employ this feature, due to the focus on studying the effects of different gas compositions on pulmonary vascular tone. This approach, in which CO2 is controlled only by the inspired gas concentration, also allowed for differential regulation of CO2 independent from minute ventilation.
A closed circuit (Figure 1) was used to recirculate the modified Krebs buffer (Table 1) to conserve BSA waste, an important cost-limiting step. Perfusion with constant flow mimicking a constant cardiac output was utilized to allow for the study of different doses of pharmacological agents on pulmonary vascular tone, independent of concomitant cardiac output changes in vivo. This can easily be modified to a constant-pressure perfusion by integrating an adjustable height pressure column in the circuit and adjusting the pump speed accordingly; for this, an additional flow probe is required, since flow, now becomes the dependent variable instead of pressure.
The method of ventilation was with positive rather than more physiological negative pressure, mimicking an unconscious or anesthetized, intubated, and ventilated rather than a spontaneously breathing subject. Changing this to negative pressure ventilation would require the use of a sealed chamber for hanging the lungs and making corresponding changes to setting up the ventilator11, but is still possible to incorporate into the apparatus with minimal effort.
The buffer was maintained at a constant temperature by a heating plate, and passed through a heating coil and heated air trap to ensure no dissolved gasses escaped the solution, causing the lung to be perfused with air emboli. To perfuse the lungs, a peristaltic pump was used. The roller pump we used had the great advantage of having both manual controls and voltage inputs and outputs for easy connection to the DAQ. This made it possible to not only record the pump speed, but also to control the pump electronically. One important aspect of in vivo lungs that is absent in the isolated model is pulsatile flow from the heart. By using sinusoidal, square, or triangle waves generated by the software as voltage outputs to the pump, the setup is able to mimic the pulsatile blood flow coming from an in vivo heart. This effect was slightly dampened by the Windkessel effect from tubing between the buffer and the lungs, but was still seen in both the PA and PV pressure recordings. Non-compliant tubing was used to minimize the Windkessel effect. A three-way stopcock placed in the circuit coming from the PV allowed for the collection of effluent for the analysis of injury markers (e.g., lactate dehydrogenase).
A concern was maintaining a consistent metabolic environment – glucose concentration – in the isolated lungs similar to what would be seen in vivo. The compositions of the buffer, as well as ventilated gasses, play an important role. Valenza et al. showed in a porcine EVLP model that glucose consumption in the lungs correlated with lung function4. As a means to keep the glucose concentration constant in the recirculated buffer despite ongoing glucose metabolism, the glucose level of the solution was checked hourly and glucose added back into the buffer to keep the composition as close to the target of 5.5 mM as possible. A modified Krebs buffer (Table 1) with an electrolyte composition that mimicked blood was used for the perfusion. However, this presented problems when trying to measure the glucose concentration; a standard blood glucose measurement device (ACCU-CHEK, Aviva) was unable to read the glucose levels of the buffer, so other options needed to be explored. Glucose test strips did not prove accurate enough to measure the glucose of the solution when compared to a more clinically relevant device (VetScan i-STAT 1)19, which worked reliably.
While other EVLP setups utilize additives such as Perfadex and Celsior20, this setup used BSA due to prior success with previous experiments and its relative affordability if recirculated. This addition was necessary to maintain a physiological oncotic pressure in the lungs, as experiments performed without BSA deteriorated much more rapidly, typically within 30 min of extraction, due to pulmonary edema. Others have also reported success when using higher concentrations of BSA and whole blood. This has the added benefit of allowing the blood to be fractionated to suit the experiment. Lungs that were extracted by a skilled technician in a timely manner were able to survive for up to 6 h under control conditions before edema formation became a problem. This was apparent when observing changes in the lungs, such as gradual changes in color from white to pink, an increase in size due to edema formation, and sharp pressure increases in the pulmonary vasculature. During the procedure, it is important to measure the electrolyte composition of the buffer after the addition of BSA; measured Ca2+, for example, decreased in the buffer following the addition of BSA, implying that BSA was chelating Ca2+, reducing the amount that was available for the lungs21.
In summary, several commercial isolated lung kits are available for purchase that require significantly less time starting up. However, these can be limited in their scope as to what they can do and might not offer the ability to adapt to different experimental parameters. Taking the DIY approach offers a system that is on par with the commercial setups, while also allowing for flexible adaption to a variety of research topics. Once the initial barrier of entry is overcome, the system can be run with minimal oversight and teaching required.
The authors have nothing to disclose.
Support was provided, in part, by a Merit Review Award (101 BX003482) from the U.S. Department of Veteran Affairs Biomedical Laboratory R&D Service, a NIH grant (5R01 HL123227), a Transformative Project Award (962204) from the American Heart Association, and by institutional funds awarded to Dr. Riess. Dr. Balzer received unrelated funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), project number BA 6287/1-1. The authors would like to thank Matthew D. Olsen, Chun Zhou, Zhu Li, and Rebecca C. Riess for their valuable contributions to the study.
1,000 mL Glass Beaker | Pyrex, Chicago, IL | ||
1,500 mL Glass Beaker | Pyrex, Chicago, IL | ||
Air Trap Compliance Chamber | Radnoti | 130149 | |
Bioamplifiers | CWE Inc | BPM-832 | |
Clamps | Fisher Scientific | S02626 | |
DAQ (Data Acquisition) | National Instruments, Austin, TX | NI USB-6343 | |
Gas Mixer | CWE Inc, Ardmore, PA | GSM-4 | |
Heating Coil | Radnoti, Covina, CA | 158822 | |
Heating Plate | Thermo Fisher Scientific, Waltham, MA | 11-100-49SH | |
Heparin | Pfizer | W63422 | |
LabVIEW Full Development System 2014 | National Instruments | ||
Pentobarbital | Diamondback Drugs | G2270-0235-50 | |
pH700 Probe | OAKTON, Vernon Hills, IL | EW-35419-10 | |
Polystat Water Bath | Cole-Parmer | EW-12121-02 | |
Rodent Ventilator | Harvard Apparatus, Holliston, MA | Model 683 | |
Roller Pump | Cole-Parmer, Wertheim, Germany | Ismatec REGLO Digital MS 2/8 | |
Sprague Dawley Rat | Charles River, Wilmington, MA | Strain code 001 | |
VetScan i-STAT | Abraxis, Chicago, IL | i-STAT 1 |
.