Here, we present a robust method for in situ perfusion of the mouse liver to study the acute and direct regulation of liver metabolism without disturbing the hepatic architecture but in the absence of extra-hepatic factors.
The liver has numerous functions, including nutrient metabolism. In contrast to other in vitro and in vivo models of liver research, the isolated perfused liver allows the study of liver biology and metabolism in the whole liver with an intact hepatic architecture, separated from the influence of extra-hepatic factors. Liver perfusions were originally developed for rats, but the method has been adapted to mice as well. Here we describe a protocol for in situ perfusion of the mouse liver. The liver is perfused antegradely through the portal vein with oxygenated Krebs-Henseleit bicarbonate buffer, and the output is collected from the suprahepatic inferior vena cava with clamping of the infrahepatic inferior vena cava to close the circuit. Using this method, the direct hepatic effects of a test compound can be evaluated with a detailed time resolution. Liver function and viability are stable for at least 3 h, allowing the inclusion of internal controls in the same experiment. The experimental possibilities using this model are numerous and may infer insight into liver physiology and liver diseases.
The liver is an essential organ in metabolism. It plays a key role in the control of whole-body energy balance by regulating glucose, lipid, and amino acid metabolism. The increase in liver diseases worldwide is emerging as a major global health burden, and more knowledge is needed about the pathophysiology and its consequences for liver functions.
Various in vitro models have been developed for research on the liver to complement in vivo studies. Isolated and cultured primary hepatocytes from rodents and humans are widely used. Non-parenchymal cells can be separated from hepatocytes using differential and gradient centrifugation, and the co-culture of different cell types is useful for studying intercellular crosstalk1. Although primary human hepatocytes are considered the golden standard for testing drug toxicity, several studies have shown that the hepatocytes rapidly dedifferentiate in tissue culture resulting in loss of hepatic functions2,3,4. Hepatocyte culture in a 3D spheroid system ameliorates the dedifferentiation, is more stable, and appears to mimic the liver in vivo to a higher degree than the traditional 2D culture systems5. Precision-cut liver slices are another well-established in vitro model that keeps the tissue architecture intact and contains the non-parenchymal cells present in the liver6. More advanced in vitro models include liver-on-a-chip7 and liver organoids8. However, with all these approaches, there is a loss of structural integrity and flow dynamics, including vectorial portal-hepatic vein flow, which likely impacts the generalizability.
The isolated perfused rat liver was first described by Claude Bernard in 18559, and is still used in various scientific fields for studies of liver biology, toxicology, and pathophysiology. Advantages of the perfused liver compared to the above-mentioned in vitro models include the maintenance of the hepatic architecture, the vascular flow, the hepatocyte polarity and zonation, and the interactions between hepatocytes and non-parenchymal cells. Compared to in vivo studies, the perfused liver allows the study of liver metabolism in an isolated manner avoiding extra-hepatic factors carried by the blood and with complete control over the experimental conditions. Several modifications have been made to improve the rat liver perfusion model over the years10,11,12,13. Although mice have been used for isolated perfused liver studies, less literature is available. Here, we present a method for in situ perfusion of the mouse liver by cannulation of the portal vein and the suprahepatic vena cava inferior to study the acute and direct metabolic responses to metabolic substrates and hormones as measured in the hepatic venous effluent from the mouse liver in real-time.
All animal experiments were conducted with permission from the Danish Animal Experiments Inspectorate, Ministry of Environment and Food of Denmark (permit 2018-15-0201-01397), and the local ethics committee in accordance with the EU directive 2010/63/EU, the National Institutes of Health (publication No. 85-3) and following the guidelines of Danish legislation governing animal experimentation (1987). This is a terminal procedure, and the cause of death is exsanguination and perforation of the diaphragm under deep anesthesia.
1. Experimental animals
2. Preoperative preparations
3. Operation and perfusion
NOTE: An illustration of the perfusion setup used in this study is provided in Figure 1.
Figure 1: An illustration of the perfusion setup. (A)The operating table is elevated on a tripod stand and heated to 37 °C. The perfusion buffer is gassed (95% O2, 5% CO2), pumped via a peristaltic roller pump, and heated in the heat exchanger with a built-in bubble trap. The system furthermore consists of a pressure gauge and spindle pump for adjustment of the perfusion pressure. The perfusion pressure is continuously recorded and visualized via a transducer on a PC, a pressure recording program. (B) The red box captures the connections of three-way stopcocks. The first three-way stopcock is open for the infusion of a test compound via a syringe pump, and the second is closed. The third is open for continuous pressure measurements. The fourth stopcock may be used to collect input samples, for e.g., gas analysis across the perfused liver. The connectors can be modified as needed for specific experiments requiring more or fewer infusion lines. Please click here to view a larger version of this figure.
Figure 2: Photos of the mouse abdominal cavity before and during liver perfusion. (A)The green dot indicates the location of the tip of the portal vein catheter. It is important that the tip of the catheter is positioned just below the branching point of the portal vein into the left and right hepatic portal veins but above the pancreato-duodenal branch to avoid leakage. The yellow dot indicates the correct location of the vessel clamp on the infrahepatic inferior vena cava between the right renal vein and the liver to avoid the backflow of blood into the perfused liver. (B, C) A perfused mouse liver with the two catheters inserted in the portal vein (B) and suprahepatic inferior vena cava (C) and the vessel clamp on the infrahepatic inferior vena cava (B). Please click here to view a larger version of this figure.
4. Experiment
5. Biochemical measurements
6. Data analysis
A steady baseline is required to determine whether a stimulus or substrate leads to the release of the molecule of interest. Figure 3A shows an example of a successful experiment. Production of urea in the perfused liver is measured in 2 min intervals and shown as mean ± SEM. The baseline periods preceding each of the two stimulation periods are steady. The mean urea production during the two stimulation periods and the respective preceding baselines are shown in Figure 3B. Statistical significance between the periods was tested using one-way ANOVA with repeated measurements. Based on these results, it can be concluded that the urea response to two consecutive stimulations with mixed amino acids is similar, which is an important control experiment in evaluating repeated stimulations with different doses or test compounds.
In addition to ureagenesis, it is possible to study hepatic glycogenolysis and lipolysis using the above protocol. Figure 3C–F show data from a separate experiment during an infusion with glucagon. The release of glucose (Figure 3C) and the non-esterified fatty acids (NEFA) (Figure 3E) was measured in 4 min periods. A steady basal release of glucose (Figure 3C) and NEFA (Figure 3E) is observed prior to a robust increase in glucose and NEFA during the glucagon infusion, respectively. Based on the mean values at the basal state and during the glucagon infusion, it can be concluded that glucagon rapidly stimulates hepatic glycogenolysis (Figure 3D) and lipolysis (Figure 3F).
Figure 4 shows an example of apparently unsuccessful experiments. The basal release of urea is unsteady, and the stimulation periods are too close to each other for the urea production to return to basal levels before the next stimulation begins. Without steady basal periods, it is not possible to distinguish one urea response from the next. From these data, it is impossible to conclude whether different glucose concentrations influence amino acid-induced ureagenesis in the perfused mouse liver. Experimental protocols should instead be designed with 20-30 min of baseline periods in-between two consecutive stimulations to reach a steady baseline before each stimulation period.
Figure 3: Good-quality data from the perfused mouse liver. (A,B) Urea total is shown during basal conditions and in response to periods of administration of mixed amino acids (10 mM) and glucose (6 mM). Data are presented as (A) mean ± SEM and (B) mean values during each stimulation and preceding basal period (each dot represents a mouse), n = 6. (C, D) Glucose and (E,F) non-esterified free fatty acids (NEFA) total outputs are shown in basal conditions and response to the administration of glucagon (10 nM). Data are presented as (C,E) mean ± SEM and (D,F) mean values during the basal and stimulation period (each dot represents a mouse), n = 6. Please click here to view a larger version of this figure.
Figure 4: Poor-quality data from the perfused mouse liver. Urea total output is shown in basal conditions in response to mixed amino acids (10 mM) and decreasing concentrations of glucose (18 mM, 12 mM, 6 mM, 3 mM, and 0 mM). Data are presented as mean ± SEM, n = 3. Please click here to view a larger version of this figure.
Supplementary Figure 1: Time-dependent partial pressure of oxygen and carbon dioxide and pH in Krebs-Henseleit bicarbonate buffer during gassing with 95% O2 + 5% CO2. (A) Partial pressure of oxygen, (B) carbon dioxide, and (C) pH in samples of Krebs-Henseleit perfusion buffer (pH not adjusted prior to the experiment) collected after having run through the perfusion system at indicated time points after the start of the gassing with 95% O2 + 5% CO2. Perfusion buffer samples were measured with a blood-gas analyzer. Please click here to download this figure.
Supplementary Figure 2: Changes in partial pressure of oxygen and carbon dioxide and pH across the perfused liver. (A) Partial pressure of oxygen, (B) carbon dioxide, and (C) pH in samples of Krebs-Henseleit perfusion buffer collected immediately before reaching the liver and after being passed the perfused liver at 1 min and 180 min into an experiment. Perfusion samples were measured with a blood-gas analyzer. Data is shown as mean ± SD. **P < 0.01, ****P < 0.0001 by multiple paired t-tests corrected for multiple testing using the Holm-Sidak method. n =14. Please click here to download this figure.
The isolated perfused mouse liver is a strong research tool for studies of the dynamics and molecular mechanisms of hepatic metabolism. The possibility of minute-to-minute sample collection provides a detailed evaluation of the direct effect of a test compound on the liver. Compared to で vivo studies, the perfused liver allows us to study liver metabolism in an isolated manner avoiding extra-hepatic factors carried by the blood and with complete control over the experimental conditions. The advantages of liver perfusion compared to in vitro studies using isolated hepatocytes are the maintenance of the hepatic architecture, polarity, zonal division, and vascular integrity. Another advantage of liver perfusion is the large sample size volume (3.5 mL/min), allowing measurement of several outcomes in the same sample. The perfused liver is, however, not ideal when studying effects dependent on transcriptional regulation since such mechanisms may take several hours to occur. In addition to metabolic research, the protocol may be applied in other research fields to study the endocrine functions of the liver (secretion of hormones and hepatokines) and liver metabolism of drugs and xenobiotics.
The most critical step for this method is the placement of the catheter in the portal vein. It is important that the tip of the catheter is positioned immediately before the branch point of the portal vein into the left and right hepatic portal veins (Figure 2). If the tip of the catheter is pushed too far, only part of the liver will be properly perfused. If it is placed too far down in the portal vein, leakage through the pancreato-duodenal branch may occur. Placement of the second catheter for collection is less urgent since the liver at this step is already perfused with the oxygenated buffer. When the operation has been successfully performed, the liver will be respiring and responsive for at least 3 h with a constant pressure and perfusion output (Supplementary Figure 2). At this stage, the avoidance of air bubbles in the perfusion system is the most important factor for a successful experiment. Air bubbles are difficult to avoid completely since the buffer is continuously gassed with oxygen, but the bubbles should be trapped in the bubble trap of the perfusion system. It is therefore important to keep an eye on the bubble trap and refill it with perfusion buffer when it is close to empty.
Krebs-Henseleit buffer (KHB) is the most widely used solution for liver perfusions13. The classical formulation contains 2.5 mmol/L CaCl2 but based on a previous study showing that calcium contributes to mitochondrial damage, we use a buffer with only 1.25 mmol/L CaCl212. The addition of erythrocytes, albumin, or glucose to the buffer is not needed11. The proposed buffer also allows molecular profiling using advanced biochemical techniques such as mass-spectrometry-based metabolomics. Different concentrations of a test compound may be tested to determine the optimal dose and to evaluate physiological vs. pharmacological effects. In the experiments conducted in this study, 10 mM AA is in the high physiological range corresponding to postprandial levels, whereas 10 nM glucagon corresponds to a pharmacological dose. As demonstrated in Figure 4, it is essential to include 20-30 min baseline period between two consecutive stimulations to be able to compare the dynamics and molecular mechanisms of the respective responses.
It is important to avoid blood in the perfusate samples as it may interfere with the subsequent analyses, e.g., colorimetric determination of urea and glucose concentrations. If blood appears in the collected samples after ~15 min into the equilibration period, when a 3-way valve is opened for infusions, or when the bubble trap is filled, the vessel clamp on the infrahepatic vena cava is likely not placed correctly.
In our experience, a successful operation results in a stable portal pressure of ~10 mmHg and an output flow rate of 3.5 mL/min within a few minutes after inserting the collecting catheter. If the pressure is considerably higher or continuously increasing, or if the collected output flow rate is not 3.5 mL/min or continuously dropping, several events should be considered: 1) Are all liver lobes equally perfused and showing a uniform pale color? If not, an occlusion may have occurred (often an air bubble; carefully "massage" the liver lobe with a cotton stick; this sometimes helps to release the air bubble whereby the pressure will drop immediately), or the portal vein catheter has been pushed too far up, thus supplying only one of the two branches of the hepatic portal vein (cautiously try to pull back the catheter a few millimeters). 2) Is the suprahepatic inferior vena cava twisted? This can occur when attaching the male part of the collecting tube to the female part of the catheter inserted in the vein; cautiously remove the collecting tube and observe if the pressure drops. 3) Is the collecting catheter pushed too far down the suprahepatic inferior vena cava so the tip of the catheter has wedged into the liver? Cautiously retract a few millimeters if needed.
In summary, the perfused mouse liver is a physiologically relevant in vitro experimental model that can be used to study the dynamic effects of a given compound on the liver. The quality of the operation is critical for the quality of data obtained, and in our experience, it takes ~2 months of training before data of satisfactory quality can be expected. Once the technique has been learned, the possibilities with this method are numerous, including the use of transgenic or knockout donor animals regarding the genes of interest as well as mouse models of liver diseases. Experiments may, however, be limited by potential toxic side effects (which are likely to be much more severe in vivo) or poor solubility of test compounds. Such compounds may be given orally to the donor animals at an appropriate time period before perfusion.
The authors have nothing to disclose.
The studies and Nicolai J. Wewer Albrechtsen were supported by Novo Nordisk Foundation Excellence Emerging Investigator Grant – Endocrinology and Metabolism (Application No. NNF19OC0055001), European Foundation for the Study of Diabetes Future Leader Award (NNF21SA0072746) and Independent Research Fund Denmark, Sapere Aude (1052-00003B). Novo Nordisk Foundation Center for Protein Research is supported financially by the Novo Nordisk Foundation (Grant agreement NNF14CC0001). Figure 1B was created with biorender.com. We thank Dr. Rune E. Kuhre (Novo Nordisk A/S) for fruitful discussions on the perfused mouse liver.