Isolation of Regenerating Hepatocytes after Partial Hepatectomy in Mice

Summary

Lipid-laden hepatocytes are inherent to liver regeneration but are usually lost upon density-gradient centrifugation. Here, we present an optimized cell isolation protocol that retains steatotic hepatocytes, yielding representative populations of regenerating hepatocytes after partial hepatectomy in mice.

Abstract

Partial hepatectomy has been widely used to investigate liver regeneration in mice, but the isolation of high yields of viable hepatocytes for downstream single-cell applications is challenging. A marked accumulation of lipids within regenerating hepatocytes is observed during the first 2 days of normal liver regeneration in mice. This so-called transient regeneration-associated steatosis (TRAS) is temporary but partially overlaps the major proliferative phase. Density-gradient purification is the backbone of most existing protocols for the isolation of primary hepatocytes. As gradient purification relies on the density and size of cells, it separates non-steatotic from steatotic hepatocyte populations. Therefore, fatty hepatocytes often are lost, yielding non-representative hepatocyte fractions.

The presented protocol describes an easy and reliable method for the in vivo isolation of regenerating hepatocytes regardless of their lipid content. Hepatocytes from male C57BL/6 mice are isolated 24-48 h after hepatectomy by a classic two-step collagenase perfusion approach. A standard peristaltic pump drives the warmed solutions via the catheterized inferior vena cava into the remnant, using a retrograde perfusion technique with outflow through the portal vein. Hepatocytes are dissociated by collagenase for their release from the Glisson's capsule. After washing and careful centrifugation, the hepatocytes can be used for any downstream analyses. In conclusion, this paper describes a straightforward and reproducible technique for the isolation of a representative population of regenerating hepatocytes after partial hepatectomy in mice. The method may also aid the study of fatty liver disease.

Introduction

The liver can regenerate itself even after major tissue loss. This unique regenerative capacity is explicitly illustrated by the experimental model of partial (70%) hepatectomy, first described in rats by Higgins and Anderson in 1931¹. In this model, 70% of the liver is surgically removed from animals by clipping off larger liver lobes. The remaining lobes then grow through compensatory hypertrophy to restore the original liver mass within about 1 week after surgery, albeit without restoration of the original liver architecture^2,3. Additional hepatectomies with varying amounts of tissue removal have been developed, such as 86%-extended hepatectomy where the liver remnant is too small to recover, eventually leading to posthepatectomy liver failure (PHLF) and subsequent death in 30%-50% of the animals^4,5,6. These models enable the study of normal and failed liver regeneration, depending on the amount of resected tissue (Figure 1).

Although mouse models of hepatectomies have been used successfully for many years, only recently have more advanced analytical methods allowed for a deeper insight at the single-cell level. For most of these methods, however, the presence of individual hepatocytes is a basic prerequisite. Most protocols for the isolation of primary hepatocytes are based on a two-step collagenase perfusion technique and subsequent density-gradient purification to separate viable hepatocytes from debris and non-parenchymal, as well as dead cells^7,8,9. This method was first described by Berry and Friend in 1969¹⁰ and adapted by Seglen and colleagues in 1972^11,12. However, as gradient centrifugation relies on the density and size of cells, lipid-laden hepatocytes are often lost during standard purification. While such loss may be negligible for many research questions, it is a crucial aspect for early liver regeneration. During the first 2 days, hepatocytes within the regenerating mouse liver accumulate lipids, thereby growing in size and dipping in density. This transient regeneration-associated steatosis (TRAS) serves to provide regenerative fuel and is temporary, but partially overlaps the major proliferative phase and is unevenly distributed within the liver lobules - the functional units of the liver^13,14. After extended 86%-hepatectomy, however, TRAS also occurs but persists, because regeneration is stalled and lipids are not being oxidized¹⁴. Therefore, gradient-purification of hepatocytes following 70%- or 86%-hepatectomies will yield non-representative fractions, as most lipid-laden hepatocytes are lost due to their low density¹⁵.

In this modified isolation protocol, hepatocytes from C57BL/6 mice are isolated 24-48 h after hepatectomy by a classic two-step collagenase perfusion approach. Usually, cannulation and perfusion of the remnant for cell isolation are done via the portal vein. However, portovascular resistance in small remnants left after major resection is high¹⁶, and thus perfusion is delicate. Because the vena cava remains unaffected by hepatectomies, perfusion can be easily performed in the retrograde direction via cannulation of the vena cava. A standard peristaltic pump drives the warmed solutions via the catheterized inferior vena cava into the liver remnant, using retrograde perfusion with outflow through the portal vein (Supplementary Figure S1). Hepatocytes are dissociated by collagenases and released from the Glisson's capsule. After washing and careful processing of viable hepatocytes by stepwise isolation using a low-speed centrifugation approach, the hepatocytes can be used for any downstream analyses.

Protocol

All animal experiments were in accordance with Swiss Federal Animal Regulations and approved by the Veterinary Office of Zurich (n° 007/2017, 156/2019) assuring human care. Male C57BL/6 mice aged 10-12 weeks were kept on a 12 h day/night cycle with free access to food and water. Each experimental group consisted of six to eight animals. See the Table of Materials for details related to all materials, equipment, and reagents used in this protocol.

1. Partial hepatectomy in mice

1. For standard hepatectomy (70%), ligate and resect the left lateral lobe, the right portion of the median lobe, and the left portion of the median lobe (Figure 1B). For extended hepatectomy (86%)⁴, also remove the caudate lobes and the right anterior lobe (Figure 1C).
2. NOTE: Standard hepatectomy is a procedure that has been used in liver regeneration research for many years. Protocols for this procedure are available^3,17, including the video-assisted protocol of Mitchell and Willenbring¹⁸. Further details for the techniques used here for hepatectomies can be found in Supplementary File 1.

Figure 1: Standard (70%) and extended (86%) hepatectomy in mice. (A) The five mouse liver lobes and their respective contributions to the total liver weight. (B) Schematic illustration of 70%-hepatectomy in mice. The dark lobes represent the future liver remnant. (C) Schematic illustration of 86%-hepatectomy in mice. The dark lobes represent the future liver remnant. (D) Precise volume of resected tissue post 70%- and 86%-hepatectomy. (E) Mouse abdomen immediately after 70%-hepatectomy; (F) mouse abdomen immediately (left) and 48 h (right) after 86%-hepatectomy. Note the pale color of the steatotic remnant (white arrow). n = 6-7/group. Abbreviations: sHx = standard hepatectomy; eHx = extended hepatectomy; LW = liver weight. Please click here to view a larger version of this figure.

2. Preparation of the perfusion solutions

1. Prepare the perfusion, digestion, and preservation buffers (see Table 1).
   1. Adjust the pH of all buffer solutions at 37 °C by adding sodium hydroxide (NaOH) or hydrogen chloride (HCl) as needed. The optimal pH for the buffers is 7.4.
   2. Place the preservation buffer and Williams' Medium E on ice.
2. Prepare the flow cytometry buffer and store it on ice.

Table 1: Solutions and buffers used for the digestion and purification of hepatocytes. Please click here to download this Table.

3. Preparation of perfusion equipment

1. Warm the water bath to 42 °C and place the perfusion buffer (50 mL) and the digestion buffer (10-20 mL) in the water bath. Do not add collagenase to the digestion buffer yet.
2. Prepare the peristaltic pump and insert the tubing. The complete perfusion setup is shown in Figure 2.
   1. Connect a 26 G IV cannula to the outlet end of the tubing using a Luer lock connector. Insert the inlet end of the tube into the prewarmed perfusion buffer tube in the water bath. Flush the tubing with 70% ethanol, followed by 50 mL of sterile sodium chloride (NaCl 0.9%). Prime the tubing with warm perfusion buffer (pump speed of 3 mL/min).
3. Sedate the mouse using isoflurane inhalation anesthesia (800 mL/min O₂, 3%-5% isoflurane for induction and 2% for maintenance during the procedure). Handle isoflurane under a laboratory hood and provide adequate ventilation.
4. Administer buprenorphine subcutaneously 30 min before surgery at a dosage of 0.1 mg/kg body weight.
5. To prevent hypothermia, place the sedated mouse on a warming pad and put a rolled cloth tissue underneath the upper abdomen to elevate the liver above the other organs and facilitate access to the inferior vena cava.
   NOTE: Do not use tissue that is too thick as kinking of vessels is possible and perfusion efficacy would be affected.
6. Add eye ointment to prevent corneal damage.
7. Before starting surgery, ensure that the animal is adequately anesthetized by testing the pedal withdrawal reflex (foot pad pinch on both hind feet). In case of a response, supply additional anesthesia and retest before starting the procedure.
8. Clean the abdomen with 70% ethanol.
   1. Reopen the midline incision by cutting the suture and gently pulling the wound edges apart. If the hepatectomy is older than 24-48 h, remove the suture and cut the skin with scissors.
   2. Fix a 5-0 polypropylene suture to the sternum, pull it cranially, and fix it in this position. Use a retractor or simple clips to keep the abdomen open. The abdominal cavity has to be exposed as much as possible to optimize access and visualization.
9. Move the intestines to the right using cotton swabs to reveal the portal vein and the vena cava. Use a wet cloth to retain the intestines.
10. Place a heavy object of approximately 2 cm height (e.g., a silicone-coated weight ring for volumetric flasks) adjacent to the mouse hind legs (Supplementary Figure S2A). Place the tubing with the connected 26 G IV cannula on the object and position the needle carefully on top of the vena cava. Adjust the length of the tubing.
11. Put the prepared collagenase stock solution into the prewarmed digestion buffer tube. Add 250 µL of the stock solution to 10 mL of digestion buffer. Prepare 10-20 mL of digestion buffer per animal. For larger animals or perfusion of whole livers, prepare up to 30 mL of digestion buffer.
    NOTE: It is recommended to add the collagenase stock solution to the warmed digestion buffer approximately 30 min prior to the start of the digestion process.
    
    Figure 2: Overview on the perfusion setup. (A) Surgical table with the equipment needed for the perfusion. (B) The materials needed for preparation of the liver, as well as hepatocyte extraction and isolation. Please click here to view a larger version of this figure.

4. Cannulation and perfusion

1. Adjust the pump speed to 3 mL/min and turn on the pump. Let the prewarmed perfusion buffer reach the needle. Discard the first 2-3 mL of perfusion buffer.
2. Perform cannulation of the inferior vena cava.
   1. While the buffer is running through the needle, insert the 26 G IV cannula at a shallow angle into the vena cava below the kidney. Ensure that the needle bevel points upward.
   2. Use a cotton swab to gently pull the vena cava caudally below the puncture site so that the provided tension facilitates the insertion of the cannula into the vein. Look for blood in the flash chamber of the catheter when the needle enters the lumen.
   3. Advance the needle an additional 2-3 mm to ensure that the tip of the plastic catheter also has entered the vein.
   4. Slide the plastic catheter over the needle and into the vena cava another 5 mm. Remove the needle slowly and very carefully.
      NOTE: Fixing the cannula with a ligature is not recommended. This step is time-consuming as the vessel must first be dissected for this purpose. If the cannula is loosely positioned and supported with an object, no further fixation is necessary (see perfusion setup in Supplementary Figure S2). To stabilize the cannulation site and prevent backflow, a single drop of monomeric n-butyl-cyanoacrylate can be added onto the cannulation site.
3. When blood drops out of the cannula, fill it with warm perfusion buffer using a syringe.
4. Reattach the tube to the cannula, still running at a pump speed of 3 mL/min. Let the perfusion buffer enter the liver.
5. After 2-3 s, look for white spots forming in the liver and/or expansion/swelling of the portal vein, which indicate that the perfusion buffer is flowing through the liver and entering the liver lobules from the central vein (Figure 3).
6. Wait for the portal vein to visibly swell within 1-2 s after the appearance of white spots on the liver surface. Cut the portal vein with scissors as distally as possible from the liver hilus. Use a micro vessel clip to label (not occlude) the cutting site (Figure 3B).
   NOTE: This simplifies assessment of the flow through the liver during the perfusion process. The liver clears of blood instantly and turns yellow-white within a few seconds (Figure 3C). An additional cut through the skin on the right side of the abdominal opening facilitates the outflow of the blood as well as the perfusion solution (Figure 3B and Supplementary Figure S2B,C).
7. Increase the flow up to 4-7 mL/min depending on the weight of the animal, liver size, and the extent of the prior hepatectomy.
8. Clamp the portal vein with tweezers or vascular clamp for 7-10 s. Make sure no fluid is passing through.
   NOTE: The liver swells visibly during clamping and relaxes upon release. This is crucial to flush the whole liver and clear it of any remaining blood.
9. Perform a second clamp after approximately 30 s and make sure that the liver swells and relaxes. Continue with flushing the animal until the buffer flowing out of the portal vein is clear, but at least for 3-4 min.
   NOTE: The pump speed depends on the tube as well as the size of the liver. It must be evaluated individually.
10. At this point of the procedure, euthanasia should have occurred secondary to exsanguination. Confirm that systemic circulation has stopped (no heartbeat or flickering of the heart). To ensure death, bilateral pneumothorax is performed at this stage of the procedure as secondary physical method of euthanasia.
    NOTE: Reduce the pump speed slightly if systemic circulation has stopped (no heartbeat or flickering of heart).

Figure 3: Perfusion process from cannulation to digestion. (A) Anatomy of the mouse liver with the inferior vena cava (white arrow) and the portal vein (yellow arrow). (B) Cannulation of the inferior vena cava. The cannula is secured with a ligature (white arrow), and the location of the outflow through the opened portal vein is marked (not clamped) with a micro vessel clamp. (C) Note the appearance of patchy structures before the perfusion buffer has cleared the liver from all remaining blood (white arrow). The skin is incised (yellow arrow) and a cotton swab is placed to ensure drainage of blood and perfusion fluid. Intermittent clamping can be performed with a vascular clamp or tweezers. (D) The liver should be cleared of all blood (*). After the collagenase-containing digestion buffer has entered the liver, it will no longer relax after clamping and the liver lobes will increase in size. (E) After a while, a bubbly appearance on the surface of the liver can be observed (*). Please click here to view a larger version of this figure.

5. Digestion

1. Pause the perfusion pump and quickly transfer the inlet tubing from the perfusion buffer to the prewarmed digestion buffer. Restart the pump.
2. Before the digestion buffer reaches the liver, clamp the portal vein one more time for 3-4 s. Ensure that the liver relaxes upon release of the clamp and the perfusion fluid remains clear.
   NOTE: The digestion buffer contains phenol red and is easily distinguishable from the clear perfusion buffer. This allows for easy tracking within the tube.
3. As soon as the digestion buffer has reached the liver, clamp the portal vein once again with a micro vessel clip.
   NOTE: When clamping, the liver swells but does not relax upon release of the clamp. This is normal.
4. To facilitate the digestion process, close the superior vena cava with a vascular clamp directly below the diaphragm to allow the digestion buffer to pass from the inferior vena cava to the liver to the outflow through the opened portal vein.
   NOTE: This clamping ensures that the systemic circulation is bypassed and unnecessary contact with residual blood components/inhibitors is prevented. This step is optional, as it is difficult to approach the superior vena cava behind the enlarged liver tissue after sham surgery, but access is much easier in hepatectomized mice.
5. Digest for approximately 4 min at a flow rate of 5 mL/min. As digestion progresses, look for signs of the liver beginning to swell and small clear/transparent sections on the surface of the liver. Furthermore, observe that the liver takes on the texture of a wet piece of cloth and appears almost soggy (Figure 3E). Probe the consistency by carefully touching with a damp cotton swab.
6. Continue with the perfusion until a marked difference in the surface texture of the liver can be observed. Observe that the liver assumes a very light color and a bubbly appearance (Figure 3E), and the Glisson's capsule (i.e., the liver sack) separates from the parenchyma. Stop the digestion process as soon as the liver has acquired these properties, as over-digestion can damage the hepatocytes. Remove the needle before air gets into the liver.
   NOTE: Usually 10-20 mL of digestion buffer is needed to reach sufficient digestion. This depends on the animal's size, extent of hepatectomy, tubing setup, and quality of the collagenase solution. If needed, increase the perfusion time rather than the perfusion speed. Too much pressure in the vascular system can cause the liver to burst and the perfusion/digestion fluid might be lost to the retroperitoneal space.

6. Preparation of the liver

1. Remove the liver gently from the abdominal cavity. Be very careful as it is now very flimsy and fragile.
2. Grasp the central connective tissue between the lobes using forceps and slightly lift it upward, using it as an anchor point.
3. Cut all the connections of the liver to other organs, remove the gall bladder, and place the liver in the ice-cold preservation buffer.
   NOTE: Ideally, hepatocyte extraction and further processing should take place immediately to maintain the viability of hepatocytes. However, if necessary, the liver can be stored for a short time at 4 °C (e.g., for transport). This delay should not exceed 30-40 min.

7. Hepatocyte extraction

1. Transfer the liver to a 10 cm Petri dish and add 10 mL of ice-cold Williams' Medium E.
2. Rupture the Glisson's capsule with fine-tip tweezers in a few locations along the liver surface. Grasp a central part (e.g., connective tissue at the liver hilus) with two pairs of tweezers and slowly pull them apart, allowing the capsule to tear without damaging the hepatocytes. Release the cells by gently shaking the capsule (Supplementary Figure S3).
   NOTE: Ideally, the liver tears apart easily and releases the cells. Do not apply force. A cell scraper can help to completely remove all cells and increase cell yield. Do not cut the liver into pieces with scissors.
3. Filter 5 mL of liver pulp through a 100 µm cell strainer into a 50 mL tube. Rinse the filter with 10 mL of fresh ice-cold medium. Filter the remaining 5 mL of pulp through the cell strainer.
   NOTE: Use a 25 mL serological pipette to transfer the liver pulp with the dissociated hepatocytes (Figure 4A). Smaller pipettes with smaller openings increase the shear stress and irreversibly damage the hepatocytes.
4. Add a total of 30 mL of cold medium to rinse the Petri dish, filter it, and add the suspension to the 50 mL tube until it is full. All isolated cells are now in suspension (Figure 4B).

Figure 4: Purification by gentle centrifugation. (A) Liver homogenate left after the extraction step. (B) Microscopic view (20x magnification) of the homogenate; note the marked contamination with debris. (C) Purification centrifugation steps and (D) microscopic views of the supernatants to be discarded. (E) Microscopic view of the purified hepatocyte fraction. Scale bars = 100 µm. Please click here to view a larger version of this figure.

8. Hepatocyte isolation

1. Spin at 50 × g for 5 min at 4 °C (lowest acceleration and lowest brake possible).
   NOTE: Hepatocytes are denser than non-parenchymal liver cells. Due to the low centrifugation force, only hepatocytes are pelleted, while other cells (e.g., immune cells, erythrocytes, and sinusoidal cells) remain in the supernatant.
2. Aspirate most of the supernatant, leaving 1 mL to resuspend the cells by gently swirling the tube.
3. Add 40 mL of cold Williams' E medium and spin again at 50 × g for 10 min at 4 °C (low acceleration, low brake) to further remove dead hepatocytes and cell debris and pellet viable and fatty hepatocytes (Figure 4C).
4. Discard most of the supernatant, leaving 1 mL to resuspend the cells by swirling the tube.
5. Add 40 mL of cold Williams' E medium and spin again at 50 × g for 5 min at 4 °C (low acceleration, low brake).
6. Aspirate most of the supernatant, leaving 1 mL to resuspend the cells by gently swirling the tube.
   NOTE: Do not stop this process until cells are fixed or analyzed. Hepatocytes are very fragile and any delay in the perfusion, digestion, and purification process can damage the cells.
7. Determine the final cell concentration after the addition of trypan blue, using a Neubauer-improved counting chamber.
   NOTE: Most debris and non-parenchymal cells have now been removed, resulting in a clean pellet of approximately 10-15 × 10⁶ hepatocytes left after 70%-hepatectomy.
8. Depending on the desired concentration, add more ice-cold medium. Use the hepatocyte cell suspension for any downstream analysis or initiate a primary cell culture.
   NOTE: At this stage, only few immune and non-parenchymal cells (<5%) remain in the suspension. If further purification is desired, perform a negative selection of CD31⁺ and CD45⁺ cells by magnetic- or fluorescence-activated cell sorting (MACS/FACS). To date, there is no reliable and robust surface marker for liver parenchymal cells.

9. Preparation of the isolated hepatocytes for flow cytometry

1. Centrifuge the hepatocytes at 100 × g for 5 min.
2. Discard the supernatant and add the desired amount of flow cytometry buffer, depending on the concentration of the cell suspension.
3. Add 1 mL of cell suspension in a flow cytometry tube. Centrifuge for 5 min at 100 × g and discard the supernatant.
4. Add 100-200 µL of the diluted Alexa Fluor 488 Zombie green viability dye (concentration 1:400) to the cells and shake them gently. Carefully resuspend the hepatocytes by manual shaking or using a vortex mixer at low speed (maximum 2-3) for 2 s.
   NOTE: Do not use small pipette tips to resuspend the cells. They are very fragile and the applied shear stress causes damage and reduces cell viability. If pipetting is not avoidable, use a 1,000 µL pipette after cutting off the smallest part from the tip to enlarge the diameter and pipette the cells very slowly.
5. Place the tubes on ice or store them at room temperature, depending on the desired staining. Incubate the cells for 20-30 min in the dark.
6. Add 2 mL of flow cytometry buffer and wash the cells 3x. Centrifuge the cells after each washing step at 100 × g for 5 min.
7. Add 2 mL of fixation buffer (1:1 4% PFA and PBS). Carefully resuspend the hepatocytes by manual shaking or using a vortex mixer at low speed (maximum 2-3) for 2 s.
8. Fix the cells for 30 min.
9. Centrifuge at 100 × g for another 5 min, discard the supernatant, and add flow cytometry buffer.
   NOTE: Cells can be stored in flow cytometry buffer prior to analysis up to 72 h after isolation.

10. Analyzing hepatocytes with flow cytometry

1. Analyze the hepatocytes using a fluorescence-activated cell sorter.
   NOTE: Consider the relatively large size of hepatocytes and adjust the voltages. Start with a low voltage and do not go higher than 350 V for forward scatter (FSC) and 220 V for side scatter (SSC).
   1. Adjust the voltages for FSC and SSC to the estimated large size of the cells. Identify the hepatocyte population and record all events using SSC-A and FSC-A (Figure 5A).
   2. Observe that debris and non-parenchymal cells are displayed at the bottom left corner of the FSC versus SSC density plot and are excluded (Figure 5A).
   3. As doublet cells can affect the analysis, construct a side scatter height (SSC-H) versus side scatter area (SSC-A) density plot to exclude doublets, as shown in Figure 5B.
   4. Select the final hepatocyte population by gating CD31^- (endothelial marker) and CD45^- (immune marker) cells (Figure 5C).

Representative Results

TRAS peaks at 16 h post hepatectomy and gradually vanishes 32-48 h after standard hepatectomy, but persists beyond 48 h after extended hepatectomy. Macroscopically, TRAS is readily visible as a pale complexion of the liver remnant (Figure 1F) and can be observed in hepatectomized mice between 16 h and 48 h after surgery.

The estimated final yield is 10-15 × 10⁶ hepatocytes after 70%-hepatectomy and 4-9 × 10⁶ after extended 86%-hepatectomy in mice, with a mean final viability of 78% and 65%, respectively. This corresponds to the expected percentage compared to a total yield of 50-70 × 10⁶ from a whole mouse liver (Figure 6A,C). After partial hepatectomies, hepatocytes show an increased cell size compared to normal livers, corresponding to their increased lipid-content (Figure 6B). The gating strategy is shown in Supplementary Figure S4.

Figure 6: Hepatocyte yield, cell size, and viability. (A) Cell count of one whole mouse liver and of remnants 24 h post 70%- and 86%-hepatectomy. (B) Calculated cell volume of isolated hepatocytes. Note the marked increase in cell size 24 h post hepatectomy, for both 70% and 86%. (C) Cell viability of isolated hepatocytes after each step of the purification process. Viability was measured by flow cytometry using Alexa Fluor 488 Zombie green viability dye. Percentages are taken from all hepatocyte singlets. For the gating strategy, see Supplementary Figure S6 and Supplementary Figure S4. Error bars refer to standard deviations with n = 6-8/group. Abbreviations: sHx = standard hepatectomy; eHx = extended hepatectomy; w/o resection = sham surgery. Please click here to view a larger version of this figure.

An increased lipid content in murine hepatocytes 24 h after partial hepatectomy can be observed by an increase in granularity (Figure 7; for gating strategy, see Supplementary Figure S5) or directly observed by the presence of lipid vesicles inside enlarged hepatocytes (Figure 8 and Supplementary File 1).

Figure 7: Increased granularity and cell size measured by flow cytometry. (A) Increased granularity among isolated hepatocytes 24 h post hepatectomy as an indirect marker for the presence of lipid vesicles. (B) Histogram showing SSC. (C) The percentages represent the fraction of cells with high cytoplasmic granular intensity, measured by the SSC signal. Error bars refer to standard deviations with n = 6-8/group. Abbreviations: sHx = standard hepatectomy; eHx = extended hepatectomy; w/o resection = sham surgery; FSC = forward scatter signal; SSC = side scatter. Please click here to view a larger version of this figure.

Figure 8: Sudan IV staining as a marker of fat content of isolated fresh hepatocytes. (A) Hepatocytes from a fresh whole mouse liver without prior surgery. Increased size and fat content of hepatocytes 24 h (B) post 70%-hepatectomy and (C) 86%-hepatectomy. Note the concomitant presence of both larger, steatotic hepatocytes and smaller, non-steatotic hepatocytes. Scale bars = 30 µm (B-D). Error bars refer to standard deviations with n = 6-8/group. Abbreviations: sHx = standard hepatectomy; eHx = extended hepatectomy; w/o resection = sham surgery. Please click here to view a larger version of this figure.

A very low percentage of immune and non-parenchymal cells remain in the final suspension. Further purification is achieved by FACS or MACS, if desired. Negative selection of CD31⁺ and CD45⁺ cells by flow cytometry is used to demonstrate and quantify the non-parenchymal cells (Figure 5).

Figure 5: Flow cytometry gating strategy and selection of CD31^- and CD45^- parenchymal cells. (A) Gating chart with all events recorded; consider the relatively large size of hepatocytes and use adjusted voltages. Do not go higher than 350 V for FSC and 220 V for SSC. (B) Singlet gating. (C) Hepatocytes are selected by gating CD31^- (endothelial marker) and CD45^- (immune marker) cells. This selection can be performed in a fluorescence-activated cell sorter to select parenchymal cells for further downstream analyses, if needed; n = 4-5/group. Abbreviations: FSC = forward scatter; SSC = side scatter. Please click here to view a larger version of this figure.

To demonstrate the efficacy of the modified protocol in retaining steatotic cells, hepatocytes were isolated using the classic density-gradient approach⁷ (Figure 9A). Unlike with the modified procedure (Figure 9B), a fatty cell layer was clearly visible on top of the density gradient. Analysis of cells confirmed a gross absence of lipid-laden hepatocytes in the pellet following density-gradient centrifugation. In contrast, cells from the fatty layer were enriched with steatotic hepatocytes and a substantial fraction of non-parenchymal and dead cells (Figure 9A). Thus, classic isolation deprives the harvest of steatotic cells, while this modified protocol omitting the density gradient retains lipid-laden subpopulations, enabling the unbiased exploration of liver regeneration without skewing toward lean hepatocytes.

Figure 9: Yields of steatotic hepatocytes following the classic density-gradient isolation protocol compared to the improved protocol. (A) With the classic density-gradient purification method (final concentration of 90% density-gradient solution in phosphate-buffered saline), dead hepatocytes are collected in a cell layer on top of the density-gradient solution. (B) This layer does not only consist of non-parenchymal cells and non-viable hepatocytes but also of viable, large, fatty hepatocytes. (C) The pellet obtained following density-gradient centrifugation contains lean hepatocytes of smaller size and is mostly devoid of steatotic cells. This is the "pure" fraction that is isolated with the classic protocol. (D) With the improved protocol, the lipid-filled hepatocytes are not lost and all hepatocytes are pelleted. The supernatant (E) consists mostly of dead and fragmented hepatocytes and non-parenchymal cells, while the pellet (F) is a mixture of hepatocytes of variable size and lipid-content. Scale bars = 100 µm. Please click here to view a larger version of this figure.

Supplementary File 1: Supplementary protocols. (A) Lipid staining with Sudan IV; (B) standard and extended hepatectomy in mice. Please click here to download this File.

Supplementary Figure S1: Overview on the perfusion setup. (1) Cannulate the inferior vena cava. (2) Perfuse the liver with warm perfusion buffer to chelate calcium. (3) Warm the digestion buffer with collagenases and Ca²⁺ to dissociate cells in vivo. Remove the liver and (4) store in ice-cold preservation buffer (maximum 30 min). Please click here to download this File.

Supplementary Figure S2: Perfusion setup. (A) Perfusion table with an isoflurane nozzle, red light heating lamp, and perfusion tube. A heavy object can be placed underneath the tube to stabilize it and reduce the risk of displacement. (B) During the first few minutes of the perfusion, some blood will outflow through the portal vein and pool in the open abdominal cavity. (C) The abdominal wall is incised on one side to drain the blood. A cotton swab facilitates the drainage. Please click here to download this File.

Supplementary Figure S3: Glisson's capsule. (A) The Glisson's capsule is ruptured with fine-tip tweezers in a few locations along the liver surface, and the liver cells are released by gently shaking the capsule. (B) The liver should tear apart easily. (C) The empty liver sac (Glisson's capsule) with released hepatocytes in suspension. Please click here to download this File.

Supplementary Figure S4: Flow cytometry gating strategy for viability screening. (A) All events were recorded. (B) Singlet gating. (C) Live/dead staining with Alexa Fluor 488 Zombie green viability dye. Please click here to download this File.

Supplementary Figure S5: Flow cytometry gating strategy for granularity measurement. (A) All events were recorded. (B) Singlet gating. (C) Dot plot of side scatter (SSC; x axis) versus forward scatter (FSC; y axis) to analyze the levels of granularity. Please click here to download this File.

Supplementary Figure S6: Hepatocyte purification. (A) Cell pellet after the first centrifugation step; note the fatty layer on top of the medium. (B) Hepatocyte pellet after the second round of centrifugation. (C) Purified hepatocytes at the end of the purification process. Please click here to download this File.

Discussion

The published protocol provides a reliable and straightforward method to isolate a high yield of normal and steatotic murine hepatocytes for single-cell downstream analyses or bulk analysis of cells following FACS sorting. The distinct advantage over density-gradient purification is that the cellular lipid content has essentially no impact on the effective yield of hepatocytes. Thus, the fraction of steatotic hepatocytes will be retained and included in downstream analyses. This is not only crucial for the study of steatogenic liver diseases but also paramount for any analysis of processes following major hepatectomies, where hepatocytes display a temporally and spatially dynamic steatosis within the first 2-3 days of regeneration. Even though (single-cell) analyses of isolated hepatocytes have already been performed after hepatectomies^19,20,21, it must be assumed that the cell population analyzed was, at least partially, deprived of lipid-laden hepatocytes and the results, therefore, were not fully representative and skewed toward lean hepatocytes.

Currently, the function of TRAS is not fully clarified. For example, the spatial differences in lipid content within liver lobules remain ill-understood. Therefore, a method is needed that allows quantitative and qualitative detection without precisely excluding these cells from analyses (e.g., for single-cell transcriptomics, flow cytometry, and metabolomics).

Overall, hepatocyte yield with this improved protocol is markedly superior over other methods for the isolation of fat-containing hepatocytes (e.g., from mice with nonalcoholic steatohepatitis, NASH) using density-gradient purification¹⁵. However, cell yield, viability, and fat content can vary between preparations. Several factors seem to be responsible.

First, different batches of collagenase or collagenase blends will vary in their enzymatic activity; however, the lot-to-lot differences are not as critical as with traditional collagenases. Nevertheless, adjusting the working concentration may be necessary. The differences can be further minimized by proper storage until use (dry lyophilizates and rehydrated stock solutions at -15 to -25 °C) and by avoiding repetitive freeze-thaw cycles, but cannot be entirely eliminated. Therefore, it is recommended to use only collagenases of a specific batch for a given experimental series. In addition, the enzymatic activity depends on the temperature at which the digestion buffer reaches the liver, with an optimal temperature between 35 °C and 37 °C for a blend of collagenases I and II²². Lower temperatures will reduce the enzymatic activity, while temperatures above 50 °C can lead to irreversible denaturation of the protein. Test this in advance and optimize the experimental conditions by adjusting the initial temperature of the digestion buffer, the length and diameter of the plastic tube, and the flow velocity. For example, a drop in buffer temperature of up to 10 °C within a tube length of 60 cm is to be expected. Correct the temperature by using warming pads and infrared lamps, if needed. Collagenases display their optimum digestion capacity at a pH of 7.4; therefore, it is advised to adjust the pH of the buffer solutions already at a working temperature of approximately 37 °C before use.

Second, it is paramount to flush the liver (or the whole mouse) sufficiently with perfusion buffer before starting the digestion process to eliminate potential inhibitors such as serum or albumin. Ideally, perfusion is started only after the systemic circulation has stopped, to ensure the digestion buffer moves only from the vena cava through the liver to the outflow via the opened portal vein. The superior vena cava can be closed with a small vascular clamp. This prevents contact with residual blood components/inhibitors. Another method of optimally flushing the liver is intermittent clamping of the portal vein. This should be done carefully and - once the digestion buffer has reached the liver - performed only once to avoid damaging the already digested cells.

Third, regenerating hepatocytes are sensitive, and experience has shown that they require shorter digestion times and more careful handling to avoid over-digestion and harm to the cells. To harvest the liver as gently as possible, it is recommended to grasp the hepatic hilus with a clamp and then, first cut the superior vena cava and free the liver from the connections to the vein and the diaphragm. Subsequently, the inferior vena cava and the portal vein can be cut, which enables mobilization of the liver and easily frees it from the gastrointestinal ligaments, the esophagus on one side, and the right kidney on the other side. In hepatectomized mice, it is possible to carefully grasp one of the ligatures from the median lobes to remove the liver. If an attempt is made to pull the liver out by force, the Glisson's capsule may rupture, and the isolated cells could get lost.

Finally, it is essential that the perfusion procedure and further processing of the hepatocytes are not interrupted or delayed, as this will immediately and inevitably affect viability. Ideally, the perfusion is performed in a team of at least two people and on one animal at a time. These requirements in time and manpower, however, are one of the limitations of this method, although this applies to alternative methods as well.

Another limitation is the final purity of the resulting cell suspension, as the advantage of not losing steatotic hepatocytes to a density gradient results in a small proportion of remaining non-parenchymal and/or immune cells. In the shown sample, the fraction of CD45⁺ (immune cells) and CD31⁺ (endothelial cells) is below 5%, and when selecting for cell size first during flow-cytometry, the percentages of CD45⁺ and CD31⁺ are 1.2% and 2.2%, respectively (Figure 5). For most applications, this level of purity is sufficient, but highly sensitive methods or the further use in primary cell cultures probably requires a higher degree. Numerous studies have used CD31 and CD45 as systemic markers to sort hepatic cell populations by MACS^23,24 or FACS²⁵.

This improved isolation method is ideally suited for the study of liver regeneration, particularly the early period that presents with significant amounts of steatotic hepatocytes. PHLF, featuring persisting steatosis, is a particularly relevant subject requiring research. Following extended hepatectomies leaving behind marginal remnants, PHLF can develop and indeed represents the most common cause of death due to liver surgery. Its pathobiology is only partially understood, and currently no treatment is available²⁶. Given that PHLF significantly limits the application of liver surgery (such as for the cure of hepatic malignancies), exploring its causes and identifying potential measures is a clear medical need.

Need also exists for the study of diseases that share hepatic steatosis as a starting point. A prominent example is nonalcoholic fatty liver disease (NAFLD), which may progress to NASH and eventually to hepatocellular carcinoma (HCC)²⁷. The spread of sedentary lifestyles in conjunction with the Western diet has led to a worldwide epidemic of NAFLD, and accordingly, HCC incidence is projected to rise^28,29. Steatosis, in turn, is also tightly linked to obesity, the metabolic syndrome, and type-2-diabetes, all diseases with a rising incidence³⁰. Therefore, steatosis represents an increasing burden for the current healthcare system, calling for effective means for its management. This improved two-step collagenase perfusion method enables the study of steatotic hepatocytes at the single-cell level, and hence should aid in exploring the causes and consequences of hepatocellular lipid accumulation.

This protocol presents an easy, fast, and reliable technique for the in vivo isolation of regenerating hepatocytes from mice after partial (70%) and extended (86%) hepatectomy. This approach does not rely on gradient-purification and can be used to investigate liver regeneration processes with the inclusion of lipid-laden hepatocytes that are usually lost during traditional isolation protocols. Moreover, this method can also be applied for the research of various liver diseases associated with steatotic changes and is not limited to the mouse species.

Materials

Name	Company	Catalog Number	Comments
Reagents
Alexa Fluor 488 Zombie green	BioLegend	423111	Amine-reactive viability dye
Attane Isoflurane ad us. vet. 99.9%	Provet AG	QN01AB06	CAUTION: needs ventilation
EDTA solution	Sigma-Aldrich	E8008-100ML	-
Ethanol	Sigma-Aldrich	V001229	Dilute with water to 70%
Fetal bovine serum (FCS)	Gibco	A5256701	-
Hanks' Balanced Salt Solution (HBSS), Ca2+, Mg2+, phenol red	Sigma-Aldrich	H9269-6x600ML	For digestion/preservation
Hanks' Balanced Salt solution (HBSS), w/o Ca2+, w/o Mg2+, no phenol red	Sigma-Aldrich	H6648-6x500ML	For perfusion buffer
HEPES solution, 1 M	Sigma-Aldrich	83264-100ML-F	-
Histoacryl tissue adhesive (butyl-2-cyanoacrylate)	B. Braun	1050052	For stabilization of cannulation site
Hoechst 33258 Staining Dye Solution	Abcam	ab228550	-
Liberase Research Grade	Roche	5401119001	Lyophilized collagenases I/II
NaCl 0.9% 500 mL Ecotainer	B. Braun	123	-
Paralube Vet Ointment	Dechra	17033-211-38	-
Phosphate buffered saline (PBS)	Gibco	A1286301	-
Sudan IV – Lipid staining	Sigma-Aldrich	V001423	-
Temgesic (Buprenorphine hydrochloride), Solution for Injection 0.3 mg/mL	Indivior Europe Ltd.	345928	Narcotics. Store securely.
Trypan blue, 0.4%, sterile-filtered	Sigma-Aldrich	T8154	For cell counting
Williams’ Medium E	Sigma-Aldrich	W4128-500ML	-
Materials
25 mL serological pipette, Greiner Cellstar	Merck	P7865	-
50 mL Falcon tubes	TPP	-	-
BD Neoflon, Pro IV Catheter 26 G	BD Falcon	391349	-
Cell scraper, rotating blade width 25 mm	TPP	99004	-
Falcon Cell Strainer 100 µm Nylon	BD Falcon	352360	-
Fenestrated sterile surgical drape	-	-	Reusable cloth material
Filling nozzle for size 16# tubing (ID 3.1 mm)	Drifton	FILLINGNOZZLE#16	To go into the tubes
Flow cytometry tubes, 5 mL	BD Falcon	352008	-
Male Luer to Barb, Tubing ID 3.2 mm	Drifton	LM41	Connection tube to syringe
Petri dishes, 96 x 21 mm	TPP	93100	-
Prolene 5-0	Ethicon	8614H	To retract the sternum
Prolene 6-0	Ethicon	8695H	For skin suture
Prolene 8-0	Ethicon	EH7470E	Ligature gall bladder
Tube 16#, WT 1.6 mm, ID 3.2 mm, OD 6.4 mm	Drifton	SC0374T	Perfusion tube
Equipment
BD LSRFortessa Cell Analyzer Flow Cytometer	BD	-	-
Isis rodent shaver	Aesculap	GT421	-
Isofluran station	Provet	-	-
Low-speed centrifuge – Scanspeed 416	Labogene	-	-
Neubauer-improved counting chamber	Marienfeld	-	-
Oxygen concentrator – EverFlo	Philips	1020007	0 – 5 L/min
Pipetboy – Pipettor Turbo-Fix	TPP	94700	-
Shenchen perfusion pump – YZ1515x	Shenchen	YZ1515x	-
Surgical microscope – SZX9	Olympus	-	-
ThermoLux warming mat	Thermo Lux	-	-
Vortex Genie 2, 2700 UpM	NeoLab	7-0092	-
Water bath – Precision GP 02	Thermo scientific	-	Adjust to 42 °C