Extended 78% Hepatectomy in a Mouse Surgical Model

Mouse and rat surgical liver resection models, first described in 1931, are the most common experimental models utilized to study the molecular basis of liver regeneration. They could also be useful in translational science research to test and develop strategies to improve outcomes following extended liver resection or transplantation of suboptimal liver grafts^1,2,3,4. Partial hepatectomies (PH) in mice entail the removal of approximately 2/3 (66%) of the total liver mass (TLM), which when performed in healthy animals have exceptional outcomes⁵. The procedure is short in duration, easily reproducible due to little variation in mouse liver anatomy, and postoperative survival typically nears 100%¹.

Partial 2/3 hepatectomy encompassing the resection of the left lobe (LL) and median lobe (ML) allows for the residual lobes to regenerate relatively unimpeded by lobar inflammation or restriction of hepatic inflow and outflow. Rather, increased portal venous flow and subsequently shear stress on liver sinusoidal endothelial cells following PH result in sustained upregulation of endothelial nitric oxide synthase (eNOS) expression and subsequent nitric oxide (NO) release, which contribute to the priming of hepatocytes for proliferation and liver regeneration³. Outcomes commonly studied after 2/3 PH in disease models such as non-alcoholic fatty liver disease or in specific genetic backgrounds include risk of acute liver failure, qualitative and quantitative measures of the liver regenerative capacity, and other biologic responses to stress or traumatic injury^1,3.

However, a mouse model mimicking functional or anatomical small-for-size syndrome, as it occurs following extended liver resection for cancer or transplantation of marginal (steatosis or prolonged ischemic time) or partial (split or from living donor liver) liver grafts, remains to be well-established. To address this need, models of more extensive liver resections that extend beyond the maintenance of a minimal (and functional) liver mass are required to model small-for-size liver syndrome and the heightened mortality that is associated with this syndrome^6,7.

Mouse liver anatomy exhibits minimal variation. The mouse liver is comprised of five lobes, each accounting for the following percentage of the total liver mass: left lobe (LL; 34.4 ± 1.9%), median lobe (ML; 26.2 ± 1.9%), right upper (also called right superior) lobe (RUL; 16.6 ± 1.4%), right lower (also called right inferior) lobe (RLL; 14.7 ± 1.4%), and caudate lobe (CL, 8.1 ± 1.0%)^1,5. Each lobe is supplied by a portal triad, including a branch of the hepatic artery, a branch of the portal vein, and a bile duct⁵. Historically, several techniques were described to perform a 2/3 PH by resecting the LL and the ML. These include 1) the classical technique that consists of a single ligature en bloc at the base of each of the resected lobes; 2) the hemostatic clip technique, using titanium clips applied at the base of the resected lobes; 3) a vessel-oriented parenchyma-preserving technique, using piercing sutures proximal to the clamp; and 4) a vessel-oriented microsurgical technique, whereby the portal vein and hepatic artery branches are ligated prior to lobe resection¹. While each technique has relative strengths and weaknesses, none yields higher lethality^1,8,9.

In this study, we present a novel method for extended 78% PH in mice. In this model, three of five liver lobes, including the LL, ML, and RUL, are removed separately using a ligature technique (Figure 1). This procedure results in the resection of approximately 78% (77.2 ± 5.2%) of the total liver mass. Our choice of removing the LL and ML separately, and not "en bloc" as in the classical PH technique, minimizes complications that are associated with en bloc resection of these two lobes, such as suprahepatic vena cava stenosis and heightened risk of necrosis of the remaining lobes when the single ligature is applied too close to the vena cava^{1,10,11,12,13,14}. This is crucial before moving to the final step of this procedure to remove the RUL. This extensive hepatectomy in 8-12 weeks old, wild-type C57BL/6 mice causes 50% lethality within 1 week of surgery due to failed liver regeneration causing fulminant liver failure^15,16. This mouse model of heightened lethality following extended 78% hepatectomy appropriately recapitulates the pathophysiology of small-for-size syndrome and enables the development and testing of novel strategies to improve outcomes.

The methods described within this procedure protocol have been approved by the Institutional Animal Care and Use Committee (IACUC) at the Beth Israel Deaconess Medical Center (BIDMC). All experiments were completed in accordance and compliance with IACUC and the BIDMC animal research facility guidelines.

1. Mouse preoperative preparation

1. Shave the mouse abdomen from the mid-sternum to the suprapubic region with clippers.
2. Induce general anesthesia with 1-4% isoflurane in 100% oxygen. When anesthetized, place the mouse supine on the surgical field with a heating pad underneath. Before making an incision, firmly pinch the toe to ensure the pedal reflex is absent (if present, the animal will respond). Adjust the anesthetic level as needed to achieve a state of general anesthesia.
   NOTE: Titrate isoflurane as needed to maintain adequate general anesthesia.
3. Administer 1.2 mg/kg buprenorphine extended-release (ER) subcutaneously for postoperative analgesia. Place the mouse supine with the forelimbs and hindlegs extended and secure the limbs with tape; then, prepare a sterile field for surgery.
   NOTE: Ensure the forelimbs are relaxed when secured so respiration is not hindered.
4. Prepare the abdomen with warm sterile saline and betadine swabs, alternating between each swab 3 times. Drape the abdomen in a sterile fashion.

2. Hepatectomy

1. Make a vertical midline laparotomy incision through the skin from the xiphoid process to the suprapubic region using a scalpel. Next, incise through the linea alba with sharp scissors to enter the peritoneal cavity and extend this incision to the length of the skin incision.
   NOTE: It is safer to first incise the linea alba at the subxiphoid region where the liver is deep to the abdominal wall to avoid injury to the underlying bowel.
2. Retract the abdominal wall laterally using appropriate retractors; then, clamp the xiphoid process with a hemostat and retract the sternum superiorly to expose the liver.
3. Retract the liver inferiorly to expose the falciform ligament and then, transect the ligament along the length of the liver using sharp scissors. Retract the liver superiorly towards the thorax to expose the hepatogastric ligament and intrahepatic lobe ligaments and transect these structures using sharp micro scissors.
   NOTE: Retraction should be performed very gently with moist cotton-tipped applicators as the liver, encapsulated by the Glisson's capsule, is very fragile and easy to bruise and lacerate.
4. Retract the median lobe superiorly while keeping the left lobe in its original anatomic position. Wrap a 5-0 silk suture around the superior-medial part of the LL. Reflect the LL superiorly towards the thorax to expose the undersurface of the lobe, bring together the suture ends at the base of the lobe, and tie the suture at the base. Ensure that the suture does not obstruct the blood flow in the inferior vena cava (IVC) or the portal vein before tying the suture to ligate the LL.
   NOTE: It is best to tie this suture while the LL is reflected superiorly towards the thorax so that the portal triad is well exposed during ligation. This facilitates resection of the lobe close to the base without compromising adjacent structures.
5. Resect the LL just distal to the suture tie using sharp scissors and ensure that a small cuff of tissue (~2 mm) separates the suture from the edge of the resected lobe. Confirm hemostasis.
6. Reflect the median lobe superiorly towards the thorax, place a 5-0 silk suture around the base of the ML, and return the ML to its original anatomic position. Approximate the suture ends over the base of the superior aspect of the ML and ensure tie them at the base of the lobe. Resect the ligated ML, leaving a small cuff of remnant tissue around the suture tie. Confirm hemostasis.
7. Mobilize the liver from right to left to expose the right upper and lower lobes and carefully reflect these lobes medially and inferiorly. Wrap a 5-0 suture over the superior-medial aspect of the RUL to ensure the suture encircles the RUL base and then, reflect the RUL towards the thorax. Wrap the suture under the RUL and tie the ends close to its base, then resect it, leaving a small cuff of remnant tissue around the suture tie.
   NOTE: Tying too proximal at the base of the RUL can compromise the blood supply to the RLL, which can result in ischemia of the RLL and death of the mouse within 24 h postoperatively. In contrast, tying too distally from the RUL base decreases the amount of resected of liver mass, thereby increasing postoperative survival rates beyond what is expected.
8. Return the remaining liver to its resting anatomic position and ensure hemostasis. If required, apply pressure with gauze to areas of minor bleeding at resected liver margins.
9. Close the midline abdominal wall (fascia and muscle layers) using a 5-0 polyglactin suture in an uninterrupted fashion. Close the skin incision with staples or 5-0 monofilament sutures.
10. Discontinue anesthesia and monitor the mouse until it regains consciousness and can ambulate normally.

3. Postoperative care

1. Observe the mouse postoperatively to ensure appropriate recovery (i.e. the mouse is awake, alert, and ambulatory within the cage) and pain control. Examine the mouse every 2 h up until 6 h after the procedure, and then daily.
   NOTE: It is expected that the mouse will move slower within the cage postoperatively. The mouse should recover in an isolated cage from other mice and only returned to the company of other mice when it is fully recovered.
2. Administer warmed normal saline injections (0.1-1.0 mL, subcutaneous or intraperitoneal) if the mouse becomes hypovolemic from insensible fluid or excessive blood loss from surgery.

A successful extended 78% hepatectomy is expected to induce 50% mortality within 1 week in healthy adult mice aged 8-12 weeks¹⁶. When properly performed, minimal blood loss is expected. Residual bleeding that persists can be controlled by manual pressure. Perioperative death within 24 h of surgery is often caused by technical errors. Technical failures include inadvertent injury to large blood vessels causing intractable intraoperative hemorrhage; significant postoperative hemorrhage often due to residual bleeding from the resected liver margins; injury to surrounding structure such as inadvertent ligation of the adjacent portal triad, the portal vein, or the IVC; ischemia of the RLL secondary to ligation of the RUL too proximally to the lobe base; and complications from general anesthesia. Signs of hepatic failure include progressive lethargy, hair clumping, anorexia, and hypoglycemia, which often become apparent within 24 h of surgery.

The two expected outcomes after extended 78% hepatectomy are either survival or death. In the first scenario, the mouse recovers appropriately after surgery, resumes normal activity within 72 h, and survives beyond 7 days. Laparotomy performed 9-10 days postoperatively demonstrates full recovery of presurgery liver mass^4,16,17,18. The second outcome would be mortality within 2-7 days postoperatively. The mouse may show initial signs of recovery in the first 12 h after surgery but worsens relatively quickly afterwards due to the development of fulminant liver failure¹⁶. The mouse exhibits signs of stress, weight loss, and progressive lethargy until death. Examples of unexpected outcomes include death from technical complications or mouse survival rates well above or below 50% after extended 78% hepatectomy.

Our laboratory pioneered this extended 78% hepatectomy and validated its usefulness in an earlier manuscript showing that extended 78% hepatectomy in 8-week-old healthy BALB/c mice results in 50% lethality, which could be abrogated if the mouse liver was preengineered to express A20 (Tumor Necrosis Factor Inducible Protein 3 [TNFAIP3]), delivered intravenously using a recombinant adenovirus¹⁶. In this study by Longo et al., a healthy BALB/c mouse underwent a 78% percent hepatectomy 5 days following the administration of a recombinant adenovirus vector expressing human A20 (rAd.A20) or the control β-Galactosidase (rAd.βGal). An additional non-treated control group was also included. Following 78% hepatectomy, Longo et al. observed that 6 of the 12 (50%) non-treated mice survived the procedure (Figure 2)¹⁶.

Two determinants of proficiency include 1) limiting technical complications, as listed above, chief of which being RLL ischemia, causing early death, and 2) ensuring sufficient mobilization of the RUL, without which one fails to appreciate the full size of the RUL and hence, fails to properly resect it. This reduces the amount of liver mass resected and therefore, improves the overall survival rate well above the expected 50%. In the initial period of training of the investigator/author, 10 of 15 (67%) healthy adult male and female CD1 and C57BL/6 mice aged 11-21 weeks survived 1 week following 78% hepatectomy (3 technical deaths excluded) (Table 1)^19,20,21. Further training in the context of an appreciative understanding of mouse liver anatomy to improve RUL mobilization and facilitate an adequate resection, measured by estimated percent hepatectomy, was essential to achieve the expected 50% survival rate following 78% hepatectomy. After technical proficiency was felt to be fully achieved, 8 of 16 mice (50%; 1 technical death excluded) survived 1 week following 78% hepatectomy.

Figure 1: Mouse hepatectomy positioning and liver anatomy. (Left) Visual representation of mouse positioning for laparotomy and extended 78% hepatectomy. The mouse is depicted supine with a midline laparotomy incision. (Right) Mouse liver anatomy depicted from below with colored lines delineating resection sites. The left lobe, median lobe, and right upper lobes are resected in a sequential fashion following suture ligation at their base. Please click here to view a larger version of this figure.

Figure 2: Survival advantage following extended (78%) hepatectomy of mice engineered to overexpress A20. Survival data following 78% extended hepatectomy in control untreated (non-infected NI), rAd.βGal, and rAd.A20-treated mice. Overexpression of A20 in mouse livers yielded a significant survival advantage compared with NI (50%) and rAd.βGal controls (13%) (n = 12 mice/group). This figure is from Longo et al.¹⁶. Please click here to view a larger version of this figure.

Table 1: Extended 78% hepatectomy training results. Survival results within 1 week following extended 78% hepatectomy in healthy adult CD1 and C57BL/6 mice aged 11-21 weeks. Technical proficiency was self-determined via observable improvements in technical success and adequate resection of the right upper lobe, as calculated by estimated percent hepatectomy (Mean ± SEM).

To successfully perform an extended 78% hepatectomy causing 50% lethality in mice, it is critical that each liver lobe is precisely resected. This level of competency and precision can only be achieved if the procedure is performed repeatedly. The training curve varies between operators but typically requires 3-6 months of practice. A liver resection that removes less than 78% of the TLM would result in higher survival rates, while a liver resection that removes greater than 78% of the TLM would result in greater lethality. Each lobe resection is challenging, albeit not to the same extent.

The left lobe is the easiest to reliably resect. The LL base is narrow, and when reflected superiorly, the operator can easily identify the liver base and tie the suture in the same position with near-identical lobe resection volume each time. The median lobe has the widest base when compared to the LL and RUL. Therefore, it requires that the operator carefully estimate where the suture needs to be placed and carefully approximate the suture ends at the base of the ML prior to resecting the lobe. When the ML is tied too proximally, the suture may compromise venous liver drainage or impede blood return from the IVC to the heart. When the ML is tied too distally, insufficient liver mass is resected, and the risk of hemorrhage at the resection margin increases since the ML base is wider. The right upper lobe is perhaps the most difficult to reliably resect. The anatomic position of the RUL posteriorly in the peritoneal cavity makes it difficult to wrap the suture completely around its base, which could result in an incomplete resection of this lobe. In contrast, if the suture is tied too proximally at the RUL base, blood supply to the right lower lobe might be jeopardized, causing ischemia of this RLL and increasing the likelihood of postoperative mortality.

Other critical elements to minimize procedure-related risks include the minimization of general anesthesia (e.g., Isoflurane) to reduce toxicity and ensuring adequate hemostasis after each lobe resection to limit postoperative hemorrhage. It is important to consider that the extended 78% hepatectomy is preferably performed in adult 8-12-week-old mice, as older mice can exhibit more variability in survival rates due to their greater body mass and reduced regenerative liver capacity, while younger mice might suffer greater technical complications due to the smaller size of their liver and a higher rate of anesthesia-related complications. We surmise that the observed 50% lethality following 78% hepatectomy corresponds to intrinsic single mouse characteristics which relate to subtle anatomical variations in the relative percentage of each liver lobe mass relative to total liver mass between individual animals. The 78% hepatectomy in mice represents an anatomic threshold at which only 50% of the animals can timely and successfully regenerate and survive while the other 50% fail to do so and die. We also acknowledge that subtle differences in liver manipulation might be associated with different degrees of liver damage, and hence skew this fine balance towards failure to regenerate and death²².

Albeit, the most limiting factor remains the mastery of the surgical procedure itself, which can only come with practice. Practice is essential to ensure a reproducible outcome through precise mapping of the sutures at the base of each liver lobe. One cautionary note is that individual variations in the liver anatomy amongst mice - which are rare - may require some modification in technique. Other interventions that should be considered on a case-per-case basis to improve success include administration of normal saline boluses in case of high insensible fluid losses or significant hemorrhage, and prolonged manual pressure or electrocautery at the liver margin in instances of persisting hemorrhage.

In summary, the extended 78% hepatectomy in a mouse model is a valuable technique for translational science research. Extensive training is critical to achieve a technically successful outcome associated with this procedure. Mice are not only a preferred small animal species that are easy-to-handle and relatively inexpensive, but also available in a number of well-studied inbred strains in addition to an ever-growing number of genetically modified lines (transgenic, knockout (KO), cell-type specific, and conditional KO), enabling fine mechanistic studies^3,9,23. In addition to genetically modified mice, various pathologies, including non-alcoholic fatty liver disease, cirrhosis, and diabetes that are known to influence survival and outcomes after liver resection, can be easily induced in mice^{24,25,26,27,28,29,30,31}.

As mentioned earlier, the classic 2/3 PH remains extremely valuable but does not recapitulate the high lethality that is associated with extended liver resection for cancer or small-for-size syndrome following liver transplantation when the liver mass is anatomically or functionally inadequate (such as in fatty livers)^7,32,33,34. When properly performed, this extended 78% hepatectomy results in 50% postoperative death, which better reflects clinical reality such as following extensive liver resections for trauma or cancer and in the context of small-for-size syndrome following transplantation of marginal liver grafts, and also after mere liver surgery in patients with severe non-alcoholic steatohepatitis (NASH) or cirrhosis^16,32. This mouse model represents a highly valuable and necessary proof-of-concept step to test novel therapeutic strategies to improve outcomes in all these conditions. Any positive results in mice are bound to substantially decrease the number of animals required to conduct pre-translational large animal studies prior to clinical translation of innovative therapies.