Experimental VX2 Rabbit Liver Tumor Model in Carbon Tetrachloride–Induced Cirrhosis of the Liver

Abstract

Liver cirrhosis is a major underlying factor in the development of hepatocellular carcinoma. Currently, there is an unmet need for midsize experimental vertebrate models that would offer reproducible implantable liver tumors in a cirrhotic liver background. This study establishes a protocol for a syngeneic rabbit model of VX2 liver cancer with underlying liver cirrhosis induced using carbon tetrachloride (CCl₄). Male New Zealand white rabbits (n = 3) received CCl₄ by intragastric administration once weekly. Concentrations started at 5% v/v CCl₄ dissolved in olive oil. CCl₄ dosing was progressively increased every week by 2.5% v/v increments for the duration of treatment (16 weeks total). VX2 tumors were then orthotopically implanted into the left hepatic lobe and allowed to grow for 3 weeks. Cross-sectional imaging confirmed the presence of hepatic tumors. Gross and histopathological evaluations showed reproducible tumor growth in the presence of liver cirrhosis in all animals.

Primary liver cancer is a leading cause of cancer-related deaths worldwide. Although the risk factors for liver cancer are highly variable, 90% of the cases arise in the setting of progressive inflammatory changes (1). In this context, many patients are diagnosed at advanced disease stages with limited treatment options and a poor prognosis (2). Advanced imaging and image-guided therapies (eg, ablation and embolization) are essential for the management of liver cancer across all disease stages (3). The development of novel imaging paradigms and image-guided therapies for liver cancer requires large animal models that would allow experiments on human-size scanners using translational protocols. Although some murine and rat models may offer faithful recreation of a cirrhotic liver background with orthotopic liver tumors (xenografts, syngeneic tumors, and genetically modified models), their small size remains a significant experimental limitation. This protocol introduces the combination of carbon tetrachloride (CCl₄)–induced liver cirrhosis (4) in a well-established VX2 rabbit tumor model and offers an experimental animal model suitable for translational imaging and image-guided interventions (5).

ANIMALS, MATERIALS, AND METHODS

Six male New Zealand white (NZW) rabbits (Charles River Laboratories, Boston, Massachusetts; age, 11–17 weeks; weight, 3.2 kg ± 0.05) were used under approved Institutional Animal Care and Use Committee protocols. Liver cirrhosis was induced by intragastric gavage of CCl₄ (diluted 1:1 with water, 99.9% pure; Sigma-Aldrich, Burlington, Massachusetts) once a week for 16 weeks (n = 3). Starting from an initial dose of 5% v/v CCl₄ (25 μL) in 1 mL of olive oil, the dose was increased by 2.5% v/v for each consecutive week until the end of the treatment period. The initial dose was modified from the published protocol by Fortea et al (6), in which the initial dose was 40 μL and mixed with water before its administration, as opposed to the use of olive oil. Untreated animals with VX2 liver tumors served as controls (n = 3).

Oral Administration by Intragastric Gavage

Animals were sedated with a single intramuscular injection of midazolam (5 mg/mL; Covetrus North America, Chicago, Illinois) and xylazine (100 mg/mL; Covetrus) at 1 and 9 mg/kg, respectively. These low dosages provided sufficient sedation for the procedure and minimized post-procedural recovery time. An 8-F feeding tube (833031, PFTM8.0P-NC; NeoMed, Woodstock, Georgia) was placed 30 cm down the esophagus using a laryngoscope equipped with a flat blade (size 0, catalog no. CIA2063790; Medline, Northfield, Illinois). A 3-mL syringe containing 1 mL of water was connected to the feeding tube. The plunger of the syringe was slowly pulled back to confirm the feeding tube location via aspiration of intragastric content. Next, 1 mL of CCl₄ in olive oil was delivered, followed by 1 mL of water.

Hepatic VX2 Tumor Implantation

At Week 16 of CCl₄ treatment, a 1–2 mm³ fragment of VX2 tumor chunks grown in an NZW donor rabbit was orthotopically implanted subcapsularily into the left anterior liver lobe as previously described (7,8) (Fig 1). After 21 days of tumor growth, animals were imaged using contrast-enhanced computed tomography (CT) and sacrificed. The livers and tumors were histopathologically evaluated for fibrosis and the presence of VX2 tumor.

Figure 1.

Timeline for carbon tetrachloride (CCl₄)–induced cirrhosis in rabbits. The schematic represents the animal experiment schedule in this study. The initial dose consisted of 5% CCl₄ (v/v) dissolved in 1 mL of olive oil (Day 0) and progressively increased by 2.5% for the duration of treatment (16 weeks). At Week 16, VX2 tumors were implanted into animals’ livers and grown for 3 weeks (Week 19).

Histological Evaluation of Liver Fibrosis

Tissue analysis included staining for hematoxylin-eosin, Masson’s trichrome, and sirius red scanned using the Aperio Scope XT Digital Slide Scanner (Aperio, Vista, California) at ×20 magnification (3.94 × 10⁶ pixels = 1 mm²). The evaluation of cirrhosis was based on the presence of architectural hepatic distortion, regenerative nodules, and fibrotic septa. A board-certified pathologist (X.Z.) confirmed the stage of cirrhosis across samples. Fibrosis index (FI) was calculated by the ratio of the fibrosis area to the total sample area, expressed in pixels, and calculated automatically using the Leica Aperio Image Scope Positive Pixel Count algorithm. The optimum threshold for positive pixels that corresponded to the areas of sirius red or trichrome staining was determined using the software by changing the hue value and color saturation using the original image for comparison. Tumor volume was calculated by the equation V = L × S × S/2, where “L” was the longest diameter and “S” was the shortest diameter (9). Tumor necrosis as a percentage of the whole tumor was measured using hematoxylin-eosin–stained tissue slides.

Contrast-Enhanced CT Acquisition

All acquisitions were performed using a human-size hybrid 64-slice single-photon emission computed tomography/CT Scanner (Discovery 570c; GE Healthcare, Chicago, Illinois). Rabbits were sedated by intramuscular injection of 30 mg/kg ketamine and were placed in supine position. After unenhanced scan acquisition, a triphasic contrast-enhanced scan was performed 15 seconds (arterial phase), 70 seconds (portal venous phase), and 5 minutes (delayed) after intravenous injection in the marginal ear vein of 1 mL/kg Omnipaque 350 at 1 mL/s followed by a saline flush (300 mg/mL; Omnipaque, GE Healthcare). The volumetric scanning parameters were: field of view = 22 × 22 cm, KVp = 120, mA = 80, and slice thickness = 0.5 mm.

Statistical Analysis

Statistical values are expressed as mean ± SD. Statistical analysis included t test, and single-factor analysis of variance for group comparison. A P value of <.05 indicates that there was a significant difference, and a P value of < .01 indicates that there was a very significant difference.

RESULTS

Pathological Outcomes

Tumor growth was confirmed on the CT scan in all animals (Fig 2a–d). Histopathological evaluation showed a single intrahepatic tumor growing in a cirrhotic liver (Fig 3a–c). Tumors were well delineated with a pseudocapsule and necrotic center containing necrotic cell debris. Mean percentage of tumor necrosis at pathological examination was 20.2% ± 3.42 for the control group and 32.23% ± 1.40 for the treated group. Large amounts of fibrous stroma and a few inflammatory cells were also present. Results of sirius red staining were similar to those of Masson trichrome staining. Control rabbits had a normal distribution of collagen (Meta-analysis of Histological Data in Viral Hepatitis score F0), whereas CCl₄-treated rabbits demonstrated fibrous septa forming nodules corresponding to the portal lobules with focal subsegmentations (Meta-analysis of Histological Data in Viral Hepatitis score F4) (Fig 4). Tumor size in the treated animals averaged 1.8 cm in the largest tumor diameter (range, 0.6 cm), which was not statistically different from that of control animals (P > .05). The FI corresponding to the blue color parameters with Masson’s trichrome staining in Image Scope revealed significantly more fibrosis in CCl₄-treated animals (31.3% ± 0.5) than in controls (7.25% ± 0.6; P < .01, using the Student t test) (Fig E1a–d, available online on the article’s Supplemental Material page at www.jvir.org). Although trichome-stained sections resulted in a higher contrast when setting thresholds for positive pixels, both methods provided clear differences in the percentage of fibrosis between the treated and control groups. In all treated rabbits, no signs of encephalopathy were observed. No obvious macroscopic portal-systemic venous collaterals were observed on imaging or gross pathological assessment on necropsy.

Figure 2.

Axial abdominal computed tomography in a cirrhotic rabbit 3 weeks after tumor implant. (a) Unenhanced, (b) 15-second arterial phase, (c) 90-second portal venous, and (d) 5-minute delayed phase images illustrate the 2.5-cm VX2 tumor mass with arterial phase hyperenhancement (oval) showing washout on portal venous phase and delayed imaging.

Figure 3.

Digitized hematoxylin-eosin histology samples of VX2 tumors from carbon tetrachloride (CCl₄)–treated animals at different magnifications. (a) Cirrhotic liver parenchyma (CL) and tumor (T) (left) and areas of viable tumor (VT), fibrotic stroma (FS), and pseudocapsule with ill-defined borders (right). (b) Histopathological evaluation of cirrhosis was assessed with Masson’s trichrome (left) and sirius red staining (right). (c) Gross pathology of CCl₄-treated rabbit showed the anterior (left) and posterior (right) images of VX2 tumor in a cirrhotic liver.

Figure 4.

Liver biopsies stained using sirius red and trichrome for both the control and carbon tetrachloride–treated rabbits. A Meta-analysis of Histological Data in Viral Hepatitis score of F0 indicated no fibrosis activity for the control, and a score of F4 indicated cirrhosis. Fibrosis index (FI) revealed higher percentage of collagen per area in the treated group using both staining methods.

DISCUSSION

This protocol induced advanced cirrhosis without mortality or complications and was based on the previous study by Brandão et al (10), who used phenobarbital in drinking water for microsomal induction of liver injury in combination with aminotransferase-level adapted dosages of CCl₄ for 16 weeks. Approximately 80% of the treated rabbits developed cirrhosis without ascites. However, in the present study, the authors were able to induce cirrhosis in all treated animals without the use of phenobarbital in drinking water. Despite well-defined cirrhosis, the animals did not develop ascites, a complication observed in rats that developed CCl₄-induced fibrosis (6). The success of this modified protocol may be attributed to the route of administration of CCl₄ and the chronic exposure to the toxin at relatively low doses. Because the intraperitoneal application of CCl₄ frequently results in higher mortality (<5%) than oral application, this protocol avoids animal loss by applying intragastrically lower dose(s) for adaptation (5% v/v CCl₄ [25 μL]), especially in cases of excessive weight loss (>10%) (11,12). Additionally, the orogastric route may ensure accumulation of CCl₄ predominantly in the liver, requiring low quantities of the agent to induce chronic liver injury compared with subcutaneous delivery (13). Furthermore, because rabbits are sensitive to CCl₄, gradual and continued exposure to the agent may reduce risk of renal injury and other morbidities, such as ascites (14). Thus, the choice of a low initial dose of the hepatotoxin with a weekly progressive dosage increase of 2.5% enabled chronic exposure to the agent without major complications.

This work addresses one of the major limitations of the current rabbit VX2 tumor model, in which the crucial influence of the cirrhotic tumor microenvironment on early carcinogenesis and tumor progression is not considered. Although other models of large animals for hepatocellular carcinoma have been described in the context of underlying liver fibrosis (15), the VX2 tumor model has the advantage of growing quickly (15 days vs 8–12 months for pig hepatocellular carcinoma) and has been extensively validated for interventional radiology because of the host’s size compatibility with devices used in humans (16). However, this model still has the limitation of using a tumor of non-hepatic origin and the spontaneous necrosis that may increase with tumor size (17). Additionally, histopathological assessment of the treated livers did not reveal the presence of CCl₄-induced autochthonous tumors. However, these data are limited considering the relatively short period of exposure to CCl₄ (16 weeks).

Future studies with this model need to be expended to a larger group of animals and could envision including laboratory parameters and measurements of portal hypertension, the most frequent complication of cirrhosis and a well-described feature associated with the CCl₄-induced model (18). Presently, this study is a “proof of principle” that successful subcapsular engraftment of VX2 tumors can be achieved in cirrhotic livers. However, additional studies investigating successful engraftment of tumors through intraportal injection could be interesting.

In conclusion, this study’s protocol successfully induced liver cirrhosis in an NZW rabbit VX2 liver tumor model, thus validating the utility of a modified version of the protocol described by Brandão et al (10). In particular, the development of advanced cirrhosis independent of phenobarbital administration is effectively achieved in a reproducible manner, without causing complications or mortality.

STUDY DETAILS.

Study type: Animal study

ABBREVIATIONS

CCl₄

carbon tetrachloride

CT

computed tomography

FI

fibrosis index

NZW

New Zealand white