Comparison of Two Protocols of Carbon Tetrachloride-Induced Cirrhosis in Rats – Improving Yield and Reproducibility

Despite being a cardinal experimental model, the induction of cirrhosis in rats by repeated exposure to carbon tetrachloride (CCl4) has low reproducibility. Here, we compared two models of cirrhosis induced by orogastric administration of CCl4 once (CCl4-1xWk) or twice a week (CCl4-2xWk) for 12 weeks in male Sprague-Dawley rats. Control rats received water instead of CCl4. Both CCl4 protocols similarly attenuated body weight gain (p < 0.01 vs. Control). Although both CCl4 protocols increased hepatic fibrosis, portal hypertension and splenomegaly, the magnitude of these alterations was higher and more consistent in CCl4-2xWk rats. Importantly, two CCl4-1xWk rats did not develop cirrhosis versus a 100% yield of cirrhosis in CCl4-2xWk rats. The CCl4-2xWk protocol consistently induced liver atrophy together with hematological, biochemical and coagulation abnormalities characteristic of advanced cirrhosis that were absent in CCl4-1xWk rats. Ascites occurred in 20% and 80% of rats in theCCl4-1xWk and CCl4-2xWk groups (p < 0.01). All rats showed normal renal function, arterial blood gases and stable systemic hemodynamics. The total dose of CCl4 and mortality rate were similar in both protocols. The CCl4-2xWk protocol, therefore, was highly reproducible and effective for the induction of experimental cirrhosis within a confined time, representing a valuable advance for liver research.

the individualization of CCl4 doses according to the body weight or to the changes in body weight of the animal 7,9 , and diverse duration and schedules of CCl4 administration such as once 6,7 , twice 10 or thrice 11,12 a week or others 9,13 . Such variety of protocols reflect that the prior problems are still present. One of the most commonly used protocols is the one reported by Runyon et al. 6 , which is based on the administration of CCl4 by oral gavage once a week at a dose adjusted to the change of body weight 48 h after the last dose. Although high yields of cirrhosis (100%) and ascites (>90%) have been reported using this protocol, the associated mortality is consistently high (40-60%) and the duration of CCl4 administration required to develop ascites is extremely variable (6 to 20 weeks) from one rat to another even within single studies 6,[14][15][16][17][18] . In 2008, Regimbeau et al. reported a "rapid" protocol for the induction of cirrhosis involving the administration of CCl4 by oral gavage twice a week, with the CCl4 dose being adjusted to the body weight on the same day of treatment 9 . Although the reported yield of cirrhosis was 100% with an associated mortality of 30%, the later protocol has only been scarcely used by researchers. Importantly, there are no studies comparing two different protocols of CCl4-induced cirrhosis in parallel.
Here, we performed a comprehensive comparison between Runyon's protocol (CCl4-1xWk) 6 and a modified version of the "rapid" protocol reported by   9 . Importantly, the CCl4-2xWk protocol showed major improvements in terms of reproducibility and yield of cirrhosis, portal hypertension and ascites without increasing mortality.

Material and Methods
Animals. Sprague-Dawley rats (Rattus norvegicus, Charles River Laboratories) were bred in our animal facilities. Male experimental (F1) animals (130-200 grams body weight, 5-6 week-old) were maintained under constant conditions of temperature, air humidity, and a 12:12-hour light:dark schedule. All rats were allowed 7-10 days of acclimation to the room and to manipulation prior to the start of oral gavaging, and they had free access to standard chow and tap water throughout the experiments. All studies were approved by the Ethics Committee for Animal Experimentation of Hospital General Universitario Gregorio Marañón, and were conducted in conformity with the European Union Directive 2010/63/EU and the RD53/2013 of Ministerio de la Presidencia of Spain.
Experimental protocols for induction of cirrhosis. The rats were randomly divided in three groups: a) Rats receiving oral gavage with water (Control group, n = 11), b) Rats receiving CCl4 once a week (CCl4-1xWk group, n = 12), and c) Rats receiving CCl4 twice a week (CCl4-2xWk group, n = 15). Phenobarbital (35 gr/dl) was added to the drinking water of all rats from 2 weeks prior to the administration of CCl4/water until the termination of the experiments. CCl4 (diluted 1:1 with water, Sigma-Aldrich − 99.9% pure) or water alone were administered by orogastric intubation with a metal cannula without anesthesia or prior fasting.
Rats were treated for 12 weeks or until the development of overt ascites according to the following protocols (see Table 1): a) Control group: Rats received 0.5 ml of tap water once weekly (on Mondays). Body weight was measured before each administration and 4 days later (Monday and Friday). b) CCl4-1xWk group: Rats received CCl4 following the protocols published by Proctor and Chatamra 7 , later modified by Runyon et al. 6 . Briefly, the initial dose of CCl4 was 0.04 ml, and subsequent doses were administered once weekly (on Mondays) and adjusted based on the change in body weight measured 48 hours after the last dose. Each CCl4 dose was mixed with 0.5 ml of water before its administration, as opposed to the undiluted administration of CCl4 in the original protocol of Runyon et al. Body weight was measured before each administration and 2 and 4 days later (Monday, Wednesday and Friday). c) CCl4-2xWk group: Rats received CCl4 twice weekly for 12 weeks, with the CCl4 dose being adjusted on the body weight of the rats before each administration (see Table 1), based on the dosing reported in the 6-wk "rapid protocol" of Regimbeau et al. 9 with some modifications. In particular, CCl4 was diluted in 0.5 ml of water, and the doses were given on Monday and Fridays (in the original Regimbeau's protocol CCl4 was diluted 1:1 in olive oil and doses were given every 4 days). Hemodynamic measurements. A hemodynamic study was performed two weeks after the last dose of CCl4, a washout period intended to avoid the interference of inflammation associated with acute hepatic injury. The rats were anesthetized with sevoflurane (Abbott Laboratories), and the right common carotid artery and external jugular vein were dissected and canalized using a 24 G Abbocath catheter (B. Braun) and a polyethylene tube (PE50), respectively, to measure the mean arterial pressure (MAP) and the central venous pressure (CVP). Thereafter, a mid-laparotomy was performed and a 24 G Abbocath catheter (B. Braun) was inserted into the ileocolic vein to measure the portal pressure (PP). After 5 minutes of stabilization, blood pressures were registered for 5 minutes using pressure transducers and a multichannel PowerLab 8/35 and Lab Chart Reader software (AD Instruments) for analysis. Body temperature was monitored with a thermometer and maintained at stable levels with a warming pad throughout the experiment.
Hematological, biochemical and coagulation blood tests. Immediately after the hemodynamic measurements, arterial blood was collected in tubes containing EDTA, lithium heparin or citrate. Blood cell counting, biochemical, coagulation, and arterial blood gases analyses were performed in automated analyzers.
To confirm the platelet count and to assess if platelet aggregates were present, blood smears stained following the May-Grünwald method (Merck Millipore) were also evaluated.
Histology and quantification of fibrosis. The diagnosis of cirrhosis was based exclusively in histological criteria, namely the presence of architectural distortion of the liver characterized by the formation of regenerative nodules of hepatocytes surrounded by fibrous tissue. After the hemodynamic study and the collection of blood samples, the liver and the spleen were quickly excised and weighed. Portions of the left lateral and the median lobes of the liver were placed in 10% neutral buffered formalin, and processed for paraffin embedding. Tissue sections (6 μmm thick) were cut in a microtome, and stained with Hematoxylin-Eosin, Sirius red (Direct Red 80, Sigma-Aldrich) or Masson Trichrome (Bio-Optica Milano SpA). The stained slides were assessed in a fashion blinded to the groups by trained hepatologists for establishing the diagnosis of cirrhosis (JV) and for the quantification of fibrosis (JIF). For the later, digital images were captured with a Nikon Digital Camera DXM1200F coupled to a Nikon Eclipse E800 microscope, and the area of fibrosis was quantified in thirty 20x-magnification fields (Masson trichrome stain) or fifteen 10x-magnification fields (Sirius Red stain) per specimen. An RGB (Red, Green, Blue) threshold was used to identify the areas of fibrosis using ImageJ NIH software. Fibrosis was expressed as percentage (%) of total area.

Statistical analysis. Quantitative variables were expressed as median [interquartile range] and qualitative
variables as proportions (%), unless otherwise noted. One-way ANOVA followed by Tukey's post-hoc tests were used to assess differences between groups. In case of heterogeneity of variances, a Kruskal Wallis test followed by Dunn's post-hoc tests were used instead. In some cases of heterogeneity of variances (e.g. Sirius and Masson), a logarithmic transformation of the variables was performed prior to analysis. Comparisons of the evolution of body weight gain were analyzed by a Mixed-Model, with group and day as fixed effects and the rat identifier as a random effect, dismissing the baseline value (zero), and including the square of the variable "day" as a covariate and the interaction day-group. All comparisons were two-tailed, and a p value < 0.05 was considered statistically significant. The analyses were performed with GraphPad Prism v7 or IBM SPSS v21.
Data availability statement. Authors agree to make materials, data and associated protocols promptly available to readers upon requirement.

Results
General characteristics and mortality of the two experimental protocols of cirrhosis. . Although the impairment of body weight gain was similar in both CCl4 protocols, the rats receiving CCl4 once a week presented a highly serrated curve of body weight gain that contrasted with the smooth curve observed in the rats of the CCl4 2xWk group (Fig. 1 inset).
Mortality was also similar in both CCl4 protocols (CCl4-1xWk: 17% vs. CCl4-2xWk: 33%, p = 0.41). None of the control rats receiving water died. In the CCl4-1xWk group, 2 rats died during the wash-out period (13 th week). In the CCl4-2xWk group, there were 3 deaths due to accidental instillation of CCl4 into the trachea, and 2 deaths at the 2 nd and 10 th week of induction of cirrhosis. At the end of the study, there were 11 rats in the Control group and 10 rats in each of the CCl4 groups. Hematological, biochemical and coagulation parameters. Only rats receiving oral gavage with CCl4 twice a week presented blood analytical alterations suggestive of advanced liver disease. Compared with the Control and CCl4-1xWk groups, the rats in the CCl4-2xWk group showed leukocytosis (p < 0.05), thrombocytopenia (p < 0.05), biochemical alterations suggestive of liver damage (increased alanine aminotransferase and aspartate aminotransferase, both p < 0.01) and of decreased liver function, including parameters of cholestasis [increased alkaline phosphatase (p < 0.001) and bilirubin (p < 0.05)] and parameters reflecting poor synthetic function, such as increased INR (p < 0.01) and decreased circulating concentrations of fibrinogen (p < 0.05), total proteins (p < 0.01) and albumin (p < 0.001) (Fig. 2 and Table 2). Rats in the CCl4-1xWk group only showed increased levels of alkaline phosphatase compared with Control rats (p < 0.01) (Fig. 2E and Table 2). There were no significant alterations of renal function parameters, electrolytes, arterial blood gases or arterial lactate in any of the CCl4 groups compared with the Control group (Table 2). Systemic hemodynamics. Both groups of rats receiving CCl4 presented stable systemic hemodynamics (heart rate, MAP and CVP) at the end of the experiments. The MAP, however, was higher in the CCl4-1xWk group compared with the Control (p < 0.05) and CCl4-2xWk (p < 0.01) groups ( Table 3), suggesting that hyperdynamic circulation was more common in the CCl4-2xWk protocol. Rats in the CCl4-1xWk group showed mild increases of body temperature and respiratory rate (both p < 0.05 vs. Control) at the time of hemodynamic measurements, although all values were within the normal range ( Table 3).

Development of portal hypertension.
In line with the analytical and histological data, rats following the CCl4-2xWk protocol also showed increased portal pressure and portal hypertension-related complications such as ascites, splenomegaly and thrombocytopenia compared with rats in the CCl4-1xWk group. Thus, ascites was present in 0 of 11 (0%) Control rats, 1 of 10 (10%) CCl4-1xWk rats, and 8 of 10 (80%) CCl4-2xWk rats at the end of the experiments (p < 0.01 CCl4-2xWk vs. Control and CCl4-1xWk groups). Compared with Control rats, both groups of CCl4-1xWk and CCl4-2xWk rats developed splenomegaly (Fig. 6A) and portal hypertension (Fig. 6.B), but both parameters were higher in the rats following the CCl4-2xWk protocol (Spleen-to-body weight ratio:

Discussion
Repeated exposition to CCl4 in the rat is a cardinal experimental animal model of cirrhosis, but its reproducibility problems and excessive mortality are well known 4,7 . In the present study, we performed a side-by-side comparison of one of the most commonly used CCl4 protocols involving oral gavage once a week (CCl4-1xWk) 6 with a protocol involving oral gavage twice a week (CCl4-2xWk) based on the protocol reported by Regimbeau et al. 9 . The main finding of our study was that the CCl4-2xWk protocol was highly reproducible and largely superior to the  CCl4-1xWk protocol in terms of development of advanced cirrhosis, portal hypertension and ascites at 12 weeks of treatment without worsening mortality.
Reproducibility is a cornerstone of experimental models that facilitates the accomplishment of solid conclusions. Despite its low reproducibility, the experimental model of CCl4-induced cirrhosis has provided relevant insight into mechanisms of liver disease because CCl4-treated rats behave very different than normal rats despite the presence of widely heterogenous degrees of liver fibrosis in the former. Such heterogeneity, however, becomes a serious problem when assessing the effect of different treatments or manipulations within CCl4-treated animals, a problem increased due to the limited sample size of most experimental studies. Importantly, two of the rats (20%) following the CCl4-1xWk protocol had liver fibrosis but did not develop cirrhosis in our study. Furthermore, different degrees of portal hypertension and histological (see schematic in Fig. 5) or biochemical alterations ranging from normal to markedly disturbed were observed within the CCl4-1xWk group, with only 20% of the rats presenting ascites after the 12-week period. A common approach to solve this issue and obtain homogeneous groups with cirrhosis consists of treating the rats with CCl4 until they develop overt ascites [14][15][16][17][18] , but this introduces additional problems. First, the total dose of CCl4 is highly different between the rats. Second, the different age and body weight of the rats when the experiments are performed influence many variables, as the duration of treatment may vary from 7 to more than 20 weeks. Third, the duration of concomitant treatments or manipulations will also largely differ within single groups. Fourth, the appearance of overt ascites is not always immediately detectable de visu, particularly in older rats. Finally, some rats following the CCl4-1xWk protocol never develop ascites or die earlier due to other complications representing a relevant bias 17 . In contrast to this scenario, all rats following the CCl4-2xWk protocol in our study developed advanced cirrhosis and portal hypertension within 12 weeks of CCl4 administration, with ascites being confirmed in 80% of them. This was further supported by the presence of liver atrophy and the characteristic alterations of cirrhosis in blood analyses, including parameters related to liver damage (transaminases, AP), liver function (bilirubin, INR, albumin) or portal hypertension and hypersplenism (thrombocytopenia).
Despite the accelerated development of advanced cirrhosis in rats following the CCl4-2xWk protocol, mortality was similar in both protocols, and compared favorably with mortality rates (30% to 60%) reported in most studies 6,[14][15][16][17][18] . Notably, mortality may be further reduced as accidental instillation of CCl4 into the trachea, which occurred at a rate higher than expected according to our prior experience, was the cause of 3 of the 5 deaths (60%) in the CCl4-2xWk group. The smoother curve of body weight gain in the CCl4-2xWk group versus the serrated curve observed in the CCl4-1xWk group also suggests that the lower and more repeated doses used in the former protocol mimicked better the development of a chronic disease versus the higher acute doses used in the CCl4-1xWk protocol. Each image represents one rat, with their position being determined by the median area of fibrosis of the rat (shown in the inset). Rats following the CCl4-1xWk and the CCl4-2xWk protocols are shown, respectively, above and below the rectangle. Note that the extent of fibrosis in rats of the CCl4-1xWk group ranged from mild to severe, whereas the majority of rats in the CCl4-2xWk group presented severe fibrosis with fully developed cirrhosis in all of them. Within the CCl4-1xWk group, the rats showing 1.53% and 5.88% area of fibrosis did not fulfill histological criteria of cirrhosis. It is important to note that not all experimental animal models of liver disease express the disturbances characteristic of the syndrome of portal hypertension or can be used in several species 4 . In this regard, the administration of CCl4 accomplishes most requirements, being a feasible and not expensive model with minimal extrahepatic damage that can be used in different species (rats, mice, rabbits) and that closely resembles the biochemical, histological and hemodynamic alterations observed in human patients 19 . Among the different routes of administration of CCl4, the orogastric route is frequently used as it presents several advantages versus other alternatives. In particular, the administration of CCl4 by oral gavage requires low quantities of CCl4 and ensures its direct delivery to the liver through the portal vein, diminishing the extrahepatic effects due to the selective accumulation of CCl4 in the liver 20 . The inhalational route is also frequently used 21,22 , but it exposes extra-hepatic organs to CCl4 and is more expensive due to the need of large amounts of CCl4 and specific facilities to perform the administration (fume-hoods, filters). Importantly, CCl4 causes liver toxicity in humans and was reasonably anticipated to be a human carcinogen in the 12 th Report on Carcinogens 23 . Finally, the intraperitoneal route is convenient, but it interferes hemodynamic measurements due to the formation of adhesions and has the risk of vessel puncture and intraabdominal bleeding particularly in advanced liver disease presenting large portal-systemic collaterals.
In conclusion, our results are important for liver research as they validate the utility of a modified version of the protocol described by Regimbeau et al. 9 compared to a more widely used but inferior protocol described by Runyon et al. 6 . In particular, the CCl4-2xWk protocol accomplished the rapid development of advanced cirrhosis, portal hypertension and ascites in rats over a 12-week period of exposition to CCl4 in a highly effective and reproducible manner, without worsening mortality.