Hepatoprotective Effect of Water Soluble Extract of Coleus Barbatus on Cholestasis on Young Rats

1 Efeito hepatoprotetor do extrato aquoso de Coleus barbatus na colestase em ratos jovens ABSTRACT Purpose: To test the effects of water extract of Coleus barbatus (WEB) on liver damage in biliary obstruction in young rats. Methods: Forty 21 day-old male Wistar rats were divided into four groups of ten 21 day old (P21) submitted to sham or actual operation (S or L) combined with WEB or Water (B or A). At P48 pentobarbital sleeping time (ST) was measured. At P49 they were submitted to euthanasia to determine of serum activities of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), liver wet weight (PFF) and, on hepatic histological slides, the frequency of mitoses (FM), the number of necrotic areas (NN), intensity of fibrosis (IF) and intensity of ductal proliferation (IPD). Two Way ANOVA, the S.N.K. test and the Wilcoxon test for paired multiple comparisons were employed to study the effects of cholestasis and those of EAB and their interactions. The Pearson's coefficient of linear correlation of between paired histological variables separately for the groups LA and LD was determined. The test results were considered statistically significant when the p of α error <0.05. The EAB decreased the TS and IM in the animals without cholestasis (sham operated animals). The EAB decreased the TS, ALT, AST, PFF, MI, NN and IF of the cholestatic animals. In the LA group there was a positive correlation between the IPD and the IF, a negative correlation between the IPD and the FM and a negative correlation between the IF and the FM. In the LD group there was a negative correlation between the NN and the IPD. Conclusions: In the absence of cholestasis the EAB reduces the pentobarbital sleeping time and decreases the frequency of mitoses. The EAB has a hepatoprotective effect on the biliary cirrhosis secondary to extra-hepatic biliary obstruction. RESUMO Objetivo: Testar os efeitos do extrato aquoso de Coleus barbatus (EAB) na cirrose biliar secundária por obstrução das vias biliares extra-hepáticas em ratos jovens. Métodos: Quarenta ratos Wistar machos com 21 dias de vida (P21), foram distribuídos em quatro grupos de 10 animais, submetidos a operação simulada ou dupla ligadura e ressecção do ducto biliar (S ou L) combinados EAB e a Água (B ou A). No P48, foi medido o tempo de sono com o pentobarbital (TS). No P49, foram submetidos a eutanásia para a determinação das atividades séricas do aspartato aminotransferase …


Introduction
The main indication for liver transplantation in infants and children is biliary atresia, a severe cause of hepatobiliary disease which presents a mortality rate of 100% when untreated 1,2 .
Ligature and resection of the common bile duct in young rats has been used as a model for obstructive cholestasis and the consequent secondary biliary cirrhosis in infants 3 .
Coleus barbatus from the Lamiaceae family 4 , commonly known as National Boldo, Brazilian Boldo or False Boldo is used in popular Brazilian medicine in the treatment of liver problems 5,6 .However, there are no reports in the literature of experimental or clinical studies on the effects of Coleus barbatus on cholestasis or other hepatic diseases.A few plants extracts have been tested in bile duct ligated adults rats: Salvia miltiorrhiza 7 ; Camellia 8 and Stephania tetrandra 9 .
Our aim was to test possible hepatoprotective effects of a water extract of Coleus barbatus in the model of obstructive cholestasis.To that effect we studied in four groups of young rats (bile duct ligated plus WEB or water and cham operated plus WEB or water) the following variables to ascertain the extent of liver disfunction and damage.
Pentobarbital sleeping time is a easily performed liver function test in rats 10 which evaluates the oxidative metabolism of drugs, which is impaired in chronic cholestatic disease that reduces the activity of enzyme cytochrome P450 11 .Pentobarbital sleeping time was used in several studies to demonstrate hepatoprotective effects of plant extracts (Calotropis procera 12 , Artemisia scoparia 13 , Cyperus scariosus 14 and Bougainvillea spectabillis 15 ).
The serum aminotransferase activities (alanine aminotransferase -AST and aspartate aminotransferase -ALT) are biochemical markers of the rate of hepatocelular lesions 16 .In experimental cholestasis, through ligature and common bile resection, a significant increase in these cellular enzymes occurs 16 .Fujii, 1997 17 summarized the toxicological correlation between serum biochemical tests and the hepatic histopathology found in his own experiments and those published by other authors, with rats and dogs, and concluded the increase in ALT and AST presents a correlation with the intensity of necrosis.
Hepatic histopathology in laminas of the liver with cholestasis showed the intensity of the following lesions 18,19 : a) hepatocyte coagulation necrosis 20 ; b) fibrosis 21 ; c) ductal proliferation 22 .The regeneration of liver cells can be observed by the presence of mitosis 23 .

Collection and preparation of the water-soluble extract of Boldo (WEB)
The plant (Coleus barbatus) was collected fresh, in its natural habitat, in the district of Rubião Jr., Botucatu, during February 2000.It was identified and registered under the number BTU24169, in the Herbarium of the Botanical Department of the Biosciences Institute, UNESP -São Paulo State University, Botucatu, São Paulo.The leaves were oven-dried at 60ºC for three days and then grounded with a pestle until they constituted a fine powder.This was stored in hermetically sealed plastic flasks and labeled.The water-soluble extract was prepared by passing one liter of water at room temperature through a funnel with 60g of the powder placed in filter paper.The filtrate was stored in 10ml amber colored glass flasks, conserved in a freezer at -20ºC for up to eight weeks, while thawed aliquots were maintained in a refrigerator at -4ºC for use in a maximum of a two day period.
Forty male Wistar rats were studied, born in the Pediatrics Experimental Laboratory of the Botucatu Faculty of Medicine -UNESP, of nulliparous females, weighing 180 to 220 g at the onset of gestation, provided by the Biotério Central of the Botucatu Campus -UNESP.From birth they were suckled in groups of six animals per dam until the 21 st day of life (P21).During this period and for the whole experimental period (P21 to P49), the animals were maintained under controlled environmental conditions -temperature between 22 to 23ºC and a 12 hour light/dark cycle.

Experimental procedure
The experimental procedure was approved by the Ethics Commission in Animal Experimentation of the Botucatu Faculty of Medicine / UNESP -São Paulo State University, protocol number 64, 25/10/1999, in accordance with the Principles of Ethics in Animal Experimentation of the Brazilian College of Animal Experimentation (COBEA) 24 .The maintenance of the animals and anesthetic, surgical and euthanasia procedures were performed according to the recommendations of the Canadian Council of Animal Care (1984) 25 .
On P21, each animal was randomly assigned to one of the four experimental groups of 10 animals each: SA Group: sham operation with daily administration, by gavage, of 5µl of water per gram of animal weight, SB Group: sham operation with daily administration, by gavage, of 5µl of water-soluble extract of Boldo (WEB) per gram of animal weight, LA Group: double ligature and resection of the common bile duct with daily administration, by gavage, of 5µl of water per gram of animal weight, LB Group: double ligature and resection of the common bile duct with daily administration, by gavage, of 5µl of WEB per gram of animal weight.
All animals had ad libitum access to water and standard feed with 18g of protein per 100g of feed prepared by the Pediatrics Experimental Laboratory of the Botucatu School of Medicine of the São Paulo State University-UNESP (Table 1).

Statistical procedures
The mean and standard deviation or median values of each variable for each group were calculated.The results of each variable for each experimental group were compared according to the following steps: 1) testing for normality and equality of variances of the values obtained for each group; 2) if the data passed these tests, a two way analysis of variance was applied, where one factor was the administration or non-administration of WEB and the other the presence (double ligature and common bile duct resection) or absence of cholestasis (sham operation).The effects of each factor were tested independently of the effect of the other, as well as the interaction between both factors; 3) when a significant interaction between factors was found (p≤0.05),multiple paired comparisons were performed using the Student-Newman-Keuls (S.N.K.) method, identifying pairs of groups with significant differences among them (p of α≤ 0.05); 4) if the measurements did not satisfy the criteria of item 1, mathematical transformations were performed and the transformed values that passed the tests where employed in the calculations as in item 2 above.Pentobarbital sleeping time and aminotransferase were substituted by their rank number.

5) histological variables (frequency of mitosis, number of areas of necrosis and intensity of ductal proliferation)
were analysed by means of a factorial delineation using a generalized linear model, presuming a Poisson's distribution and a canonic (logarithmic) connection.A 1/6 value was added to the observations for a better fit to the model.For this analysis the means of the 30 readings of each group were used, n=30 (3 readings per animal from each experimental group).However this model did not result in an adequate fit: in the case of the number of mitosis excessive dispersion occurred and in the case of the variables number of areas of necrosis, intensity of fibrosis and the intensity of ductal proliferation a sub-dispersion occurred.Even when the interaction was not significant, we opted to perform the decomposition of the effects of the drug and ligature factors in the nestling effect which were tested using the Wilcoxon test (p<0.05).
Pearson's analysis of linear correlation was applied to paired histological variables (p of α<0,05) separately for groups LA and LB..For this analysis the means of the 3 measurements per animal from each experimental group (n=10) were used.
Table 2 shows the mean and standard deviation values of the variables for which no mathematical transformations were necessary and the median of those values for which transformations were necessary.The graphs in figures 1 and 2 show the means (solid points), the standard deviations above and below the mean (lines) and the individual values (hollow points) of the measurements, or their transformations, performed on each experimental group.
Calculations were performed by the Sigma Stat program for Windows © 2.0, Jandel Corporation or the SAS program.
The ligature and resection of the common bile duct was performed according to the technique presented by Cameron and Oakley 28 with modifications 29 .
The sham operation followed exactly the same technique described above with exception of the ligatures and common bile duct resection.

Sleeping time by pentobarbital
On P48, the sleeping time test was performed using pentobarbital.The animals received pentobarbital (concentration of 3g/100ml of solution) through intraperitoneal injection at a dosage of 50 mg/kg of rat body weight (or 0.0017 ml/g of animal weight) and time was measured in minutes, from the loss of position reflex to its recuperation.
Euthanasia was performed on P49, beginning with an intraperitoneal injection of sodium pentobarbital, followed by posterior bleeding.The blood and liver were collected for biochemical tests: • aminotransferase dosage: ALT (alanine aminotransferase) and AST (aspartate aminotransferase) in IU/L by the automated colorimetric enzymatic method using Technicon ® equipment, model RAXT; • Liver wet weight in g using Mettler Toledo digital scales -model PB 3002, maximum weight = 3.100 g, minimum weight = 0.5 g, sensibility 0.1 g.

Histological lesions
A fragment of the right lobe of the liver, 0.3 to 0.35 cm thick, was removed from each animal.Histological sections of a around 4 µm thick were stained with hematoxylin and eosin stains.A semi-quantative analysis of a slide from each animal was performed and scores from zero to three for the intensity of fibrosis and ductal proliferation were attributed to each animal.Areas of necrosis were counted in the whole slide, and the number of mitosis in ten fields at high magnification were counted.Three independent measurements of each of the histopathological variables mentioned above were performed.

Pearson's correlation between the groups with cholestasis
In animals with cholestasis not receiving WEB (LA) there was a significant positive linear correlation between IDP and IF (r p = 0.7852; p= 0.0071) and a significant negative lin-ear correlation between IDP and FM (r p = -0.8198;p=0.0037), see Table 3 and Figure 3.
In animals with cholestasis with WEB administration (LB) there was a significant linear positive correlation between NN and IDP (r p = -0.8541;p=0.0017), see Table 4 and Figure 4.

TABLE 3 -Pearson's correlations between pairs of histological variables in group LA
Shaded areas indicates a statistically significant linear correlation between pairs of variables (p<0.05).

Effects of cholestasis
In order to characterize the effects of cholestasis after ligature and resection of the common bile duct (LRBD) we summarized the following alterations: an increase in the aminotransferase serum activities (AST and ALT), a prolongation of sleeping time after pentobarbitol, an increase in liver wet weight and the presence of the following histological lesions -necrosis, fibrosis, ductal proliferation and an increase in the number of hepatocyte mitoses.
The sleeping time test induced by pentobarbitol, in hypnotic doses, was proposed as a simple, non-invasive test of liver function in rats.It reflects the velocity of pentobarbitol disintoxication by the hepatocyte cytochromes P-450 10 .The hepatocytic lesion in cholestasis induced by LRBD leads to a reduction in the isoenzymes of the cytochrome P-450 11 .The sleeping time test induced by pentobarbitol, has been used to evaluate liver function in several experimental models: biliary cirrhosis secondary to extra-hepatic biliary obstruction 30,31 and by carbon tetrachloride 13,14,15,32,33 .In the literature we found no reports on sleeping time in young rats with obstructive cholestasis by LRBD.
The aminotransferase serum activities (AST and ALT) are indicators of the velocity of hepatocellular damage and have been applied in the study of several experimental models for hepatopathies 34,8 .Our results are in agreement with other findings in the literature: experimental cholestasis by anaphthylisothiolcyanate -ANIT 34,35 , experimental cholestasis by LRBD in adult rats 36,37,9,8 , cholestasis induced by endotoxins 38 and cholestasis associated with total experimental parenteral nutrition in adult rats 39 .In the literature we found no reports on aminotransferase serum activities (AST and ALT) in young rats with obstructive cholestasis by LRBD.
Hepatomegaly is an important clinical characteristic in cholestasis.The liver weight increases as a result of the depositing of conjunctive tissue, the formation of regenerative nodules and ductal proliferation 40 .Our results are in agreement with the findings of Aller et al, 1993 41 and Sokal et al., 1996 42 , who studied rats submitted to LRBD, and who also observed an increase in liver weight in these animals.
The presence of necrosis 8 , hepatic fibrosis 40,9,8 and ductal proliferation 21,8 in the biliary cirrhosis secondary to ligature of the common bile duct model has been confirmed in recent works.As our study has pointed out, Aggerbeck et al., 1983  22 , studying the model of cholestasis by LRBD in adult rats also observed an increase in the number of mitoses in cholestatic animals.
In animals with cholestasis without the administration of WEB there was a positive correlation between the intensity of the fibrosis and the intensity of the ductal proliferation.This correlation is to be expected, as in this experimental model these two lesions are intensely related to each other 21 .
The frequency of mitosis presented negative correlation with both intensity of fibrosis and ductal proliferation.Perhaps this correlation can be explained by the deviation of oval cell (originated from stem cells) transformation into cholangiocytes, diminishing the oval cell transformation into hepatocytes.Although it is known that oval cells are totipotent, in that they can transform themselves into either cholangiocytes or hepatocytes 43,44,45 , there is no evidence in the literature that the transformation into in cholangiocytes diminishes the transformation into hepatocytes, which our study suggestswith the possibility of totipotent oval cells.

WEB effects in the absence of cholestasis
The administration of WEB in animals without choletasis resulted in a shorter sleeping time, which is in agreement with the work of Fischman et al., 1991 46 , who studied the effects of water extract of Coleus barbatus in adult rats, concluding that the animals which received the extract presented a 20 to 25% reduction in sleeping time induced by pentobarbitol.Therefore, WEB must have a stimulatory effect on the cytochrome P450.No reports on the effects of WEB on the cytochrome P450 were found in the literature.
WEB also reduced the number of mitoses in the absence of cholestasis.No reports on the effects of WEB and a reduction in the number of mitoses were found in the literature.

WEB effectsf on cholestasis
We were able to show that WEB, at the dosage applied, in biliary cirrhosis secondary to extra-hepatic biliary obstruction in young rats has a hepatoprotective effect characterized by the following actions: it diminished the increase in aminotransferase serum activities (ALT and AST); shortened sleeping time induced by pentobarbitol; prevented the increase in liver wet weight; reduced the number of necrotic areas, the intensity of hepatic fibrosis and hepatocytic regeneration (measured by the number of mitoses).
This hepatoprotective action may be mediated by forskolin, which is one of the most important components of Coleus barbatus 5 .Forskolin stimulates adenylate cyclase and the production of the secondary messenger cAMP 47 , which has a hepatoprotective action against glycochenodeoxycholic acid in hepatocyte cultures 48 .The accumulation of this acid appears to be the most significant factor in the hepatocytic damage of cholestasis.We were unable to show the effect of WEB on ductal proliferation, perhaps due to the small sample size.
The effect of WEB on the reduction of the frequency of mitoses in animals with cholestasis suggests a reduction in the hepatocytic regeneration, which is in agreement with the reduced intensity of necrosis.No reports were found in the literature on the use of WEB in the reduction in the number of mitosis in cholestasis.
Although the group with cholestasis with WEB administration did not present a greater ductal proliferation than the group with cholestasis with distilled water administration, there was a negative correlation between ductal proliferation and the number of necrotic areas.This negative correlation could indicate that one process inhibits the other, which is not the case, or that a third factor inhibits one and stimulates the other, which suggests that the WEB, possibly through the action of forskolin, inhibits necrosis and stimulates ductal proliferation, as show above.
We present this hypothesis even though it was not possible to demonstrate any difference between the ductal proliferation with or without the administration of WEB, perhaps due to the sample size.Knowing that forskolin stimulates ductal proliferation in the presence of the reduction of vagal tonus 49 , further studies should be conducted to test for the existence of vagal hypotonia in this cholestasis model and if found not to be the case, if this action also occurs in the presence of normal vagus tonus.

Conclusions
The administration of water extract of Coleus barbatus exerts an effect on animals without cholestasis (sham surgery) reducing the frequency of mitoses and the sleeping time induced by pentobarbitol.In choestasis, the water extract of Coleus barbatus exerts a hepatoprotective effect, demonstrated by the reduction of: the sleeping time with pentobarbitol, aminotransferase seric activity of AST and ALT, liver wet weight, hepatocytic regeneration (measured by the number of mitoses), the number of necrotic areas, the intensity of fibrosis and, further, by the negative correlation between the number of necrotic areas and the intensity of ductal proliferation.

FIGURE 1 -
FIGURE 1 -Means, standard deviations and individual values of sleeping time (min.)(1A) of aminotransferase serum activity ALT (IU/L) (1B), aminotransferase serum activity AST (IU/L) (1C), and liver wet weight (g) (1D) in rats, in groups of 10, submitted to one of the following treatments: SA: sham operation with daily administration of water; SB: sham operation with daily administration of WEB; LA: double ligature with common bile duct resection with daily administration of water; LB: double ligature with common bile duct resection with daily administration of WEB.Letters repeated in figures 1A, 1B, 1C and 1D, of the graphs, indicate the group pairs with significant differences among them in the S.N.K. test (p<0.05):A=SA>SB; B=LA>SA; C=LB>SB; D=LA>LB.

FIGURE 2 -
FIGURE 2 -Means (of 30 readings -3 per animal), standard deviations and individual values of histological lesions: Frequency of mitosis (2A), number of areas of necrosis (2B), Intensity of fibrosis (2C) and Intensity of ductal proliferation (2D) in rats, in groups of 10, submitted to one of the following treatments: SA: sham operation with daily administration of water; SB: sham operation with daily administration of WEB; LA: double ligature with common bile duct resection with daily administration of water; LB: double ligature with common bile duct resection with daily administration of WEB.Letters repeated in figures 2A, 2B, 2C and 2D, of the graphs, indicate the group pairs with significant differences among them in the Wilcoxon test (p<0,05): A=SA>SB; B=LA>SA; C=LB>SB; D=LA>LB.

FIGURE 3 -FIGURE 4 -
FIGURE 3 -Histological variables in group LA (ligature of the common bile duct with administration of distilled water) Graphs of dispersion showing the results of analysis of Pearson's linear correlation

TABLE 1 -
Feed composition (18g/100g of proteins) prepared by the Pediatrics Department Experimental Laboratory

TABLE 2 -
Sleeping time, aminotransferase serum activities (ALT and AST) and liver wet weight in young ratscomparisons between effects of cholestasis and of WEB, interaction between these effects and descriptive analysis

TABLE 4 -
Pearson's correlation between pairs of histological variables in group LBShaded area indicate a statistically significant linear correlation between pairs of variables (p<0.05)