You are on page 1of 13

Life Sciences 70 (2001) 535–547

The mechanism of hypoglycemic action of

the semi-purified fractions of Averrhoa bilimbi
in streptozotocin-diabetic rats
Peter Natesan Pushparaja, Benny Kwong Huat Tanb,*, Chee Hong Tana
Department of Biochemistry, Faculty of Medicine, National University of Singapore, Singapore-119260
Department of Pharmacology, Faculty of Medicine, National University of Singapore, Singapore-119260
Received 6 November 2000; accepted 3 July 2001

In the present study, we have examined the possible mechanism of the hypoglycemic action of the
semi-purified fractions of an ethanolic extract of Averrhoa bilimbi Linn (Oxalidaceae) leaves (ABe) in
streptozotocin-diabetic male Sprague-Dawley (SD) rats. The ABe was partitioned with water and bu-
tanol to yield a butanol-soluble fraction (BuF) and a water-soluble fraction (AF). The AF was further
partitioned with ethyl acetate and hexane to obtain ethyl acetate (EF) and hexane (HF) soluble frac-
tions. The hypoglycemic property of each fraction was assessed by the oral glucose tolerance test
(OGTT) at a dose of 125-mg/kg-body weight in streptozotocin (STZ)-diabetic rats (STZ 60 mg/kg
i.p.). Fractions AF, BuF and the reference drug metformin (500 mg/kg body weight), produced signifi-
cant blood glucose-lowering effect in the diabetic rats when compared to the vehicle (distilled water).
In the long-term study, the diabetic rats were randomly divided into 4 groups and treated orally by gav-
age with vehicle, AF (125 mg/kg body weight), BuF (125 mg/kg body weight), and metformin (500
mg/kg body weight) respectively twice a day for 14 days. On day 7 and day 14, AF and BuF, like the
reference drug, metformin, lowered the fasting blood glucose concentration significantly (P , 0.05)
when compared with the vehicle. The serum insulin level was significantly increased in the AF-treated
rats only on day 14 when compared to that in the vehicle-treated rats on day zero (P , 0.05). The
serum insulin level in BuF-treated rats was also significantly higher (P , 0.05) on both day 7 and day
14 compared to that on day zero. Hepatic glucose-6-phosphatase activity was significantly lower
(P,0.05) in AF- and metformin-treated groups, but not in BuF-treated groups, compared to that in
vehicle-treated group. However, there was no change in hepatic glycogen content in AF-, BuF- and
metformin-treated group compared to the vehicle-treated group. These results indicate that AF is more
potent than BuF in the amelioration of hyperglycemia in STZ-diabetic rats and is a potential source for
the isolation of new orally active agent(s) for anti-diabetic therapy. © 2001 Elsevier Science Inc. All
rights reserved.
Keywords: Averrhoa bilimbi; Blood glucose; Cytochrome P450; Diabetes; Glucose-6-phosphatase; Hepatic glyco-
gen; Hypoglycemia; Serum insulin; Streptozotocin; Thiobarbituric acid reactive substances

* Corresponding author. Tel.: (165) 8743272; fax: (165) 7730579.

E-mail address: (B.K.H. Tan)

0024-3205/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S 0 0 2 4 - 3 2 0 5 ( 0 1 )0 1 4 2 3 -0
536 P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547

The polyuric states, clinically akin to diabetes mellitus, were documented in the Ebers
papyrus, as early as 1550 BC by the ancient Egyptians. The Egyptologist, Georg Ebers,
found this papyrus in a grave in Thebes in 1862 [1]. Throughout the world, many traditional
plant treatments for diabetes exist and therein lies a hidden wealth of potentially useful natu-
ral products for the control of diabetes [2]. Despite this, few traditional anti-diabetic plants
have received proper scientific screening. The World Health Organization has recommended
that this area warrants further evaluation [3]. Such an evaluation might reveal effective di-
etary adjuncts for the treatment of either diabetes mellitus or the discovery of natural prod-
ucts for developing new anti-diabetic drugs.
Averrhoa bilimbi Linn. (Oxalidaceae, common name: Bilimbi), a common plant in Asia,
which has been widely used in traditional medicine as a cure for cough, cold, itches, boils,
rheumatism, syphilis, diabetes, whooping cough, and hypertension [4]. In Indonesia it has a
considerable reputation as a potent folk remedy for diabetes mellitus [5]. A preliminary study
in our laboratory showed decreases in blood glucose levels and food intake in streptozotocin
(STZ)-diabetic rats given i.p extracts of A. bilimbi fruits and leaves [6]. We also found that
twice a day oral administration of the ethanolic leaf extract of Averrhoa bilimbi (ABe) for
2 weeks at a dose of 125 mg/kg produced a significant reduction in blood glucose and lipid
levels in STZ-diabetic rats [7,8].
The present study was undertaken with the aim of evaluating the possible mechanism (s)
of action of the semi-purified fractions of ethanolic leaf extract of Averrhoa bilimbi (ABe) on
blood glucose levels in STZ-diabetic Sprague-Dawley (SD) rats.

Materials and methods

Preparation of the extract and fractions
The plant was collected from a private garden and identified as Averrhoa bilimbi by
Dr. Ruth Kiew, Keeper of Herbarium and Library, Singapore Botanic Gardens. A dried speci-
men was deposited in the herbarium (Voucher specimen No: BT-2). The fresh leaves of A.
bilimbi (1 kg) were blended and extracted with 80% ethanol (10 L) until exhaustion. The
mixture was filtered with Whatman No 1 filter paper (Whatman International Ltd., England).
The filtrate was centrifuged for 10 minutes at 10,000g to remove particulate substances. The
clear supernatant was partitioned between butanol and water to obtain aqueous fraction (AF)
and butanol fraction (BuF). The AF was further partitioned by ethyl acetate and hexane to ob-
tain ethyl acetate (EF) and hexane soluble fractions (HF). Each fraction was concentrated at
408C, using a rotavapor (Buchi Labortechnik AG, Switzerland) and freeze-dried, to yield
about 40 g of AF, 25g of BuF, 15 g of EF and 12 g of HF. The freeze-dried powder of AF,
BuF, EF and HF were used at dose of 125 mg/kg in both OGTT and long-term study.
All experiments were performed on male Sprague-Dawley (SD) rats aged 10 weeks (200–
250 g) obtained from the Laboratory Animal Holding Unit, National University of Sin-
gapore, Singapore. The study has been carried out along the “Principles of laboratory animal
P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547 537

care” [9]. A standard pellet diets (Glen Forrest, WA, Australia) and water was given ad libi-
tum. Animals were maintained under a constant 12-h light and dark cycle and an environ-
mental temperature of 21–238C.
Induction of experimental diabetes mellitus
The overnight fasted SD rats were made diabetic with streptozotocin (STZ) (Sigma, St.
Louis, MO., USA; 60 mg/kg, i.p). The STZ was freshly dissolved in citrate buffer (0.01 M,
pH 4.5) and maintained on ice prior to use; the injection volume was 1 mL/kg [9]. Diabetes
was confirmed in the STZ-treated rats by measuring the fasting blood glucose concentration
72-h post injection. The SD rats with blood glucose level above 350 mg/dl were considered
to be diabetic and were used in the experiment [10]. Animals had free access to food and
water after the STZ injection.
Experimental procedure
The OGTT in STZ - diabetic SD rats using the semi-purified fractions of ABe
Prior to an oral glucose tolerance test (OGTT), rats were fasted for 16 h. Distilled water
(control), four different fractions of ABe viz. AF, BuF, EF and HF each at a dose of 125 mg/
kg body weight and a reference drug, metformin at a dose of 500 mg/kg body weight were
orally administered to groups of 5–6 rats each. Thirty minutes later, glucose (3 g/kg) was
orally administered to each rat with a feeding syringe [11]. Blood samples were collected
from the tail vein by tail milking at – 30 min (just before the administration of distilled water,
fractions of ABe and metformin in respective groups), 0 (just before the oral administration
of glucose), 30, 60, 120, and 180 min after glucose load for the assay of glucose [12] [Glu-
cose assay kit, Sigma Chemical Co., St. Louis, Mo., USA].
Repeated administration of AF and BuF in STZ-diabetic SD rats
One week after STZ induction of diabetes in male SD rats, the fasting blood glucose levels
were measured. The hyperglycemic rats (blood glucose .350 mg/dl) were divided on day
zero into four groups (each with 5–6 rats). The fasting blood glucose level was measured on
day zero at 9.00 am. Distilled water, AF (125 mg/kg), BuF (125 mg/kg) and metformin (500
mg/kg) were then administered orally twice a day at 9.00 am and 9.00 pm to diabetic control,
treatment and positive control groups respectively for 2 weeks. Body weight, food and water
intakes were monitored every day between 9.00 and 10.00 am for 2 weeks. On 15th day, after
16 h fasting, the rats were decapitated and the blood was collected at 9.00 am for estimation
of the fasting blood glucose. The organs such as liver and kidney were isolated, weighed and
stored at 2708C for the assay of hepatic glucose-6-phosphatase, glycogen, cytochrome P450
and thiobarbituric acid reactive substances (TBARS) in both liver and kidney.
Assay of serum insulin
The insulin in the rat serum was measured by Mercodia rat insulin ELISA kit (Mercodia
AB, Seminariegatan, Sweden). The solid phase two-site enzyme immunoassay is based on
the direct sandwich technique in which two monoclonal antibodies are directed against sepa-
rate antigenic determinants (epitopes) on the insulin molecule. During incubation insulin in
the sample reacts with peroxidase-conjugated anti-insulin antibodies and anti-insulin anti-
bodies bound to the microtitration well. A simple washing step removes unbound enzyme
538 P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547

labelled antibody. The bound conjugate is detected by reacting with 3,39, 5,59-tetramethyl-
benzidine. The reaction is stopped by adding acid to give a colorimetric end point that is read
spectrophotometrically using a microplate autoreader (Bio-tek Instruments Inc.,) at 450 nm.
Assay of glucose-6-phosphatase in liver
The hepatic glucose-6-phosphatase activity was assayed by the method of Baginski et al
[13]. In brief, the glucose-6-phosphate in the liver extract is converted into glucose and inor-
ganic phosphate. The inorganic phosphate liberated is determined with ammonium molyb-
date; ascorbic acid is used as the reducing agent. Excess molybdate is removed by the arsenite-
citrate reagent, so that it can no longer react with other phosphate esters or with inorganic
phosphate formed by acid hydrolysis of the substrate. Arsenite-citrate also stabilizes the sys-
tem. The amount of phosphate liberated per unit time, determined as the blue phosphomolyb-
dous complex at 700 or 840 nm, is a measure of the glucose-6-phosphatase activity. The
glucose-6-phosphatase activity was expressed as mmol of phosphate released /min/ mg of
Assay of liver glycogen content in liver
The liver glycogen content was determined by the enzymatic method of Murat and Serfaty
[14]. Liver was homogenized with the Potter-Elvejhem homogenizer in ice-cold citrate buffer
(0.1 mol/liter, pH 4.2). Immediately after homogenization, 10 ml of the mixture was used to
determine free glucose in the tissue by the glucose oxidase method, according to Trinder
(1969) [12]. Amyloglucosidase dry powder (Sigma Chemical Co., St. Louis, Mo. USA) was
then mixed with the homogenate and allowed to stand at room temperature for 2 h. A 10ml
sample was used to determine total glucose. Initial free glucose was subtracted from the total
glucose, and the resultant value was used to calculate the glycogen content. The glycogen
content was expressed as mg/g wet tissue.
Determination of TBARS
Measurement of MDA by TBA reactivity is the most widely used method for assessing
lipid peroxidation. Kidney and liver samples were minced finely and homogenized with a
Potter-Elvejhem homogenizer in ice-cold 1.15% KCl to make a 25% homogenate. The deter-
mination of thiobarbituric acid reactive substances (TBARS) values was performed by the
method of Uchiyama and Mihara [15]. To 0.1 mL homogenate, 0.2 mL of 8.1% dodecyl sul-
phate sodium salt (SDS), 1.5 mL of 1% phosphoric acid, 0.2 mL distilled water and 1 mL of
0.6%TBA were added. Malonaldehyde (MDA), an end product of lipid peroxidation, reacts
with TBA to form a colored substance. The mixture was heated in a boiling water bath for 45
min. After the reaction, the mixture was cooled in an ice bath; the cold TBA reactants were
extracted with 4 mL of n-butanol. After centrifugation at 1,000 g for 5 min, the optical den-
sity of the n-butanol layer was determined at 535 nm. Values are expressed as nmol MDA/25
mg wet tissue.
Determination of the microsomal cytochrome P450 content
Preparation of liver microsomes
Liver microsomes were prepared, using the ultra-centrifugation method [16]. Rat liver (2 g)
was homogenized in ice cold 1.15% KCl (20 mL) using a glass Potter-Elvejhem homoge-
P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547 539

niser. The homogenate obtained was centrifuged at 10,000g for 10 min at 48C. The resulting
supernatant obtained was recentrifuged at 100,000g for a further 45 min at 48C. The microso-
mal pellet was resuspended in a volume of glycerol buffer [200 mM potassium phosphate
buffer, pH 7.4; 50% (v/v) and 1.15% (w/v) potassium chloride (5:8:7)] equivalent to twice
the liver weight. The pellet was further homogenized with 7 strokes of the pestle after which
aliquots of the microsomal suspension were stored at 2708C.
Protein determination
Protein content in each microsomal suspension was determined by the method of Lowry
et al [17].
Assay of liver microsomal cytochrome P450 content
An aliquot of microsomal preparation of 1mg protein/mL was obtained by adding 0.5 mL
of 1M potassium phosphate buffer and the required volume of 1.15% KCl. A modified tech-
nique [18] was adopted in this assay to eliminate the absorption peak at 420 nm due to con-
tamination by hemoglobin in the sample. The microsomal preparation was placed in two
cuvettes and initially saturated with carbon monoxide. A small amount of sodium dithionite
(not more than 2 mg) was added to the sample cuvette only. The microsomal P450 content was
then determined from the difference in absorbance values between the dithionite-reduced and
control microsomal preparations using a Shimadzu UV-dual-beam spectrophotometer. The
molar extinction coefficient of microsomal P450 at the l max of 450 nm was 91 mM21 cm21.
Statistical analysis
The results are presented as means 6 SEM with n 5 5–6. The changes in body weight,
food and water intakes during the 14-day treatment period were compared by two-way
ANOVA. The data on blood glucose, serum insulin, hepatic glucose-6-phosphatase, hepatic
glycogen, TBARS and cytochrome P450 were analysed by the Student t-test (two-tailed).
P values of , 0.05 were considered to be statistically significant.

Results and discussion

In the OGTT (Fig. 1), AF (125 mg/kg BW) caused a significant hypoglycemic effect
within 30 minutes after oral administration to streptozotocin-diabetic rats. AF also produced
a significant attenuation (P , 0.01) in blood glucose level at 0 min, 120-min and 180 min
when compared with vehicle control. BuF (125 mg/kg BW) had no significant effect on
blood glucose at 0 min, but produced a significant attenuation (P , 0.05) at 120 and 180 min-
utes after oral administration. The other two-fractions, ethyl acetate (EF) and hexane (HF),
did not cause any reduction in blood glucose level at any time point.
The single high dose STZ-induced diabetic rat is one of the animal models of human
IDDM or type I diabetes mellitus. In this model, diabetes arises from irreversible destruction
of the b-islet cells of the pancreas, causing degranulation or reduction of insulin secretion. In
this type I model of diabetes, the insulin is markedly depleted, but not absent [19]. Although
insulin has become one of the most important therapeutic agents known to medicine, there is
a continuing effort to find insulin substitutes, secretagogues, or sensitizers from synthetic or
plant sources for the treatment of diabetes. Over 150 plant extracts and some of their active
principles including flavonoids are known to be used for the treatment of diabetes mellitus
540 P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547

Fig. 1. The oral glucose tolerance test (OGTT) in STZ-diabetic rats using the semi-purified fractions of Averrhoa
bilimbi leaf (ABe). The graph represents the mean percentage changes in blood glucose concentration over – 30
min level in Vehicle (565 6 28), Aqueous Fraction (608 6 42), Butanol Fraction (540 6 51), Ethyl Acetate Frac-
tion (487 6 11), and Hexane Fraction (448 6 19) of ABe, each at a dose of 125 mg/kg and Metformin (603 6 30)
–treated (500 mg/kg) diabetic rats, while bars represent SEM (n 5 5–6). The blood glucose concentration (mg/
dL) of each group at –30 min is given in brackets. * P,0.05 BuF-treated diabetic rats compared with the untreated
diabetic rats. ** P,0.01 Metformin-treated diabetic rats compared with untreated diabetic rats. # P,0.05 AF-
treated diabetic rats compared with the untreated diabetic rats. ## P,0.01 AF-treated diabetic rats compared with
the untreated diabetic rats.

[20]. The present study revealed that the semi-purified fractions of the ethanolic extract of
Averrhoa bilimbi leaves such as AF and BuF have potent hypoglycemic property when given for
2 weeks to STZ-diabetic rats. Tan et al [6] reported the hypoglycemic properties of both aqueous
and ethanolic extracts of the leaves in streptozotocin-diabetic Wistar rats when these were ad-
ministered intraperitoneally at doses of 100 mg/kg BW and 300 mg/kg BW respectively.
Although the body weight of the rats did not differ significantly, food intake and water
intakes of AF and BuF-treated diabetic rats was reduced significantly as the metformin [22]-
P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547 541

treated group when compared to the vehicle-treated diabetic rats. Similar effects were re-
ported for other hypoglycemic agents such as tungstate and vanadate [23, 24].
AF caused a significant (P,0.01) time-dependent hypoglycemic effect (Fig. 2) after
twice-daily oral administration at a dose of 125 mg/kg BW for 7 and 14 days. BuF also
showed a significant (P,0.05) hypoglycemic property on day 7 as well as on day 14 com-
pared to the vehicle-treated control group. The serum insulin level in the AF-treated group
was significantly higher on day 14 compared to both the control (P,0.05) and day zero lev-
els (P,0.05). On the other hand, the serum insulin level in the BuF-treated group was signif-
icantly higher on both day 7 (P,0.05) and day 14 (P,0.05) when compared to its day zero

Fig. 2. The blood glucose and insulin levels were measured on day 0, day 7, and day 14 at 9.00 a.m. after 16 hour
fast in the vehicle (distilled water), AF (125 mg/kg), BuF (125 mg/kg) and metformin (500 mg/kg)-treated STZ-
diabetic rats. * P,0.05 as compared with the diabetic untreated rats. ** P,0.01 as compared with the diabetic
untreated rats. # P,0.05 as compared with the corresponding day 0 value.
542 P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547

The elevation in serum insulin in the AF and BuF-treated STZ-diabetic rats could either be
due to the insulinotropic substances present in the fractions, which induce the intact func-
tional b-cells to produce insulin, or the protection of the functional b-cells from further dete-
rioration so that they remain active and produce insulin. However, except for the level in AF-
group on day 14, the insulin levels were well below the normal insulin level in control rats,
suggesting that they may not be sufficient to lower the blood glucose to its normal level in the
STZ-diabetic rats. Our results indicate a possible insulin-releasing action of ABe in STZ-
diabetic rats. Similarly the extracts of Medicago sativa [25], Eucalyptus globulus [26], Sam-
bucus nigra [27] have been shown to possess insulin-releasing action both in vitro and in vivo.
Glucose-6-phosphatase activity in the liver (given in brackets) was significantly reduced
(P,0.05) in both AF (0.4 6 0.08) and metformin-treated (0.38 6 0.03) groups when com-
pared to the vehicle-treated diabetic control (0.6 6 0.04) group. However, there was no sig-
nificant change in the glucose-6-phosphatase activity of BuF –treated (0.5 6 0.05) group.
Glucose-6-phosphatase catalyzes the final step in glucose production by the liver and kidney.
Streptozotocin increases the expression of glucose-6-phosphatase mRNA, which contributes
to the increased glucose-6-phosphatase activity in diabetes mellitus [28]. Overproduction of
glucose by the liver is the major cause of fasting hyperglycemia in both insulin-dependent
and non-insulin-dependent diabetes mellitus. 90% of partially pancreatectomized diabetic
rats have a .5-fold increase in the messenger RNA and a 3-4-fold increase in the protein
level of the catalytic subunit of hepatic glucose-6-phosphatase. Prolonged hyperglycemia
may thus result in overproduction of glucose via increased expression of this protein [29].
Normalization of the plasma glucose concentration in diabetic rats with either insulin or the
glycosuric agent phlorizin normalized the hepatic glucose-6-phosphatase messenger RNA
and protein within approximately 8 h. However, phlorizin failed to decrease hepatic glucose-
6-phosphatase gene expression in diabetic rats when the fall in the plasma glucose concentra-
tion was prevented by glucose infusion. These data indicate that in vivo gene expression of
glucose-6-phosphatase in the diabetic liver is regulated by glucose independently of insulin.
AF-fraction, like the biguanide drug, metformin, controls the increase in blood glucose in
STZ-diabetic rats by decreasing the activity of glucose-6-phosphatase in the liver. This could
be one of the mechanisms for the suppression of blood glucose concentration in the diabetic rats.
Other workers have also reported that extracts of some plants such as Zizyphus spina-christi
significantly reduced serum glucose level, liver phosphorylase and glucose-6-phosphatase
activities, and significantly increased serum pyruvate level and liver glycogen content after 4
weeks of treatment [30]. Similarly, 60 % ethanolic extract of Coccinia indica and 95 % etha-
nolic extract of Momordica charantia extracts were found to lower blood glucose by depress-
ing its synthesis, on the one-hand, through depression of the key gluconeogenic enzymes,
glucose-6-phosphatase and fructose-1, 6-bisphosphatase and on the other by enhancing glu-
cose oxidation by the shunt pathway through activation of its principal enzyme G6PDH [31].
The liver is an attractive target organ for insulin gene expression in type 1 diabetes as it
contains appropriate cellular mechanisms of regulated gene expression in response to blood
glucose and insulin. The expression of the promoter of the G6pase gene in the liver is in-
duced by glucose and suppressed by insulin [32]. Insulin suppresses hepatic glucose produc-
tion (HGP) in euglycemia by solely decreasing the G-6-P concentration; when combining
hyperinsulinemia and hyperglycemia, the suppression of HGP involves the inhibition of the
P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547 543

G-6-Pase activity; and a sustained glucose-phosphorylation flux might be a crucial determi-

nant in the inhibition of G-6-Pase and of HGP [33]. The operation of glucose 6-phosphatase
(EC (Glc6Pase) stems from the interaction of at least two highly hydrophobic pro-
teins embedded in the ER membrane, a heavily glycosylated catalytic subunit of m 36 kDa
(P36) and a 46-kDa putative glucose 6-phosphate (Glc6P) translocase (P46). P36 gene ex-
pression is increased by glucose, fructose 2,6-bisphosphate (Fru-2, 6-P2) and free fatty acids,
as well as by glucocorticoids and cyclic AMP; the latter are counteracted by insulin. P46
gene expression is affected by glucose, insulin and cyclic AMP in a manner similar to P36
[34]. The insulinotropic effect of AF might play a crucial role in the control of hyperglycemia
in STZ-diabetic rats. The insulin thus inhibits the activity of glucose-6-phosphatase in the
liver of STZ-diabetic rats and thereby controls HGP. Hence the suppression of glucose-
6-phosphate hydrolysis could also be one of the reasons for the hypoglycemic effect of AF
in STZ-diabetic rats. Similar effects were reported for other hypoglycemic agents such as
vanadate [35].
However, there was no significant difference in the level of hepatic glycogen content
(given in brackets) in AF (18 6 2.4), BuF (14 6 0.9) and also metformin-treated (16 6 1.5)
rats compared to vehicle-treated (13.5 6 1.6) rats although the hepatic glycogen content
tended to be higher in the AF-treated group when compared to the control group. Similarly
Vanadate compounds have been shown to inhibit hepatic G-6-Pase activity there by reducing
blood glucose levels in non-obese diabetic (NOD) mice. However no significant difference
was found in the hepatic glycogen stores of the treatment groups compared to control [36].
The kidney TBARS in AF- and metformin-treated diabetic rats were significantly lower
(Table 1, P , 0.05) than in the vehicle-treated rats. On the other hand, the kidney TBARS
value in BuF-treated rats was not significantly different from that in the vehicle-treated rats.
There was also no difference in liver TBARS values between AF, BuF and metformin-treated
rats and vehicle-treated control rats. The liver microsomal cytochrome P450 content was sig-
nificantly lower in the metformin-treated rats when compared to that in the corresponding
vehicle-treated rats. However, there was no significant difference in the liver cytochrome P450
content in AF- and BuF-treated rats when compared with that in the corresponding vehicle-
treated rats.

Table 1
Body weight, water and food intakes in STZ-diabetic rats before and after oral treatment with vehicle
(distilled water), Aqueous Fraction (125 mg/kg), Butanol Fraction (125 mg/kg) and metformin (500 mg/kg)
twice a day for two weeks
Body weight (g) Water (mL/rat/day) Food (g/rat/day)
Group (n 5 5–6) Before After Before After Before After
Control (distilled water) 223 6 3 221 6 11 166 6 10 195 6 10 32 6 5 50 6 6
Aqueous Fraction (125 mg/kg) 225 6 5 243 6 19 144 6 19 158 6 19* 33 6 3 43 6 4*
Butanol Fraction (125 mg/kg) 221 6 9 251 6 30 145 6 10 168 6 5* 41 6 3 46 6 1*
Metformin (500 mg/kg) 236 6 6 241 6 9 153 6 7 145 6 6* 40 6 2 42 6 2*
The body weight, water and food intakes of the STZ-diabetic rats were measured daily as described under
“Materials and Methods”. Results are expressed as means 6 SEM [n 5 5–6].
* P , 0.001 (Compared against control).
544 P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547

Table 2
Liver cytochrome P450 content and TBARS levels in the kidney and liver of STZ-diabetic rats after twice-a-day
oral treatment for 2 weeks with the vehicle (distilled water), Aqueous Fraction (125 mg/kg), Butanol Fraction
(125 mg/kg) and metformin (500mg/kg)
TBARS level
(nmol of malonaldehyde per Hepatic cytochrome
25 mg of tissue) P450 content
Group (n 5 5–6) Liver Kidney (nmol/mg protein)
Control (distilled water) 4.0 6 0.2 4.7 6 0.3 1.25 6 0.04
Aqueous Fraction (125 mg/kg) 3.8 6 0.1 3.9 6 0.2* 1.15 6 0.04
Butanol Fraction (125 mg/kg) 3.9 6 0.2 4.0 6 0.3 1.20 6 0.06
Metformin (500 mg/kg) 3.7 6 0.2 3.7 6 0.2* 1.07 6 0.03*
Results are expressed as means 6 SEM [n 5 5–6].
* P , 0.05 (Compared against control).

The cytochromes are the primary system (Phase I detoxification enzymes) responsible for
chemical defense in animals [37]. The cytochrome P450 content in the liver has been found to
be increased in diabetic animals [38]. The level of increase in the hepatic cytochrome P450 de-
pends on the duration of diabetes [39]. The reduction in the insulin levels in the diabetic state
also causes an increase in the level of cytochrome P450 enzyme [41]. The duration of diabetes
also modulates the hepatic cytochrome P450 profile, the effect being isoenzyme specific [41].
In this study, AF and BuF did not cause any change in the cytochrome P450 enzymes in the
liver. However a reduction was found in the metformin-treated group. The mechanism by
which metformin reduces the cytochrome P450 content is not known.
Hypoinsulinemia in diabetes increases the activity of fatty acyl coenzyme A oxidase,
which initiates b-oxidation of fatty acids, resulting in lipid peroxidation [42]. The enormous
increase in lipid peroxidation leads to the alteration of the transbilayer fluidity gradient [43],
which could hamper the activities of membrane-bound enzymes and receptors. The products
of lipid peroxidation such as lipid radicals and lipid peroxides are extremely harmful to most
of the cells in the body and are associated with a variety of diseases, such as atherosclerosis
and brain damage [44]. It has been suggested that oxidative stress plays an important role in
the development of chronic complications of diabetes [45]. In our present study, there was no
significant change in the level of TBARS in the liver of AF-and BuF-treated and metformin-
treated diabetic rats when compared to the vehicle-treated rats. The assessment of lipid per-
oxidation by conjugate diene levels showed a significant increase in all diabetic tissues
except the liver. This suggests a unique response of the liver in STZ-induced diabetic rats to
oxidative stress and indicates it has higher capacity to cope with the stress compared to other
organs [43]. It was also reported that in the liver of the streptozotocin-diabetic rats, there was
no change in the lipid peroxides after 2-weeks of streptozotocin-administration at a dose of
32-mg/kg iv per day [45]. Hence the lack of change in the TBARS levels in the liver of AF-
and BuF-treated and metformin-treated diabetic rats could again reflect the resistance of the
liver to the oxidative stress in the diabetic state. It is significant to note that neither AF nor
BuF affects this capacity adversely.
P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547 545

The hyperglycemia observed in the STZ-treated rats may lead to the formation of hydro-
gen peroxide, which subsequently generates free radicals such as O22 and OH.. These reac-
tive compounds can cause peroxidation of lipids, resulting in the formation of hydroperoxy
fatty acids and endoperoxides and subsequently an increase the formation of malonaldehyde
(MDA) and thromboxane-B2 (TxB2). The accumulation of TxB2 along with thromboxane-A2
(TxA2) can cause platelet aggregation and promote thrombosis [46]. Since AF has the ability
to reduce the formation of thiobarbituric reactive substances (TBARS), it could potentially
prevent platelet aggregation and thrombosis.
However, the chemical nature of potential antihyperglycemic component (s) of AF and BuF
remains to be established. The AF and BuF can be separated into different fractions by reversed-
phase HPLC and the hypoglycemic activity of each can be tested in vivo in streptozotocin-
diabetic rats or by in vitro glucose uptake studies. The molecular masses of the active fraction
can be identified by mass spectrometry followed by nuclear magnetic resonance (NMR)
spectrometric analysis to obtain the structural information about the active component (s).
In conclusion, the present study shows that the semi-purified fractions of the ethanolic ex-
tract of Averrhoa bilimbi leaves have hypoglycemic action in STZ-diabetic rats. The AF is
more potent in reducing blood glucose in STZ-diabetic rats than BuF. Hence further bio-
chemical and pharmacological studies are being carried out to purify and isolate the active
principle(s) in both AF and BuF and to elucidate their mechanism (s) of action.

The authors wish to thank the National University of Singapore for the research grant
(R-184-000-018-112) and a research scholarship awarded to Peter Natesan Pushparaj.

1. John Pickup C, Williams G editors. Textbook of Diabetes, 1st edn. Oxford: Blackwell Scientific Publications,
1991; pp 3.
2. Gray AM, Flatt PR. Nature’s own pharmacy: the diabetes perspective. Proceedings of the Nutrition Society
1997; 56: 507–17.
3. World Health Organization. Second report of the WHO Expert committee on Diabetes Mellitus. Geneva:
Technical Report Series 1980; 646: pp 66.
4. Goh SH, Chuah CH, Mok JSL, Soepadmo E. Malaysian medicinal plants for the treatment of cardiovascular
diseases. Malaysia: Pelanduk Publications, 1995; pp 62–63.
5. Wee Yeow Chin editor. A guide to medicinal plants. Singapore: Singapore Science Center, 1992; pp 21.
6. Tan BKH, Fu P, Chow PW, Hsu A. Effects of Averrhoa bilimbi on blood sugar and food intake in streptozoto-
cin-induced diabetic rats. Phytomedicine 1996; 3 [Supp1]: 271.
7. Pushparaj P, Tan CH, Tan BKH. Effects of Averrhoa bilimbi leaf extract on blood glucose and lipids in strep-
tozotocin-diabetic Sprague-Dawley rats. Diabetologia 1999; [Supp 1]: 871.
8. Pushparaj P, Tan CH, Tan BKH. Effects of Averrhoa bilimbi leaf extract on blood glucose and lipids in strep-
tozotocin-diabetic rats. Journal of Ethnopharmacology. 2000; 72: 69–76.
9. World Health Organization (WHO) Chronicle. 1985; 39: 51–56.
10. Hamilton K, Eaton EJ, Garland HO, Old S. Effects of experimental diabetes mellitus on Gentamicin-induced
acute renal functional changes in the anesthetized rats. Clinical Experimental Pharmacology and Physiology
1998; 25: 231–35.
546 P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547

11. Lisa P, Kathleen GR, Kenneth JL, Yan L, Luis Ferran J, Chow WS, Stern D, and Schmidt AM. Suppression of
accelerated diabetic atherosclerosis by soluble receptor for the advanced glycation end products. Nature
Medicine. 1998; 4: 1025–31.
12. Al-Awadi FM, Khattar MA, Gumaa A. On the mechanism of the hypoglycaemic effect of a plant extract.
Diabetologia 1985; 28: 432–34.
13. Trinder P. Determination of glucose in blood using glucose oxidase with an alternative oxygen acceptor.
Annals of Clinical Biochemistry 1969; 6:24.
14. Baginski ES, Foa PP, Zak B. In: Methods of Enzymatic Analysis, Vol 2, 2nd edn, Bergmeyer HU and Gawehn
K editors. New York: Academic Press Inc., 1974; pp 876–80.
15. Murat JC and Serfaty A. Simple enzymatic determination of Polysaccharide (Glycogen) content of animal
tissue. Clinical Chemistry 1974; 20:1576–77.
16. Uchiyama M, Mihara M. Determination of malonaldehyde precursor in the tissues by thiobarbituric acid test.
Analytical Biochemistry 1978; 86: 271–78.
17. El Defrawy S, Masry El, Cohen GM, Mannering GJ. Sex-dependent differences in drug metabolism in rat I:
Temporal changes in the microsomal drug-metabolising system of the liver during sexual maturation. Drug
Metabolism and Disposition 1974; 2: 267–78.
18. Lowry OH, Rosenbrough NJ, Farr AL, Randall JL. Protein measurement with the Folin Phenol reagent. Jour-
nal of Biological Chemistry 1951; 193: 265–75.
19. Omura T, Sato R. The carbon monoxide binding pigment of liver microsomes, I: Evidence for its hemopro-
tein nature. Journal of Biological Chemistry 1964; 239: 2370–8.
20. Junod A, Lambert AE, Atauffacher W, Renold AE. Diabetogenic action of streptozotocin: relationship of dose
to metabolic response. Journal of Clinical Investigation 1969; 48: 2129–39.
21. Olajide OA, Awe SO, Makinde JM, Morbise O. Evaluation of the anti-diabetic property of Morinda lucida
leaves in streptozotocin-diabetic rats. Journal of Pharmacy and Pharmacology 1999; 51: 1321–4.
22. Weirnsperger NF. Preclinical pharmacology of biguanides. In: Kuhlmann J, Puls W (eds). Oral Antidiabetics.
Germany: Springer., 1996. Ch. 12.
23. Barbera A, Rodriquez-Gil JE, Guinovart JJ. Insulin-like actions of tungstate in diabetic rats. Journal of Bio-
logical Chemistry 1994; 269: 20047–53.
24. Gil J, Miralpeix M, Carreras J, Bartrons R. Insulin-like effects of vanadate on glucokinase activity and fruc-
tose-2,6-bisphosphate levels in the liver of diabetic rats. Journal of Biological Chemistry 1988; 263:1868–71.
25. Gray AM, Flatt PR. Pancreatic and extra-pancreatic effects of the traditional anti-diabetic plant, Medicago
sativa (Lucerne). British Journal of Nutrition 1997; 78: 325–34.
26. Gray AM, Flatt PR. Antihyperglycemic actions of Eucalyptus globulus (Eucalyptus) are associated with pan-
creatic and extra-pancreatic effects in mice. Journal of Nutrition 1998; 128: 2319–23.
27. Gray AM, Abdel-Wahab YHA, Flatt PR. The traditional plant treatment, Sambucus nigra (elder), exhibits
insulin-like and insulin-releasing actions in vitro. Journal of Nutrition 2000; 130: 15–20.
28. Liu Z, Barrett EJ, Dalkin AC, Zwart AD, Chou JY. Effect of acute diabetes on rat hepatic glucose-6-
phosphatase activity and its messenger RNA level. Biochemical and Biophysical Research Communications
1994; 205: 680–86.
29. Massillon D, Barzilai N, Chen W, Hu M, and Rossetti L. Glucose regulates in vivo glucose-6-phosphatase
gene expression in the liver of diabetic rats. Journal of Biological Chemistry 1996; 271: 9871–74.
30. Glombitza KW, Mahran GH, Mirhom YW, Michel KG, Motawi TK. Hypoglycemic and antihyperglycemic
effects of Zizyphus spina-christi in rats. Planta Medica 1994; 60: 244–47.
31. Shibib BA, Khan LA, Rahman R. Hypoglycaemic activity of Coccinia indica and Momordica charantia
in diabetic rats: depression of the hepatic gluconeogenic enzymes glucose-6-phosphatase and fructose-1,
6-bisphosphatase and elevation of both liver and red-cell shunt enzyme glucose-6-phosphate dehydrogenase.
Biochemical Journal 1993; 292: 267–70.
32. Chen R, Meseck M, McEvoy RC, Woo SL. Glucose-stimulated and self-limiting insulin production by
glucose-6-phosphatase promoter driven insulin expression in hepatoma cells. Gene Ther2000; 7:1802–9.
33. Guignot L, Mithieux G. Mechanisms by which insulin, associated or not with glucose, may inhibit hepatic
glucose production in the rat. Am J Physiol 1999; 277:E984–9.
34. Van de Werve G, Lange A, Newgard C, Mechin MC, Li Y, Berteloot A. New lessons in the regulation of glu-
cose metabolism taught by the glucose-6-phosphatase system. Eur J Biochem 2000; 267:1533–49.
P.N. Pushparaj et al. / Life Sciences 70 (2001) 535–547 547

35. Mosseri R, Waner T, Shefi M, Shafrir E, Meyerovitch J. Gluconeogenesis in non-obese diabetic (NOD) mice:
in vivo effects of vanadate treatment on hepatic glucose-6-phoshatase and phosphoenolpyruvate carboxyki-
nase. Metabolism 2000; 49:321–5.
36. Schulz LO. Suppression of the hepatic glucose-6-phosphatase system in diabetic rats by vanadate. Annals of
Nutrition and Metabolism 1988; 32: 289–96.
37. Gillam EM. Human cytochrome P450 enzymes expressed in bacteria: Reagents to probe molecular interac-
tions in toxicology. Clinical Experimental Pharmacology and Physiology 1998; 25: 877–86.
38. Lucas D, Farenz LG, Bardou JV, Attali JR, Valensi P. Cytochrome P450 2E1 activity in diabetic and obese pa-
tients as assessed by chlorzoxazone hydroxylation. Fundamental and Clinical Pharmacology 1998; 112: 553–8.
39. Barnett CR, Flatt PR, Ioannides C. Modulation of the rat hepatic cytochrome P450 composition by long-term
streptozotocin-induced insulin-dependent diabetes. Journal of Biochemistry and Toxicology 1994; 9: 63–69.
40. Woodcraft KJ, Novak RF. Cytochrome P4502E1 activity in diabetic and obese patients as assessed by chlor-
zoxazone hydroxylation. Chemico-Biological Interactions 1997; 107: 75–91.
41. Horie S, Ishii H, Suga T. Changes in peroxisomal fatty acid oxidation in diabetic rat liver. Journal of Bio-
chemistry 1981; 90 1691–6.
42. Bruch RC, Thayer WS. Differential effect of lipid peroxidation on membrane fluidity as determined by elec-
tron spin resonance probes. Biochimica Et Biophysica Acta 1983; 733: 216–22.
43. Oberley LW. Free radicals and diabetes. Free Radical Biology and Medicine 1988; 5: 113–24.
44. Elangovan V, Shohami E, Gati I, Kohen R. Increased hepatic lipid soluble antioxidant capacity as compared
to other organs of streptozotocin-induced diabetic rats: a cyclic voltammetry study. Free Radical Research.
2000; 32: 125–34.
45. Tatsuki R, Satoh K, Yamamoto A, Hoshi K, Ichihara K. Lipid peroxidation in the pancreas and other organs in
streptozotocin diabetic rats. Japanese Journal of Pharmacology 1997; 75: 267–73.
46. Sushil Jain K, Stephen Kruegerk K, Mcvie R, John Janamillo J, Palmer M, Smith T. Relationship of the blood
thromboxane-B2 (TxB2) with lipid peroxides and effect of vitamin E and placebo supplementation on TxB2
and lipid peroxide levels in Type I diabetic patients. Diabetes Care 1998; 21: 1511–16.m