The effect of flabelliferins of palmyrah fruit pulp on intestinal glucose uptake in mice

Palmyrah (Boorsc~tsJlizbelltfrr L.) is 3 tree found growing in .rid c l i n ~ ~ t e s . It has an edlble fruit pulp. Previous work had shown that Institute of Cancer Reseal-ch (ICR) mice fed with 10°/o fruit pulp in standard feed showed reduced weight gain due ro b ~ ~ t e r sreroiddi sdponins termed flabelliferin-11. T h e effecr of this pulp and flnl~elllferin I1 o n glucose uprake was srudied. Ar a dose of 10mg/niouse, mixed fl.~hellifttrins (with 2.5 mg flabelliferin 11) reduc.cti blonil ~ l u c o s e aher glucose challenge (p < 0.001), increased f aec~ l glucose (p <0.001) 2nd inresrinal glucose Ip < 0.001), but nor FaecJ g l i ~ c o ~ , @=0.62). The latter is probably due to the removal of f l~be l l i t ' e r in t r ig lycoside n;hich is an r imic rob ia l . T h e an~irnicrol,ial component anJ t\vo other tlahellifrl-ins did no1 show the al,ove effect. As rhe pulp is on-toxic ir may find application 3s a functional food.

Flabelliferins, a family of steroidal saponins in the fruit pulp of palmyrah (Borasslrs flahellzfcr L., Family: Arecacede) were first highlighted in 1994'.O f t h e ilabelliferins isolated so far from the crude flabelliferin mixture, a tetraglycoside flabelliferin-I1 (I-11), with two rhaninosyl moieties and two glucosyl moieties, is responsible for the bitrerness in palinyrah fruit pulp.' F-I1 has also beet1 shown to inhibit ATPase of ghcst red blood cells.' ilebitrering of Palmyrah Fruit Pulp (PFP) by the enzyme naringinase causes the loss of F-I1 (and bitterness) and another flabelliferin, flabelliferin B (FB), a branched flabelliferin triglycoside.'Flabelliferin B was shown to have potent anti-yeas1 and anti-bacterial activity..' It had been obsenreci that PFP caused weight loss in Institute of Cancer Research mice (ICR) despite there being no change in isocaloric ieed intake.'Feeding of non-bitter PFP (no F-11) did not result in such weight loss thus indirectly implicating F-I1 as the causative agent for the observed weight loss.Flabelliferins are naturally associated with a uv-active binder6 which could have an effect o n the extent of their activity.
T h e objectives of this study were to determine: (i) whether PFP induced weight loss'.bwere reproducible, (ii) the cause for the weight loss, (iii) whether separation of the uv-active binder would affect bioactivity (iv) and whether therapeutic applications would be feasible. --

Biological materials
Experinzental animals : Inbred, hoinogeneous ICR mice (4 wlc old, 20k5 g body weight, weanling males for weight gain studies, and 6 w k old, 35 +5 g body weight males for other studies) were purchased from the Medical Research Institu~e, and the experiments were conducted at the animal house, University of Sri Jaycwardenepura.The mice were maintained in a temperature-controlled room (30 OC) under 12 h light/dark cycle (dark phase 6 p.m to 6 a.m).They were fed with the rat and mouse breeding feed recommended by the World Health Organization (WHO)7 or 10% I'FP containing pellet feed and water ad libitum.Feed illtake and weight gain were determined by methods previously des~ribed.'.~ Plant ~nateria!~:Fruits of palmyrah were collected from Kalpitiya in the North-West of Sri Lanka.The fruit pulp was extracted from its fibrous matrix by extracting manually with a spoon, using water in the ratio of 2:l.' Extractzon offlabellzferzn mzxture from PFP: T h e flabelliferin mixture was extracted using methanol followed by de-carotenising with petroleum ether (60-80 OC).' The sugar was separated from extracts using a dry cellulose column.'Co~~-e>poniilng author Separation of jlabell2ferins: Separation of flabelliferins was conducted by using a chromatotron8 (Harrison, California, USA).When necessary, further purification was carried o u t using preparative t h i n layer chromatography (300 m m silica gel G60 using butanol: ethanol; ammonium hydroxide, specific gravity 0.88 = 7 3 4 ) .Spraying with anisaldehyde reagent was necessary to monitor fractions and test them for purity.'All flabelliferins contain a uv-active binder.6This was separated using a mixture of iso-propanol:methanol (1:1) in a chromatotron to obtain pure flabelliferins.' Assay offlabellgerins: This was conducted using a computerized densitometer by methods described previously.'A d m i n i s t r a t i o n o f 10% PFP containing feed: P F P containing feed was prepared by substituting one-third the maize with 10% PFP on dry basis (when PFP was used) in W H O scandard rat and mouse breeding feed.T e s ~ and control diets were isocaloric.Mice (36) were divided into 6 groups (6 in each), 3 control groups and 3 test groups.
O One set (a control and a test group) was used for the weight gain study and at the end of 2 wks their fasting blood sugar levels were tested.(11) In the 2nd set, 10% PFP containing feed was given to the test group and the WHO standard rat and mouse breeding feed was given to the control group.Faeces were collected on the following day for testing of reducing substances.

@I)
The remaining set of animals were given 10°/o PFP containing feed (test) and W H O standard rat and mouse breeding feed (control) for a day.O n the following day they were given a glucose load and after 1.5 h they were sacrificed to collect blood and intestinal samples.

Admznistration offlabelllferins:
T h e dosage of mixed flabelliferins administered to the test group was 10 mg/ 50 g mouse in 0.5 mL distilled water (calculated o n the basis of the yield of flabelliferins after separation from PFP and food intake of a mouse).Controls were given the same volume of water.T h e dosage of mixed flabelliferins w i t h o u t u l t r a violet (uv) b i n d e r administered was 2 mg/ 50 g mouse in 0.5 ml distilled water.
Extractives were administered orally using a Sondi needle by the gastric garage method.
Glucose challenge: Mice were fasted overnight and glucose (Glaxo -Welcome Co.,) dissolved in distilled water, was orally administered to mice a t a dose of 1.5 g/kg body weight in avolume of 0.5 mL.
Collection and assay offaeces: Faeces were collected 24 h after administration of glucose and the total sugar content was determined by the Dinitrosalicylic acid (DNS) method.'

Estimation of glucose in blood
Fastzng blood glucose: Animals were fasted for 12 h and they were sacrificed under diethyl ether anesthesia.Blood (0.5 mL) was drawn by cardiac puncture and the blood glucose concentrations analysed immediately using a reagent kit (DMA, USA) employing the glucose oxidase method.l 3 Postprandz~zl bloodglucose: Animals were sacrificed under diethyl ether anasthesia, 90 min post glucose loading (2 h after administration of extract) and the glucose level was determined as before.
Collection ofintes~inalsanzplm and deternzination ofglucose conte?zt i n the ~ntestinal wash '!: Animals were fasted overnight and the test group was orally administered the extractive 30 min prior to the glucose challenge.The control group was orally administered distilled wawr (0.5 mL/50 g mousc) at the same time.Tlle animals in the tesr group and the control group were sacrificed 1.5 h after the administration of glucose load undcr diethyl ether anasthesia and their intestines removed.Intestines from each animal were washed with 25 m L of distilled water and the washings centrifuged at 3000 rpm for 10 min.The supernatant was rested by the glucose oxidase methodlo t o de~ermine the glucose content.

Statistical analysis
All the results are presented as (mean+S.E.) Statistical analysis was carried out in Microsoft Excel.T h e significance was tested by Student's r-test.A probability level of p c 0 .0 5 was chosen as the cut off point of statistical significance, but accurate p values were calculated using the computer software.
The dosage of purified flabelliferins (calculated on the basis of preliminary experiment) was: F-11, 1 mg/50 g mouse, FB, 3 mg/50 g mouse, FD +FE, 1.5 mg'50 g mouse.(50 mg/100 gJ.The sample was a bulk sample from for 7 days.Controls were given the same volume of water.

Kalpitiya, o n e of t h e main sources of P F P . T h e
At the end of the 7 day period of feedingwith the mixed percentages of t h e p u r e flabelliferins (per total flabelliferins, the test-animals had a lower weight gain flabelliferins) were F-11-51%, FB -3.9%, F D + F E -41.5% than (p= 0.01) the control animals.However, fasting in this sample.
(i) Animal set 1. Weanling ICR mice fed over a period of 2 weeks with 10% PFP containing feed showed a marked decrease in weight gain (p=0.007,Table I) compared to control mice fed with the PFP free diet, although food intake did not decrease significantly (p = 0.7).This had been previously reported5,'.N o significant difference was found in the fasting blood sugar levels between the test and control animals (p=0.64).
(ii) Animal set 2. The reducing sugar content in the faeces of the test group of animals was significantly higher (p < 0.0001) than in the control group, while the faecal fat content was similar (14.77% in test group and 14.47% in controls).
(iii) Animal set 3. T h e rest group of animals had significantly lower blood glucose levels (p < 0.001) with no significant changes in the intestinal glucose content (p= 0.21) after a glucose challenge (Table I).
Animal set 1. Test mice (n=6) were administered 10 mg/50 g mouse mixed flabelliferins ~n 0.5 mL of water Animal set 2. There was no significant difference between faecal weight of test animals and controls.
However, after a glucose challenge, faecal sugar (p < 0.01) and intesrinal glucose (p<0.001) in the test animals increased significantly while blood glucose showed a marked decline (p <0.0001) when compared to the corresponding values in control animals (Table 2).Separating t h e fluorescent binder f r o m mixed flabelliferins gave larger decline even when the dose was reduced t o 2 mg/50 g mouse(Tab1e 2).

Separated flabelliferins.
F-11.Oral administration of the flabelliferin F-I1 (MW = 1030) containing uv binder (h4W = 544) after a glucose challenge resulted in a significant increase in intesrinal glucose content (p< 0.001) and decrease in blood glucose level (p < 0.001) with no change in faecal sugar and faecal weight, despite the dose being only 1 mg/50 g mouse (Table 2).
FB and F D + F E .Administration of FB (3 mg/50 g mouse) and F D + F E (1.5 rng/50g mouse) resulted in a reduction in intestinal glucose content and elevation in blood glucose level after a glucose challenge.This was the reverse of the effect observed with F-II (Table 2).A: test animals = 6, were o,iven 10% PFP containing feed and the control group was given WHO standard rat and mouse breeding feed for two weeks.B and C: after feeding for one day the mice were given a glucose load and after 1.5 h they were sacrificed to collect blood and intestinal samples.It should be noted that the F-I1 constitutes 50% of the mixed flabelliferins while flabelliferins FB, F D and FE make up mosr of the rest.This study shows that FB, F D and FE-produce no decline of glucose uptake.
The overall results have implications o n the use of F-I1 as either an anti-obesity o r anti-diabetic food component.The fact that PFP is 11on-toxic t o mice even at 5 0 % incorporation in feed1' lends weight to this possible application.

RESULTS 1 .
Flabelliferin profile.Total flabelliferin content in the PFI' sample used in the present study was 0.5%/ PFP joirrnal o f the Ndtzonnl Sczmce Folindntton of Srz Lnnkn 34(l) Mean +S.EJor/mal ofthe .Mational Sc~ence Folrndatzon of Sri Lnnka 34(1)   March 2006 mlce ( n = 6 each) were fasred for 12 h, and a glucose load (1.5 g/Kg BW) was given 30 min after admlnisrr3tion of test flabelliferins.Glucose contenr in blood and intestinal wash nas determined 2 h atter ex~mcrive follow~ng sacrificing with dierhyl ether anesthesia.Faeces were collt.cted in a separate trial, 24 h aher adminis~ra~ion of F-11, and analysed for sugar in rest and conrrols.A, B, and C were separate trials with their own conrrols.C h2J two controls, one for F-I1 rr~al and another for the other flabelliferins.-DISCUSSIONknownto inhibit a wide range of b~cteria', possibly inhibits bacteria of the large intestine and leads LO h i ~hThe results confirm previous observations3~'that administration of PFP containing feed results in a significant weight loss.The insignificant effect on fasting blood sugar of mice fed o n lo0/" PFP could probably be attributed to the high sugar content (about 80% on a dry weight basis of P F P ) which may serve to counteract the hypoglycaemic effects of the active principle.In addition, gluconeogenesis that would occur during fasting may also contribute to maintaining blood glucose homeostasis.Experiments with separated flabelliferins, clearly showed that F-I1 (flabelliferin tetraglycoside ) is the compound that inhibits absorption of glucose from the intestinal lumen and this results in a reduction of blood glucose o n glucose challenge (an anti hyperglycaemic effect) and probably leads to the reduced weight gain observed p r e v i o u ~l y ' ~~.It is interesting that faecal glucose content was n o t f o u n d t o be significantly h i g h e r o n administration of pure F-11.This could be attributed to absence of the antimicrobial flabelliferin (FB)' in pure F-11.FB, that is present in mixed crude flabelliferins is glucose in the faeces of mice administered with crude flabelliferins or fed with 10°/o PFP containing feed.Thc effects of flabelliferins plus and minus the uv active binder o n glucose challenge show that separation of the binder increases the efficiency with which the flabelliferins, probably flabelliferin F-11, reduces i n ~e s ~i n a l glucose uptake even when calculations are made o n molar basis.

Table 1 :
Effect of 10% PFP containin: feed on weight gain and blood and intestinal o,lucose after a glucose load (1.5 g/Kg BW' )

Table 2 :
Effects of flabelliferins on glucose content in blood, faeces and intestinal wash after glucose challenge.