Original Paper
In vivo Metabolism of Aloemannan
A. Yagi1, *, J. Nakamori, T. Yamada, H. Iwase, T. Tanaka, Y. Kaneo, J. Qiu2, and S. Orndorff
1 Faculty of Pharmacy and Pharmaceutical Sciences, Fukuyama University, Fukuyama, Japan
2 Univera Pharmaceutical Inc., Broomfield, CO, U.S.A.
Revision accepted: March 25, 1999; Received: September 17, 1998
Abstract: The metabolism of fluoresceinyl isothiocyanate la-
beled aloemannan (FITC-AM) was examined by p.o. and i.v.
ministration in mice at a dose of 120 mg/kg. Analysis of FITC-
AM in urine and feces showed that FITC-AM (MW 500 KD) was
metabolized into smaller molecules that mainly accumulated in
the kidneys. AM was catabolized by the human intestinal mi-
croflora to catabolites 1 and 2 with molecular weights of 30
and 10 KD, respectively. Hydrolysis of AM showed hexosamine
peaks on HPAE. The findings suggest that the immunomodula-
tion of AM may come from not only neutral polysaccharides
but also contaminated hexosamine in AM.
Key words: Aloemannan, Aloe barbadensis, Asphodelaceae,
fluoresceinyl isothiocyanate aloemannan, metabolism, catabo-
lites, intestinal microflora.
Introduction
Several biological and clinical studies on aloemannan, isolat-
ed from Aloe barbadensis gel have been reported (1). The
structure was established to be a neutral polysaccharide com-
posed of partially acetylated glucomannan having a molecular
weight of more than 200 KD (2), (3). In a previous paper we
reported the characterization of a glycoprotein fraction, verec-
tin, which possessed cell proliferation activity for normal hu-
man dermal fibroblasts (4). This paper presents metabolism
of aloemannan (AM) by mice and catabolites of AM by human
intestinal bacteria.
Materials and Methods
Fractionation
AM was obtained from Aloe barbadensis freeze-dried powder
which was supplied by Aloecorp (Texas, U.S.A.), by the meth-
od of Yagi (4). Briefly, the powder was dialyzed at 4 8C against
distilled water for 36 h and the retentive fraction after dialy-
sis, dissolved in 0.02 M NH4HCO3, was applied to a column of
DEAE-Sephacel. The eluate with the same solvent was con-
Planta Medica 65 (1999) 417±420
 Georg Thieme Verlag Stuttgart New York
· ter (6:4:3) solvent was used on a cellulose plate, and aniline
ISSN: 0032-0943
417
centrated, dialyzed and lyophilized to produce a colorless
ad- neutral polysaccharide fraction (aloemannan).
FITC-labeled dextrans of various molecular weight (MW; 4.4±
485 KD) were purchased from Sigma Chemical Co., U.S.A. [FD-
4 (MW 4.4 KD), FD-10 (MW 9.4 KD), FD-20 (MW 18.9 KD), FD-
40 (MW 40.5 KD), FD-70 (MW 69 KD), FD-150 (MW 144 KD),
FD-500 (MW 485 KD)]. Fluorescein thiocarbamoyl aloeman-
Downloaded by: NYU. Copyrighted material.
nan (FITC-AM) was synthesized by the method of deBelder
and Granath (5). Briefly, aloemannan was dissolved in DMSO
containing a few drops of pyridine. Isothiocyanate fluorescein
was added, followed by dilaurate, and the mixture was heated
for 2 h at 95 8C.
By several precipitations with ethanol, free dye was removed
and FITC-AM in ethanol was filtered and dried in vacuo at
80 8C. FITC-AM, dissolved in water, was purified by gel per-
meation on a Sephadex G-25 column to provide FITC-AM in a
yield of 38.3%. Labelling efficiency was determined by absor-
bance at 495 nm of the compounds dissolved in 25 mM borate
buffer (pH 9.0). The degree of substitution of FITC in FITC-dex-
trans and -AM was estimated to be 0.01 mol FITC/glucose unit
and 0.02 mol FITC/mannose unit, respectively, based on the
calibration curve for fluorescein sodium (uranine).
High performance size-exclusion chromatography (HPSEC) was
used for the quantitative analysis of FITC-labeled dextrans
and FITC-AM (6). HPSEC was carried out using a Tosoh HPLC
system (CCPD; Tokyo) equipped with a variable-wavelength
fluorescence detector (RF-530, Shimadzu, Kyoto). The excita-
tion and emission wavelengths were set at 495 and 520 nm,
respectively. A 7.8 ” 300 mm, TSK gel G 3000SWXL column
(Tosoh) was used at ambient temperature with a mobile
phase of 0.2 M NaCl in 0.05 M phosphate buffer (pH 7.0) at a
flow rate of 1.0 ml/min. Data analysis was done using a Tosoh
super system controller SC-8010 (7).
High performance anion exchange chromatography (HPAE)
using Dionex CarboPac PA 1 column was carried out by iso-
cratic 16 mM NaOH (1.0 ml/min) with pulsed amperomeric
detector for the quantitative analysis of hexoses and amino-
hexoses after 2 M TFA hydrolysis. Desalting was carried out by
use of ion exchange-mixed resin, MB-3 (Amberlite Organo
Inc.). For TLC analysis of the hydrolysate BuOH-pyridine-wa-
hydrogen phthalate and ninhydrin reagents were used for de-
418 Planta Med. 65 (1999) A. Yagi, J. Nakamori, T. Yamada, H. Iwase, T. Tanaka, Y. Kaneo, J. Qiu, and S. Orndorff

tection. IR spectra were obtained with an FT-IR 8500 (Shimda- Scheffe test for comparison between the positive group and
zu). The 1H-NMR spectra were obtained on a JEOL Lambda the aloe-treated group.
500 spectrometer operating at 500 MHz at 25 8C. Chemical
shift values were expressed in ppm downfield from trimethyl-
silylpropionic acid Na salt as an internal standard and sam-
Results and Discussion

ples were dissolved in D2O in a 5 mm sample tube.
Preparation of catabolites with an intestinal bacterial mixture
AM (500 mg) dissolved in 0.1 M phosphate buffer, (pH 6.8,
20 ml) was incubated with a suspension (10 ml) of fresh feces
obtained from healthy man and stored at 20 8C for 5 days in
dark under anaerobic conditions. The mixture was centri-
fuged and an appropriate volume (about 1 l) of acetone was
added to the supernatant to afford precipitation. The precipi-
tate was extracted with water. The extract was applied to gel
permeation chromatography on DEAE-Sephacel using 0.02 M
NH4HCO3 to obtain unbound material. Absorbed material was
eluted with 0.3 M NaCl. This fraction was applied to a Sepha-
dex G-25 column (1.5 ” 75 cm) using distilled water as a sol-
vent to obtain catabolite fractions, 1 (yield: 1.0% of AM) and 2
(yield: less than 0.1%).
Molecular weight determination
Figure 1 shows the distribution curve for FITC-dextrans of var-
ious molecular weights as determined by HPSEC. A close cor-
relation between MW of FITC-dextran molecules and reten-
tion times was obtained without any influence by FITC-deri-
vatization. Thus, the method can be used to determine MW
changes of FITC-AM in the animal body. Size exclusion chro-
matography of FITC-AM which was monitored by fluores-
cence detection yielded an elution profile as shown in Figure
2. FITC-AM, MW 500 KD, was metabolized into low MW sub-
stances (70±10 KD) in urine 24 h after i.v. administration at a
dose of 120 mg/kg in mice. Under the same conditions FITC-
AM was converted to low molecular weight substances (5 KD)
in feces. FITC-AM was metabolized into low molecular weight
substances (less than 9 KD) in urine and feces 24 h after p.o.
administration at a dose of 120 mg/kg in male ddY mice (20±
30 g). Figures 3 and 4 show the cumulative fecal and urinary
excretion of FITC-AM in mice after i.v. and p.o. administrations
at a dose of 120 mg/kg. About 0.3% of FITC-AM was excreted
into urine 48 h after p.o. administration (see Table 1), indicat-

Downloaded by: NYU. Copyrighted material.

Catabolites 1 and 2 were analyzed according to a previous pa-
per (4).
Hydrolysis

Hydrolysis of AM and 1 was performed with different concen-

trations of TFA at various reaction times at 100 8C, up to 2 M
TFA for 2 h. The hydrolysates were evaporated, desiccated and
examined by TLC and by HPAE. Alkaline hydrolysis of AM
(100 mg) was carried out with 0.2 M Ba(OH)2 at 100 8C for 6 h
(8). The hydrolysate was applied to a Sephadex G-25 column
to obtain a glycopeptide fraction by monitoring at 280 and
490 nm (4). The glycopeptide fraction (5 mg) was hydrolyzed
with 2 M TFA, and the hydrolysate showed aminohexose
peaks on HPAE.
Immunomodulatory activity
Molecular weight calibration curve for FITC-labeled dextrans
Eight-week old (about 25 g) mice (C3H/HeN from Harlan Fig. 1

by high-performance size-exclusion chromatography on TSK gel G

Sprage Dawley) were allowed to acclimate for 1 week. UVB- 3000SWXL Vo: void volume.
induced immunosuppression (or contact hypersensitivity)
was used as model according to the method of Strickland (9).
Briefly, abdominal hair of mice were shaved, followed by ex-
posure of the shaved skin to 2000 J/m2 UVB. Immediately
after exposure, the test sample (0.5 mg/ml in aquaphor) (9)
was applied to the skin. Three days later the mice were sensi-
tized with 50 ml of 0.3% dinitrofluorobenzene (DNFB) in acet-
one. Six days later the ears of each mice were measured fol-
lowed by challenge with 5 ml of 0.2% DNFB in acetone to each
side of the ear surface. The net ear swelling was determined
24 h after challenge. Three control groups were included in
each experiment. Blank groups received no UVB, no treat-
ment, no sensitization, but were challenged. Positive groups
received no UVB, no treatment, but were sensitized and chal-
lenged (100% response). Suppressed groups received UVB, no Elution profile of authentic FITC-AM on TSK gel G 3000WXL
treatment but were sensitized and challenged (fully suppres-
Fig. 2

by high-performance size-exclusion chromatography with fluores-

sed). Statistical significance was calculated using ANOVA and cence detection (lex = 495 nm, lem = 520 nm).
In vivo Metabolism of Aloemannan Planta Med. 65 (1999) 419

Table 1 Excretion of FITC-AM into urine and feces after administra-

tion (120 mg/kg) to mice.
i.v.injection p.o. administration

urine feces urine feces

(% to dose) (% to dose) (% to dose) (% to dose)

0±24 h 63.8 11.3 0.18 91.6

24±48 h 9.31 1.48 0.12 3.72
total 73.1 12.8 0.30 95.3

Fig. 3 Cumulative
urinary (*) and fe-
cal (~) excretion of
FITC-AM after intra-
veneous injection
(120 mg/kg) in mice.
Values are given as
mean  S.D. for
groups of 3 mice.

Fig. 5 Tissue distribution of FITC-AM 2 h after intravenous injection

(120 mg/kg) in mice.

Downloaded by: NYU. Copyrighted material.

1
H-NMR and IR spectral data of aloemannan , 1 and 2. Aloeman-
nan (ppm): 2.10, 2.16, 2.19, 2.22 (CH3COO), 3.48, 3.55 (t,
Fig. 4 Cumulative CH2OH), 4.52, 4.66 (anomeric), 5.44, 5.52 (CH3COOCH); 1:
urinary (*) and fe- 1.92, 2.07, 2.16 (CH3COO, CH3CONH), 3.44, 3.55 (t, CH2OH),
cal (~) excretion of 4.39, 4.52 (anomeric), IR cm±1: 1656 (amide I); 2: 1.90, 1.92
FITC-AM after oral (CH3COO, CH3CONH), 3.56 (t, CH2OH), 4.42, 4.50 (anomeric);
administration (120 IR cm±1: 1656 (amide I). Sugar component of hydrolysates of 1
mg/kg) in mice. and 2 on HPAE. Aloemannan. Man:Glc (9:1), 1: Man:
Values are given as
mean  S.D. for Glc:Gal:GalN:GlcN (1:9:2:7:27), 2: Man:Glc:Gal:GalN:
groups of 3 mice. GlcN (2:3:24:10:1).
Alkaline (8) and acid hydrolysate of AM Man : Glc : Gal : GalN :
GlcN (22 : 4 : 6 : 2 : 1)

Protein content was measured by Lowry method and was

ing that FITC-AM was absorbed from the intestinal tract, and
metabolized and/or catabolized into smaller molecules (less
than 9 KD) in the animal body.
Tissue distribution of FITC-AM after i.v. injection in mice
showed that it was primarily located in the kidneys but was
cleared in a time-dependent fashion (see Fig. 5). This indicates
that FITC-AM was metabolized into smaller molecules that
mainly accumulated in the kidney. In a previous paper hepatic
accumulation of FITC-dextrans after administration by i.v. and
p.o. routes was reported (7), but little accumulation of FITC-
AM in liver tissue was found. The difference in tissue distribu-
tion may be the result of structural differences in dextran and
aloemannan molecules, the latter of which has acetyl groups.
A bacterial mixture from human feces in phosphate buffer
could transform aloemannan (MW over 400 KD) into small
molecules.
shown to 1 (3.25%) and 2 (10.4%). The catabolites showed the
presence of a peptide moiety in the molecule. The water-solu-
ble catabolites, 1 and 2, had no acetyl group but an amide
bond in the IR spectra, and a molecular weight of 30 KD and
10 KD, respectively, on Sephadex G-50 gel permeation. Protein
content was 3.25 % and 10.4 % in 1 and 2, respectively. Since
aloemannan indicated no peptide conjugation, the discrepan-
cy may be the result of contamination or affinity of the pep-
tide for small aloemannan molecules by microbials in feces.
The restorative activity of 1 and 2 in immunomodulation
showed 3.5% and 29.5%, respectively, together with that of
AM (41.7%) and that of the lyophilized powder (60.8 %) from
Aloe barbadensis gel. No significant difference between 2 and
AM was observed.
Aloemannan has been reported as a neutral polysaccharide
mainly composed of 1±4 b-D-mannose. This study shows that
aloemannan is metabolized into smaller molecules that accu-
mulated in the kidney in a yield of 1% and is also catabolized
into 1 and 2, MW 30 and 10 KD, respectively, by human intes-
tinal bacteria in a yield of 1%. Structural study of the catabo-
lites, 1 and 2 from aloemannan was done by 1H-NMR and IR
420 Planta Med. 65 (1999) A. Yagi, J. Nakamori, T. Yamada, H. Iwase, T. Tanaka, Y. Kaneo, J. Qiu, and S. Orndorff

spectral and HPLC analyses. It was shown that 1 and 2 were Prof. Akira Yagi
composed of (or contaminated with) sugar and peptide moi- Faculty of Pharmacy and Pharmaceutical Sciences
eties. From this preliminary experiment a metabolisation of Fukuyama University
AM by bacterial microorganism from feces may be deduced, Fukuyama, Hiroshima 729-0292
especially concerning a degradation of the molecular weight. Japan
It is not clear if the presence of peptide-linkage found in the E-mail: yagi@fupharm.fukuyama-u.ac.jp
isolated products 1 and 2 is due to a peptide-containing meta- Fax: +81-849/36-2024
bolite originating from AM or from coeluated mucin, being
part of human feces. The isolation procedures, ion exchange
and gel permeation chromatographies, used in our experi-
ments are not capable to differentiate between these two pos-
sibilities. A starting material, AM, does not contain peptide
moiety, but it showed a small amount of hexosamine after hy-
drolysis on HPAE. The finding is explained by the facts that
contamination of a small amount of hexosamine with AM oc-
curred during experimental procedure. We reported that phe-
nolic components were contaminated with glycoprotein frac-
tion in Aloe barbadensis gel powders in a previous paper (4).
The finding indicates that biological activity of aloemannan,
which was separated from the retentive fraction after dialysis
by gel permeation on DEAE Sephacel, may come from not only
the polysaccharide moiety but also the contaminated hexosa-
mine. It has been reported that high molecular weight aloe-
mannans exhibit better in immunomodulation (1). However,
contaminated hexosamine in aloemannan may also contrib-

Downloaded by: NYU. Copyrighted material.

ute to immunomodulation.
Acknowledgements

We are grateful to Aloecorp (Texas, U.S.A.) for providing the

Aloe barbadensis gel, to Dr. T. Ishizu for NMR spectral meas-
urements, and to Mr. S. Yano for his technical assistance.
References

1 Grindlay, D., Reynolds, T. (1996) J. Ethnopharmacol. 16, 117.

2 Gowda, D. C., Neelisiddaiah, B., Anjaneyalu, Y. V. (1979) Carbo-
hydr. Res. 72, 201.
3 Mandal, G., Das, A. (1980) Carbohydr. Res. 87, 249.
4 Yagi, A., Egusa, T., Arase, M., Tanabe, M., Tsuji, H. (1997) Planta
Med. 63, 18.
5 Belder, A. N., d., Granath, K. (1973) Carbohydr. Res. 30, 375.
6 Kurtzhals, P., Larsen, C., Johansen, M. (1997) J. Chromatogr. 491,
117.
7 Kaneo, Y., Uemura, T., Tanaka, T., Kanoh, S. (1997) Biol. Pharm.
Bull. 20, 181.
8 Gammon, D., Stephen, A. M., Churms, S. C. (1986) Carbohydr. Res.
158, 157.
9 Strickland, F. M., Pelley, R. P., Kripke, M. L. (1994) J. Invest. Derma-
tol. 102, 197.