Antidiabetic properties of selected Mexican copalchis of the Rubiaceae family

Article abstract of DOI:10.1016/j.phytochem.2007.05.006

The extracts prepared from the stem barks of several Mexican copalchis species, including Hintonia latiflora, Exostema caribaeum and a commercial mixture of Hintonia standleyana and E. caribaeum (CM) showed significant hypoglycemic and antihyperglycemic effects. The extracts were tested in three different in vivo models using normal and streptozotocin (STZ)-induced diabetic rats. From the active extract of H. latiflora, 25-O-acetyl-3-O-β-d-glucopyranosyl-23,24-dihydrocucurbitacin F (1), an analog of 23,24-dihydrocucurbitacin F, and several known compounds (2-8) were isolated; cucurbitacin 1 was also isolated from H. standleyana. Oral administration of H. latiflora extract [100 mg/kg of body weight (bw)] and 5-O-β-d-glucopyranosyl-7,3′,4′-trihydroxy-4-phenylcoumarin (5) (30 mg/kg of bw) to STZ-induced diabetic rats, for a 30 day duration, restored blood glucose levels to normal values. The groups treated either with the active principle 5, or the extract of H. latiflora, showed less body weight loss than glibenclamide-treated and diabetic control groups (p < 0.05). It was also demonstrated that the extract of H. latiflora regulated hepatic glycogen and plasma insulin levels (p < 0.05). These data suggest that its antihyperglycemic effect is due in part to stimulation of insulin secretion and regulation of hepatic glycogen metabolism.

Full text of DOI:10.1016/j.phytochem.2007.05.006

PHYTOCHEMISTRY
Phytochemistry 68 (2007) 2087–2095
www.elsevier.com/locate/phytochem
Antidiabetic properties of selected Mexican copalchis of the
Rubiaceae family
a,1
a
a
b
Jos e´ Guerrero-Analco , Omar Medina-Campos , Fernando Brindis , Robert Bye ,
a
a
a,
*
Jos e´ Pedraza-Chaverri , Andr e´ s Navarrete , Rachel Mata
a
Facultad de Qu ı´ mica, Universidad Nacional Aut o´ noma de M e´ xico, M e´ xico DF, 04510, Mexico
Instituto de Biolog ı´ a, Universidad Nacional Aut o´ noma de M e´ xico, M e´ xico DF, 04510, Mexico
b
Received 29 November 2006; received in revised form 16 April 2007
Available online 18 June 2007
Abstract
The extracts prepared from the stem barks of several Mexican copalchis species, including Hintonia latiflora, Exostema caribaeum and
a commercial mixture of Hintonia standleyana and E. caribaeum (CM) showed significant hypoglycemic and antihyperglycemic effects.
The extracts were tested in three different in vivo models using normal and streptozotocin (STZ)-induced diabetic rats. From the active
extract of H. latiflora, 25-O-acetyl-3-O-b-D-glucopyranosyl-23,24-dihydrocucurbitacin F (1), an analog of 23,24-dihydrocucurbitacin F,
and several known compounds (2–8) were isolated; cucurbitacin 1 was also isolated from H. standleyana. Oral administration of H. latifl-
0
0
ora extract [100 mg/kg of body weight (bw)] and 5-O-b-D-glucopyranosyl-7,3 ,4 -trihydroxy-4-phenylcoumarin (5) (30 mg/kg of bw) to
STZ-induced diabetic rats, for a 30 day duration, restored blood glucose levels to normal values. The groups treated either with the active
principle 5, or the extract of H. latiflora, showed less body weight loss than glibenclamide-treated and diabetic control groups (p < 0.05).
It was also demonstrated that the extract of H. latiflora regulated hepatic glycogen and plasma insulin levels (p < 0.05). These data sug-
gest that its antihyperglycemic effect is due in part to stimulation of insulin secretion and regulation of hepatic glycogen metabolism.
ꢀ
2007 Elsevier Ltd. All rights reserved.
Keywords: Hintonia latiflora; Hintonia standleyana; Exostema caribaeum; Copalchi; Rubiaceae; 4-Phenylcoumarins; Cucurbitacins; STZ-induced diabetic
rats; Diabetes mellitus
1
. Introduction
ex DC), H. standleyana Bullock, Exostema caribeaum
Jacq.) Roem et Schult and E. mexicanum Gray. Previous
(
In Mexico, circa 10% of the population suffers non-insu-
chemical investigation of these species resulted in isolation
and identification of several cucurbitacins and 4-phen-
ylcoumarins (Reher and Kraus, 1984; Mata et al., 1987,
1988, 1990a,b, 1992; Reguero et al., 1987; Guerrero-Analco
et al., 2005; Argotte-Ramos et al., 2006), a phenylstyrene
(Mata et al., 1992), as well as an indole monoterpenoid
alkaloid (D e´ ciga-Campos et al., 2006). From a pharmaco-
logical point of view, the antimicrobial (Rojas et al., 1992)
and antiplasmodial properties of the extracts and some 4-
phenylcoumarins from H. latiflora, E. caribaeum and
E. mexicanum were also demonstrated (Noster and Kraus,
1990; K o¨ hler et al., 2001). Additionally, the extract of
H. standleyana, as well as its cucurbitans, induced antino-
ciceptive effects in mice. We also demonstrated that the
lin-dependent diabetes mellitus (NIDDM). As in other
regions of the world, most of the Mexican people employ
a variety of plant preparations for treatment of this illness
(
Subash-Babu and Mainzen-Prince, 2004). Among the
plants highly valued for treatment of NIDDM are several
of the Rubiaceae with extremely bitter stem barks. These
species are commonly regarded as ‘‘copalchis’’ and the
most important are Hintonia latiflora (Sesse et Moci n˜ o ex
DC) Bullock (synonym: Coutarea latiflora Sesse et Moci n˜ o
*
Corresponding author. Tel.: +52 5 5622 5289; fax: +52 5 5622 5329.
E-mail address: rachel@servidor.unam.mx (R. Mata).
Taken in part from the PhD thesis of J.A. Guerrero-Analco.
1
0
031-9422/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2007.05.006

2088
J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
extracts of both Hintonia species are non-toxic to mice
when tested according to the Lorke procedure (D e´ ciga-
Campos et al., 2006; D e´ ciga-Campos et al., 2007); however,
the extract of E.caribaeum result in various toxic effects
such as inducing tremor, respiratory distress as well as a
decrease in motor activity (D e´ ciga-Campos et al., 2007).
The antidiabetic properties of H. latiflora have been
extensively investigated, with the earliest work carried out
at the beginning of the 20th century at the Instituto M e´ dico
Nacional in Mexico City (Mart ´ı nez, 1989). Later, research-
ers in Germany (Khur, 1953; Kaiser and Geyer, 1955; Sli-
jepcevic and Kraus, 1997; Korec et al., 2000), France
(
Bastien, 1961), and Italy (Pinto et al., 1997) corroborated
the work of the Mexican scientists using short term animal
experiments. Recently, Korec and co-workers (2000), using
short term experiments, reported that oral administration
of an extract prepared from a commercial mixture of the
stem barks of H. latiflora and E. caribaeum (CM) of uncer-
tain origin decreased glucose levels in streptozotocin
(
STZ)-diabetic rats. In that study, however the natural
active principles contributing to this activity were not
determined. More recently, our research group studied
the antidiabetic potential of H. standleyana (Guerrero-
Analco et al., 2005); these results demonstrated the antihy-
perglycemic effect of an organic extract of the stem bark of
H. standleyana and compounds 2 and 6 (Guerrero-Analco
et al., 2005). The relevant issues raised in this study were: (i)
the establishment of active principles of a Mexican copal-
chi, which turned out to be 4-phenylcoumarin glycosides
and cucurbitacins, and (ii) the demonstration in a long
term assay that compounds 2 and 6 reduced blood glucose
levels near to normal values in STZ-induced diabetic rats.
Continuing our work on Mexican copalchis of the Rubi-
aceae family (Guerrero-Analco et al., 2005), we describe
herein the long term antihyperglycemic effect of the
extracts prepared from H. latiflora, E. caribaeum and a
commercial mixture of E. caribaeum and H. standleyana
(
CM) that were purchased in local markets. Finally, the
effect of copalchis extracts and their antihyperglycemic
compounds on plasma insulin and hepatic glycogen levels
were assessed on STZ-induced diabetic rats. These two
experiments were performed in order to initiate the studies
regarding the antidiabetic mode of action of Mexican
copalchis herbal preparations and their active principles.
2
. Results and discussion
As shown in Tables 1 and 2, the organic extracts
(
100 mg/kg and 300 mg/kg of bw) of the stem-barks of
H. latiflora, E. caribaeum and CM caused significant
decreases in blood glucose levels, in both normal and
STZ-diabetic rats when compared with vehicle-treated
groups (p < 0.05). The results of H. latiflora are thus in
agreement with those previously described by Bastien
(
1961) and Pinto et al. (1997). In normoglycemic animals,
the hypoglycemic effect of H. latiflora was around 24%

J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
2089
and persisted throughout the experiment (Table 1). In
diabetic animals, the highest antihyperglycemic effect of
the H. latiflora extract was observed at doses of 100 and
3
2
00 mg/kg at 9 h (ꢀ21.0 and ꢀ33.4%, respectively) (Table
). In the case of the extract of E. caribaeum, the hypogly-
cemic effect was around 17% in both models (Tables 1 and
) at doses of 100 and 300 mg/kg at 7 h (ꢀ18.3% and
16.8%, respectively). Glibenclamide (10 mg/kg), used as
positive control, showed maximum hypoglycemic effect at
h in normoglycemic animals (Table 1) and at 5 h in
2
ꢀ
7
STZ-induced diabetic rats (Table 2).
In order to isolate the major active principles, the extract
of H. latiflora was subjected to an extensive chromato-
graphic fractionation. This process led to the isolation of
compounds 1–8 identified as 25-O-acetyl-3-O-b-D-gluco-
pyranosyl-23,24-dihydrocucurbitacin F (1), a previously
not described cucurbitacin glycoside, 3-O-b-D-glucopyr-
anosyl-23,24-dihydrocucurbitacin F (2), 5-O-b-D-glucopyr-
0
0
anosyl-7-methoxy-3 ,4 -dihydroxy-4-phenylcoumarin (3) 5-
0
0
O-b-D-galactopyranosyl-7-methoxy-3 ,4 -dihydroxy-4-phe-
0
0
nylcoumarin (4), 5-O-b-D-glucopyranosyl-7,3 ,4 -trihy-
droxy-4-phenylcoumarin (5), 5-O-[b-D-apiofuranosyl-
0
0
(
1 ! 6)-b-D-glucopyranosyl]-7-methoxy-3 ,4 -dihydroxy-4-
phenyl-coumarin (6), 5-O-[b-D-xylopyranosyl-(1 ! 6)-b-D-
0
0
glucopyranosyl]-7-methoxy-3 ,4 -dihydroxy-4-phenylcoum-
arin (7), and the alkaloid desoxycordifolinic acid (8). Com-
pound 1 was also isolated from the related species H. stan-
dleyana. The structure of compound 1 was elucidated
using spectrometric and spectroscopic methods including
uni- and bi-dimensional NMR spectroscopic techniques.
Compounds 2–8 were characterized by comparison with
authentic samples (Aquino et al., 1987; Mata et al., 1987,
1988, 1990a, 1992; Reguero et al., 1987; Guerrero-Analco
et al., 2005; D e´ ciga-Campos et al., 2006).
The HRFAB mass spectrum (positive mode) of 1
+
showed a quasi-molecular ion at m/z 725.8731 [M+H]
consistent with the molecular formulae C H O . The
3
8
60 13
NMR spectra (CDOD ) were very similar to 23,24-dihyd-
3
rocucurbitacin F glucoside previously isolated from
H. latiflora and H. standleyana (Mata et al., 1990a; Guer-
rero-Analco et al., 2005). Furthermore, acid hydrolysis of
compound 1 yielded 25-O-acetyl-23,24-dihydrocucurbita-
1
3
cin F (1a) (Reguero et al., 1987) and glucose. The
C
NMR chemical shifts observed in the region of the spec-
trum between d 107 and 60 confirmed the presence of glu-
C
0
0
0
0
cose [d 106.8 (C-1 ), 76.2 (C-2 ), 78.5 (C-3 ), 72.1 (C-4 ),
C
0
0
7
8.0 (C-5 ), 62.7 (C-6 )] in the molecule. The value of the
coupling constant (J = 7.5 Hz) of the signal for the ano-
meric hydrogen (dH 4.48) established a b-configuration.
On the other hand, the chemical shift values of C-3 (dC
93.8) and C-2 (dC 71.5) indicated that the sugar moiety
was linked to the cucurbitacin aglycone through the hydro-
xyl group at C-3. Furthermore, the HMBC correlation
between C-3 and the anomeric hydrogen is consistent with
this proposal. On the basis of these considerations, com-
pound 1 was characterized as 25-O-acetyl-3-O-b-D-gluco-
pyranosyl-23,24-dihydrocucurbitacin F.

2090
J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
The long term antihyperglycemic effect of the three
extract to prevent body weight loss could be attributed to
the presence of E. caribaeum.
extracts (E. caribaeum, CM and H. latiflora) and pure com-
pounds 2–6 and 8, using a classical chronic experiment with
STZ-induced diabetic rats (Verspohl, 2002; Subash-Babu
and Mainzen-Prince, 2004; Sezick et al., 2005; Jong-Yuh
and Mei-Fen, 2005), was next performed. The results dem-
onstrated that daily oral administration of the extract
(
50 mg/kg each time) of H. latiflora, H. standleyana and
compounds 2–6 and 8 (15 mg/kg each time), twice a day,
for 30 days, induced a significantly more pronounced anti-
hyperglycemic effect in the STZ-diabetic rats (Figs. 1 and
2
). The extracts of CM and E. caribaeum were less efficient
in decreasing blood glucose levels in diabetic rats (Fig. 1),
and compound 8 was inactive (Fig. 2). Compounds 3 and
4
also restored blood glucose levels to near normal values
at the end of the experiment [Fig. 2 (30 days)]. However,
the extract of H. latiflora (Fig. 1) and compound 5
(
Fig. 2) brought back normal blood glucose levels, and
the effect was comparable to that of the glibenclamide used
as a positive control (Verspohl, 2002; Subash-Babu and
Mainzen-Prince, 2004). The treatments with H. latiflora
extract, as well as those with compounds 2 and 6 (Guer-
rero-Analco et al., 2005), also prevented body weight loss
in hyperglycemic rats (Table 3); however, the effect was
higher than that of glibenclamide. E. caribaeum and CM
extracts, on the other hand, did not prevent weight loss
in diabetic rats. In our previous study on the extract of
H. standleyana and compound 2 and 6, we had found that
they prevented body weight loss in diabetic rats (Guerrero-
Analco et al., 2005); therefore, the inefficacy of the CM
6
5
4
3
2
1
00
00
00
00
00
00
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
0
5
10
15
20
25
30
Time (days)
Comparison of the antihyperglycemic activity of com-
pounds 3–5 established that the most active compound
possesses a free hydroxyl group at C-7 in the 4-phenyl-
coumarin core. On the other hand, the comparison of the
activity of compounds 3 and 4 indicates that the nature
of the sugar moiety (glucose vs. galactose) does not influ-
ence biological activity.
One of the possible mechanisms by which the copalchis
extracts and active principles exert their antihyperglycemic
action in rats could be due to an insulin-release stimulatory
Normal Control (vehicle)
Diabetic Control (vehicle)
Glibenclamide (30 mg/kg)
H. latiflora extract (100 mg/kg)
E. caribaeum extract (100 mg/kg)
CM extract (100 mg/kg)
H. standleyana extract (100 mg/kg)
Fig. 1. Long-term effect of the extracts of H. latiflora, E. caribeaum, H.
standleyana, CM (commercial mixture of copalchi) on blood glucose levels
in STZ-diabetic rats. Each value is the mean ± SEM for six rats in each
group. *p < 0.05 significantly different ANOVA followed by Dunnett’s
t-test for comparison with the diabetic control group at same time.

J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
2091
might increase the renewal of b-cells in the pancreas or per-
mit the recovery of partially destroyed b-cells and stimu-
lates pancreatic insulin secretion (Subash-Babu and
Mainzen-Prince, 2004). Since the treatments with the Hin-
tonia extracts and compounds 2, 5 and 6 did not restore to
normal values the insulin levels, it is highly possible that
they exert their antihyperglycemic through an insulinomi-
metic mode of action (Verspohl, 2002).
6
5
4
3
2
1
00
00
00
00
00
00
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
The extract of H. latiflora significantly improved hepatic
glycogen content in STZ-diabetic rats (Fig. 4) as well
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
1
0
8
0
5
10
15
20
25
30
*
Time (days)
Normal Control (vehicle)
*
Diabetic Control (vehicle)
Glibenclamide (30 mg/kg)
Compound 2 (30 mg/kg)
Compound 3 (30 mg/kg)
Compound 4 (30 mg/kg)
Compound 5 (30 mg/kg)
Compound 6 (30 mg/kg)
Compound 8 (30 mg/kg)
*
6
4
2
0
*
*
*
*
*
Fig. 2. Long-term effect of compounds 2–6 on blood glucose levels in
STZ-diabetic rats. Each value is the mean ± SEM for six rats in each
group. *p < 0.05 significantly different ANOVA followed by Dunnett’s
t-test for comparison with the diabetic control group at same time.
Samples tested
effect; therefore, their action on plasma insulin on STZ-
induced diabetic rats was investigated. This model was
used because the induction of diabetes with a lower dose
of STZ (50 mg/kg) produced an incomplete destruction
of pancreatic b-cells even though the rats become perma-
nently diabetic (Subash-Babu and Mainzen-Prince, 2004).
Thus, after treatment with 50 mg/kg of STZ there should
be some surviving b-cells. Furthermore, with some treat-
ments their regeneration is also possible. The results sum-
marized in Fig. 3 clearly demonstrate that the extracts of
H. latiflora and H. standleyana as well as compounds 2, 5
and 6 induced the secretion of insulin in a similar way to
the positive control glibenclamide. Given that the treat-
ments induced an increment in plasma insulin levels, they
Normal control (vehicle)
Diabetic control (vehicle)
Glibenclamide (30 mg/Kg)
H. latiflora extract (100 mg/Kg)
H. standleyana extract (100 mg/Kg)
E. caribaeum extract (100 mg/Kg)
CM extract (100 mg/Kg)
Compound 2 (30 mg/Kg)
Compound 3 (30 mg/Kg)
Compound 4 (30 mg/Kg)
Compound 5 (30 mg/Kg)
Compound 6 (30 mg
Fig. 3. Effect of the extracts of H. latiflora, H. standleyana, E. caribaeum,
CM (commercial mixture of copalchi), and compounds 2–6 on plasma
insulin levels in STZ-diabetic rats. Each value is the mean ± SEM for six
rats in each group. p < 0.05 significantly different ANOVA followed by
Dunnett’s t-test vs. the diabetic control group at same time.
*
Table 3
Long-term effect of the extracts of H. latiflora, E. caribeaum, CM (commercial mixture of copalchi) and compounds 3–5 and 8 on body weight, in STZ-
diabetic rats
Test samples
Body weight (g)
Dose (per os) mg/kg of bw
1st day
7th day
15th day
30th day
Normal control (vehicle)
Diabetic Control (vehicle)
Glibenclamide
Extract of H. latiflora
Extract of E. caribeaum
Extract of CM
Compound 3
Compound 4
Compound 5
Compound 8
–
–
213.4 ± 3.9*
169.8 ± 5.2
178.3 ± 7.8
182.9 ± 7.7
177.9 ± 10.6
173.3 ± 9.2
169.7 ± 4.8
169.6 ± 11.6
190.3 ± 6.6
170.3 ± 6.6
249.9 ± 3.0*
216.2 ± 5.2
210.2 ± 4.1
227.2 ± 8.0
213.4 ± 11.9
214.9 ± 5.9
213.3 ± 6.6
205.9 ± 8.9
218.8 ± 8.7
188.8 ± 7.8
274.6 ± 4.2*
194.8 ± 6.0
212.8 ± 4.9
238.3 ± 12.4*
223.3 ± 13.2
220.8 ± 15.3
219.2 ± 8.1
211.3 ± 12.0
225.2 ± 12.0*
205.2 ± 11.0
288.2 ± 5.5*
204.2 ± 9.2
216.3 ± 8.9
248.3 ± 16.4*
217.7 ± 17.6
223.1 ± 20.4
230.3 ± 10.0*
221.3 ± 14.6
241.3 ± 12.3*
201.3 ± 13.2
30
100
100
100
30
30
30
30
Each value is the mean ± SEM for six rats in each group.
*
p < 0.05 significantly different ANOVA followed by Dunnett’s t-test for comparison with the diabetic control group at same time.

2092
J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
alkaloids. The lack of efficacy of E. caribaeum, as well as
the toxic effects previously demonstrated, should cause
Mexican Health Authorities to alert local consumers about
the risks in using this plant.
*
*
4. Experimental
4.1. General experimental procedures
Melting points were determined on a Fisher–Johns
apparatus and are uncorrected. IR spectra were recorded
as KBr pellets or film on a Perkin–Elmer 59913 spectro-
photometer. Optical rotations were registered on a Per-
kin–Elmer 241 digital polarimeter. NMR spectra were
Samples tested
recorded on a Varian VXR-300S spectrometer in CD OD,
Normal Control (vehicle)
Diabetic Control (vehicle)
Glibenclamide (30 mg/Kg)
H. latiflora extract (100 mg/Kg)
H. standleyana extract (100 mg/kg)
E. caribaeum extract (100 mg/Kg)
CM extract (100 mg/Kg)
Compound 2 (30 mg/Kg)
Compound 3 (30 mg/Kg)
Compound 4 (30 mg/Kg)
Compound 5 (30 mg/Kg)
Compound 6 (30 mg/Kg)
3
1
13
either at 300 MHz ( H) or 75 ( C) MHz, using tetrameth-
ylsilane (TMS) as an internal standard. EIMS were
obtained on a JMS-AX505HA mass spectrometer, whereas
positive FABMS data were obtained using a JEOL SX 102
mass spectrometer with a NBA matrix. Open column chro-
matography (CC) was carried out on silica gel 60 (70–
230 mesh, Merck). TLC analyses were performed on silica
gel 60 F254 plates (Merck), and visualization of the plates
was carried out using a ceric sulfate (10%) solution in
H SO or anisaldehyde-H SO reagent (for sugars).
Fig. 4. Effect of the extracts of H. latiflora, H. standleyana, E. caribaeum,
CM (commercial mixture of copalchi), and compounds 2–6 on glycogen
hepatic content in STZ-diabetic rats. Each value is the mean ± SEM for
2
4
2
4
*
six rats in each group. p < 0.05 significantly different ANOVA followed
by Dunnett’s t-test vs. the diabetic control group at same time.
4.2. Plant material
The stem bark of H. latiflora was collected in Chihuahua
in June 2003 and identified by Dr. Robert Bye (Jard ´ı n Bot-
anico, UNAM). A voucher specimen (Bye No 31, 735) is
deposited at the National Herbarium (MEXU), Mexico
City. The stem bark of H. standleyana Bullock was collected
in Atenango del Rio, Guerrero, in January 2003 and identi-
fied by Celso Guerrero and Francisco Ramos from the
National Herbarium (MEXU); a voucher specimen (P.
Hersch No 824) is deposited at the ASFM-INAH Herbar-
ium, Cuernavaca. The stem bark of E. caribeaum was col-
lected in Atenango del Rio, Guerrero, in January 2003
and identified by Celso Guerrero and Francisco Ramos
from the National Herbarium (MEXU). A voucher speci-
men (Marchena 1167098) is deposited at the ASFM-INAH
Herbarium, Cuernavaca, State of Morelos. A commercial
sample (CM) of ‘‘copalchi’’ composed by the stem barks
of H. standleyana and E. caribaeum in a 1:4 ratio was pur-
chased from the popular market of Sonora in Mexico City.
suggesting that it might modulate glycogen breakdown by
inhibiting glucose-6-phosphatase (G6Pase),
a crucial
enzyme for the final step of gluconeogenesis or glycogenol-
ysis. Several antidiabetic agents, such as metformin and
p-methoxycinnamic acid, a natural product, also inhibit
glucose production by suppressing G6Pase activity. Alter-
natively, the extract could induce an increment in glucoki-
nase activity in diabetic rats, which in turn regulated
glycogen metabolism (Adisakwattana et al., 2005).
3
. Conclusions
Our results corroborated further that the active princi-
ples of the Mexican copalchis of the Rubiaceae family
are both 4-phenylcoumarins glycosides and cucurbitacins.
This is thus the first report describing the antidiabetic prin-
ciples of H. latiflora and beyond any doubt answers any
questions raised by many researchers concerning the nature
of the active principles of this Mexican plant species. The
extract of H. latiflora and compound 5 were also the most
effective in regulating blood glucose levels in hyperglycemic
conditions, using both short and long term assays; their
mode of action involves stimulatory effect on insulin secre-
tion by b-cells in the pancreas. H. latiflora extract also stim-
ulate hepatic glycogen metabolism. As for H. standleyana,
H. latiflora biosynthesizes indole-monoterpenoid type
4.3. Extraction of stem bark of H. latiflora and isolation of
compounds 1–8
The air-dried and pulverized material (3.0 kg) was mac-
erated with CH Cl –MeOH (6 l · 3, 1:1) at room tempera-
2
2
ture for a duration of 6 days, with subsequent evaporation
of the solvent in vacuo to afford a crude extract (160 g). An
aliquot of this (115 g) was then dissolved in a minimal
amount of MeOH, and from this solution compound 4

J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
2093
(
36 g) crystallized, m.p. 230 ꢁC (lit m.p. 228–231 ꢁC; Mata
1630, 1455, 1430 and 1255; HRFABMS: m/z = 725.8731
+
et al., 1987). The mother liquors were next subjected to sil-
ica gel (1.0 kg) CC in a glass column using CHCl with
increasing amounts of MeOH as eluents to yield 12 pri-
mary fractions (FH1-FH12). Fraction FH2 (2 g), eluted
[M+H] (calcd. for C H O : 725.8752); FABMS:
3
8
60 13
+
+
m/z = 748 [M+Na+H] ; 665 [M–CH CO H+H] ; 563
3 2
[M–162+H] . NMR-Data [ C (75.5 MHz, CDOD )/ H
3
+
13
1
3
(300 MHz, CDOD )]: C-1 [dC 34.1/dH 1.78 (1H, m), 2.23
3
with CHCl –MeOH (92:8), was further applied to a second
(1H, m)], C-2 [dC 71.5/dH 3.84 (1H, m)], C-3 [dC 93.8/dH
2.99 (1H, d, J = 9 Hz)], C-4 (d 43.3), C-5 (d 142.5), C-6
3
silica gel column using CHCl –EtOAc–MeOH (47.5:47.5:5)
3
C
C
as a mobile phase to give nine fractions (FH2I–FH2IX).
From fraction FH2III, compound 1 (200 mg) spontane-
ously crystallized. From fraction FH2V, compound 2
[dC 120.5/dH 5.78 (1H, bd, J = 5.4 Hz)], C-7 [dC 25.5/dH
1.21 (1H, dd, J = 13.2, 2.1 Hz), 1.37 (1H, m)], C-8 [dC
44.3/dH 1.93 (1H, m)], C-9 (dC 49.1), C-10 [dC 34.8/dH
2.46 (1H, m)], C-11 (dC 216.0), C-12 [dC 49.8/dH 3.22
(2H, m)], C-13 (dC 48.7), C-14 (51.8), C-15 [dC 46.7/dH
1.93 (2H, m)], C-16 [dC 71.5/dH 3.30 (1H, m)], C-17 [dC
59.5/dH 2.52 (1H, d, J = 7.2 Hz)], C-18 [dC 20.5/dH 0.88
(3H, s)], C-19 [dC 19.7/dH 1.36 (3H, s)], C-20 (dC 80.9),
C-21 [dC 25.6/dH 1.28 (3H, s)], C-22 (dC 216.6), C-23 [dC
(
1.0 g) crystallized, m.p. 201 ꢁC (lit m.p. 199–201 ꢁC; Mata
et al., 1990). From fraction FH5, eluted with CHCl3–
MeOH (9:1), compound 3 (4.5 g), m.p. 238 ꢁC (lit m.p.
2
37–238 ꢁC; Mata et al., 1988). Fraction FH9 eluted with
CHCl –MeOH (85:15), yielded compound 5 (6.5 g), m.p.
3
2
51 ꢁC (lit m.p. 250–252 ꢁC; Mata et al., 1990). Column
chromatography over silica gel (600 g) of fraction FH10
15 g), eluted with CHCl –MeOH (75:25), using CHCl –
32.8/dH 1.58 (2H, m)], C-24 [d 35.8/dH 2.02 (2H, m)], C-
C
(
25 (dC 83.1), C-26 [dC 26.2/dH 1.44 (3H, s)], C-27 [dC
26.2/dH 1.44 (3H, s)], C-28 [dC 23.4/dH 0.99 (3H, s)], C-
3
3
MeOH (9.5:0.5 ! 0:1) as mobile phase yielded eight sec-
ondary fractions, FH10I-FH10VIII. From fraction
29 [dC 24.8/d 1.28 (3H, s)], C-30 [dC 20.4/dH 1.1 (3H,
H
0
0
FH10IV (1 g), eluted with CHCl –MeOH (8:2), compound
s)], C-1 [dC 106.8/dH 4.48 (1H, d, J = 7.5 Hz)], C-2 [
76.2/d 3.30 (1H, m)], C-3 [d 78.5/d 3.34 (1H, m)], C-
3
C
0
6
(100 mg), m.p. > 300 ꢁC (lit m.p. > 300 ꢁC; Guerrero-
Analco et al., 2005) was obtained by crystallization
CHCl –MeOH 8:2). From fraction FH10VI (3.6 g) eluted
H
C
H
0
0
4 [d 72.1/d 3.32 (1H, m)], C-5 [dC 78.0/dH 3.22 (1H,
C
H
0
(
m)], C-6 [dC 62.7/dH 3.69 (1H, dd, J = 11.7,5.1 Hz), 4.06
3
with CHCl –MeOH (7:3), compound 7 (468 mg) crystal-
(1H, m)], CH -C@O [dC 22.3/dH 1.95 (3H, s)],
3
3
lized. Finally, column chromatography over silica gel
CH ꢀ C = O (d 172.4).
3
C
(
100 g) of FH12 (2 g), eluted with CHCl –MeOH (6:4),
3
using CH Cl –MeOH (8:2 ! 0:1) as mobile phase yielded
4.6. Acid hydrolysis of 1
2
2
five secondary fractions, FH12I-FH12V. From fraction
FH12III (10 g), eluted with CHCl –MeOH (4:6),
Compound 1 (5 mg), dissolved in 2 M HCl (10 ml) was
heated until reflux began, this being held for 2 h. A white
powder precipitated from the acid solution and after usual
work-up, aglycone 1a (2 mg) was obtained (Reguero et al.,
1987); glucose was identified by TLC following comparison
with an authentic sample.
3
compound 8 (1 g) crystallized, m.p. > 300 ꢁC (lit m.p. >
3
00 ꢁC; D e´ ciga-Campos et al., 2006).
4
.4. Extraction of 25-O-acetyl-3-O-b-D-glucopyranosyl-
23,24-dihydrocucurbitacin F (1) from the stem bark of
H. standleyana
4.7. Extraction of the stem bark of E. caribeaum and CM
(commercial mixture of copalchi)
Dried and shredded stem bark (300 g) was macerated
with CH Cl –MeOH (2 l · 3, 1:1) for 6 days at room tem-
2
2
perature. The combined extracts were evaporated in vacuo
to yield a brown residue (80 g), which was applied to a glass
column packed with silica gel (800 g). This was next eluted
with hexane–EtOAc (1:1 ! 0:1) and EtOAc–MeOH
Dried and shredded stem bark (250 g) of E. caribeaum
was macerated with CH Cl –MeOH (6 l · 3, 1:1) for 6 days
2
2
at room temperature. The combined extracts were evapo-
rated in vacuo to give a light brown residue (80 g). Dried
and shredded stem bark (300 g) of CM was macerated with
CH Cl –MeOH (1:1) (6 l · 3) for 6 days at room tempera-
(
1:0 ! 0:1) to give five fractions (FS1–FS5). Fraction
FS2 (2.0 g), eluted with EtOAc–MeOH (9:1), was further
2
2
subjected to silica gel CC using CHCl –EtOAc–MeOH
ture. The combined extracts were evaporated in vacuo to
yield a brown residue (90 g).
3
(
47.5:47.5:5) as mobile phase to give 10 fractions (FS4I–
FS4X). From fraction FH4V, compound 1 (30 mg)
spontaneously crystallized; from fraction FH4X precipi-
tated compound 2 (120 mg), m.p. 201–203 ꢁC (lit m.p.
4.8. Biological assays
1
99–201 ꢁC; Guerrero-Analco et al., 2005).
4.8.1. Test animals
Male Wistar normoglycemic rats, 55–60-day-old, weight
180–220 g, were obtained from Centro UNAM-Harlan
(Harlan M e´ xico, SA de CV). Procedures involving animals
and their care were conducted in conformity with the Mex-
ican Official Norm for Animal Care and Handing (NOM-
062-ZOO-1999) and in compliance with international rules
4
dihydrocucurbitacin F (1)
.5. 25-O-Acetyl-3-O-b-D-glucopyranosyl-23,24-
Glassy white solid; m.p. 158–161 ꢁC, [a] + 45(c = 0.2,
MeOH); IR mmax (KBr) cm : 3400, 2920, 1710, 1690,
D
ꢀ
1

2094
J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
on care and use of laboratory animals. The animals were
housed under standard laboratory conditions and main-
tained on a standard pellet diet and water ad libitum.
4.8.6. Effects of daily treatment with crude extracts and pure
compounds in STZ-induced diabetic rats
DM was induced in 66 rats by the same method
showed in the section of the induction of experimental
diabetes. Thereafter, the extracts, pure compounds and
glibenclamide were administered twice (9 a.m. and
6 p.m.) on a daily basis, for a period of 30 days, to
STZ-diabetic rats which were divided into eight groups
(A–H). A nine group of normal rats (Group I) was used
as a control. Groups H (diabetic control) and I received
4
.8.2. Preparation of the test samples
Samples were suspended in 0.05% Tween 80 in saline
ꢂ
solution. Glibenclamide (Sigma Co., St. Louis, MO,
USA) was used as a hypoglycemic model drug (Verspohl,
2
002). Control rats group only received the vehicle
ꢂ
(
0.05% Tween 80 in saline solution) in the same volume
0.5 ml of vehicle/100 g of bw) by the same route.
ꢂ
(
vehicle (0.05% Tween 80 in saline solution) per os dur-
ing 30 days. Group A received glibenclamide (15 mg/kg;
30 mg/kg daily). Groups B to G received different treat-
ments: B, H. latiflora extract (50 mg/kg; 100 mg/kg
daily); C, E. caribaeum extract (50 mg/kg; 100 mg/kg
daily); D, CM of ‘‘copalchi’’ (50 mg/kg; 100 mg/kg daily);
E, compound 3 (15 mg/kg; 30 mg/kg daily); F, compound
4 (15 mg/kg; 30 mg/kg daily) and G, compound 5
(15 mg/kg; 30 mg/kg daily). Compounds 2, 6 (15 mg/kg;
30 mg/kg daily) and H. standleyana extract (50 mg/kg;
100 mg/kg daily) were also included in the experiments
for comparative purposes. All tested materials were sus-
pended in the same vehicle and administered per os daily
for 30 days. The animals were housed under standard
laboratory conditions and maintained with free access
to water and food during all the experiment. Blood glu-
cose concentration on both diabetic (Groups A–H) and
control group (I) animals was estimated at days 0, 3, 7,
4.8.3. Collection of blood samples and determination of blood
glucose levels
Blood samples were collected from the caudal vein by
means of a little incision at the end of the tail. Blood glu-
cose levels (mg/dl) were estimated by enzymatic glucose
oxidase method using a commercial glucometer (One
Touch Ultra, Jonhson–Jonhson, California, USA). The
percentage variation of glycemia for each group was calcu-
lated with respect to initial (0 h) level according to:
%
Variation of glycemia = [(G –G )/G ] · 100%
i
t
i
where G is initial glycemia values and G is the glycemia
i
t
value after samples administration.
4
.8.4. Acute hypoglycemic assay
Normal rats were placed in single cages with wire-net
10, 12, 15, 21 and 30. The body weights of the animals
were measured concomitantly to the blood glucose anal-
yses. All experiments were carried out using six animals
per group.
floors and deprived of food for 12 h before experimenta-
tion but allowed free access to tap water throughout.
The crude extracts (at the dose of 10, 30, 100 and
3
00 mg/kg of bw) of H. latiflora and E. caribaeum were
ꢂ
4.8.7. Determination of hepatic glycogen and plasma insulin
levels
suspended in 0.05% Tween 80 in saline solution. Gliben-
clamide (10 mg/kg of bw) was suspended in the same vehi-
cle. The samples were freshly prepared immediately before
experimentation and administered by the intragastrical
route. Control rats received only the vehicle (0.05%
Tween 80 in saline solution) by the same route. Blood
samples were collected at 0, 1.5, 3, 5, 7, 9 h after drugs
administration.
After 30 days of treatment with any of the extracts (H.
stadleyana; H. latiflora; E. caribaeum and CM) or tested
compounds (2–6; 8), the animals (normal and diabetic)
were decapitated after an overnight fast. Blood was col-
lected in heparinized tubes and plasma was separated after
centrifugation. Liver tissues were excised immediately and
stored in ice-cold containers. Hepatic glycogen content
was measured according to Murat and Serfaty (1974).
Briefly, liver samples (100 mg) were homogenized in 2 ml
of ice-cold citrate buffer (0.1 M, pH 4.2). Immediately after
homogenization, 10 ll of the sample was used to determine
free glucose in the tissue. Amyloglucosidase (exo-1,4-a-glu-
cosidase, EC 3.2.1.3, Cat #A9228, Sigma Chemical, Co, St
Louis, MO, USA) dry powder was then mixed with the
homogenate (1 mg/ml) and allowed to stand at room tem-
perature for 2 h. A 10 ll sample was then used to determine
total glucose amounts. Initial free glucose values were sub-
tracted, with the difference being the value used to calculate
the glycogen content in the tissue. Glucose was determined
using a commercial kit (Glucose TR, Cat 1001190, Spinre-
act, Girona, Spain). Plasma insulin levels were estimated
by the Rat Insulin Enzyme Immunoassay Method using
ꢂ
4
.8.5. Effect on diabetic animals (non-insulin dependent
diabetes model-NIDDM)
.8.5.1. Induction of experimental diabetes. Diabetes melli-
4
tus type 2 was induced in rats by a single intraperitoneal
injection of freshly prepared STZ (50 mg/kg; Sigma. St.
Louis, MO, USA) dissolved in 0.1 M citrate buffer, pH
4
.5, in a volume of 1 ml/kg of bw. After 7 days of STZ
administration, blood glucose levels of each rat were
determined. Rats with blood glucose levels higher than
2
50 mg/dl were considered diabetic and were included in
the study (Verspohl, 2002; Guerrero-Analco et al., 2005).
4
.8.5.2. Acute antihyperglycemic assay. Diabetic rats were
treated exactly as the normal rats (see Section 4.8.4).

J. Guerrero-Analco et al. / Phytochemistry 68 (2007) 2087–2095
2095
a SpiBio Kit (Microplate Reader BIO-RAD Model 680)
Grassi and Pradelles, 1991).
Guerrero-Analco, J.A., Hersch-Martınez, P., Pedraza-Chaverri, J.,
´
Navarrete, A., Mata, R., 2005. Antihyperglycemic effect of constitu-
ents from Hintonia standleyana in streptozotocin-induced diabetic rats.
Planta Med. 71, 1099–1105.
(
4
.9. Statistical analysis
Jong-Yuh, C., Mei-Fen, S., 2005. Potential hypoglycemic effects of
Chlorella in streptozotocin-induced diabetic mice. Life Sci. 77, 980–
990.
Data are expressed as the means ± SEM for the number
Kaiser, H., Geyer, H., 1955. Zur pharmakognosie der rinde von Coutareae
latiflora DC. Arch. Pharm. 535 (Chemie), 595 (Pharmakologie).
Khur, R., 1953. Orale diabetestherapie mit einem euhporbiaceenextrakt.
Landarzt 29, 1.
(
n = 6) of animals in each group. Repeated measurement
analysis of variance (ANOVA) was used to analyze the
changes in blood glucose and other parameters. Dunnett
range post-hoc comparisons were used to determine the
source of significant differences where appropriate;
K o¨ hler, I., Jenett-Siems, K., Mockenhaupt, F.P., Siems, K., Jakupovic, J.,
Gonzalez, J.C., Hernandez, M.A., Ibarra, R.A., Berendsohn, W.G.,
´
Bienzle, U., Eich, E., 2001. In vitro antiplasmodial activity of 4-
phenylcoumarins from Exostema mexicanum. Planta Med. 67, 89–91.
Korec, R., Heinz-Sensch, K., Zoukas, T., 2000. Effects of the neoflavonoid
coutareagenin, one of the antidiabetic active substances of Hintonia
latiflora, on streptozotocin-induced diabetes mellitus in rats. Drug Res.
(
p < 0.05) was considered statistically significant. Sigma
plot and Sigma Stat (version 9.0) software were used for
statistics and plotting.
5
0, 122–128.
Mart ´ı nez, M., 1989. Las Plantas Medicinales de M e´ xico, sixth ed.
Ediciones Botas, M e´ xico DF, pp. 85–87.
Acknowledgements
Mata, R., Calzada, F., Garcia, M., Reguero, M., 1987. Chemical studies
on Mexican plants used in traditional medicine III: new 4-phen-
ylcoumarins from Exostema caribaeum. J. Nat. Prod. 50, 866–870.
Mata, R., Calzada, F., Garcia, M., 1988. Chemical studies on Mexican
plants used in traditional medicine VI: additional new 4-phenylcouma-
rins from Exostema caribaeum. J. Nat. Prod. 51, 851–856.
Mata, R., Camacho, M., Cervera, E., Bye, R., Linares, E., 1990a.
Secondary metabolites from Hintonia latiflora. Phytochemistry 29,
2037–2040.
Mata, R., Albor, C., Pereda-Miranda, R., McLaughlin, J.L., 1990b.
Cytotoxic constituents of Exostema mexicanum. Planta Med. 56, 241.
Mata, R., Camacho, M., Mendoza, S., Cruz, M.C., 1992. A phenylstirene
from Hintonia latiflora. Phytochemistry 31, 3199–3201.
Murat, J.C., Serfaty, A., 1974. Simple enzymatic determination of
polysaccharide (glycogen) content of animal tissues. Clin. Chem. 20,
This work was supported by Grants of Consejo Nac-
ional de Ciencia y Tecnolog ´ı a (CONACyT, Proyecto
C01-018) and DGAPA-UNAM IN212005. We are in-
debted to Alejandro Urzua, Victor Lemus and Chabetty
Vargas for their collaboration in this work. The technical
assistance of Isabel Rivero-Cruz, Laura Acevedo, Rosa
del Villar, Marisela Guti e´ rrez and Georgina Duarte-Lisci
is also recognized. J.A. Guerrero-Analco acknowledges fel-
lowships from CONACyT and Direcci o´ n General de
Estudios de Posgrado, UNAM, to support graduate
studies.
1
576–1577.
References
Noster, S., Kraus, L., 1990. In vitro antimalarial activity of Coutarea
latiflora and Exostema caribaeum extracts on Plasmodium falciparum.
Planta Med. 56, 63–65.
Pinto, A., Capasso, A., Sorrentino, L., 1997. Experimental animal studies
on the hypoglycemic effects of a copalchi extract. Drug Res. 47, 829–
833.
Reguero, M., Mata, R., Bye, R., Linares, E., Delgado, G., 1987. Chemical
studies on Mexican plants used in traditional medicine II: curcubitac-
ins from Hintonia latiflora. J. Nat. Prod. 50, 315–316.
Reher, G., Kraus, L., 1984. New neoflavonoids from Coutarea latiflora. J.
Nat. Prod. 47, 172–174.
Rojas, A., Hernandez, L., Pereda-Miranda, R., Mata, R., 1992. Screening
for antimicrobial activity of crude drug extracts and pure natural
products from Mexican medicinal plants. J. Ethnopharmacol. 35, 275–
283.
Adisakwattana, S., Roengsamran, S., Hsu, W.H., Yibchok-anun, S., 2005.
Mechanisms of antihyperglycemic effect of p-methoxycinnamic acid in
normal and streptozotocin-induced diabetic rats. Life Sci. 78, 406–412.
Aquino, R., D’Agostino, M., De Simone, F., Pizza, C., 1987. 4-
Arylcoumarin glycosides from Coutarea hexandra. Phytochemistry
27, 1827–1830.
Argotte-Ramos, R., Ram ´ı rez-Avila, G., Rodr ´ı guez-Guti e´ rrez, M.C.,
Ovilla-Mu n˜ oz, M., Lanz-Mendoza, H., Rodr ´ı guez, M., Gonz a´ lez-
Cortazar, M., Alvarez, Laura., 2006. Antimalarial 4-phenylcoumarins
from the stem bark of Hintonia latiflora. J. Nat. Prod. 69, 1442–1444.
Bastien, M., 1961. Recherches sur les copalchis, drogues hypoglyc e´ miantes
et en particulier sur le Coutarea latiflora DC (Rubiaceae). Dissertation,
Facult e´ de Pharmacie de l’Universit e´ , Paris.
D e´ ciga-Campos, M., Guerrero-Analco, JA., Quijano, L., Mata, R., 2006.
Antinociceptive activity of 3-O-b-D-glucopyranosyl-23,24-dihydrocu-
curbitacin F from Hintonia standleyana (Rubiaceae). Pharmacol.
Biochem. Behav. 83, 342–348.
D e´ ciga-Campos, M., Arriaga-Alba, M., Rivero-Cruz, I., Casta n˜ eda-
Corral, G., Angeles-L o´ pez, G., Navarrete, A., Mata, R., 2007. Acute
toxicity and mutagenic activity of Mexican plants used in traditional
medicine. J. Ethnopharmacol. 110, 334–342.
Grassi, J., Pradelles, Ph., 1991. Compounds labelled by the acetylcholin-
esterase of Electrophorus electricus. Its preparation process and its use
as a tracer or marker in enzymo-immunological determinations.
United States Patent, No. 1,047,330.
Sezick, E., Aslan, M., Yesilada, E., Ito, S., 2005. Hypoglycaemic activity
of Gentiana olivieri and isolation of active constituent through
bioassay-directed fractionation techniques. Life Sci. 76, 1223–1238.
Slijepcevic, M., Kraus, L., 1997. Diabetische m a¨ use als pr u¨ fmodell f u¨ r
hypoglyk a¨ mische aktivit a¨ ten von pflanzenextrakten und insulinpr a¨ -
paraten. Act. Ther. 23, 47–58.
Subash-Babu, P., Mainzen-Prince, P.S., 2004. Antihyperglycaemic and
antioxidant effect of hyponidd, an ayurvedic herbomineral formulation
in streptozotocin-induced diabetic rats. J. Pharm. Pharmacol. 56,
1435–1442.
Verspohl, E.J., 2002. Recommended testing in diabetes research. Planta
Med. 68, 581–590.