Journal of Medicinal Plants Research Vol. 3(8), pp. 576-582, August, 2009
Available online at http://www.academicjournals.org/JMPR
ISSN 1996-0875© 2009 Academic Journals
Full Length Research Paper
Multiple inflammatory and antiviral activities in
Adansonia digitata (Baobab) leaves, fruits and seeds
Vimalanathan Selvarani and Hudson James B.*
Department of Pathology and Laboratory Medicine, University of British Columbia, Vancouver, Canada.
Accepted 24 July, 2009
Adansonia digitata (Baobab) is a traditional African medicinal plant with numerous applications,
including treatment of symptoms of infectious diseases. Standardized commercial preparations of
Adansonia digitata leaves, fruit-pulp and seeds were acquired and extracted with three different
solvents, water, methanol and DMSO. The extracts were compared quantitatively for antiviral MIC100
(minimal inhibitory concentration) values against influenza virus, herpes simplex virus and respiratory
syncytial virus and for their effects on cytokine secretion (IL-6 and IL-8) in human epithelial cell
cultures. The leaf extracts had the most potent antiviral properties, especially the DMSO extracts and
influenza virus was the most susceptible virus. Pulp and seed extracts were less active but significant.
Cytotoxic activities were only evident at much higher concentrations of extract. Several of the extracts,
especially leaf extracts, were also active as cytokine modulators, some being pro-inflammatory and
others being anti-inflammatory. The results overall indicated the presence of multiple bioactive
compounds in different parts of the plant and these activities could explain some of the medical
benefits attributed to traditional leaf and pulp preparations, in the treatment of infectious diseases and
Key words: Adansonia digitata, Baobab, antiviral, inflammatory, cytokines.
The Baobab tree (Adansonia digitata L. Family
Bombacaceae) is indigenous in many African countries
(Wickens, 1982; Sidibe and Williams, 2002). Many parts
of the plant, especially leaves, fruit pulp, seeds and bark
fibers, have been used traditionally for medicinal and
nutritional purposes (Sidibe and Williams, 2002; Chadare
et al., 2009) and some commercial enterprises produce
standardized preparations derived from seeds, fruit pulp
and leaves. The medicinal applications include treat-
ments for intestinal and skin problems and various uses
as anti-inflammatory, anti-pyretic and analgesic agents.
Recent research in animals has confirmed the presence
of such activities in specific extracts (Ramadan et al.,
1994; Palombo, 2006; Ajose, 2007; Karumi et al., 2008).
In addition antibacterial, antiviral and anti-trypanosome
activities have been reported (Anani et al., 2000; Hudson
et al., 2000; Atawodi et al., 2003).
Inflammation is a common underlying cause of many
diseases, infectious and otherwise and can occur in many
*Corresponding author. E-mail: email@example.com.
Tel.: 1-604-948-2131. Fax: 1-604-875-4351.
organs and tissues, although a controlled acute inflame-
matory reaction is a normal part of our innate immune
response to infection and injury. In order to address the
prospect of medicinal plant applications to the treatment
of inflammatory conditions, we have devised cell culture
systems in which specific viruses and bacteria can induce
substantial amounts of pro-inflammatory or anti-
inflammatory cytokines. Plant extracts can be evaluated
for inflammatory properties in such a system and direct
antiviral effects can also be tested against the same
viruses (Sharma et al., 2008, 2009).
In this study we evaluated the presence and relative
potencies of antiviral and inflammatory activities (cytokine
modulating activities) in standardized commercial prepa-
rations of leaves, fruits and seeds. Extracts were
prepared in methanol, DMSO and water.
MATERIALS AND METHODS
Three standardized preparations were obtained from Baobabtek
(Laval, Quebec). They were:
1. Dried fruit pulp (code #PBA).
2. Dried seed endocarp (code #PEA).
3. Micronized dried leaves (code #PFA).
Portions of each dried material were extracted in methanol, DMSO
and water, at starting concentrations of 100 mg/mL. The extractions
were performed by intermittent shaking and vortexing over a period
of 3 days, at 20°C, in the dark. All extracts were clarified by low
speed centrifugation and filter-sterilized through 0.2 micron filters.
Aliquots were taken for dry weight measurements and the final
extracts were stored in the dark at 4°C. These values were used in
calculating anti-viral MIC (minimal inhibitory concentrations), as
A standard reference Echinacea extract (Echinaforce, acquired
from A. Vogel-Bioforce, Switzerland) was used in the antiviral tests
(see below). The composition of this extract was reported
previously (Sharma et al., 2008, 2009).
Cells and viruses
All cell lines, Vero monkey kidney cells; MDCK canine kidney cells;
Hep-2 human epithelial cells; H-1 sub clone of HeLa cells; A549
human lung epithelial cells; BEAS-2B human bronchial cells; were
obtained from ATCC (American Type culture collection, Rockville,
MD). They were propagated in Dulbecco MEM (DMEM), without
antibiotic or antimycotic agents, in cell culture flasks, supplemented
with 5 - 10% fetal bovine serum, at 37°C in a 5% CO2 atmosphere,
with the exception of the H-1 cells, which were grown at 35oC
(Sharma et al., 2008).
The following viruses were used: influenza, strain H3N2, human
isolate (from BC Centre for Disease Control), propagated in MDCK
cells; HSV (herpes simplex virus type 1, BC-CDC), propagated in
Vero cells; rhinovirus type 1A (RV 1A, from ATCC), propagated in
H-1 cells; respiratory syncytial virus (RSV, from BC-CDC) in Hep-2
cells. All the stock viruses were prepared as clarified cell-free
supernatants, with titers ranging from 106 to 108 pfu (plaque-forming
units) per mL.
The diluted extract (1:100), in 200 µL aliquots, was serially diluted
across replicate rows of a 96-well tray, in DMEM. Virus, 100 pfu in
100 µL, was added to each well and allowed to interact with the
extract for 60 min at a temperature of 22°
C. Following the
incubation period, the mixtures were transferred to another tray of
cells from which the medium had been aspirated. These trays were
incubated at 37°C until viral cpe (cytopathic effects) were complete
in control wells containing untreated virus (usually 2 days for
influenza, 4 - 5 days for the other viruses). Additional wells
contained cells not exposed to virus. The MIC100 was derived from
the maximum dilution at which cpe was completely inhibited by the
extract. In the alternative method (intracellular method), the cells
were incubated with the diluted extracts first for 60 min, before
In some experiments antiviral activity was measured in the
presence and absence of light, since we have shown previously
that many plant extracts contain photoactive compounds (Hudson
and Towers, 1999; Vimalanathan et al., 2005). In this case half the
trays were exposed to a combination of fluorescent and UVA lamps
during the virus-extract reactions and the other half were wrapped
in aluminum foil. A standardized antiviral extract of Echinacea
(Echinaforce®) was also tested in parallel as a reference in some
experiments (Sharma et al., 2008).
Vimalanathan and Hudson 577
Inflammatory cytokine culture system
Details of the test system were described previously (Sharma et al.,
2006, 2008). A549 and BEAS-2B cells were grown in DMEM, in 6-
well trays, to produce confluent monolayers. The medium was
replaced with serum free DMEM for the experiments. Rhinovirus
was added to the cells, for the anti-inflammatory tests, at 1.0
infectious virus per cell (1 pfu/cell), for 1 h at 35°C, followed by a
1:100 dilution of the test extract. Cell free culture supernatants were
harvested after 48 h and assayed for IL-6 and IL-8. Controls
included cells with no virus and cells (± virus) with equivalent
amounts of solvent only.
All cultures were in duplicates and each supernatant was
assayed in duplicates. All data presented are from individual
experiments, but all experiments were repeated at least once, with
ELISA assays were carried out with commercial kits, according to
the instructions supplied by the companies (either R & D Systems
Inc. Minneapolis, MN, USA, for IL-8, or e-Bioscience, San Diego,
CA, USA, for IL-6).
Results were expressed as means ± SEM. For statistical analysis in
the anti-inflammatory activity, one-way analysis of variance
(ANOVA) was used followed by the Dunnet’s t-test. A p value less
than 0.01 (p < 0.01) was considered statistically significant.
A total of 9 extracts were obtained, representing three
different parts of the plant, seed, pulp and leaf, in three
different solvents, water, methanol and DMSO. These
were all tested in two inflammatory cell culture models
and in antiviral tests against several different human
pathogenic viruses. All extracts were tested under
comparable conditions, at similar concentrations (in dry
mass/vol, indicated in Materials and Methods).
In the absence of RV stimulation the basal level of
cytokine secretion in epithelial cells is relatively low, in
which case extracts with pro-inflammatory activity would
enhance cytokine secretion, whereas in RV stimulated
cells anti-inflammatory activities are seen as inhibition in
cytokine secretion (Sharma et al., 2008).
Figure 1a and 1b show results for RV stimulated BEAS-
2B cells and A-549 cells respectively. The virus only
control, without extract, is shown on the far left. In both
cell lines most of the extracts showed little effect on
cytokine secretion and were therefore not anti-
inflammatory (results are shown only for IL-8; similar data
were obtained for IL-6). However the DMSO/pulp and
water/leaf extracts significantly decreased the IL-8
578 J. Med. Plant. Res.
Beas 2B Virus-stimulated
Figure 1. Anti-inflammatory activities in Adansonia extracts.
BEAS 2B epithelial cells (a) and A-549 lung epithelial cells (b) were stimulated by rhinovirus to secrete cytokine IL-8 (CXCL8).
The plant extracts, at 1:100 dilution, were added to the infected cells and incubated for 48 h, at which time cell free
supernatants were removed for assay of IL-8 and IL-6 (only IL-8 data are shown) by ELISA. Readings were converted to
pg/mL by comparison with a standard curve. P-DMSO and L-water extracts showed significant inhibition and were therefore
anti-inflammatory. Results were expressed as mean ± SEM. Statistical significance was considered at P < 0.01 (**)
L = leaf. P = fruit pulp, S = seed.
Beas 2B Non-stimulated
Figure 2. Pro-inflammatory activities in Adansonia extracts.
Unstimulated BEAS 2B epithelial cells (a) and A-549 lung epithelial cells (b) were incubated with the plant extracts, at 1:100
dilution, for 48 h, at which time cell free supernatants were removed for assay of IL-8 and IL-6 (only IL-8 data are shown) by
ELISA. Control cultures represent unstimulated cells incubated with medium (or solvent) only. Readings were converted to
pg/mL by comparison with a standard curve. All three leaf extracts and the pulp water extract, were pro-inflammatory. Results
were expressed as mean ± SEM. Statistical significance was considered at P < 0.01 (**).
L = leaf, P = fruit pulp, S = seed.
Vimalanathan and Hudson 579
Figure 3. Anti-Influenza virus MIC100 (µg/mL) for Adansonia extracts.
All extracts were evaluated for antiviral activity (cpe inhibition assay)
against Influenza virus. The end points were read in duplicate assays
from the two-fold dilution series (duplicates gave identical end points).
Data were converted to MIC100 based on extract concentrations and the
reciprocals were plotted. The higher the 1/MIC value the greater the
antiviral activity; leaf extracts, especially DMSO-leaf, showed greater
activity than the others (the open top end of the bar for DMSO leaf
extract indicates that its activity was >> 3).
secretion and were therefore anti-inflammatory. The leaf
extract was the more active based on concentration (70
µg/mL final concentration, compared with 247 µg/mL for
the pulp extract). As seen in Figure 1, the results were
the same in both cell lines.
Figure 2a and 2b show the corresponding results for
the unstimulated levels of IL-8 secretion (similar results
were obtained for IL-6). All three leaf extracts showed
substantial increases in IL-8 secretion, that is, they were
pro-inflammatory. The water/pulp extract was also
strongly pro-inflammatory, although its concentration was
several-fold greater than the leaf extracts. The other
extracts showed little effect.
Minimum inhibitory values (MIC100) were calculated for
each extract-virus combination. Influenza virus was parti-
cularly vulnerable to some of the extracts, as indicated in
Figure 3, in which antiviral activities are expressed as the
reciprocals of MIC100 in order to emphasize the differ-
rences between degrees of antiviral activity. Thus the
DMSO/leaf extract was very potent and the other leaf
extracts were also significantly anti-influenza virus, with
MIC100 values ranging from approximately 2 µg/mL to <
1.0 µg/mL. In addition some of the other extracts were
antiviral to some degree, but much less than the leaf
extracts (see also Table 1). These MIC values were
comparable to those of the reference Echinacea
preparation (Materials and Methods), a known potent
antiviral extract (Sharma et al., 2009).
Figure 4 shows results of similar tests against two other
membrane-containing viruses, HSV (herpes simplex
virus) and RSV (respiratory syncytial virus), in compa-
rison with influenza. Only the results for the leaf extracts
were shown since they were much more potent than all
the others. However none of the extracts showed activity
against rhinovirus (RV 1A), a non-membrane virus.
Similar antiviral tests were carried out in the absence of
light, to determine the possible presence of antiviral
photosensitizers in the extracts (Hudson and Towers,
1999; Vimalanathan et al., 2005). However the antiviral
MIC100 values were generally not much different, within a
factor of 4 between light and dark. These data are
summarized in Table 1.
MIC values were also obtained for the leaf and pulp
extracts in the alternative test protocol, in which extracts
were incubated first with the cells, followed by virus
infection, in parallel with the standard protocol in which
580 J. Med. Plant. Res.
DMSO-leaf MeOH-leaf Water-leaf
Figure 4. Relative antiviral activities of leaf extracts.
Reciprocal MIC100 values, obtained as described in Figure 3 legend, were plotted for the three leaf
extracts against HSV, RSV and influenza virus. Error bars are shown when the duplicates differed in
their end points. Influenza virus was the most susceptible and RSV the least.
Table 1. Antiviral MIC100.
Source Extract Influenza virus
Leaf DMSO 0.12 L & D 1.0 L
32.5 D 130
Methanol 0.72 L & D 2.9 L
46.8 D 187
Water 2.8 L
43.8 D 350
Pulp DMSO 9.9 L
77 D ND 617
Methanol 31.2 L & D 400 L & D ND > 800
Water 950 L & D 633 L & D ND ~ 1,900
Seed DMSO 4.6 L
> 290 D ND 290
Methanol 150 L & D > 300 L & D ND > 300
Water 220 L & D > 550 L & D ND 550
a = Minimum concentration (µg/mL) showing microscopically visible cytotoxic effects on the cells. ND =
L = Antiviral activity in light; D = Antiviral activity in dark.
Vimalanathan and Hudson 581
Table 2. Comparison of virucidal (standard protocol) and intracellular antiviral
Sample Ratio: MIC virucidal/MIC intracellular
Leaf methanol > 640a
Leaf DMSO 1.0
Leaf water 1.0
Pulp methanol > 160a
Pulp DMSO 4
Pulp water 4
a For the methanol extracts no antiviral activity at all was detected in the intracellular
The extracts were evaluated for anti-HSV activity, by cpe-end point inhibition, using the
standard protocol (pre-incubation of virus with extracts before adding to cells), and in
parallel the intracellular protocol (pre-incubation of extract with cells before adding virus).
MIC100 values were calculated and expressed as a ratio.
virus and extract were pre-incubated before adding to the
cells (as in all the results shown above). Ratios of these
MIC values (standard protocol/alternative protocol) are
shown in Table 2. Both the methanol extracts, leaf and
pulp, gave rise to high ratios, > 640 and > 160 respec-
tively, whereas the other four extracts gave ratios of 1 or
4, indicating that the latter antiviral activities were not
dependent on protocol.
Many of the extracts were cytotoxic at high concentra-
tions, well above their antiviral MIC values, in at
least one of the test cell lines. These values are shown in
the right hand column of Table 1 and represent the most
extreme cases of toxicity observed.
It is evident from the results described that the pattern of
bioactivities are different for the three plant sources, leaf,
fruit-pulp and seed, with the latter containing less signifi-
cant antiviral or inflammatory properties (summarized in
These had the most potent antiviral activities, based on
MIC100 values, with influenza virus being more suscep-
tible than HSV and RSV and the DMSO extract showing
more activity than the methanol and water extracts. The
MIC values observed were similar to those shown by the
reference Echinacea extract, which we have previously
determined to be an excellent antiviral extract (Sharma et
al., 2009). The relative light/dark activities for the leaf
extracts were also similar, suggesting that these extracts
might contain the same or similar antiviral compound/s,
but in different concentrations. However, the observation
that the methanol extract was exclusively active in the
standard protocol only, that is, with pre-incubation of
extract with virus and was devoid of antiviral activity when
added to the cells first, in contrast to the water and
DMSO extracts (Table 2), indicates that in fact different
compounds are involved.
The cytotoxic concentrations were substantially higher
than the antiviral levels, but approximately proportional to
antiviral MIC’s. Consequently we believe that the
Adansonia leaves contain potentially useful and safe
antiviral activity, although for practical applications the
slightly less potent water extract would be preferable.
Further evidence for multiple bioactive components in
leaves came from the different effects on cytokine
secretion. Methanol and DMSO extracts showed pro-
inflammatory properties, as indicated by their stimulatory
effects on IL-6 and IL-8 secretion in BEAS-2B bronchial
epithelial cells and A549 lung epithelial cells, whereas the
water extract did not show this effect. In contrast the
latter was anti-inflammatory, as shown by its inhibition of
rhinovirus-stimulated cytokines in the same cells.
Leaf extracts, usually aqueous, have been used for a
variety of traditional medicinal purposes, including fever,
respiratory and intestinal symptoms and a variety of skin
afflictions, some of which probably involved infectious
diseases and/or inflammation (Wickens, 1982; Ajose et
al., 2007; Karumi et al., 2008) consequently the presence
of antiviral and anti-inflammatory components could
explain beneficial uses of water extracts.
These extracts were also antiviral in a similar manner to
the leaf extracts, although at much lower potencies. Thus
influenza virus was the most susceptible virus and the
DMSO extract the most potent of the three. The methanol
extract, like its leaf counterpart showed activity exclu-
sively in the standard (pre-incubation) protocol, whereas
the DMSO and water extracts showed activity in both
protocols, as was the case for the corresponding leaf
extracts. Cytotoxic concentrations were correspondingly
much lower than in the leaf extracts. These observations
taken together suggest that the antiviral components of
582 J. Med. Plant. Res.
Table 3. Summary of distribution of activities.
Fraction Inflammatory activity(pro/anti) Antiviral activity Flu (F), HSV(H), RSV(R)
Leaf DMSO Pro- F, H, R
Leaf water Anti- F, H, R
Leaf methanol Pro- F, H, R
Pulp DMSO Anti- F, H
Pulp water - +/-
Pulp methanol +/- pro- F, H
Seed DMSO - F,H
Seed water Pro- F
Seed methanol +/- pro- F
the fruit pulp could be similar to the leaf compounds, but
at lower concentrations.
The effects on cytokines were different from corres-
ponding leaf extracts however, the DMSO pulp extract
being anti-inflammatory, while the other two were rela-
tively inactive. Fruit pulp has been traditionally a popular
material for consumption in various ways, raw or boiled in
water, including an anti-diarrhea remedy and various
uses to stimulate or counteract immune responses
(Ramadan et al., 1994; Ajose et al., 2007). Some of these
applications could involve the antiviral and cytokine-
modulatory activities described here.
These were, like the pulp extracts, relatively less
bioactive than the leaf extracts, the DMSO extract again
being the most active antiviral and influenza virus the
most susceptible virus. None of the three extracts
showed anti-inflammatory activity, although water and
methanol extracts were slightly pro-inflammatory. The
relative dearth of bioactivities in the seed extracts, in
comparison with leaf and fruit pulp extracts, could
account for their fewer medicinal applications.
Table 3 summarizes the various antiviral and cytokine
modulating properties of the different extracts. The
results overall indicate the presence of multiple bioactive
compounds in different parts of the plant and these
activities could explain some of the medical benefits attri-
buted to traditional leaf and pulp preparations, in
particular in their treatment of infectious diseases and
inflammatory conditions. However, it is also evident that
we cannot explain the combination of different properties
in terms of just one or two compounds, although the three
solvents used in the extractions tend to yield predomi-
nantly hydrophilic compounds. Chemical analyses have
reported the presence of various potentially bioactive
ingredients (Chadare et al., 2009), including triterpenoids,
flavonoids and phenolic compounds, but at present we
cannot ascribe any of the activities described in this study
to specific compounds.
Ajose FOA (2007). Some Nigerian plants of dermatologic importance.
Int. J. Derm. 46 (suppl 1): 48-55.
Anani K, Hudson JB, de Souza C, Akpagana K, Towers GHN, Arnason
JT, Gbeassor M (2000). Investigation of Medicinal Plants of Togo for
Antiviral and Antimicrobial Activities. Pharm. Biol. 38: 40-45.
Atawodi SE, Bulus T, Ibrahim S, Ameh DA, Nok AJ, Mamman N,
Galadima M (2003). In vitro trypanocidal effect of methanolic extract of
some Nigerian savannah plants Afr. J. Biotech. 2: 317-321.
Chadare FJ, Linnemann AR, Hounhouigan JD, Nout, MJR, Van Boekel
MAJ (2009). Baobab Food Products: A Review on their Composition
and Nutritional Value. Crit. Rev. Food Sci. Nutr. 49: 254-274.
Hudson JB, Anani K, Lee MK, De Souza C, Arnason JT, Gbeassor M
(2000). Further Investigations on the Antiviral Properties of Medicinal
Plants of Togo. Pharm. Biol. 38: 46-50.
Hudson JB, Towers GHN (1999). Phytomedicines as antivirals. Drugs of
the Future 24: 295-320.
Karumi Y, Augustine AI, Umar IA (2008). Gastroprotective effects of
aqueous extract of Adansonia digitata leaf on ethanol-induced
ulceration in rats. J. Biol. Sci. 8: 225-228.
Palombo EA (2006). Phytochemicals from traditional medicinal plants
used in the treatment of diarrhoea: Modes of action and effects on
intestinal function. Phytother. Res. 20: 717-724.
Ramadan A, Harraz FM, El-Mougy SA (1994). Anti-inflammatory,
analgesic and antipyretic effects of the fruit pulp of Adansonia digitata.
Fitoterapia 65: 418-422.
Sharma M, Anderson SA, Schoop R, Hudson JB (2009). Induction of
pro-inflammatory cytokines by respiratory viruses and reversal by
standardized Echinacea, a potent antiviral herbal extract. Antiv. Res.
Sharma M, Arnason JT, Burt A, Hudson JB (2006). Echinacea extracts
modulate the pattern of chemokine and cytokine secretion in
rhinovirus-infected and uninfected epithelial cells. Phytother. Res. 20:
Sharma M, Schoop R, Hudson JB (2008). Echinacea as an
antiinflammatory agent: The influence of physiologically relevant
parameters. Phytother. Res. 23: 863-867.
Sidibe M, Williams JT (2002). Baobab. Adansonia digitata. International
Centre for Underutilized Crops. University of Southampton p. 105
Vimalanathan S, Kang L, Treyvaud Amiguet V, Livesey J, Arnason JT,
Hudson J (2005). Echinacea purpurea aerial parts contain multiple
antiviral compounds. Pharm. Biol. 43: 740-745.
Wickens GE (1982). The baobab – Africa’s upside-down tree. Kew
Bulletin. 37: 173-209.