Solid State Mechanochemical Activation of Silybum
marianum Dry Extract With Betacyclodextrins:
Characterization and Bioavailability of the
Coground Systems

D. VOINOVICH,

1

B. PERISSUTTI,

1

M. GRASSI,

2

N. PASSERINI,

3

A. BIGOTTO

4

1

Department of Pharmaceutical Sciences, University of Trieste, P.le Europa 1, I-34127 Trieste, Italy

2

Department of Chemical Engineering, University of Trieste, P.le Europa 1, 34127 Trieste, Italy

3

Department of Pharmaceutical Sciences, University of Bologna, Via S. Donato 19/2, 40127 Bologna, Italy

4

Department of Chemical Sciences, University of Trieste, Via L. Giorgieri 1, Trieste I-34127, Italy

Received 30 July 2008; revised 2 December 2008; accepted 5 January 2009

Published online 18 February 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21704

ABSTRACT: Silybum marianum dry extract, whose therapeutic use is partially
restricted by the insolubility in water of its main flavonolignans, was subjected to
a mechanochemical activation process in planetary mill using betacyclodextrins as
carriers. After optimization of the operating conditions according to an established
theoretical model, the best active-to-carrier proportion was selected from the preli-
minary trials. When using the optimized conditions, the mechanochemical process
permits an improvement of the physico-chemical properties of the active, which reaches
an ‘‘activated’’ solid state, that is stable for at least 1 year. In fact, XRD, DRIFT
and Raman spectroscopy analyses showed that the main extract component, Silybin,
completely lost its crystalline structure after co-grinding with betacyclodextrins and
formed weak interactions with the carrier. The powder characteristics remarkably
changed after co-grinding, leading to a sample with a very small mean diameter and
with a twofold increase of the specific surface area in comparison to the dry extract. The
activated solid state of the coground systems remarkably enhanced the

in vitro drug

dissolution kinetics with consequent improved oral bioavailability. Furthermore, the
in vivo studies on rats revealed a 6.6-fold bioavailability increase respect to the
S. marianum Italian commercial product used as reference (Silirex

1

200 capsules).

ß

2009 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 98:4119–4129, 2009

Keywords:

Silybum marianum extract; betacyclodextrins; planetary mill; solid

state; mechanical activation; process optimization; oral absorption; bioavailability;
physical characterization

INTRODUCTION

Milk thistle is a popular botanical product
promoted for its hepatoprotective properties,

1

and ranks among the top-selling botanical supple-
ments in the United States.

2

The purported

‘‘active’’ phyto-chemicals present in milk thistle
are a series of flavonolignans, known collectively
as ‘‘silymarin,’’ which include silybin, isosilybin
(each as a pair of diastereoisomers A and B),
silydianin, silychristin, and taxifolin.

3

Bioavailability and dissolution characteristics

for silymarin-containing products have been

Correspondence

to:

B.

Perissutti

(Telephone:

þ39-

0405583106; Fax.

þ39-04052572; E-mail: [e-mail])

Journal of Pharmaceutical Sciences, Vol. 98, 4119–4129 (2009)
ß

2009 Wiley-Liss, Inc. and the American Pharmacists Association

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

4119

shown to vary widely. An evaluation of nine
separate milk thistle products found that the
amount of silybinin released over 1 h into an
aqueous buffered solution (pH 7.5, 378C) ranged
from 0% to 85%,

4

while a comparative bio-

availability study of three silybin-containing
dosage forms found that values for AUC and

C

max

varied among products by factors of 3 and 6,
respectively.

5

Moreover, several studies have

demonstrated that silymarin-containing products
exhibit especially poor bioavailability and drug-
release properties when not formulated with
solubility-enhancing agents.

4

Several methods

to improve the oral bioavailability of silymarin
have been attempted: its complexation with
phosphatidylcholine,

6,7

the

formulations

as

lipid-containing microspheres,

8

the complexation

in b-cyclodextrins,

9

the incorporation in water-

soluble matrices produced by spray drying
or lyophilizing process,

10

solid dispersions with

hydrophilic polymers such as PEG 6000 or
PVP,

11–14

the synthesis of a novel silybin prodrug

with a linear PEG and succinic ester linkage,

15

and, finally, the formulation of self-microemulsi-
fying systems.

16

Recently, simultaneous improve-

ment of solubilization kinetics of the main
flavonolignans of

Silybum marianum extract

was obtained by co-grinding the dry extract with
two crosslinked polymers, micronized crospovi-
done and sodium carboxymethylcellulose, in a
high-energy planetary mill.

17

It appeared that the

mechanical stress, given by this mechanochemical
process, induced structural changes of the main
crystalline components, changing the size and
surface area of the powders and, consequently,
remarkably enhancing the solubilization perfor-
mances and its

in vivo oral bioavailability. These

encouraging results have led to the present study,
where betacyclodextrins (BCD) were employed as
carriers using the mechanochemical process, with
the aim of increasing the

in vivo oral bioavail-

ability of

S. marianum dry extract. BCD were

selected as a carrier, due to their well known
ability to interact with poorly water-soluble drugs
and drug candidates (including silymarin

6,7

)

resulting in an increased apparent water solubi-
lity.

18

Furthermore, several studies have already

demonstrated that BCD can be successfully
employed as carriers in the mechanochemical
process.

19–23

The activation of the drug in these

composites can be obtained through mechano-
chemical activation by several mechanisms,
including the formation of an inclusion compound,
or by decreasing the particle size with grinding, or

by transformation of crystalline drugs into amor-
phous state.

19

In particular, as regards to this

latter activation mechanism, in a study upon the
suitability for the mechanical treatment of several
polysaccaridic carriers, BCD resulted the most
efficient amorphisizing agent.

19

To go into more details, the first part of this

work encompassed the optimization of the milling
process in planetary mill, in order to select the
best operating conditions for ‘‘activating’’ the dry
extract. In the second part of the research, the
chosen process variables were applied to the
preparation of a binary system with BCD. After
characterization of the solid state of the coground
product and evaluation of its

in vitro dissolution

performance, its

in vivo oral absorption on

rats was assessed in comparison to an Italian
commercial

S. marianum formulation (Silirex

1

200 capsules).

MATERIALS AND METHODS

Materials

S. marianum dry extract was kindly donated by
Indena S.p.A. (Milano, Italy). Betacyclodextrins
(BCD) with a 12% hydration water was given
by Wacker Chemie Gmbh (Munich, Germany).
Naringenine pure standard was from Extra-
synthese

(Genay,

France),

b

-glucuronidase/

arylsulfatase (Helix pomatia) was from Boehrin-
ger Mannhein Gmbh (Mannhein, Germany).
Silirex

1

200 capsules were from Lampugnani

Farmaceutici S.p.A., Nerviano, Italy. Silirex

1

200 capsules contains the following ingredients:
at

least

200

mg

of

flavoinoids,

expressed

as Silymarin, magnesium stearate, titanium
dioxide (E171), gelatin, sodium indigotindisulfo-
nate (E132).

All other chemicals, of analytical grade, and

solvents, HPLC grade, were provided by Carlo
Erba (Milan, Italy).

Process Optimization

The planetary mill used was a Fritsch P5
(Pulverisette, Contardi Fritsch s.r.l., Milan,
Italy). This apparatus is equipped with four agate
cylindrical vials (internal height

H

v

¼ 5.6 cm,

internal radius

R

v

¼ 3.7 cm, volume ¼ 250 cm

3

)

adopting agate balls (diameter

d

b

¼ 2 cm) as

grinding media. According to the Burgio et al.
model,

24

the specific energy,

E

(J/kg), a planetary

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

DOI 10.1002/jps

4120

VOINOVICH ET AL.

mill supplies via the grinding media to its load
(pure substance or mixture of more substances in
powder form) in the temporal range D

t, is given by:

E

¼ f ðd

b

; r

b

;

R

v

; v

v

;

R

p

; v

p

Þf

b

N

b

D

t

(1)

where

f is a function depending on balls (grinding

media) diameter and density (

d

b

and r

b

, respec-

tively), vials radius and angular velocities (

R

v

and

v

v

, respectively), mill plate radius and angular

velocity (

R

p

and v

p

, respectively) while

N

b

is ball

numbers and w

b

is expressed by:

f

b

¼ ð1 n

"
v

Þ;

n

v

¼

N

b

d

3
b

p

R

2

v

H

v

;

"

¼

log

10

ð5Þ 2

log

ðð2R

v

d

b

Þd

b

=3

R

2

v

Þ

(2)

where

H

v

is vials height. An inspection of Eqs. (1)

and (2) reveals that once vials (

R

v

,

H

v

, v

v

), mill

plate (

R

p

, v

p

), grinding media (

d

b

, r

b

) charac-

teristics and D

t are fixed, E

depends only on

N

b

.

Accordingly, the

N

b

value maximizing

E

is given

by:

@

E

@

N

b

¼ f Dt N

b

N

1

þ"

b

d

3
b

p

R

2

v

H

v

¼ 0 ) N

b

¼

1

ðd

3
b

=p

R

2

v

H

v

Þð1 þ "Þ

1="

(3)

On the basis of P5 geometrical characteristics,
Eq. (3) yields

N

b

18. Vial load was chosen so that

it completely occupied the ball bed void fraction
assuming the highest balls packing (around 0.26).
Accordingly, we had that mill load is 11.3 grams
per vial. In order to maximize grinding effect,
maximum mill plate angular velocity was chosen.

Preparation of Coground Mixtures

By employing the process conditions selected
during the above reported process optimization,
dry extract and BCD were first blended in the
suitable proportions with a stainless steel spatula,
then simultaneously coground. In particular,
the grinding procedure was conducted at the
maximum velocity with 18 grinding media, with a
vial load of 11.3 g and for a time of 1 h, stopping
every 15 min to homogeneously mix the mass with
a stainless steel spatula. For comparison pur-
poses, dry extract and BCD were separately
milled in identical conditions.

As for the composition of the coground mixtures,

a set of preliminary trials was carried out on
several dry extract-to-carrier weight ratios (from
2:1 to 1:2.5 wt). The choice of the final composition

was made with the aim of reducing adhesion
phenomena to the mixing bowl walls and to
the grinding media, and hence to obtain the
higher production yield. The selected dry extract-
to-carrier weight ratio was 1:2.5, whose recovery
was almost complete (98 wt%).

Preparation of Physical Mixtures

For comparison purposes, a physical mixture (PM)
was prepared by manually mixing dry extract and
BCD using the same weight ratio as the selected
coground system (1:2.5 wt).

X-Ray Powder Diffraction Studies (XRD)

Solid state of the samples was studied by means
of XRD technique using a D500 diffractometer
(Siemens, Munich, Germany) with Cu Ka radia-
tion (1.5418 A

˚ ), monochromatized by a secondary

flat graphite crystal. The current used was 20 mA
and the voltage 40 kV. The scanning angle ranged
from 58 to 308 of 2u, steps were of 0.058 of 2u, and
the counting time was of 2 s/step.

DRIFT Spectroscopy

Fourier transform-infrared spectra were obtained
on a FT-IR spectrometer (FT-IR Perkin Elmer
Spectrum One, Monza, Italy) using the diffuse
reflectance

method

(DRIFT).

The

samples

were added to anhydrous KBr (FT-IR grade) in
a 1:15 weight ratio (sample to KBr) and gently
ground, thus avoiding solid transition possibly
induced by extended grinding. The scanning
range was 450–4000 cm

1

and the resolution

was 4 cm

1

, scan number was 8 and scan speed

0.20 cm/s.

Raman Spectroscopy

Raman spectra were obtained using a Perkin-
Elmer System 2000 FT-Raman instrument with
excitation provided by 1064 nm radiation from a
diode-pumped Nd:YAG laser. The samples were
contained in capillary cells. The treatment of the
spectral data was performed using Perkin Elmer
Spectrum and Galactic GRAMS386 software.

DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

SOLID STATE MECHANOCHEMICAL ACTIVATION OF

S. MARIANUM

4121

Porosity Measurements

The porosity of the samples has been determined
with a mercury porosimeter (ThermoQuest)
equipped by Macropore Pascal 140 low pressure
porosimeter and Macropore Pascal 240 high
pressure porosimeter (CE Instruments, Italy). A
dilatometer for powders with capillary diameter of
1.5 mm was loaded with 300 mg samples. Before
measuring, a degasification procedure under
vacuum pressure for 30 min was performed.
The experiments were performed in triplicate.
The volume of mercury intruded in function of the
applied pressure was transformed in the porous
distribution applying the model of cylindrical
porosity of Washburn. The size distribution was
calculated

applying

the

Mayer

and

Stowe

method,

25

whereas the specific surface area was

calculated by using the method proposed by
Rootare and Prenzlow.

26

Dissolution Kinetic Test

The dissolution kinetics of the coground system
were determined in nonsink conditions according
to dispersed amount method.

27

The same condi-

tions were used for PM, dry extract and commer-
cial capsules formulation (as a powder). In
particular, an excess quantity of the samples
was added to 900 mL of deionized water in a USP
XXXI dissolution vessel, at a temperature of
37.0

0.58C while stirring with a paddle at

200 rpm. The tested amount for each sample
(20 mg of active, expressed as silybin) approximately
corresponded to 3 times its solubility in water.
Samples (5 mL) were collected at predetermined
time intervals, and filtered with Sartorius regen-
erated cellulose membrane 0.45 mm. The dissolved
dry extract was spectrophotometrically determined
and absorbances were recorded at 287 nm. Results
were the average of three measurements and
standard deviations did not exceed 5%.

In Vivo Absorption Studies

Experiments on animals complied with the
Italian D.L. n. 116 of January 27, 1992 and
associated guidelines in the European Com-
munities Council Directive of November 24,
1986 (86/609 ECC).

Animals used were male Wistar rats (250 g

weight) and were supplied by Centro Servizi
Polivalenti di Ateneo (University of Trieste). Rats,

with free access to water, were fasted overnight.
Experimental formulations were administered by
gastric gavage as aqueous suspensions to four
rats. Each rat received a single dose of formula-
tions: 200 mg/kg of silymarin as a c-grounded
system or as a commercial reference product
(Silirex

1

200 capsules).

Blood samples were collected from animal

abdominal aorta in heparinized tubes at 1, 2,
4 h after administration. Blood samples were
centrifuged at 1500 rpm for 10 min and plasma
was separated and immediately frozen at

208C,

and stored at this temperature till the analysis.

HPLC Sample Preparation

For the determination of silybin in the coground
system, in the dry extract and in the commercial
reference product (Silirex

1

200 capsules) the

method described in our previous study was
employed.

17

As for the determination of total silybin in

plasma samples the following procedure was
adopted. One milliliter serum sample containing
the internal standard naringenin (50 mL of an
internal standard solution containing 2 mg/mL)
was incubated with 1 mL of 1 M sodium acetate
buffer pH 5.0 and 100 mL of b-glucuronidase/
arylsulfatase for 3 h at 378C, while shaking at
55 rpm. After adding 2 mL of 0.5 M borate buffer
solution pH 8.5, silybin and the internal standard
naringenin were extracted into 5.5 mL diethyl
ether by shaking at 55 rpm for 20 min. After
centrifugation for 10 min at 2000

g (128C) the

organic phase was transferred into 5 mL micro-
reaction vessels and evaporated at 458C under a
stream of nitrogen. The residue was redissolved
in 200 mL of methanol, vortexed for 1 min,
centrifuged and 50 mL of the supernatant were
used for HPLC analysis.

Pharmacokinetic Analysis

Pharmacokinetic parameters were calculated on
the composite plasma curves. The area under the
plasma concentration-time curve extrapolated to
the last sampling time at which a quantifiable
concentration is found (AUC

last

) was calculated

using the log-linear trapezoidal method. Time and
value of maximum concentration (

t

max

and

C

max

,

respectively) were reported as observed. The
relative bioavailability after oral administration

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

DOI 10.1002/jps

4122

VOINOVICH ET AL.

(

F

rel

) was calculated in Eq. (4):

F

rel

¼

AUC

ðformulationÞ

AUC

ðcommercialÞ

(4)

Pharmacokinetic analysis were performed using
WinNonlin Version 2.1 (Pharsight Corporation,
Mountain View, CA) software. As this was a pilot
study, the results were not subjected to any formal
statistical test.

RESULTS AND DISCUSSION

The process variables previously selected by
the above mentioned optimization procedure,
were applied to the preparation of coground
systems. The first issue was the selection of the
dry extract-to-BCD wt ratio giving less adhesion
to the grinding media and the walls of the mixing
bowl. In fact, as already noticed in our previous
study,

17

during the mechanochemical process the

formation of an adhesive layer quite hard and
difficult to remove can occur, and this inhibits the
process itself. Consequently, the best composition
resulted to be 1:2.5 dry extract-to-BCD wt ratio,
whose recovery was almost complete.

Furthermore, only for comparison purposes, the

two components were singularly processed in
identical conditions.

The first solid state investigation was made

carrying out XRD analysis and comparing the
coground system with the starting materials and
the simple physical mixture (Fig. 1). The dry
extract behaved at the XRD analysis as a semi-
crystalline material, exhibiting some peaks of
medium intensity together with a strong under-
neath scattering phenomenon, due to its amor-
phous content (Fig. 1e). As for the crystalline
portion, the visible reflections were due to the
extract major components, silybin (that showed
several very intense peaks at 14.68, 16.58, 19.58,
22.38, and 24.58 of 2u, see Fig. 1c, according to
literature data

17,28

), whereas the other compo-

nents were not detectable probably because of
their amorphous nature or their little content in
the dry extract. When the dry extract was ground
alone, it completely lost its crystalline portion and
showed a halo pattern (data not shown).

In the diffractograms of native BCD, several

high intensity peaks were typically present at
11.08, 12.68, and 13.28 of 2u, and in the range
between 188 and 208 2u (Fig. 1f). As for the
BCD singularly processed (data not shown), the
grinding procedure caused the transition in a

Figure 1.

XRD patterns of the 1:2.5 (w/w) coground systems freshly prepared (b) and

1 year after the preparation (a) compared to Silybin (c), 1:2.5 (w/w) dry extract:BCD
physical mixture (d),

Silybum marianum dry extract (e), and BCD (f).

DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

SOLID STATE MECHANOCHEMICAL ACTIVATION OF

S. MARIANUM

4123

complete amorphous state, with a remarkable
scattering phenomenon and two broad bands in
correspondence to main BCD reflections.

The patterns recorded for the PM (Fig. 1d) could

be simply regarded as the superposition of those of
the active and the carrier, with no evidence of
interactions

between

the

components.

Con-

versely, from the coground composite (Fig. 1b) a
complete diffuse pattern was obtained. Only two
broad humps are visible, referable to the sum of
original main reflections of BCD and silybin.
Hence, XRD results are suggestive of a total active
amorphization probably as a consequence of the
formation of an amorphous mixture of the two
components and/or homogeneous dispersion in
the carrier and/or inclusion of the active in the
carrier cavity.

The subsequent characterization consisted in

the DRIFT spectroscopy: in particular DRIFT
spectra were recorded for raw materials, as
received and separately ground, and 1:2.5 (w/w)
dry extract-to-carrier physical mixture (PM) and
coground systems (Fig. 2).

In accordance to previous XRD findings, ground

BCD (Fig. 2f) showed modest variations of the
features, consisting in a little broadening of the
bands and some variations in the relative band
intensity in comparison to starting BCD (Fig. 2e).
Conversely, the dry extract changed remarkably
its features after milling. In general, the bands in
the ground sample were smoothened and less
intense, as typical of an amorphous sample.
Further, the original intense bands at 3608 and
3458 cm

1

(Fig. 2a) completely disappeared after

grinding (Fig. 2b). The signals at 3608 and
3458 cm

1

are attributable to –OH stretching

modes affected by intra and intermolecular
hydrogen bonds between molecules of silybin.

28

It is likely that the intermolecular hydrogen bonds
were originally involved in the stabilization of
silybin crystalline lattice. The grinding strongly
changed the status of the –OH groups involved in
these bonds: this is reflected in the modification of
the profile of the absorption between 3000 and
3700 cm

1

.

For shortness, only the 4000–2400 cm

1

spec-

tral range is reported in Figure 2. In this range,
the physical mixture was simply the sum of the
signals of both components, with no evidence of
interactions (Fig. 2c). In contrary, the coground
system (Fig. 2d) showed absence of the peaks of
the silybin crystalline lattice (visible in the PM at
3606 and 3459 cm

1

). Further, in the whole

spectrum (not reported) the peaks in the coground

Figure 2. DRIFT spectra of pure

Silybum marianum

dry extract (a), ground

Silybum marianum dry extract

(b), 1:2.5 (w/w) PM (c), 1:2.5 (w/w) coground (d), pure
BCD (e), and ground BCD (f). For the sake of brevity
only the 4000–2400 cm

1

range is reported.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

DOI 10.1002/jps

4124

VOINOVICH ET AL.

system are smoothened and less intense than
those of the simple PM, showing a typical trend of
the amorphous state.

In addition, in the carbonyl stretching region,

with respect to native dry extract, little shifts of
the carbonyl stretching band could be observed in
the binary systems (both PM and coground) and in
the ground silybin. To clarify this point and get
some additional information about the solid state
of the samples and possible interactions between
components, Raman spectroscopy was then con-
ducted on the same samples.

The Raman spectra are reported in Figure 3.

For the samples containing silybin our attention
was focused on three parts of the spectrum:
the aryl C–H stretching band (originally at
3061 cm

1

, indicated by an oval), the doublet in

the 1640–1615 cm

1

range (probably due to a

combination of the carbonyl band and to the

aromatic ring stretching band, underlined by a
rectangle) and the peaks at values less than
1200 cm

1

. The aryl C–H stretching band shifted

very remarkably from 3061 to 3075 cm

1

(ground

silybin), to 3070 cm

1

(in the PM) and to

3074 cm

1

(in the coground system), suggesting

a more interacting aryl C–H. As for the doublet,
modest variations of the bands are detected,
probably because both carbonyl group and ben-
zene ring are slightly affected by the aryl C–H
interaction. In particular, the position of the
carbonyl peak is slightly shifted from 1630 (in
pure and ground extract, and PM) to 1635 cm

1

in

the coground sample; whereas the benzene ring
peak, originally at 1618 cm

1

slightly shifted to

1614 cm

1

in the ground silybin, to 1615 cm

1

in

the PM and to 1621 cm

1

in the coground system.

Finally, in the samples subjected to mechanical
activation the intensity of the peaks at values less
than 1200 cm

1

was dramatically reduced, the

ground noise was very strong and the baseline was
progressively

increasing,

demonstrating

the

amorphous nature of the ground samples. Con-
versely, these phenomena were absent in the PM.

As for BCD, our attention was paid to the intense

band at 482 cm

1

(indicated by the rhombus), that

is a typical mode of the ‘‘skeleton’’ of these
molecules. This band appeared unchanged in the
ground BCD (481 cm

1

) and in the PM (481 cm

1

),

and substantially in an identical position in the
coground system (at 484 cm

1

). This fact shows the

lack of significant perturbation of the BCD skeleton
when mixed or coground with the active, thus
proving the lack of inclusion phenomena of
the active. Further, the increasing baseline, the
broadening and the reduction of the intensity of the
peaks are compatible with the transition in an
amorphous state of the sample.

It can be concluded that Raman and DRIFT

spectroscopy attested the absence of strong
interactions between BCD and active, confirming
the destructuration of silybin crystalline lattice
after grinding and cogrinding. It also points to the
presence of some weak interactions between
silybin and BCD, probably Van der Waals or
hydrophobic interactions, involving silybin aro-
matic ring in the binary systems, and especially in
the coground system.

Then, Hg porosimetry was employed to deter-

mine possible changes in powders’ particle size
and surface area after cogrinding. In the curves
‘‘a’’ of Figures 4 and 5, the oversize cumulative
volume distributions of the native dry-extract
and coground system are depicted, respectively,

Figure 3. RAMAN spectra of 1:2.5 PM (a), 1:2.5 (w/w)
coground (b), pure BCD (c), ground BCD (d), ground
Silybum marianum dry extract (e) and pure Silybum
marianum dry extract (f).

DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

SOLID STATE MECHANOCHEMICAL ACTIVATION OF

S. MARIANUM

4125

whereas the differential plots of the particle size of
the same samples are reported in curves ‘‘b’’ of
Figures 4 and 5, respectively. From these figures
it appeared that the cogrinding process caused a
remarkable diminution of particle diameter with
respect to the starting extract. Further, a
pronounced increase of the specific surface area
after mechanochemical activation was noticed
(Tab. 1). The surface area of the coground product
was about twice than that of the dry extract and
hence the system was theoretically more prone to
dissolution.

This assumption was confirmed by the

in vitro

solubilization

kinetic

tests

of

this

sample

in comparison to the original dry extract, the
physical mixture and a commercial

S. marianum

formulation (Silirex

1

200), that is simply com-

posed of dry extract and common excipients. In
fact, as illustrated in Figure 6, the coground
system exhibited a significant improvement of the
release behavior, both in terms of rate and extent
of dissolution, with respect to all the other
samples.

Silirex

1

, dry extract and PM solubilization

kinetics did not differ in a significant way, for
example,

there

is

no

significant

difference

between every two samples (

p > 0.05). Since also

the presence of BCD (in the PM) failed to increase
the release performance in the environment, this
seems a ‘‘drug controlled’’ solubilization kinetics,
with a typical trend of a crystalline powder.

Furthermore, these results underlined that the

mechanochemical process positively influenced
the

S. marianum solubilization performance.

Several explanations can be hypothesized for
such behavior. First, the amorphous state is
primarily responsible for the increased solubility
and hence dissolution rate of these systems.

29

Second, the powders’ superficial area increased
by twofold positively influences the dissolution
rate, as is well understood from the Noyes–
Whitney equation. Third, the dissolution rate

Figure 4.

Cumulative size distribution (a), and dif-

ferential plot (b) of the particles of the dry extract.

Figure 5.

Cumulative size distribution (a), and dif-

ferential plot (b) of the particles of the 1:2.5 (w/w) dry
extract:BCD coground.

Table 1. Specific Surface Area Calculated by the Hg
Porosimetry Analysis

Sample

Specific Surface

Area (m

2

/g)

Dry extract

1.17

1:2.5 (w/w) dry extract:BCD coground

2.37

Figure 6. Solubilization kinetics of the 1:2.5 (w/w)
dry extract:BCD coground system freshly prepared
(~) and 1 year after the preparation (- -) compared to
pure

Silybum marianum dry extract (&), Silirex

1

200 capsules (

), and 1:2.5 (w/w) dry extract:BCD phy-

sical mixture (*).

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

DOI 10.1002/jps

4126

VOINOVICH ET AL.

enhancement may be attributed to the more
intimate dispersion of the active in the carrier
as a results of the mechanical treatment.

20,22

Further, the increased wettability of the drug
obtained thanks to the presence of the BCD is
expected to be more intensive in the ground
mixture than in the simple mixture.

29

Finally, the

dissolved drug and BCD can form a dissolved
inclusion compound which enhances the absorp-
tion process due to the increased drug solubility
and dissolution rate.

Comparing the dissolution kinetics of this

coground system with the findings of our previous
experience of mechanical activation, in which
S. marianum dry extract was coground with two
insoluble crosslinked polymers, crospovidone and
sodium croscarmellose (1:3 active-to-carrier wt
ratio),

17

and taking in consideration that the

dissolution mechanism was obviously different,
the following conclusions can be made. In all three
cases, the better performance of the coground
systems compared to the dry extract and the PM
was evident, both in terms of rate and extent of
dissolution, testifying that the mechanochemical
process led to an activated status easier to be
dissolved.

Among

them,

the

best

profile

was achieved when the system was coground
with sodium croscarmellose, being able to main-
tain the highest solubilized amount of active
(81.30 mg/mL) for the entire duration of the
analysis. In contrary, in the case of the crospo-
vidone the lowest dissolution enhancement was
obtained together with a different solubilization
kinetic characterized by a supersaturation phe-
nomenon and a solubilized amount of actives
of about 63 mg/mL. This different behavior was
attributed to the presence of remarkable chemical
interactions

between

crospovidone

and

dry

extract components, and to the presence of large
agglomerates due to undesired crystal growth
during the mechanochemical process.

Finally, when the dry extract is coground with

the water soluble BCD in the 1:2.5 drug-to-carrier
wt ratio (hence with a lesser amount of carrier), a
solubilized amount of active of 72 mg/mL was
obtained, for the entire duration of the analysis
with lack of supersaturation phenomena. This

value was probably reached and maintained
thanks to the sum of the mechanisms described
above.

The next step was the evaluation of the

in vivo

oral absorption on rats of the coground system
in the relation to the commercial reference
product, Silirex

1

200 capsules, as a powder.

The oral bioavailability of the coground system
was 6.6 times higher than that of the commercial
product, reaching a maximal concentration of
33.2 mg/L compared to 4.1 mg/L obtained with
Silirex

1

200 (Tab. 2, Fig. 7), thus confirming that

the ‘‘activated’’ dry extract solid state corre-
sponded to an effective higher oral bioavailability.

With respect to the

in vivo performances of the

coground systems with sodium croscarmellose
and crospovidone,

17

once again an oral bioavail-

ability intermediate between the two crosslinked
polymers can be recorded for the system based on
BCD (the AUC is in fact 145.4 mgh/L for sodium
croscarmellose,

107.8

mgh/L

for

BCD

and

78.4 mgh/L for crospovidone).

Finally, for the period of 1 year the stability of

the ‘‘activated’’ solid state of the dry extract in the
composite was checked once a month by XRD
analysis. No evidence of recrystallization pheno-
mena of the dry extract was seen (Fig. 1a).
Furthermore, the dissolution kinetics of the

Table 2. Pharmacokinetic Data

Sample

C

max

(mg/L)

t

max

(h)

AUC (mgh/L)

Rel. Bioavail.

Silirex

1

200

4.1

1.72

16.2

1

1:2.5 (w/w) dry extract:BCD coground

33.2

2.34

107.8

6.6

Figure 7.

Mean

plasma

silybin

levels

obtained

after single oral administration of a 1:2.5 (w/w) dry
extract:BCD coground (black circles), and a commercial
tablet formulation, Silirex

1

200 (black squares).

DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

SOLID STATE MECHANOCHEMICAL ACTIVATION OF

S. MARIANUM

4127

coground sample after 1 year of ageing did not
significantly differ from that of the fresh sample
(Fig. 6). This would probably mean that the BCD
acted as a stabilizer towards the finely dispersed
state of the drug, as previously reported,

19

and

towards the amorphous state of the active
components, thanks to the hydrophobic and/or
Van der Waals interactions between the BCD and
the silybin aromatic ring.

CONCLUSIONS

S. marianum dry extract was activated through
a solid state mechanochemical process with
betacyclodextrins in a planetary mill, using
operating variables selected after a proper opti-
mization. The cogrinding with the carrier led to
the amorphization of the main extract component,
silybin, with changes in size and surface area of
the powders and with improved solubilization
kinetics in water. These results were shown not
only in comparison to the dry extract and physical
mixture but also to a marketed product (Silirex

1

200 capsules). This activated status of the dry
extract remained stable over a period of at least
1 year, probably thanks to the interactions
established at the solid state with the carrier.
Finally, the

in vivo absorption studies in rats

revealed a very remarkable improvement of the
oral bioavailability of the dry extract with respect
to the reference product.

ACKNOWLEDGMENTS

The authors wish to express their appreciation to
Lorenzo Magarotto for helpful discussions, to
Prof. Francesco Princivalle for kind hospitality
in Dept. Earth Sciences of University of Trieste,
and to Indena S.p.A. for the gift of all the phyto-
chemicals used in this study and the Regione
Friuli Venezia Giulia for the financial support
of the project ‘‘Tecnologie nella trasformazione
di piante officinali per lo sviluppo di prodotti
nel settore alimentare e zootecnico’’ (grant No
11-2003).

REFERENCES

1. Saller R, Meier R, Brignoli R. 2001. The use of

silymarin in the treatment of liver diseases. Drugs
61:2035–2063.

2. Committee on the Use of Complementary and Alter-

native Medicine by the American Public. Comple-
mentary and Alternative Medicine in the United
States. Washington, DC: The National Academies
Press; 2005.

3. Morazzoni P, Bombardelli E. 1995. Silybum maria-

num (Carduus marianum). Fitoterapia 66:3–42.

4. Schulz H-U, Schurer M, Krumbiegel G, Wachter W,

Weyhenmeyer R, Seidel G. 1995. Investigation of
dissolution and bioequivalence of silymarin pro-
ducts. Arzneim-Forsch 45:61–64.

5. Kim YC, Kim EJ, Lee ED, Kim JH, Jang SW, Kim

YG, Kwon JW, Kim WB, Lee MG. 2003. Compara-
tive bioavailability of silybinin in healthy male
volunteers. Int J Clin Pharmacol Ther 41:593–596.

6. Morazzoni P, Montalbetti A, Malandrino S, Pifferi

G. 1993. Comparative pharmacokinetics of silypide
and silymarin in rats. Eur J Drug Metab Pharma-
cokinet 18:289–297.

7. Mareshwari H, Agarwal R, Patil C, Katare OP.

2003. Preparation and pharmacological evaluation
of silibin liposomes. Drug Res 53:420–427.

8. Abrol S, Trehan A, Katare OP. 2004. Formulation,

characterization, and in vitro evaluation of sily-
marin loaded lipid microspheres. Drug Deliv 11:
185–191.

9. Arcari M, Brambilla A, Brandt A, Caponi R, Corsi

G, Di Rella M, Solinas F. 1992. Nuovo complesso di
inclusione tra silibina e la b-ciclodestrina: Velocita’
di dissoluzione in vitro e assorbimento in vivo in
confronto a formulazioni tradizionali. Boll Chim
Farm 5:205–209.

10. Benesova K, Cvak L, Stuchlik M. 2002. Increased

solubility flavolignan preparations. PCT/US 02/
27713.

11. Feng-Quian L, Jin-Hong H. 2004. Improvement of

the dissolution rate of Silymarin by means of solid
dispersions. Chem Pharm Bull 52:972–973.

12. Qiu MF, Wei J, Li S-S, Xu Z-H, Xia S, Wang X-R,

Zhang Y-Y, Xie G-X. 2005. A new silymarin pre-
paration based on solid dispersion technique. Adv
Ther 22:595–600.

13. Sun N, Wei X, Wu B, Chen J, Lu Y, Wu W. 2008.

Enhanced dissolution of Sylimarin/polyvinilpirroli-
done solid dispersion pellets prepared by a one step
fluid bed coating technique. Powder Technol
182:72–88.

14. Sun N, Zhang X, Lu X, Wu W. 2008. In vitro evalua-

tion pellets and pharmacokinetics in dogs of solid
dispersion pellets containing Sylibum marianum
extract prepared by fluid bed coating. Planta Med
74:126–132.

15. Zhang P, Ye H, Min T, Zhang C. 2008. Water soluble

Polyethylene glycol prodrug of silybin: Design
synthesis and characterization. J Appl Polym Sci
107:3230–3235.

16. Wu W, Wang Y, Que L. 2006. Enhanced bioavail-

ability of silymarin by self- microemulsifying drug

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

DOI 10.1002/jps

4128

VOINOVICH ET AL.

delivery system. Eur J Pharm Biopharm 63:288–
294.

17. Voinovich D, Perissutti B, Magarotto L, Ceschia D,

Guiotto P, Bilia AR. 2008. Solid state mechano-
chemical simultaneous activation of the constitu-
ents of the Silybum marianum phytocomplex with
crosslinked polymers. J Pharm Sci (in press).

18. Brewster ME, Loftsson T. 2007. Cyclodextrins as

pharmaceutical solubilizers. Adv Drug Del Rev 59:
645–666.

19. Shakhtshneider TP, Boldyrev VV. 1999. Mechan-

ochemical synthesis and mechanical activation of
drugs. In: Boldyreva E, Boldyrev V, editors. Reac-
tivity of molecular solids. Wiley. pp. 271–311.

20. Tozuka Y, Wongmekiat A, Sakata K, Moribe K,

Oguchi T, Yamamoto K. 2004. Co-grinding with
cyclodextrin as nanoparticle formation method of
a poorly water soluble drug. J Inclus Phenom Macr
Chem 50:67–71.

21. Magarotto L, Bertini S, Cosentino C, Torri G. 2001.

Characterization

of

Nimesulide/b-cyclodextrin

composite obtained by solid state activation. Mater
Sci Forum 360–362:643–648.

22. Mura P, Faucci MT, Parrini PL. 2001. Effect of

grinding with microcrystalline cellulose and cyclo-
dextrins on the ketoprofen physicochemical proper-
ties. Drug Dev Ind Pharm 27:119–128.

23. Corvi Mora P, Cirri M, Allolio B, Carli F, Mura P.

2003. Enhancement of dehydroepiandrosterone
solubility and bioavailability by ternary complexa-
tion with a-cyclodextrin and glycine. J Pharm Sci
92:2177–2184.

24. Burgio N, Iosanna A, Magini M, Martelli S, Padella

F. 1981. Mechanical alloying of the Fe-Zr System.
Correlation between Input energy and end pro-
ducts. Il Nuovo Cimento 13D(4): 459–476.

25. Mayer RP, Stowe RA. 1965. Mercury porosimetry-

breakthrough pressure for penetration between
packed spheres. J Colloid Interf Sci 20:893–911.

26. Rootare HM, Prenzlow CF. 1967. Surface area from

mercury porosimeter measurements. J Phys Chem
71:2733–2736.

27. Nogami H, Nagai T, Youtsuyanagi T. 1969. Disso-

lution phenomena of organic medicinals involving
simultaneous phase changes. Chem Pharm Bull 17:
499–509.

28. Yao W-W, Bai T-C, Sun J-P, Zhu CW, Zhang H-L.

2005. Thermodynamic properties for the system of
Silibin and poly(ethylene glycol) 6000. Thermochim
Acta 437:17–20.

29. Froemming K-H, Szejtl JI. 1994. Cyclodextrins in

pharmacy. Dordrecht (NL): Kluwer Academic Pub-
lishers. pp. 145–146.

DOI 10.1002/jps

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 11, NOVEMBER 2009

SOLID STATE MECHANOCHEMICAL ACTIVATION OF

S. MARIANUM

4129