European Journal of Pharmaceutical Sciences 41 (2010) 452–457

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences

journal homepage: www.elsevier.com/locate/ejps

What is modulating solubility in simulated intestinal fluids?

Giorgio Ottaviani a,∗, Daniel J. Gosling a, Celine Patissier a, Stephane Rodde a, Liping Zhou b, Bernard Faller a

a
Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, Basel, Switzerland
b
Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, Cambridge, MA, USA

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this study was to understand which parameters are responsible for the selective modulation
Received 16 April 2010 of compounds solubility in simulated intestinal fluids. The solubility of 25 chemically diverse reference
Received in revised form 1 June 2010 compounds was measured in simulated intestinal fluid (FaSSIF-V2) and in aqueous phosphate and maleate
Accepted 18 July 2010
buffers. Electrostatic interactions between compounds and the bio-relevant medium components seem to
Available online 23 July 2010
explain the different solubility behavior observed for acids and bases. The solubility of ionized acids is not
increased in FaSSIF-V2 probably due to electrostatic repulsions with the media components. Lipophilicity
Keywords:
plays an important role but mainly for charged bases with a log P > 4 (or log D6.5 > 1.9). When the aqueous
Simulated intestinal fluids
Solubility
solubility is mainly driven by lipophilicity, the FaSSIF-V2 components seem to improve the solubility of
Ionization basic compounds to a greater extent than for compounds whose solubility is limited by crystal packing.
Lipophilicity These results suggest that ionization, lipophilicity and crystal packing play important but peculiar roles
Crystal packing in controlling solubility in FaSSIF-V2 compared to that in aqueous buffer and this information could be
useful to guide medicinal chemists and formulation scientists.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction and Ca is the concentration of the drug in the basolateral GI side

Solubility is a crucial parameter for successful drug develop-
ment as poor solubility could compromise the pharmacokinetic
and pharmacodynamic properties of xenobiotics (Avdeef, 2003;
Alsenz and Kansy, 2007; Faller and Ertl, 2007; Kerns et al., 2008).
To achieve desirable systemic exposure levels after oral dosing,
drugs have to cross the GI tract, and the fraction absorbed can be
described by the Fick’s first law of permeation:
F = Pe (Cd − Ca ) (1)
where F is the flux, Pe is the permeability coefficient, Cd is con-
centration of the drug in the GI fluid (i.e. the maximum solubility)
Abbreviations: pKa , ionization constant; FaSSIF, Fasted State Simulated Intestinal
Fluid; FeSSIF, Fed State Simulated Intestinal Fluid; BCS, Biopharmaceutics Clas-
sification System; log P, octanol/water partition coefficient; log D, octanol/water
distribution coefficient; log S0 , logarithm of intrinsic solubility; GI, Gastro-Intestinal
tract; UPLC, Ultra Performance Liquid Chromatography; HPLC–UV, High Pressure
Liquid Chromatography coupled with UV detector; DAD, Diode Array Detector;
MDA, Methanol/Diethylen-glycol-monoethylether/Acetonitrile (1:1:1, v/v/v); SL,
Intrinsic Solubility Lipophilicity difference; SGF, Simulated Gastric Fluid; HT-log P,
High Throughput octanol/water partition coefficient.
∗ Corresponding author at: Novartis Pharma AG, 4056 Basel, Switzerland.
Tel.: +41 61 3240893.
E-mail address: giorgio.ottaviani@novartis.com (G. Ottaviani).
(Avdeef, 2003; Kwon, 2001).
Aqueous solubility is the reference medium to predict the
bioavailability and absorption of orally administered drugs in
accordance with the BCS (Amidon et al., 1995). In the last years,
simulated biological fluids have been extensively used in early
development to better predict the in vivo dissolution profile for
oral dosage forms, especially for BCS class II compounds (Galia et
al., 1998).
FaSSIF and FeSSIF are widely used media that simulate physio-
logical conditions in the proximal small intestine in the fasted state
(FaSSIF) and in the fed state (FeSSIF) (Galia et al., 1998; Kostewicz
et al., 2002). These media simulate the main components of human
GI tract such as bile salts and lecithin. It was suggested that drug
solubility in these media is normally increased compared to the
solubility determined in aqueous buffers, as a result of enhanced
wetting and/or micellar solubilization (Andrieux et al., 2004). Lavé’s
group (Jones et al., 2006) showed that when the solubility data mea-
sured in FaSSIF and FeSSIF were incorporated into physiologically
based absorption models, plasma concentration-time profiles were
more accurately predicted. Recently professor Dressman and col-
laborators (Jantratid et al., 2008) published revised compositions
for these simulated fluids (-V2, version 2) based on newer GI stud-
ies in healthy human volunteers (Porter et al., 2007). Table 1 shows
the composition of the new FaSSIF-V2 media compared to the old
one (Jantratid et al., 2008; Vertzoni et al., 2004). Although meth-

0928-0987/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejps.2010.07.012
 G. Ottaviani et al. / European Journal of Pharmaceutical Sciences 41 (2010) 452–457 453

ods for the solubility determination in bio-relevant media were

reported (Bard et al., 2008), the solubility of medchem compounds
in these media has never been fully studied from a physicochemical
perspective.
This work was done to investigate and better understand which
parameters modulate solubility in FaSSIF-V2 medium compared
to aqueous media. Twenty-five chemically diverse reference com-
pounds were selected with a balanced dataset of acids and bases
with generally low intrinsic solubility values (to better assess the
FaSSIF solubility enhancement compared to the aqueous media).
Solubility was determined by shake-flask HPLC–UV and the values
in FaSSIF-V2 were compared with the values measured in aqueous
media (phosphate and maleate buffers) at the same pH. Maleate Fig. 1. Comparison of solubility values (in log units, M) determined in FaSSIF-V2
and in phosphate buffer at pH 6.8 using a high-throughput equilibrium solubil-
buffer was chosen because it is the buffer component of the new
ity assay for 4148 in-house compounds. The plotted line refers to the identity line
FaSSIF-V2. In order to separate electrostatic and lipophilicity con- (log Sphosphate = log SFaSSIF-V2 ).
tributions, intrinsic solubility was measured for a selected number
of compounds by potentiometric titration (Avdeef et al., 2000).
glass vial and centrifuged again for 15 min. If floating particles were
still present in solution, samples were also filtrated using mini prep
2. Materials and methods
vials (Syringeless Filter Device Mini uni prepTM , Whatman). 75 ␮L
of supernatant was then collected and the same volume of MDA
2.1. Chemicals
was added before injecting the solution for quantification into the
UPLC-DAD (Waters, AcquityTM Ultra Performance LC) using a C18
All compounds and chemicals for buffers were purchased from
column (Waters, Acquity UPLC BEH C18 1.7 ␮m, 2.1 mm × 50 mm).
Sigma (division of Fluka Chemie AG, Buchs, Switzerland). Phos-
Twenty times diluted solutions were also prepared and injected
phate buffer at pH 6.8 (0.0355 mol/L KH2 PO4 and 0.0312 mol/L
into UPLC apparatus by adding 7.5 ␮L of supernatant into 67.5 ␮L
Na2 HPO4 ) was also purchased from Fluka and the pH was adjusted
of MDA. The solubility was then determined using calibration stan-
to pH 6.5 with HCl 1N. Maleate buffer (FaSSIF-V2 blank) was pre-
dard solutions in MDA/water (50:50, v/v).
pared using the ingredients of FaSSIF-V2 without lecithin and
sodium taurocholate (Table 1).
2.4. Intrinsic solubility measurements
2.2. Preparation of FaSSIF-V2
Intrinsic solubility (i.e. solubility of the neutral species) of
amodiaquin, chlorprothixene, meclizine, miconazole, glipizide,
To obtain 1 l of clear FaSSIF-V2 solution, the components were
diclofenac, flufenamic acid and carprofen (Table 3) was determined
added in a specific order with stirrings in between the additions
by potentiometric titration (Avdeef et al., 2000). 0.7–1.2 mg of com-
(Table 1). The sodium taurocholate and sodium chloride were first
pounds (final concentration ∼1 mM) were dissolved in aqueous and
dissolved in 400 mL of purified water followed by addition of 1 mL
FaSSIF-V2 media. The maleate buffer was removed from the latter
of 1 N HCl. After stirring for 30 min, the lecithin was added and then
in order to ensure a proper compounds titration. The intrinsic sol-
the mixture was sonicated until complete dissolution. The solution
ubility was then determined from a mathematical analysis of the
was stirred for 2 h before the addition of maleic acid and 500 mL
titration curve using measured pKa values.
water. After an overnight stirring, the pH was adjusted to 6.5 with
1N NaOH and the total volume was adjusted to 1 l. The final solution
was clear and was stored at 4 ◦ C when not used. 3. Results and discussion

3.1. Data analysis of compounds solubility in FaSSIF-V2 media

2.3. Saturation shake-flask

Fig. 1 shows the solubility determined in FaSSIF-V2 and in phos-

Samples as powder were added to aqueous buffers and FaSSIF-
phate buffer at pH 6.8 for 4148 in-house compounds. The solubility
V2 medium to have a final concentration of 2 mg/ml (final volume
was determined from evaporated stock DMSO solutions of test
∼1 ml). The suspension was sonicated for 5 min and then shaken at
compounds using a modified version of a high-throughput equilib-
room temperature for 24 h on a rotary shaker. The phases were first
rium solubility assay with miniaturized shake-flask approach and
separated by decantation using an Eppendorf 5804 lab-bench cen-
HPLC–UV analysis (Zhou et al., 2007). The solubility of most com-
trifuge for 15 min at 3600 rpm. Then they were transferred into a
pounds was higher in FaSSIF-V2 than in phosphate buffer (82%)
but interestingly 16% of compounds were more soluble in phos-
Table 1 phate buffer than in FaSSIF-V2 (they are below the identity line in
Composition of the standard FaSSIF and revised FaSSIF-V2 (Vertzoni et al., 2004;
Fig. 1). This result showed that an in depth evaluation of solubil-
Jantratid et al., 2008).
ity in FaSSIF-V2 was required to understand why only a fraction of
FaSSIF FaSSIF-V2 compounds was more soluble in this medium compared to aque-
Sodium taurocholate (mM) 3 3 ous media. This information could be useful to guide medicinal
Lecithin (mM) 0.75 0.2 chemists as well as formulation scientists in discovery projects.
NaH2 PO4 (mM) 28.66 – Accordingly a diverse dataset of reference compounds was
Maleic acid (mM) – 19.12
NaCl (mM) 106 68.62
selected and the solubility was measured in FaSSIF-V2 and aqueous
NaOH (mM) ∼13.8 34.8 buffers using the shake-flask method from powder samples. The
pH 6.5 6.5 more time consuming shake-flask assay was preferred to the high-
Osmolality (mOsmol kg−1 ) 270 180 throughput assay. It was in fact observed that for some in-house
Buffer capacity (mmol l−1 pH−1 ) 12 10
compounds the solubility determined with the high throughput
454 G. Ottaviani et al. / European Journal of Pharmaceutical Sciences 41 (2010) 452–457
Fig. 2. Solubility measured with the shake-flask method in buffers maleate and
phosphate at pH 6.5 for 25 reference compounds. Solubility values are expressed in
log units (M). The solid line refers to the identity line for solubility in both media
(log Sphosphate = log Smaleate ).
assay using DMSO evaporated solutions (Zhou et al., 2007) was
overestimated compared to the traditional shake-flask method
starting from dry powder. One possible explanation could be that
after DMSO evaporation, some compounds did not re-crystallize
during the incubation time. The lower throughput traditional
shake-flask assay starting from powder allowed to better preserve
the original solid state properties of compounds thus minimizing
the risk of supersaturation (Alsenz and Kansy, 2007).
3.2. Evaluation of phosphate and maleate buffer effects
In FaSSIF-V2 the maleate buffer replaces the phosphate buffer
present in the previous composition (see Table 1). For compounds
which are fully ionized at pH 6.5, the buffer components could
play a crucial role in solubility as the latter could be driven by
the solubility product of the corresponding salts in solution or by
the formation of soluble/insoluble aggregates. As most tested com-
pounds were ionized at pH 6.5, the potential buffer effects were
studied by comparing the solubility measured in maleate buffer
(FaSSIF blank) and in phosphate buffer at pH 6.5 using the shake-
flask, with the solubility value extrapolated at the same pH derived
from the intrinsic solubility measurements. As intrinsic solubility
is independent of buffer composition, it is possible to factor out and
evaluate buffer effect contributions (Bergstrom et al., 2004). If the
pKa and the intrinsic solubility (S0 ) are known, the solubility (S)
at a different pH can be calculated using a modified version of the
Henderson–Hasselbalch equation as shown below for monoprotic
bases (Eq. (2)) and acids (Eq. (3)) (Avdeef, 2003):
S = S0 · (10+pKa −pH + 1) (2)
−pKa +pH
S = S0 · (10 + 1) (3)
If the solubility measured at a given pH value is lower than the
value extrapolated from the intrinsic solubility at the same pH, it is
likely that compounds precipitated with salts present in the media
or formed insoluble aggregates, while if the measured solubility
is higher than the value extrapolated, soluble aggregates could be
formed (Bergstrom et al., 2004).
For most compounds the solubility measured in phosphate and
maleate buffers correlated well with the values extrapolated from
the intrinsic solubility (Figs. 2 and 3) showing that no remarkable
buffer effects were present. For some very soluble compounds such
as pindolol and quinacrine, the highest value of solubility mea-
sured in phosphate and maleate buffers at pH 6.5 was limited by
the dynamic range of the assay. Although with tested compounds
buffer effects were not present, it was important to assess it before
proceeding with further analysis as different solubility values in
maleate and in phosphate buffers have been measured for some
in-house compounds (data not showed).
Fig. 3. Intrinsic solubility extrapolated at pH 6.5 using experimental pKa values
plotted against the experimental solubility (log S) measured in buffer maleate (X)
and phosphate (♦) at pH 6.5 for tested compounds. The solid line refers to the identity
line and the diagonal dotted lines represent a threefold deviation from the identity
line.
3.3. Supersaturation in FaSSIF-V2
Supersaturation is a thermodynamic unstable phenomenon
where chemicals stay in solution at a concentration above their
equilibrium solubility. Induction or maintenance of intralumi-
nal supersaturation using specifically designed formulations is an
attractive strategy to improve intestinal absorption (Brouwers et
al., 2009). For weak bases supersaturated solutions starting from
crystalline powder may also occur in the small intestine as a result
of a pH gradient in the gastrointestinal lumen. Kostewicz et al.
(2004) showed that when weak bases were first dissolved in SGF
and then put in FaSSIF media, solubility values in FaSSIF were higher
than those measured directly in FaSSIF starting from powder as a
result of supersaturation driven by a pH gradient between SGF (pH
2) and FaSSIF (pH 6.5) media.
In our shake-flask experiments starting from dry powder, super-
saturation was not strictly controlled (the apparent solubility was
measured in FaSSIF-V2 at one time point after 24 h) but as solubility
was directly measured in FaSSIF-V2 and for a long incubation time
it is likely that thermodynamically stable solutions were obtained
at the end of the incubations.
3.4. What is modulating the solubility in FaSSIF?
Table 2 shows the solubility values measured with the shake-
flask method in phosphate buffer, in FaSSIF-V2 and in maleate
buffer (FaSSIF-V2 blank: no lecithin or taurocholic acid) media. F/B
ratio expresses the solubility enhancement of FaSSIF-V2 compared
to the FaSSIF blank and was calculated as follows:
F SFaSSIF
= (4)
B Smaleate
where SFaSSIF is the solubility measured in FaSSIF-V2 media (mM)
and Smaleate is the solubility measured in FaSSIF-V2 blank (mM).
Fig. 4 shows the F/B ratio plotted against log P for tested com-
pounds. It is interesting to observe that the solubility of acids did
not increase in FaSSIF-V2 compared to aqueous medium (F/B ratio
close to 1). On the contrary, the most lipophilic bases showed
an enhanced solubility in FaSSIF-V2 compared to that in aque-
ous buffer. Although tested acids are on average less lipophilic
than bases (log P < 5.6 for acids, log P < 7.57 for bases), the solubility
of acids does not increase in FaSSIF-V2 regardless their log P val-
ues. Conversely there are three bases with log P < 5.6 (amodiaquin,
chlorprothixene and miconazole) that show a remarkable solubility
increase in FaSSIF-V2 compared to aqueous buffer.
The fraction ionized at experimental pH (pH 6.5) calculated from
the aqueous pKa values using the Henderson–Hasselbalch equa-
tion (see Table 2), shows that most of acids and bases are strongly
 G. Ottaviani et al. / European Journal of Pharmaceutical Sciences 41 (2010) 452–457 455

Table 2
Physicochemical properties and solubility data of 25 tested compoundsa .

Compound pKa fi pH 6.5 log S0 log S6.5 extrap log P log D6.5 extrap SL Solubility Solubility Solubility F/B
FaSSIF-V2 maleate phosphate
(mM) (mM) (mM)

Amiodarone B 8.73 0.99 −8.17 −5.94 7.57 5.40 0.6 0.147 <0.006 <0.006 >23.8
Amodiaquin B 7.37 0.98 −5.88b −3.27 4.20 2.50 1.7 >4.216 1.372 1.608 >3.1
B 8.24
A 11.49
Astemizole B 5.73 0.99 −5.93 −3.88 5.70 3.80 0.2 0.074 0.046 0.068 1.6
B 8.48
Carprofen A 4.247 1 −4.63b −2.46 4.29 2.00 0.3 1.512 1.377 1.527 1.1
Chlorprothixene B 9.52 1 −5.82b −3.38 5.48 2.80 0.3 1.083 0.500 0.237 2.2
Corticosterone Not 0 −3.2c −3.20 1.90c 1.90 1.3 0.569 0.404 0.358 1.4
ionizable
Diclofenac A 4.032 1 −5.02b −2.95 4.51 2.10 0.5 4.471 3.377 1.736 1.3
Flufenamic acid 3.97A 1 −5.08b −2.82 5.56 3.00 −0.5 1.395 1.221 1.249 1.1
Folic acid B 2.33 1 −5.31 −2.37 0.20 −2.40 5.1 2.121 2.161 2.639 1.0
A 3.87
A 4.76
A 7.98
Glipizide A 5.13 0.96 −5.28b −4.10 2.58 1.20 2.7 0.049 0.058 0.036 0.8
Glyburide A 5.3b 0.94 −5.9c −4.66 4.40c 3.20 1.5 <0.008 <0.008 <0.008
Haloperidol B 8.42 0.99 −5.47 −3.54 4.30 2.40 1.2 0.314 0.330 0.325 1.0
Maprotiline B 10.33 1 −4.69 −1.76 4.85 1.90 −0.2 >5.407 >5.407 >5.407
Meclizine B 2.23 0.85 −5.55b −5.68 6.20 5.40 −0.6 0.100 <0.010 <0.010 >9.8
B 7.24
Mefenamic acid A 4.22 1 −6.34 −4.06 5.33 3.10 1.0 0.054 0.041 0.033 1.3
Miconazole B 6.3 0.39 −5.88b −5.41 5.34 5.10 0.5 0.049 <0.010 <0.010 >5.1
Pindolol B 9.54 1 −3.79 −0.85 1.83 −0.90 2.0 4.546 >6.04 >6.04 <0.8
Piroxicam A 1.87 1 −5.75 1.98 3.8 0.920 0.335 0.320 2.7
B 5.29
Progesterone Not 0 −4.4c −4.40 3.90c 3.90 0.5 0.029 0.019 0.019 1.5
ionizable
Promethazine B 8.99 1 −4.19 −1.70 4.56 2.20 −0.4 >5.274 >5.274 >5.274
Quinacrine B 7.9 1 −4.35 −1.39 5.44 2.50 −1.1 >3.75 >3.75 >3.75
B 10.26
Sulfasalazine A 2.35 1 −6.28 −2.33 3.61 −0.20 2.7 1.405 1.114 1.320 1.3
A 7.99
A 10.86
Tamoxifen B 8.48c 1 −7.55c −5.57 6.00e 4.10 1.6 0.118 0.027 0.011 4.4
Terfenadine B 9.25 1 −7.74 −4.99 5.42 2.90 2.3 0.051 0.034 0.040 1.5
Testosterone Not 0 −4.06c −4.06 3.30c 3.30 0.8 0.107 0.087 0.087 1.2
ionizable
a
pKa , log S0 and log P were taken from Box and Comer (2008) unless otherwise stated.
b
Internal data, measured using potentiometric titration.
c
From Faller and Ertl (2007).
e
Internal data, measured using HT-log P (Faller et al., 2005).

ionized at pH 6.5: acids being negatively charged, while bases pos-
itively charged. The structures of lecithin and taurocholic acid used
in FaSSIF-V2 are shown in Fig. 5. As lecithin is zwitterionic and tau-
rocholic acid is fully deprotonated at pH 6.5, the FaSSIF-V2 media
contains a net negative charge, and this could explain the different
behavior observed for acids and bases.
It is likely that acids, being negatively charged, were not able to
establish favorable interactions with FaSSIF-V2 micelles because
of electrostatic repulsion. Bases, on the contrary, being positively
charged, were better solubilized in FaSSIF-V2 due to favorable elec-
trostatic interactions.
To better understand this peculiar phenomenon, the intrin-
sic solubility of four acids and fours bases was measured using a
potentiometric approach in both aqueous and FaSSIF-V2 media.
The intrinsic solubility of an ionizable compound is defined as the
equilibrium solubility of the free acid or base form at a pH where the
compound is fully unionized. If electrostatic interactions between a
compound and the FaSSIF-V2 media are present, one should expect
a pH dependent solubility enhancement.
F0 /B0 ratio expresses the intrinsic solubility enhancement of
FaSSIF-V2 compared to aqueous media calculated as follows:
F0 S0 FaSSIF
=
B0 S0 aqueous
where S0 FaSSIF is the intrinsic solubility measured in FaSSIF-V2
media (␮M) and S0 aqueous is the intrinsic solubility measured in
aqueous media (␮M). F0 /B0 enables to factor out ionic interactions
as when compounds are neutrals mainly hydrophobic interactions
with FaSSIF-V2 components are present. The difference between
F0 /B0 ratio and the F/B ratio clearly shows that acids and bases
behave differently (Table 3). While F0 /B0 increases for most com-
pounds, for acids the solubility enhancement in FaSSIF is greater
when compounds are neutrals (F0 /B0 > F/B), for bases the solubil-
ity enhancement in FaSSIF is greater when compounds are ionized
(F/B > F0 /B0 ). It is thus very likely that when compounds are neu-
trals, hydrophobic interactions with the media components are
mainly responsible for the solubility enhancement in FaSSIF-V2
while when compounds are charged (pH 6.5), electrostatic inter-
456 G. Ottaviani et al. / European Journal of Pharmaceutical Sciences 41 (2010) 452–457

Fig. 4. Correlation between log P and ratio of solubility measured at pH 6.5 in FaSSIF-
V2 (SFaSSIF ) and solubility measured at pH 6.5 in maleate buffer (Sblank ) for: () bases;
() acids; (♦) neutrals; (X) zwitterions.
Fig. 6. Correlation between log P and solubility ratio F/B in log units (F/B: sol-
ubility ratio between solubility in FaSSIF-V2 and solubility in maleate buffer)
for bases with log P > 4. The line was obtained by the linear regression equation
log(F/B) = 0.34 log P − 1.32, where n = 9; r2 = 0.61.

Fig. 7. Correlation between log D6.5 and solubility ratio (in log units) of com-
pounds at pH 6.5 in FaSSIF-V2 (SFaSSIF ) and in maleate buffer (Sblank ) for
Fig. 5. structure of FaSSIF-V2 active ingredients: lecithin (A) and taurocholic acid bases with log P > 4. The line was obtained by the linear regression equation
(B). log(F/B) = 0.30 log D6.5 − 0.58, where n = 9; r2 = 0.71.

actions play an additional role in modulating the solubility in
FaSSIF-V2.
For acids it is very likely that the negative charge at pH 6.5 is
detrimental for their solubility in FaSSIF-V2 media because of elec-
trostatic repulsion with the media components. When acids are
unionized the solubility is instead enhanced in FaSSIF-V2, proba-
bly because the lipophilic interactions are not counter balanced by
the negative charge of compounds.
For bases the positive charge at pH 6.5 seems to further enhance
their solubility in FaSSIF-V2 media as when bases are unionized
F0 /B0 is lower than F/B.
Although it seems convincing that electrostatic interactions are
responsible for the selective modulation of compounds solubility
in FaSSIF-V2, they are not the only driving forces as lipophilicity
also might play an important role. Figs. 6 and 7 shows the F/B ratio
in log units plotted respectively against log P and log D6.5 for bases
with a log P > 4. At high log P values (log P > 4), there is a positive
relation between solubility enhancement in FaSSIF-V2 compared
to aqueous media and lipophilicity. The correlation is only mod-
estly improved if the distribution coefficient is taken into account
(log D6.5 ).
Fig. 8 shows the aqueous intrinsic solubility plotted against the
log P. For compounds close to the unity line, the intrinsic solu-
bility is mainly influenced by lipophilicity while for compounds
above the unity line, the solubility is affected by other parameters
such as crystal lattice energy or formation of insoluble aggregates.
Faller and Ertl (2007) introduced the SL parameter to quantify
the extent of lipophilicity contribution to the intrinsic solubility
(S0 ). SL is defined as the difference between −log S0 and log P.
When solubility is mainly limited by lipophilicity, the SL tends
to be close to 0, while when other limiting factors, such as inter-
molecular interactions play a dominant role, SL is in general
higher (SL > 2). For example terfenadine has a modest solubil-
ity increase in FaSSIF-V2 compared to maleate buffer (F/B = 1.5)
and this is associated with a high SL (SL = 2.3). It is thus likely
that when the solubility of compounds is limited by crystal lattice
energy, the solubilizing power of FaSSIF-V2 components is limited.
Although as seen above, compounds charge plays an important
role in FaSSIF solubility and this is not taken into account by the
SL parameter, it is interesting to note that basic compounds with
the highest F/B ratio (amiodarone F/B = 23.8, meclizine F/B = 9.8 and
miconazole F/B = 5.1) all lie in the unity line (SL < 0.6). For these

Table 3
Intrinsic solubility (S0 ) measured in aqueous and FaSSIF media with potentiometric titration for selected bases and acids. For F0 /B0 and F/B see text.

Compound fi pH 6.5 S0 aqueous (␮M) S0 FaSSIF (␮M) F0 /B0 F/B

Amodiaquin 0.98 (B) 1.3 2.6 1.9 >3.1

Chlorprothixene 1 (B) 1.5 1.5 1 2.2
Miconazole 0.39 (B) 1.3 4.3 3.1 >5.1
Meclizine 1 (B) 2.8 6.9 2.5 >9.8
Glipizide 0.96 (A) 5.2 16.6 3.2 0.8
Diclofenac 0.85 (A) 9.5 14.9 1.6 1.3
Flufenamic acid 1 (A) 8.3 17.9 2.2 1.1
Carprofen 1 (A) 23.1 47.5 2.1 1.1
 G. Ottaviani et al. / European Journal of Pharmaceutical Sciences 41 (2010) 452–457
Fig. 8. log P plotted against intrinsic solubility S0 (in antilog units) for tested ()
bases; () acids; (♦) neutrals. The solid line refers to the identity line (log P = log 1/S0 )
and the dotted line represents a twofold deviation between log P and log 1/S0
(SL = 2), see text.
basic compounds the solubility seems thus not to be limited by
intermolecular interactions but mainly by lipophilicity and this
could further explain the bigger solubility enhancement observed
in FaSSIF-V2.
4. Conclusions
Aqueous solubility is controlled by crystal lattice energy,
solute–solvent interactions (i.e. lipophilicity) and ionization. As in
addition to aqueous solubility, bio-relevant media are commonly
used in pharmaceutical research, it is essential to understand
what drives the solubility of compounds in these new media. This
work was thus devoted to identify the physicochemical parameters
affecting the solubility of compounds in the new FaSSIF-V2.
Results showed that ionization and lipophilicity play an impor-
tant but peculiar role in controlling solubility in FaSSIF-V2
compared to the aqueous buffer solubility. The charge seems
to partially control the affinity between the compounds and
the FaSSIF-V2 components. Probably due to electrostatic repul-
sion with the media components, the solubility of ionized acidic
compounds was not improved in FaSSIF-V2 compared to aque-
ous buffers. On the contrary strong basic compounds showed an
enhanced solubility in FaSSIF-V2 probably due to favorable electro-
static interactions with the negatively charged media. Lipophilicity
also plays a key role for bases, although it seems to have a pos-
itive impact only when the log P > 4 (log D6.5 > 1.9). Finally when
the aqueous solubility is lipophilicity driven, the FaSSIF-V2 com-
ponents seem to improve the solubility to a greater extent than
when solubility is limited by crystal packing.
Acknowledgments
The authors would like to thank Professor Jennifer Dressman for
useful discussions and advices.
457
References
Alsenz, J., Kansy, M., 2007. High throughput solubility measurement in drug discov-
ery and development. Advanced Drug Delivery Reviews 59 (7), 546–567.
Amidon, G.L., Lennernas, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a
biopharmaceutic drug classification: the correlation of in vitro drug product
dissolution and in vivo bioavailability. Pharmaceutical Research 12 (3), 413–420.
Andrieux, K., Forte, L., Lesieur, S., Paternostre, M., Ollivon, M., 2004. Insertion and
partition of sodium taurocholate into egg phosphatidylcholine vesicles. Phar-
maceutical Research 21 (8), 1505–1516.
Avdeef, A., 2003. Absorption and Drug Development. John Wiley & Sons Inc.
Avdeef, A., Berger, V.M., Brownell, C., 2000. pH-metric solubility. 2: Correlation
between the acid–base titration and the saturation shake-flask solubility-pH
methods. Pharmaceutical Research 7 (85–89), 2000.
Bard, B., Martel, S., Carrupt, P.A., 2008. High throughput UV method for the esti-
mation of thermodynamic solubility and the determination of the solubility
in biorelevant media. European Journal of Pharmaceutical Sciences 33 (3),
230–240.
Bergstrom, C.A.S., Luthman, K., Artursson, P., 2004. Accuracy of calculated pH-
dependent aqueous drug solubility. European Journal of Pharmaceutical
Sciences 22 (5), 387–398.
Box, K.J., Comer, J.E.A., 2008. Using measured pKa, Log P and solubility to investigate
supersaturation and predict BCS class. Current Drug Metabolism 9, 869–878.
Brouwers, J., Brewster, M.E., Augustijns, P., 2009. Supersaturating drug delivery
systems: the answer to solubility-limited oral bioavailability? Journal of Phar-
maceutical Sciences 98 (8), 2549–2572.
Faller, B., Ertl, P., 2007. Computational approaches to determine drug solubility.
Advanced Drug Delivery Reviews 59 (7), 533–545.
Faller, B., Grimm, H.P., Loeuillet-Ritzler, F., Arnold, S., Briand, X., 2005. High-
throughput lipophilicity measurement with immobilized artificial membranes.
Journal of Medicinal Chemistry 48, 2571–2576.
Galia, E., Nicolaides, E., Horter, D., Lobenberg, R., Reppas, C., Dressman, J.B., 1998.
Evaluation of various dissolution media for predicting In vivo performance of
class I and II drugs. Pharmaceutical Research 15 (5), 698–705.
Jantratid, E., Janssen, N., Reppas, C., Dressman, J.B., 2008. Dissolution media sim-
ulating conditions in the proximal human gastrointestinal tract: an update.
Pharmaceutical Research 25 (7), 1663–1676.
Jones, H.M., Parrott, N., Ohlenbusch, G., Lave, T., 2006. Predicting pharmacokinetic
food effects using biorelevant solubility media and physiologically based mod-
elling. Clinical Pharmacokinetics 45 (12), 1213–1226.
Kerns, E.H., Di, L., Carter, G.T., 2008. In vitro solubility assays in drug discovery.
Current Drug Metabolism 9 (9), 879–885.
Kostewicz, E.S., Brauns, U., Becker, R., Dressman, J.B., 2002. Forecasting the oral
absorption behavior of poorly soluble weak bases using solubility and disso-
lution studies in biorelevant media. Pharmaceutical Research 19 (3), 345–349.
Kostewicz, E.S., Wunderlich, M., Brauns, U., Becker, R., Bock, T., Dressman, J.B., 2004.
Predicting the precipitation of poorly soluble weak bases upon entry in the small
intestine. Journal of Pharmacy and Pharmacology 56, 43–51.
Kwon, Y., 2001. Handbook of Essential Pharmacokinetics, Pharmacodynamics, and
Drug Metabolism for Industrial Scientists. Springer-Verlag, New York, pp. 35–72.
Porter, C.J.H., Trevaskis, N.L., Charman, W.N., 2007. Lipids and lipid-based formu-
lations: optimizing the oral delivery of lipophilic drugs. Nature Reviews Drug
Discovery 6 (3), 231–248.
Vertzoni, M., Fotaki, N., Kostewicz, E., Stippler, E., Leuner, C., Nicolaides, E., Reppas, C.,
2004. Dissolution media simulating the intralumenal composition of the small
intestine: physiological issues and practical aspects. Journal of Pharmacy and
Pharmacology 56 (4), 453–462.
Zhou, L.P., Yang, L.H., Tilton, S., Wang, J.L., 2007. Development of a high through-
put equilibrium solubility assay using miniaturized shake-flask method in early
drug discovery. Journal of Pharmaceutical Sciences 96 (11), 3052–3071.