A novel hybrid drug between two potent anti-tubulin agents as a potential prolonged anticancer approach

Article abstract of DOI:10.1016/j.ejps.2016.05.032

We report the design, synthesis and biological characterisation of a novel hybrid drug by conjugation of two tubulin inhibitors, a hemiasterlin derivative A (H-Mpa-Tle-AHA-OH), obtained by condensation of three non-natural amino acids, and cis-3,4′,5-trimethoxy-3′aminostilbene (B). As we have previously demonstrated synergy between A and B, we used a monocarbonyl derivative of triethylene glycol as linker (L) to synthesise compounds A-L and A-L-B; via HPLC we analysed the release of its potential hydrolysis products A, A-L, B and B-L in physiological fluids: the hybrid A-L-B undergo hydrolysis in rat whole blood of the ester bond between A and L (half-life = 118.2 ± 9.5 min) but not the carbamate bond between B and L; the hydrolysis product B-L was further hydrolyzed, but with a slower rate (half-life = 288 ± 12 min). The compound A-L was the faster hydrolyzed conjugate (half-life = 25.4 ± 1.1 min). The inhibitory activity of the compounds against SKOV3 ovarian cancer cell growth was analysed. The IC₅₀ values were 7.48 ± 1.27 nM for A, 40.3 ± 6.28 nM for B, 738 ± 38.5 nM for A-L and 37.9 ± 2.11 nM for A-L-B. The anticancer effect of A-L-B was evidenced to be obtained via microtubule dynamics suppression. Finally, we stated the expression of the active efflux transporters P-gp (ABCB1) and MRP1 (ABCC1) in the human normal colon epithelial NCM460 cell line by reverse-transcription PCR. Via permeation studies across NCM460 monolayers we demonstrate the poor aptitude of A to interact with active efflux transporters (AET): indeed, the ratio between its permeability coefficients for the basolateral (B) → apical (A) and B → A transport was 1.5 ± 0.1, near to the ratio of taltobulin (1.12 ± 0.06), an hemiasterlin derivative able to elude AETs, and significantly different form the ratio of celiprolol (3.4 ± 0.2), an AET substrate.

Full text of DOI:10.1016/j.ejps.2016.05.032

European Journal of Pharmaceutical Sciences 91 (2016) 50–63
Contents lists available at ScienceDirect
European Journal of Pharmaceutical Sciences
journal homepage: www.elsevier.com/locate/ejps
A novel hybrid drug between two potent anti-tubulin agents as a
potential prolonged anticancer approach
a
a
a
Paolo Marchetti a, Barbara Pavan b, , Daniele Simoni , Riccardo Baruchello , Riccardo Rondanin ,
⁎
Carlo Mischiati c, Giordana Feriotto d, Luca Ferraro b, Lih-Ching Hsu e, Ray M. Lee f, Alessandro Dalpiaz a
Department of Chemical and Pharmaceutical Sciences, University of Ferrara, 44121 Ferrara, Italy
Department of Life Sciences and Biotechnology, University of Ferrara, 44121 Ferrara, Italy
Department of Biomedical and Speciality Surgical Sciences, Surgery and Experimental Medicine, University of Ferrara, 44121 Ferrara, Italy
a
b
c
d
Department of Morphology, Surgery and Experimental Medicine, University of Ferrara, 44121 Ferrara, Italy
School of Pharmacy, College of Medicine, National Taiwan University, Taipei 10050, Taiwan
e
f
Lestoni Corp., Richmond, VA 23219-155, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We report the design, synthesis and biological characterisation of a novel hybrid drug by conjugation of two tu-
bulin inhibitors, a hemiasterlin derivative A (H-Mpa-Tle-Aha-OH), obtained by condensation of three non-natu-
ral amino acids, and cis-3,4′,5-trimethoxy-3′aminostilbene (B). As we have previously demonstrated synergy
between A and B, we used a monocarbonyl derivative of triethylene glycol as linker (L) to synthesise compounds
A-L and A-L-B; via HPLC we analysed the release of its potential hydrolysis products A, A-L, B and B-L in physi-
ological fluids: the hybrid A-L-B undergo hydrolysis in rat whole blood of the ester bond between A and L
(half-life = 118.2 9.5 min) but not the carbamate bond between B and L; the hydrolysis product B-L was fur-
ther hydrolyzed, but with a slower rate (half-life = 288 12 min). The compound A-L was the faster hydrolyzed
conjugate (half-life = 25.4 1.1 min). The inhibitory activity of the compounds against SKOV3 ovarian cancer
cell growth was analysed. The IC50 values were 7.48 1.27 nM for A, 40.3 6.28 nM for B, 738 38.5 nM for
A-L and 37.9 2.11 nM for A-L-B. The anticancer effect of A-L-B was evidenced to be obtained via microtubule
dynamics suppression. Finally, we stated the expression of the active efflux transporters P-gp (ABCB1) and MRP1
(ABCC1) in the human normal colon epithelial NCM460 cell line by reverse-transcription PCR. Via permeation
studies across NCM460 monolayers we demonstrate the poor aptitude of A to interact with active efflux trans-
porters (AET): indeed, the ratio between its permeability coefficients for the basolateral (B) → apical (A) and
B → A transport was 1.5 0.1, near to the ratio of taltobulin (1.12 0.06), an hemiasterlin derivative able to
elude AETs, and significantly different form the ratio of celiprolol (3.4 0.2), an AET substrate.
Received 1 February 2016
Received in revised form 3 May 2016
Accepted 31 May 2016
Available online 01 June 2016
Keywords:
Hemiasterlin
Stilbene
Hybrid drug
Hydrolysis
Permeation
NCM460 cells
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
Abbreviations: A, H-Mpa-Tle-Aha-OH (hemiasterlin derivative obtained by
condensation of three non-natural amino acids) – Mpa, N-(1-methyl-1-phenylethyl)-
Some of the most important targets of anticancer drugs are the tubu-
lin-binding sites of microtubules (Jordan and Wilson, 2004; Yue et al.,
2010). Vinca alkaloid and taxane derivatives can interact with specific
binding sites, while stilbene derivatives are able to interact with the col-
chicine binding site of tubulin (Gupta and Bhattacharyya, 2003; Yue et
al., 2010). Interestingly, a new tubulin binding site has been identified
for the hemiasterlins, a family of natural tripeptides found in marine
sponges (Yue et al., 2010). Among the hemiasterlins, the synthetic de-
rivative taltobulin has been characterised as a potent antitumoral
agent (Zask et al., 2005).
(D)-alanine
- Tle, (D)-t.Leucine - Aha, (2E, 4S)-2,5-dimethyl-4-(methylamino)-2-
hexanoic acid); B, cis-3,4′,5-trimethoxy-3′-aminostilbene; C, cis-3,4′,5-trimethoxy-3′-
hydroxystilbene (internal standard for kinetic studies); L, monocarbonyl-triethylene
glycol; A-L-B, hybrid compound obtained by coupling A with B, using L as a spacer; A-L,
compound A linked to the spacer L using an ester function; L-B, compound B linked to
the spacer L using a carbamate function; AETs, active efflux transporters; ABC, ATP-
binding cassette; P-gp, P-glycoprotein; MRP1, multidrug resistance-associated protein;
BCRP, breast cancer resistance protein; MDR, multidrug resistance.
⁎
Corresponding author at: Department of Life Sciences and Biotechnology, University
of Ferrara, via L. Borsari, 46, 44121 Ferrara, Italy.
E-mail addresses: paolo.marchetti@unife.it (P. Marchetti), pvnbbr@unife.it (B. Pavan),
daniele.simoni@unife.it (D. Simoni), riccardo.barucchello@unife.it (R. Baruchello),
riccardo.rondanin@unife.it (R. Rondanin), carlo.mischiati@unife.it (C. Mischiati),
giordana.feriotto@unife.it (G. Feriotto), frl@unife.it (L. Ferraro), lihchinghsu@gmail.com
(L.-C. Hsu), richvain@gmail.com (R.M. Lee), dla@unife.it (A. Dalpiaz).
Taltobulin can be synthesised by condensing three non-natural amino
acids, although the enantioselective synthesis of one of these amino acids
has been reported to be especially difficult (Nieman et al., 2003; Simoni
et al., 2010). Recently, however, a versatile enantioselective means of
http://dx.doi.org/10.1016/j.ejps.2016.05.032
0928-0987/© 2016 Elsevier B.V. All rights reserved.

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
51
synthesising a new class of synthetic hemiasterlins has been proposed
(Hsu et al., 2012). The new hemiasterlins [an example, compound A
(H-Mpa-Tle-Aha-OH) is reported in Fig. 1] have proved to be
potent tubulin inhibitors, able to induce a strong synergism with
cis-3,4′,5-trimethoxy-3′aminostilbene (compound B, Fig. 1) (Simoni
et al., 2010).
especially ABCB1 and ABCC1, appear to be expressed in a wide variety of
cancers (Dean et al., 2001; Wu et al., 2011; Pavan et al., 2014).
In order to characterise the efficacy profiles of novel anticancer
drugs, it is therefore vital to consider their interactions with ABC efflux
transporters (International Transporter Consortium et al., 2010). Here
we report the use of normal human colon epithelial NCM460 cells as a
suitable cellular model for efflux studies referred to compound A. In par-
ticular, by RNA extraction and reverse- transcription PCR, NCM460 cells
were shown to express the P-gp and MRP1 active efflux transporters.
The permeation studies were completed by comparison of the perme-
ability coefficients of A for the basolateral (B) → apical (A) and B → A
transport with those of celiprolol and taltobulin, chosen for their ability
to interact with the efflux transporters or elude them, respectively
(Karlsson et al., 1993; Loganzo et al., 2003).
We report here the synthesis and characterisation of the hybrid
compound obtained by coupling hemiasterlin A with the stilbene deriv-
ative B, using the monocarbonyl derivative of triethylene glycol (L) as a
spacer (compound A-L-B, Fig. 1). The hydrolysis potential of this conju-
gate in physiological fluids (cell culture media, whole blood and liver
homogenates) was explored, in order to determine whether it could
feasibly act as a controlled release system for the hemiasterlin A,
whose activity is potentiated by the presence of the stilbene derivative
B. In order to complete the characterisation, the cytotoxic activity of
both the hybrid compound and its hydrolysis products was investigat-
ed. Finally, we have evaluated the potential ability of the hemiasterlin
derivative A (the main active hydrolysis product of the hybrid A-L-B)
to elude the active efflux transporters (AETs), that on the cell membrane
are the body's first line of defence against xenobiotics. These trans-
porters drive substrates across biological membranes, even against con-
centration gradients, thereby limiting xenobiotic entry into the cells.
This important system of defence is generally enormously enhanced in
cancer cells, in which overexpression of AETs, most notably the ATP-bind-
ing cassette (ABC) efflux transporters ABCB1 (for P-glycoprotein, P-gp),
ABCC1 (multidrug resistance-associated protein, MRP1) and ABCG2
(breast cancer resistance protein, BCRP), has been implicated in multi-
drug resistance (MDR) (Aller et al., 2009). These ABC efflux transporters,
2. Materials and methods
2.1. Chemistry
Commercially available solvents and reagents were used without
further purification. Reactions and product mixtures were routinely
monitored by thin-layer chromatography (TLC) on pre-coated silica
gel plates (Merck F254) using the indicated solvent systems. Flash chro-
matography was carried out with Merck silica gel (230–400 mesh). All
the final products undergoing biological testing were purified by pre-
parative reversed-phase HPLC [Waters Delta Prep LC 40 mm assembly
column C18 (30 cm × 4 cm, 15 μm particle size)] eluted at a flow rate
of 20 mL/min with mobile phase solvent A (10% CH3CN + 0.1% TFA in
Fig. 1. Chemical formulas of the analysed hemiasterlin (A, A-L) and stilbene (B, B-L) derivatives and their hybrid A-L-B. The stilbene derivative C was employed as an internal standard in
HPLC analysis. X− = CF3CO2−
.

52
P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
13C NMR (CDCl3): 13.9, 18.7, 19.5, 21.3, 21.4, 26.6, 26.7, 26.9, 30.0,
31.2, 35.3, 53.0, 54.7, 55.0, 61.8, 63.9, 69.2, 70.5, 70.7, 72.5, 110.2,
125.7, 126.8, 128.3, 132.4, 139.3, 167.7, 171.7, 173.1.
H2O v/v), and a linear gradient from 10% to 60% B (60% CH3CN + 0.1%
TFA in H2O v/v) over 25 min. Melting points were determined on a
Reichert-Kofler apparatus, and are uncorrected. NMR spectra (1H and
13C) were recorded on a Varian-Mercury Plus 400 spectrometer, and
chemical shifts are given in parts per million (™) downfield from
tetramethylsilane (TMS) as an internal standard; J values are expressed
in Hz. Molecular weights of compounds were determined by mass spec-
m/z: 606.709.
Compound A-L-B. A solution of DIEA (0.05 mL, 0.27 mM), PyBOP
(0.035 g, 0.07 mM) and compound B-L (Simoni et al., 2009)
(0.035 g, 0.08 mM) in CH2Cl2 (3 mL) was added to a suspension of
the trifluoroacetate salt of hemiasterlin analog A (Simoni et al.,
2010) (0.04 g, 0.07 mM) in CH2Cl2 (2 mL). The mixture was stirred
overnight at room temperature, and then concentrated in vacuo to
give a slurry, which was purified first by column chromatography
(PE/AcOEt, 1/3, Rf: 0.4) and then preparative HPLC, thereby yielding
the product as a colourless solid (0.044 g, 69%).
1H NMR (CDCl3): 0.77 (d, 3H, J = 6.8), 0.86 (d, 3H, J = 6.8), 0.98 (s,
9H), 1.16 (br, 3H), 1.50 (br s, 7H), 1.83–1.91 (m, 1H), 1.92 (d, 3H, J =
0.8), 2.97 (s, 3H), 3.03 (m, 1H), 3.65 (s, 6H), 3.68–3.79 (m, 8H), 3.81
(s, 3H), 4.28–4.31 (m, 4H), 4.74 (d, 1H, J = 9.4), 5.10 (dd, 1H, J = 7.3
and 7.1), 6.29–6.30 (m, 1H), 6.37 (t, 1H), 6.43–6.46 (m, 5H), 6.66 (d,
1H, J = 8.4), 6.92 (dd, 1H; J = 7.3 and 0.8), 7.22–7.27 (m, 1H), 7.31–
7.36 (m, 2H), 7.43–7.45 (m, 2H), 8.01 (br, 2H).
trometer ESI Micromass ZMD-2000, and values are expressed as MH+
.
Light petroleum (PE) refers to the fractions boiling in the range 40–
60 °C. All drying operations were performed over anhydrous sodium
sulfate, and evaporation was carried out using a rotary evaporator. The
purity of tested compounds was determined by combustion analysis,
which was conducted on a Yanagimoto MT-5 CHN recorder elemental
analyser at the University of Ferrara Pharmaceutical and Chemistry De-
partment Microanalysis Laboratory. All tested compounds yielded data
within 0.4% of calculated values.
The hemiasterlin derivative A, and the compounds cis-3,4′,5-
trimethoxy-3′-aminostilbene hydrochloride (B), B-L and cis-3,4′,5-
trimethoxy-3′-hydroxystilbene (C) (Fig. 1), were synthesised as previ-
ously described (Simoni et al., 2009, 2010), and the new compounds
as follows.
Compound A-L. A solution of TEG (76.6 mg, 0.51 mmol) and DIEA
(26.3 mg, 0.204 mmol) in CH2Cl2 (3 mL) was added dropwise to a
suspension of the trifluoroacetate salt of hemiasterlin analog A
(Simoni et al., 2010) (30 mg, 0.051 mmol), and PyBOP (26.5 mg,
0.051 mmol) in CH2Cl2 (3 mL). The reaction mixture was stirred
for 24 h at room temperature before removing the solvent under re-
duced pressure. The product was purified first by column chroma-
tography with (CH2Cl2/toluene/MeOH, 17/2.2/0.5, Rf1: 0.25) and
then by preparative HPLC, thereby yielding compound A-L as a
colourless solid (27.5 mg, 89%).
13C NMR (CDCl3): 14.0, 18.6, 18.9, 19.4, 25.0, 26.4, 27.4, 29.7, 29.8, 31.4,
35.8, 53.3, 55.3, 55.8, 56.2, 56.9, 63.6, 64.1, 64.3, 69.2, 69.6, 70.7, 99.7,
100.1, 106.8, 109.7, 119.3, 123.5, 126.3, 127.1, 129.2, 129.5, 129.9,
130.2, 130.5, 133, 137.8, 137.9, 160.6, 161.1, 167.4, 169.3, 171.0, 183.7.
m/z: 917.398.
2.2. Materials
Taltobulin was purchased from MedKoo Biosciences (North Caroli-
na, USA). Trifluoroacetic acid (TFA) and celiprolol hydrochloride were
obtained from Sigma-Aldrich Italia (Milan, Italy). High performance liq-
uid chromatography (HPLC)-grade methanol, acetonitrile, ethyl acetate
and water were acquired from Sigma-Aldrich, and McCoy's 5A medium
and L-glutamine were purchased from Invitrogen Life Technologies
(Carlsbad, CA, USA). Foetal bovine serum (FBS), Dulbecco's modified
Eagle's medium (DMEM) + Glutamax, streptomycin, penicillin, and
Rf2 (CH2Cl2/MeOH): 0.78.
1H NMR (CDCl3): 0.76 (d, 3H, J = 7.2), 0.85 (d, 3H, J = 7.1), 0.99 (s,
9H), 1.12 (br, 3H), 1.40–1.51 (br, 8H), 1.82–1.90 (m, 1H), 1.92 (d, 3H,
J = 1.2), 2.48–2.52 (m, 1H), 2.97 (s, 3H), 3.58–3.76 (m, 10H), 4.29–
4.32 (m, 2H), 4.74 (d, 1H, J = 10), 5.09 (dd, 1H, J = 8 and 9.8), 6.67
(dd, 1H, J = 8 and 1.2), 7.19–7.43 (m, 5H), 8.12 (br s, 1H).
Scheme 1. Synthesis of compound A-L.a aReagent and conditions: (a) PyBOP, DIEA, CH2Cl2, 2 h, 76%; (b) PhC(CH3)2NH2, Ag2O, toluene, 1.5 h, sonication, 90%; (c) LiOH, MeOH/H2O, 3 h, 82%;
(d) H+; (e) TEG, PyBop, DIEA, CH2Cl2, 24 h, 89%.

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
53
Scheme 2. Synthesis of the compound A-L-B.a aReagents and conditions: (a) Trichloromethyl chloroformate, dioxane, 60 °C, 3 h; (b) triethylene glycol, dioxane, 48 h, 78%; (c) PyBop, DIEA,
CH2Cl2, 24 h, 69%.
phosphate-buffered saline (PBS) were obtained from Invitrogen (Life
Technologies Italia, Milan, Italy).
(C, Fig. 1), was employed as an internal standard for the analysis of rat
blood and liver homogenate extracts.
The retention times obtained with the isocratic elutions were
4.82 min for the compounds A and B, 5.30 min for the compound A-L,
9.80 min for the compound B-L, 12.10 min for the internal standard C,
4.94 min for taltobulin, 4.50 min for celiprolol, and 4.60 min for the hy-
brid A-L-B. The retention times obtained with the elution regulated by
the gradient profile were the same as previously reported for the com-
pounds A, A-L, B, B-L and the internal standard C, and the retention
time of the hybrid A-L-B was 19.80 min. As the retention times of the
compounds A and B were the same, compound A released by the hydro-
lysis of the hybrid A-L-B was quantified as the difference between the
peak area obtained at 220 nm at 4.82 min and double the peak area ob-
tained at 270 nm at the same retention time as the same injected sam-
ple. Indeed, compound A was found to be totally undetectable at
270 nm, whereas the mean S.D. of the ratio of the peak areas of com-
pound B detected at 220 and 270 nm was 2.0 0.1 in the concentration
range between 5 μM and 50 μM (n = 6).
The chromatographic precisions, represented by relative standard
deviations (RSD), were evaluated by repeated analysis (n = 6) of the
same sample solution containing each of the examined compounds at
a concentration of 10 μM. The solutes were dissolved in water, with
the exception of the hybrid A-L-B, which was dissolved in a 50:50 mix-
ture of water and methanol (v/v). The RSD values ranged between 0.89%
and 0.96% for all analysed compounds. Calibration curves of peak areas
versus concentration (n = 9) were generated in the ranges 1 to 300 μM
for the compounds A and taltobulin, 0.5 to 100 μM for the compounds A-
L, B and B-L, 1 to 100 μM for the hybrid A-L-B, and 0.25 to 200 μM for
celiprolol. The calibration curves were linear (n = 9, r N 0.996,
P b 0.0001) in the concentration ranges investigated.
2.3. HPLC analysis
Quantification of taltobulin, celiprolol, the hybrid compound A-L-B,
and its potential hydrolysis products A, A-L, B and B-L was performed
by HPLC. The chromatographic apparatus consisted of a modular system
(model LC-10 AD VD pump and model SPD-10A VP variable wavelength
UV–Vis detector; Shimadzu, Kyoto, Japan) and an injection valve with
20 μL sample loop (model 7725; Rheodyne, IDEX, Torrance, CA, USA).
Separations were performed at room temperature on a 5-μm Hypersil
BDS C-18 column (150 mm × 4.6 mm i.d.; Alltech Italia Srl, Milan,
Italy), equipped with a guard column packed with the same Hypersil
material. Data acquisition and processing were performed on a personal
computer using CLASS-VP Software, version 7.2.1 (Shimadzu Italia,
Milan, Italy). The detector was set at 220 nm for analysis of compounds
A and A-L and the hybrid A-L-B, while a setting of 232 nm was used for
the analysis of celiprolol, and 270 nm for the compounds B and B-L. For
the analysis of compounds A, A-L, B, B-L and taltobulin, the mobile phase
consisted of an isocratic mixture of water and acetonitrile in the pres-
ence of TFA 0.1% (v/v) with a ratio of 60:40 (v/v); the ratios for analysis
of celiprolol and the hybrid A-L-B were 70:30 and 10:90 (v/v),
respectively.
Kinetic analysis of the degradation of hybrid A-L-B and the release of
it hydrolysis products in rat whole blood and liver homogenates were
performed separately via HPLC with a mobile phase consisting of a mix-
ture of water and acetonitrile in the presence of TFA 0.1% (v/v). Gradient
profile regulation was planned and applied as follows: isocratic elution
with a ratio 60:40 (v/v) for 14 min; then a 1-min linear gradient to the
ratio 10:90% (v/v); followed by maintenance of the mobile phase com-
position at the ratio 10:90% (v/v) for 6 min. After each cycle, the column
was conditioned with the ratio 60:40 (v/v) for 10 min. The flow rate was
1 mL/min, and the compound cis-3,4′,5-trimethoxy-3′-hydroxystilbene
The limits of quantification (LOQ) at a signal-to-noise ratio of 10
were 1.0 μM (5.9 ng injected) for A; 0.5 μM (3.0 ng injected) for A-L;
0.15 μM (0.5 ng injected) for B; 0.089 μM (0.4 ng injected) for B-L;
0.9 μM (9.3 ng injected) for the hybrid A-L-B; 0.25 μM (0.95 ng injected)
for celiprolol; and 0.6 μM (2.8 ng injected) for taltobulin. The limits of

54
P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
water, and left to equilibrate at room temperature for 36 h under con-
tinuous stirring. After filtration, sample A-L-B was analysed by HPLC,
whereas sample A was analysed after dilution to 1:100.
2.5. Kinetic analysis in culture medium
The compounds A, A-L, B and B-L and the hybrid A-L-B were sepa-
rately incubated at 37 °C in a mixture of DMEM supplemented with
10% FBS, 50 mg/mL streptomycin, and 50 IU/mL penicillin. The kinetic
analysis was performed as previously described in the case of drug incu-
bation in plasma or rat liver homogenates (Dalpiaz et al., 2001, 2012a)
The incubation phase (3 mL) was spiked with 10−2 M stock solutions
of the compounds in DMSO in order to obtain their final concentration
of 50 μM. During the experiments, the samples were shaken gently
and continuously in an oscillating water bath. At regular time intervals,
100-μL aliquots of samples were taken, and immediately quenched in
300 μL of ethanol (4 °C), then adding 100 μL of 50 μM internal standard
C. After centrifugation at 13,000g for 10 min, 400-μL aliquots were re-
duced to dryness under a nitrogen stream, and re-dissolved in 150 μL
of water-methanol (50:50 v/v). The solutions were then centrifuged,
and 10 μL was injected into the HPLC system for sample quantification.
All values were obtained as the mean of three independent incubation
experiments.
Preliminary analysis performed on blank samples showed that its
components did not interfere with the retention times of the com-
pounds analysed.
2.6. Kinetic analysis in rat whole blood
Fig. 2. [a] Degradation profiles of the compounds A, A-L, B and B-L, and the hybrid A-L-B in
rat whole blood. All values are reported as the percentage of the overall amount of
incubated prodrug. [b] Semi-logarithmic plots of the degradation profiles; their linearity
(n = 9, r ≥ 0.990, P b 0.0001) evidences a degradation following an apparent first order
kinetic (half-lives = 25.4 1.1 min for A-L, 288 12 min for B-L and 118.2 9.5 min
for A-L-B). No degradation was detected for compounds A and B. Data are reported as
the mean S.E. of three independent experiments.
The compounds A, A-L, B and B-L and the hybrid A-L-B were sepa-
rately incubated at 37 °C in heparinised whole blood obtained from dif-
ferent rats (male Wistar, Harlan SRC, Milan, Italy) weighing 200–250 g.
All procedures involving the rats were performed in accordance
with the guidelines issued by the Italian Ministry of Health (D.L. 116/
92 and D.L. 111/94-B), the Declaration of Helsinki, and the Guide for
the Care and Use of Laboratory Animals as adopted and promulgated
by the National Institute of Health (Bethesda, Maryland, USA). The ex-
perimental design has been approved by the Local Ethics Committee
(University of Ferrara, Ferrara, Italy). All efforts were made to reduce
the number of animals and their suffering.
The kinetic analysis was performed as previously described (Dalpiaz
et al., 2001). Three millilitres of whole blood was spiked with compound
solutions to yield a final concentration of 50 μM, obtained by adding 5 μL
of 10−2 M stock solution in DMSO for each millilitre incubated. During
the experiments, the samples were shaken gently and continuously in
an oscillating water bath. At regular time intervals, 100-μL aliquots of
samples were taken, and immediately quenched in 500 μL of ice cold
water, then adding 50 μL of 10% sulfosalicylic acid and 100 μL of
100 μM internal standard C. Sulfosalicylic acid was added in order to in-
duce the denaturation and precipitation of the proteins in the incuba-
tion media. Differently from sodium hydroxide, used for the same
purpose, sulfosalicylic acid does not induce the hydrolysis of the pro-
drugs (Dalpiaz et al., 2001). The samples were extracted twice with
1 mL of water-saturated ethyl acetate, and, after centrifugation, the or-
ganic layer was reduced to dryness under a nitrogen stream. Two hun-
dred microlitres of a water and methanol mixture (50:50 v/v) was
added, and, after centrifugation, 10 μL was injected into the HPLC sys-
tem. All values were obtained as the mean of three independent incuba-
tion experiments. Preliminary analysis performed on blank samples
showed that its components did not interfere with the retention times
of the compounds analysed.
detection (LOD) at a signal-to-noise ratio of 3 were 0.3 μM (1.8 ng
injected) for A; 0.15 μM (0.9 ng injected) A-L; 0.05 μM (0.15 ng injected)
for B; 0.027 μM (0.12 ng injected) for B-L; 0.27 μM (2.8 ng injected) for
the hybrid A-L-B; 0.08 μM (0.29 ng injected) for celiprolol; and 0.18 μM
(0.84 ng injected) for taltobulin.
As for hydrolysis studies in different media, recovery experiments
were performed by comparing the peak areas of each 25 μM compound
extracted from each medium at 4 °C (n = 3) with those obtained by in-
jection of an equivalent concentration of the analytes dissolved in the
water-methanol mixture (50:50 v/v). The average recoveries from rat
whole blood
S.E. of 10 μM of the compounds were 91
4% for A;
53 3% for A-L; 24 1% for B; 95 4% for B-L, and 39 4% for the hy-
brid A-L-B. The average recoveries from both the cell culture medium
and from rat liver homogenates ( S.E.) of 10 μM compound were
≥92 4% for A; ≥84 3% for A-L; ≥41% 3 for B; ≥91 5% for B-L
and ≥54 3% for the hybrid A-L-B. The concentrations of these com-
pounds were therefore referred to as peak area ratio with respect to
their internal standard. The precision of this peak area ratio-based
method is demonstrated by the RSD values ranging between 1.05%
and 1.64% for 10 μM compounds extracted from the different incubation
media at 4 °C, whose calibration curves were linear over the range 5–
50 μM (n = 6, r N 0.992, P b 0.0001). The accuracy of the method was
evaluated for each 25 μM of compound with respect to their calibration
curves (n = 6) as relative errors ranging between −2.64% and 1.68%.
2.4. Solubility assessment
2.7. Preparation of rat liver homogenates
To determine their respective solubility values, an excess of com-
pound A (8 mg/mL) or hybrid A-L-B (3 mg/mL) was added to 3 mL of
The livers of male Wistar rats were immediately isolated after their
decapitation, washed with ice-cold saline solution, and homogenized

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
55
Fig. 3. Degradation profiles of the compounds A-L [a], B-L [b] and A-L-B [c], with the corresponding appearance profiles of their hydrolysis products in rat whole blood. All the values are
reported as the concentrations of compounds detected in the blood during time and as the mean S.E. of three independent experiments.
in 4 volumes (w/v) of TrisHCl (50 mM, pH 7.4, 4 °C) with a Potter-
Elvehjem apparatus (Vetrotecnica, Padua, Italy). The supernatant ob-
tained after centrifugation (2000g for 10 min at 4 °C) was decanted off
and stored at −80 °C before being used in kinetics studies. The total
protein concentration in the tissue homogenate was determined using
the Lowry procedure (Lowry et al., 1951), and was found to be
31.8 1.3 μg of protein/μL.
a final concentration of 50 μM, obtained by adding 15 μL of 10−2
M
stock solution in DMSO. The kinetic analysis was performed as previous-
ly described (Dalpiaz et al., 2012a). During the experiments, the sam-
ples were shaken gently and continuously in an oscillating water bath.
At regular time intervals, 100-μL aliquots of samples were taken, and
immediately quenched in 300 μL of ethanol (4 °C). After addition of
100 μL of 100 μM internal standard C and centrifugation at 13,000g for
10 min, 400-μL aliquots were reduced to dryness under a nitrogen
stream, and re-dissolved in 150 μL of water-methanol mixture (50:50
v/v), prior to centrifugation and injection of 10 μL into the HPLC system.
All values were obtained as the mean of three independent incubation
experiments. Preliminary analysis performed on blank samples showed
2.8. Kinetic analysis in rat liver homogenates
The compounds A, A-L, B and B-L and the hybrid A-L-B were sepa-
rately incubated at 37 °C in 3 mL of rat liver homogenates, resulting in

56
P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
at least 30 min in 75% ethanol at −20 °C, and subjected to propidium io-
dide (PI) staining and flow cytometry analysis. At least 10,000 cells from
each sample were analysed using CellQuest Pro software (BD Biosci-
ences, San Jose, CA).
2.12. Cell culture
The human-derived non transformed colon mucosa NCM-460 cells,
kindly provided by Dr. Enzo Spisni (Dept. of Biological, Geological and
Environmental Sciences, University of Bologna, Italy) were grown in
DMEM/Glutamax culture medium supplemented with 10% foetal bo-
vine serum (FBS), 100 U/mL penicillin and 100 μg/mL streptomycin at
37 °C in a humidified atmosphere of 95%, with 5% CO2. For maximum vi-
ability, NCM460 cells were passaged at pre-confluent densities twice a
week in fresh and spent growth medium in a 1:1 ratio. NCM460 cells
were used between passages 29 and 39 in this study.
2.13. RNA extraction and reverse-transcription PCR
Total RNA was isolated from 10 million cells as follows. The cells
were lysed in 1 mL of TRIzol Reagent (Invitrogen, Carlsbad, CA, USA)
by pipetting. After incubation of the homogenized samples for 5 min
at room temperature, 0.2 mL of chloroform was added. The samples
were stirred vigorously, and then centrifuged at 12,000g for 15 min at
4 °C. The RNA was precipitated from the aqueous phase at room temper-
ature by adding 0.5 mL of isopropanol, and centrifuging at 12,000g for
10 min at 4 °C. The RNA pellet was washed once with 75% ethanol,
and then air-dried and dissolved in diethyl pyrocarbonate (DEPC)-treat-
ed water. The RNA concentration was determined by measuring the op-
tical absorbance at 260 nm. Reverse transcription was performed from
2 μg of total RNA using ImProm-II™ (Promega, Madison, WI, USA) and
a mixture of oligo-dT and random primers in a final volume of 20 μL.
To perform the PCR reactions, 0.5-μL aliquots of cDNA and the following
forward and reverse gene-specific primers were used: ABCB1, Homo sa-
piens ATP-binding cassette, sub-family B, member 1 (5′-ATG TTT CCG
GTT TGG AGC CT-3′/5′-TCC TTC CAA TGT GTT CGG CA-3′); ABCC1,
Homo sapiens ATP-binding cassette, sub-family C, member 1 (5′-CCT
GAA GGT GGA CGA GAA CC-3′/5′-TGT GCC TGA GAA CGA AGT CC-3′);
ABCG2, Homo sapiens ATP-binding cassette, sub-family G, member 2
(variant 1, 5′-CTC CCA TCG TGA CCT CCA GC-3′/5′-TCA TTG GAA GCT
GTC GCG GG-3′; variant 2, 5′-GGG TAA TCC CCA GGC CTC TA-3′/5′-
TGA GAT TGA CCA ACA GAC CAT CA-3′). In order to prevent amplifica-
tion from genomic DNA, the forward and reverse primers were de-
signed on exons separated by a long intronic sequence; briefly, 25 μL
of PCR mixture, containing 0.4 μM primers, 2.5 mM MgCl2, 0.4 mM
deoxynucleoside triphosphates, and 0.2 μL DyNAzyme II DNA Polymer-
ase (Finnzymes, Espoo, Finland) were amplified using a T100 thermal
cycler (Bio-Rad Laboratories, Hercules, CA). The thermal cycler condi-
tions included initial denaturation for 30 s, and then 28 cycles of 95 °C
for 10 s; annealing at 62 °C for 30 s; and at 72 °C for 45 s. PCR products
were separated on 2% agarose gel, stained with ethidium bromide, and
examined under UV light with the aid of a Gel Doc 1000 Documentation
System (Bio-Rad Laboratories, Hercules, CA). In order to confirm the
quality of cDNA, amplification of human actin, GAPDH and HMBS
housekeeping genes was performed in parallel PCR reactions. Negative
controls (no template cDNA) were also run on every experimental
plate to assess specificity and to rule out contamination. In order to
demonstrate the specificity of amplification, ABCB1 and ABCC1 PCR
products were sequenced in both forward and reverse directions by
using the PCR primers and the complete PCR sequence was determined.
Then, sequences were aligned to the human genomic plus transcript da-
tabase by web access to the BLASTN program (Zhang et al., 2000) at Na-
tional Center for Biotechnology Information, NIH. The alignments were
optimized by using the megablast algorithm that permits to retrieve
highly similar sequences. By using the ABCB1 PCR sequence as query,
only the sequences NM_000927.4, NC_018918.2 and NC_000007.14
Fig. 4. [a] Degradation profiles of the compounds A, A-L, B, B-L and hybrid A-L-B in rat liver
homogenates. All values are reported as a percentage of the overall amount of incubated
prodrug. [b] Semi-logarithmic plots of the degradation profiles; their linearity (n = 9,
r ≥ 0.973, P b 0.0001) evidences a degradation following an apparent first order kinetic
(half-life of A-L = 118.0
7.2 min). The degradation of both compounds B-L and
hybrid A-L-B was less than 30% during 8 h of incubation. No degradation was detected
for compounds A and B. Data are reported as the mean
S.E. of three independent
experiments.
that its components did not interfere with the retention times of the
compounds analysed.
2.9. Kinetics calculations
The computer program GraphPad Prism (GraphPad, San Diego, CA)
was used to calculate the half-lives of the compounds showing first-
order kinetic decay on an exponential decay plot of their concentration
versus incubation time. The same software was then used for linear re-
gression of the corresponding semi-logarithmic plots. The quality of fit
was determined by evaluating the correlation coefficients (r) and P
values.
2.10. Pharmacological measurements
An ovarian cancer cell line SKOV3 was used to compare the growth-
inhibitory effect of the hybrid compound A-L-B with that of A, A-L, and
B. SKOV3 cells were maintained in McCoy's 5A medium supplemented
with 10% foetal bovine serum and 1.5 mM L-glutamine, and seeded in
a 96-well plate at 2500–3500 cells/well one day before drug treatment.
Cells were then treated with various concentrations of compounds for
48 h, and cell viability was measured by MTT assay. Half-maximal inhib-
itory concentrations (IC50) were then calculated using Scientist soft-
ware. Experiments were performed in triplicate, and data were
presented as mean S.E. of two to five independent experiments.
2.11. Flow cytometry analysis
SKOV3 cells were treated with 100 nM of compound A, B or A-L-B
for 6 to 48 h, and harvested by trypsinisation. Cells were then fixed for

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
57
produced significant alignments (100% of identity), respectively corre-
sponding to the mRNA of Homo sapiens ATP-binding cassette, sub-fam-
ily B (MDR/TAP), member 1 (ABCB1) gene and to genomic sequences of
the Homo sapiens chromosome 7 containing the ABCB1 gene. Similarly,
the ABCC1 PCR sequence produced significant alignments with
the sequences NM_004996.3, NC_018927.2, NC_000016.10 and
NT_187607.1, the first was correspondent to the Homo sapiens ATP-
binding cassette, sub-family C (CFTR/MRP), member 1 (ABCC1),
mRNA and the others to the genomic sequences of the Homo sapiens
chromosome 16 containing the ABCC1 gene.
of 105 cells/mL in a 1:1 ratio of fresh and spent culture medium, housed
in 12-well Millicell inserts (Millipore, Milan, Italy) made of polyethylene
terephthalate (PET) filter membranes of surface area 1.12 cm2 and pore
size of 1.0 μm. Filters were pre-soaked for 24 h with fresh culture medi-
um, and then the upper compartment (apical, A) received 400 μL of the
diluted cells, whereas the lower (basolateral, B) received 2 mL of the
medium in the absence of cells. A half volume of the culture medium
in the upper and lower compartments was replaced with fresh medium
every two days. The integrity of the cell monolayers was monitored
by measuring the transepithelial electrical resistance (TEER) by
means of a voltmeter (Millicell-ERS; Millipore, Milan, Italy). The
resistance measured was multiplied by the area of the filter to obtain
an absolute value of TEER, expressed in Ω·cm2. The background
resistance of blank inserts not plated with cells, around 35 Ω·cm2,
was deducted from each value. The homogeneity and integrity of the
cell monolayers were also monitored by phase-contrast microscopy.
2.14. Differentiation of NCM460 cells into polarised monolayers
Differentiation to NCM460 cell monolayers was performed using a
modified version of the method reported by Dalpiaz et al. (2012a). Brief-
ly, after two passages, confluent NCM460 cells were seeded at a density
Fig. 5. Degradation profiles of the compounds A-L [a], B-L [b] and A-L-B [c], with the corresponding appearance profiles of their hydrolysis products in rat liver homogenates. All the values
are reported as the concentrations of compounds detected in the homogenate during time and as the mean S.E. of three independent experiments.

58
P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
Based on these parameters, cell monolayers reached confluence and
epithelial polarisation after 6 days, and monolayers with a stable
TEER value of 180 11 Ω·cm2 were used in permeation experiments,
after the medium in both the apical and basolateral compartments
had been replaced with fresh low-serum medium (1% FBS) and left
for 24 h.
For the A → B permeation studies, 0.4 mL of compound A, celiprolol, or
taltobulin solution was added to the apical side at time t = 0, and the in-
serts were placed in a cell culture plate whose basolateral compartment
was pre-filled with 2 mL of pre-warmed PBS containing 5 mM glucose.
At the pre-established time points, the inserts were removed and trans-
ferred to a new well-plate containing fresh PBS with glucose. The con-
tents of the basolateral compartment were collected after insert
removal, and 10-μL aliquots of filtered (0.45 μm) samples were immedi-
ately injected into the HPLC apparatus. For the B → A permeation stud-
ies, at time t = 0, 2 mL of compound A, celiprolol, or taltobulin solutions
were placed on the basolateral side of Millicell inserts whose apical side
contained 0.4 mL of fresh PBS with glucose. At the predetermined time
points, the apical samples were removed and replaced with fresh PBS
containing glucose. The collected apical samples were immediately fil-
tered (0.45 μm) and injected (10 μL) into the HPLC apparatus. TEER
values were monitored before and after each experiment. Permeation
studies were also conducted using cell-free inserts under the same con-
ditions described above as control. All values obtained were the means
of three independent experiments.
2.15. Permeation studies across cell monolayers
Permeation experiments were performed in triplicate in both the
apical-to-basolateral (A → B) and basolateral-to-apical (B → A) direc-
tions. To this end, solutions of compound A, celiprolol and taltobulin
were separately prepared in PBS containing 5 mM glucose at a concen-
tration of 100 μM, obtained by adding 5 μL of 2 · 10−2 M stock solution
in DMSO for each millilitre incubated. The fresh low-serum media-cul-
tured monolayers of NMC460 cells were removed from both the A and
B sides of the inserts, and both sides were washed twice with pre-
warmed PBS. During transport experiments, the Millicell systems
were continuously swirled on an orbital shaker (100 rpm) at 37 °C.
Fig. 6. Cell cycle arrest induced by the compounds A, A-B-L and B. SKOV3 cells were treated with DMSO (Control) or 100 nM of each compound for 6 to 48 h and harvested for flow
cytometric analysis. The hybrid compound A-B-L was as potent as compounds A and B in inducing mitotic arrest and apoptosis.

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
59
Apparent permeability coefficients (Papp) of the analysed com-
3. Results and discussion
pounds were calculated according to the following equation
(Artursson and Karlson, 1991; Pal et al., 2000; Raje et al., 2003):
3.1. Chemistry
It is known that hemiasterlins are capable of influencing both the
vinca alkaloid and colchicine-binding sites of tubulin (Nunes et al.,
2005; Ravi et al., 2005; Kuznetsov et al., 2009), and we have previously
demonstrated that the free compound A and the stilbene derivative B
synergize with each other in their anti-microtubule activity (Hsu et al.,
2012). As conjugation between active drugs can bring about their con-
trolled release through hydrolysis processes occurring in physiological
fluids (Dalpiaz et al., 2012b), this study proposes a conjugation of the
compounds A and B in order to investigate the potential antitumor ac-
tivity of the conjugate or its ability to be prodrug of the parent com-
pounds. Indeed, a hybrid approach in which at least one component
targets tubulin is known to improve the clinical outcome of anticancer
therapies (Breen and Walsh, 2010). Due to the issue that the water sol-
ubility of the stilbene derivative B is very poor (b1 mg/mL) (Simoni et
al., 2009) a hydrophilic spacer L (triethylenglicole) was used for conju-
gation with the hemiasterlin derivative A. The solubilities of compound
dc
Vr
dt
Papp
¼
ð1Þ
SACo
where Papp is the apparent permeability coefficient in cm/min; dc/dt
(μM/min) is the flux of drug across the filters, calculated as the linearly
regressed slope through linear data; Vr is the volume in the receiving
compartment (apical = 0.4 mL; basolateral = 2 mL); SA is the diffusion
area (1.13 cm2); and C0 is the initial compound concentration in the
donor chamber at t = 0.
The permeabilities were determined for the filters alone (Pf), and for
the filters covered with cells (Pt). The apparent permeability coefficients
(PE) of the cell monolayer were then calculated as follows (Raje et al.,
2003; Yee, 1997):
A and the hybrid A-L-B by us detected in water were 7.7
0.5 mM
1
1
1
¼
−
ð2Þ
(4.5 0.03 mg/mL) and 14.5 0.7 μM (0.015 0.001 mg/mL), respec-
tively. It is encouraging to note that the poor water solubility of A-L-B
indicates this hybrid as a good candidate for encapsulation studies, fur-
ther to the development of nanoparticulate formulations able to induce
selective targeting of tumoral tissues (Pavan et al., 2014).
PE Pt Pf
2.16. Statistical analysis
The synthesis and the cytotoxic activities of the compounds A, B and
B-L have been reported previously (Hsu et al., 2012; Roberti et al., 2003;
Simoni et al., 2009, 2010). We completed a series of potential hydrolysis
products of the hybrid A-L-B along with the synthesis of compound A-L,
to characterise its cytotoxicity and potential aptitude to be hydrolyzed
in physiologic fluids.
Statistical comparisons of permeability coefficients and cumulative
concentrations obtained from the transport studies were made by
one-way ANOVA or Student's t-test (GraphPad Prism). P b 0.05 was con-
sidered statistically significant. GraphPad Prism was employed for linear
regression of the cumulative amounts of the compounds in the receiv-
ing compartments of the Millicell systems. The quality of fit was deter-
mined by evaluating the correlation coefficients (r) and P values.
Single diastereoisomers of compounds A-L and A-L-B were therefore
synthesised, respectively, as depicted in Schemes 1 and 2. The synthesis
Fig. 7. Expression of multi-drug resistance channel genes in NCM460 cells. The RNA isolated from NCM460 cells was retro-transcribed, specific cDNAs for the beta-actin (lane 1), ABCB1
(lane 2) and ABCC1 (lane 3) genes were amplified by PCR and analysed by agarose gel electrophoresis. As negative control, the un-reverted RNA was amplified in the presence of primers
specific for ABCB1 (lane 4) and ABCC1 (lane 5). M = pUC mix marker 8 (Fermentas). PCR products were sequenced to confirm the specificity of amplification and the complete sequences
obtained have been reported in the lower side. Regions recognized by the PCR primers have been underlined.

60
P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
of hemiasterlin congener A represented the most challenging synthetic
sequence; its key step was an Ag2O-promoted nucleophilic substitution
(Maran, 1993; D'angeli et al., 1995; Marchetti et al., 1997) on the chiral
non-racemic precursor 2-bromo-derivative (R)(S)(S)-3, which was
achieved via condensation of 2-bromoacid (R)-1 with dipeptide
(S)(S)-2. Bromine displacement by 2-phenyl-2-propanamine yielded
tripeptide-ester (R)(S)(S)-4, which possessed the desired regio- and
stereochemistry to produce hemiasterlin derivative A via the subse-
quent hydrolysis with LiOH. The carboxyl group of A was then reacted
with triethylene glycol by standard procedures to yield the correspond-
ing ester A-L.
The amino-stilbene B was quite easily prepared, as previously de-
scribed and illustrated in Scheme 2, via a Wittig reaction between the
opportune phosphonium salt and aldehyde. The cis-stereoisomer yield
was lower than that of the trans-stereoisomer (1:2 ratio), but, being
the only active isomer, it was readily isolated by flash chromatography.
Reaction with trichloromethyl chloroformate derivatised the amino
group of B to isocyanate (Cis-5), which was condensed, without purifi-
cation, with triethylene glycol to obtain the carbamate Cis-6 (Simoni et
al., 2010). The hybrid compound A-L-B was then produced by esterifica-
tion of A at the C-terminus with hydroxyl group of Cis-6.
the linker was not hydrolyzed in rat whole blood, whereas the ester
bond between A and the linker was. Indeed, as evidenced in Fig. 3c,
the amounts of A released over time in rat whole blood appeared to cor-
respond with the amount of the hybrid A-L-B degraded. Hence we can
deduce that the appearance of B during incubation of A-L-B in rat
whole blood (Fig. 3c) was caused by the hydrolysis of B-L, released dur-
ing the degradation of the hybrid compound.
In rat liver homogenates, the compounds A-L, B-L and the hybrid A-
L-B all showed hydrolysis patterns similar to those observed in rat
whole blood, albeit characterised by significantly lower rates of degra-
dation. Indeed, as reported in Fig. 4, during 8 h of incubation less than
30% of the compounds B-L and A-L-B were degraded, whereas more
than 95% of compound A-L was degraded, showing a half-life of
118.0 7.2 min. Fig. 5 reveals that the degradation of these compounds
was due to hydrolysis, and, as observed in rat whole blood, that only the
ester bond between A and the linker of the hybrid A-L-B was hydro-
lyzed. To summarize, the compound with the fastest rate of hydrolysis
3.2. Hydrolysis and stability studies
We have firstly evaluate the potential hydrolysis pattern of the hy-
brid compound A-L-B in different media, namely cell culture medium,
rat whole blood, and rat liver homogenates. To this end it was necessary
to detect and quantify not only the hybrid compound A-L-B, but also its
potential hydrolysis products, A, A-L, B and B-L, in all incubation media.
In order to do so, we developed an efficacious HPLC-UV method of anal-
ysis employing a reverse-phase C-18 HPLC column and a mobile phase
constituted by a mixture of water and acetonitrile in the presence of
TFA following a gradient profile. No interference from the culture
medium, whole blood or liver homogenate components was observed,
and the gradient profile therefore allowed us to quantify, on the
same HPLC chromatograms in all incubation media investigated,
both the hybrid A-L-B and its hydrolysis products A, A-L, B and B-L.
Thanks to isocratic elution we were also able to use one HPLC
chromatogram to quantify the compounds A-L and B-L and their
potential hydrolysis products, A and B respectively, in all incubation
media investigated.
First we verified that neither A-L-B nor any of its potential hydrolysis
products were degraded by a mixture of DMEM supplemented with
FBS, streptomycin and penicillin, a medium employed for growth inhi-
bition studies of several cancer cell lines (Simoni et al., 2009). Although
neither A nor B incubated at 37 °C in either rat whole blood or rat liver
homogenates showed signs of degradation after 8 h (Figs. 2 and 4),
under the same experimental conditions the compounds A-L, B-L and
the hybrid A-L-B all did, exhibiting different half-lives according to the
compound and type of incubation medium. As shown in Fig. 2, the
half-lives in rat whole blood of compounds A-L, and B-L, and the hybrid
A-L-B were 25.4 1.1 min, 288 12 min and 118.2 9.5 min, respec-
tively. These degradation profiles were indicative of pseudo first-order
kinetics, as confirmed by the linear patterns of corresponding semi-log-
arithmic plots (n = 9, r ≥ 0.990, P b 0.0001), suggesting that the degra-
dation of A-L and B-L was governed by hydrolysis. This was confirmed
by the appearance in incubation media of their hydrolysis products, A
and B, respectively, in a time-dependent fashion (Fig. 3a and b). Specif-
ically, after 100 min of incubation in rat whole blood, the compound A-L
appeared to have been completely hydrolyzed, whereas the incubation
of B-L for 450 min resulted in the degradation of about 70% of its starting
concentration, alongside the appearance of a corresponding amount of
its hydrolysis product B. As reported in Fig. 3c, the degradation of the
hybrid A-L-B was accompanied by the appearance of the compounds
A and B-L, whose patterns increased over time. As no compound A-L
was detected, this indicates that the urethane bond between B and
Fig. 8. Permeation kinetics of 100 μM celiprolol [a], taltobulin [b], and compound A [c]
across the Millicell filters covered by monolayers obtained by NCM460 cells. The
permeations were analysed from the apical to basolateral compartments (A → B) and
from the basolateral to apical compartments (B → A). In all cases analysed, the
cumulative amounts in the receiving basolateral compartment were linear across
120 min or 150 min (r ≥ 0.995, P ≤ 0.002). All data are reported as mean S.E. of three
independent experiments.

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
61
was A-L in which the drug and the linker are joined by an ester bond,
3.4. Millicell permeation studies
whereas the hydrolysis of conjugate B-L was very slow, indicating that
endogenous enzymes have difficulty hydrolyzing the urethane bond be-
tween the B and L. This pattern was confirmed by the hydrolysis profile
of the hybrid compound A-L-B, in which only the ester bond between A
and L moieties was hydrolyzed to any great extent, at rates intermediate
between those of A-L and B-L. These results are supported by the obser-
vation that the biodegradation of poly(ester-urethane)s in blood ap-
pears to be mainly due to breakage of their ester bonds, whereas the
hydrolysis of urethane bonds is dependent on the linking agents used
in their synthesis (Liu et al., 2009). They also suggest that A-L-B may
be a useful means of achieving the prolonged release of compounds A
and B in physiological fluids. However, further in-vivo studies are neces-
sary to corroborate the current in vitro kinetic data.
The results above reported indicate that the hybrid A-L-B is hydro-
lyzed in physiologic fluids, allowing the release of the compound A,
that exhibits potent anti-microtubule activity, being a parent compound
of taltobulin, a potent anti-cancer agent (Simoni et al., 2010; Hsu et al.,
2012). However, is vital to determine whether, like its parent com-
pound (Loganzo et al., 2003; Zask et al., 2005), A is able to elude AET sys-
tems, whose overexpression by cancer cells induces MDR (Pavan and
Dalpiaz, 2011; Pavan et al., 2014). To this end, we set out to study the
potential interaction of A with the active efflux transporters on human
NCM460 cells. We evidenced that, after confluence in Millicell systems,
these cells acquire functional polarisation, thereby acting as an epitheli-
al barrier (Valerii et al., 2015; Ferretti et al., 2015). Therefore, we used
NCM460 cells as a pharmacological barrier-based approach of human
origin, where investigating interaction of drugs with efflux transporters.
We preferred to use NCM460 cells with respect to other consolidate
models, such as Caco-2 cells, because changes in growth characteristics
(monolayers/multilayers) are known to occur in Caco-2 cell line, due to
the loose of contact inhibition and polarisation in transformed cell
(Rothen-Rutishauser et al., 2000).
First we conducted control experiments to provide evidence that the
NCM460 cell monolayer is can be a useful model for drug transport
studies, initially by RT-PCR analysis of the expression of active efflux
transporters P-gp, MRP1 and BCRP in this cell line (Fig. 7). This analysis
was performed on total RNA using specific primers that recognize the
genes ABCB1 (P-gp), ABCC1 (MRP1), and ABCG2 – variant 1 and variant
2 – (BCRP). Since amplification products were obtained only for the
genes ABCB1 and ABCC1, we were able to deduce that NCM460 cells ex-
press the P-gp and MRP1 active efflux transporters. The expression and
activity of P-gp transporters in NCM460 cells was verified by perme-
ation experiments using the P-gp substrate celiprolol (Karlsson et al.,
1993). In particular, transport studies were performed in both the api-
cal-to-basolateral (A → B) and basolateral-to-apical (B → A) directions
after cell cultures reached the confluence on parallel sets of Millicell
well plates with similar TEER values (180 11 Ω cm2). The permeation
profiles of celiprolol across the Millicell filters covered with monolayers
of NCM460 cells are reported in Fig. 8a, where it is shown that the cu-
mulative amounts of celiprolol (nmol) in the receiving compartments
have a linear profile over 120 min in all cases (r ≥ 0.999, P ≤ 0.001), in-
dicating constant permeation conditions within this range of time. A lin-
ear profile has been also observed for permeations across the filters
alone (data not shown). The slopes of the linear fits were used to calcu-
late the permeability coefficients (Pt and Pf), which were in turn used to
calculate, as per eq. 2, the apparent permeability coefficients (PE) of
celiprolol specific for the cell monolayers, which are reported in
Table 1. The PE values for A → B and B → A transport of celiprolol were
2.95 0.06 × 10−4 and 9.98 0.56 × 10−4 cm/min, respectively, indi-
cating that the permeation rate of this drug from the basolateral to the
apical compartments was roughly 3.5 times higher than its permeation
rate in the opposite direction (P b 0.001), in accordance with the expres-
sion of the active efflux transporter P-gp in NCM460 cell monolayers.
Thus, even if NCM460 cells are known to show untransformed
3.3. Pharmacological measurements
Previous studies indicated that the IC50 values of compound A range
from 10 to 20 nM (Simoni et al., 2010; Hsu et al., 2012) and compound B
from 15 to 30 nM in UCI-101 human ovarian cancer cells (Simoni et al.,
2009; Hsu et al., 2012), whereas compound B-L is inactive and has an
IC50 value greater than 10,000 nM in the same cell line (Simoni et al.,
2009), probably due to its inability to fit into the colchicine-binding
pocket (Simoni et al., 2009). Here we compared the potencies of com-
pounds A, B, A-L and A-L-B in another human ovarian cancer cell line,
SKOV3, treated with various concentrations of the compounds for 48 h
and measured by MTT assay. We obtained IC50 values of 7.48
1.27 nM for compound A (n = 5), 40.3
6.28 nM for compound B
(n = 2), 738 38.5 nM for compound A-L (n = 3), and 37.9
2.11 nM for compound A-L-B (n = 2). This shows that, like B-L, A-L
was much less potent with respect to compound A. Taken together
with the above, these results suggest that the presence of a linker may
block the activity of both compounds A and B, by compromising the
fit with the hemiasterlin binding site. Furthermore, these data suggest
that the linker is not hydrolyzed in the in vitro cell culture system. In
contrast, the hybrid compound A-L-B was much more potent than A-L
or B-L, displaying an IC50 very close to that of compound B under cell
culture conditions. It may be that further investigation of the spacer
length may enable modulation of the activity of the A-B conjugates.
Moreover, further investigations on the potential hydrolysis of A-L-B
in the SKOV3 cells may indicate the contributions of the various com-
pounds to its pharmacologic activity.
Tubulin inhibitors exert their anticancer effect via suppression of
microtubule dynamics, which leads to mitotic arrest and apoptosis.
To determine the cell cycle effect of the hybrid compound A-L-B,
SKOV3 cells were treated with 100 nM of A-L-B for 6, 12, 24 and
48 h, and subjected to PI staining and flow cytometry analysis.
Compounds A and B were included for comparative purposes.
Results shown in Fig. 6 indicate that the hybrid compound A-L-B
was as potent as compounds A and B in inducing M phase arrest
and apoptosis, which, at 48 h, involved approximately 80% of G2/M
cells and 17% of the subG1 population.
Table 1
Permeability coefficients (PE) (×10–5 cm/min) of celiprolol, taltobulin and compound A transported by the millicell system NCM460 cells monolayer. The values were obtained from co-
efficients Pt and Pf, as in Eq. (2). Coefficients refer to transport from the apical compartment (A) to the basolateral compartment (B) and vice versa. The ratio between the PE values for the
basolateral (B) → apical (A) and B → A transport are reported.
Compound
A → B
B → A
Ratio (B → A)/( A → B)
Celiprolol
Taltobulin
Compound A
2.95 0.06
6.21 0.28
5.28 0.44
9.98 0.56a
6.96 0.24
7.96 0.37b
3.4 0.2
1.12 0.06c
1.5 0.1c,d
a
P b 0.001 as compared to PE value of celiprolol transported from A to B.
P b 0.01 as compared to PE value of compound A transported from A to B.
P b 0.001 as compared to ratio value of celiprolol.
b
c
d
P N 0.05 as compared to ratio value of taltobulin.

62
P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
properties, we observed that they express the same secretion activity
(basal-to-apical direction) of a model drug-effluxed such as celiprolol
as well as it can be observed in Caco-2 cell line, a well-established sys-
tem for permeation studies (Karlsson et al., 1993).
nanoparticle encapsulation studies, potentially being suitable for
targeted delivery and prolonged release of compound A, a potentially
efficacious and selective anticancer drug.
We then performed identical transport studies with taltobulin, a ref-
erence drug able to elude the P-gp transporter (Loganzo et al., 2003).
The permeation profiles of this drug from the apical to basolateral com-
partments (A → B) and vice versa are reported in Fig. 8b, and cumulative
amounts (nmol) in both showed a linear profile over 120 min in all
cases (r ≥ 0.998, P ≤ 0.002), indicating constant permeation conditions
within this range of time. In this case, however, the PE values for
Acknowledgements
Support from the University of Ferrara (F72I15000470005) in the
frame of the project FAR2014 is gratefully acknowledged.
References
A → B and B → A transport were 6.21
0.28 × 10−4 and 6.96
0.24 × 10−4 cm/min, respectively (Table 1), indicating a very similar
permeation rate in both directions (P N 0.05), and suggesting that this
drug was effectively eluding the AET systems, as reported in the litera-
ture (Nieman et al., 2003). Together, the data reported in Table 1 and
Figs. 7–8 provide compelling evidence that NCM460 cell monolayers
could represent a useful model for drug transport studies.
Aller, S.G., Yu, J., Ward, A., Weng, Y., Chittaboina, S., Zhuo, R., Harrell, P.M., Trinh, Y.T.,
Zhang, Q., Urbatsch, I.L., Chang, G., 2009. Structure of P-glycoprotein reveals a molec-
ular basis for polyspecific drug binding. Science 323, 1718–1722.
Artursson, P., Karlson, J., 1991. Correlation between oral absorption in humans and appar-
ent drug permeability coefficients in human intestinal epithelial (Caco-2) cells.
Biochem. Biophys. Res. Commun. 175, 880–885.
Breen, E.C., Walsh, J.J., 2010. Tubulin-targeting agents in hybrid drugs. Curr. Med. Chem.
17, 609–639.
Dalpiaz, A., Scatturin, A., Menegatti, E., Bortolotti, F., Pavan, B., Biondi, C., Durini, E.,
Manfredini, S., 2001. Synthesis and study of 5′-ester prodrugs of N6-
cyclopentyladenosine, a selective A1 receptor agonist. Pharm. Res. 18, 531–536.
Dalpiaz, A., Paganetto, G., Pavan, B., Fogagnolo, M., Medici, A., Beggiato, S., Perrone, D.,
2012a. Zidovudine and ursodeoxycholic acid conjugation: design of a new prodrug
potentially able to bypass the active efflux transport systems of the central nervous
system. Mol. Pharm. 9, 957–968.
Dalpiaz, A., Cacciari, B., Vicentini, C.B., Bortolotti, F., Spalluto, G., Federico, S., Pavan, B.,
Vincenzi, F., Borea, P.A., Varani, K., 2012b. A novel conjugated agent between dopa-
mine and an A2A adenosine receptor antagonist as a potential anti-parkinson multi-
target approach. Mol. Pharm. 9, 591–604.
D'angeli, F., Marchetti, P., Bertolasi, V., 1995. Stereoselective substitution in 2-bromo am-
ides in the presence of Ag+ or Ag2O. J. Organomet. Chem. 60, 4013–4016.
Dean, M., Hamon, Y., Chimini, G., 2001. The human ATP-binding cassette (ABC) transport-
er superfamily. J. Lipid Res. 42, 1007–1017.
Ferretti, V., Dalpiaz, A., Bertolasi, V., Ferraro, L., Beggiato, S., Spizzo, F., Spisni, E., Pavan, B.,
2015. Indomethacin co-crystals and their parent mixtures: does the intestinal barrier
recognize them differently? Mol. Pharm. 12, 1501–1511.
Hence we applied this model to study the transport of compound A,
whose permeation profiles from the apical to basolateral compartments
(A → B) and vice versa are reported in Fig. 8c. In all cases the receiving
compartments showed cumulative amounts (nmol) exhibiting a linear
profile across 120 min or 150 min (r ≥ 0.995, P b 0.001), indicating con-
stant permeation conditions during these ranges of time. In this case,
the PE values for A → B and B → A transport were 5.28 0.44 × 10−4
and 7.96 0.37 × 10−4 cm/min, respectively (Table 1), i.e., the perme-
ation rate of compound A in a basolateral to apical direction was about
1.5 times higher than the reverse (P b 0.01), demonstrating its behav-
ioural similarity to taltobulin, as opposed to celiprolol. Indeed, analysis
of the ratios between the PE values for the basolateral → apical and
apical → basolateral directions were 3.4
0.2 for celiprolol, 1.12
0.06 for taltobulin, and 1.5 0.1 for compound A (Table 1), the former
being statistically higher than the latter two (P b 0.001). As the PE ratios
for taltobulin and compound A were very similar (P N 0.05), compound
A appeared as a weaker AET substrate than celiprolol in this experiment.
As multidrug resistance (MDR) is often the major obstacle to the
success of cancer chemotherapy, it is currently a crucial issue. This phe-
nomenon is brought about by several factors, but among these overex-
pression of ABC drug transporters by cancer cells appears to be the most
influential. Indeed, a considerable amount of experimental data sup-
ports a role of ABC transporters, in particular ABCB1 and ABCC1, in ac-
quired MDR (Monti, 2007). The NCM640 cell monolayers revealed
that compound A is a substrate of ABC transporters weaker than
celiprolol, suggesting that it may exert its anticancer effect with very
poor AET-mediated MDR phenomena.
Gupta, S., Bhattacharyya, B., 2003. Antimicrotubular drugs binding to vinca domain of tu-
bulin. Mol. Cell. Biochem. 253, 41–47.
Hsu, L.C., Durrant, D.E., Huang, C.C., Chi, N.W., Baruchello, R., Rondanin, R., Rullo, C.,
Marchetti, P., Grisolia, G., Simoni, D., Lee, R.M., 2012. Development of hemiasterlin de-
rivatives as potential anticancer agents that inhibit tubulin polymerization and
synergize with a stilbene tubulin inhibitor. Invest. New Drugs 30, 1379–1388.
International Transporter Consortium, Giacomini, K.M., Huang, S.M., Tweedie, D.J., Benet,
L.Z., Brouwer, K.L., Chu, X., Dahlin, A., Evers, R., Fischer, V., Hillgren, K.M., Hoffmaster,
K.A., Ishikawa, T., Keppler, D., Kim, R.B., Lee, C.A., Niemi, M., Polli, J.W., Sugiyama, Y.,
Swaan, P.W., Ware, J.A., Wright, S.H., Yee, S.W., Zamek-Gliszczynski, M.J., Zhang, L.,
2010. Membrane transporters in drug development. Nat. Rev. Drug Discov. 9,
215–236.
Jordan, M., Wilson, L., 2004. Microtubules as a target for anticancer drugs. Nat. Rev. Cancer
4, 253–265.
Karlsson, J., Kuo, S.M., Ziemniak, J., Artursson, P., 1993. Transport of celiprolol across
human intestinal epithelial (Caco-2) cells: mediation of secretion by multiple trans-
porters including P-glycoprotein. Br. J. Pharmacol. 110, 1009–1016.
Kuznetsov, G., TenDyke, K., Towle, M.J., Cheng, H., Liu, J., Marsh, J.P., Schiller, S.E., Spyvee,
M.R., Yang, H., Seletsky, B.M., Shaffer, C.J., Marceau, V., Yao, Y., Suh, E.M., Campagna, S.,
Fang, F.G., Kowalczyk, J.J., Littlefield, B.A., 2009. Tubulin-based antimitotic mechanism
of E7974, a novel analogue of the marine sponge natural product hemiasterlin. Mol.
Cancer Ther. 8, 2852–2860.
4. Conclusions
The hybrid compound A-L-B obtained by the conjugation of the
Liu, Q., Cheng, S., Li, Z., Xu, K., Chen, G.Q., 2009. Characterization, biodegradability and
blood compatibility of poly[(R)-3-hydroxybutyrate] based poly(ester-urethane)s.
J. Biomed. Mater. Res. A 90, 1162–1176.
hemiasterlin derivative
A and the stilbene derivative B using
triethylenglicole (L) as a spacer, is able to induce a sustained release of
the compound A by hydrolysis of its ester bond with L, in rat whole
blood. The hydrolysis of the urethane bond between B and L has been
observed only for the compound B-L and it induced a slow release in
rat whole blood of the compound B. Easily synthesised by a versatile
enantioselective approach, unlike other antitumoral members of the
hemiasterlin family, compound A is known to synergize its anticancer
activity with the stilbene derivative B. The hybrid compound A-L-B
showed anticancer activity with a potency similar to that of compound
B and an order of magnitude lower than that of compound A. On the
other hand, the presence of the spacer L in compounds A-L and B-L in-
duced a drastic decrease of their anticancer activity. We have demon-
strated that the main hydrolysis product of A-L-B, i.e. the hemiasterlin
derivative A, is a weak substrate of active efflux proteins, suggesting
its poor aptitude to induce MDR phenomenon. Furthermore, the poor
water solubility of A-L-B indicates this hybrid as a good candidate for
Loganzo, F., Discafani, C.M., Annable, T., Beyer, C., Musto, S., Hari, M., Tan, X., Hardy, C.,
Hernandez, R., Baxter, M., Singanallore, T., Khafizova, G., Poruchynsky, M.S., Fojo, T.,
Nieman, J.A., Ayral-Kaloustian, S., Zask, A., Andersen, R.J., Greenberger, L.M., 2003.
HTI-286,
a synthetic analogue of the tripeptide hemiasterlin, is a potent
antimicrotubule agent that circumvents P-glycoprotein-mediated resistance in vitro
and in vivo. Cancer Res. 63, 1838–1845.
Lowry, O.H., Rosebrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurement with the
folin phenol reagent. J. Biol. Chem. 193, 265–275.
Maran, F., 1993. Electrochemical and stereochemical investigation on the mechanism of
the decay of 2-halo amide anions. The intermediacy of aziridinones. J. Am. Chem.
Soc. 115, 6557–6563.
Marchetti, P., D'Angeli, F., Bertolasi, V., 1997. Stereoselective synthesis of neutral and cat-
ionic 2-heterocyclically substituted propanamides. Tetrahedron Asymmetry 8,
3837–3842.
Monti, E., 2007. Molecular determinants of intrinsic multidrug resistance in cancer cells
and tumors. In: Teicher, B. (Ed.), Cancer Drug Discovery and Development: Cancer
Drug Resistance. Humana Press Inc., Totowa, N.J., pp. 241–260.
Nieman, J.A., Coleman, J.E., Wallace, D.J., Piers, E., Lim, L.Y., Roberge, M., Andersen, R.J.,
2003. Synthesis and antimitotic/cytotoxic activity of hemiasterlin analogues. J. Nat.
Prod. 66, 183–199.

P. Marchetti et al. / European Journal of Pharmaceutical Sciences 91 (2016) 50–63
63
Nunes, M., Kaplan, J., Wooters, J., Hari, M., Minnick Jr., A.A., May, M.K., Shi, C., Musto, S.,
Beyer, C., Krishnamurthy, G., Qiu, Y., Loganzo, F., Ayral-Kaloustian, S., Zask, A.,
Greenberger, L.M., 2005. Two photo affinity analogues of tripeptide, hemiasterlin, ex-
clusively label alpha-tubulin. Biochemistry 44, 6844–6857.
Simoni, D., Invidiata, F.P., Eleopra, M., Marchetti, P., Rondanin, R., Baruchello, R., Grisolia,
G., Tripathi, A., Kellogg, G.E., Durrant, D., Lee, R.M., 2009. Design, synthesis and biolog-
ical evaluation of novel stilbene-based antitumor agents. Bioorg. Med. Chem. 17,
512–522.
Pal, D., Udata, C., Mitra, A.K., 2000. Transport of cosalane, a highly lipophilic novel anti-HIV
Simoni, D., Lee, R.M., Durrant, D.E., Chi, N.W., Baruchello, R., Rondanin, R., Rullo, C.,
Marchetti, P., 2010. Versatile synthesis of new cytotoxic agents structurally related
to hemiasterlins. Bioorg. Med. Chem. Lett. 20, 3431–3435.
Valerii, M.C., Ricci, C., Spisni, E., Di Silvestro, R., De Fazio, L., Cavazza, E., Lanzini, A.,
Campieri, M., Dalpiaz, A., Pavan, B., Volta, U., Dinelli, G., 2015. Responses of peripheral
blood mononucleated cells from non-celiac gluten sensitive patients to various cereal
sources. Food Chem. 176, 167–174.
agent, across Caco-2 cell monolayers. J. Pharm. Sci. 89, 826–833.
Pavan, B., Dalpiaz, A., 2011. Prodrugs and endogenous transporters: are they suitable for
drug targeting into the central nervous system? Curr. Pharm. Des. 17, 3560–3576.
Pavan, B., Paganetto, G., Rossi, D., Dalpiaz, A., 2014. Multidrug resistance in cancer or inef-
ficacy of neuroactive agents: innovative strategies to inhibit or circumvent the active
efflux transporters selectively. Drug Discov. Today 19, 1563–1571.
Raje, S., Cao, J., Newman, A.H., Gao, H., Eddington, N.D., 2003. Evaluation of the blood-
brain barrier transport, population pharmacokinetics, and brain distribution of
benztropine analogs and cocaine using in vitro and in vivo techniques.
J. Pharmacol. Exp. Ther. 307, 801–808.
Ravi, M., Zask, A., Rush, T.S., 2005. 3rd structure-based identification of the binding site for
the hemiasterlin analogue HTI-286 on tubulin. Biochemistry 44, 15871–15879.
Roberti, M., Pizzirani, D., Simoni, D., Rondanin, R., Barucchello, R., Bonora, C., Buscemi, F.,
Grimaudo, S., Tolomeo, M., 2003. Synthesis and biological evaluation of resveratrol
and analogues as apoptosis-inducing agents. J. Med. Chem. 46, 3546–3554.
Rothen-Rutishauser, B., Braun, A., Gunthert, M., Wunderli-Allenspach, H., 2000. Formation
of multilayers in the caco-2 cell culture model: a confocal laser scanning microscopy
study. Pharm. Res. 17, 460–465.
Wu, C.P., Hsieh, C.H., Wu, Y.S., 2011. The emergence of drug transporter-mediated multi-
drug resistance to cancer chemotherapy. Mol. Pharm. 8, 1996–2011.
Yee, S., 1997. In vitro permeability across Caco-2 cells (colonic) can predict in vivo (small
intestine) absorption in man - fact or myth. Pharm. Res. 14, 763–766.
Yue, Q.X., Liu, X., Guo, D.A., 2010. Microtubule-binding natural products for cancer thera-
py. Planta Med. 11, 1037–1043.
Zask, A., Kaplan, J., Musto, S., Loganzo, F., 2005. Hybrids of the hemiasterlin analogue
taltobulin and the dolastatins are potent antimicrotubule agents. J. Am. Chem. Soc.
127, 17667–17671.
Zhang, Z., Schwartz, S., Wagner, L., Miller, W., 2000. A greedy algorithm for aligning DNA
sequences. J. Comput. Biol. 7, 203–214.