PEG prodrug of gambogic acid: Amino acid and dipeptide spacer effects

Article abstract of DOI:10.1016/j.polymer.2012.02.022

The clinical application of gambogic acid (GA), a natural component with promising antitumor activity, was limited due to its extremely poor aqueous solubility, rapid elimination in vivo, and wide biodistribution. To solve these problems, 30 poly(ethylene

Full text of DOI:10.1016/j.polymer.2012.02.022

Polymer 53 (2012) 1694e1702
Contents lists available at SciVerse ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
PEG prodrug of gambogic acid: Amino acid and dipeptide spacer effects
,
,
b
,
b
b
b
Ya Ding a b, Peng Zhang b, Xiao-Yan Tang b, Can Zhang a , Song Ding , Hai Ye , Qi-Long Ding ,
*
Wen-Bin Shen c, Qi-Neng Ping b
a State Key Laboratory of Natural Medicines, China Pharmaceutical University, Nanjing 210009, PR China
b School of Pharmacy, China Pharmaceutical University, Nanjing 210009, China
c Center for Instrumental Analysis, China Pharmaceutical University, Nanjing 210009, China
a r t i c l e i n f o
a b s t r a c t
Article history:
The clinical application of gambogic acid (GA), a natural component with promising antitumor activity,
was limited due to its extremely poor aqueous solubility, rapid elimination in vivo, and wide bio-
distribution. To solve these problems, 30 poly(ethylene glycol)-amino acid (or dipeptide)-gambogic acid
(PEG-spacer-GA) conjugates were synthesized. All polymeric prodrugs showed satisfactory aqueous
solubility (1.2 ꢀ 103e4.5 ꢀ 105 times of GA solubility). It was found that the molecular weight of PEG and
the choice of spacers played important role in controlling the drug percentage, water solubility, and drug
release properties of PEG-GA conjugates with and without spacers. Studies of pharmacokinetics, bio-
distribution, and cell cytotoxicity revealed that, employing the polymeric conjugation strategy, the
remarkably improved circulatory retention time and bioavailability, as well as reduced peripheral toxicity
were obtained in comprising with GA and its Cremophor EL formulation. The liver target character of
PEG-GA conjugates made them potential prodrugs for liver cancer treatment.
Received 8 December 2011
Received in revised form
20 January 2012
Accepted 16 February 2012
Available online 24 February 2012
Keywords:
Polymeric prodrug
Gambogic acid
Poly(ethylene glycol)
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
than 0.5 mg/mL), rapid plasma clearance, and wide distribution
in vivo [8,9], which would cause low bioavailability.
In the past decades, polymeric prodrug strategy has been
employed to solve practical application problems of many hydro-
phobic anticancer drugs such as Paclitaxel, Doxorubicin and
Camptothecin [1]. Conjugating drug molecules with biocompatible
water-soluble polymers, e.g. poly(ethylene glycol) (PEG),
will increase water solubility of lipophilic drugs, extend the half-life
(t1/2) of most drugs, accumulate in the tumor via the enhanced
permeability and retention (EPR) effect, and produce enhanced
therapy efficacy [2,3].
Gambogic acid (GA, shown in Fig. 1) is the major active ingre-
dient of gamboge [4,5], a brownish to orange resin obtained from
various Garcinia species. It has been revealed in 1970s that, by
selectively inducing apoptosis in tumor cells, GA exhibited exten-
sive and potential anti-tumor activities both in vitro and in vivo
[6,7], without significant inhibitory effect on the hemopoietic and
immune functions. However, the main obstacles for GA in clinical
applications are due to its extremely poor aqueous solubility (less
To solve these problems, the introduction of solubilizers, such as
L-arginine or Cremophor EL, has been employed to improve the
water solubility of GA [10,11]. However, these agents may cause
a series of side-effects such as hypersensitivity reactions, nephro-
toxicity, neurotoxicity, and cardiotoxicity. And rapid plasma elimi-
nation of GA cannot be avoided by these formulations [12].
The utility of polymeric carriers is an alternative way to solve the
problem. In our previous work, GA was encapsulated in micelle,
which was constructed by amphipathic chitosan derivatives (N-
octyl-O-sulfate chitosan, NOSC) [13,14]. Comparing with GA-L-
arginine formulation, NOSC-encapsulated drug showed higher
drug-loading rate (29.8 ꢁ 0.17%) and improved entrapment effi-
ciency (63.8 ꢁ 0.52%). Additionally, in the experiment in vivo, the
increased area under concentrationetime curve (AUC), prolonged
elimination half-life (t1/2b), passive liver target, and low nephro-
toxicity were also achieved.
However, unlike these self-assembled nanovehicles such as
micelles that tend to dissociate and release encapsulated drug upon
intravenous administration, polymeric prodrugs are much more
stable and may effectively prevent premature drug release. In the
polymeric prodrug family, PEG prodrugs have received tremendous
interest for targeted cancer therapy, due to the unique biocom-
patible and water-soluble properties of PEG polymer [15e17],
* Corresponding author. State Key Laboratory of Natural Medicines, China Phar-
maceutical University, Nanjing 210009, PR China. Fax: þ86 025 8327 1171.
E-mail address: zhangcan@cpu.edu.cn (C. Zhang).
0032-3861/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2012.02.022

Y. Ding et al. / Polymer 53 (2012) 1694e1702
(1)
1695
BocHN A1 COOH
1)
H2N A2 COOCH3
EDC, HOBt, DCM
HO CH2CH2O nA1 NHBoc
2) LiOH
HO CH2CH2O nH
MW = 2, 10, 20 kDa
(2)
BocHN dipeptide COOH
HO CH2CH2O ndipeptide NHBoc
DCC, DMAP, DCM
O
O
HO CH2CH2O n C GA
OH
29
(PEG-GA ester)
O
O
O
O
O
Gly, Lys, Phe, Pro, Arg (NO )
A1=
A2=
2
(GA)
OH
Gly, Phe, Pro
dipeptide = Gly-Gly
Gly-Phe
O
H
Phe-Phe
HO CH2CH2O nA1 N C GA
TFA, DCM
1)
2)
Pro-Pro
O
GA, DCC,
H
HO CH2CH2O ndipeptide N C GA
HOBt, DCM
(PEG-spacer-GA conjugates)
Fig. 1. Chemical structure of GA and synthetic scheme of PEGeGA conjugates with and without amino acid or dipeptide spacers.
which has been approved by FDA. A variety of PEG-drug conjugates
[3,18,19], including protein, peptide, and several chemotherapy
agents [20e23], such as doxorubicin, paclitaxel, gemcitabine, and
camptothecin, have been reported.
(Shimadzu, USA). HPLC was performed using an Agilent 1100 series
with a DiamonsilÒ C18 reversed-phase chromatography column
(250 ꢀ 4.6 mm)and thedetectionwavelengthwas 360 nm. Methanol/
water (93/7, v/v, pH 3.5) solution acidified by phosphoric acid was
used as the mobile phase, and the flow rate was set to be 1 mL/min.
In addition to the choice of polymeric partial, polymer linkage is
an important factor in determining therapeutic potentials. The
connection between PEG and drug can be a direct chemical bond
[20e23], including hydrazone, ester, and amide bonds, or be
a heterobifunctional spacer such as amino acid based linkages
[22,24e26]. Glycine, alanine, and small peptides are preferred due
to their chemical versatility for covalent conjugation and biode-
gradability, which can be a specific substrate for plasmin enzyme
whose concentration is high in various kinds of tumor mass [24,26].
In this work, for improving water-solubility, circulation time in
body, and pharmacokinetic profiles of GA, 30 PEGeGA conjugates
linked with and without spacers (including 5 amino acids and 4
dipeptides) were synthesized. The in vitro drug release, pharma-
cokinetics, biodistribution, and cell cytotoxicity of PEGeGA conju-
gates were systematically investigated to demonstrate the
importance of molecular weight of polymer and steric hindrance of
spacers in the design of macromolecular prodrugs.
2.3. General synthesis method of PEG-spcaer-GA conjugates
2.3.1. Isolation and purification of gambogic acid
Dry gamboge resin from the Garcinia hanburyi tree (100 g) was
suspended in pyridine (300 mL) and stirred at 80e90 ꢂC for 30 min
to form a pyridine salt of GA. After filtered through kieselguhr,
20 mL water was added into the filtrate. The mixture was cooled to
4 ꢂC overnight, and the precipitate was collected and washed with
pyridine solution (70%, v/v) and water for several times. After dried
under reduced pressure, a yellow powder was dissolved in ethyl
ether (500 mL) to reflux for 30 min, and then filtrated, concentrated
and precipitated by using petroleum ether, the yellow precipitate of
GA pyridine salt was collected and dried. The obtained solid was
dissolved in ethyl ether (250 mL) and washed with aqueous HCl
(1 M) and water. The ether solution was then dried over sodium
sulfate and evaporated to yield an orange powder of 8.2 g (GA,
purity 96.7%, w/w, analyzed by HPLC).
2. Experimental details
A general synthesis method was followed for obtaining all
PEGeGA conjugates (showed in Fig. 1), and synthesis procedure of
PEG2kDaeproeproeGA conjugate was described as an example.
2.1. Materials
PEG with molecular weight (MW) of 2, 10, and 20 kDa was
purchased from Sinopharm Chemical Reagent Co. (Shanghai,
China). All amino acids and their esters were purchased from GL
Biochem Ltd. (Shanghai, China), and gamboge resin was obtained
from Anhui Bozhou Herb Market (Anhui, China). HPLC grade
reagents were used in HPLC analysis, and other reagents were
2.3.2. N-tert-butoxycarbonyl-prolyl-proline methyl ester
(BoceproeproeCOCH3)
The synthesis of compounds 1, N-tert-butoxycarbonyl-amino
acid, was referred to a previous report [27]. The resultant N-tert-
butoxycarbonyl-L-proline (2.15 g, 0.01 mol) was dissolved in
dichloromethane (DCM, 10 mL). Slow dropping of HOBt (1.62 g,
0.012 mol) in a minimal amount of tetrahydrofuran (THF) was over
30 min. When the solution was cooled to 0 ꢂC, EDC hydrochloride
(2.11 g, 0.011 mol) in DCM (15 mL) was added, and additional stir for
30 min in ice-water bath was carried out. And then, a well-mixed
DCM solution (20 mL) containing L-proline methyl ester hydro-
chloride (1.82 g, 0.011 mol) and 4.5 mL triethylamine (TEA,
0.033 mol) was added into the reaction solution and stirred at room
analytical grade and used as received. Deionized water (>18 M
U,
Purelab Classic Corp., USA) was used in all experiments.
2.2. Measurements
1H NMR spectra were determined on a Bruker (AVACE) AV-500
spectrometer using deuterated chloroform as the solvent. UV/Vis
spectra were recorded on a UV-2401 PC UV/Vis spectrophotometer

1696
Y. Ding et al. / Polymer 53 (2012) 1694e1702
temperature overnight. After evaporation under reduced pressure,
the residue was suspended in ethyl acetate and filtrated. The filtrate
was washed with dilute citric acid, saturated sodium bicarbonate
solution, and water. Dried over anhydrous sodium sulfate and
concentrated in vacuum, the off-white BoceproeproeCOCH3 was
obtained (2.83 g, 87%).
2.4. Drug content of PEGeGA conjugates
An ultraviolet absorbance method at
a fixed wavelength
(360 nm), where PEG and spacer have no absorption, was estab-
lished for the determination of GA content in PEGeGA conjugates.
The GA content of conjugates was determined with the help of
a calibration curve of GA in methanol, range from 2 to 40 mg,mL
2.3.3. N-tert-butoxycarbonyl-prolyl-proline (BoceproeproeCOOH)
After BoceproeproeCOCH3 (1.5 g, 4.6 mmol) was dissolved in
methanol (15 mL) and cooled to 0 ꢂC, 7 ml of LiOH aqueous
solution (7 mmol) was added dropwise into the solution to react
for 1 h at 0 ꢂC and 2 h at room temperature. After reaction,
methanol was removed from the solution by evaporation. The
resultant aqueous phase became turbid when pH value was
adjusted to w3 with potassium hydrogen sulfate (1 mol/L) at 0 ꢂC.
Extracted by ethyl acetate for three times, the organic layer was
washed with saturated sodium chloride solution, dried over
sodium sulfate, and evaporated to yield white powder,
BoceproeproeCOOH (1.1 g, 77%).
with R2 ¼ 0.9991, and the GA content was calculated as follows:
ꢀ
ꢁ
GA% ¼
m
GA=mprodrug ꢀ 100%
2.5. Solubility of PEGeGA conjugates
A visual observation method was used to evaluate the solubility
of PEGeGA conjugates. An exact amount of sample (100 mg) was
weighed and placed in a volumetric flask (2 mL) and 25 ml aliquots
of distilled water were added sequentially using a pipette, followed
by sonication for 30 s. When the solution became fluidic and clear,
the total volume of water added was calculated. The experiment
was carried out at 25 ꢂC, and each sample was repeated in triplicate.
2.3.4. BoceproeproePEG2kDa
PEG (2.0 g, 2 mmol OH, MW ¼ 2 kDa), BoceproeproeCOOH
(0.9 g, 3 mmol), and DMAP (0.07 g, 0.6 mmol) were dissolved in
anhydrous DCM (30 mL), and DCC (0.66 g, 3.2 mmol) was added
into the solution and stirred overnight. Produced dicyclohexylurea
(DCU) was filtered off and DCM in filtrate was evaporated. The
residue was dissolved in ethyl acetate and cooled at 4 ꢂC to filter
trace DCU. The filtrate was concentrated to a mud-like sample,
which was added dropwise to 100 mL of vigorously stirred cold
ethyl ether. The white precipitate was filtered and dried under
2.6. In vitro drug release of PEG-GA conjugates
In vitro GA release from the conjugates was performed by both
chemical and enzymatic hydrolysis.
The chemical hydrolysis was carried out in triplicates in PBS at
pH 7.4 and 5.5 at 37 ꢂC, respectively. The solution (10 mL) of all
conjugates of 10 mmol equivalent GA concentration was stirred
mildly. At the scheduled time, 20 mL of the solution was analyzed by
reduced
pressure,
affording
the
crude
product
of
an HPLC method to determine the release of GA from conjugates
(tR,prodrug ¼ 2.67 min and tR,GA ¼ 14.48 min). The half-life of each
conjugate in different media was calculated using linear regression
analysis.
BoceproeproePEG2kDa (1.8 g, 90%) used without further
purification.
2.3.5. PEG2kDaeproeproeNH2
The method for analysis of GA release in plasma and in liver
homogenate was similar to that of hydrolysis in PBS, except for the
pretreatment process. Blank human plasma was procured from the
Nanjing Red Cross Blood Center. Rat liver tissue was accurately
weighed (300e500 mg) and homogenized in 2.5 mL of normal saline
containing 1.9% w/v NaCl (Sodium Chloride) solution prior to use.
The crude product of BoceproeproePEG2kDa (1.8 g) was dis-
solved in 10 mL of TFA/DCM (1:1, v/v) and stirred at 0 ꢂC for 1 h.
When deprotection process was completed, the solvent was
removed under reduced pressure, and the resulting slurry was
added dropwise to vigorously stirred cold ethyl ether. The
precipitate was collected and resolved in ethanol/ethyl ether
mixed solvent to make a saturated solution, and the mixture
solution was then cooled to form a white precipitate. The solid was
washed with ethyl ether for several times, and the white powder
was dried in a vacuum overnight to obtain PEG2kDaeproeproeNH2
(1.45 g, 81%).
PEGeGA conjugate (100 mmol GA equivalent content) in 100
saline was added to 10 mL of plasma/liver homogenate and incu-
bated at 37 ꢂC with mild stirring. At designated time intervals,150
of plasma or liverhomogenatesolutionwas treated with 50 L of 1 M
hydrochloride acid and extracted by 800 L of acetonitrile on
mL of
m
L
m
m
a vortex mixer for about 3 min followed by centrifugation at
10,000 rpm for 10 min. The clear supernatant liquid was analyzed by
an HPLC method and the half-life of each sample was calculated
accordingly. Each experiment was repeated in triplicate.
2.3.6. PEG2kDaeproeproeGA conjugate
GA (0.75 g, 1.2 mmol), PEG2kDa-pro-pro-NH2 (1.2 g, 1 mmol) in
anhydrous DCM (15 mL), and HOBt (0.19 g, 1.4 mmol) in a minimal
amount of THF were mixed at room temperature. A solution of DCC
(0.29 g, 1.4 mmol) in a minimal amount of DCM was added and the
reaction solution was stirred for 24 h. After filtration and evapo-
ration, an oily substance was obtained and resolved in isopropanol
to make a saturated solution, and the solution was then cooled to
form an orange precipitate. The orange powder was filtered,
2.7. In vivo pharmacokinetics
The animal experiment protocols were approved by the
University Ethics Committee for the use of experimental animals
and conformed to the Guide for Care and Use of Laboratory
Animals. The in vivo pharmacokinetics was carried out in
SpragueeDawley rats of 190e220 g and 4e6 weeks old. The rats
were supplied by the Laboratory Animal Center of Nantong
University. The animals were held in air-conditioned rooms,
provided with standard food and filtered water.
washed with ether and dried in
a
vacuum to afford
PEG2kDaeproeproeGA conjugate (0.97 g, 81%).
2.3.7. PEG2kDaeGA ester
PEG2kDa-GA ester without spacers was prepared in a similar
fashion to the above procedure, where PEG2kDa (1 equiv.) and GA (2
equiv.) were directly coupled utilizing the DCC/HOBt method in 82%
(w/w) yield. The same procedure could also be used for the
synthesis of PEG10kDa- and PEG20kDa-GA esters.
PEG10kDaeGA conjugates and a Cremophor EL preparation of GA
(GAeC) were diluted in normal saline containing 1.9% w/v NaCl and
filtrated through 0.22
mm pore-sized micropore films. Animals
were divided into eleven groups with each of five rats and were

Y. Ding et al. / Polymer 53 (2012) 1694e1702
1697
given an intravenous injection via tail vein of GAeC, PEG10kDaeGA
conjugates with and without spacers solution at equivalent dose
(4 mg/kg of GA), respectively. All animals were observed for
mortality, general condition and potential clinical signs.
reaction site for the polymeric conjugation. In our initial synthesis,
connecting GA with spacer firstly showed a relatively higher
binding efficiency of drug with spacer molecules. However, GA was
found to be unstable in the conjugation and after treatment
The blood samples were collected into the heparinized tube at
time period of 0, 5, 15, 30, 60, 90, 120, 180, and 240 min post-
treatment. The blood samples were centrifuged at 4000 rpm for
10 min and the separated plasma was stored at ꢃ20 ꢂC until anal-
ysis. Liquideliquid extraction was performed by mixing the plasma
processes, due to the instability of the a, b-unsaturated ketone at
the C-10 position [28,29]. Therefore, the synthesis of PEG-spacer
and then the coupling of GA on polymer were proceeded as
shown in Fig. 1.
t-Boc-protected amino acids (1, including Gly, Lys, Phe, Pro, and
Arg(NO2)) was synthesized firstly to act as the amino acid spacer in
PEG-GA conjugate preparation. Coupling t-Boc-protected Gly, Phe,
and Pro with their corresponding methyl esters would lead to t-
(150
300
m
L) with 50
mL of 1 M hydrochloride acid and extracted by
mL of acetonitrile on a vortex mixer for about 3 min followed by
centrifugation at 10,000 rpm for 10 min, after which the upper
supernatant was analyzed by using the standard curve-based HPLC
method. The parameters of the pharmacokinetics were calculated
by using DAS procedures.
Boc-dipeptide methyl esters (dipeptide
¼
GlyeGly, GlyePhe,
PheePhe, and ProePro). And then, t-Boc-dipeptide (2), for con-
necting PEG and GA as the dipeptide spacer, can be obtained by
careful hydrolysis with lithium hydroxide to release free carboxyl
group of dipeptide.
2.8. Biodistribution
Conjugating PEG molecule (MW ¼ 2, 10, and 20 kDa) with GA
and t-Boc-spacer (1 or 2), respectively, PEGeGA esters and t-Boc-
spacer-PEG conjugates were obtained. The removal of the t-Boc
moiety on amino acid or dipeptide spacer was achieved with acid
treatment (TFA/DCM) in an ice bath. Finally, the classical DCC/HOBt
catalyzed coupling of PEG-spacer-NH2 and GA was carried out to
yield the crude product of PEG-spacer-GA conjugates. All conju-
gates were resolved in 2-propanol to form a saturated solution, and
the solution was cooled to obtain yellow solid and wash with ethyl
ether for several times, the un-reacted GA and other undesired
substances were separated from the conjugates.
SpragueeDawley rats (50% male and 50% female) were randomly
assigned to four groups (n ¼ 100) with intravenous injection of
GAeC,
PEG10kDaePheeGA,
PEG10kDaePheePheeGA,
and
PEG10kDaeGA conjugate at the equivalent 4 mg/kg dose respectively.
Each group has five sets of five rats sacrificed at five designated time
points for the biodistribution investigation. Before drug adminis-
tration, GAeC, PEG10kDaePheeGA, PEG10kDaePheePheeGA, and
PEG10kDaeGA conjugate were diluted in normal saline containing
1.9% w/v NaCl and filtrated through 0.22 mm pore-sized micropore
films to obtain an estimated injection volume of 1e1.5 mL.
Intravenous injection was given via the tail vein. All animals
were observed for mortality, general condition and potential clin-
ical signs. Animals in each set were sacrificed by cardiac stick
exsanguinations at 5, 30, 60, 120, 240 min, respectively after the
injection and tissues (heart, liver, spleen, lung, kidney, and brain)
were collected. The tissues were then washed with saline, weighed
and homogenized. After that, 1 ml of plasma, an average weight of
organ (liver ¼ 0.9140 g, lung ¼ 0.1109 g, heart ¼ 0.0810 g,
spleen ¼ 0.1245 g, kidney ¼ 0.2302 g, and brain ¼ 0.3555 g) for each
was mixed with PBS, followed by extraction and analysis as the
blood sample was done.
All of final conjugates were characterized by FT-IR and 1H NMR
analysis in CDCl3 [30]. For instance, the typical FTeIR spectra of
PEG2kDa and PEG2kDaeProeProeGA conjugate were shown in Fig. 2
(A and B, respectively). In the spectrum of PEG2kDa, the absorption
bands located at 2860 cmꢃ1 and 1465 cmꢃ1 are, respectively, due to
the adsorption of CeH stretching and bending vibrations of
methylene groups in PEG. The broad band at 3346 cmꢃ1 arises from
OeH stretching vibration of hydroxyl ends on PEG, and the absor-
bance of 1110 cmꢃ1 is due to CeO stretching vibration in PEG
backbone. While several new IR signals attributed to the ProePro
spacer and GA structure are shown in IR spectrum of
PEG2kDaeProeProeGA conjugate in Fig. 2 (curve B) in comparison
with that of PEG2kDa. The absorption band at 1739 cmꢃ1 can be
assigned to yC¼O in ester bond between GA and ProePro spacer. And
2.9. Cell cytotoxicity assay
In vitro cytotoxicity of the conjugates was quantified by
measuring its IC50 (drug concentration inhibiting 50% of cells) on
a human hepatoma cell strains (HepG2). HepG2 cells were cultured
in RPMI-1640 containing heat-inactivated FCS (RPMI-10% FCS) at
new absorbance peaks located at 1653 cmꢃ1 and 1626 cmꢃ1
/
1593 cmꢃ1 attributed to yC¼O of amide I group and yC¼C of benzene
a density of 1 ꢀ 105 cells/180
mL/well into 96-well microtiter plates
A
and allowed to proliferate at 37 ꢂC in a humidified incubator with 5%
CO2 for 24 h. A total of 20
(final GA equivalent concentration 0.5e5000
solved by DMSO (<1&)) were added to the wells and all samples
were prepared and measured in quintuplicate for each concentra-
tion. After 24 h of incubation, 20 mL of MTT (5 mg/mL) per well was
added and the plates were incubated for a further 4 h. Then, the
m
L serial dilutions of the tested compounds
1464
mg/mL, GA was dis-
3346
2860
B
1110
supernatant was removed followed by the addition of 150 mL of
DMSO to each well. 15 min after that, the absorbance at 490 nm was
detected with a microplate reader and the IC50 values for each
compound werecalculated from absorbance vs. dilution factor plots.
1739
1593
1626
1653
3. Results and discussion
4000 3000
2000
1000
3.1. Synthesis and characterization of PEG-GA conjugates
Wavenumber / cm-1
It has been reported that the 29-carboxy group is a potential
structure modification site of GA, [16], which was chosen as the
Fig. 2. FTeIR spectra of PEG2kDa (curveA) and PEG2kDaeProeProeGA conjugate (curve B).

1698
Y. Ding et al. / Polymer 53 (2012) 1694e1702
ring, respectively, indicate the successful linkage of dipepptide
spacer and conjugation of GA molecule on PEG long chain.
1H NMR spectra of PEG-GA and PEGeProeProeGA conjugate
were shown in Fig. 3 (PEG MW ¼ 2 kDa). Resonance at 3.7 ppm
corresponding to protons on PEG appeared in both spectra, in
which characteristic peaks of GA were also showed. It displayed the
existence of PEG and GA in conjugate regardless of the spacer. By
comparison of the resonances associated with PEG-pro-pro-GA
conjugate with those associated with PEG-GA ester, additional
chemical shifts at ca. 4.5, 3.5, and 2.1 ppm (marked with stars in
Fig. 3) can be attributed to protons in ProePro spacer. It demon-
strated the successful linkage of dipeptide spacer between PEG and
GA molecules.
A0
A
B0
B
C0
C
30
25
20
15
10
5
0
3.2. Drug content of PEGeGA conjugates
Fig. 4 showed the GA content in all PEGeGA conjugates. It was
found that with PEG MW rose from 2 to 20 kDa, GA content of
conjugates decreased obviously. GA content in PEGeGA esters
decreased from 22.09 to 0.49, and the average value of spacer-
linked conjugates dropped from 20.82 to 3.17, respectively. This
phenomenon could be due to two reasons, the GA content calcu-
lation method and the steric hindrance of PEG polymers. Due to the
calculation method, GA%¼(mGA/mprodrug) ꢀ 100%, large polymeric
molecular weight leading to low theoretical GA content value was
reasonable. However, comparing the theoretical and measured
values of conjugates, several PEG2kDa-based conjugates showed
higher GA content than their theoretical values (Fig. 4, curve A0 and
A). However, measured GA content of PEG20kDa-based conjugates
(curve C) was relatively consistent with their theoretical value
(curve C0). This result indicated the existence of disubstituted GA-
Spacer
Fig. 4. The theoretical (A0eC0) and measured (AeC) values of GA content in PEGeGA
conjugates with and without different spacers (PEG MW ¼ 2 (A0 and A), 10 (B0 and B),
and 20 (C0 and C) kDa).
PEG-GA molecules in low molecular weighted PEG2kDa conju-
gates, and that high PEG MW products (PEG20kDaeGA prodrugs)
were more inclined to be the mono-substituted compound. It
implied that the terminal concentration became lower with
increasing MW of PEG, which would depress the efficiency of
conjugation reaction.
In addition, keeping PEG MW unchanged, it was found that the
existence and selection of amino acid or dipeptide spacers played
Fig. 3. 1H NMR spectra of PEG2kDaeGA ester and PEG2kDaeProeProeGA conjugate.

Y. Ding et al. / Polymer 53 (2012) 1694e1702
1699
3.3. Solubility of PEG-GA conjugates
PEG 2kDa
PEG 10kDa
PEG 20kDa
A direct observation method is used to evaluate the solubility of
PEGeGA conjugates [31], and the equivalent solubility value was
calculated depending on the measured solubility and GA content in
Fig. 4. Compared with the extremely poor solubility of parent drug
200
150
100
50
(GA, 0.5 m
g,mLꢃ1), all of conjugates exhibited satisfactory aqueous
solubility, which was 1.2 ꢀ 103-fold (PEG20kDaeGA, 0.61 mg,mLꢃ1
)
at least and 4.5 ꢀ 105-fold (PEG2kDaeGlyeGA, 225.6 mg,mLꢃ1) at
most of GA (shown in Fig. 5).
Although there was disubstituted GA on both ends of PEG in the
case of PEG2kDa-based prodrugs, PEG2kDaeGA conjugates still
showed the best water solubility due to the best aqueous solubility
of PEG2kDa and the possibility of micellar formation self-assembled
by PEG2kDaeGA conjugates in the aqueous solution. It suggested
that the solubility of prodrugs was highly depended on the water
solubility of polymer used.
0
Comparing with PEG-GA esters having the same PEG MW, the
existence of spacers had little effect in enhancing the water solu-
bility of conjugates in the case of PEG2kDa-spacer-GA sample.
However, PEG10kDa- and PEG20kDa-spacer-GA samples displayed
better aqueous solubility than those of their esters, although it
showed no obvious regularity among different spacers. It implied
that the function of spacer in controlling the polymer prodrug
properties was generally reflected in large MW cases, which was
consistent with the result of GA content.
Spacer
Fig. 5. Equivalent aqueous solubility of GA in conjugates.
an important role in adjusting GA content results. On one hand,
the GA percentage of PEG-GA ester always showed lower value
than that of conjugate containing spacers, especially in cases of
large polymer prodrugs (Fig. 4, curves B and C). It indicated that,
in comprising with hydroxyl-terminated PEG, amino acid or
dipeptide-modified PEG molecules can be seen as an activated
polymer, which showed higher conjugation efficiency with drugs.
On the other hand, those spacers having large side chain groups,
such as Arg(NO2), Phe, and Pro, would result in low GA
percentage no matter how much the PEG MW was.
3.4. In vitro drug release of PEGeGA conjugates
Considering the possible in vivo process of prodrugs, chemical
and enzymatic hydrolysis was carried out to evaluate the drug
release behavior of PEGeGA conjugates by using HPLC method. The
half-life (t1/2) values corresponding to all conjugates and hydrolysis
PEG 2kDa
A
120
B
120
PEG 10kDa
PEG 20kDa
100
100
chemical hydrolysis (pH7.4)
chemical hydrolysis (pH5.5)
80
60
40
20
0
80
60
40
20
0
C
14
D
14
12
12
human plasma
rat liver homogenate
10
8
10
8
6
6
4
4
2
2
0
0
Spacer
Fig. 6. In vitro drug release: half life of each conjugate in various media at 37 ꢂC toward chemical hydrolysis and enzymolysis in (A) buffer pH 7.4, (B) buffer pH 5.5, (C) human
plasma, and (D) rat liver homogenate (All experiments were done in triplicate. SD ¼ ꢁ10%.).

1700
Y. Ding et al. / Polymer 53 (2012) 1694e1702
conditions were shown in Fig. 6. The results implied several
important clues, demonstrating the structure-dependent proper-
ties of conjugates in this work.
hydrolysis rate. As shown in Fig. 6, PEG20kDaeGA conjugates were
the most stable compounds, of which the half-lives were 2e5 times
as those of PEG2kDa with the same spacers.
Firstly, PEG-GA esters, connecting GA and PEG molecules with
ester bond directly, displayed the highest stability towards both
chemical and enzymatic hydrolysis. It is consistent with the
previous literature [18,19,24], reporting that several PEG-drug
esters were excreted in the form of prodrug before exerting its
pharmacodynamic effect. In contrast, conjugates with amino acid
or dipeptide spacers showed relative fast and appropriate drug
release velocity, insuring the effective release of bioactive
molecules.
Secondly, the drug release behavior toward chemical hydrolysis
was evaluated at pH 7.4 and 5.5 in PBS (37 ꢂC) to mimic the phys-
iological and cell lysosomal environments, respectively (Fig. 5A and
B). The t1/2 values for products at pH 5.5 were commonly shorter
than those at pH 7.4, implying that lower pH environment led to
faster releasing speed. Namely, selective release of active drug (GA
in this case) under the acidic physical environment in cancer cells
rather than in normal cells could be achieved.
3.5. In vivo pharmacokinetics
Conjugates of PEG10kDa series were chosen and assayed for
in vivo pharmacokinetics. A Cremophor EL preparation of GA
(GAeC) was selected as the control. In this study, GA concentrations
in plasma was measured over 4 h after the i.v. injection of conju-
gates or GA-C to rats at a dose of 4 mg/kg (equivalent GA concen-
tration), and the plasma concentrationetime curves were analyzed
by two-compartment models. The pharmacokinetic parameters
were presented Table 1.
Conjugates obviously showed prolonged mean residence time
(MRT) and elimination half-life (t1/2b), in comparison with both GA
(MRT ¼ 21.05 min, t1/2b ¼ 16.07 min) [15] and GA-C. This result may
attribute to the polymeric protection role of PEG and stable drug
Thirdly, drug release by enzymatic hydrolysis was also tested to
evaluate the release properties of conjugates in the presence of
nonspecific proteases (in human plasma) or hydroamidases in the
liver (in rat liver homogenate) (shown in Fig. 6C and D). It revealed
that all conjugates were more rapidly activated to GA in human
plasma and rat liver homogenates than in PBS. For example, in PBS
at pH 5.5, the half-lives of the PEGeArg(NO2)eGA conjugates with
PEG MW of 2, 10, and 20 kDa were 42.2 h, 68.4 h, and 79.5 h,
respectively. While in human plasma, 50% of GA was released from
the same conjugates after only 2.8 h, 6.5 h, and 10.2 h. And in the
liver homogenate, the half-lives were even shorter. These results
strongly suggest that enzymatic activation may be the major
contributor to the release of GA from the polymeric conjugates
in vivo.
A
Phe
10
8
Phe-Phe
None
GA-C
6
4
2
0
heart liver lung kidney spleen brain plasma
B
Fourthly, increased steric hindrance provided by spacers with
large side chain could lead to a significantly prolonged release
character under the same environmental condition. In the case of
amino acid spacers, the half-life of PEG conjugates employing
Arg(NO2) or Phe spacer showed at most 8 or 5 times longer t1/2
values than those using Gly moiety. Similarly, for the dipeptide-
connected conjugates, such as GlyeGly, GlyePhe, and PheePhe
spacers, a gradually prolonged half-life value was found due to
7
6
5
4
3
2
1
0
Phe
Phe-Phe
None
GA-C
the increasing size of the a-substituted group in spacers. This can
be interpreted that the large side chain-induced steric hindrance
would become an obstacle for the enzyme approaching, and lead to
a decreased hydrolysis rate and prolonged half-life.
Finally, the molecular weight of PEG is another factor that can
contribute to the GA release control. The lower terminal concen-
tration arising from the increasing MW of PEG made the action site
of the proton or enzyme inaccessible and slowed down the
heart liver lung kidney spleen brain plasma
C 3.5
Phe
Phe-Phe
None
3.0
2.5
2.0
1.5
1.0
0.5
0.0
GA-C
Table 1
Pharmacokinetic study of PEG10kDa-GA conjugates (n ¼ 5).
Spacer in
conjugates
AUC
g min/ml)
MRT
(min)
t
1/2a
t
1/2b
Vd
(ml)
Cl
(
m
(min)
(min)
(ml/min)
GAeC
Gly
Lys
Phe
Pro
Arg(NO2)
GlyeGly
GlyePhe
PheePhe
ProePro
None
85.41
38.52
39.93
43.44
50.19
49.05
56.73
58.73
69.72
81.03
96.31
57.58
9.35
7.23
46.03
52.18
85.63
72.91
73.12
72.19
48.32
61.64
99.27
161.61
73.37
239.94
54.48
75.94
67.81
59.01
82.07
25.25
12.3
11.91
2.28
4.43
3.54
3.23
4.22
0.44
0.61
1.24
3.24
1.22
437.14
516.56
463.79
376.81
219.82
381.48
308.7
227.18
239.76
208.65
20.87
16.77
13.29
19.54
7.87
heart liver lung kidney spleen brain plasma
9.26
17.22
16.07
11.55
20.11
18.83
89.37
Fig. 7. Tissue distribution profile of GA in mice after i.v. administration of PEG10kDa
ePheeGA (Phe), PEG10kDaePheePheeGA (PheePhe), PEG10kDaeGA (None) and GA-C at
a dose of 4 mg/kg at (A) 30 min, (B) 60 min and (C) 120 min.

Y. Ding et al. / Polymer 53 (2012) 1694e1702
1701
Table 2
AUC(0eN) (
m
g/mL min or
m
g/g min) of GA & relative exposure (re) and AUCi(0eN) (
m
g/min)a in mice plasma and organs after iv administration of PEG-GA conjugates or
Cremophor EL preparation (GAeC) at a dose of equivalent GA 4 mg/kg (n ¼ 5).
Organ
GAeC
PEG10,000ePheeGA
PEG10,000ePheePheeGA
PEG10,000eGA
AUC(0eN)
AUC(0eN)
re
AUCi(0eN)
AUC(0eN)
re
AUCi(0eN)
AUC(0eN)
re
AUCi(0eN)
re
AUCi(0eN)
Plasma
Liver
Lung
95.35
527.56
393.72
153.25
76.42
1
1
1
1
1
1
1
95.35
482.19
43.66
12.41
9.51
138.357
824.639
623.293
177.772
111.599
204.128
47.871
1.45
1.56
1.58
1.16
1.46
1.83
1.24
138.36
753.72
69.12
14.40
13.89
46.99
17.01
137.80
862.67
572.92
144.12
116.24
221.03
88.38
1.45
1.64
1.46
0.94
1.52
1.99
2.28
137.80
788.48
63.54
11.67
14.47
50.88
31.42
152.32
1062.65
383.53
158.61
103.75
217.99
47.52
1.60
2.01
0.97
1.04
1.36
1.96
1.23
152.32
971.09
42.53
12.85
12.92
50.18
16.89
heart
Spleen
Kidney
Brain
111.18
38.74
25.59
13.77
a
AUCi(0eN) ¼ AUC(0eN) ꢀ average volume of plasma or AUC(0eN) ꢀ average weight of organ.
linkage via the covalent bond. In addition, the reduced blood
clearance values (Cl) for conjugates demonstrated that the pro-
longed drug release and prevention of rapid renal filtration in the
circulation were achieved. In addition, an obvious increase, 2.5-fold
at least and 6-fold at most, of the area under the plasma
concentrationetime curve (AUC) in contrast with that of GAeC
demonstrated a remarkably enhanced bioavailability. From the
in vivo pharmacokinetic studies, it can be found that data of
apparent distribution volume (Vd) sharply reduced, only 5%e37% of
the control group value (GAeC). This phenomenon may contribute
to a reduction in peripheral toxicity, and implied that the bio-
distribution of GA molecule has been changed and governed by the
conjugation with the macromolecular carrier. Therefore, bio-
distribution studies were carried out to further evaluate the passive
target of PEGeGA conjugates.
concentrated in these tissues. The ratio of drug concentration in
liver to that in plasma was calculated to be in the range of 6e10, and
the spacer shows no obvious influence in this trend of
biodistribution.
In the assay duration from 30 to 120 min, GA concentration
detected in liver and lung after the administration of PEG conju-
gates showed twice as high as that of the control group GAeC.
In addition, higher AUC (area under the plasma or organ
concentrationetime curve) and re values (relative exposure
compared with GAeC) showed that, after conjugated to PEG
molecule, kidney filtration rate or overall clearance rate of GA were
reduced significantly (shown in Table 2). Actually, more than 70% of
the in vivo distribution of GA, revealed by the AUCi, was found in
liver, which demonstrated a liver-targeting property of the conju-
gate. Therefore, due to the tissue distribution profile of the poly-
meric prodrug system, more effective pharmacodynamic action for
liver cancer and reduced systemic toxicity of GA may be expected.
It is noted that, after 120 min, the concentration of GA for PEG-
GA ester sample remained at most 23-fold higher in liver than
plasma. It indicates slow metabolic rate of the ester in vivo and
subsequently accumulated in liver.
3.6. Biodistribution
Three conjugates (PEG10kDaePheeGA, PEG10kDaePheePheeGA,
and PEG10kDaeGA) were selected to test the tissue distribution
profiles. The histograms and relative parameters for GAeC,
PEGeGA ester, and PEG-spacereGA conjugates were showed in
Fig. 7 and Table 2, respectively. The results demonstrated that, after
i.v. administration for 30 min, GA was found in plasma and all
organs we detected. And a high GA concentration in liver and lung
was shown no matter what the prescription was, which can be due
to the accumulation of conjugates in the endothelial system
3.7. Cell cytotoxicity assay
The in vitro biological efficacy of PEG-GA conjugates was eval-
uated against HepG2 cells using the MTT method. Data in Fig. 8
showed that all of conjugates exhibited obvious cytotoxicity
against HepG2 cells, indicating polymeric prodrugs did not lose
anti-cancer activity of GA. However, compared with native drug
120
PEG 2kDa
(GA, IC50 ¼ 3.26
m
M), the IC50 values of conjugates (for instance, IC50
PEG 10kDa
of PEG10kDaeProeProeGA ¼ 117.66
mM) was as high as 36.1 times
100
PEG 20kDa
that of GA. It suggested a relatively lower cellar cytotoxicity induced
by the polymeric derivatization of the excellent biocompatibility of
PEG, amino acids, and dipeptide molecules. The cell cytotoxicity
was mainly due to the native GA released from the PEGylated
conjugates.
80
60
40
20
0
4. Conclusions
A series of PEG conjugates of the natural antitumor agent
gambogic acid with different amino acid and dipeptide spacers was
prepared. These PEG-GA conjugates showed satisfactory water
solubility compared with gambogic acid. Based on the results
in vitro, the molecular weight of polymeric carrier and the choice of
spacers had important influence in the solubility, drug content, and
drug release behavior of polymeric prodrugs. Moreover, in vivo
studies revealed that, employing the polymeric conjugation
strategy, the circulatory retention time, biodistribution, and
bioavailability of conjugates were remarkably improved, and liver
targeting character has been achieved. These results demonstrate
Spacer
Fig. 8. In vitro cytotoxicity of PEGylated conjugates of GAa against HepG2 cell lines
after 24 h incubation (IC50
24 h incubation was 3.26
, m
mol/L). aIC50 of native drug GA against HepG2 cells after
mmol/L according to our MTT assay.

1702
Y. Ding et al. / Polymer 53 (2012) 1694e1702
[11] You QD, Guo QL, Ke X, Xiao W, Dai LL, Lin Y, CN Patent No. 03132386.3 2003.
[12] Han JY, Xiao J, Wang H, Chang HY, Ma PS. J Chem Ind Eng 2001;52:64e7.
[13] Zhu X, Zhang C, Wu XL, Tang XY, Ping QN. Drug Dev Ind Pharm 2008;
34:2e9.
[14] Qu GW, Zhu X, Zhang C, Ping QN. Drug Deliv 2009;16:363e70.
[15] Harris JM, Zalipsky S. Poly(ethylene glycol): chemistry and biological appli-
cations. Plenum Press, New York, 0-306-44078-4/408.
[16] Wang T, Wang YC, Yan LF, Wang J. Polymer 2009;50:5048e54.
[17] Knop K, Hoogenboom R, Fischer D, Schubert US. Angew Chem Int Ed 2010;49:
6288e308.
[18] Greenwald RB, Choe YH, McGuire J, Conover CD. Adv Drug Deliver Rev 2003;
55:217e50.
[19] Pasut G, Veronese FM. Prog Polym Sci 2007;32:933e61.
[20] Zhan FX, Chen W, Wang ZJ, Lu WT, Cheng R, Deng C, et al. Biomacromolecules
2011;12:3612e20.
the polymeric conjugation method, by rational design and
component selection, would solve many problems of insoluble
chemical and natural anti-cancer agents, and at the same time
achieve excellent properties of drug delivery systems.
Acknowledgements
This work is financially supported by the 111 Project from the
Ministry of Education of China and the State Administration of
Foreign Expert Affairs of China (Grant No. 111-2-07), Fundamental
Research Funds for the Central Universities of China (Grant No.
JKY2009016, JKQ2009026, and JKP2011008), and the Natural
Science Foundation of China (Grant No. 30900337).
[21] Alani AWG, Bae Y, Rao DA, Kwon GS. Biomaterials 2010;31:1765e72.
[22] Pasut G, Canal F, Via LD, Arpicco S, Veronese FM, Schiavon O. J Control Release
2008;127:239e48.
[23] Greenwald RB, Pendri A, Conover C, Gilbert C, Yang R, Xia J. J Med Chem 1996;
39:1938e40.
References
[24] Cavallaro G, Pitarresi G, Licciardi M, Giammona G. Bioconjug Chem 2001;12:
143e51.
[25] Silva M, Ricelli NL, Seoud OE, Valentim CS, Ferreira AG, Sato DN, et al. Arch
Pharm Chem Life Sci 2006;339:283e90.
[26] Min T, Ye H, Zhang P, Liu J, Zhang C, Shen WB, et al. J Appl Polym Sci 2009;111:
444e51.
[27] Keller O, Keller WE, van Look G, Wersin G. Org Synth 1985;63:160e6.
[28] Han QB, Cheung S, Tai J, Qiao CF, Song JZ, Xu HX. Biol Pharm Bull 2005;28:
2335e7.
[1] Khandare J, Minko T. Prog Polym Sci 2006;31:359e97.
[2] Greenwald RB. J Control Release 2001;74:159e71.
[3] Minko T. Drug Discov Today Technologies 2005;2:15e20.
[4] Autherhoff H, Frauendorf H, Liesenklas W, Schwandt C. Arch Pharm 1962;295:
833e46.
[5] Ollis WD, Ramsay MVJ, Sutherland IO, Mongkolsuk S. Tetrahedron 1965;21:
1453e70.
[6] Cai SX, Jiang SC, Zhang HZ. PCT Int Appl 2005;51. pp. CODEN: PIXXD2 WO
2005060663 A2 20050707.
[7] Zhao L, Guo QL, You QD, Wu ZQ, Gu HY. Biol Pharm Bull 2004;27:998e1003.
[8] Liu YT, Hao K, Liu XQ, Wang GJ. Acta Pharmacol Sin 2006;27:1253e8.
[9] Hao K, Liu XQ, Wang GJ, Zhao XP. J Drug Metab Ph 2007;32:63e8.
[10] Dai JG, CN Patent No. 03131511.9 2003.
[29] Liu WY, Feng F, Chen YS, You QD, Zhao SX. Chin J Nat Med 2004;12:75e7.
[30] Zhang P, thesis: PEG prodrug studies on natural anti-tumor active ingredient
2008.
[31] Greenwald RB, Pendri A, Bolikal D. J Org Chem 1995;60:331e6.