Synthesis and in vitro evaluation of potential sustained release prodrugs via targeting ASBT

Article abstract of DOI:10.1016/j.ijpharm.2010.06.039

The objective was to synthesize prodrugs of niacin and ketoprofen that target the human apical sodium-dependent bile acid transporter (ASBT) and potentially allow for prolonged drug release. Each drug was conjugated to the naturally occurring bile acid ch

Full text of DOI:10.1016/j.ijpharm.2010.06.039

International Journal of Pharmaceutics 396 (2010) 111–118
Contents lists available at ScienceDirect
International Journal of Pharmaceutics
journal homepage: www.elsevier.com/locate/ijpharm
Synthesis and in vitro evaluation of potential sustained release prodrugs via
targeting ASBT
Xiaowan Zheng, James E. Polli∗
Department of Pharmaceutical Sciences, School of Pharmacy, University of Maryland, Baltimore, MD 21201, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The objective was to synthesize prodrugs of niacin and ketoprofen that target the human apical sodium-
dependent bile acid transporter (ASBT) and potentially allow for prolonged drug release. Each drug was
conjugated to the naturally occurring bile acid chenodeoxycholic acid (CDCA) using lysine as a linker. Their
inhibitory binding and transport properties were evaluated in stably transfected ASBT-MDCK monolay-
ers, and the kinetic parameters Ki, Kt, normJmax, and Pp were characterized. Enzymatic stability of the
conjugates was evaluated in Caco-2 and liver homogenate. Both conjugates were potent inhibitors of
ASBT. For the niacin prodrug, substrate kinetic parameter Kt was 8.22 ␮M and normJmax was 0.0917. In
4 h, 69.4% and 26.9% of niacin was released from 1 ␮M and 5 ␮M of the conjugate in Caco-2 homogenate,
respectively. For the ketoprofen prodrug, Kt was 50.8 ␮M and normJmax was 1.58. In 4 h, 5.94% and 3.73%
of ketoprofen was released from 1 ␮M and 5 ␮M of the conjugate in Caco-2 homogenate, and 24.5% and
12.2% of ketoprofen was released in liver homogenate, respectively. In vitro results showed that these
bile acid conjugates are potential prolonged release prodrugs with binding affinity for ASBT.
© 2010 Elsevier B.V. All rights reserved.
Received 12 April 2010
Received in revised form 3 June 2010
Accepted 18 June 2010
Available online 1 July 2010
Keywords:
Bile acid transporter
MDCK cells
Prolonged release
Ketoprofen
Niacin
Stability
1. Introduction
2000; Stern et al., 1992). The conjugation metabolic pathway has
low affinity and high capacity and results in the formation of glycine
The human apical sodium-dependent bile acid transporter
(ASBT, SLC10A2) is expressed at high levels in the terminal ileum
where it mediates bile acid recovery (Dawson et al., 2003). ASBT is
critical for intestinal reabsoption of bile acid in the enterohepatic
recirculation. ASBT’s high affinity and capacity for native bile acids
affords it to be a potential prodrug target for intestinal drug absorp-
tion (Balakrishnan and Polli, 2006; Tolle-Sander et al., 2004), as well
as a contributing factor to prolong drug release via enterohepatic
recirculation. Previously, bile acid conjugates or bile acid derived
compounds showed potential to provide sustained systemic con-
centrations of the drug after oral administration (Gallop and Cundy,
2002a,b). Prolonged drug release from the drug conjugated to a bile
acid may be achieved via enterohepatic recirculation, where parent
drug is slowly released into the systemic circulation from the bile
acid conjugate.
conjugates of niacin which has been associated with flushing. The
cutaneous flushing is mediated by a G-protein-coupled receptor
(GPR109A) effect involves release of prostaglandin (PGE2, PGD2)
(Benyo et al., 2005; Pike, 2005). The amidation pathway has high
affinity and low capacity. When the amidation pathway is satu-
rated, niacin is mainly metabolized by the conjugative pathway,
resulting in flushing. The pharmacokinetic and pharmacodynamic
profiles of different niacin formulations reflect differences in their
absorption rates and metabolic depositions. Slower niacin absorp-
tion causes lower flushing as most of the drug undergoes amidation
(Pieper, 2002, 2003). Extended-release niacin has demonstrated
reduced rates of flushing with a tmax of 4–5 h, compared to the
immediate-release niacin of 30–60 min (Knopp et al., 1998; Morgan
et al., 1998).
Ketoprofen is a potent non-steroidal anti-inflammatory drug
(NSAID). It is associated with higher gastrointestinal (GI) bleed-
ing/perforation compared to other NSAIDs (Gonzalez et al., 2010).
It has a short half-life of 2 h, requiring frequent doses to maintain
therapeutic efficacy (Kantor, 1986).
The objective of this study is to design respective prodrugs
of niacin and ketoprofen by conjugating them with bile acid and
targeting ASBT for absorption. Prolonged availability of the par-
ent drug may be expected if prodrug is slowly hydrolyzed during
enterohepatic circulation. This approach could reduce flushing
from niacin and extend the pharmacological effect of ketoprofen.
Moreover, the approach could minimize ketoprofen GI side effects
Niacin (nicotinic acid) has been used in the management of
hyperlipidemia for over 50 years. Clinical use of niacin has been
limited by a relatively high incidence of cutaneous flushing. Niacin
undergoes extensive first-pass metabolism in the liver via the con-
jugation and the amidation pathways (Menon et al., 2007; Piepho,
∗
Corresponding author at: Department of Pharmaceutical Sciences, School of
Pharmacy, University of Maryland, 20 Penn Street, HFS2, Room 623, Baltimore, MD
21201, United States. Tel.: +1 410 706 8292; fax: +1 410 706 5017.
E-mail address: jpolli@rx.umaryland.edu (J.E. Polli).
0378-5173/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpharm.2010.06.039

112
X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
by preventing the parent drug direct contact with the GI tract. In
this study, niacin and ketoprofen were conjugated with the nat-
urally occurring bile acid chenodeoxycholate (CDCA) via lysine as
a linker. CDCA was chosen due to its high ASBT binding affinity
(Balakrishnan et al., 2006b). In vitro results indicated that both
conjugates were potential prodrugs with affinity for ASBT.
of fractions in TLC and MS that contained the product were collected
and evaporated to give a white fluffy solid compound 3.
The neutral compound 3 was subjected to catalytic hydrogena-
tion for the removal of the benzyl and CBZ protection groups using
10% Pd/charcoal in ethanol at 50 psi overnight. Suspension was fil-
tered through Celite® and solvent was evaporated under vacuum
to give white solid intermediate 4. MS showed an appropriate peak
[M+H]+ 521.4.
2. Materials and methods
2.2.2. Synthesis of CDCA–lysine–niacin (6)
2.1. Materials
Compound 4 (0.5 g) was coupled to 2.0 equiv. of nicotinoyl
chloride hydrochloride 5 in 5 mL of anhydrous DMF along with
3.0 equiv. of TEA. The reaction was conducted at 4 ◦C for 4 h. The
reaction was terminated by adding 30 mL of water. The reaction
mixture was extracted with EtOAc (3×). The combined organic
layer was washed with 1N HCl (3×) and water (3×). Then organic
layer was extracted with 1N NaOH (3×). The combined water layer
was washed with EtOAc (3×). HCl solution (37%) was added drop
by drop to the aqueous layer until the pH was about 2. Then the
solution was extracted with EtOAc (3×) again and the organic layer
was washed with brine (1×). The organic extract was dried with
anhydrous sodium sulfate, filtered, and evaporated under vacuum
to yield the final compound 6. MS showed an appropriate peak
[M+H]+ 626.3.
[3H]-Taurocholic acid was purchased from PerkinElmer
(Waltham, MA). Taurocholic acid, nicotinoyl chloride hydrochlo-
ride, and ketoprofen were purchased from Sigma (St. Louis, MO).
Protected lysine analogs were from Novabiochem (Gibbstown, NJ).
Fetal bovine serum (FBS), trypsin, and Dulbecco’s modified Eagle’s
medium (DMEM) were purchased from Invitrogen Corporation
(Carlsbad, CA). CDCA was obtained from TCI America (Portland,
OR). Pooled IGS Sprague–Dawley male rat liver S9 fraction was
purchased from Xenotech (Lenexa, KS).
2.2. Synthetic procedures
Among the native bile acids, CDCA was selected on the basis of its
favorable transport across ASBT-MDCK monolayers (Balakrishnan
et al., 2006b). Pyridine-3-carboxylic acid and ketoprofen were con-
jugated to CDCA using lysine as the linker. Synthetic approaches of
both prodrugs are shown in Scheme 1.
2.2.3. Synthesis of CDCA–lysine–ketoprofen (9)
Ketoprofen 7 (5 g) was activated using 20% thionyl chloride in
10 mL of toluene with a few drop of DMF as the catalyst at RT for
30 min. The reaction was terminated by evaporating the solvent
under vacuum to yield a white solid 8. The resulting compound
8 was coupled to 0.5 equiv. of 6 in anhydrous DMF along with
3.0 equiv. of TEA at RT overnight. The reaction was terminated by
adding 30 mL of water. The reaction mixture was extracted and
purified as described above (see synthesis of CDCA–lysine–niacin).
The organic extract was dried with anhydrous sodium sulfate, fil-
tered, and evaporated under vacuum to yield the final compound
9. MS showed an appropriate peak [M+H]+ 757.5.
2.2.1. Synthesis of CDCA–lysine (4)
First, 5 g of CDCA 2 were coupled to 0.5 equiv. of H-lys(Z)-OBzl
HCl 1 using HBTU (1.1 equiv.) and HOBT (0.5 equiv.) as the coupling
reagent in 10 mL of anhydrous DMF, along with 2.2 equiv. of triethy-
lamine (TEA), at 60 ◦C for overnight. The reaction was terminated
by adding 30 mL of water. The reaction mixture was extracted with
ethyl acetate (EtOAc) (3×). The combined organic layer was washed
with 1N HCl (3×), 1N NaOH (3×), and brine (1×). The organic extract
was dried with anhydrous sodium sulfate, filtered, and evaporated
under vacuum to yield orange oil. The crude product was further
purified by silica gel column chromatography using a mobile phase
of EtOAc and hexane (80:20). Fractions were collected. Mass spec-
trometry (MS) showed an appropriate peak [M+H]+ 745.3. Solvent
2.3. Characterization of the prodrugs
Identity and purity were characterized by TLC, LC–MS, 1H NMR
(proton, COSY), and LC–MS/MS. The mass spectra gave a molecular
Scheme 1. General synthetic approach to obtain CDCA–lysine–niacin (6) and CDCA–lysine–ketoprofen (9).

X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
113
Table 1
Structure and NMR results of CDCA–lysine–niacin and CDCA–lysine–ketoprofen.
.
Conjugate
R1
1H NMR (DMSO-d6)a
CDCA–lysine–niacin
11.95 (1H, s, ␥), 8.99 (1H, s, ␣), 8.69 (1H, d, ␣), 8.64 (1H, t,
␤), 8.17 (1H, d, ␣), 7.92 (1H, d, ␤), 7.48 (1H, t, ␣), 4.30 (1H,
m, ␪),
CDCA–lysine–ketoprofen
11.65 (1H, s, ␥), 7.98 (1H, d, ␤) 8.04 (4, t, ␤) 7.78-7.58 (9H,
␣), 4.13(1H, m, ␥) 3.68 (1H, m, ␣), 1.37 (3H, s, ␣)
a
Only characteristic peaks are shown. CDCA and methylene peaks are not shown. Values are in ppm relative to TMS. ␣ indicates assignment to proton peaks in R1; ␤
indicates assignment to protons in amide; ␥ indicates assignment to protons in carboxylic acid region; ␪ indicates the proton in ␣ carbon.
ion in agreement with the calculated mass of the prodrug. No CDCA
impurity was detected in any final conjugate (Gonzalez and Polli,
2008). 1H NMR results are compiled in Table 1. The chemical shifts
are reported in parts per million (ppm) relative to tetramethylsilane
(TMS) (ı = 0.0 ppm). 1H NMR spectra were recorded on a Varian
Inova 500 MHz (Varian Inc., Palo Alto, CA).
2.5.2. Stability study
To begin the stability assay, each prodrug was added into a
centrifuge tube containing 1000 ␮L Caco-2 cell homogenate or
liver S9 fraction (20 mg/mL) to yield a prodrug concentration of
1 ␮M or 5 ␮M. Tubes were placed in a 37 ◦C incubator and shaken.
Samples (100 ␮L) were removed at selected time points over 4 h
and immediately added to a new centrifuge tube prefilled with
100 ␮L ice-cold acetonitrile to precipitate cellular protein and stop
enzymatic activity. Samples were centrifuged at 12,000 rpm for
30 min at 4 ◦C. The supernatant was carefully transferred to another
centrifuge tube. The samples were stored at −80 ◦C until further
analysis.
2.4. Cell culture
Stably transfected ASBT-MDCK were cultured as described
(Balakrishnan et al., 2005). Briefly, cells were grown at 37 ◦C, 90%
relative humidity, and 5% CO2 atmosphere and fed every 2 days by
the DMEM supplemented with 10% FBS, 50 units/mL penicillin, and
50 ␮g/mL streptomycin. Geneticin was used at 1 mg/mL to main-
tain selection pressure. Cells were passaged every 4 days or after
reaching 90% confluence.
2.6. ASBT biological activity assessment
Each prodrug was subjected to ASBT inhibition and uptake stud-
ies.
Caco-2 cells were maintained in DMEM with 10% fetal bovine
serum, 1% nonessential amino acid solution, 1% antibiotic solution
(penicillin/streptomycin), and 1% glutamine solution and grown in
flasks. Cells were incubated at 37 ◦C with 5% CO2. The medium was
changed every 2 days.
2.6.1. Inhibition of taurocholate uptake by prodrugs
To characterize prodrug binding affinity to ASBT, cis-inhibition
studies of taurocholate uptake by prodrugs were conducted as pre-
viously described (Zheng et al., 2009). Briefly, cells were exposed
to donor solution containing 2.5 ␮M taurocholate (including [3H]-
taurocholate) in the presence of a prodrug at eight concentrations
(1–100 ␮M) for 10 min at 37 ◦C. At the end of the study, cell
lysate was counted for taurocholate radioactivity using a liq-
uid scintillation counter (Beckmann Instruments, Inc., Fullerton,
CA).
2.5. Enzymatic stability study
Prodrug was subjected to Caco-2 homogenate and liver S9 frac-
tion. Intact prodrug and the liberated drug were quantified.
2.6.2. Uptake of prodrug into ASBT-MDCK
2.5.1. Caco-2 homogenate
To characterize prodrug substrate properties across ASBT,
uptake studies of prodrugs were conducted as described (Zheng et
al., 2009). Briefly, cells were exposed to donor solution containing
eight different prodrug concentrations (1–500 ␮M) for 10 min at
37 ◦C. The studies were conducted in both HBSS buffer or modified
HBSS buffer (i.e. sodium-free). Since ASBT is a sodium-dependent
transporter, studies using sodium-free buffer enabled passive per-
meability measurement of the prodrug. Subsequently, to quantify
uptake, cells were lysed as pervious described by adding 0.25 mL
acetonitrile to each well (Rais et al., 2008). Following evaporation
Caco-2 cells grown in 225-cm2 culture flask for 21–25 days were
washed twice with ice-cold phosphate-buffered saline (PBS). The
cells were scraped off in 12 mL ice-cold PBS and the scraped cells
were transferred to a centrifuge tube. Tubes were stored at −80 ◦C
overnight. The next day, tubes were quickly thaw at 37 ◦C (leave
some ice), sonicated the cells twice at 1 s pulses with a sonicator
probe (F60 Sonicator, Thermo Fisher Scientific Inc., Waltham, MA).
The cell homogenate (protein content = 0.63 mg/mL) was stored at
−80 ◦C (Tantishaiyakul et al., 2002).

114
X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
Table 2
LC–MS/MS methods of analyzing CDCA–lysine–niacin and CDCA–lysine–ketoprofen and the parent drugs niacin and ketoprofen.
Compound
Exact massa
625.41
Q1 mass (m/z)b
Q3 mass (m/z)
Retention time (min)
1.26
Gradient timetable (B%)c
CDCA–lysine–niacin
626.40
189.12
0–0.2 (45%), 0.2–3.0 (45–95%),
3.0–4.0 (95%), 4.0–4.2
(95–45%), and 4.2–7.0 min
(45%)
Niacin
123.03
756.47
124.06
757.50
80.15
1.02
4.77
0–3.1 (5%), 3.1–3.2 (5–90%),
3.2–4.2 (90%), 4.2–4.5
(90%-5%), and 4.5–8.0 min (5%)
0–0.2 (40%), 0.2–3.0 (40–95%),
3.0–4.0 (95%), 4.0–4.2
(95–40%), and 4.2–7.0 min
(40%)
CDCA–lysine–ketoprofen
485.30
Ketoprofen
254.09
255.10
198.36
2.36
0–0.2 (40%), 0.2–3.0 (40–95%),
3.0–4.0 (95%), 4.0–4.2
(95–40%), and 4.2–7.0 min
(40%)
a
Calculated monoisotopic mass.
b
Q1 and Q3 mass are experimental mass [M+H]+.
c
The mobile phase was A, 0.1% formic acid in water; and B, 0.1% formic acid in acetonitrile. The gradient times were listed as the percent organic phase, B%.
at RT, 1 mL solution containing 50% water and 50% acetonitrile was
added to each well for 10 min. Sample was stored in −80 ◦C until
analyzed by LC–MS/MS.
developed modified Michaelis–Menten model (Eqs. (2) and (3))
(Balakrishnan et al., 2007):
PABL((Jmax/(Kt + S)) + Pp)
J =
J =
S
(2)
(3)
PABL + ((Jmax/(Kt + S)) + Pp)
2.7. Analytical methods
PABLPpS
Samples from uptake studies were quantified by LC–MS/MS
on a Thermo Finnigan Surveyor HPLC system, equipped with a
Thermo Finnigan Surveyor autosampler and a Thermo TSQ quan-
tum mass spectrometer (Thermo Fisher Scientific Inc., Waltham,
MA). For both prodrugs and parent ketoprofen, the column was a
Phenomenex Luna C8 (50 mm × 2.0 mm, 3 ␮m; Phenomenex; Tor-
rance, CA). For parent niacin, the column was a Thermo Aquasil C18
(50 mm × 2.1 mm, Thermo Fisher Scientific, MA). Columns were
heated to 45 ◦C. The mobile phase was A, 0.1% formic acid in
water; and B, 0.1% formic acid in acetonitrile, with a flow rate
of 0.4 mL/min. The injection volume was 10 ␮L. Detection was
achieved under positive ion electrospray [M+H]+ tandem mass
spectrometry, as positive mode provided the greatest sensitivity.
The mobile phase and retention time for both conjugates and parent
drugs are detailed in Table 2.
PABL + Pp
where J is prodrug uptake. Eqs. (2) and (3) were applied simul-
taneously to data measured with sodium and without sodium to
estimate Kt, Jmax, and Pp. Data were analyzed using WinNonlin
Professional.
3. Results
3.1. Prodrug synthesis
Bile acid conjugates CDCA–lysine–niacin and CDCA–lysine–
ketoprofen were successfully synthesized. Their structures are
depicted in Table 1. For each, the synthetic strategy was based
on coupling the carboxyl group at the C-24 position of CDCA
(Scheme 1) to the carboxyl group of ketoprofen or niacin using
lysine as a linker. Prodrug properties were characterized using NMR
and LC–MS, and the resulting data are shown in Tables 1 and 2,
respectively. Both LC–MS and NMR data were consistent with the
target structures. TLC showed a single spot.
2.8. Kinetic analysis
Inhibition data was analyzed in terms of inhibition constant
Ki as previously described (Zheng et al., 2009), where a modified
Michaelis–Menten competitive inhibition model that considers
aqueous boundary layer (ABL) resistance was used (Balakrishnan
et al., 2007).
3.2. ASBT inhibition and uptake
Inhibition of taurocholate uptake was carried out for each
prodrug using stably transfected ASBT-MDCK monolayers. Fig. 1
illustrates the concentration-dependent inhibition profile by
CDCA–lysine–niacin and niacin (Fig. 1A), and CDCA–lysine–
ketoprofen and ketoprofen (Fig. 1B), respectively. Taurocholate
uptake was reduced over 2-fold in the presence of 50 ␮M of each
prodrug. Neither ketoprofen nor niacin was ASBT inhibitor. The Ki
value indicated that each prodrug was a potent inhibitor and is
listed in Table 3, with Ki of 12.9 ␮M for niacin prodrug and 15.6 ␮M
for ketoprofen prodrug, respectively. Compared to the Ki range
of native bile acids from 0.427 ␮M to 22.6 ␮M, with taurocholate
Ki = 4.96 ␮M, both prodrugs were potent inhibitors and within the
range of native bile acids (Balakrishnan et al., 2006b).
PABL(Jmax/(Kt(1 + (I/Ki))) + S) + Pp
J =
S
(1)
PABL + (Jmax/(Kt(1 + (I/Ki))) + S) + Pp
where
J
is taurocholate uptake, Jmax and Kt are the
Michaelis–Menten constants for ASBT-mediated transport,
S
is taurocholate concentration, Pp is the passive taurocholate
permeability, I is the inhibitor concentration, Ki is the inhibition
constant, and PABL is the aqueous boundary layer permeability
(i.e. 150 × 10−6 cm/s) (Balakrishnan et al., 2007). Kt was 5.03 ␮M,
obtained from a pooled data analysis approach. Jmax was measured
from taurocholate uptake studies using 200 ␮M of taurocholate; Pp
was estimated from taurocholate uptake studies in the absence of
sodium. Ki was calculated using WinNonlin Professional (Pharsight
Corporation; Mountain View, CA). Jmax and Pp of taurocholate were
measured on each study occasion.
Given favorable binding affinity for ASBT, prodrugs were
evaluated for ASBT-mediated uptake. Fig.
2 illustrates the
For uptake studies, uptake data into monolayers was analyzed
in terms of substrate affinity Kt and was fitted to a previously
concentration-dependent transport of CDCA–lysine–niacin and
CDCA–lysine–ketoprofen into ASBT-MDCK monolayers. Table 3

X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
115
Table 4
Niacin and ketoprofen released from CDCA–lysine–niacin and CDCA–lysine–
ketoprofen, respectively in Caco-2 and liver S9 after 4 h.
Caco-2
69.4%
Liver S9
a
Percent niacin released in 1 ␮M
CDCA–lysine–niacin
–
Percent niacin released in 5 ␮M
CDCA–lysine–niacin
Percent ketoprofen released
from 1 ␮M
26.9%
5.94%
–
24.5%
CDCA–lysine–ketoprofen
Percent ketoprofen released
from 5 ␮M
3.73%
12.2%
CDCA–lysine–ketoprofen
a
Cannot be measured due to a high baseline niacin level in rat liver S9 fraction.
lists the kinetic uptake parameters for both compounds. Niacin pro-
drug had a high substrate binding affinity with a Kt = 8.22 ␮M, but
the transport capacity is significantly reduced compared to tau-
rocholate, such that normJmax of niacin prodrug was only 0.0917.
normJmax was obtained by dividing prodrug Jmax by taurocholate
Jmax. Therefore, ASBT’s capacity to transport niacin prodrug is only
about 9% of that of taurocholate.
Ketoprofen exhibited a more modest substrate binding affin-
ity of Kt = 50.8 ␮M, which is 10-fold higher than the taurocholate
Kt = 5.03 ␮M. However, the transport capacity of this prodrug was
1.87-fold of the capacity for taurocholate. The passive permeability
of CDCA–lysine–ketoprofen was 6.38 × 10−6 cm/s, which was 20-
fold higher compared to CDCA–lysine–niacin of 0.367 × 10−6 cm/s.
This difference reflects the greater lipophilicity afforded by the
ketoprofen moiety (Log P = 3.31), compared to the niacin moiety
(Log P = 0.25). When [S] = 1/2Kt, passive permeability accounts for
10.3% for total niacin prodrug uptake, and 44.1% for ketoprofen
prodrug.
Fig. 1. Concentration-dependent inhibition of taurocholate uptake into
ASBT-MDCK monolayers by (A) CDCA–lysine–niacin (᭹) and niacin (ꢀ); (B)
CDCA–lysine–ketoprofen (᭹) and ketoprofen (ꢀ) for 10 min. Circles indicate
observed data points, while the solid line indicates model fit. Taurocholate uptake
into ASBT-MDCK cells was reduced significantly by both conjugates, where Ki was
12.9 ␮M for niacin prodrug and 15.6 ␮M for ketoprofen prodrug. Taurocholate
uptake was not reduced by niacin or ketoprofen.
3.3. Enzymatic stability of prodrugs and drug release
Each prodrug was designed with the employment of two amide
linkages, which are stable in acid and base. Chemical stability of
prodrugs was evident from successful synthesis, which entailed
extracting prodrug using strong acidic and basic solutions; no
breakdown to potential degradants was observed in MS. Therefore,
general chemical stability of the prodrugs was concluded.
Prodrug hydrolysis can occur in many locations during
enterohepatic circulation. In order to assess prodrug hydrolysis,
enzymatic stability was conducted in Caco-2 homogenate and
rat liver homogenate, using prodrugs concentrations of 1 ␮M and
5 ␮M. Time-dependent prodrug degradation and corresponding
parent drug release were monitored. Table 4 summarizes the per-
centage of conversion of prodrug to the parent drug after 4 h.
3.3.1. Niacin prodrug
In Fig. 3, the niacin prodrug decreased with time in Caco-2
homogenate, while niacin increased with time. Under the same
conditions, 5 ␮M prodrug displayed a slower hydrolysis profile
compared to 1 ␮M, with 26.9% and 69.4% of niacin released in 4 h,
respectively. The mass balance during niacin prodrug hydrolysis
was over 100% for the last two sample times (i.e. 180 min and
Fig. 2. Concentration-dependent uptake of CDCA–lysine–niacin (A) and
CDCA–lysine–ketoprofen (B) into ASBT-MDCK monolayers for 10 min. The
conjugate was measured at varying concentrations in the presence and absence of
sodium to delineate ASBT-mediated and passive uptake components. Data indicate
total uptake in the presence of sodium (᭹) and passive uptake in the absence of
sodium (ꢀ), while the solid line indicates model fit of the total uptake.
Table 3
ASBT inhibition and uptake kinetics of CDCA–lysine–niacin and CDCA–lysine–ketoprofen.
a
Conjugate
Ki (␮M)
Kt (␮M)
normJmax
Pp × 106(cm/s)
CDCA–lysine–niacin
CDCA–lysine–ketoprofen
12.9 ( 2.6)
15.6 ( 2.7)
8.22 ( 3.63)
50.8 ( 23.2)
0.0917 ( 0.0102)
1.87 ( 0.36)
0.367 ( 0.023)
6.38 ( 0.44)
Data were summarized as mean ( SEM) of three measurements.
a
To accommodate variation in ASBT expression levels across studies, Jmax of each bile acid was normalized against taurocholate Jmax from the same study yielding normJmax
.

116
X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
Fig. 3. Hydrolysis of CDCA–lysine–niacin in Caco-2 homogenate. Data indicate the
time course of disappearance of 1 ␮M (ꢀ) and 5 ␮M (ꢁ) of CDCA–lysine–niacin and
appearance of niacin from 1 ␮M (᭹) and 5 ␮M CDCA–lysine–niacin (ꢀ).
Fig. 5. Hydrolysis of CDCA–lysine–ketoprofen in Caco-2 homogenate. In panel A,
data indicate the time course of the disappearance of 1 ␮M (ꢀ) and 5 ␮M (ꢁ) of
CDCA–lysine–ketoprofen and the appearance of ketoprofen from 1 ␮M (᭹) and 5 ␮M
(ꢀ) CDCA–lysine–ketoprofen. In panel B, ketoprofen appearance from 1 ␮M (᭹) and
5 ␮M (ꢀ) prodrug is re-plotted on a smaller scale.
Fig. 4. Hydrolysis of CDCA–lysine–niacin in rat liver S9 fraction. Data indicate the
time course of the disappearance of 1 ␮M of CDCA–lysine–niacin (ꢀ).
240 min). The error was possibly caused by evaporation via incuba-
tion at 37 ◦C over this time frame, which involved several sample
withdraws.
LC–MS/MS showed a high baseline level of niacin (i.e. nico-
tinic acid) in rat liver S9 fraction. Therefore, the generation profile
of niacin in liver homogenate was not able to be measured.
Fig. 4 shows hydrolysis profile of 1 ␮M niacin prodrug in liver
homogenate. About 30.3% of the prodrug was degraded after 4 h,
which was slightly less than the 41.7% degradation in Caco-2
homogenate.
with a recovery efficiency over 97% (Dawson et al., 2009; Dawson
and Oelkers, 1995; Wong et al., 1996).
In order to be efficiently enterohepatically recycled, high hep-
atic uptake and high secretion into bile are also critical which also
3.3.2. Ketoprofen prodrug
Figs. 5A and 6A show the stability profile of the ketoprofen pro-
drug in Caco-2 and liver homogenate, respectively. Like the niacin
prodrug, under identical conditions, the hydrolysis rate of 5 ␮M
prodrug was slower than that of 1 ␮M prodrug. Table 4 shows
only 5.94% and 3.73% of ketoprofen was released from 1 ␮M to
5 ␮M of prodrug in Caco-2 homogenate after 4 h; while 24.5%
and 12.2% was released in liver homogenate. The time-dependent
appearance of ketoprofen is re-plotted using a smaller scale in
Figs. 5B and 6B. The prodrug demonstrated high metabolic sta-
bility in Caco-2 homogenate. Similarly, conversion of prodrug to
ketoprofen rate was slower in liver homogenate.
4. Discussion
4.1. Enterohepatic circulation of bile acid conjugates
The prodrugs were designed to be absorbed at the terminal
ileum by the high capacity and high affinity transporter ASBT, fol-
lowed by a slow release of the parent drug via enzymatic hydrolysis
during enterohepatic circulation. Bile acid enterohepatic circula-
tion is a very efficient system where 3–5 g bile acid pool cycles six
times daily and results in a turnover of 12–18 g of bile acid each day
in humans. Only less than 0.5 g bile acid is lost in feces each day,
Fig. 6. Hydrolysis of potential prodrug CDCA–lysine–ketoprofen in rat liver
homogenate. In panel A, data points indicate the time course of the disappearance
of 1 ␮M (ꢀ) and 5 ␮M (ꢁ) of CDCA–lysine–ketoprofen and the appearance ketopro-
fen from 1 ␮M (᭹) and 5 ␮M (ꢀ) CDCA–lysine–ketoprofen. In panel B, ketoprofen
appearance from 1 ␮M (᭹) and 5 ␮M (ꢀ) prodrug is re-plotted on a smaller scale.

X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
117
involve transporters. Hepatic bile acid uptake is mainly mediated by
sodium/taurocholate co-transporting polypeptide (NTCP), as well
as members of the sodium independent organic anion-transporting
polypeptides (OATPs) family (i.e. OATP-A, OATP-C, OATP-8) (St-
Pierre et al., 2001). Bile salt exports pump (BSEP) and multidrug
resistance protein (MRP2) transport bile acid into the bile canalicu-
lus (Alrefai and Gill, 2007). MPR4 and MPR3 are able to remove
bile acid through the sinusoidal surface (Alrefai and Gill, 2007).
These bile acid transporters generally possess broader substrate
specificity than ASBT. As a result, a bile acid conjugate that is well
recognized by ASBT has high potential to be a substrate of liver bile
acid transporters.
Although numerous transporter proteins are involved in the bile
acid transport system, ASBT knockout mice exhibit 10–20-fold ele-
vated fecal bile acid levels and a 80% decrease in bile acid pool size
(Dawson et al., 2003). In humans, some inherited ASBT mutations
cause primary bile acid malabsorption (Montagnani et al., 2001;
Oelkers et al., 1997; Wong et al., 1995). These results indicate ASBT
is critical for bile acid intestinal reabsoprtion. Bile acid–drug con-
jugates that utilize ASBT have potential to be absorbed from the
intestine to the liver and excreted into the bile, to achieve a cir-
culating reservoir of conjugated drugs. The reservoir may slowly
release drug to provide sustained systemic drug concentrations.
of a HMG-CoA reductase inhibitor were active against HMG-CoA
reductase in rat liver after an oral administration (Kramer et al.,
1994). Also, a bile acid conjugate increased acyclovir bioavailability
by 2-fold in rat (Tolle-Sander et al., 2004).
In the present study, both prodrugs were ASBT substrates, with
high affinity and low capacity for the niacin prodrug and moderate
affinity and high capacity for the ketoprofen prodrug. Transporter-
mediated translocation is a complex process. It minimally requires
substrate binding to the transporter, translocation of substrate
via transporter conformation changes, dissociation of substrate,
and transporter conformation restoration. Additional steps involve
ASBT cotransport of sodium. Regarding the marked difference in
transport capacity between two prodrugs, any step in the trans-
porting process could be rate limiting, and consequently limit
transporter capacity. However, since both prodrugs were potent
inhibitors, substrate binding to the transporter was not the rate-
limiting step, even for the niacin prodrug, for which ASBT exhibited
a lower capacity.
An appropriate hydrolysis rate would be critical for prolonged
release of drug from prodrug. If hydrolysis is too rapid, then
prolonged release is not achieved. If hydrolysis is too slow, ther-
apeutic concentrations of drug are not obtained. In vitro results
suggest slow hydrolysis of niacin prodrug in Caco-2 and liver
homogenates. Considering niacin’s half-life of 20–45 min (Porter
and Kaplan, 2009), the prodrug showed potential to target ASBT
and yield a prolonged pharmacokinetic profile, which may cause
levels of the flushing side effect. Compared to the niacin pro-
drug, the ketoprofen prodrug was a weaker substrate for ASBT and
exhibited greater metabolic stability in Caco-2 homogenate. The
conversion of prodrug to ketoprofen was achieved slowly in liver
homogenate. Enzymatic stability results suggest that the prodrugs
may be hydrolyzed gradually in vivo after absorption via ASBT.
For both prodrugs, further examination is warranted, in terms of
these bile acid conjugates being substrates for other transporters
that contribute to enterohepatic recirculation, as well as in vivo
studies to assess potential for extended pharmacokinetic drug pro-
file after prodrug administration.
4.2. Advantage to target ASBT
The conventional approaches for prolonged drug release are
formulations or devices which provide a slow release/dissolution
of drug within the intestine. Such approaches require the drug
to be well absorbed in the large intestine where formulations
will largely reside for 12–24 h after oral administration. Drugs
that are amenable to conventional sustained release approaches
must be highly permeable. Moreover, it is difficult to control drug
release, and dose dumping can cause toxicity. Previously, prodrug
approaches for the oral delivery of ketoprofen have employed poly-
mer and dextran (Larsen et al., 1991; Wang et al., 2002). Prodrug
approaches for niacin have entailed dextran and diacylglycerol
esters (Cordi et al., 1995; Filip et al., 2003). Extended-release for-
mulations of ketoprofen and niacin have been marketed.
5. Conclusions
The carrier used in the prodrug approach here was CDCA,
with lysine as the linker. Lysine was selected since it is a natural
amino acid and allows for CDCA and drug to be incorporated via
amide linkages. It also provided a negative charge, which is crit-
ical for ASBT substrate translocation (Balakrishnan et al., 2006a).
As a targeting moiety for ASBT as well as transporters involved
in enterohepatic recirculation, CDCA is an endogenous compound,
as is lysine, such that the prodrug has the potential advantage of
low toxicity. In addition, as bile acids are distinctive physiological
molecules and are poorly metabolized, a bile acid conjugate may
show metabolic stability (Kramer and Wess, 1996).
Bile acid conjugates of each niacin and ketoprofen were syn-
thesized. Both prodrugs exhibited binding affinity for ASBT. Niacin
prodrug showed higher ASBT substrate binding affinity, but lower
substrate capacity compared to taurocholate. Ketoprofen prodrug
was a weaker substrate. Stability results suggest that each prodrug
exhibits potential enzymatic stability for the conjugates to provide
sustained drug release. Overall, in vitro results showed that these
bile acid conjugates are potential prolonged release prodrugs with
binding affinity for ASBT. Future in vivo pharmacokinetic assess-
ment of the conjugates is warranted.
Several other intestinal influx transporters have been proposed
as prodrug targets, such as such as nucleoside transporters (ENTs),
organic cation transporters (OCTs), organic anion-transporting
polypeptides (OATPs), and peptide transporter (Pept1). Pept1
translocates valacyclovir and increase acyclovir bioavailability 3-
fold (Burnette and de, 1994). Ki of valacyclovir is 4.08 mM in
Xenopus laevis oocytes (Balimane et al., 1998) and 1.10 mM in CHO-
PepT1 (Han et al., 1998). In comparison to this peptide prodrug, the
bile acid prodrug approach that targets ASBT has the advantage of
higher binding affinity (i.e. micromolar range).
Acknowledgments
This work was supported in part by National Institutes of Health
grants DK67530. The author gratefully acknowledges Dr. Gasirat
Tririya for her assistance in chemical synthesis and Dr. Guoyun Bai
for NMR assignment assistance.
References
4.3. Bile acid prodrugs
Alrefai, W.A., Gill, R.K., 2007. Bile acid transporters: structure, function, regulation
and pathophysiological implications. Pharm. Res. 24, 1803–1823.
Balakrishnan, A., Hussainzada, N., Gonzalez, P., Bermejo, M., Swaan, P.W., Polli, J.E.,
2007. Bias in estimation of transporter kinetic parameters from overexpres-
sion systems: interplay of transporter expression level and substrate affinity. J.
Pharmacol. Exp. Ther. 320, 133–144.
Several bile acid conjugates were previously designed to provide
site-directed liver and gallbladder delivery (Kramer et al., 1994,
1992; Petzinger et al., 1995, 1999). For example, bile acid conjugates

118
X. Zheng, J.E. Polli / International Journal of Pharmaceutics 396 (2010) 111–118
Balakrishnan, A., Polli, J.E., 2006. Apical sodium dependent bile acid transporter
(ASBT, SLC10A2): a potential prodrug target. Mol. Pharm. 3, 223–230.
Balakrishnan, A., Sussman, D.J., Polli, J.E., 2005. Development of stably transfected
monolayer overexpressing the human apical sodium-dependent bile acid trans-
porter (hASBT). Pharm. Res. 22, 1269–1280.
Balakrishnan, A., Wring, S.A., Coop, A., Polli, J.E., 2006a. Influence of charge and
steric bulk in the C-24 region on the interaction of bile acids with human apical
sodium-dependent bile acid transporter. Mol. Pharm. 3, 282–292.
Balakrishnan, A., Wring, S.A., Polli, J.E., 2006b. Interaction of native bile acids
with human apical sodium-dependent bile acid transporter (hASBT): influ-
ence of steroidal hydroxylation pattern and C-24 conjugation. Pharm. Res. 23,
1451–1459.
Kramer, W., Wess, G., Schubert, G., Bickel, M., Girbig, F., Gutjahr, U., Kowalewski, S.,
Baringhaus, K.H., Enhsen, A., Glombik, H., 1992. Liver-specific drug targeting by
coupling to bile acids. J. Biol. Chem. 267, 18598–18604.
Larsen, C., Jensen, B.H., Olesen, H.P., 1991. Bioavailability of ketoprofen from orally
administered ketoprofen–dextran ester prodrugs in the pig. Acta Pharm. Nord.
3, 71–76.
Menon, R.M., Gonzalez, M.A., Adams, M.H., Tolbert, D.S., Leu, J.H., Cefali, E.A., 2007.
Effect of the rate of niacin administration on the plasma and urine pharmacoki-
netics of niacin and its metabolites. J. Clin. Pharmacol. 47, 681–688.
Montagnani, M., Love, M.W., Rossel, P., Dawson, P.A., Qvist, P., 2001. Absence of
dysfunctional ileal sodium-bile acid cotransporter gene mutations in patients
with adult-onset idiopathic bile acid malabsorption. Scand. J. Gastroenterol. 36,
1077–1080.
Morgan, J.M., Capuzzi, D.M., Guyton, J.R., 1998. A new extended-release niacin (Nias-
pan): efficacy, tolerability, and safety in hypercholesterolemic patients. Am. J.
Cardiol. 82, 29U–34U.
Balimane, P.V., Tamai, I., Guo, A., Nakanishi, T., Kitada, H., Leibach, F.H., Tsuji, A., Sinko,
P.J., 1998. Direct evidence for peptide transporter (PepT1)-mediated uptake
of a nonpeptide prodrug, valacyclovir. Biochem. Biophys. Res. Commun. 250,
246–251.
Benyo, Z., Gille, A., Kero, J., Csiky, M., Suchankova, M.C., Nusing, R.M., Moers, A.,
Pfeffer, K., Offermanns, S., 2005. GPR109A (PUMA-G/HM74A) mediates nicotinic
acid-induced flushing. J. Clin. Invest. 115, 3634–3640.
Oelkers, P., Kirby, L.C., Heubi, J.E., Dawson, P.A., 1997. Primary bile acid malabsorp-
tion caused by mutations in the ileal sodium-dependent bile acid transporter
gene (SLC10A2). J. Clin. Invest. 99, 1880–1887.
Burnette, T.C., de, M.P., 1994. Metabolic disposition of the acyclovir prodrug valaci-
clovir in the rat. Drug Metab. Dispos. 22, 60–64.
Cordi, A., Lacoste, J.M., Duhault, J., Espinal, J., Boulanger, M., Broux, O., Husson, B.,
Volland, J.P., Mahieu, J.P., 1995. Synthesis of 1,2-diacyl-3-nicotinoyl glycerol
derivatives and evaluation of their acute effects on plasma lipids in the rat.
Arzneimittelforschung 45, 997–1001.
Dawson, P.A., Haywood, J., Craddock, A.L., Wilson, M., Tietjen, M., Kluckman, K.,
Maeda, N., Parks, J.S., 2003. Targeted deletion of the ileal bile acid trans-
porter eliminates enterohepatic cycling of bile acids in mice. J. Biol. Chem. 278,
33920–33927.
Dawson, P.A., Lan, T., Rao, A., 2009. Bile acid transporters. J. Lipid Res..
Dawson, P.A., Oelkers, P., 1995. Bile acid transporters. Curr. Opin. Lipidol. 6, 109–114.
Filip, C., Ungureanu, D., Gheorghita, N., Ghitler, N., Mocanu, G., Nechifor, M., 2003.
Hypolipidemic effect of a prodrug containing nicotinic acid in rats. Correlation
with plasmatic levels. Rev. Med. Chir. Soc. Med. Nat. Iasi 107, 179–183.
Gallop, M.A., Cundy, K.C. Bile-acid conjugates for providing sustained systemic
concentrations for drugs. PCT/US01/42628. 11-4-2002a. WO 02/28883 A1. 9-
10-2001a. Ref. type: Patent.
Petzinger, E., Nickau, L., Horz, J.A., Schulz, S., Wess, G., Enhsen, A., Falk, E., Baring-
haus, K.H., Glombik, H., Hoffmann, A., 1995. Hepatobiliary transport of hepatic
3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors conjugated with
bile acids. Hepatology 22, 1801–1811.
Petzinger, E., Wickboldt, A., Pagels, P., Starke, D., Kramer, W., 1999. Hepatobiliary
transport of bile acid amino acid, bile acid peptide, and bile acid oligonucleotide
conjugates in rats. Hepatology 30, 1257–1268.
Pieper, J.A., 2002. Understanding niacin formulations. Am. J. Manag. Care 8,
S308–S314.
Pieper, J.A., 2003. Overview of niacin formulations: differences in pharmacokinetics,
efficacy, and safety. Am. J. Health Syst. Pharm. 60, S9–14.
Piepho, R.W., 2000. The pharmacokinetics and pharmacodynamics of agents proven
to raise high-density lipoprotein cholesterol. Am. J. Cardiol. 86, 35L–40L.
Pike, N.B., 2005. Flushing out the role of GPR109A (HM74A) in the clinical efficacy
of nicotinic acid. J. Clin. Invest. 115, 3400–3403.
Porter, R.S., Kaplan, L.J., 2009. The Merck Manuals Online Medical Library. Merck
Research Laboratories, Division of Merck & Co., Inc., Whitehouse Station, NJ.
Rais, R., Gonzalez, P.M., Zheng, X., Wring, S.A., Polli, J.E., 2008. Method to screen sub-
strates of apical sodium-dependent bile acid transporter. AAPS J. 10, 596–605.
St-Pierre, M.V., Kullak-Ublick, G.A., Hagenbuch, B., Meier, P.J., 2001. Transport of bile
acids in hepatic and non-hepatic tissues. J. Exp. Biol. 204, 1673–1686.
Stern, R.H., Freeman, D., Spence, J.D., 1992. Differences in metabolism of time-release
and unmodified nicotinic acid: explanation of the differences in hypolipidemic
action? Metabolism 41, 879–881.
Gallop, M.A., Cundy, K.C. Bile-acid derived compounds for providing sustained
systemic concentrations of drugs after oral administration. PCT/US01/42513.
11-4-2002b. WO 02/28881 A1 Ref. type: Patent.
Gonzalez, E.L., Patrignani, P., Tacconelli, S., Rodriguez, L.A., 2010. Variability of risk of
upper gastrointestinal bleeding among nonsteroidal anti-inflammatory drugs.
Arthritis Rheum..
Gonzalez, P., Polli, J.E., 2008. Impact of impurity on kinetic estimates from transport
and inhibition studies. J. Pharmacol. Exp. Ther. 326, 296–305.
Han, H., de Vrueh, R.L., Rhie, J.K., Covitz, K.M., Smith, P.L., Lee, C.P., Oh, D.M., Sadee,
W., Amidon, G.L., 1998. 5ꢂ-Amino acid esters of antiviral nucleosides, acyclovir,
and AZT are absorbed by the intestinal PEPT1 peptide transporter. Pharm. Res.
15, 1154–1159.
Tantishaiyakul, V., Wiwattanawongsa, K., Pinsuwan, S., Kasiwong, S., Phadoong-
sombut, N., Kaewnopparat, S., Kaewnopparat, N., Rojanasakul, Y., 2002.
Characterization of mefenamic acid-guaiacol ester: stability and transport
across Caco-2 cell monolayers. Pharm. Res. 19, 1013–1018.
Tolle-Sander, S., Lentz, K.A., Maeda, D.Y., Coop, A., Polli, J.E., 2004. Increased acyclovir
oral bioavailability via a bile acid conjugate. Mol. Pharm. 1, 40–48.
Wang, L.F., Chiang, H.N., Wu, P.C., 2002. Kinetics and hydrolysis mechanism of poly-
meric prodrugs containing ibuprofen, ketoprofen, and naproxen as pendent
agents. J. Biomater. Sci. Polym. Ed. 13, 287–299.
Wong, M.H., Oelkers, P., Dawson, P.A., 1995. Identification of a mutation in the ileal
sodium-dependent bile acid transporter gene that abolishes transport activity.
J. Biol. Chem. 270, 27228–27234.
Wong, M.H., Rao, P.N., Pettenati, M.J., Dawson, P.A., 1996. Localization of the ileal
sodium-bile acid cotransporter gene (SLC10A2) to human chromosome 13q33.
Genomics 33, 538–540.
Kantor, T.G., 1986. Ketoprofen: a review of its pharmacologic and clinical properties.
Pharmacotherapy 6, 93–103.
Knopp, R.H., Alagona, P., Davidson, M., Goldberg, A.C., Kafonek, S.D., Kashyap,
M., Sprecher, D., Superko, H.R., Jenkins, S., Marcovina, S., 1998. Equiva-
lent efficacy of a time-release form of niacin (Niaspan) given once-a-night
versus plain niacin in the management of hyperlipidemia. Metabolism 47,
1097–1104.
Kramer, W., Wess, G., 1996. Bile acid transport systems as pharmaceutical targets.
Eur. J. Clin. Invest. 26, 715–732.
Kramer, W., Wess, G., Enhsen, A., Bock, K., Falk, E., Hoffmann, A., Neckermann, G.,
Gantz, D., Schulz, S., Nickau, L., 1994. Bile acid derived HMG-CoA reductase
inhibitors. Biochim. Biophys. Acta 1227, 137–154.
Zheng, X., Ekins, S., Raufman, J.P., Polli, J.E., 2009. Computational models for drug
inhibition of the human apical sodium-dependent bile acid transporter. Mol.
Pharm. 6, 1591–1603.