pharmaceutics

Article
Preparation and Evaluation of Amino Acid
Conjugates of Celecoxib as Prodrugs to Improve
the Pharmacokinetic and Therapeutic Properties
of Celecoxib
Yonghyun Lee 1,2,3, *, Jungyun Kim 1 , Wooseong Kim 1 , In-Soo Yoon 1 and Yunjin Jung 1, *
1 College of Pharmacy, Pusan National University, Busan 46241, Korea; [email protected] (J.K.);
[email protected] (W.K.); [email protected] (I.-S.Y.)
2 Department of Pharmacy, College of Pharmacy, Ewha Womans University, Seoul 03760, Korea
3 Graduate School of Pharmacy, Ewha Womans University, Seoul 03760, Korea
* Correspondence: [email protected] (Y.L.); [email protected] (Y.J.)




Received: 20 October 2020; Accepted: 26 October 2020; Published: 30 October 2020


Abstract: Although celecoxib is quite effective in the management of inflammation-related diseases,

especially arthritis, its use is limited by concerns including low bioavailability (BA), non-linear
pharmacokinetic (PK) profile, and peak concentration-related toxicity. To overcome these issues,
we designed and prepared hydrophilic celecoxib prodrugs, namely N-glycyl-aspart-1yl celecoxib
(N-GA1C), glutam-1-yl celecoxib (G1C), and aspart-1yl celecoxib (A1C), for the sustained release of
celecoxib in the intestine with limited systemic absorption. The celecoxib derivatives were converted
to celecoxib in the intestinal contents. The conversion rates were in order of N-GA1C > G1C > A1C.
Oral administration of the celecoxib derivatives (oral celecoxib derivatives) sustained the plasma
concentration of celecoxib for 24 h, improving the BA and linearity of the PK profile of celecoxib.
The peak concentrations (Cmax ) of celecoxib after oral celecoxib derivatives were lower than that
after oral celecoxib. In a carrageenan-induced rat paw edema model, oral N-GA1C exhibited greater
anti-inflammatory activity for a longer duration compared with oral celecoxib. The order of efficacy
of the celecoxib derivatives was N-GA1C > G1C > A1C. Taken together, the prodrug approach is
a feasible strategy to improve the PK and therapeutic properties of celecoxib, and among the celecoxib
derivatives, N-GA1C may be the most promising prodrug of celecoxib.

Keywords: celecoxib; prodrug; bioavailability; PK linearity; extended absorption; carrageenan-induced

rat paw edema model

1. Introduction
Celecoxib, a selective COX-2 inhibitor, is one of the most popular anti-inflammatory drugs for the
treatment of chronic inflammatory diseases such as arthritis [1,2]. Celecoxib is a Biopharmaceutics
Classification System (BCS) class II drug with low solubility and high permeability [3,4]. Despite
the extensive clinical use of celecoxib, some negative pharmacokinetic (PK) properties, such as
(1) low bioavailability (BA) [3–5], (2) non-linear pharmacokinetic profiles [6], and (3) peak blood
concentration-related cardiovascular toxicity [2], have hindered its broad clinical usage.
Generally, a prodrug, pharmacologically inactive, per se, is created by appropriate chemical
modification of a parent drug, and it is chemically and/or biologically converted to the active parent
drug after achieving its purpose [7,8]. This chemical approach has been widely utilized to improve the
BA of hydrophobic drugs and/or the sustainability of drug action [7,8]. Prodrugs for the pharmaceutical
purposes are likely to have PK profiles with improved linearity, low peak blood concentration, and

Pharmaceutics 2020, 12, 1043; doi:10.3390/pharmaceutics12111043 www.mdpi.com/journal/pharmaceutics

Pharmaceutics 2020, 12, 1043 2 of 13

sustained concentration of the parent drugs in the blood for a long time, resulting in extended duration
of action and reduced risk of peak blood concentration-associated toxicity [7,8]. Besides, it would
be ideal that prodrugs administered orally achieve the goal with minimal systemic absorption, thus
avoiding possible side effects associated with systemic absorption of prodrugs [7,8].
To address the challenges associated with the PK of celecoxib, we designed hydrophilic prodrugs of
celecoxib, which can release the parent drug in a controlled manner in the small and large intestine [9–12].
Hydrophilicity and the controlled release property of prodrugs are expected to not only improve the
dissolution of the BCS class II drug but also sustain the absorption of the parent drug while limiting
the systemic absorption of the prodrug in the intestine. Our previous studies have demonstrated that
the amide bond formed between the carboxylic group of amino acids and the sulfonamide group
of celecoxib is susceptible to the host and/or microbial enzymes in the intestine [10–15]. Therefore,
celecoxib was coupled with hydrophilic amino acids to yield aspart-1yl celecoxib (A1C), glutam-1yl
celecoxib (G1C), and N-glycyl-aspart-1yl celecoxib (N-GA1C). The ability of the hydrophilic celecoxib
derivatives to improve the PK properties of the parent drug was evaluated. Moreover, we examined if
such improvement in PK properties translates into an improved therapeutic action of celecoxib against
carrageenan-induced paw edema in rats.

2. Materials and Methods

2.1. Synthesis of Aspart-1-yl Celecoxib (A1C) and Glutam-1-yl Celecoxib (G1C)

Briefly, 1,10 -carbonyldiimidazole (5.80 mmol) (CDI, Sigma-Aldrich, St. Louis, MO, USA) was added
to 30 mL of acetonitrile (ACN, Junsei, Tokyo, Japan) containing 4.98 mmol of 4-benzyl N-(t-Boc)-aspartic
acid or 5-benzyl N-(t-Boc)-glutamic acid (Tokyo Kasei Kogyo Co. Tokyo, Japan). After stirring for
10 min at room temperature (RT), 1.31 mmol of celecoxib (ether-extracted from Celebrex capsules) and
0.3 mL of triethylamine (TEA, Junsei) were added to the mixture, and the reaction was allowed to
proceed while stirring for 4 h at 55 ◦ C under nitrogen gas. After evaporating the solvents, the residue
was dissolved in 40 mL ethyl acetate/ether (1:3), washed with 5% NaHCO3 , and subsequently dried
over anhydrous Na2 SO4 followed by the evaporation of the solvents. For the deprotection of the t-Boc
group, 1 M HCl/acetic acid (10 mg/mL) was treated for 3 h at RT, and for the deprotection of the benzyl
group, 1 M NaOH (5 mL) was treated for 2 h at RT. After acidification, A1C or G1C was acquired as
a white precipitate. 1H NMR spectra were obtained on a Varian AS 500 spectrometer (Varian, Palo
Alto, CA, USA) and chemical shifts represent ppm downfield from tetramethylsilane. Infrared (IR)
spectra were recorded on a Varian Fourier-transform IR spectrophotometer (Varian) and elemental
analysis was performed using FLASH 2000 series (Thermo Scientific Inc. Waltham, MA). Physical data,
including the melting point and the results of IR (Nujol), 1H-NMR (DMSO-d6), and elemental analysis,
are identical with those in a previous study [9].

2.2. Synthesis of N-Glycyl-Aspart-1-yl Celecoxib (N-GA1C)

CDI (2.21 mmol) was dissolved in 12 mL of ACN containing 1.9 mmol of N-(t-Boc)-glycine (Tokyo
Kasei Kogyo Co.). After stirring for 10 min at RT, 0.5 mmol of 4-benzyl aspart-1-yl celecoxib and
1.64 mL of TEA were added to the reaction mixture, and the mixture was stirred for 4 h at 55 ◦ C. After
the removal of solvents by evaporation, the residue was dissolved in 20 mL of ethyl acetate/ether (1:3),
washed with 5% NaHCO3 , dried over anhydrous Na2 SO4 , and then concentrated by evaporation. The
residue was deprotected in 1 M HCl/acetic acid (10 mg/mL) at RT for 3 h and in 1 M NaOH (5 mL) for
2 h. After acidification, N-glycyl-aspart-1-yl celecoxib (N-GA1C) was acquired as white precipitate.
Physical data, including the melting point and the results of IR (Nujol), 1H-NMR (DMSO-d6), and
elemental analysis, are identical with those in a previous study [10].
Pharmaceutics 2020, 12, 1043 3 of 13

2.3. HPLC Analysis

The HPLC system consisted of a model 306 pump, a 117 variable ultraviolet detector, a model
234 autoinjector, and a Model 805 manometric module from Gilson Inc. (Middleton, WI, USA).
A C18 symmetry column (Waters, Milford, MA, USA) (250 × 4.6 mm) with a guard column (Waters,
3.9 × 20 mm) was used. The samples were filtered through a membrane filter (0.45 µm), and the filtrate
(20 µL) was injected on a symmetry C18 column (Waters), which was eluted with a mobile phase at
a flow rate of 1 mL/min. The mobile phase consisted of 60% ACN (Merck, Darmstadt, German) in
0.067 M phosphate buffer (pH 4.5) containing 0.1% trifluoroacetic acid (Sigma-Aldrich), which was
filtered through a 0.45 µm membrane filter (Waters) before use. The eluate was monitored at 273 nm,
and the detection limit was about 0.2 µg/mL under our experimental conditions. Accuracy and relative
standard deviations were 98.7% and 0.43%, respectively Trilution® LC V4 software (Gilson, Middleton,
WI, USA) was used for data analysis. The retention times of celecoxib, A1C, G1C, and N-GA1C were
10.68, 3.11, 3.00, and 2.43 min, respectively.

2.4. Apparent Distribution Coefficient (Logd6.8 ) Measurement

For this study, 1-octanol and pH 6.8 isotonic phosphate buffer (IPB) were pre-saturated with IPB
and 1-octanol, respectively. 1-octanol (10 mL) was added to 500 µM of celecoxib, A1C, or N-GA1C in
10 mL IPB. The mixture was shaken for 1 h and left at 37 ◦ C for 3 h. The concentration of celecoxib, A1C,
G1C, or N-GA1C in the aqueous phase was analyzed by HPLC. The apparent partition coefficients were
calculated using the equation (Co − Cw)/Cw, where Co and Cw represent the initial and equilibrium
concentration of the drug in the aqueous phase, respectively.

2.5. Chemical Stability Study

A1C, G1C, or N-GA1C (500 µM) in hydrochloric acid (pH 1.2) or isotonic phosphate (pH 6.8) buffer
was incubated at 37 ◦ C for 10 h. At the predetermined time points, 20 µL of the solution was transferred
into an HPLC vial, and the concentrations of A1C, G1C, or N-GA1C were analyzed by HPLC.

2.6. Animals
The study animals were cared for following national and local guidelines. The animal protocol
used in this study was reviewed and approved by the Pusan National University Institutional Animal
Care and Use Committee on ethical procedures and scientific care (Approval number: PNU-2017-1525).
All animals were obtained from Samtako (Osan, South Korea) and housed under pathogen-free
conditions in the animal facility at the College of Pharmacy of Pusan National University. Rats were
assigned randomly to the experimental groups. The investigators were not blinded to allocation during
experiments and outcome assessment unless the section mainly included a blind assessment.

2.7. Drug Release Study

Drug release study was performed with the celecoxib derivatives as described previously [9,10].
Six-week-old male Sprague Dawley rats (250–260 g) were sacrificed using CO2 , and a midline incision
was made. The contents of the proximal small intestine, distal small intestine, and cecum, liver
homogenate, or blood plasma were collected and suspended in IPB (20% (w/v)). A1C, G1C, or N-GA1C
in the buffer (0.5 mL, 1 mM) was mixed with 0.5 mL of the suspension and then incubated at 37 ◦ C
under nitrogen gas. At the predetermined time points, the samples were extracted with 0.5 mL of
ethyl acetate followed by centrifugation at 6000× g for 5 min. Methanol (1.0 mL) was added to the
residue obtained from the evaporation of 0.1 mL of the organic layer. The solution was centrifuged at
20,000× g at 4 ◦ C for 10 min. The concentrations of celecoxib, A1C, G1C, and N-GA1C in 20 µL of the
supernatant were determined by HPLC.
Pharmaceutics 2020, 12, 1043 4 of 13

2.8. Pharmacokinetic Study

Six-week-old male Sprague Dawley rats (250–260 g) were administered 10 mg/kg of celecoxib,
A1C (equivalent mass of celecoxib), G1C (equivalent mass of celecoxib), or N-GA1C (equivalent
mass of celecoxib) orally or intravenously via the tail vein. Blood samples were collected from
the jugular vein using a heparinized syringe at predetermined times. Heparinized blood samples
were immediately centrifuged at 6000× g for 5 min. Five-fold volume of ethyl acetate was mixed
with plasma, centrifugation was performed at 6000× g for 5 min, and 1 mL of the organic solvent
was evaporated. The residues were mixed with 0.1 mL of IPB and centrifuged at 20,000× g for
10 min at 4 ◦ C. The concentration of celecoxib or N-GA1C in the samples was determined by HPLC.
PK profiling was performed using PKsolver [16]. Absolute BA and relative BA were calculated
using the following equations: Absolute BA (F, %) = ((AUC0–∞ )PO/dose PO)/((AUC0–∞ )IV/dose IV);
Relative BA (%) = ((AUC0–∞ )A/dose A)/((AUC0–∞ )B/dose B).

2.9. Carrageenan-Induced Inflammation Study

Six-week-old male Sprague Dawley rats (250–260 g) were orally administered celecoxib (10 mg/kg),
A1C (equivalent mass of celecoxib), G1C (equivalent mass of celecoxib), N-GA1C (equivalent mass of
celecoxib), or IPB as a control. After 1 h, 0.1 mL of carrageenan (1%) was injected into the hind paw of
the rats. Carrageenan-induced increase in paw volume (paw edema) was measured at predetermined
time points (0, 3, 6, 12, and 24 h). The paw volume of rats up to the knee joint was measured using a
plethysmometer. The degree of inhibition of paw edema (%) at each time point was calculated using
the following equation: (Control volume − treatment group volume)/Control paw volume × 100.

2.10. Statistical Analysis

All experiments were performed at least twice with duplicate repeated measures. The results are
expressed as means ± standard error of means. A one-way or two-way ANOVA, followed by Tukey’s
HSD multiple comparison post hoc test was used to determine the differences among groups. Data
were approximately normally distributed, and the variance was similar between the groups. Statistical
significance is indicated as * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. GraphPad Prism 8.0
(GraphPad Software, La Jolla, CA, USA) was used for statistical analyses.

3. Results

3.1. Preparation of Celecoxib Derivatives Coupled with Amino Acids

The synthesis of A1C, G1C (Scheme 1a) and N-GA1C (Scheme 1b) was conducted as described
previously [9,10].We conjugated the sulfonamide group of celecoxib with the carboxyl group of
protected aspartic acid and glutamic acid using CDI, a coupling reagent. The protection groups, t-Boc
and benzyl groups, were removed in acidic and basic conditions, respectively, yielding A1C and G1C.
For the synthesis of N-GA1C (Scheme 1b), the amine group of 4-benzyl aspart-1-yl celecoxib was
conjugated with the carboxyl group of N-t-Boc glycine with CDI as the coupling reagent. Subsequently,
the protection groups were removed, yielding N-GA1C. Physical data of the celecoxib-amino acid
conjugates are shown in Table 1, which are identical with those in previous studies [9,10].
Pharmaceutics 2020, 12, 1043 5 of 13
Pharmaceutics 2020, 12, x FOR PEER REVIEW 5 of 13

Synthesisscheme
Scheme1.1.Synthesis
Scheme schemeofofaspart-1-yl
aspart-1-yl celecoxib
celecoxib (A1C,
(A1C, a),a), glutam-1-yl
glutam-1-yl celecoxib,
celecoxib, (G1C,
(G1C, b), and
b), and N-
N-glycyl-aspart-1-yl
glycyl-aspart-1-yl celecoxib
celecoxib (N-GA1C,
(N-GA1C, c). c). Adapted from [9,10], published by Springer Nature, 2012
and Dovepress, 2015.
Table 1. IR spectra, NMR spectra, elemental analysis, log D6.8 values and melting point (MP) of
Table 1. IR spectra, NMR spectra, elemental analysis, log D6.8 values and melting point (MP) of
celecoxib, A1C, G1C, and N-GA1C.
celecoxib, A1C, G1C, and N-GA1C.
IR Spetra (Nujol) 1
H-NMR (DMSO-d6) Log Melting
IR Spetra 1 H-NMR Elemental Analysis
(DMSO-d6)
(δ, ppm) D6.8Melting
Point
(Nujol)
(cm−1) Elemental Analysis Log D6.8 Point
(δ, ppm)
(cm−1 ) C19H17F3N4O3S (438);
2.30 (s, 3H), 7.17 (s, 1H), 7.16– C19 H17calcd:
F3 N4 OC,
1720 (C=O, 2.30 (s, 3H), 3 S 52.05;
(438); H, 3.88; 189 °C–
Celecoxib 7.20 (m,7.174H),
(s, 1H),
7.517.16–7.20
(d, 2H, J = calcd:
7.5 C, 4.01 ◦
1720SO(C=O,
NHCO) 52.05; H, 3.88;
Celecoxib 2 (m, 4H), 7.51 (d, 2H, J = 7.5 Hz), N, 12.78; found: C,4.01 194 °C◦ C
189 C–194
SO2 NHCO) Hz), 7.86 (d, 2H, J = 8.0 Hz) N, 12.78; found: C,
7.86 (d, 2H, J = 8.0 Hz) 52.10;
52.10; H, 3.89; H,
N, 3.89;
12.75.N, 12.75.
1721 (C=O, 1.98 (dd, 1H, J = 10.5, 16.5 Hz),
1.98 (dd, 1H, J = 10.5, 16.5 Hz), C21 H19 F3 N4 O5 S
1721 (C=O,
SO2 NHCO), C H19F3N4O5S (496.10)
2.29 (s,2.29
3H),(s,2.32
3H), 2.32
(dd, 1H,(dd,
J =1H,
3.5, J = 3.5,(496.10)21 Calcd: C,
1596 (C=O, 1596
SO2NHCO), 16.0 Hz), 3.34 (m, 1H), 7.13 (s, ◦ 134 °C–◦
A1C
A1C 16.0 Hz), 3.34 (m, 1H), 7.13 (s, 1H), 50.80; H, Calcd: C, 50.80;
3.86; N, 11.29; H, 3.86;
0.6 0.6134 C–137 C
zwitterionic
(C=O, zwitterionic 1H),
7.15–7.20 7.15–7.20
(m, 4H), 7.34(m,(d,4H),
2H, 7.34 (d,
Found: C, 50.82; H,
N, 11.29; Found: C, 137 °C
carboxylate
carboxylate J = 8.5 Hz),
salt) 2H,7.80 (d, Hz),
J = 8.5 = 9.0(d,
2H, J7.80 Hz)2H, J = 3.91; N, 11.35.
50.82; H, 3.91; N, 11.35.
salt)
1719 (C=O, 9.0 Hz)
1.7 (m, 1H), 2.25 (m, 1H), 2.02 (m, C H F N O S
1719 (C=O, 2H), 2.301.7
SO2 NHCO), (m, 1H), 2.25 (m, 1H), 2.02 22 C21 H3 F4 N5 O S (510.12)
(s, 3H), 3.29 (m, 1H), 7.14 (510.12)22 Calcd:
21 3 4C, 5
1595 (C=O, (m, 2H), 2.30 (s, 3H), 3.29 (m,
G1C SO NHCO), (s,
1595 1H), 7.16–7.21 (m, 4H), 7.59 (d, 51.76; H,
Calcd: N,
4.15; 10.98; H, 4.15;
C, 51.76; 0.81 110 ◦ C–115 ◦
110 °C–C
G1C zwitterionic
2
2H, J1H),
= 8.57.14
Hz),(s,7.95
1H),(d,7.16–7.21
2H, (m,Found: C, 51.82; H, 0.81
115 °C
carboxylate
(C=O, zwitterionic
= 8.5
4H), J7.59 2H, J = 8.5 Hz), 7.95 4.11; N,
(d,Hz) 10.98; Found: C,
N, 11.20.
salt)
carboxylate salt) 51.82; H, 4.11; N, 11.20.
1700 (C=O, 2.30 (s, 3H), 2.35 (d, (dd,
2H, J1H,= 8.5
J =Hz)
7.5,
C23 H22 F3 N5 O5 S
SO2 NHCO), 15.5 Hz),2.30 (s,(dd,
2.61 3H),1H,2.35J =(dd,
5.5,1H,
12.5J = 7.5,
(553.12) Calcd:
F3N5C,
16681700
(C=O, (C=O, Hz), 4.0215.5 Hz),
(t, 2H, J =2.61 (dd, 4.4
5.0 Hz), 1H,(m,J = 5.5, C23H O5S (553.12)
N-GA1C 49.91; H, 4.01;22 N, 12.56; −0.13 154 ◦ C–158 ◦ C
amide),
SO2NHCO),1596 16681H), 7.13
12.5 (s,
Hz),1H), 7.14–7.21
4.02 (t, 2H, (m,
J = 5.0 Hz), Calcd: C, 49.91; 154 °C–
N-GA1C Found: C, 49.95; H, H, 4.01; −0.13
(C=O,
(C=O, amide), 15964H), 7.514.4(d,
(m, 2H, J = 7.13
1H), 8.5 Hz), 7.85 7.14–
(s, 1H), 158 °C
4.07; N, 12.56;
12.50. Found: C,
carboxylic) (d, 2H, J = 8.0 Hz)
(C=O, carboxylic) 7.21 (m, 4H), 7.51 (d, 2H, J = 8.5 49.95; H, 4.07; N, 12.50.
Hz), 7.85 (d, 2H, J = 8.0 Hz)
Pharmaceutics 2020,
Pharmaceutics 12,12,
2020, 1043
x FOR PEER REVIEW 6 of 1313
6 of

3.2. Celecoxib Derivatives Are Converted to Celecoxib in The Intestinal Contents of Rats
3.2. Celecoxib Derivatives Are Converted to Celecoxib in The Intestinal Contents of Rats
To examine if celecoxib derivatives can be converted to celecoxib in the intestine, A1C, G1C, and
To examine
N-GA1C if celecoxibinderivatives
were incubated can be
the small- and converted to celecoxib
large-intestinal contents andin the intestine,
changes A1C,
in the G1C, and
concentration
N-GA1C were incubated in the small- and large-intestinal contents and changes in the
of the celecoxib derivatives and celecoxib were monitored. In the proximal and distal small intestinal concentration of
the celecoxib
contents, derivatives
N-GA1C andcelecoxib
released celecoxib(27.5
were monitored.
± 1.51% at 3 h, In
PSI;the proximal
33.28% andatdistal
± 2.25% small
3 h, DSI); intestinal
the celecoxib
contents, N-GA1C released celecoxib (27.5 ± 1.51% at 3 h, PSI; 33.28% ± 2.25% at
release occurred at different rates -N-GA1C > G1C > A1C (p < 0.0001; Figure 1a,b). Upon incubation3 h, DSI); the celecoxib
release occurred
with cecal at different
contents, N-GA1C rateswas rapidly>and
-N-GA1C > A1C (p
G1Centirely < 0.0001;toFigure
converted 1a,b).within
celecoxib Upon 4incubation
h, and the
with cecal contents,
celecoxib release from N-GA1C
N-GA1C was rapidly
was and
greater thanentirely
from A1C converted
and G1C to that
celecoxib within 4 converted
were partially h, and theto
celecoxib
celecoxibrelease from 24
even after N-GA1C
h (p < was greater
0.0001; than
Figure from
1c). TheseA1Cresults
and G1C are that were partially
consistent converted
with previous to
papers
celecoxib even after 24 h (p < 0.0001; Figure 1c). These results are consistent with previous
[9,10]. To rule out the possibility of chemical cleavage of the amide bonds between celecoxib and the papers [9,10].
Toamino
rule out the the
acids, possibility
celecoxibof derivatives
chemical cleavage of the amide
were incubated bonds between
in HCl-NaCl buffer celecoxib and IPB
(pH 1.2) and the amino
for 24 h
acids, the celecoxib derivatives were incubated in HCl-NaCl buffer (pH
before analyzing the concentration of celecoxib and the cececoxib derivatives. No change 1.2) and IPB for 24 h beforein
analyzing the concentration
concentration of celecoxib
of the compounds wasand the cececoxib
observed, derivatives.
suggesting that theNo change
celecoxibin concentration
derivatives were of
the compounds was observed, suggesting that the celecoxib derivatives were chemically
chemically stable during GI transit. These results indicate that the release of celecoxib from A1C, G1C, stable during
GIand
transit.
N-GA1CTheseis results
mainlyindicate that
facilitated bythe release enzymes
intestinal of celecoxib from
in the A1C,
small G1C, and
intestine andN-GA1C
microbial is enzymes
mainly
facilitated by intestine.
in the large intestinal enzymes in the small intestine and microbial enzymes in the large intestine.

Figure 1. Amino acid conjugates of celecoxib release celecoxib in both small- and large-intestinal
Figure 1.
contents. Amino
(a–c) acid conjugates
Percentage of celecoxibofreleased
celecoxib release
from celecoxib
prodrugs in both small-
after incubation and suspension
with 20% large-intestinal
of
contents.
contents of (a–c) Percentage
the proximal of intestine
small celecoxib(a),
released from prodrugs
distal small after
intestine (b), orincubation with(c)20%
cecal contents suspension
in phosphate
of contents
buffered salineof(pH
the6.8).
proximal
Data are small intestine
presented (a), distal
as mean small
± s.e.m. fromintestine (b), or cecal
a representative contents
experiment = 5)in
(n (c)
phosphate
from buffered
3 independent experiments. < 0.05,
saline (pH* p6.8). Data p <presented
****are as mean
0.0001, analyzed ± s.e.m. ANOVA
by two-way from a representative
with Tukey’s
HSD multiple(n
experiment comparison
= 5) from 3post hoc test. experiments. * p < 0.05, **** p < 0.0001, analyzed by two-way
independent
ANOVA with Tukey’s HSD multiple comparison post hoc test.
3.3. Systemic Absorption and Exposure of Orally Administered N-GA1C Are Limited
3.3.To
Systemic
examineAbsorption and Exposure
if the systemic of Orally
absorption Administered
of the hydrophilicN-GA1C Are
celecoxib Limited could be reduced,
derivatives
we measured the apparent distribution coefficients of A1C, G1C, N-GA1C, and
To examine if the systemic absorption of the hydrophilic celecoxib derivatives celecoxib. Thebeapparent
could reduced,
distribution coefficients (logD 6.8 ) were 0.6 (A1C), 0.8 (G1C), –0.13 (N-GA1C),
we measured the apparent distribution coefficients of A1C, G1C, N-GA1C, and celecoxib. Theand 4.01 (celecoxib),
suggesting that the passive
apparent distribution transport
coefficients of the
(logD celecoxib derivatives via the intestinal epithelial layer
6.8) were 0.6 (A1C), 0.8 (G1C), –0.13 (N-GA1C), and 4.01
is(celecoxib),
less efficientsuggesting
than that of celecoxib. N-GA1C
that the passive transport was found
of thetocelecoxib
be the most effective via
derivatives derivative with
the intestinal
regard to limiting
epithelial layer isthe systemic
less absorption
efficient than thatofofcelecoxib.
celecoxib.ToN-GA1C
further prove whether
was found N-GA1C’s
to be the mostsystemic
effective
derivative with regard to limiting the systemic absorption of celecoxib. To further prove whether N-
Pharmaceutics 2020, 12, x FOR PEER REVIEW 7 of 13
Pharmaceutics 2020, 12, 1043 7 of 13
Pharmaceutics 2020, 12, x FOR PEER REVIEW 7 of 13
GA1C’s systemic absorption is limited, we monitored blood levels of N-GA1C after oral N-GA1C.
GA1C’s
Figure systemic
2 shows absorption
that N-GA1C is limited,
was we monitored
not detected blood
in theofplasma forlevels
24 of N-GA1C
h, indicative of after oral N-GA1C.
limited
absorption is limited, we monitored blood levels N-GA1C after oral N-GA1C. Figuresystemic
2 shows
Figure
absorption2 shows
and was that
exposureN-GA1C was
of N-GA1C. not detected in the plasma for 24 h, indicative of limited systemic
that N-GA1C not detected in the(Figure
plasma2)for 24 h, indicative of limited systemic absorption and
absorption and exposure of N-GA1C. (Figure 2)
exposure of N-GA1C. (Figure 2).
1.2

concentration (g/ml)
1.2

N-GA1C plasma
1.0
1.0
0.8
0.8
0.6

0.4 0.6

0.2 0.4

0.0 0.2
0
0.0 5 10 15 20 25
0 5 10 (h) 15
Time 20 25
Time (h)

N-GA1C’s
Figure2.2.N-GA1C’s
Figure systemic
systemic absorption
absorption is limited
is limited after
after oraloral administration.
administration. Plasma
Plasma concentration
concentration of
N-GA1C was monitored after oral administration of N-GA1C (10 mg/kg). Data are presented as meanasof
of N-GA1C
Figure 2. was monitored
N-GA1C’s after
systemic oral administration
absorption is limited of
after N-GA1C
oral (10 mg/kg).
administration. Data
Plasma are presented
concentration
± mean
N-GA1C
s.e.m. ±from
s.e.m.
was from a representative
monitored after
a representative experiment
oral administration
experiment (n =
(n = 5) from of25) from 2 independent
N-GA1C (10 mg/kg).
independent experiments.
Data
experiments.are presented as mean
± s.e.m. from a representative experiment (n = 5) from 2 independent experiments.
3.4. Celecoxib Derivatives are Metabolized to Celecoxib in the Liver Homogenate of Rats
3.4. Celecoxib Derivatives are Metabolized to Celecoxib in the Liver Homogenate of Rats
3.4. Additionally,
Celecoxib Derivatives are Metabolized
we examined to Celecoxib
if the celecoxib in the Liver
derivatives Homogenate
could of Rats to celecoxib in the
be metabolized
Additionally, we examined if the celecoxib derivatives could be metabolized to celecoxib in the
blood Additionally,
and liver if absorbed systemically.
we examined To do
if the celecoxib this, the celecoxib
derivatives could derivatives
be metabolizedwere to
incubated
celecoxib in the
blood and liver if absorbed systemically. To do this, the celecoxib derivatives were incubated in in
thethe
liver
bloodhomogenate of rats,
and liver ifofabsorbed and the concentration of celecoxib released from the derivatives was analyzed.
liver homogenate rats, andsystemically. To do this,
the concentration the celecoxib
of celecoxib derivatives
released from the were incubatedwas
derivatives in the
N-GA1C
liver was metabolized
homogenate of to celecoxib
rats, and the (94.87 ± 1.78% after
concentration of 10 h and released
celecoxib 100% conversion
from at 24
the h; Figure 3a),
derivatives was
analyzed. N-GA1C was metabolized to celecoxib (94.87  1.78% after 10 h and 100% conversion at 24
indicating that it was more susceptible to hepatic metabolism than A1C and G1C (Figure 3a). On
h; analyzed.
Figure 3a), N-GA1C wasthat
indicating metabolized
it was more to celecoxib (94.87
susceptible to±hepatic
1.78% after 10 h andthan
metabolism 100%A1Cconversion
and G1C at 24
incubation
h; Figure with
3a), plasma forthat
indicating 24 h,itnone
was of the derivatives
more susceptible released
to celecoxib
hepatic (Figurethan
metabolism 3b). A1C
Theseandresults
G1C
(Figure 3a). On incubation with plasma for 24 h, none of the derivatives released celecoxib (Figure
indicate
(Figure that
3a). N-GA1C
On was poorly
incubation with absorbed
plasma for and
24 a none
h, large ofportion
the of N-GA1C
derivatives absorbed
released systemically
celecoxib (Figure
3b). These results indicate that N-GA1C was poorly absorbed and a large portion of N-GA1C
would be rapidly
3b). These resultsconverted
indicate tothat
celecoxib
N-GA1C in the liver.
was poorly absorbed
absorbed systemically would be rapidly converted to celecoxib in theand
liver.a large portion of N-GA1C
absorbed systemically would be rapidly converted to celecoxib in the liver.

Figure 3. Celecoxib release profiles of A1C, G1C, and N-GA1C in liver homogenate and plasma.
Figure 3. Celecoxib release profiles of A1C, G1C, and N-GA1C in liver homogenate and plasma. (a,b)
(a,b) Percentages
Figure 3. Celecoxibof celecoxib
release releasedoffrom
profiles A1C,the prodrugs
G1C, and after incubation
N-GA1C in liver with 20% suspension
homogenate of liver
and plasma. (a,b)
Percentages of celecoxib released from the prodrugs after incubation with 20% suspension of liver
homogenate
Percentages (a)
of or blood plasma
celecoxib released(b) from
in IPB. Data
the are presented
prodrugs after as mean ±with
incubation s.e.m.20%
from a representative
suspension of liver
homogenate (a) or blood plasma (b) in IPB. Data are presented as mean ± s.e.m. from a representative
experiment
homogenate (n = 5) from 2 independent experiments. **** p < 0.0001, analyzed by two-way ANOVA
experiment (n =(a)
5) or blood
from plasma (b) inexperiments.
2 independent IPB. Data are****
presented as mean
p < 0.0001, ± s.e.m.
analyzed by from a representative
two-way ANOVA
with Tukey’s HSD
experiment 5)multiple
(n = multiple comparison
from 2comparison
independent post hoc test.
experiments.
with Tukey’s HSD post hoc test. **** p < 0.0001, analyzed by two-way ANOVA
with Tukey’s HSD multiple comparison post hoc test.
3.5. Celecoxib Derivatives Improve the PK Properties of Celecoxib
3.5. Celecoxib Derivatives Improve the PK Properties of Celecoxib
3.5. Our data Derivatives
Celecoxib suggest that N-GA1C
Improve would
the PK release
Properties celecoxib during transit in the small and large
of Celecoxib
intestine with the least systemic absorption. Next, we examinedduring
Our data suggest that N-GA1C would release celecoxib if N-GA1Ctransit in the small
improved the PKand large
properties
intestineOur datathe
with suggest
least that N-GA1C
systemic would release
absorption. Next, celecoxib
we examined during
if transit inimproved
N-GA1C the small the and PK
large
of celecoxib. First, the blood concentration of celecoxib was monitored after the oral administration of
intestine ofwith
properties the least systemic
celecoxib. absorption. Next, wecelecoxib
examined if N-GA1C improved the
oralPK
celecoxib and N-GA1C toFirst, rats. the blood
To see howconcentration
the differencesof in the releasewas monitored
rate of celecoxib after
in the the
intestinal
properties
administration of celecoxib.
of celecoxib First, the blood
andcoefficient concentration
N-GA1C (indicating
to rats. To aqueous of celecoxib
see how solubility) was
the differences monitored after the oral
contents and the distribution affect in
thethe
PKrelease rate of
properties, the
administration
celecoxib of celecoxib
in the intestinal and N-GA1C
contents with
and the to rats. To see how the differences in the release
solubility)atof
rate
same experiment was performed A1Cdistribution
and G1C. After coefficient
the oral(indicating aqueous
administration of N-GA1C
celecoxib
affect the PKin the intestinalthe contents and the distribution coefficient (indicating aqueous solubility)
10, 20, and 40properties,
mg/kg, sustained same experiment
profiles was performed
of celecoxib concentration withwere
A1C and G1C.
observed inAfter the
the blood, oral
and
affect the
administration PKof properties,
N-GA1C atthe
10, same
20, andexperiment
40 mg/kg, was performed
sustained profileswith
of A1C and
celecoxib G1C. After
concentration the
were oral
the peak concentration of celecoxib (Cmax) was reduced from 1.18 µg/mL (with oral celecoxib at
administration of N-GA1C at 10, 20, and 40 mg/kg, sustained profiles of celecoxib
observed in the blood, and the peak concentration of celecoxib (Cmax) was reduced from 1.18 μg/mL concentration were
observed in the blood, and the peak concentration of celecoxib (Cmax) was reduced from 1.18 μg/mL
Pharmaceutics 2020, 12, 1043 8 of 13

10 mg/kg) to 0.33 µg/mL (with oral N-GA1C at 10 mg/kg) (Figure 4a,b, Table 2). The blood concentration
of celecoxib after the oral administration of N-GA1C was higher than that observed after the oral
administration of celecoxib from 6 h until the end of the experiment; the blood concentration of celecoxib
rapidly decreased after the oral administration of celecoxib. Although the blood concentration of
celecoxib was sustained for 24 h with both A1C and G1C, N-GA1C afforded greater concentration
of celecoxib (0.15–0.33 µg/mL) than the other celecoxib derivatives (A1C, 0.08–0.29 µg/mL and G1C,
0.10–0.31 µg/mL) (Figure 4c,d), suggesting that the more hydrophilic and faster the bioconversion of
a celecoxib-amino acid conjugate is, the higher the ability of the conjugate to sustain a high blood
Pharmaceutics 2020, 12, x FOR PEER REVIEW 9 of 13
concentration of celecoxib.

Figure 4. Celecoxib derivatives improve the PK properties of celecoxib. (a–e), Plasma concentration
ofFigure 4. Celecoxib
celecoxib after the derivatives improve theofPKcelecoxib
oral administration properties of N-GA1C
(a), celecoxib. (b),
(a–e),A1C
Plasma
(c), concentration
or G1C (d).
of celecoxib after the oral administration of celecoxib (a), N-GA1C (b), A1C (c), or
(e), Plasma concentration of celecoxib after the intravenous administration of celecoxib. (f–g),G1C (d). (e), Plasma
Absolute
concentration(f)ofand
bioavailability celecoxib after the intravenous
relative bioavailability administration
(g) of A1C, G1C, and N-GA1C of after
celecoxib.
the oral (f–g), Absolute
administration
ofbioavailability
celecoxib, A1C(f)(equivalent
and relativemass bioavailability
of celecoxib), (g)
G1Cof(equivalent
A1C, G1C, massandof N-GA1C
celecoxib),after the oral
or N-GA1C
administration of celecoxib, A1C (equivalent mass of celecoxib), G1C (equivalent
(equivalent mass of celecoxib) to Sprague-Dawley rats. Data are presented as mean ± s.e.m. from mass of celecoxib),
aor N-GA1C (equivalent
representative experiment mass = 5)
(n of celecoxib) to Sprague-Dawley
from 3 independent ** p <are
rats. Data
experiments. *** p < 0.001,
presented
0.01, as meanand±
**** p < from
s.e.m. a representative
0.0001, experiment
analyzed by two-way (n or
(a–d) = 5)one-way
from 3 independent
(e–f) ANOVAexperiments.
with Tukey’s ** HSD
p < 0.01, *** p <
multiple
0.001, and ****
comparison postphoc
< 0.0001,
test. analyzed by two-way (a–d) or one-way (e–f) ANOVA with Tukey’s HSD
multiple comparison post hoc test.

Table 2. Pharmacokinetic profiles of celecoxib, A1C, G1C, and N-GA1C. Absolute bioavailability and
relative bioavailability were calculated using the following equations: Absolute bioavailability (F, %)
= [(AUC0–∞)PO/dose PO]/[(AUC0–∞)IV/dose IV]; Relative bioavailability (%) = [(AUC0–∞)A/dose
A]/[(AUC0–∞)B/dose B].

Dose Cmax Tmax AUC0-24 AUC0-∞ CL Vd BA % t1/2 Relative

(mg/kg) (mg/mL) (h) (mg/mL)h (mg/mL)h (mL/min/kg) (L/kg) (f) (h) Bioavailability %
Celecoxib
10 - - 5.99 ± 0.44 6.12 ± 0.52 8.01 1.04 - 1.50
(IV)
Celecoxib 1.18 ± 75.82%
10 2 4.60 ± 0.45 4.64 ± 0.45 - - -
(Oral) 0.18 ± 2.66
Celecoxib 1.23 ± 42.40%
20 2 5.06 ± 0.49 5.19 ± 0.50 - - -
(Oral) 0.24 ± 3.94
Pharmaceutics 2020, 12, 1043 9 of 13

Table 2. Pharmacokinetic profiles of celecoxib, A1C, G1C, and N-GA1C. Absolute bioavailability and
relative bioavailability were calculated using the following equations: Absolute bioavailability (F, %)
= [(AUC0–∞ )PO/dose PO]/[(AUC0–∞ )IV/dose IV]; Relative bioavailability (%) = [(AUC0–∞)A/dose
A]/[(AUC0–∞ )B/dose B].

Relative
Dose Cmax Tmax AUC0–24 AUC0–∞ CL Vd BA % t1/2
Bioavailability
(mg/kg) (mg/mL) (h) (mg/mL)h (mg/mL)h (mL/min/kg) (L/kg) (f) (h)
%
Celecoxib 5.99 ± 6.12 ±
10 - - 8.01 1.04 - 1.50
(IV) 0.44 0.52
Celecoxib 1.18 ± 4.60 ± 4.64 ± 75.82%
10 2 - - -
(Oral) 0.18 0.45 0.45 ± 2.66
Celecoxib 1.23 ± 5.06 ± 5.19 ± 42.40%
20 2 - - -
(Oral) 0.24 0.49 0.50 ± 3.94
Celecoxib 1.38 ± 6.72 ± 7.12 ± 29.08%
40 2 - - -
(Oral) 0.24 0.58 0.69 ± 2.02
A1C 0.29 ± 3.68 ± 4.68 ± 76.47%
10 4 - - 100.85% ± 7.07
(Oral) 0.05 0.39 0.43 ± 2.48
A1C 0.43 ± 6.32 ± 8.70 ± 71.07% 167.62%±
20 4 - -
(Oral) 0.05 0.53 0.72 ± 2.51 12.84
G1C 0.31 ± 5.35 ± 7.22 ± 117.97%
10 2 - - 155.59% ± 7.16
(Oral) 0.06 0.44 0.61 ± 4.032
G1C 0.49 ± 9.06 ± 13.07 ± 106.78% 251.84%±
20 2 - -
(Oral) 0.05 0.85 1.11 ± 5.80 14.15
N-GA1C 0.33 ± 6.09 ± 9.78 ± 159.80% 210.76% ±
10 2 - -
(Oral) 0.05 0.49 0.84 ± 4.31 12.07
N-GA1C 0.51 ± 10.23 ± 18.55 ± 151.55% 357.43% ±
20 2 - -
(Oral) 0.05 0.94 1.19 ± 4.75 17.74
N-GA1C 0.80 ± 15.75 ± 33.87 ± 138.35% 476.76% ±
40 4 - -
(Oral) 0.04 1.12 3.69 ± 7.079 21.15

Next, we examined if the sustained blood concentration of celecoxib after the oral administration
of N-GA1C is linked to improving oral BA of celecoxib. To calculate the relative and absolute BA of
N-GA1C, area under the curve (AUC) values were acquired in advance after the intravenous and oral
administration of various doses of celecoxib (Figure 4a,e, and Table 2). For comparison with the other
celecoxib derivatives, the relative and absolute BA of A1C and G1C was calculated at 10 and 20 mg/kg.
As the oral doses of celecoxib increased from 10 to 40 mg/kg, the absolute BA of celecoxib
dramatically decreased from 75.82 ± 2.66% to 29.08 ± 2.02% (Figure 4a,e,f, and Table 2). However,
oral N-GA1C achieved greater absolute BA of celecoxib than oral celecoxib at all doses (p < 0.0001),
and no significant differences in absolute BA were observed among the different doses of N-GA1C
(Figure 4a,b,e,f and Table 2). A1C and G1C at 20 mg/kg increased the absolute BA of celecoxib by
up to approximately two- (for A1C) and three-fold (for G1C), whereas no increase or a 30% increase
in absolute BA was observed with oral A1C and oral G1C at 10 mg/kg, respectively, which was
significantly lesser than that of N-GA1C at the same dose (Figure 4a,c–f and Table 2). Moreover,
the relative BA of oral N-GA1C was determined to evaluate the improvement in the linearity of PK
profiles of celecoxib. With oral N-GA1C, the relative BA of celecoxib increased from 210.76 ± 12.07%
to 476.76 ± 12.07% (Figure 4a,b,f and Table 2) with the elevation of dose from 10 mg/kg to 40 mg/kg.
Increased relative BA was also observed with A1C and G1C as the doses were elevated from 10 mg/kg
to 20 mg/kg, and their dose proportionality was similar to that of N-GA1C (Figure 4g). Overall,
these results suggest that the celecoxib derivatives can improve the PK properties of celecoxib, and
N-GA1C is the most promising prodrug of celecoxib to overcome the challenges associated with the
PK of celecoxib.
 10 2 6.09 ± 0.49 9.78 ± 0.84 - - 210.76% ± 12.07
(Oral) 0.05 ± 4.31
N-GA1C 0.51 ± 10.23 ± 18.55 ± 151.55%
20 2 - - 357.43% ± 17.74
(Oral) 0.05 0.94 1.19 ± 4.75
N-GA1C 0.80 ± 15.75 ± 33.87 ± 138.35%
40 4 - - 476.76% ± 21.15
(Oral)
Pharmaceutics 2020, 12, 1043
0.04 1.12 3.69 ± 7.079 10 of 13

3.6. Oral N-GA1C Prolongs and Enhances the Anti-Inflammatory Activity of Celecoxib
3.6. Oral N-GA1C Prolongs and Enhances the Anti-Inflammatory Activity of Celecoxib
Although oral N-GA1C significantly improved the PK properties of oral celecoxib, it is not
Although
certain that suchoral N-GA1C
improvement significantly
confersimproved
therapeuticthebenefits
PK properties of oral celecoxib,
on celecoxib. Thus, weit compared
is not certainthe
that such improvement
therapeutic confers therapeutic
(anti-inflammatory) activity ofbenefits on celecoxib.
oral N-GA1C with thatThus, we celecoxib
of oral comparedinthe therapeutic
a carrageenan-
(anti-inflammatory)
induced rat paw edema activity
modelof oral N-GA1C
[17,18]. Exceptwith thatlinearity
for the of oral of
celecoxib in a carrageenan-induced
the PK profiles, N-GA1C exhibited
rat paw edema
greater absolute model
BA and[17,18]. Except
higher for the concentration
sustained linearity of theofPK profiles, in
celecoxib N-GA1C
the bloodexhibited
than the greater
other
absolute
celecoxib BAderivatives.
and higherTo sustained
determineconcentration
the relevance of celecoxib
of PK data in the toblood than the
therapeutic other celecoxib
activity, the same
derivatives.
experimentTo wasdetermine
performed thewith
relevance
G1C and of PK dataThe
A1C. to therapeutic activity,activity
anti-inflammatory the same experiment
was determined wasby
performed
calculating with G1C reduction
volume and A1C. of Theratanti-inflammatory
paw edema after the activity
oral was determinedofbyeach
administration calculating
drug (10volume
mg/kg).
reduction
With oralofcelecoxib,
rat paw edema after the
the volume oral administration
reduction of paw edema of each drug
reached the(10maximum
mg/kg). With h (42.06 
oral6celecoxib,
around
the volume
4.69%) andreduction of paw
continuously edema reached
diminished to 20.99  2.791% after
the maximum around
24 h (42.06 ±
6 h(Figure 5).4.69%) and continuously
Oral N-GA1C achieved
diminished
maximal volume ± 2.791% after
to 20.99 reduction of paw24 hedema
(Figure(78.96
5). Oral  N-GA1C
1.56%) atachieved
24 h, which maximal wasvolume
greaterreduction
than that
ofachieved
paw edema by (78.96 ± 1.56%) at
oral celecoxib. 24 h, which
However, thewas greater
volume than thatachieved
reduction achieved by oral celecoxib.
N-GA1C h (32.42 
at 6 However,
the volume
2.50%) wasreduction achieved
less than that by N-GA1C
by celecoxib (Figureat5).6hOral G1C±and
(32.42 2.50%)
A1Cwas less than
showed that pattern
a similar by celecoxib
to oral
(Figure
N-GA1C 5). in
Oral G1Cofand
terms A1C
their showed
ability a similar
to reduce pattern to
the volume of oral N-GA1C
rat paw edema. in terms of their the
G1C reduced ability to
edema
reduce
volume thetovolume
a greater ofextent
rat paw edema.
than G1C reduced
oral celecoxib from the12 h,edema volume
whereas A1C to a greater
could extent than
only achieve this oral
at 24
celecoxib fromcelecoxib
h. The other 12 h, whereas A1C could
derivatives wereonlylessachieve
effective this
in at 24 h. The
reducing theother
volumecelecoxib
of ratderivatives
paw edemawere than
less effective in reducing the volume of rat paw edema than N-GA1C
N-GA1C at all time points. Compared with celecoxib and the other celecoxib derivatives, oral N-at all time points. Compared
with
GA1C celecoxib
showed andgreater
the other celecoxib derivatives,
anti-inflammatory oral(Figure
activity N-GA1C 5).showed
Moreover, greatertheanti-inflammatory
anti-inflammatory
activity (Figure
activities of the5).celecoxib
Moreover, the anti-inflammatory
derivatives correlated with activities
their PKofproperties
the celecoxib suchderivatives
as BA andcorrelated
sustained
with their PK properties
the concentration such asinBA
of celecoxib theand sustained the concentration of celecoxib in the blood.
blood.

Figure 5. N-GA1C prolongs and enhances the anti-inflammatory activity of celecoxib. (a,b), SD rats
were orally
Figure administered
5. N-GA1C celecoxib
prolongs (10 mg/kg),
and enhances theA1C (equivalent mass
anti-inflammatory of celecoxib),
activity G1C (a,b),
of celecoxib. (equivalent
SD rats
mass of celecoxib), N-GA1C (equivalent mass of celecoxib), or IPB. (a), Changes in
were orally administered celecoxib (10 mg/kg), A1C (equivalent mass of celecoxib), G1C (equivalentrat paw volume
inmass
eachofgroup for 24 N-GA1C
celecoxib), h. (b), Percentage
(equivalentchanges
mass ofincelecoxib),
edema inhibition in each
or IPB. (a), group
Changes infor
rat 24
pawh. volume
Data arein
presented
each group for 24±h.
as mean s.e.m. from a representative
(b), Percentage changes inexperiment (n = 5) from
edema inhibition 2 independent
in each group for 24experiments.
h. Data are
< 0.05, **as
* ppresented p< 0.01,±and
mean s.e.m. p < a0.0001,
****from analyzedexperiment
representative by two-way(nANOVA
= 5) fromwith Tukey’s HSD
2 independent multiple
experiments.
comparison post hoc test.
* p < 0.05, ** p < 0.01, and **** p < 0.0001, analyzed by two-way ANOVA with Tukey’s HSD multiple
comparison post hoc test.
4. Discussion
Celecoxib, sold under the brand name Celebrex, is a COX-2 selective nonsteroidal
anti-inflammatory drug widely used for the treatment of pain and inflammation in chronic
inflammatory diseases such as osteoarthritis, rheumatoid arthritis, and ankylosing spondylitis [1,2].
Despite its extensive clinical use, there are some challenges associated with the PK of celecoxib,
such as low BA, poor linearity in PK profiles, and peak blood concentration-related cardiovascular
toxicity, thereby deterring broader clinical usage [2–6]. To address these issues, we designed celecoxib
prodrugs, namely A1C, G1C, and N-GA1C, with high hydrophilicity and the ability to convert to
celecoxib in the intestine (Scheme 1).
Consistent with our previous studies, the conjugation of celecoxib with amino acids via the
amide bond of the sulfonamide group in celecoxib and the carboxyl group in amino acids yielded
Pharmaceutics 2020, 12, 1043 11 of 13

prodrugs bio-activated in the small and large intestine [10–15]. This is shown in our data demonstrating
that A1C, G1C, and N-GA1C liberated celecoxib in the small and large intestinal contents upon
incubation. The derivatives were stable in buffer solutions of pH 1.2 and 6.8, indicating chemical
stability and enzyme susceptibility of the amide bonds. The enzymatic susceptibility of the prodrugs
was dependent on the amino acid promoieties as intestinal bioconversion of the celecoxib derivatives
occurred at different rates (N-GA1C > G1C > A1C). This result also indicates that the dipeptide
promoiety (N-glycyl-aspart-1-yl) in N-GA1C is not bio-activated to celecoxib by one-by-one cleavage to
form A1C as the intermediate. Actually, the presence of A1C was not observed during the incubation
of N-GA1C in the intestinal contents.
The cleavage of the celecoxib derivatives to celecoxib in the small- and large-intestinal contents
and the hydrophilicity of the celecoxib derivatives, indicated by their lower distribution coefficient
than that of celecoxib, suggest that oral celecoxib derivatives constantly release celecoxib during transit
in the small and large intestine. The release of celecoxib occurs with limited systemic absorption of
the celecoxib derivatives, especially for N-GA1C, which has the lowest DC. Accordingly, compared
with oral celecoxib, the oral celecoxib derivatives could more effectively sustain the concentration of
celecoxib in the blood for 24 h. Furthermore, the concentration of celecoxib after the oral administration
of N-GA1C was higher than that measured after the oral administration of G1C and A1C at all time
points. This is consistent with the release profiles of celecoxib from the celecoxib derivatives in the
intestinal contents, where N-GA1C released celecoxib faster than the other derivatives. The data
showing that oral G1C achieved greater concentration of celecoxib in the blood than oral A1C further
supports our findings.
The celecoxib derivatives increased the BA of celecoxib as determined by AUC. Our data show that
the BA of the derivatives was achieved in the following order, N-GA1C > G1C > A1C, suggesting that
the BA of celecoxib depends on both the hydrophilicity and the rate of celecoxib release in the intestine.
It is likely that N-GA1C provides higher concentration of celecoxib available for systemic absorption
than the other derivatives; this is due to greater hydrophilicity, resulting in better dissolution and
faster release of celecoxib. At the same time, sustained release of celecoxib from N-GA1C in the
intestine may reduce the precipitation of celecoxib due to low solubility. This argument is reasonable
considering that celecoxib is a BCS class II drug, with low solubility and high permeability; thus,
an increase in the amount of celecoxib available for absorption increases the systemic absorption.
Furthermore, it is not surprising that the BA of celecoxib was greater with oral G1C than with A1C,
given that the bioconversion of G1C (to celecoxib) occurred faster than that of A1C and DCs of the
two derivatives were similar.
The systemic absorption of N-GA1C is lower than that of the other celecoxib derivatives owing to
its lowest DC; hence, contribution of hepatic N-GA1C (absorbed from intestine) metabolism toward
BA increase would be negligible, unlike G1C and A1C whose systemic absorption may be more than
N-GA1C. In addition, among the celecoxib derivatives, the most hydrophilic N-GA1C has the least risk
of systemic side effects. Our data showed that N-GA1C is the most susceptible to hepatic metabolism,
thus further reducing the risk of systemic side effects of the prodrug [7,8].
Our data showing that an increase in the dose of N-GA1C from 10 mg/kg to 40 mg/kg was
accompanied by proportional increase in PK parameters such as Cmax and AUC indicate that
N-GA1C improves the BA as well as the linearity of the PK profile of celecoxib. This suggests that
the pharmaceutical behavior of N-GA1C in the intestine is not significantly changed in the dose
range. Considering that a linear PK profile facilitates the prediction of concentration-time profiles and
thus enables the adjustment of dose and dose regimen in patients, N-GA1C may be more suitable
for therapeutic use than celecoxib, especially for patients who need dose adjustment for celecoxib.
Although A1C and G1C also showed similar proportionality with regard to dose change, their PK
values were less than those of N-GA1C. The difference in PK improvement shown by the celecoxib
derivatives is associated with their anti-inflammatory activities.
Pharmaceutics 2020, 12, 1043 12 of 13

The improvement in the PK properties of celecoxib positively affects its therapeutic activity and
duration of action. This is supported by the data demonstrating that, compared with oral celecoxib, oral
N-GA1C exhibited greater anti-inflammatory activity for a longer duration in the carrageenan-induced
rat paw edema model, which was consistent with the concentration profile of celecoxib in the blood
after the oral administration of N-GA1C and celecoxib. Furthermore, of all the derivatives, N-GA1C
exhibited the highest ability to reduce rat paw edema at all time points, ensuring close association of
PK improvement with therapeutic activity and duration.
Taken together, the celecoxib derivatives, N-GA1C, G1C, and A1C, are prodrugs of celecoxib
with improved PK properties that can be largely ascribed to the continuous bioconversion of the
celecoxib derivatives to celecoxib during transit through the intestine, thereby leading to the sustained
absorption of celecoxib. Compared with celecoxib and the other derivatives, N-GA1C exhibited
superior anti-inflammatory activity and lower risks of side effects due to limited systemic absorption;
hence, it is the most promising prodrug of celecoxib. The prodrug strategy may be an effective
approach to improve the PK and therapeutic properties of other BCS class II hydrophobic drugs with
pharmaceutical limitations similar to those of celecoxib.

Author Contributions: Conceptualization, Y.L. and Y.J.; methodology, Y.L., J.K., W.K., and Y.J.; validation, Y.L.,
W.K., and Y.J.; formal analysis, Y.L.; investigation, Y.L., J.K., W.K., and Y.J.; data curation, Y.L.; writing—original
draft preparation, Y.L.; writing—review and editing, Y.L., I.-S.Y., and Y.J.; visualization, Y.L.; supervision, Y.J.;
project administration, Y.L. and Y.J.; funding acquisition, Y.L. and Y.J. All authors have read and agreed to the
published version of the manuscript.
Funding: The study was supported by the Ewha Womans University Research Grant of 2020.
Conflicts of Interest: Authors declare no conflict of interest.

References
1. FitzGerald, G.A.; Patrono, C. The coxibs, selective inhibitors of cyclooxygenase-2. N. Engl. J. Med.
2001, 345, 433–442. [CrossRef] [PubMed]
2. Nissen, S.E.; Yeomans, N.D.; Solomon, D.H.; Luscher, T.F.; Libby, P.; Husni, M.E.; Graham, D.Y.; Borer, J.S.;
Wisniewski, L.M.; Wolski, K.E.; et al. Cardiovascular Safety of Celecoxib, Naproxen, or Ibuprofen for
Arthritis. N. Engl. J. Med. 2016, 375, 2519–2529. [CrossRef] [PubMed]
3. Fong, S.Y.; Ibisogly, A.; Bauer-Brandl, A. Solubility enhancement of BCS Class II drug by solid phospholipid
dispersions: Spray drying versus freeze-drying. Int. J. Pharm. 2015, 496, 382–391. [CrossRef] [PubMed]
4. Lakshmi, K.; Reddy, M.P.; Kaza, R. Design and characterization of microcrystals for enhanced dissolution
rate of celecoxib. Curr. Drug Discov. Technol. 2013, 10, 305–314. [CrossRef]
5. Paulson, S.K.; Vaughn, M.B.; Jessen, S.M.; Lawal, Y.; Gresk, C.J.; Yan, B.; Maziasz, T.J.; Cook, C.S.; Karim, A.
Pharmacokinetics of celecoxib after oral administration in dogs and humans: Effect of food and site of
absorption. J. Pharm. Exp. 2001, 297, 638–645.
6. PFIZER. CELEBREX (Celecoxib capsules). FDA Medicat. Guide 1998, 1, 1–22.
7. Rautio, J.; Kumpulainen, H.; Heimbach, T.; Oliyai, R.; Oh, D.; Jarvinen, T.; Savolainen, J. Prodrugs: Design
and clinical applications. Nat. Rev. Drug Discov. 2008, 7, 255–270. [CrossRef] [PubMed]
8. Rautio, J.; Meanwell, N.A.; Di, L.; Hageman, M.J. The expanding role of prodrugs in contemporary drug
design and development. Nat. Rev. Drug Discov. 2018, 17, 559–587. [CrossRef] [PubMed]
9. Lee, Y.; Jung, E.H.; Kim, H.; Yoon, J.-H.; Kim, D.-D.; Jung, Y. Preparation and in vitro evaluation of
celecoxib-amino acid conjugates as a colon specific prodrug. J. Pharm. Investig. 2012, 42, 115–120. [CrossRef]
10. Lee, S.; Lee, Y.; Kim, W.; Nam, J.; Jeong, S.; Yoo, J.W.; Kim, M.S.; Moon, H.R.; Jung, Y. Evaluation of
glycine-bearing celecoxib derivatives as a colon-specific mutual prodrug acting on nuclear factor-kappaB, an
anti-inflammatory target. Drug Des. Devel. 2015, 9, 4227–4237. [CrossRef]
11. Lee, Y.; Kim, H.; Kim, W.; Yoon, J.H.; Jeong, S.H.; Jung, Y. Colon-specific delivery of celecoxib is a potential
strategy to improve toxicological and pharmacological properties of the selective Cox-2 inhibitor: Implication
in treatment of familiar adenomatous polyposis. J. Drug Target. 2012, 20, 524–534. [CrossRef] [PubMed]
Pharmaceutics 2020, 12, 1043 13 of 13

12. Lee, Y.; Kim, J.; Kim, H.; Kang, S.; Yoon, J.H.; Kim, D.D.; Kim, Y.M.; Jung, Y. N-succinylaspart-1-yl celecoxib
is a potential colon-specific prodrug of celecoxib with improved therapeutic properties. J. Pharm. Sci.
2012, 101, 1831–1842. [CrossRef] [PubMed]
13. Kim, W.; Lee, Y.; Jeong, S.; Nam, J.; Lee, S.; Jung, Y. Colonic delivery of celecoxib is a potential pharmaceutical
strategy for repositioning the selective COX-2 inhibitor as an anti-colitic agent. Arch. Pharm. Res.
2015, 38, 1830–1838. [CrossRef] [PubMed]
14. Lee, Y.; Kim, J.; Kim, W.; Nam, J.; Jeong, S.; Lee, S.; Yoo, J.W.; Kim, M.S.; Jung, Y. Celecoxib coupled to
dextran via a glutamic acid linker yields a polymeric prodrug suitable for colonic delivery. Drug Des. Devel.
2015, 9, 4105–4113. [CrossRef]
15. Lee, Y.; Kim, W.; Hong, S.; Park, H.; Yum, S.; Yoon, J.H.; Jung, Y. Colon-targeted celecoxib ameliorates
TNBS-induced rat colitis: A potential pharmacologic mechanism and therapeutic advantages. Eur. J. Pharm.
2014, 726, 49–56. [CrossRef] [PubMed]
16. Zhang, Y.; Huo, M.; Zhou, J.; Xie, S. PKSolver: An add-in program for pharmacokinetic and pharmacodynamic
data analysis in Microsoft Excel. Comput. Methods Programs Biomed. 2010, 99, 306–314. [CrossRef] [PubMed]
17. Morris, C.J. Carrageenan-induced paw edema in the rat and mouse. Methods Mol. Biol. 2003, 225, 115–121.
[CrossRef] [PubMed]
18. Guay, J.; Bateman, K.; Gordon, R.; Mancini, J.; Riendeau, D. Carrageenan-induced paw edema in rat elicits a
predominant prostaglandin E2 (PGE2) response in the central nervous system associated with the induction
of microsomal PGE2 synthase-1. J. Biol. Chem. 2004, 279, 24866–24872. [CrossRef] [PubMed]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional
affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).