Novel phosphoramidate prodrugs of N-acetyl-(d)-glucosamine with antidegenerative activity on bovine and human cartilage explants

Article abstract of DOI:10.1021/jm300074y

(d)-Glucosamine and other nutritional supplements have emerged as safe alternative therapies for osteoarthritis (OA), a chronic and degenerative articular joint disease. In our preceding paper, a series of novel O-6 phosphate N-acetyl (d)-glucosamine prod

Full text of DOI:10.1021/jm300074y

Article
pubs.acs.org/jmc
Novel Phosphoramidate Prodrugs of N-Acetyl-(D)-Glucosamine with
Antidegenerative Activity on Bovine and Human Cartilage Explants
Michaela Serpi,† Rita Bibbo,† Stephanie Rat,† Helen Roberts,‡ Claire Hughes,‡ Bruce Caterson,‡
María Jose Alcaraz,§ Anna Torrent Gibert,∥ Carlos Raul Alaez Verson,∥ and Christopher McGuigan*,†
́
†Welsh School of Pharmacy, Cardiff University, Cardiff, King Edward VII Avenue, Cardiff CF10 3NB, U.K.
‡Cardiff School of Biosciences, Cardiff university, Biomedical Building, Museum Avenue, Cardiff CF10 3US, U.K.
§Bioberica SA, Plaza Francesco Macia
̀
7, Barcelona 08029, Spain
∥Department of Pharmacology and IDM, University of Valencia, Avenida Vicent A. Estelles s/n, 46100 Burjasot, Valencia, Spain
S
* Supporting Information
ABSTRACT: (D)-Glucosamine and other nutritional supplements
have emerged as safe alternative therapies for osteoarthritis (OA), a
chronic and degenerative articular joint disease. In our preceding paper,
a series of novel O-6 phosphate N-acetyl (D)-glucosamine prodrugs
aimed at improving the oral bioavailability of N-acetyl-(D)-glucosamine
as its putative bioactive phosphate form were shown to have greater
chondroprotective activity in vitro when compared to the parent agent.
In order to extend the SAR studies, this work focuses on the O-3
and O-4 phosphate prodrugs of N-acetyl-(D)-glucosamine bearing a
4-methoxy phenyl group and different amino acid esters on the
phosphate moiety. Among the compounds, the (L)-proline amino acid-
containing prodrugs proved to be the most active of the series, more
effective than the prior O-6 compounds, and well processed in
chondrocytes in vitro. Data on human cartilage support the notion that
these novel O-3 and O-4 regioisomers may represent novel promising leads for drug discovery for osteoarthritis.
importantly, it is a precursor in the biosynthesis of the GAGs1
INTRODUCTION
2-Amino-2-deoxy-(D)-glucose ((D)-glucosamine, 1) (Figure 1)
is an amino monosaccharide derivative of (D)-glucose. It is
■
that, covalently linked with a protein core, form the
proteoglycans (PGs). Proteoglycans2 are major components
of the extracellular matrix of the articular cartilage and, together
found in numerous biologically potent molecules such as cell
with the chondrocytes, the cellular components play a pivotal
surface N-glycoproteins, hyaluronic acid, glycosphingolipids
role in the functioning of joints.3 Failure of chondrocytes to
(Lewis a/x), glycosylphosphatidylinositol (GPI) anchors,
maintain the balance between synthesis and degradation of the
blood group antigens, bacterial cell wall, lipopolysaccharides,
chitin/chitosan, and glycosaminoglycan (GAGs) chains. Most
extracellular matrix may lead to osteoarthritis (OA), which is a
degenerative and progressive joint disorder characterized
by joint pain, tenderness, limitation of movement, crepitus,
occasional effusion, and variable degrees of joint inflammation.4
Cartilage is responsible for providing the shock-absorption at
the end of bones, and it is indeed the loss of this cushioning
effect due to the loss of GAGs that results in pain and loss of
range of movement. Osteoarthritis (OA) is a major cause of
disability and, being strongly associated with aging, its medical
relevance is rising given the increasing proportion of older
people. It can be treated with analgesics and nonsteroidal
anti-inflammatory drugs, but these drugs can cause serious
gastrointestinal and cardiovascular adverse events, especially
with long-term use,5 and they do not address the underylying
physiology of the disease. Disease-modifying drugs that
Figure 1. Structures of (D)-glucosamine (1), N-acetyl-(D)-glucosamine
(2), and general structure of the O-6 aryloxy phosphoramidates of 2
(3). Stereochemistry at the phosphorus and amino acid α-carbon is
omitted.
Received: January 18, 2012
Published: April 13, 2012
© 2012 American Chemical Society
4629
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
a
Scheme 1
a
Reagents and conditions: (a) Phosphorochloridate, NMI, pyridine, THF, −30 °C to ambient, 5 h.
interfere with the progression of the condition would be
strongly desired.
phosphate prodrugs, we decided to investigate the regioiso-
meric O-3 and O-4 prodrugs of 2. These were originally
identified as synthetic byproduct, but we herein report the
surprising biological activity of these analogues.
Worldwide, over the past 10 years, products containing
chondroitin (GAGs) and 1 have been increasingly recom-
mended by general practitioners and rheumatologists, and
nowadays 1 (either as sulfate or chloride salts) represents one
of the most commonly used agents to treat OA.6−11 Despite
the increased use of 1 in the treatment of OA, the mechanisms
accounting for its in vivo and in vitro activity are still poorly
understood. The most common notion is that augmenting the
intake of the precursor molecule 1 may directly stimulate
articular PG synthesis,12,13 although the results of these
investigations have been controversial.14 Different studies
have also presented experimental evidence that 1 and, to a
larger degree, N-acetyl-(D)-glucosamine (2) (Figure 1), possess
a unique range of anti-inflammatory activities.15 2 is one of the
building blocks for glycoproteins, formed from (D)-glucose
and/or 1 during the in vivo biosynthesis of PGs.1 (D)-Glucose
is a general precursor for cellular GAG biosynthesis. Inside the
cell, (D)-glucose is converted into (D)-glucose 6-phosphate and
(D)-fructose-6-phosphate. The conversion of fructose-6-
phosphate to glucosamine-6-phosphate takes place by the
enzyme glutamine fructose-6-phosphate aminotransferase.
(D)-Glucosamine-6-phosphate is rapidly converted into the
N-acetyl-(D)-glucosamine-6-phosphate by acetyl-CoA glucos-
amine-6-phosphate N-acetyltransferase. However, exogenous 1
or 2 supplemented to the cultured cells can enter this metabolic
pathway by conversion into (D)-glucosamine-6-phosphate.
N-Acetyl-glucosamine-6-phosphate is further converted via
N-acetylglucosamine-1-phosphate into uridine diphosphate
(UDP) N-acetyl-(D)-glucosamine, and by epimerase into
UDP-N-acetyl-(D)-galactosamine.1 These nucleotide-activated
sugars, together with UDP-glucuronic acid, are utilized in the
assembly of GAG chain. By analogy to 1, 2 has been found to
decrease pain and inflammation and to aid cartilage repair,
increasing the range of motion in osteoarthritic patients.16
One of the aims in the development of our phosphate
prodrugs was to deliver the putative bioactive O6 mono-
phosphate directly and to increase the lipophilicity of 1 which,
due to the highly polar character (ClogD(pH=6.50) = −3.93),17
has an inadequate oral bioavailability (∼5−20%).18−20 Our
group had previously investigated a family of aryloxy
phosphoramidates at the O-6 position of 2 (see general
structure 3 in Figure 1).21 Several of these prodrugs at
concentrations as low as 0.1 mM were noncytotoxic and ca. ≥
100-fold more active than the parent compound in a biological
assay for inhibition of interleukin 1 (IL-1) induced GAG release
from bovine articular cartilage in in vitro explant cultures.
On the basis of these encouraging results and as a
continuation of our study on N-acetyl-(D)-glucosamine mono-
RESULTS AND DISCUSSION
■
Chemistry. We have previously reported21 that reaction of
2 with different phosphorochloridates in THF/pyridine in the
presence of N-methyl imidazole (NMI) at −30 °C to ambient
yields a complex mixture of regio (mono phosphorylated at
O-3, O-4, or O-6) and stereoisomers (Rp and Sp, α and β
anomers). In the first instance, we were able to isolate from the
mixtures pure O-6 derivatives in a range of yield between ≤1−
21%. Besides the desired O-6 isomers, formation of the O-3
and O-4 regioisomers were also detected, but at such time they
were not isolated in pure form (Scheme 1).
More recently, we were able to isolate the O-3 and O-4
phosphoramidate regioisomers in addition to the O-6 prodrugs
by reacting 2 with a series of freshly prepared phosphoro-
chloridates (6a−v) in the presence of NMI in THF/pyridine at
−30 °C to ambient. At first, except for 9d and 9o, which were
isolated as pure O-3 isomers, O-4 and O-3 phosphoramidate
regioisomers were recovered as mixtures (8a−v/9a−v) in a
range of isolated yields between ≤1−27% (Scheme 2, Table 1).
According to the observed ratios of regioisomers, the formation
of O-4/O-3 isomers was in general more favored than the O-6
derivatives, which were found to be predominant only in
reactions with α,α-dimethylglycine or (L)-proline amino acid-
containing phosphorochloridates. 1H NMR spectra of the O-4/
O-3 isomer mixtures suggested that the O-3 derivatives were
always the predominant regioisomers. This result is in good
accordance with the literature, which provides much evidence
of the poor reactivity of the 4-OH22 of 2, when compared to
the 3-OH. Several explanations have been provided, including
steric hindrance at this position,22 formation of either
intermolecular or intramolecular hydrogen bonding, involving
the glucosamine amide group and lowering the reactivity of
the hydroxyl groups,23 and NH group acting as a competitive
nucleophile.24
There was no significant increase in yield or any major
change in the ratio of regioisomers when the temperature was
varied from −30 °C to room temperature.
Preliminary biological evaluation indicates that several of the
O-4/O-3 regioisomeric mixtures (8a−s/9a−s) are able to
significantly reduce the in vitro IL-1 induced GAG release (see
the later section on Biological Activity), suggesting that testing
O-3 and O-4 regioisomers individually was essential. Separation
of O-4/O-3 isomers by extensive and repeated column
chromatography was achieved in a few cases (Table 2).
4630
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
a
Scheme 2
a
Reagents and conditions: (a) Dowex 50 WXS 200, methanol, reflux, 24h; (b) NMI, pyridine, THF, −30 °C to rt, 5 h. O-6 isomers are also formed
(not shown).
Table 1. Substitution Pattern and Isolated Yields of N-Acetyl-(D)-Glucosamine O-4/O-3 Phosphoramidate Derivatives 8a−v/
9a−v and 1-O-Methyl-N-Acetyl-(D)-Glucosamine O-3 Phosphoramidate Derivatives 10a−c,m,o
compd
AA
R′
yield (%)
compd
AA
R′
yield (%)
8a/9a
8b/9b
8c/9c
9d
(L)-Ala
(L)-Val
(L)-Pro
(L)-Phe
(L)-Leu
(L)-ILe
(Me)2Gly
Gly
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Et
16
28
2
9o
(L)-Sar
(L)-Pro
(L)-Val
(L)-Val
(L)-Val
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Ala
(L)-Val
(L)-Pro
(L)-Val
(L)-Sar
Et
2
8
8p/9p
8q/9q
8r/9r
8s/9s
8t/9t
8u/9u
8v/9v
10a
nBu
tBu
9
3
cyclohex
2-Bu
iPr
5
8e/9e
8f/9f
16
13
7
5
1.5
2.4
1.8
13
3
8g/9g
8h/9h
8i/9i
2-Bu
cyclohex
Bn
6
(L)-Met
(D)-Ala
(L)-Gly
(L)-Pro
(L)-Val
(L)-Val
5
8j/9j
1
10b
Bn
8k/9k
8l/9l
1
10c
Et
3
Et
1
10m
10o
Et
0.3
1
8m/9m
8n/9n
Et
5
Et
iPr
9
Table 2. Substitution Pattern and Isolated Yields of N-Acetyl-(D)-Glucosamine O-3 Phosphoramidate Derivatives 8c,l,p,t−v and
N-Acetyl-(D)-Glucosamine O-4 Phosphoramidate Derivatives 9c,l,p,t−v
compd
AA
R′
yield (%)
compd
AA
R′
yield (%)
8c
9c
8l
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
Bn
Bn
Et
0.5
1.3
0.8
1.3
1.0
1.1
8t
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
iPr
iPr
05
9t
1.0
0.8
1.6
0.5
1.3
8u
9u
8v
9v
2-Bu
9l
Et
2-Bu
8p
9p
nBu
nBn
cyclohex
cyclohex
However, because the separation proved to be slow and
inefficient, it was decided to investigate a regioselective syn-
thetic procedure. Considering that the 4-OH is reported to be
the least reactive secondary OH group and that the O-3
phosphoramidates appeared to be the most abundant in the
mixture, our initial effort was focused on obtaining the O-3
regioisomers selectively.
Interestingly, it has been demonstrated that in a sugar moiety
the hydroxyl group reactivity pattern usually depends upon the
nature of the surrounding protecting groups. In the case of
glucosamine, many publications showed the strong influence of
the protecting groups at the O-6 and N-2 positions as well as
the configuration of the anomeric carbon on the observed
regioselectivity of reaction.25−27 In particular, it has been noted
that bulky protecting groups at O-6 position could hinder the
topside of the 3-OH,28,29 whereas large N-2 groups may impart
steric hindrance on the neighboring O-3 position,30 in both
cases preventing the possibility to approach the 3-OH. Pre-
vious studies26 have also shown that O-6 protected methyl
α-glycoside derivatives tend to give preferentially O-3 substitution,
4631
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
a
Scheme 3
a
Reagents and conditions: (a) cat. pTsOH, PhCH(OMe)2, DMF, 55 °C 24h; (b) NMI, pyridine, THF, −30 °C to rt, 5 h or 1M tBuMgCl, THF, rt,
5 h; (c) I2, MeOH at reflux or H2, Pd/C, EtOH, 55 psi rt.
whereas the corresponding β anomers gave mainly substitution
at the O-4 position.
regioisomer was produced, however, there was not a
satisfactory increase in yield for the coupling reaction (3.4−
8.4% yield).
As a first attempt to pursue selective phosphorylation at the
O-3 over the O-4 position, protection at the O-1 position of 2,
favoring mainly formation of the α anomer, was investigated.
Thus 1-O-methyl-N-acetyl-(D)-glucosamine (7) (ratio α/β 8:1),
selected for these studies, was synthesized by treatment of 2
with Dowex 50 WXS 200 in methanol at reflux, and then it
was reacted with different phosphorochloridates using the same
conditions described above (Scheme 2, Table 1). In all cases,
besides the O-6 derivatives (data not shown), exclusively O-3
regioisomers were obtained only as α anomers (10a−c,m,o) as
determined on the basis of the C-3-P and H-1−H-2 coupling
constant values (J(C‑3‑P) ∼7.0 Hz and J(H‑1, H‑2) = 3.0 Hz). No
traces of O-4 derivatives were detected. Because the O-6
isomers were predominant, regardless of which amino acid-
containing phosphorochloridate was used, it is clear how mani-
pulation of the protecting group on a sugar moiety can not only
increase the usual selectivity (3-OH ≫ 4-OH) but also reverse
it (6-OH > 3-OH). Unfortunately, no significant improvement
in the yield of O-3 phosphate regioisomers was achieved
(≤1−13%).
Therefore, in view of these results and in order to achieve
regioselective phosphorylation, a regioselective-protecting
strategy had to be applied. It was decided to introduce a 4,6-
O-benzylidene protecting group onto the sugar, the advantages
of which were to block both the O-4 and O-6 positions. To this
end, 2 was transformed into the N-acetyl-4,6-O-benzylidene-
(D)-glucosamine (12) by treatment with benzaldehyde
dimethylacetal in the presence of p-toluenesulfonic acid
(p-TsOH) as catalyst [Scheme 3].31 Then derivatives 9c and 9t
were synthesized, reacting 12 with the appropriate phosphoro-
chloridate (6c or 6t) in THF/pyridine at −30 °C to rt,
followed by either treatment with I2 in methanol at reflux32 or
catalytic hydrogenation under pressure at rt (Scheme 3, Table 2).
By following the latter strategy, only the desired O-3
As a final synthetic strategy, it was decided to protect the
1-OH, 4-OH, and 6-OH groups simultaneously. Therefore, the
reaction of 1-O-benzyl-N-acetyl-(D)-glucosamine (11) with
benzaldheyde dimethylacetal in the presence of p-TsOH as a
catalyst afforded regioselectively 1-O-benzyl-4,6-di-O-benzylidene-
N-acetyl-(D)-glucosamine (13) (ratio α/β 8:2) (Scheme 3).31
Having protected all the hydroxyl groups, which could possibly
compete with the 3-OH for phosphorylation, it was decided to
use the Grignard reagent, tert-butyl magnesium chloride
(tBuMgCl), as base. Because of the higher solubility of 13 in
organic solvent when compared to 2, it was possible to replace
pyridine with tetrahydrofuran (THF). Reaction of 13 with
different phosphorochloridates (6u−y) furnished O-4, O-6-
protected O-3 phosphoramidates 15u−y in moderate yields
(30−69%) (Scheme 3, Table 3). Both the 4,6-di-O-benzylidene
and benzyl groups were removed in one pot by hydrogenation
under pressure in the presence of a palladium catalyst. In some
cases, partial deprotection along with decomposition was
observed, causing lowering of the yield and requiring further
purification on silica gel in order to afford the final compounds
9u−y.
Biological Activity. To examine the effect of the gluco-
samine derivatives on the loss of aggrecan (cartilage
proteoglycan) from articular cartilage (an early event in the
development of matrix degradation in osteoarthritis), we used
the previously established21 in vitro culture system where
cartilage explants were exposed to IL-1 (to induce the loss of
aggrecan from the tissue) in the presence or absence of the
glucosamine derivatives. Exposure of bovine articular cartilage
explant cultures to IL-1 led to an approximately 3- to 4-fold
increase in levels of glycosaminoglycan (GAG) released (which
is representative of a loss of aggrecan from the tissue due to
degradation by matrix proteases) relative to untreated controls.
4632
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
Table 3. Substitution Pattern and Isolated Yields of Protected N-Acetyl-(D)-Glucosamine O-3 Phosphoramidate Derivatives
14c,t and 15u−y and N-Acetyl-(D)-Glucosamine O-3 Phosphoramidate Derivatives 9c,t−y
compd
R″
AA
R′
yield (%)
compd
R′′
AA
R′
yield (%)
14c
14t
H
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-NorVal
(L)-Pro
(L)-Val
Bn
8
3
9c
9t
H
H
H
H
H
H
H
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-NorVal
(L)-Pro
(L)-Val
Bn
20
70
30
12
77
10
30
H
iPr
iPr
15u
15v
15w
15x
15y
Bn
Bn
Bn
Bn
Bn
2-Bu
cyclohex
Et
30
41
69
47
38
9u
9v
9w
9x
9y
2-Bu
cyclohex
Et
tBu
tBu
Me
Me
Table 4. Efficacy and Toxicity Data of (D)-Glucosamine (1), N-acetyl-(D)-Glucosamine (2) Representative Phosphoramidate of
b
2 from O-6 Series (16) and N-Acetyl-(D)-Glucosamine Phosphoramidate Derivatives 8a−s/9a−s, 9t−y, and 10a−c,m−o in
a
Bovine Cartilage Explant
10 mM
1 mM
0.1 mM
% reduction GAG fold
( SE)
MTT
( SE)
% reduction GAG fold
( SE)
MTT
( SE)
% reduction GAG fold
( SE)
MTT
( SE)
compd
n
n
n
1
2
1
19 (30.42)
9.3
56 (0.5)
61(1.6)
5
3
2
3.4 (4.5)
−1.0 (7.6)
31.0 (6.38)
88 (2.51)
70 (1.68)
66 (1.92)
36
4
2
4
4
1
1
0.8 (5.9)
5.4 (0.4)
106 (4.18)
107 (2.03)
96 (2.91)
79
2
b
16
24.3 (16.28)
−2.7 (16.0)
26.9 (4.0)
31.6 (4.5)
8a/9a
8b/9b
8c/9c
9d
1
1
71.2
27.2
44
60
4
5
61.0 (4.4)
66.2(6.9)
100
87
94
8e/9e
8f/9f
8g/9g
8h/9h
8i/9i
8j/9j
8k/9k
8l/9l
8m/9m
8n/9n
9o
1
1
68.3
25.3
52
46
22
47
4
2
2
5
5
6
−7.0 (6.4)
−8.9 (17.2)
−5.5 (9.8)
−9.6 (10.1)
12.9 (11.3)
15.1 (4.0)
96
100
56
33
81
100
99
1
2
70.4
100
40
54.3 (13.6)
83
96
80
6
2
15.6 (12.5)
0.2 (66.5)
103
98
2
−24.9 (5.0)
−5.4 (8.0)
−14.4 (7.2)
5.1 (18.1)
41.1 (6.4)
17.5 (6.4)
−4.4 (15.7)
−9.6
32
91
92
60
84
90
96
8p/9p
8q/9q
8r/9r
8s/9s
9t
22
8
3
4
6
2
3
26
9u
7
9v
9w
2
3
3
5
2
4
4
2
11.8 (11.5)
−8.0 (9.4)
−4.4 (15.7)
−19.0 (17.2)
9.4 (25.2)
−53.7 (33.4)
−52.5 (28.4)
−23.2 (5.4)
93
91
96
78
99
62
85
100
9x
5
7
9y
10a
10b
10c
10m
10o
a
The average fold increase in GAG release into the culture media in IL-1 treated cultures is calculated using the appropriate control (minus IL-1) for
explants cultured in the absence (control GAG fold) and presence (GAG fold) of glucosamine compounds at concentrations ranging from 10 to 0.1 mM.
The percent reduction in GAG fold was calculated for each experiment as the percent difference observed for each experiment using the following
calculation: {[(control GAG fold) − (sample GAG fold)]/(control GAG fold)} × 100. The figures described above are the mean values of the
calculated results from individual experiments within the sample group. The number of experiments performed using cartilage explants from different
cows are indicated by n, where experiments were performed in triplicate for each number. The effects of different concentrations of glucosamine
compounds on chondrocyte viability were assessed using the MTT assay. The percentage cell viability was calculated compared to the control cells
(absence of glucosamine compounds, taken as 100%). Standard error SE was calculated from the data generated from the number of experiments n;
b
some high standard error readings are due to biological variability of the cartilage obtained. 3j from ref 21.
In each experiment, a control level of fold-increase in GAG
released from the explant as a result of exposure to IL-1 only
was determined. A fold-increase was also calculated for cartilage
explants cultured in the presence of glucosamine derivatives
4633
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
with or without IL-1. This GAG-fold was then taken and com-
pared to the individual control GAG-fold for each set of
experiments. From these two figures, a percent reduction in
GAG-fold release was calculated and the values are reported
herein (Table 4). Thus, in the first set of experiments, where
drug substance was dosed at 10 mM, the mean increase in GAG
released in controls (i.e., IL-1 alone, no glucosamine derivatives
added) was 2.6-fold. Drug efficacy was measured as the
reduction in this fold-increase in GAG release into the culture
media. Thus, treatment with 1 at 10 mM led to a reduction in
GAG release (before/after exposure) to 2-fold, representing a
19% reduction relative to the control (IL-1 only treated explant
culture), while 2 caused only 9% reduction relative to the
control. It is also notable that in the parallel cytotoxicity assays,
conducted using the MTT assay on chondrocyte monolayer
cultures, both 1 and 2 were cytotoxic at 10 mM with only an
approximately 60% cell viability relative to the control. Whether
the apparent inhibition of GAG release by these agents was in
part attributable to cytoxicity is at present unclear. However, we
were able to ascertain that these compounds were considerably
less cytotoxic at 1 and 0.1 mM, but at these lower con-
centrations, both of these compounds were unable to inhibit
IL-1 induced aggrecan release from the tissue (% reduction
GAG fold). Thus, it is not impossible that all of the chondro-
protective effect of (1) and (2) at 10 mM is simply due to
cytotoxicity. There was no nontoxic concentration of (1) or (2)
at which we were able to demonstrate a protective effect in the
assay.
By contrast, phosphate derivatives when tested at 0.1 mM
concentration showed low or no cytoxicity. At the same
concentration, several of the O-4/O-3 phosphoramidate
mixtures synthesized were much more active than the parent
compounds 1 and 2, indeed showing a reduction of GAG
release to the same or greater extent when compared to the
most active O-6 phosphoramidate 16 previously reported by us
(3j from reference 21). At first sight, the increased activity of
prodrugs with the phosphate attached to the “wrong” site, with
respect to the putative pharmacophore, is surprising. As
previously mentioned, N-acetyl-(D)-glucosamine-6-phosphate
is indeed the active intermediate formed during the in vivo
biosynthesis of PGs. However, the identification and cloning
of the human phosphoacetylglucosamine mutase,33 which
has been reported to be able to interconvert N-acetyl-(D)-
glucosamine-6-phosphate into N-acetyl-(D)-glucosamine-1-
phosphate during the in vivo biosynthesis of GAGs, could
explain how activity is not solely found in phosphoramidates
where the phosphate is attached to a specific hydroxyl group. It
cannot be excluded that the phosphohexose mutase may be
able to transfer the phosphate group to the O-1 position from
either O-3 or O-4 positions of the phosphoramidate or a
metabolite thereof. Among the active compounds, the
percentage reduction of GAG release varied over a range of
9−41%. Analysis of the data shows the importance of the
amino acid side chain. The (L)-valine and (L)-proline emerged
as the amino acids of choice, the compounds 8s/9s and 8c/9c
being the most active among the series. The low or no activity
found using the 1-O methylated compounds indicate that a free
1-OH is an essential requisite for maintaining activity.
Table 5. Efficacy and Toxicity Data of (D)-Glucosamine (1)
and N-Acetyl-(D)-Glucosamine O-3 Phosphoramidate
Derivatives 9c, 9l, 9p, 9t, and 9u and N-Acetyl-(D)-
Glucosamine O-4 Phosphoramidate Derivative 81 in Human
a
Cartilage Explant
%
reduction
in GAG
release
%
metabolic
activity
compd
AA
ester
isomer
10 μM
10 μM
9c
9l
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Pro
Bn
O3
O3
O4
O3
O3
O3
28.8
38
115.4
66.4
106.7
87.1
79.4
53.9
Et
8l
Et
17.4
30.4
24.4
27.1
9p
9t
9u
nBu
iPr
2-Bu
% reduction in GAG release
% metabolic activity
compd
250 μM
10 μM
250 μM
10 μM
106.7
1
14.1
10.3
85.3
a
́
The percentage of reduction in GAGs release for each compound was
calculated using the following calculation: {[(mean control GAG's
release) − (mean sample GAG's release)]/(control GAG's release)} × 100.
Compounds 9c, 9l, and 9p showed a statistically significant
activity in the reduction of GAG release at 10 μM, whereas 9t
and 9u were slightly less active at the same concentration. The
O-4 isomer 8l showed substantial decrease in activity compared
to its corresponding O-3 isomer 9l, suggesting that the O-4
isomer series is most probably less active than the O-3 isomers.
Most importantly, 1 was inactive at the same or even higher
doses (see Table 5).
Some of the compounds showed some degree of cytotoxic
effects in WST-1 assay at the studied doses. Derivative 9c, the
proline benzyl ester, showed the best chondroprotective effect/
cytoxicity ratio among the series, being significantly more active
and less cytotoxic than 1.
The CLogP values were also evaluated using the program
Marvin Sketch v. 5.2.0. Log P is an index of molecular hydro-
phobicity, a parameter that affects the compound bioavailability,
besides interaction with biological targets, metabolism as well as
the toxicity. It has become one of the key parameters used to
study the fate and behavior of bioactive compounds. According
to Lipinski's rule of five, Log P should optimally be not greater
than 5. The prediction results showed that Log P values of the
identified active compounds are between −0.99 and 0.99
(Table 6), confirming that they represent promising candidates
for further lead optimization and development.
Table 6. Predicted Log P Values for (D)-Glucosamine (1)
N-Acetyl-(D)-Glucosamine (2) and the Active Compounds
8b/9b, 9c, 9l, 9p, and 8s/9s
compd
AA
ester
Log P
1
−3.04
−3.22
0.37
2
8b/9b
9c
(L)-Val
(L)-Pro
(L)-Pro
(L)-Pro
(L)-Val
Bn
Bn
0.99
9l
Et
−0.99
−0.17
0.56
Additionally, the chondroprotective activity of several proline
amino acid-containing phosphoramidates as either pure O-3
or O-4 isomer was investigated in human cartilage explants.
Table 5 summarizes the preliminary results of the GAG release
inhibition assay for compounds 9c, 9l, 9p, 9t, 9u, and 8l.
9p
nBu
2-Bu
8s/9s
Stability Studies. The stability of several compounds (8c/
9c, 8b/9b, 9l, and 8s/9s, 9y) was studied in biological media
4634
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
using 31P NMR analysis. First, we investigated the stability in
guinea pig and human serum to assess the potential of these
compounds as drugs, and then we studied their stability in
chondrocyte cell lysate in order to verify whether or not the
phosphoramidates may be activated once delivered to the cell.
The disappearance of the phosphoramidate and the appearance
of metabolites were monitored by 31P NMR for several hours at
15 min intervals at 37 °C. Because of excess noise and poor
shimming profiles (most likely due to the biological media and
concentration), individual spectra were processed further. After
normal Fourier transform processing, each spectrum was
deconvoluted (Lorentz−Gauss deconvolution) to reveal solely
the frequency and area of spectral peaks without the baseline.
To assess the enzymatic stability, the rate of disappearance of
each prodrug was determined in guinea pig or human serum
and in chondrocytes cell lysate at 37 °C (pH = 7.0). Half-lives
of the compounds, reported in Table 7, were determined from
either in guinea pig or human serum when compared to those
containing (L)-valine amino acid. No obvious difference was
instead observed by changing the benzyl ester with the 2-butyl
in the valine analogues as 8b/9b and 8s/9s showed similar
stability profiles in both human and guinea pig serum.
Compound 9x was extremely stable in serum, with no sign of
any hydrolysis after 17 h. This stability, likely due to the poor
susceptibility of the t-butyl ester to esterase, may account for its
inactivity, confirming that the initial ester cleavage is important
in the activation of the phosphoramidates, as was previously
found within our research group working on 5′ phenoxy,
alaninyl tert-butyl ester phosphoramidate derivatives of d4T,
and other nucleosides.34 The proline series emerged as
approximately 200 times more stable than the valine one in
guinea pig or human serum, where as in the chondrocytes cell
lysate model this difference is reduced and both of the
compounds were processed. In Figure 2 is reported the 31P
NMR spectra of 8c/9c, 12 h after the addition of chondrocytes
cell lysate at 37 °C. The spectrum clearly shows the formation
of a peak at 0.84 ppm, which may correspond to the
glucosamine monophosphate.
Table 7. Half-Lives for N-Acetyl-(D)-Glucosamine O-3/O-4
Phosphoramidate Derivatives 8b/9b, 8c/9c, 8l, 8s/9s, and 8u
in Guinea Pig and Human Serum and in Chondrocyte Cell
Lysate at 37 °C
CONCLUSION
■
half-lives, t1/2 (h)
In conclusion, we prepared a new series of O-4/O-3 N-acetyl-
(D)-glucosamine phosphoramidates, the structural isomers of
our previously reported O-6 N-acetyl-(D)-glucosamine phos-
phoramidates. Several of these compounds were able to sig-
nificantly decrease the loss of GAG at noncytotoxic concen-
trations. In addition, we reported several examples of 1-O
methylated O-3 phosphoramidates prodrugs. However, in
general, these methylated prodrugs appear to be inactive in
the biological assay, suggesting that a free 1-OH is required for
activity. Finally, we elaborated a protection−deprotection
strategy, which made it possible to achieve exclusively O-3
phosphoramidate regioisomers. Although this strategy is a
longer process, in our case it was found to be the most efficient
synthetic route to afford pure O-3 regiosomer phosphate N-
acetylglucosamine prodrugs in moderate yield. We are now
guinea pig
serum
human
serum
chondrocyte cell
lysate
compd
AA
ester
8b/9b Bn
(L)-Val
(L)-Pro
(L)-Pro
(L)-Val
(L)-Pro
0.55
107
57
0.80
164
42
13
39
8c/9c
8l/9l
8s/9s
9x
Bn
a
Et
nd
2-Bu
tBu
1
1.3
nd
nd
stable
stable
a
nd = not determined.
the apparent first-order rate constant derived from linear
regression of pseudo-first-order plots of prodrug concentration
versus time.
The (L)-proline amino acid-containing phosphoramidates
tested proved to be in general significantly much more stable
Figure 2. Stability of compound 8c/9c after 12 h incubation in chondrocyte cell lysate at 37 °C, monitored by 31P NMR.
4635
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
9o, 8p, 9p, 8t, 9t, 8u, 8v, 9v, 9x, 10a−b, and 10o, and the <95% purity
of the final compounds 8c/9c, 8h−i/9h−i, 9u, and 10c was confirmed
using HPLC analysis. The ≥95% purity of the final compounds 9d,
8r/9r, 8s/9s, and 9w was confirmed by elemental analyses, which were
within 0.4 of calculated values.
Standard Procedures. For practical purposes, standard procedures
are given. Any variations from these procedures are discussed
individually. Procedures that differ from the standard ones are
described in full.
actively working toward minimizing the protecting group
manipulations in order to influence the outcome of the coupling
reaction in a desirable way. Considering the reduction in the
GAG release induced by IL-1 in both bovine and human
cartilages, in addition to good biological stability and the desired
activation in chondrocyte cell lysate, the (L)- proline amino acid-
containing phosphoramidate mixture 8c/9c emerged as a
promising candidate for further development for the treatment
of osteoarthritis and other musculoskeletal diseases. It is a
striking outcome of this research that the prodrugs of the O-3
and O-4 phosphate are more active than those of the O-6 phos-
phate, which is more closely related to the suggested bioactive
species. Thus the O-3 and O-4 regioisomers emerge as promis-
ing novel leads, with activity in human cartilage ex vivo exceeding
that of 1.
4-Methoxyphenyl Phosphorodichloridate (5). Triethylamine
(4.75 mL 33.83 mmol) was added dropwise to a −78 °C stirred
solution of phosphorus oxychloride (5.19 g, 33.83 mmol) and
4-methoxyphenol (4.20 g, 33.83 mmol) in anhydrous diethyl ether
(100 mL) under nitrogen. Following the addition, the reaction mixture
was allowed to slowly warm to room temperature and stirred for 2 h.
The mixture was filtered under nitrogen and the solvent removed
under reduced pressure to give the title compound (8.1 g, 99% yield).
1H NMR (CDCl3, 500 MHz): δ 3.70 (3H, s, −OCH3), 6.80 (2H, d J =
9.15 Hz, −OPh), 7.12 (2H, d J = 9.15 Hz, −OPh). 31P NMR (CDCl3,
202 MHz): δ 4.31.
EXPERIMENTAL SECTION
■
General Experimental Details. Chemistry. General Proce-
dures. Solvents and Reagents. The following anhydrous solvents
were bought from Sigma-Aldrich: dichloromethane (CH2Cl2),
diethyl ether (Et2O), N-methylimidazole (NMI), pyridine
(pyr), tetrahydrofuran (THF), triethylamine (TEA), para-
toluensulfonic acid (pTsOH), amino acid ester salts, N-acetyl-
Standard Procedure 1: Synthesis of Phosphochloridates 6a−v. 5
(1.0 mol) and the appropriate amino ester hydrochloride salt (1.0
mol) were suspended in anhydrous dichloromethane (61.6 mol).
Anhydrous triethylamine (2.0 mol) was added dropwise at −78 °C,
and after 15 min the reaction was left to rise to room temperature and
stirred overnight. The formation of phosphorochloridate was
monitored by 31P NMR. The solvent was removed under reduced
pressure, anhydrous ethyl ether (Et2O) was added to solubilize the
phosphorochloridate, and the triethylamine salt was removed by
filtration, the filtrate was reduced to dryness and the product purified
by flash chromatography using like eluent ethyl acetate/hexane 1/1
(V/V).
(D)-glucosamine, and any other reagents used. All reagents
commercially available were used without further purification.
Thin Layer Chromatography (TLC). Precoated aluminum backed
plates (60 F254, 0.2 mm thickness, Merck) were visualized under
both short and long wave ultraviolet light (254 and 366 nm) or by
burning using the following TLC indicators: (i) molybdate ammonium
cerium sulfate, (ii) potassium permanganate solution. Preparative TLC
plates (20 cm × 20 cm, 500−2000 μm) were purchased from Merck.
Flash Column Chromatography (FCC). Flash column chromatog-
raphy was carried out using silica gel supplied by Fisher (60A, 35−70
μm). Glass columns were slurry packed using the appropriate eluent,
with the sample being loaded as a concentrated solution in the same eluent
or preadsorbed onto silica gel. Fractions containing the product were
identified by TLC and pooled and the solvent was removed in vacuo.
High Performance Liquid Chromatography (HPLC). Analytical
and semipreparative HPLC were conducted on Varian Prostar LC
workstation, Varian Prostar 335 LC detector, Varian fraction collector
(model 701), and Prostar 210 solvent delivery system, with Varian
Polaris C18-A (10 μm) as an analytical column and Varian Polaris
C18-A (10 μm) as a semipreparative column. The software used was
Galaxie Chromatography Data System.
4-Methoxyphenyl(cyclohexoxy-(l)-prolinyl) phosphorochloridate
(6v). Prepared according to standard procedure 1 in 30% yield as a
1
mixture of two diastereoisomers (SSp and SRp). H NMR (CD3OD,
500 MHz): δ 7.08 (2H, m, −OPh), 6.76 (2H, m, −OPh), 6.87 (2H,m,
−OPh), 4.73 (1H, m, −OCHCH2), 4.35 (0.5H, m, −CHCO2, one
diastereoisomer), 4.26 (0.5H, m, −CHCO2, one diastereoisomer),
3.69 (1.5H, s, −OCH3, one diastereoisomer), 3.68 (1.5H, s, −OCH3,
one diastereoisomer), 3.66 (1.5H, s, −OCH3, one diastereoisomer),
3.40 (2H, m, −CH2N), 2.13 (1H, m, −CH2aCHCO2), 1.94 (3H, m,
−CH2bCHCO2 and CH2CH2N), 1.73 (2H, m, −CH2 (cyclohex)),
1.71 (2H, m, −CH2 (cyclohex)), 1.31 (6H, m, −CH2 (cyclohex)). 31
NMR (CDCl3, 202 MHz): δ 8.46, 8.38 (ratio 1:1).
P
Standard Procedure 2: Synthesis of Phosphoramidates 8a−v/
9a−v, 10a−o, and 14c,t using NMI. A solution of appropriate
phosphorochloridate (15.63 mmol) in anhydrous THF (15 mL) was
added to a −30 °C stirred solution of N-acetyl-(D)-glucosamine or 1-
O-methyl-N-acetyl-(D)-glucosamine or N-acetyl-4,6-O-benzylidene-
(D)-glucosamine (13.56 mmol) and NMI (6.2 mL, 78.15 mmol) in
pyridine (100 mL). After 15 min, the reaction was allowed to slowly
warm to room temperature and stirred for 3 h. Then methanol was
added, and the solvent was removed under reduced pressure, the crude
residue purified by flash chromatography using like eluent a gradient of
CH2Cl2/MeOH from 98/2 to 95/5 (V/V).
1
Nuclear Magnetic Resonance (NMR). H NMR (500 MHz), 13C
NMR (125 MHz), and 31P NMR (202 MHz) were recorded on a
Bruker Avance 500 MHz spectrometer at 25 °C. Chemical shifts (δ)
are quoted in parts per million (ppm) relative to internal CD3OD (δ
1
1
3.34 H NMR, δ 49.86 13C NMR) and CDCl3 (δ 7.26 H NMR, δ
77.36 13C NMR) or external 85% H3PO4 (δ 0.00 31P NMR). Coupling
constants (J) are measured in hertz. The following abbreviations are
used in the assignment of NMR signals: s (singlet), d (doublet),
t (triplet), q (quartet), m (multiplet), bs (broad singlet), dd (doublet of
doublet), dt (doublet of triplet). The characterization of phosphor-
amidates involves the identification of α and β sugar derivatives. The
ratio of diastereisomers was determined based on 31P NMR. The
2-Acetamido-2-deoxy-4-O-[4-methoxyphenyl-(cyclohexyloxy-(L)-
prolinyl)-phosphate-(D)-glucopyranose (8v) and 2-Acetamido-2-
deoxy-3-O-[4-methoxyphenyl-(cyclohexyloxy-(L)-prolinyl)-phos-
phate-(D)-glucopyranose (9v). The compounds were obtained as
described in standard procedure 2. However the crude product was
purified twice by flash chromatography on silica gel under gradient
elution system of CH2Cl2/MeOH from 100/0 to 97/3, affording 8v as
a mixture of Rp and Sp diastereoisomers as α and β anomers in 0.5%
assignment of the signals in H NMR and 13C NMR was done based
1
on the analysis of coupling constants and additional two-dimensional
experiments (COSY, HSQC, HMBC, PENDANT).
Mass Spectrometry (MS). Low resolution mass spectra was
performed on Bruker Daltonics microTof-LC, (atmospheric pressure
ionization, electron spray mass spectroscopy) in either positive or
negative mode. High resolution mass spectroscopy was performed as a
service by Birmingham University, using fast atom bombardment (FAB).
Elemental Analysis (CHN). CHN microanalysis was performed as a
service by the School of Pharmacy at the University of London.
Purity of Final Compounds. The ≥95% purity of the final
compounds 8a−b/9a−b, 8c, 9c, 8e−g/9e−g, 8j/9j, 81, 91, 8m/9m,
1
yield. H NMR (CD3OD, 500 MHz) δ 7.18 (1.1H, m, −OPh), 7.09
(0.9H, m, −OPh), 6.98 (1.1H, m, −OPh), 6.81 (0.9H, m, −OPh),
5.17 (1H, d J = 3.7 Hz, −CH-1), 4.78 (1H, m, −OCHCH2), 4.67 (1H,
m, −CH-4), 4.55 (1H, m, −CH2a-6), 4.49 (1H, m, −CH2b-6), 4.38
(1H, m, −NCHCO2), 4.04 (2H, m, −CH-2, −CH-5), 3.81 (3H, s,
PhOCH3), 3.62 (1H, m, −CH-3), 3.46 (1H, m, −CH2aNH), 3.38
(1H, m, −CH2bNH), 2.21 (1H, m, CH2a-CH-CO2), 2.06−1.98 (3H,
4636
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
m, −CH2b-CH-CO2, −CH2CH2NH), 1.79 (4H, m, −CH2CHO and
−CH2CH2CH2), 1.78 (3H, s, −NHCOCH3), 1.38 (6H, m,
−OCHCH2, CH2CH2CH2 and CH2CH2CH2). 13C NMR (CD3OD,
125 MHz) δ 24.68 (−NHCOCH3), 26.13 (d J(C−P) = 8.2 Hz, −CH2-
CH2-NH), 26.45 (−CH2-CH2), 32.44 (m, −CH2-CH-CO2CH2Ph,
−CH2CHO, and −OCHCH2), 48.50 (d J(C−P) = 3.7 Hz, −CH2−NH,
overlap with the solvent), 54.93 (d J(C−P) = 3.0 Hz, C-2), 56.11, 56.16
(−OCH3), 62.01 (d J(C−P) = 6.5 Hz, −CHNH), 66.84 (CH2-6), 70.18
(C-3), 71.80 (C-5), 74.41 (−OCHCH2), 79.88 (d J(C−P) = 6.4 Hz,
(CD3OD 125 MHz): δ 22.95 (−NHCOCH3), 23.00 (−NHCOCH3),
24.62 (−CH2 (cyclohex), 24.67 (−CH2 (cyclohex), 26.10 (d J(C−P)
=
8.2 Hz, −CH2CH2N), 26.48 (−CH2 (cyclohex)), 32.46 (m,
−CH2CHCO2, −CH2 (cyclohex)), 48.80 (CH2N, under the residual
solvent peak), 54.43 (d J(C−P) = 3.8 Hz, C-2), 54.35 (d J(C−P) = 3.8 Hz,
C-2), 56.12 (−OCH3), 56.17 (−OCH3), 61.62 (d J(C−P) = 7.3 Hz,
(−CHCO2), 64.42 (C-5), 69.75 (−CH2-6), 70.83 (−OCH2Ph), 74.63
(−OCHCH2), 76.64 (d J(C−P) = 7.6 Hz C-3), 76.77 (d J(C−P) = 7.6 Hz,
C-3), 81.61 (d J(C−P) = 2.0 Hz, C-4), 98.59 (C-1), 102.18 (C-1),
102.93 (−CHPh), 115.49 (−PhOCH3), 115.75 (−PhOCH3), 122.10
(d J(C−P) = 5.3 Hz, −PhOCH3), 122.29 (d J(C−P) = 5.3 Hz, −
PhOCH3), 127.77 (−CHPh), 129.20 (−CHPh), 129.22 (−CH2Ph),
129.57 (−CH2Ph), 129.62 (−CH2Ph), 130.11(−CHPh), 138.51
(“ipso” −CHPh) 138.95 (“ipso” −CH2Ph), 145.7 (d J(C−P) = 7.6 Hz
“ipso” −POPh), 158.05 (“ipso” −PhOCH3), 173.38, (−NHCOCH3),
173.47 (−NHCOCH3), 174.30 (−CO2(cyclohexyl)). 31P NMR
(CD3OD 202 MHz): 0.96, 0.58 (ratio 10:1).
Standard Procedure 4: Synthesis of Phosphoramidates 9t−y by
Hydrogenation of 14t and 15u−y. A solution of 2-acetamido-2-
deoxy-4,6-O-benzylidene-3-O-substituted-phosphate-(D)-glucopyra-
nose 14t or 1-O-benzyl-2-acetamido-2-deoxy-4,6-O-benzylidene-3-O-
substituted-phosphate-(D)-glucopyranose 15u−y (0.28 mmol) was
hydrogenated in MeOH or ethanol (5 mL) under pressure of 54 psi
for 6 h in presence of a catalyst (10% Pd/C or 10% Pd(OH)2/C). The
catalyst was removed by filtration through a Celite pad. and the solvent
was removed under reduced pressure to obtain pure compounds
unless specified otherwise.
C-4), 92.70 (C-1), 97.65 (C-1), 115.78 (−PhOCH3), 122.29 (d J(C−P)
=
4.4 Hz, −PhOCH3), 145.51 (d J(C−P) = 8.1 Hz, “ipso” −POPh), 158.43
(“ipso” −PhOCH3), 173.25 (−COCH3), 175.25, 176.31 (−CO2CH2).
31P NMR (CD3OD, 202 MHz): 3.16, 2.53, 2.24, ratio (6.1:5.9:1). MS
(E/I) 587.11 (MNa+). HPLC: (gradient H2O/CH3CN from 70/30 to
H2O/CH3CN: 0/100 in 20 min, flow = 1 mL/min, λ = 275 nm): tR
12.51 min, 12.84 min.
Further elution with CH2Cl2/MeOH 97/3 afforded 9v as a mixture
1
of Rp and Sp diastereoisomers as α and β anomers in 1.3% yield. H
NMR (CD3OD, 500 MHz): δ 7.15 (1H, m, −OPh), 7.01 (1H, d, J =
9.2 Hz, −OPh), 6.78 (1H, d, J = 9.2 Hz, −OPh), 6.70 (1H, d, J = 9.2
Hz, −OPh), 4.98 (0.8H, d, J = 3.9 Hz, −CH-1), 4.62 (1.2H, m, CH-1
and −OCHCH2), 4.39 (0.8H, m, −CH-3), 4.22 (0.2H, m, −CH-3),
4.11 (1H, m, −NCHCO2) 3.99 (1H, m, −CH-2), 3.75 (1H, m,
−CH-5), 3.67 (5H, m, −CH2-6 and PhOCH3), 3.51 (0.8H, t J = 9.6 Hz,
−CH-4), 3.44 (0.2H, d J = 9.6 Hz, −CH-4), 3.30 (1H, m, −CH2aNH),
3.20 (1H, m, −CH2bNH, overlap with the solvent), 2.10 (1H, m,
CH2a-CH-CO2) 1.91, 1.90 (3H, s, −NHCOCH3), 1.81 (3H,
CH2bCHCO2, −CH2CH2NH), 1.63 (4H, m, −CH2CHO and
−CH2CH2CH2), 1.28 (6H, m, −OCHCH2, CH2CH2CH2, and
CH2CH2CH2). 13C NMR (CD3OD, 125 MHz): δ 23.01, 23.33
(−NHCOCH3), 24.62 (−CH2CH2CH2), 26.10, 26.25 (d J(C−P) = 8.2
Hz, −CH2-CH2-NH), 26.46 (−CH2CH2CH2), 32.46 (m, −CH2-CH-
CO2CH2Ph, −CH2CHO, and −OCHCH2), 48.80 (d J(C−P) = 3.7 Hz,
−CH2-NH, overlap with the solvent), 54.59 (d J(C−P) = 3.8 Hz, C-2),
56.05, 56.09 (−OCH3), 61.96 (d J(C−P) = 7.3 Hz, −CH2CHNH),
66.46, 62.61 (−CH2-6), 70.80, 70.90 (d J(C−P) = 3.1 Hz, C-4), 74.44,
74.74 (−OCHCH2), 74.82 (C-5), 81.05 (d J(C−P) = 7.0 Hz, C-3),
82.87 (d J(C−P) = 8.7 Hz, C-3), 92.83 (C-1), 96.29 (C-1), 115.19,
115.58 (−PhOCH3), 122.32, 122.56 (d J(C−P) = 4.7 Hz, −PhOCH3),
145.94 (d J(C−P) = 7.1 Hz, “ipso” −POPh), 158.28 (“ipso” −PhOCH3),
173.46, 173.78 (−NHCOCH3), 174.50, 175.06 (d J(C−P) = 1.5 Hz,
−CO2CH2). 31P NMR (CD3OD, 202 MHz): 2.42, 1.80 ratio (4.4:1).
MS (E/I) 587.11 (MNa+). HPLC: (gradient H2O/CH3CN from
70/30 to H2O/CH3CN: 0/100 in 20 min, flow = 1 mL/min, λ = 275 nm):
tR 11.97 min, 12.24 min.
2-Acetamido-2-deoxy-3-O-[4-methoxyphenyl-(cyclohexyloxy-(L)-
prolinyl)]-phosphate-(D)-glucopyranose (9v). The compound was
prepared as described in standard procedure 4 (EtOH, 10% Pd(OH)2/
C) and obtained in 12% yield as a mixture of Rp and Sp
diastereoisomers as α and β anomers, after purification by preparative
TLC using CH2Cl2/MeOH 92/8 (V/V). 1H NMR (CD3OD 500
MHz): δ 7.15 (1.5H, m, −OPh), 7.01 (0.5H, d J = 9.2 Hz, −OPh),
6.78 (1.5H, d J = 9.2 Hz, −OPh), 6.70 (0.5H, d J = 9.2 Hz, −OPh)
4.98 (0.6H, d J = 3.9 Hz −CH-1α), 4.62 (1.3H, m, CH-1β and
−OCHCH2), 4.39 (0.6H, m, −CH-3), 4.22 (0.6H, m, −CH-3), 4.11
(0.76H, m, −CHCO2) 3.99 (0.9H, m, −CH-2), 3.75 (0.8H, m, −CH-5),
3.67 (4.8H, m, −CH2-6 and PhOCH3), 3.51 (1H, t J = 9.62 Hz,
−CH-4), 3.44 (0.5H, t J = 9.62, −CH-4), 3.30 (1H, m, −CH2aN), 3.20
(1H, m, −CH2bN, overlap with the solvent), 2.10 (1H, m,
CH2aCHCO2), 1.91 (2H, s, −NHCOCH3), 1.90 (1H, s,
−NHCOCH3), 1.81 (3H, CH2bCHCO2, −CH2CH2N), 1.63 (4H,
m, −CH2 (cyclohex)), 1.28 (6H, m, −CH2 (cyclohex)). 13C NMR
(CD3OD 125 MHz): δ 23.01 (−NHCOCH3), 23.33 (−NHCOCH3),
24.59 (−CH2 (cyclohex)), 24.62 (−CH2 (cyclohex)), 26.10 (d J(C−P)
=
9v was also obtained from 15v according to standard procedure 4
(see later section).
8.2 Hz, −CH2CH2N), 26.25 (d J(C−P) = 8.2 Hz, −CH2CH2N), 26.43
(−CH2 (cyclohex)), 26.46 (−CH2 (cyclohex)), 32.46 (m,
−CH2CHCO2, −CH2 (cyclohex), 48.80 (d J = 3.7 Hz, −CH2N,
overlap with the solvent), 54.59 (d J(C−P) = 3.8, C-2), 56.05 (−OCH3),
56.09 (−OCH3), 61.96 (d J(C−P) = 7.3 Hz, (−CHCO2), 66.46 (−CH2-6),
Standard Procedure 3: Synthesis of Protected Phosphoramidates
15u−y Using t-BuMgCl. To a solution of 1-O-benzyl-N-acetyl-4,6-O-
benzylidene-(D)-glucosamine (1.98 mmol) in THF (20 mL),
t-BuMgCl (3.97 mL, 3.96 mmol) was added at rt. Then a solution
of appropriate phosphorochloridate (5.98 mmol) in THF (5 mL) was
added dropwise. After 1 h, the solvent was removed under vacuum and
the crude purified by flash chromatography using like eluent a gradient
of EtPt/EtOAc 5/5 to EtOAc (V/V).
1-O-Benzyl-2-acetamido-2-deoxy-4,6-O-benzylidene-3-O-[4-me-
thoxyphenyl-(cyclohexyloxy-(L)-prolinyl)-phosphate-(D)-glucopyra-
nose (15v). The compound was obtained as described in standard
procedure 3 in 41% yield as a mixture of Rp and Sp diastereoisomers
as α and β anomers. 1H NMR (CD3OD 500 MHz): δ 7.38 (2H, d J =
7.4 Hz, −CHPh), 7.19 (8H, m, −CHPh and −CH2Ph), 6.95 (0.2H, d
J = 8.7 Hz, −OPh), 6.81 (1.8H, d J = 8.7 Hz, −OPh), 6.75 (0.2H, d J =
9.2 Hz, −OPh), 6.52 (1.8H, d J = 8.75 Hz, −OPh), 5.47 (1H, s,
−CHPh), 4.76 (1H, d J = 3.6 Hz −CH-1, α), 4.61 (3H, m, −CH-3,
−OCHCH2, and CH2aPh), 4.43 (1H, d J = 10.3 Hz, CH2bPh), 4.21
(1H, m, −CH-2), 4.07 (1H, dd J = 10.18 and 4.7 Hz, −CH2a6), 4.02
(1H, m, −CHCO2), 3.77 (1H, m, −CH-5), 3.66 (2H, m, −CH-4 and
−CH2b6), 3.56 (2.5H, s, −OCH3), 3.55 (0.5H, s, −OCH3), 3.14 (2H,
m, −CH2N), 1.87 (3H, s, −NHCOCH3), 1.63 (8H, −CH2CH2N,
−CH2 (cyclohex)), 1.30 (6H, m, −CH2, (cyclohex)). 13C NMR
62.610 (CH2-6), 70.80 (d J(C−P) = 3.1 Hz, C-4), 70.90 ((d J(C−P)
=
3.1 Hz, C-4), 72.95 (C-5), 74.44 (−OCHCH2), 74.74 (−OCHCH2),
77.52 (C-5), 81.05 (d J(C−P) = 7.0 Hz, C-3), 82.87 (d J(C−P) = 8.7 Hz,
C-3), 92.83 (C-1α), 96.29 (C-1β), 115.19 (−PhOCH3), 115.58
(−PhOCH3), 122.32 (d J(C−P) = 4.7 Hz, −PhOCH3), 122.56 (d J(C−P)
=
4.3 Hz, −PhOCH3), 145.94 (d J(C−P) = 7.1 Hz “ipso” −POPh), 158.28
(“ipso” −PhOCH3), 173.46, (−NHCOCH3), 173.78, (−NHCOCH3),
174.50 (d J(C−P) = 1.5 Hz, −CO2cyclohex), 175.06 (d J(C−P) = 1.5 Hz,
−CO2cyclohexyl). 31P NMR (CD3OD 202 MHz): 2.42, 1.80 (ratio
10:4). MS (ES+) 609.2 (MNa+). HPLC: (gradient H2O/CH3CN
from 100/0 to 0/100 in 15 min, flow = 1 mL/min, λ = 275 nm): tR
13.75 min, 14.19 min.
31P NMR Stability Assay in Human and Guinea Pig Serum. The
phosphoramidate (10 mg) is dissolved in DMSO (50 μL) and D2O
(150 μL), into an NMR tube. The NMR tube is warmed in the NMR
chamber at 37 °C. Then either the human serum or the guinea pig
serum (300 μL) was added to the NMR tube quickly. NMR
experiments were programmed into the computer to record data every
15 min usually for 12 h (overnight) or longer if needed. The data was
4637
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
then processed via Fourier transform and then deconvoluted to further
clarify the product signals and integrate them. The pH of the samples
was measured before and after these experiments, which remained
at 7.0.
ACKNOWLEDGMENTS
This work was supported by Bioiberica SA.
■
31P NMR Stability Assay in Chondrocyte Cell Lysate. 31P NMR
experiments were conducted using chondrocyte cell lysate obtained
from Prof. Catersons group (1 mL of the cell lysate solution contained
the contents of 10 million chondrocyte cells) to which was added
deuterated water (D2O) and the compound of interest. An NMR tube
containing 250 μL of lysate (equivalent to 2.5 million cells) in 0.5 mL
of D2O was warmed in the NMR chamber until a stable temperature
of 37 °C was met and then a blank 31P NMR was taken. Compound
was added to the lysate solution to give a 15 mM concentration, the
contents were shaken, and the tube returned to the NMR from which
the first recording was taken (ca. 3 min). The NMR experiments were
then set up to record data every half hour for up to 12 h overnight,
with a final experiment taken in the morning prior to removing the
sample (13 h 42 min).
ABBREVIATIONS USED
■
GPI, glycosylphosphatidylinositol; GAGs, glycosaminoglycan;
PGs, proteoglycans; OA, osteoarthritis; UDP, uridine diphos-
phate; NMI, N-methyl imidazole; IL-1, interleukin 1
REFERENCES
■
(1) Mobasheri, A.; Vannucci, S. J.; Bondy, C. A.; Carter, S. D.; Innes,
J. F.; Arteaga, M. F.; Trujillo, E.; Ferraz, I.; Shakibaei, M.; Martín-
Vasallo, P. Glucose transport and metabolism in chondrocytes: a key
to understanding chondrogenesis, skeletal development and cartilage
degradation in osteoarthritis. Histol. Histopathol. 2002, 17, 1239−1267.
(2) Kjellen, L.; Lindahl, U. Proteoglycans: Structures and
Interactions. Annu. Rev. Biochem. 1991, 60, 443−475.
(3) Ruoslahti, E. Structure and Biology of Proteoglycans. Annu. Rev.
Cell Biol. 1988, 4, 229−255.
Glycosaminoglycan (GAG) Release Assay. Full-depth articular
cartilage were dissected from the metacarpophalangeal joints of calves
of approximately 6 months of age or human cartilage from joint
replacement surgery (approximately 4 mm diameter punches for
bovine cartilage and 2 mm for human cartilage). After an equilibration
period of 24 h, explants were cultured for one day in 200 μL of
DMEM + hydrolyzed lactalbumin + 50 μg/mL vitamin C + penicillin/
streptomycin + ITS (insulin−transferrin−selenium) in the presence or
absence of the compounds. After one day, GAG release was induced
by addition of 10 ng of IL-1. Six days after IL-1 induction, the cartilage
and culture medium were collected for analysis. During this six-day
period, the culture medium was replaced every two days.
GAG release was analyzed using the Blyscan colorimetric assay
(Biocolor, Belfast, UK). To measure GAG content of the cartilage
explants, the tissue was digested with papain. The culture medium of
day 4, 6, and 8 was pooled before GAG analysis. GAG release was
expressed as percentage of the total GAG content of the explants (i.e.,
GAG medium + GAG in papain digest).
(4) Seed, S.; Dunica, K.; Lynch, A. Osteoarthritis: a review of
treatment options. Geriatrics 2009, 64, 20−29.
(5) Sofat, N.; Beith, I.; Anikumar, P. G.; Mitchel, P. Recent Clinical
Evidence for the Treatment of Osteoarthritis: What We have Learned.
Rev. Recent Clin. Trials 2011, 6, 114−126.
(6) Huskisson, E. C. Glucosamine and Chondroitin for Osteo-
arthritis. J. Int. Med. Res. 2008, 36, 1161−1179.
(7) Miller, K.; Clegg, D. Glucosamine and chondroitin sulfate.
Rheum. Dis. Clin. North Am. 2011, 37, 103−118.
(8) Vlad, S. C.; LaValley, M. P.; McAlindon, T. E.; Felson, D. T.
Glucosamine for pain in osteoarthritis: Why do trial results differ?
Arthritis Rheum. 2007, 56, 2267−2277.
(9) Wandel, S.; Juni, P.; Tendal, B.; Nuesch, E.; Villiger, P. M.;
Welton, N. J.; Reichenbach, S.; Trelle, S. Effects of glucosamine,
chondroitin, or placebo in patients with osteoarthritis of hip or knee:
network meta-analysis. Br. Med. J. 2010, 341, c4675.
(10) Sawitzke, A. D.; Shi, H.; Finco, M. F.; Dunlop, D. D.; Harris, C. L.;
Singer, N. G.; Bradley, J. D.; Silver, D.; Jackson, C. G.; Lane, N. E.; Oddis,
C. V.; Wolfe, F.; Lisse, J.; Furst, D. E.; Bingham, C. O.; Reda, D. J.;
Moskowitz, R. W.; Williams, H. J.; Clegg, D. O. Clinical efficacy and safety
of glucosamine, chondroitin sulphate, their combination, celecoxib or
placebo taken to treat osteoarthritis of the knee: 2-year results from
GAIT. Ann. Rheum. Dis. 2010, 69, 1459−1464.
(11) Wildi, L. M.; Raynauld, J.-P.; Martel-Pelletier, J.; Beaulieu, A.;
Bessette, L.; Morin, F. D. R.; Abram, F. O.; Dorais, M.; Pelletier, J.-P.
Chondroitin sulphate reduces both cartilage volume loss and bone
marrow lesions in knee osteoarthritis patients starting as early as
6 months after initiation of therapy: a randomised, double-blind,
placebo-controlled pilot study using MRI. Ann. Rheum. Dis. 2011, 70,
982−989.
A control condition with 0.25% DMSO was included because the
test compounds were dissolved in DMSO. A condition with
glucosamine hydrochloride was added for comparison. DMSO + IL1
control value for human explant experiment was 18.4%, whereas the
control GAG-fold increase values for bovine explant experiments at 10,
1, and 0.1 mM were respectively 2.6, 4.9, and 5.1 fold.
Each culture condition was performed with cartilage of five different
paws/patients to account for biological variation. ANOVA followed by
LSD posthoc tests were used for determining differences between
groups (p ≤ 0.05 was considered significant).
Compounds that showed a nonstatistically significant reduction of
the GAG release compared to IL-1-stimulated in the presence of
0.25% DMSO condition were considered inactive.
(12) Bassleer, C.; Rovati, L.; Franchimont, P. Stimulation of
proteoglycan production by glucosamine sulfate in chondrocytes
isolated from human osteoarthritic articular cartilage in vitro.
Osteoarthritis Cartilage 1998, 6, 427−434.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and spectroscopic data of all compounds: H, 13C
1
NMR, 31P NMR, HPLC, MS. Purity data of phosphate
prodrugs of N-acetyl-(D)-glucosamine. HPLC traces of the
prodrugs 8a/9a, 8b/9b, 8c/9c, 8c, 9c, 8e/9e, 8f/9f, 8g/9g,
8h/9h, 8i/9i, 8j/9j, 8l, 9l, 8m/9m, 8o, 8p, 9p, 8t, 9t, 8u, 9u,
8v, 9v, 9x, 10a, 10b, 10c, and 10o. This material is available
free of charge via the Internet at http://pubs.acs.org.
(13) Uitterlinden, E.; Koevoet, J.; Verkoelen, C.; Bierma-Zeinstra, S.;
Jahr, H.; Weinans, H.; Verhaar, J.; van Osch, G. Glucosamine increases
hyaluronic acid production in human osteoarthritic synovium explants.
BMC Musculoskelet. Disord. 2008, 9, 120.
3
(14) Mroz, P.; Silbert, J. Use of H-glucosamine and 35S-sulfate with
cultured human chondrocytes to determine the effect of glucosamine
concentration on formation of chondroitin sulfate. Arthritis Rheum.
2004, 50, 3574−3579.
(15) Shikhman, A., R.; Kuhn, K.; Alaaeddine, N.; Lotz, M. N-
Acetylglucosamine Prevents IL-1β-Mediated Activation of Human
Chondrocytes. J. Immunol. 2001, 166, 5155−5160.
AUTHOR INFORMATION
■
Corresponding Author
(16) Ozkan, F. U.; Ozkan, K.; Ramadan, S.; Guven, Z.
Chondroprotective effect of N-acetylglucosamine and hyaluronate in
early stages of osteoarthritis: an experimental study in rabbits. Bull.
NYU Hosp. Jt. Dis. 2009, 67, 352−357.
*Phone/Fax: +44 02920874537. E-mail: mcguigan@cardiff.
ac.uk.
Notes
The authors declare no competing financial interest.
(17) LogD values were calculated using Marvin Sketch v. 5.2.0.
4638
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639

Journal of Medicinal Chemistry
Article
(18) Aghazadeh-Habashi, A.; Sattari, S.; Pasutto, F.; Jamali, F. Single
dose pharmacokinetics and bioavailability of glucosamine in the rat.
J. Pharm. Pharm. Sci. 2002, 5, 181−184.
(19) Du, J.; White, N.; Eddington, N. D. The bioavailability and
pharmacokinetics of glucosamine hydrochloride and chondroitin
sulfate after oral and intravenous single dose administration in the
horse. Biopharm. Drug Dispos. 2004, 25, 109−116.
(20) Adebowale, A.; Du, J.; Liang, Z.; Leslie, J. L.; Eddington, N. D.
The bioavailability and pharmacokinetics of glucosamine hydrochloride
and low molecular weight chondroitin sulfate after single and multiple
doses to beagle dogs. Biopharm. Drug Dispos. 2002, 23, 217−225.
(21) McGuigan, C.; Serpi, M.; Bibbo, R.; Roberts, H.; Hughes, C.;
Caterson, B.; Gibert, A. T.; Verson, C. R. A. Phosphate Prodrugs
Derived from N-Acetylglucosamine Have Enhanced Chondroprotective
Activity in Explant Cultures and Represent a New Lead in
Antiosteoarthritis Drug Discovery. J. Med. Chem. 2008, 51, 5807−5812.
(22) Paulsen, H. Advances in Selective Chemical Syntheses of
Complex Oligosaccharides. Angew. Chem., Int. Ed. 1982, 21, 155−173.
(23) Crich, D.; Dudkin, V. Why Are the Hydroxy Groups of Partially
Protected N-Acetylglucosamine Derivatives Such Poor Glycosyl
Acceptors, and What Can Be Done about It? A Comparative Study
of the Reactivity of N-Acetyl-, N-Phthalimido-, and 2-Azido-2-deoxy-
glucosamine Derivatives in Glycosylation. 2-Picolinyl Ethers as
Reactivity-Enhancing Replacements for Benzyl Ethers. J. Am. Chem.
Soc. 2001, 123, 6819−6825.
(24) Liao, L.; Auzanneau, F.-I. Glycosylation of N-Acetylglucos-
amine: Imidate Formation and Unexpected Conformation. Org. Lett.
2003, 5, 2607−2610.
(25) Colombo, M. I.; Stortz, C. A.; Ruveda, E. A. A comparative
́
study of the O-3 reactivity of isomeric N-dimethylmaleoyl-protected
D-glucosamine and D-allosamine acceptors. Carbohydr. Res. 2011, 346,
569−576.
(26) Bohn, M. L.; Colombo, M. I.; Pisano, P. L.; Stortz, C. A.;
Ruveda, E. A. Differential O-3/O-4 regioselectivity in the glycosylation
́
of [alpha] and [beta] anomers of 6-O-substituted N-dimethylmaleoyl-
protected d-glucosamine acceptors. Carbohydr. Res. 2007, 342, 2522−
2536.
(27) Bohn, M. L.; Colombo, M. I.; Ruveda, E. A.; Stortz, C. A.
Conformational and electronic effects on the regioselectivity of the
glycosylation of different anomers of N-dimethylmaleoyl-protected
glucosamine acceptors. Org. Biomol. Chem. 2008, 6, 554−561.
(28) Cao, S.; Gan, Z.; Roy, R. Active-latent glycosylation strategy
toward Lewis X pentasaccharide in a form suitable for neo-
glycoconjugate syntheses. Carbohydr. Res. 1999, 318, 75−81.
(29) Gan, Z.; Cao, S.; Wu.; Roy, R. Regiospecific Syntheses of N-
Acetyllactosamine Derivatives and Application Toward a Highly
Practical Synthesis of Lewis X Trisaccharide. J. Carbohydr. Chem.
1999, 18, 755−773.
́
(30) Figueroa-Perez, S.; Verez-Bencomo, V. Synthesis of dimeric
Lewis X antigenic determinant with azido-type spacer arm by a
sequence of regioselective glycosylation steps. Tetrahedron Lett. 1998,
39, 9143−9146.
(31) Spijker, N. M.; Keuning, C. A.; Hooglugt, M.; Veeneman, G. H.;
van Boeckel, C. A. A. Synthesis of a hexasaccharide corresponding to a
porcine zona pellucida fragment that inhibits porcine sperm−oocyte
interaction in vitro. Tetrahedron 1996, 52, 5945−5960.
(32) Ssarek, W. A.; Zamojski, A.; Tiwari, K. N.; Ison, E. R. A new
facile method for cleavage of actals and dithioacetals in carbohydrate
derivatives. Tetrahedron Lett. 1986, 27, 3827−3830.
(33) Mio, T.; Yamada-Okabe, T.; Arisawa, M.; Yamada-Okabe, H.
Functional cloning and mutational analysis of the human cDNA for
phosphoacetylglucosamine mutase: identification of the amino acid
residues essential for the catalysis. Biochim. Biophys. Acta 2000, 1492,
369−376.
(34) McGuigan, C.; Sutton, P. W.; Cahard, D.; Turner, K.; O’Leary,
G.; Wang, Y.; Gumbleton, M.; De Clercq, E.; Balzarini, J. Synthesis,
anti-human immunodefeficency virus activity and esterase lability of
some novel carboxylic ester-modified phosphoramidate derivatives of
stavudine (d4T). Antivir. Chem. Chemother. 1998, 9, 473−479.
4639
dx.doi.org/10.1021/jm300074y | J. Med. Chem. 2012, 55, 4629−4639