Phosphate prodrugs derived from N-acetylglucosamine have enhanced chondroprotective activity in explant cultures and represent a new lead in antiosteoarthritis drug discovery

Article abstract of DOI:10.1021/jm800594c

We report the application of the phosphoramidate ProTide approach, developed by us for antiviral nucleosides, to sugar derivatives with potential chondroprotection against osteoarthritis. In particular, N-acetylglucosamine was converted to a series of 06

Full text of DOI:10.1021/jm800594c

J. Med. Chem. 2008, 51, 5807–5812
5807
Phosphate Prodrugs Derived from N-Acetylglucosamine Have Enhanced Chondroprotective
Activity in Explant Cultures and Represent a New Lead in Antiosteoarthritis Drug Discovery
Christopher McGuigan,*,† Michaela Serpi,† Rita Bibbo,† Helen Roberts,‡ Clare Hughes,‡ Bruce Caterson,‡ Ana Torrent Gibert,§
and Carlos Rau´l Alaez Verson§
Welsh School of Pharmacy, Cardiff UniVersity, King Edward VII AVenue, Cardiff CF10 3XF, U.K., Cardiff School of Biosciences,
Cardiff UniVersity, Biomedical Building, Museum AVenue, Cardiff CF10 3US, U.K., and Bioiberica S.A., Plaza Francesc Macia`,
7, Barcelona 08029, Spain
ReceiVed May 19, 2008
We report the application of the phosphoramidate ProTide approach, developed by us for antiviral nucleosides,
to sugar derivatives with potential chondroprotection against osteoarthritis. In particular, N-acetylglucosamine
was converted to a series of 06 arylaminoacyl phosphoramidates with ester and amino acid variation.
Compounds were prepared by two routes, with or without sugar protection, and were isolated as phosphate
diastereoisomers. The compounds were assayed for cellular toxicity and for inhibition of IL-1 induced
glycosaminoglycan (GAG) release (i.e., proteoglycan degradation) from bovine articular cartilage in vitro
explant cultures. By comparison to the N-acetyl glucosamine parent, some of the analogues show a significant
enhancement in efficacy in the inhibition of inflammatory cytokine-induced proteoglycan degradation.
Introduction
Osteoarthritis (OAa) is a common muscoskeletal disease
leading to pain, loss of mobility, and significant reduction in
the quality of life of those affected. At present no chemothera-
peutic intervention is proven to be fully effective at ameliorating
Figure 1. Glucosamine (1) and N-acetyl glucosamine (2).
the progression of the disease. A number of glucosamine-
containing nutraceutical preparations are widely used for both
and anticancer nucleosides,14 we wondered if similar methods
prophylaxis of and therapy in OA.1 Some studies have indicated
may be effective here. We herein report our first efforts in this
a positive impact of glucosamine in both laboratory models and
regard.
human disease.2 Other clinical trials have revealed little or no
therapeutic benefit,3-5 leaving glucosamine as an unproven
therapy at present.
Chemistry
Rather than apply phosphate prodrug methods directly to
glucosamine (1, Figure 1), we considered N-acetylglucosamine
(2) a more appropriate target for several reasons. Thus,
A number of biochemical mechanisms of action of glu-
cosamine have been suggested in vitro.6-10 None has been
convincingly accepted as relevant in the clinic, although action
of the O-6 phosphate on cartilage degradation appears to be at
least one favored putative mechanism.11
N-acetylation is linked to the putative mechanism of action of
the glucosamine phosphate,11 N-acetylation satisfactorily pro-
tects the amine function in the phosphorylation reaction, it
further enhances the lipophilicity of the prodrug and removes
the potential for protonation in vitro, and there is also the
possibility of in vivo cleavage (thus acting as a dual prodrug).
Thus, we commenced our studies on 2 (Figure 1). Early
application of our conventional phenylaminoacyl phosphora-
midate chemistry14 here indicated a degree of ProTide instability,
with phenol liberation on chromatography, that we had not
previously encountered with very extensive (>2000 analogues)
experience of nucleoside O5′-phosphate analogues. This may
indicate some neighboring group participation, leading to phenyl
loss, in the products in this case, not possible for stereoelectronic
reactions in the ribose phosphates of nucleoside. In order to
stabilize the phenyl phosphate, we investigated electron-donating
substituents and found p-methoxy to be suitable. All further
examples were thus p-methoxyphenyl analogues.
Glucosamine is a highly polar molecule (ClogP ) -2.4)12
and indeed will be very largely protonated at physiological pH,
making it poorly membrane permeable by passive diffusion.
Active transport is possible via the glucose transporters, but
administered drug will then compete with endogenous glucose.13
Bearing in mind the issues of transport and phosphorylation,
we began to wonder whether a preformed glucosamine O-6
phosphate may have any therapeutic benefit. However, on
account of its charge, high polarity, and probable instability to
dephosphorylation in vivo, the free phosphate may be of limited
direct utility. Given our longstanding expertise and interest in
the use of phosphate prodrugs (ProTides) derived from antiviral
Thus, p-methoxyphenol was reacted with POCl3 in diethyl
ether at low temperature to afford the aryl phosphorodichloridate
in quantitative yield (Scheme 1). This was reacted with a series
of amino acid ester hydrochlorides in dichloromethane in the
presence of triethylamine to give a family of phosphorochlo-
ridates. The amino acids utilized were L-Ala, Phe, Pro, Val,
* To whom correspondence should be addressed. Phone/fax: +44 029
20874537. E-mail: mcguigan@cardiff.ac.uk.
†
Welsh School of Pharmacy, Cardiff University.
Cardiff School of Biosciences, Cardiff University.
Bioiberica S.A.
‡
§
a Abbreviations: OA, osteoarthritis; IL-1, interleukin1; GAG, glycosami-
noglycan; THF, tetrahydrofuran.
10.1021/jm800594c CCC: $40.75
2008 American Chemical Society
Published on Web 08/21/2008

5808 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 18
McGuigan et al.
Scheme 1. General Synthetic Pathway for the Synthesis of N-Acetylglucosamine Aryloxy Phosphoramidates
Scheme 2. Alternative Synthetic Pathway for the Synthesis of the Aryloxy Phosphoramidates Using O-1 Benzyl Protected
N-Acetylglucosamine
Leu, Ile, Gly, Met, D-Ala, and R,R-dimethylglycine, each as
their benzyl ester. Ethyl esters were prepared of Ala, Gly, Pro,
Val, and N-methylglycine (sarcosine), and other examples were
MeAla, iPrVal, cyclohexylVal, and n-butylPro. In most cases,
and in particular when the amino acid was chiral, the phospho-
rochloridates appeared as two closely spaced 31P NMR signals,
roughly equal in intensity, corresponding to the two phosphate
diastereoisomers. Proton NMR data also confirmed their struc-
tures. In most cases they were used crude in the next step; in
some cases they were purified by quick flash column chroma-
tography on silica.
Two alternative routes were explored to gain access to the
desired target compounds. In the first case, 2 was allowed to
react with the appropriate phosphorochloridate in THF/pyridine
in the presence of N-methylimidazole at -30 °C to room
temperature (Scheme 1). In every case complex mixtures of
regio- and stereoisomeric monophosphorylated products were
formed along with minor quantities of polyphosphates. Extensive
and repeated chromatographic purification gave the desired O-6
phosphates in every case, but yields were low and variable (in
the range of approximately e1-21%). Bearing the yield in
mind, an alternative route was investigated. Thus, 2 was
converted into its O-1-benzyl derivative using benzyl alcohol
and HCl gas. The product was isolated in ∼80% yield in an
R/ꢀ ratio of 2:1 as judged by NMR. This protected sugar was
reacted with several phosphorochloridate reagents (Scheme 2)
to give the O-1-blocked O-6-phosphates in 3-8% yield. This
yield was not sufficiently improved by comparison to reaction
with free 2 to warrant further exploration, but catalytic
hydrogenation was found to yield the desired O-1 free analogues
in good yield. It may well be that alternative protection of the
O-3 and/or O-4 sugar hydroxyl groups, with or without O-1
protection, would further improve the yield of O-6 phosphates,
but the routes herein described gave sufficient stocks of products
for initial biological evaluation.
In each case the target compounds were isolated as four
stereoisomers, R/ꢀ at C1 and R/S at the phosphorus. The R forms
1
clearly predominated by H NMR, while the phosphate stere-
oisomers were often roughly 1:1. 31P NMR in general showed
four peaks in the range δ 4.0-4.5. The nature of the O-6-
phosphate link was proven by 13C NMR where a 6 Hz
P-O-CH2 coupling was observed. Phosphorus coupling to
(only) the C6 of the sugar confirmed the location of the
phosphate as did downfield shifts of C6 relative to (2). Mass
spectral, HPLC, and/or microanalytical data (Supporting Infor-
mation) further confirmed the structure and purity of the target
compounds.
Discussion
To examine the affect of the glucosamine derivatives on the
loss of aggrecan (cartilage proteoglycan) from articular cartilage
(an early event in the development of matrix degradation in
osteoarthritis), a model in vitro culture system was established
whereby 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. 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 (Table 1, control GAG-fold), with the mean
value (n ) 2-5) reported. A fold-increase was also calculated
for cartilage explants cultured in the presence of glucosamine

Prodrugs from N-Acetylglucosamine
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 18 5809
Table 1. Efficacy and Toxicity Data of Glucosamine (1), N-Acetylglucosamine (2), and N-Acetylglucosamine Phosphoramidate Derivatives (3a-r)a
10 mM
1 mM
0.1 mM
% reduction
GAG
n GAG fold fold fold ((SE)
control
GAG GAG
% reduction
GAG
AA ester n fold fold fold ((SE)
control
GAG GAG
% reduction
GAG
fold fold fold ((SE)
MTT
((SE)
MTT
((SE)
control GAG
MTT
((SE)
n
1
2
2
1
2
1
2
1
2
1
2.6
3.2
2.6
2.1
2.6
3.2
2.6
2.1
2.0 19.0 (30.42) 56 (0.5)
2.9 9.3 61 (1.59)
2.0 24.2 (36.29) 85 (3.28)
5
3
3
4.9
4.4
4.4
4.7 3.4 (4.52)
4.5 -1.0 (7.57) 70 (1.68)
4.2 5.6 (3.88)
88 (2.51)
4
2
2
5.1
4.5
4.5
5.1 0.8 (5.93)
4.3 5.4 (0.41)
4.6 -2.2 (4.44)
106 (4.18)
107 (2.03)
104 (1.2)
3a Ala
3b Phe
3c Pro
3d Val
3e Leu
3f Ile
3g Me2Gly Bn
3h Gly
3i Met
Bn
Bn
Bn
Bn
Bn
Bn
81 (2.72)
2.1 2.3
49 tox(2.23)
0.8 65.6 (11.12) 63 (0.36)
1
4.2
4.1 2.2
0.9 77.6
2.6 44.7 (8.58) 20tox(0.02) 2
1.2 70.8
4.2 -0.3
1.2 73.6
1.6 62.2
2.8 31.0 (6.38) 66 (1.92)
76 (2.34)
85 (0.75)
100 (3.7)
2
9
3.3
4.1
4.7
3.9
3.3
3.9
4.7
4.2
4.2
3.0
4.7
3.8
3.7
3.8
4.9
4.7
4.0 -21.9 (41.32) 78 (1.45)
4.0 2.5 (9.50) 104 (0.88)
4.1 12.2 (30.42) 87 (2.98)
0.5 85.3
40 tox(1.38) 1 4.2
0.9 64.0 (15.93) 39 tox(1.48) 2 4.7
0.5 74.6
56 (1.92)
1
1
1
1
2
4.2
4.2
4.2
4.2
4.0
118 (6.15)
95 (1.11)
100 (1.85)
2
2
5
3.9 5.2 (6.46)
110 (3.12)
99 (1.56)
3.4 -2.2 (1.00)
Bn
Bn
4.2 -12.2 (15.42) 99 (1.92)
49tox(2.45) 2
3.8 18.0 (3.21)
103 (5.46)
3j D-Ala Bn
4
4
3.4 24.3 (16.28) 96 (2.91)
4.4 -1.9 (10.19) 89 (1.29)
3.6 -19.3 (13.90) 78 (2.05)
5.4 -16.3 (11.23) 100 (0.9)
4.4 -21.1 (25.80) 94 (3.19)
3.2 14.0 (19.01) 93 (1.24)
4.1 -17.5 (36.25) 98 (2.95)
4.8 -6.3 (24.31) 92 (1.9)
3k Ala
3l Gly
3m Sar
3n Val
3o Val
3p Pro
3q Val
3r Ala
Et
Et
Et
Et
iPr
nBu
Chex
Me
37tox(1.98) 4
100 (0.02)
2
2
4
2
4
2
2
4.0
3.6 12.1 (29.04) 56 (1.74)
30 (2.91)
37 (3.52)
3.9 15.8 (0.75)
97 (2.23)
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;
values indicated are the average of those experiments. 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; some high standard error readings are due to biological variability of the cartilage obtained.
derivatives with or without IL-1. These values have been
tabulated as GAG-fold enumeration in Table 1. This GAG-fold
was then taken and compared 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. 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 (Table 1). Drug efficacy was
measured as the reduction in this fold-increase in GAG release
into the culture media. Thus, glucosamine (1) treatment at 10
mM led to a reduction in the increased in GAG (before/after
exposure) to 2-fold, representing a 19% reduction relative to
the control (IL-1 only treated explant culture), while N-
acetylglucosamine (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 consider-
ably less cytotoxic at 1 and 0.1 mM, but at these lower
concentrations both of these compounds were unable to inhibit
IL-1 induced aggrecan release from the tissue (% reduction GAG
fold).
MTT assay. However, on further dilution, 3a loses activity,
being inactive at 1 and 0.1 mM. Thus, parent 3a appears more
active than both glucosamine and N-acetylglucosamine, but it
is only active at approximatey 10 mM and not 1 mM. If this in
vitro assay were reflected in vivo, such levels of drug would
most likely not be feasible, and thus, we sought to enhance the
potency of 3a by structural modification of the prodrug.
In the first instance the benzyl ester was retained and the
amino acid varied. The amino acid unit in ProTides has been
extensively studied by us for antiviral nucleosides, and alanine
has emerged as the most effective motif.14 Notably D-alanine
has in general been observed to be poorly effective.15 However,
given the very different target cell type in the present assay, it
remained possible that other (amino acid) structure-activity
relationships would operate in the present case. Thus, we first
prepared the phenylalanine analogue 3b, designed to increase
the side chain bulk of 3a and also its lipophilicity. However as
the data in Table 1 reveal, 3b is less active than 3a; indeed, it
is notably cytotoxic. Particularly given the toxicity profile of
3b, it was not pursued at higher dilutions.
Other amino acid variations in this first batch were proline
(3c), valine (3d), leucine (3e), and isoleucine (3f). All of these
analogues showed some degree of cytotoxicity, but some also
appeared to show enhanced potency over 3a in the GAG release
assay. Thus, each was studied at higher dilutions (1 and 0.1
mM). In general, toxicity was lost (with the exception of the
leucine compound 3e at 1 mM) and they became less active, in
a dose-dependent manner. Thus, 3d-f were all more active than
1, 2, and 3a at 1 mM, although 3e was toxic at this concentra-
tion. At 0.1 mM all of the analogues were nontoxic but all were
poorly active at best. Although differences did not reach
statistical significance, data implied that 3e (Leu) was the most
potent analogue to this point.
The first phosphate prodrug prepared was the p-methoxyphe-
nylbenzylalanine compound (3a). This was designed on the basis
of the efficacy of this ProTide motif on a range of antiviral and
anticancer nucleosides.14 Thus, it is notable that 3a is signifi-
cantly more effective than 1 and 2 at reducing the level of GAG
released into the media after IL-1 exposure, from 2.6-fold
(control) to 2.0-fold (3a at 10 mM). Moreover, 3a is significantly
less toxic than 1 and 2 at 10 mM with 85% cell viability by the

5810 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 18
McGuigan et al.
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 in H2SO4
solution. Preparative TLC plates (20 cm × 20 cm, 500-2000 µm)
were purchased from Merck.
Flash Column Chromatography (FCC). Flash column chro-
matography was carried out using silica gel supplied by Fisher (60A,
35-70 µm). Glass columns were slurry packed using the appropri-
ate 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.
The next batch of compounds (3g-j) featured other amino
acid variations again with benzyl ester retention. Because
activity had now been observed at 1 mM and given the low
yield of compounds and to preserve sample, subsequent batches
were not tested at 10 mM. Compound 3g contains the achiral
usual amino acid R,R-dimethylglycine, which we have found
to be highly effective in some nucleoside examples, particularly
hepatitis assays in hepatoma cells.16 However, here it was
inactive (and nontoxic) at 1 mM. However, the glycine (3h),
methionine (3i), and D-alanine (3j) compound all appeared active
at 1 mM, although the Met compound was also toxic. The
D-alanine case (3j) is particularly interesting given our prior
experience as noted above. Thus, 3j is more active than 3a (L-
Ala) at 1 mM, although with some toxicity, but the D-compound
retains activity at 0.1 mM, unlike the L-isomer, and is then
nontoxic. At 0.1 mM 3j gives 96% cell viability and reduces
the GAG increase from 4.2-fold in control to 3.4-fold (treated),
representing a 24% reduction in GAG increase relative to
control. It thus becomes the most active compound to date.
We finally examined ester variation, replacing the benzyl ester
by ethyl for a series of amino acids: Ala (3k), Gly (3l),
N-methylglycine (3m), and valine (3n). No clear trends were
observed here, but in general the ethyl esters were less active
than their benzyl analogues. Other esters were also briefly
investigated in a small number of cases. The isopropylvaline
case (3o) emerged as of potential interest here, being rather
similar to 3i described above. The cyclohexyl ester on valine
was poorly active in this case (3q).
High Performance Liquid Chromatography (HPLC). Ana-
lytical 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.
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
are quoted in parts per million (ppm) downfield from tetrameth-
ylsilane, and coupling constants (J) are in hertz. Spectra were
calibrated to the residual signal of the deuterated solvent used. 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 phosphoramidates involves the identification of
R and ꢀ sugar derivatives, and in some cases the diastereoisomers
of particular compounds can be identified and are arbitrarily
assigned A and B.
Conclusion
A family of 18 phosphate prodrugs of N-acetylglucosamine
is herein reported. Parent N-acetylglucosamine is inactive in
the bioassay used at 10 mM and also toxic. Glucosamine has
some chondroprotective activity at 10 mM, but not 1 mM, but
is also toxic at 10 mM. By contrast, several of the phosphate
prodrugs are active at concentrations as low as 0.1 mM and are
noncytotoxic. They are thus approximately g100-fold more
active than glucosamine in this assay. Unusual and novel SARs
have also emerged, with prior amino acid SARs in particular
being greatly revised. Of particular note is the emergence of
D-alanine as an amino acid of choice and also the relatively
poor activity of L-alanine, the previously preferred amino acid.
It is notable that the amino acid structure-activity relationships
we observe in this series, and in particular the relatively poor
performance of L-alanine, are markedly contrasting to those we
have extensively observed for phosphoramidate ProTides of
nucleosides.14 The reasons for these differences may arise from
structural (nucleoside versus sugar) or metabolic (cell to cell)
origins and are the subject of active current investigation in our
laboratories.
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 bombard-
ment (FAB).
Elemental Analysis (CHN). CHN microanalysis was performed
as a service by the School of Pharmacy at the University of London.
Standard Procedures. For practical purposes, standard proce-
dures are given. Any variations from these procedures are discussed
individually. Procedures that differ from the standard ones are
described in full.
Standard Procedure 1: Synthesis of Phosphorodichloridate
Species. Phosphorus oxychloride (1.0 mol equiv) and the appropri-
ate substituted phenol (1.0 mol equiv) were stirred with anhydrous
diethyl ether (30 mol equiv) under argon. Anhydrous triethylamine
(2.0 mol equiv) was added at -78 °C and allowed to warm to room
temperature for 4 h with stirring while under an anhydrous
atmosphere of argon. The triethylamine salt was removed by
filtration under nitrogen and the filtrate evaporated to give the
product as a clear liquid.
Standard Procedure 2: Synthesis of Phosphorochloridate
Species. Phosphorodichloridate (1.0 mol equiv) and the appropriate
amino ester hydrochloride salt (1.0 mol equiv) were suspended in
anhydrous DCM (61.6 mol equiv) under argon. Anhydrous triethy-
lamine (2.0 mol equiv) was added dropwise at -78 °C, and after
15 min the mixture was left to rise to room temperature and stirred
for 2 h while under an anhydrous atmosphere of argon. The for-
mation of phosphorochloridate was monitored by 31P NMR. The
solvent was removed under reduced pressure, after which the
resulting triethylamine salt was removed by resuspending the crude
product in anhydrous diethyl ether and the salt removed by filtration
under nitrogen or was purified by fast flash chromatography
[hexane/ethyl acetate 1/1 (v/v)].
Several additional hits have herein emerged that now form
the basis of a hit to lead optimization in our laboratories. If the
apparent potency enhancements over glucosamine that we herein
report can be translated to an in vivo efficacy boost, this family
could have the potential to represent a paradigm shift in the
glucosamine-based therapy of osteoarthritis treatments.
General Experimental Details
Chemistry. General Procedures. Solvents and Reagents. The
following anhydrous solvents were bought from Sigma-Aldrich:
dichloromethane (DCM), diethyl ether (Et2O), N-methylimidazole
(NMI), pyridine (pyr), tetrahydrofuran (THF), triethylamine (TEA),
amino acid ester salts, N-acetylglucosamine, and any other reagents
used. All reagents commercially available were used without further
purification.
Standard Procedure 3: Synthesis of Phosphoramidate Spe-
cies. N-Acetyl-D-glucosamine was coevaporated with pyridine three
times (17 mol equiv) and dried under vacuum overnight. To a

Prodrugs from N-Acetylglucosamine
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 18 5811
stirring solution/suspension of dry N-acetyl-D-glucosamine (1.0 mol
equiv) in anhydrous pyridine (105 mol equiv) and NMI (5.0 mol
equiv) the appropriate phosphorochloridate (1.2 mol equiv) in
anhydrous THF (1 M solution) was added, at - 40 °C, dropwise
over 15 min. After 15 min the mixture was left to rise to room
temperature and stirred at room temperature for 2-5 h. The reaction
was quenched with methanol, and the solvent was removed under
reduced pressure. The crude obtained was dissolved in DCM and
purified by flash chromatography [using a gradient elution method
of 100% DCM (v/v) to DCM/MeOH 85:15 (v/v)].
4-Methoxyphenyl Phosphorodichloridate. The compound was
prepared as described in standard procedure 1, using 4-methox-
yphenol (4.2 g, 33.83 mmol), phosphorus oxychloride (5.19 g, 33.83
mmol), triethylamine (4.75 mL, 33.83 mmol) in anhydrous diethyl
ether (100 mL). The product was obtained as a clear oil (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.
two major in R form and two minor in ꢀ form): δ 7.36 (5H, m,
-OCH2Ph), 7.13 (2H, m, -OPh), 6.87 (2H, m, -OPh), 5.6 (3H,
m, -CH-1R and -OCH2Ph), 4.63 (1H, m, -CH-1ꢀ), 4.31 (2H,
m, -CH2-6), 4.04 (1H, m, -CHCH3), 3.98 (1H, m, -CH-5), 3.88
(1H, m, -CH-2R), 3.77 (4H, m, -PhOCH3 and -CH-3R), 3.63
(1H, m, -CH-2ꢀ), 3.46 (1H, m, -CH-3ꢀ), 3.45 (1H, m, -CH-4),
2.02 (3H, s, -NHCOCH3), 1.41 (3H, m, -CHCH3). 13C NMR
(CD3OD, 125 MHz): δ 20.54 (d, J(C-P)) 7.5 Hz, -CHCH3 A),
20.59 (d, J(C-P)) 7.5 Hz, -CHCH3 B), 22.67 (-NHCOCH3,
R-isomer), 22.95 (-NHCOCH3, ꢀ-isomer), 51.66 (-CHCH3 A),
51.58 (-CHCH3 B), 55.85 (C-2 R-isomer), 56.12 (-OCH3,
R-isomer), 56.22 (-OCH3 ꢀ-isomer), 67.43 (d, J(C-P)) 5 Hz,
-CH2-6 A), 67.72 (d, (J(C-P)) 5 Hz, CH2-6, B), 67.96 (-OCH2Ph
A), 68.03 (-OCH2Ph B), 71.63 (d, J(C-P)) 7.5 Hz, C-5), 71.83
(C-4 A), 72.08 (C-4 B), 72.54 (C-3 A), 72.57 (C-3 B), 92.64 (C-
1R A), 92.69 (C-1R B), 97.14 (C-1ꢀ), 115.63 (-PhOCH3 A), 115.65
(-PhOCH3 B), 122.34 (d, J(C-P) ) 4.2 Hz, -PhOCH3 A), 122.43
(d, J(C-P) ) 4.2 Hz, -PhOCH3 B), 129.33 (-CH2Ph), 129.40
(-CH2Ph), 129.61 (-CH2Ph), 137.30 (“ipso”-CH2Ph), 145.68 (d,
J(C-P) ) 6.85 Hz, “ipso”-POPh A), 145.77 (d, J(C-P) ) 7.0 Hz,
“ipso”-POPh B), 158.26 (“ipso”-PhOCH3 A), 158.32 (“ipso”-
PhOCH3 B), 173.72, (-COCH3), 175.0 (d, J ) 5.75 Hz,
-CO2CH2Ph). 31P NMR (CD3OD, 202 MHz): 4.41 (R-isomer A),
4.39 (ꢀ-isomer A), 4.25 (R-isomer B), 4.14 (ꢀ-isomer B). MS (E/
I) 591.1716 (MNa+). HPLC: H2O 80% to CH3CN 20%, λ ) 275
nm, flow 1 mL/min, tR ) 4.40 min.
1-O-Benzyl-N-acetyl-D-glucosamine. The compound was syn-
thesized by the addition of N-acetyl-D-glucosamine (2.0 g, 9.0
mmol) into a solution of protonated benzyl alcohol (20 mL, 59.1
mmol), which was prepared by purging HCl gas into benzyl alcohol
(20 mL, 59.1 mmol) for approximately 15 min. The solution was
left stirring at room temperature and monitored by TLC; after 12 h
no starting material was seen. The solution was evaporated to an
emulsion and then dissolved in a small quantity of DCM and
precipitated out by addition of diethyl ether. The solid was collected
by filtration, and triturated using chloroform. This resulted in a white
solid (2.2 g, 7.0 mmol, 78%). Ratio of R to ꢀ sugar was found to
Biological Assay. Materials. All reagents were obtained from
Sigma (Poole, U.K.) and tissue culture plastics and reagents from
Invitrogen (Karlsruhe, Germany), unless otherwise stated.
1
be 2:1. H NMR (CDCl3, 500 MHz): δ 7.45-7.25 (5H, m, -Ph),
Isolation and Culture of Bovine Chondrocytes and MTT
Cytotoxicity Assay. Monolayer chondrocyte cultures were obtained
from the metacarpophalangeal joints of mature cattle (>18 months
old). Full-depth articular cartilage tissue slices were dissected under
sterile conditions and subjected to standard pronase and collagenase
digestion to isolate the chondrocytes using well established
procedures.17-19 In brief, the tissue was digested in 0.1% (w/v)
pronase (Streptomyces griseus; Boehringer Mannheim, Lewes,
U.K.) in Dulbecco’s modified Eagle’s medium (DMEM) containing
5% FCS (v/v) and gentamicin (50 µg/mL) for 1 h at 37 °C, 5%
CO2 with agitation, washed, and then further digested with 0.04%
(w/v) collagenase (Clostridium histolyticum; Worthington, Freehold,
NJ) in the same above media and incubated overnight with agitation
as before. Cells were filtered through a 40 µm Nitex filter (Falcon;
BD Biosciences, Bedford, MA) and washed before cell numbers
were established. Isolated cells were cultured in 1 mL of serum-
free DMEM media containing gentamicin (50 µg/mL), 1% (v/v)
HEPES buffer and supplemented with or without the appropriate
glucosamine compound (10-0.1 mM). After 96 h of culture, the
cell viability of the chondrocytes in the presence or absence of each
compound was determined using the 3-(4,5-dimethylthiazol-2-yl)-
2,5-diphenyltetrazolium bromide (MTT) cell viability assay.20 This
assay is a colorimetric toxicity assay based on the detection of
intracellular mitochondrial activity within viable cells. MTT (a water
soluble tetrazolium salt) is converted to an insoluble purple
formazan product (impermeable to cell membranes of viable cells)
by cleavage of the tetrazolium ring by succinate dehydrogenase
within the mitochondria of viable cells only. In brief, 200 µL of
the media was removed from each well and replaced with 200 µL
of 5 mg/mL thiazolyl blue tetrazolium bromide to make a final 1
mg/mL concentration and cultured for 3 h. The 1 mL was then
removed, and the formed intracellular MTT formazan crystals were
dissolved with the addition of 200 µL of DMSO; 100 µL from
each well was transferred to a 96-well plate, and the optical density
was measured at 570 nm. Chondrocytes cultured in the absence of
any glucosamine compound were taken as 100% viable, and the
percentage toxicity of experimental cultures was determined as [(OD
at 540 nm treated cells)/(OD at 540 nm untreated cells)] × 100.
Articular Cartilage Explant Cultures and Analysis of Gly-
cosaminoglycan Concentration in Media. Full-depth articular
cartilage explants were dissected from mature bovine metacar-
pophalangeal joints. The explants (50-100 mg/explant) underwent
4.92 (1H, d, J ) 12.0 Hz, -C(H)HPh ꢀ), 4.87 (1H, d, J ) 3.9 Hz,
H-1R), 4.88 (1H, d, J ) 12.0 Hz, -C(H)HPh R), 4.62 (1H, d, J )
12.0 Hz, -CH(H)Ph ꢀ), 4.52 (1H, d, J ) 12.0 Hz, -CH(H)Ph R),
4.49 (1H, d, J ) 8.5 Hz, H-1ꢀ), 3.92 (1H, d, J ) 3.4 Hz, H-2R),
3.92 (1H, d, J ) 3.4 Hz, H-2R), 3.90 (1H, d, J ) 3.6 Hz, H-2 ꢀ),
3.86 (1H, d, J ) 11.6 Hz, (H)H-6ꢀ), 3.85 (1H, d, J ) 11.6 Hz,
(H)H-6R), 3.77-3.68 (2H, m, H(H)-6 R/ꢀ, H-3), 3.41-3.33 (2H,
m, H-4, H-5), 1.98 (3H, s, -NCOCH3). 13C NMR (CDCl3, 126
MHz): δ 173.43, 173.38 (-NCOCH3), 139.45, 139.23 (“ipso”-
CH2Ph), 129.56, 129.50, 129.35, 128.97, 128.90, 128.80 (-Ph),
102.07 (C-1ꢀ minor isomer), 97.64 (C-1R), 78.22 (C-5 R), 76.03
(C-5 ꢀ), 74.28 (C-3), 72.65, 72.51 (C-4), 72.65 (1-OCH2Ph ꢀ), 72.51
(1-OCH2Ph R), 62.94 (CH2-6ꢀ), 62.80 (CH2-6R), 57.40 (C-2ꢀ),
55.52 (C-2R), 23.30 (-NCOCH3 ꢀ), 22.84 (-NCOCH3 R). MS
(ES+) m/z: 335.1 (MNa+H+).
4-Methoxyphenyl(benzoxy-L-alaninyl) Phosphorochloridate.
The compound was prepared as described in standard procedure 2,
using methoxyphenyl phosphorodichloridate (8.2 g, 34.0 2mmol),
L-alanine benzyl ester hydrochloride (7.34 g, 34.02 mmol), triethy-
lamine (9.4 mL, 68.04 mmol), anhydrous dichloromethane (100
mL) under argon. The solvent was removed under reduced pressure,
the crude residue was resuspended in anhydrous ether, filtered, and
the filtrate was reduced to give the product as a clear oil (12.07 g,
92.5% yield). 1H NMR (CDCl3, 500 MHz): δ 7.20 (5H, m,
-CH2Ph), 7.10 (2H, m, -OPh), 6.77 (2H, m, -OPh), 55.14 (2H,
s, -OCH2Ph A), 5.12 (2H, s, -OCH2Ph B), 4.20 (2H, m, -CHNH
and -NH), 3.71 (3H, s, -PhOCH3), 1.44 (3H, d, J ) 6.90 Hz,
-CHCH3, A), 1.43 (3H, d J ) 6.70 Hz, -CHCH3 B). 31P NMR
(CDCl3, 202 MHz): δ 8.54, 8.21, (ratio 1: 1).
2-Deoxy-2-(acetylamino)-6-[4-methoxyphenyl(benzoyl-L-ala-
nine)]phosphate-D-glucopyranoside (3a). The compound was
prepared as described in standard procedure 3, using N-acetyl-D-
glucosamine (3 g. 13.56 mmol), NMI (6.2 mL, 78.15 mmol), a
solution of 4-methoxyphenyl(benzoxy-L-alaninyl) phosphorochlo-
ridate (6 g, 15.63 mmol) in anhydrous THF (15 mL), and pyridine
(100 mL). After 15 min the mixture was allowed to slowly warm
to room temperature and stirred at room temperature for 3 h. The
crude residue was purified by flash chromatography using a gradient
elution solvent system of DCM/MeOH (98:2 to 95:5, then 9:1 v/v).
This gave the pure product 3a as a white solid (0.6 g, 7.8% yield).
1H NMR (CD3OD, 500 MHz, mixture of four diastereoisomers,

5812 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 18
McGuigan et al.
(5) Clegg, D. O.; Reda, D. J.; Harris, C. L.; Klein, M. A.; O’Dell, J. R.;
Hooper, M. M.; Bradley, J. D.; Bingham, C. O., III; Weisman, M. H.;
Jackson, C. G.; Lane, N. E.; Cush, J. J.; Moreland, L. W.; Schumaker,
H. R., Jr.; Oddis, C. V.; Wolfe, F.; Molitor, J. A.; Yocum, D. E.;
Schnitzer, T. J.; Furst, D. E.; Sawitzke, A. D.; Shi, H.; Brandt, K. D.;
Moskowitz, R. W.; Williams, H. J. Glucosamine, chondroitin sulfate,
and the two in combination for painful knee osteoarthritis. N. Engl.
J. Med. 2006, 354, 795–808.
(6) Noyszewski, E. A.; Wroblewski, K.; Dodge, G. R.; Kudchodkar, S.;
Beers, J.; Sarma, A. V.; Reddy, R. Preferential incorporation of
glucosamine into the galactosamine moieties of chondroitin sulfate in
articular cartilage explants. Arthritis Rheum. 2001, 44, 1089–1095.
(7) Bassleer, C.; Rovati, L. C.; Franchimont, P. Glucosamine sulfate
stimulates proteoglycan production in human chondrocytes in Vitro.
Osteoarthritis Cartilage 1998, 6, 427–434.
(8) Dodge, G. R.; Jimenez, S. A. Glucosamine sulfate modulates the levels
of aggrecan and matrix metalloproteinase-3 synthesized by cultured
human osteoarthritis articular chondrocytes. Osteoarthritis Cartilage
2003, 11, 424–432.
(9) Piperno, M.; Reboul, P.; Hellio Le Graverand, M. P.; Peschard, M. J.;
Annefeld, M.; Richard, M.; Vignon, E. Glucosamine sulfate modulates
dysregulated activities of human osteoarthritic chondrocytes in vitro.
Osteoarthritis Cartilage 2000, 8, 207–212.
an equilibraion period by culture in DMEM media containing 10%
FCS (v/v), gentamicin (50 µg/mL), and 1% (v/v) HEPES buffer
for 48 h, after which explants were washed (3 × 10 min) in the
same media minus FCS. Explants were then cultured in triplicate
in individual wells of a 24-well tissue culture plate containing 1
mL of DMEM media (as previously described minus FCS) in the
presence or absence of the appropriate glucosamine compound
(10-0.1 mM), after which 10 ng/mL recombinant human IL-1R
was added to half of the wells. After a further 96 h culture period,
the sulfated GAG (representative of the proteoglycan content of
the media or explants following papain digestion) was determined
using the dimethyl-methylene blue (DMMB) dye binding assay as
previously reported21 with chondroitin sulfate (CS) from shark
chondroitin sulfate C used as a standard. The DMMB assay is a
strong metachromatic dye that binds to the negatively charged
glycosaminoglycan chains attached to the core protein of aggrecan.
This causes a change in the absorption spectrum of the dye. In
order to solubilize the cartilage explants the cysteine proteinases,
papain was used to degrade and solubilize the extracellular matrix.
Cartilage explants were digested with 1 µg/mL papain suspension
(papaya latex) in 0.05 M sodium acetate, 0.025 M EDTA, pH 5.6,
containing 5 mM cysteine HCl. The papain and cysteine were added
just prior to use. Following the appropriate dilutions of the samples
in a final 40 µL volume, 200 µL of DMMB solution was added
and the absorbance read at 525 nm. From these data a percentage
fold-increase in GAG release generated by the addition of IL-1
was calculated relative to the appropriate control culture (i.e., with
[GAG-fold, Table 1] or without [control GAG-fold, Table 1] the
addition of glucosamine, N-acetylglucosamine, or the phosphora-
midate dervivatives) as % GAG release from appropriate control
culture divided by the % GAG release from IL-1 treated culture,
where % GAG release was calculated as
(10) Sandy, J. D.; Gamett, D.; Thompson, V.; Verscharen, C. Chondrocyte-
mediated catabolism of aggrecan: aggreganase-dependent cleavage
induced by interleukin-1 or retinoic acid can be inhibited by glu-
cosamine. Biochem. J. 1998, 335, 59–66.
(11) Moberasheri, A.; Vannucci, S. J.; Bondy, C. A.; Carter, S. D.; Innes,
J. F.; Arteaga, M. F.; Trujillo, E.; Ferrez, I.; Shakibaei, M.; Martin-
Vasallo, P. Glucose transport and metabolism in chrondocytes: a key
to understanding chrondogenesis, skeletal development and cartilage
degradation in osteoarthritis. Histol. Histopathol. 2002, 17, 1239–1267.
(12) CLogP value was calculated from the following computer software
program: Chemdraw Ultra, version 9.0; CambridgeSoft: Cambridge,
MA, 1986-2004; www.cambridgesoft.com.
(13) Mroz, P. J.; Silbert, J. E. Use of 3H-glucosamine and 35S-sulphate
with cultured human chondrocytes to determine the effect of glu-
cosamine concentration on the formation of chondroitin sulphate.
Arthritis Rheum. 2004, 50 (11), 3574–3579.
Total GAG released into culture media × 100.
(14) Cahard, D.; McGuigan, C.; Balzarini, J. Aryloxy phosphoramidate
triesters as protides. Mini-ReV. Med. Chem. 2004, 4, 371–382.
(15) McGuigan, C.; Salgado, A.; Yarnold, C.; Harries, T. Y.; De Clercq,
E.; Balzarini, J. Novel nucleoside phosphoramidates as inhibitors if
HIV: studies on the stereochemical requirements of the phosphora-
midate amino acid. AntiViral Chem. Chemother. 1996, 7, 184–188.
(16) McGuigan, C.; Harris, S. A.; Daluge, S. M.; Gudmundsson, K. S.;
McLean, E. W.; Burnette, T. S.; Marr, H.; Hazen, R.; Condreay, L. D.;
Johnson, L.; De Clercq, E.; Balzarini, J. Application of phosphora-
midate pronucleotide technology to abacavir leads to a significant
enhancement in antiviral potency. J. Med. Chem. 2005, 48, 3504–
3515.
{(total GAG released into culture media) ⁄
[(total GAG released into culture media) +
(total GAG from papain digested explant)]} × 100
Acknowledgment. We thank Helen Murphy for excellent
secretarial assistance.
Supporting Information Available: Microanalytical data of
target compounds and HPLC data and spectra of target compounds.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(17) Handley, C. J., III. Modulation of proteoglycan synthesis by extra-
cellular levels of proteoglycan in cartilage cells in culture. Biochim.
Biophys. Acta 1977, 500, 132–139.
(18) Handley, C. J.; Brooks, P. R.; Lowther, D. A. Extracellular matrix
metabolism by chondrocytes. VI. Concomitant depression by exog-
enous levels of proteoglycan of collagen and proteoglycan synthesis
by chondrocytes. Biochim. Biophys. Acta 1978, 544, 441–444.
(19) Hughes, C. E.; Little, C. B.; Buttner, F. H.; Bartnik, E.; Caterson, B.
Differential expression of aggrecanase and matrix metalloproteinase
activity in chondrocytes isolated from bovine and porcine articular
cartilage. J. Biol. Chem. 1998, 273, 30576–30582.
(20) Mosmann, T. Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays. J. Immunol.
Methods 1983, 16, 55–63.
(21) Farndale, R. W.; Buttle, D. J.; Barrett, A. J. Improved quantitation
and discrimination of sulphated glycosaminoglycans by use of
dimethylmethylene blue. Biochim. Biophys. Acta 1986, 883, 173–177.
References
(1) Sutton, L.; Rapport, L.; Lockwood, B. Glucosamine: on or cure.
Nutrition 2002, 18, 534–536.
(2) Reginster, J. Y.; Bruyere, O.; Neuprez, A. Review: current role of
glucosamine in the treatment of osteoarthritis. Rheumatology 2007,
46, 731–735.
(3) Reginster, J. Y.; Deroisy, R.; Rovati, L. C.; Lee, R. L. Long-term
effects of glucosamine sulphate on osteoarthritis progression: a
randomised, placebo controlled clinical trial. Lancet 2001, 357, 251–
256.
(4) Altman, R. D.; Abramson, S.; Bruyere, O.; Clegg, D.; Herrero-
Beaumont, G.; Maheu, E.; Moskowitz, R.; Pavelka, K.; Reginster, J. Y.
Commentary: osteoarthritis of the knee and glucosamine. Osteoarthrits
Cartilage 2006, 14, 963–966.
JM800594C