Synthesis and study of alendronate derivatives as potential prodrugs of alendronate sodium for the treatment of low bone density and osteoporosis

Article abstract of DOI:10.1021/jm060398v

Alendronate derivatives were evaluated as potential prodrugs for the osteoporosis drug alendronate sodium in an attempt to enhance the systemic exposure after oral administration. An investigation of the chemical behavior of alendronate derivatives led to

Full text of DOI:10.1021/jm060398v

3060
J. Med. Chem. 2006, 49, 3060-3063
Synthesis and Study of Alendronate Derivatives
as Potential Prodrugs of Alendronate Sodium for
the Treatment of Low Bone Density and
Osteoporosis
Figure 1. Bisphosphonate drugs approved by the FDA for osteoporosis.
Petr Vachal,*,† Jeffrey J. Hale,† Zhe Lu,† Eric C. Streckfuss,‡
Sander G. Mills,† Malcolm MacCoss,† Daniel H. Yin,§
Kimberly Algayer,§ Kimberly Manser,§ Filippos Kesisoglou,§
Soumojeet Ghosh,§ and Laman L. Alani§
Department of Basic Chemistry, Merck Research
Laboratories, Merck & Co., Rahway, New Jersey 07065,
Research Operations, Merck Research Laboratories, Merck & Co.,
Rahway, New Jersey 07065, and Department Pharmaceutical
Research, Merck Research Laboratories, Merck & Co.,
West Point, PennsylVania 19486
Figure 2. Design of potential alendronate prodrugs.
ReceiVed April 4, 2006
Figure 3. Rearrangement of the 1-hydroxy-1,1-bisphosphonate to
1-phosphonate-1-phosphate observed for etindronates4c,d,l and alendr-
onates.
Abstract: Alendronate derivatives were evaluated as potential prodrugs
for the osteoporosis drug alendronate sodium in an attempt to enhance
the systemic exposure after oral administration. An investigation of
the chemical behavior of alendronate derivatives led to development
of practical synthetic strategies and prediction of each structural class’s
prodrug potential. Pharmacokinetic studies of N-myristoylalendronic
acid revealed that 25% have been converted in vivo after iv administra-
tion in rat, providing an important proof-of-concept for this strategy.
tetraalkyl alendronates, N-acylalendronates (design 2) also
abolish the zwitterionic nature of the parent drug and increase
the lipophilicity of the molecule. We primarily focused on
N-acylalendronates derived from fatty acids, contemplating the
possibility of an in vivo lipase-based deacylation (design 2, RCO
) fatty acid residue). Simple N-alkylalendronates were not
considered as a viable prodrug option because they are unlikely
to significantly alter the pharmacokinetic properties (Figure 1;
alendronate and ibandronate have identical oral bioavailability).3
(3) O-Acylalendronates are anticipated to have a limited stability
in phase I metabolism and thus are only considered in the
context of derivatives of the first two designs. In this paper,
our findings for each of the alendronate prodrug designs are
discussed in detail.
While the reactivity of tetraalkyl 1-hydroxy-1,1-bisphospho-
nates has been extensively studied,4 a general methodology for
the preparation of 1-hydroxy-1,1-bisphosphonates has not yet
been developed. Most of the published synthetic strategies are
limited to a handful of simple unfunctionalized 1-hydroxy-1,1-
bisphosphonates and not applicable to derivatives bearing any
functional groups either unprotected or protected in a manner
that would allow for further selective derivatization. It is noted
that no synthesis of a simple tetraalkyl alendronate has ever
been reported.5 To evaluate the potential of tetraalkyl alen-
dronates to be used as prodrugs for the parent alendronic acid,
we first devised a practical and scalable synthetic route for their
preparation.
Attempts to directly esterify the phosphonic acid functionality
of a readily available5 alendronic acid could not be carried out
because of rapid rearrangement of the desired tetraalkyl 1-hy-
droxy-1,1-bisphosphonate product to the 1-phosphonate-1-
phosphate byproduct under various reaction conditions (Figure
3). The rearrangement has been discovered by Fitch and
Moedritzer6 and recently studied in detail by Niemi, Turhanen,
and co-workers in the context of etidronate derivatives.4c,d,l
Niemi and Turhanen not only proposed the mechanism of the
rearrangement but also defined conditions under which it takes
place: at greater than pH 8 and/or at 60 °C and higher. Our
observations for alendronate derivatives are generally consistent
with their results for etidronates. The relatively broad scope of
Low bone density and osteoporosis represent a major health
threat for millions of people worldwide. According to the U.S.
Surgeon General, by 2020 half of all Americans over the age
of 50 could be at risk for fractures from low bone mass or
osteoporosis.1 Significant efforts devoted to prevention, treat-
ment, and impediment of osteoporosis yielded a class of highly
effective structurally related nonhormonal bisphosphonate drugs
including the FDA-approved alendronate sodium, ibandronate
sodium, and risedronate sodium. The enormous benefit of this
class could be enhanced by addressing two major downfalls
common to all three drugs: virtually identical low oral bio-
availability (F ) 0.6%) and incidents of upper gastrointestinal
tract irritation. Both shortcomings may be theoretically dealt
with concurrently by administrating the active compound as a
more orally bioavailable derivative (a prodrug) that would
release the active form of the bisphosphonate in vivo along with
innocuous byproducts, translating in a higher systemic exposure
of the parent drug. This report outlines the results of our
investigation of the potential of different classes of alendronate
derivatives to serve as such prodrugs by studying their chemical
behavior, in vitro and in vivo stability, and their ability to release
the parent alendronic acid.
Figure 2 outlines the proposed designs of alendronate
prodrugs. Derivatization of any of the four functional groups
of alendronic acid may theoretically provide compounds of the
desired properties: (1) Tetraalkyl alendronates (Figure 2, design
1) no longer have a zwitterionic character of aminobisphos-
phonic acid, which in turn has a pronounced effect on the overall
polarity of the molecule because of the lack of acidic protons.
The polarity change of the molecule and increased lipophilicity
may improve the pharmacokinetic profile.2 (2) Similar to
* To whom correspondence should be addressed. Phone: 1-732-594-
6624. Fax: 1-732-594-2210. E-mail: petr_vachal@merck.com.
† Department of Basic Chemistry.
‡ Research Operations.
§ Department of Pharmaceutical Research.
10.1021/jm060398v CCC: $33.50 © 2006 American Chemical Society
Published on Web 05/09/2006

Letters
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 11 3061
Scheme 1. One-pot Two-Stage Reaction Sequence for the
Preparation of 1-Hydroxy-1,1-bisphosphonate 4a
Scheme 2. Use of O-Acyl-1-hydroxy-1,1-bisphosphonates as
Intermediates for the Synthesis of Other Alendronate
Derivatives: Potential and Limitationsa
a Conditions: (a) P(OEt)3, 0 °C; (b) HOP(OEt)2, base, solvent (see Table
1), room temp, 1 h.
Table 1. Optimization of the Pudovik Reaction (Scheme 1, 3 f 4):
Conversion and Selectivity as Functions of the Solvent Polarity and
Base Strengtha
a Conditions: (a) RCOCl, 60 °C, 36 h, 15%; (b) NaN3, DMF, 60 °C, 15
h, 17%; (c) Pd-C, DCM, H2, 2 h, >99%.
ratio 4:5 @ 90% conversion of 3 f 4 + 5
combining both steps into a single one-pot reaction sequence
not only simplified the overall synthetic process but also
circumvented the decomposition of 3 observed upon its pro-
longed storage. Analytically pure 1-hydroxy-1,1-bisphosphate
4 was isolated in modest but fully reproducible overall yield
(Scheme 1).
in various solvents
base B
(pKa of BH+
tolu- hex-
cyclo-
in H2O)14
DMF THF CH2Cl2 ene
ane
hexane none
imidazole (7.0) 1.5:1 3:1
5:1
7:1
5:1
11:1
13:1
7:1
6:1
(41%b)
(38%b)
10:1
DMAP (9.2)
1:1
1.5:1 4:1
9:1
(37%b)
3:1
0.5:1
0:1
Once isolated in pure form, 1-hydroxy-1,1-bisphosphonates
4 are bench-stable for extended periods of time. Unfortunately,
their good stability in neat form (t1/2,4f5 > 1 year)15 does not
translate into stability in organic solvents (t1/2,4f5 ) 72 and 48
h for solutions at room temperature in cyclohexane and
dichloromethane, respectively). The rearrangement is even more
rapid in aqueous solutions; at biological pH 7.4 and 37 °C, the
rearrangement is >95% complete in less than 2 h (t1/2,4f5 < 30
min). The in vitro stability studies clearly indicate that tetraalkyl
alendronates 4 are unlikely to significantly improve the systemic
exposure of the parent drugs after oral dosing because they are
more rapidly rearranged then hydrolyzed to the parent drug.
These observations may represent a fundamental obstacle to the
application of tetraalkyl alendronates as prodrugs for alendronate
sodium.
1-Hydroxy-1,1-bisphosphonates with protected hydroxyl func-
tionality represent extremely valuable synthetic intermediates
because in the absence of a free hydroxyl group, undesired
rearrangement depicted in Figure 3 cannot take place. In this
context, Niemi and Turhanen investigated the use O-acetyletidr-
onate derivatives.4c,d,l However, we found that the O-acylation
strategy was not applicable to alendronate derivatives mainly
because of an intramolecular O f N acyl transfer of tetraalkyl
O-acylalendronate 8 to tetraalkyl N-acylalendronate 9 via an
unusually kinetically favorable 7-exo-trig mechanism (Scheme
2).16 Despite the fact that O f N acyl transfer may be utilized
for the synthesis of N-acylalendronates, a more general protec-
tive group strategy was needed for the preparation of other
alendronate derivatives. In addition, the formation of 6 and 7
was low-yielding and not practical for the preparation of large
quantities of 9 (Scheme 2).
The use of silicon-based protective groups proved to be the
most synthetically useful strategy for the preparation of alen-
dronate derivatives. We based our approach on the previously
optimized conditions outlined in Scheme 1. Dichloromethane
with DMAP was found to be the optimal condition for all of
the following elements of the overall synthetic sequence: the
Michaelis-Arbuzov reaction, suppression of the rate of the
undesired rearrangement, suppression of the rate of silylenolether
formation resulting from the base-promoted enolization of
R-ketophosphonate prior to the Pudovik reaction, the Pudovik
reaction, and the O-TBS protection of the Pudovik reaction
product.17 The optimized conditions enabled us to prepare
1-trialkylsilyloxy-1,1-bisphosphonate 10 in 45% overall yield
(three steps) via a practical one-pot sequence, overcoming all
caveats of the preexisting alternatives (Scheme 3).18,19 Silylated
chloride 10 can be subsequently converted to azide 11 in
DIPEA (11)
DBU (12c)
KHMDS (26)
1:1
na
0:1
1:1
0:1
0:1
5:1
1:1
0:1
7:1
1:1
6:1
1.5:1
7:1
1:1
0.05:1 0.05:1 0.1:1
a Ratio 4:5 was determined by 31P NMR of the crude reaction mixture.
b Isolated yield after column chromatography. c pKa in DMSO.
the conditions favorable for the rearrangement caused a failure
of all investigated methods for a direct conversion of N-
(carbobenzyloxy)alendronic acid (N-Cbz-alendronic acid, 1) to
its tetraalkyl ester.7 The presence of the highly reactive primary
amine functionality in the alendronate structure precluded the
use of a previous preparative method designed specifically to
prevent the rearrangement of related 1-hydroxybisphosphonates.8
Additionally, the position of the 4-amino group caused formation
of complex mixtures of intra- and intermolecular phosphora-
mides under many synthetic conditions. Because of these
challenges, it was determined that an alternative well-defined
synthetic strategy was needed for the preparation and isolation
of tetraalkyl alendronates.
The most commonly utilized strategies for the preparation
of 1-hydroxy-1,1-bisphosphonates4 involve some variation of
a sequence consisting of trialkyl phosphite addition to an acyl
chloride, called the Arbuzov or Michaelis-Arbuzov reaction9
(Scheme 1, 2 f 3), followed by dialkyl phosphite addition to
the R-ketophosphonate 3, commonly referred to as the Pudovik
reaction (3 f 4).10 In many cases, low overall yield and/or yield
variability was observed mainly because of the rearrangement
depicted in Figure 3. In some cases, the rearranged products
were even incorrectly reported as being the desired 1-hydroxy-
1,1-bisphosphonates.11 We conducted a thorough investigation
of the reaction sequence and found that the Michaelis-Arbuzov
reaction proceeded smoothly under solvent-free conditions at 0
°C with no excess of reagents and no catalyst or base needed.
The only reaction byproduct, alkyl chloride, does not interfere
with the majority of the subsequent reactions. For many
examples it can be removed under reduced pressure. A more
extensive optimization was needed for the second stage of the
sequence, the Pudovik reaction, which requires the use of a
stoichiometric amount of base.12 The base not only promotes
the desired transformation of 3 to 4 but also simultaneously
catalyzes the undesired rearrangement of 4 to 5. The relative
rates of both transformations are functions of two factors: the
strength of the base and the polarity of the reaction media.13
Table 1 demonstrates that as long as a base is strong enough to
effectively promote the Pudovik reaction, the weaker the base
and the less polar the solvent, the higher is the ratio of 4 to 5
and thus the higher is the overall isolated yield of 4. In addition,

3062 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 11
Letters
Scheme 3. Practical One-Pot, Three-Step Synthetic Route to
Protected N-Acyl-1-trialkylsilyloxy-1,1-bisphosphonate 10 and
Its Conversion to N-Acylalendronic Acid 15a
kinetic evidence that a prodrug strategy may be a viable option
for alendronate derivatives.
To establish the oral bioavailability of potential prodrugs, each
of the four groups of rats (n ) 3 per group) was orally dosed
with alendronic acid, 1, 15a, and 15b at 1, 2, and 5 mpk
equivalents to free alendronic acid. In the animals orally
administrated with alendronic acid and alendronate prodrugs,
0.23% of alendronic acid, 0.003% of 1, 0.02% of 15a, and
<0.001% of 15b were excreted in urine as free alendronic acid
within 48 h after dosing. None of the investigated prodrugs
provided an enhanced level of oral bioavailability. Improving
the overall pharmacokinetic profile of these derivatives including
modifications of their formulation for oral administration is
ongoing and will be a subject of future disclosures.
a Conditions: (a) P(OEt)3, 0 °C; (b) HOP(OEt)2, DMAP, CH2Cl2, room
temp, 1 h; (c) TBSCl, 15 h, 45% (three steps, one-pot); (d) NaN3, DMF,
75 °C, 2 h, 82%; (e) Pd-C, 50 psi of H2, AcOEt, 15 h, >99%; (f) RCOCl,
Et3N, CH2Cl2 or RCO2H, EDC, DIPEA, MeCN; (g) TMSI, MeCN and then
MeOH, 90%; (h) TBAF, THF, 93%.
In conclusion, we have investigated the alendronic acid
prodrugs. We have devised general synthetic strategies for the
preparation of several classes of alendronate derivatives,
producing intermediates in practical and reproducible yields.
The key steps of the syntheses include optimized Michaelis-
Arbuzov and Pudovik reactions performed in a one-pot se-
quence. The detailed knowledge of the chemical and physical
properties of alendronate derivatives enabled us to predict each
class’s potential to serve as prodrugs. We have identified
N-acylalendronates as the most promising class of alendronic
acid prodrugs where 25% of the leading example, N-myristoyl-
alendronic acid (15a), is converted to the parent produg in vivo
after iv dosing in the rat. This pharmacodynamic evidence
represents the first proof-of-concept of the prodrug strategy for
any alendronate derivative reported to this date.
significantly improved yield compared to its acetylated coun-
terpart 6 and further quantitatively reduced to amine 12. Because
of the practical protection of the hydroxyl functionality, 12 may
now be derivatized under a variety of conditions to afford
N-acylalendronates 13a-d. Standard nucleophilic dealkylation20
of both phosphonate groups afforded O-TBS-N-acylalendronic
acid 14; deprotection of 14 was accomplished using aqueous
hydrofluoric acid or tetrabutylamonium fluoride (TBAF) to
afford N-acylalendronic acid 15.
Acids 1, 14, and 15 are stable in solutions at wide ranges of
temperature and pH regardless of the solvent and do not undergo
the rearrangement outlined in Figure 3. Having established their
general stability, we then tested the potential for the amidic
functionality to undergo a hydrolytic cleavage. In in vitro
experiments, incubation of 15a with human intestinal mucosa
cell homogenate and with human and dog plasma did not
provide detectable amounts of released parent drug. A study of
uptake and conversion of 15a across rat intestinal tissue also
failed to provide any evidence of the conversion of 15a to the
parent prodrug.
Acknowledgment. The authors express their gratitude to
Falguni Patel and Dr. Bernard Choi for performing HRMS
analyses.
Supporting Information Available: Experimental procedures
and characterization of 1, 4-7, and 9-15 by 1H, 13C, and 31P NMR
and HRMS analyses and HPLC analysis traces under two diverse
conditions. This material is available free of charge via the Internet
at http://pubs.acs.org.
Because it is difficult for in vitro models to accurately reflect
the in vivo behavior of prodrug conversions, in particular, the
lack of conversion of 1 and 15 to the parent drug, we resorted
to in vivo pharmacokinetic experiments in the rat. Four groups
of rats (three rats per group) were administered iv21 with
alendronic acid, 1, 15a, and 15b, respectively, at a dose
equivalent to 0.1 mpk of free alendronic acid. In the group of
rats dosed with parent alendronic acid and alendronate prodrugs,
30% of alendronic acid, 8% of 15a, 4% of 1, and <1% of 15b
were eliminated in urine as free alendronic acid within 24 h
after dosing. The rapid excretion of alendronic acid for animals
dosed with parent drug is fully consistent with previously
published work that demonstrated that alendronic acid not
absorbed by bone tissue was rapidly excreted and served as
biomarker readout for the quantitation of the amounts actually
absorbed by the bone tissue.22 Although the amounts of
alendronic acid excreted after dosing with prodrugs were
generally lower compared to amounts from dosing with the
parent drug, the data indicate that at least two of the prodrugs,
1 and 15a, were converted to the parent alendronic acid in vivo.
Acid 15a represents a particularly promising lead because 25%
of the total amount of 15a dosed underwent an in vivo
conversion to alendronic acid. Although the mechanism of the
in vivo cleavage of the amidic functionality of 15a is presently
unknown, we speculate that it is unlikely to be a simple
unactivated hydrolysis because of the absence of such cleavage
in the case of 15b. To our knowledge, activated conversion of
15a to its parent alendronic acid is the first in vivo pharmaco-
References
(1) Carmona, R. H. Report on Bone Health and Osteoporosis. United
States Department of Human Health Services Press Release, October
14, 2004.
(2) In this report, we do not investigate tri-, di-, or monoalkyl alendr-
onates because the in vivo hydrolysis of the second alkyl ester of
either phosphonate functionality was expected to be very slow and
therefore not considered an effective prodrug based on our concept.
The ultimate goal for tetraalkyl alendronate prodrugs was to design
structures in which the hydrolysis of the first ester would automati-
cally trigger the release of the second one for each of the two
phosphonates.
(3) Study of alendronate prodrugs derived from N-linked peptides: Ezra,
A.; Hoffman, A.; Breuer, E.; Alferiev, I. S.; Moenkkoenen, J.; El
Hanany-Rozen, N.; Weiss, G.; Stepensky, D.; Gati, I.; Cohen, H.;
Toermaelehto, S.; Amidon, G. L.; Golomb, G. J. Med. Chem. 2000,
43, 3641.
(4) (a) Xie, Y.; Ding, H.; Qian, L.; Yan, X.; Yang, C.; Xie, Y. Bioorg.
Med. Chem. Lett. 2005, 15, 3267. (b) Migianu, E.; Guenin, E.;
Lecouvey, M. Synlett 2005, 425. (c) Turhanen, P. A.; Vepsalainen,
J. J. Synthesis 2005, 2119. (d) Turhanen, P. A.; Vepsalainen, J. J.
Synthesis 2004, 992. (e) Even, P.; Guenin, E.; Benramdame, M.;
Quidu, P.; El Manouni, D.; Lecouvey, M. Lett. Org. Chem. 2004, 1,
75. (f) Ye, Y.; Xu, G.; Zheng, Y.; Liu, L. Heteroat. Chem. 2003, 14,
309. (g) Orstad, E.; Hoff, P.; Skattebol, L.; Skretting, A.; Breistol,
K. J. Med. Chem. 2003, 46, 3021. (h) Mizrahi, D. M.; Waner, T.;
Segall, Y. Phosphorus, Sulfur Silicon Relat. Elem. 2001, 173, 1. (i)
Turhanen, P. A.; Ahlgren, M. J.; Jarvinen, T.; Vepsalainen, J. J.
Phosphorus, Sulfur Silicon Relat. Elem. 2001, 170, 115. (j) Turhanen,
P. A.; Ahlgren, M. J.; Jarvinen, T.; Vepsalainen, J. J. Synthesis 2001,
633. (k) Mallard, I.; Benech, J.; Lecouvey, M.; Leroux, Y. Phos-

Letters
phorus, Sulfur Silicon Relat. Elem.s 2000, 162, 15. (l) Niemi, R.;
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 11 3063
(17) While the use of DMAP afforded a ratio of the desired O-TBS-
alendronate to the rearranged 1-phosphonate-1-phosphate of 9:1, the
use of other bases provided the following ratios: imidazole, 5:1;
triethylamine, 6:1; Hunigs base, 8:1; DABCO, 6:1; DBU, 0.5:1. In
addition, the use of imidazole did not affect completion of the
silylation step and the use of the Huenigs base led to undesired
formation of silylenolethers from the R-ketophosphonate, effectively
reducing the overall yield. The use of less polar solvents such as
cyclohexane or toluene with a potential of increasing the content of
desired Pudovik adduct over undesired rearranged byproduct prior
to silylation (see Table 1) led to heterogeneous reaction mixtures
and irreproducible overall yields.
Turhanen, P. A.; Vepsalainen, J. J.; Taipale, H.; Jarvinen, T. Eur. J.
Pharm. Sci. 2000, 11, 173. Reviews of literature prior to 2000: (m)
Lecouvey, M.; Leroux, Y. Heteroat. Chem. 2000, 11, 556. (n)
McKenna, C. E.; Kashemirov, B. A. Top. Curr. Chem. 2002, 220,
201.
(5) There are several synthetic strategies for the preparation of alendronic
acid. Of these, the most notable is the work of Kieczykowski et al.,
which serves as a basis for the large-scale production of alendronate
sodium: Kieczykowski, G. R.; Jobson, R. B.; Melillo, D. G.;
Reinhold, D. F.; Grenda, V. J.; Shinkai, I. J. Org. Chem. 1995, 60,
8310.
(6) Fitch, S. J.; Moedritzer, K. J. Am. Chem. Soc. 1962, 84, 1876.
(7) Examples of our efforts to directly convert alendronic acid and
N-(carbobenzyloxy)alendronic acid (1) into their respective tetraalkyl
alendronates include alkylation under basic conditions with methyl
iodide and under acidic conditions with methyl sulfate, esterification
using coupling reagents, protic and Lewis acids, and methanol or
methanol equivalents, diazomethane and trimethylsilyldiazomethane,
and various orthoformates at elevated temperatures. In most cases,
traces of the desired product were observed by 31P NMR analysis of
the reaction mixture along with the rearranged 1-phosphonate-1-
phosphate byproduct. The desired product was never isolated in
substantial yield.
(8) Ruel, R.; Bouvier, J. P.; Young, R. J. Org. Chem. 1995, 60, 5209.
(9) Review of the Michaelis-Arbuzov reaction: Bhattacharya, A. K.;
Thyagarajan, G. Chem. ReV. 1981, 81, 415.
(10) Review of Pudovik reaction: Pudovik, A. N.; Konovalova, I. V.
Synthesis 1979, 81.
(18) O-TBS-1-hydroxy-1,1-bisphosphonates Cl(CH2)3C(OTBS)(PO(OMe)2)2,
BnO2C(CH2)3C(OTBS)(PO(OEt)2)2, and BnO2C(CH2)3C(OTBS)(PO-
(OMe)2)2, were prepared in 57%, 89%, and 71% overall isolated yield,
respectively, using the same general strategy.
(19) Experimental procedure for the preparation of 10: A 1 L oven-dried
round-bottomed flask under an inert atmosphere of nitrogen was
charged with 4-chlorobutyryl chloride (14 g, 0.10 mol). Triethyl
phosphite (16.6 g, 0.10 mol) was added dropwise at 0 °C via syringe
over 5 min with stirring. The resulting mixture was allowed to reach
room temperature and stirred an additional 15 min, after which
anhydrous methylene chloride (300 mL) was added via canella.
Diethyl phosphite (15.2 g, 0.11 mol) and 4-(dimethylamino)pyridine
(12.2 g, 0.10 mol) were added sequentially. The mixture was stirred
at room temperature for 1 h, and tert-butyldimethylsilyl chloride (16.5
g, 0.11 mol) was added. The reaction stirred at room temperature
for an additional 15 h. The mixture was washed with 0.1 M aqueous
hydrochloric acid (2 × 300 mL), dried over sodium sulfate, and
concentrated. Analytically pure 10 (22.3 g, 45%) was obtained using
Biotage 75L liquid chromatography on silica gel (eluent: 50% ethyl
(11) For examples, see the introduction of ref 6.
(12) We were unable to reproduce several published procedures that relied
on the use of catalytic amounts of secondary and tertiary amines as
sufficient to promote the reaction. When 0.1 and 0.2 equiv of various
bases were used, an average conversion of only 9.3 ( 1.2% and
21.1 ( 2.3% was observed, respectively.
(13) The role of solvent polarity on the rearrangement of monophospho-
nates to monophosphates follows a trend similar to that reported here
for bisphosphonates. Compare this to a recent disclosure for
monophosphonates: El Kaim, L.; Gaultier, L.; Grimaud, L.; Dos
Santos, A. Synlett 2005, 2335.
1
acetate in n-heptane). H NMR (400 MHz, CDCl3) δ 0.13 (s, 6H),
0.83 (s, 9H), 1.30 (t, J ) 7.0 Hz, 12H), 2.07 (m, 2H), 2.16 (m, 2H),
3.49 (t, J ) 6.4 Hz, 2H), 4.07 (m, 8H); 13C NMR (400 MHz, CDCl3)
δ 16.4, 19.0, 25.1, 27.1, 33.4, 45.4, 53.7, 64.2 (m); 31P NMR (400
MHz, CDCl3) δ 19.21; HRMS calcd m/z 494.1785, obsd m/z
494.1789.
(20) (a) Tatsuoka, T.; Imao, K.; Suzuki, K. Heterocycles 1986, 24, 617.
(b) Blackburn, G. M.; Kent, D. E. J. Chem. Soc., Perkin Trans. 1
1986, 913. (c) Streicher, W.; Werner, G.; Rosenwirth, B Chem. Scr.
1986, 26, 179.
(21) Intravenous dosing was chosen over oral administration to circumvent
potential errors in the overall biomarker readout due to any phase I
metabolism of N-acylalendronates or substantial variations in their
oral bioavailability compared to the parent drug.
(14) March, J. AdVanced Organic Chemistry, 4th ed.; John Wiley &
Sons: New York, 1992.
(15) Half-life (t1/2,4f5) determined by 31P NMR for a sample stored under
an inert atmosphere of argon at 20 ( 3 °C.
(16) The intramolecular mechanism of the 7-exo-trig O f N acyl transfer
was established by a crossover experiment in which a mixture of 3
equiv of 6b and 1 equiv of 7a was hydrogenated. No cross-acylation
to produce N-acetyl 9b was observed, and the sole product of this
reaction was N-myristoyl 9a.
(22) (a) Lin, J. H.; Russell, G.; Gertz, B. Int. J. Clin. Pract. 1999, 101,
18. (b) Lin, J. H. Bone 1996, 18, 75.
JM060398V