Pilot-plant preparation of an αvβ3 integrin antagonist. Part 2. Synthesis of N-[2-(5-hydroxy-4,6- tetrahydropyrimidine)]-3-amino-5-hydroxybenzoic acid

Article abstract of DOI:10.1021/op049968w

Studies directed toward the process research and development of a scalable method for preparing tetrahydropyrimidine 2, a key intermediate to the α_vβ₃ integrin antagonist 1, are described. A linear approach employing 3-amino-5-hydrox

Full text of DOI:10.1021/op049968w

Organic Process Research & Development 2004, 8, 571−575
Pilot-Plant Preparation of an rvâ3 Integrin Antagonist. Part 2. Synthesis of
N-[2-(5-Hydroxy-4,6-tetrahydropyrimidine)]-3-amino-5-hydroxybenzoic Acid
Jerry D. Clark,*,† Joe T. Collins,‡ H. Peter Kleine,‡ Gerald A. Weisenburger,§ and D. Keith Anderson
Global Research & DeVelopment, Pfizer Inc., 2800 Plymouth Road 28/3010E, Ann Arbor, Michigan 48105, U.S.A.,
Global Research & DeVelopment, Pfizer Inc., 700 Chesterfield Parkway, Chesterfield, Missouri 63017, U.S.A.,
Global Research & DeVelopment, Pfizer Inc., Eastern Point Road 8156-003, Groton, Connecticut 06340, U.S.A., and
Global Research & DeVelopment, Pfizer Inc., 800 North Lindbergh BouleVard, St. Louis, Missouri 63167, U.S.A.
Scheme 1. Key intermediates toward preparation of 1
Abstract:
Studies directed toward the process research and development
of a scalable method for preparing tetrahydropyrimidine 2, a
key intermediate to the rvâ3 integrin antagonist 1, are described.
A linear approach employing 3-amino-5-hydroxybenzoic acid,
methyl isothiocyanate, and 1,3-diaminopropan-2-ol as key
reagents is detailed. The results of process development re-
search, a successful pilot run, and a production campaign are
explained.
Introduction
The Rvâ3 integrin plays an essential role in angiogenesis,
the process by which new blood vessels form from preexist-
ing blood vessels.1 Angiogenesis is required for tumor
growth, and therefore, antagonists of Rvâ3 are being studied
for the treatment of cancer. In a program directed toward
the discovery of such antagonists, an improved synthesis of
1 (Scheme 1) was required.2 Accordingly, a process research
and development team was formed to evaluate strategies by
which 1 could be prepared on-scale. Envisioned was a
convergent synthesis of 1 in which tetrahydropyrimidine 2
would be linked via glycine to the â-amino acid ester 3. With
regard to the preparation of 2, a research program was
completed wherein 3-amino-5-hydroxybenzoic acid, methyl
isothiocyanate, and 1,3-diaminopropan-2-ol were employed
Scheme 2. Enabling route to 2
* To whom correspondence should be addressed: E-mail: jerry.d.clark@
pfizer.com.
† Global Research & Development, Pfizer Inc., Ann Arbor, Michigan.
‡ Global Research & Development, Pfizer Inc., Chesterfield, Missouri.
§ Global Research & Development, Pfizer Inc., Groton, Connecticut.
Global Research & Development, Pfizer Inc., St. Louis, Missouri.
(1) (a) Eliceiri, B. P.; Cheresh, D. A. Cancer 2000, 6S3, 245-249. (b) Roberto,
R.; Angelo, V.; Domenico, R.; Francesco, D. R.; Francesca, M.; Franco,
D. Haematologica 2002, 87, 836-845. (c) Ruminski, P. G.; Rogers, T. E.;
Rico, J. E.; Nickols, G. A.; Westlin, W. F.; Engleman, V. W.; Settle, S. L.;
Keene, J. L.; Freeman, S. K.; Carron, C. P.; Meyer, D. M. Abstracts of
Papers, 218th National Meeting of the American Chemical Society, New
Orleans, LA; American Chemical Society: Washington, DC, 1999. (d)
Ruminski, P. G.; Rogers, T. E.; Rico, J. E.; Nickols, G. A.; Engleman, V.
W.; Settle, S. L.; Keene, J. L.; Freeman, S. K.; Carron, C. P.; Meyer, D. M.
Abstracts of Papers, 217th National Meeting of the American Chemical
Society, Anaheim, CA; American Chemical Society: Washington, DC, 1999.
(e) Flynn, D. L.; Abood, N.; Bovy, P.; Garland, R.; Hockerman, S.;
Lindmark, R. J.; Schretzman, L.; Rico, J.; Rogers, T. Abstracts of Papers,
210th National Meeting of the American Chemical Society, Chicago, IL;
American Chemical Society: Washington, DC, 1995.
as key reagents. As part two of a three-part report3 describing
the successful pilot-plant preparation of 1, herein we detail
the chemical process research and development of a scalable
method for preparing 2.
Enabling Technology
The technology developed to supply interim amounts of
2 is shown in Scheme 2.2 This convergent approach entailed
coupling the methyl thiourea 7 with 3-amino-5-hydroxyben-
(3) Clark, J. D.; Weisenburger, G. A.; Anderson, D. A.; Colson, P.-.J.; Edney,
A. D.; Gallagher, D. J.; Kleine, H. P.; Knable, C. M.; Lantz, M. K.; Moore,
C. M. V.; Murphy, J. B.; Rogers, T. E.; Ruminski, P. G.; Shah, A. S.; Storer,
N.; Wise, B. E. Org. Process Res. DeV. 2004, 8, 51-61.
(2) (a) Collins, J. T.; Devadas, B.; Lu, H.-F.; Malecha, J. W.; Miyashiro, J.
M.; Nagarajan, S.; Rico, J. G.; Rogers, T. E. U.S. Patent 6,100,423, 2000.
(b) Bloom, B. M. U.S. Patent 2,899,426, 1959.
10.1021/op049968w CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/12/2004
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
571

Scheme 3. Linear approach to 2
zoic acid (9) affording 2 in 18% overall yield. Although this
methodology was employed on-scale, the following safety,
technical, and operational issues were identified during the
course of practicing the chemistry. The process (1) required
too many steps to product, (2) gave rise to low yields and
high waste loads, (3) entailed relatively low payload and long
cycle times, (4) required the use of carbon disulfide, (5)
required the use of a protecting group, (6) involved vacuum
distillations, and (7) revealed issues associated with the
handling and transfer of some intermediates.
zole-derivative 13 at approximately 0.5%, and the dimethyl-
guanidine-analogue 14 at levels approaching 3%.
Process Research
The aforementioned enabling technology used to prepare
2 suffered from a number of drawbacks. Thus, a linear
approach to 2 was investigated. Envisioned was the reaction
of 9, prepared from 8,4 with an isothiocyanate derivative
(Scheme 3). On the basis of in-house experience, we
expected that an isothiocyanate should react selectively with
the amino substituent. Alkylation of the resulting thiourea
derivative followed by reaction with 1,3-diaminopropan-2-
ol would then give rise to 2. Accordingly, a review of the
price and availability of isothiocyanates indicated that the
methyl analogue should be the most cost-effective. Indeed,
when 9 was subjected to a slight excess of methyl isothio-
cyanate in DMF at ambient temperature, the corresponding
thiourea 10 (Scheme 3) was prepared in near quantitative
yield as determined by HPLC analysis.
The crude solution of 10 was then subjected to methyl
iodide, giving rise to 11. Treatment of this mixture with
excess 1,3-diaminopropan-2-ol was followed by careful
heating with venting to a scrubber. Gas evolution was
observed at approximately 70 °C. Analysis of the product
mixture revealed, however, that the reaction was incomplete.
Total conversion was realized upon heating to 90 °C.
Isolation of the desired product 2 as either its HCl salt or as
the zwitterion proved troublesome at first, but conditions
were developed from which zwitterion 2 was prepared in
55-60% overall yield from 8. The product assay was 85 wt
%, with a moisture content of approximately 10 wt %.
Analysis of the impurity profile of 2 by HPLC/MS and
HPLC/NMR suggested the presence of the bis-tetrahydro-
pyrimidine 12 at a level of approximately 0.5%, the imida-
Additionally it was found that product mass balance was in
the mother liquor. Taking this into account, an outstanding
overall chemical yield of 94% of 2 from 9 was realized. Thus,
an efficient, three-step, one-pot procedure for the preparation
of 2 had been developed.
Process Development
The success of the laboratory trial to make 2 via this
methyl isothiocyanate route prompted us to further develop
this chemistry. Two noteworthy results from the aforemen-
tioned laboratory trial were (1) the impurity profile and (2)
the high level of water in the isolated solid. To address these
issues and prepare the data for transfer to the pilot plant,
three significant studies were performed: (1) development
of rationale approaches to minimize the impurities, (2) a
study directed toward the drying habits of 2, and (3) a Mettler
RC1 reaction calorimetry study of the chemical route to more
effectively understand and safely transfer the technology.
It was quickly realized that 12 and 13 were derived from
impurities present in the starting materials. In the case of
12, it was verified that 3,5-diaminobenzoic acid was present
at a level of approximately 1% in 9. Reaction of the diamine
with the reagents in excess formed 12. This was verified
through independent preparation of 12 and spiking studies
thereof. It was found, however, that 12 was eliminated in
the downstream process and did not manifest itself as the
corresponding impurity in API. Therefore, investigating ways
to diminish the level of 3,5-diaminobenzoic acid in 9 was
not necessary.
(4) For the literature preparation of 9 see: (a) Becker, A. M.; Rickards, R. W.;
Brown, R. F. C. Tetrahedron 1983, 39, 4189-4192. (b) Drake, N. L. Org.
React. 1942, 5, 105-128.
572
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

Scheme 4. Impact of 12 on API impurity profile
Physical facilitation of methylamine expulsion was ac-
complished by either the installation of a nitrogen sparger
or an increase of the reaction temperature during reagent
interaction. Independent preparation of 14 and spiking studies
thereof verified these results. Ultimately, the approach
employed in the pilot plant was to increase the pot temper-
ature containing the solution of 11, maintain a vigorous
nitrogen sweep through the reactor headspace, and add the
1,3-diaminopropan-2-ol as quickly as possible.
The investigations undertaken to discern the drying habits
of 2 led to the realization that this zwitterion was hygro-
scopic. For example, original samples contained g13%
moisture. However, when samples of 2 were dried at 100
°C under a vacuum of 27.5 in Hg, a weight decrease of 4.4%
was measured within 1 h. A loss of 11.3% was observed
after 21 h, and eventually, a total of 12% of the sample’s
original weight was lost after 8 days. On the other hand, if
these dried samples were allowed to stand under ambient
conditions, the moisture level returned to a level of 13%
within 36 h. The water in this key intermediate would also
be a parameter of interest in the chemistry used to prepare
1.7
The Mettler RC1 reaction calorimeter study of the
conversion of 9 into 10 revealed a moderate heat of reaction
of -57.3 kJ/mol of 9 (-63.7 kJ/kg of solution) and an
adiabatic temperature rise (ATR) of 33 °C for this process.
The data also indicated that the reaction was essentially
complete within 6.25 h after dosing. Similar trends were
observed in the reaction of 10 with methyl iodide. A heat of
reaction of -52 kJ/kg of solution was measured, resulting
in an ATR of 27 °C for the process. The reaction was
complete within 2 h, including the 60-min addition of methyl
iodide. For the conversion of 11 to the desired product 2,
only a small endothermic response was measured when the
room temperature 1,3-diaminopropan-2-ol was added to the
hot solution of 11. Thus, the reaction calorimetry experiments
did not reveal any significant thermal risk associated with
the chemistry.
It was proposed that 13 might be derived from 2,3-
diaminopropan-1-ol present in the 1,3-diaminopropan-2-ol.
The presence of this homologue at levels as high as 7 area
% was verified by careful GC analysis of the starting
material, but directed studies toward the preparation of 13
were not completed. Like 12, 13 did not significantly impact
the impurity profile of 1.
The dimethylguanidine impurity 14 presented a most
interesting challenge as it was generated by the chemistry
itself. Also, unlike the other impurities, 14 carried down-
stream and negatively impacted the API purity. For example,
when 2-3% of 14 was present in 2, 1 contained >0.5% of
the corresponding impurity 15 (Scheme 4). However, when
the level of 14 was <0.5% in 2, then <0.2% of 15 was
observed. Thus, understanding the origin of 14 and further
development of the process to minimize it was an important
goal.
A proposed mechanism for the preparation of 2 is shown
in Scheme 5. This suggests that 14 is derived from the
interaction of unreacted starting material, or a reactive
intermediate thereof, with the methylamine byproduct of the
reaction.5 Indeed, it was found that the level of 14 could be
diminished from 3% to less than 0.5% by either adjusting
the mode of reagent interaction or by physically facilitating
gaseous methylamine expulsion. For example, the mode of
reagent addition was changed by either adding the crude
solution of 11 to a hot solution of 1,3-diaminopropan-2-ol
in DMF (inverse addition) or by increasing the rate of 1,3-
diaminopropan-2-ol addition to the crude solution of 11.
Scheme 5. Proposed mechanism for the conversion of 11 to 2
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
573

With these data, the chemistry was scaled to 22-L
equipment. No noteworthy issues were observed. Employing
2 kg of 9, 2.27 kg of 2 was prepared in 72% overall yield
with an assay of 85 wt % and 10 wt % water. Therefore, a
new linear, three-step, methyl isothiocyanate-based synthesis
of 2 from 9 was demonstrated.
water and the pH adjusted to 6 with concentrated HCl. The
zwitterion 2 was collected by centrifugation, the cake was
washed with water and acetonitrile and was then transferred
to a tray dryer and dried at 60 ( 5 °C until the LOD was
measured at less than 1%. This gave rise to 13.0 kg of 2 as
a light-tan powder in 71.5% overall yield from 9 with an
assay of 97.4%.
Pilot-Plant Campaigns
Production Campaign. The major equipment used in the
production of 2 included two 100-gal Pfaudler reactors, four
500-gal Pfaudler reactors, a 40-in. stainless steel-clad Tol-
hurst centrifuge, a vacuum tray dryer, and a 50-gal portable,
glass-lined vacuum tank. Again the chemistry performed as
expected. The one significant operational change imple-
mented was increasing the tray dryer temperature to 70 ( 5
°C. Two batches of 2 were prepared, affording 35.4 kg
(73.1% yield) and 37.0 kg (76.2% yield) of product with
assays of 96.3 and 96.4 wt %, respectively. Thus, the average
yield for all lots was 74.1%, providing a total of 85.6 kg of
2 with an average assay of 96.6 wt %.
The hazardous reagents and potential materials of con-
struction issues associated with the new chemistry led us to
choose a specific third party manufacturer (TPM). This
facility had the high temperature and pressure reactors
necessary to prepare 9, had previous experience in handling
methyl isothiocyanate, and manufactured methyl iodide.
Thus, a strategy was implemented in which a pilot run would
be completed first, and if successful, a production campaign
composed of two batches would be completed.
Preparation of 3-Amino-5-hydroxybenzoic Acid (9).
Laboratory studies directed toward the preparation of 9
followed literature precedent.4 However, upon scale-up,
mixing the ammonium hydroxide, ammonium chloride, and
3,5-dihydroxybenzoic acid (8) as described gave rise to a
solid, cement-like mass that could not be agitated. We were
able to repeat this phenomenon in the laboratory and found
that the solution to the problem was the order of mixing. It
was discovered that addition of a premixed solution of
ammonium hydroxide and ammonium chloride to 8 avoided
the formation of the solid mass. Once this was ascertained,
51.1 kg of 8 was converted to 25.3 kg of 9 in 51% overall
yield with an assay of >96%. This procedure was success-
fully repeated. The major equipment employed was an
Inconel 600 clad reactor with a working volume of 300 gal
and a maximum pressure tolerance of 1000 psi. The balance
of the work required the use of one 100-gal glass-lined
reactor, two 200-gal glass-lined reactors, a ceramic-coated
filter, and a stainless steel filter.
The Pilot Run. The pilot run was uneventful. The major
equipment utilized were two 30-gal Pfaudler reactors, three
200-gal Pfaudler reactors, a 20-in. stainless steel-clad Tol-
hurst centrifuge, a vacuum tray dryer, and a 50-gal portable,
glass-lined vacuum tank.
A slurry of 9 (11.1 kg, 72.7 mol) in DMF was treated
with a warm solution of methyl isothiocyanate (5.4 kg, 74.1
mol) in DMF. The reaction was allowed to go to completion
by overnight agitation at ambient temperature. In the next
step, crude 10 was treated directly with methyl iodide (14.4
kg, 102 mol) and allowed to agitate overnight at ambient
temperature. Most of the DMF was removed under vacuum,
providing a mobile concentrate.
Conclusions
Studies directed toward the process research and develop-
ment of a scalable method for preparing tetrahydropyrimidine
2, a key intermediate to the Rvâ3 integrin antagonist 1, were
described. A linear approach employing 3-amino-5-hydroxy-
benzoic acid, methyl isothiocyanate, and 1,3-diaminopropan-
2-ol as key reagents was detailed. Development of this
chemistry led to a better understanding of the origin of
impurities and how to diminish them. A successful pilot run
and production campaign were completed in which a total
of 85.6 kg of 2 was prepared with an average assay of 96.6
wt % in an overall yield of 74%.
Experimental Section
General. The actual charges of substrates and reagents
are given below. The molar amounts are calculated on the
basis of the assays of the materials. Similarly, yields are
calculated on the basis of assay-corrected moles of substrates
and products. Proton (1H) nuclear magnetic resonance (NMR)
spectra were recorded on either a Unity Inova Varian 300
MHz or Unity Inova Varian 400 MHz spectrometer. 1H NMR
descriptions are reported as: s (singlet), d (doublet), t (triplet),
q (quartet), m (multiplet), or br (broad).
Melting points were determined using a Laboratory
Devices Mel-Temp instrument equipped with a Fluke 51
thermocouple. Thin-layer chromatography was performed on
EM Science 0.25 nm silica gel 60, glass-backed plates with
F254 indicator. UV light was employed for visualization.
3-Amino-5-hydroxybenzoic Acid (9).4 To a high-pres-
sure Inconel clad reactor was added 8 (51.1 kg, 332 mol).
In a separate reactor, ammonium chloride (68.0 kg) was
dissolved into ammonium hydroxide (215.5 kg). A clear
solution was obtained. The ammonium hydroxide/ammonium
chloride solution was then drawn into the reactor containing
8. The loading sequence is important because altering the
procedure could create a mass that cannot be stirred. The
contents of the reactor were agitated, heated to 180-190
°C, and held within this temperature range for 40-42 h.
In the final step, a warm solution of 1,3-diaminopropan-
2-ol (19.7 kg, 218 mol) and DMF was added to the
concentrate of 11. This mixture was gradually heated,
resulting in off-gassing of methyl mercaptan and methy-
lamine. The gases were scrubbed with solutions of caustic
and sulfuric acid. The reaction solution was then diluted with
(5) No impurity derived from the expulsion of methyl mercaptan was observed.
(6) Intermediate 2 was a proposed registered starting material.
(7) The effect of the water associated with 2 on the downstream chemistry to
prepare 1 will be a subject of discussion in part three of this series.
574
•
Vol. 8, No. 4, 2004 / Organic Process Research & Development

Pressures over the course of heating reached 365 psig at the
onset and attained a maximum pressure of 431 psig.
After cooling the contents, the reaction mass was trans-
ferred to a glass-lined 200-gal reactor. The liquid appeared
to be translucent, homogeneous, and approximately dark-
maroon in color. The excess ammonia was removed from
the system to a sulfuric acid scrubber and the material
concentrated to a volume of approximately 60 gal. A dark-
maroon to coffee-black colored, homogeneous, and fairly
opaque product mixture was produced. The product mass
was acidified with 37% HCl (127 kg) and refluxed at 100-
112 °C for 16-17 h. After reflux, the material was
precipitated by cooling to 20 °C and then collected on a
ceramic filter to give crude 9 (127 kg) as a damp, dark-
brown to maroon solid.
The crude was then purified by dissolving the material
in 2% aqueous sodium hydroxide (359 kg), adding Darco
KB activated carbon (6.8 kg), filtering the resulting slurry,
and adjusting the pH to 3.5 ( 0.1 with 37% HCl (21.3 kg).
The precipitate was collected by centrifugation, affording
33.8 kg of damp 9 with an assay of 91.7% by HPLC analysis.
Further purification was accomplished by redissolving
crude 9 into 3.7% caustic (238 kg). This was filtered (without
charcoal), the pH was adjusted to 3.5 ( 0.1 with 37% HCl
(21.3 kg), and the product was isolated via filtration. The
product was dried in a tray dryer with a vacuum of 25 in.
Hg at 60-65 °C to provided 25.3 kg of 9 (50%) as a light-
gray powder with an assay of 96.5% by HPLC analysis: mp
230.5-234 °C. This material was identical to an authentic
sample.2
3-Hydroxy-5-[[(methylamino)thioxomethyl]amino]-
benzoic Acid (10). Into a 100-gal reactor was charged DMF
(57.6 kg) and methyl isothiocyanate (14.7 kg, 202 mol). To
a separate 100-gal reactor was added 9 (28.8 kg, 199 mol)
and DMF (57.6 kg). With agitation, the methyl isothio-
cyanate/DMF mixture was charged into the reactor containing
9 and allowed to agitate for 12 h at room temperature. Thin-
layer chromatographic analysis indicated that no starting
material was present. This solution was used without further
purification.
Thiomethyl Adduct 11. To crude 10 was slowly added
methyl iodide (38.2 kg, 269 mol). The internal temperature
of the mixture was maintained below 40 °C during addition,
applying cooling to the jacket when necessary. The reaction
was shown complete by thin-layer chromatographic analysis.
The methyl iodide adduct was concentrated under reduced
pressure (27-29 in. Hg) to remove excess DMF and excess
methyl iodide. The internal temperature during the distillation
was not allowed to exceed 60 °C. Crude 11 was used without
further purification.
N-[2-(5-Hydroxy-4,6-tetrahydropyrimidine)]-3-amino-
5-hydroxybenzoic Acid (2). To a separate 100-gal reactor
vented to sequential bleach and sulfuric acid scrubbers was
charged premelted (mp 40-45 °C) 1,3-diaminopropan-2-ol
(50.8 kg, 597 mol) and DMF (41.7 kg). The 1,3-diamino-
propan-2-ol/DMF solution was transferred to the reactor
containing 11 at such a rate to keep the contents below 50
°C. Upon complete addition, the contents were heated to 90
°C. Vigorous off-gassing began at approximately 60 °C.
Once the reactor contents reached 90-95 °C, it was held
within this range for 3 h.
The reactor jacket was cooled to 40 °C. Into a separate
500-gal reactor was charged 181 kg of water. With agitation,
the solution containing 2 was transferred into the 500-gal
reactor. The contents were then cooled to 10 °C, and the pH
was carefully adjusted to pH 6 with 37% aqueous HCl (27.7
kg). The resulting precipitate was isolated via a variable-
speed 40-in. Tolhurst centrifuge. The damp cake was washed
with water (18 kg) and acetonitrile (14 kg) and allowed to
spin dry. The product was transferred to a vacuum tray dyer
and dried overnight at 70-75 °C at 25-28 in. Hg to give
37 kg (74.9%) of 2 with an HPLC assay of 96.4%: TLC:
Rf ) 0.53 (CHCl3/MeOH/H2O, 25:25:5); 1H NMR (D2O, 90
MHz) δ 7.42 (s, 2H), 7.01 (t, 1H), 4.41 (t, 1H), 3.46 (br s,
4H). This product was identical to an authentic sample as
1
determined by H NMR, 13C NMR, and HPLC analysis.2
Acknowledgment
We are indebted to the fine operators and support
personnel of R.S.A. Corporation, Danbury, CT, and we
specifically thank Jim Johnston and Jan Anthony. Gratefully
acknowledged are the helpful collaborations with discovery
chemists Peter Ruminski and Tom Rogers. Jan Gard, Jim
Murphy, Ed Sheldon, Don Stano, and Wei Kong are
recognized for their invaluable analytical work. Russ Buch-
man is acknowledged for his effort in sourcing this project.
Finally, due to the depth of this project, we express our
gratitude to the many individuals that cannot be addressed
personally.
OP049968W
Vol. 8, No. 4, 2004 / Organic Process Research & Development
•
575