Identification of a manufacturing route of novel CRF-1 antagonists containing a 2,3-Dihydro-1 H-pyrrolo[2,3-b]pyridine moiety

Article abstract of DOI:10.1021/op100147h

A case study on the synthesis of novel CRF-1 antagonists containing the 2,3-dihydro-1H-pyrrolo[2,3-b]pyridine moiety is presented. The development of ever more efficient synthetic routes allowed the progression of three candidates at the same time. A manu

Full text of DOI:10.1021/op100147h

Organic Process Research & Development 2010, 14, 895–901
Identification of a Manufacturing Route of Novel CRF-1 Antagonists Containing a
2,3-Dihydro-1H-pyrrolo[2,3-b]pyridine Moiety
Arianna Ribecai,*,† Sergio Bacchi,† Monica Delpogetto,† Simone Guelfi,† Angelo Maria Manzo,† Alcide Perboni,†
Paolo Stabile,† Pieter Westerduin,† Marie Hourdin,† Sara Rossi,† Stefano Provera,‡ and Lucilla Turco‡
Chemical DeVelopment Department and Analytical Chemistry Department, GlaxoSmithKline Medicines Research Centre,
Via Fleming 4, 37135 Verona, Italy
Abstract:
A case study on the synthesis of novel CRF-1 antagonists contain-
ing the 2,3-dihydro-1H-pyrrolo[2,3-b]pyridine moiety is presented.
The development of ever more efficient synthetic routes allowed
the progression of three candidates at the same time. A manu-
facturing route was identified and successfully demonstrated on
a pilot-plant scale to prepare 100 kg of the CRF-1 antagonist
GW876008.
Figure 1. Exploration of the 2,3-dihydro-1H-pyrrolo[2,3-b]-
pyridine template.
and identified compounds 1a-c (Figure 1) as the most
Introduction
Corticotropin releasing factor (CRF) is a 41-amino acid
interesting derivatives belonging to this new series.3
peptide that plays a central role in mediating the behavioural,
neuroendocrine, and autonomic responses to stress.1 CRF is
synthesized in the hypothalamus and liberated into the hypo-
physeal-portal system, triggering the release of adrenocorti-
cotrophic hormone (ACTH) from the pituitary gland. ACTH,
in turn, induces the synthesis and the release of glucocorticoids
from the adrenal cortex. In mammals, exposure to stress
increases the release of CRF in both hypothalamic and extra-
hypothalamic brain areas and produces the same effects
observed with exogenous administration of CRF, i.e., increase
of ACTH and cortisol (human) or corticosterone (rodent) blood
levels, enhanced autonomic nervous system activity and anxiety-
like behaviour. CRF is responsible, at least in part, for the
hypothalamic pituitary adrenal (HPA) axis hyperactivity char-
acteristic of patients with major depression.2 Several lines of
clinical evidence suggest the association between the hyper-
drive of CRF and the onset of anxiety and depressive disorders.
As a consequence, several pharmaceutical research groups have
made significant investments over the past two decades in
discovering nonpeptide CRF-1 receptor antagonists as a po-
tential treatment for stress-related illnesses, including anxiety
and depression.
Results and Discussion
Initial Supply Routes to CRF-1 Candidates 1a-c. The
Medicinal Chemistry retrosynthetic approach used for the
synthesis of the initial toxicological batches is depicted in
Scheme 1.3b
Notably, being a convergent synthesis in the last chemical
step, the same chemistry could be applied to prepare all the
chosen compounds of general formula 1, thus allowing the
progression of three candidates, 1a-c, at the same time.
The synthetic route applied by Medicinal Chemistry is described
in Scheme 2.
As described in Scheme 2, commercially available anilines
2 were transformed into the corresponding amides 3 by reaction
with 4-chlorobutyryl chloride in the presence of K2HPO4. These
compounds were treated with sodium hydride in DMF at 0 °C
to afford the γ-lactam derivatives 4 after chromatographic
purification. The reaction of 4 with POCl3 and ethyl 3-amino-
crotonate in dry dichloroethane enabled the preparation of crude
compounds 5, which were then dissolved in DMF and heated
to 100 °C in the presence of sodium hydride to induce the
cyclization. The bicyclic derivatives 6 were obtained after
filtration of the crude reaction residue on silica gel, although in
limited chemical yield of 10-15%. The resulting intermediates
As a part of a broad strategy aimed at the discovery of novel
CRF-1 antagonists, scientists at GlaxoSmithKline designed and
explored the 2,3-dihydro-1H-pyrrolo[2,3-b]pyridine template
(3) (a) Di Fabio, R.; Arban, R.; Bernasconi, G.; Braggio, S.; Blaney, F. E.;
Capelli, A. M.; Castiglioni, E.; Donati, D.; Fazzolari, E.; Ratti, E.;
Feriani, A.; Contini, S.; Gentile, G.; Ghirlanda, D.; Sabbatini, F. M.;
Andreotti, D.; Spada, S.; Marchioro, C.; Worby, A.; St-Denis, Y.
J. Med. Chem. 2008, 51, 7273–7286. (b) Di Fabio, R.; St.-Denis, Y.;
Sabbatini, F. M.; Andreotti, D.; Arban, R.; Bernasconi, G.; Braggio,
S.; Blaney, F. E.; Capelli, A. M.; Castiglioni, E.; Di Modugno, E.;
Donati, D.; Fazzolari, E.; Ratti, E.; Feriani, A.; Contini, S.; Gentile,
G.; Ghirlanda, D.; Provera, S.; Marchioro, C.; Roberts, K. L.; Mingardi,
A.; Mattioli, M.; Nalin, A.; Pavone, F.; Spada, S.; Trist, D. G.; Worby,
A. J. Med. Chem. 2008, 51, 7370–7379.
* To whom correspondence should be addressed. Telephone: 00-390458219360.
Fax: 00-390458218117. E-mail: arianna.ribecai@gsk.com.
† Chemical Development Department.
‡ Analytical Chemistry Department.
(1) Vale, W.; Spiess, J.; Rivier, C.; Rivier, J. Science 1981, 213, 1394–
1397.
(2) Owen, M. J.; Nemeroff, C. B. Pharmacol. Res. 1991, 43, 425–473.
10.1021/op100147h  2010 American Chemical Society
Published on Web 06/28/2010
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
895

Scheme 1. Medicinal Chemistry retrosynthetic analysis
Scheme 2. Medicinal Chemistry route
Scheme 3. Outsourced syntheses of 1H-pyrazolyl derivatives 9
6 were transformed with good yields into the triflate derivatives
7, which after treatment with potassium iodide in NMP at 150
°C for 18 h afforded the corresponding iodide intermediates 8.
The latter were finally coupled with the substituted pyrazole
derivatives 9 under Buchwald conditions4 to give title com-
pounds 1a-c with very low overall yields of 1-2%. The
required compounds 9, 1-(1H-pyrazol-3-yl)imidazolidin-2-one,
9a, and 2-(1H-pyrazol-3-yl)thiazole, 9b, were purchased from
external suppliers and prepared according to the processes
depicted in Scheme 3.5,6
The synthetic route described previously in Scheme 2 was
clearly unsuitable to produce multikilogram batches. In fact, it
involved very low-yielding steps, made use of a low-temperature
triflation process, and employed a large excess of toxic POCl3.
Moreover, the large number of impurities generated along the
route were difficult to remove even by chromatography (e.g.,
ethyl 3-aminocrotonate self-condensation product), thus making
(4) (a) Klapars, A.; Antilla, J. C.; Huang, X.; Buchwald, S. L. J. Am.
Chem. Soc. 2001, 123, 7727–7729. (b) Antilla, J. C.; Baskin, J. M.;
Barder, T. E.; Buchwald, S. L. J. Org. Chem. 2004, 69, 5578–5587.
(c) Ley, S. V.; Thomas, A. W. Angew. Chem., Int. Ed. 2003, 42, 5400-
5449.
(5) Bacchi, S.; Delpogetto, M.; Guelfi, S.; Perboni, A.; Ribecai, A.; Stabile,
P.; Tampieri, M. PCT Int. Appl. WO/2006/108693, 2006.
(6) Andreotti, D.; Bacchi, S.; Delpogetto, M.; Guelfi, S.; Perboni, A.;
Ribecai, A.; Spada, S.; Stabile, P.; Tampieri, M. PCT Int. Appl. WO/
2006/108689, 2006.
896
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Scheme 4. Route A to prepare CRF-1 candidates 1a-c
the cyclisation to the bicyclic pyridones 6 extremely unclean
and problematic to perform without using NaH in DMF. In
addition, it is known that the combination of NaH and DMF is
intrinsically unsafe, behaving as a self-accelerating exothermic
reaction.7 The pyridone derivatives 6 looked attractive as key
intermediates to focus on during the development of a new
route, and after considerable effort, a completely novel approach
to 6 was put in place, as depicted in Scheme 4.
the last chemical transformation, we were committed to finding
out more efficient conditions, to identifying the typical impuri-
ties, and to gaining understanding of their formation mecha-
nisms. The main issue of the Medicinal Chemistry conditions,
apart from the obvious low yield, was the high level of
impurities generated by the presence of multiple nitrogen atoms
in the pyrazolyl compounds 9, which can react with the iodides
8 under copper catalysis (Figure 2).10
According to the new process (route A), anilines 2 were
N-alkylated under basic conditions by 4-bromobutyronitrile to
generate compounds 10. The subsequent acid-promoted cy-
clization in isopropanol provided pyrrolidin-2-imines 11,8 which,
after a basic wash with a 2 N NaOH solution, were used as
crude in the following step. Intermediates 5 were prepared Via
a Michael addition of compounds 11 to 1.1 equiv of ethyl-2-
butynoate in refluxing tetrahydrofuran. This successful strategy
enabled us to achieve a much cleaner profile for the amidines
5. Potassium t-butoxide (2 equiv) in THF at reflux suitably
replaced NaH in DMF for the cyclisation step to form
intermediates 6, which were not isolated but directly subjected
to the triflation step. Moreover, the triflate derivatives 7 were
isolated by precipitation from an isopropanol/water mixture,
significantly improving the quality of the material obtained.
Noteworthy improvements were introduced also in other steps
of the process. In particular, the iodo-derivatives 8 were obtained
under much milder conditions which involved the addition of
1.2 equiv of methanesulfonic acid to activate the triflate
displacement by potassium iodide (2 equiv), which occurred at
85 °C in 1.5 h.9 Also the C-N coupling was optimized
considerably, thereby increasing the yield and guaranteeing
better-quality final drug substances (DS). In fact, with this being
In order to address this issue, we followed two strategies:
(i) identification of intermediates different from the iodides 8,
such as the corresponding bromo derivatives, less reactive and
thus potentially more selective in the C-N coupling; (ii) a
systematic screening of catalysts/ligands/bases and solvents in
the coupling involving the iodides 8. Notably, triflates 7 were
also tested under the aforementioned conditions but proved not
stable and gave quantitative hydrolysis to 6. As for the first
approach, bromo derivatives were indeed more selective in the
copper-catalysed reaction, generating lower amounts of com-
pounds 12, 13, and 14. However, the required higher temper-
atures (130-135 °C) and longer reaction times (38-43 h)
produced higher levels of aromatized compounds 15. Further-
more, the reaction became sluggish when scaling up from the
laboratory to kilo-lab scale and a certain percentage of the
bromides remained unreacted. The second approach proved
more successful, and after an extensive screening and consider-
able process optimization work, we came up with the efficient
conditions depicted in Scheme 4. In the optimized procedure,
the catalytic system was preformed by mixing 2 mol % of CuI
(7) (a) DeWall, G. Chem. Eng. News 1982, 60, 5–43. (b) Hesek, D.; Lee,
M.; Noll, B. C.; Fisher, J. F.; Mobashery, S. J. Org. Chem. 2009, 74,
2567–2570.
(8) Kwok, R.; Pranc, P. J. Org. Chem. 1967, 32, 738–740.
(9) Whitfield, H. J.; Griffin, R. J.; Hardcastle, I. R.; Henderson, A.;
Meneyrol, J.; Mesguiche, V.; Sayle, K. L.; Golding, B. T. Chem.
Commun. 2003, 2802–2803.
(10) (a) Klapars, A.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc. 2002,
124, 7421–7428. (b) Stabile, P.; Lamonica, A.; Ribecai, A.; Castoldi,
D.; Guercio, G.; Curcuruto, O. Tetrahedron Lett. 2010, 51, 3232–
3235.
Figure 2. Main impurities observed in the DS batches produced
by route A.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
897

Scheme 5. Route B to prepare CRF-1 candidates 1a-c
and 0.3 equiv of the ligand (1R,2R)-N,N′-dimethylcyclohexane-
1,2-diamine in DMF. When a clear solution was obtained, 3
equiv of K2CO3 and 2.5 equiv of pyrazolyl derivatives 9 were
added, followed by iodides 8. After 15 h at 90 °C, the final
products were precipitated directly from the reaction mixture.
Unfortunately, a crystallization step was not sufficient to
guarantee a DS of the quality needed for the clinical studies.
In fact, the quality umbrella on a clinical batch is not only tight
on the impurity control but also on the heavy metal contents
which must remain below specific levels.11 Since we were using
copper iodide in the last chemical transformation, we had to
develop a suitable treatment to ensure specifications were met.
A recrystallisation step was put in place which included
dissolution of the DS in a suitable solvent, filtration through
CUNO ZetaCarbon cartridge to retain copper and, finally,
precipitation by antisolvent addition. By applying those condi-
tions, the measured copper contents were constantly kept below
1 ppm level.
Development of a Robust Supply Route to CRF-1
Candidates 1a-c. The new route A allowed the successful
production of the first kilogram-batches of drug substances 1a
and 1b to support phase I toxicological studies and clinical
studies. However, less brilliant results were obtained for
compound 1c. In fact, the presence of a -CF3 group in place
of -CH3 substantially influenced the physical properties of
butyronitrile 10b, which was oil and not solid, accessible as a
pure material in low yield only after a challenging chromato-
graphic purification. Since the first part of the synthesis of 1c
did not have any isolated intermediate, we made significant
efforts to develop an alternative, more robust, and scalable route.
As a first option, we thought about introducing a solid version
of key intermediates 6 by screening suitable salts. A screening
of acids and solvents was run to identify the mesylate in
2-butanone as the best combination of counterion/solvent.
Furthermore, a quick win was the observation that, if stage 1
and the first step of stage 2 were telescoped and DIPEA was
not added, the hydrobromic acid formed during the alkylation
of anilines 2 could remain inside the reaction mixture, promoting
the cyclization step and also the precipitation of 11 as the
hydrobromide salts 16. As a consequence, the development
work focused mainly on the first stages of the synthesis to
efficiently prepare compounds 16 and 17, while the second part
of the process remained basically unchanged. A new route,
named route B, shorter and more efficient than the previous
one, was developed and successfully scaled up to produce
multikilogram batches of 1a-c (Scheme 5). We were delighted
to observe the total yield rising up to 20-25% for all three
considered compounds.
Development of a Manufacturing Supply Route to 1a.
Once phase I was completed, 1a was selected among the
compounds containing the 2,3-dihydro-1H-pyrrolo[2,3-b]pyri-
dine scaffold as the most promising to progress in the late
clinical development. Route B demonstrated viable on scale
but still showed limitations when critically analyzed to assess
suitability for industrialization. A few weak points were
identified: (i) toxic and potentially carcinogenic DCM was
necessary in stage 2 to extract efficiently the water-soluble
pyridone 6a during workup; (ii) the synthesis of triflate 7a
involved a low-temperature triflation reaction employing the
humidity-sensitive triflic anhydride; (iii) the copper-catalyzed
coupling in stage 4 required 0.3 equiv of the very expensive
(1R,2R)-N,N′-dimethylcyclohexane-1,2-diamine to reach com-
pleteness in less than 24 h; (iV) a large excess of 2.5 equiv of
1-(1H-pyrazol-3-yl)imidazolidin-2-one, 9a, was needed, again
in stage 4, to avoid the competing arylation on the 2-imidazo-
lidinone free nitrogen (up to 25% of bis-arylated product 14
was observed if the temperature was not strictly kept under 95
°C). Notably, (1R,2R)-N,N′-dimethylcyclohexane-1,2-diamine
and 1-(1H-pyrazol-3-yl)imidazolidin-2-one represented, respec-
tively, 20% and 40% of the cost of goods (CoGs) of the entire
process.
(11) The tolerated residual level for copper was established by considering
the permitted daily exposure according to the European Medicines
Agency: Guideline on the Specification Limits for Residues of Metal
Catalysts, CPMP/SWP/QWP/4446/00; Committee for Human Me-
dicinal Products (CHMP), European Medicines Agency (EMEA):
London, 2007; http://www.ema.europa.eu/pdfs/human/swp/444600.pdf.
A thorough process optimization work was undertaken to
address all the above-mentioned issues and to come up with a
robust and reliable manufacturing route. The new process
(named route C) is outlined in Scheme 6.
898
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Scheme 6. Route C, manufacturing route to 1a
Scheme 7. Synthesis of 1-acetyl-3-(1H-pyrrol-3-yl)-
First of all, 2-methyltetrahydrofuran12 was advantageously
introduced in stage 2 as a new solvent; being higher boiling
than THF, it shortened reaction times by working at a higher
temperature; being nonmiscible with water, it replaced DCM
during the workup, and finally, it also revealed an efficient
solvent to crystallize pyridone 6a, thus skipping the methane-
sulfonate salt formation step. Another significant accomplish-
ment was gained in the following stage 3, where we identified
a way to obtain the iodide 8a directly from pyridone 6a, thus
escaping the triflation step and introducing a new stage 3. This
difficult goal was achieved by using N-iodosuccinimide (NIS)
in the presence of triphenylphosphite,13 taking advantage of the
high affinity for oxygen atom possessed by activated phospho-
rous derivatives. According to the literature mechanism hy-
pothesis, triphenylphosphite reacts with NIS to generate a very
reactive phosphorous adduct which then adds to the oxygen
atom of pyridone 6a, releasing succinimide and creating a good
leaving group. The subsequent displacement by iodide provided
the iodo-derivative 8a and generated triphenylphosphate as a
byproduct. The reaction was conducted in acetonitrile, by
preparing NIS/P(OPh)3 adduct at -10 °C from 3.5 equiv of
both reagents. The pyridone 6a was then added and the mixture
heated to 80 °C for 8 h. A subsequent treatment with aqueous
8 N NaOH at 60 °C allowed the hydrolysis of the excess reagent
and of triphenylphosphate and triggered the precipitation upon
cooling of desired 8a in high purity and with a typical yield in
the range 75-80%.
imidazolidin-2-one, 18
dimethylglycine/DMSO/DBU was identified as the most ad-
vantageous alternative to route B conditions in terms of
conversion and impurity profile. In addition, the replacement
of the expensive (1R,2R)-N,N′-dimethylcyclohexane-1,2-di-
amine with the cheaper N,N-dimethylglycine allowed a 10%
reduction of the CoGs. Also, the introduction of DBU to replace
K2CO3 offered the advantage of a homogeneous reaction.
Another breakthrough was the idea of a selective protection of
the 2-imidazolidinone free nitrogen with an acetyl group,15 to
avoid competing arylation during the coupling. Following this
strategy it was possible not only to completely suppress the
side reaction but also to employ only 1.3 equiv of pyrazolyl
compound 18 with a further reduction of 20% in the CoGs.
The acetylated intermediate 18 was efficiently obtained in
high yield and purity from 9a by a two-step synthesis which
consisted of an exhaustive acetylation of the two reactive
nitrogen atoms, followed by an in situ selective deacetylation
of the pyrazolyl moiety promoted by water. Eventually, the
desired product was filtered off directly from the reaction
mixture (Scheme 7).
Once the C-N coupling reached completeness, the acetyl
protecting group was cleaved easily and efficiently in situ by
adding MeOH to the reaction mixture and heating to 85 °C for
3 h. In fact, the basic environment promoted nucleophilic attack
of MeOH to the acetyl group with subsequent formation of
methyl acetate. Moreover, methanol acted as an antisolvent, and
1a was isolated by direct precipitation from the reaction mixture.
The quality of the API was remarkable (99.5% a/a HPLC, with
an assay of 99% w/w), and also the copper contamination was
Stage 4, the C-N coupling, was identified as the key stage
of the process since it generated the DS skeleton and all the
main impurities which can be found in the final API. For this
reason the development of this stage deserved a special attention.
In a first instance, a new and enlarged screening (including
ligands, copper sources, bases and solvents) was performed,
giving a particular focus to cheap aminoacidic ligands and
organic bases.14 As a result, the combination of CuI/N,N-
(12) Aycock, D. F. Org. Process Res. DeV. 2007, 11, 156–159.
(13) (a) Sugimoto, O.; Mori, M.; Tanji, K.-I. Tetrahedron Lett. 1999, 40,
7477–7478. (b) Sugimoto, O.; Mori, M.; Moriya, K.; Tanji, K.-I. HelV.
Chim. Acta 2001, 84, 1112–1118.
(15) Kravchenko, A. N.; Sigachev, A. S.; Gazieva, G. A.; Maksareva, E. Y.;
Trunova, N. S.; Chegaev, K. A.; Lyssenko, K. A.; Lyubetsky, D. V.;
Struchkova, M. I.; Il’in, M. M.; Davankov, V. A.; Lebedev, O. V.;
Makhova, N. N.; Tartakovsky, V. A.; Zelinsky, N. D. Chem.
Heterocycl. Compd. 2006, 42, 365–376.
(14) (a) Ma, D.; Cai, Q. Acc. Chem. Res. 2008, 41, 1450–1460. (b) Ma,
D.; Cai, Q.; Zhang, H. Org. Lett. 2003, 5, 2453–2455. (c) Ma, D.;
Cai, Q. Synlett 2004, 128–130. (d) Zhang, H.; Cai, Q.; Ma, D. J. Org.
Chem. 2005, 70, 5164–5173.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
899

kept under control and reduced to very low level (typically in
range of 5-45 ppm) by using 1 equiv of N,N-dimethylglycine
as a scavenger. The aminoacid was dissolved in the MeOH
added during the acetyl-group cleavage/precipitation step and
kept copper dissolved in the mother liquors as a soluble
complex.
The following particle-forming stage developed ad hoc for
the poorly soluble compound 1a consisted of dissolution of the
crude API in DMSO at 85 °C and subsequent controlled
crystallization by addition of 1-propanol at 65 °C. The described
methodology gave the desired API form and guaranteed a
further reduction of the already low copper contents in order
to meet specifications. The entire process demonstrated robust-
ness and reliability in pilot plant, allowing the production up
to 100 kg of API with an overall yield of 30-35%, confirming
its potential to become the manufacturing route for 1a.
Moreover, the same chemistry could be applied successfully
to prepare small batches of compounds 1b and 1c as well.
(90 mL) at 100 °C was added 4-bromobutyronitrile (1.1 equiv,
17.2 mL), and the resulting mixture was heated to 110-115
°C for 4 h. The mixture was then allowed to cool down to 45
°C, and a seed was added. After precipitation occurred, TBME
(270 mL) was added, and the resulting suspension was cooled
down to 20 °C, aged for 2 h, and then filtered. The cake was
washed with a 3:1 TBME/NMP mixture (3 × 60 mL), and the
1
solid was dried at 70 °C for 6 h (47 g, 88% yield). H NMR
(600 MHz, DMSO-d6): δ 9.83 (br s, 1H), 8.62 (br s, 1H), 7.58
(d, J ) 8.8 Hz, 1H), 7.48 (d, J ) 1.6 Hz, 1H), 7.40 (dd, J )
8.7, 1.5 Hz, 1H), 3.92 (t, J ) 7.1 Hz, 2H), 3.09 (m, 2H), 2.24
(s, 3H), 2.23 (m, 2H). MS (m/z): 259 [M + H]+.
1-(4-Methoxy-2-methylphenyl)-6-methyl-2,3-dihydro-1H-
pyrrolo[2,3-b]pyridin-4(7H)-one (6a). To a solution of 1-(4-
methoxy-2-methylphenyl)pyrrolidin-2-imine hydrobromide
(16a, 10 kg) in 2-MeTHF (70 L) was added a 10% w/w NaOH
aqueous solution (17.5 L). After aqueous phase separation, the
organic layer was washed with a 15% w/w NaCl aqueous
solution and dried by azeotropic distillation. Ethyl-2-butynoate
(1.1 equiv, 4.6 L) was added and the solution was refluxed for
8 h, then it was cooled down to 20 °C. In another vessel,
2-MeTHF (90 L) was charged, KOt-Bu (2.5 equiv, 10 kg) was
added and the suspension stirred for at least 30 min under
nitrogen before being transferred to the first vessel containing
the solution of 5a. The transfer was completed by washing the
line with fresh 2-MeTHF (60 L). The resulting mixture was
refluxed for 9 h; then it was concentrated down to 120 L,
allowed to cool down to 20 °C, washed with a 20% w/w NH4Cl
aqueous solution (40 L) and with a 15% w/w NaCl aqueous
solution (20 L), and finally was distilled down to 40 L. Fresh
2-MeTHF was added (80 L). The mixture was concentrated
down to 40 L, and then it was cooled down to 20 °C and a
seed added. The resulting suspension was aged for 6 h at 20
°C and filtered, and the cake was washed with 2-MeTHF (20
L) and dried in Vacuo at 50 °C for 18 h (5.2 kg, 55% yield).
1H NMR (600 MHz, DMSO-d6): δ 9.86 (br s, 1H), 7.10 (d, J
) 8.5 Hz, 1H), 6.82 (d, J ) 2.5 Hz, 1H), 6.75 (dd, J ) 8.7, 2.9
Hz, 1H), 5.93 (s, 1H), 3.74 (s, 3H), 3.70 (t, J ) 8.4 Hz, 2H),
2.91 (t, J ) 8.5 Hz, 2H), 2.14 (s, 3H), 2.04 (s, 3H). MS (m/z):
271 [M + H]+.
Conclusion
In conclusion, a robust, efficient, and scalable process was
developed for the synthesis of CRF-1 antagonists of general
formula 1 in 30-35% overall yield. The approach utilized a
convergent synthesis which introduced at the last stage the
expensive side chain 9, during a selective copper-catalysed C-N
coupling reaction. This more concise and atom-efficient route
used readily available starting materials, skipped the formations
of expensive intermediates, avoided the use of toxic reagents
and extremely low-yielding steps, and also had a more robust
final coupling stage. The problem of the clinical grade DS
copper contamination was efficiently resolved on-scale by using
a cheap and readily available scavenger, N,N-dimethylglycine,
which was also the ligand employed in the C-N coupling
reaction.
Experimental Section
All materials obtained from commercial suppliers were used
1
without further purification. H NMR spectra were recorded
on Varian Inova 400 or 600 spectrometers. LC/MS analyses
were performed on an Agilent 1100 series LC/MSD equipped
with APCI source.
6-Methyl-1-[2-methyl-4-(trifluoromethoxy)phenyl]-2,3-di-
hydro-1H-pyrrolo[2,3-b]pyridin-4(7H)-one (6b). The com-
pound was prepared according to the procedure applied for
6a.1H NMR (600 MHz, DMSO-d6): δ 9.78 (br s, 1H), 7.35 (d,
J ) 8.6 Hz, 1H), 7.29 (d, J ) 2.7 Hz, 1H), 7.20 (dd, J ) 8.7,
2.8 Hz, 1H), 5.95 (s, 1H), 3.72 (t, J ) 8.4 Hz, 2H), 2.90 (t, J
) 8.5 Hz, 2H), 2.19 (s, 3H), 2.06 (s, 3H). MS (m/z): 325
[M + H]+.
4-Iodo-1-(4-methoxy-2-methylphenyl)-6-methyl-2,3-dihy-
dro-1H-pyrrolo[2,3-b]pyridine (8a). To a mixture of N-
iodosuccinimide (3.5 equiv, 2.9 kg) in CH3CN (80 L) at 5 °C
was added P(OPh)3 (3.5 equiv, 3.4 L) in 1 h. 1-(4-Methoxy-
2-methylphenyl)-6-methyl-1,2,3,7-tetrahydro-4H-pyrrolo[2,3-b]py-
ridin-4-one, 6a, (10 kg) was added at 5 °C, and the resulting
mixture was refluxed for 8 h. The mixture was then cooled
down to 60 °C, and an 8 M NaOH aqueous solution (6.4 L)
was slowly added, keeping the temperature at 60-65 °C.
The mixture was then stirred for 2 h at 60 °C and cooled
1-(4-Methoxy-2-methylphenyl)pyrrolidin-2-imine Hydro-
bromide (16a). To a solution of 4-methoxy-2-methylaniline
(2a, 20 kg) in n-butanol (160 L) at 100 °C was added
4-bromobutyronitrile (1.2 equiv, 17.4 L), and the resulting
mixture was heated to 110-115 °C for 6 h. The mixture was
then allowed to cool down to 50 °C, and a seed was added.
After precipitation occurred, TBME (80 L) was added, and the
resulting suspension was cooled down to 20 °C, aged for 6 h,
and then filtered. The cake was washed with a 1/1 n-butanol/
TBME mixture (60 L) and TBME (60 L), and the solid dried
1
overnight at 50 °C (30 kg, 75% yield). H NMR (DMSO-d6,
400 MHz): δ (ppm) 8.1-9.5 (br s, 2H), 7.32 (d, J ) 8.8 Hz,
1H), 7.00 (d, J ) 2.9 Hz, 1H), 6.94 (dd, J ) 8.7, 3.0 Hz, 1H),
3.87 (t, J ) 7.1 Hz, 2H), 3.79 (s, 3H), 3.05 (t, J ) 8.0 Hz, 2H),
2.22 (m, 2H), 2.16 (s, 3H). MS (m/z): 205 [M + H]+.
1-[2-Methyl-4-(trifluoromethoxy)phenyl]pyrrolidin-2-
imine Hydrobromide (16b). To a solution of 2-methyl-4-
(trifluoromethoxy)aniline (2b, 30 g) in 1-methyl-2-pyrrolidinone
900
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

down to 20 °C, and water (30 L) was added in 30 min. The
resulting suspension was aged for 3 h at 20 °C before being
filtered, and the cake was washed with water (80 L). The
collected solid was dried in Vacuo at 50 °C for 18 h (10.9
kg, 77% yield).1H NMR (600 MHz, DMSO-d6): δ 7.15 (d,
J ) 8.5 Hz, 1H), 6.85 (d, J ) 2.7 Hz, 1H), 6.78 (dd, J )
8.7, 2.9 Hz, 1H), 6.72 (s, 1H), 3.82 (t, J ) 8.5 Hz, 2H),
3.75 (s, 3H), 2.97 (t, J ) 8.4 Hz, 2H), 2.13 (s, 3H), 2.08 (s,
3H). MS (m/z): 381 [M + H]+.
4-Iodo-6-methyl-1-[2-methyl-4-(trifluoromethoxy)phenyl]-
2,3-dihydro-1H-pyrrolo[2,3-b]pyridine (8b). The compound
was prepared according to the procedure applied for 8a.1H NMR
(400 MHz, DMSO-d6): δ 7.40 (d, J ) 8.6 Hz, 1H), 7.31 (d, J
) 2.7 Hz, 1H), 7.22 (dd, J ) 8.7, 2.6 Hz, 1H), 6.81 (s, 1H),
3.90 (t, J ) 8.4 Hz, 2H), 3.00 (t, J ) 8.4 Hz, 2H), 2.21 (s, 3H),
2.12 (s, 3H). MS (m/z): 435 [M + H]+.
1-(1-{1-[4-Methoxy-2-methylphenyl]-6-methyl-2,3-dihy-
dro-1H-pyrrolo[2,3-b]pyridin-4-yl}-1H-pyrazol-3-yl)imida-
zolidin-2-one (1a). DMSO (30 L) was charged under nitrogen
followed by CuI (0.10 equiv, 0.5 kg) and N,N-dimethylglycine
(0.30 equiv, 0.82 kg). The suspension was stirred at 25 °C for
30 min. Compound 8a (10 kg) was added to the solution at 25
°C followed by 18 (1.3 equiv, 6.6 kg) and eventually by DBU
(2 equiv, 8 L) and DMSO (10 L). The suspension was heated
to 110 °C for 7 h. The temperature was brought down to 85
°C, and methanol (30 L) was added dropwise in 1 h. The
resulting suspension was stirred at 85 °C for 3 h. The suspension
was cooled down to 70 °C, and N,N-dimethylglycine (1 equiv,
2.8 kg) was added, followed by MeOH (70 L), that was added
dropwise in 1 h. The mixture was aged at 70 °C for 30 min,
then it was cooled down to 25 °C and stirred for 12 h at 25 °C.
The suspension was filtered under pressure in a pan-filter; the
cake was then washed four times upon the filter with MeOH
(40 L). The collected solid was dried at 50 °C in Vacuo to obtain
crude grade 1a as an off-white solid (9.6 kg, 90% yield).1H
NMR (600 MHz, DMSO-d6): δ 8.31 (d, J ) 2.7 Hz, 1H), 7.17
(d, J ) 8.8 Hz, 1H), 7.06 (s, 1H), 6.86 (d, J ) 2.7 Hz, 1H),
6.78 (d, J ) 2.5 Hz, 1H), 6.79 (dd, J ) 9.0, 2.7 Hz, 1H), 6.76
(s, 1H), 3.92 (m, 2H), 3.83 (t, J ) 8.4 Hz, 2H), 3.76 (s, 3H),
3.46 (m, 4H), 2.17 (s, 3H), 2.15 (s, 3H). MS (m/z): 405
[M + H]+.
2-(1-{1-[4-Methoxy-2-methylphenyl]-6-methyl-2,3-dihydro-
1H-pyrrolo[2,3-b]pyridin-4-yl}-1H-pyrazol-3-yl)thiazole (1b).
The compound was prepared according to the procedure applied
for 1a with 2-(1H-pyrazol-3-yl)thiazole (9b) used in place of
18 (410 g, 75% yield). 1H NMR (600 MHz, DMSO-d6): δ 8.59
(d, J ) 2.5 Hz, 1H), 7.94 (d, J ) 3.3 Hz, 1H), 7.80 (d, J ) 3.3
Hz, 1H), 7.19 (d, J ) 8.5 Hz, 1H), 7.07 (d, J ) 2.7 Hz, 1H),
6.88 (d, J ) 3.3 Hz, 1H), 6.87 (s, 1H), 6.81 (dd, J ) 8.5, 3.0
Hz, 1H), 3.90 (t, J ) 8.5 Hz, 2H), 3.77 (s, 3H), 3.50 (t, J ) 8.4
Hz, 2H), 2.23 (s, 3H), 2.17 (s, 3H). MS (m/z): 404 [M + H]+.
1-(1-{6-Methyl-1-[2-methyl-4-(trifluoromethoxy)phenyl]-
2,3-dihydro-1H-pyrrolo[2,3-b]pyridin-4-yl}-1H-pyrazol-3-
yl)imidazolidin-2-one (1c). The compound was prepared
according to the procedure applied for 1a (390 g, 73% yield).
1H NMR (600 MHz, DMSO-d6): δ 8.33 (d, J ) 2.7 Hz, 1H),
7.41 (d, J ) 8.5 Hz, 1H), 7.31 (d, J ) 2.7 Hz, 1H), 7.22 (dd,
J ) 8.8, 2.7 Hz, 1H), 7.07 (s, 1H), 6.85 (s, 1H), 6.80 (d, J )
2.7 Hz, 1H), 3.91 (m, 4H), 3.47 (m, 4H), 2.24 (s, 3H), 2.21 (s,
3H). MS (m/z): 459 [M + H]+.
1-acetyl-3-(1H-pyrazol-3-yl)imidazolidin-2-one (18). A
suspension of acetic anhydride (40 L), 1-(1H-pyrrol-3-yl)imi-
dazolidin-2-one (9a) (10 kg) and sodium acetate (1.2 equiv,
6.5 kg) was heated to 115 °C for 5.5 h until formation of 19
was complete. The temperature was then brought to 50 °C, and
water (20 L) was cautiously added in 1 h. The mixture was
heated to 95 °C and the suspension stirred for 4 h at 95 °C
until a complete hydrolysis of 19 to 18 was achieved. The
mixture was eventually cooled down to 20 °C and aged for 3 h
before filtering. The cake was washed three times upon the filter
with 20 L of water, and the wet solid was dried in Vacuo at 40
°C overnight. (10.9 kg, 86% yield).1H NMR (400 MHz,
DMSO-d6): δ 12.47 (br s, 1H), 7.70 (d, J ) 2.2 Hz, 1H), 6.54
(d, J ) 2.2 Hz, 1H), 3.83 (m, 4H), 2.41 (s, 3H). M/S (mz): 195
[M + H]+.
Received for review May 25, 2010.
OP100147H
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
901