Development of scaffold synthesis for the preparation of new insulin-like growth factor 1 receptor inhibitors

Article abstract of DOI:10.1021/op8002536

The synthesis of new insulin-like growth factor 1 receptor (IGF- 1R) inhibitors is reported. The described molecules have a new sulfonyl-indazole structure. We describe the process research and development for the scaffold synthetic procedure in order to

Full text of DOI:10.1021/op8002536

Organic Process Research & Development 2009, 13, 456–462
Development of Scaffold Synthesis for the Preparation of New Insulin-Like Growth
Factor 1 Receptor Inhibitors
Ilaria Candiani, Germano D’Arasmo,* Franco Heidempergher, and Attilio Tomasi
Pharmaceutical Science Department, Process Research and DeVelopment Group, NerViano Medical Sciences S.r.l.,
Via Pasteur 10, 20014 NerViano, Italy
Abstract:
The synthesis of new insulin-like growth factor 1 receptor (IGF-
1R) inhibitors is reported. The described molecules have a new
sulfonyl-indazole structure. We describe the process research and
development for the scaffold synthetic procedure in order to
provide the large amount of product required for lead optimiza-
tion, candidate selection, and preclinical studies.
Figure 1. Subfamily compounds.
Introduction
The insulin-like growth factor 1 receptor (IGF-1R) is a
selected subfamily 3, see Figure 1, consists of an indazole-like
scaffold 1 and a substituted benzoic moiety 2 condensed as a
side chain.
member of the insulin receptor subfamily of receptor tyrosine
kinases (RTKs) and it is expressed in a wide variety of cell
types. IGF-1R mature protein consists of two alpha chains,
which are extracellular and contain ligand-binding function, and
two beta chains, which span the cell membrane and contain
the intracellular kinase domain. Binding of its cognate ligands,
IGF-1 and IGF-2, to the extracellular domain of the receptor
triggers the activation of the intracellular tyrosine kinase domain
and receptor autophosphorylation. The activated receptors have
potent mitogenic, motogenic, and anti-apoptotic activity in a
wide range of cell types. There is much evidence linking
increased IGF-1R signaling to the development and progression
of cancer; up-regulated levels of both the receptors and their
ligands have been observed in a variety of human tumors,
particularly in association with pathological events such as
invasion and metastasis.
In a typical Medicinal Chemistry approach synthesis con-
sisted of three phases: preparation of the two key intermediates
and coupling to assemble the final molecule 3. This afforded
the flexibility of exploring a variety of side chains to optimize
pharmacokinetics properties. In view of the need for higher
amounts of compound for full preclinical characterization, and
possible successive development of the subclass, we undertook
evaluation of the synthetic procedure for 1 with the objective
of delivering a safe, robust, and reliable process suitable for
pilot-plant scale.
The original laboratory-scale procedures involved a number
of steps with conditions unsuitable for scale-up (e.g., using
concentrated hydrazine, low-yield reactions, etc.), and thus
workup and purification steps needed to be highly improved.
According to this strategy we initiated investigation of synthetic
routes for scaffold preparation.
A research effort within the IGF-1R kinase inhibitor pro-
gram1 at Nerviano Medical Sciences resulted in identification
of sulfonyl-indazole derivatives showing interesting activity.
Chemical expansion led to the identification of IGF-1R kinase
inhibitors with potent in vitro, cellular, and in vivo activity. The
Results and Discussion
The original synthetic route utilized 5-fluoro-2-nitrobenzoic
acid and 3,5-difluorobenzenesulfonylchloride as starting materi-
als as highlighted in Scheme 1. Common reactions had been
used to convert the carboxylic acid into a cyano group, while
the sulfonylchloride was reduced2 to a thioderivative.
Economic and practical considerations prompted us to
purchase 5-fluoro-2-nitrobenzonitrile 2 and 3,5-difluorothiophe-
nol 3 from suppliers (e.g., the reduction of 13 which occurs
via Vilsmeyer iminium salt formation proceeded with very low
yield). The starting material 3 must be stored in a nitrogen
atmosphere since its thio group is prone to oxidation. We
followed the shortest synthetic route, as reported in Scheme 2.
* To whom correspondence should be addressed. E-mail: germano.darasmo@
yahoo.it.
(1) (a) Bandiera, T.; Lombardi Borgia, A.; Nesi, M.; Perrone, E.; Bossi,
R.; Polucci, P. Indazole derivatives as kinase inhibitors for the
treatment of cancer. WO/2008/074749 A1, 2008. (b) Fancelli, D.
Pulici, M.; Moll, J.; Bandiera, T. Preparation of 1H-furo[3,2-
c]pyrazole-5-carboxamide derivative as protein kinase inhibitors. WO/
2007/138017, 2007. (c) Bandiera, T.; Lombardi Borgia, A.; Polucci,
P.; Villa, M.; Nesi, M.; Angiolini, M.; Varasi, M. Substituted
pyrazolo[4,3-c]pyridine derivatives as tyrosine kinase inhibitors,
particularly IGF-1R inhibitors, their preparation, pharmaceutical
compositions, and use in therapy. WO/2007/068619, 2007. (d)
Bandiera, T.; Lombardi Borgia, A.; Orrenius, S. C.; Perrone, E.; Beria,
I.; Fancelli, D.; Galvani, A. Substituted pyrrolopyrazole derivatives
as tyrosine kinase inhibitors, their preparation, pharmaceutical com-
positions, and use in therapy. WO/2007/068637, 2007. (e) Fancelli,
D.; Moll, J.; Pulici, M.; Quartieri, F.; Bandiera, T. Preparation of 1H-
thieno[2,3-c]pyrazoles as kinase, particularly Aurora kinases and IGF-
1R inhibitors for treating cancer. WO/2007/009898, 2007.
(2) Hiromi, U.; Susumu, K. Tetrahedron Lett. 1999, 40, 3179–3182.
10.1021/op8002536 CCC: $40.75  2009 American Chemical Society
456
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development
Published on Web 01/21/2009

Scheme 1. Original synthetic routea
a Reagents and conditions: i) SOCl2 (5.0 equiv), DMFcat., 85 °C, 2 h; ii) Dioxane, NH3 aq 33% (2.0 equiv); iii) POCl3 (4.0 equiv), 60 °C, 18 h; iv) solution of:
dimethyldichlorosilane (3.5 equiv), DCM, Zn, plus DMAc (3.0 equiv), reflux 1 h and 50 °C 16 h.
Scheme 2. New synthetic routea
a Reagents and conditions: i) EtOAc, DIPEA (1.05 equiv), 0 °C to rt, 0.5 h; ii) MeCN, water, Oxone (2.5 equiv), 40 °C to rt, 24; iii) THF, N2H4 35%, rt; iv) THF,
TFA, 5 °C, 15 min; v) DCM, TFAA; vi) DCM, TrCl, TEA, rt, 4 h; vii) MeOH, TEA, refluxed, 5h.
In the original procedure the nucleophilic attack of 3 to 2
was performed in heterogeneous conditions3 by using cesium
carbonate (Cs2CO3). A tetrahydrofuran (THF) solution of 3 was
added to a -40 °C solution of 2 and reacted at room temperature
overnight. Workup involved replacement of solvent with
dichloromethane (DCM), filtration to eliminate possible residual
salts, and washing with brine prior to product isolation by
Figure 2. Byproduct.
evaporation to dryness. The low temperature was recommended
for regioselective control, since both fluorine and nitro groups
are activated and undergo nucleophilic substitution giving a
significant amount of byproduct 14 formation (see Figure 2).
The structure was confirmed by 1H NMR characterization after
chromatographic isolation. We investigated the reaction seeking
best conditions to achieve minimization of 14. Temperature,
(3) (a) Ruhland, T.; Nielsen, S. D.; Holm, P.; Christensen, C. H. J. Comb.
Chem. 2007, 9 (2), 301–305. (b) Arora, V.; Salunkhe, M.; Sinha, N.;
Sinha, R. K.; Jain, S. Bioorg. Med. Chem. Lett. 2004, 14 (18), 4647–
4650. (c) Beugelmans, R.; Chbani, M. Bull. Soc. Chim. Fr. 1995, 132
(3), 290–305.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
457

Table 1. Different temperature addition
int. 2 int. 3 yield purity A%
entry (equiv) (°C) (equiv)
offer at least two main advantages. First, being the top phase
in the separation process allows quicker extraction (obviating
the need for a second reactor) and also allows the reactor to be
kept relatively dry. Second, DCM requires a very efficient
condensation plant for recovery. Usually, the recovery ratio does
not exceed 35-40% for DCM, while EtOAc could be recycled
almost entirely. We focused on the homogeneous phase reaction
because of the quicker reaction time. Following the positive
results obtained, we decided to apply this synthetic procedure.
The HEL parallel synthesis apparatus also allowed us to
obtain preliminary calorimetric data. Although the studies were
not expressly designed for this purpose (and thus must be
considered as indicative), a small exothermic effect was
nonetheless observed when DIPEA was added, increasing with
addition of thiophenol. The molar heat was estimated at 159.27
kJ or 38.10 kcal (based on dose of 1.76 g of 3,5-difluo-
rothiophenol).
We moved to the following step for the oxidation of sulfide
to sulfone, see Scheme 3. The intermediate 4 does not need to
be isolated, and it can be used directly as solution in the
oxidation process, with savings in time and resources. According
to the original synthetic procedure, it was made by using HIO4
(cat. Cr2O3). For obvious environmental issues and scale-up
considerations, this oxidation reactant was abandoned. After
having considered other reagents5 (e.g., urea/H2O2,6 H2O27), we
focused on investigating the use of Oxone8 (potassium monop-
ersulfate triple salt technical grade; KHSO5 47%) as oxidizing
agent. Oxone has the main advantage of being an inexpensive
reagent easily removed by workup, while it suffers from the
drawback of the relatively high amount of product that it is
necessary to use in relation to the oxygen content.
Different conditions of solvent, temperature, time, and molar
ratio were investigated for Oxone; the results are reported in
Table 3. Entry 1 shows that in N,N-dimethylformamide (DMF)
the reaction is relatively fast, being virtually completed after
6 h. Moving to a mixture of water/ethyl acetate, entry 2, the
sulfide is transformed relatively rapidly into the sulfoxide 4a,
but oxidation to sulfone 5 does not go to completion, as the
reaction runs in heterogeneous phase. The addition of a phase
transfer catalyst (i.e., Bu4NBr) and an extra amount of fresh
Oxone does not significantly improve the conversion. We
needed to improve the workup because, when water is added,
the product tends to crash out in lumps, trapping unreacted
Oxone salts, see entry 3. Dissolution in organic solvent was
T
14
by HPLC (A%)
%
1
2
3
4
5
1.0
1.0
1.0
1.0
1.0
-40
-3
0
1.00
1.15
1.15
1.15
1.15
95.5
98.9
100.0
94.3
90.3
98.6
98.1
98.4
98.1
98.0
0.24
1.03
0.90
1.20
1.14
0
25
solvents, bases, and molar ratios were evaluated by experiments
using the HEL parallel synthesis instrument, which allowed us
to run four reactions simultaneously.
A set of reactions were made that compared the original
procedure at different temperatures, the results of which are
reported in the Table 1. We noted that the reaction reaches
complete conversion within 2 h and 30 min, with no need to
react overnight. Entries 2-5 show that the addition between
-3 and 25 °C does not seem to significantly alter the content
of 14, the formation of which is actually more linked to the
equivalent (equiv) of thiophenol used within the reaction. In
entry 1, where we used 1.0 equiv it was significantly lower;
Cs2CO3 was used in the ratio of 1.1 equiv for all experiments.
Additional tests were performed for base and solvent
screening as reported in Table 2.
In entry 1 the reaction used for byproduct 14 isolation is
reported. The use of DCM gave promising results in terms of
yield and purity, achieving lower byproduct formation (see entry
3 versus entry 2). The calculated pKa for 3,5-difluorothiophenol
is 5.09 ( 0.11, which means that a strong base was not needed.
Upon general consideration usually 4 log units between the acid
and the base are sufficient to obtain a complete dissociation of
the species. Literature4 values show: pKa t-BuOK (19), EtONa/
MeONa (18), Na2CO3 (10), R3N (10). It is possible to reduce
the amount of cesium carbonate, and this represents a significant
economic advantage; see entries 4 and 5. We considered the
use of more economic carbonate and the switch to a homoge-
neous reaction in order to achieve a more efficient conversion.
As shown in entry 6 the homogeneous reaction obtained using
N,N-diisopropylethylamine (DIPEA) is faster and gives good
yield and quality product. The use of cheaper carbonates does
not appear to give a real advantage and notably results in a
worse impurity profile, possibly due to the higher reaction time,
see entries 7 and 8. These carbonates left a certain degree of
unreacted starting material, whereas cesium carbonate worked
quite well. This might be explained by the low polarity of DCM
versus that of the THF used previously, and possibly by the
potential impact of the particle size of the bases: Cs2CO3 is a
very soft powder with very small grain/crystals, while other
carbonates have larger particle size. In entry 9 we tested sodium
carbonate (Na2CO3) with an ethyl acetate (EtOAc)/water solvent
mixture in the attempt to speed up the reaction because of better
dissolution of the carbonate (bearing in mind, however, that
these two solvents are not completely miscible at alkaline pH).
The reaction is in fact faster, but unreacted starting material is
increased. This may be due to a moderate degradation of 3,5-
difluorothiophenol in water (e.g., oxidation effects) limiting
reaction completion. In entry 10 we further investigated the use
of ethyl acetate as solvent. On a plant scale this solvent would
(5) Caron, S.; Dugger, R. W.; Ruggeri, S. G.; Ragan, J. A.; Brown Ripin,
D. H. Chem. ReV. 2006, 106, 2943–2989.
(6) Cooper, M.; Heaney, H.; Newbold, A. J.; Sanderson, W. R. Synlett
1990, 533–535.
(7) (a) Shokrolahi, A.; Zali, A.; Pouretedal, H. R.; Mahdavi, M. Catal.
Commun. 2008, 9 (5), 859–863. (b) Bahrami, K. In Regio- and stereo-
controlled oxidations and reductions; Roberts, S. M., Whitall, J., Eds.;
Catalysts for Fine Chemical Synthesis, Vol. 5; Wiley: Chichester,
England; Hoboken, NJ, 2007; pp 283-287. (c) Shaabani, A.; Rezayan,
A. H. Catal. Commun. 2007, 8 (7), 1112–1116. (d) Velusamy, S.;
Kumar, A.; Saini, R.; Punniyamurthy, T. Tetrahedron Lett. 2005, 46
(22), 3819–3822.
(8) (a) Ward, R. S.; Roberts, D. W.; Diaper, R. L.; Richard, L. Sulfur
Lett. 2000, 23 (3), 139–144. (b) Kropp, P. J.; Breton, G. W.; Fields,
J. D.; Tung, J. C.; Loomis, B. R. J. Am. Chem. Soc. 2000, 122 (18),
4280–4285. (c) Llauger, L.; He, H.; Chiosis, G. Tetrahedron Lett. 2004,
45 (52), 9549–9552.
(4) Comer, J. Chem. Br. 1994, 983–986.
458
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Table 2. Base and solvent screening
int. 2
int. 3
base
(equiv)
yield
%
purity A% RRTc1.57
14
time not reacted
entry (equiv) (equiv)
HPLC
impurity
(A% HPLC)
(h)
int. 2
solvent
1
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
2.0
1.03
1.03
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Cs2CO3 (2.0)
Cs2CO3 (1.05) 91.0
Cs2CO3 (1.05) 94.3
Cs2CO3 (1.0)
Cs2CO3 (0.5)
DIPEA (1.0)
K2CO3 (1.0)
Na2CO3 (1.0)
Na2CO3 (0.5)
DIPEA (1.0)
-
20.0
93.7
98.9
98.5
97.4
99.2
94.6
74.8
95.2
95.9
-
80.0
3.7
0.7
0.4
0.7
0.4
0.6
-
4
-
THF
2
0.5
0.12
-
2
1.4
0.2
1.0
1.5
0.3
2.3
19.7
2.4
1.6
THF
3
2
DCM
4
92.6
93.2
95.7
94.6
51.1
96.6
100.0
2
DCM
5
0.3
-
2.32
5.0
0.91
-
2
DCM
6
0a
23
28
2
DCM
7
DCM
8b
9
DCM
1.3
2.3
AcOEt/H2O
AcOEt
10
0a
a The reaction was instantaneous. b Stirring problem experienced. c Related to instability of thiophenol derivate (e.g., oxidation process), HPLC method no.1.
Scheme 3. Oxidation to sulfone
addition of a solution of 4 in acetonitrile was performed. The
exothermal effect of the oxidation process seemed to be
balanced and compensated for by the endothermal response to
the water/acetonitrile mixing. Overall this reaction showed an
endothermic behaviour. Although values must be considered
as only indicative, the molar heat was estimated at -93.34 kJ
or -22.33 kcal (based on 1.0 g of 4).
With intermediate 5 in hand we moved to formation of the
indazole ring. The synthesis went through a nucleophilic
substitution of the nitro group by hydrazine9 followed by ring-
closure10 involving the cyano group as reported in Scheme 4.
The preparation of 6 was pursued by employing the safer
35% hydrazine (N2H4) in water at room temperature instead of
the N2H4 1.0 M in THF as in the original procedure. Following
the quench in water the hydrazine derivative precipitated and
was isolated by simple filtration. Although the nitro group, being
doubly activated by cyano and sulfonyl groups, was readily
displaced, the evolving nitrite salt was deleterious to the
2-hydrazino intermediate. The reaction run in methanol or
butanol gave the indazole in moderate yield (no more than 60%)
and in low-purity grade. The nature of the solvent is crucial; in
alcohols the reaction is extremely slow at room temperature
(rt) (and dirty upon heating), while in DMF complete degrada-
tion occurs from even -78° to rt. We found THF to be a good
compromise: addition of hydrazine hydrate to diluted THF
solutions of 5 usually brought complete conversion to 6
accompanied by variable amounts of the final indazole (cy-
clization is catalyzed by acids, even by silica upon TLC).
Eliminating any nitrite salt before acidification is of crucial
importance, since nitrous acid rapidly converts the intermediate
2-hydrazinonitrile to the corresponding 2-azidonitrile. The cake
of 6 on the filter must be washed thoroughly with water. THF
is the same solvent used in the next step. In comparison with
the original procedure we also modified the order of addition
of the starting material 5 solution. It was dropped into the
solution of N2H4 to avoid the risk of producing the dimer 6a
(see Figure 4). The crude, typically obtained with a 96% yield,
required in order to efficiently wash the product with a reducing
agent (e.g., Na2S2O5 or Na2SO3) for neutralization of remaining
oxidant.
Unfortunately, when using methanol/water (entry 4) again
the reaction was incomplete; the ratio between 4a and 5 was
30:65, possibly due to poor solubility or heterogeneous condi-
tions. In entries 5 and 6 the effect of different reaction
temperatures using acetone/water are compared. Despite the fact
that both reactions were incomplete (a significant amount of
sulfoxide remained), we observed during HPLC monitoring that
at 40 °C the reaction was very fast in the early stage and
probably stopped because Oxone started to decompose. The
main impurity in this step is the one reported as RRT 1.24 min.
By HPLC study we saw that this is related to the synthesis of
4 because of the formation in this step of the double nucleophilic
substitution byproduct 14. When sulfide with a high level of
14 is used, a high level of the impurity RRT 1.24 min is
obtained (e.g., see entries 4 and 5). In fact, not surprisingly it
can go through the process generating oxidized compounds such
as the disulfoxide 15, see Figure 3, or sulfide-sulfoxide, etc.
As yet, we have not isolated this impurity, but in entry 7 we
applied the same conditions as for entry 5 using a cleaner
intermediate 4, and we obtained confirmation of our hypothesis.
We achieved a breakthrough in the oxidation step when
using a mixture of acetonitrile/water as solvent. Running the
reaction at 40 °C we obtained virtually complete conversion,
see entries 8, 9, and 10, with a good yield and impurity profile.
In this set of reactions we optimized the amount of Oxone to
2.5 equiv and the workup procedure throughout ethyl acetate
extraction and deep washing to neutralize and eliminate residual
salts.
(9) (a) Lokhande, P. D.; Raheem, A.; Sabale, S. T.; Chabukswar, A. R.;
Jagdale, S. C. Tetrahedron Lett. 2007, 48 (39), 6890–6892. (b)
Suryakiran, N.; Prabhakar, P.; Venkateswarlu, Y. Chem. Lett. 2007,
36 (11), 1370–1371.
(10) (a) Shirtcliff, L. D.; Hayes, A. G.; Haley, M. M.; Ko¨hler, F.; Hess,
K.; Herges, R. J. Am. Chem. Soc. 2006, 128 (30), 9711–9721. (b)
Yakaiah, T.; Lingaiah, B. P. V.; Narsaiah, B.; Shireesha, B.; Ashok,
K. B.; Gururaj, S.; Parthasarathy, t.; Sridhar, B. Bioorg. Med. Chem.
Lett. 2007, 17 (12), 3445–3453. (c) Lukin, K.; Hsu, M. C.; Fernando,
D.; Leanna, M. R. J. Org. Chem. 2006, 71 (21), 8166–8172.
As in the previous step we made a gross calorimetric
evaluation (entry 10) by using HEL parallel synthesis. By adding
Oxone in water an endothermal response was detected. Next,
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
459

Table 3. Oxidation screening conditions
4
Oxone
(equiv)
yield
%
purity %
by HPLC
15
reaction
time (h)
23a
143
T
(°C)
entry
(equiv)
4a
(RRT 1.24)d
solvent
DMF
AcOEt/water
DMF
MeOH/water
acetone/water
acetone/water
acetone/water
MeCN/water
MeCN/water
MeCN/water
1
2
3
4
5
6
7
8
9
10
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
2.0
5.05
2.0
2.5
2.0
4.0
3.0
3.0
2.5
2.5
92.8
92.0
121.8
92.0
91.0
100.0
111.6
87.3
110.0
92.0
95.4
81.5
92.8
87.7
91.5
93.1
91.6
98.4
95.2
98.4
-
0.8
25
10.6
0.3
6.5
2.4
5.5
7.3
0.2
0.3
0.5
4.3
25
2.2
15.5
96
25
4.4
25
4.5
23 hb
28 hc
24 hb
24 ha,b
24 ha,b
24 ha,b
40b
25c
40b
40b
40b
40b
0.4
0.28
0.27
2.4
0.24
a After 8 h the reaction was substantially complete. b 8 h at 40 °C and left at rt overnight. c 25 h at rt and 3 h at 40 °. d HPLC method no.1 (see Supporting Information).
Figure 4. Dimer formation.
Figure 3. Possible byproducts.
Scheme 4. Indazole ring formation
tion of the product took place. The organic solution was washed
by water twice in order to cleave any residual TFAA and then
cooled at 4 °C overnight. The crystallized solid was collected
by filtration; overall yield from intermediate 6 to 8 was usually
around 63%.
The trityl group11,13 had the best results during the original
screening. It was prepared by dissolving 8 in DCM and adding
1.0 equiv of TrCl and 3.0 equiv of triethylamine (TEA) followed
by stirring at room temperature for 4 h. The reacting solution
was quenched by washing with ammonium chloride (NH4Cl)
and was carried over to the last step without product isolation.
Finally, the cleavage of the trifluoroacetamido group11 to get
the target molecule 1 was made by switching the DCM with
MeOH and using 4.0 equiv of TEA. The reaction was gently
refluxed 5 h to complete the conversion. Previously, sodium
hydroxide (NaOH) was used, but this gave a 10% byproduct
due to a monofluoro substitution with a hydroxyl group. This
byproduct was very difficult to eliminate even with chromato-
graphic purification. After workup the compound was obtained
as a pure solid, 98.21% by HPLC. The overall yield for
intermediate 8 to 1 is usually 85%. With a robust synthesis in
hand to provide a large amount of the scaffold 1, it was possible
to proceed in further development of these IGF-1R inhibitors.
was employed without further purification for the cyclization
process to obtain 7. This workup avoids the need to evaporate
the solvent, bearing in mind the problems associated with the
presence of N2H4. The conversion of 6 to 7 was pursued in
THF at 5 °C, employing 1.0 equiv of trifluoroacetic acid (TFA)
for 15 min. Alternatively, the hydrochloride salt could also be
obtained by using HCl in organic solvent.
Intermediate 7 was not isolated since the reaction mixture
was utilized for the following step. The preparation was carried
out with the procedure described above, giving a product with
86.8% A% HPLC purity. The following reactions were ordinary
protection and deprotection of amino groups11 needed to reach
the target molecule 1, and they were pursued in standard
conditions, Scheme 2. To discriminate between the primary
amino group and the indazole nitrogen, TFAc and tritylchloride
(TrCl) groups were chosen as protections. The trifluoroacety-
lation11,12 on the amino group of 7 to obtain intermediate 8 was
made by switching the THF with DCM and using 2 equiv of
trifluoroacetic anhydride (TFAA). In theory, 2 equiv of TFAA
are required, because of concomitant acylation at ring nitrogens.
Addition of TFAA had the immediate effect of dissolving the
starting material, initially as a suspension. Soon after, precipita-
Conclusion
We have synthesized 1 formally in seven linear steps starting
from readily available starting materials. Three of the intermedi-
ates do not need to be isolated, the synthesis is short and efficient
and has been scaled up to multigram preparations. It represents
a substantial simplification and safer protocol compared to the
original medicinal chemistry process, which had very low-yield
steps and required chromatographic purifications. Environmen-
tally unsuitable and hazardous reagents have been changed; for
example, concentrated hydrazine has been replaced with the
(11) Greene, T. W. Wuts, P. G. M. ProtectiVe Groups in Organic Synthesis
2nd ed.;Wiley: New York, 1991.
(12) (a) Takamura, M.; Hamashima, Y.; Usuda, H.; Kanai, M.; Shibasaki,
M. Chem. Pharm. Bull. 2000, 48 (10), 1586–1592. (b) Engqvist, R.;
Stensland, B.; Bergman, J. Tetrahedron 2005, 61 (18), 4495–4500.
(13) Claramunt, R. M.; Elguero, J.; Fruchier, A. Bull. Soc. Chim. Belg.
1985, 94 (6), 421–424.
460
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

safer 35% hydrazine in water; we introduced the safer and
environmentally friendly Oxone as the oxidizing agent. Overall,
we fulfilled our objective in order to provide a process able to
deliver the requested amount of product for development work.
8.80, 2.15 Hz, 1H), 7.02 (dd, J ) 6.45, 2.15 Hz, 2H), 6.87-6.95
(m, 1H); LC/MS: Rt ) 6.25 min, 99%; ES+, ES- not
responding.
5-(3,5-Difluorobenzenesulfonyl)-2-nitrobenzonitrile (5). In
a 1.0 L jacketed reactor, equipped with an overhead stirrer, was
charged 138 g (0.228 mol) of Oxone and 500 mL of water.
The mixture ws stirred vigorously to complete dissolution with
the temperature maintained at 25 °C. The solution of 4 obtained
in the preparation described above was then added. For the 400
mL acetonitrile solution a quantitative yield was assumed,
corresponding to 26.4 g (0.091 mol) of 4. The mixture was
heated to 40 °C for 7 h 30 min; the heating system was switched
off and left to react overnight at room temperature. After an
overall 24 h, the mixture was concentrated by evaporation of
acetonitrile at reduced pressure. To the resulting mixture was
added 400 mL of ethyl acetate and 100 mL of water. After
separation, the organic phase was washed once with NaCl 30%
(4 × 300 mL), Na2SO3 15% (1 × 150 mL), and then again
with NaCl 30% (1 × 300 mL). In the final wash the acqueous
phase pH ≈ 7. The organic phase was concentrated for product
isolation and characterization. After drying at 60 °C overnight,
27.91 g of yellow solid was obtained with 94.3% yield: DSC
mp ) 120.07 °C (steel crucible) mp ) 118.02 °C (aluminium
crucible); HPLC (method no.1) Rt ) 10.3 min. purity 95.3%;
1H NMR (CDCl3) δ 8.41 (d, J ) 8.60 Hz, 1H), 8.36 (d, J )
1.96 Hz, 1H), 8.28 (dd, J ) 8.60, 1.95 Hz, 1H), 7.41-7.50
(m, 1H), 7.04-7.12 (m, 2H); LC/MS: Rt ) 5.45 min, 100%;
ES+, ES- not responding.
Experimental Section
General. 5-Fluoro-2-nitrobenzonitrile 98% (batch F03937AD)
was purchased from ABCR; 3,5-difluorothiophenol 97% (batch
Y051) was from Fluorchem. All the other chemicals and
reagents were purchased from Sigma-Aldrich and Riede-de
Haen. The 1H NMR spectra were recorded on a Varian Mercury
Plus 400 MHz spectrometer. The high performance liquid
chromatography (HPLC) analyses were carried out using an
Agilent series 1100 system and Waters series Millennium
according to methods and columns reported in the Supporting
Information. The LC/MS system used is a Waters Alliance HT
2795 separation module with a UV-VID detector Waters 996
PDA and a MS detector Waters ZQTM (single quadrupole). Two
methods were applied, one with acid eluent and one with basic
eluent, as reported in the Supporting Information. The melting
points were obtained through a DSC analysis made with a
Mettler Toledo Star system DSC 822. The parallel reactions
trials were performed using a Hazard Evaluation Laboratory
(HEL) parallel synthesis chem-SCAN and auto-MATE system
equipped with syringe pumps (Supporting Information).
5-(3,5-Difluorophenylsulfonyl)-2-nitrobenzonitrile (4). In
a 1.0 L jacketed reactor, equipped with an overhead stirrer, were
charged 15.16 g (0.091 mol) of 5-fluoro-2-nitrobenzonitrile and
195 mL of ethyl acetate at room temperature. The yellow
solution was cooled to 0 °C, and 16.64 mL (0.096 mol) of
DIPEA was added while stirring. In the meantime in a second
round-bottom flask were charged 13.34 g (0.091 mol) of 3,5-
difluorothiophenol and 150 mL of ethyl acetate while stirring
at room temperature. This solution was transferred into a
dropping funnel and added to the main reactor. The addition
was made over ∼45 min. The mixture was left to react for 30
min after the end of the addition.
Workup. The mixture was cooled at 0 °C, and then a solution
of 150 mL brine was added. The quenched mixture was warmed
to 25 °C, and 50 mL of additional water was added. The two
layers were separated, and the organic phase was washed again
with a solution of 150 mL of brine and 50 mL of water. The
solution was concentrated to a small volume (100 mL) by
evaporation on reduced pressure. [Ethylacetate (100 mL) was
added, and the resulting mixture was concentrated to reduced
volume for product isolation and characterization. This was
dried at 50 °C for at least 6 h, and a pale-yellow solid was
obtained in a 95.0% yield.]
In the synthetic protocol the isolation process could be
avoided by operating as follows: after the first concentration to
100 mL, 100 mL of acetonitrile was added, and the volume
was reduced again by evaporation to 100 mL. Finally 400 mL
of acetonitrile was added, and the solution went to the following
step without isolation: DSC mp ) 139.84 °C (steel crucible),
mp ) 134.81 °C (aluminium crucible); HPLC (method no.1)
Rt ) 13.1 min., purity >96.0%; 1H NMR (CDCl3) δ 8.16 (d,
J ) 8.60 Hz, 1H), 7.48 (d, J ) 1.96 Hz, 1H), 7.42 (dd, J )
2,5-Bis-(3,5-difluorophenylsulfonyl)benzonitrile (14). Into
a 25 mL three-necked round-bottomed flask was charged 0.5 g
(3.0 mmol) of 5-fluoro-2-nitrobenzonitrile, and this was dis-
solved in 10 mL of THF. Cs2CO3 [2.06 g (6.0 mmol)] was
added and cooled to 0 °C while stirring. A solution of 0.88 g
(6.0 mmol) of 3,5-difluorothiophenol in 5 mL of THF was
added over 5 min.
Workup. The reaction was monitored by HPLC after 4 h
and concentrated to low volume with a rotary evaporator. DCM
(60 mL) was added; the mixture was washed with brine (2 ×
30 mL), and the organic phase was treated with Na2SO4 prior
to filtration on Gooch. After evaporation to dryness a gross
yellow solid was obtained as a mixture with 4 and the byproduct
14 as main products. The target compound was isolated and
purified by flash chromatography on Silica-gel 60 Å; eluent
hexane/ethyl acetate, 10:1, followed by hexane/ethyl acetate,
10:3. TLC was checked by Pancaldi’s reagent, Rf in hexane/
ethyl acetate, 10:1, results are as follows: (4) 0.09, (2) 0.18,
(14) 0.33, (3) 0.70. After drying at 40 °C for at least 4 h, 0.78 g
of 14 was obtained in a yield of 67.0% as a white solid: HPLC
(method no. 1) Rt ) 21.4 min; purity 99.3%; 1H NMR (CDCl3)
δ 7.54 (d, J ) 1.96, 1H), 7.38 (dd, J ) 8.41, 2.15 Hz, 1H),
7.25 (d, 1H), 6.78-6.86 (m, 4H), 6.68-6.76 (m, 2H).
5-(3,5-Difluorobenzenesulfonyl)-2-hydrazinobenzoni-
trile (6). Ten grams (30.84 mmol) of 5-(3,5-difluorobenzene-
sulfonyl)-2-nitrobenzonitrile dissolved in 70 mL of THF was
dropped at 4 °C over 15 min into a solution of 5.6 mL of 35%
N2H4 (61.6 mmol) in 30 mL of THF. The reaction was stirred
at ambient temperature for 4 h. HPLC analysis revealed
complete disappearance of the staring material. At 4.0 °C 430
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
461

mL of cold water was added, and the mixture was stirred for
15 min. The solid product was collected by filtration and washed
with 30 mL of cold water. After drying at 45 °C for at least
12 h 9.23 g of a crude (96.8% yield, strength not determined,
HPLC A% (at 254 nm) ) 90.5%) was obtained and employed
for the next step without further purification. A small aliquot
was isolated for NMR and MS characterization: HPLC (method
5-(3,5-Difluorophenyl)sulfonyl-3-trifluoroacetamido-1-tri-
tylindazole (9). 5-(3,5-Difluorophenyl)sulfonyl-3-trifluoro-ac-
etamido-indazole [3.0 g (7.4 mmol)] was dissolved in 18 mL
of DCM. The solution was cooled to 4.0 °C, and 1.67 mL (22.2
mmol) of TEA was added. At 4 °C and over 5 min was added
dropwise triphenylmethyl chloride [2.06 g (7.4 mmol)] dissolved
in 3.5 mL of DCM. HPLC monitoring after 1 h of stirring
revealed complete conversion (HPLC at 254 nm ) 97.8%).
The solvent was rotaevapotated to 1/10 of the initial volume,
and this solution was processed further in the following step
without isolation. A small aliquot was isolated for characteriza-
tion: HPLC (method no. 2) Rt ) 15.5 min; 1H NMR (DMSO-
d6) δ 8.46 (dd, J ) 1.52, 0.65 Hz, 1H), 7.77 (dd, J ) 8.60,
1.52 Hz, 1H), 7.57-7.47 (m, 2H), 7.35-7.18 (m, 10H),
6.97-6.90 (m, 7H); LC/MS: Rt) 8.07 min, 100%, (M+) 648
weak, (M-) 646.
3-Amino-[5-(3,5-difluorobenzenesulfonyl)-1-tritylinda-
zole (1). The solution obtained above was diluted with 20 mL
of MeOH, and 2.8 mL (37 mmol) of TEA was added. The
remaining DCM was eliminated by evaporation, and the reaction
was gently refluxed for 5 h. HPLC control revealed the complete
removal of the trifluoroacetamido group (A% at 254 nm )
97.0%). The reaction was left at ambient temperature overnight,
and the solid thus obtained was collected by filtration, affording
3.46 g (84.7%; overall yield of two reactions, from intermediate
8) of target as off-white powder: HPLC (method no. 2) Rt )
9.5 min. purity 94.33%; 1H NMR (DMSO-d6) δ 8.54 (dd, J )
1.52, 0.65 Hz, 1H), 7.67-7.60 (m, 3H), 7.54 (dd, J ) 8.60,
1.52 Hz, 1H), 7.33-7.20 (m, 15H), 6.34 (dd, J ) 8.60, 0.65
Hz, 1H), 6.04 (sb, 1H, NH2); LC/MS: Rt ) 7.66 min, 100%;
(M+) 552 very weak, (M-) 550 very weak, 610 (M+ + HOAc).
1
no. 2) Rt ) 1.91 min; purity 90.5%; H NMR (DMSO-d6) δ
8.57 (s, 1H), 8.09 (d, J ) 2.35 Hz, 1H), 7.86 (dd, J ) 9.19,
2.15 Hz, 1H), 7.75-7.66 (m, 2H), 7.64-7.56 (m, 1H), 4.56 (s,
2H); MS m/z (M+) 310, (M-) 308.
5-(3,5-Difluorobenzenesulfonyl)-1H-indazol-3-ylamine, Tri-
fluoroacetic Salt (7). 5-(3,5-Difluorobenzenesulfonyl)-2-hy-
drazinobenzonitrile [6.16 g (19.9 mmol, theoretical value)] was
dissolved in 120 mL of THF. The solution was cooled to 4 °C,
and 1.48 mL (19.9 mmol) of TFA was added dropwise over 3
min. HPLC control after 30 min of stirring revealed a complete
cyclization process (A% at 254 nm ) 97.3%). The solvent was
rotaevaporated to 1/20 of the initial volume. A small aliquot
was isolated for characterization, obtaining a pale-cream
powder: HPLC (method no. 2) Rt ) 1.51 min; purity 97.3%;
free base 1H NMR (DMSO-d6) δ 12.05 (s, 1H, NH), 8.59 (s,
1H), 7.77 (dd, J ) 8.99, 1.96 Hz, 1H), 7.68-7.57 (m, 3H),
7.40 (d, J ) 8.99 Hz, 1H), 5.81 (s, 2H, NH2); LC/MS: Rt )
3.84 min, 95%; MS m/z (M+) 310, (M-) 308.
5-(3,5-Difluorophenyl)sulfonyl-3-trifluoro-acetamidoinda-
zole (8). The solution obtained above was diluted with 120 mL
of DCM, and 5.6 mL (39.8 mmol) of TFAA was added
dropwise over 1 h to the stirred solution maintained at 4 °C.
After 1 h of stirring the reaction was monitored by HPLC
analysis. Examination revealed the complete disappearance of
the starting material. The reaction mixture was washed twice
with water (2 × 150 mL). The organic phase was stored at 4
°C overnight. The whitish crystallized solid was collected by
filtration, yielding 5 g of target as whitish powder, HPLC A%
at 254 nm ) 97.59%. The overall yield is calculated over the
two reactions (from intermediate 6) and results to be 62.5%:
HPLC (method no. 2) Rt ) 1.97 min; purity 99.5%; 1H NMR
(DMSO-d6) δ 13.63 (s, 1H, NH-indazole), 12.22 (s, 1H, NH-
COCF3), 8.63 (dt, J ) 1.50, 0.70, 0.65 Hz, 1H), 7.94 (dd, J )
8.60, 1.52 Hz, 1H), 7.76-7.70 (m, 3H), 7.64 (sc, J ) 9.02,
2.80, 1H); LC/MS: Rt ) 4.82 min; m/z (M+) 405, (M - H2O-)
386.
Acknowledgment
We thank Dr. Marco Cattaneo and Mr. Mario Ungari for
their support in the DSC analysis.
Supporting Information Available
Additional experimental details. This material is available
free of charge via the Internet at http://pubs.acs.org.
Received for review October 6, 2008.
OP8002536
462
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development