Cell permeability as a parameter for lead generation in the protein Tyrosine kinase inhibition field

Article abstract of DOI:10.1016/S0960-894X(01)00022-1

Based on the inverse relationship between polar surface area and cell permeability and capitalizing on the properties of pyrrolopyrimidines 1 as protein tyrosine kinase inhibitors, pyrrolopyridones 2 were designed and synthesized as potential leads for th

Full text of DOI:10.1016/S0960-894X(01)00022-1

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1549±1552
Cell Permeability as a Parameter for Lead Generation in the
Protein Tyrosine Kinase Inhibition Field
Christos Papageorgiou,* Gian Camenisch and Xaver Borer
Novartis Pharma AG, WSJ350.3.14, CH-4002 Basel, Switzerland
Received 24 November 2000; accepted 29 December 2000
AbstractÐBased on the inverse relationship between polar surface area and cell permeability and capitalizing on the properties of
pyrrolopyrimidines 1 as protein tyrosine kinase inhibitors, pyrrolopyridones 2 were designed and synthesized as potential leads for
the development of novel inhibitors with improved cell permeability properties. # 2001 Elsevier Science Ltd. All rights reserved.
Protein tyrosine kinases (PTKs) tightly regulate cellular
function such as activation, proliferation, di€erentia-
tion, and apoptosis and, therefore, represent attractive
targets for the development of novel therapeutics,
particularly in the ®elds of cancer, psoriasis, restenosis,
or transplantation.1,2 The huge majority of the structu-
rally diverse enzyme inhibitors reported so far compete
with adenosine triphosphate (ATP) for binding at the
catalytic kinase domains, which share substantial struc-
tural similarities because of both the signi®cant amino
acid sequence homology and the conserved core archi-
tecture of most enzymes.3,4 The claims on the avail-
ability of various potent and selective ATP-competitive
inhibitors are usually based on data indicating the pre-
ferred inhibition of the enzymatic activity of the target
kinase over that of a panel of control kinases in cell-free
assays. Such selectivity data are independent of the cell
permeation properties of the compounds and, therefore,
of limited relevance for intracellular targets like PTKs
because of the lack of a direct correlation between the
ecacious concentration of compounds in a cell-free
and cellular biological systems. This issue is even more
important for PTK inhibitors as compared to inhibitors
of other intracellular targets since, due to the elevated
intracellular ATP concentration (1±5 mM), only sub-
stances with good cellular uptake and high intracellular
potency can be expected to e€ectively bind the ATP-
binding site.5 Consequently, predictions on the potential
of PTK inhibitors for adverse e€ects are dicult to
make based on selectivity data generated without taking
into account the cell permeation properties of the com-
pounds. The risk of systemic toxicity resulting from
their poor discrimination between the numerous PTKs
in vivo is of concern for the long-term application of such
compounds and often restricts their ®eld of application
to life saving indications.
These considerations taken together underline the
inherent complexity associated with the development of
PTK inhibitors and re¯ect the interdependency of
selectivity and cell-permeability. Therefore, the chemical
derivation of lead structures for achieving high potency
and enzyme selectivity should be preferentially carried
out with those devoid of cell permeability drawbacks.
Among the various physicochemical parameters eval-
uated as predictors of cell membrane permeability and,
hence, passive epithelial permeability, the polar part of
the molecular surface area (PSA) was shown to best
correlate with the experimentally determined membrane
permeability.6À8 The PSA value of a compound can be
estimated by computational methods while, for the
experimental determination of its permeability, the Caco-2
cell monolayer based system is the most commonly used
predictive assay.9
Pyrrolo[2,3-d]pyrimidines (1) have been described as
selective inhibitors of the highly homologous enzymes
belonging to the Src kinase family as compared to the
EGFR kinase, the substitution pattern of the aromatic
rings serving as a handle for selectivity modulation.10À13
Aiming at the discovery of sca€olds with reduced PSA
as potential leads for the development of selective PTK
inhibitors and capitalizing on the structure of 1,
*Corresponding author. Tel.: +41-61-324-6188; fax: +41-61-324-
3036; e-mail: christos.papageorgiou@pharma.novartis.com
pyrrolopyridones
2 were designed and their cell
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00022-1

1550
C. Papageorgiou et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1549±1552
Figure 1. Generic structures and key hydrogen bonding characteristics of pyrrolopyrimidines 1 and pyrrolopyridones 2.
membrane permeation aptitudes compared with those
of the corresponding pyrrolopyrimidine derivatives.
Care was taken to avoid large di€erences in the mole-
cular weights of the compounds of the two substance
classes and to integrate in 2 the key pharmacophore
features necessary for compound±PTK recognition,
namely a fused heterocycle containing a bidentate
hydrogen bond donor±acceptor (D±A) motif (Fig.
1).10,13,14 A comparative evaluation of the 3DPSA of
1a±c and 2a±c using their CORINA generate geometries
and Connoly surfaces8 indicated a better (mean 23%)
cell permeation potential for the pyrrolopyridone deri-
vatives (Table 1).
the methoxypyridine 4a led to the azaindole 6a15 in
good yields but could not be transformed into the cor-
responding pyridone due to the high stability of the
methoxy group to the various demethylation protocols
tried. Attracted by the successful use of the tert-but-
oxide moiety as a masking group for the generation of a
pyrrolo[3,2-b]pyridone derivative, we subsequently
aimed at the synthesis of the tert-butylpyridine 4b but
were unable to obtain this compound based on the
reported procedure.16 Finally, the trimethylsilylethanol
reagent was found to be satisfactory for our purposes.
Indeed, in the presence of tert-BuOK, it cleanly dis-
placed chlorine in 3 a€ording 4c in 65% yield.17 The
latter underwent a nucleophilic substitution of hydrogen
upon treatment with the anion of 4-chlorophenoxy-
acetonitrile16 to give 5c (mp 74±76 ꢀC), which was fur-
ther cyclized into the corresponding azaindole 6c (mp
121±122 ꢀC) by catalytic hydrogenation in 76% yield.
As a side product, the aniline corresponding to 5c was
isolated which could not be further transformed into 6c.
Iodination of the indole 3-position with N-iodosuccini-
mide a€orded quantitatively 7 (mp 99±100 ꢀC), the pre-
cursor for a Suzuki cross-coupling reaction. Using
phenylboronic acid and p-methoxyphenylboronic acid,
the 3-aryl azaindoles 8a (oil) and 8b (mp 117±118 ꢀC)
were obtained in 27 and 71% yield, respectively. The
alkylation of the indole with benzylbromide and p-
methoxybenzylbromide a€orded 9a±c which upon
treatment with tri¯uoroacetic acid gave the desired pyr-
idones 2a±c as amorphous solids. No cleavage of the
trimethylsilylethyl moiety took place in the presence of
an excess of TBAF in various solvents. Evidence that
2a±c adopted the lactam tautomeric form was obtained
from their IR spectra which showed a strong absorption
at 1634±1645 cmÀ1 independent of the experimental
conditions (KBr or CH2Cl2 solution).
Aiming at the generation of the necessary lactam moiety
in 2 from the corresponding 2-alkoxy pyridine, a readily
cleavable protecting group was required that could be
eciently introduced into the commercially available
2-chloro-5-nitro-pyridine 3 at the beginning of the
synthesis (Scheme 1). Preliminary investigations using
Table 1. Calculated and experimentally determined cell penetration
parameters
a
Compound
MW
3DPSA
(A2)
Pm
(10À5 cm/min)
Substrate
anityb
1a
2a
1b
2b
1c
2c
300.4
300.4
330.4
330.4
360.4
360.4
46.6
33.4
58.2
45.0
69.2
55.1
NDND
NDND
78.2
90.5
61.9
Passive
Passive
Passive
Passive
The computed polar surface area (PSA)8 and the cell
membrane permeability data (Pm) determined in the
human Caco-2 monolayer assay of compounds 1a±c18
and 2a±c are summarized in Table 1.
83.0
aMean value of triplicates.
bDetermined as the dependence of Pm on the transport direction (i.e.,
similar or dissimilar rate of transport between apical-to-basolateral and
basolateral-to-apical directions for the selected drug concentration).
ND, not determined.
In line with the reported ®ndings,6À8 an inverse rela-
tionship between calculated PSA values and Pm was
found; the smaller the PSA of a derivative the better it

C. Papageorgiou et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1549±1552
1551
permeates the cell monolayer. Plotting of the PSA ver-
sus the Pm values obtained results in a linear, highly
signi®cant correlation with r2=0.965 (Fig. 2). Indeed,
the 23% decrease in the PSA of 2b over that of 1b cor-
responds to a 15% improvement in the cell penetration
of 2b while the 20% di€erence between 2c and 1c leads
to a 33% improvement for 2c. No changes in the sub-
strate speci®cities were found between 1b,c and 2b,c.
Therefore, these results support the use of PSA as
predictor for the cell membrane penetration potential of
a series of structurally and conformationally similar
compounds and emphasize the overall contribution of
cheminformatics in the drug-design ®eld.
The consideration of the obtained structure±perme-
ability data in the light of the selectivity±permeability
issue in the PTK ®eld indicates that pyrrolopyridones 2
can potentially be used for the development of selective
Scheme 1. Synthesis of pyrrolopyridones: (a) HOCH2CH2Si(CH3)3 (1.0 equiv), tBuOK (1.3 equiv), DMF, rt; (b) tBuOK (2.2 equiv), 4-chloro-
phenoxyacetonitrile (1.1 equiv), THF, À10 ꢀC, 3 h; (c) Pd/C, H2, EtOH (0.1 M); (d) NIS (1.1 equiv), THF, rt, 2 h; (e) PdCl2(dppf) (0.02 equiv), 2 N
Na2CO3 (3.0 equiv), ArB(OH)2 (1.5 equiv), EtOH, re¯ux, 3 h; (f) NaH (1.1 equiv), ArCH2Br (1.5 equiv), THF, rt, 18 h; (g) CF3COOH (excess), 0 ꢀC,
30 min.

1552
C. Papageorgiou et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1549±1552
3. Al-Obeidi, F. A.; Wu, J. J.; Lam, K. S. Biopolymers 1998,
47, 197.
4. Traxler, P.; Furet, P. Pharmacol. Ther. 1999, 82, 195.
5. Lawrence, D. S.; Niu, J. Pharmacol. Ther. 1998, 77, 81.
6. van de Waterbeemd, H.; Camenisch, G.; Folkers, G.;
Raevsky, O. A. Quant. Struct.±Act. Relat. 1996, 15, 480.
7. Palm, K.; Luthman, K.; Ungell, A.-L.; Straundlund, G.;
Beigi, F.; Lundahl, P.; Artursson, P. J. Med. Chem. 1998, 41,
5382.
8. Ertl, P.; Rohde, B.; Selzer, P. J. Med. Chem. 2000, 43, 3714.
9. Artursson, P. J. Pharm. Sci. 1990, 79, 476.
10. Missbach, M.; Altmann, E.; Widler, L.; Susa, M.; Buch-
dunger, E.; Mett, H.; Meyer, T.; Green, J. Bioorg. Med. Chem.
Lett. 2000, 10, 945.
11. Showalter, H. D. H.; Kraker, A. J. Pharmacol. Ther. 1997,
76, 55.
12. Traxler, P.; Furet, P.; Mett, H.; Buchdunger, E.; Meyer,
T.; Lydon, N. J. Med. Chem. 1996, 38, 2285.
13. Traxler, P.; Green, J.; Mett, H.; Sequin, U.; Furet, P. J.
Med. Chem. 1999, 42, 1018.
14. Zhu, X.; Kim, J. L.; Newcomb, J. R.; Rose, P. E.; Stover,
D. R.; Toledo, L. M.; Zhao, H.; Morgenstern, K. A. Structure
1999, 7, 651.
15. Macor, J. E.; Newman, M. E. Heterocycles 1990, 37,
805.
16. Macor, J. E.; Burkhart, C. A.; Heym, J. H.; Ives, J. L.;
Lebel, L. A.; Newman, M. E.; Nielsen, J. A.; Ryan, K.;
Schulz, D. W. J. Med. Chem. 1990, 33, 2087.
17. Roberts, D. A.; Pearce, R. J.; Bradbury, R. H. Eur. Pat.
Appl. 487252, 1992; Chem. Abstr. Regno 135900-24-2.
18. Widler, L.; Green, J.; Missbach, M.; Susa, M.; Altmann, E.
Bioorg. Med. Chem. Lett. submitted for publication. For syn-
thetic entries in the pyrrolo[2,3-d]pyrimidine class, see refs 10±13.
Figure 2. Correlation between 3DPSA and P m data.
inhibitors that would allow probing the precise cellular
function of a target enzyme and lead to safe therapeutic
agents.
Acknowledgements
The authors thank Drs. Karl-Heinz Altmann and Eva
Altmann for samples of 1a±c.
References and Notes
1. Levitzki, A. Pharmacol. Ther. 1999, 82, 231.
2. Strawn, L. M.; Shawver, L. K. Expert Opin. Invest. Drugs
1998, 7, 553.