Finding the perfect spot for fluorine: Improving potency up to 40-fold during a rational fluorine scan of a Bruton's Tyrosine Kinase (BTK) inhibitor scaffold

Article abstract of DOI:10.1016/j.bmcl.2014.11.030

A rational fluorine scan based on co-crystal structures was explored to increase the potency of a series of selective BTK inhibitors. While fluorine substitution on a saturated bicyclic ring system yields no apparent benefit, the same operation on an unsaturated bicyclic ring can increase HWB activity by up to 40-fold. Comparison of co-crystal structures of parent molecules and fluorinated counterparts revealed the importance of placing fluorine at the optimal position to achieve favorable interactions with protein side chains.

Full text of DOI:10.1016/j.bmcl.2014.11.030

Bioorganic & Medicinal Chemistry Letters xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Finding the perfect spot for fluorine: Improving potency
up to 40-fold during a rational fluorine scan of a Bruton’s Tyrosine
Kinase (BTK) inhibitor scaffold
,
⇑
Yan Lou , Zachary K. Sweeney, Andreas Kuglstatter, Dana Davis, David M. Goldstein, Xiaochun Han,
Junbae Hong, Buelent Kocer, Rama K. Kondru, Renee Litman, Joel McIntosh, Keshab Sarma, Judy Suh,
Joshua Taygerly, Timothy D. Owens
Hoffmann-La Roche Inc., pRED, Pharma Research & Early Development, Small Molecule Research, Discovery Chemistry, 3431 Hillview Ave, Palo Alto, CA 94304, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A rational fluorine scan based on co-crystal structures was explored to increase the potency of a series of
selective BTK inhibitors. While fluorine substitution on a saturated bicyclic ring system yields no appar-
ent benefit, the same operation on an unsaturated bicyclic ring can increase HWB activity by up to 40-
fold. Comparison of co-crystal structures of parent molecules and fluorinated counterparts revealed
the importance of placing fluorine at the optimal position to achieve favorable interactions with protein
side chains.
Received 2 October 2014
Revised 5 November 2014
Accepted 7 November 2014
Available online xxxx
Keywords:
Bruton’s Tyrosine Kinase
BTK
Ó 2014 Published by Elsevier Ltd.
Structure-based drug design
Fluorine scan
Rational fluorine scan
c
F
The introduction of fluorine atoms into biologically active
molecules has enjoyed tremendous success in small molecule drug
discovery, and recent reviews have highlighted the large number of
fluorinated compounds approved by the FDA for medical use.1,2
The replacement of a CAH bond with a CAF bond often increases
metabolic stability and improves membrane permeability.3 Fur-
thermore, it has been appreciated recently that the introduction
of fluorine into small molecules can result in higher binding affin-
ities.4,5 To this end, ‘fluorine scans’ have been occasionally applied
in order to identify fluorine substitutions that increase inhibitor
affinity without significantly compromising physical properties.
Notable example of this approach include Diederich’s pioneering
work exploring the impact of fluorine groups on a series of throm-
bin inhibitors6 and the importance of fluorine substitution for
Merck’s DPP-4 inhibitors.7
N
N
O
HO
O
N
O
HO
O
N
N
HN
HN
N
N
N
a
b
N
N
O
1a
1b
O
Figure 1. RN486 (1a) and lead scaffold 1b.
we detail our efforts to utilize a rational fluorine scan to further
optimize this chemical scaffold starting from prototype molecules
such as 1b. This effort led to the identification of 1a and several
other inhibitors that were suitable candidates for advanced pre-
clinical evaluation. It also revealed an unusual structure–activity
relationship, in which a single fluorine substitution increased
in vitro efficacy by up to 40-fold.
We had previously described the optimization of a series of BTK
inhibitors for the treatment of Rheumatoid Arthritis (RA), which
led to the discovery of RN486 (1a, Fig. 1).8,9 In this contribution,
Compound 1b was described previously as a selective BTK
inhibitor with reasonable biological activity (human whole blood
⇑
Corresponding author.
E-mail addresses: ylou@nurix-inc.com, yan.lou@gmail.com (Y. Lou).
Current address: Nurix, Inc., Medicinal Chemistry, 1700 Owens Street, Suite 290,
San Francisco, CA 94158, United States.
(HWB) IC50 of 100 nM) and poor aqueous solubility (0.6/2/3
lg/
mL in water, SIF and SGF).9 In addition to improving the aqueous
http://dx.doi.org/10.1016/j.bmcl.2014.11.030
0960-894X/Ó 2014 Published by Elsevier Ltd.
Please cite this article in press as: Lou, Y.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.030

2
Y. Lou et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
solubility of our inhibitors, we also sought to further improve their
biological activity, as more potent inhibitors may be expected to
have lower effective doses. Inspired by the various examples in
Roche programs where fluorine scans have resulted in improved
binding affinity, we decided to adopt this strategy to examine
our chemical scaffold. Potentially favorable interactions between
fluorine and proteins include the following: (1) F interacting with
the LUMO of carbonyl groups6 (2) F substitutions strengthening
the interaction of adjacent aromatic C–H groups with protein res-
idues;10 (3) F forming electrostatic interactions with polarized
groups;11 (4) F engaging in hydrophobic interactions.12
N
N
O
N
HO
O
O
N
HO
O
N
HN
N
HN
N
R
HN
R2
O
R1
N
O
2a R1 = H, R2 = Et; HWB IC50 1600 nM
2b R1 = F, R2 = Me; HWB IC50 3700 nM
3a R = H; HWB IC50 190 nM
3b R = F; HWB IC50 200 nM
Given the availability of co-crystal structures and knowledge of
protein–ligand interactions involving fluorine, we decided to only
investigate fluorine substitutions that could result in favorable
interactions with the BTK protein. When the crystal structure of
1b bound to BTK (1.7 Å resolution) was examined, three vectors
were identified as potential candidates (Fig. 2).
R
R
N
N
CN
O
N
HO
O
O
N
HO
O
N
HN
HN
N
N
N
Three matched pairs of fluorinated and non-fluorinated ana-
logues (2a–b, 3a–b, 4a–b, Fig. 3) were prepared in order to explore
the potential of fluorination at these vectors to improve potency
(Fig. 3, note that in some cases the comparison was between ana-
logues with methyl and ethyl groups in the solvent exposed region,
which did not impact activity). These analogs were tested simulta-
neously in repeated human whole blood (HWB) assays for mean-
ingful differentiation as their potent biochemical activity was
beyond the dynamic range of our assay. As a consequence, it can-
not be ruled out that potential changes in protein binding upon
fluorination could contribute to activity difference in HWB assays,
despite that previous work has shown a good correlation between
biochemical and HWB potency for closely related compounds. Sub-
stitution at vector a (2a = 1600 nM, 2b = 3700 nM) or b (3a = 190 n
M, 3b = 200 nM) did not result in any noticeable change in the
HWB potency, whereas fluorination at vector c gave 7-fold boost
in activity (4a = 280 nM, 4b = 40 nM). It should be noted that the
matched pair comparison for vector c was initially performed on
the unsaturated isoquinolone scaffold instead of the dihydroisoqu-
inolone scaffold.
N
N
O
O
5a R = H; HWB IC50 60 nM
5b R = F; HWB IC50 50 nM
4a R = H; HWB IC50 280 nM
4b R = F; HWB IC50 40 nM
R
R
O
N
HO
O
O
N
HO
O
N
HN
N
N
HN
N
N
N
N
N
R'
N
R'
7a R1 = H, R2 = Et; HWB IC50 38 nM
7b R1 = F, R2 = Et; HWB IC50 2 nM
6a R1 = H, R2 = Et; HWB IC50 1600 nM
6b R1 = F, R2 = Et; HWB IC50 4 nM
6c R1 = Cl, R2 = Me; HWB IC50 400 nM
8
R1 = F, R2 = Me; HWB IC50 3 nM
Encouraged by the initial data resulting from fluorination at
position c, we prepared inhibitor 5b, which features the fluorine
substitution on the partially saturated dihydroisoquinolone scaf-
fold. To our surprise, the fluorinated analogue 5b has a minimal
increase in HWB activity (5a = 60 nM, 5b = 50 nM). We hypothe-
sized that the activity increase for 4b might be specific to the iso-
quinolone ring system, although we did not have a rationale at the
Figure 3. BTK inhibitor fluorinated and non-fluorinated matched pairs.
time. Therefore, 6a and 6b on the isoquinolone ring scaffold with a
cyclopropyl group in place of the dimethylamino substituent were
prepared. Astonishingly, 6b was reproducibly found to be approx-
imately 40-fold more potent than 6a in the HWB assay
Figure 2. Co-crystal structure of compound 1b and BTK (pdb code: 4RFY). Vectors explored with fluorine substitution are highlighted with black arrows.
Please cite this article in press as: Lou, Y.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.030

Y. Lou et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
3
(6a = 160 nM, 6b = 4 nM). It should be noted that this increase in
potency was significantly more pronounced than of the potency
improvement imparted by fluorination in the 4a-b matched pair.
In order to understand the dramatically different impact of fluo-
rination on these closely related scaffolds, we obtained co-crystal
structures with BTK for 4b (1.17 Å resolution), 5b (2.5 Å) and 6b9
(1.95 Å) (Fig. 4). An overlay of the structures showed that the fluo-
rine atom in 6b sits most deeply in the binding pocket toward
F413, while the fluorine in 5b lies furthest away from this residue.
The electronegative fluorine atom in 6b is within van-der-Waals
distance of the primary amine of K430, a conserved water mole-
cule, and the aromatic hydrogen at the ortho position of F413.
On the other hand, the fluorine atom in 5b is relatively distant from
these features and cannot form such favorable interactions. The
fluorine atom in 4b is within bonding distance of these groups
(i.e., F413), but does not seem to be optimally positioned. Overall,
Figure 4. (A) Structure of 6b highlighting the interactions of the isoquinolone fluorine and carbonyl groups. (B) Overlay of the structures of 4b (pdb code: 4RFZ), 5b (pdb
code: 4RG0) and 6b.
Please cite this article in press as: Lou, Y.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.030

4
Y. Lou et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
these structural observations and the apparent requirement for
multiple simultaneous interactions of the fluorine group are con-
sistent with our biological data.
The favorable interactions and impact of halogenation at this
position appear to be quite unique to fluorine. Chlorine substitu-
tion negatively affected the biological activity of analogs in this
series, and 6c was found to be one hundred fold less active than
the fluorinated counterpart 6b.
This rationale suggests that other unsaturated bicyclic systems
could also benefit from fluorine substitution at vector c. To test this
hypothesis, pair 7a-b were prepared and profiled in the HWB
assay. We were pleased to discover that the fluorinated analogue
7b was approximately 20-fold more active than 7a (HWB IC50 of
7a = 38 nM, 7b = 2 nM). Further optimization and profiling led to
the identification of the methyl analogue 8 (HWB IC50 of 3 nM)
as a new, exceptionally potent BTK inhibitor with desirable proper-
ties for preclinical studies.
The remarkable impact of fluorination on series 6 relative to
series 5 deserved further comment. Recent publications have high-
lighted the role of synergy in structure–activity relationships in
various systems, and we believe that this is an interesting example
of substituent synergy13–16 involving a fluorine atom. The hydro-
phobic cyclopropyl and dimethylcyano groups at the para position
are anchored in a well-defined liphophilic pocket that strongly
contributes to potency of these analogs. In the case of 6b, the iso-
quinolone ring system is situated to place the both the cyclopropyl
group and fluorine atom in an optimal position. Addition of the
fluorine atom in this series does not require any reduction of the
optimal hydrophobic interactions of the cyclopropyl group, and
the full, mutually reinforcing, impact of the fluorine atom can be
realized. In the case of scaffold 5, it does not seem to be possible
to benefit from the positive interactions of the fluorine group with
the protein while preserving the interactions of the dimethylcyano
functionality. While other effects of the fluorine group (e.g., influ-
encing the solvation of the adjacent carbonyl group) are clearly
possible, these explanations do not appear to be as consistent with
overall structure–potency relationships observed in this series of
inhibitors.
Additional research teams appear to have tested and recognized
the importance of this fluorophilic vector c. In particular, patent
applications from several research groups included incorporation
of fluorine group on related scaffolds that would potentially place
fluorine atoms at this hotspot.17–19
This work demonstrates that a rational fluorine scan can be
used as a strategy to improve biological activity, even on synthet-
ically challenging scaffolds. Careful examination of co-crystal
structures reduced the number of novel fluorinated analogues pur-
sued in this effort. The recent significant advances in synthetic
fluorine chemistry facilitate employment of this strategy in medic-
inal chemistry.20–22 An apparent synergetic substitution of a bicy-
clic ring system resulted in the identification of matched pairs in
which the fluorinated analogs were dramatically more potent than
their non-fluorinated analogs. These findings continue to impact
discovery efforts targeting inhibitors of Bruton’s Tyrosine Kinase
(BTK).
The BTK inhibitors described were synthesized in a convergent
manner as shown in Scheme 1. A C–N coupling or aromatic substi-
tution of an isoquinolone, dehydroquinolone, or pthalazinone to a
dihalobenzaldehyde or protected dibromobenzyl alcohol was fol-
lowed by a Suzuki–Miyaura coupling as described previously.
Reduction or deprotection afforded the final inhibitors.23,24 For
example, the synthesis of 8 proceeded from aromatic substitution
of pthalazinone 9 with 2-chloro-6-fluorobenzaldehyde followed by
reduction with NaBH4 to produce alcohol 10. This intermediate
could be coupled with boronic ester 11 to give 8 as final product.
Phthalazinone 9 can be prepared in a four-step sequence start-
ing from commercial chemical 12. Lithiation and quenching with
DMF generated an aldehyde that could be converted to the corre-
sponding acetal 13. An additional directed deprotonation followed
by addition of N-flurorobenzenesulfonimide generated the desired
flouroaromatic. Simultaneous removal of the acetal and oxazole
protecting groups provided 14, which could be converted to 9 by
treatment with hydrazine hydrate (Scheme 2).
Acknowledgment
This work was supported exclusively by Hoffmann-La Roche.
References and notes
1. Ojima, I. J. Org. Chem. 2013, 78, 6358.
2. Hagmann, W. K. J. Med. Chem. 2008, 51, 4359.
3. Fluorine in Pharmaceutical and Medicinal Chemistry: From Biophysical Aspects to
Clinical Applications; Gouverneur, V., Muller, K., Eds.; Imperial College Press:
London, 2012.
Scheme 1. Synthesis of inhibitor 7b. Reagents and conditons: (a) 2-chloro-6-
fluorobenzaldehyde, Cs2CO3, 60 °C, 4 h, 68.2%. (b) NaBH4, i-PrOH, CH2Cl2, 4 °C, 0.5 h,
65.5%. (c) 11, K2CO3, PCy3, Pd(dba)2, dioxane, water, 88 °C, 1.5 h, 83.5%.
Scheme 2. Preparation of pthalazinone 9. Reagents and conditions: (a) i-PrMgCl, n-BuLi, DMF, THF, À10 °C, (b) trimethyl orthoformate, toluenesulfonic acid, rt, 1 h, 98.5% in
two steps. (c) sec-BuLi, CO2, THF, À70 °C to 10 °C, 55.3%. (d) hydrazine hydrate, acetic acid, ethanol, 80 °C, 86%.
Please cite this article in press as: Lou, Y.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.030

Y. Lou et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
5
4. Müller, K.; Faeh, C.; Diederich, F. Science 1881, 2007, 317.
5. Meanwell, N. A. J. Med. Chem. 2011, 54, 2529.
6. Olsen, J. A.; Banner, D. W.; Seiler, P.; Wagner, B.; Tschopp, T.; Obst-Sander, U.;
Kansy, M.; Muller, K.; Diederich, F. ChemBioChem 2004, 5, 666.
12. Clark, M. P.; Laughlin, S. K.; Laufersweiler, M. J.; Bookland, R. G.; Brugel, T. A.;
Golebiowski, A.; Sabat, M. P.; Townes, J. A.; VanRens, J. C.; Djung, J. F.; Natchus,
M. G.; De, B.; Hsieh, L. C.; Xu, S. C.; Walter, R. L.; Mekel, M. J.; Heitmeyer, S. A.;
Brown, K. K.; Juergens, K.; Taiwo, Y. O.; Janusz, M. J. J. Med. Chem. 2004, 47,
2724.
13. Baum, B.; Muley, L.; Smolinski, M.; Heine, A.; Hangauer, D.; Klebe, G. J. Mol. Biol.
2010, 1042, 397.
14. Muley, L.; Baum, B.; Smolinski, M.; Freindorf, M.; Heine, A.; Klebe, G.;
Hangauer, D. G. J. Med. Chem. 2010, 53, 2126.
7. Kim, D.; Wang, L.; Beconi, M.; Eiermann, G. J.; Fisher, M. H.; He, H.; Hickey, G. J.;
Kowalchick, J. E.; Leiting, B.; Lyons, K.; Marsilio, F.; McCann, M. E.; Patel, R. A.;
Petrov, A.; Scapin, G.; Patel, S. B.; Roy, R. S.; Wu, J. K.; Wyvratt, M. J.; Zhang, B.
B.; Zhu, L.; Thornberry, N. A.; Weber, A. E. J. Med. Chem. 2005, 48, 141.
8. Xu, D.; Kim, Y.; Postelnek, J.; Vu, M. D.; Hu, D. Q.; Liao, C.; Bradshaw, M.; Hsu, J.;
Zhang, J.; Pashine, A.; Srinivasan, D.; Woods, J.; Levin, A.; O’Mahony, A.; Owens,
T. D.; Lou, Y.; Hill, R. J.; Narula, S.; DeMartino, J.; Fine, J. S. J. Pharmacol. Exp. Ther.
2012, 341, 90.
9. Lou, Y.; Han, X.; Kuglstatter, A.; Kondru, R. K.; Sweeney, Z. K.; Soth, M.;
McIntosh, J.; Litman, R.; Suh, J.; Kocer, B.; Davis, D.; Park, J.; Frauchiger, S.;
Dewdney, N.; Zecic, H.; Taygerly, J. P.; Sarma, K.; Hong, J.; Hill, R. J.; Gabriel, T.;
Goldstein, D. M.; Owens, T. D. J. Med. Chem. 2014. Ahead of Print.
10. Sweeney, Z. K.; Dunn, J. P.; Li, Y.; Heilek, G.; Dunten, P.; Elworthy, T. R.; Han, X.;
Harris, S. F.; Hirschfeld, D. R.; Hogg, J. H.; Huber, W.; Kaiser, A. C.; Kertesz, D. J.;
Kim, W.; Mirzadegan, T.; Roepel, M. G.; Saito, Y. D.; Silva, T. M. P. C.; Swallow,
S.; Tracy, J. L.; Villasenor, A.; Vora, H.; Zhou, A. S.; Klumpp, K. Bioorg. Med. Chem.
Lett. 2008, 18, 4352.
15. Jencks, W. P. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4046.
16. Peters, J.-U.; Weber, S.; Kritter, S.; Weiss, P.; Wallier, A.; Boehringer, M.;
Hennig, M.; Kuhn, B.; Loeffler, B.-M. Bioorg. Med. Chem. Lett. 2004, 14, 1491.
17. Kawahata, W.; Asami, T.; Sawa, M.; Kiyoi, T. WO 2013/161848.
18. Heng, R.; Hoegenauer, E.; Koch, G.; Pulz, R. A.; Vulpetti, A.; Waelchli, R. WO
2013/008095.
19. Crawford, J. J.; Ortwine, D. F.; Wei, B.; Young, W. B. WO 2013/067264.
20. Lee, E.; Kamlet, A. S.; Powers, D. C.; Neumann, C. N.; Boursalian, G. B.; Furuya,
T.; Choi, D. C.; Hooker, J. M.; Ritter, T. A. Science 2011, 334, 639.
21. Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011, 473, 470.
22. Tredwell, M.; Gouverneur, V. Angew. Chem., Int. Ed. 2012, 51, 11426.
23. Hong, J.; Davidson, J. P.; Jin, Q.; Lee, G. R.; Matchett, M.; Brien, E. O.; Welch, M.;
Bingenheimer, B.; Sarma, K. Org. Process Res. Dev. 2014, 18, 228.
24. Berthel, S.; Firooznia, F.; Fishlock, D.; Hong, J.-B.; Lou, Y.; Lucas, M.; Owens, T.
D.; Sarma, K.; Sweeney, Z. K.; Taygerly, J. P. US 2010/0222325.
11. Bohm, H. J.; Banner, D.; Bendels, S.; Kansy, M.; Kuhn, B.; Muller, K.; Obst-
Sander, U.; Stahl, M. ChemBioChem 2004, 5, 637.
Please cite this article in press as: Lou, Y.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.11.030