Isoform selective PLD inhibition by novel, chiral 2,8-diazaspiro[4.5]decan-1-one derivatives

Article abstract of DOI:10.1016/j.bmcl.2018.10.033

This letter describes the on-going SAR efforts to develop PLD1, PLD2 and dual PLD1/2 inhibitors with improved physiochemical and disposition properties as well as securing intellectual property position. Previous PLD inhibitors, based on a triazaspiro[4.5

Full text of DOI:10.1016/j.bmcl.2018.10.033

Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Isoform selective PLD inhibition by novel, chiral 2,8-diazaspiro[4.5]decan-
1-one derivatives
Alex G. Watersona,e, Sarah A. Scotta, Nathan R. Kettc, Anna L. Blobauma,b, H. Alex Browna,c,d,e,f
,
⁎
Craig W. Lindsleya,b,c,d,e,
a
b
c
d
Department of Pharmacology, Vanderbilt University Medical Center, Nashville, TN 37232, USA
Vanderbilt Center for Neuroscience Drug Discovery, Vanderbilt University Medical Center, Nashville, TN 37232, USA
Department of Biochemistry, Vanderbilt University, Nashville, TN 37232, USA
Department of Chemistry, Vanderbilt University, Nashville, TN 37232, USA
e
Vanderbilt Institute of Chemical Biology, Vanderbilt University/Vanderbilt University Medical Center, Nashville, TN 37232, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
This letter describes the on-going SAR efforts to develop PLD1, PLD2 and dual PLD1/2 inhibitors with improved
physiochemical and disposition properties as well as securing intellectual property position. Previous PLD in-
hibitors, based on a triazaspiro[4.5]decanone core proved to be highly selective PLD2 inhibitors, but with low
plasma free fraction (rat, human fu < 0.03), high predicted hepatic clearance (rat CLhep > 65 mL/min/kg) and
very short half-lives in vivo (t1/2 < 0.15 h). Removal of a nitrogen atom from this core generated a 2,8-dia-
zaspiro[4.5]decanone core, harboring a new chiral center, as well as increased sp3 character. This new core
demonstrated enantioselective inhibition of the individual PLD isoforms, enhanced free fraction (rat, human
fu < 0.13), engendered moderate predicted hepatic clearance (rat CLhep ∼ 43 mL/min/kg), improved half-lives
in vivo (t1/2 > 3 h), and led to the first issued US patent claiming composition of matter for small molecule PLD
inhibitors.
PLD
Phospholipase D
Isoform
Inhibitors
Phosphodiesterase
Phospholipase D (PLD) exists as two isoforms in mammals, coined
PLD1 and PLD2, and is considered to be one of the major sources of
the signal-activated second messenger phosphatidic acid (PtdOH).
Recently, the development of isoform-selective PLD inhibitors, in
combination with biochemical and molecular genetics studies, have
suggested that PLD enzymes in mammalian cells and pathogenic or-
ganisms represent therapeutic targets for several human diseases.1–5
Work from multiple labs have demonstrated that PLD dysfunction
and/or overexpression can be modulated by small molecule, allosteric
inhibitors 1–6 (Fig. 1) with efficacy in preclinical models relevant to
oncology, viral infections and CNS disorders.1–17 Despite the progress
achieved in the past decade, small molecule tools with DMPK profiles
suitable for robust in vivo proof of concept studies have remained
elusive, and there have been no PLD inhibitor composition of matter
patents (issued or published applications), further hindering transla-
tional potential for the target.1–22 In this Letter, we describe the dis-
covery and development of a new series of potent and selective PLD2
and dual PLD1/2 inhibitors from a series of 2,8-diazaspiro[4.5]de-
canones, that display enantioselective PLD isoform inhibition, provide
improved DMPK profiles and enabled a robust intellectual property
(IP) position.23
For several years, the triazaspiro[4.5]decanone core of PLD in-
hibitors 3–6 was the major focus of optimization efforts, as both PLD2-
selective and dual PLD1/2 inhibitors could be realized within this core
based on chiral α-methyl moieties along the ethyl linker chain (e.g.,
5).7–14 Despite providing both peripherally restricted and highly CNS
penetrant (brain-to-plasma Kps 0.75–1.48) PLD inhibitors, this series
uniformly afforded compounds displaying high plasma protein binding
(rat and human fu < 0.03), high predicted hepatic clearance (rat
CLhep > 65 mL/min/kg) and very short half-lives in vivo (t1/
< 0.15 h).7–14 Work within a well-known SERM chemotype afforded
2
only weak PLD inhibitors, but they proved to be universal (mammalian
and bacterial).15 Thus, the ideal in vivo proof of concept tools were
lacking, and the program also needed novel chemical matter to advance
the target. Our strategy to potentially accomplish both goals focused on
deletion of the N-aryl nitrogen atom in 7, and replacement with a ste-
reogenic carbon center, as in 8 (Fig. 2). We were also interested to see
the impact of a chiral α-methyl moiety on the linker with regards to
⁎
Corresponding author at: Department of Pharmacology, Vanderbilt University Medical Center, Nashville, TN 37232, USA.
E-mail address: craig.lindsley@vanderbilt.edu (C.W. Lindsley).
f
This manuscript is in dedication to Prof. H. Alex Brown, on the anniversary of his death.
https://doi.org/10.1016/j.bmcl.2018.10.033
Received 31 July 2018; Received in revised form 18 October 2018; Accepted 20 October 2018
0960-894X/©2018ElsevierLtd.Allrightsreserved.
Pleasecitethisarticleas:Waterson,A.G.,Bioorganic&MedicinalChemistryLetters,https://doi.org/10.1016/j.bmcl.2018.10.033

A.G. Waterson et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Fig. 1. Halopemide (1) was the progenitor of all PLD inhibitors, and a dual PLD1/2 inhibitor. Optimization efforts led to isoform-selective PLD inhibitors: 2 (1700-
fold PLD1 selective), 3 (75-fold PLD2 selective), 4 (53-fold PLD2 selective), 5 (dual PLD1/2 inhibitor) and 6 (> 80-fold PLD2 selective).
Fig. 2. Envisioned optimization strategy for 7 to improve disposition and structural novelty with analogs 8.
PLD isoform inhibition (e.g., the ‘magic methyl’ effect)24 in combina-
tion with the phenyl bearing stereocenter, and if previous SAR for 3–6
would be maintained.7–14
(95% hexanes and 5% isopropyl alcohol) to afford (S)-12 and (R)-12,
and in agreement with the work of Thomas.25
Once in hand, (S)-12 and (R)-12 were rapidly elaborated into pu-
tative PLD inhibitors (S)-14 and (R)-14 in a two-step sequence (Scheme
2).7–14 Reductive amination with tert-butyl 2-oxoethylcarbamate and
(S)-12 and (R)-12 provided (S)-13 and (R)-13, respectively in 79–85%
yield after Boc deprotection. Acylation of the primary amine with the
requisite acid chloride delivered (S)-14 and (R)-14 in yields ranging
from 58 to 95%. To assess the impact of the (S)-methyl group on the
side chain, a similar two step route was employed (Scheme 3) wherein
(S)-tert-butyl 1-oxo-2-propanylcarbamate is substituted for tert-butyl 2-
oxoethylcarbamate with comparable yields at each step.
Fortunately, Thomas and co-workers form Roche had previously
reported on a racemic synthesis of a 4-fluorophenyl 2,8-diazaspiro[4.5]
decanone core,25 and had provided details for the assignment of the
individual enantiomers. Thus, we elected to follow their route and first
evaluate both enantiomers of this core for PLD inhibition. As prescribed
(Scheme 1), deprotonation of 9 with LDA followed by trapping with
trans-1-fluoro-4-(2-nitrovinyl)benzene afforded the desired Michael
adduct 10 in 90% yield. Reduction of the nitro moiety with Ra-Ni af-
forded primary amine 11 in 86% yield. Finally, closure to the racemic
spiro-γ-lactam ( )-12 was facilitated by refluxing 11 in toluene
overnight; however, some deprotection of the Boc was noted, so the
mixture was reprotected to afford Boc-protected ( )-12 in 83% yield.
Preparative chiral separation (2–3 g batches) was performed on a Gilson
215 Preparative HPLC with pumps capable of 200 mL/min using a
Chiraltech OJ 20u, 50 × 250 mm column at a flow rate of 177 mL/min
As shown in Table 1, clear enantioselective inhibition of the PLD
isoforms was noted, with all (R)-14 analogs typically > 10-fold weaker
at inhibiting both PLD1 and PLD2 versus the (S)-14 analogs. For ex-
ample, (S)-14a was a 70 nM PLD 2 inhibitor with ∼17-fold selectivity
versus PLD1, while (R)-14 was > 10-fold weaker at PLD2
(IC50 = 810 nM). This SAR trend was conserved across all enantiomeric
2

A.G. Waterson et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 1
Structures and PLD activities of (S)-and (R)-14 and (S)-and (R)-16.
Compound
R
PLD1 IC50 (nM)a
1200
PLD2 IC50 (nM)a
70
(S)-14a
(R)-14a
(S)-16a
(R)-16a
6400
16
810
Scheme 1. Synthesis of 4-fluorophenyl 2,8-diazaspiro[4.5]decanone cores 12.a
aReagents and conditions. (a) trans-1-fluoro-4-(2-nitrovinyl)benzene, LDA, THF,
−40 °C, 2 h, 90%; (b) H2, Ra-Ni, EtOH, 50 °C, 20 h, 86%; (c) i) toluene, reflux;
ii) Boc2O, DIPEA, dioxane-H2O, 0 °C, 2 h, 83%; d) i) Gilson 215 Preparative
150
320
4000
HPLC with pumps capable of 200 mL/min using
a Chiraltech OJ 20u,
50 × 250 mm column at a flow rate of 177 mL/min (95% hexanes and 5%
(S)-14b
(R)-14b
(S)-16b
(R)-16b
21
320
1
2.5
43
isopropyl alcohol. (S)-12 is first eluting peak; ii) TFA, CH2Cl2, rt, 5 h, 95%.
12
16
140
(S)-14c
(R)-14c
(S)-16c
(R)-16c
(S)-14d
(R)-14d
(S)-16d
(R)-16d
(S)-16e
1350
12,000
25
60
450
115
700
220
3175
210
2800
26
Scheme 2. Synthesis of 2,8-diazaspiro[4.5]decanone-based PLD inhibitors 14.a
aReagents and conditions: (a) i) tert-butyl 2-oxoethylcarbamate, MPB(OAc)3H,
DCM, rt; ii) 4.0 M HCl/dioxane, DCM, MeOH, rt, 79–85%; (b) RCOCl, DIEA,
DMF, rt, 58–95%.
200
5500
> 20,000
250
1700
820
Scheme 3. Synthesis of 2,8-diazaspiro[4.5]decanone-based PLD inhibitors 16.a
aReagents and conditions: (a) i) (S)-tert-butyl 1-oxo-2-propanylcarbamate, MPB
(OAc)3H, DCM, rt; ii) 4.0 M HCl/dioxane, DCM, MeOH, rt, 70–82%; (b) RCOCl,
DIEA, DMF, rt, 55–92%.
bPLD2 cellular assay in HEK293-gfp-PLD2 cells. IC50 values are the average of
three independent measures (n = 3).
a
Cellular PLD1 assay in Calu-1 cells.
pairs of (S)- and (R)-14 derivatives. Thus, further optimization will
maintain the (S)-configuration at the aryl bearing stereogenic center.
The SAR trend within 3–6 by which a chiral (S)-methyl moiety on the
linker not only improved PLD activity, but also enhanced PLD1 in-
hibitory activity was retained in the diazaspiro[4.5]decanone series.7–14
For example, addition of the (S)-methyl moiety as in (S)-16a increased
PLD1 activity ∼9-fold over (S)-14a, providing an ∼10-fold PLD1-
prefering inhibitor. With the analogous (R)-16a, activity was increased
at PLD1, but with a significant loss of inhibitory activity at PLD2. This
enantiospecific modulation of PLD activity was also conserved within
the entire series, with the (S)-methyl group engendering enhanced
PLD1 inhibition with a concomitant loss of PLD2 activity. Notable ex-
amples include (S)-16b, a potent dual PLD1 and PLD2 inhibitor (IC50s
of 1 nM and 12 nM, respectively) and (S)-14b, a 25-fold PLD2 pre-
ferring inhibitor, which converts to a dual PLD1/PLD2 inhibitor (S)-16d
upon incorporation of the (S)-methyl group. These data proved very
exciting, confirming that a single chemotype can provide PLD1-pre-
ferring, PLD2-preferring and dual PLD1/PLD2 inhibitors (Fig. 3) based
solely on the inclusion of either one or two easily established and
constructed chiral centers. These data also supported the first, and only
to date, issued U.S. composition of matter patent for PLD inhibitors.23
We next evaluated if these novel PLD inhibitors offer advantages in
terms of disposition relative to the earlier generation 2–6, which are
plagued by high plasma protein binding (rat and human fu < 0.03),
high predicted hepatic clearance (rat CLhep > 65 mL/min/kg) and very
short half-lives in vivo (t1/2 < 0.15 h). As shown in Table 2, analogs 14
and 16 uniformly showed improved free fraction (in both rat and
human plasma, fus > 0.10) and (S)-14b and (S)-14d displayed im-
proved predicted hepatic clearance in rat (41.2 and 47.2 mL/min/kg,
respectively) and human (10.7 and 10.1 mL/min/kg, respectively).
3

A.G. Waterson et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Fig. 3. Representative PLD1 (Calu-1) and PLD2 (HEK293-PLD2) concentration-response-curves for (S)-14b, a PLD2-preferring inhibitor and (S)-16d, a dual PLD1/
PLD2 inhibitor.
clearance (rat CLhep ∼ 43 mL/min/kg), and improved half-lives in vivo
(t1/2 > 3 h). Further, this novel chemical matter led to the first issued
US patent claiming composition of matter for small molecule PLD in-
hibitors. While this new series displayed robust and tractable SAR, the
extremely high volume of distribution in vivo precluded further ad-
vancement. Follow-up efforts continue, and work is in progress to de-
velop optimal in vivo tool compounds that selectively inhibit either
PLD1 or PLD2 as well as clinical candidates.
Table 2
In vitro DMPK data for analogs (S)-14 and (S)-16.
Property
(S)-14a (S)-14b (S)-14c (S)-14d (S)-16b (S)-16e
MW (dal)
cLogP
419.5
2.87
61.4
434.2
2.29
73.4
446.2
2.68
73.8
413.4
2.84
92.9
448.5
2.73
73.4
459.5
3.31
61.4
TPSA (Å2)
In vitro PK parameters
Rat CLHEP (mL/min/
kg)
60.7
61.7
41.2
10.7
63.8
17.2
47.2
10.1
53.4
14.0
61.6
17.4
Acknowledgments
Human CLHEP (mL/
min/kg)
Rat fu plasma
Hum fu plasma
Rat fu brain
0.160
0.138
0.013
0.198
0.124
0.004
0.314
0.122
0.026
0.246
0.108
0.027
0.130
0.136
0.004
0.061
0.045
0.003
Dr. Lindsley thanks William K. Warren, Jr. who funded the William
K. Warren, Jr. Chair in Medicine (to C.W.L.) and the William K. Warren
Foundation for support of our programs.
References
Based on the improved predicted hepatic clearance of (S)-14b and (S)-
14d, we evaluated both in a rat IV PK cassette study (0.25 mg/kg per
compound in EtOH/PEG400/DMSO vehicle). Both compounds dis-
played suprahepatic clearance in vivo (CLps of 340 and 250 mL/min/kg
for (S)-14b and (S)-14d, respectively), and thus a lack of an in vitro:in
vivo correlation (IVIVC). However, both compounds showed improved
half-lives (t1/2s of 3.63 and 3.2 h, for (S)-14b and (S)-14d, respectively)
driven by large volumes of distribution at steady state (Vss of 80 and
56 L/kg for (S)-14b and (S)-14d, respectively). Thus, despite attractive
cLogPs, these new inhibitors were clearly deporting in some tissue and
being reabsorbed into plasma, providing a sustained half-life and mean
residence time. Based on these data and associated concerns regarding
accumulation with chronic dosing and the subsequent potential for tox
findings, this series could not be further progressed. Still, this new series
of 2,8-diazaspiro[4.5]decanones provided improvements in multiple
domains over historical PLD inhibitors 2–6.
1. Brown HA, Thomas PG, Lindsley CW. Nat Rev Drug Discov. 2017;16:351–367.
2. Selvy PE, Lavieri R, Lindsley CW, et al. Chem Rev. 2011;111:6064–6119.
3. Brown HA, Henage LG, Preininger AM, et al. Methods Enzymol. 2007;434:49–87.
4. Bruntz RC, Lindsley CW, Brown HA. Pharmacol Rev. 2014;66:1033–1079.
5. Scott SA, Matthews TP, Ivanova PT, et al. Biochim Biophys Mol Cell Biol Lipids.
2014;1841:1060–1084.
6. Monovich L, Mugrage B, Quadros E, et al. Bioorg Med Chem Lett. 2007;17:2310–2311.
7. Scott SA, Selvy PE, Buck JR, et al. Nat Chem Biol. 2009;5:108–117.
8. Lewis JA, Scott SA, Lavieri R, et al. Bioorg Med Chem Lett. 2009;19:1916–1919.
9. Lavieri R, Lewis JA, Scott SA, et al. Bioorg Med Chem Lett. 2009;19:2240–2244.
10. Lavieri R, Scott SA, Selvy PE, et al. J Med Chem. 2010;53:6706–6719.
11. O’Reilly MC, Scott SA, Brown KA, et al. J Med Chem. 2013;56:2695–2699.
12. O’Reilly MC, Oguin III TH, Scott SA, et al. ChemMedChem. 2014;9:2633–2637.
13. Bruntz RC, Taylor HE, Lindsley CW, et al. J Biol Chem. 2014;289:600–616.
14. Oguin T, Sharma S, Stuart AD, et al. J Biol Chem. 2014;289:25405–25417.
15. Scott SA, Spencer CT, O’Reilly MC, et al. ACS Chem Biol. 2015;10:421–432.
16. Noble AR, Maitland NJ, Berney DM, et al. Br J Cancer. 2018;118:189–199.
17. Nelson RK, Ya-Ping J, Gadbery J, et al. Sci Rep. 2017;7:9112.
18. Stieglitz KA. Curr Drug Discovery Technol. 2018;15:81–93.
In summary, we employed a scaffold-hopping strategy, moving from
the classic triazaspiro[4.5]decanone-based PLD inhibitors to a novel
2,8-diazaspiro[4.5]decanone core, harboring a stereogenic center. This
new core demonstrated enantioselective inhibition of the individual
PLD isoforms (providing access to PLD1, PLD2 and dual PLD1/PLD2
inhibitors in the same scaffold, along with enhanced plasma free frac-
tion (rat, human fu < 0.13), engendered moderate predicted hepatic
19. Song HI, Yoon MS. Sci Rep. 2016;22:36968.
20. Henkels KM, Muppani NR, Gomez-Cambronero J. PLoS One. 2016;7:e20166663.
21. Zhou G, Yu L, Yang W, et al. Mediators Inflamm. 2016:2543070.
22. Babenko NA, Kharchenko VS. Int J Endocrinol. 2015:7948838.
23. Brown HA, Lindsley CW, Waterson AG, et al. US 9,127,005, September 8; 2015.
24. Schonherr H, Cernak T. Angew Chem Int Ed. 2013;52:12256–12267.
25. Kraft EA, Kurt A, Maier A, et al. Synthesis. 2005;19:3245–3252.
4