Development of 5-Hydroxypyrazole Derivatives as Reversible Inhibitors of Lysine Specific

Demethylase 1

Daniel P. Moulda, Ulf Brembergc, Allan M. Jordana, Matthis Geitmannc, Alba Maiques-Diazb, Alison

E. McGonaglea, Helen F. Smalla, Tim C. P. Somervailleb, and Donald Ogilviea

aDrug Discovery Unit, Cancer Research UK Manchester Institute, University of Manchester,

Wilmslow Road, Manchester, M20 4BX, UK

bLeukaemia Biology Group, Cancer Research UK Manchester Institute, University of Manchester,

Wilmslow Road, Manchester, M20 4BX, UK

cBeactica AB, Uppsala Business Park, Virdings allé 2, 75450, Uppsala, SE

ABSTRACT

A series of reversible inhibitors of lysine specific demethylase 1 (LSD1) with a 5-hydroxypyrazole

scaffold have been developed from compound 7, which was identified from the patent literature.

Surface plasmon resonance (SPR) and biochemical analysis showed it to be a reversible LSD1

inhibitor with an IC50 value of 0.23 µM. Optimisation of this compound by rational design afforded

compounds with Kd values of <10 nM. In human THP-1 cells, these compounds were found to

upregulate the expression of the surrogate cellular biomarker CD86. Compound 11p was found to

have moderate oral bioavailability in mice suggesting its potential for use as an in vivo tool

compound.

KEYWORDS:

epigenetics; LSD1; KDM1A; reversible inhibitor; stem cell differentiation; cancer therapy; epigenetic

therapy; acute myeloid leukaemia

1

The many functions of the histone demethylase lysine specific demethylase 1 (LSD1) have unravelled

over the past decade to reveal a complex network of interactions with a number of protein

complexes,1, 2 transcription factors,3-5 and nucleosomes6 in a multitude of cell types. In cancer, Harris

and co-workers found that LSD1 plays a key function in maintaining the oncogenic potential of acute

myeloid leukaemia (AML) cell lines by preventing differentiation.7 Subsequently, research has shown

LSD1 expression to be linked to a number of cancers, most notably small cell lung cancer (SCLC). 8

The development of irreversible inhibitors of LSD1 from the monoamine oxidase (MAO) inhibitor

tranylcypromine (TCP) has been well documented,9 and several compounds are now in clinical trials

for AML and SCLC, either as a monotherapy, or in combination with other pro-differentiation agents

such as all-trans retinoic acid (ATRA).10 Results from a Phase I study of AML patients with resistant

or refractory disease showed that a blast cell differentiation effect was observed in the majority of

treated patients.11

Commercial interest in the development of irreversible TCP LSD1 inhibitors has increased rapidly in

the past decade. Despite this, the rate of optimisation of tranylcypromine derivatives has not been

matched by reversible inhibitors, as evidenced by the paucity of clinical trials involving reversible

agents.12 The large size and polarity of the LSD1 active site presents significant challenges to drug

discovery. However, in the past two years, the patent literature has seen a convergence towards

optimisation of derivatives of GSK-690 (also known as GSK-354),13 a potent reversible LSD1

inhibitor compromised by a significant hERG liability.

2

Figure 1. Structure of GSK-690 (1) and examples of recently disclosed reversible inhibitors of LSD1

from the patent literature (2–7).14-19

Various bicyclic and monocyclic scaffold-hops have been disclosed (Figure 1) in numerous patents.

We became interested in a hydroxypyrazole scaffold template disclosed, but not claimed in a patent

from Quanticel Pharmaceuticals.19 The replacement of the tolyl- (1) or other aryl or heteroaryl- groups

(2–6) with an ether linkage offers a new vector to investigate. Whilst, the patent only disclosed three

pyridylmethyl isomers, the 2-pyridyl isomer (7) was preferred, and was reported to have activity of

<100 nM in biochemical assays. This compound was resynthesised via the disclosed procedure, and

profiled in our biochemical time-resolved fluorescence resonance energy transfer (TR-FRET) assay

against LSD1 where it displayed an IC50 value of 0.23 µM (SD 0.08 µM). Surface plasmon resonance

(SPR) analysis indicated 7 was a reversible inhibitors of LSD1 with a Kd value of 0.042 µM. It also

displayed activity in our previously reported CD86 cellular biomarker assay (EC50 = 2.3 µM).20 These

values suggested 7 was an attractive start point for further development.

3

Scheme 1. Synthesis of compounds 11a–q. Reagents and conditions: (a) oxaloacetic acid, acetic acid,

reflux, 1 h, 93%; (b) COMU, (R)-3-Boc-aminopiperidine, DIPEA, DMF, 30 min, 84%; (c) R-X,

K2CO3, DMF, 100 °C, 1h, then 4M HCl/dioxane, rt, 1 h, 24–71%.

To explore the SAR around the ether functionality, we synthesised hydroxypyrazole 9 from 4-

cyanophenylhydrazine (8) via reaction with oxaloacetic acid (Scheme 1). The amide was then

installed under standard coupling conditions to give Boc-protected amide 10. The hydroxypyrazole

could be alkylated under basic conditions with a variety of alkylating agents, and then deprotected

under acidic conditions to afford compounds 11a–q. From the limited SAR disclosed in the patent,

we envisaged that the pyridyl nitrogen may be acting as a H-bond acceptor, hence we focused on

replacement of the pyridyl with groups containing functionalities which have hydrogen-bonding

groups analogous to the pyridyl nitrogen (Table 1). While the primary amide 11a had only modest

affinity for LSD1 by SPR, methylation at either carbon (11b) or nitrogen (11c) resulted in a

significant improvement in potency in both assay formats, with compound 11c achieving sub-

micromolar affinity by SPR. Extending the dimethylamide further to a morpholine (11d) improved

potency as well as resulting in decreased lipophilicity. More promising were alcohols 11g–i.

Compound 11h achieved a biochemical IC50 of 50 nM and a Kd value of 6 nM (Figure 2), a significant

improvement over 7, and one of the most potent reversible inhibitors of LSD1 reported to date. The

acidifying effect of the trifluoromethyl group appears important, giving a 32-fold improvement in

potency over 11g. These compounds demonstrated that a variety of alcohol side chains could be

tolerated. Finally, we incorporated a number of different 5-membered heterocyclic groups to

investigate how the presence of different heteroatoms in varying positions would affect the activity

4

(11i–q). Most notably, methylisoxazole 11o displayed excellent affinity for LSD1 by SPR, with the

extension of the methyl group to ethyl (11p) and isopropyl (11q) also well tolerated.

Table 1a

Compound R IC50 (µM) SPR Kd

(µM)

11a >150 17.7

11b 16.0 (0.9) 3.9

11c 2.8 (0.15) 0.76

11d 1.5 (0.35) 0.21

11e 4.3 (1.2) 0.45

11f 6.3 (0.7) 1.7

11g 1.6 (0.2) 0.13

11h

CF3

OH0.050 (0.01) 0.006

11i 0.27 (0.06) 0.022

11j 0.41 (0.02) 0.066

11k 2.6 (0.54) 0.32

11l 0.75 (0.13) 0.12

5

NN

N

O

N

OH2N

R

NH2

O

NH2

O

N

O

N

O

O

HN

O

OH

OH

NN

N

NN

N

N N

11m 0.19 (0.03) 0.034

11n 1.2 (0.2) 0.33

11o 0.15 (0.09) 0.007

11p 0.079 (0.024) 0.021

11q 0.083 (0.034) 0.017

aIC50 and Kd values for selected compounds against the LSD1 enzyme in biochemical and biophysical

assays. Standard deviation given in parentheses. IC50 determined from 10 point concentration/effect

experiments. Geometric mean of at least two independent experimental determinations given.

Figure 2. Surface plasmon resonance sensorgrams of the interaction between LSD1 and selected

reversible inhibitors in two-fold dilution series (highest concentration indicated in the graph).

6

N

NO

O

N N

O N

O N

O N

In an attempt to account for the observed SAR, we performed docking studies using the previously

reported structure of LSD1 bound to the reversible ligand tetrahydrofolate (PDB accession code

4KUM).21 Protein preparation and docking were performed in Glide (Schrödinger, New York,

USA).22 Previous disclosure of the crystal structure of compound 1 bound to LSD1 has shown that the

nitrile displaces a bridging water molecule between K661 and FAD, and that the basic centre is

directed towards a pair of adjacent aspartate residues, Asp555 and Asp556. Docking studies of

compound 7 identified a binding mode in which the pyridyl moiety makes an additional hydrogen

bonding interaction with Asn535 (Figure 3). Another credible binding pose suggested that the

flexibility of the ether link allows this group to be directed towards Gln358, achieving a similar

hydrogen bonding interaction as the first binding mode. Compounds 11h and 11o displayed similar

predicted binding modes, with the possibility of additional hydrogen bonding interactions from the

ether oxygen to His564.

7

Figure 3. Predicted binding modes of compounds 7 (green carbons), 11h (yellow carbons) and 11o

(magenta carbons) in the LSD1 active site (4KUM). Visualised in PyMOL. FAD truncated for clarity.

11o 11h 11p

Caco2 A–B mean Papp (10-6

cm/s); (efflux ratio) 0.73 (12) <0.1 0.415 (111)

Hu Mic CLint (μL/min/106 cells) 6.91 5.6 5.6

Hu Mic T½ (min) 201 250 250

Mo Hepatocytes T½ (min) 68 ND 121

Table 2. In vitro DMPK properties for selected compounds. ND – not determined

In vitro DMPK experiments (Table 2) identified a potential liability in the Caco-2 membrane

permeability assay, with high levels of efflux observed, or in the case of compound 11f, levels of

permeability below detectable levels. This has also been previously observed in series of

hydroxypyrazole compounds.23

To investigate if we could improve the DMPK properties, we looked to alter the basic centre while

retaining the methyl-isoxazole functional group. These compounds were synthesised in four steps

from hydrazine 8 (Scheme 2). Cycloaddition of 8 with diethyl acetylenedicarboxylate afforded

hydroxypyrazole 12. This was alkylated with 5-(bromomethyl)-3-methylisoxazole under basic

conditions to afford ether 13. Ester hydrolysis with lithium hydroxide afforded acid 14, which was

reacted with various di-amines (Boc-protected where necessary) under standard amide coupling

conditions, then deprotected where necessary to afford compounds 15a–m (Table 3).

The stereoisomer of compound 11o (15a) was roughly equipotent by biochemical assay. Methylation

of the basic centre was slightly deleterious to activity. Interestingly, the two enantiomers of the 3-

aminopyrollidine derivatives 15i and 15j showed a significant disparity in activity, with the S-

enantiomer strongly favoured. To see how modification of the pKa of the amine would affect the

activity, we synthesised 15k with a fluorine beta- to the basic centre, however this was poorly

8

tolerated. Overall, there appeared to be little scope to optimise the basic centre to further improve

potency.

Scheme 2. Reagents and conditions: (a) diethyl acetylenedicarboxylate, K2CO3, ethanol, reflux, 16 h,

79%, (b) 5-(bromomethyl)-3-methylisoxazole, K2CO3, DMF, 80 °C, 4 h, 65%; (c) LiOH, MeOH,

H2O, 50 °C, 30 min, 80%; (d) COMU, appropriate amine, DIPEA, DMF, 1 h, rt, then 4M

HCl/dioxane (if required), 20–74%.

Table 3a

Compound R IC50 (µM) SPR KD

(µM)

15a 0.20 (0.04) 0.032

15b 0.34 (0.03) 0.042

15c 0.65 (0.21) 0.087

9

15d 0.67 (0.08) 0.099

15e 0.72 (0.08) 0.14

15f 1.1 (0.3) 0.24

15g 1.2 (0.2) 0.23

15h 0.22 (0.04) 0.037

15i 0.17 (0.04) 0.013

15j 1.4 (0.1) 0.17

15k 2.6 (0.4) 0.11

15l 0.51 (0.07) 0.097

15m 0.71 (0) 0.24

aIC50 and Kd values for selected compounds against the LSD1 enzyme in biochemical and biophysical assays. Standard deviation given in parentheses. IC50 determined from 10 point concentration/effect experiments. Geometric mean of at least two independent experimental determinations given.

Compound CD86 (µM)

7 2.3 (0.41)

11o 1.9 (0.18)

11m 0.42 (0.07)

11h 0.48 (0.08)

11p 0.52 (0.09)

Table 4. IC50 values for selected compounds in CD86 expression based cellular assays. IC50

determined from 10 point concentration/effect experiments. Geometric mean of at least three independent experimental determinations given.

10

The most active compounds were taken forwards into previously described cellular assays.12

Upregulation of the cell surface protein CD86 can be used as a reliable surrogate cellular biomarker of

LSD1 inhibition, and is superior to other cell assays such as global histone methylation levels, which

should not be trusted as an indication of LSD1 inhibition in a cellular context. Here, the optimised

derivatives showed improved activity over compound 7, with the biochemical activity correlating

well, albeit with a significant, but consistent, ~5-fold drop-off in activity in the cellular assay.

Figure 4. In vivo DMPK properties of compound 11p.

Compound 11p was subjected to in vivo pharmacokinetic assessment using serial microsampling,24, 25

to determine whether the high efflux ratio and low permeability observed would transfer into low

bioavailability in mice. We were relieved to find that 11p was moderately bioavailable and only

moderately cleared with a half-life of around 1 h in both the i.v. and p.o. dosed mice (Figure 4).

Unfortunately, when tested against the hERG ion channel in a patch clamp assay this compound was

found to have an IC50 of ~5 µM. While this was an improvement over compound 1 (IC50 = 3 µM), it

still represented a significant liability that must be overcome when developing LSD1 inhibitors that

contain basic moieties adjacent to an aryl or heteroaryl core, especially when developing a rigid

aromatic scaffold such as hydroxypyrazoles or other recently patented mono- and bicyclic scaffold

modifications of 1.

In summary, we have described herein our development of a series of reversible 5-hydroxypyrazole

based inhibitors of LSD1. These compounds are highly potent by biochemical assay and SPR, and

have been shown to induce differentiation of THP-1 cells at sub-micromolar concentrations in cellular

11

assays. Compound 11p confounded in vitro DMPK experiments that indicated low levels of

permeability and high efflux by displaying moderate levels of bioavailability and reasonable half-life

in mice. These compounds are some of the most potent cell-active LSD1 inhibitors described to date

with the potential for oral bioavailability, and should enable further research in the area.

Acknowledgments

This work was supported by Cancer Research UK (Grants C5759/A12328, C480/A11411,

C5759/A17098 and C5759/A02901). Additional support was provided to DPM by the Society of

Chemical Industry through a Messel Scholarship. We are thankful to Bohdan Waszkowycz for his

assistance with Glide Docking. In vitro pharmacokinetic and hERG data was provided by Cyprotex

Discovery (Macclesfield, UK). Images of protein-ligand docking were captured within the PyMOL

Molecular Graphics System, Version 1.7.6.2. (Schrödinger, LLC, New York).

References and Notes

Tim Somervaille has ongoing research collaborations with Oryzon Genomics and consults for Imago

Biosciences. The other authors declare no competing financial interest.

1. Shi Y-J, Matson C, Lan F, Iwase S, Baba T, Shi Y. Mol Cell. 2005;19(6): 857-864.

2. Metzger E, Wissmann M, Yin N, et al. Nature. 2005;437(7057): 436-439.

3. Liu L, Souto J, Liao W, et al, J. Bio. Chem. 2013;288(48): 34719–34728.

4. Thambyrajah R, Mazan M, Patel R, et al. Nat. Cell. Biol. 2016;18(1): 21-32.

5. Saleque S, Kim J, Rooke HM, Orkin SH. Mol. Cell. 2007;27(4): 562-572.

6. Pilotto S, Speranzini V, Tortorici M, et al. Proc. Natl. Acad. Sci. U.S.A. 2015;112(9): 2752–

2757.

12

7. Harris WJ, Huang X, Lynch JT, et al. Cancer Cell. 2012;21(4): 473–487.

8. Mohammad HP, Smitheman KN, Kamat CD, et al. Cancer Cell. 2015;28(1): 57-69.

9. Hojfeldt JW, Agger K, Helin K. Nat. Rev. Drug Discov. 2013;12(12): 917-930.

10. Schenk T, Chen WC, Gollner S, et al. Nat. Med. 2012;18(4): 605-611.

11. Somervaille T, Salamero O, Montesinos P, et al. Blood. 2016;128(22): 4060.

12. Mould DP, McGonagle AE, Wiseman DH, Williams EL, Jordan AM. Med. Res. Rev.

2015;35(3): 586-618.

13. Hitchin JR, Blagg J, Burke R, et al. Development and evaluation of selective, reversible

LSD1 inhibitors derived from fragments. MedChemComm. 2013;4(11): 1513-1522.

14. Chen YK, Kanouni T, Nie Z, Stafford JA, Veal JM, Sung LM. WO2016004105. Cellgene

Quanticel Research Inc.; 2016.

15. Chen YK, Kanouni T, Nie Z, Stafford JA, Veal JM, Sung LM. WO2016003917. Cellgene

Quanticel Research Inc.; 2016.

16. Liangxing W, Xiaozhao W, Wenqing Y, Zhang C. US20160009711. Incyte Corporation;

2016.

17. Liangxing W, Konkol LC, Lajkiewicz N, et al. US20160009712. Incyte Corporation; 2016.

18. Liangxing W, Courter JR, Chunhong H, et al. US20160009720. Incyte Corporation; 2016.

19. Chen YK, Kanouni T, Kaldor SW, Stafford JA, Veal JM. Cellgene Quanticel Research Inc.;

WO2015089192.

20. Lynch JT, Cockerill MJ, Hitchin JR, Wiseman DH, Somervaille TCP. Anal. Biochem.

2013;442(1): 104-106.

13

21. Luka Z, Pakhomova S, Loukachevitch LV, Calcutt MW, Newcomer ME, Wagner C. Protein

Sci. 2014;23(7): 993-998.

22. Friesner RA, Banks JL, Murphy RB, et al. J. Med. Chem. 2004;47(7): 1739-1749.

23. Cadieux JA, Zhang Z, Mattice M, et al. Bioorg. Med. Chem. Lett. 2012;22(1): 90-95.

24. Li C, Liu B, Chang J, et al.. Drug Discov. Today. 2013;18(1–2): 71-78.

25. Liu B, Chang J, Gordon WP, Isbell J, Zhou Y, Tuntland T. Drug Discov. Today. 2008;13(7–

8): 360-367.

14