Optimization of microtubule affinity regulating kinase (MARK) inhibitors with improved physical properties

Article abstract of DOI:10.1016/j.bmcl.2016.02.003

Inhibition of microtubule affinity regulating kinase (MARK) represents a potentially attractive means of arresting neurofibrillary tangle pathology in Alzheimer's disease. This manuscript outlines efforts to optimize a pyrazolopyrimidine series of MARK in

Full text of DOI:10.1016/j.bmcl.2016.02.003

Accepted Manuscript
Optimization of Microtubule Affinity Regulating Kinase (MARK) Inhibitors
with Improved Physical Properties
David L. Sloman, Njamkou Noucti, Michael D. Altman, Dapeng Chen, Andrea
C. Mislak, Alexander Szewczak, Mansuo Hayashi, Lee Warren, Tammy
Dellovade, Zhenhua Wu, Jacob Marcus, Deborah Walker, Hua-Poo Su, Suzanne
C. Edavettal, Sanjeev Munshi, Michael Hutton, Hugh Nuthall, Matthew G.
Stanton
PII:
S0960-894X(16)30106-8
DOI:
http://dx.doi.org/10.1016/j.bmcl.2016.02.003
BMCL 23554
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Accepted Date:
18 December 2015
2 February 2016
Please cite this article as: Sloman, D.L., Noucti, N., Altman, M.D., Chen, D., Mislak, A.C., Szewczak, A., Hayashi,
M., Warren, L., Dellovade, T., Wu, Z., Marcus, J., Walker, D., Su, H-P., Edavettal, S.C., Munshi, S., Hutton, M.,
Nuthall, H., Stanton, M.G., Optimization of Microtubule Affinity Regulating Kinase (MARK) Inhibitors with
Improved Physical Properties, Bioorganic & Medicinal Chemistry Letters (2016), doi: http://dx.doi.org/10.1016/
j.bmcl.2016.02.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Optimization of Microtubule Affinity Regulating Kinase (MARK)
Inhibitors with Improved Physical Properties
David L. Sloman*†, Njamkou Noucti†, Michael D. Altman‡, Dapeng Chen§, Andrea C. Mislak§,
Alexander Szewczak¥, Mansuo Hayashi¥, Lee Warren¥, Tammy Dellovade┬, Zhenhua Wu┬, Jacob
Marcus∆, Deborah Walker, Hua-Poo Su∂, Suzanne C. Edavettal∂, Sanjeev Munshi∂, Michael Hutton∆,
Hugh Nuthall∆, Matthew G. Stanton†
Discovery Chemistry†, Chemistry Modeling and Informatics‡, Drug Metabolism and Pharmacokinetics§, Neuroscience Drug
Discovery∆, CNS Pharmacology¥, Core Pharmacology┬, Pharmaceutical Research and Development, Structure
Determination∂ Merck and Co., Inc., 33 Avenue Louis Pasteur, Boston, MA 02215, Structural Biology, 770 Sumneytown
Pike, West Point, PA, 19486.
KEYWORDS (“MARK, Kinase, Alzheimer’s disease, amine basicity, fluorine”).
ABSTRACT: Inhibition of microtubule affinity regulating kinase (MARK) represents a potentially
attractive means of arresting neurofibrillary tangle pathology in Alzheimer's disease. This manuscript
outlines efforts to optimize a pyrazolopyrimidine series of MARK inhibitors by focusing on
improvements in potency, physical properties and attributes amenable to CNS penetration. A unique
cylcyclohexyldiamine scaffold was identified that led to remarkable improvements in potency, opening
up opportunities to reduce MW, Pgp efflux and improve pharmacokinetic properties while also
conferring improved solubility.
Alzheimer's disease remains one of the highest areas of unmet medical need.1 Despite significant
efforts, disease modifying therapies for AD remain elusive.2 Neurofibrillary tangles (NFTs) are one of
the pathological hallmarks of AD and the degree of tangle formation correlates well with cognitive
decline. NFTs are composed of hyperphosphorylated tau protein, thus inhibition of tau kinases
represents a rational approach to treating AD.3-7 Of particular interest are the KXGS microtubule
binding domains in tau as these are thought to control tau/microtubule binding affinity and kinetics.8 It
is speculated that aberrant phosphorylation of the serines in these KXGS motifs acts as a "master
switch" leading to high levels of unbound tau and subsequent hyperphosphorylation.8 Microtubule
affinity regulating kinase (MARK) has been demonstrated to be one of the kinases capable of
phosphorylation of S262 lying in one of the KXGS binding motifs.9 It has been speculated elsewhere
Figure 1. Properties of MARK inhibitor 1.

that inhibition of MARK would thus be an attractive therapeutic option for the treatment of Alzheimer's
disease.5 There are four human isoforms of MARK (1-4) and all four are highly enriched in human
brain. Additionally, it has been demonstrated that all four isoforms are capable of S262
phosporylation.10 Given the high sequence homology, particularly in the ATP binding domain, and the
potential of all four isoforms to be pathologically relevant, we sought to identify and optimize pan-
MARK inhibitors targeting the ATP binding domain.10
Initial efforts in our laboratories led to the identification of compound 1 as a reasonable lead MARK
inhibitor for optimization. The compound was potent in both biochemical and cell based assays.
Kinase selectivity was less than optimal but pharmacokinetic properties in rats were fair as a starting
point. Efforts to measure the susceptibility of 1 to Pgp transport were compromised by the compound's
poor physical properties. Of significant concern is the relatively high MW of 529 as CNS penetration
tends to be inversely proportional to MW. This manuscript describes our efforts to optimize 1 focusing
on improving potency, reducing MW and attaining attributes consistent with CNS penetration.11
Modeling of 1 suggests the piperidine group is oriented towards solvent in its bound confirmation.
This group likely serves to add solubility to the molecule, but at a significant price to MW and without
necessarily improving binding affinity. In an effort to reduce MW, we prepared a series of heterocycle
substituted pyrazolopyrimidines (Table 1). Initial truncation of the piperidine to the methoxyethoxy
derivative 2 resulted in a slight loss in potency. Bicyclic analogs, 3-5 also resulted in a slight loss in
Table 1. Heterocycle substituted pyrazolopyrimidine MARK inhibitors.

potency. Further restriction to the pyridyl or pyrazole substituents again resulted in a loss in potency,
but this could be rescued by capping the 5 membered heteroaryl groups 8-9 with a methyl.
These smaller heterocycles such as methyl imidazole 8 and N-methyl pyrazole 9, were better tolerated
and offered a much better balance of potency to MW as reflected in their increased ligand efficiencies.
Given the balance of potency, low MW, and low polar surface area, compound 9 was chosen for further
characterization (Figure 2). Despite the inferior biochemical potency of 9 vs. 1, the cell based activities
were quite similar. Selectivity of 9 versus other kinases was still sub-optimal; however the
pharmacokinetic parameters in rat were quite good. The compound had modest clearance, good
bioavailability and a low volume of distribution. Additionally, 9 was not a Pgp substrate in human or rat
and displayed good passive permeability. Upon dose escalation of 9 in rodents, saturation of both Cmax
and AUC was observed at relatively low doses, thwarting efforts to look at 9 in a rat hypothermia model
of tau S262 phosphorylation. Poor solubility (<3 μM at pH 2 and 7) of this compound was deemed the
likely culprit.
Figure 2. Properties of MARK inhibitor 9
Rather than attempt to address solubility by adding polarity to the solvent exposed region of this class
of inhibitors, a strategy that has been demonstrated to be effective but at a cost to molecular weight and
ligand efficiency, we decided to focus on the amide portion of 9 (Table 2). The simple ethyl amide 10
had poor activity and highlights the impact of the trifluoromethyl group in compound 9. Alkyl
substituents proximal to the amide bond, such as the cyclopropyl derivative 11, were able to restore
potency to single digit nanomolar range, but also negatively impacted solubility. Cyclic secondary
alcohols 12-15 were well tolerated, although these also did little to improve solubility. Cyclic amines
16-21 managed to provide significant improvements in both potency and solubility. The larger, six-
membered ring 18-19 was preferred over the smaller cyclopentyl 16-17, the cis-racemic
cycolhexylamine compound 19 demonstrating sub-nanomolar potency. Despite the increase in MW, the
potency enhancements observed by the inclusion of this functional group improved the LE to a very
respectable 0.43 value. Based on improved potency and solubility, 19 was more fully characterized
(Figure 3). We were gratified to see that the improved biochemical potency and physical properties
translated into a low single digit nanomolar cellular IC50, a first for this program. The kinase
selectivity was apparently compromised, but this may also be an artifact of the improved potency

Table 2. Polar amide substituted pyrazolopyrimidine MARK
inhibitors.
Figure 3. Properties of MARK inhibitor 18

coupled with the screening conditions (biochemical potency is reaching the bottom of assay limits). The
rat pharmacokinetic properties were not particularly impressive, with clearance rates greater than
hepatic extraction and a high volume of distribution. Not surprising given the increased polarity and
basicity, this molecule was also a potent Pgp substrate.
Although 19 was not an ideal candidate for evaluation in vivo, we did seek to better understand the
source of the potency increase obtained with the novel cyclohexyldiamine amide. Crystallization efforts
were commenced with all of the purified MARK isoforms and MARK2 crystals with 19 bound were
obtained with good resolution (Figure 4). The overall protein structure is very similar to what has been
previously reported,12 and compound 19 binds in the ATP site extending from the hinge region to the
catalytic residues (Figure 4, inset). N1 of the pyrazolopyrimidine core makes a hydrogen bond contact
to the hinge backbone NH of Ala132 and the thiophene packs against Met129, the gatekeeper residue.
The carbonyl of the amide forms a hydrogen bond to the catalytic Lys82 while the free amine forms a
double salt bridge with both Glu179 and Asp193, likely contributing to the high potency. These
interactions are accommodated in a manner that enables the cyclohexyl ring to push into the
hydrophobic glycine-rich loop. The interactions outlined above are with residues that are highly
conserved in the ATP-binding domain of kinases and thus are not likely to offer obvious means of
improving selectivity. However, this offers a reasonable means of improving potency and solubility
that, at least hypothetically, may be applied to numerous kinase targets. Given the difficulty of
attenuating physical properties of ATP-competitive kinase inhibitors, this inhibitor architecture may
present a generalizable solution.
We hypothesized that the poor pharmacokinetics of 19 may, at least in part, be attributable to the
relatively high pKa of the free amine.13-16 To test this hypothesis, a series of 3 amides (20-22) with
varying calculated amine pKa’s were synthesized and evaluated in vivo (Figure 5).17 The amines with
lower basicity (21-22) led to improved in vivo clearance and Pgp ratio. We were now confident that
Figure 4. Crystal structure of 19 bound to MARK2.
synthesis of the complex fluorinated cyclohexyldiamine analogs were well worth undertaking. (Table
3). At this point in our program, we had discovered that introduction of a methyl substituent at the 4-
position of the thiophene imparted improved kinase selectivity, albeit with decreased overall potency
(23). Our efforts to address selectivity will be covered in detail in a forthcoming manuscript, but
comparative SAR moving forward was conducted with this methyl substituent. Addition of fluoro
substituents proximal to the free amine 24, had a significant negative impact on potency. Surprisingly,
addition of the fluorines proximal to the amide NH conferred a significant increase in activity (25, R,R-
enantiomer was determined to be the most active enantiomer (data not shown). Given this unexpected
boost in activity, we reasoned that removal of the N-methylpyrazole, while maintaining good potency,
may be possible. Compounds 26 and 27 bear this out as they maintain single digit nanomolar potency

Figure 5. PK Properties of MARK inhibitor 20-22.
Table 3. Fluorinated cyclohexylamine MARK inhibitors.
with improved ligand binding efficiency. Lower amine pKa along with lower MW and polar surface
area are likely contributing factors to achieving Pgp efflux and permeability values that are known to
confer CNS penetration, particular in the case of 27.

Figure 6. Properties of MARK inhibitor 27.
Further characterization of the i.v. pharmacokinetic properties of 27 in rat and dog revealed reasonable
volumes of distribution but moderate to high clearance and short half-lives (Figure 6. Additionally, the
cell-based activity of 27 was less than optimal. While these properties represent an opportunity for
further optimization, the solubility of 27 was greatly enhanced by the cumulative modifications starting
from lead molecule 1. At pH 7 27 attained a concentration of 114 μM in our solubility assay and at pH
2 that number increased to 206 μM.
We have discussed the identification of an early MARK inhibitor, 1 with high MW and described the
optimization of this initial lead through truncation of the polar solvent front substituent. Substitution
with small heteroaryls lead to the identification of 9 which maintained potency, but had poor physical
properties. In an effort to improve solubility a polar functional cyclohexyldiamine was utilized and
shown to have superior potency and physical properties, but suffered from poor PK. By attenuating the
basicity of the amine through difluoro-substitution, we were able to improve PK, and serendipitously
saw a significant improvement in biochemical and cellular potency. Further discussion on the
optimization of compounds similar to 27, and discussions on the hemodynamic effects associated with
the target will be the subject of a future manuscript.20, 21
CORRESPONDING AUTHOR E-MAILADDRESS*: david_sloman@merck.com
ACKNOWLEDGMENT
We would like to thank Dr. Bruce Adams for NMR support.
SUPPLEMENTARY DATA
The PDB accession code for the X-ray cocrystal of MARK2 + compound 19 is 5EAK. Supplementary data associated with
this article can be found, in the online version, at “ ” 18,19
ABBREVIATIONS
AD, Alzheimer's disease; MARK, microtubule affinity regulating kinase; NFT, neurofibrillary tangles; CSF, cerebrospinal
fluid; CYP, cytochrome P450; Pgp, P-glycoprotein; LE, ligand efficiency, PSA, polar surface area.
REFERENCES
(1) Querfurth, H., W. LaFerla, F. M. N. Engl. J. Med. 2010, 362, 329.
(2) Citron, M. Nat. Rev. Drug Discov. 2010, 9, 387.
(3) Gong, C.-X.; Grundke-Iqbal, I.; Iqbal, K. Drugs Aging 2010, 27, 351-365.
(4) Lee, V. M-Y.; Brunden, K. R.; Hutton, M.; Trojanowski, J. Q. Cold Spring Harb. Perspect. Med. 2011, 1, a006437

(5) Martin, L.; Latypova, X.; Wilson, C. M.; Magnaudeix, A.; Perrin, M.L.; Yardin, C.; Terro, F. Ageing Res. Rev. 2013, 12,
289-309.
(6) Iqbal, K.; Gong, C.-X.; Liu, F. Expert Opin. Ther. Tar. 2014, 18, 307-318.
(7) Drewes, G.; Ebneth, A., Preuss, U., Mandelkow, E.M., and Eckhard Mandelkow b. Lee, V. M.-Y.; Brunden, K. R.;
Hutton, M.; Trojanowski, J. Q. Cell, 1997, 89, 297–308.
(8) Nishimura, I, Yang, Y., Lu, B., Cell, 2004, 116, 671-682.
(9) Matenia, D.; Mandelkow, E.-M. Trends Biochem. Sci. 2009, 34, 332.
(10) Timm, T., Marx, A., Panneerselvam, S., Mandelkow, E., Mandelkow, E.M. BMC Neuroscience, 2008, 9(Suppl 2): S9
(11) Pajouhesh, H., Lenz, G.R. NeuroRx, 2005, 2, 541-553.
(12) Panneerselvam, S., Marx, A., Mandelkow, E. M., Mandelkow E, Structure, 2006, 14, 173-183.
(13) Smith, D.; van de Waterbeemd, H.; Walker, D. K. Pharmacokinetics and Metabolism in Drug Design, 2nd Edition:
Mannhold, R.; Kubinyi, H.; Folkers, G. Germany: Wiley-VCH. 2006.
(14) Morgenthaler, M.; Schweizer, E.; Hoffmann-Roder, A.; Benini, F.; Martin, R. E.; Jaeschke, G.; Wagner, B.; Fischer, H.;
Bendels, S.; Zimmerli, D.; Schneider, J.; Diederich, F.; Kansy, M.; Muller, K. Chem Med Chem 2007, 2, 1100−1115.
(15) van Niel, M. B.; Collins, I.; Beer, M. S.; Broughton, H. B.; Cheng, S. K. F.; Goodacre, S. C.; Heald, A.; Locker, K. L.;
MacLeod, A. M.; Morrison, D.; Moyes, C. R.; O’Connor, D.; Pike, A.; Rowley, M.; Russel, M. G. N.; Sohal, B.; Stanton, J.
A.; Thomas, S.; Verrier, H.; Watt, A. P.; Castro, J. L. J. Med. Chem. 1999, 42, 2087−2104.
(16) Hicken E. J.; Marmsater F. P.; Munson M. C.; Schlachter S. T.; Robinson J. E.; Allen S.; Burgess L. E.; DeLisle R. K.;
Rizzi J. P.; Topalov G. T.; Zhao, Q.;Hicks J. M.; Kallan, N. C.; Tarlton, E.; Allen, A.; Callejo, M.; Cox, A.; Rana, S.;
Klopfenstein, N.; Woessner, R.; Lyssikatos, J.P. ACS Med. Chem. Lett. 2014, 5, 78−83.
(17) pKa’s calculated using ACD LABS v. 11.0
(18) * Some experimentals have been previously reported in; Lim, J.; Taoka, B. M.; Lee, S.; Northrup, A.; Altman, M. D.;
Sloman, D. L.; Stanton, M. G.; Noucti, N. Pyrazolo[1,5-a]pyrimidines as mark inhibitors. WO 2011/087999, 2011.
(19) * Some experimentals have been previously reported in; Churcher, I.; Hunt, P.A.; Stanton, M., G. Use of pyrazolo[1,5-a]
pyrimidine derivatives for the treatment of Alzheimer's disease and related conditions WO 2007085873 A1 20070802, 2007.
(20) Sloman, D. L. Abstracts of Papers, 244th ACS National Meeting & Exposition, Philadelphia, PA, United States, 2012
MEDI-100.
(21) Lamore, S.D.; Kamendi, H.W.; Scott, C.W.; Dragan, Y.P.; Peters, M.F. Toxicol. Sci. 2013, 135, 402-413.
GRAPHICALABSTRACT