Discovery, synthesis, and biological evaluation of thiazoloquin(az)olin(on)es as potent CD38 inhibitors

Article abstract of DOI:10.1021/jm502009h

A series of thiazoloquin(az)olinones were synthesized and found to have potent inhibitory activity against CD38. Several of these compounds were also shown to have good pharmacokinetic properties and demonstrated the ability to elevate NAD levels in plasm

Full text of DOI:10.1021/jm502009h

Subscriber access provided by SUNY DOWNSTATE
Article
Discovery, Synthesis and Biological Evaluation of
Thiazoloquin(az)olin(on)es as Potent CD38 Inhibitors.
Curt D. Haffner, David Becherer, Eric E Boros, Rodolfo Cadilla, Tiffany Carpenter, David
Cowan, David Deaton, Yu Guo, Wallace Harrington, Brad Richard Henke, Michael Jeune,
Istvan Kaldor, Naphtali Milliken, Kimberly Petrov, Frank Preugschat, Christie Schulte, Barry
G Shearer, Todd Shearer, Terry Smalley, Eugene L. Stewart, Darren Stuart, and John Ulrich
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm502009h • Publication Date (Web): 01 Apr 2015
Downloaded from http://pubs.acs.org on April 7, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery, Synthesis and Biological Evaluation of Thiazoloquin(az)olin(on)es as Potent CD38 Inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Curt D. Haffner*, J. David Becherer, Eric E. Boros, Rodolfo Cadilla, Tiffany Carpenter, David Cowan,
David N. Deaton, Yu Guo, Wallace Harrington, Brad R. Henke, Michael R. Jeune, Istvan Kaldor,
Naphtali Milliken, Kim G. Petrov, Frank Preugschat, Christie Schulte, Barry G. Shearer, Todd Shearer,
Terrence L. Smalley, Jr., Eugene L. Stewart, J. Darren Stuart, John C. Ulrich.
GlaxoSmithKline Research and Development, 5 Moore Drive, P.O. Box 13398, Research Triangle Park,
North Carolina 27709
*To whom correspondence should be addressed. Phone: 919ꢀ483ꢀ6247. Fax: 919ꢀ315ꢀ6787. Eꢀmail:
curt.d.haffner@gsk.com.
KEYWORDS: Nicotinamide adenine dinucleotide, NAD, CD38, pyridine nucleotides, nicotinamide,
NAADP, NADH, nicotinamide mononucleotide
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
A series of thiazoloquin(az)olinones were synthesized and found to have potent inhibitory activity against
CD38. Several of these compounds were also shown to have good pharmacokinetic properties and
demonstrated the ability to elevate NAD levels in plasma, liver and muscle tissue. In particular,
compound 78c was given to diet induced obese (DIO) C57Bl6 mice elevating NAD > 5ꢀfold in liver and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
>
1.2ꢀfold in muscle versus control animals at a 2 hour time point. The compounds described herein
possess the most potent CD38 inhibitory activity of any small molecules described in the literature to
date. The inhibitors should allow for a more detailed assessment of how NAD elevation via CD38
inhibition affects physiology in NAD deficient states.
Introduction
Nicotinamide adenine dinucleotide (NAD) is a biochemical found in all cells that was first
discovered over 80 years ago. Cellular NAD is produced by either the de novo synthesis pathway from
tryptophan or by a salvage synthesis pathway from precursors such as nicotinic acid (niacin) and
1
nicotinamide, both of which are obtained from dietary sources. NAD and its related pyridine nucleotides
NADH, nicotinamide adenine dinucleotide phosphate (NADP), and NADPH have been recognized as the
major redox carriers in all organisms. These pyridine dinucleotides determine the cytosolic and
mitochondrial redox state and are key participants monitoring the metabolic status of the cell. NAD and
NADH act as hydride accepting and donating cofactors for metabolic enzymes involved in glycolysis, the
tricarboxylic acid cycle, and in electron transport. By transfering reducing equivalents generated from
catabolic processes into the de novo synthesis of new biomolecules, pyridine nucleotides are central
2
ꢀ5
regulators of sugar, fat and protein metabolism
Recent research demonstrates that NAD is also a substrate for various enzymes, where it is
consumed in the process of donating ADP ribose to acceptor molecules. Major consumers of NAD
ACS Paragon Plus Environment
2

Page 3 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
7
include the ADP ribosyl transferases (i.e. PARP and ART family of enzymes), the sirtuins (Sirt1ꢀ7), and
8
the ADP ribosyl cyclases/hydrolases (CD38/CD157). These enzymes are involved in pathways that can
regulate calcium signaling, gene transcription, DNA repair, cell survival, energy metabolism, and
oxidative stress. Thus, NAD and its phosphorylated relatives NADP and nicotinic acid adenine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
dinucleotide phosphate (NAADP), both of which are derived from NAD, also act as signaling molecules.
NAD levels are subject to circadian rhythms, with daily oscillations that tie cellular metabolism to
10
chromatin remodeling and gene transcription. It is known that exercise and caloric restriction elevate
1
1,12
NAD levels while aging and obesity decrease cellular NAD levels.
Restoring NAD levels in disease
states that consume significant amounts of NAD will likely have medical benefits as the cell strives to
survive and maintain its energy status during redox stress.
Increasing steadyꢀstate cellular levels of NAD can be achieved by two fundamental approaches:
upregulating production of NAD or downregulating its consumption. CD38 is one such consumer of
NAD. Also known as ADP ribosyl cyclase, CD38 is a type II membraneꢀanchored enzyme. It efficiently
catalyzes the breakdown of NAD to nicotinamide and ADPR and hydrolyzes NAADP to adenosineꢀ5'ꢀOꢀ
diphosphoribose phosphate (ADPRP). CD38 can also act as a cyclase converting NAD to cADPR,
although it is 100ꢀfold less efficient as a cyclase than as a hydrolase. CD38 was first characterized as a
surface antigen on immune cells and is broadly distributed throughout most tissues in the body. It exists
on the plasma membrane and on the membranes of intracellular organelles such as the nucleus and
13
mitochondria. As predicted from its function as a NAD glycohydrolase, CD38 knockꢀout (KO) mice
1
4
have elevated NAD levels relative to wildꢀtype controls. These animals are protected against weight
gain upon being fed a high fat diet. They also have an increase in energy expenditure. A genetic linkage
has been shown in humans between chromosome 4 near marker D4S403 where the CD38 gene is located
15
and metabolic syndrome. This data collectively suggest diseases linked to metabolic syndrome may be
possible targets of CD38 inhibitors. Diseases where CD38 is over expressed, or where cellular NAD
levels are depressed or desynchronized, could also be viable intervention venues.
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
16
There have been several reports in the literature of small molecule inhibitors of CD38.
However, many of these compounds possess moderate to weak inhibitory activity against the enzyme
IC50’s > 1 µM) and/or have limited selectivity. Arguably, more potent and selective inhibitors are
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
needed to better understand the biology around inhibition of CD38 and what, if any, potential uses these
inhibitors might possess with regards to treating various human diseases tied into NAD biology.
Our initial efforts to identify novel potent inhibitors of CD38 started with internal screening
campaigns of preꢀexisting compound collections, which produced a number of novel hits. An attractive
exemplar from these screening efforts was compound 1 demonstrating a human IC50 = 1.21 µM (Figure
1
).
Figure 1. CD38 thiazole screening hit.
Based on the modeling, the synthetic work was initially directed towards the bicyclic core and
substitution at the 4ꢀposition of compound 1. The modeling studies suggested that the primary
carboxamide at the 3ꢀposition was detrimental in terms of potency due to interactions with W176. When
compound 2 was synthesized to test this hypothesis the potency did improve by ~ 4ꢀfold (IC50 = 298 nM)
against the human enzyme (Figure 2).
ACS Paragon Plus Environment
4

Page 5 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Thiazoloquinoline and thiazoloquinazoline CD38 inhibitors.
The corresponding quinazoline 3b (Figure 2) was synthesized as well and found to have similar
potency to the original hit compound 1 (IC50 = 1.20 µM), which is a 4ꢀfold loss in CD38 potency relative
to 2; however we felt this to be an acceptable tradeꢀoff given the improved synthetic tractability of the
quinazoline versus the quinoline for rapid SAR expansion. As the SAR became more refined, additional
synthetic work was done on a quinolinꢀ2ꢀone core, which arose from the SAR established on the
quinazoline template.
Described herein are our efforts to identify, synthesize and biologically evaluate a series of novel
potent small molecule inhibitors of CD38 that arose from screening hit 1. Several of these inhibitors were
shown to elevate NAD in various tissues upon oral administration. The identification of these inhibitors
should allow for a more thorough understanding of CD38 pharmacology in animals and how elevating
this important cofactor, in NAD deficit animal models, affects whole body physiology.
Chemistry
The following chemistry sections describe the compounds that were made as part of a thorough
investigation of the SAR around the original screening hit 1. The biochemical data for these compounds is
described in the Results and Discussion section. As mentioned earlier, to more rapidly probe the SAR
around compound 1, chemistry focused on final targets that incorporated a quinazoline ring in place of the
quinoline. Synthetic work started on C4 substituted analogs containing amino substituted aryl, heteroaryl
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
or saturated 6ꢀthiazoloquinazolines via a two step sequence whereby 6ꢀbromoꢀ4ꢀchloroquinazoline was
treated with the requisite aniline or amine followed by a palladium mediated Stille coupling with
commercially available 5ꢀtributylstannylthiazole (Scheme 1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 1. Synthesis of 4ꢀamino substituted thiazoloquinazolines.
a
Reagents and conditions: (a) RNH (see Tables 1 and 2 for specifics with regards to what R groups were
2
incorporated into the final compounds 3aꢀyy), rt, CH CN; (b) CuI, K CO , rt or heating to 100ꢀ140 °C,
3
2
3
PdCl ·dppf·CH Cl , 5ꢀtributylstannyl thiazole, DMF.
2
2
2
Chemistry was then expanded to investigate C8 and/or C2 substituted quinazolin(on)e analogs.
The C8 analogs started with either the 3ꢀmethoxy or 3ꢀtrifluoromethylanthranilic acids 4 and 5,
halogenation generated the respective 5ꢀbromo or 5ꢀiodoanthranilic derivatives 6, 7 or 8. The quinazolinꢀ
4ꢀone ring was synthesized from these intermediates or from the commercially available 3ꢀmethylꢀ5ꢀ
bromoanthranilic acid 9 via heating with formamide to 165 °C yielding compounds 10ꢀ12 (anthranilic
acid 8 was used to synthesize analogs found in Scheme 4). Treatment of the quinazolinꢀ4ꢀones with
thionyl chloride under thermal conditions generated the 4ꢀchloroquinazolines 13ꢀ15 in good yields. The
last two steps were performed under the same conditions as described in Scheme 1 to provide the desired
final targets (19aꢀb, 20aꢀb and 21).
a
Scheme 2. Synthesis of C8 substituted quinazolines.
ACS Paragon Plus Environment
6

Page 7 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (a) when R = OMe, Br , CHCl , 0 °C to rt; (b) when R = CF , 48% HBr, H O,
1
2
3
1
3
2
3
0% H O , 75 °C or NIS, rt, DMF; (d) HCONH , 165 °C; (e) SOCl , DMF ; (f) R NH (R equals 3ꢀClꢀ
2 2 2 2 (cat) 2 2 2
4
ꢀFꢀPh or trans 4ꢀCONHMeꢀcyclohexyl), rt, CH CN; (g) CuI, K CO , rt or heating to 100ꢀ140 °C,
3
2
3
PdCl ·dppf·CH Cl , 5ꢀtributylstannyl thiazole, DMF.
2
2
2
The chemistry then targeted C2 substituted compounds (several of these final targets also
incorporated C8 functionality as well) on the quinazolin(on)e ring. It was found that incorporation of a
primary amino moiety at that position provided potent CD38 inhibitors (both mouse and human),
although many of these compounds suffered from either poor solubility and/or permeability (data not
shown). A more productive functional group turned out to be the primary hydroxyl residue (drawn as the
keto tautomer in Scheme 3). The synthesis to get into these compounds started with either 5ꢀbromoꢀ3ꢀ
methylanthranilic acid 9 or 5ꢀiodoanthranilic acid 22 (commercial) which upon melting with urea
provided the 2,4ꢀquinazolinones 23 or 24. Chlorination with POCl then cleanly provided the 2,4ꢀ
3
dichloroquinazolines 25 and 26. Reaction of these intermediates with a substituted aniline or saturated
amine generated the desired 4ꢀaminoꢀ2ꢀchloroquinazolines 27aꢀb and 28aꢀb. Hydrolysis with acetic acid
at 90 °C, provided compounds 29aꢀb or 30aꢀb. These compounds could then be converted to the final
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
product using the same coupling conditions as previously described to give 31aꢀb or 32aꢀb. Compound
30 could also be Nꢀalkylated before the thiazole was installed to generate compounds 33aꢀb followed by
thiazole incorporation (Scheme 3). Unfortunately, attempted Nꢀalkylation of compound 29 provided
predominantly the Oꢀalkylated product presumably due to the steric crowding imparted by the C8 methyl
group. An alternative route into these compounds was devised (Scheme 5).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 3. 2ꢀQuinazolinone synthesis.
a
Reagents and conditions: (a) urea, melt (b) POCl , 105 °C; (c) iPr NEt, R NH (see Table 3 for the
3
2
2
2
specific R groups that were incoroporated into the final targets), iPrOH; (d) HOAc, 90 °C; (e) CuI,
2
K CO , 100ꢀ140 °C, PdCl ·dppf·CH Cl , 5ꢀtributylstannyl thiazole, DMF; (f) NaH, MeI, DMF.
2
3
2
2
2
The remaining C2 and C8 substituted quinazolines were synthesized as outlined in Scheme 4. The
starting materials for all of these were the 3ꢀsubstituted anthranilic acids 7ꢀ9. Treatment of these acids
ACS Paragon Plus Environment
8

Page 9 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with either acetic or trifluoroacetic anhydride yielded the corresponding 2ꢀtrifluoromethyl or 2ꢀmethylꢀ
4
Hꢀbenzo[d][1,3]oxazinꢀ4ꢀones 34ꢀ36. Heating these intermediates in the presence of either NH OAc or
4
NH OH, afforded the 2ꢀsubstituted quinazolinꢀ4ꢀones 37ꢀ39 in good yield. The C4 amine could then be
4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
incorporated either by chlorination (POCl ), followed by thermal addition of the desired amine (step d) or
3
via a direct amine coupling reaction (CH CN, DBU, pyBOP, 75 °C – step g), to afford the 4ꢀ
3
aminoquinazolines (it should be noted that for these final targets the trans 4ꢀaminoꢀNꢀ
methylcyclohexylcarboxamide was used as the amine based on potent biochemical data for a close analog
compound 3oo shown in Table 2). The final step was a palladium mediated coupling of either 5ꢀ
1
7
tributylstannylthiazole or 5ꢀ(4,4,5,5ꢀtetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)thiazole with the 6ꢀbromo or 6ꢀ
iodoꢀ4ꢀaminoquinazoline to give the final targets 42ꢀ44.
Scheme 4. C2 and C8 substituted thiazoloquinazoline synthetic routes.
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
N
O
X
X
CO H
X
b
2
NH
a
O
NH2
N
R2
R2
R3
7 R = CF ; R = Me; X = I
R3
R3
3
3
3
3
3
4 R = CF ; R = Me; X = I
3
3
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7 R = CF , X = Br
8 R = CF ; X = I
9
3
3
2
3
3
8 R = Me; R = CF ; X = Br
5 R = Me; R = CF ; X = Br
36 R = Me; R = Me; X = Br
3
2
3
3
2
3
3
3
9 R = Me; R = Me; X = Br
3
2
R = Me, X = Br
3
2
3
CONHMe
O
Cl
N
S
HN
N
N
X
X
d; e or f
NH
c
N
N
N
R2
R2
R2
R3
R3
R3
3
3
7 R = CF ; R = Me; X = I
8 R = Me; R = CF ; X = Br
39 R = Me; R = Me; X = Br
40 R = CF ; R = Me; X = I
41 R = Me; R = Me; X = Br
3
3
2
3
3
2
42 R = CF ; R = Me
3 3 2
43 R = Me; R = CF
3 2 3
44 R = Me; R = Me
3 2
3
2
3
3
2
3
2
g
compound 38
only
CONHMe
f
HN
N
X
N
R2
R3
45 R = Me; R = CF ; X = Br
3
2
3
a
Reagents and conditions: (a) Ac O or TFAA, reflux; (b) NH OAc or NH OH, 150 °C; (c) POCl ,
2
4
4
3
toluene, iPr NEt; (d) trans 4ꢀNH ꢀcyclohexylꢀCONHMe, iPr NEt, CH CN, 70 °C; (e) CuI, K CO , 100ꢀ
2
2
2
3
2
3
1
40 °C, PdCl ·dppf·CH Cl , 5ꢀtributylstannyl thiazole, DMF; (f) CuI, CsF, (Ph P) Pd, 5ꢀ(4,4,5,5ꢀ
2
2
2
3
4
tetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)thiazole, DMF, 90 °C; (g) DBU, pyBOP, trans 4ꢀNH ꢀcyclohexylꢀ
2
CONHMe, CH CN.
3
The final synthetic efforts were directed towards the thiazoloquinolinꢀ2ꢀones. The quinolinꢀ2ꢀone
core was synthesized from appropriately substituted and halogenated anilines. Acylation of the starting
aniline with methyl malonylchloride provided the carboxamides 50ꢀ53, which upon hydrolysis generated
the corresponding carboxylic acids 54ꢀ57. The acids were cyclized under thermal acidic conditions
affording the quinolinꢀ2,4ꢀdiones 58ꢀ61, The 2,4ꢀdiones were then converted to the 2,4ꢀ
ACS Paragon Plus Environment
10

Page 11 of 101
Journal of Medicinal Chemistry
1
2
3
dichloroquinazolines 62ꢀ63 upon treatment with POCl when R equaled H. Selective hydrolysis (HCl,
3
2
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
CH CN, 80 °C) then provides the 4ꢀchloroquinolinꢀ2ꢀones 64ꢀ66. The C4 amine was then attached under
3
thermal conditions (step h) followed by incorporation of the 6ꢀthiazole moiety as before (step g)
generating final targets 76aꢀc and 77aꢀd. It should be noted that incorporation of the amine (step h) via
this route was limited to the thermal conditions due to chemoselective issues between the C4 chloro
moiety and the 6ꢀhalo substituent under the palladium catalyzed conditions used in step i. Alternatively,
the last two steps could be switched (step g first followed by either step h or i). This sequence then
allowed the amine to be attached either under thermal or palladium catalyzed conditions. When the
quinolineꢀ2,4ꢀdiones were Nꢀsubstituted (compounds 60ꢀ61), they could be converted directly to the 4ꢀ
chloroquinolinꢀ2ꢀones 69ꢀ70 upon treatment with either PCl or POCl . The 4ꢀchloroquinolinꢀ2ꢀones were
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
then reacted under the aforementioned conditions (step g) generating the 6ꢀthiazoloꢀ4ꢀchloroquinolinꢀ2ꢀ
ones 73ꢀ74. Compound 75 was synthesized from compound 71 upon treatment with EtI, NaH, DMF –
step j. The two synthetic pathways (NH and Nꢀsubstituted) merged at this point and were converted to
final products, 78aꢀf, 79aꢀd and 80aꢀb (Scheme 5). In some cases compounds, where R equals H, could
1
be directly alkylated to provide their N1 alkylated products (the aforementioned compound 75 is one
example of this chemistry) prior to attachment of the thiazole functionality at C6. As mentioned earlier,
when a substituent larger than hydrogen was placed at C8 the Oꢀalkylated product predominated.
Therefore, for the final targets where C8 was not hydrogen, the desired Nꢀalkyl group was installed on the
starting aniline before cyclization.
Scheme 5. Synthetic route into the 6ꢀthiazoloquinazolinꢀ2ꢀones.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
X
X
X
O
O
b
O
O
c
a
N
OMe
N
OH
NH
R1 R2
R1
R2
R1 R2
50 R1 = H, R2 = H
51 R1 = Me, R2 = H
52 R1 = H, R2 = Me
53 R1 = Me, R2 = Me
X = Br or I
5
5
5
5
4 R = H, R = H
1 2
5 R = Me, R = H
4
6 R = H, R = H
1
2
4
4
4
7 R = Me, R = H
8 R = H, R = Me
9 R = Me, R = Me
1
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
6 R = H, R = Me
1
2
1
2
7 R = Me, R = Me
1
2
1
2
X = Br or I
X = Br or I
R3
Cl
HN
O
Cl
X
X
X
X
d
e
h
R = H
N
Cl
X = Br
N
H
O
N
O
2
N
O
H
R1 R2
R1
R1
R1
62 R = H
63 R = Me
X = Br or I
64 R = H, X = I
65 R = Me, X = Br
66 R = Me, X = I
67a-c R = H
68 R = Me
5
5
6
6
8 R = H, R = H
9 R = Me, R = H
0 R = H, R = Me
1 R = Me, R = Me
1
1
1
1
2
1
1
1
1
2
1
1
2
1
2
g
X = Br or I
g
d or f
R not H
2
R3
Cl
S
N
Cl
S
N
HN
N
X
h or i
g
N
O
N
O
O
R1 R2
R1
R2
R1 R2
71 R = H, R = H
72 R = Me, R = H
73 R = H, R = Me
1 2
4 R = Me, R = Me
1 2
76a-c R1 = H, R2 = H
77a-d R1 = Me, R2 = H
78a-f R1 = H, R2 = Me
79a-d R1 = Me, R2 = Me
80a-b R1 = H, R2 = Et
6
7
9 R = H, R = Me
0 R = Me, R = Me
1
2
1
2
1
2
1
2
j
X = Br or I
7
7
5 R = H, R = Et
1
2
a
Reagents and conditions: (a) methylmalonyl chloride, TEA, EtOAc; (b) THF, MeOH, 2.0 N NaOH; (c)
CH SO H, P O , 70 °C; (d) POCl , 100 °C; (e) HCl 80 °C, CH CN or MsOH, dioxane, H O, 105 °C;
3
3
2
5
3
(conc),
3
2
(
f) PCl , 80 °C; (g) CuI, CsF, (Ph P) Pd, 5ꢀ(4,4,5,5ꢀtetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)thiazole, DMF,
3 3 4
9
0 °C or CuI, K CO , 100ꢀ140 °C, PdCl ·dppf·CH Cl , 5ꢀtributylstannyl thiazole, DMF; (h) CH CN,
2 3 2 2 2 3
R NH (see Table 4 for the specific R groups that were incorporated into the final compounds), 150ꢀ160
3
2
3
°
C, NMP; (i) R NH (see Table 4 for the specific R groups that were incorporated into the final
3
2
3
compounds), Pd dba , BrettPhos, NaOtBu, 110 °C; dioxane; (j) EtI, NaH, DMF
2
3
ACS Paragon Plus Environment
12

Page 13 of 101
Journal of Medicinal Chemistry
1
2
3
Results and Discussion
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
As we were unable to obtain a liganded crystal structure within this class of compounds with
either the human or mouse enzyme under numerous conditions, a binding model was constructed utilizing
a crystal structure of hCD38 E226Q mutant bound with nicotinamide mononucleotide (NMN) (code:
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
4
OGW). For our modeling and docking calculations, the hCD38 structure was prepared by
19
systematically adding hydrogens to both the protein and NMN using a standard inꢀhouse procedure.
Previously, we collected and reported CD38 biochemical and kinetic evidence on compounds similar to
20
this series suggesting that the nitrogen of the thiazole ring may be involved in a reversible, shortꢀlived
catalytic reaction with substrate intermediates produced by CD38. Therefore, we modeled the ribose
substrate intermediate in the hCD38 binding site by removing the nicotinamide ring from
crystallographically bound NMN. With a model of this hCD38/ribose intermediate complex in hand; the
thiazole ring of an initial 3D model of compound 1 was preꢀorientated in the hCD38 catalytic site such
that the nitrogen of that ring was in close proximity to the 1’ꢀposition of the riboseꢀ5ꢀphosphate
21
intermediate. The initial structure of 3b was generated using the CONCORD program. After orientating
the compound, an inꢀhouse docking procedure, MVP, was used to model 1 in the active site of hCD38 in
1
9
the presence of the riboseꢀ5ꢀphosphate intermediate.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 3. Docking model of 3b (in orange) in the hCD38 E226Q mutant protein (in green) bound with
riboseꢀ5ꢀphosphate (PDB code: 4OGW). The modeled ribose substrate intermediate is shown in cyan.
This modeling suggests that the majority of the binding affinity of this compound may be due to a πꢀπ
stacking interaction between the thiazolophenyl ring of the bicyclic core and the indole of W189 (Figure
2). Although not explicitly evident in our binding model, T221 may form a hydrogen bond with the 4ꢀ
amino nitrogen, providing additional binding affinity as several compounds that removed this hydrogen
through the addition of a methyl substituent on this nitrogen were significantly less active, as were
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
oxygen and sulfur heteroatom replacements (data not shown). This model of hCD38 E226Q bound with a
riboseꢀ5ꢀphosphate intermediate was utilized for further docking studies to facilitate inhibitor design,
improve potency and better interpret the SAR of synthesized compounds.
Mouse enzymatic data was collected in parallel with the human data in an effort to identify an
inhibitor that could be used for both acute and chronic rodent efficacy studies. Based on the KO data
mentioned earlier, the mouse was considered a good preꢀclinical species with which to conduct target
validation studies. The biochemical assays were the initial filtering mechanism by which compounds were
progressed with a goal of identifying inhibitors which had IC50’s < 100 nM, although inhibitors which
had potency less than this were still considered if they possessed other properties condusive to in vivo
efficacy studies (e.g. good physiochemical properties and/or good selectivity profiles).
Table 1.
a
a
human IC50 std dev mouse IC50 std dev
compound
R
(
µM) (µM) (µM) (µM)
ACS Paragon Plus Environment
14

Page 15 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
2
ꢀꢀ
ꢀꢀ
1.21
0.298
0.84
1.20
1.32
0.70
3.61
2.33
0.97
4.68
1.23
1.16
12.8
2.15
0.70
0.97
2.69
0.56
0.84
0.46
1.46
12.3
3.94
0.95
0.29
0.037
0.34
0.38
0.27
0.14
0.63
0.76
0.12
0.88
0.15
0.33
3.19
0.4
0.205
0.019
0.046
0.032
0.026
0.019
0.058
0.031
0.031
0.045
0.022
0.017
0.065
0.042
0.027
0.050
0.031
0.024
0.013
0.026
0.026
0.097
0.059
0.024
0.22
b
0.004
0.078
0.010
0.005
0.005
0.025
0.010
0.012
0.010
0.009
0.006
0.022
0.020
0.005
0.012
0.017
0.009
0.003
0.004
0.006
0.034
0.010
0.005
3a
Ph
3b
3ꢀClꢀ4ꢀFꢀPh
4ꢀClꢀPh
3ꢀClꢀPh
2ꢀClꢀPh
4ꢀFꢀPh
3ꢀFꢀPh
2ꢀFꢀPh
4ꢀMeꢀPh
3ꢀMeꢀPh
2ꢀMeꢀPh
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3c
3d
3e
3f
3g
3h
3i
3j
3k
3l
3ꢀCF ꢀPh
3
3
m
4ꢀSO
4ꢀNHSO
3ꢀSO MeꢀPh
2
NH
2
ꢀPh
0.05
0.09
1.27
0.16
0.32
0.08
0.24
2.85
0.41
0.28
3
n
o
p
q
2
MeꢀPh
3
2
3
3
4ꢀCONH ꢀPh
2
4ꢀ(5ꢀoxazole)ꢀPh
4ꢀ(1ꢀimidazole)ꢀPh
4ꢀ(Nꢀγꢀlactam)ꢀPh
2ꢀpyridyl
3
r
s
3
3t
3
u
v
3ꢀpyridyl
3
4ꢀpyridyl
a
b
The standard deviation was determined where the N ≥ 4. N = 3.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1 outlines the initial exploration of the C4 position of the thiazoloquinazolines, where the
focus was probing anilines. The SAR for this set of compounds shows that none have equivalent human
potency when compared to compound 2, although two compounds were within a 2ꢀfold difference
(compounds 3p and 3r). In fact, the potency against the human enzyme stayed fairly flat with respect to
compound 3b, with the exception of compound 3t where there was ~10ꢀfold difference. It was also
noteworthy that many compounds were close in activity to phenyl derivative 3a. The pyridyl analogs 3tꢀv
showed that as the nitrogen atom moved from the 2 to the 4ꢀposition potency against the human enzyme
increased. The docking models support this finding as the nitrogen of the 4ꢀpyridyl compound was much
more solvent exposed than either the 2ꢀ or 3ꢀpyridyl analogs. Another finding was that as the substituent
on the aromatic ring was moved to the ortho position from the para position the potency decreased (see
compounds 3c versus 3e, 3f versus 3h and 3i versus 3k). According to the modeling results, the ortho
substituents result in a nonꢀplanar conformation of the anilino ring which leads to a detrimental
interaction between the ortho group and the hydrophilic cleft near T221 of the hCD38 protein.
Interestingly, the meta substituted compounds 3d and 3g were found to be more potent that their para
counterparts against the human enzyme. It was also noted that the protein tolerated more elaborate
functionality as evidenced by compounds 3mꢀs. Modeling suggests that functionalty can extend towards
solvent coming off the 4ꢀposition perhaps explaining why such a diverse set of functional groups appear
to be tolerated. All of the compounds had IC50’s ≤ 0.050 µM, other than 3e, 3k, 3t and 3u against the
mouse enzyme. The difference between the mouse and human potency could be explained by the fact that
the human enzyme assay used for our high throughput biochemical screening contained nonꢀglycosylated
protein; whereas the mouse enzyme was fully glycosylated (there is 75% identity between the human
mouse protein sequences as calculated based on an alignment of the primary fullꢀlength amino acid
sequences of human and mouse CD38 downloaded from UNIPROT (http://www.uniprot.org)). The
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
9
alignment was calculated using standard procedures from the aforementioned MVP program. It is
unclear why the glycosylation state of the enzyme provides such a wide potency difference between the
two biochemical assays. One hypothesis is that the nonꢀglycosylated human enzyme has more hydrolytic
ACS Paragon Plus Environment
16

Page 17 of 101
Journal of Medicinal Chemistry
1
2
3
activity with the exchange reaction product. Not having the glycosyl residues on the protein opens it up
and allows water into the active site producing less of the exchange reaction product leading to less
inhibition. However, we were able to obtain human kinetic data from fully glycosylated human protein
albeit in a nonꢀhigh throughput assay on select compounds (Table 7).
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. Thiazoloquinazoline biochemical data.
human IC50 std dev mouse IC50 std dev
a
a
a
a
compound
R
ꢀꢀ
(µM)
(µM)
(µM)
(µM)
1
1.21
0.29
0.205
0.019
0.144
0.031
0.012
0.013
0.086
0.051
0.045
0.275
0.296
0.034
0.085
0.007
0.010
0.22
b
2
ꢀꢀ
0.298
7.83
0.037
2.82
0.19
0.083
0.020
0.44
0.37
0.16
0
0.004
0.075
0.014
0.003
0.002
0.022
0.027
0.026
0.094
0.180
0.005
0.007
0.002
0.003
3w
3x
cyclopropyl
cyclobutyl
1.93
3y
cyclopentyl
0.281
0.132
1.68
3z
cyclohexyl
3aa
3bb
3cc
3dd
3ee
2ꢀtetrafuranyl
4ꢀpyranyl
1.06
3ꢀpyranyl
1.03
4ꢀpiperidinyl
3ꢀpiperidinyl
4ꢀpiperidinylꢀNꢀformamide
4ꢀthiopyranyl dioxide
4ꢀmethylcyclohexyl
transꢀ2ꢀmethylcyclohyhexyl
> 30.2
> 30.2
0.149
1.22
0
3
ff
gg
hh
ii
0.036
0.21
0.027
0.011
3
3
0.117
0.140
3
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
jj
kk
ll
cisꢀ2ꢀmethylcyclohexyl
3ꢀmethylcyclohexyl
4ꢀhydroxycyclohexyl
0.049
0.142
0.406
0.078
0.971
0.092
> 30.2
0.498
0.959
8.54
0.006
0.013
0.112
0.026
0.172
0.027
0
0.012
0.015
0.025
0.009
0.037
0.012
0.459
0.040
0.135
1.11
0.002
0.005
0.015
0.002
0.004
0.002
0.094
0.015
0.036
0.20
3
3
3
mm
transꢀ4ꢀCONH
2
2
ꢀcyclohexyl
ꢀcyclohexyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
nn
transꢀ3ꢀCONH
3
oo
pp
qq
trans 4ꢀCONHEtꢀcyclohexyl
cis 4ꢀCO Hꢀcyclohexyl
3
2
3
trans 4ꢀCO Hꢀcyclohexyl
trans ꢀ4ꢀSO Meꢀcyclohexyl
2
2
0.149
0.155
1.35
0.57
4.12
0.30
0.42
4.78
0
3
rr
3
ss
tt
uu
vv
ww
cisꢀ4ꢀ(Nꢀγꢀlactam)cyclohexyl
transꢀ4ꢀ(Nꢀγꢀlactam)cyclohexyl
ethyl
3
1.96
0.077
0.145
0.075
0.066
0.189
0.298
0.024
0.046
0.023
0.022
0.060
0.089
3
11.4
3
isopropyl
3.35
3
tꢀbutyl
2.3
c
3
xx
yy
2ꢀhydroxyethyl
2ꢀmethoxyethyl
23.4
3
> 30.2
a
During curve fitting if the desired IC50 value was outside the test concentration range, ActivityBase XE
reports the value with a < or > modifier which was not included in the statistics calculation. The IC50 and
standard deviation were determined where the N ≥ 4. N = 3. N = 3. Here there were 2 values reported by
ActivityBase XE as > 30.2 µM.
b
c
A structureꢀactivity relationship for analogs in which the C4 position was an aliphatic amine is
outlined in Table 2. In general, these compounds achieved equivalent if not superior potency to
compound 3b. In fact, compounds 3yꢀz, 3ff, 3hhꢀkk, 3mm and 3oo had potency equivalent or better than
compound 2. Many of these compounds were very hydrophobic in nature. Interestingly, the urea or
carboxamides (3ff, 3mm and 3oo) showed very good potency as well. The trans/cis stereochemistry also
appeared to be important on the 4ꢀposition of the cyclohexyl ring as demonstrated by comparing
compounds 3pp to 3qq and 3ss to 3tt, although when the substituent was at the 2ꢀposition and
ACS Paragon Plus Environment
18

Page 19 of 101
Journal of Medicinal Chemistry
1
2
3
hydrophobic, (e.g. 3ii and 3jj) the same trend no longer held. The primary amide functionality found in
compound 3mm, which proved to be one of the most potent inhibitors within this series, was consisitent
with the SAR from the aromatic series (Table 1, 3p). In fact, our docking models of these 4ꢀamido
substituted compounds suggest that this primary amide may be more solvent exposed than others and
could potentially hydrogen bond with the alcohol side chain of S224 in the hCD38 protein, a residue near
its surface and important for binding some substrates. The directionality of the amide was also important
based on a > 10ꢀfold potency difference between compounds 3mm and 3nn. The heterocyclic compounds
exhibited mixed potency. Compounds 3aaꢀcc and 3gg possessed potency similar to compound 3b,
however, the piperidines, 3dd and 3ee, were completely inactive against the human enzyme. This result
suggests that CD38 does not tolerate a positive charge in the space with which these rings reside;
interestingly our model did not predict this result. All nonꢀcyclic amines turned out to be less potent than
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3b. All of the saturated compounds, with the exception of 3ss, were found to have IC50’s < 0.5 µM
against the mouse enzyme with many showing low nanomolar potency. As mentioned earlier, a potent
mouse inhibitor was necessary in order to conduct in vivo efficacy studies. Many of these inhibitors
achieved the desired potency (IC50 < 100 nM).
Having established some SAR at the 4ꢀposition on the quinazoline ring, it was expanded to
explore both the C2 and C8 positions. The biochemical data for these compounds is shown in Table 3.
Table 3. C2 and C8 substituted thiazoloquinazoli(no)ne biochemical data.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
human
IC50
mouse
IC50
std dev
std dev
a
a
a
a
compound R1
R2
R3
H
X
N
N
N
N
N
N
(µM)
(µM)
(µM)
(µM)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
19a
19b
20a
20b
OMe
OMe
3ꢀClꢀ4ꢀFꢀPh
4.71
1.48
0.22
0.009
0.004
0.003
0.048
0.003
0.002
0.0008
b
c
4ꢀCONH
ꢀPh
H
1.16
0.004
2
CF
CF
3ꢀClꢀ4ꢀFꢀPh
4ꢀCONH ꢀPh
H
1.59
0.24
0.139
0.008
0.010
0.004
3
H
0.263
0.206
0.028
0.114
0.036
0.009
3
2
2
1
Me
Me
3ꢀClꢀ4ꢀFꢀPh
3ꢀClꢀ4ꢀFꢀPh
H
3
1a
OH
trans 4ꢀMeOꢀ
cyclohexyl
3
1b
Me
OH
N
0.026
0.007
0.001
0.0003
trans 4ꢀ
OCH CH OMeꢀ
2
2
d
3
2a
H
H
cyclohexyl
OH
NMe
NMe
N
0.064
0.053
0.042
0.245
0.050
0.012
0.013
0.006
0.083
0.003
0.008
0.008
0.072
0.037
0.007
0.0002
0.001
0.025
0.013
0.002
trans 4ꢀMeOꢀ
cyclohexyl
d
32b
42
OH
trans 4ꢀCONHMeꢀ
cyclohexyl
CF
Me
Me
Me
3
trans 4ꢀCONHMeꢀ
cyclohexyl
43
CF
3
N
trans 4ꢀCONHMeꢀ
cyclohexyl
44
Me
N
a
During curve fitting if the derived IC50 value was outside the test concentration range,
ActivityBase XE reports the value with a < or > modifier which was not included in the
statistics calculation. The standard deviation was determined where the N ≥ 4. N = 2. N = 3.
b
c
d
Drawn as a hydroxyl residue, but in the keto form in the final product due to the N1 methyl
substituent.
The anilines/amines, which provided some of the more potent inhibitors in the earlier examples
described in Tables 1 and 2 or those that were were deemed to have more desirable physiochemical
properties were used to assess how substitution at C2 and/or C8 impacted the potency against the human
ACS Paragon Plus Environment
20

Page 21 of 101
Journal of Medicinal Chemistry
1
2
3
enzyme. Compounds 19a and b, where R = OMe, both had IC50’s greater than compounds 3b and 3p
1
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
respectively. However, compounds 20a and b, where R = CF , had potencies either similar to or better
1
3
than their comparator compounds 3b and 3p. Compound 21, containing the C8 methyl, was the most
potent C8 substituent examined within the 3ꢀClꢀ4ꢀFꢀphenyl aniline series with an IC50 = 206 nM (> 5ꢀ
fold potency increase versus 3b). Compounds 31aꢀb and 32aꢀb, containing the C2 hydroxy substituent
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
drawn as the enol tautomer for clarity), dramatically increased the human potency with compound 31a
demonstrating a > 40ꢀfold improvement versus compound 3b. The remaining compounds in Table 3 (42ꢀ
4), all had human IC50’s ≤ 245 nM. The placement of a trifluoromethyl or methyl substituent at C8,
4
compounds 42 and 44, provided potent inhibitors (≥ 24ꢀfold increase versus 3b against the human
enzyme). Interestingly, compound 43, where the CF moiety was placed at the C2 position, lost ~ 5ꢀfold
3
in potency when compared to compounds 42 and 44.
Table 4. 6ꢀThiazoloquinolinꢀ2ꢀone biochemical data.
human IC50 std dev mouse IC50 std dev
a
a
a
a
compound R1
R2
R3
(nM)
(nM)
(nM)
(nM)
76a
77a
78a
H
Me
H
H
trans 4ꢀCONHMeꢀ
cyclohexyl
3.4
0.9
0.9
0.19
b
c
H
trans 4ꢀCONHMeꢀ
cyclohexyl
1.4
0.53
2.5
2.6
34
< 0.18
1.1
0
Me
trans 4ꢀCONHMeꢀ
cyclohexyl
4.8
5.3
58
0.52
0.42
4.0
7
8
9a
0a
Me Me
trans 4ꢀCONHMeꢀ
cyclohexyl
0.78
6.1
H
Et
trans 4ꢀCONHMeꢀ
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cyclohexyl
7
7
7
7
6b
7b
8b
9b
H
Me
H
H
H
4ꢀpyranyl
4ꢀpyranyl
4ꢀpyranyl
4ꢀpyranyl
61
55
89
52
6
16
16
22
15
1.4
11
4.5
15
6
4.2
1.5
Me
7.2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Me Me
3.7
7
7
7
7
6c
7c
8c
9c
H
Me
H
H
H
trans 4ꢀ
1.3
0.47
OCH CH OMeꢀ
2
2
cyclohexyl
d
trans 4ꢀ
3.9
7.3
10
2.1
1.8
2.9
15
0.62
1.9
0.46
0.68
0.38
1.8
OCH CH OMeꢀ
2
2
cyclohexyl
Me
trans 4ꢀ
OCH CH OMeꢀ
2
2
cyclohexyl
Me Me
trans 4ꢀ
1.6
OCH CH OMeꢀ
2
2
cyclohexyl
80b
77d
H
Et
trans 4ꢀ
54
5.2
OCH CH OMeꢀ
2
2
cyclohexyl
d
Me
H
H
trans 4ꢀOMeꢀ
cyclohexyl
3.8
13
1.3
2.1
7.1
0.65
1.6
0.47
0.49
0.38
7
7
8d
9d
Me
trans 4ꢀOMeꢀ
cyclohexyl
e
f
Me Me
trans 4ꢀOMeꢀ
cyclohexyl
15
1.4
7
8e
8f
H
H
Me trans 4ꢀOHꢀcyclohexyl
Me
14
5.6
2.5
1.4
7
H
10773
1836
163
28.1
a
During curve fitting if the derived IC value was outside the test concentration range, ActivityBase
5
0
XE reports the value with a < or > modifier which was not included in the statistics calculation. The
b
standard deviation was determined where the N ≥ 4. N = 7 and 2 additional values were reported by
c
ActivityBase XE as < 0.51 nM. N = 7 and 2 additional values not included in the stats reported by
d
ActivityBase XE as 0.18 and 0.36 nM. Here there were 2 values reported by ActivityBase XE as <
e
f
0
.18 nM. 1 value was reported by ActivityBase XE as < 1.51 nM. Here there was 1 value reported
by ActivityBase as < 0.18 nM.
ACS Paragon Plus Environment
22

Page 23 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
We felt that compound 31a now had the requisite potency at both human and mouse enzymes,
particularly given the C4 substituent was not optimal based on our SAR to date. However, 31a did not
have good physicochemical properties, suffering from poor solubility and permeability (see Table 5 for
specifics). Thus we sought to incorporate substituents that maintained potency while improving the
physiochemical properties necessary for conducting in vivo studies. In particular, we identified four
amines which we found to possess not only very good inhibitory properties, but good drugꢀlike properties
as well. Although many aromatic amines showed very good potency, none had sufficiently good
physiochemical properties needed for in vivo studies. All of the 6ꢀthiazoloquinolinꢀ2ꢀones shown in Table
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
demonstrated very good potency against both the human and mouse enzymes with the exception of
compound 78f, which clearly shows the importance of the substitution off the amine at C4. In terms of the
R substituent, both hydrogen and methyl proved to be equipotent at CD38, although the methylated
2
compounds were in general better physiochemically. For example, compounds 76a and 76c had poorer
permeability and solubility than their methylated analogs 78a and 78c (see Table 5). When R was ethyl,
2
potency against the human enzyme was reduced when compared to the NH compound (see 80a versus
7
6a and 80b versus 76c). The C2 and C8 methyl analogs were also very potent against both enzymes (see
compounds 79aꢀd).
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 4. Docking model of 78c (in magenta) in the hCD38 E226Q mutant protein (in green) bound with
riboseꢀ5ꢀphosphate (PDB code: 4OGW). The modeled riboseꢀ5ꢀphosphate substrate intermediate is
shown in cyan.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
As was found from earlier compounds, the trans 4ꢀNꢀmethylcarboxamide cyclohexylamine
containing compounds 76aꢀ79a and 80a provided very potent inhibitors of both human and mouse
enzyme (IC50’s ≤ 58 nM versus human and ≤ 6.1 nM versus mouse). Several of the trans 4ꢀalkoxy
cyclohexylethers were found to provide potent inhibitors as well, with compounds 77cꢀd showing IC50’s
<
4 nM on the human enzyme. The trans 4ꢀhydroxycyclohexylamine analog 78e was similar in potency to
compounds 78cꢀd, demonstrating that the additional functionality was not necessary for binding affinity.
The C4 pyranyl amines 76bꢀ79b, although providing potent inhibitors, were in general less potent than
either the ether or carboxamide C4 amines. The docking of compound 78c (Figure 4) suggests that some
of the additional binding affinity seen with compounds having the 4ꢀmethoxyethoxyl moiety may arise
from increased hydrophobic interactions between this C4 group and W176. The 2ꢀketo residue’s ability to
form a hydrogen bond with the side chain amide of N183 appears unlikely as the distance is too far (4.8
and 4.7 angstroms to the N183 amide and carboxy groups, respectively), the angle is bad, and the amide
group of N183 is tied up in a critical loopꢀstabilizing hydrogen bond with a backbone residue (C180) and
thus, is not available for interacting with the inhibitor. As our goal was to identify a molecule which was
suitable for in vivo PK studies, those compounds with sufficient potency for physiochemical property
profiling had their cell permeability, solubility and hydrophobicity measured. These data are shown in
Table 5.
Table 5. Measured physiochemical properties on a select group 6ꢀthiazoloquin(az)olinꢀ2ꢀones.
a
a
a
compound permeability (nm/sec)
CLND (µM)
11
ChromLogD pH = 7.4
31a
32a
ND
30
4.19
2.47
≥ 429
ACS Paragon Plus Environment
24

Page 25 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
2b
47
< 3.0
52
≥ 410
74
2.51
1.45
2.40
7
6a
6c
7
39
b
b
c
7
7d
8a
8b
8c
202
36
2.85
2.03
2.37
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
9.9
361
350
7
200
d
d
e
7
162
≥ 364
3.22
7
9a
9c
21
339
2.34
f
b
f
7
195
≥ 323
3.54
3.47
f
f
f
79d
300
≥ 340
a
b
c
d
e
f
N = 1. N = 4. N = 3. N = 6. N = 5. N = 2.
The criteria for progression of compounds into PK studies were as follows: permeability (Madinꢀ
Darby canine kidney (MDCK) cells) ≥ 100 nm/sec, solubility (measured by chemiluminescent nitrogen
detection (CLND)) ≥ 100 µM and ChromLogD values < 4.0. All of the compounds in Table 5 met the
ChromLogD criteria (compound 31a was inserted as a comparator and did not meet these criteria as
indicated earlier). However, compounds 32aꢀb, 76a and c, 77d, 78a and 79a failed to meet either the
permeability and/or solubility filters prescribed. The remaining four compounds 78b, 78c, 79c, and 79d
passed all three criteria and thus were evaluated in mouse PK studies. The key in vivo data are shown in
Table 6.
Table 6. Mouse in vivo PK data for compounds 78b, 78c, 79c and 79d.
dose
Cl
Vdss
t½
Cmax
DNAUC
compound
8b
(mg/kg) route
(mL/min/kg)
(L/kg) (hr) %F (ng/mL)
(ng/hr/mL)
a
c
7
30
1
po
ND
1.6
ND
ND
b
iv
44.7
1.2
1110
371
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
8c
9c
30
2
po
iv
3.5
87
46
89
10800
2980
3810
2500
6250
1480
685
779
930
2030
443
497
21
8.9
1.3
0.32
1.1
5.3
1.3
7
10
1
po
iv
0.46
3.2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
79d
30
2
po
iv
33.4
1.2
a
b
p.o. formulation conditions: 2% DMSO:5% solutol in 93% (20%SBEꢀCD, CB pH 3.0); i.v. formulation
conditions: 10% DMSO:10% PEG400:80% (20% SBEꢀCD, citrate buffered pH 6.0); A DNAUC could
not be calculated as the compound exceeded the analytical range.
c
The PK data for the four compounds showed that compounds 78b and 79d had clearances that
were in the moderate to high range, and thus were not chosen for the in vivo studies. Compounds 78c and
79c had better clearance values with, 79c being the best. However, 78c had a higher volume of
distribution at steady state (Vdss) (1.3) and a longer terminal halfꢀlife (both i.v. and p.o.) than 79c. Based
on its overall profile, compound 78c was selected to carry out in vivo studies. In an effort to maximize the
oral exposure of 78c, a number of formulations were evaluated and it was found that a 5% solutol:95%
(
(
20% SBEꢀCD, citrate buffered pH 3.0) vehicle increased the dose normalized area under the curve
DNAUC) from 685 ng/hr/mL to 10525 ng/hr/mL (> 15ꢀfold increase). Also, human wild type (WT)
enzymatic data was collected in an attempt to better define the binding affinity of these compounds versus
the earlier reported biochemical data in Table 4. As can be seen in Table 7, compounds 78b, 78c, 79c and
7
9d all demonstrated low to subnanomolar Ki’s against the fully glycosylated protein.
Table 7. Potency data for compounds 78b, 78c, 79c and 79d against both the biochemical and wild
type (WT) human CD38 enzyme.
human IC50 human Ki WT
a
b
compound
78b
(nM)
enzyme (nM)
89
14
ACS Paragon Plus Environment
26

Page 27 of 101
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
7
8c
9c
7.3
10
15
0.3
0.7
7
9d
0.45
a
b
Values from Table 4. N = 1.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Because we did not have a robust cellular assay that allowed determination of CD38ꢀmediated
increases in intracellular NAD, we developed an in vivoꢀbased method for determining changes in the
levels of NAD in various tissues. The diet induced obese (DIO) mouse (60% high fat diet) was utilized
because this model demonstrated a deficit in steadyꢀstate NAD levels relative to lean controls. Thus an in
vivo study was performed with compound 78c looking at changes in NAD levels versus control animals
in liver and muscle (gastrocnemius) tissue at 2 and 6 hours after a 30 mg/kg dose. As can be seen in
Figure 5, NAD levels were significantly increased at both the 2 and 6 hour time points in liver (536% and
246% respectively over control animals). The NAD levels were also elevated in gastrocnemius muscle,
although to a smaller extent (121% and 141% respectively versus lean controls). The lower NAD levels in
muscle could be attributed to lower drug levels in this tissue versus liver (data not shown). The NAD
levels returned to controls once drug levels dropped below the IC50 (appproximately 12ꢀ15 h in liver and
8ꢀ10 h in muscle – data not shown). This study validates the hypothesis that inhibition of CD38 leads to
modulation of NAD, the endogeneous substrate, in an in vivo setting. Having demonstrated that CD38
inhibition can elevate NAD levels in vivo, compound 78c should provide a valuable tool for evaluation of
CD38ꢀmediated pharmacology in animal models of disease in which a NAD deficiency exists.
Figure 5. NAD levels in liver and gastrocnemius tissue versus lean controls after treatment with
compound 78c.
2
hr postꢀdose
6 hr postꢀdose
78c (30 mg/kg)
Tissue
Vehicle
78c (30 mg/kg)
pꢀvalue
Vehicle
pꢀvalue
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 101
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Gastrocnemius
Liver
100 ± 18
100 ± 44
121 ± 17
536 ± 31
0.2122
0.0001
100 ± 3
141 ± 25
246 ± 14
0.0520
0.0003
100 ± 15
Values presented as percent of vehicle control ± standard deviation.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
5
4
3
2
1
00
00
00
00
00
00
0
*
Vehicle Control
8c
7
* p < 0.05
*
Gastroc 2 hr
Gastroc 6 hr
Liver 2 hr
Liver 6 hr
Conclusion
In summary, we were able to identify a series of 6ꢀthiazoloquin(az)olinꢀ2ꢀoneꢀbased CD38
inhibitors with heretofore unprecedented potency through detailed SAR around screening hit 1. In
particular, compound 78c was found to not only be a very potent inhibitor, but also had suitable
physiochemical properties to serve as a robust tool for in vivo studies. In addition, we demonstrated that
compound 78c was able to elevate NAD in both liver and gastrocnemius tissue when dosed orally in a
ACS Paragon Plus Environment
28

Page 29 of 101
Journal of Medicinal Chemistry
1
2
3
DIO mouse. The identification and development of these potent small molecule inhibitors will enable a
detailed analysis of how elevating NAD in NAD deficit states affects whole body physiology and more
importantly whether this approach may be a viable strategy to treat diseases where elevation of NAD may
provide therapeutic benefit.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
All commercial chemicals and solvents were reagent grade and were used without further purification
unless otherwise specified. The following solvents and reagents have been abbreviated: tetrahydrofuran
(
THF), diethyl ether (Et O), dimethyl sulfoxide (DMSO), ethyl acetate (EtOAc), dichloromethane
2
(CH Cl or DCM), trifluoroacetic acid (TFA), formic acid (FA), N,Nꢀdimethylformamide (DMF),
2
2
methanol (MeOH), 1,2ꢀdimethoxyethane (DME), Nꢀmethylpyrrolidine (NMP), acetonitrile (CH CN or
3
ACN or MeCN), chloroform (CHCl ), phosphorous oxychloride (POCl ), magnesium sulfate (MgSO ),
3
3
4
triethylamine (Et N), 2ꢀpropanol (iPrOH), N,Nꢀdiisopropylethylamine (iPr NEt), sodium hydroxide
3
2
(
NaOH), tꢀbutylmethylether (TBME), acetic acid (AcOH or HOAc), ethanol (EtOH), diꢀtꢀ
butyldicarbonate (BOC O), sodium sulfate (Na SO ), N,Nꢀdimethylacetamide (DMA), sodium
2
2
4
bicarbonate (NaHCO ), potassium carbonate (K CO ), 1ꢀ[bis(dimethylamino)methylene]ꢀ1Hꢀ1,2,3ꢀ
3
2
3
triazolo[4,5ꢀb]pyridinium 3ꢀoxid hexafluorophosphate (HATU), azobisisobutyronitrile (AIBN), 4ꢀ(2ꢀ
hydroxyethyl)piperazineꢀ1ꢀethanesulfonic acid (HEPES), dithiothreitol (DTT). All reactions except those
in aqueous media were carried out with the use of standard techniques for the exclusion of moisture.
Reactions were monitored by thinꢀlayer chromatography (TLC) on 0.25 mm silica gel plates (60Fꢀ254, E.
Merck) and visualized with UV light, cerium molybdate, or 5% phosphomolybdic acid in 95% ethanol.
Final compounds were typically purified either by flash chromatography on silica gel (E. Merck 40ꢀ63
mm), radial chromatography on a Chromatotron using prepared silica gel plates, on a Biotage Horizon or
ISCO Combiflash pump and fraction collection system utilizing prepacked silica gel.
ACS Paragon Plus Environment
29