Structure-based design of potent and selective inhibitors of the metabolic kinase PFKFB3

Article abstract of DOI:10.1021/acs.jmedchem.5b00352

A weak screening hit with suboptimal physicochemical properties was optimized against PFKFB3 kinase using critical structure-guided insights. The resulting compounds demonstrated high selectivity over related PFKFB isoforms and modulation of the target in

Full text of DOI:10.1021/acs.jmedchem.5b00352

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN
Article
Structure-based design of potent and selective
inhibitors of the metabolic kinase PFKFB3
Scott Boyd, Joanna Brookfield, Susan E. Critchlow, Iain Cumming, Nicola Curtis, Judit Debreczeni, Sébastien
Louis Degorce, Craig Donald, Nicola J Evans, Sam Groombridge, Philip Hopcroft, Neil Jones, Jason Grant
Kettle, Scott Lamont, Hilary Lewis, Philip A MacFaul, Sheila Mcloughlin, Laurent Rigoreau, James Smith,
Steve St-Gallay, Julie Stock, Andrew Turnbull, Edward Wheatley, Jon Winter, and Jonathan Wingfield
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00352 • Publication Date (Web): 07 Apr 2015
Downloaded from http://pubs.acs.org on April 11, 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 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structureꢀbased design of potent and selective
inhibitors of the metabolic kinase PFKFB3
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
Scott Boyd,† Joanna L. Brookfield,‡ Susan E. Critchlow,† Iain A. Cumming,‡ Nicola J. Curtis,†
Judit Debreczeni,† Sébastien L. Degorce,† Craig Donald,† Nicola J. Evans,§ Sam Groombridge,†
Philip Hopcroft,† Neil P. Jones,§ Jason G. Kettle,†* Scott Lamont,† Hilary J. Lewis,† Philip
MacFaull,† Sheila B. McLoughlin,‡ Laurent J. M. Rigoreau,‡ James M. Smith,‡ Steve StꢀGallay,†
Julie K. Stock,§ Andrew P. Turnbull,§ Edward R. Wheatley,§ Jon Winter,† Jonathan Wingfield†
†Oncology Innovative Medicines Unit, AstraZeneca, Mereside, Alderley Park, Macclesfield,
Cheshire SK10 4TG, United Kingdom
‡CRT Discovery Laboratories, Jonas Webb Building (B910), Babraham Research Campus,
Cambridge, CB22 3AT, United Kingdom
§CRT Discovery Laboratories, Wolfson Institute for Biomedical Research, University College
London, The Cruciform Building, Gower Street, London, WC1E 6BT, United Kingdom
ABSTRACT
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 46
1
2
3
4
5
6
7
8
9
A weak screening hit with suboptimal physicochemical properties was optimized against
PFKFB3 kinase using critical structureꢀguided insights. The resulting compounds demonstrated
high selectivity over related PFKFB isoforms and modulation of the target in a cellular context.
A selected example demonstrated exposure in animals following oral dosing. Examples from
this series may serve as useful probes to understand the emerging biology of this metabolic
target.
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
INTRODUCTION
In the transition to a malignant phenotype and in response to an increased energy demand driven
by increased rates of proliferation, cancer cells generate energy by glycolysis in a process
marked by increased glucose uptake and lactate production. This switch from the oxidative
phosphorylation observed in normal cells was first described by Otto Warburg over half a
18
century ago,1 and is the basis on which the nonꢀinvasive diagnostic imaging technique FDGꢀ
PET relies.2 Phosphofructokinase 1 (PFK1) catalyses the conversion of fructoseꢀ6ꢀphosphate
(F6P) and ATP to fructoseꢀ1,6ꢀbisphosphate (Fꢀ1,6ꢀBP) and ADP. This irreversible reaction is
one of the rate limiting steps of glycolysis, and the activity of this enzyme is regulated by a
number of cellular metabolites,3 including ATP/AMP ratio, citrate, pH and the allosteric
activator fructoseꢀ2,6ꢀbisphosphate (Fꢀ2,6ꢀBP),4 which is the product of phosphofructokinase 2
(PFK2). PFK2 is a biꢀfunctional 6ꢀphosphofructoꢀ2ꢀkinase/fructoseꢀ2,6ꢀbiphosphatase.5,6 Both
kinase and phosphatase reactions are catalysed within the same polypeptide chain, which
contains an Nꢀterminal kinase domain and a Cꢀterminal phosphatase domain. Several isoforms
of PFK2 have been identified, and are encoded by 4 genes, PFKFB1ꢀ4. Of these, PFKFB3 is
ACS Paragon Plus Environment
2

Page 3 of 46
Journal of Medicinal Chemistry
1
2
reported to have the highest kinase to phosphatase ratio,7 approximately 700 times in favour of
the production of Fꢀ2,6ꢀBP.
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
PFKFB3 is overꢀexpressed in a wide variety of cancers including breast, prostate, colon,
astrocytoma and ovarian cancers, and its expression and/or activity is found to correlate strongly
with aggressiveness/poor prognosis in colon, breast, ovarian and thyroid tumors.8 PFKFB3
protein is highly phosphorylated in human cancer tissue when compared with corresponding
normal tissue, suggesting differential regulation in a cancer setting.9 Further evidence reveals
that PFKFB3 is essential for Rasꢀdependent glycolysis and transformation in lung fibroblast
models.10 High glycolytic flux is essential for tumor growth in hypoxic conditions and a further
mode of PFKFB3 regulation is through the hypoxiaꢀinducible factorꢀ1 (HIFꢀ1) pathway. HIFꢀ1
binds to the PFKFB3 promoter and upꢀregulates expression in response to hypoxia both in vitro
and in vivo.11,12 High glycolytic activity is also observed in certain normal tissue or organs with
high glucose demands (i.e. brain) and there is some evidence that PFKFB3 may have roles in
rapidly proliferating normal cells (i.e. astrocytes and Tꢀlymphocytes).13,14 Therefore as with
many metabolism targets it is important to consider the impact of inhibition of PFKFB3 in
normal tissues.
There are limited reports on potent and selective inhibitors of PFKFB3 kinase. The most widely
studied agent is undoubtedly 3ꢀ(3ꢀpyridinyl)ꢀ1ꢀ(4ꢀpyridinyl)ꢀ2ꢀpropenꢀ1ꢀone 1 (termed 3PO),15
and optimization has recently led to a more potent analogue 2 being reported (PFK15).16
A
range of biological activity has been ascribed to this series including reduction in Fꢀ2,6ꢀBP
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 46
1
2
3
4
5
6
7
8
9
levels, inhibition of glucose uptake and lactate production, and induction of apoptosis in cancer
cell lines both in vitro and in vivo. It is unclear however whether all of these effects are
attributable solely to direct effects on PFKFB3 kinase since the potency of the observed cellular
phenotypes appear in some cases at odds with its modest kinase inhibitory activity. 3PO itself
has been shown to display both competitive and uncompetitive inhibition with respect to isolated
PFKFB3 kinase, and whilst it is true that expression of PFKFB3 leading to increased
intracellular levels of Fꢀ2,6ꢀBP can protect Jurkat cells from the cytotoxic effects of 3PO,
additional offꢀtarget effects cannot be ruled out. A recent report disclosing pyridazinone
inhibitors of PFKFB3 such as 3,17 highlights that neither 1 nor 3 showed glycolytic pathway
cellular activity that could be separated from cytotoxicity at the concentrations used, and that the
reported enzyme inhibitory activity of 2 itself could not be validated (IC50 > 1000 µM versus 0.2
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
µM reported). In our hands, 3PO itself is inactive in a PFKFB3 kinase assay (IC50 > 100 µM)
and no crystal structure is available confirming binding of 3PO or analogues to PFKFB3 kinase.
Herein we report our own efforts to obtain potent and selective inhibitors of PFKFB3 kinase
which may serve as useful tools to further understand the biology of this interesting target.
, 3PO
1
, PFK15
2
3
Figure 1. Chemical structures of known PFKFB3 inhibitors
ACS Paragon Plus Environment
4

Page 5 of 46
Journal of Medicinal Chemistry
1
2
3
4
CHEMISTRY
5
6
7
8
9
A number of diverse approaches were taken to synthesize the compounds studied, with routes to
representative examples summarized in Schemes 1ꢀ5 (full details are available for all compounds
in the Supplementary Information). The syntheses of 2ꢀaminoꢀ5ꢀphenoxyindoles such as 11 is
illustrated in Scheme 1 with the initial step being a SNAr reaction between 5ꢀchloroꢀ2ꢀ
nitroaniline 4 and the phenol 5. This reaction proceeds in good yield using sodium
hydride/dimethylacetamide conditions. Standard diazotization conditions give rise to the oꢀ
chloronitro compound 7 which can then undergo reaction with malononitrile which upon in situ
sodium dithionite reduction cyclizes to generate the key indole 8. Deacetylation under acidic
conditions gave the lateꢀstage intermediate 9 which could then be converted to the final
compounds by Nꢀalkylation of the indole and aniline acylation – these steps may be carried out
in either order.
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
Scheme 1. Synthesis of 2ꢀAminoꢀ5ꢀphenoxyindole 11.
(i)
(iii)
4
5
6 R = NH
7 R = Cl
2
8 R = AcNH
9 R = NH
(ii)
(iv)
2
(v)
(vi)
11
10
Reagents and conditions: (i) NaH, DMA, 5ꢀ80°C; (ii) conc. HCl, NaNO2, H2O, AcOH, CuCl, 5ꢀ50°C; (iii)
malononitrile, NMP, 10M NaOH, room temperature; (iv) conc. HCl, MeOH, H2O, 55°C; (v) MeI, Cs2CO3, DMF,
80ꢀ120°C. (vi) NꢀBOCꢀglycine, HATU, DCM, DIPEA, room temperature, then TFA, DCM, room temperature.
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 46
1
2
3
4
5
6
7
8
9
Deletion of the amino functionality from the indole required a different approach, illustrated for
compound 26 (Scheme 2) which begins with the 3ꢀcyanoindole 12. Protection of the indole
nitrogen with a 2ꢀ(trimethylsilyl)ethoxymethyl (SEM) protecting group followed by subsequent
debenzylation using standard conditions yielded the 5ꢀhydroxyindole 20 which could then be
used in a subsequent SNAr reaction to afford the nitro compound 21. Reduction to aniline 22
was followed by amide formation to give doubly protected indole 23. Selective removal of the
SEM group allowed introduction of the alkyl indole Nꢀsubstituent, with compound 26 then
isolated after removal of the BOC protecting group.
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
Scheme 2. Synthesis of desꢀaminoindole 26.
N
N
N
BnO
HO
O
(ii)
(iii)
N
N
N
R
R
SEM
SEM
20
R = H
R = SEM
R = NO2
R = NH2
12
19
21
22
(iv)
(i)
(v)
N
N
O
O
O
O
(vii)
N
N
N
H
N
H
R
N
N
O
R
O
R = Boc
R = H
R = SEM
R = H
25
26
23
24
(viii)
(vi)
Reagents and conditions: (i) SEMꢀCl, NaH, THF, 5ꢀ25°C; (ii) 10% Pd/C, H2, MeOH, room temperature; (iii)1ꢀ
ACS Paragon Plus Environment
6

Page 7 of 46
Journal of Medicinal Chemistry
1
2
3
4
+
fluoroꢀ4ꢀnitrobenzene, K2CO3, DMF, room temperature; (iv) NH4 Clꢀ, Zn, MeOH, room temperature; (v) BocꢀProꢀ
5
6
OH, HATU, DMF, DIPEA, room temperature; (vi) TBAF, reflux; (vii) iꢀBuBr, Cs2CO3, DMF, room temperature;
7
8
(viii) TFA, DCM, room temperature.
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 scaffold change to the indazole series required the development of a different synthesis
(Scheme 3) which involves diazotization of aniline 27 under standard conditions followed by
cyclization to give indazole 28. Protection of the indazole NH was similarly required prior to
ether formation under copper catalyzed conditions to give 31 where the proline side chain was
already installed. Similar deprotection and alkylation conditions to those used in the indole
series were then employed to yield indazole 33.
Scheme 3. Synthesis of indazole 33.
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 46
1
2
3
4
5
6
7
(i)
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
27
R = H
R = SEM
28
29
(ii)
(iii)
(v)
R = SEM
R = H
30
31
R = Boc
R = H
32
33
(iv)
(vi)
Reagents and conditions: (i) conc. HCl, NaNO2, H2O, ꢀ50ꢀ0°C; (ii) SEMꢀCl, NaH, THF, ꢀ20ꢀ25°C; (iii) (S)ꢀ2ꢀ(4ꢀ
hydroxyꢀphenylcarbamoyl)ꢀpyrrolidineꢀ1ꢀcarboxylic acid tertꢀbutyl ester, CuI, Cs2CO3, N,Nꢀdimethyl glycine HCl
salt, DMF, 1,4ꢀdioxane, 140°C; (iv) TBAF, 40°C; (v) 4ꢀ(chloromethyl)ꢀ3,5ꢀdimethylꢀisoxazole, NaH, THF, ꢀ5ꢀ
25°C; (vi) TFA, DCM, room temperature.
Replacement of the indole nitrile group with Nꢀmethylpyrazole required the synthesis of a late
stage bromo intermediate 38 which could then undergo a Suzuki coupling to introduce the
required functionality (Scheme 4).
5ꢀHydroxyindole 34 was condensed with 4ꢀ
fluoronitrobenzene to give phenoxy indole 35. 3ꢀBromoindole 38 was obtained in a sequence
involving reduction to aniline 36, amide capping to give 37, then bromination with NBS to yield
38. Suzuki coupling and deprotection allowed access to target compound 40. For synthesis of
ACS Paragon Plus Environment
8

Page 9 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
Nꢀmethyl analogue 43, the Suzuki coupling was conducted after prior methylation of the indole
to give 41.
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
Scheme 4. Synthesis of 3ꢀpyrazole indoles 40 and 43.
(i)
(iii)
34
R = NO
R = NH2
35
36
2
R = H
R = Br
37
38
(ii)
(iv)
(vii)
(v)
(vi)
43
R = Boc
R = H
39
40
(vi)
R = Br
R = 1-methylpyrazol-3-yl
41
42
(v)
Reagents and conditions: (i)1ꢀfluoroꢀ4ꢀnitrobenzene, Cs2CO3, DMF, room temperature; (ii) 10% Pd/C, H2, MeOH,
room temperature; (iii) BocꢀProꢀOH, HATU, DMF, DIPEA, room temperature; (iv) NBS, DMF, room temperature.
(v) 1ꢀmethylꢀ4ꢀ(4,4,5,5ꢀtetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)pyrazole, Pd(PPh3)4, 1M K3PO4, 1,4ꢀdioxane, 150°C;
(vi) TFA, DCM, room temperature; (vii) MeI, Cs2CO3, DMF, room temperature.
RESULTS AND DISCUSSION
Screening of the AZ compound collection identified phenoxyindole 44 as a novel inhibitor of
PFKFB3 with good selectivity over related isoforms PFKFB1 and 2 (Table 1). A crystal
structure of 44 bound in the ATP pocket of PFKFB3 was solved and this highlighted several key
interactions between ligand and enzyme (Figure 1). The indole occupies the adenosine binding
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 46
1
2
3
4
5
6
7
8
9
pocket sitting coplanar, and the phenoxy group occupies a lipophilic pocket and displaces water
molecules seen when adenosine binds. Neither the nitrile nor the indoleꢀ2ꢀamine appear to make
any hydrogen bonds, although a water in the ribose pocket hydrogen bonds to the indole NH.
Compared with the structure of ADP in complex with PFKFB3 there is a small amount of protein
movement that allows the amide to bind, making a hydrogen bond between the amide amine and
the backbone carbonyl oxygen of Leu238 (see Figure 2 for this interaction). The pocket for the
amide tail is made accessible by protein movement, and although the electron density is not clear
for this part of the molecule there are hydrophilic interactions available for the ligand. The
pocket has polar side chains (His242 and Asn240) and a number of exposed backbone carbonyl
oxygen and nitrogen atoms, and although the ligand does not interact directly with them these
groups would be solvated and the ligand sits in this solvated environment. Residue Asn163 has
also moved slightly relative to its position in the ADP structure. Subsequent testing of close
analogues of 44 indicated key aspects of the SAR for this series. Compound 45, the enantiomer
of 44 was inactive, whereas deletion of the chiral alcohol side chain to give glycine amide 46
resulted in only a modest 4ꢀfold drop in potency relative to 44. This was perceived as an
important result since it allowed us to dispense with one of seven hydrogen bond donors in 44,
features which may contribute to reduced permeability and increased efflux. Further deletion of
the amino group in 46 to give acetamide 8 resulted in complete abrogation of activity illustrating
the importance of this group. Although there are hydrophilic groups to interact with in this
region, the amine does not appear to be held in a particularly tight hydrogen bonding network.
Examination of the overlay of 44 with ATP indicated that the ribose pocket of ATP is not
utilized by the inhibitor, but that alkylation of the indole nitrogen might allow access to this
region of the enzyme. Alkylation with the progressively more lipophilic groups methyl, ethyl,
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
ACS Paragon Plus Environment
10

Page 11 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
benzyl and isobutyl (compounds 11 and 48 through 50) resulted in increasing potency against
PFKFB3, with benzyl analogue 49 exhibiting greatest potency. These changes also had the
consequence of increasing potency against PFKFB isoforms 1 and 2, although broadly, relative
levels of selectivity versus PFKFB3 were maintained. A crystal structure of benzyl analogue 49
was solved in PFKFB3 highlighting an identical binding mode to that of 44, with the benzyl ring
directed into the ribose pocket although the phenyl ring does not directly overlay with the sugar
moiety (Figure 2). The phenyl ring packs against Val234 making a hydrophobic interaction and
Asn163 moves again compared to the structure with ADP and the structure of compound 49
overlays well with compound 44, showing a consistent binding mode. Compound 50 which
places a more hydrophilic acetamide group into the ribose pocket (logD = 0.5) shows a
somewhat broader tolerance to groups in this region other than simple alkyl groups.
Examination of these structures suggested no obvious role for the exocyclic indoleꢀ2ꢀamino
group in binding to the enzyme, and as highlighted we were concerned with the significant
number of hydrogen bond donors present in the series. Gratifyingly, deletion of the amino group
entirely had no significant impact on the observed levels of potency. Compounds 51 (desꢀamino
analogue of 46) and 52 (desꢀamino analogue of 47) had, in each case, effectively equivalent
levels of potency and isoform selectivity but for an increase in logD of 0.5 in each case. Selected
compounds were assessed for evidence of inhibition of PFKFB3 in a cellular context. A549 cells
(high expressers of PFKFB3) were treated with compound for 4 hours and then analyzed for a
decrease in intracellular levels of Fꢀ1,6ꢀBP by MSMS analysis, and for a decrease in secretion of
lactate into the media. This assay also included an assessment of cytotoxicity to ensure that
readouts were not adversely influenced by a cellꢀkilling effect. Perhaps due to a combination of
insufficient potency coupled with an excessive number of hydrogen bond donors potentially
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
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 46
1
2
3
4
5
6
7
8
9
limiting permeability, none of the compounds assayed in Table 1 showed target modulation in
this assay. Compounds 48 and 49 showed no direct effect on Fꢀ1,6ꢀBP levels but weak effects
on lactate production which was associated with cell cytotoxicity. These two compounds both
have high lipophilicity, and generally we observe effects on lactate production that are associated
such cytotoxicity for most compounds with a logD above around 3.0.
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
Table 1. Cyanoindole inhibitors of PFKFB3 kinase
Cell
Cell
PFKFB1 PFKFB2 PFKFB3 logD7.4
Fꢀ1,6ꢀBP
IC50
Lactate
IC50
a
a
a
b
c
ID
R1
H
R2
R3
IC50
IC50
IC50
NH2
51.1
26.9
0.575
1.9
> 30
> 30
44
H
H
NH2
NH2
61.5
> 100
53.8
> 100
2.32
1.6
1.6
45
46
> 100
H
NH2
NH2
24.0
38.2
> 100
21.1
> 100
0.815
2.8
1.9
8
Me
> 30
11
ACS Paragon Plus Environment
12

Page 13 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
Et
iꢀBu
Bn
NH2
NH2
NH2
NH2
H
25.0
3.18
2.66
2.53
37.6
18.4
9.18
2.52
1.48
6.84
63.2
22.8
0.514
0.177
0.104
0.343
2.01
2.0
3.2
2.9
0.5
2.1
2.5
> 30
> 30
> 30
> 30
> 30
16.4d
12.6d
> 30
47
48
49
50
51
52
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
H
Et
H
0.562
> 30
a,b,cAll IC50 data are reported as micromolar and are the mean of at least n=2 independent measurements. Each has a
b
c
SEM ± 0.2 log units. Inhibition of fructoseꢀ1,6ꢀbisphosphate in A549 cells. Inhibition of lactate production in
d
A549 cells. Compounds show evidence of cytotoxicity in this assay.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 46
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 1. Crystal structure of 44 bound to the catalytic domain of PFKFB3 (shown in green, pdb code
5ajv), overlaid with the structure of PFKFB3 in complex with ADP (shown in orange, pdb code 2axn).
Figure 2. Crystal structure of 49 bound to the catalytic domain of PFKFB3 (shown in green, pdb code
5ajw), overlaid with the structure of PFKFB3 complexed with ADP (shown in orange, pdb code 2axn).
Table 2. Variations to the amino acid side chain
ACS Paragon Plus Environment
14

Page 15 of 46
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
Cell
PFKFB1
IC50
PFKFB2
IC50
PFKFB3
IC50
logD7.4
2.7
Fꢀ1,6ꢀBP
Cell
Lactate IC50
a
a
a
b
c
ID
R
IC50
20.4
62.9
24.89
76.1
0.830
9.83
> 30
> 30
53
54
55
22.1
1.82
0.112
70.9
2.9
2.8
1.9
3.05
12.0
> 30
> 100
> 100
56
72.3
3.74
86.3
76.5
NT
23.1
0.086
2.20
> 30
19.6
> 30
57
58
59
22.0
2.7
2.6
> 100
23.7
0.531
> 30
60
a,b,cAll IC50 data are reported as micromolar and are the mean of at least n=2 independent measurements. Each has a
b
c
SEM ± 0.2 log units. Inhibition of fructoseꢀ1,6ꢀbisphosphate in A549 cells. Inhibition of lactate production in
A549 cells.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 46
1
2
3
4
5
6
7
8
9
Having observed relatively sharp SAR around the phenoxy amide tail in initial testing, we
embarked on further specific modifications to this region, since the glycine tail was perceived to
be a potential DMPK risk. Simple Nꢀmethylation of 52 to give 53 was tolerated with only a
slight drop in potency for a slight increase in logD, but again removing one more hydrogen bond
donor (Table 2). However exhaustive methylation to give dimethylamino substituted 54 caused
a significant loss of potency indicating that at least one hydrogen bond donor was required for
binding in this region. Capping of the pharmacophore with proline allowed ready access to a
cyclized analogue of 53 in which the NꢀMe is conceptually tethered back onto the glycine side
chain. Use of Sꢀproline resulted in 55 which showed potent inhibition with an IC50 of 0.112 µM,
although it is unclear if this gain in potency is a result of conformational restriction of the side
chain, or simply due to the increased lipophilicity (logD = 2.9). This compound was the first to
show appreciable on scale activity in the cell assay with an IC50 of 3 µM against the Fꢀ1,6ꢀBP
endpoint, and also activity against lactate secretion with no evidence of cellular toxicity.
Compound 56, synthesized using Rꢀproline was completely inactive however, which is entirely
consistent with the serine side chain enantiomeric preference seen in initial hits 44 and 45. For
the first time, the importance of the aromatic amide NH was illustrated by compound 57, which
is the methylated analogue of 55. Methylation in this way resulted in over 200ꢀfold loss in
potency, to 23 µM due to removal of the key hydrogen bond between this amide NH and the
carbonyl of Leu238 (see Figure 2). Further cyclic amino side chains were explored although
again SAR appeared relatively narrow. Azetidine analogue 58 was the first analogue synthesized
to breach 0.1 µM potency, although some stability issues were observed in handling of this group
which precluded more detailed optimization. Expansion to a 6 membered ring of the preferred
stereochemistry as in 59, showed a 20ꢀfold drop in potency relative to proline 55. Final
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
ACS Paragon Plus Environment
16

Page 17 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
confirmation that it was the presence of a donor that was required in this region, rather than
simply a basic primary or secondary amino group, came in the form of lactam 60. This neutral
compound retained significant amounts of potency against PFKFB3.
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
Table 3. Optimization of the ribose pocket group
Cell
PFKFB1
IC50
PFKFB2
IC50
PFKFB3
IC50
logD7.4
3.5
Fꢀ1,6ꢀBP
Cell
Lactate IC50
a
a
a
b
c
ID
R
IC50
2.06
16.9
53.8
3.47
1.93
0.384
2.35
0.023
0.111
0.256
0.048
0.027
0.343
14.0
5.77d
> 30
> 30
> 30
8.71d
26
1.4
2.1
61
62
63
64
> 100
0.429
0.374
7.48
2.8
0.633
> 4.3
0.701
0.108
0.008
3.0
0.140
2.18
65
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 46
1
2
3
4
5
6
7
8
9
1.83
1.13
0.188
0.159
0.020
0.011
3.9
2.8
6.95d
> 30
66
67
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
0.281
a,b,cAll IC50 data are reported as micromolar and are the mean of at least n=2 independent measurements. Each has a
b
c
SEM ± 0.2 log units. Inhibition of fructoseꢀ1,6ꢀbisphosphate in A549 cells. Inhibition of lactate production in
d
A549 cells. Compounds show evidence of cytotoxicity in this assay.
Having optimized the phenoxy tail, we next revisited the indole Nꢀsubstituent since in the earlier
screening of analogues, this region had clearly been shown to impact potency most significantly.
As previously observed, potent compounds could be obtained by appropriate use of lipophilic
alkyl side chains, with isobutyl 26 demonstrating an IC50 of 0.023 µM (Table 3) and also having
sub µM cellular activity albeit with some evidence of cytotoxicity due to the greater obersved
lipophilicy (logD = 3.5). As potency against the primary target is enhanced, activity against the
other isoforms also increases although a degree of selectivity is maintained. Typically across a
wide range of compounds made, selectivity versus PFKFB2 was of the order of 10ꢀfold, and
100ꢀfold versus PFKFB1. Substituent tolerance in this position proved to be quite broad, with
more polar examples such as amide 61, bases as in 62, and tetrahydropyran 63, designed to more
closely mimic the ribose group it replaces, all showing potent inhibition. Amide 61 showed
weak cellular activity, likely due to a combination of weaker potency and additional hydrogen
bond donor. Diꢀbasic analogue 62 also had weaker cell activity in line with its enzyme potency.
In addition to alkyl, direct aromatic linkers are also tolerated. Tetrahydropyran 63, despite
showing subꢀmicromolar activity against Fꢀ1,6ꢀBP showed no activity against the more distal
ACS Paragon Plus Environment
18

Page 19 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
lactate endpoint. NꢀPhenyl analogue 64 is a potent inhibitor of PFKFB3 and a variety of
substituted aryl groups, including heterocycles are tolerated. The isoindolone compound 65 is
representative of one of the more potent aromatic rings that can efficiently bind in the ribose
pocket, with an IC50 of 8 nM, a level of activity that translated well to effects in the cell assay,
with an IC50 of 140 nM, and low micromolar effects on lactate, with no evidence of cytotoxicity.
Analogue 66 again contains a simple benzyl group directed to the ribose pocket and in
comparison to the earlier incarnation in 49 is 5ꢀfold more potent and with 3 fewer hydrogen bond
donors. A wide variety of different benzylic groups could be tolerated in this region, including
heterocyclic analogues such as isoxazole 67, which also showed good activity in the cell assay,
albeit with no apparent effects on lactate production. A crystal structure of 67 was obtained
bound to PFKFB3 (Figure 3). The binding mode of compound 67 is identical to compound 49.
The basic amines both occupy the same space with a hydrogen bond to a conserved water
molecule. Asn163 is disordered in this structure, suggesting the lack of an amine at the 2ꢀ
position of the indole makes this side chain more flexible.
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
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 46
1
2
3
4
5
6
7
8
9
Figure 3. Crystal structure of 67 bound to the catalytic domain of PFKFB3 (pdb code 5ajx).
Table 4. Indazole analogues
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
Cell
PFKFB1
IC50
PFKFB2
IC50
PFKFB3
IC50
logD7.4
3.6
Fꢀ1,6ꢀBP
Cell
Lactate IC50
a
a
a
b
c
ID
R
IC50
0.294
0.191
0.015
0.025
0.004
0.003
0.297
0.067
15.4d
68
2.5
33
a,b,cAll IC50 data are reported as micromolar and are the mean of at least n=2 independent measurements. Each has a
b
c
SEM ± 0.2 log units. Inhibition of fructoseꢀ1,6ꢀbisphosphate in A549 cells. Inhibition of lactate production in
A549 cells. dCompounds show evidence of cytotoxicity in this assay.
As part of the broader exploration of this template, we were motivated to synthesize indazole
analogues 68 and 33, which were exact matched pairs to potent indoles 26 and 67 respectively.
Proximity of Asn163 to the indole Cꢀ2 position was noted, and offered the possibility of a
hydrogen bonding interaction to Nꢀ2 of the corresponding indazole. Comparing Nꢀisobutyl
compounds 26 and 68, the indazole gains around 5ꢀfold potency for equivalent logD (IC50
improves to 0.004 µM from 0.023 µM), but notably, selectivity over the other isoforms is
diminished (selectivity over PFKFB2 drops from 17ꢀfold for the indole to less than 4ꢀfold for
ACS Paragon Plus Environment
20

Page 21 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
indazole). For isoxazole analogues 67 and 33 a similar pattern is observed, although it should be
noted that at 0.003 µM, the potency of 33 is at the limit of the enzyme assay sensitivity and so
absolute conclusions about changes in potency and selectivity are not possible. These
improvements in enzyme potency also translated through into improved cellular activity with
isoxazole substituted indazole analogue 33 showing the greatest cellular potency of all
compounds in this study with an IC50 of 67 nM.
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
Table 5. Indoles with a flipped binding mode
Cell
ID
R1
H
R2
PFKFB1
IC50
PFKFB2
IC50
PFKFB3 logD7.4
IC50
Fꢀ1,6ꢀBP
IC50
Cell
Lactate IC50
a
a
a
b
c
13.5
27.4
11.8
3.57
0.496
0.114
1.6
2.3
> 30
> 30
> 30
> 30
69
H
40
Me
2.35
0.818
0.021
1.9
0.979
9.71
43
a,b,cAll IC50 data are reported as micromolar and are the mean of at least n=2 independent measurements. Each has a
b
c
SEM ± 0.2 log units. Inhibition of fructoseꢀ1,6ꢀbisphosphate in A549 cells. Inhibition of lactate production in
A549 cells.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 46
1
2
3
4
5
6
7
8
9
The crystal structures generated to date had highlighted that the nitrile present in all analogues
did not appear to form any specific interactions with the protein. In the course of exploration of
alternatives to the nitrile, Nꢀmethylpyrazole 69 was synthesized as a direct analogue of earlier
nitrile 51, and this compound showed around a 4ꢀfold improvement in potency relative to the
nitrile but with lower lipophilicity (2.1 to 1.6). A crystal structure of 69 was obtained which
highlighted an intriguing observation. Instead of acting as the nitrile mimic for which it was
designed, the indole moiety was observed to flip through 180 degrees to allow binding of the
pyrazole in the ribose pocket. Figure 4 shows the overlay of these two compounds and the subtle
shift in small molecule architecture that allows this alternate binding mode, without appreciable
change in the protein itself. The protein’s binding site is slightly larger than the compounds in
either binding mode, and along with the ability to maintain the hydrophobic interactions in both
binding modes, the indoles can fit in either way round. Synthesis of the analogue with preferred
proline solvent tail gave compound 40 which showed the anticipated improvement in potency in
line with its increased lipophilicity. Since this new binding mode highlights the pocket
previously occupied by the nitrile is empty, Nꢀmethylation was undertaken to give 43, which
showed a 5ꢀfold increase in potency relative to 40, with concomitant improvement in cellular
activity despite a favourable drop in logD. Compound 43 shows around a 50ꢀfold drop off
between its enzyme potency and effects on Fꢀ1,6ꢀBP modulation in cells, and a further 10ꢀfold
dropꢀoff to effects on lactate secretion, with no evidence of cytotoxicity. Here again, a crystal
structure determination of 43 revealed a most unexpected observation (Figure 5). The helix from
Asp155 to Asn163 has rearranged, partially unwinding towards the Cꢀterminus and forming a
tighter alpha helix towards the Nꢀterminus. This has an effect on the position of the side chains
presenting towards the binding site, especially Val159. Interestingly the size and shape of the
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
ACS Paragon Plus Environment
22

Page 23 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
binding site does not change much after this rearrangement, and both compounds fit comfortably
into both binding sites. It is difficult to make a convincing explanation for this protein
movement, and a subtle balance stabilizing the protein conformation with the binding energy will
dictate the system’s overall energy.
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 4. Crystal structure of 51 (orange, pdb code 5ajz) and 69 (green, pdb code 5ajy) overlaid in the
catalytic domain of PFKFB3.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 46
1
2
3
4
5
6
7
8
9
Figure 5. Crystal structure of 43 (blue, pdb code 5ak0) and 69 (magenta, pdb code 5ajy) overlaid in the
catalytic domain of PFKFB3. The amino acids sequence from Asp155 to Asn163 are colored differently
in each complex (compound 69 is colored magenta and compound 43 is colored blue) showing the
difference between the two structures. Val159 is shown, highlighting the twist in location of this side
chain.
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
Table 6. Changes to the linking atom
N
X
O
N
N
H
N H
Cell
PFKFB1
IC50
PFKFB2
IC50
PFKFB3
IC50
logD7.4
Fꢀ1,6ꢀBP
IC50
Cell
Lactate IC50
a
a
a
b
c
ID
X
O
2.06
36.6
5.0
0.384
6.15
0.023
1.57
3.5
2.7
0.343
5.77d
> 30
13.7d
14d
26
NH
NMe
S
70
71
72
73
74
0.141
0.190
> 100
1.45
0.075
0.111
37.7
3.8
> 30
1.66
0.321
5.92
> 4.3
2.9
SO2
> 30
> 30
24.7
CH2
0.199
3.9
> 30
a,b,cAll IC50 data are reported as micromolar and are the mean of at least n=2 independent measurements. Each has a
b
c
SEM ± 0.2 log units. Inhibition of fructoseꢀ1,6ꢀbisphosphate in A549 cells. Inhibition of lactate production in
A549 cells.
ACS Paragon Plus Environment
24

Page 25 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The impact of changing the heteroatom linker between phenyl and indole was also investigated
(Table 6). A simple NH linker 70 (matched pair to Oꢀlinked 26) lowered LogD but was poorly
tolerated, showing a sharp drop in activity. However Nꢀmethylation as in 71 gave a profile
comparable to O, albeit with increased lipophilicity. Weaker activity and higher lipophilicity
was also a feature of Sꢀlinked analogue 72. However, intriguingly when 72 was oxidized to
sulfone 73, a complete reversal of isoform selectivity was observed. This compound has an IC50
for PFKFB1 of 0.321 µM, and has selectivity of over 300ꢀ and 100ꢀfold over PFKFB2 and 3
respectively. Finally, introduction of a methylene linker 74 saw a slight drop in PFKFB3 activity
at greater lipophilicity penalty. Only the initial Oꢀlinked analogue demonstrated meaningful
cellular effects however.
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
Compound 26 was selected for more detailed pharmacokinetic profiling (Table 7). Solubility is
modest at 48 µM, and plasma protein binding is high in both rat and human, presumably linked
to high lipophilicity. Mouse has lower levels of protein binding however. In vitro clearance in
hepatocytes is remarkably consistent across all three species, with half lives in the range 21ꢀ28
minutes. When dosed to Wistar Han rats, 26 was observed to have very low in vivo clearance
(although given high protein binding, unbound clearance is high), and high bioavailability, with a
moderate half life of 3.4 h. In mouse, clearance was somewhat higher and bioavailability a more
modest 42%, with a shorter half life of 1.4 h. In both species, the in vivo clearance is observed to
correlate well with that predicted from the in vitro data, indicating that the primary clearance
route in vivo is via hepatic metabolism. The fraction absorbed in both species is high and
indicates that the absorption is driven by good solubility and high passive permeability, despite
evidence in Caco2 cells that the compound is an efflux substrate. Good exposure was also
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 46
1
2
3
4
5
6
7
8
9
observed when 26 was dosed at a high dose of 124 µMol/kg to Nude mice, where an AUC of
259 µM.h and Cmax of 35 µM were measured, indicating the potential for this compound to be an
in vivo probe.
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
Table 7. Selected Pharmacokinetic parameters for compound 26
Parameter Value
Aqueous solubility pH7.4
pKa (pyrrolidine) 8.1
48 µM
Plasma protein binding % free (mouse, rat , human) 1.64, < 0.03, 0.33 %
Intrinsic hepatocyte clearance t1/2 (mouse, rat , human) 21, 21, 28 mins
Caco2 intrinsic permeability Papp, pH6.5 13.4 1Eꢀ6.cm/s
Caco2 efflux ratio 4.6
Bioavailability (mouse,a ratb) 42, 100 %
Fraction absorbed (mouse,a ratb) 0.6, 1.0
Vss (mouse, rat) 3.3, 0.4 l/kg
Cl (mouse, rat) 38, 2 ml/min/kg
Predicted Cl (mouse, rat)c 26, 2 ml/min/kg
iv t1/2 (mouse, rat) 1.4, 3.4 h
b
aPK parameters in CD1 mice when dosed po/iv at 2.5 µMol/kg. PK parameters in 10ꢀ12 week old male Wistar Han
rats (n=2) when dosed po at 3.1 µMol/kg and iv at 2.5 µMol/kg (formulation 5% DMSO, 95% Captisol (30% in
water). cClearance was predicted using intrinsic clearance measurements determined from fresh hepatocytes
together with plasma protein binding assumed to be 0.01% free, using the well stirred model and an empirical
correction factor derived inꢀhouse.
Compound 26 was also assessed in a large panel of externally available kinases assays in order to
assess its selectivity, and therefore utility as a probe of PFKFB3 biology. When screened at a
single concentration of 1 µM against a diverse panel of 267 kinases, 26 showed no significant
ACS Paragon Plus Environment
26

Page 27 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
inhibition of any kinase (no kinase was inhibited at a level greater than 25%). This data is
available as supplementary information.
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
CONCLUSIONS
A weak screening hit with suboptimal physicochemical properties was optimized against
PFKFB3 kinase using critical structureꢀguided insights. Central to this optimization was
progressive deletion of hydrogen bonding groups deemed to be nonꢀessential for binding.
Crystallographic studies highlighted binding at the ATP site, and featured a flexible helix that
was able to shift in response to compound binding, and by appropriate substitution, a propensity
for the ligand to flip in the active site was also observed. Unusually for a kinase inhibitor,
despite binding in the ATP pocket, these compounds do not utilize the donor or acceptor
interactions required by ATP itself. The resulting compounds demonstrated high selectivity over
related PFKFB isoforms and potent modulation of the target in a cellular context, both in terms
of direct impact on Fꢀ1,6ꢀBP levels and weaker effects on lactate secretion. Potent examples
were also able to demonstrate modulation of target and its pathway without undue effects on cell
viability. The series disclosed here represents a novel hexose kinase inhibition template which
as a consequence does not display appreciable kinase inhibition outside of the PFKFB3 family.
A selected example demonstrated exposure in rodents following oral dosing. Examples from this
series may serve as useful probes to understand the emerging biology of this interesting
metabolic target, in particular to expand upon the mostly biological si/shRNA validation
undertaken to date.
ACKNOWLEDGMENTS
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 46
1
2
3
4
5
6
We would like to thank Ian Sinclair for helpful chromatography support in development of the
cell assay.
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
AUTHOR INFORMATION
Corresponding Author
*Phone: +44 (0)1625 517920. Email: jason.kettle@astrazeneca.com.
Experimental Methods
Chemistry Unless otherwise stated, commercially available reagents were used as supplied. All
reactions requiring anhydrous conditions were conducted in dried apparatus under an atmosphere
1
of nitrogen. H NMR spectra were recorded using a Bruker AV400 or AV500 NMR. Chemical
shifts δ are reported in ppm and multiplicity of signals are denoted s = singlet, d = doublet, t =
triplet and m = multiplet respectively, with coupling constants (J) reported in hertz (Hz). HRMS
were recorded using a Thermo Accela CTC ꢀ LTQ FT instrument (ESI+). Reactions and
intermediates were also characterised by mass spectroscopy following liquid chromatography
(LCMS or UPLC); UPLC was carried out using a Waters UPLC fitted with Waters SQ mass
spectrometer (Column temp 40, UV = 220ꢀ300 nm, MS = ESI with pos/neg switching) at a flow
rate of 1ml/min using a solvent system of 97% A + 3% B to 3% A to 97% B over 1.50 minutes
(total runtime with equilibration back to starting conditions etc 1.70min), where A = 0.1%
Formic acid in water (for acid work) or 0.1% Ammonia in water (for base work) B =
Acetonitrile. For acid analysis the column used was Waters Acquity HSS T3 1.8um 2.1 x 50 mm,
for base analysis the column used was Waters Acquity BEH 1.7um 2.1 x 50mm. LCMS was
ACS Paragon Plus Environment
28

Page 29 of 46
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
carried out using a Waters Alliance HT (2795) fitted with a Waters ZQ ESCi mass spectrometer
and a Phenomenex Gemini –NX (50x2.1 5um) column at a flow rate of 1.1mL/min 95% A to
95% B over 4 min with a 0.5 min hold. The modifier is kept at a constant 5% C (50:50
acetonitrile:water 0.1% Formic acid) or D (50:50 acetonitrile:water 0.1% ammonium hydroxide
(0.88 SG) depending on whether it is an acidic or basic method. Ion exchange purification was
generally performed using a SCXꢀ2 (Biotage, Propylsulfonic acid functionalized silica.
Manufactured using a trifunctional silane. Non endꢀcapped) cartridge. Individual purification
methods referred to here are detailed in the Supplementary section.
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
N-[4-(3-Amino-4-nitro-phenoxy)phenyl]acetamide (6). Nꢀ(4ꢀHydroxyꢀphenyl)ꢀacetamide 5
(1.14 g, 6.62 mmol) was stirred in DMA (10 mL) and the mixture was cooled to 5°C under an
atmosphere of nitrogen. Sodium hydride (60% dispersion in oil, 0.26 g, 6.62 mmol) was added
portionwise over 10 minutes. The reaction mixture was allowed to warm to room temperature
and stirred for 1 hour. 5ꢀChloroꢀ2ꢀnitroꢀphenylamine 4 (1 g, 6.615 mmol) was added in one
portion and the mixture was heated to 80°C overnight. The mixture was cooled on an ice/water
bath and then poured into 2M HCl (100 mL) before being extracted into ethyl acetate (200 mL).
The organic layer was washed with water (3 x 200 mL) then brine (100 mL) before being dried
(MgSO4) and evaporated to afford a brown solid. The crude product was triturated with DCM to
1
afford 6 (1.1 g, 58%). LCMS (ES+) 288.21 (M+H)+. H NMR δ (d6ꢀDMSO) 10.05 (br s, 1H),
7.99 (d, J = 9.2Hz, 1H), 7.67 (d, J = 9.0Hz, 2H), 7.47 (br s, 2H), 7.11 (d, J = 9.0Hz, 2H), 6.26 –
6.32 (m, 2H), 2.06 (s, 3H).
ACS Paragon Plus Environment
29