Development of Fluorinated Analogues of Perhexiline with Improved Pharmacokinetic Properties and Retained Efficacy

Article abstract of DOI:10.1021/acs.jmedchem.6b01592

We designed and synthesized perhexiline analogues that have the same therapeutic profile as the parent cardiovascular drug but lacking its metabolic liability associated with CYP2D6 metabolism. Cycloalkyl perhexiline analogues 6a-j were found to be unsuitable for further development, as they retained a pharmacokinetic profile very similar to that shown by the parent compound. Multistep synthesis of perhexiline analogues incorporating fluorine atoms onto the cyclohexyl ring(s) provided a range of different fluoroperhexiline analogues. Of these, analogues 50 (4,4-gem-difluoro) and 62 (4,4,4′,4′-tetrafluoro) were highly stable and showed greatly reduced susceptibility to CYP2D6-mediated metabolism. In vitro efficacy studies demonstrated that a number of derivatives retained acceptable potency against CPT-1. Having the best balance of properties, 50 was selected for further evaluation. Like perhexiline, it was shown to be selectively concentrated in the myocardium and, using the Langendorff model, to be effective in improving both cardiac contractility and relaxation when challenged with high fat buffer.

Full text of DOI:10.1021/acs.jmedchem.6b01592

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Development of Fluorinated Analogues of Perhexiline with
Improved Pharmacokinetic Properties and Retained Efficacy
Chih-Chung Tseng, Hannah Noordali, Monica Sani, Melanie Madhani,
Denis M. Grant, Michael P Frenneaux, Matteo Zanda, and Iain R. Greig
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01592 • Publication Date (Web): 09 Mar 2017
Downloaded from http://pubs.acs.org on March 12, 2017
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 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
Development of Fluorinated Analogues of Perhexiline with
Improved Pharmacokinetic Properties and Retained Efficacy
7
8
9
ChihꢀChung Tseng,a Hannah Noordali,b Monica Sani,c,d Melanie Madhani,b Denis M. Grant,e
Michael P. Frenneaux,a,f Matteo Zanda,a,c,* Iain R. Greiga,*
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 University of Aberdeen, Kosterlitz Centre for Therapeutics, Foresterhill, Aberdeen, AB25 2ZD (UK)
b University of Birmingham, Institute of Cardiovascular Sciences, Edgbaston, Birmingham, B15 2TT (UK)
c C.N.R. – I.C.R.M., via Mancinelli 7, 20131 Milan (Italy)
d KemoTech s.r.l., Parco Scientifico della Sardegna, Edificio 3, Loc. Piscinamanna, 09010 Pula, CA, Italy
e University of Toronto, Department of Pharmacology & Toxicology, Toronto M5S 1A8 (Canada)
f University of East Anglia, Norwich Medical School, Norwich NR4 7UQ
ABSTRACT
We designed and synthesized perhexiline analogues that have the same therapeutic profile
as the parent cardiovascular drug, but lacking its metabolic liability associated with CYP2D6
metabolism. Cycloalkyl perhexiline analogues 6a-j were found to be unsuitable for further
development as they retained a pharmacokinetic profile very similar to that shown by the
parent compound. Multiꢀstep synthesis of perhexiline analogues incorporating fluorine atoms
onto the cyclohexyl ring(s) provided a range of different fluoroꢀperhexiline analogues. Of
these, analogues 50 (4,4ꢀgemꢀdifluoro) and 62 (4,4,4’,4’ꢀtetrafluoro) were highly stable and
showed greatlyꢀreduced susceptibility to CYP2D6ꢀmediated metabolism. In vitro efficacy
studies demonstrated that a number of derivatives retained acceptable potency against
CPTꢀ1. Having the best balance of properties, 50 was selected for further evaluation. Like
perhexiline, it was shown to be selectively concentrated in the myocardium and, using the
Langendorff model, to be effective in improving both cardiac contractility and relaxation when
challenged with high fat buffer.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 33
1
2
INTRODUCTION
3
4
5
6
7
8
The therapeutic potential of perhexiline 1 (PHX, Fig. 1) in the treatment of cardiovascular
disorders has been known for almost 50 years1,2 and the drug was introduced in 1973 for the
treatment of angina,3 for which it has considerable efficacy and indeed continues to be used
in some centres for patients with refractory angina. Perhexiline shifts myocardial metabolism
from fatty acid to carbohydrate utilisation, thereby increasing metabolic efficiency (i.e. it
produces relatively more ATP for a given oxygen consumption). By increasing myocardial
efficiency, perhexiline has the potential to treat a range of disorders in which impaired
cardiac energetics or oxygen deficiency play a role. In addition to angina, perhexiline has
been investigated and used offꢀlabel for the treatment of heart failure, giving unprecedented
improvements in VO2 max, quality of life improvement and leftꢀventricular ejection fraction;4 for
acute coronary syndromes;5 inoperable aortic stenosis6 and hypertrophic cardiomyopathy.7
The latter study has led the FDA to award orphan drug status to perhexiline. Furthermore, by
its inhibitory action on fatty acid βꢀoxidation, it has been hypothesised that perhexiline may
have potential in antiꢀneoplastic therapies to target tumour metabolism, either alone or in
combination with existing therapeutic regimens (e.g., to increase susceptibility to
chemotherapy).8,9
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. Perhexiline 1 and its enantiomers.10
However, perhexiline was removed from most markets by 1988, due to sideꢀeffects including
hepatotoxicity and neurotoxicity.11 The sideꢀeffects leading to withdrawal of the drug were
ACS Paragon Plus Environment

Page 3 of 33
Journal of Medicinal Chemistry
1
2
due to sustained high plasma levels of the drug, in a small subgroup of patients, as a result
of variable (apparently unpredictable) metabolism. This led to phospholipid accumulation in
the liver and peripheral nerves, due to complete blockade of mitochondrial long chain fatty
acid uptake. Metabolism of perhexiline, hydroxylation at the 4ꢀpositions of either of the
cyclohexyl rings, is mediated primarily by CYP2D6. CYP2D6 is polymorphic, giving
substantial interꢀindividual variation in the rate of metabolism of drugs upon which it acts. An
approximate 100ꢀfold range is seen in the rate of perhexiline clearance and the required
dosage ranges from 100 mg/week to 500 mg/day.12 These issues can be avoided with close
patient monitoring,11 but such monitoring is not practical outside wellꢀcontrolled patient
groups, and thus it is unlikely that perhexiline can ever again be used in a wider patient
population. Recent studies investigating the potential of using pure enantiomers 1a,b (Fig. 1)
of perhexiline have suggested differences in myocardial uptake, clearance, route of
metabolism and, potentially, activity and toxicity; however, it is not yet clear as to whether
these data will translate to therapeutic relevance.13,14,15
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
Perhexiline has been described as a drug of “low potency, low specificity, and low selectivity”
and as having actions on a diverse range of molecular targets, some of which remain to be
properly defined, which contribute to its therapeutic benefit in heart failure (HF).16 However,
at present, the therapeutic action of perhexiline is generally still thought to be due to
inhibition of CPTꢀ1 and to a lesser extent CPTꢀ2.11,17 CPT is an enzyme responsible for
mediating mitochondrial uptake of long chain fatty acids, by coupling them to carnitine; the
acylated carnitine is then taken into the cell by a carnitine acylꢀcarnitine translocase.
Preventing free fatty acid transport across mitochondrial cell membranes causes a
corresponding shift to greater carbohydrate usage and an increase in myocardial efficiency
(10 ꢀ 40%; more in some studies,11 in terms of ATP generated per unit of oxygen). This
oxygenꢀsparing effect is of particular significance in the treatment of angina and heart failure,
allowing cardiac output to remain constant in spite of reduced cardiac oxygen extraction.
Perhexiline inhibits CPTꢀ1 with modest potency (IC50 = 77 ꢁM and 148 ꢁM in rat cardiac and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 33
1
2
liver mitochondria) and inhibits CPTꢀ2 (79 ꢁM) and fatty acid oxidation (βꢀoxidation) in
hepatocytes at 22 ꢁM, resulting in increased lactate and glucose utilisation.18 Whilst the
effective contribution of CPT inhibition has also been questioned,15,16,18 a further argument in
support of the importance of CPTꢀ1 inhibition as the mechanism of action, in spite of the
apparent lack of potency, is the tendency for perhexiline to become concentrated in tissues,
including the myocardium.18 Thus the concentration in the vicinity of mitochondrial CPTꢀ1 is
thought to exceed the modest IC50. This characteristic is attributed to its amphipathic
nature,11 though this has also been highlighted as the source of the proposed diverse
actions of perhexiline16 and may also complicate attempts to determine the role of
stereoselectivity in the therapeutic activity of perhexiline.
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
In spite of the therapeutic potential and the wellꢀunderstood toxicity issues, which led to its
removal from clinical usage throughout most of the world, few attempts have been made to
modify the chemical structure of perhexiline. To date, only a small number of reports on
derivatives have been published19,20 and derivatives with improved properties (relative to
perhexiline) have not been reported. It is clear that compounds which have the same
therapeutic profile as perhexiline, but which lack the metabolic liability associated with
CYP2D6 metabolism, could be highly effective therapeutic agents for the treatment of a
range of cardiovascular conditions.
In the absence of complex and variable
pharmacokinetics, such a compound would not be restricted to limited patient populations or
to use in a clinical environment, and thus could benefit a wider patient population for a range
of cardiovascular diseases. In this report we describe the successful blocking of CYP2D6ꢀ
mediated metabolism by addition of fluorine atoms at the sites of metabolism and
demonstrate that these compounds have broadly equivalent potency to that of perhexiline
against CPTꢀ1.
RESULTS AND DISCUSSION
ACS Paragon Plus Environment

Page 5 of 33
Journal of Medicinal Chemistry
1
2
Chemistry and in vitro pharmacokinetics. In the absence of satisfactory hypotheses
concerning how perhexiline exerts its therapeutic and toxicological actions, there is no clear
guidance as to how the properties could be optimised. For example, there is no evidence
that a more potent CPTꢀ1 inhibitor would necessarily be preferred. Thus, our primary aim for
this programme was to develop compounds having the same activity profile as perhexiline,
but lacking the CYP2D6 liability. The obvious route to achieve this was the development of
compounds which were structurally as similar to perhexiline as possible, in the hope that the
therapeutic effects could be decoupled from the metabolic liability. There have been
surprisingly few attempts made previously to generate even simple perhexiline analogues.
These have included Nꢀsubstituted derivatives,20 derivatives in which the piperidine has
been replaced by other amineꢀbearing groups,19 derivatives in which the carbon separating
the two cyclohexyl groups has been substituted with a carbinol group21 and the synthesis of
a deuterated derivative has also been described.22
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
As perhexiline can be hydroxylated on either cyclohexyl ring by CYP2D6, our initial focus
was on the synthesis of analogues in which one or both of the cyclohexyl rings had been
replaced by cycloalkyl rings of similar size (Scheme 1).
δꢀValerolactone was treated with methyl lithium and the resulting 6ꢀhydroxyꢀhexanꢀ2ꢀone
was transformed into the corresponding TBSꢀether. The latter was treated with LDA and
subjected to aldol reaction with aldehydes affording the βꢀhydroxyꢀketones 2a-d. The
crotonic adducts 3a-d were obtained by DBUꢀpromoted βꢀhydroxy elimination of the
intermediate mesylates. Introduction of the second group of the perhexilineꢀlike structure
was achieved by Michael reaction upon treatment of 3 with the corresponding cuprates,23
then the intermediate silyl enol ethers 4 were deprotected and the primary carbinol group
was oxidized with DessꢀMartin periodinane (DMP) to provide the aldehydes 5a-h (62 – 81%
yields). The latter were subjected to reductive amination/cyclization24 affording the small
library of perhexiline analogues 6a-h, which were isolated as mixtures of stereoisomers in
good to excellent yields (78 – 91%).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 33
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
Scheme 1.Synthesis of cycloalkyl perhexiline analogues 6a-h. Reagents and conditions: (a)
MeLi, ꢀ78 °C, then TBSCl and imidazole (93%); (b) LDA, ꢀ78 °C, then R2CHO; (c) MsCl, TEA
then DBU; (d) RM (M = Mg or Li), Cu(I)X (see experimental section), TMSCl, ꢀ78 °C; (e)
PTSA, MeOHꢀ CH2Cl2; (f) DMP (62ꢀ81% over steps dꢀf); (g) NH4OAc, NaBH3CN, MeOH (78ꢀ
91%).
ACS Paragon Plus Environment

Page 7 of 33
Journal of Medicinal Chemistry
1
2
The THP analogue 6i and the gemꢀdimethylcyclohexyl analogue 6j (Fig. 2) were prepared
3
4
5
following different synthetic routes (see Supporting Information for details).
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.
Disappointingly, in human and rat liver microsomal (HLM / RLM) stability studies,
compounds 6a-j all retained a stability profile very similar to that shown by the parent
compound. We therefore turned our attention to the addition of fluorine atoms onto the
cyclohexyl ring, as another potential route towards either physically blocking the site of
metabolism, or to deactivating the ring towards oxidation by reduction of electron density.25
With the view of introducing fluorine by dehydroxyfluorination, we first synthesised 4ꢀ
hydroxyꢀperhexiline 24 (Scheme 2) from ethyl 4ꢀhydroxycyclohexaneꢀcarboxylate.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 33
1
2
Scheme 2. Synthesis of 4ꢀhydroxyꢀperhexiline 24. Reagents and conditions: (a) MOMCl,
DIPEA, CH2Cl2 (93%); (b) Aq. NaOH, then NaBH4, THF, H2O, then Swern oxidation (87%
over three steps); (c) Cyclohexyl magnesium bromide, Et2O (91%); (d) Py2Cr2O7, CH2Cl2
(83%); (e) nBuLi, 2ꢀpicoline then SOCl2; (f) PtO2, H2 (1 atm), MeOH; (g) TFA, CH2Cl2 then
NaOH (82% over steps eꢀg).
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
The synthesis of fluorinated perhexiline derivatives proved unexpectedly difficult and the
standard lateꢀstage fluorination approaches were not successful in introducing a fluorine
onto position 4 of the cyclohexyl ring. In fact, nucleophilic attack on a 4ꢀcyclohexyl mesylate
or triflate derived from dicyclohexylꢀcarbinol 14 (see Supporting Information, Scheme S2)
using TBAF or KF led to elimination rather than substitution, while use of DAST to effect the
dehydroxyfluorination of 24 (Scheme 2) – also in Nꢀprotected form ꢀ led to dehydration. An
alternative approach, hydrogenation of the potential 4ꢀfluorophenyl precursor 6h (Scheme 1)
using a number of different catalysts led predominantly to concomitant reductive
defluorination and gave perhexiline as the end product. Attempts to access vinylꢀfluoride 27a
(See Scheme S3, Supporting Information) via Agꢀmediated ipsoꢀfluorodestannylation of the
corresponding vinyl stannane with electrophilic fluorinating agents resulted in impractical
outcomes.26 Eventually, we were able to develop multiꢀstep synthetic strategies capable of
producing enough pure fluoroꢀperhexilines 31, 42, 50, 55 and 62 (Fig. 3) for performing
biological tests.
ACS Paragon Plus Environment

Page 9 of 33
Journal of Medicinal Chemistry
1
2
Figure 3. Fluoro perhexiline derivatives synthesised in this work.
3
4
5
6
7
8
9
As an example, the synthesis of 4,4ꢀdifluoroꢀperhexiline 50 (FPERꢀ1) is shown in Scheme 3
(for detailed information on the synthesis of all fluoroꢀperhexilines see the Supporting
Information). Commercially available ethyl 4ꢀhydroxycyclohexaneꢀcarboxylate was first
oxidized to ketone 43 and then submitted to deoxofluorination with DAST, affording an
inseparable mixture of the desired gemꢀdifluoroꢀester 44a and vinylꢀfluoride byꢀproduct,
which – for ease of separation ꢀ was oxidized with mCPBA to give the fluoroꢀoxirane 44b.
This mixture was then reduced with LiAlH4, then FC purification afforded the difluoroꢀalcohol
45, which was transformed into the target gemꢀdifluorinated molecule 50 in 5 steps.
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 3. Synthesis of 4,4ꢀdifluoroꢀperhexiline 50 (FPERꢀ1). Reagents and conditions: (a)
PCC (88%); (b) DAST; (c) mCPBA; (d) LiAlH4 (62% over steps bꢀd); (e) Swern oxidation
(64%); (f) Cyclohexyl magnesium chloride; (g) DMP (86% over steps fꢀg); (h) nBuLi, 2ꢀ
picoline then SOCl2 (88%); (i) Pd(OH)2/C, H2 (1 atm), AcOH, 50 °C (94%).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 33
1
2
3
4
5
6
7
8
Racemic perhexiline 1 shows reasonable stability in the HLM assay, with a halfꢀlife of 100
minutes, but with a short halfꢀlife in RLM of just 21 minutes (Table 1). The relevance of this
was validated in the presence of CYP2D6ꢀexpressing BactosomesTM, in which perhexiline
was very rapidly turned over, with a halfꢀlife of just 14 minutes. As these rat liver microsomes
are prepared from SpragueꢀDawley rats, regarded as a model for CYP2D6 extensive
metabolisers,12 this is a logical correlation. The closest analogue of perhexiline 6c, in which
one cyclohexyl had been replaced by a cyclopentyl ring, showed near identical stability to
the parent compound in both the RLM and CYP2D6 assays. As shown in Table 1, none of
these unsubstituted cycloalkyl analogues 6a-j showed significantly improved RLM stability
compared to perhexiline 1. The improved stability of “(+)ꢀperhexiline” (absolute
stereochemistry not defined at that stage) has previously been proposed.27 The results
shown in Table 1 may be taken to suggest that (R)-(+)ꢀperhexiline 1a was more stable than
(S)ꢀ(ꢀ)ꢀperhexiline 1b in the HLM assay (98 and 65 min, respectively), whilst the two
enantiomers had the same stability in the RLM assay (10 and 11 min, respectively).
However, in view of the apparent lower stability of both of the two enantiomers in
comparison with the racemic parent compound 1 in the RLM assay (21 min; a highly
unlikely, if not impossible occurrence) we do not feel that this assay can be used to draw any
conclusions concerning relatively small differences in potency. However, we were able to
conclude that replacement of one or both cyclohexyl groups with other cycloalkyl groups did
not remove the CYP2D6 liability. Other modifications, including the addition of a gemꢀ
dimethyl onto one of the cyclohexyl rings (6j), or replacement of one of the rings with a 4ꢀ
tetrahydropyranyl (6i), with a phenyl (6g), or with a 4ꢀfluorophenyl (6h), did not lead to
significant improvements in the PK profile. The 4ꢀtetrahydropyran derivative 6i gave an
apparent gain in RLM stability, but this was not reflected by any improvement in stability to
CYP2D6. These results suggested that it would be necessary to either replace both
cyclohexyl rings, or to more strongly deactivate a remaining cyclohexyl ring in order to
remove the metabolic liability. Disappointingly, and unexpectedly, the addition of one
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
ACS Paragon Plus Environment

Page 11 of 33
Journal of Medicinal Chemistry
1
2
fluorine (compound 31) gave no improvement in stability and, equally surprisingly, the
addition of one fluorine on each cyclohexyl (compound 42) gave a compound which was
chemically unstable on storage in pure form. All other compounds showed no decomposition
on incubation for 1 h in pH 7.4 PBS at 37 °C. It is not clear as to what caused this instability,
but this compound was not further evaluated. At the other end of the scale, as predicted, the
tetrafluoro derivative 62, in which all of the potential sites for metabolism, as identified by
previous studies28 had been blocked, was extremely stable. Perhaps most interestingly, the
addition of a 4,4ꢀdifluoro group onto just one of the cyclohexyl rings (compound 50) also
gave a very substantial increase in stability over the parent compound, and stability in the
HLM, RLM and CYP2D6 assays had increased to >500, 72 and 104 min respectively (Table
1). Our hypothesis, aimed at leaving as much of the parent structure intact as possible, was
that one, but perhaps not both, of the cyclohexyl rings formed an essential binding
interaction with the target. Thus a derivative might be more likely to retain activity if only one
ring had been modified. With this in mind, we also prepared the 3,3ꢀdifluorocyclohexyl
derivative 55, feeling that the reduction in electron density from a more distant group could
be sufficient to reduce metabolism, whilst at the same time perhaps interfering less with a
binding interaction. However, this compound did not show the same improvement in stability.
These data show that certain fluorinated perhexiline analogues are much less heavily
metabolised by CYP2D6 than perhexiline and thus should avoid the interꢀindividual
variability which compromises perhexiline therapy. Therefore, these compounds are
expected to show more reliable and predictable pharmacokinetics and hence may also avoid
the toxicity issues found with perhexiline.
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
Table 1. Human and rat liver microsomal (HLM / RLM) stability studies.
HLM
t1/2 (min)
RLM
t1/2 (min)
CYP2D6
t1/2 (min)
R1
R2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 33
1
2
3
4
5
6
7
8
9
± SEM
± SEM
21 ± 1
± SEM
14 ± 3
(±)ꢀPHX 1
(+)ꢀ(R)ꢀPHX 1a
(−)ꢀ(S)ꢀPHX 1b
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclopentyl
Cyclopentyl
Cyclopentyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclopropyl
Cyclobutyl
Cyclopentyl
Cyclopropyl
Cyclobutyl
Cyclopentyl
Phenyl
4ꢀFluorophenyl
4ꢀTetrahydro
pyranyl
100 ± 1
98 ± 25
65 ± 9
124 ± 18
>200
10 ± 0.8
11 ± 0.4
3 ± 0.5
10 ± 0.6
19 ± 2
3 ± 0.04
7 ± 0.5
14 ± 1
ꢀ
ꢀ
ꢀ
ꢀ
6a
6b
6c
6d
6e
6f
>150
16 ± 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
27 ± 2
51 ± 5
50 ± 4
56 ± 7
79 ± 8
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
6g
6h
4 ± 0.3
7 ± 0.2
6i
Cyclohexyl
84 ± 21
31 ± 1
7 ± 0.8
4,4ꢀDimethyl
cyclohexyl
4ꢀFluoro cyclohexyl
4,4ꢀDifluoro
cyclohexyl
6j
31
50
Cyclohexyl
Cyclohexyl
Cyclohexyl
77 ± 11
65 ± 7
>500
20 ± 1
11 ± 0.9
72 ± 6
ꢀ
16 ± 0.5
104 ± 17
3,3ꢀDifluoro
cyclohexyl
55
42
62
Cyclohexyl
>150
ꢀ
35 ± 1
ꢀ
ꢀ
ꢀ
4ꢀFluoro
cyclohexyl
4,4ꢀDifluoro
cyclohexyl
4ꢀFluoro cyclohexyl
4,4ꢀDifluoro
cyclohexyl
>4000
>1500
>150
Biology: We compared our compounds to perhexiline 1 in a suitable efficacy model in
which perhexiline has shown potency. Compound selection was guided by first determining
in vitro inhibitory potency against the most widely accepted target for perhexiline, CPTꢀ1
(Table 2). Inhibitory potency against rat mitochondrial CPTꢀ1 in vitro was measured by
determining the ability of derivatives to inhibit the conversion of palmitoyl CoA and Lꢀ
carnitine into palmitoylcarnitine, measuring the disappearance of palmitoyl CoA
spectrophotometrically, as described previously.29 In this assay the ~50% of the CPT activity
that is sensitive to inhibition by malonyl CoA is defined as CPTꢀ1,29 and under the conditions
of the assay we obtained an IC50 of 1.2 ꢁM for malonyl CoA. Since all tested inhibitors also
showed the same maximal inhibition of activity as malonyl CoA, we assumed that this also
represents inhibition of CPTꢀ1.
We found that the gemꢀdifluoro derivative 50 showed only a small loss in inhibitory potency
(45 ꢁM, P<0.05) compared to racemic perhexiline 1 (11 ꢁM) and its (S) enantiomer 1b (14
ꢁM). The apparent increased potency of the (S) enantiomer 1b matches the published
ACS Paragon Plus Environment

Page 13 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
suggestion that “(−)ꢀperhexiline” might have better therapeutic properties.14 The (R)
enantiomer 1a shows a significant reduction in potency compared to racemic perhexiline 1
(32 ꢁM, P<0.05) and approaches significance when compared to the (S) enantiomer 1b.
This compares to a literature value of 77 ꢁM for perhexiline 1.17 However, the difficulties in
obtaining precise and consistent data for CPTꢀ1 inhibition have been noted and discussed
by Ceccarelli.18 The exceptionally stable tetrafluoro derivative 62 had lost some of its activity
against CPTꢀ1 (88 ꢁM, P<0.05 compared to 50). Replacement of one cyclohexyl with a
cyclopentyl in compound 6c had no impact on potency (11 ꢁM), whilst replacement of both
cyclohexyls with cyclopentyl groups, in compound 6f, gave a nonꢀsignificant reduction in
potency (19 ꢁM). However, the hypothesis that retaining one intact cyclohexyl might be
sufficient to retain activity, whilst a wide range of modifications on the other might be
tolerated, was disproved by the loss of potency seen with the 3,3ꢀdifluoro 55 (215 ꢁM) and
tetrahydropyran 6i (715 ꢁM) derivatives. We also investigated the stereoisomeric mixture of
4ꢀhydroxy derivatives 24 to study the possibility that the activity of perhexiline was inꢀpart
mediated through a metabolite,28 but instead found it to be very poorly active (475 ꢁM).
Having the best balance of pharmacokinetic and inhibitory properties, plus relative ease of
synthesis, we selected the 4,4ꢀgemꢀdifluoro 50 derivative to take forward for in vivo PK and
efficacy study.
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 2. CPTꢀ1 inhibitory activities.
IC50 for CPTꢀ1
inhibition (ꢁM)*
11 (7 ꢀ 17)
R1
R2
(±)ꢀPHX 1
(+)ꢀ(R)ꢀPHX 1a
Cyclohexyl
Cyclohexyl
Cyclohexyl
Cyclohexyl
32 (24 ꢀ 41)
(−)ꢀ(S)ꢀPHX 1b
Cyclohexyl
Cyclohexyl
14 (8 ꢀ 25)
6c
Cyclohexyl
Cyclopentyl
11 (5 ꢀ 24)
6f
6i
50
55
62
24
Cyclopentyl
Cyclohexyl
Cyclohexyl
Cyclopentyl
19 (7 ꢀ 40)
715 (510 ꢀ 1050)
45 (34 ꢀ 60)
215 (160 ꢀ 290)
88 (61 ꢀ 127)
475 (290 ꢀ 745)
4ꢀTetrahydropyranyl
4,4ꢀDifluorocyclohexyl
3,3ꢀDifluorocyclohexyl
4,4ꢀDifluorocyclohexyl
4ꢀHydroxycyclohexyl
Cyclohexyl
4,4ꢀDifluorocyclohexyl
Cyclohexyl
* mean and pooled 95% confidence interval of two independent experiments.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 33
1
2
3
4
5
6
7
8
In vivo pharmacokinetics: It has been proposed that the unexplained activity of perhexiline
1, in spite of modest in vitro potency against all proposed targets, is due to selective
accumulation in the myocardium. Thus, in a continued attempt to ensure that our
compounds mirrored perhexiline in as many aspects as possible, we conducted studies to
determine the in vivo stability and myocardial exposure. The distribution and stability of 50,
following daily oral dosing, were compared to perhexiline in male BALB/c mice (samples
taken at 1, 4, 8 and 24 h following dosing; n = 3 per timepoint – see supplementary
information for full details). Plasma and myocardial concentrations, and area under the
curve, following 10 mg/kg oral gavage of 50 and perhexiline were investigated. In the mouse
model, perhexiline 1 is very rapidly metabolised and thus no direct comparisons can be
drawn between the absolute concentrations of the two compounds. However, Table 3
shows that there is a significant accumulation of perhexiline in the myocardium (5.7ꢀfold at 1
h, 4.9ꢀfold at 4 h and 4.8ꢀfold at 8 h, as defined by drug concentration; 2.4ꢀfold as defined by
AUC), whilst 50 shows an even larger selective accumulation (10.7ꢀfold at 1 h, 10.9ꢀfold at 4
h and 9.5ꢀfold at 8 h, as defined by drug concentration; 8.6ꢀfold as defined by AUC).
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
Table 3. In vivo pharmacokinetics in a mouse model
50
Perhexiline
(10 mg/kg)
(10 mg/kg)
Plasma
Myocardium
3726.8 ± 2808.5 116.4 ± 39.2
1877.3 ± 606.1*#
1826.7 ± 498.7*#
32360 ± 7030**##
Plasma
Myocardium
664.5 ± 157.5#
125.8 ± 62.5
29.2 ± 25.2
1 h
4 h
8 h
346.5 ± 257.4
171.9 ± 44.1*
191.8 ± 48.0*
Conc
(ng/mL)
28.3 ± 8.8
6.0 ± 5.6
AUC 0ꢀ24
(hr*ng/mL)
2114 ± 265##
3763 ± 1501*
876 ± 123
#
*P<0.05 relative to perhexiline; **P<0.01 relative to perhexiline; P<0.05 relative to plasma;
##P<0.01 relative to plasma (Student’s tꢀtest).
Further studies were conducted, showing that selective myocardial accumulation was also
seen in male SpragueꢀDawley rats (used as a model for CYP2D6ꢀmediated metabolism)12
(Table 4). The distribution and stability of 50 following daily oral dosing were compared to
ACS Paragon Plus Environment

Page 15 of 33
Journal of Medicinal Chemistry
1
2
perhexiline 1 (plasma samples taken at 1, 2, 4, 8 and 24 h following dosing; n = 3 per
timepoint – see Supporting Information for full details). Again perhexiline is metabolised very
quickly, which is not representative of the situation in humans, but similar myocardium :
plasma ratios can be seen in comparing perhexiline 1 (19ꢀfold at 1 h and 11.5ꢀfold at 8 h, as
defined by drug concentration) and 50 (15.3ꢀfold at 1 h and 7.2ꢀfold at 8 h, as defined by
drug concentration).
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
Table 4. In vivo pharmacokinetics in a rat model
50
Perhexiline
(10 mg/kg)
(10 mg/kg)
Plasma
Myocardium
Plasma
Myocardium
5913.4 ± 1350.2#
1901.2 ± 887.1#
1 h
2 h
4 h
8 h
386.3 ± 28.4*
375.4 ± 139.7*
297.1 ± 145.2
178.6 ± 93.4
99.9 ± 43.5
112.9 ± 32.6
97.44 ± 24.3
31.4 ± 5.7
ꢀ
ꢀ
ꢀ
ꢀ
Conc
(ng/mL)
361.2 ± 104.7#
1285.9 ± 822.8
*P<0.05 relative to perhexiline; #P<0.05 relative to plasma (Student’s tꢀtest)
Thus, assuming that this accumulation is responsible for the therapeutic effect, rather than
presenting a toxicological liability, 50 appears to match our requirements as an optimised (in
terms of CYP2D6 liability) orallyꢀbioavailable analogue of perhexiline, suitable for efficacy
studies.
Ex vivo efficacy studies: The ex vivo Langendorff murine model was used to assess the
effects of perhexiline 1 and analogue 50 on indicators of cardiac contractility (LVdP/dt max)
and relaxation (LVdP/dt min) following perfusion with a high fat buffer. In recent years it has
been shown that HF can be characterised by severe metabolic disturbances, intrinsic to the
myocardium, which include fatty acid overload and buildꢀup within the cardiomyocyte cytosol
and mitochondria.30 As such, energy production within the cardiomyocyte is impaired, which
in turn disrupts cardiac function and contractile activity over time. By perfusing hearts with a
high fat buffer, we can simulate the fatty acid overload observed during HF and can measure
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 33
1
2
3
the effects of perhexiline 1 and compound 50 on cardiac function, such as LVdP/dt max and
4
5
min, under these conditions.31
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. Perhexiline 1 and 4,4ꢀgemꢀdifluoroꢀperhexiline 50 improve cardiac contractility and
relaxation in a high fat perfusate. A: The Langendorff perfusion protocol. Hearts were
stabilised for 10 minutes with BSAꢀPalmitate and then perfused for 30 minutes with either 1)
Control (BSAꢀPalmitate), 2) 2 ꢂM perhexiline, 3) 2 ꢂM 50 or 4) 10 ꢂM 50. Cardiac
contractility (LVdP/dt max) and cardiac relaxation (LVdP/dt min) were calculated from the left
ventricular pressure. B: LVdP/dt max. C: LVdP/dt min. Data are presented as mean ± SEM;
n = 12ꢀ19 mice. **P<0.01 vs Control (BSAꢀPalmitate); ***P<0.001 vs Control (BSAꢀpalmitate)
(1ꢀway ANOVA, postꢀhoc Bonferroni test.
The results demonstrate that 2 ꢁM perhexiline 1 shows a strong trend for enhancing LVdP/dt
max (2.1 fold increase when compared to the control; n=16; P=0.07) and is highly effective
for enhancing LVdP/dt min (3.2 fold increase when compared to the control; n=16; P<0.001)
in the presence of high fat. At a concentration of 10 ꢁM, like perhexiline, 50 showed a similar
trend to increasing LVdP/dt max (1.8 fold increase when compared to the control; n=16
ACS Paragon Plus Environment

Page 17 of 33
Journal of Medicinal Chemistry
1
2
P=0.07) and was almost as effective as perhexiline in increasing LVdP/dt min (2.7 fold
increase when compared to the control; n=16; P<0.01). These results also demonstrate that
2 ꢁM perhexiline and 10 ꢁM 50 have significant lusitropic effects, a feature that current
pharmacological treatments for HF lack. It can be noted that the apparent 5ꢀfold loss of
potency of 50 compared to perhexiline is in accord with the relative potencies vs CPTꢀ1;
however ꢀ as it cannot be assumed that CPTꢀ1 is the only target, or even the major target, or
that the ratio of potencies for perhexiline and 50 vs other targets will remain the same – the
importance of this observation cannot be determined. At present, most drug therapies for
HF only improve systolic function such as cardiac contractility, and therefore only benefit
patients with systolic HF.32 As such, perhexiline and 50 show promising potential for use in
diastolic HF, a condition in which cardiac relaxation is impaired, as opposed to contractility,
and where there are currently no wellꢀdefined existing treatments.32ꢀ34 Furthermore, 50 at 2
ꢁM shows a small improvement on both cardiac parameters with clear evidence of dose
dependency.
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
CONCLUSION
In summary, these results reveal that 4,4ꢀgemꢀdifluoroꢀperhexiline 50 (FPERꢀ1) holds a
similar therapeutic profile to perhexiline 1 and thus 50, and possibly other analogues of
perhexiline, may have the same potential for the treatment of HF and cardiovascular
diseases, but without the metabolic liability of the parent compound. As a consequence,
compounds with the properties exemplified by 50 are expected to be suitable for use across
a broader spectrum of the patient population and not to require additional monitoring of drug
plasma concentrations or patient phenotyping, both of which limit drug use and convenience.
However, it is also clear from these studies that elucidation of the mechanism of action of
perhexiline will be required before it will be possible to conduct an informed drug discovery
programme to generate a perhexilineꢀlike drug for clinical development.
EXPERIMENTAL SECTION
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 33
1
2
General Methods: All reactions were carried out under a nitrogen or argon atmosphere with
dry solvents under anhydrous conditions, unless otherwise mentioned. Solvents and
reagents: anhydrous tetrahydrofuran (THF), diethyl ether (Et2O), methylene chloride
(CH2Cl2) and toluene were purchased from commercial suppliers. Reagents were purchased
at the highest commercial quality, used without further purification and handled in
accordance with COSHH regulations. Chromatography: flash chromatography (FC) was
carried out on silica gel (Merck Silica gel Si 60, 40ꢀ63 ꢁm) according to the method
described by Still.35 Thinꢀlayer chromatography (TLC) was carried out on glassꢀbased 0.25
mm Merck silica gel plates (60Fꢀ254) which were developed with UV irradiation (254 nm and
365 nm), an aqueous solution of KMnO4 and an ethanol solution of ammonium molybdate,
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
1
and heat as developing agents. H NMR spectra: these were recorded at 400 MHz on a
Bruker ADVANCEIII 400 instrument. Chemical shifts (δH) are given in parts per million (ppm)
as referenced to the appropriate residual solvent peak. 13C NMR spectra: these were
recorded at 101 MHz on a Bruker ADVANCE III 400 instrument. Chemical shifts (δC) are
given in parts per million (ppm) as referenced to CHCl3. 19F NMR spectra: these were
1
recorded at 376 MHz on a Bruker ADVANCE III 400 instrument without H decoupling.
Chemical shifts (δC) are given in parts per million (ppm) as referenced to CFCl3 as 0 ppm. All
NMR spectra were recorded at 298K unless otherwise stated. The following abbreviations
were used to explain the multiplicities: s = singlet, d = doublet, t = triplet, q = quartet, quint =
quintet, m = multiplet, br = broad. Purities of all tested compounds were determined by
reverse phase HPLCꢀMS using an Agilent 1200 series chromatograph system coupled with
an Agilent G6120 signal quadrupole detector equipped with an electrospray ionisation
source and detection in positive mode. HPLC conditions for purity analysis: Zorbax Eclipse
Plus C18 column, 4.6 × 100 mm, 3.5 ꢁm; mobile phase: 75% MeOH / 25% H2O with 0.1%
formic acid; flow rate: 1 mL/min. All tested compounds were found to be at least 96% purity.
HRMS: analyses were performed using an LTQ Orbitrap XL MS spectrometer.
ACS Paragon Plus Environment

Page 19 of 33
Journal of Medicinal Chemistry
1
2
Representative procedures for carbocyclic perhexiline analogues.
3
4
5
6
6-((tert-Butyldimethylsilyl)oxy)hexan-2-one
7
8
9
A solution of MeLi (14.8 mL, 23.71 mmol, 1.6 M in Et2O) was added at ꢀ78 °C to an Et2O (10
mL) solution of δꢀvalerolactone (2.158 g, 21.55 mmol). The resulting reaction mixture was
stirred for 1 h at ꢀ78 °C before quenching with NH4Cl (10 mL, sat. aq.) and dilution with
EtOAc (20 mL). The layers were separated and the organic layer was washed with brine (20
mL), dried (Na2SO4) and concentrated under reduced pressure to give the crude hydroxy
ketone, which was used directly without further purification.
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
TBSCl (3.92 g, 26.0 mmol) and imidazole (2.31 g, 33.93 mmol) were added at 0 °C to a
CH2Cl2 (10 mL) solution of the crude hydroxy ketone. The resulting mixture was allowed to
warm to 23 °C and stirred for 3 h prior to quenching with a saturated solution of NaHCO3 (10
mL) and dilution with CH2Cl2 (10 mL). The layers were separated, and the aqueous layer
was extracted with CH2Cl2 (2 × 5 mL). The combined organic layers were dried over Na2SO4
and concentrated under reduced pressure. FC (silica gel, hexanes:EtOAc 5:1) afforded 6ꢀ
1
((tertꢀButyldimethylsilyl)oxy)hexanꢀ2ꢀone (4.64 g, 93% for two steps) as a colourless oil. H
NMR (400 MHz, CDCl3) δ = 3.60 (t, J = 6.3 Hz, 2H), 2.44 (t, J = 7.3 Hz, 2H), 2.12 (s, 3H),
1.69 – 1.58 (m, 2H), 1.56 – 1.45 (m, 2H), 0.88 (s, 9H), 0.03 (s, 6H); 13C NMR (101 MHz,
CDCl3) δ = 209.1, 62.8, 43.5, 32.2, 29.8, 25.9, 20.3, 18.3, ꢀ5.3. m/z (ESI) 231 [M+H]+.
7-((tert-Butyldimethylsilyl)oxy)-1-cyclohexyl-1-hydroxyheptan-3-one (2a)
Method A: a solution of freshly prepared LDA (3.80 mL, 3.80 mmol, 1.0 M in THF/hexanes)
was added at ꢀ78 °C to a THF (5 mL) solution of 6ꢀ((tertꢀbutyldimethylsilyl)oxy)hexanꢀ2ꢀone
(730 mg, 3.17 mmol). The resulting mixture was stirred for 1 h at ꢀ78 °C before
cyclohexanecarboxaldehyde (426 mg, 460 ꢁL, 3.80 mmol) was added. The reaction mixture
was stirred at ꢀ78 °C for further 30 min before quenching with NH4Cl (10 mL, sat. aq.) and
dilution with EtOAc (10 mL). The layers were separated and the organic layer was washed
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 33
1
2
with brine (20 mL), dried (Na2SO4) and concentrated under reduced pressure. FC (silica gel,
3
4
5
1
hexanes:EtOAc 2:1) furnished the aldol adduct 2a (955 mg, 88 %) as a pale yellow oil. H
6
7
8
9
NMR (400 MHz, CDCl3) δ = 3.85 – 3.73 (m, 1H), 3.60 (t, J = 6.2 Hz, 2H), 3.04 (brs, 1H), 2.67
– 2.39 (m, 4H), 1.89 – 1.78 (m, 1H), 1.78 – 1.70 (m, 2H), 1.69 – 1.57 (m, 4H), 1.54 – 1.47
(m, 2H), 1.41 – 1.29 (m, 1H), 1.29 – 1.09 (m, 3H), 1.09 – 0.92 (m, 2H), 0.88 (s, 9H), 0.04 (s,
6H); 13C NMR (101 MHz, CDCl3) δ = 212.6, 71.7, 62.7, 46.1, 43.4, 43.0, 32.1, 28.8, 28.3,
26.4, 26.2, 26.1, 25.9, 20.1, 18.3, ꢀ5.3. m/z (ESI) 343 [M+H]+.
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
(E)-7-((tert-Butyldimethylsilyl)oxy)-1-cyclohexylhept-1-en-3-one (3a)
Method B: Et3N (540 ꢁL, 3.87 mmol) and MsCl (150 ꢁL, 1.94 mmol) were added at 0 °C to a
CH2Cl2 solution of aldol adduct 2a (440 mg, 1.28 mmol). The solution was stirred for 1 h at 0
°C followed by addition of DBU (290 ꢁL, 1.94 mmol). The resulting mixture was warmed to
23 °C and stirred for further 2 h prior to quenching with NH4Cl (10 mL, sat. aq.) and dilution
with EtOAc (20 mL). The layers were separated and the organic layer was washed with brine
(20 mL), dried (Na2SO4) and concentrated under reduced pressure. FC (silica gel,
hexanes:EtOAc 5:1) afforded the enone 3a (410 mg, 98%) as a pale yellow oil. 1H NMR (400
MHz, CDCl3) δ = 6.76 (dd, J = 16.0, 6.8 Hz, 1H), 6.05 (dd, J = 16.0, 1.4 Hz, 1H), 3.63 (t, J =
6.2 Hz, 2H), 2.57 (t, J = 7.3 Hz, 2H), 2.19 – 2.08 (m, 1H), 1.81 – 1.73 (m, 4H), 1.73 – 1.60
(m, 3H), 1.60 – 1.49 (m, 3H), 1.36 – 1.26 (m, 2H), 1.22 – 1.09 (m, 2H), 0.89 (m, 9H), 0.04
(m, 6H); 13C NMR (101 MHz, CDCl3) δ = 201.1, 152.1, 127.8, 62.9, 40.6, 39.8, 32.3, 31.8,
25.95, 25.92, 25.7, 20.8, 18.3, ꢀ5.3. m/z (ESI) 325 [M+H]+.
7-Cyclohexyl-7-cyclopropyl-5-oxoheptanal (5a)
Method C:23 a solution of cyclopropylmagnesium bromide (1.67 mL, 1.67 mmol, 1.0 M in 2ꢀ
MeTHF) was added at ꢀ78 °C to a suspension of CuI (159 mg, 0.835 mmol) in THF (2 mL).
The resulting mixture was allowed to warm to 0 °C and stirred for further 1 h to give the
ACS Paragon Plus Environment

Page 21 of 33
Journal of Medicinal Chemistry
1
2
organocuprate reagent. The solution was then cooled to ꢀ78 °C followed by addition of a
THF solution (0.5 mL) of enone 3a (180 mg, 0.56 mmol) and TMSCl (142 ꢁL, 1.12 mmol).
The reaction mixture was stirred at ꢀ78 °C for 1 h before quenching with a MeOH solution of
PTSA (2 N, 5 mL, 10 mmol). The solution was stirred at 23 °C for 2 h prior to dilution with
EtOAc (20 mL). The layers were separated and the organic layer was washed with brine (10
mL), dried (Na2SO4) and concentrated under reduced pressure. The mixture was passed
through a plug of silica gel (EtOAc) to give the corresponding alcohol, which was used
without further purification.
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
Dess−Martin periodinane (285 mg, 0.67 mmol) was added at 23 °C to a CH2Cl2 solution (5
mL) of the alcohol above. The reaction mixture was stirred for 6 h before quenching with
Na2SO3 (5 mL, sat. aq.). The resulting mixture was extracted with CH2Cl2 (3 × 5 mL),
combined organic layers were dried over Na2SO4 and concentrated under reduced pressure.
The residue was purified by FC (silica gel, hexanes:EtOAc 3:1) afforded the dicarbonyl
compound 5a (101 mg, 73 % for three steps). 1H NMR (400 MHz, CDCl3) δ = 9.82 (t, J = 1.5
Hz, 1H), 2.59 – 2.39 (m, 5H), 1.78 – 1.42 (m, 8H), 1.42 – 1.26 (m, 1H), 1.26 – 0.91 (m, 6H),
0.68 – 0.49 (m, 1H), 0.49 – 0.37 (m, 1H), 0.37 – 0.22 (m, 1H), 0.12 (tt, J = 14.5, 7.2 Hz, 1H),
ꢀ0.09 (td, J = 9.3, 5.3 Hz, 1H); 13C NMR (101 MHz, CDCl3) δ = 212.1, 201.3, 46.0, 45.5, 43.4,
42.8, 32.1, 30.6, 30.1, 26.8, 26.74, 26.68, 19.8, 14.21 5.6, 3.4. m/z (ESI) 251 [M+H]+.
2-(2-Cyclohexyl-2-cyclopropylethyl)piperidine (6a)
Method D:24 NH4OAc (53 mg, 0.69 mmol), NaBCNH3 (41 mg, 0.65 mmol) and 3 Å molecular
sieves (100 mg) were added at 23 °C to a MeOH solution (2 mL) of dicarbonyl compound 5a
(78 mg, 0.31 mmol). The resulting mixture was stirred at 23 °C for 16 h before filtration
through a pad of Celite®. The filtrate was concentrated under reduced pressure. The residue
was purified by FC (silica gel, MeOH:CH2Cl2 0:100 → 10:90) to afford compound 6a (58 mg,
78 %, mixture of diastereoisomers, purity > 97%) as a pale yellow oil. NMR spectra were
recorded by using the HCl salt of the compound, which was prepared by treating 6a (10 mg)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 33
1
2
3
4
5
6
7
8
9
1
with HCl (200 ꢁL, 1.25 M in MeOH) followed by concentration under reduced pressure. H
NMR (400 MHz, CDCl3) δ = 9.24 (brs, 2H), 3.51 – 3.38 (m, 1H), 3.27 (dd, J = 13.1, 7.0 Hz,
0.5H), 3.21 – 3.07 (m, 0.5H), 2.83 (dtd, J = 15.9, 12.7, 3.2 Hz, 1H), 2.22 – 1.53 (m, 12H),
1.53 – 0.96 (m, 7H), 0.85 (ddd, J = 16.2, 13.6, 6.8 Hz, 0.5H), 0.72 – 0.37 (m, 7.5H), 0.24 –
0.07 (m, 3H), 0.29 – 0.08 (m, 1.5H), 0.06 – ꢀ0.03 (m, 0.5 H); 13C NMR (101 MHz, CDCl3) δ =
55.9, 55.6, 44.8, 44.7, 44.5, 44.29, 42.9, 42.0, 36.40, 36.35, 30.9, 30.7, 30.0, 29.0, 28.4,
27.1, 26.94, 26.87, 26.75, 26.72, 22.5, 22.4, 22.3, 22.2, 13.61, 13.56, 4.9, 4.5, 4.0, 3.7. m/z
(ESI) 236 [M+H]+; HRMS (ESI) calcd for C16H30N+ [M+H]+ 236.2373, found 236.2369.
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
Ethyl 4-(methoxymethoxy)cyclohexanecarboxylate (20)
DIPEA (992 ꢁL, 5.8 mmol) and MOMCl (330 ꢁL, 4.35 mmol) were added to a CH2Cl2 (7 mL)
solution of ethyl 4ꢀhydroxycyclohexanecarboxylate (500 mg, 2.9 mmol) cooled to 0 °C. After
stirring at r.t. for 6 h H2O (5 mL) was added, the layers were separated and the organic
phase was dried (Na2SO4). The solvent was evaporated and the crude was purified by FC
1
(hexane:EtOAc 8:2) to give 20 (590 mg, 93% yield) as a colorless oil. H NMR (400 MHz,
CDCl3) δ = 4.47 (s, 1 H), 3.94 (q, 2 H, J = 7.0 Hz, 2 H), 3.58ꢀ3.51 (m, 1H), 3.17 (s, 3 H), 2.23
– 2.21 (m, 1 H), 1.93ꢀ1.78 (m, 2H), 1.76ꢀ1.65 (m, 2H), 1.63ꢀ1.52 (m, 2H), 1.51ꢀ1.41 (m, 3H),
1.07 (t, J = 7.0 Hz, 3 H); 13C NMR (101 MHz, CDCl3) δ = 175.0, 94.5, 74.7, 60.0, 54.9, 42.2,
29.6, 23.8, 14.0. m/z (ESI) 217 [M+H+]
4-(Methoxymethoxy)cyclohexanecarbaldehyde (21)
NaOH (210 mg, 5.24 mmol) was added to a solution of 20 (570 mg, 2.62 mmol) in a
THF:H2O 4:1 mixture (10 mL). The reaction was stirred overnight, then H2O (5 mL) was
added. The layers were separated, the aqueous phase was acidified with 1N HCl until pH 1
was reached and then extracted with EtOAc (10 mL). The organic phase was dried
(Na2SO4), filtered and the solvent was removed under reduced pressure to give the
ACS Paragon Plus Environment

Page 23 of 33
Journal of Medicinal Chemistry
1
2
corresponding acid (413 mg ,83% yield) as a white solid. Et3N (250 ꢁL, 1.80 mmol) and ethyl
chloroformate (172 ꢁL, 1.80 mmol) were added to a dry THF (5 mL) solution of acid (300
mg, 1.6 mmol) cooled to 0 °C. The resulting suspension was stirred at the same temperature
for 1h and then filtered through Celite. The filtrate was then added to a solution of NaBH4
(116 mg, 3.1 mmol) in water (2 mL) cooled to 0 °C. The reaction mixture was stirred at the
same temperature for 30 min, quenched with H2O (5 ml) and extracted with EtOAc (10 mL).
The solvent was removed under reduced pressure and the crude was purified by FC (Hex:
EtOAc 1:1) to give 298 mg of the corresponding alcohol (268 mg , 96% yield) as a colourless
oil. A solution of (COCl)2 (133 ꢁL, 1.5 mmol) in CH2Cl2 (3 mL) was added dropwise to a
solution of DMSO (204 ꢁL, 2.87 mmol) in CH2Cl2 (5 mL), cooled to ꢀ78 °C. The reaction was
stirred at the same temperature for 15 min, then a solution of the alcohol above (200 mg,
1.15 mmol) in CH2Cl2 (5 mL) was added. After stirring for 1 h, Et3N (800 ꢁL, 5.75 mmol) was
added. After 30 min the reaction was warmed to r.t, quenched with 1N HCl (aq) and
extracted with EtOAc (10 mL). The solvent was dried (Na2SO4), filtered and concentrated
under reduced pressure. The crude was purified by FC (hexane:EtOAc 8:2) to give 58 (165
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
1
mg, 83%) as a yellow oil. H NMR (400 MHz, CDCl3) δ = 9.51 (s, 1 H), 4.52 (s, 2 H), 3.66ꢀ
3.55 (m, 1H), 3.23 (s, 3 H), 2.20 – 2.23 (m, 1 H), 1.98ꢀ1.85 (m, 2H), 1.82ꢀ1.70 (m, 1H), 1.67ꢀ
1.40 (m, 4H), 1.30 – 1.21 (m, 1H); 13C NMR (101 MHz, CDCl3) δ = 204.2, 94.6, 74.6, 71.9,
55.1, 48.2, 31.1, 23.7, 21.2. m/z (ESI) 173 [M+H+].
Cyclohexyl(4-(methoxymethoxy)cyclohexyl)methanol (22)
A small crystal of iodine was added to a suspension of magnesium shavings (63 mg, 2.58
mmol) in Et2O (2 mL, nitrogen atmosphere). Bromocyclohexane (317 ꢁL, 2.58 mmol) was
then added slowly (over 5 min), so as to maintain a gentle reflux. The mixture was heated to
reflux for 30 min. After cooling to r.t, aldehyde 21 (370 mg, 2.15 mmol) in Et2O (2 mL,), was
slowly added. After stirring at r.t overnight, NH4Cl (5 mL, sat. aq.) was added and the
reaction was extracted with EtOAc (10 mL). The organic layer was dried (Na2SO4), filtered
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 33
1
2
and concentrated under reduced pressure. The crude was purified by FC (hexane:EtOAc
3
4
5
1
7:3) to give 22 (233 mg, 91%) as a yellow oil. H NMR (400 MHz, CDCl3) δ = 4.59 (s, 2 H),
6
7
8
9
3.77 ꢀ 3.69 (m, 1H), 3.28 (s, 3H), 3.10ꢀ2.93 (m, 1H), 2.07ꢀ 1.92 (m, 1H), 1.88ꢀ1.77 (m, 1H),
1.74ꢀ1.73 (m, 3H), 1.62ꢀ1.26 (m, 8H), 1.25ꢀ 0.87 (m, 8H); 13C NMR (101 MHz, CDCl3) δ =
94.5, 79.7, 75.9, 71.3, 55.1, 40.2, 39.8, 39.0, 32.4, 30.2, 29.9, 28.0, 26.5, 24.0, 21.8. m/z
(ESI) 257 [M+H+]
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
Cyclohexyl(4-(methoxymethoxy)cyclohexyl)methanone (23)
PCC (846 mg, 2.25 mmol) was added to a solution of 22 (470 mg, 1.87 mmol) in CH2Cl2 (40
mL). The reaction was stirred overnight, then filtered through Celite. The filtrate was
concentrated under reduced pressure and purified by FC (hexane:EtOAc 9:1) to give 23
1
(395 mg, 83%) as a colourless oil. H NMR (400 MHz, CDCl3) δ = 4.56 (s, 2 H), 3.75 ꢀ 3.64
(m, 1H), 3.27 (s, 3H), 2.50ꢀ2.35 (m, 2H), 1.84ꢀ1.74 (m, 2H), 1.73ꢀ1.62 (m, 6H), 1.61ꢀ1.35 (m,
5H), 1.33ꢀ 1.05 (m, 5H); 13C NMR (101 MHz, CDCl3) δ = 216.4, 94.6, 71.0, 55.2, 48.7, 48.0,
29.8, 28.7, 25.8, 25.7, 23.1. m/z (ESI) 277.2 [M+Na+]
4-(1-Cyclohexyl-2-(piperidin-2-yl)ethyl)cyclohexanol (24)
A solution of nBuLi (680 ꢁL, 1.7 mmol 2.5 M in hexanes) was added dropwise to a stirred
solution of 2ꢀpicoline (168 ꢁL, 1.7 mmol) in THF (4 mL) at ꢀ78 °C. The resulting red solution
was stirred at that temperature for 10 min before being warmed to ꢀ10 °C for 20 min, after
which the mixture was cooled to ꢀ78 °C and a THF (2 mL) solution of 23 (395 mg, 1.55
mmol) was added. The resulting mixture was stirred at that temperature for 30 min before
SOCl2 was added (186 ꢁL, 2.55 mmol). The milky reaction mixture was warmed to 0 °C for
30 min prior to quenching with NH4Cl (10 mL, sat. aq.) and dilution with EtOAc (10 mL). The
layers were separated and the organic layer was washed with brine (20 mL), dried (Na2SO4)
and concentrated under reduced pressure. FC (hexane:EtOAc 8:2) gave the intermediate
ACS Paragon Plus Environment

Page 25 of 33
Journal of Medicinal Chemistry
1
2
unsaturated 2ꢀpyridines (442 mg, 75%) as a yellow oil. To a solution of these 2ꢀpyridines
(210 mg, 0.63 mmol) in MeOH (7 mL) a catalytic amount of PtO2 was added. The reaction
was stirred under a hydrogen atmosphere for 24 h, filtered through a Celite® pad and
concentrated under reduced pressure. The crude was purified by FC (CH2Cl2:MeOH 9:1) to
give 2ꢀ(2ꢀcyclohexylꢀ2ꢀ(4ꢀ(methoxymethoxy)cyclohexyl)ethyl)piperidine (180 mg, 84%) as a
yellow oil.
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
TFA (120 ꢁL, 1.55 mmol) was added to a CH2Cl2 (4 mL) solution of 2ꢀ(2ꢀcyclohexylꢀ2ꢀ(4ꢀ
(methoxymethoxy)cyclohexyl)ethyl)piperidine (175 mg 0.51 mmol). After stirring for 4 h the
solvent was removed under reduced pressure and the crude dissolved in THF (3 mL). 1N
aq. NaOH (1.5 mL) was added and the reaction was stirred for 1 h. H2O (5 mL) and EtOAc
(10 mL) were added and the layers were separated. The organic layer was dried (Na2SO4)
and concentrated under reduced pressure to give 24 (130 mg, 86%, purity > 96%) as a
yellow amorphous solid. .1H NMR (400 MHz, CDCl3) δ = 3.93ꢀ3.83 (m, 1H), 2.98 (d, J = 12.1
Hz, 1H), 2.53 (td, J = 9.6 and 2.6 Hz, 1H), 2.39 – 2.27 (m, 1 H), 2.26 – 2.22 (m, 1 H), 1.68ꢀ
0.81 (m, 30 H); 13C NMR (101 MHz, CDCl3) δ = 66.2, 57.0, 47.2, 44.5, 39.8, 39.4, 36.4, 33.1,
32.9, 32.9, 31.9, 29.6, 27.0, 26.8, 26.7, 26.3, 25.2, 24.9, 23.3. m/z (ESI) 294 [M+H+]
In vitro CPT-1 Inhibition Assay:29
Preparation of Rat Heart Mitochondrial Suspension (MS): All of the steps below were
performed on ice or at 4 °C. Three male Wistar rats (8 weeks of age) were sacrificed by
cervical dislocation, hearts were excised, rinsed twice with iceꢀcold Homogenization Buffer
(HB: 10 mM TrisꢀHCl, 250 mM sucrose, 1 mM EDTA, pH 7.4), blotted, weighed, and finely
minced with scissors for 5 min. Minced hearts were homogenized in 5 volumes of HB (by
tissue weight) using a Polytron homogenizer. The homogenate was centrifuged for 10 min at
600 x g, and the supernatant was removed and saved. The pellet was reꢀsuspended in HB,
homogenized again, spun for 10 min at 600 x g, and the supernatants from the first and
second homogenizations were combined. The combined supernatant was centrifuged for 10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 33
1
2
min at 7000 x g, the supernatant was discarded, the pellet was reꢀsuspended in HB to the
original homogenization volume and centrifuged again for 10 min at 7000 x g. The
supernatant was discarded and the pellet of washed mitochondria was reꢀsuspended in HB
to 0.5 volumes relative to original heart weight. Protein content was determined using the
Bradford protein assay, the solution was diluted with HB to a protein concentration of 10
mg/mL, and small individual aliquots of this mitochondrial suspension (MS) were shockꢀ
frozen in LN2 and stored at ꢀ80 °C. Aliquots of MS were thawed and diluted with HB to 2
mg/ml protein just prior to use in CPTꢀ1 assays.
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
Each reaction (final volume 200 ꢁL, pH 6.8) contained 0.04 mg of MS protein, 25 mM Trisꢀ
HCl, 150 mM sucrose, 1 mM EDTA, 60 mM KCl, 1.2 mg/mL lipidꢀfree BSA, 0.1 mM 4,4'ꢀ
dithiodipyridine, 0.4 mM Lꢀcarnitine, 0.04 mM palmitoyl CoA and varying concentrations of
inhibitor. Reactions were performed in a quartz glass microcuvette, and inhibitors were preꢀ
incubated for 3 min at 37 °C with all assay components except for Lꢀcarnitine and palmitoyl
CoA. Each inhibitor was dried down from a CH2Cl2 stock, redissolved in the assay buffer
described above, and appropriate serial dilutions were made. A solution of Lꢀcarnitine and
palmitoyl CoA (in water) was added to initiate the enzyme reaction, and the cuvette was
placed in a spectrophotometer programmed in kinetic mode to read the increase in
absorbance at 324 nm for 3.5 min. The slopes of the kinetic curves were calculated and
compared to the mean of positive controls generated in the absence of any inhibitor (defined
as 100 % activity), to determine the % control activity for each inhibitor concentration.
Controls and each inhibitor concentration were analyzed with 5 replicates, and mean +/ꢀ SD
were calculated for each point. IC50 estimates and 95% confidence intervals for the malonyl
CoAꢀinhibitable component of the CPT activity were determined following subtraction of the
nonꢀinhibited activity by nonꢀlinear regression of inhibition curves using GraphPad Prism 5
(GraphPad Software Inc), and each value listed in Table 2 represents the mean IC50 value
and pooled confidence interval from two independent experiments using each inhibitor.
ACS Paragon Plus Environment

Page 27 of 33
Journal of Medicinal Chemistry
1
2
Ex vivo Langendorff Model:
3
4
5
6
7
8
Briefly, C57BL/6 (25 ꢀ 30 g) murine hearts were isolated and mounted onto the Langendorff
perfusion system and retrogradely perfused with a modified Krebs buffer consisting of 1.2
mM sodium palmitate preꢀconjugated to 3% bovine serum albumin (BSA).36 The buffer was
perfused at a constant pressure of 80 mmHg at 37 °C and continuously gassed with 95% O2
/ 5% CO2. Following 10 min stabilisation in control (BSAꢀPalmitate) Krebs buffer, hearts were
perfused for 30 min with 1) control (BSAꢀPalmitate), 2) 2 ꢁM perhexiline, 3) 2 ꢁM 50, or 4) 10
ꢁM 50. The maximal and minimal change in left ventricular pressure over time (LVdP/dt max
and min, respectively) were calculated from the left ventricular pressure recorded throughout
the 40 minutes of perfusion. All data are expressed as mean ± SEM and statistical
comparison between groups were performed using 1ꢀway ANOVA, followed by the postꢀhoc
Bonferroni test. Values of P<0.05 were deemed as significant.
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
Pharmacokinetics assays:
Routine in vitro stability studies were conducted by Cyprotex Ltd (Macclesfield, UK) and in
vivo pharmacokinetics were conducted by Sai Life (Pune, India) using standard
methodology.
ASSOCIATED CONTENT
The Supporting Information is available free of charge on the ACS Publications website at
DOI: XXXXX. Experimental procedures, characterization of all intermediates and target
compounds, and copies of NMR spectra of compounds 6a-j, 31, 42, 50, 55, 62.
AUTHOR INFORMATION
Corresponding authors:
*Eꢀmail: m.zanda@abdn.ac.uk; telephone: +44 (0)1224 437513
*Eꢀmail: i.greig@abdn.ac.uk; telephone: +44 (0)1224 437370
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 33
1
2
3
4
5
Notes:
6
7
8
9
M.P.F., M.Z. and I.R.G. are coꢀfounders of Signal Pharma Limited that owns the I.P. of some
of the molecules described in this article.
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
ACKNOWLEDGEMENTS.
We thank Massimo Frigerio for technical support with the syntheses. We gratefully thank
Signal Pharma for financial support and fellowship (C.C.T), and the EPSRC National Mass
Spectrometry Service Centre (Swansea, UK) for performing HRMS analyses.
ABBREVIATIONS USED
CYP2D6, cytochrome P450 2D6; CPTꢀ1, carnitine palmitoyltransferaseꢀ1; CPTꢀ2, carnitine
palmitoyltransferaseꢀ2; HF, heart failure; DMP, DessꢀMartin periodinane; HLM, human liver
microsomes; RLM, rat liver microsomes; LVdP, Left Ventricular developed Pressure; SEM,
Standard Error of Mean.
REFERENCES
(1) Hudak, W. J.; Lewis, R. E.; Kuhn, W. L. Cardiovascular Pharmacology of Perhexiline. J.
Pharmacol. Exp. Ther. 1970, 173, 371–382.
(2) Cho, Y. W.; Belej, M.; Aviado, D. M. Pharmacology of a New Antianginal Drug:
Perhexilineꢃ: I. Coronary Circulation and Myocardial Metabolism. Chest 1970, 58, 577–
581.
(3) Singlas, E.; Goujet, M. A.; Simon, P. Pharmacokinetics of Perhexiline Maleate in
Anginal Patients with and without Peripheral Neuropathy. Eur. J. Clin. Pharmacol.
1978, 14, 195–201.
(4) Lee, L.; Campbell, R.; ScheuermannꢀFreestone, M.; Taylor, R.; Gunaruwan, P.;
Williams, L.; Ashrafian, H.; Horowitz, J.; Fraser, A. G.; Clarke, K.; Frenneaux, M.
ACS Paragon Plus Environment

Page 29 of 33
Journal of Medicinal Chemistry
1
2
Metabolic Modulation With Perhexiline in Chronic Heart Failure. Circulation 2005, 112,
3280–3288.
3
4
5
6
7
8
9
(5) Willoughby, S. R.; Stewart, S.; Chirkov, Y. Y.; Kennedy, J. A.; Holmes, A. S.; Horowitz,
J. D. Beneficial Clinical Effects of Perhexiline in Patients with Stable Angina Pectoris
and Acute Coronary Syndromes Are Associated with Potentiation of Platelet
Responsiveness to Nitric Oxide. Eur. Heart J. 2002, 23, 1946–1954.
(6) Unger, S. A.; Robinson, M. A.; Horowitz, J. D. Perhexiline Improves Symptomatic
Status in Elderly Patients with Severe Aortic Stenosis. Aust. N. Z. J. Med. 1997, 27,
24–28.
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) Abozguia, K.; Elliott, P.; McKenna, W.; Phan, T. T.; NallurꢀShivu, G.; Ahmed, I.; Maher,
A. R.; Kaur, K.; Taylor, J.; Henning, A.; Ashrafian, H.; Watkins, H.; Frenneaux, M.
Metabolic Modulator Perhexiline Corrects Energy Deficiency and Improves Exercise
Capacity in Symptomatic Hypertrophic CardiomyopathyClinical Perspective. Circulation
2010, 122, 1562–1569.
(8) Galluzzi, L.; Kepp, O.; Heiden, M. G. V.; Kroemer, G. Metabolic Targets for Cancer
Therapy. Nat. Rev. Drug Discovery 2013, 12, 829–846.
(9) Vander Heiden, M. G. Targeting Cancer Metabolism: A Therapeutic Window Opens.
Nat. Rev. Drug Discovery 2011, 10, 671–684.
(10) Tseng, C.ꢀC.; Greig, I.; Harrison, W.; Zanda, M. Asymmetric Synthesis and Absolute
Configuration of (+)ꢀ and (–)ꢀPerhexiline. Synthesis 2015, 48, 73–78.
(11) Ashrafian, H.; Horowitz, J. D.; Frenneaux, M. P. Perhexiline. Cardiovasc. Drug Rev.
2007, 25, 76–97.
(12) Licari, G.; Somogyi, A. A.; Milne, R. W.; Sallustio, B. C. Comparison of CYP2D
Metabolism and Hepatotoxicity of the Myocardial Metabolic Agent Perhexiline in
SpragueꢀDawley and Dark Agouti Rats. Xenobiotica 2015, 45, 3–9.
(13) Chong, CꢀR.; Drury, N. L.; Licari, G.; Frenneaux, M. P.; Horowitz, J. D.; Pagano, D.;
Sallustio, B. C. Stereoselective handling of perhexiline: implications regarding
ACS Paragon Plus Environment