Bexarotene prodrugs: Targeting through cleavage by NQO1 (DT-diaphorase)

Article abstract of DOI:10.1016/j.bmcl.2014.03.003

Bexarotene, a retinoid X receptor (RXR) agonist, is being tested as a potential disease modifying treatment for neurodegenerative conditions. To limit the peripheral exposure of bexarotene and release it only in the affected areas of the brain, we designe

Full text of DOI:10.1016/j.bmcl.2014.03.003

Bioorganic & Medicinal Chemistry Letters 24 (2014) 1944–1947
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Bexarotene prodrugs: Targeting through cleavage by NQO1
(DT-diaphorase)
Anja Schäfer a, Ethan S. Burstein b, Roger Olsson a,b,c,
⇑
a Department of Chemistry and Molecular Biology/Medicinal Chemistry, University of Gothenburg, SE-412 96 Gothenburg, Sweden
b ACADIA Pharmaceuticals Inc., 11085 Torreyanna Road, Suite 100, San Diego, CA 92121, USA
c Chemical Biology & Therapeutics, Department of Experimental Medical Science, Lund University, SE-221 84 Lund, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
Bexarotene, a retinoid X receptor (RXR) agonist, is being tested as a potential disease modifying
treatment for neurodegenerative conditions. To limit the peripheral exposure of bexarotene and release
it only in the affected areas of the brain, we designed a prodrug strategy based on the enzyme
NAD(P)H/quinone oxidoreductase (NQO1) that is elevated in neurodegenerative diseases. A series of
indolequinones (known substrates of NQO1) was synthesized and coupled to bexarotene. Bexarotene-3-
(hydroxymethyl)-5-methoxy-1,2-dimethyl-1H-indole-4,7-dione ester 7a was cleaved best by NQO1. The
prodrugs are not cleaved by esterase.
Received 23 January 2014
Revised 25 February 2014
Accepted 1 March 2014
Available online 13 March 2014
Keywords:
Prodrugs
Ó 2014 Elsevier Ltd. All rights reserved.
Parkinson’s disease
Alzheimer’s disease
dt diaphorase
Disease targeting
Bexarotene
O
Bexarotene (Fig. 1) is a retinoid X receptor (RXR) agonist
approved for T cell lymphoma.1 In 2012, bexarotene was reported
to have neuroprotective properties in rodent models of Alzheimer’s
Disease (AD)2 and a year later in Parkinson’s Disease (PD).3 This
caused a lot of excitement in the AD field, and clinical phase II trials
started early 2013 (ClinicalTrials.gov identifier: NCT01782742).
However, systemic exposure leads to peripheral side effects such
as hypertriglyceridemia, hypothyroidism and leukopenia.4 In
addition, bexarotene undergoes extensive cytochrome P450
metabolism,5,6 leading to potential drug–drug interactions if it is
co-administered with other CYP450 substrates, inducers or inhibi-
tors. It would therefore be highly desirable to find selective pro-
drugs for brain targeting of bexarotene to avoid those side
effects. In addition, it was recently shown that bexarotene displays
neuroprotective activity at low doses, making a prodrug strategy
feasible.3 Herein we disclose the first investigation of prodrugs
for pathophysiological release of bexarotene, in an ongoing re-
search project of Pathology Activated Therapies (PATs).
NH
N
O
OH
H2N
O
O
O
O
(a)
(b)
H2N
Figure 1. Structural formula of (a) bexarotene and (b) mitomycin C.
are subjected to oxidative stress, as in tumors, inflammation or
neurodegenerative diseases, detoxification enzymes, such as
NQO1 are induced.10–12 For instance, NQO1s activity is elevated
in several tumor tissues, for example up to 80 fold in lung
tumors.13 Also, it is overexpressed in neurodegenerative diseases
like in active multiple sclerosis lesions12 and in the substantia ni-
gra in Parkinson’s Disease (PD).14,15 In Alzheimer’s disease, neurons
containing diaphorase are selectively spared from degeneration.16
In addition to catecholamines, other quinones can be substrates
for NQO1.17,18 They have been used for tumor targeting; for exam-
ple diaphorase plays a key role in mitomycin C (see Fig. 1) activa-
tion in tumor tissue.19 To function as a promoiety for disease
targeting, indolequinones have a hydroxymethyl substituent in
C2 or C3 which can be connected to the active drug (see Fig. 2).20
In tissues with high diaphorase levels, the quinones are reduced
and the drug is released.
NAD(P)H/quinone oxidoreductase (QR1, NQO1, formerly
DT-diaphorase; EC 1.6.99.2) is a two electron transfer obligate oxi-
doreductase. NQO1 is induced when levels of L
-dopa are raised.7
Physiologically, it keeps
L-dopa, dopamine and other easily
oxidizable catecholamines in their reduced state.8,9 When tissues
With the potential toxicity of the quinones taken into account,
the indolequinone structures shown in Figure 2 were chosen
⇑
Corresponding author. Tel.: +46 768 87 42 17; fax: +46 40 601 34 01.
E-mail address: roger.olsson@med.lu.se (R. Olsson).
http://dx.doi.org/10.1016/j.bmcl.2014.03.003
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

A. Schäfer et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1944–1947
1945
O
O
O
O
O
O
O
h
i
N
N
NH
O
8
9
88% y
65% y
Cl
Cl
HO
O
O
O
O
j
k
O
10
68% y
11
100% y
N
N
Scheme 2. Reagents and conditions: Precursors of 11: h: MeI, KOH, acetone; i: 2-
chloroacetyl chloride, DCM, AlCl3; j: mCPBA, Na2HPO4, CHCl3; k: NaOH, MeOH, rt,
1 h.
Figure 2. Mechanism of drug release from indolequinones by NQO1: Cleavage from
(a) the 3-hydroxymethyl, (b) the 2-hydroxymethyl position.
gradient, 0.1% TFA). Injections were performed every 60 min.
Experiments were carried out in triplicate.
because they combine reasonable affinity to NQO1 with reduced
toxicity. For example, an aziridine instead of the methoxy group
in C6 results in higher affinity, but also much higher toxicity.21 In
this study, both C2- and C3-derivatives were synthesized and
coupled to bexarotene and evaluated. The free carboxylic acid is
crucial for interaction with RXR, so the esterified prodrug is not
expected to show any activity.
Bexarotene is cleaved from compound 7a, a 3-hydroxymethyl
ester (Fig. 3). 50% bexarotene is released after 120 min, after which
no further cleavage could be observed. This might be due to NQO1
inhibitory properties of compound 7a, as some 3-hydroxymethyl
derivatives are inhibitors of NQO1.25,30 Compound 7b was pre-
pared to evaluate if the positive inductive effect of the additional
methyl group might facilitate the release of bexarotene. For 7b
and 7c, free bexarotene can also be observed, but cleavage is
slower and plateaus at a much lower level (Fig. 4). A reason could
be that compound 7b is too sterically demanding to be cleaved by
NQO1 because of the additional methyl group. The behavior of
compound 18 is displayed in Figure 3. No release of bexarotene
occurred during the observation period of 4 h. Even after 24 h of
incubation, no cleavage could be observed (data not shown). This
was unexpected, as 2-hydroxymethyl indolequinones were previ-
ously reported to be the better substrates for NQO1.20,21 The
reduced peak area of compound 18 might be due to association
with the enzyme without cleavage. The behavior of compounds
7a–c and 18 at 37 °C in 1 M HCl (to mimic gastric acid) was also
monitored. There was no release of bexarotene during 4 h.
To investigate if the stagnation in the formation of the peak area
of bexarotene was due deterioration of NADH over time, additional
NADH was added to the test solutions after 30 and 60 min.
However, no change in peak area was observed (data not shown).
The reasons could be either substrate inhibition or product inhibi-
tion. Two model compounds, a 2-hydroxy- and a 3-hydroxyindole
conjugated to 3-methyl-4-nitrophenol, were used to elucidate
whether the indolequinones inhibit the diaphorase enzyme (see
Supplementary information). As the model compounds were
cleaved completely within minutes, this seems unlikely. In addi-
tion, in a more physiological environment, neither the reaction
products or starting materials would accumulate in the way they
do in a reaction vial.
The synthesis of the test compounds 7a, 7b, 7c and 18 was
accomplished by three different routes (Schemes 1–4). The route
for 7a and b started from methyl acetoacetate which was con-
densed with methylamine to form 1 (Scheme 1).22 A subsequent
Nenitzescu reaction with quinone or methyl quinone formed the
indoles 2a or 2b.23 Synthesis of 7c could not be accomplished this
way. We therefore started out from commercial methyl 1H-indole-
3-carboxylate (Scheme 2), which was methylated to 8 and then
subjected to Friedel–Crafts acylation with chloroacetyl chloride
to yield 9 (the 6-chloroacetyl isomer which was formed in small
amounts was removed by column chromatography).24 Compound
9 was further oxidized to the ester 10 with mCPBA and hydrolyzed
to 11. Compounds 2a, 2b and 11 were methylated to form 3a–c and
then nitrated with acetic acid/HNO3 to 4a–c (Scheme 3).17 Reduc-
tion with Sn/HCl yielded amines 5a–c. A subsequent reduction of
the esters using LiAlH4 yielded the corresponding alcohols, evapo-
rated and the resulting residue was immediately oxidized with
Fremy’s salt to the quinones 6a–c.25 The 3-hydroxymethyl bearing
indolequinones 6a–c were coupled to bexarotene under Mitsunobu
conditions to form 7a–c. The synthesis of the 2-hydroxymethyl
indolequinones is shown in Scheme 4. Methyl a-ethylacetoacetate,
sodium nitrite and p-anisidine reacted to 13 in a Fischer indole
cyclization via the Japp–Klingmann azo-ester intermediate
12.26,27 13 was nitrated with HNO3 in DCM to give 14 and methyl-
ated to form 15. It was then reduced with tin/HCl to yield 16.
Reduction and quinone formation was also accomplished with
LiAlH4 and Fremy’s salt successively to form 17, before it was
coupled to bexarotene to form compound 18.28
Lastly, compound 19 was prepared with an additional methyl
group on the 3-hydroxymethyl (Fig. 4, for synthesis see Supple-
mentary information). A compound coupled to the same indolequi-
none had been released more than 10 times faster than the one
without the methyl in a reductive elimination study because of
its positive inductive effect on the release position.17 The conjugate
of this indolequinone with bexarotene (19) was not cleaved at all
(data not shown).
To evaluate the compounds as prodrugs, 7a, 7b, 7c and 18 were
solved in DMSO and then diluted 100 fold with tris buffer. 5
solutions of compounds were incubated at 37 1 °C with or
without diaphorase (75 protein/mL; 50 U/mL), 0.07% BSA,
300
M NADH in tris buffer (pH 7.4).29 Cleavage of the conjugates
lM
l
g
l
was monitored with HPLC (RP-18 column, acetonitrile/water
O
O
Compounds were also tested for enzymatic stability using por-
cine liver esterase using a modified method from Bonina,31
HN
O
O
O
HO
R1
g
f
employing a final concentration of 76
in tris buffer pH 7.4. The conjugates (5
l
l
g/mL (1.3 U/mL) esterase
M) showed no cleavage
N
O
O
1
1
2a
: R = H, 64% y
of the ester over 180 min and only 13% free bexarotene after
24 h (see Supplementary information for compound 7a as
: R1 = Me, 78% y
95%y
2b
example), whereas ethylbenzoate (100
was cleaved completely after 7 min. Hence, even though cleavage
lM) as positive control
Scheme 1. Reagents and conditions: Precursor of 2a and 2b: f: MeNH2, rt,
overnight, g: quinone (2a) or methylquinone (2b), nitromethane, 24 h, rt.

1946
A. Schäfer et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1944–1947
O
O
O
O
NO2
O
O
O
NH2
O
HO
R1
O
O
O
R2
R2
R2
a
b
c
R2
R1
1
R1
N
N
N
R1
N
: R = H, R2 = Me
: R = H, R2 = Me, 73% y
1
2a
2b
3a
3b
4a,b,c
72, 48, 79 % y
5a,b,c
84, 62, 71% y
: R1 = R2 = Me
: R1 = R2 = Me, 97% y
3c: R1 = R2 = H, 55% y
11: R1 = R2 = H
O
O
O
OH
O
R1
O
R2
d
e
O
R1
O
N
N
R2
O
6a,b,c
59, 28, 88% y
7a,b,c
100, 54, 27% y
Scheme 3. Reagents and conditions: a: MeI, KOH, DMSO, rt, 3 h; b: HNO3, acetic acid, À10 °C, 2 h; c: Sn, HCl, EtOH, 30 min reflux; d: (1) LiAlH4, THF, 0 °C, 30 min, (2)
K4[ON(SO3)2]2, acetone, rt, 1 h; e: bexarotene, DEAD, PPh3, THF, 0 °C ? rt, overnight.
O
O
O
O
O
NO2
O
NO2
O
l
c
m
n
h
N
HN
HN
HN
O
NH2
N
O
O
O
O
O
12
13
14
66% y
15
100% y
44% y (2 steps)
O
O
O
NH2
O
O
O
O
O
O
d
e
N
O
N
OH
N
16
91% y
17
63% y
O
18
76% y
Scheme 4. Reagents and conditions: l: (1) NaNO2, HCl, 0 °C; (2) KOH, methyl 2-ethylacetoacetate; m: HCl, reflux; n: HNO3, DCM, À20 °C; h: MeI, KOH, acetone, rt, time; c: Sn,
HCl, EtOH, 30 min reflux; d: (1) LiAlH4, THF, (2) K4[ON(SO3)2]2, acetone; e: bexarotene, DEAD, PPh3, THF, 0 °C ? rt.
Figure 3. Release of bexarotene from (A) 7a, (B) 7b and (C) 7c by NQO1, (D) no release of bexarotene from 18 could be observed.
by NQO1 appears to be rather slow, there is no competing cleavage
by esterase.
2-Hydroxymethyl derivatives had been found to be better
substrates for NQO1 and are also preferentially cleaved by it.20
3-Hydroxymethyl indolequinones like compounds 7a–c may even
be inhibitors of NQO.25,30 In summary, the 2-hydroxymethyl indol-
equinone derivative 18 should have released the drug better than
the 3-hydroxymethyl derivatives 7a–c. Contrary to the literature,
the 2-hydroxymethyl indolequinone 18 was not cleaved at all,
but 7a was cleaved best. Others studied radiolytic cleavage of

A. Schäfer et al. / Bioorg. Med. Chem. Lett. 24 (2014) 1944–1947
1947
5. Shirley, M. A.; Wheelan, P.; Howell, S. R.; Murphy, R. C. Drug Metab. Dispos.
1997, 25, 1144.
6. Howell, S. R.; Shirley, M. A.; Grese, T. A.; Neel, D. A.; Wells, K. E.; Ulm, E. H. Drug
Metab. Dispos. 2001, 29, 990.
O
O
O
O
O
7. van Muiswinkel, F. L.; Riemers, F. M.; Peters, G. J.; LaFleur, M. V. M.; Siegel, D.;
Jongenelen, C. A. M.; Drukarch, B. Free Radic. Biol. Med. 2000, 29, 442.
8. Klegeris, A.; Korkina, L. G.; Greenfield, S. A. Free Radic. Biol. Med. 1995, 18, 215.
9. Segura-Aguilar, J.; Lind, C. Chem.-Biol. Interact. 1989, 72, 309.
10. Jenner, P. Ann. Neurol. 2003, 53, S26.
N
Figure 4. Compound 19 with an additional methyl group (bold) on the release
position.
11. Coussens, L. M.; Werb, Z. Nature 2002, 420, 860.
12. van Horssen, J.; Schreibelt, G.; Bö, L.; Montagne, L.; Drukarch, B.; van
Muiswinkel, F. L.; de Vries, H. E. Free Radic. Biol. Med. 2006, 41, 311.
13. Malkinson, A. M.; Siegel, D.; Forrest, G. L.; Gazdar, A. F.; Oie, H. K.; Chan, D. C.;
Bunn, P. A.; Mabry, M.; Dykes, D. J.; Harrison, S. D.; Ross, D. Cancer Res. 1992, 52,
4752.
14. van Muiswinkel, F. L.; de Vos, R. A. I.; Bol, J. G. J. M.; Andringa, G.; Jansen Steur,
E. N. H.; Ross, D.; Siegel, D.; Drukarch, B. Neurobiol. Aging 2004, 25, 1253.
15. Graumann, R.; Paris, I.; Martinez-Alvarado, P.; Rumanque, P.; Perez-Pastene, C.;
Cardenas, S. P.; Marin, P.; Diaz-Grez, F.; Caviedes, R.; Caviedes, P.; Segura-
Aguilar, J. Pol. J. Pharmacol. 2002, 54, 573.
aspirin from 2- and 3-hydroxymethyl indolequinones and also
found that only the 3-hydroxymethyl derivatives were cleaved,
but no explanation was given.32 However, these studies were done
in the absence of O2 to mimic hypoxia. O2 was not excluded in our
experiments, but we still found the 3-hydroxymethyl derivatives
to be cleaved instead of 2-hydroxymethyl.
Further research is being conducted on other prodrugs for
pathology activated therapies in Parkinson’s disease.
16. Tao, Z.; van Gool, D.; Lammens, M.; Dom, R. Dement. Geriatr. Cognit. Disord.
1999, 10, 460.
17. Everett, S. A.; Naylor, M. A.; Barraja, P.; Swann, E.; Patel, K. B.; Stratford, M. R. L.;
Hudnott, A. R.; Vojnovic, B.; Locke, R. J.; Wardman, P.; Moody, C. J. J. Chem. Soc.,
Perkin Trans. 2 2001, 843.
18. Brunmark, A.; Cadenas, E.; Lind, C.; Segura-Aguilar, J.; Ernster, L. Free Radic. Biol.
Med. 1987, 3, 181.
Acknowledgment
We thank the Michael J. Fox Foundation for financial support.
19. Cummings, J. Biochem. Pharmacol. 1998, 56, 405.
Supplementary data
20. Hernick, M.; Borch, R. F. J. Med. Chem. 2003, 46, 148.
21. Phillips, R. M.; Naylor, M. A.; Jaffar, M.; Doughty, S. W.; Everett, S. A.; Breen, A.
G.; Choudry, G. A.; Stratford, I. J. J. Med. Chem. 1999, 42, 4071.
22. Plath, P.; Rohr, W. Synthesis 1982, 1982, 318.
23. Patrick, J. B.; Saunders, E. K. Tetrahedron Lett. 1979, 4009.
24. Nakatsuka, S.-I.; Asano, O.; Ueda, K.; Goto, T. Heterocycles 1987, 26, 1471.
25. Beall, H. D.; Winski, S.; Swann, E.; Hudnott, A. R.; Cotterill, A. S.; O’Sullivan, N.;
Green, S. J.; Bien, R.; Siegel, D.; Ross, D.; Moody, C. J. J. Med. Chem. 1998, 41,
4755.
26. Xing, C.; Wu, P.; Skibo, E. B.; Dorr, R. T. J. Med. Chem. 2000, 43, 457.
27. Liu, R.; Zhang, P.; Gan, T.; Cook, J. M. J. Org. Chem. 1997, 62, 7447.
28. Newsome, J. J.; Swann, E.; Hassani, M.; Bray, K. C.; Slawin, A. M. Z.; Beall, H. D.;
Moody, C. J. Org. Biomol. Chem. 2007, 5, 1629.
29. Zhang, Z.; Tanabe, K.; Hatta, H.; Nishimoto, S.-I. Org. Biomol. Chem. 2005, 3,
1905.
Supplementary data (experimental procedures for all compounds,
diagrams for incubation with esterase) associated with this article
can be found, in the online version, at http://dx.doi.org/10.1016/
j.bmcl.2014.03.003.
References and notes
1. Gniadecki, R.; Assaf, C.; Bagot, M.; Dummer, R.; Duvic, M.; Knobler, R.; Ranki, A.;
Schwandt, P.; Whittaker, S. Br. J. Dermatol. 2007, 157, 433.
2. Cramer, P. E.; Cirrito, J. R.; Wesson, D. W.; Lee, C. Y. D.; Karlo, J. C.; Zinn, A. E.;
Casali, B. T.; Restivo, J. L.; Goebel, W. D.; James, M. J.; Brunden, K. R.; Wilson, D.
A.; Landreth, G. E. Science 2012, 335, 1503.
3. McFarland, K.; Spalding, T. A.; Hubbard, D.; Ma, J. N.; Olsson, R.; Burstein, E. S.
ACS Chem. Neurosci. 2013, 4, 1430.
4. Duvic, M.; Martin, A. G.; Kim, Y.; Olsen, E.; Wood, G. S.; Crowley, C. A.; Yocum,
R. C. Arch. Dermatol. 2001, 137, 581.
30. Hernick, M.; Flader, C.; Borch, R. F. J. Med. Chem. 2002, 45, 3540.
31. Bonina, F. P.; Montenegro, L.; Guerrera, F. Int. J. Pharm. 1993, 100, 99.
32. Jaffar, M.; Everett, S. A.; Naylor, M. A.; Moore, S. G.; Ulhaq, S.; Patel, K. B.;
Stratford, M. R. L.; Nolan, J.; Wardman, P.; Stratford, I. J. Bioorg. Med. Chem. Lett.
1999, 9, 113.