Design and synthesis of reboxetine analogs morpholine derivatives as selective norepinephrine reuptake inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.09.007

As part of a discovery effort aimed at identifying novel norepinephrine reuptake inhibitors (NRIs), a number of substituted morpholines were designed and synthesized. The target compounds contain vicinal stereogenic centers, and the program was greatly facilitated by the adoption of efficient synthetic routes which allowed for the late stage incorporation of structural and physicochemical diversity into the targets. Structure-activity relationships were developed by optimizing individual ring components of the structure for NRI potency and for selectivity against other monoamine reuptake transporters. Several novel morpholine derivatives with a potent and selective NRI profile are described.

Full text of DOI:10.1016/j.bmcl.2008.09.007

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5550–5553
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and synthesis of reboxetine analogs morpholine derivatives
as selective norepinephrine reuptake inhibitors
a,
b
c
Wenjian Xu a, , David L. Gray , Shelly A. Glase , Nancy S. Barta
*
*
a Department of Chemistry, Pfizer Global Research & Development Groton Laboratories, Ann Arbor, MI 48105
b Department of Chemistry, Pfizer Global Research & Development Ann Arbor Laboratories, 2800 Plymouth Road, Ann Arbor, MI 48105, USA
c Berry & Associates, 2434 Bishop Circle East, Dexter, MI 48130, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 July 2008
Accepted 3 September 2008
Available online 6 September 2008
As part of a discovery effort aimed at identifying novel norepinephrine reuptake inhibitors (NRIs), a num-
ber of substituted morpholines were designed and synthesized. The target compounds contain vicinal
stereogenic centers, and the program was greatly facilitated by the adoption of efficient synthetic routes
which allowed for the late stage incorporation of structural and physicochemical diversity into the tar-
gets. Structure–activity relationships were developed by optimizing individual ring components of the
structure for NRI potency and for selectivity against other monoamine reuptake transporters. Several
novel morpholine derivatives with a potent and selective NRI profile are described.
Keywords:
NRI
Norepinephrine
Norepinephrine reuptake inhibitor
Ó 2008 Elsevier Ltd. All rights reserved.
NET
Morpholine
Selective norepinephrine reuptake inhibitor
Modulating central norepinephrine (NE) levels via its reuptake
transporter (NET) has been shown to be an effective pharmaco-
therapy strategy for treating a variety of illnesses including
ADHD,1–4 depression,5 fibromyalgia,6 peripheral neuropathy, and
other types of pain.7 Consequently, there is much interest in devel-
oping agents which selectively inhibit the norepinephrine reuptake
transporter (NRI’s). The precise mechanism by which an NRI might
have beneficial effects on each of these disease states varies – for
instance, atomoxetine (1, StratteraTM), a relatively selective NRI re-
cently approved for the treatment of ADHD, is proposed to provide
efficacy for ADHD symptoms by increasing synaptic NE and DA lev-
els via blockade of the NET in cortical regions associated with
attention and memory.8 Similarly, elevated concentration of NE
in the synaptic cleft is considered to increase or maintain the activ-
ity of the descending inhibitory bulbospinal pathway, which is
compromised in chronic pain conditions, thereby leading to anal-
gesic effects in some patients.9 Previous work in the area of small
molecule monoamine reuptake inhibitors has shown that it is chal-
lenging to identify molecules with high affinity for inhibition of
NET which simultaneously exhibit high selectivity (>100Â) versus
the related serotonin reuptake transporter (SERT) and dopamine
reuptake transporter (DAT). In this paper, we describe elements
of a program which was aimed at identifying centrally acting NRI’s
with the potential for treating disease, specifically, the synthesis
and structure–activity relationships of a series of morpholine-
based NRIs with general structure 3.
Certain substituted morpholines such as 2 have been shown to
exhibit excellent binding selectivity for NET versus SERT and con-
tinue to attract attention to their synthesis and biological activ-
ity.10–12 We developed an interest in evaluating structures of
general form 3 (Fig. 1) for NRI potency, with the aim to access com-
pounds with varied physicochemical properties while retaining
selectivity and NRI activity. Initially, we looked to literature
describing the preparation of related morpholine derivatives for
synthetic inspiration.13 The first published route to 2 is highly reli-
able, but lengthy. It requires a 10-step linear sequence where the
desired diversity at R1 and R2 (3, Fig. 1) is incorporated into the mol-
ecules at the very beginning of the synthesis and R2 is primarily lim-
ited to phenyl. Despite its excellent utility, this route did not
incorporate suitable speed or design flexibility to allow for a rapid
and complete exploration of the broader chemical space embodied
within generic structure 3. Our first attempt at a general synthesis
of some target morpholines is outlined in Scheme 1. We expected
Me
Me
NH
O
O
O
O
O
NH
O
NH
R1
R2
1
2
3
* Corresponding authors. Tel.: +1 860 686 9158; fax: +1 860 686 7444.
E-mail address: david.l.gray@pfizer.com (D.L. Gray).
Figure 1. General structure of compounds targeting selective NRI activity.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.09.007

W. Xu et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5550–5553
5551
OH
N
R2
OH
O
O
N
HO
HO
O
O
O
c, d
O
a,b
c
a, b
O
O
d
O
N
N
Boc
N
N
N
Boc
Boc
Boc
Boc
f
Boc
4
5
6
11
12
4
O
O
O
R2
e
R2
O
HO
O
f
e
R1
R1
R1
O
3
O
N
NH
N
Boc
N
Boc
Boc
7
8
13
14
Scheme 1. Reagents and conditions: (a) IBX; (b) C6H5MgBr, À78 °C, 67% for two
steps; (c) C6H5COOH, PPh3, DIAD; (d) 5% NaOH, 60 °C, 81% for two steps; (e)
C6H5OH, PPh3, DIAD, 70–85%; (f) CF3COOH or 4 M HCl in Dioxane, 0 °C, 80–90%.
Scheme 2. Reagents and conditions: (a) NaOCl, TEMPO, KBr, TBACl;
(b) 1-propanephosphonic acid cyclic anhydride, 91% for two steps; (c) C6H5Li,
À78 °C–rt, 45–80%; (d) H2, trans-RuCl2[(R)-xylbinap][(R)-daipen], 85%; (e) C6H5OH,
PPh3, DIAD, 70–85%; (f) CF3COOH or 4 M HCl in Dioxane, 0°C, 80–90%.
that hydroxymorpholine 4 would be a useful common intermediate
from which to access our targets. This important starting material
was prepared in high optical purity according to a literature proce-
dure.14 In this approach, IBX was used to oxidize (S)-hydrox-
ymorpholine 4 to deliver its aldehyde derivative in high yield.15,16
For R2 = phenyl (Fig. 1), Grignard addition of phenyl magnesium
bromide to this sensitive aldehyde afforded a 6:1 mixture of diaste-
reoisomeric secondary alcohols 5 which were chromatographically
separated. A purified single isomer was subjected to a two-step
inversion procedure consisting of Mitsunobu reaction with benzoic
acid followed by benzoate hydrolysis. This process effectively con-
verged the mixture to a single isomer (either one as desired) and
provided a common intermediate for exploring the SAR of phenoxy
ring substitution. Mitsunobu conditions were employed to couple
the appropriate phenol with 6 (Scheme 1, (S,S) series shown),
affording the Boc-protected targets 7 which were deprotected upon
exposure to HCl in dioxane, providing the products 8 as their corre-
sponding HCl salts in good yield. In order to verify the stereochem-
ical identity of the chromatographically separated intermediates 5,
each isomer was carried independently through the outlined se-
quence to 2 and co-injected with standards of the same on chiral
HPLC. This procedure indicated that the (2S,3S) alcohol (5) was
the major isomer arising from the addition of phenyl magnesium
bromide to the aldehyde. For most nucleophiles, the substrate-
based stereocontrol of addition to the aldehyde was between 3:1
and 8:1 favoring the (2S, 3S) isomer. The new route had certain
advantages, however, the addition of lithium-based organometallic
nucleophiles to the aldehyde resulted in moderate to poor yields of
the desired adducts; some Grignard reagents were also problem-
atic. With benzyl magnesium bromide, we observed rearrangement
to the scrambled toluene adducts (9 ? 10, Fig. 2).
buffered bleach,17 and the derived acid subsequently converted
to the desired intermediate N,O dimethyl amide 11 via cyclic phos-
phinic acid anhydride promoted amidation (Scheme 2).18 With the
target electrophile (11) in hand, various organometallic reagents
were added to the amide, typically affording the functionalized
ketone 12 in moderate to good yield. The next task, namely reduc-
tion of ketone 12 in a stereocontrolled fashion, was critical to the
success of this synthetic route, and several approaches were
adopted to successfully address the stereocontrol in subclasses of
these molecules.19–21 Some substitutions required chromatogra-
phy at this stage to upgrade the diastereoisomeric purity of the
products. After establishing the required stereocenters, Mitsunobu
coupling of selected substituted phenols with the secondary alco-
hols 13 proceeded without event, affording the Boc-protected
ethers 14. Removal of the carbamate protecting group was
smoothly effected by treatment with HCl in dioxane as before,
affording the morpholine products 3 in good yield.
Catalytic hydrogenation was initially effective in setting the
second stereocenter adjacent to the morpholine ring for
R2 – phenyl, however, one drawback of this method was that a
small amount (<10%) of racemization was sometimes observed at
the morpholine center during the chiral reduction (12 ? 13). Addi-
tionally, in some of the later cases, the resultant isomeric mixture
of heterocyclic alcohol isomers (13) was not separable by flash
chromatography.
An inelegant, but workable solution to this problem was found
by carrying these mixtures through the remaining steps of the syn-
thesis and separating the final target isomers by chiral preparatory
HPLC. For such compounds, lacking positive proof of the stereo-
chemical identity of the separated isomers, both isomers were
screened for activity. The binding affinities for some of the more
potent morpholine derivatives are given in Tables 1–4.
As a result of these and other issues with the generality and reli-
ability of this route, we chose to seek an alternative method for
generating diverse morpholine derivatives (3, Fig. 1), in particular,
an approach which would be more amenable to generating non-
phenyl substitutions at R2 and allow isolation of single isomers
at the R2 stereocenter. Weinreb amide 11 was deemed to be a good
alternate intermediate from which the target morpholine deriva-
tives might be readily accessed.
Table 1
Monoamine transporter binding of compounds with (R,S) and (S,S) configurations
Structures
Entry
Stereo chem
NET (nM)
SERT (nM)
This key compound was obtained in a two-step procedure
wherein hydroxymorpholine
4
was oxidized by exposure to
O
O
1
2
(2S,3S)
(2R,3S)
2
35
482
75
CH3
MgBr
OH
O
NH
NH
HO
O
Br
O
N
N
Boc
Boc
3
4
(2S,3S)
(2R,3S)
3
17
477
105
10
9
O
Figure 2. Undesired rearrangement observed during Grignard addition.

5552
W. Xu et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5550–5553
Table 2
Table 4
Comparison of NET activities of compounds with (S,R) and (S,S) configurations
MAT binding activities for non-phenyl (S,S) morpholine derivatives
Structures
Entry
Stereo chem
NET (nM)
SERT (nM)
Structures
Entry
R2
NET (nM)
SERT (nM)
6909
Me
O
O
R2
N
F
N
1
7
1a
2
(2S,3R)
(2S,3S)
19
6940
>10,000
>10,000
O
O
O
O
F
NH
N
NH
2
3
4
5
114
27
>10000
3680
8210
676
F
N
O
3a
4
(2S,3S)
(2S,3R)
20
2570
3703
2530
Me
O
F
NH
Et
N
N
49
Absolute stereochemistry assignments change because of convention.
a
1 and 3 represent the same stereochemical configuration.
OMe
28
Table 3
F
R2
O
Binding affinities of NET active (S,S) morpholine analogs with various aryl
N
6
6
8980
substitutions
O
F
NH
O
N
S
R1
7
8
19
18
>10000
>4500
O
NH
Me
8
N
O
Entry
R1
NET (nM)
SERT (nM)
N
1
2
3
4
5
6a
7
8
9
2,6-diF
7
2
4
1210
>3500
2570
617
1620
13
6390
>10,000
>10,000
9
10
11
9
20
13
>10000
3703
445
2-F,6-OMe
2-F,6-OEt
2,4-diF
4-F
3-F
O
16
21
11
7
18
1360
2-F
O
2-Cyano
2-Methylsulfonyl
a
Entry 6 contains 20% of (R,R) isomer.
12
13
66
912
Published studies have suggested that (2R,3R) analogs of 2 are
not selective NRI’s.22 In order to further define the importance of
stereochemistry for NRI binding and MAT selectivity, some
(2R,3S) and (2S,3R) analogs were targeted. All synthesized isomers
were tested in the monoamine transporter binding assays.23 The
results in Tables 1 and 2 show that the (R, S) and (S,R) isomers lack
the desired selectivity and potency compared to the (S,S) isomers.
Having established the advantages of the (S,S) series for our pur-
poses, specific effort to pursue other diasteromers was
discontinued.
Next, we turned our attention to the phenoxy moiety and de-
signed analogs to probe both the electronic and steric effects of
modifications in this region. We observed that a number of analogs
with ortho substituents (i.e. alkyl, alkoxy, and halogen groups) on
the phenoxy ring demonstrated good NET potency and maintained
selectivity over the other MATs. A selection of morpholine analogs
displaying low nanomolar potency at NET is presented in Table 3,
along with the corresponding SERT binding affinities. In profiling
these analogs, it appeared that SERT binding affinity was affected
by phenoxy substitution, most strikingly illustrated by the 500-
fold gain in SERT potency observed in going from 2-F to 3-F substi-
tution in regard to the compounds corresponding to entries 7 and 6
in Table 3.
O
O
N
N
457
>8400
NH
one example of further elaborating an active series of compounds,
these phenoxy ring substitutions were then held constant together
with the unsubstituted morpholine ring and (2S,3S) stereochemis-
try during additional rounds of design and synthesis targeting ana-
logs in the series where R2 – phenyl (3, Fig. 1).
A number of such compounds were synthesized using the de-
scribed chemistry and several of these morpholine derivatives
are potent and selective NRI’s as demonstrated by the measured
MAT binding affinities for representative examples given in Table
4. Early in this SAR development, we replaced the phenyl group
(R2, Fig. 1) with a 2-pyridyl ring. We were pleased to find that
the R2 = 2-pyridyl analogs (entries 1 and 6, Table 4) had favorable
binding potency and NET selectivity. These compounds also had
significantly lower logD’s and an altered ADME profiles as com-
pared to analogs in Table 1. At the same time, 3-pyridyl analogs
were noticeably weaker at NET in the particular examples tested
(for example entry 2, Table 4). This result led us to design and syn-
thesize additional, related heterocyclic derivatives that were also
potent NRIs (entries 6 and 9, Table 4). Various analogs also demon-
strated the feasibility of using non-aromatic surrogates for the R2
phenyl and found, for example, that potency could be retained
Additional in-vivo and functional binding assessment of NET-
potent compounds in this series (data not shown) led to the confir-
mation that the 2- substituted and 2,6-di-substituted phenoxy ring
systems were among those present in the better NET inhibitors. As

W. Xu et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5550–5553
5553
3. Pelham, W. E.; Aronoff, H. R.; Midlam, J. K.; Shapiro, C. J.; Gnagy, E. M.; Chronis,
A. M.; Onyango, A. N.; Forehand, G.; Nguyen, A.; Waxmonsky, J. J. Pediatrics
1999, 103, e43.
4. Zamekin, A. J.; Ernst, M. N. Engl. J. Med. 1999, 340, 40.
5. Montgomery, S. A. J. Psychopharmacol. 1997, 11, 9.
6. Krell, H. V.; Leuchter, A. F.; Cook, I. A.; Abrams, M. Psychosomatics 2005, 46, 379.
7. Max, M. B.; Lynch, S. A.; Muir, J.; Shoaf, S. E.; Smoller, B.; Dubner, R. N. Engl. J.
Med. 1992, 326, 1250.
8. Bymaster, F. P.; Katner, J. S.; Nelson, D. L.; Hemrick-Luecke, S. K.; Threlkeld, P.
G.; Heiligenstein, J. H.; Morin, S. M.; Gehlert, D. R.; Perry, K. W.
Neuropsychopharmacology 2002, 27, 699.
9. Ardid, D.; Jourdan, D.; Mestre, C.; Villanueva, L.; Le Bars, D.; Eschalier, A. Brain
Res. 1995, 695, 253.
10. Zeng, F.; Jarkas, N.; Stehouwer, J.; Voll, R.; Owens, M.; Kilts, C.; Nemeroff, C.;
Goodman, M. Bioorg. Med. Chem. 2008, 16, 783.
11. Fish, P.; Deur, C.; Gan, X.; Greene, K.; Hoople, D.; Mackenny, M.; Para, K.;
Reeves, K.; Ryckmans, T.; Stiff, C.; Stobie, A.; Wakenhut, F.; Whitlock, G. Bioorg.
Med. Chem. Lett. 2008, 18, 2562.
with compounds that contained a bulky cycloalkyl such as cyclo-
pentenyl (see entry 3, Table 2 and entry 10, Table 4). Taken to-
gether with prior work, these results further define structural
features for the binding of morpholine derivatives 3 to NET and
SERT.
We have designed, synthesized and tested a number of novel,
potent, and selective NRIs. Favorable substitution patterns were
identified in our systematic SAR study of the phenoxy- and phenyl
portions of the lead structure 2. Several analogs showed promising
binding potency and selectivity for NET over SERT. A redesigned
chemical synthesis which allowed access to an expanded set of
analogs was key in enabling the discovery program, as it provided
potent NRIs which cover a range of physicochemical properties.
12. Melloni, P.; Carniel, G. C.; Torre, A. D.; Bonsignori, A.; Buonoamici, M.; Pozzi, O.;
Ricciardi, S.; Rossi, A. C. Eur. J. Med. Chem. 1984, 235.
Acknowledgments
13. Brenner, E.; Baldwin, R. M.; Tamagnan, G. Org. Lett. 2005, 7, 937.
14. Berg, S.; Larsson, L-G.; Rényi, L.; Ross, S.; Thorberg, S-O.; Thorell-Svantesson, G.
J. Med. Chem. 1998, 41, 1934.
15. Frigerio, M.; Santagostino, M.; Sputore, S. J. Org. Chem. 1999, 64, 4537.
16. More, J. D.; Finney, N. S. Org. Lett. 2002, 4, 3001.
17. van Well, R. M.; Overkleeft, H. S.; van Boom, J. H.; Coop, A.; Wang, J. B.; Wang,
H. Y.; van der Marel, G. A.; Overhand, M. Eur. J. Org. Chem. 2003, 9, 1704.
18. Oppolzer, W.; Cunningham, A. F. Tetrahedron Lett. 1986, 27, 5467.
19. Cho, B. T.; Yang, W. K.; Choi, O. K. J. Chem. Soc., Perkin Trans. 1 2001, 10, 1204.
20. Naraimhan, S.; Balakumar, R. Aldrichimica Acta 1998, 31, 19.
21. Chen, C. Y.; Reamer, R. A.; Chilenski, J. R.; McWilliams, C. J. Org. Lett. 2003, 5,
5039.
We are grateful to Paul Fish, Malcolm MacKenny, Kim Para, Cor-
ey Stiff, Thayalan Navaratham, and Greg Reichard for valuable dis-
cussions and suggestions. We would like to thank Ziqiang Wang,
Anne Akin and Frank Riley for analytical support and Mark Lovdahl
for his assistance with hydrogenations.
References and notes
1. Goldman, L. S.; Genel, M.; Bezman, R. J.; Slanetz, P. J. J. Am. Med. Assoc. 1998,
22. Serolin-Benedetti, M.; Frigerio, E.; Tocchetti, P.; Brianceschi, G.; Grazia-Castelli,
M.; Pellizzoni, C.; Dostert, P. Chirality 1995, 7, 285.
279, 1100.
2. Davids, E.; Zhang, K.; Kula, N. S.; Tarazi, F. I.; Baldessarini, R. J. J. Pharmacol. Exp.
Ther. 2002, 301, 1097.
23. NET and SERT binding were carried out in transfected HEK 293 cells measuring
displacement of radiolabelled RTI-55.