The synthesis and preliminary biological evaluation of a novel steroid with neurotrophic activity: NGA0187

Article abstract of DOI:10.1021/jo051556d

A full account of the total synthesis of neurotrophic compound NGA0187 is provided. A key feature of the synthesis involved the direct selective oxidation of 6α,7β-diol to install the unusual 6,7-ketol moiety and stereoselective conjugate addition of viny

Full text of DOI:10.1021/jo051556d

The Synthesis and Preliminary Biological Evaluation of a Novel
Steroid with Neurotrophic Activity: NGA0187
Zihao Hua,† David A. Carcache,† Yuan Tian,‡ Yue-Ming Li,‡ and Samuel J. Danishefsky*,†,§
Laboratory for Bioorganic Chemistry and Laboratory of Biochemistry and Pharmacology, Sloan-Kettering
Institute for Cancer Research, 1275 York Avenue, New York, New York 10021, and Department of
Chemistry, Columbia University, Havemeyer Hall, 3000 Broadway, New York, New York 10027
s-danishefsky@ski.mskcc.org
Received July 26, 2005
A full account of the total synthesis of neurotrophic compound NGA0187 is provided. A key feature
of the synthesis involved the direct selective oxidation of 6R,7â-diol to install the unusual 6,7-ketol
moiety and stereoselective conjugate addition of vinyl cuprate to enone. A preliminary evaluation
of its ability to stimulate the neurite outgrowth was also evaluated with PC12 cell.
Introduction
neurotrophins since NGF was first discovered 50 years
ago by Levi-Montalcini and Hamburger,1c,d neurotrophic
factors have been the focus of interdisciplinary research
for their potential in the treatment of neurologically
centered disorders. Peptidyl neurotrophic factors have
been tested in the context of neurodegenerative diseases.
However, to date, none has been demonstrably effective,
presumably because of the poor pharmacokinetic profiles
exhibited by such compounds, resulting in marginal
stability and inadequate delivery. In principle, the field
could gain significantly from the discovery and develop-
ment of nonpeptidic1e CNS-permeable molecules with
neurotrophic or neurotropic activity.
A pressing contemporary medical challenge is that of
finding effective means for the treatment of neurodegen-
erative failures such as are found in Alzheimer’s, Par-
kinson’s, Huntington’s, and Lou Gehrig’s diseases. Nerve
cells die by murder or suicide through programmed cell
death (apoptosis)1a and require trophic factors to survive.
The best characterized neurotrophic factors, or neurotro-
phins (cf. NGF, BDNF, NT-3, GDNF, NTN, ART, PSP,
NT-4, -5, -6), are naturally occurring polypeptides or
proteins which can prevent neuronal death (neurotro-
phism) or even promote axonal growth (neurotropism)
after injury.1b Because of the enormous interest in
The goal of evaluating putative small molecule neu-
rotrophic factors has been of great interest in our
laboratory in the past few years. Our approach, as per
custom, focuses on natural product leads for which
various claims of neurotrophic activity have been regis-
tered. Our discovery progression sets as its first mile-
stone, the achievement of access to these natural products
through the medium of total chemical synthesis. With
the natural product in hand, we can investigate ana-
logues synthetically derivable from the target system and
can begin to build an SAR profile. Moreover, of increasing
importance are the potentialities inherent in diverted
† Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for
Cancer Research.
‡ Laboratory of Biochemistry and Pharmacology, Sloan-Kettering
Institute for Cancer Research.
§ Columbia University.
(1) (a) Lodish, H.; Berk, A.; Zipursky, S. L.; Matsudaria, P.;
Baltimore, D.; Darnell, J. Molecular Cell Biology, 4th ed.; W. H.
Freeman & Co., New York: Section 23.8, 1999. (b) Kaneko, M.; Saito,
Y.; Saito, H.; Matsumoto, T.; Matsuda, Y.; Vaught, J. L.; Dionne, C.
A.; Angeles, T. S.; Glicksman, M. A.; Neff, N. T.; Rotella, D. P.; Kauer,
J. C.; Mallamo, J. P.; Hudkins, R. L.; Murakata, C. J. Med. Chem.
1997, 40, 1863-1869. (c) Levi-Montalcini, R.; Hamburger, V. J. Exp.
Zool. 1951, 116, 321-362. (d) Bennett, M. R.; Gibson, W. G.; Lemon,
G. Auton. Neurosci. 2002, 95, 1-23. (e) Hefti, F. Annu. Rev. Pharmacol.
Toxicol. 1997, 37, 255-257.
10.1021/jo051556d CCC: $30.25 © 2005 American Chemical Society
Published on Web 10/22/2005
J. Org. Chem. 2005, 70, 9849-9856
9849

Hua et al.
FIGURE 1. Structure of natural product neurotrophic factors.
total synthesis. By interdicting the natural product
focused total synthesis before its conclusion, we are in a
position to explore valuable chemical space which might
not be available from the target system itself. Our earlier
targets have been terpenoid compounds with claimed
neurotrophic activities. At this early stage of the work,
we do not necessarily insist on a defined molecular target
for initiation of the synthetic program. Assuming that
useful and promising activity can be corroborated and
optimized, it is expected that one or more molecular
targets will in time be identified.
Previous to the effort documented herein, our labora-
tory had accomplished the total syntheses of tricycloil-
licinone (2),2a merrilactone A (3),2b,c jiadifenin (4),2d and
scabronine G methyl ester (5)2e (Figure 1). Recently we
have initiated collaborative studies to learn about the
consequences and mechanisms of action of these small
molecule neurotrophic factors. Given our current lack of
insight at the mechanistic level, the possibilities for
failure of any particular initiative are not improbable.
Nonetheless, in the light of their records in a range of
whole cell functional assays, such molecules might
represent major opportunities at the interface of chem-
istry and the neurosciences. In short, the conditions for
admission to our total synthesis-enabled, admittedly
elitist, library are (i) promising functional properties and
(ii) a natural products lineage.
neurite outgrowth, at doses of 30 µg/mL, corresponding
to NGF at 10 ng/mL.3
To the best of our knowledge, NGA0187 is the first
discovered steroid with significant neurotrophic activity.
Furthermore, compound 1 struck us as bearing a rough
homology with compounds 2-5 in its dense confluence
of polar functions all residing on the same â-face at
carbons 3,7,11,16. The ketone at C6 provides additional
polar functionality in the core sector. Thus, we came to
look upon compound 1 in the context of a stretched out
version of the more compacted structures 2-5. Accord-
ingly, we set out to accomplish a total synthesis of
compound 1. As matters transpired, this turned out to
be a nontrivial goal, but with considerable persistence,
it was accomplished.
Results and Discussions
Overall Synthetic Strategy. The main hurdles to
overcome in a synthesis of NGA0187 were seen to include
(i) the stereoselective introduction of the side chain at
C17 with provision for the proper stereochemistry at the
C24, (ii) accommodation of the 16-acetoxyl function in the
context of the C3, C7, and C11 hydroxyl groups, (iii)
introduction and maintenance of the mutual 6-keto-7â-
hydroxy subunit, and (iv) the proper provision for the 11-
hydroxy group. While none of these implementations
appeared to be unmanageable, their accommodation in
the face of interactivity issues presented significant
problems of functional group orchestration.
Unlike the terpenoid neurotrophins 2-5 wherein the
synthesis of each presented uncertainties at the level of
skeletal architecture, the primary challenge to a synthe-
sis of compound 1 was that of functional group manipu-
lation. As will be seen in the synthesis of NGA0187, the
management of this type of problem, requiring exploita-
tion of what are often subtle reactivity differentials, may
be daunting in its own way.
The focusing target of our research herein was a
molecule termed NGA0187 (1). This compound was
isolated, and its structure was determined by Nozawa
et al. (via X-ray crystallography), in the course of a
neurotrophically based screening program directed to the
microorganism Acremonium sp. TF-0356.3 The neurite
outgrowth and cell survival activities of NGA0187 were
reported by her research group. Although compound 1
showed little ability to protect rat cerebral cortical
neuronal cells against damage, it did significantly induce
Considerable thought was given to selection of a
suitable starting material. Of course, we would begin
with a steroid. However, several possibilities still present
themselves in this context. Following careful study of the
issues and the precedents, it was decided that it would
be easier to introduce the entire steroid side chain (C20
onward) by attachment to C17 than to deal with introduc-
(2) (a) Pettus, T. R. R.; Chen, X.-T.; Danishefsky, S. J. J. Am. Chem.
Soc. 1998, 120, 12684-12685. (b) Birman, V. B.; Danishefsky, S. J. J.
Am. Chem. Soc. 2002, 124, 2080-2081. (c) Meng, Z.; Danishefsky, S.
J. Angew. Chem., Int. Ed. 2005, 44, 1511-1513. (d) Cho, Y. S.;
Carcache, D. A.; Tian, Y.; Li, Y.-M.; Danishefsky, S. J. J. Am. Chem.
Soc. 2004, 126, 14358-14359. (e) Waters, S. P.; Tian, Y.; Li, Y-M.;
Danishefsky, S. J. J. Am. Chem. Soc. 2005, 127, 13514-13515.
(3) Nozawa, Y.; Sakai, N.; Matsumoto, K.; Mizoue, K. J. Antibiot.
2002, 55, 629-634.
9850 J. Org. Chem., Vol. 70, No. 24, 2005

Novel Steroid with Neurotrophic Activity: NGA0187
SCHEME 1. Synthesis of 6r,7â-diol Steroid 11 and Attempted Acetalizationa
a Conditions: (a) AcCl, Ac2O, reflux, 17 h; (b) HC(OMe)3, (CH2OH)2, p-TsOH‚H2O, CH2Cl2, reflux, 1.5 h; (c) NaBH4, tBuOH-THF-H2O,
0 °C to room temperature, 48 h, 73% (3 steps); (d) TBSOTf, 2,6-lutidine, CH2Cl2, reflux, 12 h, 84%; (e) CrO3, 3,5-dimethylpyrazole, CH2Cl2,
-25 °C, 4 h, 51%; (f) NaBH4, CeCl3‚7H2O, THF-MeOH, -40 °C, 1 h, 98%; (g) BH3-THF, THF, 0 °C to room temperature, 12 h; NaOH,
H2O2, 2 h, rt, 81%.
SCHEME 2. Synthesis of 6r-OH-7â-OTBS Steroid
tion of the C22,23-double bond, the C-methyl at C24, and
the oxygen at C16 from a starting material which already
14a
contained some part of the side chain. Hence, our starting
steroid would contain simply a C17 keto function. Simi-
larly, we thought it more straightforward to engineer the
novel C6-keto-C7-â hydroxy diad de novo than to exploit
available steroids with existing but nonoptimal handles
already in the B-ring. Hence, we would settle for a ∆4-3
keto moiety in our precursor. In the case of C11, it was
proposed that it would actually be easier to sustain a
preexisting oxygen function at this carbon center in the
a Conditions: (a) TBSCl, Im., DMAP, DMF, 86%; (b) BH3-THF,
starting material than to install it there by de novo
THF, phosphate buffer; NaOH, H2O2, 13%.
means later in the synthesis (although this could pre-
sumably have been accomplished via a C9-C11 olefin).
Given these considerations, we settled on 11-keto-an-
drostene-3,17-dione (6) as our starting material. It was
appreciated that the distance to be traversed in getting
from 6 f 1 is quite substantial. We felt that this starting
material offered larger advantages of flexibility which
more than compensated for passing up more elaborated
structures whose added complexity could have served,
in our judgment, to narrow our available options.
Construction of 6-keto-7â-OH Ketol. A critical final
stage of our proposed synthesis would require a poten-
tially nontrivial installation of the C6-keto-C7-â-OH
arrangement. Early in the project, we sought to establish
suitable possible reaction parameters for this type of
objective, using as a model the structurally simplified
synthetic intermediate, 11. The latter was prepared from
the commercially available adrenosterone (6) in six steps,
as shown (Scheme 1). Thus, 6 was subjected to a three-
step sequence of acetylation, selective ketalization of the
C17 ketone, and NaBH4 reduction to afford the 3â,11â-
diol 7.4 Upon TBS protection of the free alcohols, inter-
mediate 8 was in hand. Treatment of 8 with CrO3-
dimethylpyrazole led to allylic oxidation at C7, furnishing
ketone 9 in moderate yield.5 Selective 1,2-reduction of
enone 9 under Luche conditions6 gave the 7â-OH steroid
10 as the only detectable epimer in high yield.7 Finally,
treatment of 10 with BH3-THF and subsequent oxidative
workup with H2O2 led to the formation of key intermedi-
ate 11.
Our first attempts to install the C6-keto-C7-â-OH ketol
focused on the strategy of selective MPM ether protection
of the 7â-OH of 11. Unfortunately, however, all of our
efforts to form p-anisyl acetal 12 from 6R,7â-OH steroid
11 were unsuccessful. This failure may be attributed to
the significant steric hindrance of the trans disposed 6,7-
diol as well as the potential susceptibility of the p-anisyl
acetal to acidic conditions.
We next investigated a strategy wherein the 7â-OH
group would be protected prior to introduction of the C6-
hydoxyl group. Thus, intermediate 10 was protected as
the TBS ether to afford 13 (Scheme 2). Unfortunately,
hydroboration of 13 gave rise to intermediate 14 in only
13% yield. The low yield may be ascribed to possible
migration of the TBS functionality from the 7â-OH to the
6R-OH group. Indeed, a similar silyl migration pattern
with a 6,7-dihydroxysteroid has been reported by Jung
et al. in their synthesis of xestobergsterols.8
We next considered the possibility of forming the ketol
through an intermediate epoxide. Thus, 10 was treated
(4) (a) Mase, T.; Ichita, J.; Marino, J. P.; Koreeda, M. Tetrahedron
Lett. 1989, 30, 2075-2078. (b) Moon, S. S.; Stuhmiller, L. M.; McMorris,
T. C. J. Org. Chem. 1989, 54, 26-28. (c) Moon, S.; Stuhmiller, L. M.;
Chadha, R. K.; McMorris, T. C. Tetrahedron 1990, 46, 2287-2306.
(5) Shen, Y.; Burgoyne, D. L. J.Org. Chem. 2002, 67, 3908-3910.
(6) Luche, J. L. J. Am. Chem. Soc. 1978, 100, 2226-2227.
(7) Di Filippo, M.; Izzo, I.; Savignano, L.; Tecilla, P.; De Riccardis,
F. Tetrahedron 2003, 59, 1711-1717.
(8) Jung, M. E.; Johnson, T. W. Tetrahedron 2001, 57, 1449-1481.
J. Org. Chem, Vol. 70, No. 24, 2005 9851

Hua et al.
SCHEME 3. Attempted Epoxide Rearrangement to 6,7-ketol Steroid 16a
a Conditions: (a) mCPBA, CH2Cl2, 0 °C to room temperature; (b) BF3‚Et2O, CH2Cl2, rt.
SCHEME 4. Synthesis of 6-keto-7â-OH Steroid 16
of the 6,7-trans diol, we returned to the synthesis of the
natural product. From intermediate 11, the next task
would be that of introducing the C17 side chain with the
correct C17 and C20 stereochemistry. We sought to ac-
complish this subgoal through cuprate addition to a
structure of the type 25. Thus, treatment of 11 with acetic
acid led to removal of the C17 ketal with concomitant
cleavage of the C3 TBS silyl ether to afford 20 (Scheme
6). Fortunately, the C6 and C7 hydroxyl groups could be
engaged as an acetonide containing the isopropylidene
protecting group.
Wittig olefination of 21 proceeded with excellent
selectivity to afford the (Z)-17(20) ethylidene 22 in 66%
yield.10 The C3 hydroxy group was then reprotected as
the TBS ether, affording 23 in quantitative yield. The
latter compound was subjected to SeO2 catalyzed allylic
oxidation to give rise to 24, which, upon MnO2 oxidation,
afforded the geometrically defined enone intermediate 25.
(17E) and (17Z)-17(20)-en-16-one steroids are configura-
tionally unstable in solution.11 However, immediate
chromatographic purification of the crude compound on
silica gel column gave the desired (17E) steroid (25) as
the major product with no observed isomerization. Thus
purified, 25 is configurationally stable and can be stored
in the solid state.
Synthesis of the Side Chain and Conjugate Ad-
dition. Vinyl iodide 29, which would serve as the
coupling partner in the side chain installation, was
prepared in three steps from commercially available
ergosterol (26). Thus, acetylation of 26, followed by
ozonolysis, provided the (R)-2,3-dimethylbutanal 28 in
moderate yield (Scheme 7).12 Although this protocol for
the preparation of aldehyde 28 is hardly “atom economi-
cal,”13 the synthesis requires only two steps and uses
inexpensive starting materials (ergosterol: $126/25 g).
Aldehyde 28 was then converted to vinyl iodide 29
according to the Takai olefination protocol.14 Compared
to the same reaction of 3-methylbutanal,15 the E/Z ratio
of the vinyl iodide product was significantly improved
from 4:1 to 93:7. The boiling point of vinyl iodide 29 is
rather low (∼60 °C), which makes it impractical to
evaporate the solvent completely. Column chromatogra-
phy on 29 was performed with the low boiling solvent,
pentane, as the eluant, and the resultant product solution
was concentrated and could be stored as a pentane
with mCPBA to afford 15, wherein the allylic 7â-hydroxyl
group directed exclusive â-face epoidation of the molecule
(Scheme 3). We now hoped to effect a rearrangement to
the ketol 16; however, treatment of 15 with BF3‚Et2O led
only to decomposition.
Finally, we decided to attempt the direct selective
oxidation of the 6,7-diol, 11. We were encouraged by a
protocol developed by Krafft in their studies toward the
synthesis of xestobergsterol A9 which allowed exclusive
benzylation of 6R-OH in the presence of 7â-OH. Our
initial attempt, using PCC as the oxidizing agent, was
unsuccessful, leading only to recovery of the starting
material (Scheme 4). Treatment of 11 with silver carbon-
ate and Celite yielded a mixture of products. Gratifyingly,
however, the reaction of 11 with Dess-Martin periodi-
nane afforded a single ketol product, which we believed
to be the desired 16. Our provisional assignment was
based on a working hypothesis that an appropriate
oxidizing reagent would approach the hydroxyl group
only from the sterically less hindered R face, resulting
in the selective oxidation of the 6R-OH group.
To confirm the structure of 16, we prepared the 6R-
OH-7-keto steroid (19). Thus, selective benzoylation of
11 under Krafft’s conditions afforded the monobenzoy-
lated intermediate 17, as expected (Scheme 5). Oxidation
of the 7â-OH with Dess-Martin periodinane afforded 18,
which, upon deprotection, gave rise to the 6R-OH-7-keto
steroid 19. Comparison of the two ketol steroids, 19 and
1
16, by H NMR revealed two different compounds, thus
confirming that the product formed in the direct oxidation
was indeed the desired keto steroid 16. We could now be
confident that we had established suitable conditions for
the late-stage installation of the 6-keto-7â-OH function-
ality directly from the 6R,7â-diol.
(10) Konno, K.; Ojima, K.; Hayashi, T.; Takayama, H. Chem. Pharm.
Bull. 1992, 40, 1120-1124.
(11) Schmuff, N. R.; Trost, B. M. J. Org. Chem. 1983, 48, 1404-
1412.
Synthesis of ∆17(20)-16-one Steroid 25. Having
established, in principle, a means for late-stage oxidation
(12) Tsuda, K.; Kishida, Y.; Hayazu, R. J. Am. Chem. Soc. 1960,
82, 3396-3399.
(13) Trost, B. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 259-281.
(14) Takai, K.; Nitta, K.; Utimoto, K. J. Am. Chem. Soc. 1986, 108,
7408-7410.
(9) Krafft, M. E.; Dasse, O. A.; Fu, Z. J. Org. Chem. 1999, 64, 2475-
2485.
(15) Coleman, R. S.; Liu, P.-H. Org. Lett. 2004, 6, 577-580.
9852 J. Org. Chem., Vol. 70, No. 24, 2005

Novel Steroid with Neurotrophic Activity: NGA0187
SCHEME 5. Synthesis of 6r-OH-7-keto Steroid 19a
a Conditions: (a) BzCl, Et3N, DMAP, CH2Cl2, 0 °C to room temperature, 4.5 h; (b) Dess-Martin periodinane, CH2Cl2, rt, 1 h; (c) 1%
KOH, MeOH, rt, 2.5 h.
SCHEME 6. Synthesis of ∆17(20)-16-one Steroid 25a
a Conditions: (a) AcOH-H2O (4:1), rt, 2.5 h, 82%; (b) Me2C(OMe)2, CSA, DMF, rt, 6h, 93%; (c) EtPPh3Br, KOtBu, THF, reflux, 66%; (d)
TBSOTf, 2,6-lutidine, CH2Cl2, rt, quant.; (e) SeO2, TBHP, CH2Cl2, 0 °C to room temperature, 6 h; (f) MnO2, CH2Cl2, reflux, 3.5 h, 43%
(two steps).
SCHEME 7. Synthesis of Vinyl Iodide 29a
to afford the 16â-OH steroid 32, which was subsequently
acylated to produce 33 (Scheme 9).16 Standard aqueous
acidic conditions (TsOH, PPTS, Pd(CH3CN)2Cl2, etc.) for
acetonide removal resulted in concomitant cleavage of the
C3 TBS ether. However, a slightly modified Williams’
procedure17 of ketal exchange with 1,2-ethanedithiol in
the presence of catalytic amounts of CSA in anhydrous
chloroform selectively removed the 6,7-acetonide to give
steroid 34 containing free hydroxy groups at C6 and C7.
The previously developed protocol for selective oxidation
of the 6R,7â-diol to the 6-keto-7â-OH steroid proceeded
smoothly to afford 35 as a single detectable product.
Unfortunately, however, the removal of the TBS protect-
ing groups proved troublesome. Although the C3 TBS
ether could be easily cleaved, the C11 TBS ether remained
intact under all conditions attempted. The difficulty in
removing the C11 TBS ether may be explained by the
significant steric hindrance caused by the axial methyl
groups projecting from C10 and C13. To complete the
synthesis of NGA0187, it would be necessary to recon-
sider our protecting group selections.
a Conditions: (a) Ac2O, Et3N, DMAP, CH2Cl2, rt, 12 h; (b) O3,
pyridine, CH2Cl2, -78 °C; then Me2S, rt. 38% (two steps); (c) CrCl2,
CHI3, THF, 0 °C to room temperature, 4 h, 45% (E:Z ) 93:7).
solution. However, best yields in subsequent coupling
reactions were obtained with freshly prepared material.
With enone 25 and vinyl iodide 29 in hand, the
installation of the C17 side chain via cuprate conjugate
addition was studied. Treatment of vinyl iodide 29 with
t-BuLi afforded the vinyllithium agent which, upon
treatment with CuCN, yielded high-order vinyl cuprate
species, 30. Stereoselective addition of the vinyl cuprate
from the Si face of the ∆17(20)-16-one, followed by kinetic
protonation of the resultant enolate from the same face,
furnished steroid 31 with the desired stereochemistry at
both C17 and C20 (Scheme 8).
First Attempted Synthesis of NGA0187. With the
C17 side chain stereoselectively installed, several func-
tional group modifications would be required for the
completion of the synthesis of NGA0187. First, the C16
ketone of 31 was stereoselectively reduced with LiAlH4
In light of these findings, we decided to replace the C11
protecting group with a more labile TES function. We
first attempted to change the protecting groups on the
relatively late stage intermediate 22; however, the C11
(16) Bruttomesso, A. C.; Doller, D.; Gros, E. G. Synth. Commun.
1998, 28, 4043-4057.
(17) Williams, D. R.; Sit, S. Y. J. Am. Chem. Soc. 1984, 106, 2949-
2954.
J. Org. Chem, Vol. 70, No. 24, 2005 9853

Hua et al.
SCHEME 8. Conjugate Addition
SCHEME 9. First Attempted Synthesis of NGA0187a
a Conditions: (a) LiAlH4, THF, -78 °C; (b) Ac2O, DMAP, pyridine, 0 °C to room temperature; (c) (CH2SH)2, CSA, CHCl3, reflux; (d)
Dess-Martin periodinane, CH2Cl2, rt.
SCHEME 10. Synthesis of C-11-TES Steroid 44a
a Conditions: (a) TESOTf, pyridine, CH2Cl2, 0 °C, 1 h, 87%; (b) CrO3, 3,5-dimethylpyrazole, CH2Cl2, -25 °C, 4 h, 51%; (c) NaBH4,
CeCl3‚7H2O, THF-MeOH, -70 to -30 °C, 1 h, 96%; (d) Ac2O, DMAP, pyridine, CH2Cl2, rt, 4 h, 98%; (e) BH3-THF, THF, 0 °C to room
temperature, 12 h; NaOH, H2O2, 2 h, rt, 76%; (f) AcOH-H2O, rt, 2.5 h; (g) 2, 2-dimethoxypropane, CSA, DMF, rt, 8 h; (h) EtPPh3Br,
KOtBu, THF, reflux, 3 h, 67% (three steps); (i) TBSOTf, 2,6-lutidine, CH2Cl2, rt, 1 h, 98%; (j) SeO2, t-BuOOH, CH2Cl2, 0 °C, 5 h, 72%; (k)
Dess-Martin periodinane, CH2Cl2, rt, 1 h, 98%.
ether was resistant to cleavage under several conditions
attempted. We were then forced to return to the early
stage diol intermediate 7. Thus, compound 7 was con-
verted to the di-TES ether 36. The latter was subjected
to allylic oxidation to afford 37, which was selectively
reduced to give rise to 38. Subsequent hydroboration of
38 with oxidative workup provided diol 40 in moderate
yield (47%). The observed yield is lower than that
obtained with the TBS ether substrate (81%) because of
formation of the 6â,7â-diol byproduct. Alternatively,
alcohol 38 was first acetylated to afford 39. Upon
exposure to the hydroboration-oxidation conditions, in-
9854 J. Org. Chem., Vol. 70, No. 24, 2005

Novel Steroid with Neurotrophic Activity: NGA0187
SCHEME 11. Completion of the Synthesis of NGA0187 (1)a
a Conditions: (a) 29, t-BuLi, -78 °C, Et2O, 0.5 h, then CuCN, -40 °C, 0.5 h; 44, -70 °C, 2 h, 61% (93% brsm); (b) LiAlH4, THF, -78
°C, 1 h; (c) Ac2O, DMAP, Py., 0 °C, 4 h; (d) (CH2SH)2, CSA, CHCl3, reflux, 1 h, 61% (three steps); (e) Dess-Martin periodinane, CH2Cl2,
rt, 3 h; (f) HF-CH3CN-H2O, rt, 5 h, 61% (two steps).
FIGURE 2. ORTEP drawing of NGA0187.
termediate 39 was converted to diol 40 in 76% yield.
Treatment of 40 with acetic acid led to removal of the
C17 ketal and cleavage of the C3 TES functionality. The
resultant triol was converted to intermediate 41 through
treatment with 2,2-dimethoxypropane in the presence of
catalytic amounts of CSA, followed by selective Wittig
olefination. The C3 hydroxy group was then converted to
the TBS ether, 42. SeO2 catalyzed oxidation furnished
allylic alcohol 43, which was subsequently oxidized to 44
under Dess-Martin oxidation conditions.
Completion of the Synthesis of NGA0187. The
TES-protected steroid 44 was further advanced, accord-
ing to the chemistry described above for the conversion
from 31 to 35. Thus, conjugate addition of the vinyl
cuprate generated from 29 to steroid 44 afforded 45 with
good yield (Scheme 11). Reduction of the C16 ketone,
acetylation of the resultant 16â-OH, and removal of
acetonide through ketal exchange with thiol afforded
6R,7â-diol steroid 46 in 61% yield over three steps.
Subjection of 46 to Dess-Martin periodinane in CH2Cl2
at room temperature led to the selective oxidation of the
6R-OH to the ketone. Finally, treatment of the crude ketol
steroid with 48% HF in CH3CN (1:9) led to cleavage of
both TBS at C3 and TES at C11 to give the natural
product, NGA0187 (1). The overall yield for 20 steps from
adrenosterone (6) is 3%. The structure of synthesized
1
NGA0187 was analyzed by H, 13C, DEPT, COSY, and
HMQC NMR techniques (CDCl3 was used as solvent
instead of the originally used d6-DMSO in isolation
paper). Moreover, to avoid any chance for error, the
structure was unambiguously confirmed by X-ray dif-
fraction on the fully synthetic material (shown in Figure
2).
Biological Evaluation of Synthesized NGA0187
and Analogues. We were now prepared to verify the
neurotrophic activity previously claimed for NGA0187.
The ability of synthetic NGA0187 to stimulate NGF-
mediated neurite outgrowth was evaluated by treating
rat pheochromocytoma cells (PC12) with NGF (50 ng/mL)
and NGA0187 (30 µM) for 96 h. The cells were then
compared to a similarly prepared control, which lacked
only the NGA0187. As shown in Figure 3, although no
neurite growth was observed in the control sample
(picture A), the sample treated with 30 µM NGA0187
demonstrated considerable sprouting (picture B).
Having corroborated the earlier reported neurotrophic
activity, we next attempted to establish a very prelimi-
nary SAR map for NGA0187. As the ultimate objective
J. Org. Chem, Vol. 70, No. 24, 2005 9855

Hua et al.
FIGURE 3. Images of neurons after treatment with (A) DMSO + NGF, (B) NGA0187 in DMSO (30 µM) + NGF.
of drug discovery is the development of biologically active,
structurally simplified analogues, we decided to examine
the effects of removing the C17 side chain. In addition,
we hoped to evaluate in a preliminary way the impact of
manipulating the oxidation states of the C6 and C16
hydroxyl components. Thus, three structurally simplified
analogues, 47, 48, and 49, were prepared and evaluated
at 30 µM under conditions similar to those employed in
the biological evaluation of NGA0187 itself.
FIGURE 4. Evaluation of neurite growth. Value represents
mean ( SE for 10 views from 5 wells.
Certainly, the prospect of using a steroid framework as
a matrix for attaching pharmacophores to enhance neu-
rotrophic activities is an attractive one. The loss of
activity with some of the derivatives described above
lacking the C17-side chain underscores the need to
pinpoint the structural specificity required for activity.
This will be necessary to foster additional potency before
nominating one of this family of compounds for serious
clinical development.
Acknowledgment. Support for this research was
provided by the National Institute of Health (Grant No.
AI16943) D.A.C. gratefully acknowledges the Novartis
Foundation and Roche Research Foundation for a post-
doctoral fellowship. We thank Dr. Atsushi Endo and Dr.
William F. Berkowitz for helpful discussions, Dr. Louis
Todaro (Hunter College New York) for X-ray structure
analysis, and Dr. George Sukenick and Ms. Hui Fang
(NMR Core Facility, Sloan-Kettering Institute, CA-
02848) for NMR and mass spectral analyses, respec-
tively. We are grateful to Dr. Yuriko Nozawa for
providing an authentic sample of NGA0187.
Interestingly, none of these analogues promoted neu-
rite outgrowth to any appreciable extent (Figure 4). The
lack of activity of compound 48 is especially informative,
as this analogue differs from NGA0187 only in the C17
side chain. These results suggest that the C17 side chain
possibly plays a critical role in target recognition.
Compound 1 (NGA0187) significantly promoted (269%)
neurite outgrowth relative to the DMSO control (*P <
0.001). Compound 49 promoted differentiation about
123% comparing to control, but it is not significant (P >
0.5). Compounds 47 and 48 do not have any effect.
In retrospect, the decision to begin with compound 6
eventually worked out very nicely, allowing us to exploit
the reasonably well understood steroidal framework to
selectively process functionality at different points with
high stereoselection. We are now confident that synthe-
sis, while nontrivial, will provide for us access to sub-
strates with which to map an informative SAR profile.
Supporting Information Available: Experimental pro-
cedures and characterization data for compounds 1, 8-11, 16-
19, 22, 23, 25, 29, 31, 34, and 36-46 and CIF file for compound
1. This material is available free of charge via the Internet at
http://pubs.acs.org.
JO051556D
9856 J. Org. Chem., Vol. 70, No. 24, 2005