A non-calcemic sulfone version of the vitamin D3 analogue seocalcitol (EB 1089): chemical synthesis, biological evaluation and potency enhancement of the anticancer drug adriamycin

Article abstract of DOI:10.1016/S0968-0896(01)00159-6

Novel side-chain diene sulfones 5, analogues of the natural hormone 1α 25-dihydroxyvitamin D₃ (calcitriol, 1), were designed to incorporate some of the therapeutically most favorable structural features of the Leo Pharmaceutical Companys drug candidate diene EB 1089 (seocalcitol, 4) and of the Hopkins' non-calcemic side-chain sulfone analogues 2 and 3. Synthesis of diene sulfones 5 features selective Swern oxidation of a primary silyl ether in the presence of a secondary silyl ether (9→10) and Horner Wadsworth Emmons aldehyde addition by a 1-phosphonyl-3-sulfonyl stabilized carbanion regiospecifically at the 1-position to form E,E-diene sulfone 11. Sulfone diene analogue 5a with natural 1α,3b-diol functionality, but not its diastereomer 5b with unnatural A-ring stereochemistry, is antiproliferative in vitro toward murine keratinocytes and malignant melanoma cells, as well as toward MCF-7 human breast cancer cells. Combining diene sulfone 5a with the currently used anticancer drug adriamycin (ADR) caused a noteworthy 3-fold enhancement of ADR antiproliferative potency in MCF-7 cells. Sulfone diene analogue 5a is weakly active transcriptionally in MCF-7 and ROS 17/2.8 cells, binds poorly but measurably to the vitamin D receptor (VDR), and desirably is non-calcemic in vivo at a daily dose (7 days) of 10 mg/kg of rat body weight. Copyright

Full text of DOI:10.1016/S0968-0896(01)00159-6

Bioorganic & Medicinal Chemistry 9 (2001) 2365–2371
A Non-Calcemic Sulfone Version of the Vitamin D Analogue
3
Seocalcitol (EB 1089):Chemical Synthesis, Biological Evaluation
and Potency Enhancement of the Anticancer Drug Adriamycin
a,
a
b
c
c
Gary H. Posner, * Kenneth R. Crawford, Sara Peleg, Jo-Ellen Welsh, Saara Romu,
e
d
d
e
David A. Gewirtz, Mona S. Gupta, Patrick Dolan and Thomas W. Kensler
aDepartment of Chemistry, School of Arts and Sciences, The Johns Hopkins University, Baltimore, MD 21218, USA
bDepartment of Medical Specialties, The University of Texas, M.D. Anderson Cancer Center, Houston, TX 77030, USA
cDepartment of Biological Sciences, University of Notre Dame, Notre Dame, IN 46556, USA
dDepartment of Pharmacology and Toxicology, Virginia Commonwealth University, Medical College of Virginia,
Richmond, VA 23298, USA
e
Department of Environmental Health Sciences, School of Hygiene and Public Health, The Johns Hopkins University,
Baltimore, MD 21205, USA
Received 1 February 2001; accepted 30 April 2001
In honor of Peter Dervan, winner of the 2000 Tetrahedron Prize.
3
Abstract—Novel side-chain diene sulfones 5, analogues of the natural hormone 1a,25-dihydroxyvitamin D (calcitriol, 1), were
designed to incorporate some of the therapeutically most favorable structural features of the Leo Pharmaceutical Company’s drug
candidate diene EB 1089 (seocalcitol, 4) and of the Hopkins’ non-calcemic side-chain sulfone analogues 2 and 3. Synthesis of diene
sulfones 5 features selective Swern oxidation of a primary silyl ether in the presence of a secondary silyl ether (9!10) and Horner–
Wadsworth–Emmons aldehyde addition by a 1-phosphonyl-3-sulfonyl stabilized carbanion regiospecifically at the 1-position to
form E,E-diene sulfone 11. Sulfone diene analogue 5a with natural 1a,3b-diol functionality, but not its diastereomer 5b with
unnatural A-ring stereochemistry, is antiproliferative in vitro toward murine keratinocytes and malignant melanoma cells, as well as
toward MCF-7 human breast cancer cells. Combining diene sulfone 5a with the currently used anticancer drug adriamycin (ADR)
caused a noteworthy 3-fold enhancement of ADR antiproliferative potency in MCF-7 cells. Sulfone diene analogue 5a is weakly
active transcriptionally in MCF-7 and ROS 17/2.8 cells, binds poorly but measurably to the vitamin D receptor (VDR), and
desirably is non-calcemic in vivo at a daily dose (7 days) of 10 mg/kg of rat body weight. # 2001 Elsevier Science Ltd. All rights
reserved.
Introduction
hormone’s tertiary alcohol group located toward the
end of the side chain for effective binding to the vitamin
D receptor (VDR). In 1999, we showed for the first
4
Using vitamin D analogues for chemotherapy against
3
various human diseases (e.g., cancer, osteoporosis,
psoriasis) requires separating desirable antiproliferative
and prodifferentiating activities from undesirable calce-
time that replacing the side-chain tertiary hydroxyl
(OH) group by an unorthodox terminal sulfone group
produced analogues (e.g., 2) that maintained the natural
hormone’s antiproliferative potency in vitro but without
1ꢀ3
mic activity.
vitamin D (calcitriol, 1), unfortunately, is too strongly
The natural hormone 1a,25-dihydroxy-
5
significant calcemic activity in vivo. In 2000, we
3
calcemic for such general medicinal applications. As
molecular architects, synthetic organic chemists worldwide
have structurally altered small portions of the natural
hormone’s skeleton to produce therapeutically valuable
showed for the first time that sulfone analogue 3 with a
different type of side-chain modification had similarly
desirable separation of antiproliferative from calcemic
6
activities while being relatively stable metabolically.
4
analogues. Most of these analogues retain the natural
One of the current leading calcitriol (1) analogue drug
candidates is Leo Pharmaceutical Company’s side-chain
conjugated diene allylic alcohol EB 1089 (seocalcitol, 4);
*Corresponding author. Fax: +1-410-516-8420; e-mail: ghp@jhu.edu
0
968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00159-6

2366
G. H. Posner et al. / Bioorg. Med. Chem. 9 (2001) 2365–2371
such side-chain unsaturation was incorporated, at least
in part, to retard metabolic deactivation (i.e., P-450
Biology
7
,8
oxidation) near the 26-hydroxyl group. Incorporating
some of the therapeutically best structural features of
this successful Leo analogue EB 1089 (4) and also of our
promising sulfone analogues 2 and 3, we have now
designed, prepared, and evaluated side-chain diene sul-
fones 5.
The in vitro antiproliferative potencies of sulfones 5
and of calcitriol (1), determined using our previously
1
2
described murine keratinocyte protocol, are shown in
Figure 1. Diastereomer 5a having the natural A-ring
hydroxyl group stereochemistry (i.e., 1a,3b) is much
more potent than the corresponding unnatural A-ring
diastereomer 5b, a trend found previously in related
5
,6
diastereomeric pairs of sulfone analogues. At 100-nM
concentration, diastereomer 5a and calcitriol (1) have
comparable antiproliferative potencies in murine kera-
tinocytes. Similar results were observed in vitro in mur-
ine malignant melanoma cells (data not shown). In
MCF-7 human breast cancer cells, analogue 5a at
100 nM reduced cell proliferation by 40% after 96 h,
compared directly with EB 1089 (4) at 100 nM causing a
50% reduction in cell proliferation. At a concentration
of only 10 nM, however, diene 5a inhibited MCF-7 cell
growth only weakly, by approximately 10%. Com-
bining this analogue 5a (10 nM) with adriamycin
Chemistry
9
Arbusov substitution on allylic bromide 6 cleanly pro-
duced phosphonate sulfone 7 (Scheme 1). Selective
1
0
Swern oxidation of the primary triethylsilyl ether
group of bis-silyl ether 9 produced aldehyde 10 without
disturbing the secondary silyl ether functionality. Hor-
ner–Wadsworth–Emmons (HWE) condensation of
aldehyde 10 with the conjugate base of phosphonate
sulfone 7 proceeded via attack of this allylic carbanion
regiospecifically by the carbanionic center adjacent to
phosphorus to form E,E-diene sulfone 11.11 Final HWE
condensation of the conjugate base of racemic A-ring
allylic phosphine oxide 13, prepared as described pre-
13
(ADR), a currently used anticancer drug, shifted the
growth inhibition dose–response curve (Fig. 2) for ADR
favorably by approximately 3-fold (from IC50 of 230 to
85 nM). The magnitude of this desirable shift is similar
to that reported previously for the combination of ADR
with the 16-ene-23-yne-25-OH drug candidate ILX-23-
1
2
viously, with enantiomerically pure C,D-ring ketone
2 and HF-promoted desilylation gave the desired side-
1
chain analogues 5 as a mixture of diastereomers. This
desilylation step was successful with HF, without dis-
rupting the conjugated triene system, whereas more
commonly used nucleophilic tetrabutylammonium fluo-
ride was avoided because of its tendency to add fluoride
1
4
7553. Whether this synergistic effect of diene 5a plus
ADR is due to the capacity of diene 5a to confer sus-
ceptibility to apoptosis, as shown previously for other
1
4,15
calcitriol (1) analogues,
remains to be determined.
5
anion in a Michael fashion to a,b-unsaturated sulfones.
When administered to rats orally for seven consecutive
days at a 20-times higher dose than calcitriol (1),
diene 5a was not statistically different from vehicle
control in terms of levels of urinary calcium excretion
(Fig. 3). Furthermore, unlike calcitriol and unlike EB
Diene sulfones 5 were separated by semipreparative
HPLC to afford pure target compounds 5a and 5b, each
diastereomer enantiomerically pure, characterized fully
5
,6
and distinguished from each other as in similar cases
1
7
by their characteristic H NMR resonances and optical
rotations (Table 1).
1089 (4), non-calcemic diene sulfone 5a did not com-
promise animal weight gain when administered orally

G. H. Posner et al. / Bioorg. Med. Chem. 9 (2001) 2365–2371
2367
Scheme 1.
1
ꢁ
Table 1. H NMR (d) and optical rotation ( ) characteristics of
sulfones 5
the various vitamin D analogues also in MCF-7 human
3
breast cancer cells, we transiently transfected MCF-7
cells with a vitamin D reporter construct derived from
2
5
3
Analogue
C6-H
C19a
[a]
D
the rat 24-hydroxylase (24-OHASE) promoter. This
construct, which consists of approximately 300 bp of the
native rat 24-OHASE promoter linked to a luciferase
5
5
a
b
6.37
6.38
5.32
5.31
+96.6
+31.6
reporter gene, is one of the most potent vitamin D
1
3
7
responsive reporter systems characterized to date.
Transactivation of the 24-OHASE reporter gene in
for seven consecutive days at a dose of 10 mg/kg of body
weight.
response to increasing doses of calcitriol (1) and two
vitamin D analogues is shown in Figure 4. As expected,
3
The in vitro vitamin D receptor-mediated transcrip-
tional potencies of diene analogues 5 were determined,
the natural hormone 1 dose-dependently enhanced
reporter gene activity, with an 18-fold induction relative
to vehicle-treated cells at the 100-nM dose. Analogue 4
(EB 1089) exhibited a higher potency than the natural
hormone 1, with significant induction of reporter gene
activity (10-fold over vehicle-treated cells) at a dose as
low as 1 nM. However, higher doses of analogue 4 (10–
100 nM) did not further enhance reporter gene activity,
suggesting that analogue 4 (EB 1089) is a partial
agonist. Thus, despite the higher potency of analogue 4
at low doses, the maximal reporter gene activity for
analogue 4 was not higher than that of the natural hor-
mone 1 (compare 100-nM doses for each compound).
Although low (1–10 nM) doses of analogue 5a were
inactive, this analogue did up-regulate vitamin D3
responsive reporter gene activity 5-fold compared to
1
6
as described previously, in rat osteosarcoma ROS 17/
.8 cells using a reporter gene containing the osteocalcin
vitamin D responsive element. The ED50 values for
2
transcriptional activities are as follows: calcitriol (1),
0
.4 nM; 5a, 20 nM; 5b, no activity. The ED50 for sulfone
a was calculated from a dose–response plot ranging
5
from 0.01 to 1000 nM. This compound 5a induces 80%
of calcitriol (1) maximal response at concentrations of
1
1
00 nM and 100% of calcitriol (1) maximal response at
000 nM. These data mean that sulfone 5a has the same
efficacy as that of calcitriol (1) but the potency of sul-
fone 5a is lower than that of calcitriol (1). These results
imply that sulfone 5a is not a partial agonist in ROS 17/
2
.8 cells. To determine the transcriptional potency of

2368
G. H. Posner et al. / Bioorg. Med. Chem. 9 (2001) 2365–2371
Figure 1. Dose–response effects of analogues on keratinocyte proliferation (96 H).
with these transcriptional potencies, relative binding
5
affinities to VDR are as follows: calcitriol (1) 100%;
a, 2%.
5
In conclusion, diene 5a represents a new type of side-
chain sulfone analogue that, despite lacking the classical
5
,6
25-OH group,
including especially considerable in vitro anti-
has selective biological activities,
proliferative potency; unlike calcitriol (1) and unlike EB
1089 (4), diene 5a is non-calcemic in vivo. This new
sulfone 5a has potential as a sensitive molecular probe
of ligand–receptor interactions. Further biological eva-
luation of this new chemical entity will reveal its full
biological profile and its possible medicinal value, alone
or combined with established anticancer drugs, espe-
cially in comparison with Leo’s promising side-chain
diene drug candidate seocalcitol (EB 1089, 4).
Experimental
Figure 2. Growth inhibition in MCF-7 cells.
Unless otherwise noted, all reactions were performed in
oven-dried glassware stirred under an atmosphere of
ultra-high-purity Ar. Et O and THF were distilled from
ethanol-treated cells when used at the 100-nM dose
level. By extrapolation, we estimate that the transcrip-
tional potency of analogue 5a was approximately equal
to that which would be observed with 5 nM of the nat-
ural hormone 1 in this system. In comparing sulfone
analogue 5a to Leo analogue 4, we conclude that the
structural modifications in 5a clearly reduced, but
did not eliminate, its ability to transactivate the
vitamin D3 responsive reporter gene. Consistent
2
Na/benzophenone ketyl immediately prior to use.
DMSO, CH Cl , and Et N were distilled from CaH .
2
2
3
2
Organolithiums were titrated prior to use following
known methods. All other reagents were used as
received from commercial suppliers. Analytical TLC
analysis was conducted on precoated glass-backed silica
gel plates (Merck Kieselgel 60 F254, 250 mm thickness)
and visualized with p-anisaldehyde or KMnO stains.
4

G. H. Posner et al. / Bioorg. Med. Chem. 9 (2001) 2365–2371
2369
Polarimeter. NMR spectra were obtained on a Varian
1
XL-400 spectrometer, operating at 400 MHz for H and
1
3
1
00 MHz for C. Chemical shifts are reported in ppm
(
d) and are referenced to CDCl (7.26 and 77.0 ppm).
3
1
UV spectra were obtained using a Beckman DU -70
spectrophotometer at ambient temperature. IR spectra
were obtained using a Perkin-Elmer 1600 Series FT-IR
instrument. HRMS were obtained with electronic or
chemical ionization (EI or CI) at the University of Illi-
nois at Urbana—Champaign on a Finnigan-MAT CH5,
a Finnigan-MAT 731, or a VG Instruments 70-VSE
spectrometer run at an ionizing voltage of 70 eV for EI
and run with methane (CH ) as a carrier gas for CI.
4
Elemental analyses were performed by Atlantic Micro-
lab Inc., Norcross, GA.
E-3-Bromo-1-propenyl t-butyl sulfone 6. To a solution of
sodium hydroxide (2.40 g, 60.0 mmol) in ethanol
ꢁ
(
50 mL) at 50 C was added dropwise via syringe 2-
methyl-2-propanethiol (6.76 mL, 60 mmol) over 2 min.
The resulting solution was stirred for 30 min while
cooling to rt. Upon dropwise addition of allylbromide
(5.19 mL, 60 mmol), the reaction mixture was heated to
reflux and stirred overnight. The resulting suspension
was filtered to remove NaBr, which was rinsed thor-
3
Figure 3. Effects of vitamin D analogues on urinary calcium excre-
tion in rats. Animals were treated with 0.5–1.0 mg/kg body weight of
test compound po for seven consecutive days, and urinary excretion of
calcium was measured during days 3–7. Values are meanꢃSE from
three animals in each group. *Signifies statistically different (p>0.05)
from vehicle.
oughly with Et O. The filtrate was neutralized with 1 M
2
HCl (50 mL), diluted with H O (100 mL), extracted with
2
Et O (3ꢂ25 mL), washed with H O, dried over Na SO ,
2
2
2
4
filtered, and concentrated to a pale-yellow oil by
removing EtOH and Et O via gentle distillation.
2
This pale-yellow oil (7.81 g theor) was dissolved in
ꢁ
2
15 mL of methanol and cooled to 0 C. To this solution
1
was added a 26% oxone solution (55.3 g, 90.0 mmol in
50 mL H O) dropwise via addition funnel. The result-
1
2
ing suspension was allowed to stir with gradual warm-
ing to room temperature overnight. The reaction
mixture was then diluted with 150 mL H O, extracted
2
with CHCl (3ꢂ75 mL), washed with H O and brine
3
2
(150 mL each), dried over Na SO , filtered, and con-
centrated to give essentially pure allyl t-butyl sulfone as
2 4
a clear oil.
Figure 4. Data are expressed as relative luciferase activity (i.e., values
of treated cells relative to that of ethanol-treated control cells, which
were normalized to 1). Each bar represents the meanꢃSEM of values
obtained from three independent cell extracts. *p<0.001, treated ver-
sus ethanol control values. Similar data were obtained with these three
compounds in independent experiments utilizing a distinct clone of
MCF-7 cells.
A portion of this sulfone (1.00 g, 6.16 mmol) was dis-
solved in CCl (40 mL), treated with bromine (0.317 mL,
6.16 mmol) for 3 h, and concentrated to give the desired
dibromide (1.46 g, 74%) as a pale-orange solid (recryst
4
ꢁ
from cold Et O/hexanes; mp 75–80 C) clean enough for
2
the following transformation.
Column chromatography was performed using flash
silica gel (particle size 230–400 mesh). HPLC was car-
ried out using a Rainin HPLX system equipped with
two 25-mL/min preparative pump heads using Rainin
Dynamax 10ꢂ250-mm (semipreparative) columns
The dibromide was dissolved in THF (50 mL), cooled to
ꢁ
0 C, and treated with Et N (0.945 mL, 6.78 mmol)
3
overnight, while gradually warming to room tempera-
ture. The reaction mixture was neutralized with dilute
HCl, diluted with H O, extracted with Et O (3ꢂ25 mL),
˚
packed with 60 A silica gel (8 mm pore size) as C-18-
2
2
bonded silica and a Rainin Dynamax UV-C dual-beam
variable-wavelength detector set at 264 nm. Yields are
reported for pure products (>95% based on their
chromatographic and spectroscopic homogeneity) and
are unoptimized. Melting points were determined in
open capillaries using a Mel-Temp metal-block appara-
tus and are uncorrected. Optical rotations were mea-
sured at the Na line using a Perkin-Elmer 141
dried over Na SO , filtered, concentrated, and purified
2 4
by column chromatography (10–20% Et O/pentane) to
2
give 6 (385 mg, 26% over two steps from crude allyl t-
ꢁ
1
butyl sulfone) as white crystals: mp 60–62 C; H NMR
(CDCl ) d 7.00 (dt, J=15, 6.8 Hz, 1H), 6.55 (dt, J=15,
3
1.2 Hz, 1H), 4.07 (dd, J=6.8, 1.2 Hz, 2H), 1.38 (s, 9H);
1
3
C NMR (CDCl ) d 143.8, 127.6, 58.9, 27.4, 23.2; IR
3
(neat) 3068, 3053, 2976, 1636, 1305, 1275, 1109, 982.

2370
G. H. Posner et al. / Bioorg. Med. Chem. 9 (2001) 2365–2371
Anal. calcd for C H BrO S: C, 34.86; H, 5.43. Found:
7
HRMS: calcd for C H O SSi: 468.3093, found
26 48 3
13
2
C, 34.98; H, 5.32.
468.3094.
Di-n-butylphosphonate 7. A stirred solution of tri-n-
butyl phosphite (5 mL) and sulfone
1
C,D-Ring ketone (+)-12. An aqueous solution of HF
(0.500 mL, 10% wt) was added dropwise to a stirred
solution of (+)-11 in H O:THF (1.00 mL H O:few
6 (370 mg,
.53 mmol) was heated to 130 C for 12 h, cooled to
ꢁ
room temperature, concentrated and purified on silica
2
2
drops of THF to effect solution) and the resulting solu-
tion stirred at room temperature for 2 days [two addi-
tional 0.500 mL portions of 10% HF(aq) were added
over this period to consume starting material]. The
reaction mixture was then carefully neutralized (satd.
NaHCO ), diluted with H O, and extracted with
gel (75% EtOAc/hexanes) to give 519 mg (95%) of 7 as
1
a clear oil: H NMR (CDCl ) d 6.74 (dt, J=15, 8.0 Hz,
3
1
4
4
6
H), 6.38 (ddt, J=15, 4.6, 1.2 Hz, 1H), 4.03–3.92 (m,
H), 2.76 (ddd, J=25, 8.0, 1.2 Hz, 2H), 1.62–1.53 (m,
H), 1.38–1.30 (m, 4H) 1.29 (s, 9H), 0.85 (t, J=7.2 Hz,
3
2
1
3
H); C NMR (CDCl ) d 140.2 (d, J=11 Hz), 128.2 (d,
CH Cl (3ꢂ10 mL). The combined organics were dried
3
2
2
J=14 Hz), 66.2 (d, J=6.8 Hz), 58.7, 32.5 (d, J=6.1 Hz),
(Na SO ), filtered, and concentrated to give the desired
2 4
3
2
1
8
0.0 (d, J=140 Hz), 23.2, 18.7, 13.5; IR (neat) 2961,
935, 2874, 1633, 1477, 1464, 1289, 1246, 1114, 1066,
024, 983. Anal. calcd for C H O PS: C, 50.83; H,
hydroxy sulfone as a white solid.
To this crude hydroxy sulfone in CH Cl (3 mL) was
2
1
5
31
5
2
.82; S, 9.05. Found: C, 50.64; H, 8.65; S, 8.91.
added 4-methymorpholine N-oxide (NMO, 58.0 mg,
˚
.492 mmol) and powdered (4 A) molecular sieves, fol-
0
ꢁ
Acetaldehyde (+)-10. To a cold (ꢀ78 C) solution of diol
lowed by tetrapropylammonium perruthenate (TPAP,
4.30 mg, 0.0123 mmol). After stirring for 15 min, the
reaction mixture was filtered, concentrated and passed
through a short pad of flash silica gel (20% EtOAc/
hexanes) to afford the desired keto sulfone (+)-12 as a
6
(
2
(
+)-8 (0.751 g, 3.54 mmol) in THF (50 mL) was added
,6-lutidine (2.10 mL, 18.0 mmol) followed by TESOTf
1.76 mL, 7.78 mmol). After 10 min, the reaction mixture
was quenched with H O (10 mL), warmed to room tem-
2
2
5
perature, extracted with CH Cl (3ꢂ10 mL), dried
white foam [79 mg, quantitative from (+)-11]: [a] +15
2
2
D
1
(MgSO ), filtered, and concentrated in vacuo. The residue
was filtered through a short plug of silica (20% EtOAc/
(c 1.9, CHCl ); H NMR (CDCl ) d 7.08 (dd, J=15,
4
3
3
10 Hz, 1H), 6.15 (t, J=15 Hz, 1H), 6.17–6.00 (m, 2H),
2.43 (dd, J=11, 8 Hz, 1H), 2.29–2.15 (m, 3H), 1.32 (s,
9H), 1.08 (d, J=6.4 Hz, 3H), 0.63 (s, 3H); C NMR
6
hexanes) to give the desired bis-triethylsilyl ether (+)-9
ꢁ
of DMSO (0.664 mL, 9.35 mmol) in CH Cl (3.00 mL) was
13
(1.51 g, 97%) as a colorless oil. To a cold (ꢀ60 C) solution
(CDCl ) d 211.4, 152.0, 146.6, 124.4, 121.2, 61.6, 58.6,
2
2
3
added oxalyl chloride (0.424 mL, 4.86 mmol) in CH Cl
2
55.6, 49.9, 40.9, 40.1, 38.7, 27.5, 23.9, 23.3, 19.6, 19.1,
12.7; IR (neat) 2955, 2934, 2872, 1708, 1638, 1587, 1461,
1296, 1278, 1111, 1001, 824; HRMS: calcd for
C H O S: 352.2072, found 352.2066.
2
ꢁ
(
22.1 mL). After 2 min, a cold (ꢀ60 C) solution of the
aforementioned bis-triethylsilyl ether (+)-9 (1.65 g,
.74mmol) in CH Cl (3.00 mL) was added via cannula.
3
2
2
20 32
3
ꢁ
quenched with Et N (2.81 mL, 20.2 mmol), and warmed
The resulting mixture was stirred at ꢀ60 C for 1 h,
Diene-26-tert-butyl sulfone analogues 5a and 5b. Race-
mic phosphine oxide (ꢃ)-13 (78.0 mg, 0.134 mmol) was
3
to room temperature. Upon dilution with H O (10 mL),
2
ꢁ
the reaction mixture was extracted with CH Cl
2
dissolved in 1.34 mL of THF and cooled to ꢀ78 C
2
(
purified by flash column chromatography (2–5%
3ꢂ10 mL), dried (MgSO ), filtered, concentrated, and
under argon. To this solution was added 97.0 mL of
4
PhLi (0.134 mmol, 1.38 M in cyclohexane/Et O) drop-
2
EtOAc/hexanes) to give the desired aldehyde (+)-10
(
wise via syringe. The deep orange solution was stirred
for 30 min, at which time a cold solution of C,D-ring
ketone (+)-11 (31.0 mg, 0.0880 mmol) in 1.00 mL of
THF was added dropwise via cannula. The resulting
1.03 g, 85%) which was carried forward directly.
26-tert-Butyl sulfone (+)-11. A solution of lithium tert-
butoxide (1.0 M in THF, 0.491 mL) was added via syr-
ꢁ
solution was stirred in the dark at ꢀ78 C for approxi-
ꢁ
175 mg, 0.493 mmol) in THF (1.0 mL). The mixture
ꢁ
inge to a cold (ꢀ78 C) solution of phosphonate 7
mately 4 h, then slowly warmed to ꢀ40 C over 2 h. The
(
was warmed slightly to effect solution and returned to
reaction mixture was quenched with 3 mL of a 2:1 (v/v)
mixture of 2 N sodium potassium tartrate and 2 N
potassium carbonate. Upon warming to room tempera-
ꢁ
ꢀ
78 C. The resulting yellow solution was delivered via
cannula to a stirred solution of aldehyde (+)-10
64.0 mg, 0.197 mmol) in THF (1.5 mL) at rt. After
0 min, the solvent was removed and the residual brown
oil was flash chromatographed (8% EtOAc/hexanes) to
give 86 mg (93%) of (+)-11 as a moist solid: [a] +78
c 4.0, CHCl ); H NMR (CDCl ) d 7.16–7.06 (m, 1H),
ture, the reaction mixture was diluted with H O,
2
(
1
extracted with EtOAc (4ꢂ20 mL), dried over MgSO ,
4
filtered, concentrated and purified by silica gel column
chromatography (10–50% EtOAc/hexanes/<0.1%
2
5
ꢁ
Et N) to afford the coupled product [33 mg, 90% based
3
on recovered (+)-11] as a yellow oil.
D
1
(
3
3
6
1
.14 (t, J=15 Hz, 1H), 6.11–6.06 (m, 2H), 4.05–4.00 (m,
H), 2.27–2.17 (m, 1H), 1.96–1.89 (m, 1H), 1.89–1.74
This oil was immediately dissolved in ethanol (1.50 mL),
ꢁ
(
3
qt, J=13, 3.8 Hz, 1H), 1.35 (s, 9H), 1.04 (d, J=6.8 Hz,
H), 0.93 (s, 3H), 0.93 (t, J=8.0 Hz, 9H), 0.54 (q,
cooled to 0 C, and treated with HF (0.100 mL, 49%
aqueous). The solution was slowly warmed to room
temperature and treated with additional HF (0.100 mL)
to complete the deprotection. After aqueous workup the
resulting white film was flash chromatographed (1%
1
3
J=8.0 Hz, 6H); C NMR (CDCl ) d 153.7, 147.2,
3
1
2
2
23.7, 120.4, 69.2, 58.6, 55.8, 52.9, 42.4, 40.6, 40.1, 34.5,
7.4, 23.3, 22.9, 19.4, 17.6, 13.8, 6.9, 4.9; IR (neat)
950, 2935, 2873, 1638, 1458, 1300, 1113, 1018, 1004;
Et N/EtOAc) to afford 17.3 mg (77% for deprotection)
3

G. H. Posner et al. / Bioorg. Med. Chem. 9 (2001) 2365–2371
2371
of diastereomers 5a and 5b. This diastereomeric mixture
was purified by reversed-phase HPLC (C-18 semi-
wells were treated with ethanol vehicle (EtOH), calci-
triol (1), or vitamin D analogues 4 or 5a at 1-, 10-, or
3
preparative column, 46% MeCN/H O, 3 mL/min) giv-
2
100-nM concentration for 22 h. Luciferase activity was
analyzed with the Dual Luciferase Assay Kit (Pro-
mega) on a Victor 2 microplate reader. Luciferase
activity of the pGL3/24-OHASE was corrected for
transfection effeciency using the luciferase values
obtained with the co-transfected control vector (pRL
SV40). Data were statistically evaluated by ANOVA
followed by Dunnetts post-hoc test using GraphPad
Instat software.
ing 8.8 mg of 5a (20%, tR 115 min) and 2.1 mg of 5b
2
5
(
5%, t 111 min). 5a (1a, 3b): [a] +97 (c 7.5, CHCl3);
R
D
1
H NMR (CDCl ) d 7.18–7.06 (m, 1H), 6.37 (d,
3
J=11 Hz, 1H), 6.16 (t, J=15 Hz, 1H), 6.13–6.05 (m,
2
4
H), 6.01 (d, J=11 Hz, 1H), 5.35–5.29 (m, 1H), 5.01–
.95 (m, 1H), 4.46–4.40 (m, 1H), 4.26–4.19 (m, 1H), 2.83
(
1
dd, J=12, 3.7 Hz, 1H), 2.59 (dd, J=14, 3.2 Hz, 1H),
.36 (s, 9H), 1.09 (d, J=6.8 Hz, 3H), 0.57 (s, 3H); 13
NMR (CDCl ) d 153.1, 147.6, 147.0, 142.4, 133.2, 124.8,
C
3
1
4
1
1
2
24.0, 120.7, 117.3, 111.8, 70.8, 66.8, 58.6, 56.1, 55.6,
6.1, 45.2, 42.8, 40.7, 40.3, 29.0, 27.5, 23.5, 23.4, 22.2,
9.7, 12.3; IR (neat) 3598–3148 (br), 3013, 2935, 2870,
637, 1588, 1457, 1295, 1108, 756; UV (MeOH) lmax
87 nm (E 7640); HRMS: calcd for C H O S 488.2960,
Acknowledgements
The authors thank the NIH (CA 44530 to G.H.P. and
T.W.K., CA 69700 to J.W., and DK 50583 to S.P.) and
the American Institute for Cancer Research (99A091 to
D.A.G.) for support of this research and H. Brem and
S. Zand for the malignant melanoma assay.
2
9
44
4
2
5
1
found 488.2950. 5b (1ß, 3a): [a]D 32 (c 1.0, CHCl ); H
NMR (CDCl ) d 7.18–7.07 (m, 1H), 6.38 (d, J=11 Hz,
1H), 6.16 (t, J=15 Hz, 1H), 6.15–6.06 (m, 2H), 6.01 (d,
J=11 Hz, 1H), 5.33–5.29 (m, 1H), 5.01–4.96 (m, 1H),
3
3
4
3
1
.47–4.40 (m, 1H), 4.26–4.18 (m, 1H), 2.84 (dd, J=12,
.8 Hz, 1H), 2.61 (dd, J=14, 3.2 Hz, 1H), 1.36 (s, 9H),
.09 (d, J=6.8 Hz, 3H), 0.57 (s, 3H); C NMR
References and Notes
1
3
(
1
4
1
1
2
CDCl ) d 153.1, 147.3, 147.0, 142.5, 133.1, 124.8, 124.0,
3
1
. Holick, M. F. Vitamin D: Molecular Biology, Physiology,
and Clinical Applications; Humana: Totowa, 1999.
. Feldman, D.; Glorieux, F.; Pike, J. W.; Vitamin D; Aca-
20.8, 117.3, 112.6, 71.3, 66.8, 58.6, 56.1, 55.6, 46.1,
5.4, 42.8, 40.7, 40.3, 29.0, 27.4, 23.45, 23.38, 22.3, 19.7,
2.3; IR (neat) 3568–3087 (br), 3016, 2954, 2928, 2872,
637, 1588, 1457, 1295, 1109, 755; UV (MeOH) lmax
86 nm (E 4442); HRMS: calcd for C H O S 488.2960,
2
demic: San Diego, 1997.
3. Kutner, A., Ed., Curr. Pharm. Des. 2000, 6, 701.
4. Bouillon, R.; Okamura, W. H.; Norman, A. W. Endocrine
Rev. 1995, 16, 200.
2
9
44
4
found 488.2944.
5
. Posner, G. H.; Wang, Q.; Han, G.; Lee, J. K.; Crawford,
K.; Zand, S.; Brem, H.; Peleg, S.; Dolan, P.; Kensler, T. W. J.
Med. Chem. 1999, 42, 3425.
MTT assay
6
. Posner, G. H.; Crawford, K.; Siu-Caldera, M.-L.; Reddy,
The influence of the vitamin D analogues on the anti-
3
G. S.; Sarabia, S. F.; Feldman, D.; van Etten, E.; Mathieu, C.;
Gennaro, L.; Vouros, P.; Peleg, S.; Dolan, P.; Kensler, T. W.
J. Med. Chem. 2000, 43, 3581.
7. Hansen, C. M.; Hamberg, K. J.; Binderup, E.; Binderup, L.
Curr. Pharm. Des 2000, 6, 803.
proliferative activity of adriamycin in MCF-7 cells was
determined using the MTT tetrazolium dye assay as
1
3
described previously. Cells were seeded in 96-well
ꢁ
plates at 37 C for 24 h, treated with the vitamin D3
analogues for 48 h and washed twice with 200 mL of PBS
before exposure to various concentrations of adriamy-
cin for 2 h. Adriamycin was removed and cells were
washed prior to an additional incubation period of 72 h.
MTT (2 mg/mL in PBS) was added, cells were incubated
8. Diaz, G. D.; Paraskeva, C.; Thomas, M. G.; Binderup, L.;
Hague, A. Cancer Res. 2000, 60, 2304.
9
4
1
. Bhattacharya, A.; Thyagarajan, G. Chem. Rev. 1981, 81,
15.
0. Tolstikov, G. A.; Miftakhov, M. S.; Adler, M. E.;
Komissarova, N. G.; Kuznetsov, O. M.; Vostrikov, N. S.
Synthesis 1989, 940.
11. Corey, E. J.; Roberts, B. E.; Dixon, B. R. J. Am. Chem.
Soc. 1995, 117, 193.
12. Posner, G. H.; Nelson, T. D.; Guyton, K. Z.; Kensler,
T. W. J. Med. Chem. 1992, 35, 3280.
ꢁ
at 37 C for 3 h, MTT was removed, 100 mL of DMSO
added to dissolve the crystallized dye and absorbance
was determined at 490 nm. Each condition involved 6–8
replicate samples.
1
3. Munger, C.; Ellis, A.; Woods, K.; Randolph, J.; Yano-
vich, S.; Gewirtz, L. Cancer Res. 1988, 48, 2404.
4. Chaudry, M.; Sundaram, S.; Gewirtz, D. Cancer Che-
MCF-7 human breast cancer cells
1
These were plated in phenol red free Hams F12 media
containing 5% charcoal stripped serum at a density of
mother. Pharmacol. 2001, 47, 429.
15. Sundaram, S.; Chaudry, M.; Reardon, D.; Gupta, M.;
Gewirtz, D. Breast Cancer Res. Treat 2001, 63, 1.
5
2
(
.5ꢂ10 cells per well in six well plates. After plating
24 h), cells were co-transfected using Fugene (Gibco)
1
6. Posner, G. H.; Lee, J. K.; Wang, Q.; Peleg, S.; Burke, M.;
Brem, H.; Dolan, D.; Kensler, T. W. J. Med. Chem. 1998, 41,
008.
with 0.75 mg of pGL3/24-OHASE, a vitamin D3
responsive 24-hydroxylase luciferase reporter construct
(
3
generously provided by Dr. J. Omdahl, University of
17. Kerry, D. M.; Dwivedi, P. P.; Hah, C. N.; Morris,
H. A.; Omdahl, J. L.; May, B. K. J. Biol. Chem. 1996, 271,
29715.
New Mexico, Albuquerque, USA) and 0.25 mg of pRL
SV40, a control luciferase vector (Promega). Triplicate