Novel Silicon-Containing Analogues of the Retinoid Agonist Bexarotene: Syntheses and Biological Effects on Human Pluripotent Stem Cells

Article abstract of DOI:10.1002/cmdc.201100156

Twofold sila-substitution (C/Si exchange) of the clinically used RXR-selective retinoid agonist bexarotene leads to disila-bexarotene, which displays pharmacological potency similar to that of the parent carbon compound, as shown in a HeLa-cell-based RXR assay. Formal exchange of the SiCH₂CH₂Si group in disila-bexarotene with a SiCH₂Si or SiOSi moiety leads to the disila-bexarotene analogues 8 and 9. The silicon compounds 8 and 9 were synthesized in multistep syntheses, starting from HC≡C(CH₃)₂SiCH₂Si(CH₃)₂C≡CH and HC≡C(CH₃)₂SiOSi(CH₃)₂C≡CH, respectively. The key step in the syntheses of 8 and 9 is a cobalt-catalyzed [2+2+2] cycloaddition reaction that affords the 1,3-disilaindane and 2-oxa-1,3-disilaindane skeletons. Disila-bexarotene and its analogues 8 and 9 were studied for their biological effects relative to all-trans retinoic acid in cultured human pluripotent stem cells. The parent carbon compound bexarotene was included in some of these biological studies. Although the silicon-containing bexarotene analogues disila-bexarotene, 8, and 9 appear not to regulate the differentiation of TERA2.cl.SP12 stem cells, preliminary evidence indicates that these compounds may possess enhanced functions over the parent compound bexarotene, such as induction and regulation of cell death and cell numbers. The biological data obtained indicate that bexarotene, contrary to the silicon-containing analogues disila-bexarotene, 8, and 9, may partially act to induce cell differentiation.

Full text of DOI:10.1002/cmdc.201100156

DOI: 10.1002/cmdc.201100156
Novel Silicon-Containing Analogues of the Retinoid
Agonist Bexarotene: Syntheses and Biological Effects on
Human Pluripotent Stem Cells
Jennifer B. Bauer,[a] W. Peter Lippert,[a] Steffen Dçrrich,[a] David Tebbe,[a] Christian Burschka,[a]
Victoria B. Christie,[b, c] Daniel M. Tams,[b, c] Andrew P. Henderson,[c, d] Bridgid A. Murray,[b, c]
Todd B. Marder,[d] Stefan A. Przyborski,[b, c] and Reinhold Tacke*[a]
Twofold sila-substitution (C/Si exchange) of the clinically used
RXR-selective retinoid agonist bexarotene leads to disila-bexar-
otene, which displays pharmacological potency similar to that
of the parent carbon compound, as shown in a HeLa-cell-
based RXR assay. Formal exchange of the SiCH2CH2Si group in
disila-bexarotene with a SiCH2Si or SiOSi moiety leads to the
disila-bexarotene analogues 8 and 9. The silicon compounds 8
and 9 were synthesized in multistep syntheses, starting from
biological effects relative to all-trans retinoic acid in cultured
human pluripotent stem cells. The parent carbon compound
bexarotene was included in some of these biological studies.
Although the silicon-containing bexarotene analogues disila-
bexarotene, 8, and 9 appear not to regulate the differentiation
of TERA2.cl.SP12 stem cells, preliminary evidence indicates that
these compounds may possess enhanced functions over the
parent compound bexarotene, such as induction and regula-
HCꢀC(CH3)2SiCH2Si(CH3)2CꢀCH and HCꢀC(CH3)2SiOSi(CH3)2Cꢀ tion of cell death and cell numbers. The biological data ob-
CH, respectively. The key step in the syntheses of 8 and 9 is a
cobalt-catalyzed [2+2+2] cycloaddition reaction that affords
the 1,3-disilaindane and 2-oxa-1,3-disilaindane skeletons. Disila-
bexarotene and its analogues 8 and 9 were studied for their
tained indicate that bexarotene, contrary to the silicon-contain-
ing analogues disila-bexarotene, 8, and 9, may partially act to
induce cell differentiation.
Introduction
There is increasing pressure for the development of innovative
new drugs. In this context, silicon chemistry is being used as a
novel source of chemical diversity in drug design (see refer-
ence [1] for reviews). The group 14 elements carbon and silicon
show many similarities, one being that they readily form four
covalent bonds with themselves and many other elements.
However, there are also some striking differences (for example,
different covalent radii and electronegativities), which can be
exploited in medicinal chemistry for the design and synthesis
of innovative new drugs. Sila-substitution (C/Si exchange) of
well-known drugs is one of the strategies that have been suc-
cessfully used for the development of new silicon-based drugs
(see references [2] and [3] for recent examples). In this context,
we have reported the synthesis and pharmacological charac-
terization of a series of disila-analogues of retinoids,[3] such as
disila-bexarotene (1b), disila-TTNPB (2b), disila-3-methyl-TTNPB
(3b), disila-SR11237 (4b), the disila-SR11237 derivative 5b,
disila-AM80 (6b), and disila-AM580 (7b). All of these silicon
compounds were found to be at least equipotent with the cor-
responding carbon-based drug 1a–7a, and in some cases the
twofold carbon/silicon exchange even led to a significant in-
crease in pharmacological potency. For example, twofold sila-
substitution of the RXR-selective retinoid agonist 5a (!5b) re-
sulted in a tenfold increase in biological activity. Notably, at
low concentrations, the silicon-containing rexinoid 5b is even
more potent than 9-cis-retinoic acid (9-cis-RA), which is often
considered to be the natural RXR ligand. In our ongoing
search for highly potent and receptor-selective retinoids, we
have now succeeded in synthesizing the disila-bexarotene ana-
logues 8 and 9 that contain SiCH2Si and SiOSi fragments in-
stead of the SiCH2CH2Si group of 1b. Herein we report the syn-
theses of compounds 8 and 9 and the biological effects of 1b,
8, and 9 relative to all-trans retinoic acid (ATRA) in cultured
human pluripotent stem cells. The parent carbon compound
[a] Dr. J. B. Bauer, Dr. W. P. Lippert, S. Dçrrich, Dr. D. Tebbe, Dr. C. Burschka,
Prof. R. Tacke
Universitꢀt Wꢁrzburg, Institut fꢁr Anorganische Chemie
Am Hubland, 97074 Wꢁrzburg (Germany)
Fax: (+49)931-31-84609
E-mail: r.tacke@uni-wuerzburg.de
[b] Dr. V. B. Christie, D. M. Tams, B. A. Murray, Prof. S. A. Przyborski
Biological and Biomedical Sciences, Durham University
South Road, Durham, DH1 3LE (UK)
[c] Dr. V. B. Christie, D. M. Tams, A. P. Henderson, B. A. Murray,
Prof. S. A. Przyborski
Reinnervate Limited, NETPark Incubator
Thomas Wright Way, Sedgefield, TS21 3FD (UK)
[d] A. P. Henderson, Prof. T. B. Marder
Department of Chemistry, Durham University
South Road, Durham, DH1 3LE (UK)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201100156.
ChemMedChem 2011, 6, 1509 – 1517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1509

R. Tacke et al.
MED
A2.cl.SP12 cells have a propensity for forming neural deriva-
tives in response to both natural and synthetic retinoids.[6–8] As
a consequence, TERA2.cl.SP12 cells upregulate neural markers
in response to these compounds. It is currently not known
whether the silicon-containing bexarotene analogues 1b, 8,
and 9 regulate cell differentiation. In this study, we have as-
sessed the ability of compounds 1b, 8, and 9 to induce the dif-
ferentiation of pluripotent EC stem cells in comparison to
ATRA. We report that although these particular bexarotene an-
alogues do not appear to regulate the differentiation of TER-
A2.cl.SP12 stem cells, preliminary evidence indicates that these
compounds regulate cell death.
Results and Discussion
Syntheses of the bexarotene analogues 8 and 9
The title compounds 8 and 9 were synthesized according to
Scheme 1. The crucial step in these syntheses was the con-
struction of the 1,3-disilaindane and 2-oxa-1,3-disilaindane
skeleton by using a [2+2+2] cycloaddition, which is catalyzed
by a catalytic system consisting of cobalt(II) iodide and zinc
powder in acetonitrile.[9] Thus, compound 8 was prepared in a
multistep synthesis, starting from the silicon-containing diyne
bis(ethynyldimethylsily)methane (10), which was treated with
the monoyne 4,4,5,5-tetramethyl-2-prop-1-ynyl-1,3,2-dioxabor-
olane (11) to give the cycloaddition product 4,4,5,5-tetrameth-
yl-2-(1,1,3,3,6-pentamethyl-1,3-disilaindan-5-yl)-1,3,2-dioxaboro-
lane (12) in 46% yield. Compound 12 was then treated with
methyl 4-[1-(trifluoromethylsulfonyloxy)vinyl]benzoate (13), in
the presence of catalytic amounts of (Ph3P)2PdCl2, to yield
methyl
4-[1-(1,1,3,3,6-pentamethyl-1,3-disilaindan-5-yl)vinyl]-
benzoate (14) in 75% yield. Treatment of 14 with potassium
hydroxide in methanol/water and subsequent acidification
with hydrochloric acid afforded the target compound 4-[1-
(1,1,3,3,6-pentamethyl-1,3-disilaindan-5-yl)vinyl]benzoic acid (8)
in 90% yield. The synthesis of the second target compound, 4-
[1-(1,1,3,3,6-pentamethyl-2-oxa-1,3-disilaindan-5-yl)vinyl]benzo-
ic acid (9), was performed analogously to that of 8, starting
from the diyne 1,3-diethynyl-1,1,3,3-tetramethyldisiloxane (15)
and the monoyne 11 (11/15!4,4,5,5-tetramethyl-2-(1,1,3,3,6-
bexarotene (1a) was included in some of these biological stud-
ies.
Cultured human embryonic carcinoma (EC) stem cells are
widely recognized and used as models to study the early
stages of embryogenesis in humans.[4] They are also considered
to be the malignant counter-
parts of pluripotent embryonic
stem cells and serve as an ap-
propriate surrogate and useful
screening tool to test regulators
of cell differentiation. The
human EC stem cell line TER-
A2.cl.SP12 is a clonal lineage de-
rived from the well-known
TERA2 cell population.[5] TER-
A2.cl.SP12 stem cells express
markers characteristic of pluripo-
tent stem cells and respond to
exogenous reagents that are
known to induce stem cell differ-
entiation. In particular, TER- Scheme 1. Syntheses of compounds 8 and 9.
1510
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1509 – 1517

Disila-Bexarotene Analogues
pentamethyl-2-oxa-1,3-disilaindan-5-yl)-1,3,2-dioxaborolane
(16) (31%)!methyl 4-[1-(1,1,3,3,6-pentamethyl-2-oxa-1,3-disi-
laindan-5-yl)vinyl]benzoate (17) (75%)!9 (93%)).
Compounds 8, 9, 12, 14, and 16–19 were isolated as crystal-
line solids, and their identities were established by elemental
analyses (C, H) and NMR spectroscopic studies (1H, 11B, 13C, 29Si).
Compounds 8, 12, 14, 16, and 17 were additionally character-
ized by crystal structure analyses.
The key step in the syntheses of 8 and 9 is the cobalt-cata-
lyzed [2+2+2] cycloaddition reaction to afford the 1,3-disi-
laindane and 2-oxa-1,3-disilaindane skeletons. The yields of the
respective intermediates 12 (46%) and 16 (31%) are rather
poor and can be explained by the formation of by-products
that derive from a [2+2+2] cycloaddition of the precursors
10 and 15, respectively, in which 10 and 15 act both as a
diyne and as a monoyne. To prove this hypothesis, the corre-
sponding products of these side reactions, compounds 18 and
19, were prepared directly from 10 and 15, respectively, in a
targeted synthesis according to Scheme 2 (yields: 18, 47%; 19,
39%).
Crystal structure analyses of 8, 12, 14, 16, and 17
Compounds 8, 12, 14, 16, and 17 were studied by single-crys-
tal X-ray diffraction. The crystal data and the experimental pa-
rameters used for these studies are given in Table 1. The
molecular structures of 8 and 17 are depicted in Figure 1 and
Figure 2, respectively; selected bond lengths, bond angles, and
torsion angles of these compounds are given in the figure leg-
ends. For further information on the crystal structure analyses,
see the Supporting Information. The bond lengths and angles
of 8, 12, 14, 16, and 17 lie within the typical ranges expected
for the molecular skeletons in question and thus are not dis-
cussed herein. Inspection of the relevant torsion angles shows
that the five-membered rings of these compounds are nonpla-
nar, the minimum deviation from planarity being observed for
12, 16, and 17, whereas 8 and 14 display a typical envelope
conformation. The different torsion angles of the 1,3-disila-
Scheme 2. Syntheses of compounds 18 and 19.
Table 1. Crystallographic data and experimental parameters for the crystal structure analyses of 8, 12, 14, 16, and 17.
8
12
14
16
17
Formula
Mr [Da]
T [K]
C21H26O2Si2
366.60
283(2)
C18H31BO2Si2
346.42
193(2)
C22H28O2Si2
380.62
173(2)
C17H29BO3Si2
348.39
98(2)
C21H26O3Si2
382.60
173(2)
l(Mo Ka) [ꢁ]
Crystal system
Space group (No.)
a [ꢁ]
b [ꢁ]
c [ꢁ]
0.71073
monoclinic
P21/c (14)
5.8745(5)
8.9747(5)
40.113(3)
90.102(10)
2114.8(3)
4
0.71073
monoclinic
Cc (9)
16.808(3)
11.646(2)
12.675(3)
119.71(3)
2154.8(7)
4
0.71073
orthorhombic
Pccn (56)
23.855(2)
30.6679(18)
5.8944(3)
90
4312.3(5)
8
1.173
0.71073
orthorhombic
Pbca (61)
12.0327(10)
12.4760(12)
54.505(5)
90
8182.3(13)
16
1.131
0.71073
orthorhombic
Pccn (56)
23.394(3)
30.321(5)
6.1857(8)
90
4387.8(10)
8
1.158
b [8]
V [ꢁ3]
Z
1calcd [gcmꢁ3
]
1.151
0.178
784
1.068
0.170
752
m [mmꢁ1
F(000)
]
0.177
1632
0.183
3008
0.178
1632
Crystal dimensions [mm]
2q range [8]
Index ranges
0.5ꢂ0.3ꢂ0.2
4.06–49.86
ꢁ6ꢂhꢂ6,
ꢁ10ꢂkꢂ10,
ꢁ47ꢂlꢂ47
17808
0.5ꢂ0.3ꢂ0.3
7.92–58.10
ꢁ22ꢂhꢂ22,
ꢁ15ꢂkꢂ15,
ꢁ17ꢂlꢂ17
14920
0.5ꢂ0.3ꢂ0.15
3.42–48.00
ꢁ27ꢂhꢂ27,
ꢁ35ꢂkꢂ35,
ꢁ6ꢂlꢂ6
34815
0.18ꢂ0.17ꢂ0.15
2.98–61.08
ꢁ16ꢂhꢂ17,
ꢁ17ꢂkꢂ17,
ꢁ77ꢂlꢂ77
196736
12514
0.4ꢂ0.4ꢂ0.2
4.40–56.02
ꢁ29ꢂhꢂ29,
ꢁ39ꢂkꢂ39,
ꢁ8ꢂlꢂ8
60196
Collected reflections
Independent reflections
Rint
3669
0.0429
5629
0.0447
3367
0.0342
5164
0.0817
0.0793
Reflections used
Restraints
3669
0
5629
6
3367
0
12514
0
5164
0
Parameters
235
1.048
0.0542/0.1247
0.0334
217
1.061
0.0773/0.9465
0.0468
241
1.057
0.0535/0.9287
0.0313
433
1.167
0.0465/6.6147
0.0494
241
1.043
0.0750/0.4875
0.0453
S[a]
Weight parameters a/b[b]
R1[c] [I>2s(I)]
wR2[d] (all data)
Abs. structure parameter
Max./min. residual
0.0905
–
+0.135/ꢁ0.142
0.1264
0.0868
–
+0.232/ꢁ0.206
0.1290
–
+0.573/ꢁ0.529
0.1328
–
+0.346/ꢁ0.417
0.00(10)
+0.373/ꢁ0.316
electron density [eꢁꢁ3
]
2
[a] S={S[w(Fo ꢁFc2)2]/(nꢁp)}0.5; n=number of reflections; p=number of parameters. [b] wꢁ1 =s2(Fo2)+(aP)2 +bP, with P=[max(Fo2,0)+2Fc2]/3. [c] R1 =
2
SjjFo jꢁjFc jj/SjFo j. [d] wR2 ={S[w(Fo ꢁFc2)2]/S[w(Fo2)2]}0.5
.
ChemMedChem 2011, 6, 1509 – 1517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1511

R. Tacke et al.
MED
indanes 12 (almost planar) and
8/14 (envelope conformation)
indicate a high degree of con-
formational flexibility of the 1,3-
disilaindane skeleton.
Biological assessment of the
bexarotene analogues 1b, 8,
and 9 on cell growth and dif-
ferentiation
To compare the biological ef-
fects of disila-bexarotene (1b)
and its analogues 8 and 9 on
the growth and differentiation
of EC stem cells, cultures of TER-
A2.cl.SP12 EC cells were incubat-
ed with 1 mm and 10 mm con-
centrations of each compound
Figure 3. Phase micrographs of TERA2.cl.SP12 EC stem cells treated with ATRA and test compounds disila-bexaro-
tene (1b), 8, and 9. Images show examples of A) untreated, undifferentiated control cells (UNDIFF) at three days,
B) cultures exposed to 1 mm 9 for seven days, and cells treated with C) ATRA, D) 9, E) 8, and F) 1b at 10 mm for
seven days. Under all conditions, cells continued to proliferate at different rates, but in each case lower than that
of the untreated control. Some cultures became over-confluent, and areas of overgrowth are indicated (*). The
morphology of cells treated with ATRA became heterogeneous after seven days, indicating induction of cell differ-
entiation (C). Scale bar: 100 mm.
tested and maintained for up to seven days. Experiments with
ATRA were performed alongside as a positive control of in-
duced cell differentiation.
Morphological analysis of phase micrograph images was
used as an initial indicator of changes to the cultures
(Figure 3). Cultures treated with ATRA contained cells with dif-
fering morphologies suggesting induction of cell differentia-
tion. In comparison, cells treated with compounds 1b, 8, and 9
did not respond in the same manner and generally maintained
a similar appearance to the untreated, undifferentiated con-
trols. However, cell numbers were influenced by the different
treatments tested (Figure 4). As expected, cultures of untreated
Figure 1. Molecular structure of 8 in the crystal. Selected bond lengths [ꢁ],
bond angles [8], and torsion angles [8]: Si1ꢁC3 1.873(2), Si2ꢁC3 1.869(3),
Si1ꢁC8 1.883(2), Si2ꢁC7 1.8827(19), C7ꢁC8 1.412(3); Si1ꢁC3ꢁSi2 105.13(12),
Si1ꢁC8ꢁC7 113.74(14), Si2ꢁC7ꢁC8 114.94(16), C3ꢁSi1ꢁC8 99.78(11), C3ꢁSi2ꢁ
C7 100.11(10); C7ꢁC8ꢁSi1ꢁC3 20.21(19), C8ꢁSi1ꢁC3ꢁSi2 ꢁ24.25(16), Si1ꢁ
C3ꢁSi2ꢁC7 21.19(16), C3ꢁSi2ꢁC7ꢁC8 ꢁ9.6(2), Si2ꢁC7ꢁC8ꢁSi1 ꢁ6.9(2).
Figure 4. Analysis of cell images shows a decrease in cell numbers in cul-
tures treated with ATRA and test compounds disila-bexarotene (1b), 8, and
9 relative to control. Quantification of the number of cells per field of view
was performed on control untreated, undifferentiated (UNDIFF) cultures at
three days (&) and treated cultures at three (&) and seven (&) days. Due
to complete overgrowth, it was not possible to acquire data for seven-day
control cultures. After three days, all cultures treated with either ATRA, 1b,
8, or 9 showed a decrease in cell numbers relative to the untreated control,
with the greatest effect observed following treatment with ATRA. Similarly,
cultures treated with ATRA for seven days contained fewer cells than those
treated with the bexarotene analogues. Data shown represent mean ꢃSD,
n=4.
Figure 2. Molecular structure of 17 in the crystal. Selected bond lengths [ꢁ],
bond angles [8], and torsion angles [8]: Si1ꢁO3 1.6467(16), Si2ꢁO3
1.6563(18), Si1ꢁC8 1.8879(17), Si2ꢁC7 1.8871(19), C7ꢁC8 1.413(3); Si1ꢁO3ꢁ
Si2 117.61(9), Si1ꢁC8ꢁC7 111.76(12), Si2ꢁC7ꢁC8 112.16(13), O3ꢁSi1ꢁC8
99.29(8), O3ꢁSi2ꢁC7 98.96(8); C7ꢁC8ꢁSi1ꢁO3 4.33(14), C8ꢁSi1ꢁO3ꢁSi2
ꢁ4.73(12), Si1-O3-Si2-C7 3.54(13), O3ꢁSi2ꢁC7ꢁC8 ꢁ0.28(14), Si2-ꢁC7ꢁC8ꢁSi1
ꢁ2.52(16).
1512
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1509 – 1517

Disila-Bexarotene Analogues
undifferentiated stem cells continued to proliferate and
became over-confluent within 4–5 days of culture. Cultures
treated with 10 mm of either compound 1b, 8, or 9 did not
overgrow as extensively as the untreated control and were
maintained for the full seven days. As a consequence, cell
numbers in these cultures were decreased relative to the
untreated control, but cell numbers were greater than in the
equivalent ATRA-treated cultures. Unlike cultures of cells ex-
posed to 1 mm ATRA, cells treated with the lower concentration
of either compound 1b, 8, or 9 (1 mm) became confluent and
displayed evidence of significant overgrowth. Compound activ-
ity at the 1 mm concentration was not investigated further.
Flow cytometric analysis was performed to measure the
expression of protein markers characteristic of stem cells or dif-
ferentiated derivatives (Figure 5). Two markers indicative of the
stem cell phenotype were analyzed, including the globo-series
antigen stage-specific embryonic antigen-3 (SSEA-3) and the
keratin-sulfate-associated glycoprotein antigen surface marker
TRA-1-60. Typically, these markers are expressed at high levels
in cultures of TERA2.cl.SP12 stem cells but not their differenti-
ated derivatives.[5] After treatment for seven days with 10 mm
ATRA, both stem markers were almost completely deactivated,
whereas both markers maintained a good level of expression
in cells treated with bexarotene (1a) and its analogues 1b, 8,
and 9. Although the expression of SSEA-3 was partially de-
creased relative to control untreated cells, levels of TRA-1-60
did remain strong in 61–92% of cells tested. Treatment by
ATRA induced high levels of A2B5 expression (a ganglio-series
antigen associated with early-stage neurogenesis) indicating
induction of cell differentiation. Cells did partially respond to
10 mm bexarotene (1a) and showed a modest increase in A2B5
expression but this was almost threefold less than the A2B5
response induced by ATRA. No similar pattern was observed in
cultures exposed to the bexarotene analogues 1b, 8, and 9
where levels of A2B5 remained low. Decreased expression of
SSEA-3 in response to 1b, 8, and 9 is unlikely to be the result
of induced differentiation as TRA-1-60 and A2B5 levels showed
minimal changes compared to those initiated by ATRA and be-
cause of the absence of morphological changes typically asso-
ciated with the differentiation response. Such a decrease in
SSEA-3 expression is more likely to be due to suboptimal
growth conditions as the culture became over-confluent.
Measurement of Annexin V was performed by flow cytome-
try to assess induction of apoptosis in cells treated with test
compounds (Figure 6A). In each growth condition, the per-
centage of cells positive for Annexin V was low, with only com-
pound 9 (containing a polar SiOSi group) showing any sub-
stantial increase in expression above the undifferentiated con-
trol. In contrast, the percentage of cells that stained positive
for 7-AAD was significantly higher in cultures treated with the
silicon-containing bexarotene analogues 1b, 8, and 9 than in
undifferentiated populations or for those cells exposed to
ATRA, although compound 9 was less responsive (Figure 6B).
Cultures treated with bexarotene (1a) also showed higher
numbers of 7-AAD-positive cells than the undifferentiated con-
trol, but the increase was less pronounced than that observed
for the silicon analogue disila-bexarotene (1b). Similar trends
were noted with cultures stained for propidium iodide (PI) and
the proportion of live/dead cells (Figure 6C), again indicating
differences between the C/Si analogues bexarotene (1a) and
disila-bexarotene (1b).
Collectively, the data obtained in this study suggest that the
silicon-containing bexarotene analogues 1b, 8, and 9 are
unable to induce robust differentiation of human pluripotent
stem cells as observed for ATRA, which acted as a positive con-
trol. However, the bexarotene analogues did decrease the
number of cells present in each culture compared to the un-
treated control when used at a concentration of 10 mm. It
appears that these compounds are moderately toxic at this
concentration, which could explain the decreased cell number
and the presence of significant proportions of dying and dead
cells. The ability to lower cell numbers is a feature of any
promising anticancer agent. Indeed, the parent compound
bexarotene (1a) is already routinely used as part of treatment
regimes for certain types of cancer, such as cutaneous T-cell
lymphoma.[10] Bexarotene is a known selective RXR agonist and
is understood to exert its effects in blocking cell cycle progres-
sion and inducing apoptosis and differentiation.[11] Consistent
Figure 5. Protein expression analysis reveals differential effects of A) the reti-
noids ATRA, disila-bexarotene (1b), 8, and 9 for 3 or 7 days as indicated,
and B) bexarotene (1a) for 7 days, at 10 mm on pluripotent stem cells. Flow
cytometry data show the percentage of positive cells expressing markers of
stem cells (SSEA-3 and TRA-1-60) and a protein marker indicative of cell dif-
ferentiation (A2B5). Treatment by ATRA shuts down expression of the stem
cell markers almost completely and induces high levels of A2B5 expression.
Conversely, treatment with bexarotene (1a) or with either of its silicon-con-
taining analogues (1b, 8, 9) showed a milder decrease in SSEA-3 expression
and had even less dramatic impact on the expression levels of TRA-1-60 and
A2B5. Data represent the mean ꢃSD, n=3.
ChemMedChem 2011, 6, 1509 – 1517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1513

R. Tacke et al.
MED
Experimental Section
Chemistry
General procedures: All syntheses were carried out under dry ni-
trogen or argon. The organic solvents used were dried and purified
according to standard procedures and stored under dry nitrogen.
A Bꢃchi GKR-50 apparatus was used for the bulb-to-bulb distilla-
tions. Melting points were determined with a Bꢃchi Melting Point
1
B-540 apparatus by using samples in sealed capillaries. The H, 11B,
13C, and 29Si NMR spectra were recorded at 238C on a Bruker
Avance 500 (1H, 500.1 MHz; 11B, 160.5 MHz; 13C, 125.8 MHz; 29Si,
99.4 MHz) or a Bruker DRX-300 NMR spectrometer (1H, 300.1 MHz;
13C, 75.5 MHz; 29Si, 59.6 MHz) using C6D6 or CD2Cl2 as the solvent.
Chemical shifts (d, ppm) were determined relative to internal
C6HD5 (1H, d=7.28 ppm; C6D6), internal C6D6 (13C, d=128.0 ppm;
C6D6), internal CHDCl2 (1H, d=5.32 ppm; CD2Cl2), internal CD2Cl2
(13C, d=53.8 ppm; CD2Cl2), external BF3·OEt2 (11B, d=0 ppm; C6D6),
or external (CH3)4Si (29Si, d=0 ppm; C6D6, CD2Cl2). Assignment of
1
1
the H NMR data was supported by H,1H gradient-selected COSY,
13C,1H gradient-selected HMQC and gradient-selected HMBC, and
29Si,1H gradient-selected HMQC experiments (optimized for JSi,H
=
2
7 Hz). Assignment of the 13C NMR data was supported by DEPT 135
and the aforementioned 13C, H correlation experiments.
1
4-[1-(1,1,3,3,6-Pentamethyl-1,3-disilaindan-5-yl)vinyl]benzoic
acid (8): A mixture of MeOH (13 mL), H2O (4 mL), KOH (583 mg,
10.4 mmol), and 14 (404 mg, 1.06 mmol) was heated under reflux
for 4 h. Subsequently, the mixture was cooled to 08C (ice bath),
and CH2Cl2 (15 mL) was added. The aqueous layer of the resulting
two-phase system was acidified to pH 1 by addition of 1m HCl
(13 mL), the ice bath was removed, and the two-phase system was
stirred at 208C for 5 min. The aqueous phase was separated, ex-
tracted with CH2Cl2 (3ꢂ15 mL), and discarded. The combined or-
ganic extracts were dried over anhydrous Na2SO4, and the solvent
was removed under reduced pressure. The solid residue was re-
crystallized from Et2O/n-pentane (slow diffusion of n-pentane via
the gas phase into a solution of the residue in Et2O (10 mL) over a
period of 14 d at 208C) to afford compound 8 in 90% yield as a
Figure 6. Flow cytometric analysis of cell viability and activation of apoptosis
in cultures treated with the retinoids ATRA, bexarotene (1a), disila-bexaro-
tene (1b), 8, and 9 at 10 mm on pluripotent stem cells. Annexin V staining
(A) indicates early-stage apoptosis, whereas 7-AAD staining (B) shows the
presence of dead cells. The proportion of living cells (PI ꢁve) is also shown
(C). The values for live/dead cell staining for 7-ADD and PI ꢁve are roughly
the opposite of each other and follow a similar trend. Necrotic cells induced
by freezing show high levels of 7-ADD staining as anticipated and low levels
for PI ꢁve. Data represent the mean ꢃSD, n=3.
1
colorless crystalline solid (349 mg, 952 mmol); mp: 2108C; H NMR
(500.1 MHz, C6D6): d=0.11 (s, 2H, SiCH2Si), 0.45 (s, 6H, Si(CH3)2),
2
0.47 (s, 6H, Si(CH3)2), 2.12–2.14 (m, 3H, CCH3), 5.30 (d, JHH =1.2 Hz,
2
1H, C=CHH), 5.73 (d, JHH =1.2 Hz, 1H, C=CHH), 7.26–7.31 (m, 2H,
H-3/H-5, Benz (=benzoic acid)), 7.57–7.58 (m, 1H, H-7, Ind (=
1,1,3,3,6-pentamethyl-1,3-disilaindan-5-yl)), 7.69 (brs, 1H, H-4, Ind),
8.08–8.11 ppm (m, 2H, H-2/H-6, Benz), COOH not detected;
13C NMR (125.8 MHz, C6D6): d=ꢁ2.2 (SiCH2Si), 0.7 (Si(CH3)2), 0.8
(Si(CH3)2), 20.4 (CCH3), 117.2 (C=CH2), 126.8 (C-3/C-5, Benz), 128.7 (C-
1, Benz), 130.8 (C-2/C-6, Benz), 133.4 (C-4, Ind), 134.1 (C-7, Ind),
136.8 (C-6, Ind), 142.0 (C-5, Ind), 146.0 (C-4, Benz), 148.0 (C-7a, Ind),
149.7 (C=CH2), 150.5 (C-3a, Ind), 170.5 ppm (COOH); 29Si NMR
(99.4 MHz, C6D6): d=8.5, 8.6 ppm; Anal. calcd (%) for C21H26O2Si2
(366.61): C 68.80, H 7.15; found: C 68.6, H 7.2.
with this observation, our data indicate that bexarotene (1a)
may partially act to induce cell differentiation and cell death
pathways in pluripotent stem cells.
However, it appears that the silicon-containing bexarotene
analogues 1b, 8, and 9 tested in this study lack the ability to
induce stem cell differentiation but they may possess some
interesting enhanced functions over the parent carbon com-
pound bexarotene (1a), such as cell death induction and regu-
lation of cell numbers. The differences observed for the C/Si
pair 1a/1b represent a further interesting example for biologi-
cal sila-substitution effects. In this context it is interesting to
note that bexarotene (1a) and disila-bexarotene (1b) showed
a very similar agonist potency in a HeLa cell-based RXR recep-
tor assay.[3c]
4-[1-(1,1,3,3,6-Pentamethyl-2-oxa-1,3-disilaindan-5-yl)vinyl]ben-
zoic acid (9): A mixture of MeOH (12 mL), H2O (4 mL), KOH
(293 mg, 5.22 mmol), and 17 (200 mg, 523 mmol) was heated
under reflux for 2 h. Subsequently, the mixture was cooled to 08C
(ice bath), and CH2Cl2 (20 mL) was added. The aqueous layer of the
resulting two-phase system was acidified to pH 1 by addition of
2m HCl (3.4 mL), followed by stirring at 08C for 10 min. The aque-
ous phase was separated, extracted with CH2Cl2 (3ꢂ20 mL), and
discarded. The combined organic extracts were dried over anhy-
drous Na2SO4, and the solvent was removed under reduced pres-
sure. The solid residue was dried under vacuum (0.001 mbar, 208C,
1514
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1509 – 1517

Disila-Bexarotene Analogues
4 h) to afford compound 9 in 93% yield as a colorless crystalline
solid (179 mg, 486 mmol); mp: 245–2478C; 1H NMR (500.1 MHz,
CD2Cl2): d=0.35 (s, 6H, Si(CH3)2), 0.36 (s, 6H, Si(CH3)2), 2.04 (dd,
tracts were dried over anhydrous Na2SO4, the solvent was removed
under reduced pressure, and the residue was purified by column
chromatography on silica gel (silica gel, 35–70 mm; eluent, n-
hexane/EtOAc (10:1 v/v)). The relevant fractions (GC analysis) were
combined, the solvent was removed under reduced pressure, and
the solid residue was recrystallized from n-hexane (7 mL; slow
cooling of a hot solution to ꢁ208C) to afford compound 14 in
75% yield as a colorless crystalline solid (411 mg, 1.08 mmol); mp:
157–1598C; 1H NMR (500.1 MHz, CD2Cl2): d=0.00 (s, 2H, SiCH2Si),
5
2
4JHH =0.7 Hz, JHH =0.4 Hz, 3H, CCH3), 5.36 (d, JHH =1.1 Hz, 1H, C=
2
CHH), 5.92 (d, JHH =1.1 Hz, 1H, C=CHH), 7.37–7.41 (m, 2H, H-3/H-5,
4
5
Benz), 7.42 (dq, JHH =0.7 Hz, JHH =0.7 Hz, 1H, H-7, Ind’ (=1,1,3,3,6-
pentamethyl-2-oxa-1,3-disilaindan-5-yl)), 7.43 (dq, 5JHH =0.7 Hz,
5JHH =0.4 Hz, 1H, H-4, Ind’), 7.99–8.04 ppm (m, 2H, H-2/H-6, Benz),
COOH not detected; 13C NMR (125.8 MHz, CD2Cl2): d=1.1
(Si(CH3)2), 1.2 (Si(CH3)2), 20.4 (CCH3), 117.7 (C=CH2), 127.0 (C-3/C-5,
Benz), 128.5 (C-1, Benz), 130.6 (C-2/C-6, Benz), 132.6 (C-4, Ind’),
133.1 (C-7, Ind’), 137.2 (C-6, Ind’), 142.1 (C-5, Ind’), 145.9 (C-3a, Ind’),
146.3 (C-4, Benz), 148.4 (C-7a, Ind’), 149.5 (C=CH2), 170.5 ppm
(COOH); 29Si NMR (99.4 MHz, CD2Cl2): d=14.8, 15.0 ppm; Anal.
calcd (%) for C20H24O3Si2 (368.58): C 65.17, H 6.56; found: C 65.4, H
6.6.
0.29 (s, 6H, Si(CH3)2), 0.31 (s, 6H, Si(CH3)2), 2.01 (s, 3H, CCH3), 3.88
2
(s, 3H, OCH3), 5.32 (d, 2JHH =1.2 Hz, 1H, C=CHH), 5.87 (d, JHH
=
1.2 Hz, 1H, C=CHH), 7.33–7.37 (m, 2H, H-3/H-5, Benz), 7.38–7.396
(m, 1H, H-7, Ind), 7.398–7.41 (m, 1H, H-4, Ind), 7.92–7.96 ppm (m,
2H, H-2/H-6, Benz); 13C NMR (125.8 MHz, CD2Cl2): d=ꢁ2.2
(SiCH2Si), 0.6 (Si(CH3)2), 0.7 (Si(CH3)2), 20.3 (CCH3), 52.3 (OCH3), 117.2
(C=CH2), 126.8 (C-3/C-5, Benz), 129.6 (C-1, Benz), 129.9 (C-2/C-6,
Benz), 133.4 (C-4, Ind), 133.9 (C-7, Ind), 136.6 (C-6, Ind), 141.8 (C-5,
Ind), 145.5 (C-4, Benz), 148.2 (C-7a, Ind), 149.7 (C=CH2), 150.6 (C-3a,
Ind), 167.0 ppm (C(O)OCH3); 29Si NMR (99.4 MHz, CD2Cl2): d=8.6,
8.7 ppm; Anal. calcd (%) for C22H28O2Si2 (380.63): C 69.42, H 7.41;
found: C 69.3, H 7.5.
Bis(ethynyldimethylsilyl)methane (10): This compound was syn-
thesized according to reference [12].
4,4,5,5-Tetramethyl-2-prop-1-ynyl-1,3,2-dioxaborolane (11): This
compound was synthesized according to references [9c] and [13].
1,3-Diethynyl-1,1,3,3-tetramethyldisiloxane (15): This compound
was synthesized according to reference [16].
4,4,5,5-Tetramethyl-2-(1,1,3,3,6-pentamethyl-1,3-disilaindan-5-
yl)-1,3,2-dioxaborolane (12): Compound 11 (1.66 g, 10.0 mmol)
and a 0.1m solution of CoI2 in CH3CN (1.25 mL, 125 mmol CoI2)
were added sequentially in single portions at 208C to a stirred sus-
pension of zinc powder (40 mg, 612 mmol) in CH3CN (10 mL).[14]
The reaction mixture was heated to 508C, and then 10 (902 mg,
5.00 mmol) was added dropwise under stirring within 5 min. The
reaction mixture was stirred at 508C for 15 min and then at 208C
for a further 30 min, and was subsequently filtered through a pad
of silica gel (35–70 mm, 50 g), followed by elution with Et2O/n-
hexane (4:1 v/v, 180 mL). The solvent of the filtrate (including the
eluate) was removed under reduced pressure, Et3N (1 mL) was
added to the residue, and the resulting mixture was then purified
by column chromatography on silica gel (silica gel, 35–70 mm;
eluent, n-hexane/EtOAc (96:4 v/v)). The relevant fractions (GC anal-
ysis) were combined, the solvent was removed under reduced
pressure, and the solid residue was recrystallized from n-hexane
(8 mL; slow cooling of a hot solution to 208C) to afford compound
4,4,5,5-Tetramethyl-2-(1,1,3,3,6-pentamethyl-2-oxa-1,3-disilain-
dan-5-yl)-1,3,2-dioxaborolane (16): Compound 11 (1.66 g,
10.0 mmol) and a 0.1m solution of CoI2 in CH3CN (1.25 mL,
125 mmol CoI2) were added sequentially in single portions at 208C
to a stirred suspension of zinc powder (32 mg, 489 mmol) in CH3CN
(20 mL).[14] The reaction mixture was heated to 408C, and then 15
(912 mg, 5.00 mmol) was added dropwise under stirring within
10 min. The reaction mixture was stirred at 408C for 1 h and then
at 208C for a further 18 h, and was subsequently filtered through a
pad of silica gel (35–70 mm, 80 g), followed by elution with Et2O/n-
hexane (4:1 v/v, 170 mL). The solvent of the filtrate (including the
eluate) was removed under reduced pressure, Et3N (1 mL) was
added to the residue, and the resulting mixture was then purified
by column chromatography on silica gel (silica gel, 35–70 mm;
eluent, n-hexane/Et2O (3:1 v/v)). The relevant fractions (GC analysis)
were combined, the solvent was removed under reduced pressure,
and the solid residue was recrystallized from n-pentane (3 mL;
slow cooling of a hot solution to 208C) to afford compound 16 in
31% yield as a colorless crystalline solid (546 mg, 1.57 mmol); mp:
98–1008C; 1H NMR (500.1 MHz, C6D6): d=0.42 (s, 6H, Si(CH3)2),
0.48 (s, 6H, Si(CH3)2), 1.26 (s, 12H, C(CH3)2), 2.91–2.93 (m, 3H, CCH3),
7.55–7.57 (m, 1H, H-7, Ind’), 8.65–8.68 ppm (m, 1H, H-4, Ind’);
11B NMR (500 MHz, C6D6): d=31.1 ppm; 13C NMR (125.8 MHz,
C6D6): d=1.1 (Si(CH3)2), 1.2 (Si(CH3)2), 22.9 (CCH3), 24.9 (C(CH3)2),
83.4 (C(CH3)2), 132.8 (C-7, Ind’), 139.5 (C-4, Ind’), 144.2 (C-3a, Ind’),
145.8 (C-6, Ind’), 152.1 ppm (C-7a, Ind’), BC (C-5, Ind’) not detected;
29Si NMR (99.4 MHz, C6D6): d=14.1, 14.9 ppm; Anal. calcd (%) for
C17H29BO3Si2 (348.40): C 58.61, H 8.39; found: C 58.7, H 8.4.
12 in 46% yield as
a colorless crystalline solid (805 mg,
2.32 mmol); mp: 153–1558C; 1H NMR (500.1 MHz, C6D6): d=0.05
(s, 2H, SiCH2Si), 0.38 (s, 6H, Si(CH3)2), 0.43 (s, 6H, Si(CH3)2), 1.25 (s,
12H, C(CH3)2), 2.91–2.93 (m, 3H, CCH3), 7.59–7.65 (m, 1H, H-7, Ind),
8.68 ppm (brs, 1H, H-4, Ind); 11B NMR (160.5 MHz, C6D6): d=
31.0 ppm; 13C NMR (125.8 MHz, C6D6): d=ꢁ2.2 (SiCH2Si), 0.5
(Si(CH3)2), 0.7 (Si(CH3)2), 22.9 (CCH3), 24.9 (C(CH3)2), 83.3 (C(CH3)2),
133.6 (C-7, Ind), 140.4 (C-4, Ind), 145.4 (C-6, Ind), 146.0 (C-3a, Ind),
154.1 ppm (C-7a, Ind), BC (C-5, Ind) not detected; 29Si NMR
(99.4 MHz, C6D6): d=8.3, 8.4 ppm; Anal. calcd (%) for C18H31BO2Si2
(346.42): C, 62.41; H, 9.02; found: C, 62.5; H, 9.0.
Methyl 4-[1-(trifluoromethylsulfonyloxy)vinyl]benzoate (13): This
compound was synthesized according to reference [15].
Methyl 4-[1-(1,1,3,3,6-pentamethyl-2-oxa-1,3-disilaindan-5-yl)vi-
nyl]benzoate (17): (Ph3P)2PdCl2 (42.0 mg, 59.8 mmol), 16 (500 mg,
1.44 mmol), and a 2m solution of Na2CO3 in H2O (4.15 mL,
8.30 mmol Na2CO3) were added sequentially at 208C to a stirred
solution of 13 (373 mg, 1.20 mmol) in THF (20 mL). The resulting
mixture was stirred at 208C for 2 h, another portion of 13 (175 mg,
564 mmol) was added, and stirring was continued at 208C for a fur-
ther 17 h. H2O (20 mL) and Et2O (40 mL) were added to the reac-
tion mixture, and the aqueous layer was separated, extracted with
Et2O (3ꢂ20 mL), and discarded. The combined organic extracts
were dried over anhydrous Na2SO4, the solvent was removed
Methyl
4-[1-(1,1,3,3,6-pentamethyl-1,3-disilaindan-5-yl)vinyl]-
benzoate (14): (Ph3P)2PdCl2 (42 mg, 59.8 mmol), 12 (500 mg,
1.44 mmol), and a 2m solution of Na2CO3 in H2O (4.18 mL,
8.36 mmol Na2CO3) were added sequentially at 208C to a stirred
solution of 13 (375 mg, 1.21 mmol) in tetrahydrofuran (20 mL). The
resulting mixture was stirred at 208C for 2 h, another portion of 13
(175 mg, 564 mmol) was added, and stirring was continued at 208C
for a further 18 h. H2O (20 mL) and Et2O (40 mL) were added to the
reaction mixture, and the aqueous layer was separated, extracted
with Et2O (3ꢂ20 mL), and discarded. The combined organic ex-
ChemMedChem 2011, 6, 1509 – 1517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1515

R. Tacke et al.
MED
under reduced pressure, and the residue was purified by column
chromatography on silica gel (silica gel, 35–70 mm; eluent, n-
hexane/EtOAc (10:2 v/v)). The relevant fractions (GC analysis) were
combined, the solvent was removed under reduced pressure, and
the solid residue was recrystallized from n-hexane (4 mL; slow
cooling of a hot solution to ꢁ208C) to afford compound 17 in
75% yield as a colorless crystalline solid (412 mg, 1.08 mmol); mp:
Crystal structure analyses: Suitable single crystals of compounds
8, 12, 16, and 17 were isolated as described in the respective syn-
thetic protocols. Suitable single crystals of compound 14 were ob-
tained by slow evaporation of a saturated solution of 14 in ben-
zene. The crystals were mounted in inert oil (perfluoropolyalkyl
ether, ABCR) on a glass fiber and then transferred to the cold nitro-
gen gas stream of the diffractometer (Bruker Nonius KAPPA APEX II
(16; Montel mirror, Mo Ka radiation, l=0.71073 ꢁ) and Stoe IPDS
(8, 12, 14, and 17; graphite-monochromated Mo Ka radiation, l=
0.71073 ꢁ)). The structures were solved by direct methods
(SHELXS-97).[17] The non-hydrogen atoms were refined anisotropi-
cally (SHELXL-97).[17] A riding model was employed in the refine-
ment of the CH hydrogen atoms. Crystallographic data (excluding
structure factors): CCDC-824053 (8), CCDC-824054 (12), CCDC-
824055 (14), CCDC-824056 (16), and CCDC-824057 (17) contain the
supplementary crystallographic data for this paper. These data can
be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
1
123–1258C; H NMR (500.1 MHz, CD2Cl2): d=0.35 (s, 6H, Si(CH3)2),
0.36 (s, 6H, Si(CH3)2), 2.03–2.04 (m, 3H, CCH3), 3.88 (s, 3H, OCH3),
5.33 (d, 2JHH =1.2 Hz, 1H, C=CHH), 5.89 (d, 2JHH =1.2 Hz, 1H, C=
CHH), 7.33–7.37 (m, 2H, H-3/H-5, Benz), 7.40–7.419 (m, 1H, H-7,
Ind’), 7.420–7.43 (m, 1H, H-4, Ind’), 7.93–7.96 ppm (m, 2H, H-2/H-6,
Benz); 13C NMR (125.8 MHz, CD2Cl2): d=1.1 (Si(CH3)2), 1.2 (Si(CH3)2),
20.4 (CCH3), 52.3 (OCH3), 117.3 (C=CH2), 126.8 (C-3/C-5, Benz), 129.7
(C-1, Benz), 129.9 (C-2/C-6, Benz), 132.6 (C-4, Ind’), 133.1 (C-7, Ind’),
137.2 (C-6, Ind’), 142.2 (C-5, Ind’), 145.4 (C-4, Benz), 145.9 (C-7a,
Ind’), 148.3 (C-3a, Ind’), 149.6 (C=CH2), 167.0 ppm (C(O)OCH3); 29Si
NMR (99.4 MHz, CD2Cl2): d=14.7, 15.0 ppm; Anal. calcd (%) for
C21H26O3Si2 (382.61): C 65.92, H 6.85; found: C 65.9, H 6.7.
Biological evaluations
Bis[dimethyl(1,1,3,3-tetramethyl-1,3-disilaindan-5-yl)silyl]me-
thane (18): Compound 10 (902 mg, 5.00 mmol) and a 0.1m solu-
tion of CoI2 in CH3CN (1.25 mL, 125 mmol CoI2) were added sequen-
tially in single portions at 208C to a stirred suspension of zinc
powder (40 mg, 612 mmol) in CH3CN (10 mL).[14] The reaction mix-
ture was heated under reflux for 18 h, cooled to 208C, and then fil-
tered through a pad of silica gel (63–200 mm, 80 g), followed by
elution with Et2O/n-hexane (4:1 v/v, 500 mL). The solvent of the fil-
trate (including the eluate) was removed under reduced pressure,
and the residue was purified by bulb-to-bulb distillation (200–
2408C/0.02 mbar) to afford compound 18 in 47% yield as a color-
less crystalline solid (422 mg, 780 mmol); mp: 63–648C; 1H NMR
(500.1 MHz, C6D6): d=0.06 (s, 4H, SiCH2Si), 0.41 (s, 12H, Si(CH3)2),
0.427 (s, 12H, Si(CH3)2), 0.430 (s, 2H, SiCH2Si), 0.45 (s, 12H, Si(CH3)2),
7.68–7.72 (m, 4H, H-6 and H-7, Ind), 8.07–8.09 ppm (m, 2H, H-4,
Ind); 13C NMR (125.8 MHz, C6D6): d=ꢁ2.4 (SiCH2Si), 0.0 (Si(CH3)2),
0.6 (Si(CH3)2), 0.7 (Si(CH3)2), 2.3 (SiCH2Si), 131.5 (C-6 or C-7, Ind),
134.0 (C-6 or C-7, Ind), 136.9 (C-4, Ind), 141.3 (C-5, Ind), 149.4 (C-7a,
Ind), 151.0 ppm (C-3a, Ind); 29Si NMR (99.4 MHz, C6D6): d=ꢁ4.1,
8.5, 8.7 ppm; Anal. calcd (%) for C27H48Si6 (541.19): C 59.92, H 8.94;
found: C 59.9, H 8.8.
Cell cultures: All cells were cultured in a humidified atmosphere of
5% CO2 in air at 378C. Cell cultures were handled using standard
aseptic techniques. Cultures with the retinoids bexarotene (1a),
disila-bexarotene (1b), 8, 9, and ATRA were incubated under re-
duced light conditions. Phase contrast images of growing cultures
were obtained using a light microscope (Nikon Diaphot 300), and
photomicrographs were captured using digital photography
(Nikon). TERA2.cl.SP12 cells were cultured in Dulbecco’s modified
Eagle’s medium (DMEM; Sigma) supplemented with 10% fetal calf
serum (FCS; Gibco), 2 mm l-glutamine, and 100 active units each
of penicillin and streptomycin (Gibco). To set up cell differentiation
tests, single cell suspensions were obtained using 0.25% trypsin/
EDTA, and 500000 cells per T25 flask were seeded and left to
settle and proliferate for 24 h. The test retinoid was then added to
the cultures at a concentration of either 1 mm or 10 mm. Cultures
were maintained for up to seven days.
Flow cytometry: Flow cytometric analysis was carried out on live
cells using cell surface markers. A single suspension of cells was
produced by the addition of 1 mL of 0.25% trypsin/EDTA solution
(Cambrex), and cell numbers were determined using a hemocy-
tometer (500000 per antibody). Suspended cells were pelleted and
resuspended in wash buffer (PBS containing 0.1% bovine serum
albumin (BSA; Sigma)). To an appropriate number of wells in a 96-
well plate 500000 cells were added in a 200 mL volume. Plates
were lightly centrifuged, and the supernatant was removed. Each
cell pellet was resuspended in 50 mL of primary antibody solution,
followed by incubation for 1 h at 48C, after which the antibody
was removed by three washes in wash buffer. Cells were then sus-
pended in 50 mL of FITC-conjugated secondary antibody (Cappell,
1:100) and incubated in the dark for 30 min at 48C. Secondary anti-
body was removed by a further three washes in wash buffer. Final-
ly, cell pellets were resuspended in 200 mL of wash buffer, and sam-
ples were analyzed in the flow cytometer (Guava Flow Cytometer).
Thresholds determining the number of positively expressing cells
were set against the negative control antibody P3X. Cytometric
analyses were carried out on three individual cultures per treat-
ment of cultured cells. Stem cell markers SSEA-3 and TRA-1-60
(used neat, generous gift from Prof. P. Andrews, University of Shef-
field, UK) and neural differentiation marker A2B5 (1:100, Abcam)
were analyzed. Cell viability and induction of apoptosis was also
determined by flow cytometry using the Guava Nexin assay ac-
cording to the manufacturer’s protocol (Millipore). In brief, an auto-
1,1,3,3-Tetramethyl-1,3-bis(1,1,3,3-tetramethyl-2-oxa-1,3-disilain-
dan-5-yl)disiloxane (19): Compound 15 (912 mg, 5.00 mmol) and
a 0.1m solution of CoI2 in CH3CN (1.25 mL, 125 mmol CoI2) were
added sequentially in single portions at 208C to a stirred suspen-
sion of zinc powder (40 mg, 612 mmol) in CH3CN (10 mL).[14] The re-
action mixture was heated under reflux for 20 h, cooled to 208C,
stirred for a further 72 h, and then filtered through a pad of silica
gel (63–200 mm, 80 g), followed by elution with Et2O/n-hexane (4:1
v/v, 500 mL). The solvent of the filtrate (including the eluate) was
removed under reduced pressure, and the residue was purified by
bulb-to-bulb distillation (190–2008C/0.02 mbar) to afford com-
pound 19 in 39% yield as a colorless crystalline solid (356 mg,
651 mmol); mp: 73–748C; 1H NMR (500.1 MHz, C6D6): d=0.46 (s,
12H, Si(CH3)2), 0.47 (s, 12H, Si(CH3)2), 0.54 (s, 12H, Si(CH3)2), 7.64
3
5
3
(dd, JHH =7.3 Hz, JHH =1.0 Hz, 2H, H-7, Ind’), 7.84 (dd, JHH =7.3 Hz,
4JHH =1.1 Hz, 2H, H-6, Ind’), 8.15 ppm (dd, 4JHH =1.1 Hz, JHH
=
5
1.0 Hz, 2H, H-4, Ind’); 13C NMR (125.8 MHz, C6D6): d=1.0 (Si(CH3)2),
1.138 (Si(CH3)2), 1.144 (Si(CH3)2), 130.7 (C-7, Ind’), 133.9 (C-6, Ind’),
135.9 (C-4, Ind’), 140.5 (C-5, Ind’), 147.8 (C-3a, Ind’), 149.9 ppm (C-
7a, Ind’); 29Si NMR (99.4 MHz, C6D6): d=ꢁ0.7, 14.4, 14.6; Anal.
calcd (%) for C24H42O3Si6 (547.11): C 52.69, H 7.74; found: C 52.4, H
7.8.
1516
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 1509 – 1517

Disila-Bexarotene Analogues
nemeyer, R. Tacke, Organometallics 2005, 24, 3192–3199; d) M. W. Bꢃtt-
ner, C. Burschka, J. O. Daiss, D. Ivanova, N. Rochel, S. Kammerer, C.
Peluso-Iltis, A. Bindler, C. Gaudon, P. Germain, D. Moras, H. Gronemeyer,
R. Tacke, ChemBioChem 2007, 8, 1688–1699; e) W. P. Lippert, C. Bursch-
ka, K. Gçtz, M. Kaupp, D. Ivanova, C. Gaudon, Y. Sato, P. Antony, N.
Rochel, D. Moras, H. Gronemeyer, R. Tacke, ChemMedChem 2009, 4,
1143–1152; f) R. Tacke, V. Mꢃller, M. W. Bꢃttner, W. P. Lippert, R. Berter-
mann, J. O. Daiß, H. Khanwalkar, A. Furst, C. Gaudon, H. Gronemeyer,
ChemMedChem 2009, 4, 1797–1802.
mated single-cell analysis to monitor early apoptosis was achieved
by measuring Annexin V binding to phosphatidylserine which is
translocated to the outer surface of the membrane during activa-
tion of the apoptotic pathway. Dead and dying cells show loss of
membrane integrity and can be distinguished using the mem-
brane-impermeant dye 7-AAD. Similarly, staining with propidium
iodide was used to identify the proportions of dead and live cells
using standard methods.
[4] S. A. Przyborski, V. B. Christie, M. W. Hayman, R. Stewart, G. M. Horrocks,
Stem Cells Dev. 2004, 13, 400–408.
[5] S. A. Przyborski, Stem Cells 2001, 19, 500–504.
[6] V. B. Christie, J. H. Barnard, A. S. Batsanov, C. E. Bridgens, E. B. Cartmell,
J. C. Collings, D. J. Maltman, C. P. F. Redfern, T. B. Marder, S. Przyborski,
A. Whiting, Org. Biomol. Chem. 2008, 6, 3497–3507.
[7] R. Stewart, V. B. Christie, S. A. Przyborski, Stem Cells 2003, 21, 248–256.
[8] S. A. Przyborski, I. E. Morton, A. Wood, P. W. Andrews, Eur. J. Neurosci.
2000, 12, 3521–3528.
[9] a) L. Doszczak, P. Fey, R. Tacke, Synlett 2007, 753–756; b) L. Doszczak, R.
Tacke, Organometallics 2007, 26, 5722–5723; c) M. W. Bꢃttner, J. B.
Nꢄtscher, C. Burschka, R. Tacke, Organometallics 2007, 26, 4835–4838;
d) S. Metz, J. B. Nꢄtscher, C. Burschka, K. Gçtz, M. Kaupp, P. Kraft, R.
Tacke, Organometallics 2009, 28, 4700–4712; e) R. Mizojiri, R. Conroy, J.
Daiss, E. Kotani, R. Tacke, D. Miller, L. Walsh, T. Kawamoto, Tetrahedron
2010, 66, 7738–7742.
Keywords: bioinorganic chemistry · cell differentiation · cell
proliferation · retinoids · sila-drugs · silicon · stem cells
[1] Reviews: a) R. Tacke, H. Linoh in The Chemistry of Organic Silicon Com-
pounds, Part 2 (Eds.: S. Patai, Z. Rappoport), John Wiley & Sons, Chiches-
ter, 1989, pp. 1143–1206; b) W. Bains, R. Tacke, Curr. Opin. Drug Discov-
ery Dev. 2003, 6, 526–543; c) G. A. Showell, J. S. Mills, Drug Discovery
Today 2003, 8, 551–556; d) J. S. Mills, G. A. Showell, Expert Opin. Invest.
Drugs 2004, 13, 1149–1157; e) P. K. Pooni, G. A. Showell, Mini-Rev. Med.
Chem. 2006, 6, 1169–1177; f) S. McN. Sieburth, C.-A. Chen, Eur. J. Org.
Chem. 2006, 311–322; g) S. Gately, R. West, Drug Dev. Res. 2007, 68,
156–163; h) A. K. Franz, Curr. Opin. Drug Discovery Dev. 2007, 10, 654–
671.
[2] Recent original publications dealing with sila-substituted drugs: a) J. O.
Daiss, C. Burschka, J. S. Mills, J. G. Montana, G. A. Showell, J. B. H. War-
neck, R. Tacke, Organometallics 2006, 25, 1188–1198; b) G. A. Showell,
M. J. Barnes, J. O. Daiss, J. S. Mills, J. G. Montana, R. Tacke, J. B. H. War-
neck, Bioorg. Med. Chem. Lett. 2006, 16, 2555–2558; c) R. Ilg, C. Bursch-
ka, D. Schepmann, B. Wꢃnsch, R. Tacke, Organometallics 2006, 25,
5396–5408; d) R. Tacke, F. Popp, B. Mꢃller, B. Theis, C. Burschka, A. Ha-
macher, M. U. Kassack, D. Schepmann, B. Wꢃnsch, U. Jurva, E. Wellner,
ChemMedChem 2008, 3, 152–164; e) J. B. Warneck, F. H. M. Cheng, M. J.
Barnes, J. S. Mills, J. G. Montana, R. J. Naylor, M.-P. Ngan, M.-K. Wai, J. O.
Daiss, R. Tacke, J. A. Rudd, Toxicol. Appl. Pharmacol. 2008, 232, 369–375;
f) T. Johansson, L. Weidolf, F. Popp, R. Tacke, U. Jurva, Drug Metab.
Dispos. 2010, 38, 73–83; g) R. Tacke, B. Nguyen, C. Burschka, W. P. Lip-
pert, A. Hamacher, C. Urban, M. U. Kassack, Organometallics 2010, 29,
1652–1660.
[10] F. Lansigan, F. M. Foss, Drugs 2010, 70, 273–286.
[11] L. Qu, X. Tang, Cancer Chemother. Pharmacol. 2010, 65, 201–205.
[12] a) D. J. Miller, G. A. Showell, R. Conroy, J. Daiss, R. Tacke, D. Tebbe,
Amedis Pharmaceuticals Ltd., Cambridge (UK), PCT Int. Appl. WO/2005/
005443A1, 2005; b) T. Kusumoto, T. Hiyama, Chem. Lett. 1988, 17, 1149–
1152.
[13] E. C. Hansen, D. Lee, J. Am. Chem. Soc. 2005, 127, 3252–3253.
[14] A suspension of zinc powder in acetonitrile was treated with catalytic
amounts of iodine and then heated until a colorless mixture resulted.
[15] F.-L. Qing, J. Fan, H.-B. Sun, X.-J. Yue, J. Chem. Soc. Perkin Trans. 1 1997,
3053–3057.
[16] W.-Y. Wong, C.-K. Wong, G.-L. Lu, J. Organomet. Chem. 2003, 671, 27–
34.
[17] G. M. Sheldrick, Acta Crystallogr. Sect. A 2008, 64, 112–122.
[3] a) J. G. Montana, G. A. Showell, I. Fleming, R. Tacke, J. Daiss, Amedis
Pharmaceuticals Ltd., Cambridge (UK), PCT Int. Pat. Appl. WO/2004/
048390A1, June 10, 2004; b) J. G. Montana, G. A. Showell, R. Tacke,
Amedis Pharmaceuticals Ltd., Cambridge (UK), PCT Int. Pat. Appl. WO/
2004/048391A1, June 10, 2004; c) J. O. Daiss, C. Burschka, J. S. Mills,
J. G. Montana, G. A. Showell, I. Fleming, C. Gaudon, D. Ivanova, H. Gro-
Received: March 24, 2011
Revised: May 12, 2011
Published online on July 1, 2011
ChemMedChem 2011, 6, 1509 – 1517
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1517