Water-soluble carbosilane dendrimers: Synthesis biocompatibility and complexation with oligonucleotides; evaluation for medical applications

Article abstract of DOI:10.1002/chem.200600594

Novel amine- or ammoniumterminated carbosilane dendrimers of type nG-[Si{OCH₂(C₆H₃)-3,5-(OCH₂-CH ₂NMe₂)₂}]_x, nG-[Si(O(CH ₂)₂N(Me)-(CH₂

Full text of DOI:10.1002/chem.200600594

FULL PAPER
DOI: 10.1002/chem.200600594
Water-Soluble Carbosilane Dendrimers: Synthesis Biocompatibility and
Complexation with Oligonucleotides; Evaluation for Medical Applications
Jesus F. Bermejo,[a, e] Paula Ortega,[b, e] Louis Chonco,[a] Ramon Eritja,[c]
Rafael Samaniego,[d] Matthias Müllner,[a] Ernesto de Jesus,[b] F. Javier de la Mata,*[b]
[a]
Juan Carlos Flores,[b] Rafael Gomez,*[b] andAngeles MuÇoz-Fernandez*
Abstract: Novel amine- or ammonium-
terminated carbosilane dendrimers of
quaternization with HCl. Quaternized
carbosilane dendrimers are soluble in
water, although degradation is appa-
profiles over extended periods. In addi-
tion, we describe a study on the inter-
actions between the different carbosi-
lane dendrimers and DNA oligodeoxy-
nucleotides (ODNs) and plasmids
along with a comparative analysis of
their toxicity. They can form complexes
with DNA ODNs and plasmids at bio-
compatible doses via electrostatic inter-
action. Also a preliminary transfection
assay has been accomplished. These re-
sults demonstrate that the new ammo-
nium-terminated carbosilane dendri-
type nG-[Si
CH2NMe2)2}]x, nG-[Si{O
(CH2)2NMe2}]x and nG-[Si
NH2}]x or nG-[Si{OCH2(C6H3)
(OCH nG-[Si{O-
2CH2NMe3+IÀ)2}]x,
(CH2)2N(Me) and
(CH2)2NMe3+IÀ}]x,
nG-[Si
{(CH2)3NH3+ClÀ}]x have been
A
ACHTREUNG
À
A
G
A
rent due to hydrolysis of Si O bonds.
À
A
G
A
However, dendrimers containing Si C
A
N
G
N
bonds are water-stable. The biocompat-
ibility of the second-generation den-
drimers in primary cell cultures of pe-
ripheral blood mononuclear cells
(PBMCs) and erythrocytes have been
analyzed, and they show good toxicity
ACHTREUNG
A
ACHTREUNG
ACHTREUNG
synthesized and characterized up to the
third generation by two strategies:
À
1) alcoholysis of Si Cl bonds with
amino alcohols and subsequent quater-
nization with MeI, and 2) hydrosilyla-
Keywords: carbosilanes
mers · DNA · drug delivery · oligo-
nucleotides
· dendri-
A
AHCTREUNG
À
tion of allylamine with Si H bonds of
considered for biomedical applications.
AHCTREUNG
the dendritic systems and subsequent
Introduction
[a] Dr. J. F. Bermejo, L. Chonco, M. Müllner, Dr. A. MuÇoz-Fernandez
Laboratorio de Inmunobiología Molecular
Hospital General Universitario Gregorio MaraÇón, Madrid (Spain)
Fax : (+34)91-586-8018
E-mail: mmunoz.hgugm@salud.madrid.org
There is currently significant interest in dendrimers as a
result of their potential applications, including light harvest-
ing and energy transfer, nanoscale catalysis, chemical sen-
sors, unimolecular micelles, enzyme mimics, encapsulation
of guest molecules, molecular recognition, diagnostic agents,
and gene and drug delivery.[1] In particular, the structural
precision of dendrimers has motivated numerous studies
aimed at biomedical applications. One of the most active re-
search areas in dendrimer-based therapeutics is DNA trans-
fection. This application is currently based on viral vectors
as the most efficient gene-transfer agents. However, the use
of viral vectors suffers from adverse effects, such as immune
reaction against the viral vector or lymphoproliferative syn-
dromes associated with oncogene dysregulation.[2] To over-
come these drawbacks, nonviral vehicles such as cationic li-
[b] P. Ortega, Dr. E. de Jesus, Dr. F. J. de la Mata, Dr. J. C. Flores,
Dr. R. Gomez
Departamento de Química Inorgµnica
Universidad de Alcalµ, Campus Universitario
28871 Alcalµ de Henares (Spain)
Fax : (+34)91-885-4683
E-mail: javier.delamata@uah.es
rafael.gomez@uah.es
[c] Dr. R. Eritja
Instituto de Biología Molecular de Barcelona, CSIC
Jordi Girona, Barcelona (Spain)
[d] Dr. R. Samaniego
Unidad de Microscopía Confocal
Hospital General Universitario Gregorio MaraÇón, Madrid (Spain)
[e] Dr. J. F. Bermejo, P. Ortega
A
U
Paula Ortega and Jesus F. Bermejo contributed equally to the devel-
opment of this work.
Once in contact with negatively charged DNA oligodeoxy-
nucleotides (ODNs) or plasmids, cationic systems form elec-
trostatic complexes with the nucleic acids, known as lipo-
Supporting Information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
483

+
plexes, polyplexes or dendriplexes. The use of liposomes for
transfection purposes was first described in 1987.[3] Cationic
lipids prepared for this purpose are commercially available
(e.g., Cytofectin and Lipofectin). However, they have side-
effects such as inflammatory lung reactions,[4] and transfec-
tion may fail in the presence of serum. The main drawback
of the use of conventional degradable polymers as delivery
agents, beside their polydispersity, is their thermodynamic
instability, which results in a short in vivo lifetime of the
active species.[5]
Dendrimers represent an alternative approach for the
transfection of nucleic acids in a wide range of cells. The
major advantage of dendrimers over other nonviral vehicles
is their uniform structure and the versatility with which their
skeletons and surfaces can be modified, which allows precise
characterization of the nucleic acid/vector complex and a
more accurate and systematic analysis of the transfection
process. The first study on using dendritic macromolecules
for transfection, reported in 1993, used polyamidoamine
(PAMAM) dendrimers,[6] and since then extensive studies
have been performed.[7] Good results are typically achieved
with sixth- or seventh-generation dendrimers. However, the
transfection efficiency can be increased two- to threefold
when PAMAM is activated by heat treatment (e.g., Super-
fect (SF)).[8]
[Si(OCH2CH2NMe3+IÀ)]8 (1) and 2G-[Si(OCH2CH2NMe3
IÀ)2]8 (2).[16] Here we describe the synthesis of new water-
soluble carbosilane dendrimers up to the third generation
and analyze their biocompatibility in primary cell cultures
of peripheral blood mononuclear cells (PBMCs) and eryth-
rocytes. These data allow a comparative analysis of the tox-
icity of the different carbosilane macromolecules. In addi-
tion, we describe a study on complex formation between
these dendrimers and ODNs. These studies provide essential
insights into the potential of these carbosilane dendrimers
as drug delivery systems, and in particular for their use in
delivery of DNA (ODNs, plasmids).
Results andDiscussion
We previously prepared a family of new peripheral amine-
and ammonium-terminated carbosilane dendrimers by alco-
holysis of the well-known chlorosilane-terminated dendri-
mers[17] nG-(SiCly)x, with N,N-dimethylethanolamine and
subsequent quaternization with MeI to afford dendrimers of
type nG-[Si(OCH2CH2NMe2)y]x and nG-[Si(OCH2CH2N-
Me3+IÀ)y]x, respectively.[16] However, all these systems are
À
sensitive to hydrolysis by slow cleavage of the Si O bonds.
To study such behavior and to get better insight into the po-
tential biomedical applications of the carbosilane systems,
new families of dendrimers were prepared.
Another class of potential transfecting agents are phos-
phorus-containing dendrimers,[9] which can be synthesized
up to the twelfth generation. The dendritic surface has been
grafted with protonated or methylated terminal tertiary
amines and examined as transfecting agent for the luciferase
gene in 3T3 cells. The efficiency increased as a function of
the dendrimer generation, although a constant value was
reached between generations three and five. Furthermore,
these dendrimers exhibit improved transfection efficiency
when serum is present.
Amine-terminatedcarbosilane dendrimers : We studied the
synthesis of new dendrimers with amino groups at their
periphery. For this purpose two general strategies were de-
veloped: 1) an extension of the alcoholysis of dendritic
Si Cl bonds by using modified terminal amine fragments,
and 2) hydrosilylation of allyl amines with dendritic Si H
À
À
bonds.
Other kinds of dendritic macromolecules, such as polypro-
pylenimine (PPI),[10] and polylysine[11] dendrimers, have also
been studied as potential DNA or ODN carriers. For in-
stance, low-generation PPI dendrimers have also shown
gene-transfection ability in vitro with low cytotoxicity, al-
though at higher generations increased toxicity prohibits
their use.
To our knowledge, no studies concerning the use of water-
soluble carbosilane-based dendrimers as potential carriers
for DNA or ODNs or for other biomedical applications
have been published, although an in vitro biocompatibility
study on poly(ethylene oxide)-grafted carbosilane dendri-
À
Dendrimers formed by alcoholysis of dendritic Si Cl bonds:
Chlorosilane-terminated dendrimers of type nG-(SiCl)x (n=
1, 2 and 3; x=4, 8 and 16) were synthesized as reported pre-
viously[17] and formed the starting materials for the prepara-
tion of new dendrimers by alcoholysis. Two different amino
alcohols were used: 3,5-(NMe2CH2CH2O)2C6H3CH2OH (I)
and Me2NCH2CH2N(Me)CH2CH2OH (II). The latter was
obtained from commercial sources, and its selection was mo-
tivated by presence of a second amino group that, if it is not
quaternized, may trap endosomal protons and reducing nu-
cleases. The former was elected to restrict the hydrolysis
process and was synthesized previously (see Experimental
Section).
mers has recently been reported.[12] Only three synthetic
studies of polycationic silane dendrimers have been reported
previously,[13–15] and only one of them has medical implica-
tions.[15]
We recently developed a synthetic strategy for the prepa-
ration of new peripheral amine- and ammonium-terminated
carbosilane dendrimers of type nG-[Si(OCH2CH2NMe2)y]x
and nG-[Si(OCH2CH2NMe3+IÀ)y]x, and a preliminary evalu-
ation of their toxicity and complexation with ODNs was car-
ried out, mainly on second-generation dendrimers 2G-
Chlorosilane-terminated dendrimers were treated with
stoichiometric amounts of amino alcohol I or II in diethyl
ether in the presence of an excess of NEt3 to afford the cor-
responding amine-terminated dendrimers nG-[Si
(C6H3)-3,5-(OCH2CH2NMe2)2}]x (n=1, x=4 (3); n=2, x=8
(4); n=3, x=16 (5)) and nG-[Si{O(CH2)2N(Me)-
(CH2)2NMe2}]x (n=1, x=4 (6); n=2, x=8 (7); n=3, x=16
(8)) in high yields as colorless or yellow oils (Scheme 1). All
A
AHCTREUNG
AHCTREUNG
AHCTREUNG
484
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 483 – 495

Water-Soluble Carbosilane Dendrimers
FULL PAPER
Scheme 1.
these dendrimers are soluble in all common organic solvents
but insoluble in water.
Ammonium-terminated carbosilane dendrimers: The ammo-
nium-terminated dendrimers were prepared by adding MeI
to parent dendrimers 3–8 in diethyl ether (see Scheme 1).
Dendrimers 3–5 reacted with an excess of MeI to cleanly
afford the corresponding quaternized derivatives nG-[Si-
À
Dendrimers formed by hydrosilylation with dendritic Si H
bonds: SiH-terminated carbosilane dendrimers of type nG-
(SiH)x (n=1 and 2; x=4 and 8) were synthesized as report-
ed previously[17d,e] and used in the hydrosilylation of allyl-
A
N
n=2, x=8 (12); n=3, x=16 (13)) as white solids. In the
conversion of amine-terminated dendrimer 5 to 13, some di-
amine. The reactions were performed in ampoules with J
Young valves using THF as solvent and the Speier catalyst[18]
methylamino groups remained unquaternized. The H N MR
1
and heating the mixture at 1208C for 4 h to afford the corre-
spectrum revealed that roughly 85% of the amino groups
were quaternized, even if we used a large excess of MeI and
prolonged reaction times. The stoichiometric reaction of
sponding dendrimers nG-[Si(CH2)3NH2]x (n=1, x=4 (9);
A
n=2, x=8 (10)) in high yields as colorless oils (see
Scheme 2). 1H NMR spectroscopy was used to follow the
dendrimers 6–8 containing 4,
8
and 16 terminal [O-
(CH2)2NMe2]
A
ACHTREUNG
units, respectively, with same
number of equivalents of MeI
gave the corresponding outer-
most
amine compounds nG-[Si{O-
(CH2)2N(Me)
(CH2)2NMe3+}IÀ]x
quaternized
terminal
A
ACHTREUNG
(n=1, x=4 (14); n=2, x=8
(15); n=3, x=16 (16)), albeit
Scheme 2.
with small amounts of doubly
methylated branches (less than
À
progress of the reactions by monitoring the loss of the Si H
resonance. The processes occurred exclusively by b-addition
1–3% depending on the generation). Hence, minor chemical
competition between the two types of amino groups per pe-
ripheral unit was observed, although the selectivity in-
creased with increasing generation or peripheral hindrance
of the dendrimers.
À
to give the SiCH2CH2CH2NH2 group, and neither a-addi-
tion products nor N-silylated byproducts were detected in
the H NMR spectra of the crude products. All of these den-
1
drimers are soluble in all common organic solvents but in-
soluble in water.
The NMR spectroscopic, MALDI-TOF MS and analytical
data for compounds 3–10 are consistent with their proposed
structures (Schemes 1 and 2).
Dendrimers 9 and 10 were quaternized by adding stoi-
chiometric amounts or an excess of HCl to produce nG-[Si-
A
white solids (see Scheme 2).
All compounds 11–18 are water-soluble, although solubili-
ty decreases with increasing generation. However, the den-
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
485

R. Gomez, A. MuÇoz-Fernandez, F. J. de la Mata et al.
À
drimers with Si O bonds decomposed slowly by hydrolysis
temperature for dendrimers 11–16, and in D2O for dendri-
mers 17 and 18. In these solvents the line widths of these
A
of these bonds. This behavior was observed in dendrimers 1
[16]
À
and 2 with OCH2CH2NMe2 terminal units. However, the
spectra tended to be broader than those of derivatives solu-
1
hydrolysis rates of dendrimers 11–13 containing quaternized
groups derived from amino alcohol I are considerably atte-
nuated with respect to dendrimers with fragment II as pe-
ripheral groups, which are similar to those in 1 and 2. In
contrast, dendrimers 17 and 18 based on Si C bonds are
completely stable towards hydrolysis.
The NMR spectroscopic and analytical data of 11–18 are
consistent with their proposed structures (Schemes 1–3).
The H NMR spectra were recorded in [D6]DMSO at room
ble in common organic solvents. The H and 13C NMR spec-
tra of the quaternized dendrimers exhibit identical reso-
nance patterns to those observed in their neutral counter-
parts 3–10 for the carbosilane framework, although broader
signals are seen with increasing generation (see Experimen-
tal Section and Supporting Information). In general for the
1H NMR spectra, the quaternization of the amine groups re-
sults in deshielding of about Dd=1 ppm for the geminal
methylene and methyl groups directly bound to the charged
À
1
Scheme 3. Molecular representations of ammonium-terminated carbosilane dendrimers 12, 15, and 18.
486
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 483 – 495

Water-Soluble Carbosilane Dendrimers
FULL PAPER
nitrogen atoms, whereas small downfield shifts of around
Dd=0.3–0.4 ppm are found for the vicinal methylene
groups. Beyond these positions, no displacement is observed
for the chemical shift due to the positive charge on the ni-
trogen atoms. However, this effect is more evident in the
compounds quaternized by MeI than those prepared with
HCl. Analogous behavior is observed for the carbon atoms
in the 13C NMR spectra. Dendrimers 14–16 with outermost
1
quaternized amines exhibit different H and 13C NMR pat-
terns for singly and doubly methylated branches that facili-
tate their discrimination (see Experimental Section and Sup-
porting Information). In any case, signals attributed to the
monomethylated forms of the innermost quaternized amine
were observed, even when a deficit of reagent was added.
This strongly suggests that the quaternization process starts
at the outermost amine and subsequently proceeds to the in-
nermost when a slight excess of MeI is added. Attempts to
carry out MALDI-TOF MS of these dendrimers failed prob-
ably due, inter alia, to solubility problems.
Toxicity evaluation of dendrimers: Quaternized second-gen-
eration carbosilane dendrimers 12, 15, and 18 were tested
on a primary cell culture of peripheral blood mononuclear
cells (PBMCs) from healthy donors as an initial screening
for biocompatibility, and the values compared with prelimi-
nary data shown by dendrimers 1 and 2.[16] First-generation
dendrimer 14 was too water sensitive for toxicity evaluation,
while third-generation dendrimers were not tested due to
solubility problems.
Toxicity was evaluated by challenging PBMCs with in-
creasing concentrations of the free quaternized carbosilane
dendrimers, in order to obtain a range of biocompatibility.
Toxicity was initially evaluated by 1) visual examination
under a phase-contrast light microscope and 2) MTT toxicity
assay.
According to microscopy studies (see Supporting Infor-
mation), the carbosilane dendrimers best tolerated by
PBMCs were 12 and 15, whereas the dendrimer 18 showed
the highest toxicity and dendrimers 1 and 2 had intermedi-
ate toxicity profiles. Superfect (SF) and a 4G-PAMAM
showed higher toxicity than all second-generation carbosi-
lane dendrimers. However, this is not surprising because of
it is known that dendrimer toxicity increases with increasing
generation.[1i,19]
In the MTT assay, cells treated with 12 and 15 showed
higher mitochondrial activity (MA) (Figure 1A). Dendrimer
18 was more toxic at 1 mm, but the toxicities of 1 and 2
strongly increase from 5 mm and beyond. However, taking
into account the data from visual observation along with the
MTT values as a whole, it is deduced that whilst 1 and 2 in-
duced diminution of birefringence or formation of cell ag-
gregates, 12 and 15 did not and are thus the most biocom-
patible systems for PBMCs. In addition, and in view of the
results of this initial screening, one can conclude that the
upper limit of the carbosilane dendrimer concentration to
be further considered for biological assays should be be-
tween 1 and 5 mm.
Figure 1. A) Quantification of mitochondrial activity (MA) of PBMCs by
MTT test after 48 h of incubation with different dendrimer concentra-
tions. MA is expressed as percentage with respect to the MA of the con-
trol (untreated cells). B) Flow cytometry of PBMCs after 72 h of incuba-
tion with dendrimers: B1) representation of cells treated with dendrimer
15; B2) representation of cells treated with SF dendrimer; B3) evaluation
of the percentage of cells with FW and SD corresponding to live cells
against the percentage of cells corresponding to dead/apoptotic cells.
Black areas denote dead cells, and grey areas live cells. C) Percentage of
cells positive for Trypan Blue after 72 h of incubation with dendrimers.
All experiments were assayed three times.
Because of the apparent discrepancy between MTT assays
and the microscopic observations, which were also found for
other polycationic dendrimers, such as those based on the
phosphoramidothioate backbone,[20] toxicity of the carbosi-
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
487

R. Gomez, A. MuÇoz-Fernandez, F. J. de la Mata et al.
lane dendrimers was studied by additional methods: flow cy-
tometry, Trypan Blue (TB) uptake, DAPI staining and in
vivo microscopy. In these studies, the incubation time was
increased to 72 h, and the doses assayed for carbosilane den-
drimers were in the range of 2–4 mm, while less than 1 mm
was used for the reference SF PAMAM dendrimer.
ic changes were observed even at 1 mm concentration. Anal-
ogous results were detected for 4G PAMAM dendrimer
(see Supporting Information).
The interaction of cationic dendrimers with negatively
charged membranes was also studied by hemolysis experi-
ments. The release of hemoglobin was used to quantify the
membrane-damaging properties of the dendritic macromole-
cules. Lysis of red blood cells precludes direct intravenous
delivery of the desired agents and often enhances their tox-
icity when administered by other routes. Hemolysis was re-
corded spectrophotometrically at 1 h and considered 100%
in the positive control (cells treated with Triton X-100) and
7% in the negative control of untreated cells. The absorban-
ces of the different wells were expressed as percentages with
respect to the absorbance of the positive control well, and
from this percentage, the 7% corresponding to the negative
control was subtracted. A value of 10% was considered as
the cut-off for toxicity.[19b] Carbosilane dendrimers 12 and 15
induced less hemoglobin release in the 1–5 mm concentration
range, although at 5 mm, these two dendrimers slightly
exceed 10% (see Figure 2). In spite of this good result, com-
Flow cytometry (FC): PBMCs treated with carbosilane den-
drimers at these doses did not show significant changes in
their forward (FW) and side (SD) light-scattering character-
istics after 72 h of incubation. Moreover, the percentage of
unviable cells was similar in all cases assayed to that of un-
treated cells (around 20%, see Figure 1B). On the other
hand, PBMCs treated with SF dramatically increased their
mortality.
Trypan Blue (TB)uptake : TB is excluded by viable cells but
can penetrate cell membranes of dying or dead cells. When
TB staining is negative, membrane integrity is present. The
percentage of cells positive for TB staining for the wells
treated with dendrimers 1, 2, 12, 15, and 18 was similar to
that found in untreated cells (Figure 1C). However, cells
treated with SF displayed significant mortality compared
with control cells.
DAPI staining: An additional test of cell viability was pro-
vided by the staining of cell nuclei with the vital dye DAPI.
Cell nuclei undergoing apoptosis or necrosis show reduced
nuclear size, chromatin condensation and nuclear fragmen-
tation, processes that can be easily detected by DAPI stain-
ing. Cells treated with all carbosilane dendrimers had similar
appearances to untreated cells, showing rounded nuclei with
homogeneously distributed chromatin. In contrast, cells
treated with SF decreased dramatically in number. Further-
more, the cell pellet obtained for DAPI staining after centri-
fugation of PBMCs treated with SF was notably small. This
reduction in cell number was confirmed by microscopy (see
Supporting Information).
Figure 2. Hemolysis test: hemoglobin release after 1 h with different den-
drimer concentrations. The experiment was assayed three times.
pound 12 induced hemagglutination, while 15 did not at this
concentration (see Supporting Information). It can be said
that polycation/DNA complexes are usually less cytotoxic
than uncomplexed polycations.[21] Thus, the experimental
conditions chosen here reflect the worst-case scenario for
cytotoxicity. In conclusion, taking into account both hemag-
glutination and hemoglobin release from erithrocytes, as
well as toxicity in lymphocytes, dendrimer 15 is the least
harmful for both types of cell at higher concentrations and
thus a good candidate as delivery agent.
Time-lapse video microscopy: We monitored cell behavior
by time-lapse and in vivo microscopy for cells treated with
carbosilane dendrimers. In all cases, cells showed similar
patterns of movement and migration to untreated control
cells (for an example, see Supporting Information).
Therefore, from these four additional methods, all carbo-
silane dendrimers 1, 2, 12, 15, and 18 assayed under these
conditions showed good biocompatibility on PBMC cells.
Toxicity was also evaluated on erythrocytes challenged
with increased dendrimer concentrations by means of induc-
tion of hemagglutination along with hemoglobin release
from the red blood cells. Dendrimers 1, 2, 12, and 15 were
used for this assay, and dendrimer 18 was not included be-
cause it was shown to be the most toxic for lymphocytes in
some of the methods and concentrations employed before.
A fourth-generation PAMAM dendrimer was also used for
comparison. From visual examination, dendrimer 15 showed
the lowest agglutination, while for dendrimer 12 morpholog-
Antigenicity: lymphoproliferative assay: When a new mac-
romolecule is under consideration for a potential biological
application, it is important that it does not constitute an un-
specific antigenic stimulus (unless desired for development
of immunogens). Again we challenged PBMCs with differ-
ent concentrations of carbosilane dendrimers 1, 2, 12, and
15 looking for induction/noninduction of proliferation, com-
pared with that achieved with an unspecific mithogenic stim-
488
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 483 – 495

Water-Soluble Carbosilane Dendrimers
FULL PAPER
uli such as phytohemagglutinin (PHA). From the data
shown in Figure 3, none of the carbosilane dendrimers con-
stituted an antigenic stimulus for PBMCs at the tested con-
centrations (2 and 5 mm).
Figure 3. Lymphoproliferative assay: antigenicity of the carbosilane den-
drimers in the proliferation of PBMCs exposed to two different concen-
trations of each dendrimer and compared with that achieved with phyto-
hemagglutinin (PHA). Data expressed as the Napierian logarithm of the
number of counts per minute. The experiment was assayed three times.
Dendriplex formation: Having evaluated the biocompatibili-
ty range for the new water-soluble ammonium carbosilane
dendrimers, we studied their capacities for complexation
with oligodeoxynucleotides (ODN) to give so-called dendri-
plexes. Quaternized second-generation dendrimers 12, 15,
and 18 as well as elsewhere-prepared 1 and 2,[16] where used
for potential dendriplex formation. Dendriplex formation
was studied by electrophoresis in agarose gels with fluores-
ceinated phosphorothioate ODNs. Mixtures were prepared
with an excess of positive charge to provide a sufficient
number of charges to interact with the ODN. Preliminary
results[16] for dendrimer 1 at different electrostatic charge
ratios (+)/(À) showed retardation of ODNmigration in all
cases, that is, the ODNis associated with the dendrimer,
even at 2:1 charge ratio, for which the dendrimer concentra-
tion is in the aforementioned range of biocompatibility.[16]
Consequently, the carbosilane dendrimers of this work were
complexed with ODNat an electrostatic charge ratio ( +)/
(À) of 2/1. Figure 4 shows agarose gel electrophoresis of the
mixtures of dendrimers 12, 15, or 18 and ODNin compari-
son with those formed with 1 and 2. In the gel presented in
Figure 4A1, the dendrimers were diluted in water and im-
mediately mixed with ODNand run in an agarose gel. In all
cases the different carbosilane dendrimers retained the
ODNduring its migration, and this demonstrates successful
complex formation. With regard to the hydrolysis problems
Figure 4. A) Electrophoresis of carbosilane dendrimer/ODNdendriplexes
with (+)/(À)=2/1 on 3% agarose gel: 1) 1/ODN; 2) 18/ODN; 3) 15/
ODN; 4) 12/ODN; 5) ODN only. A1) Dendrimers dissolved in water and
immediately mixed with ODN; A2) dendrimers dissolved in water for
24 h and then mixed with ODNprior to electrophoresis. Asterisk denotes
100 bp DNA ladder as reference. B) Electrophoresis of carbosilane den-
drimer/ODNdendriplexes with ( +)/(À)=2/1, on 3% agarose gel at
pH 7.4 at different incubation times: 1) 1/ODN; 2) 2/ODN; 3) 15/ODN;
4) ODNonly. Asterisk denotes 100 bp DNA ladder as reference. C) Elec-
trophoresis of carbosilane dendrimer/ODNdendriplexes with ( +)/(À)=
2/1 on 3% agarose gel at different pHs: 1) pH 2.8; 2) pH 3.7; 3) pH 4.7;
4) pH 5.7; 5) pH 6.4; 6) pH 7.4; 7) pH 8.0. Asterisk denotes 100 bp DNA
ladder as reference. The experiments were assayed three times.
spermine and low-molecular weight dendrons based on
Newkome-type branching.[22] This finding confirms that the
hydrolysis processes suffered by the carbosilane dendrimers
are slow under highly dilute conditions. It is noteworthy that
our results show that second-generation carbosilane den-
drimers are capable of complexing ODNs to give [ODN/
dendrimer] through electrostatic interactions, whereas
higher generations of other dendrimer topologies have been
used in the literature.[7,20,21] For the gel visualized in Figure
4A2, the dendrimers were dissolved in water for 24 h and
then mixed with ODNbefore electrophoresis. After this
time, dendrimers 1 and 15 were not able to retain the ODN
À
presented by the Si O-containing carbosilane dendrimers,
we recently demonstrated that the terminal units released
after hydrolytic scission were not able to retard ODNmigra-
tion, and that the whole functionalized dendrimer was nec-
essary to form complexes with DNA.[16] An analogous result
was recently obtained in binding experiments on DNA with
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
489

R. Gomez, A. MuÇoz-Fernandez, F. J. de la Mata et al.
during its migration, that is, the time dissolved in water af-
fects the ability of the dendrimer to form complexes with
the ODN. On the other hand, 12 and 18 were not affected
to the same degree and preserved the capacity to retain the
majority of the ODNduring its migration. Thus, under these
À
conditions dendrimers containing Si O bonds lost their abil-
ity to bind ODNs after 24 h in water in dilute concentra-
tions, except for dendrimer 12. Likewise, dendrimer 18 is
À
also water-stable because of the presence of robust Si C
bonds.
For water-sensitive dendrimers 1, 2, and 15, a second type
of electrophoretic migration assay was developed consisting
of running dendriplexes in a gel after 0, 6, and 24 h of incu-
bation in an atmosphere of 5% of CO2 at 378C. As can see
from Figure 4B, the three dendrimers released the ODNs
progressively. It is possible to conclude that 1, 2, and 15
have the ability to release the ODNin a time-dependant
way when they are dissolved in water. This feature suggests
potential use in controlled release of ODNs and perhaps of
other polyanionic drugs. The controlled release of active
substances based on the chemical stability of the linker to-
wards hydrolysis has been described, for example, in the
case of phosphorus dendrimers on the basis of slow degrada-
tion of imine bonds.[23]
Dendriplex stability was also tested at different pH
values. Blood physiological pH is 7.4, but anatomical or cel-
lular locations with more acidic (stomach, endosome/lyso-
some) or basic (duodeni) pH exist. Dendriplexes were
formed as usual and exposed to different solutions from
acid to basic pH (2.8, 3.7, 4.7, 5.7, 6.4, 7.4 and 8.0) prior
electrophoresis (see Figure 4C). Dendriplexes formed be-
tween 12 or 18 and ODNwere stable at all tested pH
values, whereas dendriplexes formed between 15 and ODN
released the ODNat acid pH ( <5.7). For dendrimers 1 and
2 the results were similar: dendriplexes 1/ODNdissociated
at pH<4.7, and dendriplexes 2/ODNat pH <5.7. All den-
driplexes were stable at basic pH (up to pH 8.0). Thus, in an
acidic environment dendriplexes 1/ODN, 2/ODNand 15/
ODNwould release the ODN, whereas under basic condi-
tions they would remain stable. This opens new perspectives
for applications that need pH-controlled ODNrelease.
Toxicity profiles of dendriplexes formed by carbosilane
dendrimers 1, 2, 12, 15, and 18 were studied by some of the
methods used for dendrimers alone: flow cytometry and
Trypan Blue (TB) uptake (see Supporting Information). The
data showed very similar values to those obtained for den-
drimers without complexation to ODN, as has been reported
elsewhere.[19]
Figure 5. Electrophoresis of dendrimer 1/NfkB plasmid dendriplexes on
3% agarose gel: 1) (+)/(À)=2/1; 2) (+)/(À)=6/1; 3) (+)/(À)=10/1;
4) (+)/(À)=100/1; 5) and 6) plasmid only. Asterisk denotes 5000 bp
DNA ladder as reference. The experiment was assayed three times.
that, regardless of the low generation, the carbosilane den-
drimers have the capacity to bind large DNA molecules.
Finally, a preliminary study of the capacity of the dendri-
plex formed by 15 and the fluoresceinated ODNto pene-
trate into PBMCs was performed by confocal microscopy.
Figure 6 shows the internalization and intracellular distribu-
tion of the nucleic material, with which the dendrimer does
not seem to interfere. Research is in progress to elucidate
the transfection process.
In addition, the ammonium-terminated second-generation
carbosilane dendrimers were able to complex with plasmids.
As a demonstrative example, dendrimer 1 formed dendri-
plexes with the NfkB plasmid (which codifies Nf-kappaB
protein involved in regulation of immune or inflammation
responses). This was again true even at 2:1 charge ratio (see
Figure 5). This plasmid has an approximate length of 5000
base pairs, determined on the basis of its comparative migra-
tion with the DNA ladder as reference. This result shows
Figure 6. A) Confocal micrograph of internalization of 15/ODNdendri-
plex after 48 h. B) Image of an isolated cell; white line denotes a section
through the median plane XY. C) Plots of fluorescence emission through
the section: green (fluoresceinated ODN), blue (cell nucleus) and red
(cell membrane).
490
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 483 – 495

Water-Soluble Carbosilane Dendrimers
FULL PAPER
ceived. nG-(SiClx)y and nG-(SiH)x dendrimers were prepared according
to reported methods.[17] Superfect (SF) is an activated PAMAM dendri-
mer with 140 terminal primary amino groups and M=35000 (Qiagen,
Conclusion
New families of amine- and ammonium-terminated carbosi-
lane dendrimers have been synthesized by two strategies
Crawley, UK). G4 PAMAM dendrimer has M=14215 and 64 surface pri-
N
mary amino groups (Aldrich Chemical Co., Milwaukee, WI).
À
1H, 13C, and 29Si NMR spectra were recorded on Varian Unity VXR-300
and Varian 500 Plus Instruments. Chemical shifts (d, ppm) were meas-
ured relative to residual 1H and 13C resonances for CDCl3, [D6]DMSO
and D2O used as solvents, and 29Si chemical shifts were referenced to ex-
ternal SiMe4 (0.00 ppm). C, H and Nanalyses were carried out with a
Perkin-Elmer 240 C microanalyzer. ESI and APCI samples were pre-
pared in acetonitrile or methanol, and spectra were recorded on a Termo-
quest Finnigan Automass Multi Mass Spectrometer. MALDI-TOF MS
and fully characterized. The first entails alcoholysis of Si Cl
bonds by amino alcohols and subsequent quaternization
with MeI. The second involves hydrosilylation of allylamine
À
with Si H bonds of dendrimers and subsequent quaterniza-
tion with HCl. Quaternized carbosilane dendrimers are solu-
ble in water, although degradation is apparent by hydrolysis
of Si O bonds. However, hydrolysis can be prevented or re-
duced, under very dilute conditions, by introduction of a
rigid phenyl group as linker between the amino or ammoni-
um groups and the Si O bonds. Similarly, dendrimers con-
taining Si C bounds proved to be water-stable. For toxicity
À
samples were prepared in
a 1,8,9-trihydroxyanthracene (dithranol)
matrix, and spectra were recorded on a Bruker Reflex II spectrometer
equipped with a nitrogen laser emitting at 337 nm and operated in the re-
flection mode at an accelerating voltage in the range 23–25 kV.
À
À
Synthesis of [3,5-(NMe2CH2CH2O)2](C6H3)CH2OH (I): 2-Chloro-N,N-di-
G
evaluation, distinct (but complementary) approaches were
employed to evaluate membrane integrity, metabolic activi-
ty, apoptosis, morphology and cell movement. This approach
allowed us to obtain a global picture of the potential toxic
effects. Novel quaternized dendrimers of the second genera-
tion show good toxicity profiles in cell cultures over extend-
ed periods. Taking together the toxicity data for PBMCs and
toxicity results for erythrocytes, the most biocompatible
compound proved to be second-generation dendrimer 15.
Remarkably, in spite of their low generation, the carbosilane
dendrimers are able to form complexes with DNA ODNs or
even with plasmids by electrostatic interaction at biocom-
methylethylamine hydrochloride (2.98 g, 20.78 mmol), K2CO3 (6.43 g,
46.75 mmol), KI (1.72 g, 10.39 mmol), and [18]crown-6 (0.54 g, 2 mmol)
were added to an acetone solution (100 mL) of 3,5-dihydroxybenzyl alco-
hol (1.47 g, 10.39 mmol). The reaction mixture was refluxed for 48 h and,
after removal of the solvent, the residue was extracted with CH2Cl2/H2O
(2”50 mL). The organic layer was dried with MgSO4, the resulting solu-
tion evaporated under reduced pressure, and the residue washed with
hexane (2”10 mL) to give I as a pale yellow oil (1.53 g, 50%). 1H N MR
(CDCl3): d=6.48 (m, 2H; C6H3), 6.36 (m, 1H; C6H3), 4.57 (s, 2H;
CH2OH), 3.99 (t, 4H; CH2OC6H3), 2.65 (t, 4H; CH2N), 2.60 (s, 1H;
CH2OH), 2.28 ppm (s, 12H; NMe2); 13C NMR (CDCl3): d=159.9 (C6H3,
Cipso bonded to OCH2CH2NMe2), 143.8 (C6H3, Cipso bonded to CH2OSi),
105.1, 100.6 (C6H3), 65.8 (CH2OC6H3), 64.9 (CH2OH), 58.2 (CH2N),
45.8 ppm (NMe2); elemental analysis calcd (%) for C15H26N2O3: C 63.80,
H 9.28, N9.92; found: C 63.50, H 9.17, N9.83.
À
patible doses. In addition, the presence of Si O bonds in the
Synthesis of 1G-[Si
excess of NEt3 (0.12 mL, 0.87 mmol) and 3,5-(OCH2CH2NMe2)2-
(C6H3)CH2OH (0.22 g, 0.76 mmol) were added to a diethyl ether solution
A
G
dendrimer architecture opens the way to use these dendritic
macromolecules as drug delivery systems via an electrostatic
approach with subsequent release by means of the hydrolyt-
ic process.
AHCTREUNG
(50 mL) of first-generation chloro-terminated dendrimer 1G-Cl4 (0.11 g,
0.19 mmol). The reaction mixture was stirred for 12 h at room tempera-
ture and then evaporated to dryness to remove residual NEt3. The resi-
due was extracted with Et2O (50 mL) and filtered through Celite to
remove the ammonium salt NEt3·HCl. The resulting solution was evapo-
rated under reduced pressure to give 3 as pale yellow oil (0.23 g, 80%).
1H NMR (CDCl3): d=6.45 (m, 8H; C6H3), 6.36 (m, 4H; C6H3), 4.56 (s,
8H; CH2OSi), 3.99 (t, 16H; CH2OC6H3), 2.66 (t, 16H; CH2N), 2.28 (s,
48H; NMe2), 1.33 (m, 8H; SiCH2CH2CH2SiO), 0.68 (m, 8H;
SiCH2CH2CH2SiO), 0.55 (m, 8H; SiCH2CH2CH2SiO), 0.08 ppm (s, 24H;
SiMe2); 13C NMR (CDCl3): d=159.8 (C6H3, Cipso bonded to
OCH2CH2NMe2), 143.2 (C6H3, Cipso bonded to CH2OSi), 104.9, 100.2
(C6H3), 65.9 (CH2OC6H3), 64.6 (CH2OSi), 58.2 (CH2N), 45.8 (NMe2), 21.2
(SiCH2CH2CH2SiO), 17.9, 17.2 (SiCH2CH2CH2SiO), À1.89 (SiMe2);
29Si NMR (CDCl3): d=0.49 (G0-Si), 18.7 ppm (G1-Si); elemental analysis
calcd (%) for C80H148N8O12Si5: C 61.81, H 9.60, N7.21; found C 62.10, H
9.82, N7.30.
These results demonstrate that the new ammonium-termi-
nated carbosilane dendrimers are a good base molecule to
be considered for biomedical applications, such as drug car-
riers, controlled drug liberation (pH- or time-dependant, as
in the case of 1, 2, and 15), as antigen carriers or as vehicles
for nucleic acids. Dendrimers 12 and 18 demonstrated very
stable ODNbinding, and hence they could be used to devel-
op DNA-based devices such as microarrays.
The results of our biocompatibility and preliminary inter-
nalization assays highlight the possibility of developing
transfection and immunomodulation experiments in
PBMCs, and of studying the ability to interfere with the rep-
lication of pathogens such as viruses, bacteria or prions.
These assays are currently being developed in our laborato-
ries and will be the subject of future papers.
Synthesis of 2G-[Si
N
N
drimer was prepared using a similar method to that described for 3, start-
ing from 2G-Cl8 (0.27 g, 0.19 mmol), 3,5-(OCH2CH2NMe2)2C6H3CH2OH
(0.43 g, 1.52 mmol) and NEt3 (0.22 mL, 1.62 mmol) to obtain compound
1
4 as a pale yellow oil (0.54 g, 82%). H NMR (CDCl3): d=6.45 (m, 16H;
C6H3), 6.36 (m, 8H; C6H3), 4.56 (s, 16H; CH2OSi), 3.99 (t, 32H;
CH2OC6H3), 2.67 (t, 32H; CH2N), 2.29 (s, 96H; NMe2), 1.33 (m, 24H;
Experimental Section
SiCH2CH2CH2SiO
and
SiCH2CH2CH2Si),
0.69
(m,
16H;
SiCH2CH2CH2SiO), 0.55 (m, 32H; rest of CH2 bonded to Si), 0.09 (s,
48H; OSiMe2), À0.09 ppm (s, 12H; SiMe); 13C NMR (CDCl3): d=159.0
(C6H3, Cipso bonded to OCH2CH2NMe2), 143.3 (C6H3, Cipso bonded to
CH2OSi), 104.8, 100.1 (C6H3), 65.9 (CH2OC6H3), 64.6 (CH2OSi), 58.3
(CH2N), 45.9 (NMe2), 21.2 (SiCH2CH2CH2SiO), 17.9, 17.2 and overlap-
ping signals (SiCH2CH2CH2SiO and SiCH2CH2CH2Si overlapping), À1.7
(OSiMe2), À4.9 ppm (SiMe); 29Si NMR (CDCl3): d=0.93 (G1-Si),
General remarks: All manipulations of oxygen- or water-sensitive com-
pounds were carried out under an atmosphere of argon using standard
Schlenk techniques or an argon-filled glove box. Solvents were dried and
freshly distilled under argon prior to use: hexane from sodium/potassium,
toluene from sodium, tetrahydrofuran and diethyl ether from sodium
benzophenone ketyl and dichloromethane over P4O10. Unless otherwise
stated, reagents were obtained from commercial sources and used as re-
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
491

R. Gomez, A. MuÇoz-Fernandez, F. J. de la Mata et al.
18.7 ppm (G2-Si), G0-Si was not observed; elemental analysis calcd (%)
for C176H332N16O24Si13: C 61.78, H 9.78, N6.55; found: C 62.51, H 9.90, N
6.75; MALDI-TOF-MS: m/z 3422.1 [M+H]+ (calcd 3422.2).
Synthesis of 1G-[Si{(CH2)3NH2}]4 (9): Allylamine (1.5 mL, 19.99 mmol)
G
and two drops of Speierꢁs catalyst were added to a solution of first-gener-
ation hydrogen-terminated dendrimer 1G-H4 (0.54 g, 1.23 mmol) in the
minimum amount of THF (1 mL). The reaction mixture was heated at
1208C for 4 h and then evaporated to dryness to remove excess allyl-
Synthesis of 3G-[Si
N
N
dendrimer was prepared using a similar method to that described for 3,
starting from 3G-Cl16 (0.07 g, 0.02 mmol), 3,5-(OCH2CH2NMe2)2-
amine. The residue was dissolved in CH2Cl2 (10 mL) and filtered through
Celite and active carbon to remove Pt. The resulting solution was evapo-
rated under reduced pressure to give 9 as a colorless oil (0.23 g, 80%).
1H NMR (CDCl3): d=2.63 (t, 8H; CH2N), 1.34 (m, 16H;
SiCH2CH2CH2Nand SiCH 2CH2CH2Si), 1.13 (s, 8H; NH2), 0.54–0.44 (m,
24H; CH2Si), À0.06 ppm (s, 24H; SiMe2); 13C NMR (CDCl3): d=45.7
(CH2N), 28.3 (Si CH2CH2CH2N), 20.2, 18.6, 17.6, (SiCH2CH2CH2Si) ,12.4
(Si CH2CH2CH2N), À3.3 ppm (SiMe2); 29Si NMR (CDCl3): d=0.50 (G0-
Si), 1.96 ppm (G1-Si); elemental analysis calcd (%) for C32H80N4Si5: C
58.14, H 12.21, N8.48; found: C 57.63, H 12.27, N8.78; ESI MS: m/z
660.52 [M+H]+ (calcd 661.52).
(C6H3)CH2OH (0.10 g, 0.35 mmol) and NEt3 (0.06 mL, 0.43 mmol) to
1
obtain compound 5 as a pale yellow oil (0.11 g, 69%). H NMR (CDCl3):
d=6.46 (m, 32H; C6H3), 6.35 (m, 16H; C6H3), 4.56 (s, 32H; CH2OSi),
3.97 (t, 64H; CH2OC6H3), 2.67 (t, 64H; CH2N), 2.28 (s, 192H; NMe2),
1.33 (m, 56H; SiCH2CH2CH2SiO and SiCH2CH2CH2Si), 0.69 (m, 32H;
SiCH2CH2CH2SiO), 0.55 (m, 80H; rest of CH2 bonded to Si), 0.09 (s,
96H; OSiMe2), À0.09 ppm (s, 36H; SiMe); 13C NMR (CDCl3): d=160.0
(C6H3, Cipso bonded to OCH2CH2NMe2), 143.2 (C6H3, Cipso bonded to
CH2OSi), 104.3, 100.8 (C6H3), 65.9 (CH2OC6H3), 64.6 (CH2OSi), 58.3
(CH2NMe2), 45.8 (NMe2), 21.1 (SiCH2CH2CH2SiO), 17.8, 17.2 and over-
lapping signals (SiCH2CH2CH2SiO and SiCH2CH2CH2Si overlapping),
À1.8 (OSiMe2), À4.9 ppm (SiMe); 29Si NMR (CDCl3): d=0.94 (G1-Si and
G2-Si), 18.6 ppm (G3-Si), G0-Si was not observed; elemental analysis
calcd (%) for C368H700N32O48Si29: C 61.76, H 9.86, N6.26; found: C 62.43,
H 9.90, N6.80.
Synthesis of 2G-[Si{(CH2)3NH2}]8 (10): This dendrimer was prepared
A
using a similar method to that described for 9, starting from 2G-H8
(0.42 g, 0.36 mmol), allylamine (2 mL, 26.7 mmol) and two drops of Spei-
erꢁs catalyst to obtain compound 10 as a colorless oil (0.32 g, 55%).
1H NMR (CDCl3): d=2.63 (t, 16H; CH2N), 1.41–1.28 (m, 40H;
SiCH2CH2CH2Si and SiCH2CH2CH2N), 0.54–0.44 (m, 64H; CH2Si),
À0.06 (s, 48H; SiMe2), À0.10 ppm (s, 12H; SiMe); 13C NMR (CDCl3):
Synthesis of 1G-[Si{O
N
N
was prepared using a similar method to that described for 3, starting
from 1G-Cl4 (0.31 g, 0.54 mmol), 2-{[2-(dimethylamino)ethyl]methylami-
no}ethanol (0.35 mL, 2.16 mmol), and NEt3 (0.5 mL, 3.58 mmol) to
d=45.7
(CH2N),
28.4
(SiCH2CH2CH2N),
20.2,
18.6,
17.6
(SiCH2CH2CH2Si), 12.5 (SiCH2CH2CH2N), À3.1 (SiMe2), À4.8 ppm
(SiMe); 29Si NMR (CDCl3): d=0.92 (G1-Si), 2.00 ppm (G2-Si). The NMR
data were assigned by analogy with the previously prepared model com-
pound Et3SiCH2CH2CH2NH2;[24] elemental analysis calcd (%) for
C80H196N8Si13: C 58.75, H 12.08, N6.85; found: C 58.16, H 11.95, N6.58;
ESI MS: z=1 not observed, z=2: m/z 818.44 [M/2+H]+ (calcd 818.8).
MALDI-TOF-MS: m/z 1636.3 [M+H]+ (calcd 1636.3).
1
obtain compound 6 as a colorless oil (0.3 g, 57%). H NMR (CDCl3): d=
3.64 (t, 8H; CH2O), 2.51 (m, 16H; CH2N(Me)CH2), 2.35 (t, 8H;
CH2NMe2), 2.26 (s, 12H; NMe), 2.19 (s, 24H; NMe2), 1.29 (m, 8H;
SiCH2CH2CH2SiO), 0.62 (m, 8H; SiCH2CH2CH2SiO), 0.59 (m, 8H;
SiCH2CH2CH2SiO), 0.05 ppm (s, 24H; SiMe2); 13C NMR (CDCl3): d=
60.8 (CH2O), 59.9, 56.2 (CH2N(Me)CH2), 57.5 (CH2NMe2), 45.9 (NMe2),
43.3 (NMe), 21.2 (SiCH2CH2CH2SiO), 17.8 (SiCH2CH2CH2SiO), 17.2
(SiCH2CH2CH2SiO), À2.0 ppm (OSiMe2); 29Si NMR (CDCl3): d=0.49
(G0-Si), 17.59 ppm (G1-Si); elemental analysis calcd (%) for
C48H116N8O4Si5: C 57.09, H 11.58, N11.10; found: C 57.60, H 11.72, N
11.20.
Synthesis of 1G-[Si
N
N
solution of MeI in Et2O (2m, 0.35 mL, 0.70 mmol) was added to a diethyl
ether (10 mL) solution of 3 (0.10 g, 0.06 mmol). The resulting solution
was stirred for 48 h at room temperature and then evaporated under re-
duced pressure to remove residual MeI. The residue was washed with
Et2O (2”5 mL) and dried under vacuum to give 11 as a white solid
(0.14 g, 90%). 1H NMR ([D6]DMSO): d=6.55 (m, 12H; C6H3), 4.58 (s,
8H; CH2OSi), 4.42 (t, 16H; CH2OC6H3), 3.77 (t, 16H; CH2N), 3.18 (s,
72H; NMe3+), 1.35 (m, 8H; SiCH2CH2CH2SiO), 0.70 (m, 8H;
SiCH2CH2CH2SiO), 0.57 (m, 8H; SiCH2CH2CH2SiO), 0.08 ppm (s, 24H;
SiMe2); 13C NMR ([D6]DMSO): d=157.8 (C6H3, Cipso bonded to
OCH2CH2NMe2), 143.2 (C6H3, Cipso bonded to CH2OSi), 104.9, 99.6
(C6H3), 63.5 (CH2NMe2), 63.0 (CH2OSi), 61.3 (CH2OC6H3), 52.7 (NMe3+),
19.9 (SiCH2CH2CH2SiO), 16.9, 16.1 (SiCH2CH2CH2SiO), À2.4 ppm
(SiMe2); elemental analysis calcd (%) for C88H172I8N8O12Si5: C 39.29, H
6.44, N4.17; found: C 38.90, H 6.24, N4.09.
Synthesis of 2G-[Si{O
N
N
was prepared using a similar method to that described for 3, starting
from 2G-Cl8 (1.12 g, 0.77 mmol), 2-{[2-(dimethylamino)ethyl]methylami-
no}ethanol (1 mL, 6.16 mmol), and NEt3 (1 mL, 7.17 mmol) to obtain
1
compound 7 as a pale yellow oil (1.3 g, 72%). H NMR (CDCl3): d=3.65
(t, 16H; CH2O), 2.51 (m, 32H; CH2N(Me)CH2), 2.30 (t, 16H;
CH2NMe2), 2.26 (s, 24H; NMe), 2.20 (s, 48H; NMe2), 1.31 (m, 24H;
SiCH2CH2CH2SiO
and
SiCH2CH2CH2Si),
0.63
(m,
16H;
SiCH2CH2CH2SiO), 0.59 (m, 32H; SiCH2), 0.06 (s, 48H; SiMe2),
À0.10 ppm (s, 12H; SiMe); 13C NMR (CDCl3): d=60.8 (CH2O), 59.9,
56.2 (CH2N(Me)CH2), 57.5 (CH2NMe2), 45.9 (NMe2), 43.3 (NMe), 21.2
(CH2SiO), 18.7–17.9 (SiCH2CH2CH2SiO and SiCH2CH2CH2Si overlap-
ping), À1.79 (OSiMe2), À4.8 ppm (SiMe); 29Si NMR (CDCl3): d=0.38
(G0-Si), 0.93 (G1-Si), 17.58 ppm (G2-Si); elemental analysis calcd (%) for
Synthesis of 2G-[Si
N
N
dendrimer was prepared by a similar method to that described for 11,
starting from 4 (0.08 g, 0.023 mmol) and a solution of MeI in Et2O(2m,
0.30 mL, 0.6 mmol). Compound 12 was isolated as a white solid (0.11 g,
85%). 1H NMR ([D6]DMSO): d=6.58 (m, 24H; C6H3), 4.58 (s, 24H;
CH2OSi), 4.44 (t, 32H; CH2OC6H3), 3.79 (t, 32H; CH2N), 3.20 (s, 144H;
NMe3+), 1.33 (m, 24H; SiCH2CH2CH2SiO and SiCH2CH2CH2Si overlap-
ping), 0.67 (m, 24H; SiCH2CH2CH2SiO), 0.54 (m, 32H; rest of CH2
bonded to Si), 0.07 (s, 48H; OSiMe2), À0.07 ppm (s, 12H; SiMe);
C
112H226N16O8Si13: C 57.67, H 11.58, N9.61; found: C 57.20, H 11.40, N
9.52; MALDI-TOF-MS: m/z 2332.8 [M+H]+ (calcd 2332.8).
Synthesis of 3G-[Si{O(CH2)2N(Me)(CH2)2NMe2}]16 (8): This dendrimer
A
ACHTREUNG
was prepared using a similar method to that described for 3, starting
from 3G-Cl16 (0.49 g, 0.15 mmol), 2-{[2-(dimethylamino)ethyl]methylami-
no}ethanol (0.39 mL, 2.43 mmol) and NEt3 (0.40 mL, 2.86 mmol) to
13C NMR
([D6]DMSO):
d=157.9
(C6H3,
Cipso
bonded
to
1
obtain compound 8 as a pale yellow oil (0.51 g, 67%). H NMR (CDCl3):
OCH2CH2NMe2), 143.2 (C6H3, Cipso bonded to CH2OSi), 105.2, 99.6
(C6H3), 63.5 (CH2NMe2), 63.0 (CH2OSi), 61.4 (CH2OC6H3); 52.7 (NMe3+),
19.8 (SiCH2CH2CH2SiO), 17.5, 16.8 and overlapping signals
(SiCH2CH2CH2SiO and SiCH2CH2CH2Si overlapping), À2.4 (OSiMe2),
d=3.65 (t, 32H; CH2O), 2.51 (m, 64H; CH2N(Me)CH2), 2.36 (t, 32H;
CH2NMe2), 2.26 (s, 48H; NMe), 2.21 (s, 96H; NMe2), 1.30 (m, 56H;
SiCH2CH2CH2SiO
and
SiCH2CH2CH2Si),
0.63
(m,
32H;
SiCH2CH2CH2SiO), 0.53 (m, 80H; rest of SiCH2), 0.06 (s, 96H; SiMe2),
À0.10 ppm (s, 36H; SiMe); 13C NMR (CDCl3): d=60.8 (CH2O), 60.0,
56.2 (CH2N(Me)CH2), 57.4 (CH2NMe2), 45.9 (NMe2), 43.3 (NMe), 21.1
(CH2SiO), 18.7–17.8 (SiCH2CH2CH2SiO and SiCH2CH2CH2Si overlap-
ping), À1.9 (OSiMe2), À4.8 ppm (SiMe); 29Si NMR (CDCl3): d=0.93
(G1-Si and G2-Si), 17.58 ppm (G3-Si), G0-Si not observed; elemental anal-
ysis calcd (%) for C240H572N32O16Si29: C 57.91, H 11.58, N9.00; found: C
57.32, H 11.38, N8.72.
À5.5 ppm (SiMe); elemental analysis calcd (%) for C192H380I16N16O24Si13
:
C 40.51, H 6.73, N3.94; found: C 41.20, H 7.02, N4.10.
Reaction of 3G-[Si
N
N
MeI: The reaction of 5 with MeI by a similar method to that described
+
for 11 afforded the dendrimer 3G-[Si
N
G
IÀ)2}]16 (13). However, the NMR analysis revealed that not all the amino
groups were quaternized: roughly 85% of the groups were methylated
492
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 483 – 495

Water-Soluble Carbosilane Dendrimers
FULL PAPER
based on integration of the corresponding signal of the outer
sense (complementary) sequence of the HIV polypurine tract (PPT) ele-
ment mRNA: 5’-fluoresceine-AAT TTT CTT TTC CCC CCT-3’. For
treatment of PBMCs and oligonucleotide synthesis, see Supporting Infor-
mation.
À
À
OCH2CH2N branch of the amino or ammonium groups. The NMR
data of the quaternized branches are analogous to those given for den-
drimer 12.
Synthesis of 1G-[Si{O
C
G
Formation of ODN/dendrimer complexes: Complex formation between
dendrimers and ODNs was performed by an electrostatic approach.
Ratios of ODNto dendrimer were based on the calculation of the elec-
trostatic charge present on each component, for example, the number of
phosphate groups in the ODNversus the number of terminal ammonium
groups on the dendrimer. Dendrimers were diluted in sterile distilled
water at 2 mgmLÀ1. The ODNconcentration for complexes with carbosi-
lane dendrimers was 0.88 mm (2.57 mg), and for complexes with SF
0.34 mm (1 mg); the concentrations of the dendrimers were 3.93 mm
(2.96 mg) for 1, 1.99 mm (2.35 mg) for 2, 1.96 mm (2.80 mg) for 12, 3.94 mm
(3.42 mg) for 15, 3.98 mm (1.92 mg) for 18 and 0.68 mm for SF (final concen-
tration in well). All complexes were formed in 60 mL of serum-free
RPMI medium, with an incubation time of 20 min at room temperature.
The concentration of DNA and SF in the complex was chosen according
to the manufacturers instructions. In the same way, complexes between
ODNand SF were formed according to these instructions.
drimer was prepared using a similar method to that described for 11,
starting from 6 (0.043 g, 0.047 mmol) and 0.095 mL of a 2m solution in
Et2O of MeI (0.19 mmol). Compound 14 was isolated as a white solid
(0.49 g, 95%). 1H NMR ([D6]DMSO): d=3.61 (t, 8H; OCH2), 3.47 (t,
8H; CH2NMe3+), 3.11 (s, 36H; NMe3+), 2.76 (t, 8H;
N(Me)CH2CH2NMe3+IÀ), 2.49 (t, 8H; OCH2CH2N), 2.22 (s, 12H; NMe),
1.30 (m, 8H; SiCH2CH2CH2SiO), 0.61 (m, 8H; SiCH2CH2CH2SiO), 0.53
(m, 8H; SiCH2CH2CH2SiO), 0.04 ppm (s, 24H; OSiMe2); 13C N MR
([D6]DMSO): d=61.1, 59.5, 58.3, 50.9 (methylene groups of
OCH2CH2N(Me)CH2CH2NMe3+), 52.2 (NMe3+), 41.3 (NMe), 20.0
(SiCH2CH2CH2SiO), 16.9, 16.1 (SiCH2CH2CH2SiO), À2.5 ppm (OSiMe2).
Synthesis of 2G-[Si{O
drimer was prepared by a similar method to that described for 11, start-
ing from 7 (0.19 g, 0.08 mmol) and a solution of MeI in Et2O (2m,
0.34 mL, 0.64 mmol). Compound 15 was isolated as a white solid (0.49 g,
95%). 1H NMR ([D6]DMSO): d=3.60 (t, 16H; OCH2), 3.42 (t, 16H;
CH2NMe3+), 3.11 (s, 72H; NMe3+), 2.76 (t, 16H; N(Me)CH2CH2NMe3+),
2.49 (t, 16H; OCH2CH2N), 2.22 (s, 24H; NMe) 1.30 (m, 24H;
Evaluation of ODN/dendrimer complex formation: Complex formation
was assessed by evaluation of migration retardation of fluoresceinated
ODNs or alternatively NFkappa-B plasmid during electrophoresis on
3% agarose gels. A 100 or 5000 bp DNA ladder was used respectively as
reference (Gibco BRL).
SiCH2CH2CH2SiO
and
SiCH2CH2CH2SiO), 0.51 (m, 32H; rest of CH2Si groups), 0.03 (s, 48H;
OSiMe2), À0.11 ppm (s, 12H; SiMe); 13C NMR ([D6]DMSO): d=61.0
(CH2NMe3+), 59.6 (OCH2), 58.3 (NCH2CH2O), 52.3 (NMe3+), 50.9
(NCH2CH2NMe3+), 41.3 (NMe), 20.0–16.8 (CH2 groups of the carbosi-
lane skeleton), À2.5 (OSiMe2), À5.5 ppm (SiMe).
pH gradient: To check stability of dendriplexes at different pH values,
we employed different phosphate and acetate buffer solutions. For acid
extreme of pH 2.8 an acidic solution (0.05m glycine/HCl/0.1m NaCl) was
used.
Synthesis of 3G-[Si{O
A
Phase-contrast light microscopy: After incubation with dendrimers,
changes in morphology and characteristics of PBMCs, such as cell mem-
brane birefringence, were observed through a phase-contrast inverted mi-
croscope (Nikon TMS, Nikon, Japan) equipped with a 100X objective
(Plan 10/0.30DL/Ph1, Nikon, Japan). Live PBMCs are bright, with a de-
fined spherical shape, and float in the culture medium. Dead cells have a
darker appearance and are mostly present in the bottom of the well. In
addition, we assessed the presence or absence of cell aggregation.
drimer was prepared by a similar method to that described for 11, start-
ing from 8 (0.084 g, 0.017 mmol) and a solution of MeI in Et2O (2m,
0.13 mL, 0.27 mmol). Compound 16 was isolated as a white solid (0.49 g,
95%). 1H NMR ([D6]DMSO): d=3.60 (t, 32H; OCH2), 3.44 (t, 32H;
CH2NMe3+), 3.12 (s, 144H; NMe3+), 2.76 (t, 32H; N(Me)CH2CH2NMe3+),
2.49 (t, 32H; OCH2CH2N), 2.22 (s, 48H; NMe), 1.28 (m, 56H;
SiCH2CH2CH2SiO
and
SiCH2CH2CH2SiO and SiCH2CH2CH2Si), 0.03 (s, 96H; OSiMe2),
À0.11 ppm (s, 36H; SiMe); 13C NMR ([D6]DMSO): d=61.0 (CH2NMe3+),
59.6 (OCH2), 58.3 (NCH2CH2O), 52.3 (NMe3+), 50.8 (NCH2CH2NMe3+),
41.3 (NMe), 20.0–16.8 (CH2 groups of the carbosilane skeleton), À2.5
(OSiMe2), À5.5 ppm (SiMe).
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)
assay: This method was selected to analyze detrimental intracellular ef-
fects on mitochondria and metabolic activity. The colorimetric MTT test,
based on the selective ability of viable cells to reduce MTT to purple for-
mazan, relies on intact metabolic activity and is frequently used for cyto-
toxicity screening. After 48 h of incubation of PBMCs with different con-
centrations of dendrimers in a 96-well plate, culture medium containing
the dendrimers was replaced with 200 mL of serum-free Optimem. 20 mL
of sterile, filtered MTT (Sigma) stock solution in PBS (pH 7.4,
5 mgmLÀ1) was added to each well to achieve a final concentration of
0.5 mg MTT per millilitre. After 4 h, unconverted dye was removed by
aspiration and the formazan crystals were dissolved in dimethyl sulfoxide
(200 mL per well; Merck, Darmstadt, Germany). The concentration of
formazan was then determined spectrophotometrically in a plate reader
at a wavelength of 570 nm (test) and 690 nm (reference). The spectropho-
tometer was calibrated to zero absorbance by using Optimem medium
without cells. The percentage cell viability relative to control wells (cells
with no dendrimer) was calculated by ([A]test/[A]control)”100. Each den-
drimer concentration was tested in triplicate, according to American
Type Culture Collection (ATCC) directives.
Synthesis of 1G-[Si
{(CH2)3NH3+ClÀ}]4 (17): A solution of HCl in Et2O
A
(1m, 1.2 mL, 1.2 mmol) was added to a diethyl ether (40 mL) solution of
9 (0.17 g, 0.26 mmol). The resulting solution was stirred for 2 h at room
temperature and then evaporated under reduced pressure to give 17 as a
white solid in a quantitative yield. 1H NMR (D2O): d=2.74 (t, 8H;
CH2N), 1.45 (m, 8H; SiCH2CH2CH2N), 1.19 (m, 8H; SiCH2CH2CH2Si),
0.38 (m, 24H; CH2Si), À0.19 ppm (s, 24H; SiMe2); 13C NMR (D2O): d=
42.0 (CH2N), 21.3 (SiCH2CH2CH2N), 18.8, 18.0, 16.6 (SiCH2CH2CH2Si),
11.2 (SiCH2CH2CH2N), À4.4 ppm (SiMe2); elemental analysis calcd (%)
for C32H84N4Cl4Si5: C 47.61, H 10.49, N6.94; found: C 48.57, H 10.46, N
6.82.
Synthesis of 2G-[Si
{(CH2)3NH3+ClÀ}]8 (18): This dendrimer was prepared
G
by a similar method to that described for 17, starting from 10 (0.09 g,
0.05 mmol) and a solution of HCl in Et2O (1m, 0.6 mL, 0.6 mmol). Com-
pound 18 was isolated as a white solid (0.06 g, 55%). 1H NMR (D2O):
d=2.74 (t, 16H; CH2N), 1.47 (m, 16H; SiCH2CH2CH2N), 1.16 (m, 24H;
SiCH2CH2CH2Si), 0.39 (m, 64H; CH2Si), À0.16 (s, 48H; SiMe2),
À0.25 ppm (s, 12H; SiMe); 13C NMR (D2O): d=42.6 (CH2N), 21.8
(SiCH2CH2CH2N), 20.5–17.0 (SiCH2CH2CH2Si), 12.1 (SiCH2CH2CH2N),
À3.23 (SiMe2), À4.17 ppm (SiMe); elemental analysis calcd (%) for
C80H204N8Cl8Si13: C 49.86, H 10.67, N5.81; found: C 49.79, H 10.18, N
5.76.
Hemolysis test: The hemolytic and hemagglutinating activity of the car-
bosilane dendrimers was evaluated according to Parnham and Wetzig[25]
and compared with that induced by a 4G PAMAM dendrimer. Erythro-
cytes were obtained from the bottom of the tube after PBMC extraction
following blood centrifugation in Ficoll gradient. Erythrocytes were dilut-
ed with cold PBS (pH 7.4) to a convenient volume to make feasible their
visualisation. This suspension of red blood cells was always freshly pre-
pared and used within 24 h after collection. Carbosilane dendrimer solu-
tions of different concentrations, also prepared in PBS buffer, were
added to erythrocytes and were incubated for 60 min at 378C in a shak-
ing water bath. The presence/absence of hemagglutination was observed
PBMCs andoligonucleoteid
: Peripheral blood mononuclear cells
(PBMCs) were derived from healthy voluntary donors, and obtained
from leukophoresed blood by Ficoll gradient and elutriation centrifuga-
tion. The ODNsequence was 18 bases long and corresponded to an anti-
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
493

R. Gomez, A. MuÇoz-Fernandez, F. J. de la Mata et al.
under a phase-contrast inverted microscope. In a second step, the release
of hemoglobin was determined after centrifugation (1500 rpm) by photo-
P. W. N. M. van Leeuwen, Angew. Chem. 2001, 113, 1878–1901;
Angew. Chem. Int. Ed. 2001, 40, 1828–1849; e) D. Astruc, F. Char-
dac, Chem. Rev. 2001, 101, 2991–3023; f) S. M. Grayson, J. M. J. FrØ-
chet, Chem. Rev. 2001, 101, 3819–3867; g) S. E. Stiriba, H. Frey, R.
Haag, Angew. Chem. 2002, 114, 1385–1390; Angew. Chem. Int. Ed.
2002, 41, 1329–1334; h) F. Aulenta, W. Hayes, S. Rannard, Eur.
Polym. J. 2003, 39, 1741–1771; i) U. Boas, P. M. H. Heegaard, Chem.
Soc. Rev. 2004, 33, 43–63; j) S. Svenson, D. A. Tomalia, Adv. Drug
Delivery Rev. 2005, 57, 2106–2129.
metric analysis of the supernatant at 540 nm. Complete hemolysis was
A
achieved with 0.2% Triton X-100 to give the 100% control value. Less
than 10% hemolysis was regarded as no-toxic-effect level in our experi-
ments. The experiments were run in triplicate and were repeated twice.
Lymphoproliferative assay: PBMCs were incubated for one week with
two different concentrations of each dendrimer in
a 96-well plate
(100000 cells per well seeded in 200 mL of complete RPMI medium with
antibiotics, glutamine, and 10% of human AB serum). A well with un-
treated cells was included along with a positive control for proliferation
(cells treated with 1 mgmLÀ1 of phytohemagglutinin). PBMC prolifera-
tion was evaluated by incorporation of [3H]thymidine into DNA during
the last 16 h of culture. The cells were pulsed with 1 mCi of [3H]thymidine
and harvested in glass-fiber filters by using an automatic cell harvester,
and radioactivity incorporation was measured in a liquid scintillation
spectrometer. The assay was carried out for triplicate cultures.
[2] S. Hacein-Bey-Abina, C. von Kalle, M. Schmidt, F. Le Deist, N.
Wulffraat, E. McIntyre, I. Radford, J. L. Villeval, C. C. Fraser, M.
Cavazzana-Calvo, A. Fischer, N. Engl. J. Med. 2003, 348, 255–256.
[3] P. L. Felgner, T. R. Gadek, M. Holm, R. Roman, H. W. Chan, M.
Wenz, J. P. Northrop, G. M. Ringold, M. Danielsen, Proc. Natl.
Acad. Sci. USA 1987, 84, 7413–7417.
[4] G McLachlan, B. J. Stevenson, D. J. Davidson, D. J. Porteous, Gene
Ther. 2000, 7, 384–392.
[5] T. V. Chirila, P. E. Rakoczy, K. L. Garrett, X. Lou, I. J. Constable,
Biomaterials 2002, 23, 321–342.
Flow cytometry (FC): Dead or dying cells were identified by their typical
diminished forward (FW) and increased side (SD) light-scattering charac-
teristics. In the FC analysis of PBMCs treated with the different dendri-
[6] J. Haensler, F. C. Szoka, Jr., Bioconjugate Chem. 1993, 4, 372–379.
[7] For some examples, see: a) A. U. Bielinska, C. Chen, J. Johnson,
J. R. Baker, Jr., Bioconjugate Chem. 1999, 10, 843–850; b) J.
Dennig, E. Duncan, Rev. Mol. Biotechnol. 2002, 90, 339–347; c) J.
Dennig, Top. Curr. Chem. 2003, 228, 227–236.
A
to live cells and another one around cells showing FW and SD of dead or
dying cells. We then compared the percentage of cells included in the
two gates. Cells with FW and SD corresponding to monocytes (less than
5%) were not included in the analysis. FC was performed in a Beck-
mann–Coulter flow cytometer.
[8] M. X. Tang, C. T. Redemann, F. C. Szoka, Jr., Bioconjugate Chem.
1996, 7, 703–714.
[9] a) C. Loup, M. A. Zanta, A. M. Caminade, J. P. Majoral, B. Meunier,
Chem. Eur. J. 1999, 5, 3644–3650; b) A. M. Caminade, J. P. Majoral,
Progr. Polym. Sci. 2005, 30, 491–505.
[10] B. H. Zinselmeyer, S. P. Mackay, A. G. Schatzlein, I. F. Uchegbu,
Pharm. Res. 2002, 19, 960–967.
Trypan Blue (TB) uptake: TB dye is excluded by viable cells but can
penetrate cell membranes of dying or dead cells. When TB staining is
negative, membrane integrity is present. Cells were treated with 0.6%
TB (Sigma) for 5 min and then washed twice with PBS. At least 200 cells
were counted under the microscope for each condition.
[11] a) T. Niidome, M. Wakamatsu, A. Wada, T. Hirayama, H. Aoyagi, J.
Pept. Sci. 2000, 6, 271–279; b) M. Ohsaki, T. Okuda, A. Wada, T.
Hirayama, T. Nidome, H. Aoyagi, Bioconjugate Chem. 2002, 13,
510–517.
[12] N. Malik, R. Wiwattanapatapee, K. Klopsch, K. Lorenz, H. Frey,
J. W. Weener, E. W. Meijer, W. Paulus, R. Duncan, J. Controlled Re-
lease 2000, 65, 133–148.
[13] S. W. Krska, D. Seyferth, J. Am. Chem. Soc. 1998, 120, 3604–3612.
[14] B. Lühmann, H. Lang, K. Brüning, Phosphorus Sulfur Silicon Relat.
Elem. 2001, 168, 481–484.
[15] A. W. Kleij, R. van de Coevering, R. J. M. Klein Gebbink, A. M.
Noordman, A. L. Spek, G. van Koten, Chem. Eur. J. 2001, 7, 181–
192.
Microscopy
DAPI labeling: Cells were seeded on glass slides coated with Poly-l-
lysine, fixed with 4% paraformaldehyde and treated with DAPI for
10 min. They were then washed three times with PBS and observed
under a Leica TCS SP2 confocal microscope with excitation at 405 nm.
Time-lapse in vivo imaging: After 72 h of incubation with dendrimers or
dendriplexes, cells were seeded in special chambers for in vivo microsco-
py at 378C and 5% CO2. One image was captured every 30 s over a
period of 10 min. A time-lapse video was prepared from the images, and
cell viability and motility were examined.
[16] P. Ortega, J. F. Bermejo, L. Chonco, E. de Jesffls, F. J. de la Mata, G.
Fernµndez, J. C. Flores, R. Gómez, M. J. Serramía, M. A. MuÇoz-
Fernµndez, Eur. J. Inorg. Chem. 2005, 1388–1396.
Acknowledgements
We thank the Ministerio de Ciencia y Tecnología (Project CTQ2005-
00795/BQU), the DGI-Comunidad de Madrid (Project GR/MAT/0733/
2004), the Plan Nacional de Salud (grant SAF-2004-06778, SAF-2003-
09209), the Red Temµtica Cooperativa de investigación en sida y genØtica
(grant RIS G03/173 and grant RIG C03/07, respectively) of Fondos de
Investigación Sanitaria (FIS), FIPSE (36514/05), the Dirección General
de Investigación Científica y TØcnica (grant BQU2004-02048), and Fun-
dació LA CAIXA (BM04-52-0) for financial support. J.F.B. is supported
by a grant of Fondos de Investigación Sanitaria Madrid (CM04/00136).
The Leica TCS SP2 confocal microscope was acquired with the grants
donated by the “Fondo de Investigaciones Sanitarias” to the “Fundación
para la Investigación del Hospital Gregorio MaraÇón”, Madrid. R.S. is
supported by “Fondo de Investigaciones Sanitarias” (FIS-CA05/0043).
[17] a) A. W. van der Made, P. W. N. M. van Leeuwen, J. Chem. Soc.
Chem. Commun. 1992, 1400–1401; b) A. W. van der Made,
P. W. N. M. van Leeuwen, J. C. de Wilde, R. A. C. Brandes, Adv.
Mater. 1993, 5, 466–468; c) L. L. Zhou, J. Roovers, Macromolecules
1993, 26, 963–968; d) D. Seyferth, D. Y. Son, A. L. Rheingold, R. L.
Ostrander, Organometallics 1994, 13, 2682–2690; e) I. Cuadrado, M.
Morµn, J. Losada, C. M. Casado, C. Pascual, B. Alonso, F. Lobete in
Advances in Dendritic Macromolecules, Vol. 3 (Ed.: G. R. New-
kome), JAI press Inc., Greenwich, CT, 1996, pp. 151–195.
[18] a) J. L. Speier, J. A. Webster, G. H. Barnes, J. Am. Chem. Soc. 1957,
79, 974–979; b) R. A. Benkeser, J. Kang, J. Organomet. Chem. 1980,
185, C9-C12.
[19] a) R. Jevprasesphant, J. Penny, R. Jalal, D. Attwood, N. B.
McKeown, A. D’Emanuele, Int. J. Pharm. 2003, 252, 263–266; b) D.
Fischer, Y. Li, B. Ahlemeyer, J. Krieglstein, T. Kissel, Biomaterials
2003, 24, 1121–1131.
[20] M. Maszewska, J. Leclaire, M. Cieslak, B. Nawrot, A. Okruszek,
A. M. Caminade, J. P. Majoral, Oligonucleotides 2003, 13, 193–205.
[21] a) A. Bielinska, J. F. Kukowaska-Lantallo, J. Johnson, D. A. Tomalia,
J. R. Baker, Jr., Nucleic Acids Res. 1996, 24, 2176–2182; b) H. Yoo,
R. L. Juliano, Nucleic Acids Res. 2000, 28, 4225–4231.
[1] a) M. Fischer, F. Vçgtle, Angew. Chem. 1999, 111, 934–955; Angew.
Chem. Int. Ed. 1999, 38, 884–905, and references therein. b) Den-
drimers and other Dendritic Polymers (Eds.: J. M. FrØchet, D. A.
Tomalia), Wiley Series in Polymer Science, Wiley, New York, 2001;
c) Dendrimers and Dendrons: Concepts, Syntheses, Applications
(Eds.: G. R. Newkome, C. N. Moorefield, F. Vçgtle), Wiley-VCH,
Weinheim, 2001; d) G. E. Ossterom, J. N. H. Reek, P. C. J. Kamer,
494
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 483 – 495

Water-Soluble Carbosilane Dendrimers
FULL PAPER
[22] M. A. Kostiainen, J. G. Hardy, D. K. Smith, Angew. Chem. 2005,
117, 2612–2615; Angew. Chem. Int. Ed. 2005, 44, 2556–2559.
[23] R. Gçller, J. P. Vors, A. M. Caminade, J.-P. Mayoral, Tetrahedron
Lett. 2001, 42, 3587–3590.
0.46 ppm (m, 8H; CH2Si and CH2CH3); 13C NMR (CDCl3): d=45.6
(CH2N), 27.9 (SiCH2CH2CH2N), 8.3 (CH2Si), 7.4 (CH2CH3), 3.2 ppm
(CH2CH3).
[25] M. J. Parnham, H. Wetzig, Chem. Phys. Lipids 1993, 64, 263–274.
[24] Synthesis of Et3Si
A
in the presence of Speier catalyst. Selected data for Et3Si-
1
A
Received: April 28, 2006
Published online: September 27, 2006
Chem. Eur. J. 2007, 13, 483 – 495
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
495