Macrocyclic nonviral vectors: High cell transfection efficiency and low toxicity in a lower rim guanidinium calix[4]arene

Article abstract of DOI:10.1021/ol801326d

(Figure Presented) New multivalent cationic lipids, one of them showing high efficiency and low toxicity in cell transfection, have been obtained by attaching guanidinium groups at the lower rim of calix[4]arenes.

Full text of DOI:10.1021/ol801326d

ORGANIC
LETTERS
2008
Vol. 10, No. 18
3953-3956
Macrocyclic Nonviral Vectors: High Cell
Transfection Efficiency and Low Toxicity
in a Lower Rim Guanidinium
Calix[4]arene
Valentina Bagnacani,† Francesco Sansone,*,† Gaetano Donofrio,‡ Laura Baldini,†
Alessandro Casnati,† and Rocco Ungaro†
Dipartimento di Chimica Organica e Industriale, UniVersita` di Parma, V.le G. P. Usberti
17/a, 43100 Parma, Italy and Consorzio INSTM, Via Giusti 9, 50121 Firenze, Italy, and
Dipartimento di Salute Animale, UniVersita` di Parma, Via del Taglio 8,
43100 Parma, Italy
francesco.sansone@unipr.it
Received June 11, 2008
ABSTRACT
New multivalent cationic lipids, one of them showing high efficiency and low toxicity in cell transfection, have been obtained by attaching
guanidinium groups at the lower rim of calix[4]arenes.
Since the first report in 1987,1 cell transfection mediated by
cationic lipids (Lipofection) has become a very useful
methodology for inserting therapeutic DNA into cells, which
is an essential step in Gene Therapy.2 Among the numerous
variants of this procedure reported in the literature, the most
promising for both in vitro and in vivo applications is
considered the combination of a cationic lipid (cytofectin)
with a neutral, biologically available colipid (called “helper
lipid”) since it allows a good balance between efficiency and
toxicity.3 Several scaffolds have been used for the synthesis
of cationic lipids, and they include polymers,4 dendrimers,5
nanoparticles,6 Gemini surfactants,7 and, more recently,
macrocycles.8 It is well-known that oligoguanidinium com-
pounds (polyarginines and their mimics, guanidinium-
modified aminoglycosides, etc.) efficiently penetrate cells,
(3) (a) Aissaoui, A.; Oudrhiri, N.; Petit, L.; Hauchecorne, M.; Kan, E.;
Sainlos, M.; Julia, S.; Navarro, J.; Vigneron, J.-P.; Lehn, J.-M.; Lehn, P.
Curr. Drug Targets 2002, 3, 1–16. (b) Dass, C. R. J. Mol. Med. 2004, 82,
579–591. (c) Fletcher, S.; Ahmad, A.; Price, W. S.; Jorgensen, M. R.; Miller,
A. D. ChemBioChem 2008, 9, 455–463.
† Dipartimento di Chimica Organica e Industriale.
(4) (a) Haag, R. Angew. Chem., Int. Ed. 2004, 43, 278–282. (b) Kichler,
A. J. Gene Med. 2004, 6, S3–S10. (c) Li, H.-Y.; Birchall, J. Pharm. Res.
2006, 23, 941–950.
‡ Dipartimento di Salute Animale.
(1) Felgner, P. L.; Gadek, T. R.; Holm, M.; Roman, R.; Chan, H. W.;
Wenz, M.; Northrop, J. P.; Ringold, G. M.; Danielsen, M. Proc. Natl. Acad.
Sci. U.S.A. 1987, 84, 7413–7417.
(5) (a) Tziveleka, L.-A.; Psarra, A.-M. G.; Tsiourvas, D.; Paleos, C. M.
J. Controlled Release 2007, 117, 137–146. (b) Guillot-Nieckowski, M.;
Eisler, S.; Diederich, F. New J. Chem. 2007, 31, 1111–1127.
(6) (a) Thomas, M.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A. 2003,
100, 9138–9143. (b) Dobson, J. Gene Ther. 2006, 13, 283–287. (c) Eaton,
P.; Ragusa, A.; Clavel, C.; Rojas, C. T.; Graham, P.; Dura´n, R. V.; Penade´s,
S. IEEE T. Nanobiosci. 2007, 6, 309–317.
(2) (a) Zabner, J. AdV. Drug. DeliVery ReV. 1997, 27, 17–28. (b) Goun,
E. A.; Pillow, T. H.; Jones, L. R.; Rothbard, J. B.; Wender, P. A.
ChemBioChem 2006, 7, 1497–1515. (c) Wasungu, L.; Hoekstra, D. J.
Controlled Release 2006, 116, 255–264. (d) Pietersz, G. A.; Tang, C.-K.;
Apostolopoulos, V. Mini-ReV. Med. Chem. 2006, 6, 1285–1298.
10.1021/ol801326d CCC: $40.75
Published on Web 08/23/2008
2008 American Chemical Society

delivering a large variety of cargos.2b,9 We have previously
reported8c that calix[n]arenes bearing guanidinium groups
directly attached on the aromatic nuclei (upper rim) are able
to condense plasmid and linear DNA and perform cell
transfection in a way which is strongly dependent on the
macrocycle size, lipophilicity, and conformation. Unfortu-
nately, these compounds are characterized by low transfection
efficiency and high cytotoxicity especially at the vector
concentration required for observing cell transfection (10-20
µM), even in the presence of the helper lipid DOPE
(dioleoylphosphatidylethanolamine).2c,10
Interestingly, we have now found that attaching guani-
dinium moieties at the phenolic OH groups (lower rim) of
the calix[4]arene through a three carbon atom spacer results
in a new class of cytofectins. One member of this family
(1b), when formulated with DOPE, performs cell transfection
quite efficiently and with very low toxicity, surpassing a
commercial lipofectin widely used for gene delivery. We
report in this communication the basic features of this new
class of cationic lipids, in comparison with a nonmacrocyclic
(Gemini-type) model compound (2).
We synthesized (Scheme 1 in the article and Schemes S1
and S2 in the Supporting Information) three new lower rim
guanidinium calix[4]arenes 1a-c in the fixed cone confor-
mation and compound 2, which is a linear model of
compound 1b. All compounds are water soluble in the range
1 × 10-3-7 × 10-2 M, 1c being the least and Gemini 2 the
most soluble. The 1H NMR spectra of compounds 1a,b and
2 in D2O (see Figures S1, S3 and S7 in Supporting
Information) at room temperature are sharp, whereas broad-
ening occurs for 1c (see Figure S5 in Supporting Informa-
tion), indicating in the latter case significant aggregation in
water. However, at the very low concentration (e10-5 M)
used in the following studies, no evidence for aggregation
was obtained for all the investigated compounds.
The ability of compounds 1a-c and 2 to bind plasmid
DNA pEGFP-C1 (4731 bp) was assessed through gel
electrophoresis (Figure 1) and ethidium bromide displace-
Scheme 1. Synthesis of the Lower Rim Guanidinium
Calix[4]arenes 1a-c and Molecular Formula of the
Nonmacrocyclic Compound 2
Figure 1. Electrophoresis mobility shift assays performed with
plasmid DNA (pEGFP-C1) (25 nM) incubated with guanidinium
calix[4]arenes 1a, 1b, and 1c and compound 2 at increasing
concentration (indicated above the gel). The control is plasmid
without ligand.
ment assays11 (Figure S9 in Supporting Information). Both
experiments evidenced that the macrocyclic derivatives 1a-c
bind to plasmid more efficiently than Gemini 2. In particular,
the electrophoretic data revealed that the DNA mobility is
strongly affected by the macrocyclic compounds 1a,c already
at 50 µM. The highly condensed plasmids remain in the well
and are not visible in the gel because the staining reagent
(ethidium bromide) can not penetrate them. On the contrary,
no change is observed with compound 2 even at 200 µM.
Interesting information on the DNA condensation proper-
ties of our ligands was obtained by atomic force microscopy
(AFM) studies performed in the tapping mode (Figure 2). A
3954
Org. Lett., Vol. 10, No. 18, 2008

chains at the upper rim, only charge-charge interactions,
which are strengthened in the ethanol/water mixture, control
the DNA condensation process. Overall, the former “double
interaction” mechanism in DNA condensation is more
efficient than the latter “single interaction” pathway, and this
explains the AFM results. No DNA condensation at all was
observed with Gemini 2 (up to 50 µM, N/P ) 20) with or
without ethanol (Figure 2f).
Encouraged by the biophysical evidence of DNA binding
and condensation, we performed transfection experiments
using plasmid DNA pEGFP-C1 (1 nM), which expresses a
green fluorescent protein detectable in the cell by fluores-
cence microscopy, and RD-4 human rhabdomyosarcoma
cells. This cell line was chosen because, beyond the medical
relevance, it is easy to grow and not very well transfectable
by traditional methods compared to other cell lines like HEK
293 cells and therefore useful to judge the real transfection
capability of our compounds. No transfection occurs when
either DOPE (Figure S11 in Supporting Information) or
ligands 1a-c are used alone, whereas the formulation ligand/
DOPE (1/2 molar ratio) especially at ligand 10 µM (N/P )
4) is quite effective. In these conditions, we were pleased to
see (Figures 3 and 4 in the article and Figure S12 in
Figure 2. AFM images showing the effects induced on plasmid
DNA by guanidinium ligands 1a-c and 2. All images were obtained
with the microscope operating in tapping mode in air and with
supercoiled pEGFP-C1 plasmid deposited onto mica at a concentra-
tion of 0.5 nM. (a) Plasmid without guanidinium derivative. Plasmid
incubated with (b) 1a 0.6 µM; (c) 1c 2.5 µM; (d) 1b 1.8 µM; (e)
1b 1.8 µM + 10% EtOH; and (f) 2 50 µM. Each image represents
a 2 × 2 µm scan.
rather intriguing and subtle dependence of the DNA con-
densation ability of compounds 1a-c on the alkyl substituent
at the upper rim was observed. The p-tert-butyl derivative
1a forms highly condensed, nanometric condensates of single
DNA plectonemes even at 0.6 µM concentration (Figure 2b)
with N/P ) 0.5 (N/P ) guanidinium/phosphate ratio),
whereas the hexyl derivative 1c condenses DNA only at
concentration of 2.5 µM (N/P ) 2) and forms condensates
constituted by more filaments (Figure 2c). On the contrary,
the upper rim unsubstituted compound 1b does not form tight
condensates, although the single plectonemes are much more
constrained (Figure 2d) with respect to their relaxed state (Figure
2a), due to their interaction with the charged guanidinium
groups. At concentrations of 1b higher than 5 µM (N/P g 4),
no DNA results deposited on the mica surface. Evidently, in
these conditions, 1b causes the complete masking of the DNA
negative charges preventing its deposition.
Figure 3. Transfection experiments performed with pEGFP-C1
plasmid 1 nM, guanidinium calixarene/DOPE (1/2 molar ratio, 10/
20 µM), 2/DOPE (1/1 molar ratio), and lipofectamine LTX
formulations to rhabdomyosarcoma cells. Transfected cells are
visualized with fluorescence microscopy (upper row, in light green
because they express the enhanced green fluorescence protein
EGFP) and phase contrast microscopy (lower row).
The compact condensates formed by 1a and 1c are
partially relaxed (Figure S10 in Supporting Information) upon
addition of ethanol (10% in the buffer solution), while in
the case of the para- unsubstituted 1b at 1.8 µM (N/P )
1.5), the presence of the alcohol favors the plasmid conden-
sation (Figure 2e). This indicates that in para-alkyl substi-
tuted macrocycles 1a and 1c the primary electrostatic
interaction between the guanidinium groups and the DNA
phosphate anions is followed by hydrophobic interactions
between the alkyl chains. The latter interactions, which
induce the formation of condensates, are partially lost in
hydroalcoholic solution. In the case of 1b, which lacks alkyl
Supporting Information) that compound 1b is a very efficient
transfectant for RD-4 human rhabdomyosarcoma cells. The
amount of transfected cells (48%, Figure 4) is higher than
that achieved by the commercially available lipofectamine12
LTX (30%, Figures 3 and 4) and by the previously
investigated upper rim tetraguanidinium calix[4]arenes (less
than 20%, Figure 4). Moreover, little transfection activity is
observed for compounds 1a (3-4%) and 1c (6-7%) and
for 2 (ca. 6% at 20 µM).
Quite rewarding was the finding that the most active
compound 1b has a very low cytotoxicity (Figure 5) showing
(7) (a) Kirby, A. J.; Camilleri, P.; Engberts, J. B. F. N.; Feiters, M. C.;
Nolte, R. J. M.; So¨derman, O.; Bergsma, M.; Bell, P. C.; Fielden, M. L.;
Garc´ıa Rodr´ıguez, C. L.; Gue´dat, P.; Kremer, A.; McGregor, C.; Perrin,
C.; Ronsin, G.; van Eijk, M. C. P. Angew. Chem., Int. Ed. 2003, 42, 1448–
1457. (b) Fisicaro, E.; Compari, C.; Duce, E.; Donofrio, G.; Ro´z˙ycka-
Roszak, B.; Woz´niak, E. Biochim. Biophys. Acta 2005, 1722, 224–233.
(8) (a) Srinivasachari, S.; Fichter, K. M.; Reineke, T. M. J. Am. Chem.
Soc. 2008, 130, 4618–4627. (b) Horiuchi, S.; Aoyama, Y. J. Controlled
Release 2006, 116, 107–114. (c) Sansone, F.; Dudic`, M.; Donofrio, G.;
Rivetti, C.; Baldini, L.; Casnati, A.; Cellai, S.; Ungaro, R. J. Am. Chem.
Soc. 2006, 128, 14528–14536. (d) Lalor, R.; DiGesso, J. L.; Mueller, A.;
Matthews, S. E. Chem. Commun. 2007, 4907–4909.
Org. Lett., Vol. 10, No. 18, 2008
3955

cell used, we tested our lead compound 1b in a varying cell
line setting using again LTX as a reference (see Figure S13 in
Supporting Information). Comparable results with the two
formulations are obtained with AUBEK and BoMAK cells,
whereas LTX is definitely more active in the case of hMSC
and N2a cell lines. The reverse is true for Vero cells which are
not transfected at all by the commercial product LTX, while
1b/DOPE gives transfection (ca. 12%).
In summary, we have found that attaching guanidinium
groups at the phenolic oxygen atoms (lower rim) of calix[4]arenes
discloses the possibility to significantly enhance the cell
transfection ability of these synthetic cationic lipids formulated
with DOPE and reduce their toxicity to cells, if compared to
the same macrocycles with the charged groups directly linked
to the aromatic nuclei (upper rim). The DNA binding and
condensation properties, cell transfection ability, and toxicity
shown by the macrocyclic cytofectins 1a-c strongly depend
on the substituents at the upper rim of the calixarene. Com-
parison of the DNA condensation and cell transfection efficiency
of the most active compound 1b with its open chain analogue
2, having a very similar lipophilic/hydrophilic ratio, suggests a
possible positive role of the macrocyclic scaffold on gene
delivery to cell, which needs to be investigated in more detail.
Notably, the transfection efficiency of the little toxic 1b/DOPE
formulation is even higher than that of the commercially
available LTX in the case of RD-4 human rhabdomyosarcoma
and Vero cell lines. The results obtained warrant further studies
which are ongoing in our laboratory aimed at elucidating the
mechanism of cellular uptake and intracellular trafficking of
this novel DNA delivery system and its cargo to establish a
structure-activity relationship.
Figure 4. In vitro transfection efficiency as percentage of transfected
cells observed upon treatment of RD-4 human rhabdomyosarcoma
cells with guanidinium derivative/DOPE formulations and lipo-
fectamine LTX.
Acknowledgment. This work was supported by Fondazione
Cassa di Risparmio di Parma and Ministero dell’Universita´ e
della Ricerca (MUR) PRIN (project no. 2006034123). We are
indebted to the Centro Interdipartimentale Misure of the
University of Parma for use of NMR and AFM facilities.
Figure 5. Post-transfection cell viability in the presence of ligand
alone (yellow bars), ligand/DOPE formulation (orange bars), DOPE
alone (light blue bars), without any treatment (blue bars), and in
the presence of LTX (gray bar).
75-80% of cell viability at 48 h from transfection. Similar
results were obtained with 1a, while 1c is more toxic (60%
of cell viability).
It is not surprising that we did not find a strict correlation
between the transfection efficiency of the 1a-c/DOPE formula-
tions and the ability of the ligands to condense DNA as
disclosed by AFM studies. In fact, cell transfection achieved
with cytofectin/DOPE formulations is a complex supramolecular
function, and many important steps outside and inside the cell
must be positively affected to achieve the goal.13
Supporting Information Available: Experimental section
and characterization for all new intermediates and compounds
1a-c and 2, 1H and 13C NMR spectra of 1a-c and 2, Schemes
S1 and S2, and Figures S1-S13. This material is available free
of charge via the Internet at http://pubs.acs.org.
OL801326D
(10) (a) Farhood, H.; Serbina, N.; Huang, L. Biochim. Biophys. Acta
1995, 1235, 289–295.
(11) The first additions of 1a-c to the ethidium bromide/DNA mixture
result in a drastic and regular fluorescence quenching (Fig. S9 in Supporting
Information). However, at higher ligand concentrations (at N/P ratio of 4.2 for
1a, 8.5 for 1b, and 5 for 1c), a marked change in the slope of the titration
curve occurs. This is probably due to the interaction between ethidium bromide
and the DNA-calixarene complex and prevents a quantitative evaluation of
the ligand affinity for plasmid.
Since it is well-known in the field of synthetic nonviral
vectors that transfection efficiency may depend on the type of
(9) (a) Nishihara, M.; Perret, F.; Takeuchi, T.; Futaki, S.; Lazar, A. N.;
Coleman, A. W.; Sakai, N.; Matile, S. Org. Biomol. Chem. 2005, 3, 1659–
1669. (b) Takeuchi, T.; Kosuge, M.; Tadokoro, A.; Sugiura, Y.; Nishi, M.;
Kawata, M.; Sakai, N.; Matile, S.; Futaki, S. ACS Chem. Biol. 2006, 1,
299–303. (c) Sainlos, M.; Hauchecorne, M.; Oudrhiri, N.; Zertal-Zidani,
S.; Aissaoui, A.; Vigneron, J.-P.; Lehn, J.-M.; Lehn, P. ChemBioChem 2005,
6, 1023–1033. (d) Elson-Schwab, L.; Garner, O. B.; Schuksz, M.; Crawford,
B. E.; Esko, J. D.; Tor, Y. J. Biol. Chem. 2007, 282, 13585–13591. (e)
Wender, P. A.; Galliher, W. C.; Goun, E. A.; Jones, L. R.; Pillow, T. AdV.
Drug ReV. 2008, 60, 452–472.
(12) Lipofectamine LTX Reagent is a cationic lipid-based, animal-
origin free formulation for the transfection of DNA into eukaryotic cells
with low cytotoxicity. See: http://www.invitrogen.com/downloads/F-069843-
LpofctLTX_FHR.pdf.
(13) (a) Kostarelos, K.; Miller, A. D. Chem. Soc. ReV. 2005, 34, 970–994.
(b) Boeckle, S.; Wagner, E. AAPS J. 2006, 8, E731–E742. (c) Zuhorn, I. S.;
Engberts, J. B. F. N.; Hoekstra, D. Eur. Bipophys. J. 2007, 36, 349–362.
3956
Org. Lett., Vol. 10, No. 18, 2008