Multifunctional core-shell silica nanoparticles for highly sensitive 19F magnetic resonance imaging

Article abstract of DOI:10.1002/anie.201308500

¹⁹F magnetic resonance imaging (¹⁹F MRI) is useful for monitoring particular signals from biological samples, cells, and target tissues, because background signals are missing in animal bodies. Therefore, highly sensitive ¹⁹

Full text of DOI:10.1002/anie.201308500

.
Angewandte
Communications
DOI: 10.1002/anie.201308500
Molecular Imaging
Multifunctional Core–Shell Silica Nanoparticles for Highly Sensitive
19F Magnetic Resonance Imaging**
Hisashi Matsushita, Shin Mizukami, Fuminori Sugihara, Yosuke Nakanishi,
Yoshichika Yoshioka, and Kazuya Kikuchi*
Abstract: 19F magnetic resonance imaging (19F MRI) is useful
for monitoring particular signals from biological samples, cells,
and target tissues, because background signals are missing in
animal bodies. Therefore, highly sensitive 19F MRI contrast
agents are in great demand for their practical applications.
However, we have faced the following challenges: 1) increasing
the number of fluorine atoms decreases the solubility of the
molecular probes, and 2) the restriction of the molecular
mobility attenuates the 19F MRI signals. Herein, we developed
novel multifunctional core–shell nanoparticles to solve these
issues. They are composed of a core micelle filled with liquid
perfluorocarbon and a robust silica shell. These core–shell
nanoparticles have superior properties such as high sensitivity,
modifiability of the surface, biocompatibility, and sufficient
in vivo stability. By the adequate surface modifications, gene
expression in living cells and tumor tissue in living mice were
successfully detected by 19F MRI.
be delivered with high density per voxel to give signals of
sufficient intensity.[2] Thus, highly sensitive 19F MRI probes
are required for their practical applications. To addressing this
sensitivity problem, it is important to increase the number of
fluorine atoms in the MRI probes. However, simple multi-
plication of the fluorine atoms does not provide higher
sensitivity because of the following two reasons: 1) increasing
the number of fluorine atoms decreases the solubility of the
molecular probes[3] and 2) suppression of the molecular
mobility induced by the increase in the molecular size
shortens the transverse relaxation time (T2), resulting in
attenuation of the MRI signal.[4] To solve these two problems
simultaneously, we designed a novel 19F MRI probe consisting
of a core–shell nanoparticle that involves a perfluoro-
[15] crown-5 ether (PFCE).
PFCE has various attractive characteristics as a 19F MRI
contrast agent. First, PFCE has 20 equivalent fluorine atoms,
which show a sharp single 19F NMR peak to yield a strong
signal. Second, a large number of liquid PFCE are expected to
maintain their high molecular mobility in a nanoparticle.
Taking these advantages, PFCE nanoemulsions have been
reported for in vivo cell tracking.[5] However, the instability of
PFCE nanoemulsions in organic solvents significantly limits
their application through surface modifications. Even under
neutral buffer conditions, Ostwald ripening, which is a molec-
ular diffusion phenomenon resulting in the gradual growth of
larger particles, can be a major problem for the stability.[6] To
overcome this limitation, we considered to cover the PFCE
nanoemulsion with a silica shell (Figure 1a). Silica nano-
particles possess attractive features such as biological inert-
ness and favorable colloidal properties.[7] In addition, the
silica surface can be variously modified with functional groups
to realize the targeted delivery and specific functions. Herein,
we report a novel multifunctional 19F MRI contrast agent,
fluorine accumulated silica nanoparticle for MRI contrast
enhancement (FLAME), and demonstrate the superior
properties for 19F MRI such as high sensitivity, stability,
modifiability of the surface, biocompatibility, and biomedical
applications such as reporter assay and in vivo tumor imaging.
The synthetic protocol of FLAME is presented in Fig-
ure 1b. To cover the PFCE-phospholipid nanoemulsion with
silica gel, we developed a novel surfactant, PAP (Scheme S1),
which consists of an alkyl part to interact with phospholipids[8]
and a basic pyridinyl group. When template nanoemulsions
are formed in the presence of PAP, the basic site of PAP was
expected to be displayed on the nanoemulsion surface and to
initiate the sol–gel process of tetraethyl orthosilicate.[9] As
a result, the silica polymerization reaction occurred only near
the nanoemulsion surface, and silica coating of the PFCE
M
agnetic resonance imaging (MRI) can provide valuable
information about deep tissues in animal bodies with high
spatial resolution without using radioactivity. 19F MRI is an
especially powerful method for in vivo imaging of particular
biomolecules, cells, and target tissues, because of negligible
background signals.[1] In 19F MRI, fluorinated agents need to
[*] H. Matsushita, Dr. S. Mizukami, Y. Nakanishi, Prof. K. Kikuchi
Graduate School of Engineering, Osaka University
Osaka 565-0871 (Japan)
E-mail: kkikuchi@mls.eng.osaka-u.ac.jp
Homepage: http://www-molpro.mls.eng.osaka-u.ac.jp/English/
indexeng/indexENG.html
Dr. F. Sugihara, Prof. Y. Yoshioka
Immunology Frontier Research Center
Osaka University, Osaka 565-0871 (Japan)
[**] This research was supported by the Ministry of Education, Culture,
Sports, Science, and Technology (Japan) (grant numbers 24108724,
24685028, 24115513, 24651259, 25620133, and 25220207), by the
Japanese Society for the Promotion of Science (JSPS) through its
Funding Program for World-Leading Innovative R&D on Science
and Technology (FIRST Program), and by CREST from JST. The
authors acknowledge the Asahi Glass Foundation and the Magnetic
Health Science Foundation. The authors thank Prof. Tsutomu Ono
(Okayama University) for helpful discussion to synthesize nano-
materials, and Dr. Takayuki Kato and Dr. Takao Sakata (Osaka
University) for their support in TEM measurements. Some of the
experiments were carried out at the Research Center for Ultra-High
Voltage Electron Microscopy, Osaka University. The authors also
thank Dr. Yuko Kamikawa (Osaka University) for the valuable
discussion.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201308500.
1008
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1008 –1011

Angewandte
Chemie
Figure 1. Design and synthetic strategy of FLAME. a) The components of FLAME. b) The chemical structure of PAP and synthesis of FLAME.
nanoemulsion was achieved (Scheme S2). Rhodamine B iso-
thiocyanate (RITC) was covalently attached to the silica shell,
which allowed for the fluorescence detection.
The core–shell structures of FLAME were clearly
observed by transmission electron microscopy (TEM; Fig-
ure 2a). The average diameter of the particles was 76 Æ 9 nm
À16.4 ppm (Figure 2c). In the 19F MRI measurements, strong
19F MRI signals were observed from FLAME (Figure 2d).
The 19F MRI signal intensity was proportional to the PFCE
concentration (Figure 2e). FLAME showed the typical
fluorescence emission of RITC (Figure S3).
To demonstrate the high 19F MRI sensitivity, FLAME was
applied to a 19F MRI-based reporter assay. We chose a BL-tag
protein as a reporter protein,[10] which is a noncatalytic b-
lactamase mutant that covalently binds various b-lactam
structures. Thus, ampicillin-modified FLAME (FLAME-
Amp) was prepared for the specific 19F MRI detection of
the reporter protein (Figure 3a, Scheme S3, and Table S1). To
confirm the specific binding of FLAME-Amp to the BL-tag
protein, we used a fusion protein of BL-tag and maltose-
binding protein (MBP), which noncovalently bind to an
amylose with high affinity (Scheme S4). After FLAME-Amp
was incubated with an MBP-BL-bound amylose resin, the
nonbound components were removed by washing with
a buffer solution (Figure S4). Then, the fractions eluted with
a maltose solution were analyzed by 19F NMR spectroscopy.
A sufficiently high and sharp 19F NMR peak was detected
from MBP-BL incubated with FLAME-Amp (Figure 3b).
The 19F NMR peak was detected even at a MBP concen-
tration of 30 nm.
A small molecule-based probe, F-Amp, was also synthe-
sized to evaluate the 19F MRI sensitivity (Scheme S5 and
Figure S5). The 19F NMR signal of F-Amp was not observed
at a MBP-BL concentration of 300 nm. At least concentra-
tions of the target protein in the micromolar range were
required for NMR detection by the small-molecule probe
because of the low 19F density per probe. In contrast,
FLAME-Amp encapsulating a large amount of fluorine
atoms enabled the detection of the target proteins in the
nanomolar range (Figure S6 and Table S2). Another impor-
tant factor for sensitivity improvement is the rotational
mobility of fluorine atoms. The mobility of small-molecule
probes is largely decreased by the binding to proteins;
furthermore, it causes the T2 reduction, which decreases the
NMR signal. The T2 of F-Amp drastically shortened from
0.434 s to 0.119 s by the covalent binding to MBP-BL and the
19F NMR signal was decreased (Figure S7). On the other
hand, the T2 of FLAME-Amp conjugated to MBP-BL was
0.210 s, which is comparable to that of FLAME-Amp (0.238 s;
Figure 2. Characterization of FLAME. a) TEM image of FLAME. b) Par-
ticle size distribution histogram calculated from TEM images.
c) 19F NMR spectrum of FLAME in PBS (pH 7.4). d) MRI of phantoms
(1–6) filled with FLAME in PBS (500 mL). PFCE concentrations: 1.30,
0.650, 0.325, 0.163, 0.0813, and 0.0406 mm for phantoms 1–6, respec-
tively. Phantom 7 contains PBS. e) Plot of the normalized 19F MRI
signal intensity versus PFCE concentration.
(Figure 2b). On the other hand, the silica shell was not well
formed in the absence of PAP (see Figure S1 in the Support-
ing Information). Dynamic light scattering (DLS) measure-
ments also indicated the silica coating of nanoemulsion. The
hydrodynamic diameter was changed from 80 nm of the
nanoemulsion to 131 nm of FLAME, and the z potential
decreased from + 29.5 mVof the nanoemulsion to À4.5 mVof
FLAME (Figure S2 and Table S1). Inclusion of PFCE in the
particle was confirmed by NMR spectroscopy. The 19F NMR
spectroscopy of FLAME dispersed in phosphate-buffered
saline (PBS) solution showed a PFCE-derived single peak at
Angew. Chem. Int. Ed. 2014, 53, 1008 –1011
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 1009

.
Angewandte
Communications
Finally, the potential of FLAMEs for in vivo targeting
were evaluated in tumor-bearing mice. The nanomaterials
were delivered to the tumor using the enhanced permeability
and retention (EPR) effect, which promotes nanoparticle
accumulation in cancerous tissues by a combination of leaky
angiogenic blood vessels and deficient tumor lymphatics.[12]
However, the naked nanoparticles should be trapped
immediately by the reticuloendothelial system (RES) and
show a decreased circulation time. Therefore, FLAME was
modified with polyethylene glycol (PEG) for the effective
delivery to tumors (Figure 4a). PEGylation reduces the
uptake by increasing the electrostatic interactions with
proteins and small molecules.[13]
Figure 4. In vivo accumulation of FLAME at tumor site. a) Structure of
FLAME-PEG and FLAME-COOH b) In vivo MRI of FLAME-PEG and
FLAME-COOH in tumor-bearing mice. The positions of the liver and
tumor are represented as L and T, respectively.
Figure 3. 19F NMR/MRI detection of BL-tag protein. a) The chemical
structure of FLAME-Amp. b) Highly sensitive 19F NMR detection of
MBP-BL proteins with FLAME-Amp. c) 19F MRI detection of gene
expression using BL-EGFR and FLAME. d) MRI of BL-EGFR-expressing
HEK293T cells treated with FLAME-Amp, FLAME-COOH, and EGFR-
expressing cells treated with FLAME-Amp.
PEGylated FLAME (FLAME-PEG) was synthesized
from FLAME-NH2 (Scheme S6 and Table S1). DLS analysis
to determine the stability of FLAME-PEG showed that there
was almost no change in the particle size for 1 week (Fig-
ure S9). The biocompatibility of FLAME-PEG was evaluated
in the MTT assay. The viability of colon-26 cells was not
affected by 24 h exposure of FLAME-PEG and up to
900 mgmLÀ1 of PFCE (Figure S10).
Passive targeting and accumulation of FLAME-PEG
using the EPR effect were demonstrated by MRI. Mice
bearing a tumor were given intravenous injections of
FLAMEs (Figure 4b). Strong 19F MRI signals of FLAME-
PEG at the tumor site indicated passive targeting of the
nanoparticles. The 19F NMR measurements of homogenized
tissue samples showed that NMR signals were detected only
Table S3). These results demonstrate that the rotational
motion of PFCE was maintained even after FLAME was
attached to a protein.
Next, 19F MRI detection of gene expression was
attempted using a cell surface-displayed reporter protein
(Figure 3c). In this experiment, the BL-tag was fused to the
epidermal growth factor receptor (EGFR), which is a mem-
brane-associated glycoprotein;[11] the recombinant protein
was expressed on the surface of HEK293T cells.[10] After
FLAME-Amp was incubated with cells expressing BL-
EGFR, MRI experiments were performed. The strong
19F MRI signal was clearly observed from BL-EGFR-express-
ing cells incubated with FLAME-Amp, whereas almost no
signal was detected from the control EGFR-expressing cells
(Figure 3d and Figure S8). As a result, specific and highly
sensitive 19F MRI detection of gene expression in living cells
was successfully demonstrated by FLAME-Amp.
from liver (CPFCE = 113 nmolgÀ1) and tumor (CPFCE
=
28.8 nmolgÀ1), but not from spleen, lungs, brain, kidneys,
and heart (Figure S11). However, the 19F MRI signals of non-
PEGylated FLAME-COOH (Scheme S3) was detected only
in the liver, indicating that FLAME-COOH was immediately
1010
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1008 –1011

Angewandte
Chemie
trapped by the RES. These results indicate that the 19F MRI
signals of FLAMEs are sufficiently strong for in vivo studies.
The accumulation of FLAME-PEG in tumor and liver
sections was verified by the RITC fluorescence (Figure S12).
Meanwhile, in the case of FLAME-COOH, fluorescence
signals were detected only in the livers and spleens, not at the
tumors (Figure S13). These results demonstrate that the 19F
MRI signal localization reflects the actual accumulation
position of FLAMEs in animal bodies.
[1] a) M. Higuchi, N. Iwata, Y. Matsuba, K. Sato, K. Sasamoto, T. C.
Saido, Nat. Neurosci. 2005, 8, 527 – 533; b) S. Mizukami, R.
Takikawa, F. Sugihara, Y. Hori, H. Tochio, M. Wꢀlchli, M.
Shirakawa, K. Kikuchi, J. Am. Chem. Soc. 2008, 130, 794 – 795;
c) Y. Takaoka, T. Sakamoto, S. Tsukiji, M. Narazaki, T. Matsuda,
H. Tochio, M. Shirakawa, I. Hamachi, Nat. Chem. 2009, 1, 557 –
561; d) M. Srinivas, A. Heerschap, E. T. Ahrens, C. G. Figdor,
I. J. M. de Vries, Trends Biotechnol. 2010, 28, 363 – 370.
[2] J. Ruiz-Cabello, B. P. Barnett, P. A. Bottomley, J. W. M. Bulte,
NMR Biomed. 2011, 24, 114 – 129.
In conclusion, we have successfully developed novel
contrast agents with a core–shell structure, FLAMEs, for
highly sensitive 19F MRI. By coating the surface with silica,
FLAMEs demonstrated practical properties such as a chemi-
cally modifiable surface, dispersibility in water, biocompati-
bility, and high stability. FLAMEs were proven to be useful
19F MRI contrast agents to overcome two major limitations of
current 19F MRI probes, that is, an impractical modifiability of
the surface of nanoemulsions and a low sensitivity of small-
molecule-based probes. As an extension of the 19F MRI-based
reporter assay shown in this study, in vivo evaluation of gene
transplantation efficiency is one of the promising future
applications. FLAME could also lead to various biomedical
applications such as 19F MRI of atherosclerosis plaques with
a clot-binding peptide[14] or bone neoplasms with bisphosph-
onate.[15]
[3] J. Liu, L. S. Lee, Environ. Sci. Technol. 2007, 41, 5357 – 5362.
[4] K. Tanaka, N. Kitamura, K. Naka, Y. Chujo, Chem. Commun.
2008, 6176 – 6178.
[5] E. T. Ahrens, R. Flores, H. Xu, P. A. Morel, Nat. Biotechnol.
2005, 23, 983 – 987.
[6] a) J. G. Riess, M. Postel, Biomater. Artif. Cells Immobilization
Biotechnol. 1992, 20, 819 – 830; b) M. G. Freire, A. M. A. Dias,
M. A. Z. Coelho, J. A. P. Countinho, I. M. Marrucho, J. Colloid
Interface Sci. 2005, 286, 224 – 232.
[7] J. L. Vivero-Escoto, R. C. Huxford-Phillips, W. Lin, Chem. Soc.
Rev. 2012, 41, 2673 – 2685.
[8] S. Watanabe, H. Oaki, K. Mitsuhashi, S. Nakahara, Makromol.
Chem. 1992, 193, 2781 – 2792.
[9] S. Wang, M. Zhang, D. Wang, S. Liu, Microporous Mesoporous
Mater. 2011, 139, 1 – 7.
[10] S. Mizukami, S. Watanabe, Y. Hori, K. Kikuchi, J. Am. Chem.
Soc. 2009, 131, 5016 – 5017.
[11] S. H. Roy, Int. J. Radiat. Oncol. Biol. Phys. 2004, 59, 21 – 26.
[12] Y. Matsumura, H. Maeda, Cancer Res. 1986, 46, 6387 – 6392.
[13] S. Li, L. Huang, Biochim. Biophys. Acta Biomembr. 2009, 1788,
2259 – 2266.
Received: September 30, 2013
[14] D. Peters, M. Kastantin, V. R. Kotamraju, P. P. Karmali, K.
Gujraty, M. Tirrell, E. Ruoslahti, Proc. Natl. Acad. Sci. USA
2009, 106, 9815 – 9819.
[15] C. Clementi, K. Miller, A. Mero, R. Satchi-Fainaro, G. Pasut,
Mol. Pharm. 2011, 8, 1063 – 1072.
Keywords: fluorine · imaging agents ·
magnetic resonance imaging · nanoparticles
.
Angew. Chem. Int. Ed. 2014, 53, 1008 –1011
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1011