Integration of Indocyanine Green Analogs as Near-Infrared Fluorescent Carrier for Precise Imaging-Guided Gene Delivery

Article abstract of DOI:10.1002/smll.201906538

Codelivery of diagnostic probes and therapeutic molecules often suffers from intrinsic complexity and premature leakage from or degradation of the nanocarrier. Inspired by the “Y” shape of indocyanine green (ICG), the dye is integrated in an amphiphilic lipopeptide (RNF). The hydrophilic segment is composed of arginine-rich dendritic peptides, while cyanine dyes are modified with two long carbon chains and employed as the hydrophobic moiety. They are linked through a disulfide linkage to improve the responsivity in the tumor microenvironment. After formulation with other lipopeptides at an optimized ratio, the theranostic system (RNS-2) forms lipid-based nanoparticles with slight positive zeta potential enabling efficient condensation of DNA. The RNS-2 displays glutathione responded gene release, activatable fluorescence recovery, and up to sevenfold higher in vitro transfection than Lipofectamine 2000. Compared with a Cy3 and Cy5 labeled fluorescence resonance energy transfer indicator for gene release, the “turn-on” indocyanine green analogs exhibit longer emission wavelength and better positive correlation with the dynamic processes of gene delivery. More importantly, the RNS-2 system enables efficient near infrared imaging guided gene transfer in tumor-bearing mice and thus provides more precise and accurate information on location of the cargo gene and synthesized carriers.

Full text of DOI:10.1002/smll.201906538

Full PaPer
www.small-journal.com
Integration of Indocyanine Green Analogs as Near-Infrared
Fluorescent Carrier for Precise Imaging-Guided Gene Delivery
Hong Liang, Xiaobing Chen, Rongrong Jin, Bowen Ke,* Matthias Barz, Hua Ai, and Yu Nie*
the aim by themselves, for the nucleic
Codelivery of diagnostic probes and therapeutic molecules often suffers from
intrinsic complexity and premature leakage from or degradation of the nanocar-
rier. Inspired by the “Y” shape of indocyanine green (ICG), the dye is integrated
in an amphiphilic lipopeptide (RNF). The hydrophilic segment is composed of
arginine-rich dendritic peptides, while cyanine dyes are modified with two long
carbon chains and employed as the hydrophobic moiety. They are linked through
a disulfide linkage to improve the responsivity in the tumor microenvironment.
After formulation with other lipopeptides at an optimized ratio, the theranostic
system (RNS-2) forms lipid-based nanoparticles with slight positive zeta poten-
tial enabling efficient condensation of DNA. The RNS-2 displays glutathione
responded gene release, activatable fluorescence recovery, and up to sevenfold
higher in vitro transfection than Lipofectamine 2000. Compared with a Cy3 and
Cy5 labeled fluorescence resonance energy transfer indicator for gene release,
the “turn-on” indocyanine green analogs exhibit longer emission wavelength
and better positive correlation with the dynamic processes of gene delivery.
More importantly, the RNS-2 system enables efficient near infrared imaging
guided gene transfer in tumor-bearing mice and thus provides more precise and
accurate information on location of the cargo gene and synthesized carriers.
acids are anionic biomacromolecules
that cannot be spontaneous translocated
through cell membrane. Cellular entry,
lysosome degradation, and nuclear pore
traverse are great obstacles during the
transport processes. Thus, gene delivery
vehicles are critical. Positively charged
synthetic vectors (such as cationic poly-
mers, lipids and amino acids) are gen-
erally used to condense nucleic acid via
electrostatic interaction into appropriate
sized nanoparticles for efficient endocy-
tosis. More and more cationic carriers
with well-defined structures have been
fabricated to explore the structure–activity
relationship, which was closely associated
with the gene transfection efficiency.[1]
In addition, microenvironment (such as
pH,[2,3] reduction,[4,5] reactive oxygen spe-
cies,[6,7] and adenosine triphosphate[8,9]
)
stimuli-responsive moieties were applied
to induce sensitive deshielding,[10–12]
charge conversion,[13,14] disassembly,[15]
and configuration change[16] for better biomimic and delivery
process optimization. We have also designed and synthesized a
series of amino acid derivatives, including amphiphilic lysine/
histidine/arginine dendrimers,[17] amino acid functionalized
polysaccharides,[18] and arginine-rich nanohybrids with hierar-
chical response or targeting,[13] which have been demonstrated
to achieve enhanced delivery efficiency of genes or drugs. How-
ever, there are still many unclear processes, such as undesir-
able accumulation, inappropriate elimination, serious immune
response, and premature degradation of the delivery system,
which may reduce the therapeutic effect, especially the final
transfection in vivo. Therefore, we need to develop a technique
for the in vivo fate understanding of gene delivery systems.
What happens to them and how the therapeutic agents is really
delivered?
In order to answer these questions, accurate and precise in vivo
traceability of delivery system is required. Ideally a tracer should
allow for an investigation of biodistribution as well as intracel-
lular localization at the same time. In the past decade, theranostic
concept provides a promising strategy for simultaneous drug or
gene delivery and molecular imaging in a single system[19] and
gives a capability to visualize and quantify the pharmacokinetic
processes of cellular and molecular therapies (adsorption, dis-
tribution, metabolism, and excretion) at both cellular and tissue
levels in a real-time fashion.[20,21] Particularly, imaging-guided
drug delivery can monitor the path of drugs in vivo, evaluate the
1. Introduction
Gene therapy has made remarkable progress in the cure of
genetic disorders and cancer. Successful gene therapy requires
transport of sufficient amount of nucleic acids (pDNA, mRNA,
siRNA, etc.) inside cell, reaching to the nuclei or cytosolic
compartments. However, genetic substances hardly realize
Dr. H. Liang, X. Chen, Dr. R. Jin, Prof. H. Ai, Prof. Y. Nie
National Engineering Research Center for Biomaterials
Sichuan University
Chengdu 610064, P. R. China
E-mail: nie_yu@scu.edu.cn
Prof. B. Ke
Laboratory of Anesthesiology and Critical Care Medicine
Department of Anesthesiology
Translational Neuroscience Center
West China Hospital
Sichuan University Chengdu
Sichuan, Chengdu 610041, P. R. China
E-mail: bowenke@scu.edu.cn
Dr. M. Barz
Institute of Organic Chemistry
Johannes Gutenberg-University Mainz
Duesbergweg 10-14, 55099 Mainz, Germany
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/smll.201906538.
DOI: 10.1002/smll.201906538
©
1906538 (1 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538

www.advancedsciencenews.com
www.small-journal.com
Scheme 1. Schematic diagram of near-infrared fluorescence theranostic carrier for image-guided gene delivery.
efficiency of tumor-targeted delivery, and further provide informa-
tion regarding pharmacokinetics of the treatment.[22]
skeleton structure as hydrophobic moiety for fluorescence
imaging, and a disulfide bond as trigger for reduction-responsive
cargo release at tumor. Furthermore, the fluorescence signal
will grow from nothing by the aggregation-caused quenching
(ACQ) effect, to gradually increase with the disassembly of
nanocarrier systems.
Typically, the probes used for imaging are physically
encapsulated or chemically bonded to the residual groups
of nanocarriers. However, these loading modes for fluores-
cent nanomaterials often suffer from high nonspecific back-
ground and low signal-to-background ratio due to the prema-
ture leakage,[23–25] even fluorescence resonance energy transfer
(FRET) was used for imaging. Compared with “always-on”
probes, the design of activatable theranostic nanomaterials with
“turn-on” feature offers an opportunity to reduce the background
interference and improve the spatial resolution,[26–28] for it could
only switch from “OFF” to “ON” for selective molecules or
events. In addition, it was found that the imaging capacity and
gene delivery procedure often implemented separately, resulting
in low theranostic outcome.[29] Thus, the conjugation linker and
the position of probe are factors that cannot be ignored. Predeg-
radation and following nonspecific adsorption events may lead
to false positive results.[30] Little information was provided about
real carrier metabolic process during the whole therapy, so that
timely feedback from the theranostic system is not available.
Therefore, researchers attempted to rationally integrate
the fluorescent probe as “main” portion of the carriers. Some
amphiphilic drug delivery carriers that taking the advantage
of strong hydrophobic properties of boron dipyrromethene
(BODIPY) have developed.[20,31,32] And platinated-BODIPY
molecules have been designed to combine chemical, photo-
dynamic, and photothermal therapy together for tumor abla-
tion,[33] showing encouraging prospect.
2. Results and Discussion
2.1. Synthesis of Arginine-rich and Near-Infrared Fluorescence
Dye-Containing Amphiphilic Dendritic Lipopeptide (RNF)
Cationic poly-l-lysine, polyethylene imine (PEI), (2,3-dioleoy-
loxy-propyl)-trimethylammonium lipid, and commercial kits
from the lipofectamine family are the most commonly used
systems for gene transfection in vitro, but restricted for in vivo
application due to their poor serum resistance capacity during
the complicated biological environment. Viruses are natural
perfect assemblies with nanoscaled core–shell structure and
smart configuration adjustment during infection. Therefore,
with the development of supramolecular science and inspira-
tion from virus, chemists have designed many new artificial
vectors for gene delivery over the last decade.[36] Some amphi-
philic molecules, which can self-assemble into nanoparticles
and response to microenvironment, provide a convenient
approach to mimic viral architectures and functions, indeed
improving the gene delivery efficiency.[37] In our previous
studies, a series of cationic arginine-rich amphiphilic lipopep-
tides have been optimized for viral cell-penetrating peptides
mimicking, achieving high delivery efficiency in vitro and
vivo.[15,38] Meanwhile, we have successfully imitated the viral
intelligent regulation by introduction of gradient pH or reduc-
tion responsive moieties in the artificial vectors, obtaining
high selectivity in gene transfection.[13,18] In order to go further
understanding and optimizing virus-inspired nanomaterials,
from our point of view, there is a tremendous requirement for
precise and accurate bioimaging on organ and cellular levels.
In view of current challenges and inspired by chemical struc-
ture of indocyanine green (ICG) which possesses “Y” shape
symmetrical group and approvement by The Food and Drug
Administration,[34,35] we herein developed a redox-induced
“turn-on” system for gene delivery and imaging with real-time
visualization of carrier metabolism (Scheme 1). The designed
amphiphilic molecules consist of three segments: an arginine-
rich hydrophilic moiety for efficient gene condensation and
cytomembrane penetration, a derivative of cyanine dye in the
©
Small 2020, 1906538
1906538 (2 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Since biodistribution and subcellular distribution are dynamic
processes, flourescence imaging seems to be the method of
choice in noninvasive optical techniques to study these processes.
Among them, near infrared (NIR) fluorescence is widely applied
due to its enhanced penetration and weak absorption/scattering
properties in tissues, and the ICG has been approved by Food
and Drug Administration as an NIR fluorescent contrast agent
for in vivo imaging. In order to visualize nanosized delivery
systems, organic dyes are typically encapsulated into nano-
carriers[39,40] or conjugated to the periphery of materials.[41–43]
However, recent studies confirmed that these patterns probably
failed to monitor the in vivo fate of carriers because the dyes
tend to fall off before reaching the targeted site.[23,44] Therefore,
we fabricated an amphiphilic molecule composed of dendritic
arginine and cyanine derivative dye (IR-780 molecule), which
could implement the fluorophore in the carrier systems itself
for combination of bioimaging and gene delivery. In prospecting
for better assembly and cellular fusion properties, we introduced
two long alkyl chains to IR-780 improving its hydrophobicity,
flexibility, and capacity of gene carrier.[45] After self-assembly, the
fluorescence would be quenched due to the ACQ effect. While
under the reductive condition, the whole amphiphile would dis-
assembly and restore its fluorescent signal, owing to cleavage of
the disulfide bond.
2 was used for the condensation reaction with prepared interme-
diation indolium salt (compound 1) in anhydrous sodium acetate.
By employing side chain on the indocyanine backbone, chloro-
substituted fluorophore (compound 3) was successfully synthe-
sized for the next step. The addition of phenoxypropionic acid to
the base chlorocyanine dye (compound 3) occurred by nucleo-
philic substitution of the choloride via an SN1 mechanism. This
reaction generated the labeling dye (compound 4) as green solid.
The disulfide bond containing cystamine (compound 5) as linker
was first conjugated with hydrophobic fluorescent molecule
to afford intermediate compound 6 with moderate yield in the
presence of O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium
tetrafluoroborate (TBTU) and N,N-diisopropylethylamine (DIEA).
After removal of protective group, the dendron with arginine-rich
periphery was covalently bonded to the other end of the linker
to generate compound 9 via a condensation between the amino
group and carboxyl group. After deprotection, the amphiphilic
molecule (RNF) with arginine (R) periphery and near-infrared
fluorophore was obtained (Figure 1). More details about the
synthesis of target molecule were presented in the Supporting
Information. And the crude products were purified by silica
column chromatography with gradient elution (methanol/dichlo-
romethane). The chemical structure of each compound was char-
acterized by nuclear magnetic resonance spectroscopy (1H) and
mass spectrum (Figures S1–S8, Supporting Information). The
arginine-rich and oleic acid-containing amphiphilc dendritic lipo-
peptide (RLS), which did not contain fluorophore, was prepared
according to the literature.[38]
The hydrophilic part, containing dendritic arginine, was syn-
thesized using divergent method with lysine as a core,[15,38] while
hydrophobic part of cyanine dye was prepared according to the
established methodologies.[46] As depicted in Figure 1, compound
Figure 1. Synthetic routes of indocyanine green analogs integrated amphiphilic dendritic lipopeptide (RNF) as near-infrared fluorescence theranostic carrier.
©
1906538 (3 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538

www.advancedsciencenews.com
www.small-journal.com
Table 1. The size and zeta potential of various assemblies and their
gene complexes.
and RNS-2/DNA (N/P = 30) complexes was observed by trans-
mission electron microscopy (Figure 2A,B), displaying a uniform
ellipsoidal shape, although in line with DLS DNA lipoplexes are
slightly larger in diameter.
Sample
Ratio of
RLS/RNF
N/P
PDI
Z-average
[nm]
Zeta potential
[mV]
Moreover, we investigated the spectral characteristic of var-
ious molecule and assemblies in organic or aqueous solutions
to mimic flourescent “ON” and “OFF”, respectively. RNF mole-
cule showed the maximal ultraviolet absorption of RNF at
770 nm in methanol (Figure S9A, Supporting Information),
while it displayed maximum fluorescence emission wave-
length of 780 nm in methanol with excitation wavelength of
750 nm (Figure S9B, Supporting Information). The fluores-
cence spectra of other assemblies with different formulation
ratios can be found in Figure S9B in the Supporting Informa-
tion. It was obviously that fluorescence of RNF was quenched
after being neither self-assemblies alone nor mixed with RLS
into nanoparticles, due to ACQ effect. Interestingly, the RNS-2
assemblies have better quenching effect than the others, con-
firming a tighter and better assembly at this mixing ratio. To
demonstrate the turn-on property of fluorescence, the RNS-2
assemblies were incubated with 10 × 10−3 m glutathione and
dithiothreitol (DTT) for 1 h at 37 °C. As expected, fluorescence
recovery was observed, almost reaching the level of RNF in
methanol, owing to the disassembly caused by disulfide bond
breakage (Figure 2C). Similar phenomenon of the fluorescence
recovery has also been confirmed by other studies,[48] in which
fluorescent dye could be cleaved from the periphery of nano-
particles and the intensity could be restored from the original
quenching state.
RNF
—
—
1
—
—
—
—
—
40
10
30
30
30
0.28 0.06
0.16 0.01
0.19 0.04
0.18 0.05
0.13 0.02
0.19 0.05
0.15 0.01
0.16 0.01
0.17 0.02
0.18 0.03
165.8 1.9
140.1 2.7
140.4 3.1
105.9 1.2
126.0 2.1
201.6 6.1
225.5 2.6
193.1 1.1
224.1 2.4
198.3 5.9
25.7 1.1
22.6 0.1
22.8 0.1
20.5 0.7
21.6 1.3
22.7 1. 6
17.5 0.4
20.4 0.8
17.6 0.1
20.8 1.2
RLS
RNS-1
RNS-2
2
RNS-3
3
RNF/DNA
RLS/DNA
RNS-1/DNA
RNS-2/DNA
RNS-3/DNA
—
—
1
2
3
2.2. Characterization of Physicochemical Properties
Based to the size and zeta potential detected by dynamic light
scattering (DLS, Table 1), RNF and RLS assemblies containing
only single molecule showed the Z-average diameter of ≈165
and 140 nm, with zeta potential of 25.7 and 22.6 mV, respec-
tively. These slight differences may be attributed to the higher
hydrophobicity and rigidity of NIR dye in RNF, which limited
flexibility in the self-assembly process.[47] Mixture of RLS an
RNF molecules at different mass ratios (RNS-1, RNS-2 and
RNS-3) could form nanosized particles with diversified charac-
ters. Among the mixed assemblies, the size of RNS-1 and RNS-3
assemblies were about 140 and 126 nm, with surface potential
of 22.8 and 21.6 mV, respectively. While RNS-2 displayed the
minimum size of 105.9 nm and zeta potential of 20.5 mV. The
smaller size of assemblies may be attributed to a better packing
of individual molecules in the self-assembly process, according
to our experience in previous work.[38] The morphology of RNS-2
Next, the disassembly process induced by reductive con-
ditions was verified through the changes in size and zeta
potential with time. Various assemblies were incubated with
10 × 10−3 m glutathione for different time period (0–3 h) at
37 °C. Once the disulfide bond cleaves, the size of assemblies
would significantly increase due to the changes in hydrophilic–
hydrophobic properties and aggregation of fragments.[49,50]
Figure 2. The morphology of A) RNS-2 assemblies and B) RNS-2/DNA (N/P = 30) complexes observed by transmission electron microscope.
C) Fluorescence emission spectra of RNF in methanol (black line), RNS-2 assemblies in methanol (orange line), RNS-2 assemblies in water (blue line),
RNS-2 assemblies incubating with 10 × 10−3 m glutathione (GSH, red line), and dithiothreitol (DTT, green line) in water for 1 h, excited by the laser with
750 nm wavelength.
©
Small 2020, 1906538
1906538 (4 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Figure 3. A) The size (blue histogram) and zeta potential (red curve) changes of various assemblies after incubation with 10 × 10−3 m glutathione.
Date are presented as means standard deviation (SD, n = 3). B) The gene condensation ability of assemblies with different N/P ratios (N/P = 0, 1,
2, 5, 10, and 20). C) The gene release ability of assemblies with 10 × 10−3 m glutathione (GSH) in HBG buffer (pH = 7.4) and DMEM culture media
containing 10% FBS.
These results showed that the order of reductive sensitivity was
RLS > RNS-2 > RNF (Figure 3A). The z-average size of RNF
assemblies remained invariant for about 2 h, while size of RLS
assemblies could rapidly respond within 30 min and diam-
eter of RNS-2 assemblies gradient increased within 1 h. The
longer response time of RNF or RNF-containing assemblies
might due to the rigidity structures of dyes, which reduced the
flexibility of assemblies like the function of cholesterol in lipo-
some.[51,52] Lower flexibility could prevent the penetration of
small molecule (like GSH). Naturally, the response speed of the
mixed assemblies (RNS-1, RNS-2, and RNS-3) increased with
increasing RLS ratio (Figure S10A,B, Supporting Information).
Meanwhile, the zeta potential of all assemblies went down
gradually and the variation trend was similar as that of size.
The results were consistent with previous reports that the zeta
potential of cationic assemblies would be sharply reduced with
destruction of the structure.[13,18]
verified by gel electrophoresis. It was clear that all assemblies
including RNF, RLS, and RNS-2 could efficiently condense
gene at low N/P ratio (Figure 3B). And then, the gene com-
plexes were incubated with 10 × 10−3 m glutathione for 2 h at
37 °C either in the HBG buffer (HEPES 20 mM, pH 7.4, 5%
glucose) or 10% fetal bovine serum (FBS) medium, in order
to simulate biological body fluid and emulate the environment
of blood fluid in vivo (Figure 3C). As expected, DNA could be
easily released from RLS assemblies in both two situations. In
the HBG buffer solution, gene was hardly released from the
RNF assemblies, while the release from RNS-2 assemblies was
inconspicuous. In 10% FBS medium, gene was still relatively
hard to release from RNF assemblies, while for RNS-2 assem-
blies a complete release was visible.[53] These results indicated
that addition of RNF had enhanced the stability of assemblies
but did not affect the gene release after reaching reductive con-
ditions in vivo. Meanwhile, the size changes of RNF, RLS, and
RNS-2/gene complexes in serum-containing medium were
detected, showing mild increasement to ≈350 nm with zeta
potential reduction (Figure S10C,D, Supporting Information),
Gene release is considerable vital for efficient transfec-
tion.[53] Whether the disassemble of particles would simulta-
neously result in a release of gene like the viral systems was
©
1906538 (5 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538

www.advancedsciencenews.com
www.small-journal.com
Figure 4. A) In vitro gene transfection effects on HeLa cells under the condition of culture medium without FBS or with 10% FBS. pEGFP plasmid
DNA were condensed by RNF, RNS-1, RNS-2, and RNS-3 assemblies at different N/P ratios (N/P = 10, 20, 30, and 40). Lipofectamine 2000 and RLS
assemblies were used as control groups. The N/P ratio of RLS/DNA complexes were 10. The scale bar was 200 µm. B) The semi-quantitative evaluation
of transfection efficiency in vitro on HeLa cell lines. Data are presented as means SD (n = 5), “*” p < 0.05, “**” p < 0.01, “–” no significant difference.
vdue to adsorption of negatively charged proteins and forma-
tion of protein corona on the surface of the particles.
Information). RNS-2 assemblies at the optimal N/P ratio have
similar efficiency as RLS, which showed much more bright
green fluorescence than the commercial Lipofectamine 2000
and other formulations (RNS-1 and RNS-3) on all cell lines
with or without serum (Figure 4A and Figure S12A,C,D, Sup-
porting Information). The differences between individual sys-
tems are more pronounced in the semi-quantitative analysis of
gene transfer (Figure 4B and Figure S12B, Supporting Informa-
tion). Under the serum-free conditions, HeLa cells transfected
by RNS-2 complexes (N/P ratio = 20) showed the best protein
expression, reaching up to 5.3-fold higher than that transfected
by Lipofectamine 2000 (**p < 0.01). While the superiority was
enhanced to sevenfold higher transfection for RNS-2 group
at N/P ratio of 30 (**p < 0.01) in 10% FBS, which was even
better than that under serum-free conditions. Compared with
other RNS groups, RNS-2 (N/P = 30) displayed approximately
tenfold and fivefold higher transfection efficiency than RNS-1
and RNS-3 (at the optimal N/P ratio) under serum-containing
conditions, respectively. Previous reports have found that rigid
2.3. In Vitro Cytotoxicity Assay and Transfection Efficiency
Cytotoxicity of carriers limits the applicability in drug or gene
delivery. Thus, cell counter kit-8 (CCK-8) assay was used to eval-
uate the biocompatibility of various assemblies against HeLa,
NIH3T3, and L929 cells (Figure S11, Supporting Information),
respectively. According to the assembly concentration of
20–80 µg mL−1 for in vitro gene transfection, the concentration
range was set between 10 and 100 µg mL−1. It was demonstrated
that all assemblies including RNF, RNS-1, RNS-2, and RNS-3
did not alter cell viability in the applied concentration range.
pEGFP as a plasmid for enhanced green florescent protein
expression was employed as model gene to assess the trans-
fection efficiency on Hela, HepG2, breast cancer cell line
MCF-7, and B16 cell lines (Figure 4 and Figure S12, Supporting
©
Small 2020, 1906538
1906538 (6 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
linkages could promote tighter and more stable formation and
offer sequent benefit for gene condensation.[54] That might
illustrate the obviously decreased green fluorescent protein
(GFP) expression by RNS-1 gene complexes from medium
without serum to 10% serum condition. However, on the other
hand, a tight condensation in the RNF assemblies would pre-
vent penetration of reductive molecules (Figure 3), and the
excessive rigidity of carriers (RNS-3) would decrease interaction
with cell membrane, hinder endocytosis, obstruct gene release,
and finally decrease transfection.[55] Overall, the RNS-2 assem-
blies exhibited the best gene transfection effect especially in cell
culture medium with 10% FBS.
scanning microscope (CLSM) with Hoechst33342 labeled
nucleus, Cy3 labeled gene, and NIR channel for recovered indo-
cyanine green observation (Figure 5A,B). At 2 h, some green
fluorescence from Cy3 channel and almost no NIR fluores-
cence could be detected inside cells, suggesting the successful
cellular uptake and ACQ effect for intact carrier systems. With
time ongoing, the fluorescent signal of Cy3 became stronger,
indicating more and more gene complexes entering into the
living cells, ensuring efficient internalization of RNS-2/DNA
complexes. Meanwhile, the signal of NIR fluorescence from
RNS-2 was also detectable and enhanced from 4 to 8 h, owing
to the fast cleavage of disulfide bonds and disintegration of
assemblies. As shown in the initial 4–6 h images, orange flu-
orescent spots obviously appeared in the merged channel, as
a result of an overlay of Cy3-labeled pGL3 (green) and near-
infrared fluorescence from cleaved indocyanine green analogs
(red). Subsequently, at 8 h, some of the pattern turned red and
accumulated around the nuclear periphery, due to the prox-
imity of the nucleus by endosome migration and successful
2.4. In Vitro Imaging for Disintegration of Assemblies
In order to trace cellular uptake behavior and study fluores-
cence recovery inside the tumor cells, fluorescence imaging
was carried out on HeLa and HepG2 cells by confocal laser
Figure 5. In vitro imaging with different fluorescent labeling by confocal laser scanning microscope (CLSM). Incubation of RNS-2 assemblies with
Cy3-labeled plasmid DNA (N/P = 30) on A) HeLa and B) HepG2 cells for 2–8 h. Green: fluorescent signal of Cy3. Red: fluorescent signal of NIR. Blue:
fluorescent signal of Hoechst33342 stained for nucleus. C) HeLa cells incubated with RLS/Cy3 and Cy5-DNA (N/P = 10) for 2–8 h. Cy3 and Cy5 were
used as donor and acceptor of fluorescence resonance energy transfer (FRET) effect. Green: fluorescent signal of Cy3 (FRET donor Cy3). Red: FRET-
mediated Cy5 fluorescent signal. Blue: signal of Hoechst33342 for cell nucleuses staining. The scale bar was 25 µm. Analysis of correlation between
various fluorescent signals with the dynamic process of gene complexes disintegration/gene release by D) heatmap, E) histogram, and F) fitted curve
analysis. Data are presented as means SD (n = 5).
©
1906538 (7 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538

www.advancedsciencenews.com
www.small-journal.com
gene release from the compression situation. On the contrary,
fluorescein isothiocyanate (FITC) labeled PEI together with Cy3
labeled gene complexes constantly showed well colocalization
of Cy3 and FITC signals due to tight condensation of DNA by
PEI during the time period (Figure S13A, Supporting Informa-
tion). It was more intuitively that the semi-quantified R (Man-
der’s overlap coefficient) value was gradually decreased over
time, inflecting the colocalization of DNA and RNS-2 carriers
more directly (Figure S13B,C, Supporting Information) (from
0.94 to 0.59 on Hela cells and 0.96 to 0.54 on HepG2 cells,
**p < 0.01).
2.5. In Vivo Imaging and Gene Transfection
The in vivo imaging of the RNS-2 assemblies without DNA
was evaluated before (Figure S15A, Supporting Information)
and after intratumoral injection (0.5–24 h) (Figure S15B–E,
Supporting Information). As expected, the assemblies in
aqueous solution did not show fluorescent signal, while addi-
tion of methanol led to a disassembly of aggregates and strong
detectable fluorescence (Figure S15A, Supporting Informa-
tion). In the animal experiment, the region of interest was
set at tumor based on the bright field. It was observed that
the fluorescence intensity was gradually enhanced with time
(Figure S15B, Supporting Information). At the beginning
(before 1 h), there was weak fluorescence signal at tumor site,
while it reached the highest intensity at 9 h and maintained
strong until 24 h postinjection (Figure S15B,D, Supporting
Information). During the whole imaging process, other tissues
apart for tumor showed hardly any signal, which was addition-
ally confirmed by the ex vivo imaging of the isolated tissues
at 24 h end point of the study (Figure S15C,E, Supporting
Information).
Above results have underlined the desired fluorescence
“Off” and “On” property of RNS-2 in vivo and arouse further
studies on the “signal to noise” (STN) for the gene delivery
imaging. It was clearly visible that signal from Cy5 channel
was relatively dispersed, while that from the NIR channel was
more localized at tumor (Figure 6A). Considering the similar
metabolic pathway of Cy5 and indocyanine green deriva-
tives,[56,57] the disseminated signal from Cy5 channel that out
of the tumor site may result from degraded DNA-Cy5 frag-
ments. A semi-quantificative comparing analysis of the signal
from tumor to total signal from the whole body could give
a preliminary assessment about the “signal to noise” of Cy5
and ICG analogs, respectively (Figure 6B). The STN values
of NIR always remained very high (0.85–0.93) from 3 to 24 h
after injection, which means almost no signal was detected
outside the tumor. While the STN value of Cy5 was much
lower (0.13–0.30), implying serious interference with back-
ground fluorescence. From the isolated tissues imaging at
24 h, obvious strong signal was detected from Cy5 channel in
normal tissues (heart, liver, spleen, lung, and kidney), but rel-
atively weak from the NIR channel (Figure 6C). Similarly, in
the semi-quantificative assessment, tumor showed up to 200-
fold higher fluorescent intensity than other tissues from NIR
channel, while only 30-fold was obtained from Cy5 channel
(Figure 6D,E).
In previous studies, we exploited Cy3 and CY5 double labeled
DNA as a donor/acceptor with FRET effect to investigate the
gene release process inside cells.[15,18] The dynamic change in
fluorescent signals (FRET Cy3 and FRET Cy5) could timely
indicate the compression or decompaction status of gene, while
the only drawback was the intensity of FRET Cy5 existed nega-
tive relationship with the amount of gene release. Herein, we
speculated the carriers with ACQ effect and reductive triggered
fluorescent recovery can provide a better detectable and visual
strategy for the bioimaging. RLS assemblies (without indocya-
nine green modification) and PEI were used as contrasts for
condensation of Cy3 and Cy5 labeled gene (N/P = 10), respec-
tively (Figure 5C and Figure S13E, Supporting Information).
The fluorescence of Cy3 was pseudo-colored green to distin-
guish it from the red fluorescence of FRET-mediated Cy5. In
the RLS/Cy3 and Cy5 double-labeled DNA group (Figure 5C),
we could observe the strong signal of Cy5 and very weak signal
of Cy3 at the beginning (2 h), due to the close distance of donor
and acceptor fluorescence in condensed DNA. As time went on,
DNA was gradually released from the complexes, the signal of
Cy3 became stronger and that of Cy5 get dimmer, attributing
to the weakened FRET effect. The semi-quantified R value
for merged images of Cy3 and Cy5 always kept constant
(Figure S13D, Supporting Information). On the other hand,
strong FRET-mediated Cy5 signal could be observed for a long
time in the PEI group, which demonstrated the tight condensa-
tion status of DNA (Figure S13E, Supporting Information). The
comparison of the fluorescent signal changes from confocal
laser scanning microscope imaging (Figure 5A,C) was visualized
in a heatmap (Figure 5D) and histogram analysis (Figure 5E).
Heatmaps represented the level of fluorescent intensity, where
green and red color represented the relative fold change of fluo-
rescent intensity decrease and increase, respectively. We could
clearly see that the signal intensity of NIR, Cy3, and FRET-Cy3
dyes were enhanced with time, while the signal of FRET-Cy5
reduced gradually. It was exciting that the recovered intensity
of NIR dye was the strongest among them with ≈4.2-fold higher
increasement than at the beginning, while the signals of Cy3
and FRET-Cy3 only showed about ≈1.5-fold enhancement. On
the other hand, when the change of the FRET-Cy5 signal exhib-
ited a negative correlation with time, the relationship between
NIR dye and time was positive (Figure 5F). The fitted curve
analysis of Cy3 and FRET-Cy3 were presented in Figure S14A–D
in the Supporting Information. In conclusion, the fluorescent
recovery of indocyanine green analogs could appropriately mon-
itor disintegration of assemblies, trace gene release, and allow
for a much visualization of the dynamic process compared with
Cy3 and Cy5 double labeled FRET indicator.
Such high signal-to-noise radio has confirmed the precision
advantages of the developed indocyanine green analog based
gene carrier, the further accuracy for the gene expression loca-
tion remained to be addressed. Therefore, pEGFP (Figure 7A–C)
or Cy3 labeled pGL3 (Figure 7D–F) was condensed by RNS-2
assemblies at N/P = 30 and the mixed gene complexes were
intratumorally injected, respectively. Isolated tissues from
the sacrificed mice were observed by Cambridge Research &
Instrumentation (CRI) Maestro Imaging System (Figure 7A,D)
and semi-quantitative analysis was performed 48 h later
(Figure 7C,E). Considerable green fluorescence signal from
the GFP channel could be observed in the isolated tumor
(Figure 7A), indicating pronounced protein expression and
©
Small 2020, 1906538
1906538 (8 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Figure 6. Fluorescence imaging for the RNS-2/Cy5-DNA complexes (N/P = 30) through intratumoral injection. A) In vivo imaging detected from Cy5
and NIR channel at different time after injection, respectively. The tumor site was marked by red circle. B) The semi-quantificative analysis of signal
to noise of Cy5 and ICG analogs by comparing the Cy5 and NIR signals from tumor with total signal from the whole body, respectively. C) The 24 h
imaging of isolated tissues detected from Cy5 and NIR channel after injection (heart, liver, spleen, lung, kidney, and tumor), respectively. The semi-
quantitative analysis for the fluorescence intensity of D) Cy5 signal and E) NIR signal in isolated tissues. Data are presented as means SD (n = 5).
efficient gene transfection by RNS-2 complexes. Other tissues
(including heart, liver, spleen, lung, and kidney) showed hardly
detectable signals due to poor gene expression. Excitingly,
fluorescence signal from NIR channel corresponded well to
the result from GFP channel, providing accurate location for
gene expression. On the other hand, when we observed the
RNS-2/Cy3-DNA complexes (N/P = 30) group, not only tumor
but also other organs (especially liver, lung and kidney) showed
obvious signals from channel Cy3 (Figure 7D,F). That means
Cy3-labeled DNA probably provide a false positive indication for
the gene expressed site. The frozen sections of isolated organs
also showed the same results (Figure 7B,E). Previous studies
have found direct gene labeling could reduce the transfection
by altering the conformation of nuclear acid,[58] while physical
encapsulation or peripheral residues bonding of probe to the
nanocarrier may be suffered from leaking.[30] Our designed
RNF molecules, in which the fluorescence dye was used as
skeleton structure of carriers has proved to be a more accurate
indication strategy.
2.6. Biodistribution and Biocompatibility Assays
In order to assess the biodistribution/premature leakage of
assemblies in vivo, the complexes of RNS-2 assemblies con-
densing 15 µg Cy5-labeled DNA with N/P = 30 were intrave-
nously injected into mice. From 5 to 24 h after injection, Cy5
signals were dispersed and could observed in liver, lung, spleen,
and kidney, while NIR signal mainly concentrated on the liver,
indicating that the ICG-based assemblies did not leakage pre-
maturely (Figure S16, Supporting Information).
The in vivo biocompatibility/biosafety was evaluated by acute
toxicity test via intravenously injecting high dose of RNS-2/
DNA complexes. All indicators in both hematological analysis
(Table S1, Supporting Information) and serum biochemistry
analysis (Table S2, Supporting Information) were in the normal
range, showing no obvious difference with saline group. In
addition, obvious pathological changes could not detectable in
the major organs by hematoxylin and eosin (H&E) staining
(Figure S17, Supporting Information).
©
1906538 (9 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538

www.advancedsciencenews.com
www.small-journal.com
Figure 7. In vivo green fluorescent protein expression and fluorescent probe signals at 48 h after intratumoral injection of A–C) RNS-2/pEGFP or
D–E) RNS-2/Cy3-labeled pGL3, respectively (N/P = 30). A,D) The fluorescence imaging of isolated tissues from different channels (GFP, NIR, and Cy3
channels) observed by CRI Maestro Imaging System, respectively. B,E) Frozen section imaging of normal tissues and tumor from GFP, NIR, and Cy3
channels, respectively (observed by CLSM). The scale bar was 200 µm. C,F) Semi-quantitative analysis of the fluorescence intensity of isolated tissues,
respectively. Data are presented as means SD (n = 5).
exhibited activable “OFF” and “ON” properties, which being
quenched by ACQ effect in the assembly state and light up in
3. Conclusion
In this study, we have introduced a novel molecular design con-
cept for theranostic gene delivery system. We report the syn-
thesis and optical properties of a novel amphiphile named RNF,
which rationally integrated indocyanine green analogs and argi-
nine-rich dendritic polypeptides in a single defined molecule.
By assembly with other lipopeptides at optimal conditions, a
theranostic system (RNS-2) for gene delivery and fluorescence
imaging can be constructed. The obtained gene delivery system
possessed nanoscaled size around 106 nm, ellipsoidal shape, and
zeta potential of +20 mV. The RNS-2/pDNA complexes exhibited
significant variation in size distribution and DNA liberation
once disassembling in the presence of reducing agents as visual-
ized by the dynamic laser scanning and gel electrophoresis assay.
Meanwhile, the fluorescent signal of the RNS-2 assemblies
the presence of GSH. In vitro transfection confirmed that RNS-2
gene complexes achieved excellent gene transfection efficacy
in different cell lines (HeLa, HepG2, MCF7, and B16 cells) at
noncytotoxic concentration. Observation from confocal laser
scanning microscopy showed that the fluorescent recovery could
appropriately indicate disintegration of assemblies and gene
release, which enable tracing of the dynamic processes of gene
delivery. In vivo imaging showed that RNS-2 assemblies lead to a
highly localized signal with enhanced signal-to-noise ratio com-
pared to conventional FRET pairs. Therefore, not only efficient
gene transfer in vivo has been confirmed but also imaging accu-
racy for gene expression location could be ensured. Collectively,
RNS-2 assemblies could be a promising theranostic system for
efficient local gene delivery and precise real-time imaging.
©
Small 2020, 1906538
1906538 (10 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
dried with anhydrous magnesium sulfate and concentrated by rotary
evaporation. The obtained crude product was purified by silica column
chromatography to obtain compound 4 as green solid (0.4 g, 27%).
Compound 5 (0.14 g, 0.54 mmol) was prepared as the previous work[15,38]
and compound 4 (0.4 g, 0.36 mmol) were dissolved in 10 mL anhydrous
DMF. Appropriate amount of TBTU (0.23 g, 0.72 mmol) and DIEA
(0.09 g, 0.72 mmol) were added into the above solution. The mixture
was stirred at 30 °C for 48 h under nitrogen atmosphere in dark for
complete reaction. After that, DMF was removed in vacuum, and other
100 mL dichloromethane was added for redissolution and successively
washed by saturated sodium bicarbonate, sodium dihydrogen
phosphate, and brine, respectively. The combined organic phase was
dried with anhydrous magnesium sulfate and concentrated in vacuum.
The crude product was purified by silica column chromatography to
obtain compound 6 as green solid (0.3 g, 63%). Compound 6 (0.3 g,
0.22 mmol) was dissolved in 10 mL dichloromethane together with
the same volume of trifluoroacetic acid, stirring in ice-water bath. And
then the mixture was stirred at 30 °C for 5 h in dark. The redundant
solvent was removed in vacuum. The product was precipitated by diethyl
ether and concentrated by centrifugation. The residual diethyl ether
was removed to obtain compound 7 as green solid (0.51 g, 91%). 1H
NMR (400 MHz, CDCl3), δ (ppm): 7.93 (d, J = 12.0 Hz, 2H), 7.34 (d,
J = 7.6 Hz, 2H), 7.32 (d, J = 8.3 Hz, 2H), 7.22 (d, J = 7.6 Hz, 2H), 7.03
(d, J = 7.9 Hz, 2H), 6.92 (d, J = 8.6 Hz, 2H), 5.94 (d, J = 14.2 Hz, 2H),
3.93 (t, J = 6.0 Hz, 4H), 3.45 (m, 2H), 3.27 (m, 2H), 2.93 (t, J = 6.0 Hz,
4H), 2.65 (m, 6H), 2.04 (t, J = 4.0 Hz, 2H), 1.78 (m, 2H), 1.33 (s, 12H),
1.25 (s, 56H), 0.88 (t, J = 6.0 Hz, 6H). MS (MALDI-TOF-MS): calcd. For
4. Experimental Section
Materials: All chemicals and solvents were reagent grade.
2,3,3-Trimethylindolenine, 1-bromopropane, cyclohexanone, sodium
acetate anhydrous, phloretic acid, di-tert-butyl dicarbonate, l-lysine methyl
ester dihydrochloride, and TBTU were purchased from Aladdin (Shanghai,
China). Cystamine dihydrochloride, anhydrous N,N-dimethylformamide
(DMF), and anhydrous dimethyl sulfoxide were purchased from J&K
Scientific Ltd. (Beijing, China). Acetic anhydride and 1-bromohexadecane
were purchased from Adamas-beta (Shanghai, China). Trifluoroacetic acid
(TFA) and DIEA were purchased from Astabiochem Ltd. (Chengdu, China).
Boc-Arg(Pbf)-OH was purchased from GL Biochem Ltd. (Shanghai,
China). CCK-8 was purchased from Dojindo Laboratories (Kumamoto,
Japan). The nucleic acid labeling kit Label IT Cy3 and Cy5 were obtained
from Mirus. Hoechst 33342 and FITC were purchased from Solarbio S&T
Co. (Beijing, China). l-Glutathione (GSH) was purchased from Beyotime
(Nantong, China). Dulbecco’s modified Eagle’s medium (DMEM) with
high glucose, penicillin, streptomycin, and FBS were purchased from
Hyclone (USA). Lipofectamine 2000 and SYBR Safe DNA Gel Stain were
obtained from Thermofisher (USA). All reagents were purchased from
commercial suppliers and used without further purification unless other
stated. Deionized water and triple distilled water were used throughout
the experiments. NMR spectra were recorded at 1H (400 MHz) on a
Bruker instrument. Chemical shifts (δ values) and coupling constants
(J values) are given in ppm and hertz, respectively, using solvents as the
internal standard. Mass spectral analyses were performed on an ABI API
3200 (electrospray ionization-Triple Quadruple).
Synthesis of Oleic Acid-Containing (RLS) and Near-Infrared (NIR)
Fluorescence Dye-Containing (RNF) Amphipathic Dendritic Lipopeptide:
The oleic acid-containing amphipathic dendrimer (RLS) was synthesized
according to the previous work,[15,38,59] and the NIR fluorescence dye-
containing amphipathic dendrimer (RNF) was prepared according to
the classic method with some modification. Briefly, 2,3,3-trimethylindolenine
(5 g, 30 mmol) and 1-bromohexadecane (14.3 g, 47 mmol) were dissolved
in toluene (100 mL). And then the reaction was heated to 130 °C and
stirred for 72 h under nitrogen atmosphere. The solvent was concentrated
under reduced pressure to give the crude product. The crude product was
purified by silica column chromatography to yield compound 1 (10.77 g,
37%). The solution of phosphorus oxychloride (30.6 g, 200 mmol) in
8 mL dichloromethane was slowly added into anhydrous DMF (20 mL,
14.6 g, 200 mmol) with stirring in ice-water bath. After that, cyclohexanone
(5 g, 50 mmol) was added into the solution and the mixture was refluxed
overnight under nitrogen atmosphere. The mixture was cooled and slowly
poured into 100 mL ice-cold water. The solution was kept in −20 °C to yield
precipitation (compound 2, yellow solid with 5.86 g, 68% yield). Next, a
mixture of compound 1 (3 g, 6.4 mmol), compound 2 (0.44 g, 2.6 mmol),
and sodium acetate anhydrous (0.79 g, 7.7 mmol) were dissolved in 20 mL
acetic anhydride. The reaction was stirred overnight at 65 °C under nitrogen
atmosphere in dark. After that, the mixture was dissolved in 300 mL
dichloromethane and washed by saturated brine (3 × 50 mL). The organic
phase was dried with anhydrous magnesium sulfate and concentrated in
vacuum to give crude product. The crude product was purified by silica
column chromatography to yield compound 3 as green solid (1.87 g, 74%).
1H NMR (400 MHz, CDCl3), δ (ppm) (Figure S1, Supporting Information):
8.36 (d, J = 12.0 Hz, 2H), 7.38 (d, J = 7.5 Hz, 2H), 7.26 (d, J = 7.6 Hz, 2H),
7.18 (d, J = 7.9 Hz, 2H), 6.29 (d, J = 16.0 Hz, 2H), 4.21 (t, J = 6.0 Hz, 4H),
2.74 (t, J = 6.0 Hz, 4H), 1.99 (m, 2H), 1.72 (s, 12H), 1.25 (s, 56H), 0.87 (t,
J = 6.0 Hz, 6H). Matrix-assisted laser desorption/ionization-time of flight-
mass spectrum (MALDI-TOF-MS): calcd. For C62H96ClN2+ m/z 903.73 [M];
found m/z 903.78 [M].
C
75H115N4O2S2+ m/z 1167.85 [M]; found m/z 1167.88 [M].
Compound 8 (0.7 g, 0.6 mmol) prepared according to previous
work[15,38] and compound 7 (0.5 g, 0.4 mmol) were dissolved in 10 mL
anhydrous DMF. Appropriate amount of TBTU (0.26 g, 0.8 mmol) and
DIEA (0.1 g, 0.8 mmol) were added. The mixture was stirred at 30 °C for
48 h under nitrogen atmosphere in dark. The residual DMF was removed
in vacuum and then 100 mL of dichloromethane was added. The mixture
was washed with aqueous solution as mentioned above. The organic
phase was dried with anhydrous magnesium sulfate and concentrated
by rotary evaporation. The crude product was purified by silica column
chromatography to obtain compound 9 as green solid (0.67 g, 70%).
Then compound 9 (0.17 g, 0.07 mmol) was removed the protective
groups of amino to obtain compound 10 as green solid (0.095 g, 79%).
1H NMR (400 MHz, CD3OD), δ (ppm): 8.03 (d, J = 16.0 Hz, 2H), 7.42
(d, J = 7.4 Hz, 2H), 7.38 (d, J = 8.0 Hz, 2H), 7.28 (d, J = 7.8 Hz, 2H),
7.23 (d, J = 7.5 Hz, 2H), 7.07 (d, J = 8.2 Hz, 2H), 6.18 (d, J = 12.0 Hz,
2H), 4.32 (m, 1H), 4.12 (t, J = 8.0 Hz, 4H), 3.50 (m, 4H), 3.42 (m, 2H),
3.25 (m, 6H), 2.87 (t, J = 6.0 Hz, 4H), 2.75 (m, 6H), 2.07 (t, J = 6.0 Hz,
2H), 1.82 (m, 8H), 1.43 (m, 6H), 1.36 (s, 12H), 1.30 (s, 58H), 0.90 (t,
+
J = 6.0 Hz,6H). MS (MALDI-TOF-MS): calcd. For C93H151N14O5S2 m/z
1609.15 [M]; found m/z 1609.34 [M].
Self-Assembly and Characterization of Nanoparticles: The self-assembly
nanoparticles RLS, RNF, RNS-1 (RLS/RNF = 1/1, m/m), RNS-2 (RLS/
RNF = 2/1, m/m), and RNS-3 (RLS/RNF = 3/1, m/m) were prepared by
injection method. Briefly, 1 mg of sample in 100 µL dimethylsulfoxide
was dropped into 1 mL ultrapure water with rapid stirring. The assembly
solution was obtained after 1 h and the absorption spectra were
performed on a UV–vis spectrophotometer (HITACHI, U-2910, Japan)
at room temperature in methanol. The excitation and emission spectra
in methanol and water were detected by fluorescence spectrometer
(HITACHI, F-7000, Japan), respectively. The size and zeta potential were
detected by DLS with a Zetasizer Nano ZS (Malvern, Worcestershire, UK)
at 25 °C, and the morphologies of every assemblies were observed by
transmission electron microscopy (JEM-2100 Plus, JEOL, Japan).
Compound
3 (1.4 g, 1.42 mmol) was dissolved in 20 mL
dimethylsulfoxide and the phloretic acid sodium salt (1.5 g, 7.14 mmol)
was added. And then the reaction was heated to 65 °C for 5 h in dark.
When it was cooled to room temperature, 100 mL dichloromethane
was added. The mixture was washed by saturated brine and the organic
phase was concentrated in vacuum. The precipitate was redissolved
in methanol and the pH was adjusted to 2 with 1 m hydrochloric acid.
100 mL dichloromethane was added and the mixed solution was
washed twice by saturated brine. Organic phase was combined and
Gene Condensation and Release: The assemblies were mixed with
pDNA (200 ng) at various N/P ratios (N/P = 0, 1, 2, 5, 10, 20) to form
the gene complexes after 20 min incubation at room temperature. The
gel electrophoresis assay was used to evaluate the gene condensation
and release ability of assemblies under the condition of with or without
10 × 10−3 m glutathione. All samples were loaded into 1% agarose gel
for electrophoresis (85 V, 1 h). The gels were stained with SYBR and
visualized by the Molecular Imager ChemiDoc XRS+ (Bio-Rad, USA).
©
1906538 (11 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538

www.advancedsciencenews.com
www.small-journal.com
In Vitro Cytotoxicity: The cytotoxicity of assemblies was tested by
Cell Counting Kit-8 (CCK-8) assay. Briefly, HeLa, NIH3T, and L929
cells were seeded into 96-well plates with a density of 1 × 104 per well
and incubated for 24 h, respectively. The assemblies with different
concentrations were added for another 4 h coincubation. After then, the
media were replaced with fresh media containing 10% serum. Twenty
four hours later, the cell culture medium was removed and the cells were
washed with phosphate buffered saline (PBS) buffer (pH 7.4). Following,
the cells were incubated with fresh serum-free media containing 10%
CCK-8 for 2 h, and the absorbance was detected by microplate reader
(Bio-Rad 550, USA) under 450 nm.
In Vitro Transfection Efficiency: In vitro gene transfection efficiency was
detected on various cells, including HeLa, HepG2, MCF7, and B16 cells,
respectively. The cells were seeded in 96-well plates with a density of
1 × 104 cells per well. After 24 h, the medium were removed and 100 µL
of fresh medium with or without 10% serum were added. Simultaneously,
complexes of gene/assemblies containing 200 ng pEGFP were added per
well with different N/P ratios (10, 20, 30, and 40). After 4 h incubation
at 37 °C, the media were replaced by the fresh DMEM containing 10%
serum, and the green fluorescent protein expression was observed by
fluorescence microscope (Leica, Germany) after other 44 h. Lipofectamine
2000 and RLS assemblies were acted as control groups.
Intracellular Fluorescence Recovery: The fluorescence recovering assay
in cells was observed by CLSM (Leica TCS SP5, Germany). HeLa cells or
HepG2 cells were seeded at a density of 1 × 104 cells per well in 35 mm
confocal dish (Φ = 15 mm) and incubated overnight for attachment. Then
the medium was replaced by fresh complete culture medium containing
RNS-2/DNA complexes with N/P = 30 (300 ng of Cy3-labeled pGL3). After
coincubation for different time (2, 4, 6, and 8 h), the cells were washed
by cold PBS (pH = 7.4) twice to remove the assemblies which did not be
taken. Then the nucleus was stained by Hoechst 33342 for 15 min prior to
observation. The modified NIR dyes were excited by 663 nm laser and the
channel (dichroic mirror 700–800 nm) was used to detect the emission
signals. The intracellular localization of DNA complexes was observed by
excitation wavelength at 545 nm laser and the emission signals of Cy3
were received in Cy3 channel (dichroic mirror 560–620 nm). The FITC-
labeled PEI condensed Cy3-labeled pGL3 gene complexes (N/P = 10) were
set as control. FITC was excited at 488 nm and the emission wavelength
of was 520 nm. The excitation and emission wavelengths of Hoechst
33342 were 350 and 461 nm, respectively.
shortest (W) diameters. The volume of tumor was calculated by formula
of V [mm3] = 1/2 × LW2. When the tumor reached 150–200 mm3, the
nude mice were randomly divided into two groups (n = 3), which received
gene/assemblies complexes treatment. For intratumoral injection, mice
were treated with 65 µg RNS-2 assemblies. Moreover, the complexes of
RNS-2/Cy5-labeled DNA with dosage of 5 µg DNA were also observed
in vivo by intratumoral injection. After injection for 24 h, the mice were
sacrificed and the isolated tissues containing heart, liver, spleen, lung,
kidney, and tumor were observed by CRI Maestro Imaging System.
For in vivo transfection efficiency, when the volume of tumor grew
to 150–200 mm3, the nude mice were randomly divided into two
groups (n = 5). They were treated with the RNS-2/pEGFP complexes
(N/P = 30) and RNS-2/Cy3-pGL3 complexes (N/P = 30) with a dosage
of 15 µg DNA through intratumoral injection, respectively. The mice
were sacrificed after 48 h and the isolated tissues including liver, lung,
kidney, and tumor were washed with ice-cold PBS buffer (pH = 7.4).
These isolated tissues were observed by CRI Maestro Imaging System
first and then were immediately froze and cut into 10 µm thick sections.
These frozen sections were observed by fluorescence microscope
(Leica, Germany).
In Vivo Distribution and Biocompatibility: In order to assess the
premature leakage of assemblies and distribution in vivo, the complexes
of RNS-2 assemblies condensing 15 µg Cy5-labeled DNA with N/P = 30
were intravenously injected into mice. The premature leakage profile was
detected using CRI Maestro Imaging System. At 24 h postintravenous
injection, the mouse organs were removed. Fluorescence signals for Cy5
and NIR were detected using CRI Maestro Imaging System, respectively.
A high dose of RNS-2/pEGFP complexes (N/P = 30) with a dosage
75 µg DNA was injected via tail vein using saline treated as control group.
After 48 h, blood was collected through retro-orbital bleeding in the two
treatment groups of nude mice. Plasma was obtained by centrifugation at
3000 rpm for 15 min twice for blood biochemistry analysis. The plasma
levels of aspartate aminotransferase and alanine aminotransferase were
measured to evaluate liver function. Renal function was investigated
by quantifying blood urea nitrogen and creatinine levels. Blood routine
and serum biochemical indicators were measured by auto hematology
analyzer (TEK-II MINI). Organs including heart, liver, spleen, lung, and
kidney were drawn out, fixed immediately in paraformaldehyde and
embedded in paraffin, and serially cut for histopathological examination
by the hematoxylin and eosin (H&E) staining.
Intracellular Gene Release: The FRET was used to observed
intracellular gene release by confocal laser scanning microscope
(CLSM, Leica TCS SP5, Germany). HeLa cells were seeded in a 35 mm
confocal dish (Φ = 15 mm) with a density of 1 × 104 cells per well and
incubated overnight and continue to incubate with RLS/DNA complexes
with N/P = 10 (300 ng of Cy3 and Cy5 colabeled pGL3) in complete
culture medium. After coincubation for different time (2, 4, 6, and 8 h),
the cells were washed with PBS (pH = 7.4) to remove the assemblies
which did not be taken. Then the nucleus was stained by Hoechst
33342 for 15 min prior to observation. The excitation wavelength of Cy3
was 545 nm and the emission signals of Cy3 and Cy5 were observed
in the Cy3 (dichroic mirror 565–620 nm) and Cy5 (dichroic mirror
655–755 nm) channels, respectively. The excitation and emission
wavelengths of Hoechst 33342 were 350 and 461 nm, respectively. PEI
condensed Cy3 and Cy5 colabeled pGL3 gene complexes were also used
as control.
Statistical Analysis: Data were performed at least in triplicate and
presented as the mean standard deviation (mean S. D.) for three or
five independent samples. Statistical analysis was carried out by analysis
of variance tests with Microsoft office software. The relative fold change
of fluorescent intensity was pseudocolored using the heatmap feature in
GraphPad Prism 7.03.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
In Vivo Imaging and Transfection Efficiency: In vivo imaging was
observed by CRI Maestro Imaging System (Cambridge Research
and Instrumentation Inc., USA). All the animal experiments were
conducted with the approval of the Institutional Animal Care and Use
Committee of the Animal Experiment Center of Western China Second
University Hospital (Chengdu, China) as well as the Regulations for
the Administration Affairs Concerning Experimental Animals. Healthy
BALB/c nude mice (18–20 g) were lived in a pathogen-free environment
according to NIH guidance and fed with sterilized food and distilled
water. HeLa cells (1 × 107) were injected into the right flank region of
female Balb/c mice to form a xenograft tumor model. The vernier caliper
was exploited to measure the size of tumor for the longest (L) and
Acknowledgements
H.L. and X.C. contributed equally to this work. This study was
supported by the National Key Research and Development Plan of
China (2017YFC1104601), the National Natural Science Foundation of
China (NSFC, Nos. 81571794 and 81873921), Sino-German cooperation
group project (GZ1512) and Sichuan Science and Technology Program
(2019JDJQ0027 and 2019JDJQ0004), the Fundamental Research Funds
for the Central Universities (2012017yjsy222) and Spark innovation
Project of Sichuan University (2018SCUH0045 and 2018SCUH0033),
Miaozi Project in Science and Technology Innovation Program of
Sichuan Province (2019071).
©
Small 2020, 1906538
1906538 (12 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
[26] S. U. Hettiarachchi, B. Prasai, R. L. McCarley, J. Am. Chem. Soc.
2014, 136, 7575.
Conflict of Interest
The authors declare no conflict of interest.
[27] R. Blau, Y. Epshtein, E. Pisarevsky, G. Tiram, S. Israeli Dangoor,
E. Yeini, A. Krivitsky, A. Eldar-Boock, D. Ben-Shushan,
H. Gibori, A. Scomparin, O. Green, Y. Ben-Nun, E. Merquiol,
H. Doron, G. Blum, N. Erez, R. Grossman, Z. Ram, D. Shabat,
R. Satchi-Fainaro, Theranostics 2018, 8, 3437.
Keywords
gene delivery, indocyanine green analogues, near-infrared fluorescence
imaging, precise real-time image, theranostic systems
[28] Q. Liu, S. Xu, C. Niu, M. Li, D. He, Z. Lu, L. Ma, N. Na, F. Huang,
H. Jiang, J. Ouyang, Biosens. Bioelectron. 2015, 64, 119.
[29] H. Kim, G. Kwak, K. Kim, H. Y. Yoon, I. C. Kwon, Biomaterials 2019,
213, 119207.
[30] M. Gao, F. Yu, C. Lv, J. Choo, L. Chen, Chem. Soc. Rev. 2017, 46, 2237.
[31] L. Quan, S. Liu, T. Sun, X. Guan, W. Lin, Z. Xie, Y. Huang, Y. Wang,
X. Jing, ACS Appl. Mater. Interfaces 2014, 6, 16166.
[32] X. Liu, B. Chen, X. Li, L. Zhang, Y. Xu, Z. Liu, Z. Cheng, X. Zhu,
Nanoscale 2015, 7, 16399.
Received: November 12, 2019
Revised: December 17, 2019
Published online:
[1] T. Frohlich, D. Edinger, R. Klager, C. Troiber, E. Salcher,
N. Badgujar, I. Martin, D. Schaffert, A. Cengizeroglu, P. Hadwiger,
H. P. Vornlocher, E. Wagner, J. Controlled Release 2012, 160, 532.
[2] X. Guan, Z. Guo, T. Wang, L. Lin, J. Chen, H. Tian, X. Chen,
Biomacromolecules 2017, 18, 1342.
[3] Y. Li, Y. Li, X. Zhang, X. Xu, Z. Zhang, C. Hu, Y. He, Z. Gu,
Theranostics 2016, 6, 1293.
[4] X. Cai, R. Jin, J. Wang, D. Yue, Q. Jiang, Y. Wu, Z. Gu, ACS Appl.
Mater. Interfaces 2016, 8, 5821.
[33] Z. Guo, Y. Zou, H. He, J. Rao, S. Ji, X. Cui, H. Ke, Y. Deng, H. Yang,
C. Chen, Y. Zhao, H. Chen, Adv. Mater. 2016, 28, 10155.
[34] H. J. Yoon, H. S. Lee, J. H. Jung, H. K. Kim, J. H. Park, ACS Appl.
Mater. Interfaces 2018, 10, 6118.
[35] H. Pan, C. Zhang, T. Wang, J. Chen, S.-K. Sun, ACS Appl. Mater.
Interfaces 2019, 11, 2782.
[36] U. Lächelt, E. Wagner, Chem. Rev. 2015, 115, 11043.
[37] Y. Tu, F. Peng, A. Adawy, Y. Men, L. K. Abdelmohsen, D. A. Wilson,
Chem. Rev. 2016, 116, 2023.
[5] J. Li, C. Zheng, S. Cansiz, C. Wu, J. Xu, C. Cui, Y. Liu, W. Hou,
Y. Wang, L. Zhang, I. T. Teng, H. H. Yang, W. Tan, J. Am. Chem. Soc.
2015, 137, 1412.
[6] X. Liu, J. Xiang, D. Zhu, L. Jiang, Z. Zhou, J. Tang, X. Liu, Y. Huang,
Y. Shen, Adv. Mater. 2016, 28, 1743.
[7] Y. Li, H. Bai, H. Wang, Y. Shen, G. Tang, Y. Ping, Nanoscale 2017, 10,
203.
[8] M. Naito, N. Yoshinaga, T. Ishii, A. Matsumoto, Y. Miyahara,
K. Miyata, K. Kataoka, Macromol. Biosci. 2018, 18, 1700357.
[9] Z. Zhou, M. Zhang, Y. Liu, C. Li, Q. Zhang, D. Oupicky, M. Sun,
Biomacromolecules 2018, 19, 3776.
[10] C. Y. Quan, J. X. Chen, H. Y. Wang, C. Li, C. Chang, X. Z. Zhang,
R. X. Zhuo, ACS Nano 2010, 4, 4211.
[11] Y. Nie, D. Schaffert, W. Rödl, M. Ogris, E. Wagner, M. Günther,
J. Controlled Release 2011, 152, 127.
[12] Y. Nie, M. Günther, Z. Gu, E. Wagner, Biomaterials 2011, 32, 858.
[13] Q. Jiang, Y. Nie, X. Chen, Y. He, D. Yue, Z. Gu, Adv. Funct. Mater.
2017, 27, 1701571.
[14] H. Zhao, J. Duan, Y. Xiao, G. Tang, C. Wu, Y. Zhang, Z. Liu, W. Xue,
Chem. Mater. 2018, 30, 3438.
[38] H. Liang, A. Hu, X. Chen, R. Jin, K. Wang, B. Ke, Y. Nie, J. Mater.
Chem. B 2019, 7, 915.
[39] Y. Chen, Z. Li, H. Wang, Y. Wang, H. Han, Q. Jin, J. Ji, ACS Appl.
Mater. Interfaces 2016, 8, 6852.
[40] F. L. Tansi, R. Ruger, C. Bohm, R. E. Kontermann, U. K. Teichgraeber,
A. Fahr, I. Hilger, Biomaterials 2016, 88, 70.
[41] F. He, Y. Tang, S. Wang, Y. Li, D. Zhu, J. Am. Chem. Soc. 2005, 127,
12343.
[42] C. A. Lin, R. A. Sperling, J. K. Li, T. Y. Yang, P. Y. Li, M. Zanella,
W. H. Chang, W. J. Parak, Small 2008, 4, 334.
[43] A. Yamada, Y. Hiruta, J. Wang, E. Ayano, H. Kanazawa,
Biomacromolecules 2015, 16, 2356.
[44] C. Yan, Z. Guo, Y. Shen, Y. Chen, H. Tiana, W. Zhu, Chem. Sci. 2018,
9, 4959.
[45] I. Koltover, T. Salditt, J. O. Rädler, C. R. Safinya, Science 1998, 281, 78.
[46] H. S. Choi, K. Nasr, S. Alyabyev, D. Feith, J. H. Lee, S. H. Kim,
Y. Ashitate, H. Hyun, G. Patonay, L. Strekowski, M. Henary,
J. V. Frangioni, Angew. Chem., Int. Ed. 2011, 50, 6258.
[47] M. Piest, J. F. J. Engbersen, J. Controlled Release 2010, 148, 83.
[48] Y. Xue, J. Tian, Z. Liu, J. Chen, M. Wu, Y. Shen, W. Zhang,
Biomacromolecules 2019, 20, 2796.
[49] G. Cheng, Y. He, L. Xie, Y. Nie, B. He, Z. Zhang, Z. Gu,
Int. J. Nanomed. 2012, 7, 3991.
[50] Y. He, G. Cheng, L. Xie, Y. Nie, B. He, Z. Gu, Biomaterials 2013, 34,
1235.
[15] X. Chen, J. Yang, H. Liang, Q. Jiang, B. Ke, Y. Nie, J. Mater. Chem. B
2017, 5, 1482.
[16] G. Shim, S. Ko, D. Kim, Q. V. Le, G. T. Park, J. Lee, T. Kwon,
H. G. Choi, Y. B. Kim, Y. K. Oh, J. Controlled Release 2017, 267, 67.
[17] Q. Jiang, D. Yue, Y. Nie, X. Xu, Y. He, S. Zhang, E. Wagner, Z. Gu,
Mol. Pharmaceutics 2016, 13, 1809.
[51] S. Raffy, J. Teissié, Biophys. J. 1999, 76, 2072.
[52] M. L. Briuglia, C. Rotella, A. McFarlane, D. A. Lamprou, Drug
Delivery Transl. Res. 2015, 5, 231.
[18] Y. He, Y. Nie, G. Cheng, L. Xie, Y. Shen, Z. Gu, Adv. Mater. 2014, 26,
1534.
[19] H. Wang, X. Li, B. W.-C. Tse, H. Yang, C. A. Thorling, Y. Liu,
M. Touraud, J. B. Chouane, X. Liu, M. S. Roberts, Theranostics 2018,
8, 1227.
[53] Y. Wen, Z. Zhang, J. Li, Adv. Funct. Mater. 2014, 24, 3874.
[54] B. Wang, R. M. Zhao, J. Zhang, Y. H. Liu, Z. Huang, Q. Y. Yu,
X. Q. Yu, Eur. J. Med. Chem. 2017, 136, 585.
[20] X. Liu, M. Wu, Q. Hu, H. Bai, S. Zhang, Y. Shen, G. Tang, Y. Ping,
ACS Nano 2016, 10, 11385.
[21] Z. Sheng, D. Hu, M. Zheng, P. Zhao, H. Liu, D. Gao, P. Gong,
G. Gao, P. Zhang, Y. Ma, L. Cai, ACS Nano 2014, 8, 12310.
[22] H. Ding, F. Wu, Theranostics 2012, 2, 1040.
[23] Z. Ma, X. Jia, J. Bai, Y. Ruan, C. Wang, J. Li, M. Zhang, X. Jiang,
Adv. Funct. Mater. 2017, 27, 1604258.
[24] D. Gyawali, S. Zhou, R. T. Tran, Y. Zhang, C. Liu, X. Bai, J. Yang,
Adv. Healthcare Mater. 2014, 3, 182.
[55] C. L. Grigsby, K. W. Leong, J. R. Soc., Interface 2010, 7, S67.
[56] J. Park, J. Park, Y. Pei, J. Xu, Y. Yeo, Adv. Drug Delivery Rev. 2016, 104,
93.
[57] J. T. Alander, I. Kaartinen, A. Laakso, T. Pätilä, T. Spillmann,
V. V. Tuchin, M. Venermo, P. Välisuo, Int. J. Biomed. Imaging 2012,
2012, 1.
[58] Y.-P. Ho, H. H. Chen, K. W. Leong, T.-H. Wang, J. Controlled Release
2006, 116, 83.
[59] Q. Jiang, X. Chen, H. Liang, Y. Nie, R. Jin, M. Barz, D. Yue, Z. Gu,
Nanoscale Adv. 2019, 1, 498.
[25] J. Croissant, J. I. Zink, J. Am. Chem. Soc. 2012, 134, 7628.
©
1906538 (13 of 13)
2020 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2020, 1906538