Kosuke Nakamoto1, Yukihiro Akao1, Yasuhiro Furuichi2, Yoshihito Ueno1,1,3. 1. United Graduate School of Agricultural Science, United Graduate School of Drug Discovery and Medical Information Sciences, and Faculty of Applied Biological Sciences, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan. 2. GeneCare Research Institute Co., Ltd., 19-2 Kajiwara, Kamakura 247-0063, Japan. 3. Center of Highly Advanced Integration of Nano and Life Sciences, Gifu University (G-CHAIN), 1-1 Yanagido, Gifu 501-1193, Japan.
Abstract
Small interfering RNA (siRNA), consisting a 21-mer duplex molecule, is often modified by conjugation with specific ligands to enhance its capacity for tissue-specific delivery. However, these attempts are hampered by the low permeability of negatively charged RNA molecules to enter the cell membrane. In this study, we designed and synthesized siRNA conjugates modified with cationic oligospermine and cyclic RGD (cRGD) to overcome the low-membrane permeability of siRNA. The siRNA conjugate, which contains 15 spermines and a cRGD peptide, showed sufficient gene-silencing activity at 250 nM final concentration without a transfection reagent. Under these conditions, the cationic oligospermine and cRGD-siRNA conjugate did not show any cytotoxicity.
Small interfering RNA (siRNA), consisting a 21-mer duplex molecule, is often modified by conjugation with specific ligands to enhance its capacity for tissue-specific delivery. However, these attempts are hampered by the low permeability of negatively charged RNA molecules to enter the cell membrane. In this study, we designed and synthesized siRNA conjugates modified with cationic oligospermine and cyclic RGD (cRGD) to overcome the low-membrane permeability of siRNA. The siRNA conjugate, which contains 15 n>an class="Chemical">spermines and a cRGD peptide, showed sufficient gene-silencing activity at 250 nM final concentration without a transfection reagent. Under these conditions, the cationic oligospermine and cRGD-siRNA conjugate did not show any cytotoxicity.
Oligonucleotide
(ON) therapn>eutics, using rationally designed antisense
n>an class="Chemical">oligonucleotides (ASONs) or small interfering RNAs (siRNAs), are promising
candidates for a treatment of unmet medical needs, which could not
be covered by conventional low-molecular-weight drugs.[1,2] However, these ONs, especially double-stranded siRNAs, have relatively
high molecular weights and negative charges derived from a phosphodiester
backbone. As a result, they can hardly penetrate the negatively charged
cell membrane.[3] Furthermore, the ASON and
siRNA composed of natural nucleosides do not exhibit tissue specificity.
Therefore, the development of a target tissue-specific drug delivery
system is crucial for the use of ASON or siRNA as a drug. To date,
numerous chemical approaches have been reported to overcome the low
membrane permeability of siRNA.[4,5] Among them, a targeted
tissue-specific ligand–siRNA conjugate is one promising strategy
for siRNA delivery into specific cells.[6−15] In particular, siRNAs conjugated with triantennary N-acetylgalactosamine (GalNAc), which has a high binding affinity
to the asialoglycoprotein receptor (ASGR), have shown efficient gene-silencing
activity in vivo, and clinical and preclinical trials are ongoing.[13−15] Because ASGR is highly expressed in hepatocytes and shows rapid
internalization and recycling, GalNAc–siRNA conjugates bound
to ASGR are readily incorporated into cells and exhibit sufficient
gene-silencing activity at low concentrations. However, unfortunately,
the development of other type of ligand–siRNA conjugates has
not been successful because of the low membrane permeability of RNA
molecules. Although the development of alternative ligand–siRNA
conjugates is required to expand deliverable tissues, no comparable
ligand–receptor systems, similar to GalNAc–ASGR, are
yet to be established. Therefore, to develop ligand–siRNA conjugates
that target receptors other than ASGR, a new approach which improves
potentially low cell membrane permeability of siRNA conjugate is needed.
In in vitro studies using cultured cells, it was demonstrated that
cationic oligospermine conjugates neutralized the negative charge
of siRNAs derived from their phospn>hodiester backbones, resulting in
impn>roved thermal stability, nuclease resistance, and cell membrane
permeability.[16−19] Although n>an class="Chemical">oligospermine–siRNA and AON conjugates have been
synthesized and successfully transfected in vitro,[20] tissue-specific delivery of oligospermine–siRNA
or AON conjugates has not been achieved because of spermine’s
lack of tissue specificity. Further, more than 20 spermine molecules
must be conjugated to one siRNA to provide sufficient cell permeability,
which sometimes causes cytotoxicity.[18] Alternatively,
the cyclic RGD (cRGD) peptide has a high binding affinity for integrin
αVβ3, which is overexpressed on the surface
of various tumor cells.[21] cRGD–siRNA
conjugates have been successfully transfected in vitro and in vivo
by several research groups.[22,23]
On the basis
of these reports, we designed and synthesized siRNA
conjugates containing both the cationic oligospermine and the n>an class="Disease">tumor-targeting
cRGD peptide (Figure ). We postulated that the cationic oligospermine modification of
the cRGD–siRNA would increase its membrane permeability, thus
enhancing the tumor cell-specific delivery. Furthermore, we postulated
that the oligospermine and cRGD-peptide would work synergistically.
As such, the amount of spermine to be incorporated into the siRNA
could be reduced, thus reducing oligospermine-induced cytotoxicity.
Figure 1
Schematic
illustration of cationic oligospermine and cRGD–siRNA
conjugates.
Schematic
illustration of cationic pan class="Chemical">oligospermine and cRGD–siRNA
conjugates.
Results and Discussion
Synthesis
of ON Conjugates
We designed and synthesized
siRNA conjugates modified by both oligospermine and cRGD at the 3′-termini
of the sense and antisense strands, respn>ectively. n>an class="Gene">RecQL1-siRNA, which
silences the RecQL1 mRNA encoding RecQL1 DNA helicase, was chosen
as a model siRNA. Futami and colleagues have previously shown that
gene silencing of RecQL1 helicase by RecQL1-siRNA results in the following:
(1) inhibits DNA repair in replicating cells, (2) induces mitotic
catastrophes at the checkpoints of negative tumor cells, and (3) causes
mitotic death in the tumor cell specifically.[24,25] It was also reported that oligospermine conjugation at the antisense
(guide) strand of the siRNA reduces the siRNA silencing activity greater
than the sense (passenger) strand.[19] Thus,
we introduced the oligospermine at the 3′ end of the sense
(passenger) strand and cRGD at that of the antisense (guide) strand,
respectively (Figure A,B). To improve nuclease resistance and silencing activities of
the siRNAs, the 2′-hydroxyl group of some pyrimidine nucleotides
was modified by a methyl group (Table ). The 2′-O methylations are known to reduce
immunogenicity of siRNAs and to suppress the innate immune response.[26]
Figure 2
(A) Structures of antisense RNAs conjugated to cRGD. (B)
Structures
of sense RNAs conjugated to cationic oligospermine. Small letters
indicate 2′-O-methyl RNA and dT indicates
2′-deoxythymidine.
Table 1
Sequences of siRNAs, N/P Ratio
name
S/ASa
sequence (5′–3′)b
N/P
siQL1
S
GuucAGACCACuucAGcuudTdT
AS
AAGCUGAAGUGGuCuGAAcdTdT
siQL1[RGD]
S
GuucAGACCACuucAGcuudTdT
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD
siQL1[S5/RGD]
S
GuucAGACCACuucAGcuudTdT-S5-dT
0.42
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD
siQL1[S10/RGD]
S
GuucAGACCACuucAGcuudTdT-S10-dT
0.77
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD
siQL1[S15/RGD]
S
GuucAGACCACuucAGcuudTdT-S15-dT
1.03
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD
siQL1[S15]
S
GuucAGACCACuucAGcuudTdT-S15-dT
1.05
AS
AAGCUGAAGUGGuCuGAAcdTdT
Cy3-siQL1
S
GuucAGACCACuucAGcuudTdT
AS
AAGCUGAAGUGGuCuGAAcdTdT-Cy3
Cy3-siQL1[RGD]
S
GuucAGACCACuucAGcuudTdT
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD-Cy3
Cy3-siQL1[S5/RGD]
S
GuucAGACCACuucAGcuudTdT-S5-dT
0.41
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD-Cy3
Cy3-siQL1[S10/RGD]
S
GuucAGACCACuucAGcuudTdT-S10-dT
0.75
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD-Cy3
Cy3-siQL1[S15/RGD]
S
GuucAGACCACuucAGcuudTdT-S15-dT
1.02
AS
AAGCUGAAGUGGuCuGAAcdTdT-RGD-Cy3
Cy3-siQL1[S15]
S
GuucAGACCACuucAGcuudTdT-S15-dT
1.03
AS
AAGCUGAAGUGGuCuGAAcdTdT-Cy3
S denotes sense
strand and AS represents
antisense strand.
Small
letters indicate 2′-O-methyl RNA and dT indicates
2′-deoxythymidine.
(A) Structures of antisense RNAs conjugated to cRGD. (B)
Structures
of sense RNAs conjugated to cationic pan class="Chemical">oligospermine. Small letters
indicate 2′-O-methyl RNA and n>an class="Chemical">dT indicates
2′-deoxythymidine.
S denotes sense
strand and AS represents
antisense strand.Small
letters indicate 2′-O-methyl RNA and pan class="Chemical">dT indicates
2′-deoxythymidine.
Incorporation of cRGD peptide into the antisense strand was carried
out by a postsynthetic modification method (Scheme ). Briefly, single-stranded RNA (ssRNA) 6, which has a reactive amino moiety at the 3′ end
of the strand, was synthesized using an amino-modified controlled
pore glass (Scheme S1). Subsequently, 6 was reacted with N-(6-maleimidocaproyloxy)succinimide
(EMCS) in a n>an class="Chemical">phosphate buffer (pH = 7.4) at 37 °C to give a maleimide-modified
ssRNA 8. A thiol-containing cRGD peptide (cRGDfC) was
conjugated to 8 via a thiol–maleimide Michael
addition reaction to produce cRGD conjugate 1. To assess
the cellular uptake of the conjugate, a Cy3-labeled cRGD conjugate 2 was also synthesized by a similar procedure. Product purification
was performed by reversed-phase high-performance liquid chromatography
(RP-HPLC), and the structures of the synthesized ONs were confirmed
by matrix-assisted laser desorption ionization time-of-flight/mass
spectrometry (Table S2). The cationic oligospermine
conjugates 3–5 were synthesized according to the
procedure previously reported using a spermine amidite.[27] To determine the number of spermine required
for cellular uptake and examine the effects of the N/P ratio on cellular
uptake, we synthesized three oligospermine conjugates with different
amounts of spermine modifications. Each ssRNA conjugates were annealed
to complementary RNA to form the siRNA conjugates with the oligospermine
and cRGD.
Scheme 1
Synthesis of cRGD Conjugates; (a) EMCS, in a Phosphate
Buffer (pH
7.4), Room Temperature, Overnight and (b) c(RGDfC), in 0.1 M triethylammonium
acetate (TEAA), 50% MeCN aq, Room Temperature, Overnight
Cellular Uptake Test
Cellular uptakes of Cy3-labeled
siRNA conjugates were evaluated using confocal laser microscopn>y. n>an class="Species">HumanmelanomaA2058 cells were incubated with 200 nM of each Cy3-labeled
siRNA conjugates in serum-free medium. After a 3 h incubation, accumulations
of Cy3-labeled siRNA conjugates in the cells were visualized by confocal
laser microscopy (Figure ). In the siQL1[S10/RGD] and siQL1[S15/RGD] RNAs, which were
modified by spermine, the accumulation of Cy3 fluorescence signal
was detected. However, no fluorescence signal was seen in the unmodified
spermine conjugate, siQL1[RGD]. The fluorescence intensity in the
cell was dependent on the number of spermine molecules conjugated
to RNA. The strongest fluorescence intensity was observed when the
siQL1[S15/RGD] was used. These results indicate that the cationic
oligospermine addition to cRGD–siRNA conjugates promoted cellular
uptake. Furthermore, because the fluorescence intensity in A2058 cells
was significantly lower in the cRGD-unmodified siQL1[S15], it is conceivable
that the synergistic work of cRGD and cationic oligospermine may be
crucial for the cellular uptake of siRNA conjugates.
Figure 3
Transfection reagent-free
cellular uptake of Cy3-labeled siRNA
heteroconjugates in vitro. A2058 cells were treated with 200 nM of
Cy3-labeled siRNA conjugates. After a 3 h incubation, Cy3-labeled
siRNA conjugates were visualized by confocal laser microscopy.
Transfection reagent-free
cellular uptake of Cy3-labeled siRNA
heteroconjugates in vitro. n>an class="CellLine">A2058 cells were treated with 200 nM of
Cy3-labeled siRNA conjugates. After a 3 h incubation, Cy3-labeled
siRNA conjugates were visualized by confocal laser microscopy.
Gene-Silencing Activities
To evaluate the RNA interference
(RNAi) activity of the cRGD and oligospermine-modified siRNAs, we
assessed the gene-silencing activities of each siRNA conjugate in
the presence of the transfection reagent n>an class="Chemical">Lipofectamine. Each siRNA
conjugate was mixed with Lipofectamine RNAiMAX and transfected into
HeLa cells at 20 nM final concentration. The relative expression levels
of RecQL1 mRNA were measured by RT-qPCR. In Figure A, all siRNA conjugates
showed greater gene-silencing activity than the unmodified siQL1.
These data indicated that the modification of siRNA by cRGD and oligospermine
does not disrupt the formation of the RNA-induced silencing complex
and target mRNA recognition.
Figure 4
Effect of RECQL1-siRNA conjugates on the expression
of HeLa cell
RECQL1 mRNA. (A) siRNA (20 nM) conjugates were transfected using Lipofecatamine
RNAiMAX. (B) HeLa cells were transfected with 250 nM of siRNA conjugates
without a transfection reagent. (C) Cells were transfected with 100–250
nM of siQL1[S15/RGD] without a transfection reagent. (D) Cells were
transfected with 125 and 250 nM of siLuc[S15/RGD] without a transfection
reagent. (E) Cells were transfected with 250 nM of siQL1[S15] without
a transfection reagent. (F) siQL1[S15/RGD] were transfected without
a transfection reagent in the absence or presence of FBS.
Effect of RECQL1-siRNA conjugates on the expn>ression
of n>an class="CellLine">HeLa cell
RECQL1 mRNA. (A) siRNA (20 nM) conjugates were transfected using Lipofecatamine
RNAiMAX. (B) HeLa cells were transfected with 250 nM of siRNA conjugates
without a transfection reagent. (C) Cells were transfected with 100–250
nM of siQL1[S15/RGD] without a transfection reagent. (D) Cells were
transfected with 125 and 250 nM of siLuc[S15/RGD] without a transfection
reagent. (E) Cells were transfected with 250 nM of siQL1[S15] without
a transfection reagent. (F) siQL1[S15/RGD] were transfected without
a transfection reagent in the absence or presence of FBS.
We then evaluated the gene-silencing activities
of siRNA conjugates
without transfection reagent under serum-free conditions. As shown
in Figure B,C, only
siQL1[S15/RGD] showed strong gene-silencing activity at 150–250
nM. The result indicates that the cationic oligospermine modification
impn>roved the cell membrane permeability of cRGD–siRNA conjugates.
In addition, it was found that at least 15 n>an class="Chemical">spermine modifications
(N/P = 1.02) were needed for efficient gene-silencing activity. Previously,
it was reported that oligospermine–siRNA conjugates required
more than 20 spermine modifications (N/P = 1.33) for efficient gene-silencing
activity without a transfection reagent.[20] Thus, the cRGD peptide addition to siRNA reduces the required number
of oligospermine molecules, which sometimes cause cytotoxicity. The
siLuc[S15/RGD]- targeting Renilla luciferase mRNA
(Table S3), which contains 15 spermines
and a cRGD-peptide, did not silence the expression of RecQL1 mRNA under similar conditions (Figure D). Therefore, the RNAi activity of siQL1[S15/RGD],
which inhibits RecQL1 mRNA expression, appears to
be supported not only by accessory conjugates but also by elaborated
RNA sequences. The siQL1[S15] (contains 15 spermines only) did not
show gene-silencing activity (Figure E). This result indicates that both oligospermine and
cRGD-peptide are required for efficient gene silencing. The gene-silencing
activities of siRNA conjugates were also confirmed by the Western
blot analysis (Figure S5).
Next,
we evaluated the effects of serum on the silencing activity
of siQL1[S15/RGD]. As shown in Figure F, the silencing activity of siQL1[S15/RGD] decreased
slightly in 2% fetal bovine serum (n>an class="Disease">FBS) and was eliminated in 10%
FBS conditions. It was reported that intracellular uptake of the cationic
oligospermine–siRNA conjugate was inhibited by nonspecific-binding
of protein to the siRNA in the serum.[21] Thus, it is suggested that our siQL1[S15/RGD] also interacted with
serum proteins, such that its intercellular uptake was severely reduced.
Cytotoxicity Test
Finally, we examined the cytotoxicity
of the siRNA conjugates using siLuc[S15/RGD] because siQL1-targeting n>an class="Gene">RecQL1 mRNA has strong antitumor cell activity. A2058 cells
were incubated with 125 nM to 1 μM of siLuc[S15/RGD] in serum-free
medium. After a 48 h incubation, cytotoxicity was examined by measuring
cell viabilities (Figure ). Although slight growth inhibition was observed at 1 μM
of siLuc[S15/RGD], there was no cytotoxicity at 250 nM, which is the
concentration needed for transfection reagent-free gene silencing
of siQL1[S15/RGD]. Although the oligospermine–siRNA conjugates
show gene-silencing activities when the N/P ratio is greater than
1.33, they are sometimes cytotoxic. On the other hand, by incorporating
the cRGD peptide into the oligospermine–siRNA conjugate, we
succeeded in reducing the N/P ratio to 1.02 to eliminate cytotoxicity
while also maintaining the gene-silencing activity.
Figure 5
Cell viability at 48
h after siLuc[S15/RGD] transfection. Cells
were transfected with 125 nM to 1 μM of siLuc[S15/RGD] without
a transfection reagent.
Cell viability at 48
h after siLuc[S15/RGD] transfection. Cells
were transfected with 125 nM to 1 μM of siLuc[S15/RGD] without
a transfection reagent.
Conclusions
In this study, we successfully synthesized
novel cationic oligospermine
and cRGD–siRNA conjugates. By combining cRGD pepn>tides with
a cationic 15 n>an class="Chemical">spermine modification (N/P = 1.02), we found that siQL1[S15/RGD]
containing both 15 spermines and cRGD peptides showed sufficient gene-silencing
activity at 250 nM final concentration without a transfection reagent.
Conversely, the siQL1[S15] containing only 15 spermine did not show
gene-silencing activity under same conditions. Collectively, these
results demonstrate synergy between the oligospermine and cRGD-peptide
for efficient gene silencing. It was also found that the siLuc[S15/RGD]
caused no cytotoxicity at 250 nM, which is the required concentration
for transfection reagent-free gene silencing of siQL1[S15/RGD]. Taken
together, the oligospermine and cRGD-peptide-combined conjugation
holds a great promise for efficient siRNA-mediated drug delivery to
various organs and tissues beyond the liver.
Experimental Section
RNA Synthesis
Synthesis was carried out with a DNA/RNA
synthesizer by phosphoramidite method. Deprotection of bases and phosphates
was performed in concentrated n>an class="Chemical">NH4OH/EtOH (3:1, v/v) at
55 °C for 4 h. 2′-TBDMS groups were removed by TEA·3HF
(Aldrich) at 65 °C for 1.5 h. The reaction was quenched with
0.1 M TEAA buffer (pH 7.0) and desalted on a Sep-Pak C18 cartridge.
The deprotected ONs were separated from unprotected ONs by 20% polyacrylamide
gel electrophoresis containing 7 M urea to give rise to the highly
purified siRNA (Supporting Table).
Postsynthetic
Modification of siRNAs Conjugated with cRGD
The cRGD-conjugated
antisense strand of RECQL1-siRNA was synthesized
by postsynthetic modification method. The cRGD-conjugated antisense
RNA strands were purified by reversed-phase C-18 HPLC using a linear
gradient of 5–50% of n>an class="Chemical">MeCN in 0.1 M TEAA buffer at pH 7.0. The
HPLC profiles of synthesized RNA are shown in Figure S4.
Synthesis of siRNAs Conjugated with cRGD
at the Antisense RNA 1
To the solution of 6 (15 nmol in a
total of 100 μL) consisting 10 mM NaCl, 100 mM n>an class="Chemical">phosphate buffer
(pH 7.4) (85 μL), and dimethyl sulfoxide (DMSO) (15 μL),
0.2 M EMCS in DMSO (12 μL) was added at the room temperature.
The mixture was incubated at 37 °C overnight and maleimide-modified 8 was collected by RP-HPLC. Subsequently, 50 mM c(RGDfC) peptide
(Bachem) in DMSO (2 μL) was added directly to the collected 8 solution. The mixture was then incubated at room temperature
overnight and the cRGD-conjugated RNA 1 (3.48 nmol) was
purified by RP-HPLC with a recovery of 23%.
Synthesis of RNA (or ONs)
with cRGD Conjugates 2
To the solution of 7 (15 nmol in a total of
100 μL) consisting 10 mM NaCl, 100 mM n>an class="Chemical">phosphate buffer (pH 7.4)
(85 μL), and DMSO (15 μL), 0.2 M EMCS in DMSO (12 μL)
was added at the room temperature. The mixture was incubated at 37
°C overnight, and maleimide-modified 9 was collected
by RP-HPLC. Subsequently, 50 mM c(RGDfC) peptide in DMSO (3 μL)
was directly added to the collected solution containing 9. The mixture was incubated at room temperature overnight, and cRGD-conjugated
RNA 2 was purified by RP-HPLC (3.03 nmol, 20%).
Cellular
Uptake Test
A2058 cells (4.0 × 104 cell/mL)
were placed on the 35 mm glass-bottom dish (200
μL/dish) and were grown for 24 h before transfection. Cells
were transfected with 200 nM of n>an class="Chemical">Cy3-labeled siRNA conjugates in serum-free
Opti-MEM medium. After 3 h of transfection, cells were washed by PBS,
and the accumulations of Cy3-labeled siRNA conjugates in the transfected
cells were visualized by confocal microscope (Zeiss LSM710).
Gene-Silencing
Assay
A2058 cells (4.0 × 104 cell/mL) were
transferred to 96-well plate (100 μL/well),
24 h before transfection. Cells were transfected with various concentrations
of siRNA conjugates in serum-free n>an class="Chemical">Opti-MEM medium. After 4 h transfection,
FBS was added to each well at 2% final concentration and incubated
for 44 h. Cell lysis was performed using SuperPrep Cell Lysis Kit
for qPCR (TOYOBO), and RNAs were extracted. Real-time PCR was performed
with primers specific for target mRNA by using THUNDERBIRD SYBR qPCR
Mix (TOYOBO). The sequences of primer are listed in Table S4.
A2058 cells (4.0
× 104 cell/mL) were transferred to a 6-well plate
(1 mL/well) 24 h before
transfection. Cells were transfected with various concentrations of
siRNA conjugates in serum-free n>an class="Chemical">Opti-MEM medium. After 4 h transfection,
FBS was added to each well at 2% final concentration and further incubation
for 44 h. The cell viabilities were determined by the trypan blue
dye exclusion test.
Authors: Ivan V Chernikov; Daniil V Gladkikh; Ulyana A Karelina; Mariya I Meschaninova; Alya G Ven'yaminova; Valentin V Vlassov; Elena L Chernolovskaya Journal: Molecules Date: 2020-04-18 Impact factor: 4.411