Open Access Article
Minh-Duc
Hoang
a,
Marie
Vandamme
b,
Gueorgui
Kratassiouk
b,
Guillaume
Pinna
*b,
Edmond
Gravel
*a and
Eric
Doris
*a
aService de Chimie Bioorganique et de Marquage (SCBM), CEA, Université Paris-Saclay, 91191 Gif-sur-Yvette, France. E-mail: edmond.gravel@cea.fr; eric.doris@cea.fr
bPlateforme ARN Interférence, Service de Biologie Intégrative et de Génétique Moléculaire (SBIGeM), I2BC, CEA, CNRS, Université Paris-Saclay, 91191 Gif-sur-Yvette, France. E-mail: guillaume.pinna@cea.fr
First published on 24th September 2019
Polydiacetylene micelles were assembled from four different cationic amphiphiles and photopolymerized to reinforce their architecture. The produced micelles were systematically investigated, in interaction with siRNAs, for intracellular delivery of the silencing nucleic acids. The performances of the carrier systems were rationalized based on the cell penetrating properties of the micelles and the nature of their cationic complexing group, responsible for efficient siRNA binding and further endosomal escape.
As nucleic acids do not freely pass the cell plasma membrane, delivery of siRNA can be promoted using nanometric vectors. Taken as a whole, the efficacy of gene silencing is, for a large part, dependent on the cellular entry of siRNAs, also called transfection. This can be commonly achieved by complexing siRNAs with nanocarrier systems, providing protection to the siRNA and favoring cellular uptake/payload delivery.3 Transfection is a stepwise process which starts with the initial crossing of the plasma membrane through endocytosis and accumulation of the siRNA-loaded nanoparticles in endosomes/lysosomes from which they need to escape to reach the nucleus.4 The design of nanocarriers is thus of prime importance to maximize the amount of siRNA to be delivered into the cells. At the cellular level, examples of siRNA delivery systems include cationic polymers,5 lipids,6 peptides,7 carbon nanotubes,8 nanofibers9 and micelles.10 Although the latter systems are efficient and could behave in a synergistic fashion,11 there is a lack of understanding on how the chemical structures of the cationic region (responsible for electrostatic interactions with the nucleic acids) impacts the overall transfection process, and a rationale is yet to be formulated.
Our group has long been involved in the self-assembly of amphiphilic units into micelles for biomedical applications.12 For this study, a series of polydiacetylene micelles of identical structures, but incorporating different terminal ammonium groups, were assembled. The carrier systems were methodically studied with respect to transfection efficacy in order to better understand the role played by the substitution pattern of the complexing nitrogen atoms on the binding, transport, and release of the siRNA payload. The four micelle-forming amphiphiles discussed in this article are all based on the same N-(2-aminoethyl)pentacosa-10,12-diynamide backbone, yet incorporating a variable alkylation degree on the terminal amino-group, leading to primary (1), secondary (2), tertiary (3) and quaternary (4) ammoniums (Fig. 1a).
![]() | ||
| Fig. 1 (a) Structure of the four DA-AM amphiphiles; (b) micelle assembly, photopolymerization and complexation with siRNA. | ||
Amphiphile 2, incorporating a terminal secondary amine (–NHMe), was obtained in 85% yield (over two steps) by the coupling of activated ester 5 with N-Boc-N-methylethylenediamine, followed by acidic deprotection of the methyl-amino group. Amphiphile 3 was produced in 95% yield from the reaction of 5 with N,N-dimethylethane-1,2-diamine, to give access to the amphiphile terminated by a tertiary amine (–NMe2).
The above amphiphile (3) also acted as a precursor to the synthesis of the quaternary ammonium 4 which was obtained by initial alkylation of 3 in methyl iodide followed by counter-ion exchange (I → Cl) over Amberlyst-26 resin. The quaternary ammonium chloride amphiphile 4 (–+NMe3) was isolated in 88% yield, over two steps.
The critical micelle concentration (CMC) of each of the newly synthesized amphiphiles was then measured by the pyrene encapsulation technique which indicated values of 0.06, 0.03, 0.03, and 0.01 mg mL−1 for DA-AM amphiphiles 1, 2, 3, and 4, respectively.
We next proceeded with the assembly of the corresponding micelles (Fig. 1b). Each amphiphile, in its ammonium form, was dispersed in a slightly acidic aqueous medium (10 mM HCl) at a concentration of 10 mg mL−1 ([c] > CMC) and was probe-sonicated for 30 min. The resulting clear suspension was then irradiated under UV (254 nm) for 5 h, producing a pale yellow colloid. The polymerization involved the formation of a ene-yne conjugated network through a topochemical 1,4-addition mechanism. The resulting pDA-AM micelles were further treated by dialysis for one week to remove non-polymerized amphiphilic units which are classically known to impart some cytotoxicity by cellular membrane destabilization.13 The polymerization step thus not only reinforced the stability of the colloidal micelles in highly dilute conditions (for example below the CMC), but also lowered their potential cytotoxicity, as dialyzed micelles interfered with HeLa cells proliferation/survival only at very high concentrations (Fig. S1†).
The micellar solution was then freeze dried and taken back in pure water before the pDA-AM micelles were characterized by Dynamic Light Scattering (DLS) analysis which indicated an average hydrodynamic diameter comprised between 6 and 9 nm, depending on the micelle type (Fig. S2†). Zeta potentials were also measured and all found to be of positive values (comprised between +15 and +30 mV). These results confirmed the overall cationic character of the micelles.
Electrostatic adherence of siRNA to the cationic micelles, which is a prerequisite to transfection, was investigated by mixing increasing quantities of the pDA-AM 1–4 micelles to siRNA duplexes (20 pmol). Each sample was then analyzed by agarose gel electrophoresis, to separate free-migrating from micelle-bound siRNAs. As seen in Fig. 2, full retardation was observed at a ratio of micelle–nitrogen atoms (N) to siRNA phosphates (P) slightly above 5 for pDA-AM 1, ca. 5 for pDA-AM 2, slightly above 25 for pDA-AM 3, and ca. 5 for pDA-AM 4.
As the pDA-AM micelles were able to electrostatically bind siRNAs, we next evaluated their ability to efficiently deliver functional siRNAs to cells. To this end, we used a commercial pool of cytotoxic siRNAs specifically designed to target genes that are essential to cell proliferation/survival (AllStars Death Control, Qiagen). Transfection efficiency was assessed by measuring the cell proliferation/survival of HeLa cells, 72 h after transfection with the cytotoxic siRNA@micelle complexes. Control experiments were conducted using a “scrambled” siRNA sequence (UNR, negative control). siRNA@micelle complexes were tested at variable N/P ratios ranging from 5 to 50, while maintaining the siRNA concentration at 20 nM. In contrast to some of the previously described siRNA delivery procedures,14 in which cationic nanoparticles are transiently applied to cells in a medium without serum, we selected more stringent conditions by adding the siRNA@micelle complexes to cell culture medium containing serum. In the latter case, lipophilic components such as seric albumin could potentially interfere with the transfection process.
Cell proliferation/survival was monitored upon cell fixation and counterstaining of cellular DNA by a fluorescent probe (Hoechst 33342), which allowed cell counting by High Content Imaging analysis. Data normalized to the “Untreated” control are reported in Fig. 3.
The first observation is that there is a strong effect of the cytotoxic siRNA pool when complexed with pDA-AM 1 and pDA-AM 2 micelles, starting at N/P ratios of 10 and 5, respectively. pDA-AM 2 micelles appeared to be the most potent in terms of transfection efficiency, whereas pDA-AM 3 and pDA-AM 4 micelles did not promote any transfection, whatever the N/P ratio considered. In fact, in the latter cases, we observed neither silencing effect nor intrinsic toxicity of the micelles. It is to be noted that some moderate unspecific cytotoxic (or -static) effect was associated to pDA-AM 1 and 2 micelles when in interaction with the negative control siRNA (UNR), especially at higher N/P ratios. Nevertheless, this effect is not detrimental per se, as strong silencing could already be detected for lower N/P ratios.
The two micellar carriers that were found to be efficient for siRNA transfection (i.e. pDA-AM 1 and pDA-AM 2) were further tested at a fixed N/P ratio but with varying siRNA concentrations (from 1 to 10 nM), to determine the optimal transfection conditions, with minimal non-specific cytotoxicity (Fig. 4). A N/P ratio of 50 was selected for pDA-AM 1 micelles and transfection experiments were carried out with increasing amounts of the siRNA@pDA-AM 1 complex. The cytotoxic siRNA pool had a moderate effect on cell survival (80% cell survival) with 1 nM siRNA. Increasing the amount of siRNA@pDA-AM 1 complex to achieve a final siRNA concentration equal or above 5 nM led to a more pronounced effect, albeit with some associated minor side toxicity. Yet, when considering a N/P ratio fixed at 50, siRNA concentration of 10 nM appeared to be optimal for transfection, while maintaining non-specific cytotoxicity to an acceptable level.
Based on our initial experiments, we set the N/P ratio of pDA-AM 2 micelles at a value of 10. Under these conditions, 1 nM of siRNA led to a moderate, yet significant, cytotoxic effect (cell viability < 70%). Increasing the siRNA@pDA-AM 2 complex concentration up to 10 nM siRNA led to an enhancement of the transfection efficiency without any non-specific toxicity. Thus, pDA-AM 2 micelles at N/P = 10 can safely be used to efficiently transfect siRNAs down to 5 nM concentration. Performances of the cationic micelles were compared to that of a reference lipidic carrier system (Lipofectamine RNAi Max, Fig. 4). The results obtained with pDA-AM 1–2 were compared to that of Lipofectamine for which we observed a silencing effect in the same range as that of the micelles, upon complexation with 1, 5, or 10 nM of siRNA. Yet, the exact chemical structure of Lipofectamine RNAi max is unknown, thus preventing the calculation of the N/P ratio of the Lipofectamine/siRNA complex. This fact puts into perspective the relative effectiveness of the commercially available transfection agent.
It should be noted that, upon addition of the siRNA to the most active pDA-AM 2 micelle, we observed an increase of the hydrodynamic diameter of the micelles to ca. 80 nm (Fig. S3a†) and a decrease of the zeta potential value to 8 mV (N/P = 10) (Fig. S3b†). These observations are in line with the formation of the siRNA@micelle complex.
The results obtained in this study show that primary and secondary amines seem to be more efficient than their tertiary and quaternary counterparts when it comes to siRNA transfection. In the case of quaternary ammoniums, these findings are in agreement with common understanding of the siRNA release process. In fact, it is established that the so called “proton sponge effect” plays a central role in the siRNA escape from lysosomes inside of which the siRNA@micelle complexes are trapped.15 The phenomenon mostly relies on the buffering capacity of free amines, which induces an increase in lysosomal pH and influx of Cl− ions. As a consequence, osmotic swelling induces disruption of the lysosome and release of siRNAs into the cytoplasm.16 This phenomenon cannot be operative in the case of per-methylated quaternary ammoniums (pDA-AM 4 micelles) as the amino groups borne by the micelle are no longer basic. On the contrary, the amino groups of pDA-AM 1–3 micelles are still likely to behave as proton sponges, and the cause of the lack of efficiency of pDA-AM 3 micelles has to be found elsewhere. We hypothesize that, in the case of pDA-AM 3, transfection is less active because of poor siRNA binding to the micelle, which was supported by the formation of the siRNA@micelle complex at only a high N/P ratio of ca. 25.
Both pDA-AM 1 and pDA-AM 2 micelles behaved in a very satisfactory fashion, with the highest transfection efficiency and negligible toxicity observed for pDA-AM 2 micelles. Optimization of the N/P ratio and tuning of the siRNA concentration allowed us to improve the delivery system in such a way that the siRNA achieved >80% inhibition of cell proliferation/survival, with no significant static effect of either the micelle alone or in combination with an irrelevant siRNA sequence. The difference in transfection efficiency between pDA-AM 1 and pDA-AM 2 micelles could lie in the higher basicity of the secondary amine of 2 where the methyl substituent adds stabilization to the cationic charge through polarization. It is likely that the buffering effect in the proton sponge phenomenon increases with basicity of the amine, leading to more efficient release of the siRNA payload into the cytoplasm, thus resulting in a higher silencing effect. Cellular internalization of the siRNA@pDA-AM 1 and siRNA@pDA-AM 2 complexes was further confirmed by confocal microscopy (Fig. 5) using a siRNA labeled with Alexa Fluor 488 fluorescent probe.
In fact, for the two active micelles (i.e. pDA-AM 1 and pDA-AM 2), we observed strong intracellular fluorescence signals at the vicinity of the nucleus. On the contrary, only weak signals were detected by confocal microscopy for the inactive micelles pDA-AM 3 and pDA-AM 4 (epifluorescence and confocal microscopy data are provided in the ESI for the four micelle–siRNA complexes, see Fig. S4 and S5,† respectively). These results suggest that, in addition to the siRNA binding/release properties of the micelles, cellular internalization also plays a central role in the overall efficacy of the transfection system.
1H NMR (400 MHz, CDCl3) δ 2.84 (s, 4H), 2.60 (t, J = 7.5 Hz, 2H), 2.24 (t, J = 6.9 Hz, 4H), 1.79–1.69 (m, 2H), 1.56–1.46 (m, 4H), 1.45–1.19 (m, 26H), 0.88 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 169.32, 168.78, 77.72, 77.55, 65.40, 65.31, 32.03, 31.03, 29.76, 29.74, 29.72, 29.59, 29.46, 29.21, 29.01, 28.97, 28.90, 28.81, 28.46, 28.38, 25.70, 24.64, 22.80, 19.31, 19.29, 14.24 ppm.
:
10
:
1) affording compound 1 (417 mg, 94%).
1H NMR (400 MHz, CDCl3) δ 5.95 (s, 1H), 3.30 (q, J = 5.8 Hz, 2H), 2.83 (m, 2H), 2.25–2.16 (m, 6H), 1.62–1.25 (m, 32H), 0.87 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 173.58, 77.71, 77.56, 65.38, 65.31, 42.00, 41.54, 36.91, 32.02, 29.75, 29.73, 29.71, 29.58, 29.45, 29.34, 29.27, 29.20, 29.02, 28.96, 28.86, 28.45, 28.39, 25.85, 22.79, 19.30, 19.28, 14.23 ppm.
:
30) affording BOC-protected 2 (216 mg, 96%).
1H NMR (400 MHz, CDCl3) δ 6.38 (s, 1H), 3.47–3.28 (m, 4H), 2.86 (s, 3H), 2.28–2.06 (m, 6H), 1.64–1.12 (m, 41H), 0.86 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 173.79, 157.42, 80.12, 77.70, 77.55, 65.40, 65.33, 53.55, 47.46, 39.05, 36.88, 34.87, 32.03, 29.76, 29.74, 29.72, 29.59, 29.46, 29.39, 29.27, 29.21, 29.04, 28.97, 28.89, 28.52, 28.47, 28.42, 25.76, 22.80, 19.32, 19.30, 14.24 ppm.
000×g, 3 min) in Et2O to afford compound 2 (156 mg, 89%).
1H NMR (400 MHz, CDCl3) δ 8.32 (s, 2H), 7.72–7.62 (m, 1H), 3.75–3.63 (m, 2H), 3.35–3.23 (m, 2H), 2.80 (t, J = 5.0 Hz, 3H), 2.39–2.28 (m, 2H), 2.23 (t, J = 7.0 Hz, 4H), 1.67–1.13 (m, 32H), 0.87 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 178.50, 77.80, 77.56, 65.41, 65.31, 49.96, 36.90, 36.09, 33.95, 32.05, 29.78, 29.76, 29.74, 29.61, 29.48, 29.23, 29.15, 29.09, 29.00, 28.94, 28.85, 28.49, 28.39, 25.55, 22.82, 19.32, 14.23 ppm.
:
10) affording compound 3 (446 mg, 95%).
1H NMR (400 MHz, CDCl3) δ 6.11 (s, 1H), 3.39–3.25 (m, 2H), 2.42 (t, J = 6.0 Hz, 2H), 2.30–2.11 (m, 12H), 1.67–1.14 (m, 32H), 0.86 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 173.43, 77.70, 77.57, 65.39, 65.33, 58.05, 45.14, 36.80, 36.60, 32.03, 29.76, 29.74, 29.72, 29.59, 29.46, 29.35, 29.30, 29.22, 29.06, 28.97, 28.89, 28.46, 28.42, 25.84, 22.80, 19.32, 14.24 ppm.
:
10) affording iodinated 4 (465 mg, 88%).
1H NMR (400 MHz, CDCl3) δ 7.71 (t, J = 5.5 Hz, 1H), 3.94–3.72 (m, 4H), 3.46 (s, 9H), 2.34–2.15 (m, 6H), 1.69–1.12 (m, 32H), 0.87 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 174.99, 77.75, 77.62, 65.61, 65.37, 65.32, 54.67, 36.51, 34.29, 31.99, 29.72, 29.70, 29.69, 29.56, 29.42, 29.34, 29.29, 29.18, 29.03, 28.95, 28.91, 28.45, 28.42, 25.53, 22.76, 19.29, 14.21 ppm.
1H NMR (400 MHz, CDCl3) δ 8.61 (t, J = 4.9 Hz, 1H), 3.91–3.64 (m, 4H), 3.40 (s, 9H), 2.31–2.14 (m, 6H), 1.65–1.12 (m, 32H), 0.85 (t, J = 6.9 Hz, 3H) ppm.
13C NMR (100 MHz, CDCl3) δ 174.91, 77.74, 77.61, 65.61, 65.37, 65.33, 54.18, 36.32, 34.37, 32.01, 29.74, 29.72, 29.71, 29.58, 29.44, 29.40, 29.35, 29.20, 29.06, 28.97, 28.93, 28.46, 28.43, 25.54, 22.78, 19.28, 14.23 ppm.
Fluorescence spectra were recorded at 339 nm UV excitation wavelength at 5 nm band pass. The relative intensities at 373 nm and 384 nm were recorded. The ratios of the relative fluorescence intensities I373nm/I384nm were plotted against log of mM concentrations. CMC is deduced from the inflexion point.
:
20 and 1 μL of HCl (37%). The solvent was evaporated under reduced pressure and a white solid was formed. The solid was then solubilized in 1 mL of 10 mM HCl and sonicated with an ultrasonic probe for 30 min. The solution was then subjected to UV irradiation at 254 nm for 5 h to yield a pale yellow solution of photo-polymerized product. Deionized water was added to replace the volume that was lost by evaporation during the photo-polymerization process. The dialyses were performed in 3000 MWCO dialysis membranes (Thermo Fisher) against a 1000 times larger volume of slightly acidic water (0.1% v/v of HCl 37%) over 7 days.
Footnote |
| † Electronic supplementary information (ESI) available: Copies of NMR spectra, epifluorescence and confocal microscopy images, and DLS profiles. See DOI: 10.1039/c9na00571d |
| This journal is © The Royal Society of Chemistry 2019 |