Journal of Controlled Release 153 (2011) 163–172

Contents lists available at ScienceDirect

Journal of Controlled Release

j ourna l homepage: www.e lsev ie r.com/ locate / jconre l

GENEDELIVERY

Insights into the cellular trafficking of splice redirecting oligonucleotides complexedwith chemically modified cell-penetrating peptides

Fatouma Said Hassane a, Rachida Abes a, Samir EL Andaloussi b,c,1, Taavi Lehto c, Rannar Sillard b,Ülo Langel b,c, Bernard Lebleu a,⁎a University Montpellier 2, UMR 5235 CNRS, place E. Bataillon, 34095 Montpellier cedex 5, Franceb University of Tartu, Institute of Technology, Nooruse 1-517, 50411 Tartu, Estoniac Stockholm University, Department of Neurochemistry, 10692 Stockholm, Sweden

⁎ Corresponding author at: UMR 5235 CNRS, CC1Montpellier cedex 5, France. Tel.: +33 467149203; fax:

E-mail address: blebleu@univ-montp2.fr (B. Lebleu)1 Present affiliation: Karolinska Institutet, Departmen

86 Huddinge, 10692 Stockholm, Sweden.

0168-3659/$ – see front matter © 2011 Elsevier B.V. Aldoi:10.1016/j.jconrel.2011.04.013

a b s t r a c t

a r t i c l e i n f o

Article history:Received 17 December 2010Accepted 13 April 2011Available online 20 April 2011

Keywords:Cell penetrating peptidesOligonucleotides deliverySplicing redirectionEndocytosisEndosomal escape

Conjugates of cell-penetrating peptides (CPP) and splice redirecting oligonucleotides (ON) display clinicalpotential as attested by in vivo experimentation in murine models of Duchenne muscular dystrophy.However, micromolar concentrations of these conjugates are required to obtain biologically relevantresponses as a consequence of extensive endosomal sequestration following endocytosis. Recent work fromour group has demonstrated that appending stearic acid to CPPs increases their efficiency and that theinclusion of pH titrable entities leads to further improvement. Moreover, these modified CPPs form noncovalent complexes with charged ON analogs or siRNAs, which allows decreasing the concentrations of ONsby nearly one log.These modified CPPs and the parent peptides have been compared here in the same in vitro model in terms ofcell uptake, trafficking and splicing redirection activity. The increased splicing redirection activity of ourmodified CPPs cannot be explained by differences in cell uptake but rather by their enhanced ability to escapefrom endocytic vesicles. Accordingly, a clear correlation between membrane destabilizing activity andsplicing redirection was observed using a liposome leakage assay.Studies of cellular trafficking for the most active PF6:ON complexes indicate uptake by clathrin-mediatedendocytosis using either FACS cell uptake or a splicing redirection functional assay. Acidification ofintracellular vesicles and membrane potential were found important for splicing redirection but not for celluptake. These results do confirm that the increased potency of PF6:ON complexes is not due to the use of a nonendocytic route of cell internalization as proposed for some CPPs.

07, place E. Bataillon, 34095+33 467149201..t of Laboratory Medicine, 141

l rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Cell penetrating peptides (CPPs) have attracted considerableinterest as delivery vectors for cargoes ranging from low molecularmass drugs to large biomolecules such as antisense oligonucleotides(ONs) or siRNAs [1]. CPPs are arginine-rich and/or amphipathicpeptides which can be conjugated to neutral ONs or complexed withcharged ones. Antisense ONs are powerful tools with widespreadapplications either in gene down-regulation or in pre-mRNA splicingredirection. As an example, skipping mutated exons through thehybridization of steric block ON analogs such as 2′OMethyl phosphor-othioate (2′OMePS) or phosphorodiamidate morpholino oligomers

(PMO) are considered a promising strategy for the treatment ofDuchenne muscular dystrophy [2].

Clinical applications of ON-based strategies have however beenhampered by the poor bioavailability of ONs, a problem which can beat least partly circumvented by CPP-assisted delivery. Various cellularuptake mechanisms have been described for CPP-cargo conjugatesand complexes, ranging from direct receptor- and energy-independenttranslocation across the plasma membrane to endocytosis. The uptakemechanism of CPPs has remained a controversial issue although mostreports suggest that arginine-rich CPP-ON conjugates preferentiallyutilize endocytosis after initial binding to cell surface glycosaminogly-cans, at least at low peptide concentrations [3–6]. Endocytosis itself isa multifaceted cellular process involving several pathways (clathrin-,caveolin-, and clathrin/caveolin independent endocytosis as well asmacropinocytosis), often resulting in endosomal sequestration of theinternalized material and subsequent lysosomal degradation. Promot-ing endosomal escape is therefore considered pivotal for efficientdelivery of CPP ON constructs to the cytosol (for siRNAs) or nucleus (forsplice redirecting ONs).

164 F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

Improved endosomal release through a proton sponge effect hasbeen first proposed to explain the relatively high plasmid DNAtransfection efficiency of polyethylene imine, knowing its pHbuffering capacity [7]. Other groups have described the use offusogenic peptides such as Influenza HA2 in association with CPP toimprove endosomal release [5,8]. Including or appending pH-titrablegroups such as histidines to CPPs [9] has been shown to facilitateendosomal escape. Appending fatty acid chains has also beenexplored. Stearylation of oligo Arg was first shown beneficial [10]and this strategy has been extensively used since then to increase theefficiency of nucleic acids transfection [11].

Our groups have recently capitalized on similar strategies toimprove the transfection efficiency of Transportan10 (TP10), anamphipathic CPP, and of RXR (with X for aminohexanoic acid), anarginine-rich CPP. The interest of these modifications was attested bythe increased splicing redirection activity of stearylated RXR ascompared to the parent RXR [12]. Likewise, PF3, a stearylatedderivative of TP10, showed more activity than the parent CPP forthe delivery of splice redirecting 2′OMe PS ONs [13]. Furthermodification of PF3 (through the inclusion of pH titrable trifluor-omethyl quinoline derivative groups) gave rise to PF6, which wasrecently shown to be superior to other tested CPPs as well ascommercial transfection reagents for the delivery of siRNAs intovarious types of cells [14]. Importantly, these modified CPPs formnanoparticles upon complexation in a non-covalent manner witheither 2′OMe PS ONs [13] or siRNA [14] thus avoiding costly and time-consuming covalent conjugation.

We here compare these modified CPPs using the splicingredirection assay used by many groups for the characterization ofON-delivery vectors. We also establish that the increased biologicalactivity of stearylated RXR, PF3 and PF6 is not due to increased celluptake or to the use of an alternative endocytic pathway as comparedto the parent CPPs. Finally, a correlation between biological activityand membrane destabilization ability is established using a liposomalmodel mimicking endosomes in terms of membrane composition andpH.

2. Materials and methods

2.1. Synthesis of ONs and CPPs

Desalted 2′OMethyl phosphorothioate (2′OMe PS) oligonucleo-tides (ON) and their Cy5-tagged derivatives (CCU CUU ACC UCA GUUACA) were obtained from Eurogentec (Seraing, Belgium).

The peptides were synthesized on a Syro II multiple peptidesynthesizer (MultiSynTech, Germany) on Fmoc-Rink-amide-4-methylbenzhydrylamine resin (0.67 mmol/g, IRIS Biotech, Germany)by a standard Fmoc solid-phase peptide synthesis protocol using 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate/1-hydroxybenzotriazole (TBTU/HOBt, IRIS Biotech, Germany) activationin dimethylformamide (DMF) [15]. Stearic acid (Sigma-Aldrich) wascoupled to the N-terminus of the peptides by using TBTU/HOBtactivation in dimethylformamide/dichloromethane (DMF/DCM) essen-tially as described [13].

For PF6 synthesis, the pH titrable moiety 4-chloro-7-(trifluoro-methyl)quinoline and N-methyl-2,2′-diaminodiethylamine were stir-red in a PEG400 bath at 130 °C for 6 h. After cooling to 22 °C, DCMwasadded and the precipitate was removed by filtration. The organicphase was washed with 5% NaHCO3 in water, and dried withanhydrous NaSO4. DCM was evaporated in vacuum to give N-(2-aminoethyl)-N-methyl-N′-[7-(trifluoromethyl)-quinolin-4-yl]eth-ane-1,2-diamine (compound 1, N90% purity by HPLC), which wasused for coupling to the peptide.

PF6 was synthesized using a N-ε-4-methyltrityl (Mtt) protectedlysine at position 7. The N-terminus of the peptide was stearylatedand the Mtt group removed by repetitive washes with 1% trifluor-

oacetic acid (TFA) in DCM. A Lys-tree was built on the free ε-aminogroup using two rounds of Fmoc-Lys(Fmoc)-OH coupling and Fmocgroup removal. After the second Fmoc removal, the four generatedamino groups were treated with excess succinic anhydride in thepresence of DIEA. Finally, compound 1 was coupled using benzo-triazole-1-yl-oxy-tris-pyrrolidino-phosphonium hexafluoropho-sphate (PyBOP) activation in DMF.

The peptides were cleaved using 95% TFA/2.5% water/2.5%triisopropylsilane for 3 h and precipitated in diethylether. The crudepeptides were dried in vacuum overnight and purified by HPLC on aDiscovery® C-18 Supelco® column (Sigma-Aldrich, Sweden) using agradient of acetonitrile/water containing 0.1% TFA. The identity of thepurified product was verified by analytical HPLC and by matrix-assisted laser desorption ionization time-of-flight mass spectrometryon a Perkin Elmer prOTOF™ 2000 (Perkin Elmer, Sweden). The mass-spectrum was acquired in positive ion reflector mode using 2,6-dihydroxyacetophenone/diammonium hydrogen citrate as matrix(Sigma-Aldrich) (20:40 mg/ml, 50:50 methanol:water). The molarityof the peptide was determined based on dilutions of accuratelyweighed substances.

2.2. CPP:ON complex formation

30 μl of 20 μMdesalted 2′OMe PS were mixed with 60 μl of 100 μMpeptide (TP 10 or its stearylated versions PF3 and PF6) in a 1.5 mlEppendorf tube to obtain a molar ratio of 10/1 (peptide/ON). RXR andstearylated-RXR (St-RXR) were complexed with ON at molar ratios of3/1 or 6/1. Complexation was carried out in MQwater for 1 h in a finalvolume of 300 μl. Serial dilutions were conducted to obtain final ONconcentrations ranging from 200 nM to 50 nM.

2.3. Luciferase assay of splicing redirection

5×104 exponentially growing HeLa pLuc 705 cells were seededovernight on 24-well plates. Cells were then washed twice withPBS and incubated with 450 μl optiMEM. 50 μl of preformed CPP:ON2′OMe PS complexes was added in eachwell and the plate was furtherincubated for 4 h at 37 °C. Thereafter, 1 ml of DMEM supplementedwith 10% FBSwas added to cells and incubationwas continued for 20 hat 37 °C. Supernatants were removed and cells were washed twicewith PBS before lysis with RLB buffer (Promega). Total protein wasquantified using a BCA assay (Pierce) and luciferase activity wasdetermined using the luciferase assay kit (Promega) on a BertholdCentro LB 960 luminometer.

In experiments involving the use of chloroquine, cells wereincubated with the CPP:ON complexes for 1 h at 37 °C. Cells werethen washed twice with PBS and further incubated at 37 °C with100 μM chloroquine in optiMEM for 1 h. Cells were washed twice,1 ml of 10% FBS supplemented DMEM was added in each well andincubation was continued at 37 °C for 22 h. Luciferase activity wasanalyzed as previously described.

To investigate the influence of temperature on splicing redirectionactivity, cells were pre-incubated for 30 min in optiMEM at 4 °C andthen treated with CPP:ON complexes for 1 h at the same temperature.For ATP depletion, cells were pre-incubated with 10 mM NaN3 and6 mM 2-deoxyglucose in optiMEM for 1 h at 37 °C, and then treatedwith CPP:ON complexes for 1 h. Cells were washed twice with PBS,1 ml of complete DMEM was added in each well and incubation at37 °C was continued for a total time of 24 h. Luciferase activity wasanalyzed as previously described.

2.4. RT-PCR analysis of splicing redirection

Total RNA was extracted from aliquots of the cellular lysatesobtained as described in the previous section using the RNA isolationkit Tri reagent (Sigma). Amplification was conducted in the presence

165F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

of forward (TTGATATGTGGATTTCGAGTCGTC) and reverse (TGTCAAT-CAGAGTGCTTTTGGCG) primers using the superscript III One-Step RT-PCR System (Invitrogen) in a MJ Research PTC200 Peltier Thermalcycler. The cDNA products were separated by electrophoresis on a 2%agarose gel. Images were recorded and analyzed on a Lumi Imager F1(Roche) with Lumianalyst software.

2.5. Flow cytometry analysis of cell uptake

Exponentially growing HeLa pLuc 705 cells were seeded on 24-well plates 1 day before treatment (2×105cells/well). Cells werewashed twice with PBS and incubated with 450 μl optiMEM. 50 μl ofpreformed ON:CPP complexes was added in each well to obtain a finalconcentration of 200 nM Cy5-labeled ON 2′OMe PS and plates wereincubated for 2 h at 37 °C. Cells were washed twice with PBS andtrypsinized with 100 μl EDTA-trypsin per well for 5 min at 37 °C.Trypsinationwas stopped by adding 500 μl of 5% FBS in PBS. Cells werebriefly centrifuged (5 min at 900×g) and resuspended in 0.5% FBSbefore FACS analysis. Fluorescence was quantified by flow cytometryusing a BD FACS Canto system and counting at least 10,000 events persample.

2.6. Mechanism of endocytosis studies

The influence of specific inhibitors for different endocyticpathways on the uptake of CPP:ON complexes was investigatedboth by monitoring cell internalization and splicing redirectionactivity. Cell uptake was monitored by FACS analysis using Cy5-labeled 2′OMe PS ON in the absence or presence of the inhibitors.HeLa pLuc 705 cells were seeded 20 h prior to treatment in 24-wellplates (2×105 cells/well). Cells were washed twice with 500 μl PBSand pre-incubated in OptiMEM at 37 °C for 20 min with the followingendocytosis inhibitors.

Chlorpromazin (30 μM) or K+ depletion was used to inhibitclathrin-mediated endocytosis, nystatin (50 μg/ml) for caveolaeendocytosis, and methyl-β-cyclodextrin (MBCD) (2.5 mM) or 5-(N-ethyl-N-isopropyl) amirolide (EIPA) (10 μM) to inhibit macropinocy-tosis, NH4Cl (50 mM) was used to prevent acidification of endocyticvesicles and isotonic KPBS (KaCl/169.6 mM, Na2HPO4/1.76 mM,KH2PO4/0.1%, glucose/pH7.4) buffer was used to decrease membranepotential. Cells were then washed twice with PBS and incubated withCy5-labeled 2′OMe PS:CPP complexes at 37 °C for 30 min in OptiMEMusing the same concentrations of inhibitors. Reactions were stoppedby washing cells twice with PBS and cells were processed for FACSanalysis as described above.

The effect of these inhibitors on the splicing redirection activitywas monitored using unlabelled 2′OMe PS ONs. HeLa pLuc 705 cellswere seeded 20 h prior to treatment in 24-well plates (1.75×105

cells/well). Cells were washed twice with 500 μl PBS and pre-incubated in OptiMEM at 37 °C for 20 min in the presence or in theabsence of endocytosis inhibitors as above. Cells were washed twicewith PBS and incubated in OptiMEM with the same concentrations ofinhibitors in the presence of 2′OMePS:CPP complexes at 37 °C for30 min. Cells were then assayed as previously described for luciferaseexpression.

2.7. Liposome leakage assay

To mimic the lipid composition of late endosomes, largeunilamellar liposomes (LUV) were prepared with the following lipidcomposition: dioleoylphosphatidylcholine (DOPC)/dioleoyl-phospha-tidylethanolamine(DOPE)/phosphatidylinositol from soybean (PI)/bis(monooleoylglycero) phosphate (LBPA) (Avanti polar lipids) [16].

10 μmoles of the phospholipids mixture (DOPC/DOPE/PI/LBPA: 5/2/1/2) in organic solvent (chloroform/ethanol: 9/1) was dried byevaporation under nitrogen atmosphere and rehydrated using 1 ml of

20 mM MES buffer containing 12.5 mM ANTS (8-aminonaphthalene-1,3,6-trisulfonic acid), disodium salt fluorochrome (Invitrogen),45 mM DPX (p-xylene-bispyridinium bromide) quencher (Invitro-gen) and 75 mM NaCl, pH 5.5. The lipid suspension was vigorouslyagitated with a Vortex and submitted to 5 cycles of freeze/thawing(liquid nitrogen/water bath 37 °C) before passaged 21 times througha 100 nm polycarbonate filter (Nucleopore, Whatman) at roomtemperature using a mini extruder (Avanti polar lipids). Free dyeand quencherwere then removed by gel filtration on a Sephadex G-50column (Amersham Biosciences) and liposomes were eluted with20 mM MES buffer, 145 mM NaCl, pH 5.5. The size distribution wasanalyzed on a submicron particle size analyzer (Coulter N4, Beckman)and phosphatidylcholine was quantified using a LabAssay phospho-lipids kit (Wako) according to manufacturer's instructions.

Liposomes mimicking the lipid composition of the plasmamembrane were prepared with DOPC, DOPE, sphingomyelin (SM)and cholesterol (Chol) (1/1/1/1.5) and processed as described above[17].

Liposomal leakage was measured on a luminescence spectrometer(Perkin Elmer LS 55) at pH 5.5 or 7.4 as an increase in fluorescenceintensity of ANTS upon addition of free CPPs or CPP:ON complexes(2 to 10 μM final concentrations) to 2 ml of the liposome solution(100 μM). Infinite dilution of the probe was used to determinefluorescence of the completely unquenched probe and was achievedby adding 0.1% (v/v) Triton X-100.

3. Results

3.1. N-terminal stearylation and inclusion of pH titrable groups increasessplicing redirection by CPP:ON complexes

It has been previously shown that N-terminal stearylation of TP10[13] or RXR [12] increases their ability to deliver charged spliceredirecting ON. In addition, the inclusion of pH titrable groups (i.e.trifluoromethyl quinoline derivatives) in stearylated TP10 increasesits efficacy dramatically for siRNA delivery in various biologicalmodels [14]. However, the PF6:ON formulation had not beenvalidated for the delivery of splice redirecting antisense ONs.

We thus compared the efficacies of all modified CPPs (Table 1)with the parent peptides and with RXR-PMO conjugates in the samesplicing redirection model [18]. In brief, a mutated β-globin intron 2has been introduced in the coding region of a luciferase reporter geneand the construct has been stably transfected in HeLa cells (HeLapLuc705). Incomplete removal of the intron from the pre-mRNA givesrise to an aberrant and non functional luciferase mRNA. Nucleardelivery of an RNaseH-incompetent ON analog (such as a 2′OMe PSON) masks the cryptic splice site and allows the production of acorrectly spliced and functional mRNA. This splicing redirection assayis indeed advantageous over other antisense ON assays in providing apositive read-out over a low background and in having a largedynamic range.

CPPs were merely mixed with a desalted 2′OMe PS spliceredirecting ON at various peptide/ON ratios. Stable complexes wereformed in deionized water at room temperature during 1 h and wereused to transfect HeLa pLuc 705 cells. Luciferase up-regulation was 18and 90 times higher for PF3 and PF6 respectively in comparison to TP10 at the optimal 10/1 peptide:ON ratio using 200 nM ON.Remarkably, significant luciferase expression was already achievedat doses as low as 50 nM with a dose–response effect with PF3 andPF6:ON complexes (Fig. 1A and B). Likewise, the stearylation of RXRpeptide (i.e. st-RXR) increased splicing redirection at the optimal 6/1molar ratio at 200 nM ON as compared with the non stearylatedparent RXR (Fig. 1C). It is noteworthy that luciferase expression withthese CPP:ON non-covalent complexes was higher than with RXR-PMO conjugates (Fig. 1C). To further support these results, RT-PCRexperiments, that allow discriminating correctly spliced products

Table 1Structures of TP 10, RXR and their derivatives.

166 F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

from aberrant ones, were performed using aliquots of the cellularextracts used for luciferase assays. The same hierarchy was foundfor these different peptidic vectors with PF6NPF3NTP10 and St-RXRNRXR (Fig. 1D). Noticeably, using 200 nM ON complexed with PF6lead to complete splicing correction while PF3:ON complexes onlycorrected half of the misspliced mRNA at this concentration inkeeping with our previous data on siRNA delivery [14]. Likewise, StRXR lead to significant splicing correction at 200 nM ON whileunmodified RXR (Fig. 1D) or RXR-PMO conjugates [19] were inactiveat these submicromolar concentrations.

Altogether, these data fully confirm our previous observationsshowing that (i) the addition of a stearyl moiety on TP10 and RXRincreases their activity and that (ii) the inclusion of a pH titrablegroup has an additional benefit also in this splicing redirectionassay.

3.2. Increased splicing redirection activity is not due to enhanced celluptake

In order to assess if enhanced internalization could explain theincreased activity of these modified CPPs, cellular uptake wasmonitored in HeLa pLuc 705 cells. Peptides were mixed with a Cy5fluorochrome-labeled ON at their optimal molar ratios. Cells wereincubated with the preformed complexes for 2 h at 37 °C. Cellmembrane adsorbed material was carefully eliminated by cellwashing and trypsination before flow cytometry analysis [3].

PF3 or PF6 derivatives were taken up slightly more efficiently thanTP10 (Fig. 2A). Moreover PF3 and PF6:ON complexes were taken up tothe same extent and have significantly different splicing redirectionactivity. Likewise, the stearylated RXR derivative was internalizedslightly better than RXR (Fig. 2B). Altogether these data do indicatethat the largely improved biological activity of ONs delivered ascomplexes with these modified CPPs is only marginally caused bydifferences in cellular uptake.

3.3. CPP:ON complexes are internalized by an energy-dependentmechanism

The mechanism by which CPPs enter cells is still a controversialissue even though endocytosis appears to be the prevailing pathwayfor Arg-rich CPP-ON conjugates [20]. It is worth pointing that

nanoparticles formed by the non covalent association ofMPG peptideswith nucleic acids are claimed to be internalized by direct transloca-tion across the plasma membrane [21]. We thus felt important todetermine whether increased biological activity of these new CPP:ONcomplexes could be due to the use of a non-endocytic route. We firstexamined the splicing redirection activity of these peptides at low(conditions preventing any form of endocytosis) and normaltemperature. We also measured the effect of ATP depletion on theactivity. Low temperature incubations and ATP deprivation werecarried out for short periods of time (30 min preincubation and 1 hincubation in the presence of the pre-formed CPP:ON complexes) tolimit cell stress. The correction levels mediated by PF3, PF6 and St-RXRwere significantly reduced by lowering the temperature or bydepleting ATP (Fig. 3A) in keeping with a predominant energy-dependent mechanism as previously shown for the parent CPPs[22,23].

Incubation at low temperature (Fig. 3A) does not completelyinhibit cell uptake or splicing redirection. This is most probably due tothe fact that trypsin treatment is unable to fully eliminate membrane-bound CPP/ON complexes. We therefore performed control experi-ments in which cells were incubated with Sr-RXR:ON complexes(data not shown) or with PF6:ON complexes (Fig. 3B) for increasingperiod of times before FACS analysis. A background level was rapidlyreached at 37 °C or at 4 °C in endocytosis-inhibitory conditions whichis probably due to non specific adsorption to the plasma membrane.Increased accumulation was indeed observed upon incubation at37 °C while no increase was observed at 4 °C in keeping with anenergy-dependent pathway of internalization for these CPP:ONcomplexes.

The use of a panel of endocytosis inhibitors and the evaluation oftheir effects on cell uptake and on splicing redirection activity of PF6:ON complexes (as will be documented below) fully confirmed the useof an endocytic pathway.

3.4. The increased splicing redirection activity of PF6- and St-RXR:ONcomplexes is due to an increased membrane destabilization potential

The increased splicing redirection activity of PF6- and St RXR:ONcomplexes is neither due to increased cell uptake nor to the use of anon-endocytic pathway.

Fig. 1.Up regulation of luciferase expression by TP10, RXR and their derivatives complexed to 2′OMe PS: A) Comparison of PF3 (black bars) and TP10 (open bars) in dose–response at10/1 molar ratio. B) Comparison of PF6 (black bars) and TP10 (open bars) in dose–response at 10/1 molar ratio. C) Comparison of St-RXR (black bars) and RXR (white bars) in dose–response at 6/1 ratio. Splicing redirection by a RXR-PMO conjugate at 1 μM (white bar). CPP:2′OMePS complexes or RXR-PMO conjugate were incubated at the indicatedconcentrations for 4 h and assayed for luciferase expression as described in Methods. Experiments were made in triplicate and error bars are indicated. Data are expressed in relativeluciferase units per microgram protein. D) RT-PCR analysis of splicing products. An aliquot of each cell extract was analyzed by RT PCR as described in Methods. 1: Negative control;2: TP10 10/1 200 nM; 3: TP10 10/1 100 nM; 4: PF3 10/1 200 nM; 5: PF3 10/1 100 nM; 6: PF6 10/1 200 nM; 7: PF6 10/1 100 nM; 8: St-RXR 6/1 200 nM; 9: St-RXR 6/1 100 nM; 10: RXR6/1 200 nM; 11: RXR 6/1 100 nM.

167F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

We thus examined whether chloroquine, a well-known andwidely used endosomolytic agent, affects splicing redirection effi-ciency to the same extent for unmodified and for modified CPPs. Inthese experiments, cells were incubated with CPP:ON complexes for1 h to allow cell internalization. Cells were then washed and further

incubated for 1 h with 100 μM chloroquine. Chloroquine treatmentincreased splicing redirection significantly more in the cases of PF3 orstearylated RXR than for PF6 (Fig. 4), in line with a higher intrinsicendosome destabilization potential for PF6 and with its increasedsplicing redirection activity.

Fig. 2. Flow cytometry analysis of CPP:ON cellular uptake: A: TP10, PF3 and PF6. B: RXRand St-RXR. CPP:Cy5-2′OMePS complexes (200 nM) were incubated at the indicatedmolar ratios for 2 h and processed for FACS analysis as indicated in Methods.Experiments were made in triplicate and error bars are indicated.

Fig. 3. Cell uptake is energy-dependent: A: Effect of temperature on splicing redirection.Cells were preincubated at 37 °C (black bars), 4 °C (gray bars) or at 37 °C in thepresence of NaN3 and deoxyglucose (white bars) for 1 h and further incubated in thesame conditions with CPP:2′OMePS complexes at the indicated concentrations andmolar ratios. Luciferase expression assays were made after 24 h as indicated inMethods. Experiments were made in triplicate and error bars are indicated. Data areexpressed as percent of splicing redirection residual activity. B: Effect of temperature oncell uptake in time-course experiments. Cells were preincubated at 37 °C (squares) or4 °C (triangles) for 30 min, further incubated for the indicated times in the sameconditions with PF6:2′OMePS (10/1 ratio) and processed for FACS analysis as indicatedin Methods. Experiments were made in triplicate and error bars are indicated.

Fig. 4. Effect of chloroquine on splicing redirection: Cells were incubatedwith CPP:Cy5-2′OMePS complexes at the indicated concentrations (ratio 10/1 for PF3 and 6/1 for St-RXR)for 1 h, washed and further incubated without (white bars) or with (black bars) 100 μMchloroquine for 1 h. Luciferase expression assays were made after 24 h as indicated inMethods. Experiments were made in triplicate and error bars are indicated. Data areexpressed as relative luciferase units per microgram protein.

168 F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

In order to test whether chemical modifications of CPPs have aninfluence on membrane destabilization, we made use of a modelassay for endosomal release. The characterization of late endoso-mal membranes has revealed that in addition to phosphatidylcho-line, phosphatidylethanolamine and phosphatidylinositol,lysobisphosphatidic acid is an abundant and specific component(accounting for 15% of the total phospholipids) of late endosomes[16]. We have thus developed a liposomal model (Fig. 5A) whichconsists of large unilamellar vesicles (LUV), with a lipidiccomposition reflecting late endosomes membrane. A fluorescentANTS (8-aminonaphthalene-1,3,6-trisulfonic acid) dye was encap-sulated in the liposomes and quenched by DPX (p-xylene-bispyridinium bromide).

Membrane destabilization by Triton X-100 (0.1% v/v for 100 μMlipids) leads to dequenching of the fluorophore and to a signalincrease corresponding to 100% leakage.

This model is advantageous when comparing the membranedestabilization potential of several free CPPs or CPP:ON complexeson the same liposome preparation. In addition, the leakage effect offree peptides and peptide:ON complexes can be investigated atneutral or acidic pH (pH 5.5), thus mimicking endosomalconditions.

TP10, PF3 and PF6 were first complexed with ONs at the optimal10/1 ratio. 50 μl of the pre-formed complexes (2 μM CPP:200 nM ON)was incubated with 100 μM lipids and the kinetics of leakage wererecorded during 30 min at acidic or neutral pH. Liposomes wereperfectly stable and the ON alone had no effect on the liposomes (datanot shown). The PF6:ON complex had a higher destabilizationpotential as compared to its PF3 analog while the impact of TP10was barely measurable (Fig. 5B and C). The leakage was clearly pH-

dependant for PF6 (Fig. 5C) in keeping with its chemical structureincluding four proton-accepting trifluoromethylquinoline-derivativeresidues (Table 1).

169F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

St-RXR was also tested at a 6/1 ratio with ON (Fig. 5D) and lead tosignificant leakagewhile the RXR-PMO conjugates gave no signal evenwhen used at high concentration (10 μM).

Taken together, these data indicate a good correlation between thesplicing redirection activity and the leakage effect of all scanned CPP:ONcomplexes, suggesting that N-stearylation of peptides and inclusion ofpH-titrable groups induce greater endosomal release.

Whether themembrane destabilizing potential of our constructs atneutral pH could reflect their ability to cross the plasma membranecould not be eliminated by these experiments. We thereforecapitalized on recent studies indicating that CPPs as Tat had nodestabilizing effect on liposomes with a lipid composition reflectingthe plasma membrane while being active on liposomes mimickingendosomes [17]. Similar data were obtained with PF6:ON complexesupon incubation at neutral pH (data not shown).

3.5. PF6:ON complexes are internalized via clathrin-mediatedendocytosis

All three major routes of endocytosis namely clathrin- [4] andcaveolae-mediated endocytosis as well as macropinocytosis [5] havebeendocumented for various CPP-cargo constructs. Although TP10 [24],Tat [4] and RXR [19] mainly use the clathrin-coated pits pathway todeliver chemically conjugated neutral ONs, the TP10- and RXR-basednanoparticles could possibly traffic through other routes. Moreover,using an alternative pathway could hypothetically contribute to thelargely increased activity of our modified CPPs as compared to theparent peptides. Since PF6-based complexes were the most efficient todeliver splice redirecting ON, we have limited our investigations to thisformulation.

The role of each endocytic pathway was probed with specificinhibitors [25] using both splicing redirection and cell uptake assays.For this purpose, we used nystatin as inhibitor of caveolae-mediatedendocytosis, methyl-β-cyclodextrin and EIPA 5-(N-ethyl-N-isopro-pyl) amirolide as inhibitors of macropinocytosis, and chlorpromazineor potassium depletion as inhibitors of clathrin-mediated endocytosis[4]. NH4Cl and an isotonic KPBS buffer were employed to assess theimportance of endocytic vesicles-acidification and membrane poten-tial, respectively [26].

For these experiments, incubation times with PF6:ON complexeswere reduced in order to avoid cytotoxicity of the inhibitors [4]. Cellswere pre-incubated for 20 min with the inhibitors and then for30 min more in the presence of PF6:ON complexes. ON uptake and, asa consequence, luciferase activity levels were naturally reduced withthese shorter incubations.

Potassium depletion and chlorpromazine reduced the uptake andthe splicing redirection activity of the PF6:ON complexes (Fig. 6A andB). In contrast, neither caveolae nor macropinocytosis inhibitors had asignificant effect on cell uptake or on splicing redirection.

Taken together, these results indicate that PF6:2′OMe PS com-plexes are internalized preferentially by a clathrin-dependentendocytic pathway and do not require raft/caveolin-dependentendocytosis or macropinocytosis, in line with our previous data onthe trafficking of CPP-ON conjugates [19].

As pointed in Section 3.3, incubation in endocytosis-inhibitoryconditions does not completely inhibit cell uptake or splicingredirection and this probably results frommembrane-bound materialwhich has not been removed by trypsinization. As a control, we haveverified that the uptake of PF6:ON complexes does not increase withincubation time in chlorpromazine-treated cells as opposed to controlsamples (data not shown).

Several publications have shown that membrane potential isrequired for CPP-based delivery [27,28].Whethermembrane potentialis required for translocation across the plasma- or the endosomemembrane has, however, remained controversial. As an example,Zhang et al. [26] have recently shown that the incubation of HeLa cells

in an isotonic/high K+ buffer such as KPBS (e.g. in conditions leadingto membrane depolarization) did not affect Tat endocytosis butreduced its escape from endocytic vesicles. As shown here, KPBStreatment had no effect on cell uptake but reduced splicing redirection(Fig. 6A and B), thus supporting the data on Tat translocation.

NH4Cl, which is known to neutralize the acidification of endocyticcompartments [26], caused a significant reduction in PF6:ON splicingredirection activity in keeping with their passage through acidiccompartments and a requirement for endosomal acidification(Fig. 6A). The uptake of the complexes was, however, not affected(Fig. 6B).

4. Discussion

Previous work in our groups has evaluated the potential ofN-stearylation [12,13] and of pH-titrable groups [14] on CPP activity.These modified RXR and TP10 peptides have all been compared here inthe splicing redirection assay, which is now used as a standard bymanygroups in the ON-delivery field [18]. As seen in Fig. 1A and C, thecomplexation of st-RXR or PF3 (stearylated TP10) with 2′OMet PS ONleads to splicing redirection at much lower concentrations than theparent CPPs. The inclusion of pH-titrable groups in PF3 (i.e. PF6)increases splicing redirection activity further (Fig. 1B). As sake ofcomparison, splicing redirection has been monitored for an RXR-PMOconjugate [19]. All compounds have also been compared within thesame experiment, monitoring the production of correctly splicedluciferase mRNA (Fig. 1D). Again, PF6:ON turned out as the mostefficient formulation with nearly completed splicing correction at100 nMON concentration. As a point of comparison, RXR-PMO [19] andPip-2b-PNA [27] which are considered as the most efficient deliveryvectors for steric-block ON derivatives have EC50 in the micromolarrange in this assay. The PF6:ON formulation thus appears as the mostefficient formulation so far for splice redirecting ONs and offersinteresting prospects of applications, for example in the treatment ofDuchenne muscular dystrophy. Since sequestration within endocyticvesicles is considered as amajor limitation for the CPP-based delivery ofbiomolecules, the increased activity of st-RXR, PF3 and PF6 as comparedto our referenceRXR-PMOconjugate and to the parent unmodified CPPscould possibly be attributed to amore efficient escape from endosomes.Alternative explanations could however be proposed and several havebeen addressed here.

Cellular uptake, as monitored by FACS, cannot explain the observeddifferences in splicing redirection activity as seen in Fig. 2A and B. Sincethe uptake of CPPs involves binding to cell surface glycosaminoglycansthrough electrostatic interactions [4,5], N-stearylation and/or inclusionof trifluoromethylquinolone groups were not expected to alter GAG-binding and subsequent cell uptake.

Most recent studies using basic CPPs such as Tat [3] and RXR [19]or more hydrophobic CPPs such as TP10 [23] agreed on theinvolvement of an endocytic pathway as the major entry route forCPPs and their conjugates. An alternative pathway involving directtranslocation across the plasma membrane has however beendescribed [29]. Interestingly, increasing local concentration of CPPsabove a threshold value is often proposed as an explanation for theshift between endocytosis and direct translocation [6]. It is also worthpointing here that a non-endocytotic cell internalization route hasbeen proposed for amphipathic peptide (MPG or CADY)-basednanoparticles [21]. Avoiding the problems of trafficking throughendocytic vesicles could have been an explanation for the improvedbiological activity of the non-covalent complexes formed between ourmodified CPPs and splice correcting ONs. As seen in Fig. 3, lowtemperature incubation or ATP deprivation decreased splicingredirection by our CPP:ON complexes, in line with the use of anenergy-dependent mechanism. Along the same lines, clathrin-mediated endocytosis inhibitors (chlorpromazine or K+ depletion)also inhibited cell uptake (Fig. 6A) and splicing redirection (Fig. 6B).

A

B

C

D

170 F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

Fig. 6. Effect of endocytosis inhibitors on cell uptake (A) and splicing redirection activity (B) of PF6:2′OMePS complexes: Cells were preincubated at 37 °C in the presence ofendocytosis inhibitors for 30 min (except for KPBS for which there was no preincubation and for K+ depletion buffer for which the preincubation was shortened to 5 min) andfurther incubated for 30 min in the same conditions with CPP:2′OMePS complexes at 200 nM and a 10/1molar ratio. (A) For cell uptake, CPP:Cy5-2′OMePS complexes wereprocessed for FACS analysis as indicated in Methods. Experiments were made in triplicate and error bars are indicated. Data are expressed as mean fluorescence. (B) Luciferaseexpression assays were made after 24 h as indicated in Methods. Experiments were made in triplicate and error bars are indicated. Data are expressed as relative luciferase units permicrogram protein.

171F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

Altogether, the data indicate that the higher splicing redirectionactivity of these CPP:ON complexes was neither due to increaseduptake nor to the use of a non-endocytic route thus suggesting anincreased efficiency to escape endocytic vesicles. In line with thisassumption, chloroquine treatment (which is known to favor therelease of entrapped material from endocytic vesicles) increasedsplicing redirection significantly more for PF3- or St-RXR-basedcomplexes than for PF6-based ones (Fig. 4).

We then made use of a liposome leakage assay aiming atquantifying the membrane destabilization potential of these CPP:ONcomplexes. In this model, large unilamellar vesicles were made with alipid composition similar to that of late endosomes. Quenching of theentrapped ANTS fluorochrome with DPX leads to a low background offluorescence unless released upon membrane destabilization(Fig. 5A). As seen in Fig. 5C and D, unmodified TP10-based complexesor RXR-PMO conjugates did not lead to any detectable fluorescentsignal, supporting their low splicing redirection efficiency. Incubationwith St-RXR- (Fig. 5D) or PF3-based complexes (Fig. 5B and C) gaverise to a significant increase in fluorescence, and PF6 was by far themost potent peptide in this assay in keeping with its stronger spliceredirecting activity. Importantly, no significant pH dependency wasfound except for PF6-based complexes in line with the inclusion ofpH-titrable groups in this modified CPP. Altogether these experimentsindicate that appending fatty acid chains favor membrane destabili-zation in a pH-independent mode while the inclusion of pH-titrablegroups further increases membrane destabilization at the slightly

Fig. 5. Membrane destabilization potential in a liposome leakage assay: A: Liposomal modeencapsulated in large unilamellar vesicles (LUV). Incubation at neutral or acidic pH leads, indequenching. B: Membrane destabilization effect of TP10 and derivatives PF3 and P 6 wcomparison with RXR-PMO conjugate.

acidic pH representative of endosomes. Although these lipid vesiclesremain a highly simplified model system, the assay sheds light on aplausible explanation for the differences in activity observed betweenour modified CPPs. Additional biophysical studies will be needed todelineate interactions between lipid membranes and these CPP-basednanoparticles. The assay remains nevertheless useful to screenmodified CPPs in future structure-activity studies.

As already pointed, controversies have remained concerning thetrafficking of free CPPs or CPP-cargo conjugates. No data have beenprovided so far (at least to our knowledge) concerning the route ofendocytosis for non covalent CPP: cargo complexes. Using a panel ofinhibitors of the three main pathways of endocytosis, we show herethat PF6:ON complexes are mainly taken up by clathrin-coated pits aspreviously shown for TP10-PNA conjugates [23] or for RXR-PMOconjugates [19]. Importantly our present conclusions were drawnfrom the analysis of cellular uptake (Fig. 6A) and from a functionalassay (Fig. 6B) namely splicing redirection. We however cannotexclude the possible use of several pathways of internalization withdifferences occurring with cell lines and CPP cargoes.

Treatment with NH4Cl did not affect cell uptake as expected butinhibited splicing redirection in line with a trafficking through acidicvesicles and a requirement for acidification to escape endosomes [26].

As pointed by Zhang et al. [26] in a recent study on the traffickingof Tat, few studies [30,28] have investigated the effect of membranedepolarization on CPP delivery. As shown here, cell treatment with anisotonic KPBS buffer did not inhibit the uptake of PF6:ON complexes

l of leakage. ANTS fluorochrome (gray circles) and DPX quencher (black circles) werethe presence of CPPs or CPP:ON complexes, to membrane destabilization and to ANTShen complexed with ON. C: Destabilization effect of St-RXR complexed with ON in

172 F. Said Hassane et al. / Journal of Controlled Release 153 (2011) 163–172

GENEDELIVERY

(Fig. 6A) but reduced splicing redirection (Fig. 6B). As proposed byZhang et al. [26], the membrane potential would be required forendosomal escape (and consequently for splicing redirection as shownhere) more than for translocation across the plasma membrane asoriginally suggested [30].

Altogether the present studies support the view that ourchemically-modified CPP:ON complexes are mainly internalizedusing a clathrin-coated pits endocytic pathway. However the additionof fatty acid chains and, more so, the inclusion of pH titrable groupsease their escape from endocytic vesicles and as a consequence lead toincreased splicing redirection activity of the transported ONs.

Acknowledgments

This work was supported by the Swedish Research Council (VR-NT); by the Center for Biomembrane Research, Stockholm; by Knutand Alice Wallenberg's Foundation; by the EU through the EuropeanRegional Development Fund through the Center of Excellence inChemical Biology, Estonia; by the targeted financing SF0180027s08from the Estonian Government to U.L. It was also funded by CNRS,AFM and ARC grants to B.L. F.S.H was supported by an AFMpostdoctoral fellowship, R.A. by a PhD fellowship from Ligue Françaisecontre le Cancer and S.ELA. by the Swedish Society of MedicalResearch (SSMF).

References

[1] F. Said Hassane, A.F. Saleh, R. Abes, M.J. Gait, B. Lebleu, Cell penetrating peptides:overviewandapplications to the delivery of oligonucleotides, Cell.Mol. Life Sci. 67 (5)(2010) 715–726.

[2] M. Wood, M.J. Gait, H. Yin, RNA-targeted splice-correction therapy for neuro-muscular disease, Brain 133 (4) (2010) 957–972.

[3] J.P. Richard, K. Melikov, E. Vives, C. Ramos, B. Verbeure, M.J. Gait, L.V.Chernomordik, B. Lebleu, Cell-penetrating peptides. A reevaluation of themechanism of cellular uptake, J. Biol. Chem. 278 (1) (2003) 585–590.

[4] J.P. Richard, K. Melikov, H. Brooks, P. Prevot, B. Lebleu, L.V. Chernomordik, Cellularuptake of unconjugated TAT peptide involves clathrin-dependent endocytosis andheparan sulfate receptors, J. Biol. Chem. 280 (15) (2005) 15300–15306.

[5] J.S. Wadia, R.V. Stan, S.F. Dowdy, Transducible TAT-HA fusogenic peptideenhances escape of TAT-fusion proteins after lipid raft macropinocytosis, Nat.Med. 10 (3) (2004) 310–315.

[6] F. Duchardt, M. Fotin-Mleczek, H. Schwarz, R. Fischer, R. Brock, A comprehensivemodel for the cellular uptake of cationic cell-penetrating peptides, Traffic 8 (7)(2007) 848–866.

[7] B. Demeneix, J.P. Behr, Polyethylenimine (PEI), Adv. Genet. 53 (2005) 217–230.[8] P. Lundberg, S. El-Andaloussi, T. Sutlu, H. Johansson, U. Langel, Delivery of short

interfering RNA using endosomolytic cell-penetrating peptides, FASEB J. 21 (11)(2007) 2664–2671.

[9] S. Lo, S. Wang, An endosomolytic Tat peptide produced by incorporation ofhistidine and cysteine residues as a nonviral vector for DNA transfection,Bimaterials 29 (15) (2008) 2408–2414.

[10] S. Futaki, W. Ohashi, T. Suzuki, M. Niwa, S. Tanaka, K. Ueda, H. Harashima, Y.Sugiura, Stearylated arginine-rich peptides: a new class of transfection systems,Bioconjug. Chem. 12 (6) (2001) 1005–1011.

[11] I.A. Khalil, S. Futaki, M. Niwa, Y. Baba, N. Kaji, H. Kamiya, H. Harashima, Mechanismof improved gene transfer by the N-terminal stearylation of octaarginine:

enhanced cellular association by hydrophobic core formation, Gene Ther. 11 (7)(2004) 636–644.

[12] T. Lehto, R. Abes, N. Oskolkov, J. Suhorutsenko, D.M. Copolovici, I. Mager, J.R. Viola,O.E. Simonson, K. Ezzat, P. Guterstam, E. Eriste, C.I. Smith, B. Lebleu, A. Samir El, U.Langel, Delivery of nucleic acids with a stearylated (RxR)4 peptide using a non-covalent co-incubation strategy, J. Control. Release 141 (1) (2009) 42–51.

[13] M. Mae, S. El Andaloussi, P. Lundin, N. Oskolkov, H.J. Johansson, P. Guterstam, U.Langel, A stearylated CPP for delivery of splice correcting oligonucleotides using anon-covalent co-incubation strategy, J. Control. Release 134 (3) (2009) 221–227.

[14] S. EL Andaloussi, T. Lehto, I. Mäger, K. Rosenthal-Aizman, II. Oprea ll, O.E.Simonson, H. Sork, K. Ezzat, D.M. Kopolovici, K. Kurrikoff, J.R. Viola, E.M. Zaghloul,R. Sillard, H.J. Johansson, F. Said Hassane, J. Suhorutsenko, P.M. Moreno, N.Oskolkov, J. Hälldin, U. Tedebark, A. Metspalu, B. Lebleu, J. Lehtiö, C.L. Smith, Ü.Langel, Design of a peptide-based vector, PepFect6, for efficient delivery of siRNAin cell culture and systemically in vivo. Nucl. Acids Res. (2011, Jan 17).

[15] I. Coin, M. Beyermann, M. Bienert, Solid-phase peptide synthesis: from standardprocedures to the synthesis of difficult sequences, Nat. Protoc. 2 (12) (2007)3247–3256.

[16] T. Kobayashi, Y. Hirabayashi, Lipidmembrane domains in cell surface and vacuolarsystems, Glycoconj. J. 17 (3–4) (2000) 163–171.

[17] S.T. Yang, E. Zaitseva, L.V. Chernomordik, K. Melikov, Cell-penetrating peptideinduces leaky fusion of liposomes containing late endosome-specific anionic lipid,Biophys. J. 99 (8) (2010) 2525–2533.

[18] S.H. Kang, M.J. Cho, R. Kole, Up-regulation of luciferase gene expression withantisense oligonucleotides: implications and applications in functional assaydevelopment, Biochemistry 37 (18) (1998) 6235–6239.

[19] S. Abes, H.M. Moulton, P. Clair, P. Prevot, D.S. Youngblood, R.P. Wu, P.L. Iversen, B.Lebleu, Vectorization of morpholino oligomers by the (R-Ahx-R)4 peptide allowsefficient splicing correction in the absence of endosomolytic agents, J. Control.Release 116 (3) (2006) 304–313.

[20] B. Lebleu, H.M. Moulton, R. Abes, G.D. Ivanova, S. Abes, D.A. Stein, P.L. Iversen, A.A.Arzumanov, M.J. Gait, Cell penetrating peptide conjugates of steric blockoligonucleotides, Adv. Drug Deliv. Rev. 60 (4–5) (2008) 517–529.

[21] E. Eiriksdottir, K. Konate, U. Langel, G. Divita, S. Deshayes, Secondary structure ofcell-penetrating peptides controls membrane interaction and insertion, Biochim.Biophys. Acta 1798 (6) (2010) 1119–1128.

[22] R. Abes, A.A. Arzumanov, H.M. Moulton, S. Abes, G.D. Ivanova, P.L. Iversen, M.J.Gait, B. Lebleu, Cell-penetrating-peptide-based delivery of oligonucleotides: anoverview, Biochem. Soc. Trans. 35 (Pt 4) (2007) 775–779.

[23] K. Padari, P. Saalik, M. Hansen, K. Koppel, R. Raid, U. Langel, M. Pooga, Celltransduction pathways of transportans, Bioconjug. Chem. 16 (6) (2005)1399–1410.

[24] P. Lundin, H. Johansson, P. Guterstam, T. Holm, M. Hansen, U. Langel, S.El Andaloussi, Distinct uptake routes of cell-penetrating peptide conjugates,Bioconjug. Chem. 19 (12) (2008) 2535–2542.

[25] D. Vercauteren, R.E. Vandenbroucke, A.T. Jones, J. Rejman, J. Demeester, S.C. DeSmedt, N.N. Sanders, K. Braeckmans, The use of inhibitors to study endocyticpathways of gene carriers: optimization and pitfalls, Mol. Ther. 18 (3) (2010)561–569.

[26] X. Zhang, Y. Jin, M.R. Plummer, S. Pooyan, S. Gunaseelan, P.J. Sinko, Endocytosisand membrane potential are required for HeLa cell uptake of R.I.-CKTat9, a retro-inverso Tat cell penetrating peptide, Mol. Pharm. 6 (3) (2009) 836–848.

[27] G.D. Ivanova, A. Arzumanov, R. Abes, H. Yin, M.J. Wood, B. Lebleu, M.J. Gait,Improved cell-penetrating peptide-PNA conjugates for splicing redirection inHeLa cells and exon skipping in mdx mouse muscle, Nucleic Acids Res. 36 (20)(2008) 6418–6428.

[28] D. Terrone, S.L. Sang, L. Roudaia, J.R. Silvius, Penetratin and related cell-penetrating cationic peptides can translocate across lipid bilayers in the presenceof a transbilayer potential, Biochemistry 42 (47) (2003) 13787–13799.

[29] G. Ter-Avetisyan, G. Tunnemann, D. Nowak, M. Nitschke, A. Herrmann, M. Drab,M.C. Cardoso, Cell entry of arginine-rich peptides is independent of endocytosis,J. Biol. Chem. 284 (6) (2009) 3370–3378.

[30] J.B. Rothbard, T.C. Jessop, R.S. Lewis, B.A. Murray, P.A. Wender, Role of membranepotential and hydrogen bonding in the mechanism of translocation of guanidi-nium-rich peptides into cells, J. Am. Chem. Soc. 126 (31) (2004) 9506–9507.