key: cord-0769470-eoi2wpod authors: Lebleu, Bernard; Moulton, Hong M.; Abes, Rachida; Ivanova, Gabriela D.; Abes, Said; Stein, David A.; Iversen, Patrick L.; Arzumanov, Andrey A.; Gait, Michael J. title: Cell penetrating peptide conjugates of steric block oligonucleotides() date: 2008-03-01 journal: Adv Drug Deliv Rev DOI: 10.1016/j.addr.2007.09.002 sha: 13acb6ab959fe6d2a349ff27900d41eb8726ac46 doc_id: 769470 cord_uid: eoi2wpod Charge neutral steric block oligonucleotide analogues, such as peptide nucleic acids (PNA) or phosphorodiamidate morpholino oligomers (PMO), have promising biological and pharmacological properties for antisense applications, such as for example in mRNA splicing redirection. However, cellular uptake of free oligomers is poor and the utility of conjugates of PNA or PMO to cell penetrating peptides (CPP), such as Tat or Penetratin, is limited by endosomal sequestration. Two new families of arginine-rich CPPs named (R-Ahx-R)(4) AhxB and R(6)Pen allow efficient nuclear delivery of splice correcting PNA and PMO at micromolar concentrations in the absence of endosomolytic agents. The in vivo efficacy of (R-Ahx-R)(4) AhxB PMO conjugates has been demonstrated in mouse models of Duchenne muscular dystrophy and in various viral infections. The concept of a synthetic oligonucleotide (ON) as a potential therapeutic agent was first demonstrated experimentally through oligodeoxyribonucleotide targeting of a Rous sarcoma virus RNA translation initiation site. The ON was used to form an RNA-DNA duplex that sterically blocked the RNA, resulting in inhibition of gene expression and consequently of viral replication [1] . Later it was found that a second inhibitory mechanism operates in cells. Recognition of the RNA:DNA duplex by the cellular enzyme RNase H results in subsequent RNA cleavage, and hence prevents gene expression. This second mechanism of action became the one primarily pursued and known generally as "antisense". This led to the industrial development of therapeutic DNA-based ON and their analogues for control of gene expression, but sadly so far to little clinical success [2, 3] . The original concept of inhibition of protein translation through steric block continued to be studied [4, 5] , and in recent years ON that are not substrates for RNase H when duplexed to RNA have begun to be developed as therapeutic agents [6] . In addition to inhibition of mRNA translation initiation, ON have also been designed to sterically inhibit critically important targets in RNA processing, such as for example nuclear splicing, a series of events that involve numerous RNA-protein interactions [7] . Further, other small RNAs involved in gene control have been targeted both in vitro and in vivo, such as endogenous microRNAs [8] . The effective target region of a steric block ON for inhibiting translation is generally limited to the 5'-UTR and the start codon region of a mRNA. One important advantage of the steric block approach of modulating gene expression is its greater specificity, and thus potentially lower off-target effects, compared to conventional antisense, since binding of an ON to a partially matched, inappropriate RNA sequence is unlikely to have biological consequences. A second advantage is the ability to use a considerably wider range of synthetic ON analogues than is possible with conventional antisense, since there is no requirement for molecular recognition by cellular RNase H. Instead, the only key initial necessities for a suitable steric block ON are tight binding to the RNA target as well as good resistance to nuclease degradation. The greater flexibility to manipulate ON chemistry makes it easier also to focus on other essential requirements for therapeutic development, such as good delivery to cells and tissues, low toxicity, and other pharmacological parameters. The first type of steric block ON to become widely established consisted of fully 2'-O-methylated (OMe) nucleosides [9] and also often contained phosphorothioate (PS) linkages [10] [11] [12] . PS linkages improve cellular uptake, metabolic stability and pharmacology, but also give rise to reduced RNA binding. Several other types of negatively charged, phosphatecontaining analogues have better binding to RNA and higher resistance to nuclease degradation, for example 2'-O-methoxyethyl (MOE) (with or without PS linkages) [5] and N3'-P5'phosphoramidates (NP) [13] . When bound to an RNA target, such analogue types adopt an A-like conformation, similar to that of an oligoribonucleotide. Locked nucleic acids (LNA) also adopt an A-like conformation and bind very tightly to RNA. For example, LNA ONs have been used for in vivo tumour growth inhibition [14] . LNA is generally used in combination with another nucleotide derivative, for example with 2'-deoxyribonucleotides [15] or with OMe residues [16] , to obtain optimally balanced binding and specificity characteristics. An alternating LNA/DNA 16-mer containing all PS linkages was found to give efficient splice switching in mice [17] . A luciferase reporter assay was used to show that OMe/LNA mixmers bound strongly to the transactivation responsive element (TAR) of HIV-1 viral RNA to block Tat-dependent trans-activation in HeLa cells [18, 19] . However, two types of uncharged ON analogues, peptide nucleic acids (PNA) [20] (Fig. 1 ) and phosphorodiamidate morpholino oligomers (PMO, also known as morpholino) [21] ( Fig. 2 ) have come to dominate steric block applications recently. Although departing significantly from the sugar-phosphate backbone found in regular DNA, oligomers of both types retain very strong and sequence-specific RNA binding characteristics [22, 23] . For example, PNAs targeting the 5'-UTR of luciferase mRNA were shown to inhibit translation of protein synthesis [24] . PMOs also have been found to be remarkable steric block ONs for inhibiting translation [25] , altering pre-mRNA [26] and blocking miRNA activity [27] , as demonstrated in embryos, cells and animals. Now PMOs have been taken to pre-clinical studies for treatment of cardiovascular diseases, viral diseases and genetic disorders, such as Duchenne muscular dystrophy (DMD) (see Section 5) . Despite being charge neutral, PNA and PMO do not enter cells in culture any more readily than do negatively charged ON. For example, many methods of PNA delivery have been devised, such as electroporation, microinjection, transfection in complex with a DNA ON, lipofection of acridine or other polyheteroaromate functionalized PNA, and photochemicallyinduced delivery. Cell transfection efficiencies are highly variable and sensitive to small variations in conditions and the type of transfection method used, with the best achievable IC 50 values usually around 200 nM in a model splice redirection system [28] . The relatively high molar levels of these oligomers required to obtain significant steric block action, compared to ON with an RNase H-dependent mechanism of action or short interfering RNA (siRNA), are probably due to the need to deliver at least a stoichiometric amount of the ON to the RNA target into the right cell compartment to obtain the desired biological effect. One idea to enhance cell delivery was first proposed 20 years ago, the covalent conjugation of ON to cationic peptides [29] . Later, conjugates of an antisense ON with a peptide derived from helix 3 of the Antennapaedia homeodomain protein from Drosophila melanogaster (residues 43-58, Penetratin) were shown to have biological activity in neuronal cells, which was attributed to the cell penetrating properties of the peptide [30] [31] [32] . Another cell penetrating peptide (CPP) was discovered in the basic region of HIV-1 Tat protein (residues 48-60) [33] . Many CPPs (also known as "protein transduction domains") with remarkable properties of cell penetration either on their own, or as conjugates with a range of other types of biomolecules, such as peptides, proteins, liposomes and nanoparticles have now been characterised (reviewed in [34] ). However, very few examples of strong delivery and biological activity of negatively charged ON or siRNA cargoes into cells through covalent attachment of a CPP have been published (reviewed in [35] [36] [37] ), and in many cases, the conjugation of a peptide does not allow sufficient cargo delivery through the endocytotic pathway and into the cell compartments necessary for biological activity. By contrast, CPP conjugates have shown significant promise towards improving cell delivery of charge neutral PNA and PMO. For PNA-peptides, an early example was the blocking of expression of the galanin receptor mRNA in human Bowes cells by a 21-mer PNA disulfide-coupled to the CPPs Penetratin or Transportan (a hybrid of a section of the neuropeptide galanin and the wasp venom peptide mastoparan), which constituted a step towards the modification of pain transmission [38] . Remarkably potent inhibition (down to 70-80 nM EC 50 ) in the same model was claimed with shorter PNAs targeting different regions of the mRNA and conjugated to a truncated Transportan version TP10 [39] . This is exceptional, since most cellular inhibition of protein translation by PNA has been achieved only in the μM range. For example a 13-mer PNA disulfide-linked to Penetratin was able to inhibit telomerase activity in JR8 melanoma cells with IC 50 ∼ 7 μM [40] . In a different model system, an amphipathic peptide, MAP, disulfide-conjugated to a PNA complementary to the nociceptin/orphanin FQ receptor mRNA was shown to impart improved cell uptake and steric block biological effects in both CHO cells and neonatal rat cardiomyocytes again in the μM range [41] . When linked to a D (AAKK) 4 peptide, a 19-mer PNA complementary to human caveolin 1 mRNA inhibited expression in HeLa cells (IC 50 = 2 μM) and in primary endothelial cells at concentrations N 15 μM [42] . CPP-PMO conjugates have been used successfully in cell culture and animal models to inhibit the replication of several viruses that are harmful to humans, including for example Coxsackievirus B3, SARS coronavirus, influenza, dengue, West Nile and Ebola viruses (see Section 6) [43] [44] [45] [46] [47] [48] [49] . In addition to the applications of CPP-PNA and CPP-PMO in translational blocking mentioned above, PMOs and PNAs delivered by CPP have been used to alter pre-mRNA splicing. 3-10 μM of a PNA 18-mer with 4 added Lys residues (PNA-4 K) complementary to a splice site (654) within an enhanced green fluorescent protein (EGFP) reporter gene containing an aberrant human β-globin intron was enough to result in some splice correction and up-regulation of EGFP in a HeLa cell culture assay [12] . In transgenic mice, PNA-4 K 654 showed activity in several tissues (heart, kidney lung, liver, muscles and small intestine) when delivered intraperitoneally at 50 mg/kg for up to 4 days, with up to 40% splicing correction observed [50] . CPP-PMO conjugates have also been used effectively to alter pre-mRNA splicing in DMD models. DMD is usually caused by nonsense or frame-shifting exon mutations in the dystrophin gene that result in an absence of functional protein. (R-Ahx-R) 4 AhxB-PMO (R = arginine, Ahx = 6-aminohexanoic acid and B = betaalanine) conjugates have been used in several models to promote skipping of the mutated exon to produce internally shortened but functional dystrophin protein. An (R-Ahx-R) 4 AhxB-PMO conjugate was much more effective to cause desired exonskipping in primary muscle cells isolated from a golden retriever muscular dystrophy dog than the corresponding non-conjugated PMO or OMe ON. A 300 nM extracellular concentration of the conjugate caused complete exon-skipping in the muscle cells. Ten days after a single treatment, conjugate-treated cells continued to show abundant exon-skipped RNA product [51] . Another (R-Ahx-R) 4 -PMO conjugate caused desired exonskipping in both mdx mouse muscle explants [52] and in several muscles of the intact mdx mice [53] (see Section 5) . (R-Ahx-R) 4 AhxB-PMO altered pre-mRNA splicing in primary murine splenic CD4 + T-cells. The PMOs were targeted to splice acceptor sites of the pre-mRNAs of the cell surface signalling protein OX-40 and the nuclear transcription factor Foxp3. Treatment of the T-cells with 4 μM of the conjugates resulted in marked reduction of OX-40 and Foxp3 protein expressions up to 24 h after removing the conjugates, and detection of the splice-altered OX-40 and Foxp3 gene products 48 h after removing the conjugates [54] . Several studies have used a model cell culture system involving splicing redirection of murine CD40 pre-mRNA to cause a down-regulation of activity. A PNA complementary to the splice site with 8 Lys residues on the N-terminus was significantly more active in cell culture than a similar PNA with 4 Lys residues, but still 3-10 μM concentrations were needed for strong down-regulation of this target in BCL 1 cells or primary murine macrophages [55] . A variety of stably linked PNA conjugates with variously spaced oligo-Lys peptides showed some small enhancements of splicing redirection activity in some cases compared to Lys 8 [56] , and a series of 28 cationic peptides based on Lys or Arg that had amphipathic character showed even stronger splicing redirection activity at 1-6 μM concentrations, and without toxicity [57] , but in vivo efficacies for these constructs have not yet been reported. A particularly useful splice redirection model system, developed by Kole and colleagues is based on the use of a stably integrated plasmid containing a gene coding for firefly luciferase interrupted by an aberrant β-globin intron with an inframe stop codon [58] . Up-regulation of luciferase resulted when an 18-mer PS/OMe oligonucleotide complementary to the aberrant 705 splice site is introduced into the HeLa cells to redirect splicing [12, 50] . Use of this sensitive splicing redirection model allows optimization of CPP conjugate technology for delivery of PNA or PMO into the nuclear compartment. Since in our experience the conjugation of CPP to negatively charged ON did not result in a level of delivery into cells sufficient for biologically activity [59] , we concentrate here on PNA-and PMO-peptide conjugates which have been much more successful. General reviews describing methods of synthesis of peptide-ON conjugates have been published [60, 61] . Peptides and PNA are synthesized through similar sequential coupling reactions of monomer units on a solid support to form amide bonds. Strategies for manual or automated PNA synthesis by 9-fluorenylmethoxycarbonyl/benzhydryloxycarbonyl (Fmoc/Bhoc) or tert-butyloxycarbonyl/carbobenzyloxy (Boc/ Cbz) routes are well established [62] [63] [64] . Thus peptide-PNA conjugates may be assembled by continuous synthesis with activated amino acids or activated PNA monomers being added in the coupling reactions. In general there are no restrictions in combining amino acids with PNA residues and these may be added in any order. It is common to include at least one Lys residue on the end of a PNA synthesis to aid water solubility. Fmoc/Bhoc chemistry is more often preferred for PNA assembly and we have reported protocols for PNA and PNApeptide syntheses using a robotic Apex 396 Synthesizer [65] . A short polyether spacer (AEEA, 8-amino-3,5-dioxo-octanoic acid, also known as an O-linker) is often added between the peptide and PNA moieties in continuous synthesis (Fig. 1a ) [66, 67] . Alternatively, fragment conjugation of pre-synthesized peptide and PNA components in solution is usually more convenient for structure-activity (SAR) studies and the overall yields and purities of the final conjugate are generally higher. Each fragment is synthesized separately on solid-phase and contains a specific functionality on one end. Each fragment is purified prior to the conjugation step. Usually, an excess of one component is added over the other to drive the conjugation reaction to completion and the product conjugate is purified by HPLC and characterized by mass spectrometry. No general rules have yet emerged as to optimal linkage types, since the factors affecting biological activity are often complex. For example, in some studies, use of labile disulfide linkages has resulted in higher biological activities [67, 68] . By contrast, a stable thioether linkage or a disulfide linkage gave equally good results in a splicing redirection system [69] . Several types of stable linkage have been reported. For example, a thiol-maleimide linkage requires addition of a Cys residue to the CPP and preparation of a N-maleimide derivative of the PNA, which may be obtained by reaction of the terminal primary amino group with the hetero-bifunctional cross-linker succinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate ( Fig. 1b) [70] . A thioether linkage may be formed between a bromoacetyl ON and a Cys-containing peptide [71] and we have adapted this reaction for conjugation of a CPP containing a Cys residue at the C-terminus with N-terminally bromoacetylated PNA (Fig. 1c ) [69] . An ester-containing conjugate is synthesized with a PNA bearing a glycyl(hydroxymethyl)benzoic acid residue at the Nterminus, and this linkage was designed to be labile to esterases within cells (Fig. 1d ) [68] . However, the most popular labile linkage is a disulfide bond between the peptide and the CPP (Fig. 1e) , which was proposed to be cleavable within the reducing environment of the cell, and which was first used for studies of Transportan-PNA and Penetratin-PNA conjugates [38] . Both CPP and PNA are functionalized with a Cys residue, and the disulfide bond is formed between the Cys thiol groups. One of the thiols is activated by reaction with either a pyridylsulfenyl or a 3-nitropyridylsulfenyl group. We have used the disulfide linkage predominantly in studies of HIV Tatdependent trans-activation inhibition and in splicing redirection [67, 69] . The disulfide linkage is convenient for in vitro and cell studies, even though it may be less suitable for in vivo use. PMO syntheses have been described previously [21, 72] , while the CPPs may be synthesized by standard Fmoc chemistry. The conjugation of CPPs and PMOs through a thioether (maleimide), disulfide or amide linker have been described previously [73] [74] [75] and the structures of these conjugates are shown in Fig. 2 . The effect of linker type on the activity [74, 75] and stability [76] of CPP-PMO conjugates has been determined. The nuclear antisense activities of the CPP-PMOs with the three linkage types were similar ( [74] and Moulton, H. M, unpublished results). Synthetically, the amide linkage has two advantages over disulfide and thioether linkages. First a PMO can easily be conjugated to a CPP via an amine at either end of the oligomer. Second, formation of an amide bond requires only a one-step synthetic and purification scheme as compared to a two-step strategy for other linkages. Therefore, the amide linkage leads to an overall greater yield of CPP-PMO. In addition, amide bonds appear to be more stable in vivo, since an unhindered disulfide-linked CPP-PMO was found to be unstable in human serum [76] . Although a CPP can be conjugated to either end of a PMO, the position of the CPP affects the activity of the attached PMO. A PMO having a CPP conjugated at the 5'-end was more functionally active than that with the CPP attached at the 3'-end [74] . Adding a bulky fluorochrome such as carboxyfluorescein at the 3'-end of PMO decreased the activity of the CPP-PMO by about 30% (unpublished data). As pointed out in Section 1, at variance with numerous reported successes for peptide payloads [77] , successful delivery of antisense ON by CPPs has not often been documented. Nielsen and colleagues [78] investigated the cellular uptake of several PNA-CPP constructs in live cells to avoid artefacts of redistribution caused by cell fixation. The PNA-Tat and PNA-Penetratin constructs were taken up by an energy-dependent mechanism and essentially accumulated in endocytotic vesicles but not in nuclei, in keeping with our own data [79] . Although not encouraging at first sight, these data could not predict whether biological effects would be achievable or not. For example, even though fusion constructs between Cre recombinase and Tat CPP seemed to remain localized in endocytotic vesicles, they gave rise to significant biological activity [80] . As mentioned in Section 1, many groups in the field of ON delivery are now relying on the splicing redirection model of Kole [58] to assess the nuclear delivery of steric block ON. This model is useful by virtue of being highly sensitive and sequence-specific, by providing a positive read-out over a low background level, and in giving rise to a linear response over a large range of concentrations. Luciferase expression is however highly sensitive to cell density and other variables as pointed by Bendifallah et al. [68] . RT-PCR evaluation of splicing correction is easy to implement, gives an estimation of the proportion of mis-versus correctly-spliced luciferase mRNA, and it may therefore be preferred to luciferase assays. Tat-and Arg-PMO constructs gave rise to a sequencespecific correction of mis-spliced luciferase pre-mRNA. However high concentrations (5 μM and higher) of the correcting steric block ON were required to achieve luciferase expression [73, 75] . As a point of comparison, OMe and MOE ON rapidly accumulated in nuclei and led to splice correction at doses as low as 10 nM when transfected with cationic lipids [81] . Likewise, Tat-, Arg-and Penetratin-PNA constructs were poorly efficient in the splice redirection assay at low concentrations [69] . A series of publications reported successful splicing correction with PNA appended to short oligolysine chains, as reviewed in Section 1. Intriguingly, splicing correction was claimed to be energy-independent [12] , at variance with most recent data on the mechanism of cellular uptake of basic amino acid-rich CPP-conjugated ON [82] . In our hands, (Lys) 8 -PNA conjugates were efficiently taken up in cells as monitored by FACS analysis and by fluorescence microscopy, but their biological activity remained rather low, in keeping with our previous data with Tat-or Arg-rich steric block ON conjugates [83] . Interestingly, cellular uptake was clearly energy-dependent at low concentration, but became temperature-independent at high concentration, in parallel with membrane destabilization as monitored by propidium iodide uptake [83] . Endocytosis may be thus the predominant mechanism of uptake at low concentrations, while direct translocation linked to CPP-induced membrane destabilization could take place at higher concentrations. CPP-induced membrane permeabilization may itself be rather dependent on peptide, cell type and experimental conditions, providing a possible explanation for the ongoing controversy concerning CPP mechanism of cell uptake. Sequestration of internalized material in endocytotic vesicles could also explain the observed low biological activity of these various CPP conjugates. Indeed, treatment with endosomolytic agents, such as chloroquine or high sucrose, gave rise to a large increase in splice correcting activity even when added after CPP-PNA internalization [83] . Similarly, in a HeLa cell assay containing integrated luciferase reporter plasmids, cationic lipid-transfected OMe/LNA ONs targeting the HIV-1 TAR element were able to dosedependently inhibit HIV-1 Tat-dependent trans-activation [18] . However, most CPP conjugates (such as Tat and Pen) of PNA targeted to TAR when incubated with the HeLa cells were unable to inhibit trans-activation unless chloroquine was added, and even the few CPP-PNA that did show some inhibition required 24 h incubation to show activity [67] . Endosomal sequestration appeared to be the explanation for these results. Alternative endosome disrupting strategies have been evaluated using the splicing correction assay. Addition of Ca 2+ at high concentration (6 mM) to the culture medium strongly increased splicing correction by Tat-or (Arg) 6 PNA constructs [84] . Evidence for a partial redistribution from endocytotic vesicles to cytoplasm was obtained but no nuclear staining was observed. Likewise chloroquine treatment improved functional activity of CPP-PNA or CPP-PMO conjugates, but we were unable to detect nuclear accumulation of conjugate or oligomer (Abes et al., unpublished observations). In any event, none of the above endosome-disrupting strategies is likely to be feasible in vivo. Therefore several groups have explored alternative strategies. Co-treatment with endosome-disruptive peptides has been studied extensively to improve plasmid DNA transfection with non-viral delivery vectors. Both synthetic (mostly amphipathic) and natural (mostly viral) endosome-disruptive peptides have been described, the most popular being the N-terminal peptide of Influenza virus HA2 protein ( [85] and references therein). A HA2-Pen fusion construct only slightly increased splicing correction by Tat-, Pen-or Transportan-PNA conjugates [86] . Likewise, the addition of several natural and synthetic membrane destabilizing peptides to Tat-PNA had only a modest effect on splicing correction at non-cytotoxic doses [87] . Photochemical internalization (PCI) capitalizes on the production of reactive oxygen species, which damage endosomes upon irradiation, thus facilitating escape of endosomeentrapped drugs [88] . PCI has been exploited successfully to increase the uptake of various biomolecules from low molecular weight drugs to plasmid DNA. Bøe and Hovig [89] have documented PCI-induced gene silencing with several PNA-CPP constructs together with re-localization of the PNA-peptides from endosomes to nuclei. Shiraishi et al. provided similar data with PNA-CPP constructs aimed at splicing correction [90] . PCI thus represents an interesting strategy to overcome endosomal entrapment of CPP-ON, since it is amenable to in vivo development [91] . Structure-activity studies of CPPs derived from proteins (Tat, Penetratin) and synthetic CPPs have emphasized the crucial role played by basic amino acids and particularly by the guanidinium side-chain of arginine. Such basic amino acids are responsible for binding to cell surface heparan sulfate proteoglycans (HSPG) through electrostatic interactions and for subsequent cell internalization. The RXR family of synthetic CPPs was first proposed by J. Rothbard, P.A. Wender and their colleagues on the assumption that not all Arg side chains in an oligo Arg peptide would point in the same direction and would therefore be available for HSPG (or membrane lipid polar heads) binding. They accordingly designed and synthesized a large collection of (RXR) n R, (RX) n R, (RXX) n R and (RXXX) n R transport peptides in which X represents various natural or non-natural amino acids [92] . All peptides were conjugated to fluorescein and assayed for cellular uptake in a Jurkat T cell line. These studies emphasized the crucial role of Arg side chain spacing for cellular uptake, and the most efficient peptide was (R-Ahx-R) 4 R in which Ahx represents a 6-atom aminohexanoic acid spacer. This strategy of interspersing Arg residues with non-natural, uncharged amino acids also appeared interesting in decreasing the number of natural peptide bonds (thus providing increased metabolic stability) and in decreasing the positive charge density often responsible for cytotoxicity. However, no studies of biological responses were reported, although the synthesis of conjugates to cyclosporin or acyclovir has been described [93] . (R-Ahx-R) 4 was chosen therefore as a lead peptide and has been stably conjugated to a PMO oligomer via a Ahx-β-Ala linker. At variance with Tat-PMO and Arg 6 -PMO conjugates, (R-Ahx-R) 4 AhxB-PMO leads to a dose-dependent and specific increase of luciferase expression in the absence of endosomolytic agents. Importantly a significant increase could already be observed at sub-micromolar concentration [75] . As mentioned above, measurement of luciferase expression allows the functional comparison of various CPP constructs but does not allow the quantification of splicing redirection. We therefore made use of RT-PCR which easily discriminates correctly from aberrantly-spliced luciferase transcripts and confirmed that highly significant splicing correction could be achieved at μM concentrations of (R-Ahx-R) 4 AhxB-PMO. (R-Ahx-R) 4 has also been stably conjugated to a splice correcting PNA oligomer and found to be effective in the same concentration range [87] . Interestingly for in vivo applications, the (R-Ahx-R) 4 constructs did not show membrane permeabilization at doses up to 50 μM (i.e. concentrations at which Tat-, (Lys) n -or (Arg) n -constructs lead to significant propidium iodide uptake). We then compared the mechanism of cellular internalization of fluorochrome-labelled (R-Ahx-R) 4 AhxB-PMO, Tat-PMO and (Arg) n -PMO in a search for a mechanistic explanation for their different efficiencies. Surprisingly, the (R-Ahx-R) 4 construct was taken up less efficiently than Tat-or (Arg) n ones [75] , at variance with the data reported by Rothbard et al. [93] . An increased activity could be explained by an non-endocytotic uptake route, but studies involving energy deprivation or use of endocytosis inhibitors did not support this hypothesis [75] and further studies are needed to clarify this issue. Along the same lines, fluorescence microscopy studies in live cells revealed that the majority of (R-Ahx-R) 4 AhxB-PMO distributes as dotted cytoplasmic spots, characteristic of accumulation in endocytotic vesicles [75] . Similar conclusions are reached when using a functional assay such as splicing redirection to assess the energy-dependence of the (R-Ahx-R) 4 AhxB-PMO conjugates (Fig. 3) . As expected, (R-Ahx-R) 4 AhxB-PMO binds to cell surface HSPG as revealed by both genetic and biochemical tools [72] . Of possible interest, (R-Ahx-R) 4 constructs bind to Sepharoseimmobilized heparin (taken as a model heparan sulfate) with a 4 ·AhxB-PMO conjugates are internalized by an energydependent mechanism. A splicing correction assay was used to assess the energy-dependence by ATP deprivation (lanes 5-8) or by low temperature incubation (lanes 9 to 12). HeLa pLuc 705 cells were pre-incubated for 1 h at 37°C in OptiMEM, supplemented with 10 mM sodium azide and 6 mM 2deoxy-D-glucose or at 4°C in OptiMEM. Cells were then incubated with the PMO conjugates at the indicated concentrations for 4 h in the same conditions. Cells were then washed and incubation was continued at 37°C for 20 h in DMEM. Total RNA was extracted and amplified by RT-PCR. PCR products from incorrectly (268 bp band) and correctly (142 bp band) spliced luciferase pre-mRNA were separated on a 2% agarose gel. Lanes 1, 5 and 9: controls in the absence of (R-Ahx-R) 4 AhxB-PMO conjugate, 2-4: cells incubated at 37°C and treated with 0.5, 1 or 2.5 μM (R-Ahx-R) 4 AhxB-PMO, respectively, 6-8: cells incubated at 37°C in the presence of 10 mM sodium azide and 6 mM 2deoxy-D-glucose and treated with 0.5, 1 or 2.5 μM (R-Ahx-R) 4 AhxB-PMO, respectively, and 9-12: cells incubated at 4°C and treated with 0.5, 1 or 2.5 μM (R-Ahx-R) 4 AhxB-PMO, respectively. significantly lower affinity than Tat-or (Arg) n -constructs [72] . This could lead to improved efficiency, if one postulates that HSPG-bound ligands need to dissociate after cellular internalization in order to escape from endocytotic vesicles. SAR studies with a series of (RXR) 4 AhxB-PMO constructs in which X varies in terms of length, affinity for heparin or hydrophobicity have been conducted (Abes et al., in preparation) . These studies demonstrate that increased hydrophobicity or affinity for heparin improves cellular uptake, but is detrimental to the efficiency of splicing correction. As briefly mentioned above, (R-Ahx-R) 4 has a lower number of adjacent natural amino acids than other CPPs, a factor which could increase its metabolic stability and consequently its biological efficiency. Although serum stability was indeed increased, analysis of intracellular cleavage products did not support this hypothesis [94] . Furthermore, the introduction of D-Arg residues to give (DR-Ahx-R) 4 -R-PMO, for the purpose of increasing resistance to proteolysis, did not lead to increased splicing correction efficiency, thus indicating that metabolic stability did not limit functional efficacy (Abes et al., in preparation) . Altogether, no definitive explanation for the superior activity of (R-Ahx-R) 4 over other Arg-rich CPPs can be provided at present. Further SAR and endosomal escape studies will hopefully shed light on this mystery and will contribute to the design of more active steric block ON delivery vectors. We note recent results from use of fluorescent microscopy and the splice correction assay, which show that a large percentage of the (R-Ahx-R) 4 AhxB-PMO that enters cells appears to localize in endocytotic vesicles, even at doses that produce nearly complete splicing correction. These results are not incompatible, since steric block ON conjugates were applied at μM concentrations, while the mis-spliced pre-mRNA target is expressed at a far lower molar concentration. A second class of Arg-rich CPPs that we have been studying, known as R 6 Pen, is based on the well known Penetratin peptide [95] to which six Arg residues are added at the N-terminus. First synthesized for conjugation to an OMe/LNA oligonucleotide [59] , R 6 Pen was disulfide-conjugated to a 16-mer PNA complementary to the HIV-1 trans-activation responsive element TAR and was found to enter the HeLa cell nuclei and to inhibit Tat-dependent trans-activation in a luciferase plasmid assay system at sub-micromolar concentrations both dose-dependently and PNA sequence-dependently [67] . Activity could be achieved more readily when the R 6 Pen-PNA was incubated with HeLa cells in the presence of the endosomolytic reagent chloroquine, suggesting an endocytotic mode of cell uptake [67] . Similarly to (R-Ahx-R) 4 -PNA or PMO conjugates, an R 6 Pen-PNA conjugate targeted to the 705 splice acceptor site showed potent and sequence-specific correction in the HeLa 705 cell assay at low concentrations in the absence of endosomolytic agents [69] . The mechanism of cell internalization of R 6 Pen-PNA705 has not yet been thoroughly investigated, but it is clearly energy-dependent and likely to involve endocytosis (Abes et al., unpublished results) . When employing the splicing redirection assay using RT-PCR to measure the levels of aberrantly and correctly-spliced RNA transcript, R 6 Pen is the most active of all CPP-PNA conjugates investigated so far by our group (Fig. 4a) . It is not yet clear why R 6 Pen is more efficient than other CPPs, such as Penetratin. For example, an essential difference has been documented using a mutant of the R 6 Pen peptide in which a Trp residue known to be essential for Pen peptide to penetrate membranes has been replaced by a Leu residue. This R 6 Pen (W → L) mutant is even more active than the wild-type R 6 Pen version (Fig. 4a) [66] . This suggests that the mechanism of cell entry and release into cytosol and nucleus must differ significantly between the R 6 Pen-PNA conjugate and the Pen peptide. Further, the importance of the appended oligo-Arg tail is affirmed by varying the number of Arg residues. Neither shorter nor longer tails are as efficient as the original R 6 Pen peptide [66] . Further SAR studies are underway aimed at understanding better the roles of various domains of the R 6 Pen peptide, at enhancing conjugate activity and serum stability, and at decreasing cytotoxicity. An often-debated issue is the nature of the linker between the CPP delivery vector and the transported payload, as mentioned in Section 1. The role of a delivery vector is obviously to transport a cargo to its intracellular site. Intracellular dissociation between the two entities would intuitively be preferred, since it would be predicted to avoid steric hindrance or other potential complications. However, the situation is more complex for steric block ON analogues, since a bound, highly charged delivery peptide should increase affinity for the targeted RNA sequence and favour invasion of secondary structures, thus leading to enhanced activity [96] . We thus compared splicing correction efficiency and the specificity of PNA sequences conjugated to R 6 Pen through a stable (thioether) or a reducible (disulfide) linker. The disulfide-linked material is only slightly more efficient than the stably linked one (Fig. 4b) . All data gathered so far with stably linked (R-Ahx-R) 4 or R 6 Pen conjugates indicate that the appended CPPs do not prevent target recognition and do not alter oligomer sequence specificity. Alternatively, the similar activity of the stably-and disulfide-linked material could be due to equal rates of intracellular degradation of the peptidic delivery vector of both compounds. In vivo data on splicing modulation, bioavailability and toxicity DMD is an X-linked recessive disease causing muscle degeneration and decreased span and quality of life in affected humans. A mouse model of muscular dystrophy (mdx) has been used to study the effectiveness of free PMO and CPP-PMO conjugate. Mdx mice carry a nonsense mutation in exon 23. Excision of exon 23 in mdx mouse cells can restore mostlyfunctional dystrophin expression [26] . Unconjugated PMO targeted to a donor site of exon 23 partially restored dystrophin function in skeletal muscles of mdx mice at 100 mg/kg/dose intravenously [26] or at 25 mg/kg/dose intra-peritoneally [97] , both with 7 consecutive weekly doses. The PMO conjugated to (R-Ahx-R) 4 AhxB CPP produced high efficacy when administered intra-peritoneally at 5 mg/kg /dose once a week for 4 weeks [53] . The dystrophin level in diaphragm muscles of mdx mice, detected two weeks after a single intra-peritoneal injection of (R-Ahx-R) 4 AhxB-PMO, was similar to the levels in normal mice. Four doses of the conjugate at 5 mg/kg/dose, administered once a week for four weeks to neonatal mice, resulted in normal levels of dystrophin in the diaphragms of 6-week-old mice. At 22 weeks post the fourth injection, dystrophin was still detectable in the diaphragm. The conjugate effect was lower in the tibialis anterior, gluteus maximus, and triceps brachialis, weaker still in the colon and stomach, and absent in the heart. Western blots detected normal levels of dystrophin in the diaphragm at six weeks, lower levels in the tibialis and gluteus, but no dystrophin was detected in cardiac muscle. Animals of one day, four weeks or one year of age were treated to test the effect of age. Whilst splice-modified dystrophin expression increased for all ages treated, muscle architecture was most improved in mice treated at younger age. Normalized pathology and central nucleation were observed for the sections taken from the 6-weekold mice treated with (R-Ahx-R) 4 AhxB-PMO. The study revealed that treatment of DMD mice at a young age is more beneficial than at an older age [53] . Tissue distribution and pharmacokinetics of a PMO and its corresponding (R-Ahx-R) 4 AhxB conjugate, administrated intravenously, have been determined in rats [98] . Conjugation of a PMO to the (R-Ahx-R) 4 AhxB peptide increased the tissue uptake and retention [98] . The CPP conjugation increased uptake of PMO in all organs assayed, except in brain, with greater increases being seen in liver, spleen and lungs. Remarkably, a similar amount of PMO was still detected in the liver 5 days after the last dose, indicating that the conjugated PMO had a long retention time in this organ. Interestingly, the concentration of conjugated PMO in rat spleens assayed 5 days after the last dose was twice the concentration in rat spleens assayed 1 day after the last dose. These results suggest that efflux of the conjugated PMO from tissues to the vascular space is slow, which is an important consideration for its therapeutic applications. Conjugation of PMO to (R-Ahx-R) 4 AhxB improved the kinetic behaviour of PMO as demonstrated by a 4-fold increased in volume of distribution, 2-fold increase in the estimated elimination half life, and an increased area under the plasma concentration versus time curve. The stabilities of 12 CPP-PMO conjugates varying in CPP sequences, amino acid compositions and/or linkers were studied in cells and in human serum. The stability of the peptide moiety in serum ranked in the order D-CPPs = (R-Ahx) 8 4 AhxB N R 9 F 2 C = Tat. Stabilities of peptide moieties within cells were found to be D-CPPs N (RB) 8 B N (R-Ahx) 8 B = (R-Ahx-R) 4 AhxB = R 9 F 2 C = Tat. Stabilities of linkers in serum or in cells were found to be amide = maleimide N disulfide [76] . The stability of a (R-Ahx-R) 4 AhxB-PMO was determined in rat serum and organ lysates. The CPP portion of the conjugate exhibited time-and tissue-dependent degradation, with biological stability ranked in the order of liver Nheart = kidney N plasma. Up to 6 h, the CPP portion of CPP-PMO was partially degraded in the plasma but had no apparent degradation in tissues [98] . The PMO portion of the conjugates was completely stable in cells, serum, plasma and tissues [76, 98] . The toxicity of (R-AhxR) 4 AhxB-PMOs is caused by (R-Ahx-R) 4 AhxB while the PMO portions of the conjugates are essentially non-toxic [48, 98] . The degree of toxicity depends on the dose, dose frequency and route of administration. Mice tolerated (R-Ahx-R) 4 AhxB-PMOs well with repeated doses at ≤10 mg/kg [43, 48, 53, 98, 99] . Four daily i.p. injections into mice at 3.7 mg/kg of an (R-Ahx-R) 4 AhxB-PMO conjugate did not produce any observable toxicity [99] . No changes in mice behaviour, weight and serum chemistry were observed for the mdx mice injected intraperitoneally with the conjugate at 5 mg/kg once a week for four weeks [53] . No histopathological abnormalities were detected in heart, liver, kidney, spleen and pancreas after two i.v. doses of an (R-Ahx-R) 4 AhxB-PMO conjugate administered 48 h apart at 10 and 15 mg/kg [43] . However, at higher doses and dosing frequency, animals experienced weight loss. Mice lost weight after five daily doses (i.p.) at 15 mg/kg of an (R-Ahx-R) 4 AhxB-PMO conjugate despite maintaining their normal organ weights and appearances. Fibrosis was found in the livers of treated animals. These mice regained their normal weights after the treatment was stopped [48] . Five daily intravenous injections of an (R-Ahx-R) 4 AhxB-PMO to rats at 15 mg/kg did not cause any detectable toxicity, as measured by behaviour, appearance, eating/drinking habits, body weight and serum chemistry. Rats treated with a single 150 mg/kg dose appeared lethargic immediately after the injection and proceeded to lose weight, accompanied by elevated serum BUN and creatinine, indicating that kidney function was most likely affected. The LD 50 of an (R-Ahx-R) 4 AhxB-PMO in rats was around 220-250 mg/kg [98] . The ability of CPPs to deliver steric block PNA and PMO for treatment of viral and genetic diseases has been demonstrated in a number of publications both in cell culture and in vivo. Very few studies have been carried out so far in vivo with CPP-PNA conjugates. For example, PNA-Lys 4 showed some activity in several tissues (heart, kidney lung, liver, muscles and small intestine) when delivered i.p. at 50 mg/kg for up to 4 days in transgenic mice with the EGFP 654 splice correction model without showing acute toxicity [50] . An 18-mer PNA complementary to the enhancer Eμ intronic sequence and stably linked to a 7-mer NLS peptide inhibited c-myc oncogene expression in Burkitt's Lymphoma (BL) cells [100] and showed very little toxicity in a human tumour SCID mice model [101] . When inoculated subcutaneously, the NLS-PNA showed inhibition of BL cell-induced tumour growth in SCID mice [102] . A PNA conjugated to K 8 showed broad tissue distribution in mice after i.v. injection and a relatively slow excretion level [56] , whereas two Lys/Arg-rich amphipathic D-peptides stably conjugated to a PNA distributed substantially differently in mice [57] . By contrast CPP-PMO conjugates have been taken into numerous pre-clinical studies. For example, CPPs have been shown to be powerful vectors to deliver PMOs into the cytosol of cells. There have been a number of reports of CPP-PMO inhibition of virus replication in cell cultures and in mouse models. Coxsackievirus B3 is a picornavirus and a significant cause of myocarditis, especially in children and adolescents. An (R-Ahx-R) 4 AhxB-PMO targeting the Internal Ribosome Entry Site (IRES) in the 5' UTR of the Coxsackievirus B3 viral genome reduced viral titer by 4 logs in virus infected cells in culture. Two 10 mg/kg doses of the IRES-targeted compound, administered intravenously once 3 h before and once 2 days after infection with virus, reduced heart tissue damage and viral titer by about 100-fold, when measured at one week postinfection [43] . The recently emerged severe acute respiratory syndrome (SARS) coronavirus raised the threat of a worldwide pandemic in 2002-2003. Experiments in cultured cells showed that antisense R 9 F 2 C-PMO dramatically reduced virus-induced cytopathic effects, viral titers, and cell to cell spread of the virus [44] . Recently, in mouse models with a variety of murine coronaviruses, an (R-Ahx-R) 4 AhxB-PMO, administrated intraperitoneally, reduced liver damage and viral titers, reflecting the improved survival observed [99] . Influenza A virus (FLUAV) causes a considerable impact on human health every year, and also poses a pandemic threat. A panel of CPP-conjugated PMO targeted against various locations in FLUAV gene segments involved in viral RNA synthesis showed that a PB1-and/or a NP-directed compound were found to markedly inhibit the replication of all viral strains tested (including H1N1, H3N2, H7N7, H3N8 and H5N1). Both effective compounds target a sequence that is highly conserved across FLUAV strains and subtypes [45] . Dengue virus (DENV) causes 50-100 million cases of dengue fever yearly, with about 300,000 requiring hospitalization, and 15,000 deaths. About 40% of the world's population is currently exposed to potential infection. Steric block R 5 F 2 R 4 C-PMO specifically inhibited DENV replication by over 10,000-fold in cell culture experiments [46] . Similarly (R-Ahx-R) 4 AhxB-PMO potently reduced viral titers of another flavivirus, West Nile virus (WNV), in cell culture experiments, and improved survival over 50% compared to controls in a WNV mouse model [47, 48] . Ebola virus causes severe hemorrhagic fever with fatality rates up to 90% in humans. In an in vivo mouse model, pre-or post-infection treatment with a R 9 F 2 C-PMO conjugate rescued mice from an otherwise-lethal Ebola virus infection [49] . We expect CPPs will continue to be a very useful research tool to deliver PMO and PNA for studying various biological processes, including splicing correction, where the leading potential therapeutic application is treatment of DMD [53] . However, CPP toxicity and endosomal entrapment appear to be limiting factors to the potential for systemic delivery of steric blocking ONs to humans. Future studies on the nature of the toxicity, endosomal entrapment and uptake mechanism of CPP-ONs will hopefully instruct the design of the next generation of CPPs having a higher therapeutic index. At the current time, CPP-ONs are more suitable for diseases amenable to either local treatments, infrequent systemic dosing, such as DMD, or perhaps a small number of systemic treatments, such as some acute viral infections. An initial Phase Ib/II clinical trial to evaluate the safety and efficacy of a CPP-PMO conjugate for the ex-vivo treatment of vein tissue in coronary artery bypass grafts has started in Poland and Ukraine. This clinical trial will evaluate AVI-5126, a (R-Ahx-R) 4 AhxB-PMO targeted to human c-myc, and is designed to prevent inappropriate cell proliferation that can occur after the grafting procedure and frequently causes eventual failure of the grafted vein. Excised vein tissue will be immersed in a solution containing 10 μM AVI-5126 for 30 min and then implanted as a bypass graft (www.clinicaltrials.gov). Hopefully, further clinical trials with CPP-PMO will become possible soon, and further commercial development of CPP-PNA as well. Inhibition of Rous sarcoma virus replication and cell transformation by a specific oligodeoxynucleotide Antisense technologies. Improvement through novel chemical modifications Antisense therapy for cancer Inhibition of translation initiation by antisense oligonucleotides via an RNase H-independent mechanism -methoxy)ethyl-modified anti-intercellular adhesion molecule 1 (ICAM-1) oligonucleotides selectively increase the ICAM-1 mRNA level and inhibit formation of the ICAM-1 translation initiation complex in human umbilical vein endothelial cells Morpholino antisense oligomers: the case for an RNase H-independent structural type Modification of alternative splicing by antisense therapeutics Silencing of microRNAs in vivo with 'antagomirs' Comparative evaluation of seven oligonucleotide analogues a potential antisense agents Antisense oligonucleotides with different backbones. Modification of splicing pathways and efficacy of uptake Effect of 2'-O-methyl antisense ORNs on expression of thymidylate synthase in human colon cancer RKO cells Nuclear antisense effects of neutral, anionic and cationic analogs Phosphoramidate oligonucleotides as potent antisense molecules in cells in vivo In vivo tumor growth inhibition and biodistribution studies of locked nucleic acid (LNA) antisense oligonucleotides Antisense inhibition of gene expression in cells by oligonucleotides incorporating locked nucleic acids: effect of mRNA target sequence and chimera design Oligonucleotide directed misfolding of RNA inhibits Candida albicans group I intron splicing Efficient and persistent splice switching by systemically delivered LNA oligonucleotides in mice Inhibition of HIV-1 Tat-dependent trans-activation by steric block chimeric 2'-O-methyl/LNA oligoribonucleotides A structure-activity study of the inhibition of HIV-1 Tat-dependent trans-activation by mixmer 2'-Omethyl oligoribonucleotides containing locked nucleic acid (LNA), a-LNA or 2'-thio-LNA residues Oligonucleotide analogues with an achiral backbone Uncharged Morpholino-based polymers having phophorus containing chiral intersubunit linkages PNA hybridizes to complementary oligonucleotides obeying the Watson-Crick hydrogen bonding rules A specificity comparison of four antisense types: morpholino, 2'-O-methyl RNA, DNA, and phosphorothioate DNA Inhibition of gene expression inside cells by peptide nucleic acids: effect of mRNA target sequence, mismatched bases, and PNA length Lon protease preferentially degrades oxidized mitochondrial aconitase by an ATP-stimulated mechanism Systemic delivery of morpholino oligonucleotide restores dystrophin expression bodywide and improves dystrophic pathology Targeted inhibition of miRNA maturation with morpholinos reveals a role for miR-375 in pancreatic islet development Evaluation of transfection protocols for unmodified and modified peptide nucleic acid (PNA) oligomers Specific antiviral activity of poly (L-lysine)-conjugated oligodeoxyribonucleotide sequence complementary to vesicular stomatitis virus N protein mRNA initiation site Downregulation of amyloid precursor protein inhibits neurite outgrowth in vitro Cell Internalization of the third helix of the Antennapaedia homeodomain is receptor-independent Downregulation of Cu/Zn superoxide dismutase leads to cell death via the nitric oxide-peroxynitrite pathway A truncated HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus Peptide-mediated cell delivery: application in protein target validation Peptide-oligonucleotide conjugates for the delivery of antisense and siRNA Peptide conjugates of oligonucleotide analogs and siRNA for gene modulation RNA targeting with peptide conjugates of oligonucleotides, siRNA and PNA Cell penetrating PNA constructs regulate galanin receptor levels and modify pain transmission in vivo Targeting of antisense PNA oligomers to human galanin receptor type 1 mRNA Inhibition of telomerase activity by a cell-penetrating peptide nucleic acid construct in human melanoma cells Enhancement of intracellular concentration and biological activity of PNA after conjugation with a cell-penetrating model peptide Intracellular uptake and inhibition of gene expression by PNAs and PNA-peptide conjugates Inhibition of coxsackievirus B3 in cell cultures and in mice by peptide-conjugated morpholino oligomers targeting the internal ribosome entry site Inhibition and escape of SARS-CoV treated with antisense morpholino oligomers Inhibition of multiple subtypes of influenza A virus in cell cultures with morpholino oligomers Inhibition of dengue virus serotypes 1 to 4 in Vero cell cultures with morpholino oligomers Inhibition of flavivirus infections by antisense oligomers specifically suppressing viral translation and RNA replication In vitro resistance selection and in vivo efficacy of morpholino oligomers against West Nile virus VP35 knockdown inhibits Ebola virus amplification and protects against lethal infection in mice Systemically delivered antisense oligomers upregulate gene expression in mouse tissues Antisense oligonucleotide-induced exon skipping restores dystrophin expression in vitro in a canine model of DMD Induced dystrophin exon skipping in human muscle explants Morpholino oligomer-mediated exon skipping averts the onset of dystrophic pathology in the mdx mouse Arginine-rich cell-penetrating peptides facilitate delivery of antisense oligomers into murine leukocytes and alter pre-mRNA splicing Identification and functional validation of PNAs that inhibit murine CD40 expression by redirection of splicing Structure-activity relationship study on a simple cationic peptide motif for cellular delivery of antisense peptide nucleic acid Evaluation of basic amphipathic peptides for cellular delivery of antisense peptide nucleic acids Up-regulation of luciferase gene expression with antisense oligonucleotides: implications and applications in functional assay development Synthesis, cellular uptake and HIV-1 Tat-dependent trans-activation inhibition activity of oligonucleotide analogues disulphide-conjugated to cell-penetrating peptides Peptide conjugates of oligonucleotides: synthesis and applications Conjugates of oligonucleotides and analogues with cell penetrating peptides as gene silencing agents Targeting peptide nucleic acid-protein conjugates to structural features within duplex DNA Automated synthesis of peptide nucleic acids and peptide nucleic-acid-peptide conjugates Synthesis and purification of peptide nucleic acids Disulfide conjugation of peptides to oligonucleotides and their analogues Bactericidal antisense effects of peptide-PNA conjugates Cell-penetrating peptide conjugates of peptide nucleic acids (PNA) as inhibitors of HIV-1 Tat-dependent transactivation in cells Evaluation of cell-penetrating peptides (CPPs) as vehicles for intracellular delivery of antisense peptide nucleic acid (PNA) Efficient splicing correction by PNA conjugation to an R6-Penetratin delivery peptide Inhibition of human telomerase by PNA-cationic peptide conjugates Synthesis and antiviral activity of a peptide-oligonucleotide conjugates prepared by using Na-(bromoacetyl)peptides Morpholino antisense oligomers: design, preparation, and properties HIV Tat peptide enhances cellular delivery of antisense morpholino oligomers Cellular uptake of antisense morpholino oligomers conjugated to arginine-rich peptides Vectorization of morpholino oligomers by the (R-Ahx-R) 4 peptide allows efficient splicing correction in the absence of endosomolytic agents Stability of cell-penetrating peptide-morpholino oligomer conjugates in human serum and in cells Delivery of bioactive molecules into the cell: the Trojan horse approach Cell-dependent differential cellular uptake of PNA, peptides and PNA-peptide conjugates Cell-penetrating peptides. A re-evaluation of the mechanism of cellular uptake Cationic TAT peptide transduction domain enters cells by macropinocytosis Comparison of basic peptides-and lipid-based strategies for the delivery of splice correcting oligonucleotides Cellular uptake of unconjugated TAT peptide involves clathrin-dependent endocytosis and heparan sulfate receptors Endosome trapping limits the efficiency of splicing correction by PNAoligolysine conjugates Calcium ions effectively enhance the effect of antisense peptide nucleic acids conjugated to cationic Tat and oligoarginine peptides Influenza fusion peptides Induction of splice correction by cell-penetrating peptide nucleic acids Peptide-based delivery of nucleic acids: design, mechanism of uptake and applications to splice-correcting oligonucleotides Photochemical internalization (PCI) -a novel technology for release of macromolecules from endocytic vesicles Photochemically induced gene silencing using PNApeptide conjugates Cellular delivery of polyheteroaromate-Peptide Nucleic Acid conjugates mediated by cationic lipids Nucleic acid based therapeutics for tumor therapy The design, synthesis, and evaluation of molecules that enable or enhance cellular uptake: peptoid molecular transporter Transporters comprising spaced arginine moieties Cell-penetrating peptides as transporters for morpholino oligomers: effects of amino acid composition on intracellular delivery and cytotoxicity Getting hydrophilic compounds into cells: lessons from homeopeptides 48000-fold acceleration of hybridisation by chemically modified oligonucleotides Dystrophin expression in the mdx mouse after localised and systemic administration of a morpholino antisense oligonucleotide Pharmacokinetics, biodistribution, stability and toxicity of a cell-penetrating peptide-morpholino oligomer conjugate Antiviral effects of antisense morpholino oligomers in murine coronavirus infection models Effects in live cells of a c-myc anti-gene PNA linked to a nuclear localization signal Therapeutically promising PNA complementary to a regulatory sequence for c-myc: pharmacokinetics in an animal model of human Burkitt's lymphoma Inhibition of Burkitt's lymphoma cells growth in SCID mice by a PNA specific for a regulatory sequence of the translocated c-myc We thank Philippe Clair and Martin Fabani for their helpful discussions, Paul Prevot for assistance with fluorescence microscopy, Donna Williams and David Owen with synthesis of PNA and peptides, and the Chemistry department of AVI Biopharma for synthesis of PMO and peptides. This work was funded by a grant from EC framework 5 (contract QLK3-CT-2002-01989 to BL and MG) and from IFCPAR (contract 3205-A to BL). S. Abes and R. Abes are recipients of pre-doctoral fellowships from Ligue contre le Cancer and Région Languedoc-Roussillon.