Problems and limitations of CPPs for drug delivery in vivo

Read Also

Although highly efficient in mediating the in vitro cell uptake of different molecules into most cell lines, the in vivo use of CPPs appears much more complicated mainly because of a complete lack of cell specificity. Indeed, CPPs, and the therapeutic molecules attached to them, are dispersed almost all over the body independently of the way of administration (intravenous, IV, or intraperitoneal, IP). Such spreading has been highlighted by a seminal work that also underlines the CPP potential as a therapeutic approach [65]. After IP injection, the short Tat cell-penetrating sequence [6] fused to b-galactosidase was found in the lung, liver, kidney and other tissues, and, more surprisingly, in the brain, demonstrating that this fusion protein could cross also the blood brain barrier [65]. To our knowledge, the brain translocation of a Tat-fusion protein was confirmed in a single example with Bcl-xL, a well-characterized death-suppressing molecule of the Bcl-2 family, also following IP injection [66]. An other CPP related to the protegrins family has been also shown to pass across the blood brain barrier [67]. However, no study has been published so far to explain precisely how CPPs and their cargoes could cross this highly selective membrane. More importantly, no complete bio-distribution has been provided for such CPP-mediated in vivo transports.

Much effort has been put into improving the cell uptake of antibodies coupled to various CPPs since the tumor-targeting efficiency of most anticancer antibodies is severely limited by their poor penetration into the tumor mass [68]. As already discussed in a previous review [69], the coupling of cell specific antibodies to CPPs did increase their cellular internalization in vitro as expected [70]. However, it also reduced significantly their selectivity in vivo since non-targeted tissues took up the chimeric construct through non-specific internalization mediated by the CPP. This study revealed that the higher the number of CPP copies grafted on a cell-specific monoclonal antibody (Mab), the lower was the specificity of the Mab for the native cell. Significant localization and retention of CPP-Mab fragments in non-targeted tissues in vivo have been confirmed more recently by Jain et al., especially with a high peptide to antibody ratio [19]. In this study, the cell-penetrating “driving force” predominated over the specific antigen binding of the Mab fragments. Similarly, Niesner et al. observed a complete loss of in vivo cellular specificity when a tumor-targeting scFv fragment was coupled to the Tat peptide [68].

CPP-mediated delivery of an antibody raised against p21, an intracellular target inducing G1-S phase cell cycle arrest, has also been reported [71]. Compared to the uncoupled antibody, Tat increased the cellular accumulation of radio-labeled anti-p21 of about 10 folds, and the in vivo distribution study demonstrated that the Tat-conjugated anti-p21 Mab was more concentrated in tumors than the un-conjugated Mab [71]. However, accumulation of the Tat modified Mab in other tissues was also observed. In conclusion, the tumor to normal tissue ratio did not improve upon conjugation of the anti-p21 Mab with the Tat peptide.

Stein et al. conjugated a longer version of the Tat peptide spanning along residues 37 to 72 of the HIV-1 Tat protein to a Fab fragment of anti-tetanus toxin antibodies either by thioether or disulfide linkage [72]. They demonstrated that only the disulfide conjugates effectively neutralized the toxin. Surprisingly, the disulfide conjugate also showed a higher nuclear accumulation compared to the thioether conjugate [72]. One would have expected the opposite result since the intracellular reducing environment should be strong enough to cause the disulfide bridge reduction as shown in a previous study with a CPP-peptide conjugate [73]. Therefore, as a nuclear localization sequence (NLS) is present within the Tat peptide sequence [74], it is difficult to understand why the Fab fragment enters more efficiently the nucleus with a reducible linker than with a stable one. We already discussed extensively in a previous review [40] the features of reducible and stable links in CPP-morpholino-nucleic acid chimeras [75] and also in that case argued for an intracellular reduction of the chimeras in the cellular environment.

Moreover, Kameyama et al. showed some important differences in the body distribution, retention and metabolic fate of the conjugates depending on the type of CPP used (Tat, Antennapedia and Rev peptides coupled to the same Fab fragment) [76]. In another study, Tat or Antennapedia peptides were co-administrated with sc(Fv)₂ fragments [19]. In such a situation, there was no stable link between the CPPs and the sc(Fv)₂ fragment. However, a higher tumor retention was observed with the Antennapedia peptide and, to a lower extent, with the Tat CPP than with the sc(Fv)₂ fragment alone. However, the Tat peptide used in this study lacked one arginine residue [19]. Therefore, the lower tumor retention observed with Tat in this work should be reconsidered as the replacement of just a single cationic residue in Tat can greatly decrease its cell uptake [11, 13]. In another work a mixture of CPPs (Antennapaedia or Tat peptides) and cargo molecules were used to improve internalization of replication-deficient viruses for the therapeutic gene delivery, both in vitro and in vivo [18], but unfortunately, no bio-distribution study was performed. To conclude this section and despite the lack of a clear understanding of how CPPs deliver cargo molecules into the cell, there are evidences that CPPs induce a strong non-specific binding by sticking any molecule attached to them to non-targeted cells, leading to a dramatic loss of very “precious” material. Therefore, this issue is probably one of the major drawbacks for the use of these peptides as in vivo delivery vehicles if alternative strategies are not developed to promote cell-specific delivery