The objective of this study was to investigate the effect of the CPP penetratin on the cell and tissue interactions of exendin-4 and its antagonistic analogue exendin(9–39). We were especially interested in the opportunity to improve exendin(9–39) retention in GLP1R-expressing tissues. We found that penetratin increases binding and internalisation of exendin(9–39) in vitro, and specific tumour uptake in vivo. Our results provide proof of concept that CPP conjugation can be used to turn a non-internalising antagonist into an internalizing tracer molecule for molecular imaging and theranostics.
In vitro, penetratin led to increased binding and uptake of exendin(9–39) in GLP1R-expressing cells. In contrast to binding, internalisation could not be completely blocked by an excess of unlabelled exendin(9–39). This observation is consistent with the presence of receptor unspecific internalisation triggered by the CPP. However, uptake of the exendin-CPP conjugates was low in cells that do not express GLP1R (Fig. 1A-B). This shows that efficient uptake only occurs when interaction of the ligand with the receptor is present. We have previously observed similar combined effects for penetratin and the nanobody 7D12 [22].
In contrast to what we found for exendin(9–39), conjugation of exendin-4 to penetratin did not lead to an increase in internalisation. A slight increase in binding was observed, which is probably due to interaction of the CPP with the cell membrane. Exendin-4 internalisation was very efficient, as has been reported before [9].
The observations we made for exendin(9–39)-Pen raise new questions for the CPP field. A current working hypothesis proposes that CPPs trigger internalisation by cross-linking of glycosaminoglycans [26]. Observations supporting this hypothesis were typically made by following the uptake of fluorescently-labelled CPPs. Fluorescence-based assays require concentrations in the lower micromolar range, while the 50 pM concentration used in our radioactivity-based assay is about five orders of magnitude smaller. Extensive cross-linking of glycosaminoglycans seems unlikely at pM concentrations, even if exendin(9–39)-mediated receptor binding will certainly contribute to some enrichment of the peptide at the plasma membrane. To our knowledge it is the first time that CPP internalisation is studied at such low concentrations, and the scope of the work we present does not allow speculation on possible mechanisms. We hope these observations will motivate new interdisciplinary research, complementing fluorescence-based assays with other approaches, to gain a deeper understanding of the capacity of CPPs to trigger endocytosis.
The results in vivo reflected the differences observed in vitro between agonist and antagonist conjugates. Exendin(9–39)-Pen reached a higher tumour uptake in comparison to exendin(9–39) (Fig. 2C). Importantly, exendin(9–39)-Pen showed higher tumour-to-kidney ratios than unconjugated exendin(9–39) (Fig. 2D). In contrast, exendin-4-Pen did not show higher tumour accumulation than unconjugated exendin-4. Considering the in vitro data, increased tumour uptake of exendin(9–39)-Pen is likely due to internalization of the tracer and subsequent intracellular trapping of the residualizing complex [111In]In-DTPA. This confirms our hypothesis that CPP-mediated internalisation leads to higher tissue accumulation of the antagonist.
Exendin-4 showed considerable uptake into the pancreas which, however, was lower for the exendin-4-Pen conjugate. This difference in uptake may be explained by sequestration of the CPP conjugate in the liver. In contrast, neither exendin(9–39) nor exendin(9–39)-Pen showed increased pancreatic uptake. However, one must be cautious in drawing conclusions from mouse pancreatic uptake. The mouse exocrine pancreas takes up exendin in a GLP1R unspecific manner, which does not reflect the human situation. Rats are a more suitable model for pancreatic uptake, as our group previously reported [27].
Both exendin-4-Pen and exendin(9–39)-Pen showed unspecific liver uptake, which was not observed for the exendin analogues without penetratin. For exendin-4, the increased hepatic sequestration of the penetratin conjugate correlated with a decreased distribution to the pancreas and tumours. By comparison, for the exendin(9–39)-Pen conjugate this was not the case. Liver sequestration is a common phenomenon among CPPs [26,28], and we also observed it upon conjugation of penetratin to a nanobody. The positive charges of the CPPs are often named as a possible reason, but the mechanism has not been defined in detail. For tumour imaging, the only negative influence of high liver uptake could be in the detection of tumours close to the liver or of liver metastases. Otherwise, it would not pose serious health risks. However, decreasing liver uptake could lead to even higher tumour uptake. We showed for the nanobody 7D12 that CPPs with different physicochemical properties differentially affect the interaction of their conjugates with cells [22,29]. A next challenge will be to identify CPP-ligand pairs that enhance accumulation at the target site with little hepatic accumulation. It would be interesting to assess if less charged CPPs, or shielded activatable CPPs [30–32] can achieve this goal. Furthermore, as the effects of N-terminal conjugation instead of C-terminal conjugation are conjugate dependent [33], it would be worthwhile to test N-terminal CPP conjugation of exendin(9–39). Finally, CPPs containing D-amino acids could be beneficial for uptake, as they have higher proteolytic stability [17,34,35].
A tracer with the characteristics of exendin(9–39)-Pen is likely to quickly find a way to application. Our group provided proof of concept that exendin can be used for image-guided surgery (IGS) and targeted Photodynamic Therapy (tPDT) [36,37]. These techniques have great theranostic potential but require high pharmacological doses of the tracer, increasing the risk of side-effects when using an agonist. This complication underlines the need for effective antagonistic tracers to avoid side-effects. Furthermore, as for tPDT and IGS fluorophores are locally activated by light application, tracer accumulated in other organs (e.g. the liver) remains inactivated and will thus neither disturb detection of tumour tissues nor cause off-target toxicity.
Importantly, this is the first study that investigated the impact of CPP-conjugation for an antagonistic G-protein-coupled receptor peptide ligand from in vitro to in vivo. Previously, nona-arginine was shown to increase the in vitro uptake of a peptide conjugate consisting of bombesin and an endosome-disrupting peptide, aiming at the cytosolic delivery of plasmid DNA [38]. However, for these conjugates no in vivo data have been presented. The previous investigation that most resembled our approach was the N- and C-terminal conjugation of several CPPs to the agonistic peptide PTH(1–34), derived from the parathyroid hormone (PTH) [33]. The IC50 values and in vitro epithelial permeability were assessed, but not cellular internalisation or biodistribution. Interestingly, that study showed that C- and N- terminal conjugation of the CPP changed the properties of the PTH(1–34) conjugate, in a different way for each CPP [33]. This ties back to the point discussed above.
In conclusion, we showed for the first time that a CPP efficiently causes cellular internalisation of an antagonistic, non-internalising peptide ligand, thereby increasing tumour retention of the tracer in vivo. This result opens the door to further unleashing the great potential of exendin, as a research tool and as a theranostic agent. Future research into CPP conjugates should be extended to other non-internalising peptide antagonists.