Choosing an Optimal Solvent Is Crucial for Obtaining Cell-Penetrating Peptide Nanoparticles with Desired Properties and High Activity in Nucleic Acid Delivery

Abstract

Cell-penetrating peptides (CPPs) are highly promising transfection agents that can deliver various compounds into living cells, including nucleic acids (NAs). Positively charged CPPs can form non-covalent complexes with negatively charged NAs, enabling simple and time-efficient nanoparticle preparation. However, as CPPs have substantially different chemical and physical properties, their complexation with the cargo and characteristics of the resulting nanoparticles largely depends on the properties of the surrounding environment, i.e., solution. Here, we show that the solvent used for the initial dissolving of a CPP determines the properties of the resulting CPP particles formed in an aqueous solution, including the activity and toxicity of the CPP-NA complexes. Using different biophysical methods such as dynamic light scattering (DLS), atomic force microscopy (AFM), transmission and scanning electron microscopy (TEM and SEM), we show that PepFect14 (PF14), a cationic amphipathic CPP, forms spherical particles of uniform size when dissolved in organic solvents, such as ethanol and DMSO. Water-dissolved PF14, however, tends to form micelles and non-uniform aggregates. When dissolved in organic solvents, PF14 retains its α-helical conformation and biological activity in cell culture conditions without any increase in cytotoxicity. Altogether, our results indicate that by using a solvent that matches the chemical nature of the CPP, the properties of the peptide-cargo particles can be tuned in the desired way. This can be of critical importance for in vivo applications, where CPP particles that are too large, non-uniform, or prone to aggregation may induce severe consequences.

Keywords: cell-penetrating peptides; nanoparticle formation; nucleic acid delivery; solvent.

Figure 1

Hydrodynamic diameter and polydispersity indexes (PDI) of particles formed by differently dissolved PF14. (A) Particles formed by PF14 dissolved in MQ water, alcohols, and the mixture of EtOH and DMSO. (B) Particles formed by PF14 dissolved in a mixture of EtOH and DMSO with addition of TMC at different percentages. The 1 mM PF14 stock solutions in different solvent mixtures were diluted 10× with MQ water and incubated at room temperature for 30 min before measurement with Malvern Zetasizer Nano. Each dataset mean ± SD of at least three measurements with individual measurements indicated as black dots. The polydispersity indexes are presented as the average of at least three independent measurements. All proportions are given as v:v.

Figure 2

Effect of nanoparticles prepared from SCO or siRNA and differently dissolved PF14 on splicing correction and luciferase silencing, respectively. The 1 mM stock solutions of PF14 were prepared by dissolving the peptide in different solvents and their mixtures, as indicated in legends (proportions are given as v:v). (A) HeLa pLuc 705 cells were incubated for 24 h with solutions containing SCO alone (100 nM), nanoparticles of SCO and PF14 taken at MR 5 (PF14:SCO) with or without the addition of 3 mM CaCl₂ or MgCl₂. (B) U87 MG-Luc2 cells were incubated for 48 h with solutions containing siRNA alone (15 nM), nanoparticles of siRNA and PF14 taken at MR 34 (PF14:siRNA) with or without the addition of 3 mM CaCl₂ or MgCl₂. As a positive control, siRNA was transfected with Lipofectamine RNAiMAX. In all cases, as a negative control, the cells were incubated with a medium containing 10% (v:v) of MQ water (“Untreated”). siNEG—negative, i.e., non-targeting siRNA. Each dataset represents mean ± SD of technical replicates (shown as black dots) from three independent experiments. Data were analysed using one-way ANOVA with post-hoc Tukey’s test. Asterisks indicate statistically significant difference compared to the same solution from “MQ water” group (A) or between indicated datasets (B), * p-value < 0.05, ** p-value < 0.005, *** p-value < 0.0005, **** p-value < 0.0001.

Figure 3

Morphology of particles formed by differently dissolved PF14 as analysed by TEM. The 1 mM stock solutions of PF14 were prepared by dissolving the peptide in different solvents and their mixtures: MQ water (A), EtOH (B), EtOH90/DMSO10 (C) or in EtOH90/DMSO9.6/TMC0.4 (D), and resulting particles were analysed by TEM. All proportions are given as v:v.

Figure 4

Internalisation and biological effect of nanoparticles containing SCO, differently dissolved PF14 and CaCl₂. HeLa EGFP 654 reporter cells were incubated with nanoparticles of 180 nM SCO-654, 20 nM Cy5-SCO-654, 1 μM PF14 and 3 mM CaCl₂ for 24 h. Cells were fixed, and specimens were analysed with Olympus FluoView FV1000 confocal microscope. Cells were either left untreated (A), incubated with the two SCOs (B), the SCO-PF14 nanoparticles (C–E) or the SCO-PF14-Ca²⁺ nanoparticles (F–H). The following PF14 stocks were used: PF14 dissolved in MQ water (C,F); PF14 dissolved in EtOH90/DMSO10 mixture (D,G); and PF14 dissolved in EtOH90/DMSO9.6/TMC0.4 mixture (E,H). Rescued EGFP expression is shown in green. Cell nuclei were visualised with DAPI (blue). For tracking cellular association and internalisation of oligonucleotide, Cy5-labelled SCO was used (red). The merged images of all confocal layers are presented. Scale bar: 50 µm.

Figure 5

Morphology of particles formed by PF14 dissolved in EtOH90/DMSO9.6/TMC0.4 and its complexes with SCO and divalent metal ions added in the form of CaCl₂ and MgCl₂. Nanoparticles were prepared by mixing SCO and PF14 at MR 10 and adding CaCl₂ and MgCl₂ after 15 min. After a total of 30 min of incubation, solutions were analysed by AFM, field emission SEM (FE-SEM), or high-resolution TEM (HR-TEM).