Polymer synthesis and characterization
The diblock and triblock copolymers were synthesized via sequential RAFT polymerizations with purification after each block synthesis step (Fig. 1A). (Propanoic acid)yl butyl trithiocarbonate (PABTC) was used as the chain transfer agent (CTA), since it is suitable to control the polymerization of acrylates and acrylamides . The core-forming hydrophobic block was synthesized by polymerization of n-butyl acrylate (nBA) and was used as a macroCTA for the subsequent RAFT polymerizations. This block was chain extended (i) with the amine-functional cationic monomer dimethylaminoethyl acrylamide (DMAEAm) yielding the diblock copolymer P(nBA-b-DMAEAm), or (ii) with the carboxyl-functional anionic monomer tert-butyl acrylate (tBA), followed by chain extension with DMAEAm to give the triblock terpolymer P(nBA-b–tBA-b-DMAEAm).
A PDMAEAm homopolymer was synthesized as a control polymer as described previously . Three different shielding polymers were used: (i) 2 kDa poly(acrylic acid) (PAA) which is commercially available, while (ii) poly(N-acryloylmorpholine) (PNAM) and (iii) a diblock copolymer comprising NAM and acrylic acid, P(NAM-b-AA), were synthesized by RAFT polymerization, starting with the hydrophilic PNAM followed by chain extension with tBA to obtain the diblock copolymer.
The size exclusion chromatography (SEC) curves in Fig. 1 display monomodal populations with relatively narrow molar mass distributions. The experimental number-average molar masses differ from the theoretical values, since the hydrodynamic radii of the applied polymers and standards utilized for calibration of the SEC systems were different (Table 1). Following characterization, P(NAM-b–tBA) and P(nBA-b–tBA-b-DMAEAm) were deprotected with TFA to expose the anionic carboxyl-group of PtBA, obtaining the anionic PAA block, as confirmed by 1H NMR spectroscopy (Additional file 1: Figures S3, S4). The molar mass distribution of P(NAM-b-AA) after deprotection obtained by aqueous SEC can be found in Additional file 1: Figure S4. It was not possible to measure the molar mass distribution by aqueous SEC for the diblock copolymer P(nBA-b-DMAEAm) and triblock terpolymer P(nBA-b-AA-b-DMAEAm) due to their amphiphilic characters.
To investigate the influence of different pH-values present in the biological system (e.g., pH 7.4 extracellularly and pH 5.5 in endolysosomes)  on the behavior and interactions of PAA and PDMAEAm, both polymers were titrated against HCl or NaOH, respectively, and the proportion of protonated amine groups (degree of charge) was calculated using Additional file 1: Equations S2, S3 (Fig. 2) . At pH 7.4, PDMAEAm shows a moderate degree of charge (60%), whereas PAA is nearly completely negatively charged (94%). By contrast, the degrees of charge are reversed at pH 5.0 (PDMAEAm: 99%, PAA: 56%). These results indicate on the one hand that PAA would neutralize a majority of the remaining positive charges of PDMAEAm at pH 7.4 in the extracellular environment, which could lead to reduced cytotoxicity. On the other hand, when the pH is decreasing in the endolysosomal pathway, the reducing amount of negative charge would unleash further positive charges which is beneficial for endosomal escape.
Micelle formation and pH dependence
Micelles of P(nBA-b-DMAEAm) and P(nBA-b-AA-b-DMAEAm) were prepared by gradually adding 150 mM NaCl solution as selective solvent to the polymer dissolved in the good solvent THF/MeOH (80/20 v/v). During this process, the block copolymers eventually underwent microphase separation, where PnBA formed the core (H—hydrophobic), and PAA (A—anionic) and/or PDMAEAm (C—cationic) the corona, resulting in the assembly of HC- and HAC-micelles, respectively. The polymer solutions were dialyzed against 50 mM sodium acetate buffer solution (pH 5.0) to replace the solvent mixture and to enhance the stability of the micelles by avoiding neutralization of the cationic charges by PAA at higher pH-values. The formation of micelles was verified first by DLS measurements (Additional file 1: Figure S6). In addition, the reproducible formulation and stability at RT for at least 1 year indicate high application potential (Additional file 1: Figures S9, S14).
The effect of a post-assembly addition of shielding to the HC-mic was investigated by mixing the HC-mic solution with shielding polymer solutions 3 + 1 (v/v) resulting in carboxy to amine (COOH/NH) ratios of 0.5 for the HCA and HCAS assemblies, which was similar to the HAC-mic. The assembly of HC-mic and PNAM (S—stealth, HCS) was prepared with a PNAM/PnBA molar ratio of 1.0. To investigate the behavior of the (layered) micelles at pH-values relevant for biological studies, the different assemblies were tested in 100 mM acetate-HEPES buffer of pH 5.0 or pH 7.4. At pH 5, cryo-TEM, DLS and ELS measurements showed no considerable differences regarding the size, morphology or the surface charge of the assemblies compared to the naked HC-mic (Fig. 3). Due to the measurement of the hydrodynamic diameter by DLS, the assemblies appeared to be slightly larger (55–66 nm) than in the cryo-TEM images (35–51 nm) but both methods provided comparable tendencies: The HAC micelles were slightly larger than the HC micelles, which can be attributed to the presence of the additional third PAA block. This partially uncharged anionic PAA block in the HAC-mic at pH 5.0 might form intramicellar interpolyelectrolyte complexes (im-IPECs) with protonated PDMAEMA or collapse to the micelle core , leading to a slightly increased size of the micelle core (21 ± 3 nm) in the cryo-TEM images compared to the HC-mic (14 ± 1 nm). With ζ-potentials in the order of 24 mV, all assemblies showed strong positive surface charges.
However, when DLS measurements were performed at pH 7.4, the hydrodynamic diameter decreased by about 7 ± 2 nm for HC-mic, HCS and HCAS. By contrast, HAC-mic and HCA formed large aggregates and turbid suspensions, which could be due to neutralization of PDMAEAm by PAA (Fig. 2). Moreover, the ζ-potential of all assemblies was decreased when measured in 100 mM acetate-HEPES buffer of pH 7.4, with the greatest difference being observed for the anionic polymer-containing assemblies HAC-mic, HCA and HCAS. Interestingly, HCAS exhibited a ζ-potential close to zero and comparable to the HAC-mic, but still formed stable and defined nanostructures instead of aggregates at pH 7.4. This indicates a beneficial contribution of the additional hydrophilic PNAM block to the micelle stability, in particular at neutral pH-values.
Polyplex formation and characterization
Since the micelles were designed for the purpose of transporting genetic material into cells, their interaction with genetic material was investigated with the combined ethidium bromide binding and heparin release assay (EBA, HRA) as described previously . The assay uses the increase in the relative fluorescence intensity (rFI) of ethidium bromide (EtBr) upon (re)intercalation into the pDNA as a fluorescence indicator for unbound pDNA. A decreasing rFI relates to displacement of EtBr from the pDNA due to the binding of polymers. The polysaccharide heparin was used to investigate the stability of the polymer–pDNA complexes (polyplexes) against polyanions outside the cells, since it competes with the pDNA for the binding to the polymers. As a starting point, an N*/P ratio (active amines of the polymer to phosphates of the pDNA) of 30 was chosen. To investigate the influence of shielding on the obtained polyplexes, the respective shielding polymers were added post-polyplex formation at an COOH/NH-ratio of 0.5 (HCA, HCAS) or at a PNAM/PnBA molar ratio of 1.0 (HCS). To increase biocompatibility and avoid aggregation of HAC-mic and HCA but still enable micelle-shielding interaction, a less strong buffer system was used for this and all further assays involving complexes of pDNA and polymer (polyplexes). The buffer contained 20 mM HEPES and 5% (w/v) glucose at pH 7.4 (HBG-buffer), resulting in pH-values of the polyplex solutions of pH 6.3 (HAC-mic) and 7.2 (remaining assemblies), which will be adjusted to pH 7.4 upon 1:10 dilution of the polyplexes with growth medium of pH 7.4 for cell treatment.
The results showed a high proportion of bound pDNA for all assemblies from 73 ± 2% (PDMAEAm, C) to 86 ± 1% (HAC-mic) being comparable to the commercial control linear poly(ethylene imine) (LPEI, 85 ± 3%, Fig. 4A). While slight increases were observed for the introduction of a hydrophobic core to the cationic PDMAEAm homopolymer (HC-mic: 79 ± 1%) and for the addition of the anionic block (HAC-mic), the addition of the shielding polymers did not change the proportion of bound pDNA compared to the naked micelle (HC-mic). Regarding the polyplex stability, the HC-mic required the highest concentration of heparin to release 50% of the pDNA, (HC50: 41.6 U mL−1) indicating a strong polymer–pDNA-interaction. The addition of the shielding polymers led to decreased HC50-values for the layered micelle assemblies compared to the naked micelle (HC-mic) with the HCA showing the lowest values (23.5 U mL−1). However, they were still higher than the HAC-mic (13.2 U mL−1), which was comparable to LPEI (8.7 U mL−1), and therefore represents a good candidate for transfection efficiency assays.
The decreased polyplex stability of the HAC-mic can be caused by the covalent connection of the anionic PAA with the cationic PDMAEAm blocks, effectively reducing the amount of excess positively charged amines available after polyplex formation (N*/P ratio). Therefore, a lower amount of heparin can be trapped by excess cationic charges (see also Fig. 4C). By contrast, the polyplex in the HCA assembly was formed in absence of PAA, which can therefore interact only with the remaining, free positively charged amines not occupied by the phosphate groups of the pDNA. Since the negatively charged heparin is repelled by the negatively charged PAA, increased concentrations of heparin were required to release the same amount of pDNA as the HAC-mic.
To further investigate the polyplex properties, DLS and ELS measurements were performed with (layered) polyplexes at N*/P 30 to determine the hydrodynamic diameter of the assemblies. Compared to the assemblies without pDNA in HBG buffer (Additional file 1: Figure S10), the formation of polyplexes led to slightly increased sizes (Fig. 4B), as observed in other studies before [23, 24, 62]. All polyplexes were slightly below 100 nm in diameter being optimal for endocytotic uptake . The HC-mic and the HCS and HCAS assemblies exhibited the smallest polyplexes (≈ 70 nm), whereas the other assemblies showed sizes of 90 to 100 nm. Regarding the HCA assembly, two distinct peaks were observed in the intensity plot (Additional file 1: Figure S8). Together with the increased PDI values, this could indicate the aggregation of several micelles due to the strong attraction between the positively charged shell of the HC-mic and the negatively charged shielding polymer PAA. Regarding the electrical potential, all assemblies containing micelles exhibited similar ζ-potentials of about 25 mV, which was slightly higher compared to the homopolymer polyplexes (C, 21 ± 3 mV) and could be explained with an increased charge density in the micellar corona .
As polycations are known for their interaction with cell membranes, different cytotoxicity assays were performed: (i) the PrestoBlue assay determining the metabolic activity, (ii) the LDH release assay determining the membrane integrity and (iii) flow cytometry determining the cell viability due to their appearance in the FSC/SSC plot. For all assays, the cells were incubated with the above described (layered) polyplexes.
The evaluation of the metabolic activity revealed differences between the assemblies. Polyplexes of the HC-mic and the HCS assembly caused the highest reduction in cell viability (≈ 50%), whereas cells incubated with the PDMAEAm polyplexes (C) showed no cytotoxicity (Fig. 5A). This indicates that the hydrophobic block and the micellar structure of the HC-mic contributed to cytotoxic effects, which have been also observed with other hydrophobic-cationic micelle systems and can be explained by a high local concentration of cationic moieties . The addition of anionic polymer, either post-polyplex formation (HCA) or within the micelle (HAC-mic), showed medium to low toxicity, while the combination with the stealth polymer PNAM (HCAS) eliminated the cytotoxic effect. Layering with only PNAM in the HCS assembly did not reduce cytotoxicity, since this assembly lacks an anionic counterpart for ionic interaction with the micelle. Furthermore, the HAC-mic also showed toxicity alleviating effects compared to the HC-mic in concentration dependent studies without pDNA in L-929 cells (Additional file 1: Figure S12A), which might be due to the decreased degree of positive charges in the HAC-mic.
In contrast to the PrestoBlue assay, there were only slight differences between the different polyplex assemblies in the LDH-release and flow cytometry assays. All assemblies led to viabilities above 80% (Fig. 5B, C). This indicates only slight or no influence of the polymers on the cell membrane and morphology, which is supported by hemolysis and aggregation assays with human erythrocytes and in microscopic investigations of HEK293T cells (Additional file 1: Figures S12C, D, S13). However, cells incubated with the HCAS assembly or the HAC-mic exhibited the highest viabilities and, therefore, represent promising candidates for further studies.
EGFP expression with (layered) micelles
Since all polymers exhibited high interaction with pDNA and led to low to moderate cytotoxicity, their transfection efficiency (amount of EGFP expressing cells) was investigated. For a better understanding of possible shielding effects, different incubation periods were examined. Therefore, HEK293T cells were incubated with the (layered) polyplexes at N*/P 30 for 24 h, 48 h or for 24 h followed by 1:2 splitting of the cells with fresh medium and an additional 24 h observation period (Fig. 6A). Subsequently, flow cytometry was used to determine the amount and relative mean fluorescence intensity (rMFI) of EGFP expressing cells. The results showed different EGFP expression for all assemblies in serum-containing media with slightly higher efficiencies than LPEI and mostly consistent within different assembly batches (Fig. 6, Additional file 1: Figure S16). A fourfold increase in transfection efficiency was observed when the incubation time (48 h) or the observation time (24 + 24 h) were extended, leading to transfection efficiencies of up to 95% viable EGFP positive cells with no changes in cytotoxicity (Additional file 1: Figure S12B). After 48 h, the homopolymer PDMAEAm resulted in nearly no transfected cells, whereas the HC-mic showed the highest transfection efficiency (95 ± 4% viable EGFP positive cells). The layering with PNAM alone (HCS) did not influence the effect of the HC-mic, but in combination with the anionic block (HCAS) the transfection efficiency was significantly reduced (47 ± 9%), but was still comparable with the commercial control LPEI. The two PAA containing assemblies without PNAM (HAC-mic and HCA) were comparable to the HC-mic (85 ± 14 and 85 ± 18%, respectively).
Furthermore, LPEI could only reach these values when the pDNA concentration was increased to 3.0 µg mL−1 and the polymer concentration remained the same (Additional file 1: Figure S16), whereas the transfection efficiencies of the assemblies did not change. This indicates a higher efficacy of these systems, requiring only half of the pDNA to achieve similar transfection rates as LPEI. The reason might be seen in the differences between polyplexes of micelles and pDNA compared to those of homopolymers with pDNA, leading to increased stability and the preservation of the pDNA structure as shown by Tan and coworkers . Although the HC-mic showed higher efficiency than the HAC-mic under these conditions, it is worth noticing, that HAC-mic outperformed the HC-mic at higher polymer concentrations due to cytotoxicity issues (Additional file 1: Figure S17).
Transfection mechanism of (layered) micelles
To gain an in-depth look into the transfection mechanism, the (layered) polyplexes were investigated regarding their performance at two crucial steps of the transfection process: (i) polyplex uptake and (ii) endosomal escape. For the polyplex uptake, the HEK293T cells were incubated with (layered) polyplexes of YOYO-1 labeled pDNA and polymers at N*/P 30 in serum-containing media for different time periods, before they were measured and analyzed via flow cytometry regarding the amount of YOYO-1 positive cells and the relative MFI (rMFI) of viable single cells (Fig. 7A, B). All assemblies exhibited a time-dependent increase of the uptake in nearly all cells after 24 h. Following 1 h of incubation, 14 to 36% of the cells were YOYO-1 positive with comparable rMFI values. Although, all cells showed polyplex uptake after 24 h, the anionic polymer containing assemblies led to slightly increased rMFI values compared to the pure cationic assemblies (HAC/HCA: 86 vs. HC/HCS: 70, p = 1.000). The HCAS assembly exhibited the lowest proportion of YOYO-1 positive cells and rMFI values. This could be due to the “stealth-dilemma” or due to a decreased aggregation number of micelles within one polyplex as it was shown for PEG-b-PDMAEMA-b-PnBMA micelles . Interestingly, the rMFI values for the successful assemblies (HC, HCA, HCS, HAC) increased although the cells and medium were split after 24 h and the observation time was increased (24 + 24 h, Fig. 6AII). This could point towards an interaction between these polymers and either cells or the culture vessel which is not disturbed by trypsinization.
The uptake of polyplexes following incubation of 1 h was further confirmed by CLSM (Additional file 1: Figure S20) showing a similar trend for the number of polyplexes per cell as observed via flow cytometry. By staining the plasma membrane and membrane originating organelles with CellMask Deep Red Plasma membrane stain, a colocalization analysis of polyplexes could be performed, which revealed the presence of about 10% free polyplexes for all treatments. Therefore, the second crucial cellular hurdle for transfection efficiency, the endosomal escape, was investigated using calcein as a non-permeable fluorescence dye leading to (i) a dotted pattern inside the cells following uptake via endocytosis or (ii) a diffuse fluorescence pattern upon its release into the cytosol if the endosomal membrane was disrupted, e.g., by a polymer. The HEK293T cells were incubated as described above with (layered) polyplexes at N*/P 30 and calcein. Following washing of the cells and staining of nuclei with Hoechst 33342, images were acquired by CLSM (Fig. 7D). The quantification of the number of cells showing calcein release was performed with ImageJ (see Additional file 1: Methods), and revealed that only the micellar assemblies displayed remarkable calcein release (Fig. 7C). The absence of calcein release by LPEI and PDMAEMA could point towards different endosomal escape mechanisms for linear polymers and micellar systems. As an example, different types of calcein leakage from vesicles (graded, all-or-none) have already been reported for detergents . Moreover, differences were also observed for the fluorescence pattern of the polyplexes in the CLSM uptake study (Additional file 1: Figure S21, weak fluorescence of linear polymers vs. bright, large spots for micellar systems), which could be an indication of different amounts of YOYO-1 labeled pDNA. However, the exact mechanism and quantification of the endosomal escape remain to be elucidated. All in all, the values of the micelles correlated well with those observed for transfection efficiency, with HC-mic and HCS leading to the highest (13 ± 5 and 14 ± 4% cells, respectively) and HCAS exhibiting the lowest calcein release (1.8 ± 1.6% cells). Calculating the escape efficiency [proportion of cells positive for calcein release (Fig. 7C) normalized to the proportion of YOYO-1 positive cells (Fig. 7A, 1 h)] demonstrated a superior escape efficiency for the HCA assembly. This could be due to an enhanced surface-accessibility of the PAA outside the micelle compared to the HAC-mic leading to a pH-dependent unmasking of cationic charges, and to an increased concentration of molecules which could also be beneficial for endosomal escape.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.