Synthesis and characterization of nylon-3 random copolymer

NM0.2/CP0.8 polymer was prepared with hydrophobic and hydrophilic β-lactams via anionic ring-opening polymerization (ROP) as described above. The synthesis led to a nylon-3 copolymer that contains randomly arranged hydrophobic and cationic subunits in a 1:4 ratio. The hydrophobic monomer was cyclopentadienyl β-lactam (CP), and the cationic monomer was a β-lactam without a methyl group (NM). Molecular weight and subunit ratios were determined by 1H-NMR spectroscopy as reported previously [42]. In this study, performed by our group, suitability for siRNA delivery into glioblastoma cells in comparison to various other nylon-3 polymers has been described and NM0.2/CP0.8 polymer was shown most promising regarding siRNA delivery into glioblastoma cells due to the high hydrophobic content [42]. Furthermore, since it was shown that hydrophobic moieties are able to induce adsorption of ApoE, the highly hydrophobic NM0.2/CP0.8 polymer was selected in this study to investigate the ApoE adsorption on NM0.2/CP0.8 PXs in comparison to cationic b-PEI PXs in an effort to examine the influence of ApoE adsorption on siRNA delivery efficiency of PXs across the BBB.

Size and zeta potential analysis of PXs by dynamic light scattering and laser Doppler anemometry

In order to investigate the hydrodynamic diameters, polydispersity indices, and zeta potentials of uncoated particles in comparison to coated PXs, DLS and LDA measurements were performed. Particle size and charge are two major parameters of NPs that affect intracellular uptake and transfection ability. Therefore, the first step of our study was to investigate the influence of modifying the particles with ApoE or PS80 + ApoE on the physicochemical characteristics of the PXs. For PS80 coating, we initially tested the optimal PS80 concentration by incubating the PXs with various PS80 concentrations ranging from 0.01 to 0.5% PS80 with subsequent measurement of sizes and PDIs of PS80-PXs by DLS. As illustrated in Fig. S1 (Supplementary Material), the sizes of NM0.2/CP0.8 PXs sizes were not affected at all by PS80 coating at the selected concentrations, whereas b-PEI PXs were destabilized by PS80. Only at PS80 concentrations of 0.1% and higher, particle sizes with low PDI values, which were comparable to those of the reference particles were obtained. In the case of NM0.2/CP0.8 PXs, we suggest that interactions between hydrophobic CP parts of the polymer and oleic acid groups of PS80 molecules resulted in immediate surface coating, reflected by colloidally stable particles over a wide PS80 concentration rage. However, in the case of b-PEI, it is conceivable that low surfactant concentrations led to destabilization and aggregation due to interactions between cationic amino groups of the PEI molecule with carboxyl groups of the PS80 molecule before the optimal concentration for polyplex stabilization was reached. Moreover, excess surfactant at concentrations above the optimum contributed to self-assembly of the molecule chains in micelles, which increased the PDI values in the DLS measurement [52]. Therefore, 0.1% PS80 was chosen as optimal concentration for further experiments.

For b-PEI PXs, as shown in Fig. 1, coating with ApoE and PS80 + ApoE resulted in a slight increase in particle size from 108.6 nm for the uncoated sample to 116.0 nm and 118.2 nm for the ApoE-coated sample and the PS80 + ApoE-coated sample, respectively. The PDI values were in the same range for all b-PEI samples; in detail, they amounted 0.095 for uncoated b-PEI PXs, 0.081 for ApoE b-PEI PXs, and 0.120 for PS80 + ApoE b-PEI PXs. The finding for the uncoated b-PEI PXs is in line with previously published results [53, 54]. The uncoated and ApoE-coated NM0.2/CP0.8 PXs revealed similar sizes of 123.9 nm and 123.3 nm, respectively, whereas the size for PS80 + ApoE-coated PXs increased in comparison to other samples to 194.8 nm. The same trend was also observed for the PDI values of NM0.2/CP0.8 PXs as they increased from 0.260 (uncoated PXs) and 0.272 (ApoE-coated PXs) to 0.401 for PS80 + ApoE-coated samples. The finding for uncoated NM0.2/CP0.8 PXs is in line with former studies [42]. The zeta potential of uncoated b-PEI PXs was + 30.9 mV and decreased with ApoE coating to + 28.23 mV and to + 26.10 mV after PS80 + ApoE coating. Uncoated NM0.2/CP0.8 PXs showed a zeta potential of + 21.7 mV that also slightly decreased after ApoE and PS80 + ApoE coating to + 20.27 mV and + 19.9 mV, respectively. All DLS measurements were performed in buffer, which does not reflect natural protein corona formation but allowed for investigating the impact of coating, washing, and centrifuging the PXs.

Fig. 1

Hydrodynamic diameters (left y-axis) and polydispersity indices (PDI, right y-axis) (A) and zeta potentials (B) of uncoated PXs, ApoE PXs, and PS80 + ApoE PXs formed with b-PEI and NM0.2/CP0.8 polymers at N/P ratio of 7 and 5, respectively (data points indicate mean ± SD, n = 3, one-way ANOVA with Bonferroni post hoc test, n.s not significant, *p < 0.05, **p < 0.01, and ***p < 0.005)

Since human ApoE has an isoelectric point of 5.55 [55] and is thus negatively charged at a pH of 7.2, it was concluded that ApoE molecules are adsorbed on the surface of positively charged PXs by electrostatic interactions and, in the case of NM0.2/CP0.8 PXs, additionally by hydrophobic interactions what might lead to the measurable size increase. ApoE is responsible for the redistribution of lipids among cells and tissues as parts of lipoproteins in the body and therefore bears a lipid binding region in the C-terminal domain [56]. For NM0.2/CP0.8 PXs, precoating with PS80 resulted in a higher and statistically significant size increase than for b-PEI PXs (no significant change), which might occur due to more efficient interactions between PS80 and NM0.2/CP0.8 PXs. The significant zeta potential decrease after coating with ApoE and PS80 + ApoE supports the assumption that negatively charged ApoE is adsorbed on the PXs’ surface. In summary, coating of ApoE and PS80 + ApoE showed little effect on the particle size, size distribution and zeta potentials of the PXs, resulting in particles with appropriate sizes and surface charges for further experiments.

Evaluation of bound ApoE by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS PAGE)

SDS-PAGE is a commonly used electrophoretic technique for separation and analysis of proteins based on their molecular weight. Protein bands within the gel can be visualized by a colorimetric staining such as Comassie [57]. SDS-PAGE was hence performed to investigate whether and to which extent ApoE is indeed adsorbed on different formulations. Therefore b-PEI and NM0.2/CP0.8 PXs were coated with ApoE and PS80 + ApoE, respectively, as described above. Coated PXs were purified via centrifugation as described above to remove free and unbound ApoE. Purified PXs were resuspended after the last centrifugation step and 25 µl of each sample (uncoated, ApoE-coated, and PS80 + ApoE-coated b-PEI and NM0.2/CP0.8 PXs) and ApoE RS (34.2 kDa) as reference, were loaded and run on a polyacrylamide gel for visualization of proteins by staining with Comassie Brilliant Blue G.

As displayed in Fig. 2, no bound ApoE was detected by SDS-PAGE on ApoE-coated b-PEI PXs (0%), whereas a slight band of ApoE was visible for PS80-precoated b-PEI PXs (6%). For NM0.2/CP0.8 PXs a slight band of ApoE appeared for ApoE-coated NPs (4%), while a more distinct ApoE band was observed for PS80-precoated samples (22%), indicating that a larger amount of ApoE was attached to PS80 + ApoE NM0.2/CP0.8 PX samples. These results are supported by the more visible size increase of PS80-precoated samples and led to the conclusions that the coating of PXs with ApoE was generally possible but the binding affinities of ApoE strongly depend on the underlying NP material and its modification, which is in line with previous reports. A former study of Blank assessed bound proteins on model polystyrene carriers as a function of their modification after incubation in plasma and described that on hydrophobically modified particles mainly apolipoproteins were found [58]. Subsequently, it was described for a wide variety of NPs that hydrophobicity facilitates the binding of specific proteins, such as ApoE [59]. Herein, we suggest that the C-terminal domain of the ApoE molecule that contains a lipid binding region [60] can directly interact with the hydrophobic CP subunits of the NM0.2/CP0.8 PX. Moreover, several other studies described that successful binding of ApoE was achieved by a surfactant-based approach in which poloxamers or PS were used as hydrophobic anchor [3]. Our results for PXs are in line with the literature, as adsorption of ApoE was not detected on cationic b-PEI PXs, while ApoE was found, albeit to a small extent, on more hydrophobic NM0.2/CP0.8 PXs. Moreover, PS80 precoating led for both, b-PEI and NM0.2/CP0.8 PXs to an adsorption of ApoE. Herein, we hypothesize that the higher amount of ApoE on NM0.2/CP0.8 PXs can be explained by more efficient interactions with PS80. PS80 might adsorb with its hydrophobic part (oleic acid) to the CP subunits of the NM0.2/CP0.8 PXs, whereas the hydrophilic part ((poly(ethylene oxide) (PEO) sorbitan) protrudes into the dispersion medium and facilitates the binding of ApoE, as similarly described for poloxamers by the group of Blunk et al. [58]. Taken together, SDS-PAGE results illustrated that functionalization of PXs with ApoE was successful by direct coating of NM0.2/CP0.8 PXs and by utilizing the surfactant-based approach with PS80 for both, b-PEI and NM0.2/CP0.8 PXs.

Fig. 2

ApoE bound to NM0.2/CP0.8 and b-PEI polyplexes with and without PS80 precoating evaluated by SDS-PAGE performed with a polyacrylamide-gel 10%. The slots of the gel were loaded with a molecular marker (10 to 250 kDa), ApoE RS (34.2 kDa) as reference, uncoated PXs, ApoE-PXs, and PS80 + ApoE-PXs samples. Proteins were visualized by staining with Comassie Brilliant Blue G. Gel was analyzed densitometrically and results for ApoE content are given in percentage of intensity in comparison to ApoE RS representing 100% intensity

In vitro experimentsQuantification of cellular uptake into glioblastoma cells by flow cytometry

As successful coating of PXs with ApoE was confirmed, the next step was to investigate whether this may lead to selective uptake of NPs in an LDL and LRP1-receptor bearing human glioblastoma cell line. ApoE is a component in lipoprotein classes VLDL and chylomicrons and mediates lipid transport and delivery into cells mainly via LDL and LRP1 receptor-mediated pathways [61]. The N-terminal domain of the ApoE molecule contains a lysine- and arginine-rich receptor binding site, and interactions of this domain with the respective receptor initiate endocytotic uptake of the particles into cells [62]. LDL receptor expression of the U87 glioblastoma cell line used in this study was measured elsewhere [63] and expression of LRP1 receptors was confirmed by antibody staining and subsequent flow cytometry measurements as described above and shown in Fig. S2 (Supplementary Material). Trypan blue quenching, which was additionally performed in order to exclude extracellular fluorescent signals caused by cell surface-bound siRNA, resulted in insignificantly lower MFI values for all tested b-PEI PXs, indicating that inconsiderable amounts of PXs were only bound to the outer cell membranes, whereas significantly lower MFIs values were detected for all tested NM0.2/CP0.8 PXs, pointing out that NM0.2/CP0.8 PXs s adhered more strongly to the cells due to hydrophobic interactions (Fig. S3, Supplementary Material). In order to exclude signals caused by surface-bound siRNA when quantifying cellular uptake of formulations, MFI values are presented after trypan quenching. As illustrated in Fig. 3, PS80-coated particles showed, against our expectations, no differences in MFI values in comparison to negative control samples with free siRNA and therefore no measurable cellular internalization into glioblastoma cells. The MFI value for negative control was 646.5, for PS80 + ApoE- b-PEI PXs 680.1 and for PS80 + ApoE-NM0.2/CP0.8 PXs 651.8, respectively. Uncoated NM0.2/CP0.8 PXs (MFI = 14,852) revealed a slightly higher cell entry capability than b-PEI PXs (MFI = 8548), what goes in line with literature [64]. The most efficient siRNA delivery in glioblastoma cells was observed by the ApoE-coated PXs, where significantly increased cellular uptake was detected in comparison to the uncoated particles. This effect, however, was even more pronounced for NM0.2/CP0.8 PXs than for b-PEI PXs. The MFI value for ApoE-coated b-PEI PXs was 9957 (uncoated 8548) and for ApoE-coated NM0.2/CP0.8 PXs an average MFI value of 25,216 (uncoated 14,852) was observed.

Fig. 3

Cellular uptake of uncoated PXs, ApoE PXs, and PS80 + ApoE PXs (b-PEI PXs: N/P ratio = 7 and NM0.2/CP0.8 PXs: N/P ratio = 5) after 24-h incubation as quantified by flow cytometry performed with trypan quenching and presented as median fluorescence intensitiy (MFI). Negative control: untreated cells (blank) and with free siRNA-treated cells (results are shown as mean ± SD, n = 3, one-way ANOVA with Bonferroni post hoc test; *p < 0.05 and ***p < 0.005)

The obtained results regarding the cellular uptake of PS80 + ApoE PXs are not in line with our expectations that arose from the SDS-PAGE result, which indicated successful binding of ApoE on PXs especially on PS80-precoated particles as presented in “Evaluation of bound ApoE by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS PAGE).” As no cellular uptake was detected for both, b-PEI and NM0.2/CP0.8 PS80-precoated samples, it may be speculated that PS80-ApoE-modified PXs were destabilized in the complex environment of cell culture medium that was required for the uptake experiment, so that no functional particles were present to enter the cells. To exclude the possible influence of ApoE coating, we repeated the experiment exemplarily using PS80-precoated b-PEI PXs and uncoated b-PEI PXs. As displayed in Figure S4 (Supplementary Material), PS80-precoated b-PEI PXs were not able to enter glioblastoma cells in comparison to uncoated b-PEI PXs confirming that PS80 coating tremendously reduced the cellular uptake ability of PXs. Our hypothesis that in particular hydrophobic NM0.2/CP0.8 PXs constitute a promising system for the PS80-ApoE targeting approach that were supported by promising SDS-PAGE and DLS results was hereby disproved. Consequently, PS80 coating does not seem to be a suitable approach for PXs in general, whereby the exact impact of PS80 coating on PX destabilization remains unclear and gives rise to investigate this question in future studies. Consequently, PS80-precoated samples were not considered further in the following experiments. Notwithstanding, the uptake data revealed that modification of PXs with ApoE alone significantly enhanced the cellular internalization ability of both b-PEI and NM0.2/CP0.8 PXs, whereas this effect was more pronounced for NM0.2/CP0.8 PXs. Therefore, we suggest that adsorbed ApoE on the surface of the PXs interacted with LDL and LRP1-receptors and consequently induced selective receptor-mediated endocytosis of PXs into glioblastoma cells [65]. This goes in line with recently described data in the literature. Particularly, the group of Kreuter et al. has described that coating of analgesic hexapeptide dalargin-loaded PBCA NPs with ApoE or ApoB alone mediated the delivery of the cargo across the BBB after intravenous injection. However, in their studies, ApoE/B overcoating after PS80 precoating induced an even higher effect in the in vivo experiment [20]. Further, the group of Mulik et al. has shown that PBCA NPs loaded with the drug curcumin were more efficiently taken up by SH-SY5Y cells after coating them with ApoE in comparison to plain PBCA NPs [66, 67]. These results underline our finding that ApoE coating of NPs alone could induce receptor-mediated endocytosis and that this approach can be translated from solid NPs to dynamic systems such as PXs at least in an in vitro setting. Interestingly, and in line with SDS-PAGE results, ApoE-NM0.2/CP0.8 PXs were more efficiently taken up in comparison to ApoE-treated b-PEI PXs what may be an indication that a higher amount of ApoE bound on NM0.2/CP0.8 PXs is able to induce more efficient receptor-mediated cellular uptake. In conclusion, our results indicate that a surfactant-based targeting approach with PS80 precoating for enhanced ApoE binding does not constitute an appropriate strategy for dynamic systems such as PXs due to suspected destabilization under in vitro conditions. However, direct coating of PXs with ApoE resulted in remarkably increased cell uptake, particulary for NM0.2/CP0.8 PXs, rendering them into highly promising candidates for ApoE-induced brain delivery. Considering the high LRP-1 expression of U87 cells, the ApoE-dependent uptake is hypothesized to be receptor-mediated.

GAPDH knockdown measurements by PCR

Next, we investigated whether the significantly higher internalization of ApoE-modified PXs correlated with the gene silencing of a targeted gene. Therefore, glioblastoma cells were transfected with ApoE-b-PEI PXs and ApoE-NM0.2/CP0.8 PXs prepared with siRNA against GAPDH (siGAPDH) or scrambled negative control siRNA (siNC), respectively. Uncoated samples were used as negative control, and Lipofectamine 2000 lipoplexes were utilized as positive control samples. GAPDH gene expression was quantified by real time PCR (RT-qPCR) and normalized by β-actin gene expression and normalized to the values obtained after transfection with negative control siRNA for each sample. As displayed in Fig. 4, significant downregulation of GAPDH expression was observed for uncoated NM0.2/CP0.8 PXs (knockdown of 34%) and more pronouncedly for ApoE-NM0.2/CP0.8 PXs (knockdown of 46%) compared with negative control samples. In contrast, there was no significant difference of GAPDH expression among b-PEI PX (uncoated and ApoE-coated)-treated cells and their respective negative controls. Positive control samples displayed a GAPDH gene silencing effect of 36% for Lipofectamine lipoplexes, reflecting the poor transfectability of U87 glioblastoma cells in general.

Fig. 4

GAPDH knockdown in the human glioblastoma cell line U87 as quantified by RT-PCR 48 h after transfection with uncoated and ApoE-coated b-PEI and NM0.2/CP0.8 PXs prepared with GAPDH siRNA or scrambled control siRNA as respective negative control (NC). The positive control consisted of Lipofectamine (LF) 2000 lipoplexes. The expression of GAPDH was normalized to the expression of β-actin and to cell samples transfected with the respective negative control PX (data points indicate mean ± SD, n = 3, one-way ANOVA with Bonferroni post-hoc test, n.s not significant; p > 0.05, *p < 0.05, and **p < 0.01)

Unmodified b-PEI PXs can undergo endocytosis by electrostatic interactions between their positive surface charge and the negative charge of the cellular membrane, however, the internalization of b-PEI PXs was not sufficient to induce significant GAPDH gene silencing effects. A lack in GAPDH knockdown ability was also observed for ApoE modified b-PEI particles, implying that even receptor-mediated endocytosis of b-PEI particles did not lead to significant gene knockdown effects. This goes in line with literature, as high molecular weight b-PEI was shown to exhibit, besides low toxicity, poor transfection efficiencies [68]. In contrast, unmodified NM0.2/CP0.8 PXs demonstrated efficient gene knockdown, leading to the assumption that additional hydrophobic interactions with the outer cell membrane as well as with the endosomal membrane might have beneficial effects on the siRNA amount present in the cytoplasm, what is a prerequisite for successful gene silencing. As described in literature, endosome disruption can be caused by cationic as well as hydrophobic moieties of polymeric NPs, which are both present in NM0.2/CP0.8 polymers [69]. However, in accordance with cellular uptake results, ApoE modification of NM0.2/CP0.8 PXs mediated the most efficient gene knockdown effect, comparable to the commercially available transfection reagent Lipofectamine 2000, indicating that receptor-mediated endocytosis in combination with favorable characteristics of the NM0.2/CP0.8 PXs provided most suitable conditions to induce sequence-specific gene silencing in a cell line, which is generally described to be hard to transfect [42]. Overall, the results presented here demonstrated efficient knockdown ability of ApoE-modified NM0.2/CP0.8 PXs, providing further evidence for the applicability of the targeted siRNA delivery system for potential treatment of brain diseases.

Cytotoxicity measurements Cytotoxicity measurements of PXs by CTB assay

In order to test the cytotoxicity of PXs, CTB assays were conducted with U87 cells that had been incubated for 24 h with uncoated b-PEI and NM0.2/CP0.8 PXs and ApoE NM0.2/CP0.8 PXs at two different N/P ratios. Thereby, lower N/P ratio of 5 (b-PEI) and 7 (NM0.2/CP0.8 PXs) represented treatment relevant conditions in in vitro experiments. The CTB assay is based on the ability of viable cells to reduce the nonfluorescent resazurin to fluorescent resorufin mainly by mitochondrial and cytosolic enzymes, while dead cells rapidly lose this capacity once their membrane integrity has been compromised [70]. All tested formulations resulted in favorable toxicity profiles, even at the higher N/P ratio of 15, as no considerable influence on cell viability was observed after PX treatment (Fig. 5).

Fig. 5

Cell viability as determined by CTB assay for formulated polyplexes at relevant N/P ratios after an incubation period of 24 h. DMSO 25% was used as positive control (Results are shown as mean ± SD as percentage of viable cells in comparison to untreated cells representing 100% viability, n = 3)

A very low but still negligible effect on cell viability was demonstrated solely by b-PEI PXs at N/P 15 (97.82%) and ApoE NM0.2/CP0.8 PXs at N/P 5 (99.28%). The positive control of 25% DMSO resulted in a survival rate of 55.54%. Summing up, these data showed that neither b-PEI PXs nor NM0.2/CP0.8 PXs with or without ApoE coating are expected to have a noticeable effect on U87 cell viability and, therefore, are well tolerated in in vitro experiments. As nylon-3 polymers have been investigated in order to design an advanced delivery system with excellent compatibility, our results are in line with previous studies that revealed high tolerability combined with efficient transfection efficacies especially for highly hydrophobic polymers [42, 64]. The functionalization of the NM0.2/CP0.8 PXs with ApoE was shown to not affect the cell tolerability in a significant manner. Although broadly used high molecular weight b-PEI is known for its high cellular toxicity due to its high cationic charge density [68], our findings demonstrated that b-PEI PXs do not exhibit significant cellular toxicity in our specific experimental setting. Altogether, these observations are especially important for future in vivo experiments as they demonstrate that our delivery systems are well tolerated after internalization into cells.

Hemocompability of polymers evaluated by hemolytic and erythrocyte aggregation assay

The hemocompability of b-PEI and NM0.2/CP0.8 polymers was investigated by hemolysis and erythrocyte aggregation assays. These assays are an indispensable initial step in evaluating the blood compatibility of NPs in advance of administering the materials intravenously in animals or humans. Several studies have reported a good correlation between the results of in vitro hemolysis assays and in vivo toxicity studies identifying hemolysis as a toxic effect [71]. Hemolysis occurs due to disruption of erythrocytes leading to the release of intracellular components such as hemoglobin, which can be measured spectroscopically. We tested 1 mg/ml as the highest polymer concentration assuming that the test concentration in vivo will not exceed this concentration. For comparison, in in vitro studies polymer concentrations of 0.011 mg/ml of b-PEI polymer to obtain PXs with N/P 7 and 0.203 mg/ml of NM0.2/CP0.8 polymers to obtain PXs with N/P 5 were used. The results of the hemolysis assays are represented in Fig. 6 and show that both formulations revealed no hemolysis at concentrations from 0.5 to 0.00781 mg/ml as the measured values were in the same range as the Hb0 value that represents the amount of basal hemoglobin found in the negative control sample. As shown in Fig. 6A, b-PEI polymer displayed slight hemolytic activity at the highest concentration of 1 mg/ml, namely a final hemolysis of 0.58%. At the same polymer concentration, as shown in Fig. 6B, NM0.2/CP0.8 polymers exhibited slightly higher hemolysis of 1.60%. Nevertheless, these values are still tolerable since substances are classified as non-hemolytic when hemolysis remains below 2% [72].

Fig. 6

Hemolysis of human erythrocytes at pH 7.4 induced by b-PEI (A) and NM0.2/CP0.8 polymers (B) as a function of log concentration values (results are shown as mean ± SD as percentage of hemolysis in comparison to Triton-X treated cells representing 100% hemolysis, Hb0 represents the amount of basal hemoglobin found in the negative control samples, n = 3)

The erythrocyte aggregation assay allowed only semiquantitative estimations about the hemocompatibility of the polymers. Results in this study are displayed in Fig. 7. Microscopic pictures of the erythrocytes are exemplarily shown for the concentrations of 1 mg/ml, 0.25 mg/ml, and 0.01563 mg/ml for b-PEI and NM0.2/CP0.8 polymers, as well as for PBS as negative control. PBS control samples did not induce any aggregation of the erythrocytes, whereas polymer solutions, depending on the concentration, caused a slight (0.25 mg/ml and 0.01563 mg/ml) to strong (1 mg/ml) aggregation of the RBCs, respectively.

Fig. 7

Erythrocyte aggregation profiles exemplarily shown for b-PEI and NM0.2/CP0.8 polymer concentrations of 1 mg/ml, 0.25 mg/ml and 0.01563 mg/ml in comparison to PBS used as negative control

These results suggest that a higher amount polymer and thus positive charges induced stronger aggregation effects due to interactions with negatively charged components of the RBC membrane. Additional hydrophobic interactions with the CP subunits of the NM0.2/CP0.8 polymers might lead to slightly more aggregation effects as visible for NM0.2/CP0.8 polymer concentrations of 0.25 mg/ml and 1 mg/ml. Taken together, the favorable hemolytic profiles and the low tendency to induce erythrocyte aggregation indicated that both polymers are well tolerated by RBCs, emphasizing the safe profile of these materials with regard to following in vivo experiments.

In vivo biodistribution experiments

After it was shown that ApoE-NM0.2/CP0.8 PXs in particular mediated targeted and efficient cellular uptake and in vitro knockdown in combination with favorable toxicity profiles, they were subsequently investigated in vivo in SWISS mice to evaluate their biodistribution behavior compared to non ApoE-coated NM0.2/CP0.8 PXs and in addition to b-PEI PXs. Therefore, siRNA was covalently coupled with DTPA to enable the labeling with 177Lu as radioactive marker for biodistribution studies of formulations following an adjusted protocol previously described for 111Indium-labeling of siRNA [49]. ApoE-NM0.2/CP0.8 PXs, NM0.2/CP0.8 PXs, and b-PEI PXs were formed with 177Lu-radiolabeled siRNA and intravenously administered through the tail vein, and biodistribution was investigated 1 h post injection in comparison to 177Lu-labeled free siRNA as control. As measured by gamma scintillation counting of resected organs (Fig. 8A), free siRNA exhibited a different biodistribution profile in comparison to PXs and accumulated preferentially in the kidney (50.65% ID/g), as reported earlier [50]. Small amounts of siRNA were in addition found in the liver (7.44% ID/g) and in the spleen (5.48% ID/g ratio). The results indicated that 177Lu-labeled siRNA encapsulated with b-PEI accumulated mainly in the liver (106.41% ID/g) and spleen (83.08% ID/g), which is in good agreement with published data [73]. In contrast, with 177Lu-labeled siRNA encapsulated with NM0.2/CP0.8 was preferentially detected in the lung, as 104.41% ID/g and 138.87% ID/g were found for NM0.2/CP0.8 and ApoE-NM0.2/CP0.8 PXs, respectively. In addition, NM0.2/CP0.8 and ApoE-NM0.2/CP0.8 complexed siRNA accumulated, although to a smaller extent, in the spleen (NM0.2/CP0.8 PXs 75.74% ID/g and ApoE-NM0.2/CP0.8 PXs 72.98% ID/g) and in the liver (NM0.2/CP0.8 PX 75.92% ID/g and ApoE-NM0.2/CP0.8 PXs 66.54% ID/g ratio). As illustrated in Fig. 8B, only low concentrations of free 177Lu-labeled siRNA and complexed with both, b-PEI and NM0.2/CP0.8 were detected in the brain. In fact, the lowest radioactive signal in the brain was observed in mice treated with ApoE-coated NM0.2/CP0.8 PXs 1 h post injection.

Fig. 8

Biodistribution of 177Lu-labeled siRNA, b-PEI PXs, NM0.2/CP0.8 PXs and ApoE-NM0.2/CP0.8 PXs in dissected organs (A) and the brain (B) of SWISS mice 1 h post injection (results are presented as % ID/g and shown as mean ± SD, n = 3, two-way ANOVA with Bonferroni post hoc test, n.s not significant; p > 0.05, *p < 0.05, **p < 0.01, and.***p < 0.005, mice that urinated during the incubation period were excluded from the study to avoid falsification of the results due to undefined loss of radioactive material)

Several in vivo studies performed with polyplexes have shown that a variety of parameters influence the biodistribution of siRNA formulations. Physicochemical properties of the NPs such as particle size, surface charge, and surface hydrophobicity further determine the stability of the polymer-siRNA complexes as well as interactions with proteins within the blood stream. Protein corona formation in turn can lead to recognition of the particles by the reticuloendothelial system (RES) determining the fate of the particles in the body after intravenous injection. Free labeled siRNA administered to mice was shown to be distributed to kidneys and the liver within minutes after injection, and a minor fraction of the siRNA was rapidly excreted via the urine so that levels of siRNA within the body decreased markedly after 24 h [74]. It is also known that siRNA is rapidly degraded by serum nucleases upon injection [75] implying that small fragments bearing the label might additionally circulate in the blood pool and show a different behavior than macromolecules, which needs to be considered as well when interpreting biodistribution data. Nevertheless, it was previously ascertained that renal clearance might occur more rapid than the degradation processes [76]. Moreover, free siRNA did not selectively accumulate in the other organs or in the brain, reflecting the poor ability of siRNA to penetrate the blood–brain barrier [50]. Taken together, our data for the biodistribution of 177Lu-labeled free siRNA in mice were consistent with previous reports. Labeled siRNA complexed with PEI demonstrated accumulation in the liver and the spleen. This goes in line with earlier findings, which suggest that PEI-particles are cleared from the blood stream after opsonization very rapid by phagocytosing macrophages ending up in the organs of the RES, namely, the liver and spleen [