The use of short-interfering RNAs (siRNAs) is an established method to silence gene expression [1]. Effective delivery to targets, including tumors, requires balancing high potency with low carrier toxicity [2,3,4,5,6,7,8,9]. Cationic polymers, such as polyethyleneimine (PEI) and polylysine [2], are widely used as nucleic acid carriers; however, application of these materials to in vivo disease models is often limited by their cytotoxicity. Many strategies have been explored to translate formulations from in vitro to in vivo to overcome issues including toxicity, aggregation, degradation, and target site enrichment [10]. PEGylation and attachment of carbohydrates, alkyl chains, and cell targeting moieties are frequently applied strategies [11]. Modification of PEI with neutral or anionic moieties has been shown to reduce cytotoxic effects, sometimes without loss of the endosomal rupture abilities. In particular, hydrophobic modifications have been shown to improve siRNA delivery [12,13,14,15,16]. In addition, further functionalization of PEI with primary amines showed improved transfection efficiency and reduced cytotoxicity for nucleic acid delivery [17]. Given the smaller size of siRNA versus pDNA, one could reason that lower MW PEI (or BPEI) might improve delivery. For example, although oliogoethylenimine (OEI) 800 is inactive for siRNA delivery, hydrophobic modification of this construct with ten hexyl acrylate residues per OEI molecule yielded a promising carrier for siRNA delivery [12]. In an effort to improve delivery, not only intracellularly but also in vivo, full deacylation of PEI increased gene delivery to mouse lungs [18]. Hydrophobic modification of similar amidoamine carriers increased stability and efficacy [19]. Oligomers, such as guanidinium-rich amphipathic oligocarbonates, also serve to balance cationic and hydrophobic interactions with lower MW polymeric architectures to achieve effective delivery [20]. In vivo siRNA delivery to endothelial cells has also been achieved using alkyl-modified low molecular weight PEIs formulated with PEG lipids [21]. The effect of polymer composition and architecture on delivery was shown using triblock copolymers of poly(LPEI-b-(propylene glycol)-b-LPEI). Whereas LPEI50-b-PPG36-b-LPEI50 showed poor efficacy, decreasing the LPEI block length and increasing the hydrophobic block to LPEI14-b-PPG68-b-LPEI14 greatly improved delivery [22]. In these examples, it appears that incorporation of siRNA into ordered nanoparticles provides increased electrostatic interactions. Also, the addition of hydrophobic interactions further stabilizes the structures above the amine pKas.

There are many reports to illustrate that fluorination is an effective approach to improve the pharmacokinetic drug properties in terms of stability and effectiveness [23, 24]. In addition, fluorocarbon modification of delivery carriers benefits gene therapy with enhanced biocompatibility and serum stability [25,26,27,28,29,30,31,32,33,34,35,36,37]. Recent studies show that fluorination of PEIs or PAMAM dendrimers reduces the cytotoxicity and improves the potency of siRNA delivery to cells [38, 39]. However, the role of fluorine modification has not been fully recognized; particularly, effects on in vivo behavior of siRNA nanoparticles have not been extensively investigated.

In this research, we used branched polyethyleneimine (PEI) as a model cationic polymer to explore the effect of fluorination on siRNA delivery both in vitro and in vivo. We chose two different kinds of fluorination strategies by amidation with two different lengths of fluorine chains, ethyl trifluoroacetate, and perfluorobutyryl chloride. By varying the modification degree, we obtained two series of polymers, CF3PEI (1-6) and F7PEI (1-6). Fluorinated PEI carriers mediated successful siRNA binding and luciferase reporter gene knockdown in the triple negative breast cancer (TNBC) cell line MDA-MB-231, while simultaneously decreasing cytotoxicity and aggregation versus unmodified control PEI. Interestingly, we found that PEI fluorination significantly shifted the biodistribution of siRNA from the lungs to the liver (Fig. 1), which provides insights for rational design of other fluorinated cationic polymers for liver targeting.

Fig. 1
figure 1

Fluorination of branched PEI (Mw = 25,000 g/mol) altered the biodistribution of siRNA from the lungs to the liver after intravenous administration of siRNA polyplexes to mice

Full size image

Guan and Chen et al. reported the fluorination of PEI using the reaction between the primary amine and epoxy or anhydride groups [26, 38]. Here, we proposed a facile approach to the fluorination of branched PEI (MW = 25,000 g/mol) with trifluoroacetate or perfluorobutyryl chloride under mild conditions (Scheme 1). By varying the mole ratio of PEI to fluorine compound from 1 to 6, we prepared two series fluorinated polymers with varying fluorination degree and the length of fluorine side chains (Supplementary Material Table S1). Fluorinated PEIs were characterized by gel permeation chromatography (GPC), 1H nuclear magnetic resonance spectroscopy (NMR), and 19F NMR (Supplementary Material Figure S1 and Supplementary Material Figure S2)

Scheme 1
scheme 1

Fluorination of branched PEI via amidation with ethyl trifluoroacetate or perfluorobutyryl chloride in dichloromethane

Full size image

Successful siRNA delivery was demonstrated by the decrease in relative luciferase expression in MDA-MB-231 cells after treatment with polyplex NPs containing siRNA against luciferase (Fig. 2). The delivery efficacy increased with polymer/siRNA weight ratio for both polymer series implying that more fluorinated polymers may benefit the complexation with siRNA and facilitate endosomal release of siRNA inside MDA-MB-231 cells. At a weight ratio of 50:1 (Fig. 2c), siRNA delivery efficacy decreased with an increase of fluorination degree for CF3PEI polymers, while F7PEI polymers exhibited efficacious delivery even at high fluorination degree. Lead fluorinated polymers showed comparable siRNA delivery efficacy with the benchmark PEI under the same conditions. Importantly, there was a remarkable reduction in cytotoxicity for fluorinated PEI polyplexes comparing with unmodified PEI polyplexes, suggesting that fluorination maybe an approach to tackle the toxicity challenge of PEI-based carrier in nucleic acid delivery.

Fig. 2
figure 2

Luciferase expression and viability of MDA-MB-231 cells after treatment with siRNA polyplexes of a CF3PEI or b F7PEI at polymer/siRNA ratio (wt/wt) of 10, 20, and 50. c In vitro delivery of siRNA with fluorinated polymers at polymer/siRNA ratio (wt/wt) of 50. CF3PEI-2-10 denotes the sample of trifluoroacetate modified PEI (trifluoroacetate chloride/PEI = 2/1, mol/mol) and siRNA with a polymer/siRNA ratio (wt/wt) of 10/1. PEI-10 denotes the sample with PEI/siRNA ratio (wt/wt) of 10. Unmodified PEI was used as a control under the same conditions. Gray bars indicate the relative luciferase expression and black dots denote the cell viability compared with untreated cells

Full size image

The binding affinity of the starting PEI and modified PEIs to siRNA was quantified using the Ribogreen binding assay at the optimized weight ratio of 50:1 (polymer:siRNA) (Fig. 3a). The result showed that the fluorination did not significantly change the binding affinity of siRNA to PEI under the test conditions. The size of siRNA polyplex nanoparticle with fluorinated PEIs varied with the fluorine modification ratio. For the F7PEI series, the diameter of most siRNA NPs was smaller than that for unmodified PEI-based NPs (Fig. 3b). Besides the decrease in particle size, the intrinsic hydrophobicity of fluorine groups in PEI should improve the stability of NPs, implying that these polymers may have better potential for siRNA delivery over unmodified PEI.

Fig. 3
figure 3

Complexation between siRNA and fluorinated PEI. a siRNA binding and b size of siRNA polyplexes formed at a siRNA/polymer ratio of 50:1 (wt/wt)

Full size image

In vitro delivery results showed that fluorination benefited siRNA delivery by significantly reducing the cytotoxicity of higher molecular weight PEI, which is in agreement with a previous report on fluorinated low molecular weight PEI (Mn = 600) [38]. Here, we carried out animal experiments to further investigate how fluorination affects in vivo behavior of siRNA NPs (Fig. 4). It is now appreciated that the chemical structure of delivery carriers determines the physicochemical properties of corresponding siRNA NPs, which in turn affect serum protein-NP interactions and the fate of siRNA NPs in vivo [40, 41]. To examine this behavior, we employed fluorescent Cy5.5-labeled siRNA to track biodistribution of fluorinated and non-fluorinated siRNA polyplexes. We choose CF3PEI-5 polymer because it has lowest toxicity with moderate delivery efficacy (~ 60% luciferase knockdown). Mice were injected intravenously with 2.5 mg/kg Cy5.5-siRNA complexes with PEI and CF3PEI-5 at a polymer/siRNA ratio of 50 (wt/wt). Ex vivo organ imaging of mice injected with unmodified PEI Cy5.5-siRNA polyplexes showed that the majority siRNA polyplexes accumulated in the lungs, while there was considerable siRNA retention in the liver and kidneys. This is consistent with previous reports (Fig. 4a) [42, 43]. In contrast, Cy5.5 siRNA-fluorinated PEI polyplexes mainly accumulated in the liver and to lesser extent, the kidneys for CF3PEI-5-50 (Fig. 4b). There was no detectable signal in the lungs. This effect could be beneficial to reduce RES and endothelial cell uptake to promote liver delivery of PEI-based carriers. The fluorination of PEI altered the biodistribution of siRNA NPs from the lungs to the liver. In future work, we plan to examine the detailed effects of fluorination on in vivo siRNA delivery, additional chemical modifications, and the potential of fluorinated PEI for the treatment of liver diseases.

Fig. 4
figure 4

Representative ex vivo images of organs after tail vein injection of polyplexes formed with Cy5 siRNA and a unmodified PEI or b CF3PEI-5-50. Mice were injected with 2.5 mg/kg Cy5.5-siRNA (i.v.). Images were taken 4 h after systemic administration of siRNA polyplexes

Full size image

In this study, we developed a facile approach to synthesize fluorinated branched PEI and investigated the effect of fluorination on siRNA delivery both in vitro and in vivo. All fluorinated polymers showed above 90% binding affinity with siRNA and the resulting siRNA polyplexes were stable at a polymer/siRNA ratio (wt/wt) of 50. In vitro delivery studies showed that by fluorination, the cytotoxicity of PEI was significantly reduced, and the lead fluorinated polymers were still shown to be effective for siRNA delivery to MDA-MB-231 cells. The organ level biodistribution of siRNA was evaluated by ex vivo imaging following tail vein injection of Cy5.5 siRNA polyplexes in mice. Our results suggest that the fluorination can substantially alter the distribution of siRNA polyplexes, i.e., siRNA was mainly accumulated in the liver for siRNA-fluorinated PEI NPs in contrast to the major distribution of siRNA in the lungs for NPs with non-fluorinated PEI. Fluorination may provide a useful strategy to reduce the toxicity of cationic polymer-based siRNA carriers, while simultaneously increasing the liver targeting in vivo. Such chemical reactions could thereby reduce the adverse effect of lung-related toxicity of PEI.