*3.4. Therapeutic In Vivo Application of P5Y*/*siRNA Complexes Leads to Profound Anti-Tumor E*ff*ects*

Pivotal in the development of nanoparticles for siRNA delivery is their applicability in vivo. In previous biodistribution assays, we showed that PEI/siRNA or P10Y/siRNA complexes become systemically available upon i.p. injection, and thus are able to reach organs/tissues remote from the injection site including tumors [50,55]. These studies were now extended towards P5Y/siRNA complexes. In a first experiment, P5Y/siRNA complexes were tested for reporter gene knockdown in mice bearing s.c. HCT116-Luc tumor xenografts. Prior to the in vivo application, the P5Y complexes were first evaluated in vitro for their potential to transfect this cell line. The luciferase expression was

decreased to 60% at the lowest siRNA concentration of 15 pmol and further to 25% of control levels at 30 pmol siRNA (Figure 4A).

**Figure 4.** (**A**) Evaluation of luciferase knockdown efficacies of P5Y/siRNA complexes for in vivo use, in stably expressing HCT116-Luc cells over a range of different siRNA concentrations as indicated in the figure. (**B**) Luciferase activity of lysed HCT116-Luc tumors after in vivo treatment with P5Y/siRNA complexes. Bars represent the luciferase activities normalized for total protein concentrations from individual tumors of siCtrl-treated (blue) and siLuc-treated (red) mice.

Next, mice were injected with 5 <sup>×</sup> 106 tumor cells and, upon establishment of tumor xenografts, treated by repeated systemic application of 10 μg P5Y-complexed siRNA (3× every second day, intraperitoneal injection). Upon termination of the experiment at day 6 after treatment start, tumors were harvested and luciferase activities were analyzed from tumor lysates. Some variations were found between individual samples, especially in the negative control group. Still, the comparison of luciferase activities between the specific (P5Y/siLuc3) and the non-specific negative control (P5Y/siCtrl) group revealed a ~25% reduction of luciferase expression already after three injections (Figure 4B), and thus confirm that the treatment of mice with P5Y/siRNA complexes leads indeed to the knockdown of a given target gene in the tumors.

Switching to a more relevant therapy model, PC3 cells were s.c. injected into mice. Upon establishment of tumor xenografts of ~100 mm<sup>3</sup> in size and with clear growth kinetics, mice were randomized into different groups. Specific treatment relied on the P5Y-complexed siRNAs targeting the oncogenic proteins PLK-1 or survivin. For P10Y complexation, only PLK1 was selected as specific siRNA, while untreated mice or mice treated with siCtrl-containing complexes served as negative controls. The comparison of the latter groups revealed the absence of non-specific effects of repeated treatments (3×/week), as indicated by identical tumor growth curves (Figure 5A,B). In contrast, the treatment with the P5Y/siPLK-1 or P5Y/siSurv (Figure 5A) or P10Y/siSurv complexes (Figure 5B) led to profound antitumor effects.

During the experiment and upon its termination, mice were also screened and tested for possible adverse effects due to the repeated i.p. administration of P5Y/siRNA complexes. Mice showed no alterations in behavior, no obvious weight loss, and no other signs of adverse effects. In addition, possible toxic effects were determined by measuring important serum parameters. Healthy mice were repeatedly i.p. injected with P5Y complexes comprising 10 μg siCtrl four times over eight days. The serum analyses revealed no increase in liver enzymes (ASAT, ALAT), blood glucose (Gluc), urea, albumin, or LDH (Figure 6A). Other side effects may include the unwanted stimulation of the innate immune system. To test for this, immunocompetent mice were treated twice within 24 h with P5Y/siCtrl complexes. In this experiment, i.v. injection was preferred over i.p. administration to achieve instantaneous 100% bioavailability, and serum levels were analyzed at 4 h after the last injection, as described previously [50]. As shown in Figure 6B, no increase in TNFα or IFNγ levels were observed. We thus conclude that P5Y/siRNA complexes are efficient and safe in vitro and in vivo.

**Figure 5.** Therapy studies in PC3 tumor bearing mice using P5Y/siRNA (**A**) or P10Y/siRNA (**B**) complexes targeting either the oncogenes PLK1 or survivin. Mice were intraperitoneally treated with the complexes equivalent to 10 μg siRNA every 2–3 days. Tumor growth curves (upper panels) demonstrate the growth inhibiting effects of the specifically treated groups as compared to a negative control treated or untreated group. Lower panels: representative examples of mice.

**Figure 6.** (**A**) Determination of blood serum markers for hepatotoxicity, cardiac or skeletal muscle damage, and kidney dysfunction. Healthy mice were repeatedly i.p. injected with P5Y/siCtrl complexes and blood serum levels were compared to untreated mice. ALAT: Alanine aminotransferase; ASAT: Aspartate aminotransferase; Gluc: Glucose; Alb: Albumin; LDH: Lactate deghydrogenase. (**B**) Absence of immunostimulation of P5Y/siCtrl complexes. Complexes were i.v. injected into immunocompetent mice twice within 24 h and blood was collected 4 h after the last injection. Treatment with lipopolysaccharides (LPS) served as positive control. #, not significant.

#### **4. Discussion**

PEIs and other cationic polymers are promising with regard to nucleic acid delivery, but they are often associated with toxicity. This is particularly true for higher molecular weight PEIs, while their lower molecular weight counterparts suffer from poor efficacy [21–23]. The toxicity of high molecular weight PEIs occurs at later stages. An early toxic effect is largely due to a disturbance of the cell membrane after contact with the complexes, leading to increased LDH release and induction of necrosis [56]. The late-stage cytotoxicity was reported to be caused by a damage of the mitochondrial membrane and subsequent activation of caspases and cytochrome C release [57–59]. While the latter findings have been frequently reported, the genotoxic potential of PEI is controversially discussed. Both, the presence and absence of DNA-damaging potential has been reported [60–63] and, possibly explaining these discrepancies, may be dependent on the tested concentrations and selected cell lines.

The fact that PEIs offer relatively easy access to chemical modification is a particularly attractive approach for further improvement of physicochemical and biological properties of the polymer and its complexes. Tyrosine may be considered as among the best candidate amino acids for chemical modification, due to its ability to balance between hydrophilicity (primary amines necessary for nucleic acid complexation) and hydrophobicity (favoring cellular internalization and endosomal escape [44,46]). Many chemical modifications of low molecular weight PEIs in order to increase their bioactivity rely on the cross-linking to create larger PEI derivatives, the coupling of lipidic groups to facilitate micellar structures for improved nucleic acid binding, or modifications with cationic groups to increase the electrostatic interactions (see Introduction). The aromatic amino acid tyrosine plays a key role in protein/DNA interactions. Several studies have reported that tyrosine is essential for the binding to nucleic acids by hydrogen bonding and by π–π-stacking [64,65], which can provide an explanation for the substantial increase in very low molecular weight PEI bioactivity upon tyrosine modification. Generally, phenol-containing molecules are an interesting motif to positively affect a given nucleic acid delivery system. For example, the pre-incubation of siRNA with EGCG, a polyphenol from green tea, and subsequent complexation with low molecular weight polyamines ≤5 kDa (PEI, PLL, PAMAM) strongly increased the knockdown efficacy [22].

Serum stability is an important factor for nucleic acid delivery systems, and we demonstrate high serum compatibility of our complexes based on tyrosine-modified PEIs (see Results and [50]). This may be attributed to the hydroxyl group of tyrosine. In contrast to phenylalanine, tyrosine shows negligible interactions with bovine serum albumin (BSA), as shown previously by the incubation of thymine-dityrosine or thymine-diphenylalanine with BSA solutions [66]. Still, P5Y/siRNA complex stability was even slightly enhanced upon incubation in serum (see Figure S3C). The stabilizing effect of hydroxyl groups was demonstrated in another study, by the modification of PEI with a hydroxyl-containing cyclic carbonate (EHDO). This modification improved serum-tolerance and reduced cytotoxicity as compared to the parent 25 kDa bPEI [67]. Similar results were reported for the Tris-modification of 25 kDa bPEI for pDNA and siRNA delivery [68].

Indeed, we could demonstrate in this study that tyrosine-modification allows for using very low, otherwise inactive PEIs. High complexation efficacies even at very low mass ratios are particularly relevant in the case of small oligonucleotides like siRNAs, considering the lack of binding cooperativity and their high rigidity [69]. Our results are in line with other studies demonstrating low toxicity and high biocompatibility of other (non-PEI) tyrosine-based polymers [70], as well as the safety of tyrosine in humans [71]. The beneficial effects of tyrosine modification thus allowed for using very small amounts of (derivatized) non-toxic low molecular weight PEIs whose bioactivity and biocompatibility critically relies on the grafting with tyrosine. Optimal complex stabilities, which need to be sufficiently high for efficient siRNA complexation/protection and sufficiently low for efficient intracellular siRNA release from the complex, are additional requirements. The potential relevance of the approach of using tyrosine-modified PEIs for translation into the clinic is emphasized by the fact that these nanoparticles were found to be efficacious also in vivo.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1999-4923/11/11/600/s1, Figure S1: Degrees of tyrosine grafting and original 1H-NMR spectra, Figure S2: Additional data to Figure 1, Figure S3A,B: Cellular siRNA uptake, Figure S3C: complex stabilities, Table S1: siRNA sequences, Table S2: RT-qPCR primer sequences.

**Author Contributions:** Conceptualization, A.E. and A.A.; methodology, A.E., S.N., and M.K.; investigation, A.E., S.N., and M.K.; writing—original draft preparation, A.E. and A.A.; writing—review and editing, A.E. and A.A.; supervision, A.A.; project administration, A.A.; funding acquisition, A.A.

**Funding:** This research was funded by the Deutsche Forschungsgemeinschaft (DFG), grant number AI 24/21-1, and the Deutsche Krebshilfe (grant number 111616) to A.A.

**Acknowledgments:** The authors are grateful to Markus Böhlmann and Anne-Kathrin Krause for expert mouse maintenance, the core unit "fluorescence technologies" (Kathrin Jäger) for cell sorting and Gabi Oehme for technical assistance in tissue culture experiments. Aileen Heinze is acknowledged for proofreading the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
