br Statistical analysis br Differences between
2.8. Statistical analysis
Differences between groups were considered to be significant at a P value of < 0.05. Statistical analyses (One-way ANOVA) were performed with GraphPad Prism 7.0 (GraphPad Software, Inc., San Diego, CA).
Effect of storage in water on the size and charge of nanoparticles with variable polycation:HA weight ratio.
Polycation:HA a Size (nm) b
ζ-potential (mV) b
As prepared Stored c As prepared Stored c
a Please note that siRNA was always used at a ratio 2.3%wt. in respect to the polycation weight, hence nanoparticles were prepared also in the absence of HA. b Concentration: 1 mg/mL in deionized water c Storage: 1 week, 4 °C in deionized water.
3. Results and discussion
3.1. Nanoparticles characterization
3.1.1. Nanoparticle physical stability
As prepared, all NP showed a rather comparable hydrodynamic size, although typically lower for Nanocin (Table 1, Fig. 1a and b). This is likely due to the higher charge density of this polymer, which allows a higher ionic cross-link density and therefore also a lower water content in the particles. As expected, the ζ-potential (surface charge) depended on the polycation:HA weight ratio, and shifted from positive to negative values with increasing HA content. With both polycations, a ‘charge inversion’ occurred at a polycation:HA weight ratio between 1:1 and 1:2; these samples also exhibited the largest sizes. At stoichiometry ratios close to the effective complexation between positive and negative charges, polyelectrolyte complexes form and keep aggregating due to the reduced repulsion, until a size is reached. At this equilibrium, a small imbalance in either of the charged components leads to a suffi-cient surface coverage to grant electrostatic stabilization. In deionized water, chitosan-based NP showed a mild agglomeration upon storage (size variation less than 5%), independent of the amount of HA. On the contrary, Nanocin ML385 depended on HA: Better stability at high HA content which decreased at low (polycation:HA < 1:2), and showed significant increases in size and ζ-potential post storage. This effect is probably due to the strong interactions between the high-cationic density, small-size Nanocin and HA, which significantly reduce the electrostatic stabilization due to excess (uncomplexed) negative charges of HA. Due to their better stability, all further experiments were con-ducted with a 1:4 polycation/HA ratio.
3.1.2. Protection of siRNA against RNase
We have assessed the stability of siRNA in nanoparticles (siRNA:polycation 1:4 wt ratio), when they were exposed to different concentrations of RNase I (Fig. 1e). siRNA in solution was already partially degraded at 0.01 U/μL RNase (∼10% degradation), with > 50% degradation at 0.1 U/μL RNase, whereas siRNA liberated from nanoparticles was intact even after exposure to 1 U/μL RNase, with no statistically relevant difference with the two polycations and with/out HA.
3.2. Targeting and internalization of HA-coated nanoparticles
Colorectal cancers have a high incidence of mutation of KRAS, and International Journal of Pharmaceutics 561 (2019) 114–123
we have therefore chosen a model of human colorectal cancer, HCT-116 cells, known to present KRASG13D mutation (Alves et al., 2015). HCT-
116 also have high CD44 expression (Rios de la Rosa et al., 2017b), and unsurprisingly this is higher than in the cells (HDFa) which we used as a model for stromal component of the tumor microenvironment (Fig. 2a). Importantly, HCT-116 cells were positive to CD44 variants commonly
3.2.2. Nanoparticle internalization: mono-culture vs. co-culture
We have followed the kinetics of both NP internalization via flow cytometry using a fluorescently labelled siRNA (L3-DY547-NP) for 24 h. NP internalization was firstly investigated on cancer cells and fibro-blasts in mono-culture, then cells were co-cultured. In mono-culture of both cell types, chitosan/HA NP showed a more rapid internalization, however eventually reaching a plateau, as already seen in previous works (Lallana et al., 2017; Rios de la Rosa et al., 2017a). At the later time points the fluorescence intensities produced by the siRNA were comparable in both carriers (Fig. 3). Qualitatively, in co-culture we observed a similar kinetic behavior, i.e. an earlier plateau for chitosan/ HA. However, assuming that the siRNA fluorescence always provides quantitative and comparable estimates of NP internalization, we no-ticed another more interesting phenomenon: in mono-culture the up-take in HDFa and HCT-116 appeared to be comparable (Fig. 3a) be-tween the tested NP, whilst in co-culture HCT-116 always internalized NP in much larger amounts (Fig. 3b). Of note, where internalization in HCT-116 always increased in co-culture, the internalization in HDFa increased for Nanocin/HA and decreased for chitosan/HA, which means that although this HCT-116/HDFa differential internalization (and targeting) was clear for both NP, it was much larger with chitosan/ HA systems. Mechanistically, this may be due an increased expression of CD44 and/or its variants: we have demonstrated that the overall CD44 expression depends on the HCT-116 environment (3D culture increasing it) (Rios de la Rosa et al., 2018a,b), whereas interactions with cancer associated fibroblasts have been reported to increase CD44v expression in colon cancer cells (Misra et al., 2011). Ther-apeutically, this is a very promising result, which indicates that HA-coated formulations may be able to preferentially target and treat po-pulations of tumoral cells, thereby effectively reducing potential off-target effects. Additionally, preliminary in vitro toxicology experiments did not indicate increased cytotoxicity (see Supplementary Information, section SI.10 and Fig. 5SI) nor increased levels of cellular stress (see Supplementary Information, section SI.11 and Fig. 6SI) following ex-posure over a range of nanoparticle concentrations. These results in-dicate that both nanoparticles are suitable for further in vivo studies.