In Vitro and in Vivo Analyses of the Effects of Source, Length, and Charge on the Cytotoxicity and Immunocompatibility of Cellulose Nanocrystals

Abstract

Cellulose nanocrystals (CNCs) are an emergent, sustainable nanomaterial that are biosourced, abundant, and biodegradable. On account of their high aspect ratio, low density, and mechanical rigidity, they have been employed in numerous areas of biomedical research including as reinforcing materials for bone or tissue scaffolds or as carriers in drug delivery systems. Given the promise of these materials for such use, characterizing and understanding their interactions with biological systems is an important step to prevent toxicity or inflammation. Reported herein are studies aimed at exploring the in vitro and in vivo effects that the source, length, and charge of the CNCs have on cytotoxicity and immune response. CNCs from four different biosources (cotton, wood, Miscanthus x Giganteus, and sea tunicate) were prepared and functionalized with positive or negative charges to obtain a small library of CNCs with a range of dimensions and surface charge. A method to remove endotoxic or other impurities on the CNC surface leftover from the isolation process was developed, and the biocompatibility of the CNCs was subsequently assayed in vitro and in vivo. After subcutaneous injection, it was found that unfunctionalized (uncharged) CNCs form aggregates at the site of injection, inducing splenomegaly and neutrophil infiltration, while charged CNCs having surface carboxylates, sulfate half-esters, or primary amines were biologically inert. No effect of the particle source or length was observed in the in vitro and in vivo studies conducted. The lack of an in vitro or in vivo immune response toward charged CNCs in these experiments supports their use in future biological studies.

Keywords: biocompatibility; biomaterials; cellulose nanocrystals; nanomedicine; tissue scaffolds.

Figure 1.

Overview of CNC synthesis. CNCs were obtained from tunicate, MxG, cotton, and wood to allow for the study of CNCs with four different lengths. CNCs were then functionalized using epichlorohydrin-mediated amination or TEMPO oxidation to achieve CNCs with positive and negative charge modifications, respectively, on their surface.

Figure 2.

Microscopy of CNCs used in this study. (A) AFM image of t-CNC–OSO₃⁻. (B) AFM image of MxG-CNC–OSO₃⁻. (C) TEM image of MxG-CNC–OH. (D) AFM image of c-CNC–OSO₃⁻. (E) TEM image of MxG-CNC–OH_(ls). (F) AFM image of w-CNC–OSO₃⁻_(ls). For AFM imaging, samples were drop-cast from a dilute aqueous solution onto a poly-lysine coated mica substrate, and for TEM imaging, samples were casted onto a carbon-coated copper grid and stained with uranyl acetate.

Figure 3.

Endotoxin removal from CNC samples was validated using the HEK mTLR4 reporter assay (n = 3/group). CNC samples were washed three times with endotoxin-free dH₂O by subsequent sonication and ultracentrifugation. The cleaned samples were freeze-dried, redissolved in endotoxin-free dH₂O, and incubated with cells at 10–100 μg/mL overnight. HEK mTLR4 activity was indicated using HEK-Blue detection medium (InvivoGen), and the absorbance was quantified at 620 nm. LPS was used as a positive control, and an uncleaned sample (far left in red) with low-level endotoxin contamination is shown as a reference. Error bars indicate standard deviation.

Figure 4.

In vitro assays assessing biocompatibility of the prepared CNCs (n = 3/group). CNCs were incubated with the RAW-Blue macrophage cell line at 10–100 μg/mL for 24 h, and (A) immune system activity and (B) cytotoxicity were assayed using RAW-Blue and MTT Assays, respectively. No aberrant immune system activity or cytotoxicity were observed using CNCs from any source or functionalization. Error bars indicate standard deviation.

Figure 5.

(A) In vivo experimental design to test inflammation induced by the CNCs (n = 3–5/group). No changes in (B) TNF-α or (C) IL-6 were observed in the CNC samples at either concentration or time point analyzed. (D) Mice injected with 0.5%MxG-CNC–OH had significant splenomegaly after 2 d, suggesting an acute immune response toward the CNCs. Error bars indicate standard deviation. Significant differences between treatments were analyzed by two-way ANOVA with residual multiple comparisons testing. **p < 0.01.

Figure 6.

Hematoxylin and eosin staining of tissue collected from the injection sites of uncharged (A) MxG-CNC–OH and negatively charged (B) MxG-CNC–COOH 2 d after injection. Significant neutrophil infiltration and the presence of crystallite aggregates was observed in the MxG-CNC–OH samples, suggesting a pro-inflammatory immune response.