Targeting cell surface glycans with lectin-coated fluorescent nanodiamonds

Abstract

Glycosylation is arguably the most important functional post-translational modification in brain cells and abnormal cell surface glycan expression has been associated with neurological diseases and brain cancers. In this study we developed a novel method for uptake of fluorescent nanodiamonds (FND), carbon-based nanoparticles with low toxicity and easily modifiable surfaces, into brain cell subtypes by targeting their glycan receptors with carbohydrate-binding lectins. Lectins facilitated uptake of 120 nm FND with nitrogen-vacancy centers in three types of brain cells - U87-MG astrocytes, PC12 neurons and BV-2 microglia cells. The nanodiamond/lectin complexes used in this study target glycans that have been described to be altered in brain diseases including sialic acid glycans via wheat (Triticum aestivum) germ agglutinin (WGA), high mannose glycans via tomato (Lycopersicon esculentum) lectin (TL) and core fucosylated glycans via Aleuria aurantia lectin (AAL). The lectin conjugated nanodiamonds were taken up differently by the various brain cell types with fucose binding AAL/FNDs taken up preferentially by glioblastoma phenotype astrocyte cells (U87-MG), sialic acid binding WGA/FNDs by neuronal phenotype cells (PC12) and high mannose binding TL/FNDs by microglial cells (BV-2). With increasing recognition of glycans having a role in many diseases, the lectin bioconjugated nanodiamonds developed here are well suited for further investigation into theranostic applications.

Fig. 1. (A) Schematic representation of EDC/NHS chemistry to bioconjugate carboxylated FNDs to lectins using PEG (poly(ethylene glycol) 2-aminoethyl ether acetic acid). (B), (C), (D) and (E) are stained (UAR negative staining) TEM images of raw, AAL–, WGA– and TL–FNDs, respectively (no BSA added). Scale bars 100 nm. (F) Particle size distribution (PSD) by intensity of colloidal dispersion of raw and lectin bioconjugated FNDs in water after sonication. (G) Particle size distribution histogram of raw FNDs as supplied with an average size of 100 nm.

Fig. 2. (A) Confocal microscopy images of raw–FNDs (negative control) and lectin–FITC conjugated FNDs at 100× magnification. Colocalization (yellow; bottom row) is shown of FND fluorescence (red; middle row) with FITC fluorescence of conjugated lectins (green; top row) as a demonstration of successful bioconjugation of lectin to FNDs. Scale bars 5 μm. (B) The object channel (FND) was overlapped on the background channel (FITC) each with respectively red and green arbitrary colour manually creating a third channel (merge) containing the filtered voxels from the co-localized sites. The intensity from these filtered regions was used to calculate the percentage of co-localized sections (n = 3 images from different areas of samples with magnification of 100× for each condition), which is significantly higher (p < 0.0001 in all cases) than raw diamonds for AAL–FND, WGA–FND and TL–FND. (C) The concentration of protein in lectin–FND solutions measured by micro BCA assay is significantly higher for AAL–FND (p < 0.0001), WGA–FND (p < 0.0001) and TL–FND (p < 0.05) than on the surface of raw nanodiamonds obtained by paired t-tests.

Fig. 3. Colorimetric MTS cell proliferation assays (n = 5). (A), (B) and (C) cell proliferation viability following treatment with increasing concentrations of each raw or lectin bioconjugated FNDs in glioblastoma astrocytes (U87-MG), neuronal phenotype cells (PC12), and microglial cells (BV-2). Absorbance readings were normalized to the mean values obtained for untreated cells. * for p ≤ 0.05, ** for p ≤ 0.01, obtained by non-parametric Kolmogorov–Smirnov tests.

Fig. 4. (A). Uptake of raw or lectin bioconjugated FNDs by brain cell subtypes. Confocal microscopy of U87-MG (left column), PC12 (middle column) and BV-2 (right column) cells treated with raw FNDs (first row), AAL–FNDs (second row), WGA–FNDs (third row) and TL–FNDs (fourth row). ActinGreen stain for f-actin filaments (green – ex 488 nm; em 500–550 nm), NucBlue stain of nuclei (blue) and FND fluorescence (red – ex 561 nm; em 650–750 nm) were observed. Inset zoomed areas selected from representative cells highlight FND uptake (indicated by arrows). Scale bars = 20 μm. (B) Quantification of raw and lectin bioconjugated FND uptake by U87-MG (blue), PC12 (red) and BV-2 cells (green) by fluorescence intensity of nanodiamonds per cell (n = 30 cells). The black line represents the mean fluorescent intensity bar for each column, (a.u.) stands for arbitrary intensity unit. **** for p ≤ 0.0001 and not significant for p ≥ 0.05, obtained from non-parametric Kolmogorov–Smirnov tests.

Fig. 5. (A) Untreated U87-MG MUSE flow cytometry cell viability test at 4 °C after 4 hours and 8 hours of incubation, showing 70% of U87-MG cells remain viable in this low temperature (n = 3 dishes of cultured cells per time point). (B) Comparison of raw and lectin bioconjugated FND (AAL–FND, WGA–FND, TL–FND) uptake by U87-MG astrocyte cells (n = 30) following 8 hours of incubation at 4 °C or 37 °C. Imaris software was used to calculate the fluorescent intensity of the mean fluorescent intensity of FNDs of each randomly selected and manually contoured cell. p ≤ 0.0001, obtained from non-parametric Kolmogorov–Smirnov tests.