Quantifying nanodiamonds biodistribution in whole cells with correlative iono-nanoscopy

Abstract

Correlative imaging and quantification of intracellular nanoparticles with the underlying ultrastructure is crucial for understanding cell-nanoparticle interactions in biological research. However, correlative nanoscale imaging of whole cells still remains a daunting challenge. Here, we report a straightforward nanoscopic approach for whole-cell correlative imaging, by simultaneous ionoluminescence and ultrastructure mapping implemented with a highly focused beam of alpha particles. We demonstrate that fluorescent nanodiamonds exhibit fast, ultrabright and stable emission upon excitation by alpha particles. Thus, by using fluorescent nanodiamonds as imaging probes, our approach enables quantification and correlative localization of single nanodiamonds within a whole cell at sub-30 nm resolution. As an application example, we show that our approach, together with Monte Carlo simulations and radiobiological experiments, can be employed to provide unique insights into the mechanisms of nanodiamond radiosensitization at the single whole-cell level. These findings may benefit clinical studies of radio-enhancement effects by nanoparticles in charged-particle cancer therapy.

Fig. 1. Ionoluminescence nanoscopy.

a Schematic illustration of the basic beam optics and experimental setup. Note that ionoluminescent photons emitted from the nanodiamonds induced by the α-beam are collected with a parabolic-mirror-based system and detected with a photomultiplier tube for ionoluminescence mapping (see Supplementary Fig. 3). b Nitrogen-vacancy (NV) center in the diamond crystalline structure. c Ionoluminescence image of the nanodiamonds through α-particle excitation. Scale bar, 2 μm. d Confocal image of the same region of the nanodiamond sample taken by using 543-nm laser excitation. Scale bar, 2 μm. e Cross-sectional line profile extracted along the arrow shown in the inset of a high-magnification image that corresponds to the region of interest marked in c, demonstrating the discrimination of two single nanodiamonds with a size of 38 and 35 nm, respectively. f Cross-sectional line profile extracted along the arrow depicted in the inset of a high-magnification image that hosts two single nanodiamonds with a separation distance approximating the resolving limit, directly indicating a sub-40 nm resolution of the ionoluminescence image.

Fig. 2. Mechanistic investigation and characterization of ionoluminescence in nanodiamonds.

a Photoluminescence (PL) spectrum of nanodiamonds excited with a 532-nm laser. b Photoluminescence spectrum of nanodiamonds excited with a 405-nm laser. c Ionoluminescence (IL) spectrum of nanodiamonds excited with a beam of 1.6 MeV α-particles. d Illustration of α-particle-induced atomic ionization in producing secondary electrons (top), through energy deposition (ΔE) of the bombarding α-particles (energy of E₀), and calculated energy distribution of the secondary electrons (bottom) in a nanodiamond. e Proposed mechanism of ionoluminescence through α-particle excitation. Process 1 represents NV-defect-assisted recombination which results in the emission of NV⁰ and NV¯. Process 2 represents interband recombination which results in the conversion of NV¯ to NV⁰ through ground-state ionization of NV¯, forming NV⁰ in its excited state of (NV⁰)*. f Time-resolved ionoluminescence measurement. Note that the instrumental response function (IRF) was determined by measuring the ionoluminescence response of a fast-decay material (Supplementary Fig. 5 and Note 3). g Relative ionoluminescence yield measurement. Note that the ionoluminescence yields of nanodiamonds, upconversion nanocrystals (NaYF₄: Yb/Tm), CdSe/ZnS quantum dots (QDs), and fluorescein dyes (FITC-1907) were normalized with a perovskite-QDs scintillator (Supplementary Figs. 6 and 7 and Note 4). h Ionoluminescence intensity profile as a function of the accumulated fluence of α-particles, showing a considerable iono-bleaching resistance of the nanodiamonds. The inserted images, taken at different time intervals (5, 40, and 75 min), indicate that the emission brightness of the nanodiamonds remains essentially unaltered over time. Scale bars are 2 μm.

Fig. 3. Correlative ultrastructure and ionoluminescence mapping towards quantitative localization of single nanodiamonds in a whole HeLa cell.

a Demonstration of correlative ultrastructure and ionoluminescence imaging of the HeLa cell with endocytosed nanodiamonds, realized in a single imaging experiment. The basic experimental design (top right) enables simultaneous acquisition of luminescence image of the nanodiamonds and 3D rendering of cell ultrastructure, by capturing α-particle-induced photons with a photomultiplier tube (PMT) and by detecting the energy loss (ΔE) of the transmitted α-particles through a Si surface barrier detector, respectively. b Overlay of the structural and luminescent images presented in a, showing localization of nanodiamonds in the cell. Scale bar, 2 μm. c Measurement of the number of intracellular nanodiamonds displayed in b in terms of their distance to the cell nuclear boundary.

Fig. 4. Evaluation of nanodiamonds as proton radiosensitizers.

a Histogram showing the distribution of nanodiamond-to-nucleus distance that was experimentally measured within 12 HepG2 cells. The magnified inset indicates that zero nanoparticles were found within 457 nm of the nuclear boundary. b Plausible mechanisms of nanodiamond-mediated proton radiosensitization in damaging nuclear DNA. Secondary electrons can be induced by the 2-MeV protons when they penetrate a nanodiamond. The secondary electrons that emanate from the nanodiamond, as well as subsequent generations of induced electrons, can either damage nuclear DNA directly or ionize intracellular water molecules in producing reactive oxygen species (ROS) to react with nuclear DNA indirectly. c Calculated energy distribution of the secondary electrons escaping the nanodiamond at its surface, simulated with Geant4-DNA. d Image showing range distributions of the secondary electrons in liquid water, simulated with Geant4-DNA. Note that we assumed the protons travel only within the nanodiamond (100 nm diameter). Scale bar, 100 nm. e Image showing range distributions of hydroxyl radicals (•OH) in liquid water, simulated with Geant4-DNA. Scale bar, 100 nm. f Measurement of the number of hydroxyl radicals and secondary electrons per proton impact as a function of their distances from the nanodiamond-surface. Note that only those traveling more than 457 nm (the minimum nanodiamond-to-nucleus distance determined in a) were taken into account.

Fig. 5. Experimental evaluation of nanodiamond radiosensitization upon proton irradiation of live HepG2 cells.

Confocal microscopic images of a, unirradiated (0 Gy) control cells without nanodiamonds (−), and b unirradiated control cells with 100 μg/ml nanodiamonds (+), in contrast to c, irradiated (2 Gy) cells without nanodiamonds, and d irradiated (2 Gy) cells with nanodiamonds. Dual-staining for γH2AX (red) and 53BP1 (green) indicates DNA double-strand breaks in the nuclei (blue). Note that the red clusters in the perinuclear regions in b and d are nanodiamonds. All scale bars in a–d are 10 μm. e Bar graph depicting red foci counts (γH2AX) per nucleus in terms of proton dose (0, 1, and 2 Gy). n = 3 biologically independent samples. f Bar graph depicting green foci counts (53BP1) per nucleus in terms of proton dose (0, 1, and 2 Gy). n = 3 biologically independent samples. Note that in e and f, each data point represents the measurement of foci counts in one 3D confocal image that contains 10–20 cells. The value of foci per nucleus was obtained by averaging total foci counts with total nuclear volume in the 3D confocal image, and subsequently normalizing with the volume of a typical nucleus (10 × 10 × 10 μm³). For each sample, 150–200 cells were analyzed. The data in e and f are shown as the mean ± s.d. Mann–Whitney test (two-tailed) was used to determine statistical significance. Foci counts between control or cells containing nanodiamonds were not statistically significantly different for γH2AX foci (p = 0.94 and p = 0.35 at 1 and 2 Gy, respectively) or 53BP1 foci (p = 0.28 and p = 0.24 at 1 and 2 Gy, respectively).