Pitfalls in methods to study colocalization of nanoparticles in mouse macrophage lysosomes

Abstract

Background: In the field of nanoscience there is an increasing interest to follow dynamics of nanoparticles (NP) in cells with an emphasis on endo-lysosomal pathways and long-term NP fate. During our research on this topic, we encountered several pitfalls, which can bias the experimental outcome. We address some of these pitfalls and suggest possible solutions. The accuracy of fluorescence microscopy methods has an important role in obtaining insights into NP interactions with lysosomes at the single cell level including quantification of NP uptake in a specific cell type.

Methods: Here we use J774A.1 cells as a model for professional phagocytes. We expose them to fluorescently-labelled amorphous silica NP with different sizes and quantify the colocalization of fluorescently-labelled NP with lysosomes over time. We focus on confocal laser scanning microscopy (CLSM) to obtain 3D spatial information and follow live cell imaging to study NP colocalization with lysosomes.

Results: We evaluate different experimental parameters that can bias the colocalization coefficients (i.e., Pearson's and Manders'), such as the interference of phenol red in the cell culture medium with the fluorescence intensity and image post-processing (effect of spatial resolution, optical slice thickness, pixel saturation and bit depth). Additionally, we determine the correlation coefficients for NP entering the lysosomes under four different experimental set-ups. First, we found out that not only Pearson's, but also Manders' correlation coefficient should be considered in lysosome-NP colocalization studies; second, there is a difference in NP colocalization when using NP of different sizes and fluorescence dyes and last, the correlation coefficients might change depending on live-cell and fixed-cell imaging set-up.

Conclusions: The results summarize detailed steps and recommendations for the experimental design, staining, sample preparation and imaging to improve the reproducibility of colocalization studies between the NP and lysosomes.

Keywords: Colocalization; LysoTracker probes; Lysosomes; Macrophages; Nanoparticles; Pitfalls.

Fig. 1

Comparison of the normalized raw integrated densities of three different LysoTracker probes (red, blue and green) obtained by cell imaging in cRPMI with phenol red (white bars) and without phenol red (gray bars). Error bars are standard deviation, n = 10 cells per treatment. *p < 0.05 by one-way ANOVA

Fig. 2

Stability of LysoTracker Red probe under different imaging conditions. A Cell viability after incubation with LysoTracker Red probe for 1 h, 4 h and 24 h assessed via membrane rupture assay (lactate dehydrogenase; LDH) and presented as fold change over untreated cells (control). Data shows the mean of three replicates. Cells exposed to 0.2% Triton X-100 (v/v) served as positive control for LDH assay. B Schematic representation of staggered experimental set-up (cells were stained for 1 h with LysoTracker Red probe, the supernatant containing LysoTracker Red was removed and cells were further incubated and imaged at 1 h, 4 h and 24 h as hyperstacks) and continuous set-up (cells were stained for 1 h with LysoTracker Red probe and imaged continuously for 24 h as hyperstacks). C Comparison of raw integrated densities of LysoTracker Red probe versus time for continuous and staggered imaging. Each time-point (hour) shows the raw integrated density measured for each cell individually. The whiskers are the standard deviations. **** (P < 0.05) is the significant difference analyzed by One-way ANOVA

Fig. 3

Effect of spatial resolution, optical slice thickness, pixel saturation and bit depth on the Pearson’s correlation coefficients (PCC, white boxes) and Manders’ M1 coefficient (gray boxes). The data correspond to the experimental set-up where cells were exposed to 59 nm SiO₂-BDP FL NP for 13 h, fixed and stained with Alexa Fluor 647-conjugated LAMP-2 antibody. The error bars represent the standard deviations. All the images were analysed using ImageJ with the JACoP plugin comparing the correlation coefficients after each specific simulation

Fig. 4

The effect of continuous live cell imaging on NP colocalization with lysosomes. A Experimental workflow: Cells were first stained with LysoTracker Red probe for 1 h, and then exposed to 20 μg/mL of 59 nm SiO₂-BDP FL NP in phenol free cRPMI. Cells were imaged continuously from 4 to 8 h. All the images were analyzed using ImageJ with a script (Script S2). Each cell was analyzed individually. B Pearson’s correlation coefficient (PCC) demonstrating correlation between NP and LysoTracker Red probe (lysosomes) at different time points. C Manders’ correlation coefficient M1 corresponds to the fraction of 59 nm SiO₂-BDP FL NP within lysosomes (stained with LysoTracker Red) and M2 corresponds to the fraction of lysosomes filled with NP. The corresponding histograms of fluorescence intensities of each channel are shown in Additional file 1: Fig. S5A

Fig. 5

Colocalization of different-sized SiO₂ particles with lysosomes in J774A.1 macrophages. A Experimental workflow: Cells were exposed to SiO₂ particles of different sizes for 24 h followed by LysoTracker Red staining for 1 h and imaged live. B Representative bright field and confocal images of single cells with the corresponding cytofluorograms show colocalization of 119 nm SiO₂-Cy5 particles (green) with lysosomes (red). C Colocalization of 920 nm SiO₂-Cy5 particles (green) with lysosomes (red). Co-localized pixels appear in yellow. Scale bar: 10 µm. The corresponding histograms of fluorescence intensities of each channel is shown in Additional file 1: Fig. S5A and the bright field images identifying cell morphology in Additional file 1: Figure S6. Quantification of colocalization of D. 119 nm SiO₂-Cy5 particles and E 920 nm SiO₂-Cy5 particles with LysoTracker Red probe was performed using the ImageJ software with JACoP plugin (n = 10 cells). Data is presented as Pearson’s correlation coefficient (PCC) and Manders’ overlap coefficients, where M1 represents the fraction of particles-associated pixels overlapping with lysosomes and M2 the fraction of lysosome-associated pixels overlapping with particles. Statistical analysis was performed using unpaired t-test in GraphPad Prism software: ***p < 0.001, ns: not significant

Fig. 6

Colocalization between NP with different fluorophores and lysosomes. A Experimental workflow. Representative images of single cells exposed to two different SiO₂ NP (green), B SiO₂-RhoB NP and C SiO₂-BDP FL NP. Lysosomes were stained with LAMP-2 antibody (shown in red). Scale bar: 10 µm. The corresponding histograms of fluorescence intensities of each channel are shown in Additional file 1: Fig. S5B and the F-actin staining identifying cell morphology in Additional file 1: Fig. S6. The plots represent main parameters (PCC, M1 and M2) to estimate colocalization between different NP and lysosomes. The colocalization between the D 59 nm SiO₂-RhoB NP and E 59 nm SiO₂-BDP FL NP with lysosomes, labelled with LAMP-2 antibody. PCC: Pearson correlation coefficient, M1 and M2: Manders’ correlation coefficients. The whiskers represent the standard deviations. Analysis was performed in individual cells for each experiment (n = 10–13 cells). All the images were analysed using ImageJ, with the JACoP plugin comparing the correlation coefficients. Statistical analysis was performed using unpaired t-test in GraphPad Prism software. * p < 0.05, ns: not significant

Fig. 7

NP colocalization with lysosomes between fixed-and live-cell imaging. A Experimental workflow. Representative images of a single cell exposed to two different NP types, B SiO₂-RhoB NP (green) and C SiO₂-BDP FL NP (green). Lysosomes were stained with LAMP-2 antibody (red). The corresponding histograms of fluorescence intensities of each channel are shown in Additional file 1: Fig. S5B. The graphs depict the main parameters (Pearson’s correlation coefficient—PCC, Manders’ correlation coefficients—M1 and M2) to estimate the colocalization between D SiO₂-RhoB NP and lysosomes (LysoTracker Green) and E SiO₂-BDP FL NP and lysosomes (LysoTracker Red) in live cells (white box) and fixed cells stained with LAMP-2 (grey box). The whiskers are the standard deviations. The analysis was done in individual cells for each experiment (n = 8–13 cells). All the images were analysed using ImageJ with the JACoP plugin comparing the correlation coefficients. Statistical analysis was performed using unpaired t-test in GraphPad Prism software. ***p < 0.001, ns: not significant