Influence of shear stress and size on viability of endothelial cells exposed to gold nanoparticles

Screening nanoparticle toxicity directly on cell culture can be a fast and cheap technique. Nevertheless, to obtain results in accordance with those observed in live animals, the conditions in which cells are cultivated should resemble the one encountered in live systems. Microfluidic devices offer the possibility to satisfy this requirement, in particular with endothelial cell lines, because they are capable to reproduce the flowing media and shear stress experienced by these cell lines in vivo. In this work, we exploit a microfluidic device to observe how human umbilical vein endothelial cells (HUVEC) viability changes when subject to a continuous flow of culture medium, in which spherical citrate-stabilized gold nanoparticles of different sizes and at varying doses are investigated. For comparison, the same experiments are also run in multiwells where the cells do not experience the shear stress induced by the flowing medium. We discuss the results considering the influence of mode of exposure and nanoparticle size (24 and 13 nm). We observed that gold nanoparticles show a lower toxicity under flow conditions with respect to static and the HUVEC viability decreases as the nanoparticle surface area per unit volume increases, regardless of size. Electronic supplementary material The online version of this article (10.1007/s11051-017-3993-5) contains supplementary material, which is available to authorized users.


SUPPLEMENTARY MATERIAL
High Resolution TEM of the gold nanoparticles Figure S1: TEM image of Au NPs in bright field mode (on the left) and histogram of the size distribution (on the right) of Batch 24 nm.

Determination of hydrodynamic radius by means of Fluorescence Correlation Spectroscopy
Fluorescence Correlation Spectroscopy records the fluctuations of fluorescence intensity emitted by a few fluorescent probes inside the focal volume sampled by the microscope objective in a confocal fluorescence microscope. Two independent avalanche photodiodes record simultaneously these fluctuations, and then the signals are cross-correlated to obtain the autocorrelation curve for fluorescence fluctuation, which is the FCS signal. When the fluorescent probes are subject to freediffusion, the FCS curve is fitted using the following formula, containing both rotational and The parameters are: the rotation contrast C, the characteristic rotational time τ rot , the particle number N in the focal volume, the time τ trasl employed by the tracer to pass through the focal volume through a process of translational diffusion, and S the focal shape-factor.
Adopting a classical hydrodynamic description, τ rot and τ trasl are proportional to the volume and the radius of the nanoparticle, respectively (Sauer et al. 2011): where η is the water viscosity, K B is the Boltzman constant, T is the temperature, w 0 is the radial focal volume radius, and R h is the NP hydrodynamic radius. Assuming spherical shaped Au NPs, an estimate of the average Au NP hydrodynamic radius is calculated from extrapolated τ rot values.
The main advantage of using this parameter instead of τ trasl for the determination of the hydrodynamic size is its independence with respect to the dimension of the focal volume. Figure S2 shows a typical FCS curve recorded for Au NP in water.

FLIM: Fluorescence Lifetime Imaging Experiments
In this technique the image maps the fluorescence lifetime of the sample under investigation.
For the preparation of FLIM samples, HUVEC are seeded on a glass coverslip, incubated for 24h with different Au NPs concentrations, and fixed in formaldehyde 4% in PBS 1X. For FLIM analysis the fs laser beam (820 nm and 10 mW average power) is focused on the sample through a 60× water immersion microscope objective. The emission signal, filtered by a 750nm-shortpass and a 572/35 bandpass filters is sent to the avalanche photodiode connected to PicoHarp300 from PicoQuant. The scan area is 256×256 px. The resulted FLIM data are analyzed with the Symphotime software (PicoQuant) to generate the fluorescence lifetime map. The fluorescence decay curve from the overall map is fitted with a double-exponential model using the formula: F(t) = A1×exp(t/τ1)+A2×exp(-t/τ2), where τ1 and τ2 are the fluorescence lifetimes. Figure S3 shows the fluorescence decay curves for Au NP in water, HUVEC only and HUVEC together with Au NP.  collected at 3 µm from the coverglass. In FLIM image, the red signal is the amplitude of short lifetime contribution (assigned to Au NP emission) and the blue signal is the amplitude of long fluorescence lifetime term (assigned to residual cell autofluorescence).
Specimens are observed in a Hitachi H-300 Electron Microscope. Figure S5: Optical microscopy images of HUVEC treated for 24 h in microfluidic device (left) and in multiwells (right) with the lower concentration of Au NP (4.3 ×10 11 NPs/mL).