Nano assessing nano: Nanosensor-enabled detection of cell phenotypic changes identifies nanoparticle toxicological effects at ultra-low exposure levels

Abstract

Industrial use of nanomaterials is rapidly increasing, making the effects of these materials on the environment and human health of critical concern. Standard nanotoxicity evaluation methods rely on detecting cell death or major dysfunction, and will miss early signs of toxicity. In this work, we report the use of rapid and sensitive nanosensors that can efficiently detect subtle phenotypic changes on the cell surface following nanomaterial exposure. Importantly, our method revealed significant phenotypic changes at dosages where other conventional methods show normal cellular activity. This approach holds promise in toxicological and pharmacological evaluations to ensure safer and better use of nanomaterials.

Graphical Abstract

Phenotypic changes of cells as a result of low exposure to a series of gold nanoparticles were efficiently detected by the nanosensor.

Nanomaterial technologies are widely used today in multiple aspects of biomedicine, including tissue engineering,^[1,2] diagnostic techniques,^[3] drug delivery,^[4] and biosensing.^[5] Concomitantly, nanomaterials are also being increasingly incorporated into consumer products, leading to increased release of nanomaterials into the environment, making human exposure a major health concern.^[6]

Nanotoxicity, as in toxicology in general, relies predominantly on detecting substantial cellular changes, such as damage to the cell membrane integrity, ^[7] effects on metabolic activities,^[8] and genetic alterations.^[9] However, subtle phenotypic changes or slight cellular abnormalities can be an early indication of potential toxicity, at a threshold below what is detectable using standard methods. Recent studies demonstrated that multiple cellular pathways and processes are affected by low dosage of metal-based nanoparticles.^[10,11] As an example, several studies have reported significant changes in microviscosity, morphology, and cytoskeletal network as a result of nanomaterial exposure.^{[12,13,14,15]} Clearly, detection of early signs of nanotoxicity is of great importance to human health.

Hypothesis-free sensing provides a promising way to assess the effects arising from exposure to low dosages of nanomaterial. This approach detects changes without bias, making them ideal for detection of low-dose toxicological effects. Nanosensors are powerful tools in for applying this detection strategy to identifying subtle phenotypic changes on cell surfaces. This array-based nanosensing strategy works through selective interactions between the sensor elements and cell surface functionalities, avoiding the limitations of current endpoint evaluation methods.^[16,17,18] Another advantage of using a nanosensor approach lies in its highly sensitive nature; the small size and high surface to volume ratio of gold nanoparticles (AuNPs) enable the nanosensor to have multiple interaction sites with target analytes. Successful examples of hypothesis-free nanosensing include the detection of phenotypic changes in mammalian cells, ^[19,20] bacteria, ^[21] and bacteria inbiofilms.^[22]

We hypothesized that array-based nanosensing could be used to determine the level of nanomaterial required to generate cellular changes, and important question in assessing the health effects of nanomaterial exposure. To this end, we employed a robust nanosensor composed of AuNP and efficient green fluorescent protein (EGFP) to rapidly detect early signs of abnormality in cells under low exposure to nanomaterials. To provide a rigorous testing of the sensor system we employed AuNPs as model toxicological agents. Since surface functionality is a strong indicator of nanotoxicity, ^[23,24] we selected a series of AuNPs with varying degrees of hydrophobicity. By comparing the sensing results observed using this sensor array with commonly used toxicity evaluation methods, we observed significant phenotypic alterations at a nanomolar dosage range, where traditional methods showed normal cellular activity. Our results demonstrate a complex dose-response sensing curve that indicates multiple cellular processes occurred upon exposure to nanomaterials. This observation presents an opportunity to apply hypothesis-free sensing as the frontier checkpoint in toxicological and pharmacological evaluations to ensure safer and better use of nanomaterials.

The nanosensor employed in this work was comprised of a positively charged AuNP with a benzyl terminal group (BzNP) supramolecularly complexed with a negatively charged EGFP. BzNP was selected due to its high sensitivity in interacting with cell surface. ^[25] Through electrostatic interactions, the fluorescence of EGFP is initially quenched. Upon interacting with cells, the competitive binding between BzNP and cell surface functionalities releases EGFP into solution, restoring the fluorescence. Changes in the fluorescence are hence correlated with changes in the cell surface (Figure 1a). In practice, a suitable sensor concentration of 150 nM EGFP and 100 nM BzNP was selected from the fluorescence titration curve, a value where EGFP intensity is efficiently quenched (Figure S1).

Figure 1.

a) Fabrication of the nanosensor. The sensor is formed by incubating BzNP with EGFP at concentrations determined through fluorescent titration. The association constant between BzNP and EGF, K_a = 1.52 × 10⁷ M⁻¹ was calculated using the non-linear least-squares curve fitting analysis (Figure S1). b) Chemical structure of the R groups of AuNPs used to treat cells. c) Schematic illustration of the workflow. Cells were first stimulated with C2–C10 NPs for 48 h and then subjected to analysis along with three types of commonly used cytotoxicity assays for comparison.

A library of four positively charged AuNPs engineered to vary only in hydrophobicity were used as model (Figure 1b). Hydrodynamic size, charge, and physical properties of these NPs were validated through DLS, zeta potential, and TEM (Figure S2–5). A conceptual scheme of our work is depicted in Figure 1c, where cells were first treated with C2–C10 NPs at a low concentration ranging from 5 nM to 75 nM for 48 h. Next, the treated cells were assayed using the nanosensor as well as three frequently used types of cytotoxicity assays.

For this study we chose a non-malignant human mammary epithelial cell line MCF10A as our model to mimic human exposure to nanomaterials. At the end of AuNP treatment, an equal number of 10⁴ cells were added to the nanosensor and the fluorescence signals were measured. As shown in Figure 2a, a trend of an increasing sensor response was observed as the concentration of C2 NP increases. However, we also noted a slight change in cell number, reflected in the signal of DNA binding dye Hoechst 33342. The EGFP signal was hence normalized to the Hoechst signal (Figure 2b). After this normalization, a significant change in sensor response was observed even at the lowest concentration of 5 nM, indicating phenotypic changes indeed occurred on cell surface. The sensing trend was different for each tested AuNP (Figure S6). The complex dose-response curve suggests that multiple cellular processes were affected by the nanomaterials.

Figure 2.

Representative sensor output of MCF10A cells after exposing to C2 NP for 48 h. a) Fluorescence intensities of EGFP/AuNP nanosensor (green) and Hoechst 33342 dye (orange) were normalized to sensor only without any cells. Each data point is the average of six replicates. b) Additional normalization of sensor signal to Hoechst dye to ensure equal number of cells were being sensed. Curve was plotted to lead the eye.

Next, we compared the sensing results with three commonly used cytotoxicity assays in which cell membrane integrity (Trypan Blue exclusion assay), mitochondrial metabolism (AlamarBlue assay), and cell proliferation (DNA staining Hoechst dye) were evaluated. For comparison purpose, each assay was reported at the scale of 1 and all treatment groups were normalized to cell only group. Membrane integrity was consistent throughout all tested concentrations of the four NPs while mitochondrial activities showed a slight increase at 5 nM for C4, C6 and C10 NPs (Figure 3). This phenomenon could be explained by hormesis effect where cells simultaneously increase their performance and defense mechanism in response to challenges in their environment. ^[26,27] As concentration increases, cell viability stayed consistent. Similarly, Hoechst dye stained DNA content of all tested NPs show normal or slight increase suggesting these NPs did not induce dramatic changes in cell proliferation. However, nanosensor signals in all tested concentrations deviated significantly from the control, indicating phenotypic alternations had occurred as a result of low NP exposure. Our previous results show that the nanosensor is sensitive to changes in glycosylation, ^[28] suggesting the composition of glycoproteins on cell surface is likely being altered under the tested conditions. This is both an exciting and alarming finding since the nanosensor detected abnormal changes on cell surface whereas other methods show normal cellular activity. It raises the concern of potential false negatives at the exposure concentration when precautions should be taken into consideration.

Figure 3.

Heatmap of response from nanosensor and other cytotoxicity assays when cells were exposed to low dosages of C2–C10 NPs. These plots show strong response from the nanosensor and weak or null responses from the other methods. For comparison purpose, each treatment group was normalized to the cell only group.

In conclusion, we have demonstrated a simple and robust method that rapidly detects subtle phenotypic changes in cells arising from low exposure to nanomaterials. Since the nanosensor works through a hypothesis-free strategy, any changes on the cell surface functionalities could interact with the nanosensor, generating a signal. By comparing the signal from NP treated cells to control, we observed significant changes at the nanoparticle exposure concentration where other standard methods report normal cell behavior. These results demonstrate a serious shortcoming in other contemporary methods for evaluating nanomaterial cytotoxicity, especially at low dosages. We believe that nanosensor-enabled hypothesis-free sensing will provide a tool complementary to current nanotoxicology methods, enabling better evaluation of the safety of nanomaterials.