(HepG2) Cells Exposed to Subtoxic Levels of ... - ACS Publications

Mar 2, 2018 - NMR Metabolomics Reveals Metabolism-Mediated Protective Effects in Liver (HepG2) Cells Exposed to Subtoxic Levels of Silver Nanoparticle...
0 downloads 0 Views 2MB Size
Subscriber access provided by Kaohsiung Medical University

Article

NMR metabolomics reveals metabolism-mediated protective effects in liver (HepG2) cells exposed to sub-toxic levels of silver nanoparticles Joana Carrola, Ricardo J. B. Pinto, Maryam Nasirpour, Carmen Sofia Rocha Freire, Ana M Gil, Conceição Santos, Helena Oliveira, and Iola F Duarte J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.7b00905 • Publication Date (Web): 02 Mar 2018 Downloaded from http://pubs.acs.org on March 2, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

NMR metabolomics reveals metabolism-mediated protective effects in liver (HepG2) cells exposed to sub-toxic levels of silver nanoparticles

Joana Carrola†, Ricardo J.B. Pinto†, Maryam Nasirpour†, Carmen S.R. Freire†, Ana M. Gil †, Conceição Santos‡, Helena Oliveira†,§, Iola F. Duarte*, †



CICECO – Aveiro Institute of Materials, Department of Chemistry, University of

Aveiro, 3810-193 Aveiro, Portugal. ‡

Department of Biology, Faculty of Sciences, University of Porto, 4169-007 Porto

Portugal. §

CESAM & Laboratory of Biotechnology and Cytomics, Department of Biology,

University of Aveiro, 3810-193 Aveiro, Portugal.

*

Corresponding author: Tel: +351 234401424; Fax: +351 234401470; Email:

[email protected]

1 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 39

ABSTRACT The expansion of biomedical and therapeutic applications of silver nanoparticles (AgNPs) raises the need to further understand their biological effects on human cells. In this work, NMR metabolomics has been applied to reveal the metabolic effects of AgNPs towards human hepatoma (HepG2) cells, which are relevant with respect to nanoparticle accumulation and detoxification. Cellular responses to widely disseminated citrate-coated AgNPs (Cit30) and to emergent biogenic AgNPs prepared using an aqueous plant extract as reducing and stabilizing agent (GS30) have been compared with a view to assess the influence of nanoparticle coating on the metabolic effects produced. Sub-toxic concentrations (IC5 and IC20) of both nanoparticle types caused profound changes in the cellular metabolome, suggesting adaptations in energy production processes (glucose metabolism and the phosphocreatine system), antioxidant defenses, protein degradation and lipid metabolism. These signatures were proposed to reflect, mainly, metabolism-mediated protective mechanisms, and found to be largely common to Cit30 and GS30 AgNPs, although differences in the magnitude of response, not captured by conventional cytotoxicity assessment, were detected. Overall, this study highlights the value of NMR metabolomics for revealing sub-toxic biological effects and helping to understand cell-nanomaterial interactions.

KEYWORDS: Liver (HepG2) cells, silver nanoparticles (AgNPs), ionic silver, NMR metabolomics, cell metabolism

2 ACS Paragon Plus Environment

Page 3 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

INTRODUCTION Silver nanoparticles (AgNPs) exhibit a multitude of interesting biological properties, from well-documented antibacterial action1 to recently reported antiviral and anticancer activities.2,3 Thus, in addition to their established use in consumer goods (e.g. cosmetics, textiles) and biomedical products (e.g. wound dressings), the range of possible therapeutic applications of AgNPs is likely to be expanded. Biogenic AgNPs synthetized through eco-friendly biological methods, using for example plant extracts, emerge as particularly promising, as they can potentially take advantage from the synergistic effects of silver and the natural compounds used in nanoparticle synthesis and stabilization. At the same time, given the high propensity for human exposure, there is a recognized need to better define the potential adverse effects of AgNPs on human health, and to understand their mechanisms of action in eukaryotic cells and higher organisms. While many reports have addressed the acute toxic effects of AgNPs towards various cultured cells, there is still insufficient knowledge on possible sub-toxic alterations of cell function, more representative of real exposure scenarios. Untargeted omic technologies, namely genomics, transcriptomics, proteomics and metabolomics, constitute a valuable approach in this respect, as they offer the possibility to detect unsuspected subtle changes in genes, transcripts, proteins or metabolites before cytotoxic events become detectable by conventional methods.4 Assessing the metabolic changes induced by nanomaterials on cellular or animal models, through Nuclear Magnetic Resonance (NMR) or Mass Spectrometry (MS)-based metabolomics, is increasingly proving valuable for revealing unforeseen biological effects and helping to understand the mechanisms of nanomaterials toxicity.5,6 Recently, we have reported the metabolic effects induced by chemically synthetized AgNPs of different sizes and

3 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 39

coatings on human skin keratinocytes (HaCaT cells), using NMR metabolomics.7 While skin is a major entry portal for AgNPs, it is well known that, whether exposure occurs through ingestion, inhalation or dermal contact, AgNPs may enter central blood circulation and translocate to several organs.8,9 Liver in particular is a major site of AgNPs accumulation10,11 and an important detoxification organ, which plays a central role in body metabolism and homeostasis. Therefore, assessing the impact of AgNPs on the metabolism of liver cells, as newly reported herein, is of utmost relevance to understand the toxicity and biological effects of these nanoparticles. In the present work, we have investigated the metabolic effects of two types of AgNPs towards human hepatoma (HepG2) cells, at sub-toxic concentrations. The HepG2 cell line is a well characterized in vitro model for assessing liver toxicity, having a stable phenotype and being easy to culture, while maintaining antioxidant defenses and several liver-specific metabolic functions normally present in hepatocytes.12,13 The two types of AgNPs tested, synthetized either by chemical reduction with citrate or with an aqueous plant extract (rich in sugars and phenolic compounds), had identical metal core diameter (ca. 30 nm), and differed in the coating layer (citrate or plant sugars/phenols). With this comparison we aimed at assessing the influence of AgNPs coating on cellular metabolic responses and at improving current knowledge on the biological effects of biogenic AgNPs. Cellular responses to the plant extract used in the synthesis of biogenic AgNPs, as well as to citrate alone and to ionic silver (administered as AgNO3) were also characterized with a view to determine their influence on the metabolic changes observed in AgNPs-exposed cells. To our knowledge, this is the first comprehensive metabolomic study of AgNPs on HepG2 liver cells, which explores the comparison between well-characterized biogenic AgNPs, citrate-stabilized AgNPs and ionic silver.

4 ACS Paragon Plus Environment

Page 5 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

EXPERIMENTAL Silver nanoparticles Sterile, purified and endotoxin-free silver nanoparticles (Biopure AgNPs 1.0 mg/mL), with a diameter of 30 nm and a citrate surface, were purchased from Nanocomposix Europe (Prague, Czech Republic). Biogenic AgNPs were prepared through the reduction of silver nitrate with an aqueous extract of Eucalyptus globulus (EG) Labill. bark, according to the procedure previously described,14 with minor modifications. In brief, autoclaved E. globulus bark extract was added to an AgNO3 aqueous solution (1 mM), at a 1:3 volume ratio, and allowed to react in a sealed container (121ºC, 1 bar) for 20 min. Then, the nanoparticle suspension was centrifuged, the supernatant was replaced with ultra-pure water, and centrifuged three times to wash the biosynthesized AgNPs. The final nanoparticles suspension was stored at 4ºC, protected from light. Nanoparticle characterization was performed as previously reported.7 Briefly, AgNPs’ size and morphology were assessed by transmission electron microscopy (TEM), after evaporating dilute suspensions of AgNPs on a copper grid coated with an amorphous carbon film. The hydrodynamic diameter (Dh) and the zeta potential were measured for colloidal solutions of AgNPs dispersed in ultrapure distilled water, by dynamic light scattering (DLS) and electrophoretic mobility, respectively. Silver quantification measurements were performed by inductively coupled plasma optical emission spectrometry (ICP-OES).

Cell culture

5 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 39

The human hepatoma cell line (HepG2) was obtained from European Collection of Authenticated Cell Cultures (ECACC) and supplied by Sigma Aldrich (St. Louis, MO, USA). The cells were grown in high glucose DMEM supplemented with 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 100 U/mL penicillin, 100 µg/mL streptomycin and 250 µg/mL fungizone, in 75 cm2-flasks, at 37 ºC in 5% CO2 humidified atmosphere. For routine subculturing, the cells were allowed to grow for 3-4 days (when they reached 80% confluency), tripsinized and replated at a 1:4 split ratio in complete DMEM.

MTT cell viability assay Cells were seeded at a density of 5.6 x 104, in 96-well plates, and after 24 h the medium was replaced with fresh medium containing: AgNPs (concentration range 1-25 µg/mL), ionic silver (Ag+), added as AgNO3 (concentration range 0.5-2 µg/mL), sodium citrate (15 and 20 µM), or E. globulus (EG) Labill. bark extract (0.20 and 0.50 mg/mL). After additional 24 h, 50 µL of MTT (1 mg/mL) in PBS were added to each well and cell viability measured as previously described.15 Three independent assays, with three technical replicates each, were performed.

Cell exposure for metabolomics assays HepG2 cells were seeded at a density of 5.6 x 104/cm2 onto 10 cm diameter Petri dishes and allowed to adhere for 24 h. Then, the medium was replaced by fresh complete medium containing AgNPs or Ag+, at the IC5 or IC20 concentrations (selected based on the cell viability results). Specifically, the IC5/IC20 concentrations used were 6.4/11.0 µg/mL, 5.4/14.0 µg/mL and 0.35/0.64 µg/mL for Cit30 AgNPs, GS30 AgNPs and Ag+, respectively. It should also be noted these concentrations are above the minimum

6 ACS Paragon Plus Environment

Page 7 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

inhibitory concentrations (MIC) and minimum bactericidal concentrations (MBC) reported for citrate-stabilized16 and biogenic AgNPs17, which makes them meaningful in the context of AgNPs antibacterial applications. Fresh medium without AgNPs was added to control cells. Furthermore, cells were also incubated with medium containing citrate or EG bark extract, at concentrations equivalent to those added when cells were exposed to AgNPs. Five independent assays were performed.

Sample collection and preparation for NMR analysis Medium aliquots were collected from each dish, centrifuged (1000×g, 5 min) and the supernatant stored at -80 ºC until NMR analysis. Aliquots from medium without cells, incubated in the same conditions, were also collected. At the time of analysis, 540 µL of thawed medium were mixed with 60 µL of D2O containing 0.25% TSP-d4 and 550 µL of each sample were transferred into 5 mm NMR tubes. To collect cell samples, the remaining medium was aspirated and each dish was washed twice with PBS. Then, intracellular aqueous metabolites were extracted using a dual phase extraction, with methanol:chloroform:water (1:1:0.7), as described in our previous report.7 At the time of NMR analysis, dried polar extracts were reconstituted in 600 µL of deuterated phosphate buffer (100 mM, pH 7.4) containing 0.1 mM TSP, and lipophilic extracts were reconstituted in deuterated chloroform containing 0.03% TMS. For NMR analysis, 550 µL of each sample were transferred into 5 mm NMR tubes.

NMR data acquisition Standard 1D 1H NMR spectra of culture medium supernatants, polar extracts and lipophilic extracts were acquired on a Bruker Avance DRX-500 spectrometer operating at 500.13 MHz for 1H observation, at 298 K, using a 5 mm probe. Two-dimensional 1H-

7 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Page 8 of 39

H total correlation (TOCSY) spectra, 1H-13C heteronuclear single quantum correlation

(HSQC) spectra and J-resolved spectra were also registered for selected samples to assist spectral assignment. The main acquisition and processing parameters for these experiments are provided in Supplementary Table S1.

Multivariate analysis and spectral integration of 1D 1H NMR spectra Spectra were normalized by total spectral area to compensate for differences in cell numbers, and aligned in MATLAB version 7.9.0 (The MathWorks, Inc., Natick, MA, USA), using recursive segment-wise peak alignment18 to minimize chemical shift variations. PCA and PLS-DA were applied to UV scaled data, in the SIMCA-P 11.5 software (Umetrics, Umeå, Sweden), with a default seven-fold internal cross validation, from which Q2 and R2 values, respectively reflecting predictive capability and explained variance, were extracted. PLS-DA model robustness was further assessed by Monte Carlo cross validation (MCCV) with 500 iterations and permutation testing (class membership randomly assigned). The PLS-DA loadings plots were back-transformed by multiplying the loading weight w by the standard deviation of each variable, and colored according to variable importance to the projection (VIP) using the R-statistical software (R Foundation for Statistical Computing, Vienna, Austria). Selected signals in the normalized 1D spectra were integrated using the AmixViewer software. In cases of significant overlap, spectral deconvolution was employed, using a variable Gaussian/Lorentzian curve fitting, with a default line width of 1 Hz. For each metabolite, the percentage of variation and respective error were calculated, along with the effect size (ES), corrected for small sample numbers, and respective standard error, according to equations provided in the literature.19 Metabolites presenting variations with absolute ES larger than 0.8 (and with standard error < ES)

8 ACS Paragon Plus Environment

Page 9 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

were expressed in a heatmap colored as a function of the percentage of variation, using the R statistical software. Moreover, the statistical significance of the difference between the means of two groups (control and exposed) was assessed using the twosample t-test or the non-parametric Wilcoxon rank sum test with continuity correction, and considering the False Discovery Rate (FDR)-corrected p values (confidence level 95%).20 Correlation analysis was also performed to seek biochemical relations among metabolite variations. To this end, matrices with metabolite integrals were imported to the Matlab software, where the Pearson correlation coefficients (r), and respective significance, were calculated. A threshold of |r| > 0.75, with p < 0.005, was applied.

9 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 39

RESULTS Physicochemical properties and cytotoxicity of AgNPs The main properties of the AgNPs aqueous colloidal solutions used in this work are summarized in Table 1. Transmission electron microscopy (TEM) (Figure 1) showed that AgNPs were approximately spherical in shape, with average core diameters of 29.1±3.9 nm and 32.9±4.0 nm for Cit30 and GS30 AgNPs, respectively. DLS measurements further showed reasonably monodisperse size distributions for both AgNPs, as indicated by PdI values below 0.3. The hydrodynamic diameter (Dh) of GS30 AgNPs (76.1±3.3 nm) was higher than that of Cit30 AgNPs (43.3±0.5 nm), which is consistent with the coating of the silver core by an organic layer comprising mainly sugars and phenolic compounds present in the bark extract used for the synthesis of the former NPs.14 The negative zeta-potential value for these nanoparticles further indicates the presence of negatively charged coating compounds, which confer high colloidal stability. After 24h incubation in complete culture medium, the Dh of Cit30 and GS30 AgNPs increased to 92.2±1.7 nm and 110.3±2.4 nm, respectively. The dose-response curves of HepG2 cells exposed for 24 h to Cit30 or GS30 AgNPs, obtained through the MTT assay, are shown in Figure 1. Both nanoparticles decreased the viability of HepG2 cells in a dose-dependent way, inducing a significant decrease in cell viability at a concentration of 10 µg/mL or higher. Based on these results, we have selected two concentration levels for each NP type corresponding to equivalent effects on cell viability, namely IC5 (no significant difference from controls) and IC20 (20% decrease in viability). The choice of exposing cells to these sub-toxic concentrations was based on the hypothesis that cell metabolism would respond even in the absence of major toxicity.

10 ACS Paragon Plus Environment

Page 11 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Intracellular metabolic changes induced by AgNPs The 1H NMR spectra of cellular polar extracts (Figure 2) showed contributions from about 50 metabolites (Table S2), unambiguously identified with the aid of 2D spectra (Figure S1) and the spectral databases BBIOREFCODE-2–0–0 (Bruker Biospin, Rheinstetten, Germany) and HMDB.21 The profiles of control cells and of cells exposed to Cit30 or GS30 AgNPs showed subtle quantitative differences which were highlighted through multivariate analysis (Figure 3). For both types of AgNPs, PCA scores showed a concentration-dependent distribution along PC1, while PLS-DA discriminated control and exposed cells with high robustness (Q2 0.8 and 0.7 for Cit30 and GS30 models, respectively), as confirmed through MCCV and permutation testing (Figure S2). The corresponding loadings (Figure 3, C and F) showed that several of the metabolic features responsible for class discrimination were common to the two nanoparticle types, although with varying importance (different VIP coloring of loadings). Spectral integration of individual metabolite signals further allowed the magnitude of those changes to be studied in detail. The more relevant variations (absolute effect size > 0.8), expressed as the percentage variation in relation to controls, are presented in Figure 4A and in Table S3. The variation of most intracellular polar metabolites was, for both NP types, clearly dependent on the nanoparticle concentration, showing higher magnitude at the higher concentration of exposure. Additionally, while Cit30 AgNPs significantly altered the levels of several metabolites already at the IC5 concentration, the impact of an equivalent concentration of GS30 on the metabolic profile was smaller (19 vs. 9 metabolites with relevant variations); at the IC20 concentration, the overall impact of the two nanoparticle types was qualitatively similar (23 metabolite levels consistently affected upon exposure to either Cit30 or GS30 AgNPs), although there were clear

11 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 39

quantitative differences in the response profiles. For instance, the increase in intracellular lactate was especially prominent in cells exposed to GS30 AgNPs (1528%), while Cit30 AgNPs caused a smaller increase (9%) in lactate levels, seen only at the IC5 concentration. On the other hand, fumarate was more markedly decreased upon exposure to Cit30 AgNPs (-46% vs. -25% for GS30 AgNPs). In regard to amino acids, the two nanoparticle types induced opposite variations in the levels of intracellular glutamine: cells exposed to the IC20 concentration of Cit30 AgNPs increased their glutamine levels (10% higher in relation to controls) while exposure to GS30 AgNPs caused a significant decrease (-14%) in glutamine at the IC5 concentration (and no change at IC20). As for other amino acids, both nanoparticle types induced significant decreases in glutamate, aspartate, N-acetylaspartate and glycine and significant increases in isoleucine, leucine, valine, phenylalanine, methionine and an unassigned Nacetylated amino acid; these changes were generally more prominent upon exposure to Cit30 AgNPs. Moreover, these particles additionally caused significant increases in tyrosine and histidine, not seen upon exposure to GS30 AgNPs. Nanoparticle exposure also affected the levels of several metabolites known to be involved in energy generation and transfer processes, namely creatine, phosphocreatine, ADP and NAD+ (decreased in relation to control cells), along with ATP and UTP (increased compared to control cells). Also of notice was the specific increase in UDP and the relatively larger increases in ATP and UTP upon exposure to GS30 AgNPs (near 40% increases vs. 17-21% increases with Cit30 AgNPs). The intracellular levels of reduced glutathione (GSH) increased 17-19% upon exposure to the IC5 of either Cit30 or GS30 AgNPs, and 23-25% in cells exposed to the IC20 of each nanoparticle type; thus, this variation showed little dependency on either the concentration or type of AgNPs. Finally, upon exposure to AgNPs, the

12 ACS Paragon Plus Environment

Page 13 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

endometabolome of HepG2 cells showed significantly decreased levels of phosphocholine (-24 and -32% in Cit30 and GS30 AgNPs exposed cells, respectively) and of taurine (-22 and -31% in Cit30 and GS30 AgNPs exposed cells, respectively). Analysis of lipophilic extracts further revealed a few changes in the cellular lipid composition (Figure 4B and Table S4), most of which were common to cells exposed to Cit30 and GS30 AgNPs. These included increased levels of esterified cholesterol and decreased levels of polyunsaturated fatty acids and phosphatidylcholine. Additionally, cells exposed to Cit30 AgNPs presented significantly lower levels of total cholesterol, while cells exposed to GS30 AgNPs showed a non-significant decrease in phosphatidylethanolamine.

Extracellular metabolic changes induced by AgNPs The metabolic variations in culture medium supernatants of AgNP-exposed cells relatively to control cells are summarized in Figure 4C and in Table S5. Most changes were concentration-dependent, with IC5 generally inducing fewer changes in medium composition than IC20. Also, extracellular metabolic signatures of Cit30 and GS30 AgNPs shared several common features. Both NP types induced higher consumption of pyruvate (as indicated by decreased levels compared to the medium of control cells), although in the case of GS30 AgNPs this decrease was only relevant at the IC20 concentration. Moreover, the excretion of lactate, citrate and fumarate increased. In this respect, it should be noted that while in the case of exposure to Cit30 AgNPs, increased citrate levels in the medium could be partially due to citrate release from the nanoparticles surface, this was not the case for GS30 as no citrate was present in the coating layer; thus, some citrate excretion must have originated from the cells altered metabolic activity. Nanoparticle exposure also induced alterations in the extracellular

13 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 39

levels of several amino acids. Changes common to the two nanoparticle types comprised increased excretion of glutamate and decreased consumption of the amino acids phenylalanine, leucine and isoleucine. This variation in branched chain amino acids was in turn highly correlated to decreased release of their catabolic products 2oxoisoleucine and α-ketovaline (|r| > 0.9, p < 0.001). Interestingly, GS30, but not Cit30 AgNPs, caused a higher consumption of glutamine at the IC5 concentration (as compared to non-exposed control cells). Furthermore, both AgNPs induced decreased formate excretion and choline consumption, this latter change being clearly more marked upon exposure to GS30 AgNPs. Finally, the medium of GS30-exposed cells, but not of Cit30-exposed cells, showed significantly higher acetate levels compared to the medium of control cells.

Metabolic effects of compounds used for nanoparticle coating Cell incubations with citrate (stabilizer for Cit30 AgNPs) and the EG extract used to prepare the GS30 AgNPs were also performed to assess the possible influence of these coating substances on the metabolic responses. As shown in Figure 4, citrate did not induce any significant changes in the endo- or exometabolome of HepG2 cells. On the other hand, the EG extract alone caused relevant changes in several metabolites. Regarding intracellular polar metabolites (Figure 4A), some effects were common to those observed in cells exposed to either Cit30 or GS30 AgNPs, namely the increases in lactate, an N-acetylated amino acid and UTP, together with the decreases in phosphocholine and taurine. Hence, it may be postulated that these variations did not reflect nanoparticle-specific effects. On the other hand, a number of alterations seen upon exposure to AgNPs were not observed in cells exposed to the EG extract; for instance fumarate, glutamine, glutamate, creatine and phosphocreatine levels were

14 ACS Paragon Plus Environment

Page 15 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

unaffected in those cells. Additionally, the variation of several other metabolites was in the opposite direction between cells exposed to AgNPs and the EG extract. To cite a few examples, cells exposed to the EG extract showed increased levels of aspartate, glycine and NAD+, compared to control cells, and decreased levels of ATP and GSH. In regard to the lipid composition, the EG extract caused decreases in total and esterified cholesterol and a slight but significant increase in phosphatidylethanolamine, again exhibiting a mode of action different from the nanoparticles (Figure 4B). Cellular incubation with the EG extract also induced a few significant changes in the exometabolome (Figure 4C). These included decreased pyruvate consumption and increased formate release, as compared to control cells, variations which were actually in the opposite direction to those induced by GS30 AgNPs. Additionally, the medium of cells incubated with the EG extract showed increased levels of lactate (although not statistically significant), citrate and acetate.

Metabolic effects of ionic silver The intracellular metabolic signature of Ag+ in HepG2 cells was somewhat similar to that of AgNPs (Figure 4), although a few differences could be noted. In regard to intracellular polar metabolites, these included larger increases of branched chain and aromatic amino acids, histidine, methionine and threonine, especially at the IC20 concentration, along with a non-significant decrease in NAD+ and smaller increases in GSH (< 10%, as compared to ~20% induced by AgNPs). Additionally, UTP levels were not increased upon exposure to Ag+ while pantothenate registered a non-significant increase, not seen with AgNPs. In what concerns the lipid profile, the most marked effects were the increase in cholesterol esters and the decrease in phosphatidylcholine,

15 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 39

similarly to the observations made for AgNPs. However, Ag+ seemed to produce smaller effects on total cholesterol and polyunsaturated fatty acids. At the exometabolome level, ionic silver enhanced the consumption of pyruvate, downregulated the consumption of branched chain amino acids and choline, and increased the excretion of lactate, fumarate and glutamate, similarly to AgNPs. On the other hand, differently from AgNPs, citrate levels were not altered, whereas nonsignificant decreases in alanine and acetate and a significant increase in allantoin were noted.

DISCUSSION Exposure of HepG2 cells to AgNPs with a metallic core diameter of 30 nm coated either with citrate or a plant-derived phenolic/sugar layer, at sub-toxic concentrations, was found to induce profound changes in the cellular metabolome, reflecting metabolic adaptations within different pathways, as proposed in Figure 5. While glucose levels were not affected under AgNPs exposure, exposed cells appeared to intensify pyruvate use and TCA cycle activity, as supported by increased pyruvate consumption from the culture medium, together with decreased intracellular levels of anaplerotic amino acids (glycine, N-acetylaspartate, aspartate and glutamate) and polyunsaturated fatty acids, possibly reflecting their oxidation for energy production. The observed increase in intracellular ATP levels is also in line with this hypothesis. These results are, however, in contrast with those reported by Chen and coworkers whereby HepG2 cells (and other cell types) exposed to non-cytotoxic concentrations of AgNPs (up to 8 µg/mL) decreased intracellular ATP.22 In that work, ATP depletion was related to a shift in energy metabolism from oxidative phosphorylation and fatty acid oxidation to glycolysis, based on increased intracellular

16 ACS Paragon Plus Environment

Page 17 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

pyruvate, lactate, triglyceride and cholesterol levels. Although the size of the PVPcoated AgNPs used in the cited work (core diameter of 25 nm and Dh of 76 nm) was comparable to those of the AgNPs used in this work, differences in other AgNPs physicochemical properties (e.g. coating layer, surface charge) could contribute for this discrepancy. Another possible explanation could relate to the supplementation of the culture medium used in this work with sodium pyruvate, according to the specific recommendations for culturing HepG2 cells. It is known that pyruvate may serve as an energy substrate and regenerate NAD+ through conversion into lactate.23 Indeed, we have also seen increased production of lactate accompanying intensified pyruvate metabolism. Similarly, Sheline and co-authors reported that adding pyruvate to the medium of neuronal cells attenuated the toxic effects induced by zinc ions, namely by mitigating ATP loss and cell death.23 Importantly, in the specific case of hepatocytes, the medium supplementation with pyruvate should better match the in vivo situation, as the liver is largely responsible pyruvate recycling (produced from lactate by other organs) and gluconeogenesis. The ability of AgNP-exposed cells to increase their ATP production, observed in this work, may be viewed as an indicator of low cytotoxicity, in consonance with the small decreases (5-20%) in MTT-assessed cell viability (Figure 1). The opposite effect, i.e. ATP depletion, has been implicated in the cytotoxicity of AgNPs24, as well as of other nanoparticles.25,26,27 In particular, a low ATP yield in human fibroblasts and glioblastoma cells exposed to AgNPs has been associated to mitochondrial dysfunction caused by structural damage induced by nanoparticle deposition and/or the induction of oxidative stress.28 Furthermore, it has been demonstrated that AgNPs can directly inhibit the activity of liver mitochondrial ATPase, thus reducing ATP synthesis29, and that AgNPs can impair oxidative phosphorylation in liver mitochondria.30,31 Our results

17 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 39

further showed significantly decreased levels of creatine and phosphocreatine, possibly reflecting reprogramming of energy storage and transmission. A similar observation has been made for HepG2 cells exposed to graphene nanosheets.32 Another notable effect upon AgNPs exposure was the increase in intracellular levels of GSH, a key player against oxidative stress.12 These results agree with earlier works where HepG2 cells33 or primary mouse liver cells34 increased GSH levels (~1.1 fold), along with superoxide dismutase activity, upon exposure to AgNPs at a concentration causing ~10 % decrease in cell viability. In contrast, exposure to AgNPs at more cytotoxic concentrations was reported to induce decreased levels of GSH and postulated to reflect persistent oxidative stress.35,36,37 Additionally, AgNPs-treated cells showed significant increases in branched chain amino acids, aromatic amino acids, histidine and methionine, which were especially pronounced with Cit30 AgNPs. Increases in those amino acids have been previously identified as part of the metabolic signature of drug-induced autophagy in tumor cells.38 Autophagy is a catabolic process by which unnecessary or dysfunctional cellular components are sequestered into double-membrane vesicles (autophagosomes) and targeted for lysosomal degradation.39 Recently, a variety of nanoparticles have been shown to induce autophagy, leading either to increased cell death or, paradoxically, to promotion of cell survival.40 PVP-coated silver nanoparticles were shown to induce a size-dependent autophagic response in HepG2 cells, even at relatively low or noncytotoxic doses. Furthermore, AgNPs-induced autophagy was shown to play a cytoprotective role at sub-toxic concentrations in HepG2 and HeLa cells.41,42 We may thus hypothesize that the observed increase in several amino acids could relate to an autophagic response, thus providing additional substrates for energy production and aiding cell survival, an hypothesis which needs to be verified in further studies. Also of

18 ACS Paragon Plus Environment

Page 19 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

notice is that HepG2 cells appeared to use glutamine differently depending on exposure to Cit30 or GS30 AgNPs. While this amino acid increased in Cit30-treated cells (possibly reflecting protein catabolism, as postulated for other amino acids), GS30 induced a higher glutamine consumption from the medium and decreased intracellular glutamine levels, suggesting higher glutaminolytic activity. Phospholipid metabolism also seemed to be affected upon exposure to both nanoparticle types, as suggested by decreased choline consumption from the medium along

with

decreased

levels

of

phosphocholine

and

phosphatidylcholine.

Phosphatidylcholines are major components of cellular membranes, synthesized mainly through the CDP-choline pathway, which is part of the Kennedy pathway.43 The CDPcholine pathway begins with the uptake of exogenous choline into the cell and proceeds with the phosphorylation of choline to form phosphocholine (Figure 6A). In AgNPstreated cells, we found a significant negative correlation between choline uptake and phosphatidylcholine (Figure 6B), suggesting the downregulation of phospholipid synthesis. This effect was especially pronounced in GS30-treated cells, where significant correlations were found between several compounds of the CDP-choline pathway. Also, in GS30-treated cells, the intermediates of this pathway additionally showed a strong correlation with the increase in UTP, possibly reflecting its role as a CTP precursor. Furthermore, cholesterol esters increased in exposed cells, which could reflect a form of biologically inert storage (detoxification) of fatty acids. As for the influence of coating substances on cellular responses to AgNPs, while citrate produced no effect, the EG extract used to prepare and stabilize the GS30 showed a totally different response profile from the nanoparticles themselves. For instance, ATP and GSH decreased, most amino acids varied to a lesser extent and in opposite sense, creatine and phosphocreatine remained unaltered, and so did the levels

19 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 39

of polyunsaturated fatty acids and phosphatidylcholine, while cholesterol esters decreased. This could reflect not only the modified chemical nature of the compounds coating the GS30 AgNPs (and eventually released into the medium) compared to those present in the pristine EG extract, due to oxidation during NP synthesis, but also the expectedly different uptake and interaction of nanoparticles and soluble organic compounds with cells. On the other hand, AgNPs and Ag+ metabolic signatures displayed several similar features, although a few differences could be noted. Specifically, the higher increase in several amino acids, the less pronounced increases in ATP or GSH in Ag+-exposed cells, together with the increase in the oxidative stress marker allantoin (product of uric acid oxidation) suggest that cells were less able to cope with the amount of ionic silver administered at once than with the gradual intracellular release from AgNPs. Indeed, this is consistent with the much larger cytotoxicity of Ag+ compared to AgNPs (IC50 1.19 and 18.9 µg/mL, respectively). Overall, the metabolic signature of AgNPs in HepG2 cells suggests that, under the conditions tested (IC5 and IC20 exposures), these cells were able to activate metabolism-mediated protective mechanisms. Notably, a protective response has also been found in vivo in the liver of mice intravenously injected with PEG-coated AgNPs (core diameter of 30 nm).44 Specifically, the liver of treated mice displayed upregulated synthesis of glycogen, lipolysis, fatty acid oxidation and production of antioxidative molecules (including GSH). Thus, our results demonstrate that in vitro metabolomics shows good potential in aiding to explain and predict the in vivo outcomes of nanoparticle exposure.

CONCLUSIONS

20 ACS Paragon Plus Environment

Page 21 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

This study has revealed, for the first time to our knowledge, the metabolic signatures of differently coated 30 nm AgNPs and ionic silver in human liver HepG2 cells, an especially relevant cell type in the context of nanoparticle toxicity. Sub-toxic concentrations of AgNPs were proposed to induce metabolic adaptations in energy production processes (glucose metabolism and the phosphocreatine system), antioxidant defenses, protein degradation and lipid metabolism, most of which were suggested to reflect activation of metabolism-mediated protective mechanisms. These findings highlight a major role for intermediary metabolism in protecting cells and mitigating toxic effects, which paves the way for the development of novel strategies to achieve improved control over cellular responses to nanoparticles/toxicants. On the other hand, cells exposed to ionic silver appeared to be less able to activate such protection mechanisms, emphasizing that more toxic outcomes may potentially result from direct ionic silver administration as compared to gradual release of silver ions from AgNPs. Furthermore, this work revealed subtle differences in the signatures of the two types of AgNPs compared, not captured by conventional cytotoxicity assessment, suggesting that the fine tuning of AgNPs properties, such as those influenced by their coating, may be used to modulate cellular metabolic responses and, thus, the resulting biological outcomes.

Supporting Information. The following supporting information is available free of charge at ACS website http://pubs.acs.org. Figure S1. Expansions of 1H-1H TOCSY, 1H-13C HSQC, and J-resolved spectra of a polar extract from HepG2 control cells.

21 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 39

Figure S2. ROC space and Q2 histogram obtained by MCCV and permutation testing of the PLS-DA models of polar extracts, corresponding to exposure to Cit30 and GS30 AgNPs. Table S1. Main parameters used for the acquisition and processing of 500 MHz 1D 1H and 2D NMR spectra of culture media, polar extracts, and lipophilic extracts. Table S2. Assignment of resonances in the NMR profile of polar extracts from HepG2 cells. Table S3. Main metabolite variations in polar extracts of HepG2 cells exposed to Cit30 and GS30 AgNPs at respective IC5 and IC20 concentrations. Table S4. Main metabolite variations in lipophilic extracts of HepG2 cells exposed to Cit30 and GS30 AgNPs at respective IC5 and IC20 concentrations. Table S5. Main metabolite variations in medium supernatants of HepG2 cells exposed to Cit30 and GS30 AgNPs at respective IC5 and IC20 concentrations.

Acknowledgments. This work was developed in the scope of the projects CICECOAveiro Institute of Materials (Ref. FCT UID/CTM/50011/2013) and CESAM (Ref. FCT UID/AMB/50017/2013), financed by national funds through the FCT/MEC and when applicable co-financed by the European Regional Development Fund (FEDER) under the PT2020 Partnership Agreement. Funding to the project FCOMP-01-0124-FEDER021456 (Ref. FCT PTDC/SAU-TOX/120953/2010) by FEDER through COMPETE and by national funds through FCT, financial support from the European Union Framework Programme for Research and Innovation HORIZON 2020, under the TEAMING Grant agreement No 739572 - The Discoveries CTR, and the FCT-awarded grants (SFRH/BD/79494/2011, SFRH/BPD/111736/2015 and SFRH/BPD/89982/2012) are acknowledged. I.F.D and C.S.R.F. acknowledge the FCT/MCTES for research contracts

22 ACS Paragon Plus Environment

Page 23 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

under the Program ‘Investigador FCT’. Tiago Pedrosa is thanked for laboratory assistance, Dr António Barros for helpful advice regarding statistical analysis, and Prof. Ruth Duncan for insightful discussions. Dr Manfred Spraul, Bruker BioSpin (Germany) is thanked for access to software and spectral database. The Portuguese National NMR (PTNMR) Network, supported with FCT funds, is also acknowledged.

Declaration of Interest. No conflict of interest to declare. Funding institutions had no role in study design, data collection or analysis, decision to publish, or manuscript preparation.

23 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 39

REFERENCES

(1) Rai, M. K.; Deshmukh, S. D.; Ingle, A. P.; Gade, A. K. Silver nanoparticles: the powerful nanoweapon against multidrug-resistant bacteria. J. Appl. Microbiol. 2012, 112 (5), 841-852. (2) Galdiero, S.; Falanga, A.; Vitiello, M.; Cantisani, M.; Marra, V.; Galdiero, M. Silver nanoparticles as potential antiviral agents. Molecules 2011, 16 (10), 8894-8918. (3) Wei, L.; Lu, J.; Xu, H.; Patel, A.; Chen, Z.-S.; Chen, G. Silver nanoparticles: synthesis, properties, and therapeutic applications. Drug Discovery Today 2015, 20 (5), 595-601. (4) Costa, P. M.; Fadeel, B. Emerging systems biology approaches in nanotoxicology: Towards a mechanism-based understanding of nanomaterial hazard and risk. Toxicol. Appl. Pharmacol. 2016, 299, 101-111. (5) Duarte, I. F. Following dynamic biological processes through NMR-based metabonomics: A new tool in nanomedicine? J. Controlled Release 2011, 153 (1), 34-39. (6) Lv, M. Y.; Huang, W. Q.; Chen, Z. P.; Jiang, H. L.; Chen, J. Q.; Tian, Y.; Zhang, Z. J.; Xu, F. G. Metabolomics techniques for nanotoxicity investigations. Bioanalysis 2015, 7 (12), 1527-1544. (7) Carrola, J.; Bastos, V.; Jarak, I.; Oliveira-Silva, R.; Malheiro, E.; Daniel-da-Silva, A. L.; Oliveira, H.; Santos, C.; Gil, A. M.; Duarte, I. F. Metabolomics of silver nanoparticles toxicity in HaCaT cells: structure-activity relationships and role of ionic silver and oxidative stress. Nanotoxicology 2016, 10 (8), 1105-1117.

24 ACS Paragon Plus Environment

Page 25 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

(8) Oberdorster, G.; Oberdorster, E.; Oberdorster, J. Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environ. Health Perspect. 2005, 113 (7), 823-839. (9) Johnston, H. J.; Hutchison, G.; Christensen, F. M.; Peters, S.; Hankin, S.; Stone, V. A review of the in vivo and in vitro toxicity of silver and gold particulates: particle attributes and biological mechanisms responsible for the observed toxicity. Crit. Rev. Toxicol. 2010, 40 (4), 328-346. (10) Dziendzikowska, K.; Gromadzka-Ostrowska, J.; Lankoff, A.; Oczkowski, M.; Krawczynska, A.; Chwastowska, J.; Sadowska-Bratek, M.; Chajduk, E.; Wojewodzka,

M.;

Dusinska,

M.;

Kruszewski,

M.

Time-dependent

biodistribution and excretion of silver nanoparticles in male Wistar rats. J. Appl. Toxicol. 2012, 32 (11), 920-928. (11) Xue, Y. Y.; Zhang, S. S.; Huang, Y. M.; Zhang, T.; Liu, X. R.; Hu, Y. Y.; Zhang, Z. Y.; Tang, M. Acute toxic effects and gender-related biokinetics of silver nanoparticles following an intravenous injection in mice. J. Appl. Toxicol. 2012, 32 (11), 890-899. (12) Alia, M.; Ramos, S.; Mateos, R.; Bravo, L.; Goya, L. Response of the antioxidant defense system to tert-butyl hydroperoxide and hydrogen peroxide in a human hepatoma cell line (HepG2). J. Biochem. Mol. Toxicol. 2005, 19 (2), 119-128. (13) Maier, K.; Hofmann, U.; Reuss, M.; Mauch, K. Dynamics and control of the central carbon metabolism in hepatoma cells. BMC Syst. Biol. 2010, 4 (1), 1-28. (14) Santos, S. A. O.; Pinto, R. J. B.; Rocha, S. M.; Marques, P. A. A. P.; Neto, C. P.; Silvestre, A. J. D.; Freire, C. S. R. Unveiling the chemistry behind the green synthesis of metal nanoparticles. ChemSusChem 2014, 7 (9), 2704-2711.

25 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 39

(15) Carrola, J.; Bastos, V.; Ferreira de Oliveira, J. M.; Oliveira, H.; Santos, C.; Gil, A. M.; Duarte, I. F. Insights into the impact of silver nanoparticles on human keratinocytes metabolism through NMR metabolomics. Arch. Biochem. Biophys. 2016, 589, 53-61. (16) Zhou, Y.; Kong, Y.; Kundu, S.; Cirillo, J. D.; Liang, H. Antibacterial activities of gold and silver nanoparticles against Escherichia coli and bacillus CalmetteGuerin. J. Nanobiotechnology 2012, 10, 19 (19 pages). (17) Ramalingam, B.; Parandhaman, T.; Das, S. K. Antibacterial effects of biosynthesized silver nanoparticles on surface ultrastructure and nanomechanical properties of Gram-negative bacteria viz. Escherichia coli and Pseudomonas aeruginosa. ACS Appl. Mater. Interfaces 2016, 8 (7), 4963-4976. (18) Veselkov, K. A.; Lindon, J. C.; Ebbels, T. M. D.; Crockford, D.; Volynkin, V. V.; Holmes, E.; Davies, D. B.; Nicholson, J. K. Recursive segment-wise peak alignment of biological 1H NMR spectra for improved metabolic biomarker recovery. Anal. Chem. 2009, 81 (1), 56-66. (19) Berben, L.; Sereika, S. M.; Engberg, S. Effect size estimation: methods and examples. Int. J. Nurs. Stud. 2012, 49 (8), 1039-1047. (20) Benjamini, Y.; Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Series B 1995, 57 (1), 289300. (21) Wishart, D. S.; Jewison, T.; Guo, A. C.; Wilson, M.; Knox, C.; Liu, Y.; Djoumbou, Y.; Mandal, R.; Aziat, F.; Dong, E.; Bouatra, S.; Sinelnikov, I.; Arndt, D.; Xia, J.; Liu, P.; Yallou, F.; Bjorndahl, T.; Perez-Pineiro, R.; Eisner, R.; Allen, F.; Neveu, V.; Greiner, R.; Scalbert, A. HMDB 3.0 - The human metabolome database in 2013. Nucleic Acids Res. 2013, 41 (D1), D801-D807.

26 ACS Paragon Plus Environment

Page 27 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

(22) Chen, Y.; Wang, Z.; Xu, M.; Wang, X.; Liu, R.; Liu, Q.; Zhang, Z.; Xia, T.; Zhao, J.; Jiang, G.; Xu, Y.; Liu, S. Nanosilver incurs an adaptive shunt of energy metabolism mode to glycolysis in tumor and nontumor cells. ACS Nano 2014, 8 (6), 5813-5825. (23) Sheline, C. T.; Behrens, M. M.; Choi, D. W. Zinc-induced cortical neuronal death: contribution of energy failure attributable to loss of NAD+ and inhibition of glycolysis. J. Neurosci. 2000, 20 (9), 3139-3146. (24) Zhang, T. L.; Wang, L. M.; Chen, Q.; Chen, C. Y. Cytotoxic potential of silver nanoparticles. Yonsei Med. J. 2014, 55 (2), 283-291. (25) Dong, X.; Mattingly, C. A.; Tseng, M. T.; Cho, M. J.; Liu, Y.; Adams, V. R.; Mumper, R. J. Doxorubicin and paclitaxel-loaded lipid-based nanoparticles overcome multidrug resistance by inhibiting P-glycoprotein and depleting ATP. Cancer Res. 2009, 69 (9), 3918-3926. (26) Bhattacharjee, S.; Rietjens, I. M.; Singh, M. P.; Atkins, T. M.; Purkait, T. K.; Xu, Z.; Regli, S.; Shukaliak, A.; Clark, R. J.; Mitchell, B. S.; Alink, G. M.; Marcelis, A. T.; Fink, M. J.; Veinot, J. G.; Kauzlarich, S. M.; Zuilhof, H. Cytotoxicity of surface-functionalized silicon and germanium nanoparticles: the dominant role of surface charges. Nanoscale 2013, 5 (11), 4870-4883. (27) Yu, K. N.; Yoon, T. J.; Minai-Tehrani, A.; Kim, J. E.; Park, S. J.; Jeong, M. S.; Ha, S. W.; Lee, J. K.; Kim, J. S.; Cho, M. H. Zinc oxide nanoparticle induced autophagic cell death and mitochondrial damage via reactive oxygen species generation. Toxicol. in Vitro 2013, 27 (4), 1187-1195. (28) AshaRani, P. V.; Low Kah Mun, G.; Hande, M. P.; Valiyaveettil, S. Cytotoxicity and genotoxicity of silver nanoparticles in human cells. ACS Nano 2009, 3 (2), 279-290.

27 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 39

(29) Chichova, M.; Shkodrova, M.; Vasileva, P.; Kirilova, K.; Doncheva-Stoimenova, D. Influence of silver nanoparticles on the activity of rat liver mitochondrial ATPase. J. Nanopart. Res. 2014, 16 (2), 2243, 14 pages. (30) Costa, C. S.; Ronconi, J. V.; Daufenbach, J. F.; Goncalves, C. L.; Rezin, G. T.; Streck, E. L.; Paula, M. M. In vitro effects of silver nanoparticles on the mitochondrial respiratory chain. Mol. Cell. Biochem. 2010, 342 (1-2), 51-56. (31) Teodoro, J. S.; Simoes, A. M.; Duarte, F. V.; Rolo, A. P.; Murdoch, R. C.; Hussain, S. M.; Palmeira, C. M. Assessment of the toxicity of silver nanoparticles in vitro: a mitochondrial perspective. Toxicol. in Vitro 2011, 25 (3), 664-670. (32) Jiao, G. Z.; Li, X.; Zhang, N.; Qiu, J. Q.; Xu, H. Y.; Liu, S. M. Metabolomics study on the cytotoxicity of graphene. RSC Adv. 2014, 4 (84), 44712-44717. (33) Jain, J.; Arora, S.; Rajwade, J. M.; Omray, P.; Khandelwal, S.; Paknikar, K. M. Silver nanoparticles in therapeutics: development of an antimicrobial gel formulation for topical use. Mol. Pharmaceutics 2009, 6 (5), 1388-1401. (34) Arora, S.; Jain, J.; Rajwade, J. M.; Paknikar, K. M. Interactions of silver nanoparticles with primary mouse fibroblasts and liver cells. Toxicol. Appl. Pharmacol. 2009, 236 (3), 310-318. (35) Piao, M. J.; Kang, K. A.; Lee, I. K.; Kim, H. S.; Kim, S.; Choi, J. Y.; Choi, J.; Hyun, J. W. Silver nanoparticles induce oxidative cell damage in human liver cells through inhibition of reduced glutathione and induction of mitochondriainvolved apoptosis. Toxicol. Lett. 2011, 201 (1), 92-100. (36) Vrček, I. V.; Žuntar, I.; Petlevski, R.; Pavičić, I.; Dutour Sikirić, M.; Ćurlin, M.; Goessler, W. Comparison of in vitro toxicity of silver ions and silver nanoparticles on human hepatoma cells. Environ. Toxicol. 2016, 31 (6), 679692.

28 ACS Paragon Plus Environment

Page 29 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

(37) Avalos, A.; Haza, A. I.; Mateo, D.; Morales, P. Cytotoxicity and ROS production of manufactured silver nanoparticles of different sizes in hepatoma and leukemia cells. J. Appl. Toxicol. 2014, 34 (4), 413-423. (38) Lin, G.; Troy, H.; Andrejeva, G.; Fong, A.-C. L. W. T.; Koh, D.-M.; Robinson, S. P.; Judson, I. R.; Griffiths, J. R.; Leach, M. O.; Chung, Y.-L. Abstract B56: Treatment-induced autophagy increases amino acid uptake and switches glucose addiction to amino acid catabolism in cancer. Mol. Cancer Res. 2016, 14 (1 Supplement), B56-B56. (39) Jiang, P.; Mizushima, N. Autophagy and human diseases. Cell Res. 2014, 24 (1), 69-79. (40) Zabirnyk, O.; Yezhelyev, M.; Seleverstov, O. Nanoparticles as a novel class of autophagy activators. Autophagy 2007, 3 (3), 278-281. (41) Mishra, A. R.; Zheng, J.; Tang, X.; Goering, P. L. Silver nanoparticle-induced autophagic-lysosomal disruption and NLRP3-inflammasome activation in HepG2 cells is size-dependent. Toxicol. Sci. 2016, 150 (2), 473-487. (42) Lin, J.; Huang, Z.; Wu, H.; Zhou, W.; Jin, P.; Wei, P.; Zhang, Y.; Zheng, F.; Zhang, J.; Xu, J.; Hu, Y.; Wang, Y.; Li, Y.; Gu, N.; Wen, L. Inhibition of autophagy enhances the anticancer activity of silver nanoparticles. Autophagy 2014, 10 (11), 2006-2020. (43) Chang, Y. F.; Carman, G. M. CTP synthetase and its role in phospholipid synthesis in the yeast Saccharomyces cerevisiae. Prog. Lipid Res. 2008, 47 (5), 333-339. (44) Jarak, I.; Carrola, J.; Barros, A.; Gil, A. M.; Pereira, M. L.; Corvo, M. L.; Duarte, I. F. Metabolism modulation in different organs by silver nanoparticles: an NMR metabolomics study of a mouse model. Toxicol. Sci. 2017, 159 (2), 422-435.

29 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 39

Figure Captions

Figure 1. TEM images and particle size histograms of Cit30 (A,B) and GS30 (D,E) AgNPs; Viability (%) of HepG2 cells, measured by the MTT assay, after 24 h of exposure to: Cit30 AgNPs (C) or GS30 (F) AgNPs, Ag+ (G), citrate (H) and EG extract (I). Data expressed as mean ± standard deviation of three independent assays with three replicates each (n 9). Statistically significant (p < 0.05) differences relatively to controls are indicated by *.

Figure 2. 500 MHz 1H NMR spectra of polar extracts from HepG2 cells grown for 24 h: (A) in the absence of AgNPs (controls); exposed to the IC5 concentrations of (B) Cit30 AgNPs, (C) GS30 AgNPs.

Figure 3. Multivariate analysis of 1H NMR spectra from polar extracts of HepG2 control cells and cells exposed to (A-C) Cit30, and (D-F) GS30 AgNPs: (A and D) PCA; (B and E) PLS-DA scores scatter plots generated by pairwise comparisons (control and AgNPs); (C and F) LV1 loadings w, colored as a function of variable importance to the projection (VIP).

Figure 4. Heatmap of the main metabolite variations in (A) polar extracts, (B) lipophilic extracts, and (C) culture media, from HepG2 cells exposed to Cit30 AgNPs, GS30 AgNPs, citrate, EG extract, or Ag+, colored according to % variation in relation to controls. aConcentrations that reflected the maximum amounts present when IC20 concentrations of AgNPs were administered to cells. * Uncorrected p-value < 0.05; ** FDR-corrected p-value < 0.05. Three letter code used for amino acids; NAA, N-

30 ACS Paragon Plus Environment

Page 31 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

acetylaspartate; Un. NAaa, unidentified N-acetylated amino acid; Cr, creatine; PCr, phosphocreatine; ADP/ATP, adenosine di/triphosphate; NAD+, nicotinamide adenine dinucleotide; GSH, reduced glutathione; PC, phosphocholine; UDP/UTP, uridine di/triphosphate; PTC, phosphatidylcholine; PTE, phosphatidylethanolamine; FA, fatty acyl.

Figure 5. Schematic diagram of the putative effects of AgNPs on glycolysis and the TCA cycle; Regular and circled arrows indicate effects caused, respectively, by Cit30 AgNPs and GS30 AgNPs, in relation to controls; Dashed arrows indicate |% variation| < 10%.

Figure 6. (A) Schematic diagram of phosphatidylcholine biosynthesis via the CDPcholine pathway. Regular and circled arrows indicate effects caused, respectively, by Cit30 AgNPs and GS30 AgNPs, in relation to controls; dashed arrows indicate |% variation| < 10%; thick arrows indicate |% variation| > 50%. (B) Pearson correlation coefficients (r) between metabolites involved in phosphatidylcholine biosynthesis, found to vary after HepG2 exposure to Cit30 (grey) or GS30 (black) AgNPs. ext, extracellular; * p < 0.05; ** p-value < 0.005.

31 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 39

Figure 1

32 ACS Paragon Plus Environment

Page 33 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Figure 2

33 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 39

Figure 3

34 ACS Paragon Plus Environment

Page 35 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Figure 4

35 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 39

Figure 5

36 ACS Paragon Plus Environment

Page 37 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Figure 6

37 ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 39

Table 1. Physicochemical properties of the citrate-coated AgNPs (Cit30) and the biogenic AgNPs obtained by green synthesis (GS30), dispersed in ultrapure distilled water. aDiameter indicated by the manufacturer; c

b

Diameter measured by TEM;

Hydrodynamic diameter and polydispersity index (PdI) measured by DLS; dZeta

potential assessed by electrophoretic mobility; eWavelength of maximum absorbance peak in the UV-Vis spectrum; fPercentage of ionic silver in the AgNPs suspension. Standard deviations calculated for Dh, ζ and %Ag+ correspond to 3 replicate measurements. n.a. not available (NPs synthesized in-house). AgNP

D (nm)a

D (nm)b

Dh (nm)c

PdIc

ζ (mV)d

λmaxe (nm)

%Ag+f

Cit30

32.7±4.8

29.1±3.9

43.3±0.5

0.25-0.26

-42.7±2.7

408

3.32±0.04

GS30

n.a.

32.9±4.0

76.1±3.3

0.20-0.22

-37.1±8.1

434

1.68±0.01

38 ACS Paragon Plus Environment

Page 39 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

For TOC only

39 ACS Paragon Plus Environment