Modulation of the thiol redox proteome by sugarcane ash-derived silica nanoparticles: insights into chronic kidney disease of unknown etiology

Abstract

Introduction

Chronic kidney disease of unknown etiology (CKDu) is an epidemic which is increasingly prevalent among agricultural workers and nearby communities, particularly those involved in the harvest of sugarcane. While CKDu is likely multifactorial, occupational exposure to silica nanoparticles (SiNPs), a major constituent within sugarcane ash, has gained increased attention as a potential contributor. SiNPs have high potential for generation of reactive oxygen species (ROS), and their accumulation in kidney could result in oxidative stress induced kidney damage consistent with CKDu pathology.

Methods

In order to characterize the impact of sugarcane ash derived (SAD) SiNPs on human kidney proximal convoluted tubule (PCT) cells and identify potential mechanisms of toxicity, HK-2 cells were exposed to treatments of either pristine, manufactured, 200 nm SiNPs or SAD SiNPs and changes to cellular energy metabolism and redox state were determined. To determine how the cellular redox environment may influence PCT cell function and toxicity, the redox proteome was examined using cysteine-targeted click chemistry proteomics.

Results

Pristine, 200 nm SiNPs induced minimal changes to energy metabolism and proteomic profiles in vitro while treatment with SAD SiNPs resulted in mitochondrial membrane hyperpolarization, inhibited mitochondrial respiration, increased reactive oxygen species generation, and redox proteomic trends suggesting activation of aryl hydrocarbon receptor (AHR) and other signaling pathways with known roles in mitochondrial inhibition and CKD progression.

Conclusion

Results suggest that PCT cell exposure to SAD SiNPs could promote glycolytic and fibrotic shifts consistent with CKDu pathology via oxidative stress-mediated disruption of redox signaling pathways.

Introduction

Chronic kidney disease (CKD) is a growing threat to public health, estimated to affect 850 million people worldwide in 2017 and projected to be the 5th highest cause of years of life lost by 2040 [1,2,3]. Cases continue to increase each year, the majority of which occur in poor communities with limited medical resources [4]. While many of these kidney diseases are linked to growing rates of diabetes and cardiovascular conditions, of particular concern is the rapidly expanding epidemic of chronic kidney disease of unknown etiology (CKDu).

CKDu predominantly occurs among agricultural communities, with hot spots occurring in Latin America, Sri Lanka, and India [5,6,7]. Regions in Mesoamerica have significantly outpaced global trends in CKD-associated disability-adjusted life-years, increasing by over 50% between 1990 and 2017 alone [8]. Sri Lanka and India have reported an estimated 60,000 and 34,000 yearly cases respectively [9]. However determining the true extent of this disease has proven extremely challenging due, in part, to the limited availability of medical resources and records in regions with highest prevalence as well as intensive diagnostic criteria (i.e., kidney biopsy). Sugarcane workers, particularly those involved in the harvest, are among those with the highest risk of CKDu development [10, 11]. As this disease occurs independent of common risk factors such as age, diabetes, and hypertension, it is suspected that environmental factors may be contributing to pathogenesis [12,13,14]. Exposure to agricultural toxicants, heat stress, dehydration, and exertional injury have all been identified as major occupational concerns [15,16,17,18]. Sugarcane ash is one such occupational exposure that has been gaining increased attention as a potential contributor to CKDu [11]. Sugarcane is often burned to facilitate harvest, exposing workers both directly through particulate matter and indirectly through ash deposition/groundwater contamination. Sugarcane ash is comprised of ~ 80% amorphous silica, primarily silica nanoparticles (SiNPs) with a size of ~ 200 nm [19, 20]. Elemental silicon has been found to be highly elevated in both the groundwater of communities and urine of individuals with increased risk of CKDu [21,22,23]. Recently, SiNPs have also been identified in kidney biopsies of those with confirmed CKDu [24]. Such findings suggest that sugarcane ash and constituent components are a major hazard for those that live or work in or near sugarcane fields. This has concerning implications for public health, as sugarcane ash, and SiNPs derived from ash, have been found to have deleterious properties both in vivo and in vitro [18, 25,26,27,28]. Thus, it becomes increasingly urgent to elucidate the mechanisms of toxicity in order to support accurate risk assessment and potential medical interventions.

Mitochondrial dysfunction, altered central carbon metabolism, and oxidative stress are key hallmarks of CKD which have also been identified to occur following exposure to sugarcane ash — some of which our group has explored previously [25, 29,30,31,32,33]. One potential source of this oxidative stress may be the SiNPs within the ash, which are known to be capable of catalyzing the production of reactive oxygen species (ROS) through exposed silanol groups [34,35,36]. SiNP accumulation in the kidney via proximal convoluted tubule (PCT) cell mediated uptake could result in a perturbed redox state, consistent with kidney biopsy findings and metabolomic profiles of agricultural workers [24, 37,38,39]. Disrupted redox signaling is a common event in many kidney diseases, capable of causing direct oxidative damage as well as indirectly through inhibition of protective signaling pathways [40, 41]. Because such pathways are highly dependent on the redox status of thiols in key signaling proteins, interrogation of the cysteine proteome via proteomic analysis is a powerful approach to identify which key proteins and associated pathways may be involved in initiation and progression of disease state [42,43,44]. Previous research has paired this approach with click chemistry-based labeling of reduced cysteine residues, allowing for quantification of redox state and clarification of the mechanisms through which specific exposures can induce redox-mediated toxicity [45]. As changes to the general thiol redox state have been found to occur in CKD and end-stage renal disease (ESRD), investigating specific redox proteomic changes following PCT cell exposure to sugarcane ash derived (SAD) SiNPs could yield valuable insight into the mechanisms of toxicity, potential contributions to CKDu, and any similarities to traditional CKD [46].

The goal of this manuscript is to characterize the short-term effects of acute SAD SiNP exposure on cellular redox state and investigate proteomic changes. To accomplish this a human PCT cell line (HK-2) was treated with environmentally relevant SAD SiNPs or pristine manufactured SiNPs of comparable size. Following treatment, changes to cell viability, mitochondrial/glycolytic function, ROS generation, and redox proteome state were investigated to identify mechanisms of toxicity related to CKD pathophysiology.

Materials and methods

Treatments

Sugarcane ash was collected in December 2019 from the Ingenio San Antonio-Nicaragua sugarcane field near Leon, Nicaragua. SiNPs were derived from SA using a method previously utilized and briefly described below [47]. Pristine homogenous nano-sized (200 nm) SiNPs were purchased from nanoComposix/Fortis Life Sciences (SISN200-25 M). 1 mg of particles were weighed out using a XP6 Microbalance (Mettler Toledo), added to 1mL of endotoxin-free water (Thermo Fisher Scientific, 1897399) and sonicated for 10 min with a Fisherbrand™ Model 705 Sonic Dismembrator (Thermo Fisher Scientific) to generate a 1 mg/mL stock solution. Each stock solution was exposed to UV light overnight to reduce risk of contamination and inactivate potential endotoxin. These stock solutions were then diluted to chosen exposure concentrations of 2.5–25 µg/mL. This working solution was generated with cell culture media immediately preceding treatment. Following dilution, each solution was found to be stable and remain dispersed without signs of precipitation for up to 24 h.

Dose calculation

Dose determination calculations have been discussed in prior publications and are based on estimated exposure rates in previous literature [25]. Briefly, PM₁₀ air sampling values during sugarcane burning (1 h time weighted averages of 1807 µg/m³) were multiplied by EPA standard breathing rates (1.4m³/hour) and estimated SiNP kidney biodistribution rates of ~ 2% [48,49,50]. This results in an estimated burden of 25 µg per kidney for each hour exposed to burning conditions and is the rationale for an upper dose of 25 µg/mL. As this calculation is an approximation, it carries several limitations which must be considered and are likely to result in it being relatively high — particularly under in vitro conditions. Average exposure conditions, total exposure time, breathing rates, and individual biodistribution values will vary significantly across real world exposure events.

Particle characterization

As demonstrated and discussed in our previous publication, there are significant morphological differences between SiNP populations [25]. Pristine synthesized SiNPs maintain the expected homogenous spherical morphology associated with nanomaterials, while environmentally relevant SAD SiNPs are far more irregular in structure. Previous dynamic light scattering and transmission electron microscopy analyses have found SAD SiNP particle populations to be larger on average than pristine SiNPs. This was confirmed by repeated analyses performed for this study (Fig. 1) with mean hydrodynamic diameters of 421 nm vs. 225 nm. SAD SiNPs are also more variable in size and demonstrate higher potential for agglomeration than pristine particles, as indicated by respective polydispersity indices of 0.393 and 0.009.

Fig. 1

Particle Characteristics. Particles were imaged via TEM and further analyzed via DLS. Hydrodynamic diameters and polydispersity indices indicate a larger more homogeneous population among SAD SiNPs when compared to pristine 200 nm SiNPs. This distinct morphology is confirmed by TEM imaging, demonstrating perfectly spherical particles in pristine 200 nm SiNP populations rather than the more variable structure present in environmentally relevant SAD SiNPs

Sugarcane ash derived silica nanoparticle (SAD SiNPs) generation

SAD SiNPs were generated as discussed in previous publications and summarized below [25]. Cetyltrimethylammonium bromide was added to a 1:1 H₂O/n-butanol solution and heated to 60 °C. Na₂SiO₃ was introduced under stirring to form a SiO₂ nanogel, followed by the addition of 0.5 M H₂SO₄ to adjust the pH to 4. The gel was aged for 8 h, washed with distilled water three times, filtered, and dried overnight at 120 °C. The SAD SiNPs were washed again three times with double distilled water, dried at 120 °C, and stored at room temperature.

Cell culture

Immortalized human proximal convoluted tubule cells (HK-2) were acquired from ATCC (CRL-2190) and grown under standard (37 °C and 5% CO2) conditions. HK-2 cells originate from the renal cortex of a healthly kidney from an adult male donor, immortalized via transduction with human papilloma virus (HPV 16) E6/E7 genes [51]. Cells were cultured in complete Keratinocyte SFM (Serum Free Media) (Gibco, 17005042) supplemented with 5% FBS (Fetal Bovine Serum) (Corning, 35-011-CV) and 1% penicillin/streptomycin (Gibco, 15140122). Cells were harvested at 85% confluency with 0.25% Trypsin-EDTA (Gibco, 25200056) for subculture. For all subsequent experiments, cells were stained with acridine orange (Invitrogen, A1301) and counted using a Cellometer Auto X4 (Nexcelom) to ensure accurate seeding density.

Cell treatment protocol

The following general protocol was used for viability, mitochondrial activity, mitochondrial membrane potential, and reactive oxygen species quantification. A graphical overview has been provided of overall experimental design and specific timelines for each assay (Fig. 2). Cells were seeded at a density of 1.5 × 10⁴ cells per well in a 96-well plate with 100µL of complete supplemented Keratinocyte SFM and allowed to rest overnight. The type of plate used was dependent on assay: black for fluorescence and clear for absorbance. The next day, culture media was replaced with 100µL of media without FBS supplement and serum starved prior to 24-hour groups receiving treatment. This process was repeated 18 h later for the 6-hour group cells and then again 5 h later for the 1-hour group cells. This was done to ensure that each timepoint would have comparable cell growth time and media conditions. Following assay specific procedures outlined below, signal was normalized to healthy control cells and presented relative to control. Each figure is comprised of at least three independent experiments with a minimum of three technical replicates per experiment.

Fig. 2

Graphical overview of experimental timeline. Treatment timing and specific timepoints included in each assay parameter are included as discussed in detail within the methodology. Figure was created via Biorender.com

Viability

A LIVE/DEAD™ Viability/Cytotoxicity Kit (Invitrogen, L3224) was used to quantify viable cells. Following treatment, media was removed, cells were washed with DPBS (Gibco, 14190144), and 100µL of DPBS supplemented with 2µM calcein AM and 4µM ethidium homodimer-1 was added to each well. Plates were then incubated in the dark for 30 min at room temperature before fluorescence signal was read on a microplate reader (BioTek Synergy HT) at a wavelength of 494/517nm for calcein to determine live cells and 528/617nm for ethidium homodimer-1 to determine dead cells. Cell free wells containing supplemented DPBS were used to calculate and subtract background signal.

Mitochondrial dehydrogenase activity

A CellTiter 96^® Aqueous One Solution Cell Proliferation Assay (Promega, G3580) was used to determine mitochondrial activity. Following treatment, media was removed, cells were washed with PBS (Gibco, 10010023), and 100µL of DPBS supplemented with 20µL of the MTS solution was added to each well. Plates were then incubated in the dark for 30 min at 37 °C before absorbance was read on a microplate reader (BioTek Synergy HT) at a wavelength of 490 nm. Cell free wells containing MTS supplemented media were used to calculate and subtract background signal.

Real-time mitochondrial membrane potential

Image-iT™ TMRM Reagent (Invitrogen, I34361) was used to determine dynamic changes to mitochondrial membrane potential. Media was removed from cells and 100µL of phenol red free media containing 100nM TMRM was used to stain the cells. Plates were then incubated in the dark for 30 min at 37 °C. Staining media was removed, cells were washed with DPBS, and 100µL of phenol red free treatment media was added. Plates were then read on a pre-heated 37 °C microplate reader (BioTek Synergy HT) at 548/574nm every 5 min for 3 h. Cell free wells containing treatment media were used to calculate and subtract background signal.

Reactive oxygen species generation

CellROX™ Green Reagent (Invitrogen, C10444) is cell permeable dye which emits a strong fluorescent signal following oxidation. It is not specific to any particular species (e.g., hydroxyl radical, superoxide, etc.) and was used to characterize general levels of reactive oxygen species (ROS). MitoSOX™ Red Reagent (Invitrogen, M36008) is a mitochondrially targeted dye with increased susceptibility to oxidation by superoxide, increasing in fluorescence following oxidation, and was used to characterize the mitochondrial redox environment. Following treatment, media was removed, cells were washed with PBS (Gibco, 10010023), and 100µL of HBSS (Gibco, 14025092) supplemented with 5µM CellROX™ or MitoSOX™ and 0.1 µg/mL Hoechst 33,342 (Invitrogen, H3570) was added to each well. Plates were then incubated in the dark for 30 min at 37 °C. Staining media was removed, cells were washed with HBSS, and 100µL of HBSS was added. Plates were then read on a microplate reader (BioTek Synergy HT) at 485/520nm and 361/486nm. Cell free wells were used to calculate and subtract background signal. Hoechst 33,342 was used to normalize signal to account for cell death.

Cell imaging

To acquire high quality images, the general procedure as above was used, with some slight modifications. Briefly, cells were seeded on black 96-well plates (Revvity, 6005182). Following treatment and wash step, cells were stained with 100µL of HBSS containing 5µM MitoSOX™/CellROX™, 0.1 µg/mL Hoechst 33,342, and CellMask™ Deep Red Actin Tracking Stain (Invitrogen, A57245) for 30 min in the dark. Staining buffer was removed, cells were washed with HBSS, and 100µL of HBSS was added. Plates were then imaged with an Opera Phenix Plus High-Content Screening System (Revvity, HH14001000).

Real-time cellular metabolism

Cell metabolism analyses (mito stress test and glycolytic function test) were performed using a Seahorse XFe96 Analyzer (Agilent). Analyses were performed as previously described following the addition of an acute (real-time) injection [25]. Additional detail can be found in the supplementary methods (Supplemental File 4). Cells were seeded at a density of 1.5 × 10⁴ cells per well in a XF Pro Cell Culture Microplate (Agilent, 103,792–100) with 100µL of complete supplemented Keratinocyte SFM and allowed to rest overnight. The next day, culture media was replaced with 100µL of media without FBS supplement and serum starved prior to treatment. The day before the assay, an XF pro sensor cartridge (Agilent, 103,792–100) was filled with 200µL double distilled H2O (ddH2O) and placed in a non-CO2 incubator overnight. DdH2O was removed and replaced with 180µL seahorse calibrant 1 h prior to running the cell metabolism assay. Following completion of cell serum starvation, media was removed, wells were washed, and then filled with 180µL of supplemented Seahorse XF DMEM media (Agilent, 103,575–100) specific to the type of test. Each plate was read on a brightfield plate reader (BioTek Cytation 5) to determine cell count. All measurements were normalized to cell count and each parameter was then normalized to healthy control cells and presented relative to control. Each figure is comprised of three independent experiments with at least three technical replicates per experiment.

Labeling of reduced cys proteome

Cells were seeded at a density of 3 × 10⁵ cells per well in a 6-well plate with 2mL of complete supplemented Keratinocyte SFM and allowed to rest overnight. The next day, culture media was replaced with 2mL of media without FBS supplement and serum starved prior to 6-hour groups receiving treatment. This process was repeated 5 hours later for 1-hour groups to ensure that total growth and serum starvation time was the same across all groups by endpoint measurement. Following completion of treatment, media was removed, and wells were washed with 500µL of cold PBS. 200µL of a homogenization buffer (pH 7.4) containing 25mM Tris (Invitrogen, 15504020), 0.5mM EDTA (Invitrogen, 15575020), 0.5% v/v NP-40 (Thermo Scientific, 85124), 1X Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific, 78441), 200units/mL Catalase (Sigma-Aldrich, SRE0041), and 100µM Iodoacetamide Alkyne (Vector Laboratories, CCT-1568) was added to each well to label reduced thiols during cell lysis. Cells were scraped on ice, added to tubes, and lysed via sonication before being incubated for 1 h at 25 °C in the dark. 8mM Dithiothreitol (Invitrogen, 15508013) was added to tubes and incubated in the dark for 15 min at 75 °C to quench the reaction and reduce unlabeled thiols. 100mM Iodoacetamide (Thermo Scientific, 122270250) was then added to the tubes and incubated at 25 °C in the dark for 20 min. Protein was precipitated with a 4:1 methanol: chloroform solution for 1 h at -20 °C prior to being centrifuged at 1,700 g for 20 min at 4 °C. Samples were washed with 1:1 methanol: chloroform, the protein pellet was resuspended with 50mM ammonium bicarbonate (Thermo Scientific, 370930250) and 0.1 M urea (Invitrogen, AM9902), and a BCA assay (Thermo Scientific, 23225) was used to determine protein concentration.

Immunoblotting

PRX-1/3 oxidation ratio

Cells were seeded at a density of 3 × 10⁵ cells per well in a 6-well plate with 2mL of complete supplemented Keratinocyte SFM and allowed to rest overnight. The next day, culture media was replaced with 2mL of media without FBS supplement and serum starved prior to 24-hour groups receiving treatment. This process was repeated 18 h later for the 6-hour group cells and then again 5 hours later for the 1-hour group cells. Following treatment, media was removed, wells were washed with warm PBS, and then 200µL of warm buffer containing 100mM NEM (Thermo Scientific, 23030) was added to the well and allowed to incubate at room temperature for 10 min. Cells were lysed by the addition of 1% w/v SDS, scrapped on ice, and lysate was collected. Protein concentration was determined via BCA and 20 µg was loaded onto a non-reducing, denaturing 12% polyacrylamide gel. Gels were run at 120v for 90 min before being transferred to a nitrocellulose membrane by TransBlot Turbo (BIO-RAD) using the mixed molecular weight setting. A Ponceau stain was used to confirm transfer. Membranes were blocked with 5% NFDM (BIO-RAD, #1706404) at room temperature for 1 h and then incubated with either PRDX1 (proteintech, 15816-1-AP) or PRDX3 (proteintech, 10664-1-AP) at a 1:2000 dilution overnight at 4 °C. The next day, membranes were washed, incubated with anti-rabbit HRP linked secondary antibody (Invitrogen, # 31460) at a 1:10,000 dilution for 1 h at room temperature. Membranes were developed via enhanced chemiluminescence (Thermo Scientific, 32106), imaged on a ChemiDoc system (BIO-RAD), and band intensity was quantified via ImageJ.

Reduced cysteine quantification

10 µg of labeled protein samples were dried in a speed vac without heat. Protein was resuspended in 15µL of 50mM Tris-HCl (Thermo Scientific, J22638.AE) and 0.5% w/v SDS (Invitrogen, 15525017) at pH 6.6. A click reaction was performed using the following reagents: 2mM N3-UV-biotin (Kerafast, EVU102), 2.5mM ascorbic acid (Thermo Scientific, A15613.36), 0.1mM TBTA (Thermo Scientific, H66485.MD), and 1mM copper sulfate pentahydrate (Thermo Scientific, A11262.0B). Samples were incubated in the dark at 25 °C for 2 h with constant rotation. Samples were denatured with 4x Laemmli sample buffer (BIO-RAD, #1610747) before being loaded into a reducing, denaturing 10% polyacrylamide gel and run at 120v for 90 min. Gels were transferred to a nitrocellulose membrane using a TransBlot Turbo using the mixed molecular weight setting. A Ponceau stain was used to confirm transfer and normalize protein loading across samples. Membranes were blocked with 5% NFDM at room temperature for 1 h and then incubated with anti-biotin, HRP-linked antibody (Cell Signaling, #7075) at a 1:1000 dilution overnight at 4 °C. Membranes were developed via enhanced chemiluminescence, imaged on a ChemiDoc system, and band intensity was quantified via ImageJ.

Proteomics

Click chemistry and sample cleanup

300 µg of labeled protein per sample was digested with 6 µg of sequencing grade trypsin (Promega, V5113) overnight at 37 °C with constant rotation. Samples were centrifuged at 17,000 g for 10 min, resuspended with an additional 3 µg of sequencing grade trypsin, and incubated for 2 h at 37 °C with constant rotation. Samples were centrifuged at 17,000 g for 20 min and supernatant was collected. Samples were desalted using HLB columns (Waters Corp, WAT094225), washed with ultrapure water (Thermo Scientific, 022934.K7), and eluted with 80% v/v acetonitrile (Thermo Scientific, 610010040) and 10% v/v methanol (Thermo Scientific, 610090040). 10% of peptides were set aside for general proteomic analysis and all samples were dried via vacuum centrifugation. Samples were resuspended in 60µL 30% acetonitrile (pH 6.0). The following reagents were added in order to perform the click reaction: 1mM N₃-UV-biotin, 10mM sodium ascorbate (Thermo Scientific, AAA1775936), 1mM TBTA, and 10mM copper sulfate pentahydrate. Samples were incubated in the dark at 25 °C for 2 h with constant rotation. The reaction was quenched with 4x volumes of SCX buffer (25% acetonitrile and 5mM sodium phosphate monobasic (Thermo Scientific, 447982500) at pH 3.0) and then centrifuged at 17,000 g for 5 min. Supernatant was collected, cleaned with SCX columns (Thermo Scientific, 90008), washed with SCX buffer, and eluted with SCX buffer containing 0.4 M sodium chloride (Thermo Scientific, 447302500). Samples were diluted with 50mM sodium acetate (Thermo Scientific, A13184.30) at pH 4.5 and allowed to bind to streptavidin beads (cytive, 28985738) overnight at 4 °C with constant rotation. These beads were washed 50mM sodium acetate buffer twice, washed with the same buffer containing 2mM sodium chloride twice, washed with water twice, and resuspended in 500µL of 25mM ammonium bicarbonate (Thermo Scientific, 393212500). Samples were transferred to a glass scintillation vial, exposed to 365 nm UV light for 2 h with vortex mixing every 15 min. Supernatant was collected, desalted with HLB columns, and dried with vacuum centrifugation. Samples underwent a final cleanup with C18 spin tips (Thermo Scientific, 84850) and were stored at -80 °C until nHPLC-MS/MS analysis.

General protein quantitation (GPQ) analysis by LC MS/MS

Samples were analyzed using LCMS methods similar to a previous publication [52]. The method is discussed in detail in the supplementary methods (Supplemental File 4).

Click chemistry analysis by LC MS/MS

Samples were re-suspended in 3% Acetonitrile 0.1% formic acid solution. Half of the sample was injected and LC MS/MS data was acquired using the same method parameters as the GPQ data acquisition with the following exceptions: A shorter gradient was used for peptide separation, the ion trap scan rate was set to Normal, the mass range used was 350–1700 m/z, the quad isolation window was set to 1.2 m/z, and the intensity threshold cut-off was raised to 15,000 counts. The shorter gradient used is described as follows: Samples were loaded onto the trapping column at 5.0 µL/min for 5 min at initial condition before being chromatographically separated at a flow rate of 300 nl/min using a gradient of 3–7% B over 3 min, 7–22% B over 32 min, and 22–33% B over 6 min for a total 41 min gradient at 40⁰C. The gradient method was followed by a column wash at 70% B for 4 min.

Data were searched and extracted using the same parameters as general protein quantitation data with the following exceptions: The samples were searched with a fully tryptic enzyme search with 2 missed cleavages allowed, and instead of an alkynl (C) dynamic modification allowed a dynamic cysteine click C12 modification was allowed with a mass shift of 252.122 da. After data was searched, extracted and aligned the Cysteine Click C12 containing peptides abundances were adjusted for protein abundance differences between each sample from the GPQ data set. Control Sample 15 was used as the reference data file to generate the GPQ normalization factors for each protein within each GPQ data file. 5980 Click C12 peptides came from proteins that had GPQ values in every corresponding data file and 1602 Click C12 peptides came from proteins that were either not detected or the protein was not detected in at least one sample of each Sample group in the GPQ data set and therefore these peptides were not normalized. There were 338 other click C12 peptides that were from proteins where there was at least one sample in each sample group with a GPQ value and therefore the Average protein abundance of each group was used to the adjust the abundances of these peptides using the control group as the reference group for generating the GPQ normalization factors.

Statistical analysis was then performed in mass profiler professional software V. 15.1 (Agilent Technologies) on adjusted Click C12 peptide data using an ANOVA with benjamini Hochberg multiple testing corrections and a Tukey HSD post hoc test with a fold change cut-off of|2.0|.

Pathway analysis

General and redox proteomic data was analyzed using Ingenuity Pathway Analysis (IPA) software (QIAGEN) to identify significantly impacted canonical and toxicity associated pathways. Data was condensed to the protein level by combining modified peptide abundance values and analysis was performed on proteins based on abundance fold change with a Tukey HSD post doc p value cutoff of 0.05. A Benjamini–Hochberg corrected p value (q value) < 0.05 was used to identify significantly enriched pathways to account for multiple testing and control false discovery rate. General proteomics includes a z-score, determined by IPA, to denote activation or inhibition of specific pathways where available. Scores above 0 indicate a pathway is activated and below 0 indicate inhibition, a change ≥ 2 is considered significant by IPA. Database for Annotation, Visualization and Integrated Discovery (DAVID) version 2021 cellular component functional annotation analysis was used to identify subcellular location for significantly affected proteins across treatment groups [53].

Statistical analysis

All other statistical analyses and graphics were performed and generated in GraphPad Prism (Version 10.2.3). Statistical comparisons were done using a one-way ANOVA, Tukey multiple comparison test, and Dunnet T3 multiple comparison test with a p-value threshold of 0.05 used to determine statistical significance.

Results

SAD SiNPs alter mitochondrial function and promote glycolytic shift prior to cell death

Treatment of HK-2 cells with SAD SiNPs resulted in a time and dose dependent reduction in cell viability (Fig. 3A) preceded by an even greater decrease in mitochondrial activity (Fig. 3B). Pristine 200 nm SiNPs did not result in any significant changes to viability or mitochondrial activity, but did reduce mitochondrial membrane potential (Fig. 3C). Regarding membrane potential, treatment with SAD SiNPs resulted in the opposite response, hyperpolarizing the mitochondrial membrane in a dose dependent manner. Data in Fig. 3A-C clearly show that SAD SiNPs have a significant impact on the mitochondria; therefore, we next employed Seahorse extracellular flux analyses to evaluate mitochondrial oxygen consumption.

To investigate mitochondrial alterations, we first studied the impact of particle exposure on simple mitochondrial oxygen consumption without manipulation by inhibitors. Following particle injection, Seahorse analyses of cells exposed to SAD SiNPs demonstrated a rapid, dose-dependent reduction in mitochondrial respiration (Fig. 3D). When analyzed using a mitochondrial stress assay, cells maintained a strong oligomycin and FCCP response (Fig. 3E) at lower concentrations of SAD SiNPs (2.5 µg/mL), demonstrating expected changes to oxygen consumption rate (OCR). Increasing this to 25 µg/mL resulted in a severe blunting to the oligomycin response, but cells did still maintain ~ 50% of the expected response to FCCP. Cells treated with pristine 200 nm SiNPs did not demonstrate any significant changes in OCR. Glycolytic function tests (Fig. 3F) indicated an increase in extracellular acidification rate (ECAR) following particle injection in all groups. This response was greatest among cells treated with SAD SiNPs, however ECAR began to drop after 40 min in cells treated with 25 µg/mL. Treatment with oligomycin had minimal effect on ECAR in cells treated with 2.5 µg/mL SAD SiNPs (Fig. 3G) but was followed by a precipitous drop in cells treated with 25 µg/mL.

Fig. 3

Cytotoxicity and cellular energetic assays. SAD SiNPs demonstrated rapid cytotoxic effects on HK-2 cells (A) which were preceded by significant reductions to mitochondrial activity (B) and hyperpolarization of mitochondrial membranes (C). Although pristine 200 nm SiNPs had minimal effects on viability and mitochondrial activity, treatment did result in some degree of mitochondrial membrane depolarization. SAD SiNPs were found to inhibit mitochondrial respiration (D) as soon as 3 min post injection. Cells treated with higher concentrations (25 µg/mL) demonstrated a loss of oligomycin response (E) while still retaining some degree of response to FCCP, potentially indicating a loss of ATP production dependent on mitochondrial membrane permeability. All treatments resulted in increased glycolytic activity (F), while SAD SiNP exposure also resulted in a loss of glycolytic reserve (G) as indicated by a reduced/reversed response to oligomycin. Data is presented as mean and SEM with values normalized to vehicle treated control cells in figures A-C and normalized to cells prior to treatment in figures D-G (N = 3). Data was analyzed with one-way ANOVA (P value: <0.0001) and Dunnett’s multiple comparison test for group-to-control comparisons (**** indicates a P value: <0.0001). In time course experiments, **** indicates the earliest timepoint that experienced significant differences

SAD SiNPs induce a ROS response and prompt PRX-1/3 shift to an oxidized state

Treatment of HK-2 cells with SAD SiNPs increased general ROS levels (CellRox) over time up to ~ 150%. Mitochondrial ROS (MitoSox) was also significantly increased in SAD SiNP treated cells and was disproportionate to general ROS, reaching ~ 600% of control levels by 6 h of 25 µg/mL, further supporting a mitochondrial etiology for SAD SiNP toxicity. Treatment of HK-2 cells with pristine 200 nm SiNPs resulted in a slight decrease in ROS generation (Fig. 4A) by 6 h, which returned to baseline by 24 h; and a brief increase in mitochondrial superoxide (Fig. 4B) which again returned to baseline by 24 h. Immunoblot analysis of cytosolic PRX-1 (Fig. 4C) demonstrated a slight trend towards dimerization following treatment with SAD SiNPs, indicating some degree of oxidation, but mitochondrial PRX-3 was found to respond more strongly (Fig. 4D) with a 16-fold increase in dimerization relative to control. Treatment with pristine 200 nm particles resulted in minimal changes.HK-2 cells exposed to SAD SiNPs demonstrated a time and treatment dependent decrease in the amount of reduced cysteines (Supplemental Fig. 1). Treatment with 5mM TCEP (a potent reducing agent) for 1 h resulted in an increased amount of reduced cysteines, confirming the labeling protocol was performing as expected. Based on these results, the higher dose of 25 µg/mL SAD SiNPs was selected for subsequent proteomic analyses.

Fig. 4

Cellular redox environment changes across treatment groups. SAD SiNPs induced a strong general ROS response (A) which was found to be disproportionately high in mitochondria (B). Higher concentrations of SAD SiNPs were also found to induce greater morphological changes in HK-2 cells. Treatment with SAD SiNPs promoted changes in PRX1 indicative of increased oxidation in the cytosolic compartment (C). PRX3 results (D) indicate that this response was particularly pronounced in the mitochondria. Pristine 200 nm SiNPs induced minor changes in early timepoints which returned to control levels by 24 h. Data is presented as mean and SEM (N = 4). Data was analyzed with one-way ANOVA (P value: <0.0001) and Dunnett’s multiple comparison test for group-to-control comparisons (* indicates a P value: <0.05). Imaging was performed in live cells with the following stains: CellROX (green) for general ROS, MitoSOX (red) for mitochondrial superoxide, Hoechst 33,342 (blue) for nuclei, and CellMask (yellow) for F-actin. Images are representative of 3 independent experiments and band intensity was quantified via ImageJ. SAD HD: SAD High Dose (25 µg/mL)

SAD SiNP treatment results in distinct general proteomic trends over time

Following cellular SiNP treatment for 1 to 6 h, quantitative proteomic analyses detected 4,368 proteins found in at least 1 treatment condition. 3,322 proteins had 2 or more unique peptides, and 2,625 proteins underwent ANOVA with a Tukey HSD post Hoc (p-value cutoff of 0.05) testing and were selected for subsequent analysis (Supplemental File 3). Principal component analysis (PCA) demonstrated clear group separation (Fig. 5A) between cells treated with SAD SiNPs for 6 h and all other groups. Ingenuity toxicity pathway analysis (Fig. 5B) found significant changes in pathways associated with cellular defense signaling and energy metabolism in cells following 1 h of treatment with pristine 200 nm SiNPs, most of which were no longer significantly affected by 6 h. Treatment with SAD SiNPs did not result in any significant changes to toxicity pathways by 1 h but did result in deleterious changes by 6 h. Affected pathways included those associated with mitochondrial dysfunction, cell death, energy metabolism, oxidative stress, aryl hydrocarbon receptor signaling, and peroxisome proliferator-activated receptors (PPAR). The top 30 most significantly altered proteins (determined by ANOVA with Tukey HSD) that were present across all sample groups (Fig. 5C) included those associated primarily with cell adhesion, remodeling, cell cycle oversight, transcription, mitochondrial transport, ubiquitination, and heat shock proteins. Analysis of IPA canonical pathways that were most significantly affected (Table 1) found further signs of increased transcription, perturbed cellular energetics, and drastically increased SUMOylation pathway activation. The complete list of canonical pathways identified can be found in the supplemental materials (Supplemental Table 1).

Fig. 5

General proteomic analysis. Principal component analysis (PCA) plot of sample groups (A) showed a high degree of clustering among control and all treatment groups other than 6-hour SAD SiNPs. IPA toxicity analysis (B) shows significantly modified pathways primarily included those involved in cellular defense and energy production in pristine 200 nm SiNPs. SAD SiNP treated cells did not show significant changes by 1 h but did result in major indications of cytotoxicity by 6 h. The most significantly altered proteins (C) include a variety of proteins central to cell signaling networks. Original proteomic data was analyzed with an ANOVA and Tukey Honestly Significant Difference (HSD) test with a p value threshold of 0.05 to determine significance for subsequent analysis (N = 3). Dotted lines indicate the point of statistical significance (-log10 q value of 1.3)

Table 1 IPA analysis of canonical pathways in general proteomic samples. Z-score is a measurement of activation or inhibition of this pathway, with 2 and − 2 considered significant alteration respectively. A Benjamini-Hochberg corrected p value (q value) was used to correct for false discovery rate and -log10 converted such that a q value ≤ 0.05 is considered statistically significant. Bolded values are indicative of statistical significance

SAD SiNPs alter the redox proteome

Redox proteomic analysis detected 3,026 proteins in 100% of at least 1 treatment condition, 1,872 had 2 or more unique peptides, and 823 proteins underwent an ANOVA with a Tukey HSD post Hoc (p value cutoff of 0.05) and were selected for subsequent analysis (Supplemental File 3). Redox affected proteins were mapped on a volcano plot (Fig. 6A) by significance and fold change, with increasing fold change corresponding to tendency for Cys reduction and decreasing fold change corresponding to Cys oxidation. Treatment with pristine 200 nm SiNPs resulted in relatively few overall proteins significantly affected, with an early shift from reduction at the 1-hour timepoint to more oxidation by 6 h. Treatment with SAD SiNPs caused an expected oxidative shift by 1 h, but with more reduction than was observed in 6-hour pristine 200 nm SiNPs. By 6 h this perturbation became even more severe, with an increase in significantly oxidized and reduced proteins in SAD SiNP treated cells indicative of pathologically-relevant shifts in redox proteome state. The 30 most significantly affected proteins present in all groups and their respective cysteine sites that underwent modification were sorted on a heat map (Fig. 6B), revealing that most significant changes were primarily oxidation of molecules associated with transcription, translation, and DNA damage sensing. Functional annotation via DAVID identified subcellular location across significantly impacted proteins that underwent a 2-fold change relative to control (Table 2). Distribution of greatest redox changes were unevenly distributed across subcellular locations, with nucleus, mitochondrion, and cytoskeleton-associated proteins found to be among the most prominent (Fig. 6C).

Table 2 Database for annotation, visualization and Integrated Discovery (DAVID) functional annotation analysis of significantly oxidized or reduced proteins that experienced a ≥ 2 fold change relative to control and their reported subcellular location. The total column indicates the number of proteins found in each respective sample group

Fig. 6

Redox proteomic analysis. Proteins found in each treatment group were filtered by p value/fold change and visualized via volcano plot (A) to demonstrate protein tendency to be oxidized (decreased fold change) or reduced (increased fold change) relative to control samples. The most significantly altered proteins and specific cysteine sites (B) consisted primarily of oxidized proteins associated with inflammatory signaling and ubiquitination. The subcellular location of significantly affected proteins (C) indicated uneven distribution across compartments, particularly the mitochondria. Redox proteomic data was normalized to general proteomic data at the protein level to account for changes in abundance unrelated to redox. Data was analyzed with an ANOVA and Tukey Honestly Significant Difference (HSD) test with a p value threshold of 0.05 to determine significance for subsequent analysis (N = 3)

SiNPs alter redox pathway signaling prior to general proteomic changes

IPA toxicity pathway analysis identified significant proteomic changes in cell death and aryl hydrocarbon receptor signaling (AHR) among all SiNP treated cells (Fig. 7). Treatment with SAD SiNPs for 1 h (Fig. 7B) resulted in an overall oxidative shift, with significant pathways including AHR, cell death, and proliferation pathways. These pathways remained significantly altered by 6 h, becoming markedly more reduced. At the 6-hour timepoint there were several more pathways that underwent significant redox alteration, including mitochondrial dysfunction and oxidative stress, in line with observed functional changes. Changes observed in cells treated with pristine 200 nm SiNPs were comparatively minor, with early changes observed at 1 h including primarily reduced proteins that dropped below significance by 6 h (Fig. 7A). Analysis of canonical IPA pathways identified additional pathways (Table 3) experiencing significant redox differences. Included pathways ranged from RNA processing, changes to energy metabolism, stress response, cell cycle modulation, cellular remodeling, and SUMOylation. The complete list of redox canonical (Supplemental Table 2) and redox toxicity (Supplemental Table 3) can be found in supplemental material.

Fig. 7

IPA toxicity pathway analysis of redox proteomics. Treatment with pristine 200mn SiNPs (A) showed minimal significant toxicity pathways, with early redox changes primarily involving reduction of proteins associated with cell death and aryl hydrocarbon receptor (AHR) signaling and falling below significance by 6 h. Treatment with SAD SiNPs (B) resulted in early indications of oxidation, with significant changes in proliferation, cell death, and AHR pathways. Changes became more pronounced by 6 h, including indications of mitochondrial perturbation and oxidative stress response in addition to 1-hour pathways. Original redox proteomic values were normalized to general proteomic results and analyzed via IPA with a q value cutoff of 0.05 to determine statistical significance (N = 3). Dotted lines indicate the point of statistical significance (-log10 q value of 1.3). Green, red, and gray bars are representative of proportional percents of the proteins in the pathway that were found to be oxidized, reduced, or unchanged relative to control respectively

Table 3 IPA analysis of canonical pathways in redox proteomic samples. Molecules in pathway indicates number of proteins known to be associated with the respective pathway. Coverage indicates the percent represented in each sample group. Oxidized and reduced columns show percent of these proteins found to be reduced or oxidized relative to control. A Benjamini-Hochberg corrected p value (q value) was used to correct for false discovery rate and -log10 converted such that a q value ≤ 0.05 is considered statistically significant. Bolded values are indicative of statistical significance

Discussion

SiNPs are a known occupational exposure among sugarcane workers, and these particles have been hypothesized to contribute to the growing epidemic of kidney disease among agricultural communities [24, 25]. SiNPs derived from sugarcane ash have a high potential for generating oxidative stress, a key hallmark of chronic kidney disease [30, 54]. As demand for sugarcane grows and industries consider new commercial applications of sugarcane ash, it is increasingly urgent to understand the potential health burden of such exposures [55,56,57,58]. Unfortunately, current literature regarding SiNP toxicity is primarily based on pristine, homogenous, manufactured particles and may be of limited relevance to exposures of environmentally derived SiNPs [27, 28, 49]. Thus, there is a need to characterize the effects environmentally relevant particles have on cellular redox environments and investigate signaling pathways through which SiNPs may exert deleterious effects. Here we report for the first time that sugarcane ash derived (SAD) SiNPs induced significant oxidative stress and altered the redox state of proteomic signaling pathways with known roles in mitochondrial dysfunction and CKD.

Mitochondrial dysfunction plays a critical role in the pathogenesis of various kidney diseases, with PCT cells particularly susceptible to such injuries due to high density of mitochondria [31, 59, 60]. When such an injury is induced by toxicant exposure, it can result in a substantial loss of mitochondrial energy production [61]. This is accompanied by a compensatory increase in glycolytic activity to mitigate damage due to insufficient ATP production [62]. This glycolytic shift is precisely what is seen in HK-2 cells following exposure to SAD SiNPs. One explanation for this could be inhibition of ATP synthase, a known mechanism of nephrotoxicants, which would explain both the minimal response to oligomycin and mitochondrial hyperpolarization [63, 64]. When PCT cells are unable to recover from injury, such as through continuous SAD SiNP exposure, extended ATP deficiency will result in permanent kidney damage and transition to CKD [65]. Concurrently there is an increase in oxidative stress, generated in part by mitochondrial dysfunction [66]. These trends align with what was seen following exposure to SAD SiNPs, with ROS disproportionately present in the mitochondrial compartment. If persistent generation of ROS is ongoing, it may overwhelm cellular ability to maintain redox homeostasis, as indicated by perturbations to cellular defense systems, such as glutathione or peroxiredoxins [67, 68]. This is in line with previous metabolomic findings, as well as the observed shift in peroxiredoxin towards the oxidized form [25, 29, 69]. Cellular response seems to be centered in the mitochondrial compartment, followed by functional changes in the rest of the cell. This is consistent with our findings that SAD SiNPs induce mitochondrial dysfunction and oxidative stress which, if not resolved by cellular protective systems, can result in perturbed redox signaling, proteomic changes, and cytotoxicity associated with CKD pathophysiology.

Quantitative proteomic analysis was performed to identify changes in total protein abundance and normalize subsequent thiol redox proteomic results. Overall, SAD SiNP treatment of HK-2 cells resulted in significant cytotoxicity, while minimal changes were found following exposure to pristine 200 nm SiNPs. In the case of pristine SiNPs, significantly impacted protective pathways, including modification of retinoid receptor function, oxidative stress response, and xenobiotic metabolism signaling [70]. There were also pathway indications of altered fatty acid metabolism, which may be indicative of increased energy demand to respond to SiNP induced stress. This is consistent with previously discussed indications of slightly altered energetics, increased mitochondrial activity, and heightened mitochondrial superoxide in cells exposed to pristine 200 nm SiNPs. As the majority of these features begin to return to expected values by 6 h of exposure, it appears that cells are responding to SiNP exposure but, in the case of pristine SiNPs, are able to induce an appropriate adaptive response to ensure cell survival. Interestingly, certain pathways demonstrated a trend towards increased significance by 6 h (i.e., oxidative stress and cell cycle regulation). Although these pathways did not reach statistical significance by this timepoint, this change may be indicative of some of the protective pathways contributing to cell survival — cell cycle regulation is essential to allow cells time to repair damage prior to division [71]. This is in stark contrast to the cells which have been exposed to SAD SiNPs for 6 h and demonstrate signs of severe mitochondrial dysfunction, cell death, major oxidative stress, cell cycle regulation and retinoid receptor activation. These widespread regulatory changes are expected in the context of apoptosis induction, but interestingly there are few such proteomics indications at the earlier 1-hour time point. This is notable given that severe mitochondrial impacts were observed within 15 min of SAD SiNP exposure, indicating that early mechanistic initiators are upstream of these pathways.

Thiol redox proteomics provides the opportunity to interrogate subtle changes to thiol redox signaling that may proceed general proteomic trends. While significant alterations in redox state were found across all groups, SAD SiNP treatment resulted in the greatest perturbation, including a large degree of relative reduction indicative of a destabilized redox environment. These changes could be associated with cysteine modifications that drastically alter protein activity even without any indications of altered protein abundance. Although significantly altered proteins covered a wide range of cellular processes (e.g., remodeling, transcription, ubiquitination, transport, etc.), there were several proteins that were identified to play key roles in perturbed pathways and were considered of particular interest. For example, one of the most significantly impacted proteins, Prostaglandin reductase 1 (PTRG1) was specifically reduced on a NADP + binding site, suggesting that treatments may have a major impact on lipid metabolism [72]. Similarly, a major site of oxidation on SUMO-activating enzyme subunit 2 (UBA2) was on the active site of thioester bond formation, suggesting modulation of SUMOylation pathway activity by 1 h of SAD SiNP despite significant pathway changes not being identified at the general proteomic level until the 6-hour timepoint [73, 74]. Identification of cellular sensors that respond to early redox changes is essential for determining mechanisms of toxicity. Another notable site of oxidation was found in Voltage-dependent anion-selective channel protein 3 (VDAC3), a marker of mitochondrial redox status, with an essential role in protecting mitochondria from oxidative stress [75, 76]. This pore forming membrane protein has several cysteine residues in the intermembrane space which are sensitive to oxidation, including the cysteine at site 229, which was found to be highly oxidized following SAD SiNP exposure [77]. VDAC3 is known to have a role in regulating mitochondrial membrane potential, ATP generation, and compartmentalization of ROS to the inner mitochondrial membrane [78]. It has also been identified as central to certain renal injury models and a potential target for therapeutic intervention in CKD [79, 80]. If VDAC3 is responding to the increased ROS generated by SAD SiNP exposure by inhibiting mitochondrial membrane permeability, it may be a component of the cellular defense response to limit mitochondrial contribution to ROS production. This would also be consistent with the apparent ATP synthase inhibition and increased oxidation of mitochondrial inner membrane proteins observed. However, it is important to note that the mechanism of oxidative stress induced damage is wide ranging and likely to include a variety of pathways rather than a single critical one.

Significantly impacted cellular pathways included those that play a role in the development of CKD. For example, transforming growth factor beta (TGF-β) signaling, a central player in driving wound healing, fibrosis, and epithelial mesenchymal transition, was highly represented among redox altered proteins [81]. Among the most significant were DNA-dependent protein kinase catalytic subunit (PRKDC) and promyelocytic leukemia protein (PML). PRKDC alteration is particularly interesting as it was one of the few proteins found to be reduced in all treatments. Beyond its role in driving CKD progression, it has also been found to modulate mitochondrial activity in an oxidative stress-dependent manner via direct interaction with VDACs [82, 83]. PML also demonstrated an interesting redox response, with cysteine site 389 found to be reduced following 200 nm SiNP exposure, but oxidized following SAD SiNP exposure. PML is a modulator of SMAD proteins (involved in TGF-β signaling), but also is involved with aryl hydrocarbon receptor (AHR) signaling and inhibition of peroxisome proliferator-activated receptor gamma (PPARγ) [84,85,86,87]. The nexus of these pathways in this exposure model is not unexpected given that they are known to be influenced by oxidative stress, but their consistent overrepresentation in CKD pathology warrants further investigation in sugarcane ash exposure [88,89,90]. AHR signaling is known to be intertwined with tryptophan metabolism, with 3-Hydroxyanthranilic acid (HAA) an AHR activator [91, 92]. Previous in vitro investigations have found significant alteration of tryptophan metabolism following sugarcane ash exposure and metabolomic investigation of sugarcane workers found evidence of increased HAA specifically in urine of workers with evidence of kidney function decline [25, 29]. Upon activation, AHR signaling induces a variety of downstream changes which may include the inhibition of PPARγ [93,94,95]. PPARγ inhibition results in reduced ATP generation, mitochondrial hyperpolarization, and perturbed lipid metabolism [96, 97]. Over the short term, this can be advantageous for reducing endogenous ROS production and extending survival while cells restore redox homeostasis. However, if prolonged by persistent oxidative stress, such as in the case of SAD SiNP exposure, these changes can lead to a glycolytic shift and progressive fibrotic phenotype [65]. These outcomes, in line with our own findings, are hallmarks of the transition to CKD and why modulation of PPARγ function is a promising potential target for kidney disease therapeutics [98,99,100].

While these findings are of substantial interest and provide valuable information regarding potential mechanisms of toxicity underlying SAD SiNP exposure, there are several limitations which must be acknowledged. The cell model used, HK-2, has been long established as a valuable model for exploring PCT toxicity which retains the phenotype of well differentiated PCT cells [51]. However, it is still an immortalized cell line with limited physiological relevance and is likely to demonstrate somewhat distinct responses from primary cells [101]. Ideally, future studies will be able to make use of kidney organoids or in vivo models to help bridge this gap between experimental and real-world conditions. Analysis of ROS generation is valuable for determining toxicity timelines and targeted subsequent analyses, CellRox and MitoSOX are commonly used for characterizing the redox environment of the cell but have limitations in the specificities of their signals. CellRox is cell permeable, not isolated to a specific compartment, and greatly increases in fluorescence following oxidation — as such, it does not distinguish between oxidizing species. In the case of MitoSOX, beyond the signal generated from the specific superoxide product (Mito-2-OH-E+) there are other products which contribute to overall fluorescent signal (i.e., Mito-HE/E dimers and Mito-HE-E heterodimes) [102]. Thus, it may be more reasonable to view this signal as an indication of oxidant formation rather than solely superoxide quantification. An additional consideration is the potential for SAD SiNPs to act as a carrier for other toxicants. The burning process is known to generate a variety of potentially harmful compounds (i.e., volatile organic compounds and polycyclic aromatic hydrocarbons) which may use SAD SiNP particles as a carrier to evade biological barriers and exert additional toxicity [11]. It is possible that such “contaminants” are key contributors to observed toxicity, particularly in real-world conditions, and require additional compositional analysis to help untangle such factors. Future work may benefit from subjecting pristine synthetic SiNPs to environmental conditions such as burning or ash exposure in order to provide a more relevant control treatment.

This study highlights the importance of using environmentally relevant particles and demonstrates the value of redox proteomics in identifying mechanisms of toxicity. Analysis of the cellular energetics and redox environment following exposure to SAD SiNPs revealed a variety of deleterious effects centered around perturbed energy metabolism and redox signaling which correspond with patterns in kidney disease (CKDu). General protein quantitation revealed minimal variation across treatments until 6 h of SAD SiNP exposure, despite SAD SiNPs inducing severe mitochondrial dysfunction within 1 h. By utilizing a proteomic approach to specifically investigate thiol redox changes, early signaling pathways were able to be identified and linked to observed outcomes. Highly redox affected pathways included known CKD associated pathways such as TGF-β and AHR which exert downstream effects on energy metabolism consistent with outcomes of SAD SiNP exposure (Fig. 8). Future research is needed to investigate mechanisms by which these and associated pathways contribute to the pathogenesis of CKDu in select populations of sugarcane agricultural workers.

Fig. 8

Hypothesized mechanism of oxidative stress induced cellular dysfunction. Exposed silanol groups on SAD SiNPs catalyze the generation of ROS, enhanced by other toxicants present, which induce activation of redox signaling pathways with known roles in inhibiting mitochondrial activity and CKD development. Under conditions of persistent ROS generation this signal will be continuously reinforced, prompting a shift to a glycolytic and fibrotic phenotype. Figure was created via BioRender.com