Aristolochic acid I induces proximal tubule injury through ROS/HMGB1/mt DNA mediated activation of TLRs

Abstract Aristolochic acids (AAs) are extracted from certain plants as folk remedies for centuries until their nephrotoxicity and carcinogenicity were recognized. Aristolochic acid I (AAI) is one of the main pathogenic compounds, and it has nephrotoxic, carcinogenic and mutagenic effects. Previous studies have shown that AAI acts mainly on proximal renal tubular epithelial cells; however, the mechanisms of AAI‐induced proximal tubule cell damage are still not fully characterized. We exposed human kidney proximal tubule cells (PTCs; HK2 cell line) to AAI in vitro at different time/dose conditions and assessed cell proliferation, reactive oxygen species (ROS) generation, nitric oxide (NO) production, m‐RNA/ protein expressions and mitochondrial dysfunction. AAI exposure decreased proliferation and increased apoptosis, ROS generation / NO production in PTCs significantly at 24 h. Gene/ protein expression studies demonstrated activation of innate immunity (TLRs 2, 3, 4 and 9, HMGB1), inflammatory (IL6, TNFA, IL1B, IL18, TGFB and NLRP3) and kidney injury (LCN2) markers. AAI also induced epithelial‐mesenchymal transition (EMT) and mitochondrial dysfunction in HK2 cells. TLR9 knock‐down and ROS inhibition were able to ameliorate the toxic effect of AAI. In conclusion, AAI treatment caused injury to PTCs through ROS‐HMGB1/mitochondrial DNA (mt DNA)‐mediated activation of TLRs and inflammatory response.

urinary tract uroepithelial cancer. 5,10 The incidence of AA-induced nephropathy (AAN) is not limited to any specific geographic region and cases occur regularly throughout the world, despite warnings from many health agencies. Kidney injury and fibrosis are predictable features of AAN, such that it is now used as a kidney injury/ fibrosis model in vitro and in vivo studies. [11][12][13] Although studies have provided considerable insights into mechanisms of AA-induced nephropathy, the pathophysiology of AAN remains incompletely understood. [14][15][16][17][18][19] The toll-like receptors (TLRs) are pattern recognition receptors (PRRs) that have a unique and essential function in innate immunity.
There are ten known TLRs in humans 20  Endogenous damage signals can stimulate an inappropriate innate immune response that can initiate a pro-inflammatory cytokine cascade. 21,22 The presence of extensive fibrosis suggests a significant role for inflammatory pathways. Specifically, the role of TLRs, mediators of innate immunity, is unknown. Since the TLRs act as first responders for danger signals and regulate downstream inflammatory and profibrotic cytokines, we investigated the role of TLRs in AAI-induced HK-2 cell damage.
We exposed cells to AAI at varying doses and time intervals. We collected cell supernatants for ELISA and cell pellets for gene expression studies. Based on our dose-time screening data, we used a concentration of AAI (40 μM) for an exposure time of 24 h for most of the experiments.

| Cell proliferation assay
We used CellTiter 96 AQueous One Solution Cell Proliferation Assay (MTS) (Promega, catalogue G3580) to check proliferation of HK2 cells. We seeded the cells with the density of 10 K cells per well in 96-well plates and exposed them to AAI for the indicated time periods followed by the MTS assay according to the manufacturer's instructions.

| ApoToxGloTM triplex assay
We purchased the ApoToxGloTM Triplex Assay from Promega (Fitchburg, WI, USA, catalogue G6320) and performed according to the manufacturer's instructions.

| Flow cytometry
We used Beckman coulter kit for propidium iodide (PI) staining and Annexin V FITC kit (sigma) for apoptosis assay. We followed kit protocols as per manufacturer's instructions. For analysing flow cytometry data, we used ModFitLT software (verity Software House, ME, USA) and Kaluza flow analysis software (Beckman Coulter, Inc., CA, USA). Cell cycle and apoptosis were analysed in control and AAexposed samples at different time intervals.

| Real-Time quantitative polymerase chain reaction (qPCR)
We isolated total RNA from cultured HK2 cells using the RNeasy and estimated comparative CT values. We calculated relative gene expression through 2 -(ΔΔCT) method and expressed in arbitrary units (a.u.) relative to paired controls. We generated heatmaps of gene expressions by using the 'Heatmapper' web server. 27

| Western blotting
We isolated protein from cells and quantified it by using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, catalogue 23,225). We We blocked the membranes with PBS, TBS or 5% milk in Tris-Buffered Saline-Tween (0.05%) solution based on primary antibody-specific recommendations of the manufacturer. After overnight incubation with primary antibody, we incubated the blots with LI-COR secondary antibody (Odyssey IRDye 680RD, catalogue 926-68,070 or 800CW, catalogue 926-32,211) for 1 h. We visualized Western blots by Odyssey CLx Imaging System (LI-COR Biotechnology) using near-infrared fluorescence capture. We performed Western blot image analysis using Image Studio software (LI-COR Biotechnology) to obtain the integrated intensities. We analysed data after normalizing with endogenous control protein (β-actin).

| ELISA
We determined the levels of IL6 (Human IL-6 ELISA Ready-SET-Go,

paraformaldehyde-fixed HK2 cells. We permeabilized cells with
Triton X-100 (0.1%) and BSA (2%) in PBS for intra-cellular proteins but not for cell surface proteins. We blocked cells in normal goat serum (Cell Signalling Technology, catalogue 5425S) plus 0.1% Triton X-100 (PBST) for 1 h. We incubated slides in LCN2 primary antibody

| Reactive oxygen species (ROS) generation assay
We estimated general oxidative stress and production of ROS in cells using the 5 μM of the redox-sensitive fluorescent probe chloromethyl derivative of 2′,7′-dichlorodihydrofluorescein diacetate (CM-H2DCFDA; catalogue #c6827, Thermo Fisher Scientific; Pittsburgh, PA, USA). This probe freely enters the cell and is hydrolysed by intracellular esterase into its non-fluorescent form 2′,7′-dichlorofluorescin (DCFH). In the presence of ROS, DCFH is oxidized into the highly fluorescent compound DCF (2′,7′-dichlorofluorescein), which is no longer membrane permeant. We seeded HK2 cells (3 × 10 5 ) and allowed to attach for 24 h in 6-well plates. After replacing culture media by fresh media, we exposed cells to AAI (40 μM for 2, 6 CTCF values were calculated as described previously.

| TMRE assay
To quantify changes in mitochondrial membrane potential in HK2 cells, we used TMRE-Mitochondrial Membrane Potential Assay Kit (Abcam catalogue ab113852) and measured fluorescence by microplate spectrophotometry. To label active mitochondria, we used TMRE (tetramethylrhodamine, ethyl ester, which is a cell permeant, positively charged, red-orange dye that readily accumulates in active mitochondria due to their relative negative charge. Depolarized or inactive mitochondria have decreased membrane potential and fail to sequester TMRE. As a positive control, we used FCCP (carbonyl cyanide 4-[trifluoromethoxy] phenylhydrazone), which is an ionophore uncoupler of oxidative phosphorylation. FCCP eliminates mitochondrial membrane potential and TMRE staining. We seeded HK2 cells (20 × 10 3 ) and allowed to attach for 24 h in 96-well plates.
We then replaced the culture media by fresh media with or with-

| Statistical analysis
We analysed data by comparing mean values through either un-
We quantified apoptosis by the ratio of caspase 3/7 and found that apoptosis significantly increased at and above 40 μM concentration of AA (one-way anova ****p < 0.0001, n = 6; Figure 1M) and at 24 h exposure (one-way anova ****p < 0.0001, n = 6; Figure 1N). Increased AA-induced apoptosis was further confirmed through flow cytometry using propidium iodide (PI) and Annexin V-FITC ( Figure 1O,P). We also found that AA-exposed HK2 cells significantly underwent G2M cell cycle arrest ( Figure 1Q,R). Although AA exposure increased apoptosis and G2/M cell cycle arrest in same time-dependent manner, however, superimposing both data showed that cell cycle arrest at G2/M phase was not the causation of apoptosis initiation ( Figure 1S).

| AAI enhanced expression of TLRs and inflammatory cytokines
We performed candidate gene expression analysis for AKIassociated genes based on our previous publication 23 in HK2 cells exposed to AAI in comparison with unexposed cells (control). We found significantly increased expressions of TLRs (2-4, 9, and) and inflammatory cytokines (TNFα, IL6, IL1B, IL18, TGFβ and NLRP3) in
Immunofluorescence confirmed that another KI marker LCN2, and TLR9 protein expression was increased in HK2 cells after exposure to AAI (40 μM) for 24 h (Figure 3H-K; corrected for total cell fluorescence p < 0.0001). We found that iNOS expression increased in HK2 cells exposed to AAI (CTCF for iNOS, unpaired t-test p < 0.0001; Figure 4C,D) along with a significant increase in NO production ( Figure 4E).

| ROS production and mitochondrial dysfunction in HK2 cells upon AAI exposure
Increased ROS and NO production may cause mitochondrial dysfunction; therefore, we measured mitochondrial membrane potential through TMRE assay and found that AAI (40 μM) exposure causes loss of mitochondrial membrane potential significantly ( Figure 4F; one-way anova Dunnett's multiple comparisons test ****p < 0.0001).

| Pretreatment with 1 mM tempol and TLR 9 knock-down ameliorated AA-induced overexpression of KI markers in HK2 cells
Our data indicated a possible mechanism of AA-induced toxicity in HK2 cells via ROS, mt -DNA, TLR9 axis. Therefore, we tried to res-

cue HK2 cells by inhibiting ROS generation through tempol (1 mM)
pre-treatment and TLR 9 siRNA. We found that both tempol and TLR 9 siRNA ameliorated AA-induced overexpression of KI markers ( Figure 6A). We also found that TLR9 KD and tempol pretreatment can reduce AA-induced mitochondrial membrane potential loss ( Figure 6B) and apoptosis in HK2 cells ( Figure 6C).

| DISCUSS ION
Aristolochic acid nephropathy (AAN) causes a rapidly progressive interstitial nephritis, which leads to end-stage renal disease. is widely present, 30,31 although there is some controversy. [32][33][34] What is noncontroversial, however, is that AA can cause a rapidly progressive kidney disease often associated with a high incidence of upper F I G U R E 2 Candidate gene expressions that contribute to injury (A-O) in HK2 cells exposed to aristolochic acidI (40 μM) compared to unexposed cells (control) showed increased expressions of TLRs (2)(3)(4)9) and inflammatory cytokines (TNFα, IL6, IL1B, IL18, TGFβ and NLRP3)); in AA-mediated nephrotoxicity. 43,44 Recent evidence indicates that AA is concentrated within proximal tubular epithelial cells via the basolateral OAT (OAT1), which exchanges organic anions such as paraaminohippurate (PAH) and AA for α-ketoglutarate. 45,46 The integrative omics analysis clearly demonstrated that AAI damage to HK-2 cells involves three pathways: the amino sugar and nucleotide sugar metabolism pathway, the amino acid biosynthetic pathway and the carbon metabolism pathway. Through catabolism of amino acids, cells can use carbon skeletons to form tricarboxylic acid cycle (TCA) cycle intermediates used in the central metabolic pathway. 42 Carbon metabolism is one of the basic metabolic pathways of organisms and can affect many physiological functions. 47,48 TLRs are well-known for their role as first responders to exogenous/ endogenous damage signals; however, the involvement of TLRs in AAN has not been systematically investigated. In the present study, we demonstrated that AAI exposure increased ROS generation in HK2 cells, which induced mitochondrial dysfunction and activation of endogenous alarmins (DAMPs) such as HMGB1 and mt DNA. TLRs 2, 4 and 9 were activated by their ligands HMGB1 and mt DNA, respectively, and activated downstream inflammatory pathways. ROS inhibition by tempol and TLR9 knock-down through siRNA transiently ameliorated the toxic effect of AAI.
We detected significantly increased mRNA expressions of TLRs 2,3,4 and 9 along with HMGB1 in HK2 cells exposed to AAI. HMGB1 is a known ligand of TLRs 2,4, and previously, we have shown that HMGB1 regulates activation of TLRs 2, 4 and 6 in proximal tubule cells. 23 Therefore, AAI-induced nephropathy may be initiated through HMGB1-mediated TLR activation. Furthermore, in the present study, we identified markedly higher expression of TLR 9 (>30-fold) in AAI-exposed HK2 cells compared to unexposed cells.  52 Since the cells in this study were endotoxin free and experiments were performed in aseptic conditions, we suspected mitochondrial DNA as a ligand for TLR 9 activation and detected mitochondrial membrane potential loss in HK2 cells upon AA1 exposure through TMRE assay. Thus, not only HMGB1 but mt DNA also acts as a DAMP ligand (and probably to a higher extent) to activate TLRs in AAN.
Renoprotective effect of Stat1 deletion was shown in a murine model of aristolochic acid nephropathy. 53 Similarly, our previous study 23 showed STAT1 knockdown can protect PTCs through decreasing HMGB1 release. In the present study, we found that AAinduced significant overexpression of HMGB1 mRNA ( Figure 2B) compared to control HK2 cells along with TLRs 2,3,4,9. This suggests that AA-induced overexpression of HMGB1 and TLRs 2,3,4,9 is mediated via STAT1 activation. A recent report showing the macrophage interferon regulatory factor 4 deletion ameliorates aristolochic acid nephropathy is consistent with our observation on the role of innate immunity in AA nephropathy. 54 The study confirms the expectation that anti-TLR strategy may be of therapeutic value. in AAI-exposed HK2 cells (Figure 4), leading to proximal tubule injury confirmed by the markedly increased expression of kidney injury marker LCN2 in AAI-exposed cells ( Figure 2E, 3H). Proximal tubules contain abundant mitochondria to perform the energy-demanding processes of reabsorption and secretion. Mitochondria are the most active organelles in the cell where approximately 90% of the total oxygen content in the cell is consumed to enable oxidative phosphorylation and ATP synthesis 55 and therefore more susceptible to ROS-mediated injury. Damaged or dysregulated mitochondria F I G U R E 4 ROS production and mitochondrial dysfunction in HK2 cells upon AAI (40 μM) exposure. A: AA (40 μM) increased ROS generation in HK2 cells (20x images) as early as 2 h and persisted for up to 24 h compared with unexposed HK2 cells (control) and H2O2 (50 μM for 10 mins); B: ROS fluorescence measurement data; *p = 0.0127; One-way anova test with Dunnett's multiple comparisons test; iNOS expression and NO production was increased in HK2 cells exposed with AAI (40 μM); C: Immunofluorescence of iNOS (green), phalloidin (red) and nucleus (blue) D: corrected cell fluorescence (CTCF) for iNOS, Unpaired t-test ****p < 0.0001. E: NO production, oneway anova Dunnett's multiple comparisons test **p = 0.0011, ****p < 0.0001; N = 6; F: AA (40 μM) exposure causes loss of mitochondrial membrane potential (TMRE assay); one-way anova with Dunnett's multiple comparisons test ****p < 0.0001; N = 6; Comparison of immunofluorescence between control vs AA (40 μM, 24 h) exposed HK2 cells: GRP75 (G), CTCF for GRP75, unpaired t-test ***p = 0.0003 (M); MTCO2 (H), CTCF for MTCO2, unpaired t-test ****p < 0.0001 (N); CHOP (I), CTCF for CHOP, unpaired t-test **p = 0.0051 (O), Beclin1 (J), CTCF for Beclin1, unpaired t-test ***p = 0.0003 (P); LC3A/B (K), CTCF for LC3A/B, unpaired t-test ***p = 0.0008 (Q) and GRP78 (L), CTCF for GRP78, unpaired t-test *p = 0.017 (R) F I G U R E 5 AAI induces fibrosis/ epithelial-to-mesenchymal transition (EMT) in HK2 cells in 24 h; A: AA (40 μM) increased α-SMA expression (green) in HK2 cells compared to unexposed HK2 cells; B: E-cadherin expression decreased in AA-treated cells; C: corrected total cell fluorescence (CTCF) for α-SMA in AA-exposed vs unexposed HK2 cells, Unpaired t-test **p = 0.0035; D: CTCF for E-cadherin in AAexposed vs unexposed HK2 cells, Unpaired t-test *p = 0.0121 | 4289 UPADHYAY AnD BATUMAn generate excessive ROS and start a chain reaction damaging other mitochondria. 56 NO is a reactive nitrogen species (RNS) that is generated from injured mitochondria by inducible isoform of nitric oxide synthase (iNOS or NOS2) enzyme, and it can alter respiration, mitochondrial biogenesis and oxidative stress through increased production of ROS/RNS and thus can impact cell physiology. 56,57 Our data suggest that AAI induces DAMPS that bind to specific TLRs and initiate a pro-inflammatory cascade by damaging mitochondria and generating ROS as well as RNS.
Additionally, we found that AA exposure leads to decreased expression of E-cadherin and increased expression of α-SMA in PTCs, indicating phenotypic transformation of epithelial cells to a myofibroblast phenotype through epithelial-to-mesenchymal transition (EMT). Although fibrosis has generally been associated with increased expression of TGFβ-dependent ECM components, 58 it can also be linked to overexpression of TLR 9. TLR 9 depletion can attenuate renal tubulointerstitial fibrosis after I/R injury. 59 Similarly, TLR9 overexpression is associated with pulmonary fibrosis and is an indicator of fibrosis. 60 Our findings showed that HK2 cells could be rescued through ROS inhibitor (tempol 1 mM) and TLR 9 siRNA by disrupting the ROS-DAMP-TLR9 axis. These observations suggest potential antifibrotic strategies in the kidney and other organs. In the future, other known inhibitors of TLR 9 could be used to inhibit downstream cytokine cascade or FDA-approved drugs could be screened for their potential role in AAN.
In conclusion, this study confirms that AAI is a potent nephrotoxin primarily affecting PTCs. Our study suggests the role of oxygen radicals, DAMPs (HMGB1/mt DNA) mediated TLR activation and inflammatory response in AA-induced cytotoxicity, which could be ameliorated by ROS inhibitor (tempol 1 mM) and TLR 9 siRNA.
Extensive fibrosis is a hallmark of AA nephrotoxicity and our study showing a prominent TLR 9 effect in AA-exposed PTCs indicate a major role for innate immunity in AA-induced fibrogenesis in the kidney.

AUTH O R CO NTR I B UTI O N S
RU and VB conceived and designed research; RU performed experiments, analysed data and prepared figures; RU and VB interpreted results of experiments, drafted manuscript, edited/revised manuscript and approved the final version of manuscript.

ACK N OWLED G EM ENT
Graphical summary was created with BioRender software at BioRe nder.com (Paid subscription for Publication and Licensing Rights; Agreement number: YK23I1VRZ1).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.