Predicting nucleic acid drug-induced nephrotoxicity using a 3D human renal proximal tubule spheroid model

Abstract

Nucleic acid drugs hold considerable promise; however, their toxicological profiles are often difficult to assess in animal models. Clinical studies have reported adverse effects, including thrombocytopenia, complement activation, hepatotoxicity, and nephrotoxicity. While human cell-based models for hepatotoxicity are advancing, nephrotoxicity assessment remains limited by the scarcity of physiologically relevant kidney cells. In this study, a three-dimensional spheroid model of human primary renal proximal tubule epithelial cells (3D-RPTEC, Nikkiso) was employed to evaluate the nephrotoxicity of nucleic acid drugs. Proteomic profiling revealed enhanced expression of drug transporters and endocytic machinery in 3D-RPTEC compared with two-dimensional cultures. Lipofection enabled efficient intracellular delivery of nucleic acids. Toxicity was assessed using ATP quantification, biomarker analysis (LDH, KIM-1, NGAL), and high-content analysis (HCA). Significant ATP depletion was observed only after prolonged exposure to SPC5001, a nephrotoxic antisense oligonucleotide. In contrast, biomarker expression and HCA facilitated early detection of compound-specific toxicity and implicated endoplasmic reticulum and mitochondrial stress as underlying mechanisms. These findings establish 3D-RPTEC as a sensitive and physiologically relevant platform for predicting the nephrotoxic potential of nucleic acid drugs.

INTRODUCTION

Early identification of potential toxicities is critical in drug development. Approximately 30% of candidate compounds are withdrawn due to safety concerns (Hornberg et al., 2014), and the high costs associated with clinical trials represent a major bottleneck in the drug discovery process (DiMasi et al., 2016). Nonclinical safety evaluations have traditionally relied on animal models; however, despite the rejection of nearly 50% of compounds during preclinical toxicity studies, toxicity is still observed in approximately 40% of candidates during clinical trials (Morgan et al., 2018). These discrepancies highlight the limited translational predictivity of animal models, likely due to interspecies differences. Recent advances have led to the development of novel therapeutic modalities, including nucleic acid drugs, gene therapies, and cell-based treatments, in addition to conventional small-molecule drugs. Nucleic acid drugs, in particular, offer high target specificity through sequence-based interactions with gene transcripts, enabling direct modulation of gene expression. However, their distinct mechanisms of action pose challenges for traditional toxicity assessment approaches, which are largely based on animal studies (Monticello et al., 2017; Goyenvalle et al., 2023). Nucleic acid drugs exert pharmacological effects by hybridizing to specific nucleotide sequences, giving rise to three primary categories of potential toxicity: (1) sequence-dependent on-target toxicity due to exaggerated pharmacodynamic effects; (2) sequence-dependent off-target toxicity arising from unintended hybridization to non-target transcripts; and (3) sequence-independent off-target toxicity attributable to the physicochemical properties of nucleic acid molecules. A white paper by the Japanese Working Group on ICH S6 and Related Issues (WGS6) outlines strategies for evaluating each of these risks (Hirabayashi et al., 2021). Traditionally, on-target toxicity has been assessed via repeat-dose studies in at least one rodent and one non-rodent species. In addition, homologous surrogate oligonucleotides and disease model animals have been used to assess target-specific effects. Toxicity testing strategies are selected on a case-by-case basis, depending on the target gene, therapeutic sequence, and affected organ system.

Sequence-dependent off-target toxicity is increasingly being investigated through in silico analyses using genomic databases and transcriptomic profiling via microarray analyses of human-derived samples. In the context of hepatotoxicity, human cell-based systems have shown promise for predictive assessment (Dieckmann et al., 2018); however, several mechanistic aspects remain unresolved. Sequence-independent off-target toxicities, including thrombocytopenia, complement activation, hepatotoxicity, and nephrotoxicity, are commonly reported class effects of nucleic acid therapeutics. Among these, nephrotoxicity is of particular concern, as nearly half of all approved nucleic acid drugs have been associated with kidney adverse effects (Wu et al., 2022). Drug-induced nephrotoxicity from small-molecule compounds typically arises through one or more of the following mechanisms: proximal tubular damage, acute tubular necrosis, tubular obstruction, or interstitial nephritis (Kwiatkowska et al., 2021), with the proximal tubule being the most frequently affected region. This susceptibility stems from its central role in substance reabsorption and elimination. Nucleic acid drugs are similarly implicated, as exemplified by the withdrawal of SPC5001 due to its accumulation in the proximal tubule and subsequent nephrotoxicity in clinical trials (van Poelgeest et al., 2015). To improve predictive accuracy, various cell-based systems have been developed for nephrotoxicity assessment (Petreski et al., 2023; Varda et al., 2025); however, progress has been hindered by the absence of physiologically functional kidney cells. The HK-2 cell line, an immortalized human proximal tubular epithelial cell line, remains the most widely used model in kidney research (Ryan et al., 1994). While HK-2 cells express apical efflux transporters such as ABCB1 (MDR1) and members of the ABCC family, they lack expression of key uptake transporters in the SLC22 family (e.g., OAT1, OAT3, OCT2), suggesting that critical kidney functions may be inadequately represented (Jenkinson et al., 2012). A prior study by Arakawa et al. evaluated the nephrotoxicity of 32 compounds known to affect proximal tubular cells, revealing poor predictive performance in HK-2-based assays (sensitivity: 52.9%, specificity: 66.7%) (Arakawa et al., 2024). Human primary renal proximal tubule epithelial cells (RPTEC) have also been used in two-dimensional (2D) cultures. RPTEC, in contrast to HK-2 cells, are not immortalized, and they preserve renal functions more than HK-2 cells. However, several compounds failed to yield predictable outcomes (Secker et al., 2019). In contrast, a similar evaluation conducted by the same group demonstrated improved accuracy (sensitivity: 82.4%, specificity: 86.7%) (Arakawa et al., 2024). However, unlike 3D-RPTEC cultures, 2D cultures retain proliferative capacity. Consequently, it is difficult to distinguish between compound-induced cytotoxicity and proliferation-associated effects, thereby complicating the determination of robust and interpretable assay endpoints. As with other organ-derived cells, renal proximal tubular cells rapidly lose their function in planar culture conditions (Hughes et al., 2006), posing challenges for evaluating functional nephrotoxicity in vitro. Spheroid cultures of proximal tubular epithelial cells formed in low-adhesion plates have yielded substantially improved specificity (sensitivity: 82.4%, specificity: 100%), with further gains observed upon extended exposure durations (sensitivity: 88.2%, specificity: 93.3%) (Arakawa et al., 2024). These findings suggest that three-dimensional cultures more accurately replicate kidney physiology and are particularly suitable for evaluating the nephrotoxicity of nucleic acid drugs under prolonged exposure conditions. Accordingly, in this study, we first sought to identify an optimal in vitro model for nephrotoxicity evaluation. A preliminary report by Ishiguro et al. demonstrated that an early version of the 3D-RPTEC model more closely resembled in vivo kidney tissue than standard planar cultures, based on comparison with human renal cortex samples (Ishiguro et al., 2023). Here, we compared the commercial version of 3D-RPTEC with conventional monolayer cultures using proteomic profiling. The analysis employed data-independent acquisition (DIA) mass spectrometry, which enables comprehensive and reproducible peptide fragmentation by acquiring MS/MS data for all precursor ions (Wolf-Yadlin et al., 2016). Based on these results, we systematically examined differences in the expression of key drug transporters and endocytic receptors involved in nucleic acid drug uptake. Functional protein categories were further assessed using Gene Ontology (GO) enrichment analysis to elucidate relevant biological processes.

Next, a nephrotoxicity evaluation system for nucleic acid drugs was developed. Commonly used nucleic acid transfection reagents (Cardarelli et al., 2016) were employed to determine the cellular uptake of labeled nucleic acid drugs and to assess their intracellular retention time.

Subsequently, nephrotoxicity was evaluated using three nucleic acid drugs, SPC5001, Viltolarsen, and Givosiran, that have been associated with kidney adverse effects in clinical trials (van Poelgeest et al., 2013; Alhamadani et al., 2022; Lazareth et al., 2021). Evaluation methods included ATP-based cell viability assays (Xu et al., 2025), measurement of cytotoxicity biomarkers (Coca et al., 2008; Xiao et al., 2023), and high-content analysis (Wardwell-Swanson et al., 2020). Through this integrative approach, we identified suitable evaluation methods for assessing the nephrotoxic potential of nucleic acid drugs.

MATERIALS AND METHODS

Cell culture

Human primary renal proximal tubule epithelial cells (RPTEC) were obtained from Lonza (Walkersville, MD, USA; Lot: 18TL117405; donor: Asian male, 57 years) and cultured in REGM medium (Lonza, Walkersville, MD, USA). For 2D culture, cells were seeded in 100-mm dishes and maintained at 37°C in a humidified incubator with 5% CO₂. The medium was replaced every 2–3 days until the cells reached subconfluence. These monolayer-cultured cells are hereafter referred to as "2D-RPTEC." For three-dimensional (3D) culture, we utilized 3D-RPTEC (Nikkiso, Tokyo, Japan), a commercially available product comprising human primary RPTEC cultured in a spheroid configuration to better simulate physiological conditions. Cells were maintained in a proprietary 3D-RPTEC medium (Nikkiso, Tokyo, Japan) under identical incubation conditions. Medium changes were performed every 2–3 days. For experimental comparisons, equal cell numbers were used in both 2D and 3D cultures. The spheroid-cultured cells are hereafter referred to as "3D-RPTEC."

Test compounds

SPC5001 (Ajinomoto, Tokyo, Japan), conjugated with Alexa Fluor™ 647, was used to verify intracellular uptake. For toxicological evaluation, unlabeled SPC5001 (Ajinomoto, Tokyo, Japan), Viltolarsen (MedChemExpress, Monmouth Junction, NJ, USA), and Givosiran (MedChemExpress, Monmouth Junction, NJ, USA) were employed. The properties of each compound, including nucleic acid type, catalog number, 5'-terminal modification, molecular weight (MW), and chemical modifications, are summarized in Table 1. Additionally, Table 2 summarizes the maximum plasma concentration (C_max), routes of administration (RoA), and nephrotoxicity data from nonclinical and clinical studies.

Table 1. Characteristics of test compounds.

Test Compound	Type	Catalog No.	5'-Terminal Addition	MW (g/mol)	Modification	Manufacturer
SPC5001	Antisense oligo	None (custom synthesis)	Alexa Fluor™ 647	5726.02	PS, LNA, 5mC	Ajinomoto Bio-Pharma Services, Inc.
SPC5001	Antisense oligo	None (custom synthesis)	None	4689.85	PS, LNA, 5mC	Ajinomoto Bio-Pharma Services, Inc.
Viltolarsen	Antisense oligo	HY-132586A	None	7386.42	PMO	MedChemExpress
Givosiran	siRNA	HY-132610	None	16300.6	PS, 2'-OMe, 2'-F	MedChemExpress

Table 2. Nonclinical and clinical information of test compounds

Test Compound	Cmax	RoA	Nonclinical study	Clinical study
SPC5001	unknown	subcutaneous	No clear nephrotoxicity was observed	· Withdrawn due to nephrotoxicity · Increased serum creatinine and decreased estimated glomerular filtration rate (eGFR) · Multiple foci of tubular necrosis and oligonucleotide deposition were observed in the renal biopsy
Viltolarsen	0.043 µmol/L	intravenous	Increased serum blood urea nitrogen (BUN) and creatinine	Increased β2-microglobulin and N-acetyl-β-D-glucosaminidase (NAG)
Givosiran	0.021 µmol/L	subcutaneous	No clear nephrotoxicity was observed	Increased serum creatinine and decreased eGFR

Proteomic sample preparation and analysis

2D-RPTEC cells were harvested using a cell scraper, transferred to 15-mL tubes, centrifuged at 100 ×g for 3 min, and washed with Dulbecco's Phosphate-Buffered Saline (D-PBS) (–). Cell pellets were snap-frozen in liquid nitrogen and stored at –80°C. For 3D-RPTEC, 96 spheroids were collected from a single plate using wide-bore pipette tips, transferred to 15-mL tubes, and processed in parallel with the 2D-RPTEC samples. Protein extraction was performed using Phase Transfer Surfactant (PTS) buffer (Masuda et al., 2009). Samples were heated at 95°C for 5 min and sonicated using the BIORUPTOR II (Sonic Bio, Kanagawa, Japan) under high-power settings (30 sec per cycle, 15 cycles). Protein concentrations were determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA), and 20 µg of protein was used for downstream processing. Protein purification followed a published protocol (Kawashima et al., 2022). Briefly, proteins were reduced with Tris (2-carboxyethyl) phosphine Hydrochloride (TCEP-HCl; Fujifilm Wako, Tokyo, Japan), alkylated with iodoacetamide (Nacalai Tesque, Kyoto, Japan), and quenched with L-cysteine (Fujifilm Wako, Tokyo, Japan). Samples were purified using Sera-Mag Speed Beads Carboxylate-Modified Magnetic Particles (SP3 beads; GE Healthcare, Chicago, IL, USA), followed by enzymatic digestion with trypsin (Promega, Madison, WI, USA), LysC (Fujifilm Wako, Tokyo, Japan), and ammonium bicarbonate. Peptides were desalted using C18 GL-Tip SDB tips (GL-Sciences, Tokyo, Japan) and dried under vacuum. Dried peptides were reconstituted in 1% trifluoroacetic acid (TFA) and 2% acetonitrile (ACN) prior to injection into a liquid chromatography–tandem mass spectrometry (LC-MS/MS) system comprising an UltiMate 3000 RSLCnano LC system coupled to an Orbitrap Fusion Lumos Tribrid mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Peptide separation was performed using a 3-µm C18 nano-HPLC capillary column (Nikkyo Technos, NTCC-360/75-3) with the following gradient: 5–35% buffer B (0.1% formic acid in 100% ACN) over 90 min, 35–95% over 3 min, and re-equilibration to 5% over 27 min. The flow rate was set to 280 nL/min, and the column temperature was maintained at 40°C. DIA parameters, MS acquisition settings, and spectral library generation followed established protocols (Kawashima et al., 2022). Proteomic data were analyzed using DIA-NN (Lou and Shui, 2024, Demichev et al., 2020), and quantitative outputs were further processed with DIA-Analyst for clustering, violin plot generation, and GO enrichment analysis.

Evaluation of nucleic acid drug delivery into 3D-RPTEC

To assess nucleic acid drug delivery into 3D-RPTEC spheroids, Lipofectamine™ 3000 Transfection Reagent and P3000 Enhancer Reagent (Thermo Fisher Scientific, Waltham, MA, USA) and Opti-MEM medium (Thermo Fisher Scientific, Waltham, MA, USA) were employed following the manufacturer’s instructions. Reagent concentrations were optimized to minimize cytotoxicity (nucleic acid drugs: 100 ng/µL/well; Lipofectamine™ 3000: 0.3 µL/well; P3000: 0.3 µL/well in Opti-MEM medium). To confirm intracellular uptake, SPC5001 labeled with Alexa Fluor™ 647 was used. After transfection, the medium was completely replaced on Day 3, followed by medium changes every 2–3 days until Day 14. Fluorescence imaging was performed using an inverted fluorescence phase-contrast microscope (BZ-X710, KEYENCE, Osaka, Japan), with excitation/emission wavelengths of 651 nm/672 nm for Alexa Fluor™ 647.

ATP measurement

Intracellular ATP levels were measured on Days 3, 7, and 14 following exposures of 3D-RPTEC to SPC5001, Viltolarsen, or Givosiran at final concentrations of 0.05 µmol/L, 0.5 µmol/L, or 5 µmol/L. ATP content was quantified using the CellTiter-Glo^® 3D Cell Viability Assay (Promega, Madison, WI, USA). For each measurement, a single spheroid was collected using a wide-bore pipette tip and transferred to a white-walled 96-well plate containing 100 µL of medium. An equal volume (100 µL) of CellTiter-Glo^® 3D reagent was added, and samples were incubated at room temperature for 30 min. Luminescence was measured using the Nivo^® multimode plate reader (Revvity, Waltham, MA, USA). Data analysis was performed using GraphPad Prism (GraphPad Software, Boston, MA, USA; version 10.5.0).

Biomarker measurement

To evaluate early nephrotoxicity, 3D-RPTEC spheroids were exposed to SPC5001, Viltolarsen, or Givosiran at the indicated concentrations. Supernatants (10 µL) were collected on Days 1, 2, 3, and 7, and diluted to 40 µL in Storage Buffer consisting of 200 mmol/L Tris-HCl (Roche, Basel, Switzerland), pH 7.3, 10% glycerol (Fujifilm Wako, Tokyo, Japan), 1% BSA (Fujifilm Wako, Tokyo, Japan), then stored at −80°C until analysis.

LDH measurement: Lactate dehydrogenase (LDH) release was quantified using the LDH-Glo™ Cytotoxicity Assay (Promega, Madison, WI, USA) per the manufacturer’s instructions. Frozen samples were thawed and 25 µL of each was transferred to a white half-area 96-well plate (Revvity, Waltham, MA, USA). An equal volume (25 µL) of LDH Detection Reagent was added, and the plate was incubated for 60 min at room temperature (about 21-24°C) in the dark. Luminescence was measured using the Nivo^® multimode plate reader.

KIM-1 and NGAL measurement: Kidney injury molecule-1 (KIM-1) and neutrophil gelatinase-associated lipocalin (NGAL) levels were quantified using the AlphaLISA^® KIM-1 (Human) and AlphaLISA^® NGAL (Human) Detection Kits (Revvity, Waltham, MA, USA), respectively, following the manufacturer’s protocol. Briefly, 5 µL of each sample was added to a white half-area 96-well plate, followed by 20 µL of 2.5× AlphaLISA mix containing acceptor beads (final concentration: 10 µg/mL) and biotinylated anti-KIM-1 or anti-NGAL antibodies (final concentration: 1 nmol/L). After 60 min of incubation at room temperature, 25 µL of 2× streptavidin donor beads (final concentration: 40 µg/mL) was added. Plates were incubated for an additional 30 min in the dark. Signals were detected using AlphaScreen settings on the Nivo^® plate reader. Data were analyzed and visualized using GraphPad Prism (version 10.5.0).

High-content analysis

To evaluate subcellular stress responses, 3D-RPTEC spheroids were exposed to SPC5001, Viltolarsen, or Givosiran at varying concentrations. On Days 7 and 14, the medium was carefully removed to prevent spheroid aspiration, and 100 µL of staining solution (prepared in 3D-RPTEC-specific medium) was added to each well. Two multiplex staining protocols were employed: Staining A: CellROX Green (Thermo Fisher Scientific, Waltham, MA, USA; 5 µmol/L, excitation/emission: 485/520 nm) and LysoTracker Red (Thermo Fisher Scientific, Waltham, MA, USA; 100 nmol/L, excitation/emission: 577⁄590 nm). Staining B: MitoTracker Deep Red (Thermo Fisher Scientific, Waltham, MA, USA; 100 nmol/L, excitation/emission: 644⁄665 nm) and ER-Tracker Green (Thermo Fisher Scientific, Waltham, MA, USA; 500 nmol/L, excitation/emission: 504/511 nm). Following staining, cells were washed three times with Hank's Balanced Salt Solution (HBSS; Thermo Fisher Scientific, Waltham, MA, USA, 125 µL/well per wash) and transferred to black, U-bottom 96-well plates with transparent bottoms (Corning, Corning, NY, USA) for imaging. Image acquisition was performed using a Confocal Quantitative Image Cytometer (CQ-1; (Yokogawa Electric, Tokyo, Japan), and image analysis was conducted using Cell Pathfinder software (Yokogawa Electric, Tokyo, Japan), and fluorescence intensities were quantified based on mean intensity. Data were analyzed and visualized using GraphPad Prism (version 10.5.0).

Statistical analyses

Statistical analyses were performed using an unpaired Student’s t-test for comparisons between two groups (Fig. 1). Dose–response relationships were evaluated using Williams’ test following one-way analysis of variance (ANOVA), comparing each treatment group with the control in ascending dose order (Figs. 3–6). A P value of less than 0.05 was considered statistically significant. All analyses were conducted using GraphPad Prism (version 10.5.0).

Fig. 1

Comparative proteomic analysis of 2D-RPTEC and 3D-RPTEC. (A) Volcano plot illustrating differential protein expression between 2D- and 3D-RPTEC. Proteins significantly upregulated or downregulated in 3D-RPTEC are highlighted. (B) Gene Ontology (GO) enrichment analysis of differentially expressed proteins, indicating associated biological functions. (C) Cluster analysis of global proteomic profiles of 2D- and 3D-RPTEC. (D) Comparative expression of key drug transporters and proximal tubule markers (e.g., AQP1, CADH6) between 2D- and 3D-RPTEC.

RESULTS AND DISCUSSION

Proteomic analysis of 2D- and 3D-RPTEC

Proteomic profiling of 2D- and 3D-RPTEC was performed, and cluster analysis using DIA-Analyst (https://analyst-suites.org/apps/dia-analyst/) revealed distinct clustering patterns, even among cells derived from the same donor lot (Fig. 1A), suggesting substantial differences in their proteomic landscapes.

A volcano plot was generated using a false discovery rate (FDR) threshold of 0.05 to identify significantly differentially expressed proteins (Fig. 1B). Of the 7,617 proteins detected, 210 were significantly upregulated (highlighted in red) in 3D-RPTEC. These included glycerol-3-phosphate dehydrogenase (GPDA), implicated in carbohydrate metabolism and mitochondrial proton translocation, and UD19, a glucuronosyltransferase involved in xenobiotic metabolism. Conversely, 51 proteins (highlighted in blue) were significantly downregulated, including replication protein A (RPA), which binds single-stranded DNA, and mini-chromosome maintenance complex component 6 (MCM6), a helicase essential for DNA replication.

GO enrichment analysis indicated that the upregulated proteins in 3D-RPTEC were associated with fatty acid and pyruvate metabolism and the tricarboxylic acid (TCA) cycle (Fig. 1B). These findings suggest a shift toward enhanced oxidative metabolism and gluconeogenesis, closely resembling the in vivo metabolic phenotype of human renal proximal tubule epithelial cells (Tian and Liang, 2021). In contrast, the downregulation of DNA replication-associated proteins is consistent with a quiescent, non-proliferative state characteristic of mature proximal tubule cells. Expression of key proximal tubule-specific proteins and drug transporters was then compared between 2D- and 3D-RPTEC (Fig. 1D). In 2D cultures, major drug transporters such as organic anion transporter 1 (OAT1), OAT4, multidrug and toxin extrusion proteins MATE1 and MATE2-K, and endocytic receptors critical for nucleic acid uptake, including Megalin and Cubilin, were not detected (Nielsen et al., 2016). Among ATP-binding cassette (ABC) transporters, only MDR1 was expressed.

In contrast, 3D-RPTEC demonstrated robust expression of these transporters and receptors, reflecting similarities to in vivo physiological functions. These results indicate that 2D-RPTEC rapidly lose key characteristics of renal proximal tubule cells, whereas 3D-RPTEC retain functions such as drug transport and endocytosis, while concurrently exhibiting a more differentiated metabolic and proliferative profile. Collectively, these findings support the utility of 3D-RPTEC as a physiologically relevant model for assessing the nephrotoxicity of nucleic acid drugs.

Fluorescence imaging was conducted on Days 1, 2, 3, and 7 to evaluate intracellular localization and uptake efficiency based on fluorescence intensity and distribution (Fig. 2). To assess the uptake kinetics of SPC5001 in 3D-RPTEC, Alexa Fluor™ 647–labeled SPC5001 was administered under the three delivery conditions described above. Fluorescence imaging revealed that SPC5001 was efficiently internalized as early as Day 1 when delivered with Lipofectamine or Lipofectamine plus P3000, and the signal persisted through Day 7, indicating sustained intracellular retention and suitability for long-term studies. Interestingly, even in the absence of transfection reagents, gradual uptake of SPC5001 was observed from Day 2 onward. This uptake is likely mediated by receptor-dependent endocytosis via Megalin and Cubilin, both of which were detected in 3D-RPTEC by proteomic analysis and are known to be upregulated in spheroid cultures. Spatial distribution analysis showed that fluorescence was initially restricted to the peripheral regions of the spheroids (up to Day 3), but by Day 7, the signal was uniformly distributed throughout the spheroid core, suggesting progressive internalization and diffusion over time. Based on these observations, the combination of Lipofectamine and P3000, with a 3-day delivery period, was determined to be the optimal condition for efficient and sustained delivery of nucleic acid drugs in the 3D-RPTEC model.

Fig. 2

Cellular uptake of Alexa Fluor™ 647–labeled nucleic acid drug in 3D-RPTEC. SPC5001 was labeled with Alexa Fluor™ 647 and introduced into 3D-RPTEC under three conditions: (1) SPC5001 alone, (2) SPC5001 with Lipofectamine, and (3) SPC5001 with Lipofectamine and P3000.

Results of toxicity tests with nucleic acid drugs

ATP assay results

The impact of long-term exposure to nucleic acid drugs on cell viability was evaluated via ATP quantification in 3D-RPTEC. Cells were treated with SPC5001, Viltolarsen, or Givosiran at concentrations of 0.05–5 µmol/L, delivered using a lipofection mixture containing Lipofectamine and P3000. For all nucleic acid drugs, exposure lasted 3 days, followed by washout and culture in drug-free medium with full medium replacement every 2–3 days. A modest reduction in ATP levels was noted by Day 7 for SPC5001 and Viltolarsen at higher concentrations. Notably, SPC5001 induced a pronounced ATP decline at 5 µmol/L by Day 14 (higher than 50%), indicating cumulative cytotoxicity (Fig. 3A).

In contrast, Viltolarsen and Givosiran did not cause significant ATP depletion (lower than 20%) on Day14 (Fig. 3B–C). SPC5001 was previously withdrawn from clinical development due to multifocal tubular necrosis linked to intracellular accumulation during prolonged administration (van Poelgeest et al., 2013). The ATP depletion observed in 3D-RPTEC aligns with these findings and underscores the model’s capacity to detect nephrotoxicity associated with nucleic acid drugs. Importantly, these results emphasize that such toxicity may be underrepresented in short-term assays, highlighting the necessity of extended exposure studies for nucleic acid drugs.

Fig. 3

ATP levels in 3D-RPTEC exposed to nucleic acid drugs. ATP levels were assessed following exposure of 3D-RPTEC to SPC5001 (A), Viltolarsen (B), and Givosiran (C) at 0.05, 0.5, and 5 µmol/L. Measurements were taken on Days 3, 7, and 14 to evaluate time- and dose-dependent effects on cellular energy metabolism.

Biomarker analysis

LDH release analysis revealed an increase between Days 1 and 3 for Viltolarsen and Givosiran, suggesting the induction of sub-lethal cellular stress during the early phase of exposure. This response occurred in the absence of significant ATP depletion, indicating pre-lethal injury (Fig. 4). KIM-1 levels were markedly elevated on Day 1 in SPC5001-treated cells, whereas no substantial increase was observed with Viltolarsen or Givosiran (Fig. 5A). These findings are consistent with clinical observations reporting a ~60-fold rise in urinary KIM-1 following SPC5001 administration (van Poelgeest et al., 2013), supporting the translational relevance of the 3D-RPTEC model for detecting early kidney injury. NGAL expression exhibited compound-specific dynamics. Viltolarsen induced significant NGAL elevation on Days 2 and 3 (Fig. 5B), whereas SPC5001 triggered a delayed increase on Day 14 (Fig. 5A). These patterns suggest mechanistic differences in nephrotoxicity among antisense oligonucleotides. Importantly, NGAL elevation is associated with “subclinical acute kidney injury (AKI),” wherein patients exhibit tubular injury without changes in serum creatinine but remain at increased risk for adverse kidney outcomes (Zou et al., 2022). Thus, NGAL may serve both as a biomarker of early tubular damage and a predictor of clinical progression. Animal studies have reported Viltolarsen-induced tubular degeneration, and the FDA has issued warnings regarding its nephrotoxic potential and risk of fatal glomerulonephritis (NS Pharma, VILTEPSO injection). Given that drug concentrations used in this study exceeded the clinical C_max by over 100-fold, the observed responses may reflect adverse effects associated with high systemic exposure. Givosiran elicited minimal changes in LDH, KIM-1, and NGAL, in line with clinical data indicating a transient, non-progressive decline in kidney function in ~90% of treated patients, without progression to AKI or chronic kidney disease (Lazareth et al., 2021).

Fig. 4

LDH levels in supernatants of 3D-RPTEC exposed to nucleic acid drugs. LDH release was quantified in the culture supernatant of 3D-RPTEC treated with SPC5001 (A), Viltolarsen (B), and Givosiran (C) at 0.05, 0.5, and 5 µmol/L. Samples were collected on Days 1, 2, 3, 7, and 14 to assess time- and dose-dependent cytotoxicity.

Fig. 5

KIM-1 and NGAL levels in supernatants of 3D-RPTEC exposed to nucleic acid drugs. KIM-1 and NGAL were measured in supernatants from 3D-RPTEC exposed to 5 µmol/L of SPC5001 (A), Viltolarsen (B), and Givosiran (C). Time points included Days 1, 2, 3, 7, and 14.

High-content analysis (HCA)

HCA was conducted to assess subcellular responses to SPC5001, Viltolarsen, and Givosiran at three concentrations (0.05, 0.5, and 5 µmol/L) in 3D-RPTEC at Days 7 and 14 (Fig. 6). On Day 7, all compounds induced organelle-specific stress responses, with a concentration-dependent increase in endoplasmic reticulum (ER) stress. By Day 14, cells treated with Viltolarsen and Givosiran exhibited marked mitochondrial alterations, suggesting a temporal progression of organelle dysfunction. These findings indicate that nucleic acid drug-induced toxicity may involve a sequential cascade, wherein ER stress precedes and contributes to subsequent mitochondrial impairment. This mechanistic progression mirrors that observed in cisplatin-induced nephrotoxicity, in which ER damage triggers the release of Ca²⁺, disrupting intracellular calcium homeostasis (Tang et al., 2023). The elevated cytosolic Ca²⁺ is taken up by mitochondria, leading to impaired ATP production and activation of Ca²⁺-dependent enzymes such as phospholipases and proteases. These events collectively diminish cellular repair capacity and viability, a pathological mechanism also implicated in diabetic and hypertensive kidney injury (Wang et al., 2020).

Fig. 6

High-content analysis (HCA) of 3D-RPTEC exposed to nucleic acid drugs. SPC5001 (A, D), Viltolarsen (B, E), and Givosiran (C, F) were administered to 3D-RPTEC at concentrations of 0.05, 0.5, and 5 µmol/L. Cells were stained and analyzed on Days 7 and 14 using high-content imaging. Bar plots represent mean fluorescence intensities, reflecting drug-induced cellular stress in a time- and concentration-dependent manner.

In conclusion, these findings demonstrate that Nikkiso’s 3D-RPTEC is a highly suitable in vitro model for evaluating the nephrotoxicity of nucleic acid drugs. It enables sensitive, mechanism-based, and clinically relevant toxicity prediction, outperforming conventional viability assays in both sensitivity and translational value. However, the 3D-RPTEC model consists only of proximal tubule cells and therefore cannot fully represent all types of drug-induced nephrotoxicity. Additionally, the nucleic acid drugs tested in this study represent only a small subset, and due to the limited clinical information available, a uniform dose was used. These factors constitute the limitations of this study. In the future, we aim to construct models using other types of kidney cells beyond proximal tubules, including those derived from animal cells, and set concentrations based on clinical C_max values. We also aim to expand the range of compounds evaluated. Through these efforts, we hope to contribute to the efficient identification of nucleic acid drugs with minimal nephrotoxicity.

ACKNOWLEDGMENTS

We express our sincere gratitude for the excellent technical assistance provided by Dr. Rieko Kojima of the Toyama Prefectural Research and Development Center for Pharmaceutical Affairs in proteomic analysis, and by Ms. Nagisa Kato of Yokogawa Electric Corporation in high-content analysis. We would also like to thank Editage (www.editage.jp) for English language editing.

Funding

This research was conducted with a research fund from NIKKISO Co., Ltd., which the author belongs to.

Conflict of interest

The authors declare that there is no conflict of interest.

Data availability

We confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

Author contributions

Conceptualization: Kaoru Morimura

Funding acquisition: Kaoru Morimura, Etsushi Takahashi, Yoichi Jimbo

Investigation: Kaoru Morimura, Ayano Araki, Yukiko Nishioka

Supervision: Kaoru Morimura, Etsushi Takahashi, Yoichi Jimbo

Visualization: Kaoru Morimura

Writing – original draft: Kaoru Morimura

Writing – review & editing: Kaoru Morimura, Etsushi Takahashi, Hayata Maeda

Ethical approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

REFERENCES

•  Alhamadani,  F.,  Zhang,  K.,  Parikh,  R.,  Wu,  H.,  Rasmussen,  T.P.,  Bahal,  R.,  Zhong,  X.B. and  Manautou,  J.E. (2022): Adverse Drug Reactions and Toxicity of the Food and Drug Administration-Approved Antisense Oligonucleotide Drugs. Drug Metab. Dispos., 50, 879-887.
•  Arakawa,  H.,  Higuchi,  D.,  Takahashi,  E.,  Matsushita,  K.,  Nedachi,  S.,  Peng,  H.,  Kadoguchi,  M.,  Morimura,  K.,  Araki,  A.,  Kondo,  M.,  Ishiguro,  N.,  Jimbo,  Y. and  Tamai,  I. (2024): Three-dimensional culture of human proximal tubular epithelial cells for an in vitro evaluation of drug-induced kidney injury. J. Pharm. Sci., 113, 3255-3264.
•  Cardarelli,  F.,  Digiacomo,  L.,  Marchini,  C.,  Amici,  A.,  Salomone,  F.,  Fiume,  G.,  Rossetta,  A.,  Gratton,  E.,  Pozzi,  D. and  Caracciolo,  G. (2016): The intracellular trafficking mechanism of Lipofectamine-based transfection reagents and its implication for gene delivery. Sci. Rep., 6, 25879.
•  Coca,  S.G.,  Yalavarthy,  R.,  Concato,  J. and  Parikh,  C.R. (2008): Biomarkers for the diagnosis and risk stratification of acute kidney injury: a systematic review. Kidney Int., 73, 1008-1016.
•  Demichev,  V.,  Messner,  C.B.,  Vernardis,  S.I.,  Lilley,  K.S. and  Ralser,  M. (2020): DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput. Nat. Methods, 17, 41-44.
•  Dieckmann,  A.,  Hagedorn,  P.H.,  Burki,  Y.,  Brügmann,  C.,  Berrera,  M.,  Ebeling,  M.,  Singer,  T. and  Schuler,  F. (2018): A Sensitive In Vitro Approach to Assess the Hybridization-Dependent Toxic Potential of High Affinity Gapmer Oligonucleotides. Mol. Ther. Nucleic Acids, 10, 45-54.
•  DiMasi,  J.A.,  Grabowski,  H.G. and  Hansen,  R.W. (2016): Innovation in the pharmaceutical industry: new estimates of R&D costs. J. Health Econ., 47, 20-33.
•  Goyenvalle,  A.,  Jimenez-Mallebrera,  C.,  van Roon,  W.,  Sewing,  S.,  Krieg,  A.M.,  Arechavala-Gomeza,  V. and  Andersson,  P. (2023): Considerations in the Preclinical Assessment of the Safety of Antisense Oligonucleotides. Nucleic Acid Ther., 33, 1-16.
•  Hirabayashi,  Y.,  Maki,  K.,  Kinoshita,  K.,  Nakazawa,  T.,  Obika,  S.,  Naota,  M.,  Watanabe,  K.,  Suzuki,  M.,  Arato,  T.,  Fujisaka,  A.,  Fueki,  O.,  Ito,  K. and  Onodera,  H. (2021): Considerations of the Japanese Research Working Group for the ICH S6 & Related Issues Regarding Nonclinical Safety Assessments of Oligonucleotide Therapeutics: Comparison with Those of Biopharmaceuticals. Nucleic Acid Ther., 31, 114-125.
•  Hornberg,  J.J.,  Laursen,  M.,  Brenden,  N.,  Persson,  M.,  Thougaard,  A.V.,  Toft,  D.B. and  Mow,  T. (2014): Exploratory toxicology as an integrated part of drug discovery. Part I: why and how. Drug Discov. Today, 19, 1131-1136.
•  Hughes,  R.D.,  Mitry,  R.R.,  Dhawan,  A.,  Lehec,  S.C.,  Girlanda,  R.,  Rela,  M.,  Heaton,  N.D. and  Muiesan,  P. (2006): Isolation of hepatocytes from livers from non-heart-beating donors for cell transplantation. Liver Transpl., 12, 713-717.
•  Ishiguro,  N.,  Takahashi,  E.,  Arakawa,  H.,  Saito,  A.,  Kitagawa,  F.,  Kondo,  M.,  Morinaga,  G.,  Takatani,  M.,  Takahashi,  R.,  Kudo,  T.,  Mae,  S.I.,  Kadoguchi,  M.,  Higuchi,  D.,  Nakazono,  Y.,  Tamai,  I.,  Osafune,  K. and  Jimbo,  Y. (2023): Improvement of protein expression profile in three-dimensional renal proximal tubular epithelial cell spheroids selected based on OAT1 gene expression: a potential in vitro tool for evaluating human renal proximal tubular toxicity and drug disposition. Drug Metab. Dispos., 51, 1177-1187.
•  Jenkinson,  S.E.,  Chung,  G.W.,  van Loon,  E.,  Bakar,  N.S.,  Dalzell,  A.M. and  Brown,  C.D. (2012): The limitations of renal epithelial cell line HK-2 as a model of drug transporter expression and function in the proximal tubule. Pflugers Arch., 464, 601-611.
•  Kawashima,  Y.,  Nagai,  H.,  Konno,  R.,  Ishikawa,  M.,  Nakajima,  D.,  Sato,  H.,  Nakamura,  R.,  Furuyashiki,  T. and  Ohara,  O. (2022): Single-Shot 10K Proteome Approach: Over 10,000 Protein Identifications by Data-Independent Acquisition-Based Single-Shot Proteomics with Ion Mobility Spectrometry. J. Proteome Res., 21, 1418-1427.
•  Kwiatkowska,  E.,  Domański,  L.,  Dziedziejko,  V.,  Kajdy,  A.,  Stefańska,  K. and  Kwiatkowski,  S. (2021): The mechanism of drug nephrotoxicity and the methods for preventing kidney damage. Int. J. Mol. Sci., 22, 11.
•  Lazareth,  H.,  Poli,  A.,  Bignon,  Y.,  Mirmiran,  A.,  Rabant,  M.,  Cohen,  R.,  Schmitt,  C.,  Puy,  H.,  Karras,  A.,  Gouya,  L. and  Pallet,  N. (2021): Renal Function Decline With Small Interfering RNA Silencing Aminolevulinic Acid Synthase 1 (ALAS1). Kidney Int. Rep., 6, 1904-1911.
•  Lou,  R. and  Shui,  W. (2024): Acquisition and Analysis of DIA-Based Proteomic Data: A Comprehensive Survey in 2023. Mol. Cell. Proteomics, 23, 100712.
•  Masuda,  T.,  Saito,  N.,  Tomita,  M. and  Ishihama,  Y. (2009): Unbiased quantitation of Escherichia coli membrane proteome using phase transfer surfactants. Mol. Cell. Proteomics, 8, 2770-2777.
•  Monticello,  T.M.,  Jones,  T.W.,  Dambach,  D.M.,  Potter,  D.M.,  Bolt,  M.W.,  Liu,  M.,  Keller,  D.A.,  Hart,  T.K. and  Kadambi,  V.J. (2017): Current nonclinical testing paradigm enables safe entry to First-In-Human clinical trials: the IQ consortium nonclinical to clinical translational database. Toxicol. Appl. Pharmacol., 334, 100-109.
•  Morgan,  P.,  Brown,  D.G.,  Lennard,  S.,  Anderton,  M.J.,  Barrett,  J.C.,  Eriksson,  U.,  Fidock,  M.,  Hamrén,  B.,  Johnson,  A.,  March,  R.E.,  Matcham,  J.,  Mettetal,  J.,  Nicholls,  D.J.,  Platz,  S.,  Rees,  S.,  Snowden,  M.A. and  Pangalos,  M.N. (2018): Impact of a five-dimensional framework on R&D productivity at AstraZeneca. Nat. Rev. Drug Discov., 17, 167-181.
•  Nielsen,  R.,  Christensen,  E.I. and  Birn,  H. (2016): Megalin and cubilin in proximal tubule protein reabsorption: from experimental models to human disease. Kidney Int., 89, 58-67.
• NS Pharma, VILTEPSO (Viltolarsen) injection, for intravenous
•  Petreski,  T.,  Varda,  L.,  Gradišnik,  L.,  Maver,  U. and  Bevc,  S. (2023): Renal Proximal Tubular Epithelial Cells: From Harvesting to Use in Studies. Nephron, 147, 650-654.
•  Ryan,  M.J.,  Johnson,  G.,  Kirk,  J.,  Fuerstenberg,  S.M.,  Zager,  R.A. and  Torok-Storb,  B. (1994): HK-2: an immortalized proximal tubule epithelial cell line from normal adult human kidney. Kidney Int., 45, 48-57.
•  Secker,  P.F.,  Schlichenmaier,  N.,  Beilmann,  M.,  Deschl,  U. and  Dietrich,  D.R. (2019): Functional transepithelial transport measurements to detect nephrotoxicity in vitro using the RPTEC/TERT1 cell line. Arch. Toxicol., 93, 1965-1978.
•  Tang,  C.,  Livingston,  M.J.,  Safirstein,  R. and  Dong,  Z. (2023): Cisplatin nephrotoxicity: new insights and therapeutic implications. Nat. Rev. Nephrol., 19, 53-72.
•  Tian,  Z. and  Liang,  M. (2021): Renal metabolism and hypertension. Nat. Commun., 12, 963.
•  van Poelgeest,  E.P.,  Swart,  R.M.,  Betjes,  M.G.,  Moerland,  M.,  Weening,  J.J.,  Tessier,  Y.,  Hodges,  M.R.,  Levin,  A.A. and  Burggraaf,  J. (2013): Acute kidney injury during therapy with an antisense oligonucleotide directed against PCSK9. Am. J. Kidney Dis., 62, 796-800.
•  van Poelgeest,  E.P.,  Hodges,  M.R.,  Moerland,  M.,  Tessier,  Y.,  Levin,  A.A.,  Persson,  R.,  Lindholm,  M.W.,  Dumong Erichsen,  K.,  Ørum,  H.,  Cohen,  A.F. and  Burggraaf,  J. (2015): Antisense-mediated reduction of proprotein convertase subtilisin/kexin type 9 (PCSK9): a first-in-human randomized, placebo-controlled trial. Br. J. Clin. Pharmacol., 80, 1350-1361.
•  Varda,  L.,  Petreski,  T.,  Gradišnik,  L.,  Maver,  U. and  Bevc,  S. (2025): Narrative review of proximal tubular epithelial cell in-vitro co-culture models. Biofabrication, 17, 1.
•  Wang,  Z.,  do Carmo,  J.M.,  da Silva,  A.A.,  Fu,  Y. and  Hall,  J.E. (2020): Mechanisms of Synergistic Interactions of Diabetes and Hypertension in Chronic Kidney Disease: Role of Mitochondrial Dysfunction and ER Stress. Curr. Hypertens. Rep., 22, 15.
•  Wardwell-Swanson,  J.,  Suzuki,  M.,  Dowell,  K.G.,  Bieri,  M.,  Thoma,  E.C.,  Agarkova,  I.,  Chiovaro,  F.,  Strebel,  S.,  Buschmann,  N.,  Greve,  F. and  Frey,  O. (2020): A Framework for Optimizing High-Content Imaging of 3D Models for Drug Discovery. SLAS Discov., 25, 709-722.
•  Wolf-Yadlin,  A.,  Hu,  A. and  Noble,  W.S. (2016): Technical advances in proteomics: new developments in data-independent acquisition. F1000 Res., 5.
•  Wu,  H.,  Wahane,  A.,  Alhamadani,  F.,  Zhang,  K.,  Parikh,  R.,  Lee,  S.,  McCabe,  E.M.,  Rasmussen,  T.P.,  Bahal,  R.,  Zhong,  X.B. and  Manautou,  J.E. (2022): Nephrotoxicity of marketed antisense oligonucleotide drugs. Curr. Opin. Toxicol., 32.
•  Xiao,  X.,  Zhang,  J.,  Lang,  Y.,  Cai,  L.,  Yang,  Q.,  Liu,  K.,  Ji,  S.,  Ju,  X. and  Liu,  F. (2023): Associations of lactate dehydrogenase with risk of renal outcomes and cardiovascular mortality in individuals with diabetic kidney disease. Diabetes Res. Clin. Pract., 203, 110838.
•  Xu,  G.,  Ni,  Y.,  Cui,  C.,  Du,  J.,  Duan,  M.,  Yang,  Y. and  Yu,  C. (2025): Development and validation of a cell-based assay to determine the bioactivity of antibodies against CD39 targets. J. Pharm. Biomed. Anal., 265, 117021.
•  Zou,  C.,  Wang,  C. and  Lu,  L. (2022): Advances in the study of subclinical AKI biomarkers. Front. Physiol., 13, 960059.