2025 Volume 50 Issue 9 Pages 483-491
Ionic liquids (ILs) are salts with melting points below 100°C. These materials are promising novel solvents in organic reactions, as new electrolytes, and in protein stabilization, able to refold enzymes and aid in drug discovery. IL properties are strongly influenced by the types of their constituent cations and anions. To date, many ILs with unique solvent properties not found in water or organic solvents are reported. However, IL toxicity evaluations often focus on trends associated with widely used imidazolium cations. Consequently, knowledge regarding the toxicity of individual ionic structures and their combinations is limited. This study investigated alkylphosphonium and alkylammonium cation derivatives, evaluating their cytotoxicity against mouse macrophage J774.1 cells using dihydrogen phosphate ([dhp]) or bromide (Br) ions as anions. Experiments were conducted using alkylphosphonium cation derivatives ((C4H9)3RP+, [P444R]) with varying alkyl chain lengths (R: 4; C4H9, 8; C8H17, and 12; C12H25) and on tetrabutylammonium cation derivatives [N4444]. This study revealed that [P44412][dhp] yielded the highest toxicity, which decreased with shorter alkyl chains. A similar trend was observed with Br as the anion. For [P444R], anion variation did not significantly affect toxicity. Comparing [P4444][dhp] and [N4444][dhp], the phosphonium cation showed lower ED50 values, indicating higher toxicity. This tendency persisted when Br was used as the anionic species. In summary, for the ILs evaluated, the quaternary cationic species of the IL effects the greatest influence on IL toxicity.
Ionic liquids (ILs) are salts with melting points below 100ºC. These novel solvents consist solely of cations and anions, and through constituent ion selection, their chemical and physical properties may be tailored to specific purposes. In 1992, Wilkes and Zaworotko (1992) reported on ILs stable in both water and air. Since then, IL research underwent rapid progress. ILs are considered environmentally friendly for their low vapor pressure, flame resistance, and high stability (Rogers and Seddon, 2003). However, their environmental persistence and potential for bioaccumulation remain subjects of active investigation (Petkovic et al., 2011). Due to these and other unique properties not found in water or organic solvents, they have various applications, including as electrolytes in lithium-ion batteries and in the dissolution of poorly soluble cellulose. ILs are also of interest in pharmaceutical applications. For example, paclitaxel injections use Cremophor EL (polyethoxylated castor oil), a non-ionic surfactant, as a solubilizing agent. However, Cremophor EL is associated with hypersensitivity reactions upon intravenous administration (Singla et al., 2002). Thus, patients require premedication with antiallergic drugs before paclitaxel administration. To address hypersensitivity complications, Chowdhury et al. (2018) evaluated an IL composed of a choline cation and a glycine anion as a potential alternative solubilizing agent to completely replace Cremophor EL. Their findings indicate that this IL-based formulation may prevent the hypersensitivity reactions associated with Cremophor EL while preserving drug solubility. Additionally, by using ionic liquid-in-oil (IL/O) microemulsions with imidazolium-based ILs, acyclovir, a poorly soluble antiviral agent, gained improved solubility and enhanced skin permeability (Moniruzzaman et al., 2010).
Additionally, ILs gained attention as useful solvents in bioscience applications. Hydrated ILs, prepared by adding small amounts of water to the IL matrix, are capable of dissolving various biomolecules while preserving their native structures (Fujita et al., 2007a). The addition of water creates a microenvironment that promotes biomolecular stability, with the water content optimized to maintain both solubility and bioactivity of the dissolved species. Studies demonstrated that biomolecular stability may be tuned by IL ion selection (Fujita et al., 2007b). In such biocompatible hydrated ILs, water molecules exhibit similarities to the intermediate water observed at the solvation layer for biocompatible polymer surfaces (Tanaka et al., 2013). Notably, the dihydrogen phosphate anion ([dhp], H2PO4−) has been reported as one key component in the forming of intermediate water, indicating that [dhp] is associated with biomolecular stabilization (Rajapriya Inbaraj et al., 2023).
Despite increased activity developing IL applications across various fields, concerns remain regarding their environmental stability and toxicity. Presently, most toxicity studies have primarily examined ILs with imidazolium- and pyridinium-based cations, investigating the relationship between alkyl chain length and toxicity. However, limited information is available on the influence of the cationic structure and whether combinations of anionic and cationic species alter toxicity. Moreover, mechanisms of IL toxicity remain largely unexplored.
This study elucidates relationships among alkyl chain length, cation core structure, anion species, and their combined impact on toxicity. ILs typically consist of organic cations, which commonly bear alkyl chains of varying lengths attached to a core structure such as imidazolium, phosphonium, or ammonium, and are paired with either organic or inorganic anions. The length of the alkyl chains and the nature of the cationic core are key determinants of the physicochemical properties and biological activity of ILs. This study focused on the cytotoxicity of selected ILs composed of alkylphosphonium or alkylammonium cations in combination with either dihydrogen phosphate ([dhp]) or bromide (Br) anions. The cytotoxicity of various ILs was assessed using J774.1 mouse macrophage cells, and the median effective dose (ED50) values were determined. This cell line was chosen because macrophages are among the first immune cells to encounter foreign substances in biological systems and are widely used as a representative model for preliminary toxicity screening. To further investigate the mechanisms of cytotoxicity, lactate dehydrogenase (LDH) assays were performed. LDH release is a well-established marker of cell membrane integrity and is widely employed to evaluate membrane damage as a primary indicator of cytotoxic effects.
Dulbecco’s Modified Eagle’s Medium (DMEM) was purchased from Fujifilm Wako Pure Chemicals Corporation (Osaka, Japan). Antibiotic-Antimycotic (100X), fetal bovine serum (FBS), GlutaMAXTM supplement, and Opti-MEMTM I Reduced Serum Medium were purchased from Thermo Fisher Scientific (Waltham, MA, USA). The 100-mm cell culture dishes and 96-well plates were obtained from Violamo (Osaka, Japan). The Cell Counting Kit-8 and Cytotoxicity LDH Assay Kit-WST were purchased from Dojindo Laboratories (Kumamoto, Japan). J774.1 cells were acquired from RIKEN BioResource Research Center (Tsukuba, Japan). All other reagents were of the highest grade available.
ILsStructures of the ILs synthesized in this study are shown in Fig. 1. All chemicals and solvents for IL synthesis were purchased from Tokyo Chemical Industry Co., Ltd. (TCI), (Tokyo, Japan) and Kanto Chemical Co. Inc. (Tokyo, Japan), and were used as received. IL with Br anions, namely tetrabutylammonium bromide ([N4444]Br, TCI, >99% pure), tetrabutylphosphonium bromide ([P4444]Br, TCI, >99% pure), tributyl-n-octylphosphonium bromide ([P4448]Br, TCI, >98% pure) and tributyldodecylphosphonium bromide ([P44412]Br, TCI, >97% pure) were also evaluated. ILs with [dhp], tetrabutylammonium dihydrogen phosphate ([N4444][dhp]), tetrabutylphosphonium dihydrogen phosphate ([P4444][dhp]), tributyloctylphosphonium dihydrogen phosphate ([P4448][dhp]), and tributyldodecylphosphonium dihydrogen phosphate ([P44412][dhp]) were synthesized according to a previously published procedure (Fujita et al., 2007a). ILs with Br anions were dissolved in water and stirred with silver oxide for 2 hr in the dark at room temperature, and the resultant solution was filtered. The filtered solution was then mixed with a phosphoric acid solution. The solvent was evaporated, and the resulting product was dried under vacuum. The synthesized ILs were analyzed by 1H NMR spectroscopy, and the spectra showed only the expected signals, with no observable peaks corresponding to organic impurities.
Structures and abbreviations of the ILs used in this study. ILs containing either an alkylammonium ion ([N4444]) or an alkylphosphonium ion ([P444R], with R representing C4H9, C8H17, or C12H25) as the basic backbone cation, and [dhp] or Br as the anion, were used. Names and abbreviations of ILs are described in the Materials and Methods section.
J774.1 cells were cultured in 100-mm cell culture dishes using DMEM supplemented with 1% (v/v) Antibiotic-Antimycotic, 1% (v/v) GlutaMAXTM supplement, and 10% (v/v) heat-inactivated FBS. Cells were seeded in dishes and maintained at 37°C in a 5% CO2 incubator. When J774.1 cells reached approximately 80% confluency in a 100-mm cell culture dish, they were transferred to a 96-well plate at a density of 4 x 104 cells per well in 100 µL and cultured in a CO2 incubator for 24 hr.
ILs treatmentAfter J774.1 cells were cultured in a 96-well plate, the medium was removed and replaced with various concentrations of ILs (0–50 mM) in Opti-MEM, followed by incubation over another 24 hr. All tested ILs exhibited complete solubility in Opti-MEM at the concentrations evaluated. Stock solutions were prepared at high concentrations and subsequently diluted into the culture medium in small volumes (≤10% v/v) to minimize alterations to the final medium composition. For LDH activity measurements, cells were incubated for 0.5, 1, 2, 4, 8, and 24 hr.
Cell viabilityCell viability was assessed following the instructions provided in the Cell Counting Kit-8 manual. Specifically, after treating cells with IL, the medium was removed, and Opti-MEM containing 10% Cell Counting Kit-8 was added at 100 µL/well. Cells were then incubated in a CO2 incubator for 1.5 hr. Following the incubation, absorbance at 450 nm was measured using a microplate reader SH-9000 (Corona Electric Co., Ltd., Ibaraki, Japan) to quantify the amount of reduced formazan. Both control samples and background samples untreated with IL were prepared. For controls, after cell seeding, the medium was removed and 100 µL of Opti-MEM containing Cell Counting Kit-8 was added. For the background samples, no cells were seeded and 100 µL of Opti-MEM containing Cell Counting Kit-8 was added. Absorbance at 450 nm was measured in the same manner as that for cells treated with IL.
LDH activityLDH activity was assessed following the instructions provided in the Cytotoxicity LDH Assay Kit-WST. Briefly, after treating cells with IL, 100 µL of the treated medium was transferred to another 96-well plate. High and low controls, a background, and a volume correction control were prepared. For the high and low controls, after cell seeding, the medium was removed and 100 µL of Opti-MEM was added. For the background and volume correction controls, no cells were seeded and 100 µL of Opti-MEM was added. Following a specific treatment period, 20 µL of lysis buffer was added only to the high control and volume-correction control while the low control and background were left untreated. After incubation in a CO2 incubator for 0.5 hr, the medium was transferred to another 96-well plate and 100 µL of substrate solution was added to all wells. After incubation at room temperature for 0.5 hr in the dark, 50 µL of reaction stop solution was added to all wells. Absorbance at 490 nm was measured using a microplate reader SH-9000 to determine the amount of reduced formazan produced. The LDH activity of control cells was calculated by subtracting the absorbance of the low control from that of the high control, and the LDH activity of cells treated with IL was calculated by subtracting the absorbance of the low control from that of the IL-treated cells. The background absorbance was used for the absorbance of the IL-treated cells and the low control and the value corrected by subtracting the absorbance of the volume-corrected control was used for the absorbance of the high control.
Logistic analysisLogistic regression models in SAS were employed to calculate the ED50 values from the cellular viability data for each IL (Omori et al., 1998).
Statistical analysisStatistical analyses were performed using JMP Pro version 16.2.1 (SAS Institute Inc., Cary, NC, USA). Differences between groups were compared using Student’s t-test. Statistical significance was defined as p < 0.05 and p < 0.001. Data are presented as mean ± standard deviation (S.D.).
The cytotoxicity of various ILs towards J774.1 cells was evaluated by cell viability according to the protocol described in the Materials and Methods section. The viability of untreated cells served as the control and was considered 100%. The relative viability of IL-treated cells compared to untreated controls is shown in Fig. 2A and 2B. Among phosphonium-based ILs, those with shorter alkyl chains ([P4444][dhp] and [P4444]Br) maintained high cell viability up to 781 µM. However, beyond this concentration, cell viability decreased rapidly, reaching approximately 0% at 6250 µM. [P4448][dhp] and [P4448]Br, which have intermediate chain lengths, reduced cell viability to approximately 0% at 95.3 μM. The most hydrophobic phosphonium-based ILs, [P44412][dhp] and [P44412]Br, exhibited the highest cytotoxicity. Cell viability decreased to approximately 0% at 12.2 µM for [P44412][dhp] and 24.4 µM for [P44412]Br. The anion type (dhp versus Br) did not appear to influence the cytotoxicity of these ILs. The ammonium-based ILs ([N4444][dhp] and [N4444]Br) maintained high cell viability up to 2400 µM. However, beyond this concentration, cell viability decreased rapidly, reaching approximately 0% at 25000 µM.
Cell viability of J774.1 cells after 24-hr treatment with various ILs. IL concentrations ranged from 0-50 mM. In Panel A, the anion is [dhp] and the cations are [N4444] (○), [P4444] (●), [P4448] (▲) and [P44412] (■). In Panel B, the anion is Br and the cations are [N4444] (○), [P4444] (●), [P4448] (▲) and [P44412] (■). Cell viability was measured using the Cell Counting Kit-8 as described in the Materials and Methods section. Data are presented as mean ± S.D. (n=12-16 for Panel A, n=12 for Panel B).
The ED50 values for [P4444] and [N4444] in combination with [dhp] as an anion were 2020.7 ± 89.3 µM and 4612.8 ± 610.7 µM, respectively (Table 1). Despite having identical alkyl chain lengths, [P4444], with phosphorus as the core cation, exhibited significantly greater toxicity than [N4444] (p < 0.01, Student's t-test). Similarly, when combined with the Br anion, the ED50 values were 2435.0 ± 164.4 µM for [P4444] and 5019.3 ± 256.9 µM for [N4444]. As with [dhp], [P4444] showed significantly higher toxicity (p < 0.01, Student's t-test) when phosphorus was the cationic core atom.
| Cation | ED50 (μM) | |||||
|---|---|---|---|---|---|---|
| [dhp] | Br | |||||
| [N4444] | 4612.8 | ± | 610.7† | 5019.3 | ± | 256.9‡ |
| [P4444] | 2020.7 | ± | 89.3†, * | 2435.0 | ± | 164.4‡, * |
| [P4448] | 70.9 | ± | 7.8 | 72.8 | ± | 16.9 |
| [P44412] | 6.6 | ± | 1.1 | 5.2 | ± | 0.8 |
ED50 values were determined according to the methods described in the Materials and Methods section. Data are presented as the mean ± S.D. (n = 3-4). Statistical significance was determined by Student's t-test. †p < 0.01 between [N4444][dhp] and [P4444][dhp], ‡p < 0.01 between [N4444]Br and [P4444]Br, *p < 0.05 between [P4444][dhp] and [P4444]Br
The ED50 value of [P44412][dhp], which has the longest alkyl side chain among the ILs based on the phosphonium cation structure, was 6.6 ± 1.1 µM (Table 1, Fig. 2). Additionally, the ED50 value of [P4448][dhp], with the second-longest chain length, was 70.9 ± 7.8 µM, which was lower than that of the shortest [P4444][dhp]. This trend was similarly observed when the anion was Br, with ED50 values increasing in the order [P44412]Br, [P4448]Br, and [P4444]Br. This indicates a clear relationship between alkyl chain length and toxicity, implying a mechanistic influence of alkyl chain upon toxicity. Additionally, while cell viability for both [P44412][dhp] and [P44412]Br appear to increase at concentrations of 5000 µM or higher, this observed increase may not accurately reflect the true cell survival rate due to potential experimental limitations in the approach employed (Fig. 2). Trypan blue exclusion assay conducted under the same experimental conditions indicated that most cells were non-viable at these concentrations, suggesting that the apparent increase in viability at higher concentrations may not accurately reflect actual cell survival.
Cytotoxicity evaluation -Combination of anion species-When comparing the relationship between toxicity and different anion species with identical alkyl chain lengths, [N4444] showed no significant difference in toxicity (Table 1). ED50 values for [P4448][dhp] and [P4448]Br were 70.9 ± 7.8 µM and 72.8 ± 16.9 µM, indicating no significant variations in toxicity between [dhp] and Br for [P4448]. ED50 values for [P44412][dhp] and [P44412]Br were 6.6 ± 1.1 µM and 5.2 ± 0.8 µM, also indicating no significant variation in toxicity between [dhp] and Br for [P44412]. However, for [P4444], the [dhp] anion combination was significantly more toxic than Br (p < 0.05, Student's t-test). These results suggest that, except for [P4444], for cations with the same alkyl chain length, there is no major influence of anion species on cytotoxicity.
Concentration and exposure time influence on toxicityThe relationship between toxicity and IL exposure time for J774.1 cells, attributed to differences in alkyl chain length, was investigated (Fig. 3). Regarding the investigation of ILs in this study, cell viability and LDH activity were examined at three concentrations: a concentration with negligible toxicity after 24 hr of exposure (0.35 µM), then at approximately the ED50 values for each IL, and finally at a concentration for which cell survival was nearly 0% after 24 hr of exposure (25000 µM). For [P4444][dhp], no decrease in cell viability was observed at a concentration of 0.35 µM for any of the exposure times (0.5, 1, 2, 4, 8 and 24 hr). At a concentration of 2020.7 µM, approximately the ED50 value, cell viability decreased in a time-dependent manner, accompanied by increased LDH activity. After 24 hr of exposure, both the cell survival rate and LDH activity were recorded at 50%. At a concentration of 25000 µM, cell viability dropped to 60% at 0.5 hr, 30% at 1 hr, and almost 0% for exposure times exceeding 2 hr. LDH activity ranged from approximately 65% to 90% at exposure times exceeding 2 hr. For [P4448][dhp], cell viability did not decrease up to 8 hr after treatment at a concentration of 0.35 µM but showed a slight decreasing trend after 24 hr. When treated at a concentration of 70.9 µM, approximately the ED50 value, cell viability decreased in an exposure time-dependent manner similar to [P4444][dhp]. At a concentration of 25000 µM, the survival rate was nearly 0% at all times, with LDH activity remaining around 90% throughout the exposure period. For [P44412][dhp], almost no decrease in cell viability was observed up to 8 hr after treatment at a concentration of 0.35 µM. However, cell viability decreased to approximately 80% after 24 hr. When treated at a concentration of 6.6 µM, approximately the ED50 value, there was a time-dependent decrease in cell viability and a corresponding increase in LDH activity. At a concentration of 25000 µM, cell viability consistently dropped to approximately 0% at all times, but LDH activity remained at nearly 0% throughout the exposure period.
Time course of cell viability and LDH release in J774.1 cells after treatment with [P4444][dhp], [P4448][dhp] or [P44412][dhp]. J774.1 cells were exposed to three different concentrations of ILs for 0.5, 1, 2, 4, 8, and 24 hr: a non-toxic concentration (0.35 μM) showing no cytotoxicity after 24-hr exposure (white bars), a concentration at approximately the ED50 value (gray bars) and a toxic concentration (25000 μM) that reduced cell viability to approximately 0% after 24-hr exposure (black bars). Under certain conditions, cell viability occasionally exceeded 100%, which was likely attributable to experimental variability or potential hormetic effects. Cell viability and LDH activity were measured as described in the Materials and Methods section. Data are presented as mean ± S.D. (n = 12).
This study evaluated the correlation between quaternary cation core atoms, alkyl chain length, anion species, and cytotoxicity reflected by ED50 values. Cytotoxicities of [P4444], [P4448], and [P44412] were compared based on the alkyl chain length. Results showed that ED50 values for [P4444] and [P4448] differed by approximately 30-fold, with ED50 values of [P4448] and [P44412] differing by approximately 10-fold, independent of the anion species (Figs. 2 and 3, Table 1). These findings indicate that the primary influence of increasing IL toxicity is increasing cation alkyl chain length. The observed sharp decrease in cell viability at higher IL concentrations may be attributed to increased hydrophobicity resulting from longer alkyl chains, which promote stronger interactions with lipid bilayers and potentially lead to membrane disruption.
Ranke et al. (2004) investigated the toxicity of imidazolium cation-based ILs with varying alkyl chain lengths on IPC-81 leukemia cells and C6 rat glioma cells, reporting a correlation between longer alkyl chains and increased toxicity. This observation aligns with our findings, despite the use of different cells and IL types. Similarly, Jing et al. (2016) suggested that a longer alkyl chain enhances hydrophobicity for imidazolium-based ILs, which may increase affinity for cell lipid bilayer membranes, potentially inducing membrane damage. As a result, this may increase the affinity of the cells to the lipid bilayer membrane, potentially causing membrane damage (Jing et al., 2016).
Comparing the toxicity of [N4444] and [P4444] considering the cation core atom, phosphonium cations ([P4444]) ILs were significantly more toxic than ILs with ammonium cations ([N4444]) (Figs. 2 and 3, Table 1). This indicates that phosphorus, when used as the core atom in the cation structure, induces greater toxicity than nitrogen, despite identical alkyl chain lengths. Supporting this finding, Couling et al. (2006) reported similar results in their toxicity studies against Vibrio fischeri, a marine luminescent bacterium. Although, precise mechanisms driving these differences in cations core atom toxicity remain to be fully elucidated, this correlation is critical for the use of ILs. Further research is necessary to elucidate how variations in core atoms influence IL toxicity.
When comparing relative toxicity of [dhp] and Br given the same cation, [dhp] was significantly more toxic than Br for [P4444]. However, when the cations were [N4444], [P4448], and [P44412], no significant difference in toxicity between the two anions was observed (Figs. 2 and 3, Table 1). Even among cations with the same alkyl chain length, different cell viability reduction patterns were observed depending on the accompanying anion. For example, an increase in cell viability was observed when the concentration of the IL with the Br anion was increased for [P4444], [P4448], and [P44412], suggesting different cell viability curves compared to [dhp]. This suggests that, for Br paired with [P4444], [P4448], and [P44412], cells were possibly exposed to a subtoxic stimulus or underwent an hormetic response (Weaver et al., 2010). A behavior that was not observed with [dhp] or with Br for [N4444]. Therefore, identification of a consistent trend in toxicity or ED50 values between [dhp] and Br was not possible.
ILs containing Br and chloride ions are reportedly less toxic than those containing fluoride ions, such as tetrafluoroborate and tetrafluorophosphate (Ranke et al., 2004). Therefore, although no significant differences in toxicity were observed for anionic species in this study, toxicity variations may arise for different anionic species, such as fluoride ions.
The relationship between the onset of toxicity and exposure time for different alkyl chain lengths to J774.1 cells was investigated at three concentrations. No toxicity was observed at low concentrations of any ILs regardless of exposure time. When the treatment concentration was at approximately ED50 values, toxicity occurred reflecting exposure time. At higher concentrations, [P4444][dhp] and [P4448][dhp], but not [P44412][dhp], showed a correlation between viability and LDH activity (Fig. 3). Therefore, it is strongly suggested that cell death, except at high concentrations of [P44412][dhp], is caused by the failure of the cell membrane. It is speculated that at high concentrations of [P44412][dhp], cell death occurs through a mechanism unrelated to cell membrane failure. Currently, the mechanism of cell death in the high concentration range is unknown.
Notably, LDH activity was markedly low in these samples. Follow-up experiments revealed that [P44412][dhp] directly inhibits LDH enzymatic activity (data not shown), indicating that the reduced LDH signal likely results from assay interference rather than a true absence of membrane damage. Although the extent of LDH release remains uncertain, viability assays clearly demonstrate that [P44412][dhp] induces cell death. This suggests the involvement of alternative cytotoxic mechanisms, such as apoptosis, necroptosis, oxidative stress-induced cell death, or mitochondrial dysfunction. Further studies are currently underway to elucidate these pathways.
While cell viability was similar at 0.5 and 1 hr for high concentrations of [P4444][dhp] and at 24 hr for ED50, LDH activity was lower at 0.5 and 1 hr for high concentrations. This suggests that LDH may not be adequately released into the medium following cell membrane damage during short exposure times. To address this, future studies must account for the possibility that measurement limitations imposed by brief exposure durations influence the observed results.
To our knowledge, no previous study has elucidated the mechanism of toxicity development based on both exposure time and concentration, as observed with [P44412][dhp]. Future research should explore the relationship between exposure time and concentration, as well as toxicity mechanisms for other ILs used in this study, including [P44412][dhp], and for ILs based on imidazole cations, which are widely used.
In this study, it was observed that the toxicity of ILs with phosphonium cations increased with longer alkyl chain lengths. Additionally, phosphonium cations exhibited higher toxicity compared to nitrogen-based cations. The results also strongly suggest that [P44412][dhp] exhibits different toxicity mechanisms depending on concentration. As ILs are attracting increasing interest in pharmaceutical applications, including drug discovery and delivery, further research is warranted to elucidate their effects on metabolic enzyme induction and DNA integrity, since both are critical parameters for evaluating toxicological risk.
We thank Mr. Kaname Hasegawa for his expert guidance on ED50 value calculations, which was invaluable for the analytical aspects of this research. We also thank Dr. Kenichiro Ogura and Dr. Tomokazu Ohnuma for their helpful advice. This study was partially supported by JSPS KAKENHI (Grant Numbers JP22K05221 and JP22H04561 (K. F.)) and the Izumi Science and Technology Foundation.
Conflict of interestThe authors declare that there is no conflict of interest.
