Analysis of sex differences in cisplatin-induced conditioned taste aversion

Abstract

Chemotherapy-induced taste disorders and food aversion are important adverse effects that lead to reduced food intake and a deteriorated quality of life. Sex differences have been proposed to influence their development, but findings remain inconsistent, and no clinical consensus has been established. Further basic research is required to clarify these discrepancies underlying mechanisms. In this study, conditioned taste aversion (CTA) was evaluated in C57BL/6J mice after intraperitoneal administration of cisplatin, prepared in physiological saline, using a two-bottle choice test with distilled water and a saccharin-containing solution as drinking options. To assess the influence of saccharin on sweet taste sensitivity, four concentrations of saccharin-containing solutions were tested. A clear sex difference emerged during CTA development. Male mice showed reduced aversion at lower saccharin concentrations, that did not elicit preference, and no aversion at the lowest concentration. In contrast, female mice exhibited the onset of CTA even at concentrations that did not induce preference. Given reports of higher sweet-taste sensitivity in females, these findings suggest that CTA is driven by taste stimulation rather than taste preference. Additionally, female mice showed a lower propensity to develop CTA at 0.02% saccharin concentration than male mice. To investigate whether female sex hormones regulate CTA development, estradiol (E2) was administered subcutaneously before cisplatin treatment, followed by the same behavioral assessments. E2 administration ameliorated CTA, suggesting that E2 may facilitate the extinction of aversive learning. In summary, cisplatin-induced CTA appears to be triggered by taste stimulation itself, and E2 may represent a potential therapeutic approach for mitigating cisplatin-induced CTA.

Highlights

• •Male mice showed reduced conditioned taste aversion at low saccharin levels.
• •Female mice developed conditioned taste aversion at nonpreferred concentrations.
• •Female mice showed lower conditioned taste aversion at 0.02% saccharin.
• •Estradiol may mitigate cisplatin-induced conditioned taste aversion.

1. Introduction

Taste disorders and food aversion associated with anticancer chemotherapy are important adverse effects that lead to reduced food intake and a decline in quality of life (QOL) in patients with cancer [1,2]. In recent years, several studies have suggested that sex differences may contribute to the development of chemotherapy-related food aversion, with some clinical reports indicating a higher incidence i women [3], whereas others have reported no clear association with sex nor a failure to detect sex differences [4,5]. Thus, the clinical evidence remains inconsistent.

In contrast, basic research using conditioned taste aversion (CTA) paradigms has suggested the possibility of sex-based differences in aversion formation. Classical studies have reported that the minimum effective dose required to induce aversion differs between male and female [6]. More recent reports have shown that sex differences in nausea- and aversion-related behaviors can be present or, in some cases, limited [[7], [8], [9]]. Therefore, a consensus has not yet been reached, even at the basic research level. Given the insufficient evidence in both clinical and preclinical fields, experimental validation is required to support clinical observations regarding sex differences.

Cisplatin is a widely used chemotherapeutic agent for solid tumors and is known to induce taste disturbances, nausea, vomiting, and reduced food intake [[10], [11], [12]]. However, whether sex differences contribute to cisplatin-induced CTA remains unclear. Clarifying these sex-dependent mechanisms is crucial for understanding the pathophysiology of chemotherapy-induced CTA and for developing effective therapeutic strategies. We have previously established a mouse model of cisplatin-induced CTA and demonstrated that a single administration of cisplatin induces a robust and long-lasting aversive response [13]. In the present study, we examined whether sex differences contribute to the development of CTA by comparing the responses of male and female mice.

2. Materials & methods

2.1. Animals

All experiments were performed using C57BL/6J mice (Japan SLC, Shizuoka, Japan) (135 male and 84 female mice). Mice were housed at 23–25 °C with free access to food and water. Cisplatin (30 μg; FUJIFILM Wako Chemicals Corporation, Osaka, Japan) was prepared in 100 μL of physiological saline (Otsuka Normal Saline, 20 mL ampoule; Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan) and administered intraperitoneally to each mouse.

In addition, 100 μL of 17β-estradiol (E2; FUJIFILM Wako Chemicals Corporation) suspension, prepared in corn oil at a concentration of 20 μg/100 μL, was administered subcutaneously to each mouse 1 h before cisplatin administration. Although physiological circulating estradiol concentrations in mice are substantially lower [14,15], this dose was intentionally selected to induce pharmacological estrogenic effects rather than physiological hormone replacement, based on previous literature [16,17]. The 1 h interval was chosen to permit the initiation of rapid estrogen signaling before cisplatin exposure [18,19], while behavioral effects were assessed over the subsequent 5-day drinking test.

All procedures were approved by the Animal Ethics Committee of the University of Osaka (approval number 24-085-004) and were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study followed the ARRIVE guidelines, and all efforts were made to reduce the number of animals used and to minimize distress.

2.2. Two bottle tests

A two-bottle test was performed according to previously published methods [13]. Saccharin (FUJIFILM Wako Chemical Corporation; cat. no. 199-08665) was dissolved in distilled water (DW) to prepare solutions at concentrations of 0.2, 0.02, 0.002, and 0.0002%. In the two-bottle test, mice were provided free access to two bottles: one containing DW and the other containing a saccharin solution. The bottles used in the test were microvolume drinking bottles (SN–950H; Shinano Seisakusho Co., Ltd., Tokyo, Japan). The test was initiated immediately after intraperitoneal injection of cisplatin or saline. To avoid place-dependent preferences, the positions of the two bottles were changed daily. Daily fluid consumption was recorded by measuring the weight of the bottles. Fluid consumption was determined as the total volume consumed by three mice housed in each cage. The number of cages assigned to each experimental group was calculated as follows: In the sex-difference analysis of CTA, the saline-treated group consisted of three male and three female cages for all saccharin concentrations. In the cisplatin-treated group, the numbers were as follows: 0.2% saccharin (five male cages, five female cages), 0.02% saccharin (six male cages, five female cages), 0.002% saccharin (three male cages, three female cages), and 0.0002% saccharin (four male cages, four female cages). In the E2 administration experiment, the E2-treated groups consisted of three cages in the cisplatin-treated group and three cages in the saline-treated group. In the non–E2-treated condition, the cisplatin-treated group consisted of six cages, whereas the saline-treated group consisted of three cages. The number of cages differed depending on saccharin concentration and sex. Variability tended to change according to both variables, and inter-cage variability in particular was influenced by saccharin concentration and sex. When greater variability was observed, additional animals were included to ensure adequate statistical reliability.

In this study, CTA was assessed by examining changes in the preference for saccharin-containing solutions. The preference score for saccharin-containing solution was defined as the proportion of saccharin-containing solutions consumed relative to total fluid intake. The preference score ratio was calculated using the following formula:

Preference score ratio = (preference score in the cisplatin-treated group)/(mean preference score in the control group).

2.3. Statistical analysis

For all studies, statistical analyses were performed using a two-way analysis of variance (two-way ANOVA) to evaluate differences among the four experimental groups. When significant main effects or interactions were detected, post hoc multiple comparisons were conducted using unpaired t-tests to compare individual group differences. Statistical significance was defined as p < 0.05. Data are presented as the standard error of the mean (SEM).

3. Results

3.1. Concentration-dependent induction of cisplatin-induced conditioned taste aversion

As mice show a strong preference for a 0.2% saccharin -containing solution [20], we first performed a two-bottle choice test using a 0.2% saccharin-containing solution. When analyzed as daily averages, saccharin-containing solution intake was markedly reduced in cisplatin-treated mice compared with saline controls. In the order presented in the figures, cisplatin-treated males (Cis/male) consumed 3.32 ± 2.89 mL/day (n = 5) and cisplatin-treated females (Cis/female) consumed 2.28 ± 2.89 mL/day (n = 5), whereas saline-treated males (Saline/male) and females (Saline/female) consumed 16.13 ± 2.89 mL/day (n = 3) and 13.57 ± 2.89 mL/day (n = 3), respectively (Fig. 1b). Two-way ANOVA revealed a significant main effect of treatment (F(1,12) = 65.1, p < 0.0001), whereas neither the main effect of sex (F(1,12) = 1.47, p = 0.2506) nor the treatment × sex interaction (F(1,12) = 0.26, p = 0.6186) was significant.

Fig. 1.

(a–d) Bar graph showing average daily intake of saccharin-containing solution (dots) and total amount of fluid consumed (white). Saccharin concentration: 0.2% (a), 0.02% (b), 0.002% (c), and 0.0002% (d). Data are presented as mean ± SEM. Statistical significance was defined as *p < 0.01, ***p < 0.001, and ****p < 0.0001, determined by two-way ANOVA followed by post hoc unpaired t-tests. The numbers are: 0.2% saccharin (5 male cages, 5 female cages), 0.02% saccharin (0.02% Sac: 6 male cages, 5 female cages), 0.002% saccharin (0.002% Sac: 3 male cages, 3 female cages), and 0.0002% saccharin (4 male cages, 4 female cages). The saline‐treated group consisted of three male and three female cages for all saccharin concentrations.

(e–h) Bar graph of mean preference scores for each group in the two-bottle choice test. Saccharin concentration: 0.2% (e), 0.02% (f), 0.002% (g), and 0.0002% (h). Data are presented as mean ± SEM. Statistical significance was defined as *p < 0.05 and ***p < 0.001, determined by two-way ANOVA followed by post hoc unpaired t-tests.

Preference score analysis revealed a marked effect of treatment on aversive behavior. Cisplatin-treated males exhibited markedly reduced preference scores (24.44 ± 16.9%, n = 5), and cisplatin-treated females showed similarly low scores (16.32 ± 16.9%, n = 5). In contrast, saline-treated males and females exhibited high preference scores (81.87 ± 16.9%, n = 3 and 71.87 ± 16.9%, n = 3, respectively) (Fig. 1e). Two-way ANOVA demonstrated a significant main effect of treatment (F(1,12) = 41.86, p = 0.0003), whereas neither the main effect of sex (F(1,12) = 1.08, p = 0.3196) nor the treatment × sex interaction (F(1,12) = 0.01, p = 0.9160) was significant.

Next, the two-bottle choice test was performed using three saccharin concentrations (0.02%, 0.002%, and 0.0002%) to evaluate CTA induced by cisplatin treatment.

At 0.02% saccharin, cisplatin treatment markedly reduced saccharin intake. Two-way ANOVA revealed a significant main effect of treatment on saccharin intake (F(1,13) = 33.87, p < 0.0001), whereas neither sex (F(1,13) = 0.90, p = 0.3607) nor the treatment × sex interaction (F(1,13) = 2.16, p = 0.1647) was significant. Mean intake values were 3.97 ± 1.02 mL/day in cisplatin-treated males (n = 6) and 7.04 ± 1.12 mL/day in cisplatin-treated females (n = 5), compared with 13.23 ± 1.44 mL/day in saline-treated males (n = 3) and 12.57 ± 1.44 mL/day in saline-treated females (n = 3). Preference score analysis showed a similar pattern, with a significant main effect of treatment (F(1,13) = 24.9, p = 0.0002) but no significant effects of sex (F(1,13) = 0.56, p = 0.4637) or interaction (F(1,13) = 1.77, p = 0.2079). Preference scores were reduced in cisplatin-treated mice (30.77 ± 7.42% in males, n = 6; 49.98 ± 8.13% in females, n = 5) relative to saline controls (89.13 ± 10.49% in males, n = 3; 83.87 ± 10.49% in females, n = 3).

At 0.002% saccharin, saccharin intake did not significantly differ among groups. Two-way ANOVA detected no significant main effects of treatment (F(1,8) = 0.92, p = 0.3631) or sex (F(1,8) = 0.001, p = 0.9738), and no interaction effect (F(1,8) = 0.12, p = 0.7438). Intake values were 5.57 ± 1.97 mL in cisplatin-treated males (n = 3), 6.17 ± 1.97 mL in cisplatin-treated females (n = 3), 8.13 ± 1.97 mL in saline-treated males (n = 3), and 7.40 ± 1.97 mL in saline-treated females (n = 3). Preference scores were likewise comparable across groups (cisplatin males: 45.67 ± 14.28%, n = 3; cisplatin females: 49.73 ± 14.28%, n = 3; saline males: 62.07 ± 14.28%, n = 3; saline females: 47.83 ± 14.28%, n = 3), with no significant effects of treatment (F(1,8) = 0.26, p = 0.6254), sex (F(1,8) = 0.13, p = 0.7311), or interaction (F(1,8) = 0.41, p = 0.5397).

At 0.0002% saccharin, saccharin intake was also similar among groups. Two-way ANOVA revealed no significant effects of treatment (F(1,10) = 1.15, p = 0.3107), sex (F(1,10) = 0.34, p = 0.5746), or treatment × sex interaction (F(1,10) = 0.15, p = 0.7082). Intake values were 6.60 ± 1.21 mL in cisplatin-treated males (n = 6), 4.95 ± 2.09 mL in cisplatin-treated females (n = 2), 7.77 ± 1.71 mL in saline-treated males (n = 3), and 7.43 ± 1.71 mL in saline-treated females (n = 3). Preference scores similarly showed no significant main effects of treatment (F(1,10) = 0.001, p = 0.9779), sex (F(1,10) = 0.10, p = 0.7539), or interaction (F(1,10) = 0.67, p = 0.4308), with values of 58.73 ± 10.35% (cisplatin males, n = 6), 42.00 ± 17.93% (cisplatin females, n = 2), 47.13 ± 14.64% (saline males, n = 3), and 54.43 ± 14.64% (saline females, n = 3).

Collectively, these results demonstrate that cisplatin induced robust CTA n at higher saccharin concentration (0.2% and 0.02%), whereas lower concentrations (0.002% and 0.0002%) failed to detect significant treatment- or sex-dependent effects, indicating a concentration-dependent sensitivity of the behavioral assay.

3.2. Concentration-dependent changes in CTA incidence and sex differences

In our previous study, cisplatin-induced CTA was confirmed in male mice using a 0.2% saccharin solution. In the present study, because all animals exhibited preference score ratios below 0.7 under the 0.2% condition (Fig. 2a), a preference score ratio of 0.7 was defined as the threshold for CTA induction.

Fig. 2.

(a) Dot plot of preference score ratios for each cage in the two-bottle choice test of the cisplatin-treated groups. Black = male; white = female. If the subjects drank equal amounts of saccharin-containing solution and DW, the preference score ratio would be 1 (a vertical dot line). CTA onset was defined as a preference score ratio of 0.7 or lower (red line). The numbers were: 0.2% saccharin (5 male cages, 5 female cages), 0.02% saccharin (6 male cages, 5 female cages), 0.002% saccharin (3 male cages, 3 female cages), and 0.0002% saccharin (4 male cages, 4 female cages).

(b) Incidence of CTA in each group during the two-bottle choice test. Animals with a preference score ≥ 0.7 (red dot line) were considered not to have developed CTA. Black = male; white = female.

In male mice, all cages showed preference score ratios below 0.7 at the 0.02% saccharin concentration. In contrast, at 0.002%, two of three cages, and at 0.0002%, all cages exhibited ratios exceeding 0.7. In female mice, although all cages showed ratios below 0.7 at 0.2%, more than half of the cages exceeded the 0.7 threshold at all lower concentrations.

Defining CTA induction as a preference score ratio ≤0.7, the number of cages exhibiting CTA was used to calculate the CTA incidence rate. In male mice, CTA incidence remained 100% up to the 0.02% concentration but sharply declined to 33.3% at 0.002% and was completely absent at 0.0002% (Fig. 2b). In contrast, female mice exhibited CTA incidence rates of ≤50% at concentrations of 0.02% and below.

Collectively, these findings indicate that although males tended to be less susceptible to CTA than females under high saccharin concentrations, females were still capable of developing CTA under low-concentration conditions in which males no longer exhibited CTA.

3.3. Estradiol attenuates cisplatin-induced taste aversion

As female mice showed a lower incidence of CTA in the experiment using a 0.02% saccharin-containing solution, E2 was considered a potential modulator capable of ameliorating CTA. To examine this possibility, E2 was administered subcutaneously to male mice 1 h before cisplatin injection. Subsequently, a two-bottle choice test was conducted using a solution containing 0.02% saccharin.

For saccharin-containing solution intake, two-way ANOVA revealed a significant overall model effect (F(3,11) = 7.98, p = 0.0042), with a significant main effect of treatment (F(1,11) = 8.12, p = 0.0157), no main effect of E2 administration (F(1,11) = 0.08, p = 0.7791), and a significant treatment × E2 interaction (F(1,11) = 10.89, p = 0.0071). Mean intake values were 3.97 ± 1.14 mL/day in cisplatin-treated mice without E2 (n = 6) and 9.37 ± 1.61 mL/day with E2 (n = 3), compared with 13.23 ± 1.61 mL/day in saline-treated mice without E2 (n = 3) and 8.70 ± 1.61 mL/day with E2 (n = 3). Post hoc analysis confirmed significantly greater intake in E2-treated cisplatin mice than untreated cisplatin mice (p = 0.0194), indicating that cisplatin-induced reduction of saccharin intake was attenuated by E2 (Fig. 3a).

Fig. 3.

(a)Bar graph showing average daily intake of 0.02% saccharin-containing solution (0.02% Sac: dots) and total fluid consumed (white) E2-and cisplatin-treated group (E2/Cisplatin: 3 cages), E2-and saline-treated group (E2/Saline: 3 cages), E2-untreated and cisplatin-treated group (Cisplatin: 6 cages), and E2-untreated and saline-treated group (Saline: 3 cages). Data are presented as mean ± SEM. Statistical significance was defined as *p < 0.01, and ***p < 0.001, determined by two-way ANOVA followed by post hoc unpaired t-tests.

(b) Bar graph of the mean preference scores for each group in the two-bottle choice test using a 0.02% saccharin-containing solution (0.02% Sak). Data are presented as mean ± SEM. Statistical significance was defined as *p < 0.01, and **p < 0.01, determined by two-way ANOVA followed by post hoc unpaired t-tests.

Preference score analysis similarly showed a significant overall model effect (F(3,11) = 5.77, p = 0.0128), a significant main effect of treatment (F(1,11) = 6.35, p = 0.0283), no main effect of E2 (F(1,11) = 0.63, p = 0.4464), and a significant treatment × E2 interaction (F(1,11) = 6.11, p = 0.0312). Preference scores were higher in E2-treated cisplatin mice (68.87 ± 12.50%, n = 3) than untreated cisplatin mice (30.77 ± 8.84%, n = 6), whereas saline-treated mice showed scores of 89.13 ± 12.50% without E2 (n = 3) and 69.50 ± 12.50% with E2 (n = 3). Post hoc testing confirmed a significant increase in preference score following E2 administration in cisplatin-treated mice (p = 0.0301; Fig. 3b).

Together, these findings demonstrated that E2 administration suppressed the development of cisplatin-induced CTA.

4. Discussion

To investigate the influence of sex on cisplatin-induced CTA, a two-bottle choice test was performed in male and female mice using saccharin-containing solutions at four concentrations. The results revealed that sex differences emerged as saccharin concentration decreased.

In males, the incidence of CTA declined as the saccharin concentration decreased to 0.002% below the detection threshold. Previous studies using two-bottle choice tests in male C57BL/6J mice reported a saccharin detection threshold of 0.02% [21]. Consistent with this, the preference score in the control group approached 0.5 at concentrations of 0.002% or lower, indicating the absence of saccharin preference (Fig. 1g and h). Importantly, the total daily fluid intake remained stable across groups, indicating that changes in the preference score reflected altered taste preference rather than differences in overall fluid consumption. Therefore, the reduced incidence of CTA in males treated with the 0.002% saccharin-containing solution is likely attributable to weakened taste stimulation, and the absence of aversion to the 0.0002% saccharin-containing solution suggests that this concentration fell at or below the detection threshold.

In contrast, female mice exhibited a decrease in CTA incidence at 0.02%, but the incidence remained unchanged at concentrations of 0.002% and below. These findings indicate that only female mice were able to detect taste stimulation from the 0.0002% saccharin-containing solution, resulting in the development of CTA. Indeed, sex differences have been reported in taste preference, detection thresholds, and responsiveness to taste stimuli [22]. Human studies have similarly shown that females have lower detection thresholds and thus higher sensitivity for sweet tastes than males [23,24]. However, detailed investigations of sex differences in sweet taste perception in mice are limited. Studies using rats have reported that estrogen increases the detection threshold for sucrose, thereby reducing sensitivity [25]. Thus, the presence and nature of sex differences in sweet taste detection thresholds remain a matter of debate.

Therefore, it is not yet clear whether the observed sex differences in CTA at low saccharin concentrations are attributable to sex-specific differences in detection thresholds. Further investigation is required to clarify this point.

In this study, we demonstrated that CTA was induced by taste stimulation rather than by preference. Clinical studies have reported that food items such as meat and chocolate are often targets of aversion in patients undergoing chemotherapy [26,27]. Given that these foods likely provide strong taste stimulation and are generally preferred, it was hypothesized that low concentrations of saccharin—expected to produce weak taste stimulation and low palatability—would not induce CTA. However, in the cisplatin-untreated groups receiving 0.002% or 0.0002% saccharin-containing solution, neither male nor female mice had a preference for saccharin; however, some female mice still developed CTA.

These findings suggest that CTA is triggered by taste stimulation rather than by the presence or absence of preference.

CTA studies have traditionally used lithium chloride (LiCl)–based classical conditioning paradigms, typically employing ∼0.1% saccharin as the conditioned stimulus [28,29]. Because LiCl rapidly induces strong visceral malaise, robust aversion can also be established at lower saccharin concentrations (<0.1%) [30,31]. In contrast, chemotherapy-induced aversion models using cisplatin generally employ higher saccharin concentrations (≥0.1%), likely reflecting differences in the physiological actions of LiCl and cisplatin [[32], [33], [34]].

LiCl rapidly induces acute gastrointestinal discomfort, generating strong interoceptive signals that facilitate efficient associative learning and lower the taste threshold for aversion formation [30]. In contrast, cisplatin-induced malaise develops more gradually through delayed systemic toxicity and neuroinflammatory mechanisms [35]. As a result, intake avoidance in chemotherapy models likely reflects not only associative CTA learning but also altered taste perception and reduced ingestive motivation [36,37], which may necessitate higher saccharin concentrations.

Importantly, the present study demonstrated that cisplatin induced clear CTA even at a low saccharin concentration of 0.02% regardless of sex (Fig. 1). Notably, more than half of the female mice showed aversion at even lower concentrations in some cages (Fig. 2). To our knowledge, few rodent studies have examined cisplatin-induced CTA under such low saccharin conditions. Therefore, these findings raise the possibility that chemotherapy-induced CTA may reflect altered taste sensitivity or reward valuation rather than simple aversive learning.

The present study revealed a clear sex-based differences in cisplatin-induced CTA under defined taste stimulus conditions. Specifically, female mice showed a lower incidence of CTA than male mice when tested with a 0.2% saccharin-containing solution, indicating reduced susceptibility to aversive learning at higher taste intensities. Several studies of LiCl–induced CTA have reported sex-related differences. In rats, females are less likely to develop CTA than males, and gonadectomy increases the susceptibility of females to the aversion [38]. Our findings extend these observations to a chemotherapy-induced CTA model and suggest that biological sex modulates drug-induced aversive learning. Importantly, stable fluid intake across groups indicates that the observed sex difference likely reflects differences in aversive taste memory rather than consumption-related factors.

Another study reported that the implantation of E2 capsules in rats ameliorated LiCl–induced CTA [17], suggesting that E2, among estrogens, may play a key role in suppressing CTA. Based on these findings, E2 was administered before cisplatin injection, followed by a two-bottle choice test. This treatment suppressed the development of CTA. Notably, E2 administration attenuated CTA under identical conditioning procedures, supporting a causal role of estrogenic signaling in cisplatin-induced aversive learning. E2 has been reported to facilitate the extinction of fear memory in CTA learning and other fear-conditioning paradigms [39,40]. Consistent with these reports, E2 may promote extinction or weaken consolidation of aversive taste memory rather than alter taste perception.

5. Conclusion

In the present study, we demonstrated that cisplatin-induced CTA occurred even at saccharin concentrations that did not elicit a preference in control mice. Furthermore, sex-related differences in the incidence of CTA were observed at low saccharin concentrations. Additionally, subcutaneous administration of E2 ameliorated cisplatin-induced CTA. Therefore, our findings suggest that CTA is triggered by taste stimulation rather than by preference, and E2 may be a potential therapeutic agent for cisplatin-induced CTA.

Disclosure of interest

The authors report no conflict of interest.

Funding details

This research was supported by the Tohmonkai Foundation and university grants of The University of Osaka in Japan.

CRediT authorship contribution statement

Shuya Iitsuka: Data curation, Formal analysis, Investigation, Validation, Visualization, Writing – original draft. Yoshihisa Koyama: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing – original draft. Koh Shinoda: Supervision, Writing – review & editing. Shoichi Shimada: Supervision, Writing – review & editing.

Declaration of competing interest

The authors report no conflict of interest.

Data availability

No data was used for the research described in the article.