Animals

This experimental protocol was approved by the Animal Care and Use Committee of Hebei Medical University (Approval No. IACUC-Hebmu-2025129; approval date: December 17, 2025). The animals used in this study were specific pathogen-free male C57BL/6 mice (n = 70), aged 8–10 weeks and weighing 20–25 g, supplied by Hebei Zizhen Biotechnology Co., Ltd. The mice were housed in an animal facility under a 12 hours (h) light/12 h dark cycle, at an ambient temperature of 20–24°C and a relative humidity of 40%–60%, with ad libitum access to food and water. All mice were acclimated for 7 days before the start of the experiment. General health status was assessed preoperatively, and only healthy mice without obvious signs of disease or behavioral abnormalities were included in the subsequent experiments. Mice with body weights outside the specified range or with unsuccessful model establishment were excluded. All animals meeting the inclusion criteria were included in the analysis, with no additional exclusions. All animal procedures were conducted in strict accordance with the institutional guidelines for animal experimentation, and the study was reported in compliance with the ARRIVE 2.0 guidelines.

In this study, humane endpoints were predefined for all animals. If an animal exhibited severe pain or distress, euthanasia was performed promptly. Animal health status and behavior were monitored throughout the experiment, with continuous observation during anesthesia and recovery, and twice-daily monitoring after surgery to ensure animal welfare. The specific humane endpoints included inability to eat or drink; severe lethargy with no response to stimulation; severe dyspnea; progressive marked swelling, ulceration, or necrosis of the affected hindlimb; evidence of self-mutilation tendency or obvious tissue loss; or a body weight loss of more than 20% relative to baseline. Once any of these endpoint criteria were met, immediate euthanasia was carried out. In this study, no animal died before meeting the euthanasia criteria.

The experimental duration for each animal consisted of 3 h of hindlimb ischemia followed by 24 h of reperfusion. At the end of the 24 h reperfusion period, all animals were euthanized at the predetermined experimental endpoint after completion of the corresponding experimental procedures. A total of 70 animals were used, and all animals ultimately underwent humane euthanasia. All surgeries and procedures were performed under isoflurane anesthesia (1.5-3%), and measures such as the use of heating pads were implemented to minimize pain and stress as much as possible. Euthanasia was performed by carbon dioxide inhalation. All experiments were conducted by trained researchers in accordance with relevant ethical and technical standards.

Experimental design and animal grouping

A total of 70 mice were included in this study. The main experimental cohort consisted of 60 mice, which were allocated into five groups using a computer-generated randomization sequence, with 12 mice per group: the Sham group, I/R group, ESK + I/R group, ESK + chloroquine phosphate (CQ) + I/R group, and CQ + I/R group. The Sham group underwent the same anesthesia and procedural manipulations but without tourniquet application, whereas the other groups were all subjected to the left hindlimb I/R model. The doses of esketamine (10 mg/kg; Hengrui Pharmaceutical, Jiangsu, China) and chloroquine phosphate (60 mg/kg; Aladdin, Shanghai, China) were determined based on previous animal studies and the results of preliminary experiments [15,16]. These agents were administered intraperitoneally 20 min and 30 min before ischemia, respectively. The I/R and Sham groups received an equal volume of normal saline.

To accommodate the assessment of different outcome measures, the mice in each group were further divided into two subgroups (n = 6 each). Subgroup 1 was used for mechanical pain threshold testing, skeletal muscle contractile force measurement, and wet-to-dry weight ratio analysis. Subgroup 2 was used for the rotarod test, serum inflammatory cytokine assays, and histological and molecular biological analyses of the gastrocnemius muscle. On the day of model establishment, laser speckle blood perfusion imaging was performed in all groups before ischemia, after ischemia, and at 24 h after reperfusion. At 24 h after reperfusion, after completion of the mechanical pain threshold and skeletal muscle contractile force assessments, the left gastrocnemius muscle from subgroup 1 was harvested for wet-to-dry ratio analysis. After completion of the rotarod test, blood samples and the left gastrocnemius muscle were collected from subgroup 2 for the analysis of serum inflammatory cytokines, hematoxylin and eosin (H&E) staining, immunofluorescence, Western blotting, and reverse transcription-quantitative polymerase chain reaction (RT-qPCR). With the exception of blood collection, all tissue sampling procedures were performed after euthanasia.

An additional 10 mice were used as an independent animal cohort, with two mice per group. The same randomization procedure, treatment regimen, ischemia/reperfusion modeling protocol, and tissue collection time points were applied to increase the number of biological replicates for Western blot and RT-qPCR analyses.

Establishment of the hindlimb I/R model

After induction of anesthesia with inhaled isoflurane, a McGivney hemorrhoidal ligator was used to apply an orthodontic rubber band to the proximal left hindlimb, thereby inducing ischemia of the left hindlimb for 3 h, followed by reperfusion for 24 h after release of the rubber band [17]. Throughout the ischemic period, anesthesia was maintained only during the application of the orthodontic rubber band, and body temperature was maintained at 37°C using a heating pad until the animals regained consciousness. Laser speckle imaging was performed after tourniquet application and before tourniquet release to confirm near-complete ischemia of the affected hindlimb, thereby verifying successful model establishment.

Blood perfusion imaging

A laser speckle contrast imaging system was used to assess hindlimb blood perfusion in mice. After anesthesia with 1.5% inhaled isoflurane, the mice were placed on a black imaging platform, and blood perfusion in the plantar regions of both hind paws was measured using a blood perfusion imaging device (RFLSI III, RWD, Shenzhen, China). During the experiment, perfusion images were acquired before ischemic treatment, after tourniquet application, and at 24 h after tourniquet release and reperfusion. The images were analyzed using the accompanying software, and the level of blood perfusion was expressed as the perfusion ratio of the ischemic hindlimb to the contralateral healthy hindlimb.

Rotarod test

Mice underwent rotarod pre-acclimation training for 3 consecutive days before model establishment, and performance was assessed using a mouse rotarod fatigue apparatus (ZS-RDM-XS, Zhongshi Dichuang, Beijing, China). The pre-acclimation training was performed in the accelerating mode, with an initial speed of 4 revolutions per minute (rpm), an acceleration of 20 rpm/min, and a maximum speed of 30 rpm for 5 min; each mouse was trained twice daily. Formal testing was conducted at 24 h after reperfusion, and the latency to fall was recorded using the same accelerating protocol. Each mouse was tested three times, and the mean value was used for statistical analysis. After the rotarod test, the mice were allowed to rest for 1 h before subsequent experiments.

Mechanical pain threshold assessment

Baseline mechanical pain threshold measurements were performed for 3 consecutive days before model establishment. Prior to testing, mice were placed in transparent chambers with a metal mesh floor and allowed to acclimate for 1 h. The plantar surface of the affected hind paw was then stimulated perpendicularly using von Frey filaments (Touch Test, North Coast Medical, Gilroy, CA, USA), and the 50% paw withdrawal threshold was determined using the up-down method [18]. Formal testing was conducted again at 24 h after reperfusion using the same procedure.

Skeletal muscle contractile force measurement

At 24 h after reperfusion, skeletal muscle contractile force was measured in mice. Under isoflurane anesthesia, the mice were placed in the prone position, and body temperature was maintained at 37°C using a heating pad. The left gastrocnemius muscle was kept continuously moistened with warm normal saline, and the distal tendon of the left gastrocnemius was connected to a high-precision force transducer (ADInstruments, Colorado Springs, CO, USA). After incision of the left biceps femoris to expose the sciatic nerve, the nerve was transected, and the distal nerve stump was stimulated using bipolar platinum electrodes to evoke gastrocnemius contraction. The parameters for nerve stimulation were as follows: twitch contraction, 10 V, 1 Hz, 1 ms; tetanic contraction, 10 V, 100 Hz, 1 ms, sustained for 2 s. For direct muscle stimulation, the electrodes were placed at the proximal end of the gastrocnemius, and the maximal force of twitch contraction (20 V, 1 Hz, 1 ms) and tetanic contraction (20 V, 100 Hz, 1 ms, sustained for 2 s) was recorded, with an interval of 2 min between each stimulation. Muscle contractile force data were recorded and analyzed using a PowerLab 8/30 data acquisition system and LabChart 8 software (ADInstruments, Colorado Springs, CO, USA).

Skeletal muscle wet-to-dry weight ratio

At 24 h after reperfusion, the intact left gastrocnemius muscle was harvested from each mouse. After gently removing surface blood and moisture, the wet weight was measured. The tissue was then dried in an oven at 80°C for 48 h until a constant weight was reached, and the dry weight was recorded. The wet-to-dry weight ratio was subsequently calculated to assess the degree of tissue edema.

Measurement of inflammatory cytokine levels

At 24 h after reperfusion, orbital blood was collected from the mice, and serum was isolated by centrifugation and stored at −80°C until further analysis. Serum levels of interleukin-1 beta (IL-1β), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α) were measured using chemiluminescent immunoassay kits purchased from Wuhan Servicebio Technology Co., Ltd. (catalog numbers: GLM0010-100T, GLM0005-100T, and GLM0004-100T). All procedures were performed strictly in accordance with the manufacturers’ instructions.

Assessment of skeletal muscle injury

At 24 h after reperfusion, the left gastrocnemius muscle was harvested from the mice for histological evaluation. The tissue was fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned transversely. After deparaffinization and rehydration, the sections were subjected to H&E staining, and images were observed and acquired under a bright-field microscope.

Western blot analysis

At 24 h after reperfusion, samples of the left gastrocnemius muscle were collected from the mice and stored at −80°C. The gastrocnemius tissue was homogenized and lysed in RIPA buffer, followed by centrifugation at 12,000 g for 20 min at 4°C to collect the total protein in the supernatant. Protein concentration was then determined using a BCA protein assay kit (Servicebio, Wuhan, China). One-fourth volume of 5 × loading buffer was added to the protein samples, which were then denatured at 95°C for 10 min. Equal amounts of total protein (20 μg/lane) were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and subsequently transferred onto polyvinylidene difluoride membranes by constant-current electrophoretic transfer. After transfer, the membranes were blocked with a protein-free rapid blocking solution for 20 min. According to the molecular weight of the target proteins, the membranes were cut and incubated separately with the following primary antibodies overnight at 4°C: rabbit anti-microtubule-associated protein 1 light chain 3 II/I (LC3-II/I) antibody (HUABIO, Hangzhou, China; ET1701−65; 1:2000), rabbit anti-sequestosome 1 (p62) antibody (HUABIO, Hangzhou, China; HA721171; 1:2000), rabbit anti-mammalian target of rapamycin (mTOR) antibody (HUABIO, Hangzhou, China; ET1608−5; 1:5000), rabbit anti-phosphorylated mTOR (p-mTOR) antibody (HUABIO, Hangzhou, China; HA60094; 1:1000), and rabbit anti-β-actin antibody (Servicebio, Wuhan, China; GB15003; 1:5000). After washing with Tris-buffered saline containing Tween, the membranes were incubated for 2 h at room temperature with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG) secondary antibody (Servicebio, Wuhan, China; GB23303; 1:10000). After further washing, protein bands were visualized using a chemiluminescent substrate (Applygen, Beijing, China; P1050), and band densitometric analysis was performed using ImageJ software, with β-actin as the internal reference for normalization.

Immunofluorescence

At 24 h after reperfusion, the left gastrocnemius muscle was collected from the mice for immunofluorescence analysis. Gastrocnemius injury was evaluated by laminin immunostaining combined with detection of endogenous IgG extravasation [19]. Fresh gastrocnemius tissue was embedded in Optimal Cutting Temperature compound and stored at −80°C, and 10 μm transverse sections of the gastrocnemius were subsequently prepared.

For staining, the sections were washed with phosphate-buffered saline (PBS) and then blocked for 1 h at room temperature with 5% bovine serum albumin containing 0.3% Triton X-100. After removal of the blocking solution, the sections were incubated overnight at 4°C with a primary antibody against laminin (HUABIO, Hangzhou, China; HA721532; 1:1000). On the following day, the sections were rewarmed to room temperature and washed with PBS. They were then incubated for 2 h at room temperature in the dark with the following fluorescent secondary antibodies: Alexa Fluor 488-conjugated goat anti-rabbit IgG (HUABIO, Hangzhou, China; HA1121; 1:500) for detection of the laminin signal, and Alexa Fluor 594-conjugated goat anti-mouse IgG (HUABIO, Hangzhou, China; HA1126; 1:500) for detection of endogenous IgG extravasation in the tissue. After PBS washing, the sections were mounted with an anti-fade mounting medium containing DAPI, and images were observed and acquired under a fluorescence microscope.

All immunofluorescence images were acquired using identical parameters. Quantitative analysis was performed using ImageJ software after background subtraction. For each mouse, two transverse gastrocnemius sections were analyzed. From each image, 30 muscle fibers with clear boundaries and intact structure were randomly selected, and the mean fluorescence intensity of endogenous IgG within the muscle fibers was measured. The average value from the two images was taken as the result for that mouse.

RT-qPCR

At 24 h after reperfusion, the left gastrocnemius muscle was collected from the mice for RT-qPCR analysis. Total RNA was extracted using the FastPure Complex Tissue/Cell Total RNA Isolation Kit (Vazyme, Nanjing, China; RC113). Genomic deoxyribonucleic acid (DNA) contamination was removed and cDNA was synthesized according to the manufacturer’s instructions using HiScript III RT SuperMix for qPCR (+gDNA wiper) (Vazyme, Nanjing, China; R323). Real-time quantitative PCR was then performed on a QuantStudio 6 Flex Real-Time PCR System (96-well, 0.2 mL block) using the SYBR Green method with ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China; Q711). The total reaction volume was 20 μL per well, and primer concentrations were prepared according to the manufacturer’s recommended conditions. The amplification protocol was as follows: 50°C for 2 min and 95°C for 30 s, followed by 40 cycles of 95°C for 10 s and 60°C for 30 seconds (s). After amplification, melt curve analysis was performed under the following conditions: 95°C for 15 s, 60°C for 1 min, and then a gradual increase to 95°C at 0.05°C/s with continuous fluorescence signal acquisition to verify amplification specificity. Gene expression levels were analyzed using the 2^^-ΔΔCt method [20], with β-actin as the internal reference gene and the Sham group as the calibrator for relative expression. The primer sequences used are listed in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Primer sequences used for RT-qPCR.

https://doi.org/10.1371/journal.pone.0358125.t001

Statistical analysis

Data are presented as the mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, USA). Under the assumptions of normal distribution and homogeneity of variance, comparisons among multiple groups were conducted using one-way analysis of variance, followed by Tukey’s post hoc multiple-comparisons test. A value of P < 0.05 was considered statistically significant. The researchers were blinded to the group allocation during outcome assessment and data analysis.

ESK attenuated hyperperfusion at 24 h after hindlimb I/R

Laser speckle imaging was used to dynamically assess blood perfusion in both hindlimbs of mice. The results showed that blood perfusion in the ischemic hindlimb was markedly reduced in all groups during tourniquet application, whereas perfusion recovered and developed into a hyperperfusion state at 24 h after reperfusion.

Before tourniquet application, there was no statistically significant difference in the hindlimb perfusion ratio (ischemic side/contralateral side) among the groups. During tourniquet application, the perfusion ratio of the ischemic hindlimb in each intervention group was significantly decreased compared with baseline, with no statistically significant difference among groups. At 24 h after reperfusion, the perfusion ratio in the I/R group was significantly higher than that in the Sham group, whereas ESK treatment significantly reduced this ratio. After co-administration of CQ, the effect of ESK showed a decreasing trend; CQ alone did not improve hyperperfusion. These findings suggest that ESK can attenuate the hyperperfusion state induced by I/R (Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Effects of esketamine on blood perfusion after hindlimb ischemia-reperfusion in mice.

(A) Laser speckle images at different stages, where red indicates high perfusion and blue indicates low perfusion.(B) Quantitative analysis before ischemia.(C) Quantitative analysis during ischemia.(D) Quantitative analysis at 24 h after reperfusion. Data are presented as the mean ± standard deviation (n = 6). Intergroup comparisons labeled with different letters indicate statistically significant differences (P < 0.05).

https://doi.org/10.1371/journal.pone.0358125.g001

ESK improved motor function at 24 h after I/R

The rotarod test showed that the latency to fall was significantly shorter in the I/R group than in the Sham group, whereas ESK treatment significantly prolonged the latency to fall. This improvement was abolished by co-administration of CQ, and CQ alone did not improve motor function. These findings indicate that ESK can improve motor function after I/R.

The von Frey test showed that the 50% paw withdrawal threshold in the Sham group was lower than that in all other groups, whereas no statistically significant differences were observed among the I/R-treated groups. These results indicate that the mechanical withdrawal threshold was elevated at 24 h after reperfusion, and that neither ESK nor CQ treatment significantly affected this parameter (Fig 2).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Effects of esketamine on motor performance and mechanical sensitivity after hindlimb ischemia-reperfusion in mice.

(A) Rotarod test results, in which motor coordination was represented by the latency to fall from the accelerating rotarod.(B) Mechanical nociception test results, in which the paw withdrawal threshold of the hind paw was measured using von Frey filaments to assess mechanical sensitivity.Data are presented as the mean ± standard deviation (n = 6). Intergroup comparisons labeled with different letters indicate statistically significant differences (P < 0.05).

https://doi.org/10.1371/journal.pone.0358125.g002

ESK improved muscle contractile force at 24 h after I/R

Under direct muscle stimulation, both twitch force and tetanic force were significantly lower in the I/R group than in the Sham group, whereas ESK treatment partially restored both parameters. This improvement was attenuated after co-administration of CQ, while CQ alone produced no significant improvement. These findings indicate that ESK can partially ameliorate the decline in muscle contractile function induced by I/R, and that CQ can weaken this effect.

Under sciatic nerve electrical stimulation, the evoked twitch force and tetanic force were extremely low in all groups except the Sham group, with no statistically significant differences among groups. These results suggest that at 24 h after reperfusion, ESK did not restore nerve stimulation-driven muscle contraction, and CQ treatment did not alter this trend (Fig 3).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Effects of esketamine on muscle contractile force after hindlimb ischemia-reperfusion in mice.

(A) Twitch contraction, tetanic contraction, and quantitative analysis under electrical stimulation of the gastrocnemius muscle.(B) Twitch contraction, tetanic contraction, and quantitative analysis under electrical stimulation of the sciatic nerve.Data are presented as the mean ± standard deviation (n = 6). Intergroup comparisons labeled with different letters indicate statistically significant differences (P < 0.05).

https://doi.org/10.1371/journal.pone.0358125.g003

ESK attenuated I/R-induced skeletal muscle histopathological injury, edema, and disruption of sarcolemmal integrity

H&E staining showed that, in the Sham group, gastrocnemius muscle fibers were regularly arranged, structurally intact, and separated by relatively small interstitial spaces. In contrast, the I/R group exhibited marked muscle injury, characterized by disorganized muscle fiber arrangement, widened interstitial spaces, and aggravated structural disruption. These histopathological changes were alleviated in the I/R + ESK group, as evidenced by relatively preserved muscle fiber morphology and reduced interstitial expansion. In the I/R + ESK + CQ group, however, structural damage and interstitial expansion were more pronounced than those in the I/R + ESK group, while the I/R + CQ group still displayed obvious histological injury.

To further quantify tissue edema, the wet-to-dry weight ratio of the gastrocnemius muscle was measured. The results showed that the wet-to-dry ratio was significantly increased in the I/R group compared with the Sham group, whereas ESK treatment significantly reduced this ratio. This beneficial effect was attenuated after co-administration of CQ, while the wet-to-dry ratio remained at a relatively high level in the CQ-alone group.

Laminin staining combined with endogenous IgG immunofluorescence was used to evaluate skeletal muscle membrane integrity. The results showed that almost no IgG-positive signal was observed within muscle fibers in the Sham group, whereas IgG fluorescence intensity was markedly increased in the I/R group. ESK treatment significantly reduced IgG extravasation, and this effect was partially attenuated by co-administration of CQ; CQ alone did not produce an obvious improvement. Taken together, these findings indicate that ESK can attenuate I/R-induced gastrocnemius tissue injury, edema, and disruption of sarcolemmal integrity, whereas CQ can partially weaken its protective effects (Fig 4).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Effects of esketamine on skeletal muscle histopathology, edema, and sarcolemmal integrity after hindlimb ischemia-reperfusion.

(A) Longitudinal and transverse H&E-stained sections of the gastrocnemius muscle.(B) Representative images of the gastrocnemius muscle from each group before and after drying.(C) Immunofluorescence images of transverse gastrocnemius sections showing DAPI, IgG, Laminin, and merged staining in each group.(D) Quantitative analysis of the wet-to-dry weight ratio of the gastrocnemius muscle (n = 6).(E) Quantitative analysis of IgG fluorescence intensity (n = 3).Data are presented as the mean ± standard deviation. Intergroup comparisons labeled with different letters indicate statistically significant differences (P < 0.05).

https://doi.org/10.1371/journal.pone.0358125.g004

ESK attenuated the systemic inflammatory response induced by I/R and downregulated the transcription of certain inflammation-related genes

To evaluate the systemic inflammatory response induced by I/R, serum levels of TNF-α, IL-6, and IL-1β were measured in each group using chemiluminescent immunoassay. The results showed that serum TNF-α and IL-6 levels were significantly elevated in the I/R group, whereas ESK treatment significantly reduced both. This anti-inflammatory effect was attenuated after co-administration of CQ. In contrast, no significant differences in IL-1β levels were observed among the groups.

RT-qPCR was further performed to assess the transcript levels of inflammation-related genes in the gastrocnemius muscle. The results showed no statistically significant difference in Casp1 mRNA expression among the groups. Compared with the Sham group, the I/R group exhibited increased mRNA expression of Ccl2 and Nlrp3. Following ESK treatment, the expression levels of these genes were reduced relative to those in the I/R group. After combined treatment with CQ, Il1b mRNA expression was significantly increased, whereas Ccl2, Il10, and Nlrp3 mRNA expression showed no consistent trend (Fig 5).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Effects of esketamine on the systemic inflammatory response and inflammation-related gene expression after hindlimb ischemia-reperfusion.

(A) mRNA expression levels of target genes in each group, normalized to the internal reference gene and presented as relative expression levels (n = 5). (B) Heatmap of the relative expression levels of target genes in each group. Data are shown as log₂(2^^-ΔΔCT), and color variation indicates the relative level of gene expression. (C) Serum levels of TNF-α, IL-6, and IL-1β in each group are also shown (n = 6).Data are presented as the mean ± standard deviation. Intergroup comparisons labeled with different letters indicate statistically significant differences (P < 0.05).

https://doi.org/10.1371/journal.pone.0358125.g005

ESK suppressed mTOR phosphorylation and induced autophagy-related changes, which were attenuated by CQ

To clarify the effect of ESK on autophagy-related pathways after I/R, the expression of autophagy-related proteins in the gastrocnemius muscle was assessed at 24 h of reperfusion. The results showed that, compared with the I/R group, ESK treatment increased the LC3 II/I ratio, decreased p62 expression, and reduced the p-mTOR/mTOR ratio, suggesting that ESK treatment is associated with inhibition of mTOR signaling and alterations in the expression of autophagy-related proteins.

Following co-treatment with CQ, the LC3-II/I ratio remained elevated, while p62 expression tended to increase compared with that in the ESK-alone group. The CQ-alone group similarly exhibited an elevated LC3-II/I ratio and marked p62 accumulation. These findings suggest that CQ may partially inhibit the ESK-associated autophagic degradation process, resulting in the accumulation of LC3-II and the autophagic substrate p62. (Fig 6).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Effects of esketamine on autophagy-related proteins in skeletal muscle after hindlimb ischemia-reperfusion in mice.

(A) Representative immunoblot bands of each protein.(B) Densitometric quantification of each protein.Data are presented as the mean ± standard deviation (n = 5). Intergroup comparisons labeled with different letters indicate statistically significant differences (P < 0.05).

https://doi.org/10.1371/journal.pone.0358125.g006