† These authors contributed equally.
We evaluated the practicability of using the rarely utilized C57BL/6N mouse as a Parkinson’s disease model established via the acute MPTP/probenecid (MPTP/p) protocol. We confirmed dopaminergic degeneration in terms of decreased expression levels of tyrosine hydroxylase in the substantia nigra and striatum of MPTP/p-lesioned mice. In addition, acute MPTP/p-lesioned mice demonstrated initial motor dysfunctions followed by spontaneous recovery. Interestingly, these MPTP/p-lesioned mice exhibited anxiolytic and antidepressive behaviors upon recovery from these motor deficits. Additionally, increased expression of norepinephrine transporters in several brain regions, including the hippocampus, medial prefrontal cortex, and striatum, and an elevated rate of adult neurogenesis (in terms of increased numbers of doublecortin-positive neuroblasts) in the hippocampus were observed after recovery from motor dysfunctions. We suggest that the emotional alterations observed under these experimental conditions may be associated with enhanced adult neurogenesis, increased levels of norepinephrine transporters, and/or a possible interplay between these two factors. Consequently, this acute MPTP/p model adequately satisfies the criteria for the validity of a Parkinson’s disease model regarding dopaminergic loss and motor impairment. However, the non-motor findings may offer novel evidence against the practicability of utilizing the acute MPTP/p-lesioned mice for modeling the emotional aberrations found in Parkinson’s disease patients.
Parkinson’s disease (PD) is one of the most common neurodegenerative disorders [1]. Unfortunately, a definitive cure for this disease is still elusive; most currently available treatment options rely on palliative symptomatic therapy. Therefore, there is an urgent need to develop medications that prevent the onset or progression of PD. Animal models play a significant role in the preliminary stages of drug development. The model’s ability to mimic the disease phenotype is vital in establishing the validity of the disease model. In general, animal models of PD should replicate as many phenotypic characteristics of the disease as possible.
A mouse model commonly used for evaluating PD-like symptoms is the
1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-lesioned model [2, 3, 4]. Upon
uptake in the brain, MPTP is transformed to its active metabolite,
1-methyl-4-phenylpyridinium (MPP
Another reason for the variability of acute MPTP regimens in inducing PD-like non-motor symptoms may be the inconsistent use of mouse strains in different studies. Evidence suggests that mice of different genetic backgrounds exhibit different physiological [21, 22, 23] and behavioral [24, 25] phenotypes. So far, the majority of studies utilizing an acute MPTP regimen have used the inbred strain C57BL/6 owing to its high susceptibility to the MPTP toxin compared to other strains [26]. However, there is growing evidence that even for the most common substrains of C57BL/6, including BL/6J and BL/6N, there are still notable differences in their metabolic [27] and behavioral [28] phenotypes. Remarkably, compared to studies using BL/6J, studies that specifically use the BL/6N substrain for establishing acute MPTP models are lacking.
To assess and validate the ability of the acute MPTP regimen to replicate the pathogenic mechanisms of PD, we sought to establish a model by using a protocol with a few variations from that usually employed by previous reports and evaluated the practicability of using the C57BL/6N substrain in creating an acute MPTP/p-lesioned model of PD. We describe the profile of motor impairments and the timeline of their onset and eventual recovery, along with the corresponding changes in DA signaling in the nigrostriatal region. Furthermore, we characterize the non-motor behaviors observed upon recovery from motor dysfunction and their possible underlying mechanisms.
C57BL/6N male mice aged 12 weeks and weighing 25–28 g were procured from
Daihan-Biolink Co. (Chungbuk, South Korea) and allowed to acclimatize for 1 week
before their use. The animals were housed in a room maintained at 23
The acute MPTP/p protocol is similar to that described previously [29, 30, 31]. Briefly, probenecid (250 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) was injected 30 min before the first MPTP injection. Then, the MPTP/p-lesioned group received four intraperitoneal injections of MPTP (Sigma-Aldrich), at 22 mg/kg per dose, an empirically derived dosage, and administered at 2-h intervals single day for a cumulative dose of 88 mg/kg. This resulted in a 50–80% survival rate while still achieving a ~50% decrease in motor function. The control group (CON), on the other hand, received saline injections. The mice were sacrificed, and brain samples were obtained for immunohistochemical analyses at 1, 2, 8, and 16 days and western blot analyses at 16 days after the last injection of MPTP or saline. Behavioral tests (open field, rotarod, elevated plus maze [EPM], and marble burying tests; grid test and tail suspension test [TST]) were performed on different days, as described in Fig. 1. A total of 74 mice were divided among three separate experiments; in experiment 1, n = 21 per group, in experiment 2, n = 7 per group, and in experiment 3, n = 9 per group. The same set of animals was used for the battery of behavioral tasks indicated in experiments 2 and 3, as shown in Fig. 1.
Schematic diagram of the experimental procedure. On day 0, C57BL/6N mice in the control group received intraperitoneal injections of saline, and those in the MPTP/p-lesioned group received multiple injections of MPTP (22 mg/kg; four times at 2-h intervals) 30 min after a single dose of probenecid (250 mg/kg). In experiment 1, the mice were sacrificed for the sampling of brain tissues at the indicated time points to be used for IHC (n = 4/group) and WB (n = 5/group) analyses. In experiment 2, the open field and rotarod tests (for evaluating motor symptoms) along with the elevated plus-maze and marble burying tests (for evaluating non-motor symptoms) were performed at the indicated time points (n = 7/group). In experiment 3, the vertical grid (for evaluating motor symptoms) and tail suspension tests (for evaluating non-motor symptoms) were performed at the indicated time points (n = 9/group). IHC, immunohistochemistry; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; WB, western blot.
Immunohistochemical analysis of free-floating sections was performed as per
protocols described in previous studies [32, 33, 34]. After sacrificing the animals,
brain tissues were collected and fixed with 4% (w/v) paraformaldehyde in
phosphate-buffered saline (PBS). Subsequently, the brain tissues were stored in
30% (w/v) sucrose for 4 days. After that, 30-
Images of immunohistochemically stained slices were observed with a Leica microscope and captured using the Leica Caption Suite software (v4.12.0; Leica Microsystems CMS GmbH). The images were converted to grayscale to measure TH immunoreactivities, and the threshold in each tissue was comparably modified using background subtraction. For further correction, the tissue sections from the negative controls (without primary antibodies) were used to identify a single threshold for all slides. The mean gray value (256 gray levels) was calculated using ImageJ software (NIH, Bethesda, MD, USA) for each area chosen. The substantia nigra (approximately 3.64 mm caudally from the bregma) and striatal sections (approximately 0.14 mm rostrally from the bregma) were selected from each brain. The intensities were measured in the substantia nigra and striatum. The relative changes in intensity levels were expressed as relative TH optical densities (ODs) after setting the mean intensity of the control to 1. Moreover, the number of DCX-positive neuroblasts in the dentate gyrus (DG) subregion of hippocampal sections (approximately 1.94 mm caudally from the bregma) were counted as previously described [35].
Western blot analysis was performed as described previously [32]. Briefly,
samples of the medial prefrontal cortex (mPFC), hippocampus, and striatum were
separately sonicated in buffer H (50 mM
To evaluate the activity of the mice, an open field test was performed as per previous studies’ protocols [32, 36]. For each test, a mouse was first placed in the center of an open field, and various parameters, including the ambulatory move time (s) and ambulatory distance (cm), were determined over a 30-min period using a TruScan Photo Beam Activity System (Coulbourn Instruments, Whitehall, PA, USA).
According to previous reports with some modifications, the motor endurance of the mice was evaluated using a rotarod test performed according to previous reports [37]. Briefly, the rotarod apparatus (Mouse Rota-Rod; Ugo Basile, Varese, Italy) was programmed to rotate with linearly increasing speed from 5 rpm to 40 rpm in 300 s. When a mouse fell off the rod, the time (s) and speed attained (rpm) were automatically recorded; a total of three trials were performed with a 20-min interval between trials. The results were expressed as the average of the three successive trials. The mice were pre-trained for 2 days before MPTP/p treatment. During pre-training, the testing day protocol was followed except for the rotation speed, which was kept at a constant speed of 5 rpm for a period of 300 s.
A vertical grid test, which is employed to measure deficits in fine motor
skills, was performed in accordance with that described in previous work [38].
Briefly, the mice were placed 3 cm from the top of the apparatus (W
An EPM test was used to evaluate the levels of anxiety in MPTP/p-lesioned mice.
The protocol used has been described previously [39]. Briefly, a maze comprising
a central square platform (10
To further confirm the results of the EPM test performed by MPTP/p-lesioned
mice, we conducted a marble-burying test. It is a defensive behavior test for
mice, and a more significant number of buried marbles indicates higher anxiety
levels. The protocol was adapted from previous work [40]. Briefly, each mouse was
introduced to a new cage (W
To assess other non-motor symptoms, such as depression, we performed a TST. This behavior test was performed as reported previously [41]. Briefly, the mice were suspended from their tails at the height of 50 cm above the surface with adhesive tape wrapped around the tail 2 cm from the tip. Immobility—defined as the absence of all limb or body movements, except those due to breathing or gravity—was measured for 6 min.
All statistical analyses were performed using Prism (GraphPad Software, San
Diego, CA, USA). Two-way analysis of variance (ANOVA) [42] tests were used to
test for the main effects of the treatment and time and their interactions on the
immunohistochemical analysis of TH expression and the behavioral analysis of
motor symptoms. When required according to significant omnibus F-statistic, a
Šidăk’s posthoc test was used to test for multiple comparisons. The
results of the ANOVA tests are presented in Table 1, and the results of the
multiple comparison tests are presented in the corresponding figures mentioned.
Unpaired Student’s t-tests were used for all other analyses. For all
statistical tests, a P-value
Figure | Treatment | Time | Interaction | |
Immunohistochemical analysis of TH expression | ||||
Relative OD in the substantia nigra | Fig. 2A | F(1, 6) = 93.06 | F(3, 18) = 1.205 | F(3, 18) = 2.276 |
P |
P = 0.3362 | P = 0.1144 | ||
Relative OD in the striatum | Fig. 2B | F(1, 6) = 255.6 | F(3, 18) = 9.971 | F(3, 18) = 5.639 |
P |
P = 0.0004 | P = 0.0066 | ||
Behavioral analysis using the open field test | ||||
Ambulatory move time | Fig. 3A | F(1, 12) = 28.31 | F(3, 36) = 16.18 | F(3, 36) = 23.92 |
P = 0.0002 | P |
P | ||
Ambulatory distance | Fig. 3A | F(1, 12) = 29.52 | F(3, 36) = 8.111 | F(3, 36) = 7.551 |
P = 0.0002 | P = 0.0003 | P = 0.0005 | ||
Behavioral analysis using the rotarod test | ||||
Rotation speed attained | Fig. 3B | F(1, 12) = 13.02 | F(3, 36) = 10.76 | F(3, 36) = 4.966 |
P = 0.0036 | P |
P = 0.0055 | ||
Latency to fall | Fig. 3B | F(1, 12) = 11.57 | F(3, 36) = 16.92 | F(3, 36) = 5.942 |
P = 0.0053 | P |
P = 0.0021 | ||
Behavioral analysis using the vertical grid test | ||||
Total time | Fig. 3C | F(1, 8) = 5.366 | F(1, 8) = 0.00335 | F(1, 8) = 2.335 |
P = 0.0492 | P = 0.9552 | P = 0.1650 | ||
Time to climb down | Fig. 3C | F(1, 8) = 2.090 | F(1, 8) = 0.09613 | F(1, 8) = 0.4998 |
P = 0.1863 | P = 0.7644 | P = 0.4997 | ||
Time to turn | Fig. 3C | F(1, 8) = 4.306 | F(1, 8) = 0.04870 | F(1, 8) = 3.441 |
P = 0.0717 | P = 0.8309 | P = 0.1007 | ||
Successful hindlimb steps | Fig. 3C | F(1, 8) = 15.87 | F(1, 8) = 0.7146 | F(1, 8) = 1.399 |
P = 0.0040 | P = 0.4225 | P = 0.2709 | ||
ANOVA, analysis of variance; MPTP/p, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine + probenecid; OD, optical density; TH, tyrosine hydroxylase. |
There is consensus that the change in DA signaling in the nigrostriatal region
is the most common outcome in PD models. Therefore, to confirm successful MPTP/p
lesioning, DA loss was evaluated using immunohistochemistry for TH expression in
both the substantia nigra and striatum at different time points (n = 4 mice/group
at 1, 2, 8, and 16 days after acute MPTP/p treatment). The semi-quantitative
results revealed that TH intensities (relative OD) in both the substantia nigra
(Fig. 2A) and striatum (Fig. 2B) in the MPTP/p-lesioned group were significantly
lower than those in the CON group at all time points (all P
Acute MPTP/p treatment persistently decreased TH
immunoreactivity in the nigrostriatal dopaminergic pathway. (Left panels in A
and B) Representative photomicrographs of TH protein expression in the substantia
nigra (A) and striatum (B). Scale bars represent 200
In MPTP/p-lesioned mice, DA loss in the nigrostriatal pathway often correlates
with motor dysfunctions [43, 44]. Motor functions in the mice were assessed using
the open field, rotarod, and vertical grid tests (Fig. 3). In the open field test
(n = 7 mice/group), the ambulatory movement time was significantly decreased in
MPTP/p-lesioned mice at 1 day (mean
Motor endurance was evaluated using the rotarod test simultaneously as those of
the open field test (n = 7 mice/group). Rotation speeds attained (Fig. 3B, left
violin plot) in the MPTP/p-lesioned group were lower than those in the CON group
at 1 day (CON: 31.38
We performed the vertical grid test at 3 and 7 days after acute MPTP/p treatment
(n = 9 mice/group). The total time (Fig. 3C, left violin plot; CON: 27.89
C57BL/6N mice spontaneously recovered from motor symptoms
induced by acute MPTP/p treatment. Motor symptoms were evaluated using open
field (A), rotarod (B), and vertical grid (C) tests. Upper and lower dashed lines
signify the upper and lower quartiles, respectively, and the median is
represented by a solid black line within a violin plot. The data are reported
from two separate experiments for the open field and rotarod tests (n = 7
mice/group) and the vertical grid test (n = 9 mice/group). *P
To investigate the occurrence of non-motor symptoms after recovery from motor
symptoms, anxiety-like behaviors were evaluated using the EPM and marble burying
tests on days 12 and 14 after MPTP treatment, respectively. In the EPM test,
MPTP/p-lesioned mice (n = 7) spent more time in the open arm areas compared to
that spent by mice of the CON group (open arms, CON: 17.30
MPTP/p-lesioned C57BL/6N mice showed anxiolytic behaviors after
recovery from motor symptoms. Anxiolytic behaviors were evaluated using
EPM (A) and marble burying (B) tests. Upper and lower dashed lines signify the
upper and lower quartiles, respectively, and the median is represented by a solid
black line within a violin plot. The data are reported from two separate
experiments for the EPM and marble burying tests (n = 7 mice/group). *P
As depression is another major non-motor feature of PD that often accompanies
anxiety, depression-like behavior was evaluated using the TST on day 15 after
MPTP/p treatment. MPTP/p-lesioned mice (n = 9 mice/group) presented significantly
lower immobility time than that seen with CON mice (CON: 151.8
MPTP/p-lesioned C57BL/6N mice showed antidepressive behaviors
after recovery from motor symptoms. Antidepressive behaviors were evaluated
using the TST. Upper and lower dashed lines signify the upper and lower
quartiles, respectively, and the median is represented by a solid black line
within a violin plot. The data are reported from two separate TST experiments (n
= 9 mice/group). *P
To identify the possible mechanism underlying the anxiolytic and antidepressive
behaviors and the recovery of motor functions in acute MPTP/p-lesioned mice, NET
expression levels in the mPFC, hippocampus, and striatum at 16 days
post-treatment (n = 5) were evaluated by western blot analysis (Fig. 6). The NET
expression levels (relative OD) in the mPFC (CON: 1.0
Acute MPTP/p treatment significantly increased the expression
levels of NET in the mPFC, hippocampus, and striatum. (Left panels in A–C)
Representative immunoblot images of NET expression in the mPFC (A), hippocampus
(B), and striatum (C) at 16 days post-treatment. (Right panels in A–C) Violin
plots showing semi-quantitative data analyses (relative OD, n = 5 mice/group).
Upper and lower dashed lines signify the upper and lower quartiles, respectively,
and the median is represented by a solid black line within a violin plot. The
relative NET expression levels were calculated after normalization to that of the
Immunohistochemical analyses showed considerable changes in the number of
neuroblasts in the hippocampus of acute MPTP/p-lesioned mice at 16 days
post-treatment. The number of DCX-positive neuroblasts in the hippocampal DG of
MPTP/p-lesioned mice was significantly higher than in the control mice (CON: 86.0
Increased rate of adult neurogenesis in the hippocampus of mice
treated with acute MPTP/p protocol. (Left panel) Representative images
of immunohistochemical stainings for DCX, a marker for neuroblasts, in the
hippocampal DG at 16 days post-treatment. The scale bar represents 50
We assessed the practicability of using C57BL/6N mice and adjunct probenecid treatment with an acute MPTP regimen for establishing PD models. We found that our treatment protocol induced DA neurodegeneration and elicited both motor and non-motor behavioral alterations. As predicted, the motor impairment was pronounced and transient, but unexpected non-motor behaviors were observed upon recovery from the motor impairments.
The majority of previous studies have used the ability of the MPTP toxin to cause a depletion of DA levels in the nigrostriatal region to replicate motor function impairments seen in patients with PD [45]. We found a substantial decrease in motor activity directly after MPTP/p injection, owing to an acute decline of about 50% in nigrostriatal DA signaling within 24 h after MPTP/p lesioning. The lesioned mice recovered spontaneously starting from 4 days post-treatment and attained full recovery by 8 days post-injection. The observed motor dysfunctions agree with previous studies that employed the acute MPTP mouse model [46, 47]. Interestingly, this spontaneous recovery of motor function was not accompanied by a significant DA rescue in the substantia nigra. Instead, the TH-positive DA signaling remained depleted in the lesioned animals. Thus, it is highly likely that other compensatory mechanisms had developed for the nigrostriatal and/or non-dopaminergic neurotransmission [48, 49]. We observed the following: (1) a small but gradually significant recovery of DA signaling in the striatum and (2) a significant increase in NET expression levels in the striatum of acute MPTP/p-lesioned brains, both of which coincided with the full recovery of motor functions at 16 days post-treatment. The slight improvement in striatal DA signaling has also been reported previously by Mitsumoto et al. [50]; they found that acute MPTP-lesioned C57BL/6 mice established without probenecid treatment spontaneously regenerated their striatal axons independent of degenerative changes in TH-positive cell bodies of the substantia nigra at 3 days post-injection.
Additionally, [51] suggested that NET-mediated re-uptake of extracellular DA plays a role in regulating the recovery of DA signaling in lesioned mice. Therefore, we suggest that the eventual recovery of motor function observed may be attributed partly to the gradual improvement of DA levels in the striatum through nigral cell axonal sprouting and the regulation of aberrant DA neurotransmission increased NET-dependent re-uptake of DA in MPTP/p-lesioned mice. However, further studies are warranted to confirm these hypotheses and identify other mechanisms possibly involved in the recovery of motor functions.
Upon recovery from motor deficits, MPTP/p-lesioned mice were evaluated for non-motor symptoms to eliminate possible artifacts due to locomotor deficits [52]. We evaluated anxiety-related behavior in the post-motor impairment period of acute MPTP/p treatment using the EPM and marble burying tests. Interestingly, the current model showed anxiolytic behavior instead of the usual anxiogenic effect of acute MPTP administration in other models [10, 53, 54, 55, 56, 57]. Anxiolytic behaviors in the current model are mainly related to (1) reduced passive aversion of possible danger (as evaluated using the EPM test) and (2) decreased defensive burying in response to a distinct danger (as evaluated using the marble burying test) [10]. As anxiety almost always concomitantly occurs with depression in clinical cases of PD [58, 59], we also investigated any possible depression-like behavior in the post-motor impairment period of our acute MPTP/p model using the TST paradigm. Together with signs of anxiolytic activity, we observed antidepressive behaviors in the current model. Again, this contrasts with previous reports that evaluated depression-like symptoms in acute MPTP-treated mice established without probenecid administration [60, 61, 62, 63, 64, 65]. Although the exact mechanisms behind the non-motor behavioral features exhibited by the models remain unclear, the involvement of extranigral pathways is highly probable [66, 67]. Our current model found molecular alterations in the mPFC-hippocampal-striatal region to increase NET expression levels in the mPFC, hippocampus, and striatum after acute MPTP/p treatment. Although the reason for this increase in NET levels is unknown, its possible involvement in the altered emotional behavior of the current model cannot be ruled out. We hypothesize that since NETs are the primary target of antidepressants and psychostimulants [68, 69], any aberrations in their expression may also translate to abnormal depression-related emotional dysregulation. Moreover, as anomalies in DA neurotransmission are also correlated with anxiety [70] and depression-related behaviors [71], a possible NET-induced alteration in the DA control of emotional processes should be considered.
Additionally, we found that acute MPTP/p treatment enhanced the rate of hippocampal neurogenesis as the number of DCX-positive neuroblasts increased upon recovery from motor dysfunction. Furthermore, acute MPTP treatment is known to induce neurogenesis in the DG, striatum, and rostral subventricular zone to contribute to the functional replacement within the nigrostriatal circuitry [72, 73]. In our current model, the behavioral changes following acute MPTP/p treatment are consistent with the evidence of hippocampal neurogenesis-induced anxiolytic and antidepressive behaviors reported by previous studies [74, 75, 76, 77]. Moreover, several classes of antidepressant drugs are known to increase hippocampal neurogenesis [78, 79, 80]. Thus, we further hypothesize that hippocampal neurogenesis-related factors may be involved in an additional mechanism by which abnormal anxiolytic and antidepressive behaviors were induced in the acute MPTP/p model.
The acute MPTP/p model utilized varies from the original protocol regarding probenecid supplementation and the strain of mice used. Probenecid administration and MPTP mainly serve to increase the degree of and prolong the progression of DA damage. We found that in the current set-up, probenecid treatment did not seem to offer any advantage in eliciting motor nor non-motor symptoms of PD, as compared with previous studies using acute MPTP treatment without probenecid treatment [26, 46, 53, 54]. Nevertheless, a possible effect of an interplay between probenecid and C57BL/6N substrain, within the frame of the acute MPTP protocol, on the development of unexpected emotional abnormalities cannot be discounted. Further research is warranted to clarify other specific factors, such as age, sex, and/or protocol of behavioral tests, which may also be responsible for the behavioral aberrations detected in our research. Additionally, more specific mechanistic gain- or loss-of-function studies exploring the underlying molecular processes involved in developing aberrant behaviors will further substantiate our findings.
The supplementation of probenecid and C57BL/6N mice for an acute MPTP PD model could evoke robust nigrostriatal DA degeneration and motor symptoms. We also report the unexpected manifestation of non-motor symptoms in this acute MPTP/p model, wherein anxiolytic and antidepressive behaviors were observed. These findings open the door to in-depth investigations on the various animal models currently used in the research on PD. Our results challenge the current evidence available in support for the use of acute MPTP/p-lesioned mice for modeling emotional dysfunction seen in PD.
BSA, bovine serum albumin; CON, control; DA, dopaminergic; DAB, diaminobenzidine; DG, dentate gyrus; DCX, doublecortin; EPM, elevated plus maze; mPFC, medial prefrontal cortex; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; MPTP/p, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine + probenecid; NET, noradrenaline transporter; NGS, normal goat serum; OD, optical density; PD, Parkinson’s disease; PBS, phosphate-buffered saline; RT, room temperature; SDS, sodium dodecyl sulfate; TH, tyrosine hydroxylase; TST, tail suspension test.
CM conceived and designed the experiments; MW, MJA, PDEW-M, S-HK, J-CK, TS and CM designed the methodology; MW, MJA and PDEW-M performed the experiments; MW, MJA, PDEW-M and CM did the formal analysis; CM provided the resources; MW, MJA, PDEW-M and CM wrote the original draft; MJA, PDEW-M and CM edited and reviewed the draft; CM supervised the work, CM was responsible for funding acquisition.
The procedures and protocols followed were approved by the Institutional Animal Care and Use Committee of Chonnam National University (CNU IACUC-YB-2019-22). Animal care was in accordance with the National Institute of Health (NIH) guide for the care and use of laboratory animals (NIH Publication No. 8023, revised 1978). Care was taken to reduce the suffering of the animals throughout the experiments.
Not applicable.
This work was supported by a grant from the National Research Foundation (NRF) of Korea funded by the Korean Government (NRF-2019R1A2C1004045).
The authors declare no conflict of interest.