Ninjin'yoeito Targets Distinct Ca2+ Channels to Activate Ghrelin-Responsive vs. Unresponsive NPY Neurons in the Arcuate Nucleus

Appetite loss or anorexia substantially deteriorates quality of life in various diseases, and stand upstream of frailty. Neuropeptide Y (NPY) in the hypothalamic arcuate nucleus (ARC) and ghrelin released from stomach are potent inducers of appetite. We previously reported that Ninjin'yoeito, a Japanese kampo medicine comprising twelve herbs, restores food intake, and body weight in cisplatin-treated anorectic mice. Furthermore, Ninjin'yoeito increased cytosolic Ca2+ concentration ([Ca2+]i) in not only ghrelin-responsive but ghrelin-unresponsive NPY neurons in ARC. The cellular lineage/differentiation of ghrelin-unresponsive neuron is less defined but might alter along with aging and diet. This study examined the occupancy of ghrelin-unresponsive neurons among ARC NPY neurons in adult mice fed normal chow, and explored the mechanisms underlying Ninjin'yoeito-induced [Ca2+]i increases in ghrelin-unresponsive vs. ghrelin-responsive NPY neurons. Single ARC neurons were subjected to [Ca2+]i measurement and subsequent immunostaining for NPY. Ghrelin failed to increase [Ca2+]i in 42% of ARC NPY neurons. Ninjin'yoeito (10 μg/ml)-induced increases in [Ca2+]i were abolished in Ca2+ free condition in ghrelin-responsive and ghrelin-unresponsive ARC NPY neurons. Ninjin'yoeito-induced [Ca2+]i increases were inhibited by N-type Ca2+ channel blocker ω-conotoxin in the majority (17 of 20), while by L-type Ca2+ channel blocker nitrendipine in the minority (2 of 23), of ghrelin-responsive neurons. In contrast, Ninjin'yoeito-induced [Ca2+]i increases were inhibited by nitrendipine in the majority (14 of 17), while by ω-conotoxin in the minority (8 of 24), of ghrelin-unresponsive neurons. These results indicate that ghrelin-unresponsive neurons occur substantially among NPY neurons of ARC in adult mice fed normal chow. Ninjin'yoeito preferentially target N-type and L-type Ca2+ channels in the majority of ghrelin-responsive and ghrelin-unresponsive neurons, respectively, to increase [Ca2+]i. We suggest ARC N- and L-type Ca2+ channels as potential targets for activating, respectively, ghrelin-responsive, and unresponsive NPY neurons to treat anorexia.

Appetite loss or anorexia substantially deteriorates quality of life in various diseases, and stand upstream of frailty. Neuropeptide Y (NPY) in the hypothalamic arcuate nucleus (ARC) and ghrelin released from stomach are potent inducers of appetite. We previously reported that Ninjin'yoeito, a Japanese kampo medicine comprising twelve herbs, restores food intake, and body weight in cisplatin-treated anorectic mice. Furthermore, Ninjin'yoeito increased cytosolic Ca 2+ concentration ([Ca 2+ ] i ) in not only ghrelin-responsive but ghrelin-unresponsive NPY neurons in ARC. The cellular lineage/differentiation of ghrelin-unresponsive neuron is less defined but might alter along with aging and diet. This study examined the occupancy of ghrelin-unresponsive neurons among ARC NPY neurons in adult mice fed normal chow, and explored the mechanisms underlying Ninjin'yoeito-induced [Ca 2+ ] i increases in ghrelin-unresponsive vs. ghrelin-responsive NPY neurons. Single ARC neurons were subjected to [Ca 2+ ] i measurement and subsequent immunostaining for NPY. Ghrelin failed to increase [Ca 2+ ] i in 42% of ARC NPY neurons. Ninjin'yoeito (10 µg/ml)-induced increases in [Ca 2+ ] i were abolished in Ca 2+ free condition in ghrelin-responsive and ghrelin-unresponsive ARC NPY neurons. Ninjin'yoeito-induced [Ca 2+ ] i increases were inhibited by N-type Ca 2+ channel blocker ω-conotoxin in the majority (17 of 20), while by L-type Ca 2+ channel blocker nitrendipine in the minority (2 of 23), of ghrelin-responsive neurons. In contrast, Ninjin'yoeito-induced [Ca 2+ ] i increases were inhibited by nitrendipine in the majority (14 of 17), while by ω-conotoxin in the minority (8 of 24), of ghrelin-unresponsive neurons. These results indicate that ghrelin-unresponsive neurons occur substantially among NPY neurons of ARC in adult mice fed normal chow. Ninjin'yoeito preferentially target N-type and L-type Ca 2+ channels in the majority of ghrelin-responsive and ghrelin-unresponsive neurons, respectively, to increase [Ca 2+ ] i . We suggest ARC N-and L-type Ca 2+ channels as potential targets for activating, respectively, ghrelin-responsive, and unresponsive NPY neurons to treat anorexia.

INTRODUCTION
Reduced appetite and body weight are associated with cancer, sarcopenia and frailty (1,2), and deteriorate the quality of life (QOL) (3,4). Hence, effective means to promote appetite have been awaited. Ninjin'yoeito, a Japanese traditional Kampo medicine, has been used clinically and demonstrated to be effective to treat anorexia, fatigue, anemia, cold limbs, persistent cough, mental disequilibrium, and to promote recovery from disease (5-7). Its ability to ameliorate this variety of symptoms is considered to result partly from counteraction of anorexia. Ninjin'yoeito comprises 12 crude drugs ( Table 1). Some of these crude drugs are known to pass through the blood brain barrier (BBB) and hence possibly access to the hypothalamus including the arcuate nucleus (ARC), while ghrelin, an orexigenic gut hormone, freely accesses to the ARC neurons without crossing the BBB (8). Hence, oral administration of Ninjin'yoeito could act on ARC via one or both of these routes. We previously reported that Ninjin'yoeito counteracted the anorexigenic and body weight-lowering effects of cisplatin, a chemo-therapy drug, in mice (9). In parallel, Ninjin'yoeito increased cytosolic Ca 2+ concentration ([Ca 2+ ] i ) in single neurons isolated from ARC, and the majority (79%) of the Ninjin'yoeito-responsive neurons were immunoreactive (IR) to neuropeptide Y (NPY). The neuron coexpressing NPY and agouti-related protein (AgRP) (NPY/AgRP neuron) in ARC of hypothalamus is considered as the principal neuron for initiating feeding behavior (10)(11)(12)(13). Selective activation of ARC NPY/AgRP neurons acutely and robustly triggers feeding (14,15), while their selective deletion in adult mice markedly reduces feeding (16,17), placing the ARC NPY/AgRP neuron as the necessary and adequate inducer of feeding. We reported previously that Ninjin'yoeito increased [Ca 2+ ] i in the ghrelin-responsive and ghrelin-unresponsive NPY neurons in ARC. Ghrelin, released from gut under fasted condition, stimulates feeding via activating NPY neurons (18)(19)(20). Hence, the ghrelin-responsive NPY neurons in ARC is considered to play a central role in stimulating physiological feeding. Compared to ghrelin-responsive NPY neurons, the role of the ghrelin-unresponsive NPY neurons is less defined. A possibility exists that ghrelin-responsive and ghrelin-unresponsive NPY neurons could be converted to each other depending on the diet/metabolic states and aging, including ghrelin resistance (21)(22)(23). The present study firstly investigated the occupancy (percentage) of the ghrelinunresponsive neurons among ARC NPY neurons in adult mice fed normal chow. Notably, it has been documented that ghrelin-resistance occurs with aging and likely contributes to appetite reduction and resultant BW decrease and frailty. In this line, our previous finding that Ninjin'yoeito interacts with and recruits ghrelin-unresponsive NPY neurons to [Ca 2+ ] i increases suggests its potential to restore appetite and counteract frailty. Hence, it is of relevance to elucidate the mechanisms underlying the [Ca 2+ ] i responses to Ninjin'yoeito in ghrelin-unresponsive NPY neurons. The present study aimed to elucidate the Ca 2+ channel type implicated in the [Ca 2+ ] i response to Ninjin'yoeito in ghrelin-unresponsive NPY neurons in comparison with that in ghrelin-responsive NPY neurons.
It has been reported that N-type (CaV2.2) and L-type (CaV1.3) voltage-dependent Ca 2+ channels (VDCCs) participate in depolarization-induced release of NPY in rat medium eminence-ARC preparation (24). In this report, 23 mM KClinduced NPY release was blunted by nitrendipine, a blocker of L-type Ca 2+ channels, and 45 mM KCl-induced NPY release was markedly attenuated by ω-conotoxin, a blocker of Ntype Ca 2+ channels, but not by nitrendipine. These data may indicate differential roles of N-type and L-type Ca 2+ channels depending on the magnitude of depolarization by KCl. In addition, N-type and/or L-type Ca 2+ channels in ARC neurons are involved in the [Ca 2+ ] i responses to several substances including ghrelin (25,26). Hence, the present study explored the link of Ninjin'yoeito to N-type and/or L-type VDCCs in ghrelin-responsive vs. unresponsive neurons. Single neurons were isolated from ARC of adult mice fed normal chow, subjected to Ca 2+ imaging to determine their responsiveness to ghrelin and Ninjin'yoeito, and subsequently immunostained for NPY. In the ghrelin-responsive and ghrelin-unresponsive NPY neurons, the effects of Ninjin'yoeito on [Ca 2+ ] i and their modulation by N-type and L-type Ca 2+ channel blockers were examined.

Animals
Male C57BL/6J mice aged 4-6 weeks were obtained from Japan SLC (Shizuoka, Japan) and housed for at least 1 week under conditions of controlled temperature (23 ± 1 • C), humidity (55 ± 5%) and lighting (light phase 7:30-19:30). Food and water were available ad libitum. Animal experiments were carried out after receiving approval from the Institutional Animal Experiment Committee and in accordance with the Institutional Regulation for Animal Experiments at Jichi Medical University (IACUC approval number; 17-229) and Kobe University (IACUC approval number; 30-10-06-R1).

Preparation of Single Neurons From ARC
The ARC was isolated from the brain of mice aged 5-7 weeks and single neurons were prepared as reported previously (25). Briefly, mice were anesthetized with intraperitoneal injection of urethane (ethyl carbamate; 1 g/kg, ip) or inhalation administration of isoflurane and decapitated, and their brain was removed. Brain slices containing ARC were prepared, and the whole ARC of the left and right sides was punched out. The dissected tissues were incubated in HKRB supplemented with 20 units/ml papain (Sigma Aldrich, St. Louis, MO), 0.015 mg/ml deoxyribonuclease, and 0.75 mg/ml BSA for 16 min at 36 • C in a shaking water bath, followed by gentle mechanical trituration for 5-10 min. The cell suspension was centrifuged at 100 × g for 5 min. The pellet was resuspended in HKRB and distributed onto coverslips. The cells were kept at 30 • C in moisture-saturated dishes till [Ca 2+ ] i measurements for up to 6 h.

Measurements of [Ca 2+ ] in Single ARC Neurons
At 2-10 h after cell preparation, [Ca 2+ ] i was measured by ratiometric fura-2 fluorescence imaging as previously reported (25). Briefly, following incubation with 2 µM fura-2AM (DOJINDO, Kumamoto, Japan) for 30 min at 30 • C, the cells were mounted in a chamber and superfused at 1 ml/min with HKRB containing 2.5 mM glucose at 30 • C. Data were taken from the single cells that were identified as neurons by the criteria reported previously (25); relatively large diameter (≥10 µm), clear and round cell bodies on phase-contrast microscopy. Ninjin'yoeito (10 µg/ml) and ghrelin (10 −8 M) were administered under superfusion conditions. Fluorescence ratio (F340/F380) images were produced by Aquacosmos ver. 2.5 (Hamamatsu Photonics, Shizuoka, Japan). When [Ca 2+ ] i increases took place within 10 min of superfusion with agents and their amplitudes were at least twice larger than fluctuations of baseline, they were considered responses. In all experiments, neurons from at least three separate preparations were analyzed.

Immunocytochemistry and Identification of NPY Neurons
After [Ca 2+ ] i measurements, cells were fixed with 4% paraformaldehyde, pretreated with 3% H 2 O 2 for 1 h, and blocked in 10% normal goat serum and in 0.1 M PBS for 1 h at room temperature. Cells were incubated overnight at 4 • C with primary antiserum to NPY (DiaSorin, Stillwater, MN) diluted 1:10,000 in PBS containing 1.5% normal goat serum. After rinsing, cells were incubated with biotinylated secondary antibody raised against rabbit IgG (Vector Laboratories Inc., Burlingame, CA; diluted 400-fold) for 1 h at room temperature. After rinsing, the sections were labeled with avidin-peroxidase complex (ABC kit, Vector) for 1 h and color-developed with 3, 3'-diaminobenzidine (DAB). [Ca 2+ ] i and immunocytochemical data were correlated to each other, based on the photographs of the single neurons subjected to [Ca 2+ ] i measurements in the microscopic field (25).

Effect of Ghrelin on [Ca 2+ ] i in ARC NPY Neurons
Single neurons isolated from ARC were superfused with HKRB containing 2.5 mM glucose for [Ca 2+ ] i imaging. After [Ca 2+ ] i was stabilized at baseline, administration of ghrelin (10 −8 M) increased [Ca 2+ ] i in some single neurons that were subsequently shown to be immunoreactive (IR) to NPY by immunocytochemistry ( Figure 1A). In contrast, administration of ghrelin failed to increase [Ca 2+ ] i in other single neurons that were IR to NPY (Figure 1B). Among 43 ARC NPY neurons, 25 (58%) neurons responded to ghrelin and 18 (42%) neurons did not respond to ghrelin ( Figure 1C). These results indicated that ghrelin-unresponsive neurons occur substantially among NPY neurons of ARC in adult mice fed normal chow.  (Figure 2A). In nine ghrelinresponsive neurons, none responded to Ninjin'yoeito in Ca 2+free HKRB (Figure 2A, Right). This result indicated that the [Ca 2+ ] i response to Ninjin'yoeito in ghrelin-responsive neurons requires the presence of extracellular Ca 2+ .
We next examined whether particular type of VDCCs could be implicated in the [Ca 2+ ] i response to Ninjin'yoeito   contrast, in the presence of nitrendipine (2 µM) Ninjin'yoeito increased [Ca 2+ ] i in the majority of single neurons (Figure 2C), and this response occurred in 21 of 23 (91.3%) of ghrelinresponsive neurons (Figure 2C, Right) Figure 3A). In nine ghrelin-unresponsive neurons, none responded to Ninjin'yoeito in Ca 2+ -free HKRB (Figure 3A (Figure 3C, Right). These results indicated that Ninjin'yoeito increases [Ca 2+ ] i in ghrelinunresponsive neurons via Ca 2+ influx to which L-type Ca 2+ channel has greater contribution than N-type Ca 2+ channel.

DISCUSSION
The present study employed [Ca 2+ ] i measurement in ARC single neurons combined with immunocytochemistry and found that among 43 ARC NPY neurons, 25 (58%) neurons responded to ghrelin and 18 (42%) neurons did not respond to ghrelin. This incidence of ghrelin-responsive neurons (58%) is not far from that in previous reports, considering different experimental conditions used among studies. The extracellular single unit recordings from in vitro slices indicated that ghrelin excited 73% of neurons in the ventromedial ARC, where NPY neurons are dominant, in adult rats (27). Ex vivo whole-cell patchclamp recordings showed that ghrelin depolarized 40% of GFPlabeled arcuate NPY neurons in brain slices from 8-12 weekold male NPY-humanized Renilla reniformis green fluorescent protein transgenic mice (28). Ghrelin increased [Ca 2+ ] i in 59% of single ARC NPY neurons (9) and in 21-41% of single ARC neurons (12,25) in 5-7 week-old male mice. The present study demonstrated that ghrelin-unresponsive neurons occur substantially among NPY neurons of ARC in 5-7 week-old male mice fed normal chow. Ninjin'yoeito-induced [Ca 2+ ] i increase was blunted in Ca 2+ -free condition in both ghrelinresponsive and ghrelin-unresponsive neurons. Ninjin'yoeitoinduced [Ca 2+ ] i increases were inhibited by N-type Ca 2+ channel blocker ω-conotoxin in the majority (17 of 20), while by L-type Ca 2+ channel blocker nitrendipine in the minority (2 of 23), of ghrelin-responsive neurons. In contrast, Ninjin'yoeitoinduced [Ca 2+ ] i increases were inhibited by nitrendipine in the majority (14 of 17), while by ω-conotoxin in the minority (8 of 24), of ghrelin-unresponsive neurons. Both N-and Ltype Ca 2+ channels are reportedly expressed and functioning in ARC (24,25,29). Our results demonstrate that Ninjin'yoeito increases [Ca 2+ ] i via Ca 2+ influx, to which N-type Ca 2+ channels have greater contribution in ghrelin-responsive neurons and L-type Ca 2+ channels have greater contribution in ghrelinunresponsive neurons. The [Ca 2+ ] i increase often results from membrane excitation and/or stimulated signal transduction and results in exocytosis, transport, gene expression, and/or protein regulation. Thus, the [Ca 2+ ] i increase, in general, reflects the neuronal activation. The present findings place N-type and Ltype Ca 2+ channels in ARC as potential molecular targets for Ninjin'yoeito to preferentially activate ghrelin-responsive and ghrelin-unresponsive NPY neurons, respectively, in ARC.
The mechanisms underlying the link of Ninjin'yoeito to distinct VDCCs in ghrelin-responsive and ghrelin-unresponsive neurons remain to be elucidated. It has been documented that ghrelin and/or GHSR influence the activities of VDCCs (30)(31)(32)(33)(34)(35)(36). Since ghrelin increases [Ca 2+ ] i via Ca 2+ influx primarily through N-type Ca 2+ channels in NPY neurons (25), it is speculated that Ninjin'yoeito could interact with the GHSR and/or downstream signaling linked to N-type Ca 2+ channels in ghrelin-responsive NPY neurons. In consistent with this, it was previously reported that GHSR coexpression with dopamine type 2 receptor (D2R) FIGURE 3 | Ninjin'yoeito-induced [Ca 2+ ] i increases were inhibited in Ca 2+ free condition and preferentially by L-type Ca 2+ channel blocker in the majority of ghrelin-unresponsive ARC NPY neurons. ARC NPY neurons that responded to Ninjin'yoeito (10 µg/ml) but not to ghrelin ( reduces the inhibition of N-type VDCC currents by D2R activation (35). However, different results were also reported that N-type VDCC currents is inhibited by constitutive GHSR activity in mouse hippocampal cultures (34) and hypothalamic neurons (33). Though the cause for the apparent discrepancy among previous documents and our finding remains unknown, we investigated the effect of short-term (∼5 min) administration of ghrelin in the current and previous studies (25), while some of previous documents observed the effect of constitutive GHSR activation (33,34). Hence, ghrelin and GHSR signaling may have dual, acute stimulatory and chronic inhibitory, actions on Ntype Ca 2+ channel activity. Compared to ghrelin-responsive NPY neurons, properties of ghrelin-unresponsive ARC NPY neurons are less characterized. However, it was reported that a ligand for the taste receptor T1R2/T1R3 increased [Ca 2+ ] i in ARC neurons, the majority of which did not respond to ghrelin (26), and that this [Ca 2+ ] i increase was inhibited by nitrendipine, but not ωconotoxin. Hence, T1R2/T1R3 is possibly linked to L-type Ca 2+ channels in ghrelin-unresponsive ARC neurons and this pathway could be involved in the action of Ninjin'yoeito. Though the Ninjin'yoeito's cellular signaling is less defined, it could interact with the excitatory signaling pathways in NPY neurons, which include the orexin-OX1R-phospholipase C pathway and low glucose-Na + ,K + -ATPase suppressiondepolarization pathway (37,38). Notably, it was shown that AMPK activator AICAR increased [Ca 2+ ] i in two types of ARC NPY neurons, one with and the other without [Ca 2+ ] i responses to ghrelin (39), and that the AICAR-induced [Ca 2+ ] i increases were blunted in Ca 2+ -free conditions (40). Thus, Ninjin'yoeito and AICAR share the common properties: they stimulate Ca 2+ influx in both ghrelin-responsive and unresponsive NPY neurons in ARC. In line with this, expression of carnitine palmitoyltransferase 1 (CPT1), a signaling molecule of AMPK, is regulated by metabolic conditions (41). Taken together, Ninjin'yoeito may elicit intracellular AMPK signaling pathway for activating ghrelin-responsive and ghrelin-unresponsive NPY neurons. However, further studies are definitely required to elucidate intracellular signaling mechanisms of Ninjin'yoeito.
The functional role of the Ninjin'yoeito-regulated VDCCs in ARC NPY neurons remains to be clarified. In ghrelinresponsive neurons, administration of Ninjin'yoeito interacts with N-type Ca 2+ channels to enhance and/or cooperate with the action of ghrelin, possibly leading to efficacious activation of ghrelin-responsive NPY neurons and consequent stimulation of appetite. Our study places ARC N-type Ca 2+ channel as a potential mediator and integrator of the actions of Ninjin'yoeito and ghrelin in ghrelin-responsive NPY neurons.
On the other hand, the ghrelin-unresponsive ARC NPY neuron has been less defined for its physiological property and cellular lineage/differentiation. It has been documented that metabolic/feeding conditions induce dynamic remodeling of NPY/AgRP neurons and differentiation of feeding related neurons in ARC (42). Hence, the ghrelin-responsive neurons and ghrelin-unresponsive neurons could be converted to each other, depending on metabolic/feeding conditions and aging. Our finding that the type of VDCC correlates with ghrelin responsiveness in NPY neurons suggests that expression of specific VDCC type may be related to remodeling of NPY neurons. In this line, the ghrelin resistance, the phenomenon that ghrelin administration cannot stimulate feeding, occurs in association with aging and diet-induced obesity (21)(22)(23). This ghrelin resistance reportedly takes place in ARC NPY neurons (21), which could result in reductions in NPY neuronal activity and appetite. The transformation of ghrelin-responsive to ghrelin-unresponsive NPY neurons may underly ghrelin resistance. Of note, we found that Ninjin'yoeito preferentially target L-type VDCC to activate ghrelin-unresponsive NPY neurons in ARC. This action of Ninjin'yoeito could serve to compensate for the ghrelin resistance and restore appetite.
The present study demonstrated that Ninjin'yoeito activates the majority of ghrelin-responsive ARC NPY neurons preferentially via N-type VDCC while the majority of ghrelinunresponsive NPY neurons preferentially via L-type VDCCs. We suggest ARC N-type VDCC as a target for activating ghrelin-responsive NPY neurons and promoting feeding while L-type VDCC as a target for activating ghrelin-unresponsive NPY neurons and possibly compensating for ghrelin-resistance to restore appetite.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
Animal experiments were carried out after receiving approval from the Institutional Animal Experiment Committee and in accordance with the Institutional Regulation for Animal Experiments at Jichi Medical University and Kobe University.

AUTHOR CONTRIBUTIONS
TY designed the study. CG, KD, and LW conducted experiments. AI and YS participated in discussion. TY and CG wrote the manuscript. TY supervised the work. All authors contributed to the article and approved the submitted version.