As described by Guo et al. (1996), third instar flies were reared on a standard medium of corn meal and molasses at 60% relative humidity in a 12 h: 12 h light/dark cycle. For whole cell patch-clamp recordings, larvae of a PTTH-Gal4 (obtained from Christian Wegener’s lab, University of Würzburg) or 103604 (NP0394-Gal4; Drosophila Genetic Resource Center (DGRC) Kyoto) line specific for PTTH-expressing neurons (Yamanaka et al., 2013a,b) were crossed with flies of the UAS-mCD8:GFP strain (BL5137 from Bloomington Drosophila Stock Centers, BDSC). Crosses of the lines PTTH-Gal4 with flies of the UAS-GCaMP6m strain and PTTH-GAL4 with flies of the UAS-Epac-camps (BL25408) strain were used for Ca²⁺ imaging and Epac-camps FRET imaging experiments in ex vivo preparations. In the Calcium imaging experiments combined with AChR RNAi, we crossed the homozygous flies (UAS-GCaMP6m; PTTH-Gal4) expressing GCaMP6m in PTTH neurons with various AChR RNAi fly stocks. The flies, PTTH-Gal4 line crossed with different neurotransmitter receptors RNAi stocks, were analyzed in the developmental experiments. The fly stocks were used in RNAi experiments as follows: The stocks from Tsinghua Drosophila stock cener: UAS-nAChRα1-RNAi (THU3068), UAS-nAChRα3-RNAi (THU2756). The stocks from Bloomington Drosophila Stock Center: UAS-mAChR-B-RNAi (67775), UAS-mAChR-C-RNAi (61306).

Whole cell patch-clamp recordings of PTTH neurons in the larval brain of Drosophila were exercised in vivo, using the following protocol. The ventral part of a third instar larva was fixed with dental glue (3 M, Vetbond) on the bottom of a recording chamber. Afterward, larval extracellular Ringer’s solution (101 mM NaCl, 3 mM KCl, 4 mM MgCl₂, 1 mM CaCl₂, 5 mM glucose, 20.7 mM NaHCO₃, 1.25 mM NaH₂PO₄, pH 7.2) was supplied, and the perineural organ was gently removed with tweezers. All electrophysiological recordings were performed using a Nikon upright microscope (NI-FLT6) equipped with a mercury lamp (OSRAM) and a GFP filter (BP 450–480). The PTTH neurons were identified by their fluorescence and morphology under a 40x objective, and glass capillaries (1–2 MΩ, pulled from the glass of the type GCF150-7.5, Harvard Apparatus, United States) were used to gently remove upper neurons and superficial glia via applying negative pressure to expose PTTH neurons without disrupting them. During all of the recordings, the recording chamber was continuously perfused with larval extracellular Ringer solution and bubbled with 95% O₂ and 5% CO₂ (2 mL/min). Whole patch-clamp recordings were carried out with borosilicate glass capillaries (B15024F, Vitalsense Scientific Instruments Co., Ltd.) of about 8∼10 MΩ using a P97 micropipette puller (Sutter Instruments, Novato, United States) and filled with internal solution (102 mM K-gluconate, 0.085 mM CaCl₂, 0.94 mM EGTA,8.5 mM HEPES, 1.7 mM MgCl₂, 17 mM NaCl, adjusted to 235 mOsm, pH 7.2). Minimal suction followed by break in into whole cell configuration using the voltage clamp mode with a holding voltage of −80 mV allowed to achieve gigaseals. Afterward, the current-clamp mode was chosen to record the spontaneous firing of the cells. An Axon-700B MultiClamp amplifier acquired and digitized signals at 25 kHz and a filter was applied at 2 kHz with 4-pole Bessel filter using a Digidata 1550 (Axon Instruments, California, United States) converter. Traces of action potentials were analyzed with Clampex 9.0 software (Molecular Devices). The maximum action potential frequencies were calculated by Spikes2.0 software.

The brain tissues of early third instar larvae were dissected in Drosophila larval extracellular Ringer’s solution. Subsequently, each dissected brain was attached to the bottom of the perfusion chamber, with the anterior side up. The bottom of the chamber was coated with polylysine (0.2 mg/ml) in advance. Larval brains were continuously superfused with Ringer solution or stimulus presentation during the entire imaging experiment. Baseline fluorescences were acquired for 2 min before applying drugs or neurotransmitters. Brains were superfused until the fluorescences were restored to the basal level after stimulus treatments. To monitor Ca²⁺- and cAMP-induced fluorescence changes of GCaMP6m and Epac-camps expressing cells, we used a ZEISS LSM 880 confocal laser scanning microscope and a 20x/1.0NA W Plan-Apochromat DIC water-immersion objective (Carl Zeiss MicroImaging GmbH). All Ca²⁺ imaging experiments were scanned with a 488-nm laser. In each larval preparation, the calcium level of one PTTH neuron was analyzed. The cell bodies were defined for the region of interest (ROI), in which the GCaMP6m fluorescence intensities were measured. For individual PTTH neurons, calcium signal per frame was calculated as an average of all pixels within its ROI. This calculation was repeated for each frame in the time series to generate a single-neuron time course. The fluorescent signal was expressed as the percent change relative to the prestimulus baseline F₀. The baseline fluorescence F₀ was defined as average fluorescence during the 1-min period right before the beginning of stimulus application. The fluorescent signal for each PTTH neuron was calculated as the change of fluorescence intensity (F–F₀ = ΔF) divided by the baseline fluorescence (ΔF/F₀).

To characterize the calcium responses to ACh, we constructed dose-response curves based upon the normalized amplitude of the response. The response amplitude was defined as the absolute value from baseline to the peak of the stimulation-induced Ca²⁺ levels. The amplitude increased with the increasing concentration of neurotransmitters. Each recorded PTTH neuron reached a similar saturated response amplitude with 500 μM neurotransmitter application. Desensitization became apparent at higher concentrations (1 mM). Therefore, response amplitude to 500 μM ACh was used as reference (=100%) to calculate the dose-response relationship. All response amplitudes were normalized to this value. The normalized values were fitted to a 4-parameter dose-response curve with variable slope and plotted on logarithmic scales using SigmaPlot. Half-maximal effective concentration (EC50) was calculated from dose-response curves.

Epac-camps FRET imaging was performed by scanning frames with a 458-nm laser and a frequency of 0.83 Hz. The CFP and YFP emission (460–520 nm/520–580 nm) were separated by a Carl Zeiss confocal splitter. Background subtraction was performed for CFP and YFP intensities. The CFP/YFP ratio was calculated for each time point.

To analyze the response of PTTH neurons to potential neurotransmitters, different doses of neurotransmitters ACh (100 nM–1 mM), OA (10 μM–1 mM) were applied for 60 s perfusion. In corresponding vehicle control groups, the application of neurotransmitters was replaced by Ringer’s solution. Next, we used the sodium channel blocker tetrodotoxin (TTX) (Taizhoukangte, 1 μM) to block synaptic transmission from other neurons. Brain tissues were incubated with TTX for 5 min at least. Afterward, TTX (1 μM) with the respective neurotransmitter was perfused simultaneously into the brain sample. We applied different specific AChR antagonists to further identify their receptor types. The brains were incubated with 100 μM scopolamine (Scop, the blocker of metabotropic acetylcholine receptors) (Gorczyca et al., 1991) or 100 μM mecamylamine (MCA, the blocker of nicotinic acetylcholine receptors) (Huang et al., 2010) before being treated with ACh (1 mM). To analyze the OA receptors, the neurons were incubated with 300 μM Ca²⁺ channel antagonist cadmium chloride (CdCl₂) for 5 min before being treated with OA (50 μM). According to the dose-response dependent curve, we perfused a low concentration of OA (100 μM) to the brain and recorded the change of cAMP levels in PTTH neurons with Ringer solution as a control group. All drugs were purchased from Sigma Aldrich.

To confirm the RNAi efficacy, RT-PCR was performed following the protocol described in a previous study (Li et al., 2014). The RNAi lines were crossed with the elav-Gal4 line (BL458 from BDSC). Actin was used as the internal control. The expression level was calculated using AlphaEaseFC. The primers used in RT-PCR were as follows: primer of nAChRα1 up: 5′-G CTGGCTGGTGGAATCTA-3′; primer of nAChRα1 low: 5′-AG CACTGCCGATGACACT-3′; primer of nAChRα3 up: 5′-GC GGGCGTCTGTGATATT-3′; primer of nAChRα3 low: 5′-TGTG CTCTCCTCTTCCTTG-3′; primer of mAChR-B up: 5′-ATCTC TGGCTTTCCGTCG-3′; primer of mAChR-B low: 5′-TGAA GAAGAGCAGAGCGG-3′; primer of mAChR-C up: 5′-GAA GAGCGTCCAGATTGT-3′; primer of mAChR-C low: 5′-CGG ATTGAGAGCAGAGTTG-3′; primer of actin up: 5′-CAGG CGGTGCTTTCTCTCTA-3′; primer of actin low: 5′-AGCTGT AACCGCGCTCAGTA-3′.

All developmental experiments were performed at 25°C in 12 h light/dark cycle conditions. To maximize the fecundity, 150 female flies were crossed with 50 male flies for 3 days. Before actual collections, adults were allowed to lay for 1 h to remove held eggs. The flies were transferred to standard cornmeal food plus yeast paste for 1.5− to 2-h egg laying. The collected embryos laid on cornmeal food were incubated at 25°C for 110 h.

The times of pupariation were scored every 3 h from 8:00 to 23:00 at 110 h after egg laying. Data were compiled and ordered by progressive pupariation time and cumulative percentage pupariation in Microsoft Excel and subsequently analyzed in GraphPad Prism.

Pupae were aligned in a petri dish with the dorsal upward and photographed under a microscope after most of the pupae turned black, at which point the gender could be distinguished according to the eye color or by dissecting the pupae. In each group, the volumes of ten male and female pupae were measured with ImageJ. The pupal length was determined along the medial line between anterior and posterior, not including the anterior and posterior spiracles. Pupal width was measured along the axial line. Pupal volume was calculated by using the formula for a prolate spheroid: (π/6) W² × L, where W was pupal width and L was the pupal length (Shimell et al., 2018; Deveci et al., 2019).

Data were given as means ± standard errors of the mean. Statistical analysis was performed with GraphPad Prism 6.0. The normality of all data groups was examined with the D’Agostino and Pearson omnibus normality test. For statistical analysis of RT-PCR data (Supplementary Figure 1), t-test was used to compare the mRNA expression in RNAi knockdown experimental groups with the corresponding control groups. For statistical analysis of developmental data (Figures 3A1,A2), one-way ANOVA was applied to analyze the mean difference among three normal distributed groups of data. Tukey test method was performed for the correction of multiple comparisons. Kruskal–Wallis test was applied to analyze the mean difference among three groups of data which include at least one non-normal distributed group. In calcium/cAMP imaging experiments, n indicated the number of tested brains. To identify the AChR (Figure 2D), the amplitudes of Ca²⁺ transients in the same cells before and after blocker treatment were compared by use of the paired t-test. For statistical analysis of calcium imaging data in Figure 4C, an unpaired t test was used to compare the difference between two normally distributed data groups. Non-normal distributed data groups were statistically analyzed by Wilcoxon signed rank test. The significance levels of the statistical tests were presented as *p < 0.05, ^**p < 0.01, ^*** p < 0.001; ns, not significant p > 0.05.

To examine the intrinsic physiological properties of PTTH neurons in Drosophila larvae, we used electrophysiological recordings and Ca²⁺ imaging. In vivo whole-cell patch-clamp recordings revealed that GFP-labeled PTTH cells exhibited spontaneous activity and fired APs in a tonic pattern (Figure 1A, n = 15). The firing frequency was 1.26 ± 0.14 Hz. To study the possible regulatory mechanisms of neurotransmitter signaling, we applied ACh, OA, TA, HA, GABA, GLU, 5-HT, and DA to GCaMP6m-expressing PTTH cells in ex vivo preparations and examined the responses using Ca²⁺ imaging. To ensure that the neurotransmitters directly modulated neuronal activity, we disrupted AP-dependent synaptic transmission with TTX. After TTX incubation, the co-application of 1 μM TTX and 1 mM ACh increased Ca²⁺ levels in PTTH neurons (n = 10). In contrast, the simultaneous treatment with 1 μM TTX and 1 mM of other tested neurotransmitters (OA, TA, HA, GABA, GLU, 5-HT, and DA) decreased Ca²⁺ concentrations in PTTH neurons (Figures 1B–I, n_OA = 5; n_TA = 4; n_HA = 8; n_GABA = 8; n_GLU = 7; n_5–HT = 6; n_DA = 7).

ACh treatment increased the intracellular Ca²⁺ levels of PTTH neurons in a dose-dependent manner (Figure 2A, n = 10). To calculate the dose dependency of the ACh responses, we measured the amplitude from the control level to the maximum intracellular Ca²⁺ increase. The threshold concentration of the dose-dependent ACh effect varied between 100 nM and 1 μM. The magnitude of the peak response was dose-dependent between 1 μM and 1 mM (Figures 2A,B; EC₅₀ = 22.65 μM). Desensitization of ACh-sensitive PTTH neurons was observed at higher concentrations (1 mM, Figures 2A,B). To determine whether ACh responses were mediated by ionotropic nAChRs or metabotropic mAChRs, the mAChR antagonist scopolamine (Scop) or the nAChR antagonist mecamylamine (MCA) was applied. Pre-incubation with Scop (100 μM) did not block ACh responses in any of the cells tested (n = 10). In contrast, MCA (100 μM) blocked the effects of ACh in all of the tested PTTH neurons (Figure 2C; n = 10). The statistic also demonstrated that the maximum ACh response almost got to zero after incubating 100 μM MCA, which was significantly lower than that response without MCA (Figure 2D, n = 8). However, there was no difference in maximum ACh response between before and after the incubation of 100 μM Scop (Figure 2D, n = 4). Thus, all ACh responses of PTTH cells were conveyed through nAChRs, and not through mAChRs.

The findings of our pharmacological experiments suggest that nAChRs, but not mAChRs, are expressed on PTTH neurons. To evaluate whether AChRs on PTTH neurons affect the larval developmental process by altering neural activity, we analyzed the two most common parameters in Drosophila metamorphosis—pupal volume and pupariation timing, which reflect the larval holistic developmental state—because PTTH neurons are known to promote larval maturation (Kawakami et al., 1990; Tomioka et al., 1995; McBrayer et al., 2007; Rewitz et al., 2009a,b). In the present study, we downregulated the expression of different kinds of AChR subunits in PTTH neurons by crossing a PTTH-Gal4 line with corresponding RNAi lines of receptor subunits. RT-PCR analysis showed that the mRNA expression of nAChRα1, nAChRα3, mAChR-B, and mAChR-C was significantly reduced in Drosophila brains with RNAi knockdown in comparison with their corresponding control groups (Supplementary Figure 1). When the expression of nAChR α1 and α3 subunits, which are homologs of vertebrate nAChR α2 subunits, were specifically decreased in PTTH neurons, pupal size was significantly increased compared with control groups in both male and female flies (Figures 3A1,A2). In contrast, the time curve of pupariation percentage was unaffected (Figure 3B). In addition to the nAChR subunits, we performed developmental experiments in metabotropic mAChR subtypes, including mAChR-B and mAChR-C. Larvae with decreased expression of these two mAChR subtypes on PTTH neurons had no significant changes in pupal size or pupariation timing (Figures 3A1,A2,B). Together, these results suggest that nAChRs in PTTH cells participate in the regulation of larval metamorphosis, and especially growth rate, but that mAChRs do not.

To further demonstrate that ACh functions by directly binding to nAChRs in PTTH cells, we applied 500 μM ACh to PTTH cells that had RNAi-induced decreased expression of different AChR subunits or subtypes in Ca²⁺ imaging experiments. In these experiments, we used the PTTH-Gal4 driver lines combined with UAS-GCaMP6m and UAS-AChR-RNAi to detect the Ca²⁺ response of PTTH neurons to ACh. When the expression of nAChR α1 and α3 subunits in PTTH neurons was decreased, intracellular Ca²⁺ levels were not affected by the perfusion of 500 μM ACh. To test the viability of the tested brains, Ca²⁺ responses to Ringer’s solution containing 20 mM KCl were examined. This solution depolarized intact cell membranes and promoted a voltage-dependent Ca²⁺ influx. The Ca²⁺ levels increased markedly with perfusion of 20 mM KCl. This finding indicates that PTTH neurons without nAChR α1 or α3 subunits are unable to produce intact and functional nAChRs, which contributes to a silent response to ACh (Figures 4A,B). In contrast, PTTH neurons with decreased mAChR-B and mAChR-C still had elevated intracellular Ca²⁺ levels when 500 μM ACh was applied, as did the w¹¹¹⁸ control group (Figures 4D–F). Moreover, the increased reaction induced by ACh was completely blocked by 100 μM of the nAChR antagonist MCA, similar to what was observed in the pharmacological experiments (Figure 2C). Statistical analysis revealed that the Ca²⁺ response to ACh application in the nAChR α1 and α3 RNAi groups was significantly lower than that in the w¹¹¹⁸ control group. In contrast, there were no significant differences between the mAChR-B/C RNAi groups and the w¹¹¹⁸ control group (Figure 4C). These results were consistent with our aforementioned developmental results (Figures 3A1,A2). Together, these data indicate that ACh regulates the activity of PTTH neurons through nAChRs, but not mAChRs, to affect pupal volume in Drosophila.

Depending on the OA concentrations that were added, the PTTH neurons had different responses. Higher concentrations of OA treatment (200 μM–1 mM) led to decreases in Ca²⁺. In contrast, the application of lower concentrations, in more common physiological ranges (10–100 μM), increased the intracellular Ca²⁺ levels of PTTH neurons (Figure 5A, n = 7). There are three classes of OA receptors in D. melanogaster: (1) OCTα-R, which is mainly associated with calcium release from the endoplasmic reticulum (ER) and an increase in intracellular Ca²⁺ level; (2) OCTβ-R, which is mainly coupled with an increase with intracellular cAMP level; (3) Octopaminergic/tyraminergic (OCT/TYR-R), which could be activated by both OCT and TYR, and associated with a decrease in cAMP level (Evans and Maqueira, 2005; Farooqui, 2012). To distinguish which receptor type was responsible for this Ca²⁺ increase, we next examined whether PTTH responses were mediated via the activation of adenylyl cyclase (AC), or by the release of Ca²⁺ from the ER. The cAMP imaging experiments demonstrated that cAMP levels increased after the bath application of 100 μM OA, whereas the control group (with the application of the bathing vehicle) had no changes (Figure 5B, n = 5). Adding the voltage-dependent Ca²⁺ channel antagonist CdCl₂ (300 μM) prevented the OA-induced Ca²⁺ increases (Figure 5C, n = 6). This result indicates that OA-dependent AC/cAMP signaling may increase the opening probability of voltage-dependent Ca²⁺ channels and induce extracellular calcium influx, but not calcium release from ER. Thus, low concentrations of OA treatment are likely to activate PTTH neurons via Octβ-type receptors.

Our neurotransmitter screening revealed that ACh treatment led to dose-dependent Ca²⁺ increases via nAChRs in all tested PTTH neurons in the Drosophila larval brain. Hyperactivation/depolarization has been reported to play a critical role in PTTH release. For example, several studies have implied that cholinergic neurons can form synapses with prothoracicotropes and promote PTTH release (Lester and Gilbert, 1986, 1987; Agui, 1989; Shirai et al., 1994; Gilbert, 2009; Gong et al., 2010; Qin et al., 2019). Reports of M. sexta and B. mori brains have noted that high K⁺ concentrations together with Ca²⁺ cause depolarization of the treated neurons and leads to PTTH release (Carrow et al., 1981; Shirai et al., 1995). Furthermore, fifth instar larval Manduca brains have been demonstrated to release PTTH and increased ecdysteroid synthesis after ACh accumulation (Lester and Gilbert, 1986, 1987), suggesting a supportive role for the interaction between cholinergic neurons and PTTH. To date, evidence for PTTH release from cholinergic stimulation can also be drawn from in vitro studies. Cultured Mamestra brains release PTTH into the medium after the application of ACh and cholinergic agonists (Agui, 1989). Moreover, an in vitro study of the brains of B. mori showed similar results and highlighted the likely role of cholinergic transmission in regulating PTTH release from neurosecretory cells. However, the PTTH-producing neurosecretory cells of B. mori function via mAChRs, but not nAChRs (Aizono et al., 1997). This discrepancy with the present results may be caused by insect species variability. ACh/AChR have been reported to play an important role in circadian clock neurons (Keene et al., 2011; Keene and Sprecher, 2012), which are essential for steroid hormone production in Drosophila (Di Cara and King-Jones, 2016). Some anatomical and physiological studies regarding the role of ACh in the larval visual and circadian circuits implied clock neurons were supposed to be the most likely presynaptic input candidates for transmitting ACh to PTTH neurons. Cholinergic fifth lateral neurons in the Drosophila larval brain probably serve as potential upstream neurons that relay circadian input or light information to PTTH neurons via nAChRs (McBrayer et al., 2007; Johard et al., 2009; Keene and Sprecher, 2012).

Our data suggest that PTTH cells receive excitatory input from upstream cholinergic neurons via nAChRs to adjust the schedule of larval metamorphosis. On the one hand, the ubiquitous excitatory neurotransmitter ACh may relay a universal excitatory drive to PTTH neurons, responsible for the generally fast exchange of information flow between neurons. Because clock neurons receive light-induced input from photoreceptors on the retina and are described to form potential cholinergic synaptic contacts with PTTH neurons, we speculate that PTTH neurons might receive cholinergic inputs from clock cells, such as fifth lateral neurons (Johard et al., 2009; Keene et al., 2011; Yuan et al., 2011). This likely supports light avoidance behavior and promotes appropriate PTTH release, which is critical for larval survival before the transition to an adult fly. However, we also demonstrated that ACh/nAChR signaling is indeed involved in larval metamorphosis by directly exciting PTTH cells. When the expression of nAChR in PTTH cells was impaired, the disrupted balance between excitatory and inhibitory inputs resulted in an increased larval growth rate and pupal size but did not affect pupariation timing, which is very similar to the results of a previous study of the neuropeptide Crz in Drosophila larvae (Imura et al., 2020). Furthermore, our study also supports the idea that PTTH signaling controls ecdysteroid biosynthesis not only at peak levels but also at basal levels (Colombani et al., 2015; Moeller et al., 2017). Basal ecdysteroid negatively affects the larval growth rate without changing the pupariation timing.

To date, little is known about the relationship between OA and PTTH neurons. However, there is some indication of their interaction in relation to larval metamorphosis. One study reported that Octβ3R in the PG plays an essential role in ecdysone biosynthesis. The knockdown of Octβ3R in PG leads to arrested metamorphosis because of a lack of ecdysone (Ohhara et al., 2015). Meanwhile, PTTH and insulin-like peptide signaling are also impaired after the downregulation of Octβ3R expression in the PG. Monoaminergic autocrine signaling in the PG is considered to regulate ecdysone biogenesis by integrating and coordinating PTTH and insulin-like peptide signaling, thus allowing metamorphosis to occur only when critical body weight is attained during larval development and nutrients are sufficiently abundant (Luo et al., 2012; Shimada-Niwa and Niwa, 2014). Additionally, another study demonstrated that a cluster of OA neurons in the subesophageal zone transmits nutrient signals to the PG via the Crz–PTTH neuronal axis to modulate ecdysteroid biosynthesis and negatively control systemic body growth during the mid-L3 in Drosophila larvae (Youn et al., 2018; Imura et al., 2020). Based on the results of these studies, we speculated that OA probably acts on PTTH cells, and affects cellular activity in a direct or indirect manner. Our Ca²⁺ imaging experiments demonstrated that OA led to either Ca²⁺ increases or decreases depending on the applied dose. Low-dose OA (≤100 μM) elevated the intracellular Ca²⁺ levels of PTTH cells, whereas high-dose OA (≥200 μM) had a clear inhibitory effect on PTTH cells. Next, we confirmed that receptor activation via low doses of OA, which were much closer to physiological concentrations in vivo, was associated with increased cAMP levels in PTTH cells (Figure 5B). We, therefore, hypothesized that low doses of OA might modulate the activity of PTTH cells via OctβRs, according to the classification and characteristics of OA receptors (Farooqui, 2012). This result is consistent with the expression pattern of Octβ1R, which has a wide distribution in the larval brain and especially high expression around the pars intercerebralis, the major region to which the dendrites of PTTH neurons project (El-Kholy et al., 2015). Thus, we consider that Octβ1R in PTTH cells has a positive effect on PTTH release. For the inhibitory effects at higher concentrations, we made preliminary speculation that high doses of OA might interfere with other signaling pathways, thus inducing some unphysiological changes; one possibility is that OA at high concentrations might bind and activate other analog receptors, such as tyramine receptors. The effect of bidirectional regulation of OA on PTTH neurons remains to be examined further.

In summary, the results of our study indicate that multiple neurotransmitters and their specific receptors contribute to maintaining the normal release of PTTH during larval development and metamorphosis by directly regulating the activity of PTTH neurons, which are also able to generate spontaneous APs in Drosophila larvae. Additionally, signal transmission to PTTH neurons is modulated by potentially opposing forces from diverse neurotransmitter inputs. Among these neurotransmitters, both ACh and low-dose OA activate PTTH neurons. The excitatory effect of ACh on PTTH neurons plays a restrictive role in the larval growth rate and pupal volume by activating PTTH neurons. In contrast, all of the other neurotransmitters that were investigated showed direct or indirect inhibitory effects on the activity of PTTH neurons. We thus speculate that the dynamic state of PTTH neurons is determined by direct and/or indirect interactions with various neurotransmitters, which determine the homeostasis of PTTH release at any given time. These input signals from different neurotransmitters probably come from respective upstream neural circuits, corresponding to continuous physiological and behavioral adjustments caused by external or internal conditional changes. Thus, PTTH cells likely maintain a low active state through a disinhibitory mechanism, and the larvae keep feeding and growing until all inhibitory conditions are removed. At this point, presynaptic neurons integrate all of the positive input signals to PTTH neurons and initiate a peak of PTTH release, which then promotes larval molting and metamorphosis. This mechanism coordinates signals from the external environment and internal homeostasis and ensures the normal process of larval development by adaptively altering the timing of steroid hormone biosynthesis. However, more evidence is needed to confirm this hypothesis.