Abstract
Activity in the dorsal vagal complex (DVC) is essential to gastric motility regulation. We and others have previously shown that this activity is greatly influenced by local GABAergic signaling, primarily because of somatostatin (SST)-expressing GABAergic neurons. To further understand the network dynamics associated with gastric motility control in the DVC, we focused on another neuron prominently distributed in this complex, neuropeptide-Y (NPY) neurons. However, the effect of these neurons on gastric motility remains unknown. Here, we investigate the anatomic and functional characteristics of the NPY neurons in the nucleus tractus solitarius (NTS) and their interactions with SST neurons using transgenic mice of both sexes. We sought to determine whether NPY neurons influence the activity of gastric-projecting neurons, synaptically interact with SST neurons, and affect end-organ function. Our results using combined neuroanatomy and optogenetic in vitro and in vivo show that NPY neurons are part of the gastric vagal circuit as they are trans-synaptically labeled by a viral tracer from the gastric antrum, are primarily excitatory as optogenetic activation of these neurons evoke EPSCs in gastric-antrum-projecting neurons, are functionally coupled to each other and reciprocally connected to SST neurons, whose stimulation has a potent inhibitory effect on the action potential firing of the NPY neurons, and affect gastric tone and motility as reflected by their robust optogenetic response in vivo. These findings indicate that interacting NPY and SST neurons are integral to the network that controls vagal transmission to the stomach.
SIGNIFICANCE STATEMENT The brainstem neurons in the dorsal nuclear complex are essential for regulating vagus nerve activity that affects the stomach via tone and motility. Two distinct nonoverlapping populations of predominantly excitatory NPY neurons and predominantly inhibitory SST neurons form reciprocal connections with each other in the NTS and with premotor neurons in the dorsal motor nucleus of the vagus to control gastric mechanics. Light activation and inhibition of NTS NPY neurons increased and decreased gastric motility, respectively, whereas both activation and inhibition of NTS SST neurons enhanced gastric motility
Introduction
Gastric function is dependent on the complex interplay between visceral sensory input to the brain, activities in brain circuitries, autonomic outflow, and function in the enteric nervous system. In the hindbrain, the dorsal vagal complex (DVC) contains nuclei essential to the regulation of vagovagal gastric activity. These are the nucleus tract of solitarius, specifically the medial subnucleus [referred to as the nucleus tractus solitarius (NTS)] and the dorsal motor nucleus of the vagus (DMV; Gillis et al., 1989; Rogers et al., 1996; Berthoud, 2004; De Jonghe et al., 2011; Gillis et al., 2022).
In a series of studies, we found a distinct projection that controls the gastric tone and motility by the local γ-aminobutyric acid (GABA) signaling in the DVC (albeit into the NTS; Herman et al., 2009, 2010, 2012). Pharmacological in vivo studies with GABA and glutamate antagonists (Herman et al., 2009) showed the evidence for NTS-DMV-projecting neurons to be both excitatory and inhibitory (Glatzer et al., 2003; Davis et al., 2004). Blockade of the GABAA receptors in NTS is accompanied by robust decreases in gastric mechanical activity (Herman et al., 2009a). Although glutamatergic signaling is essential, its blockade in the NTS per se does not affect gastric mechanical activity (Herman et al., 2009).
This GABAergic drive was further supported by the advent of transgenic mice with GFP-expressing inhibitory neurons (GIN) mice (Oliva et al., 2000), showing its presence in the DVC as essential to the inhibition of the brainstem network dynamics system (Glatzer et al., 2007; Gao et al., 2009). Furthermore, studies using sst-Cre transgenic mice revealed the abundant presence of somatostatin (SST) neurons in the NTS (Lewin et al., 2016; Thek et al., 2019) and in the DMV (Lewin et al., 2016) that likely correspond at least in part to neurons identified in GIN mice (Oliva et al., 2000). Optogenetic activation of SST neurons results in IPSCs and robust inhibition of the action potentials of retrogradely labeled gastric antrum DMV neurons. Retrograde polysynaptic tracing revealed that SST neurons are part of the vagovagal system (Wang and Bradley, 2010; Lewin et al., 2016). Moreover, the DVC afferent vagus directly activates SST neurons in NTS (Thek et al., 2019) as well as neurons in GIN mice in this nucleus (Glatzer et al., 2007), thus, mediating the visceral sensory motility input to the hindbrain.
In addition to SST neurons, neuropeptide-Y (NPY) neurons are conspicuously present in the DVC and receive direct input from sensory afferents (van den Pol et al., 2009; Chen et al., 2020), but their role in the gastric vagal reflex circuitry is unknown. NPY neurons are involved in different physiological and homeostatic processes that include, among others, feeding (Kamiji and Inui, 2007; Zhang et al., 2019), blood pressure (Lettgen et al., 1994), and stress and depression (Heilig, 2004; Morales-Medina et al., 2010; Reichmann and Holzer, 2016).
To assess the involvement of the NPY neuron in the gastric circuitry of the DVC and, in particular, its relation to SST signaling, we took advantage of distinct transgenic mice, including npy-Cre;rosa26-tdTomato (Milstein et al., 2015) and sst-Cre;rosa26-tdTomato (Taniguchi et al., 2011) mice expressing fluorescence protein and the optogenetic effectors of inhibitory and excitatory opsins. Initially, we showed the connection between SST and NPY neurons and their neural function using Cre/Flox and Flp/Frt combinations of drivers and reporter mice. Later, we used sst- or npy-Cre;rosa26-ChR2-YFP mice to assess the impact of SST or NPY neuron activation based on their synaptic connectivity and electrophysiological characteristics on gastric activity in vivo. Our goals were to (1) characterize neurons in the dorsovagal complex of in the npy-Cre;rosa26-tdTomato, (2) profile their electrophysiological characteristics, (3) assess the synaptic action of NPY and SST neurons on DMV motoneurons and their reciprocal synaptic connectivity in the vagovagal circuitry, and (4) determine their influence on end-organ function (that is, gastric mechanical function). We found that brainstem NPY and SST neurons interact and control stomach motility and tone.
Materials and Methods
Animal models
All animals were housed in a climate-controlled animal facility (22 ± 2°C) and maintained on a 12 h light/dark cycle with ad libitum access to food and water. Animal housing rooms were maintained at MP14 barrier (pathogen and opportunistic free) in the animal facility prior to experiments.
Transgenic mice examined in this study were all The Jackson Laboratory reporter strains NPY-Cre (RRID:IMSR_JAX:027851), a gift from Boris V. Zemelman, Center for Learning and Memory, University of Texas at Austin (Milstein et al., 2015); Rosa-ChR2-EYFP reporter (RRID:IMSR_JAX:012569); Bac transgenic NPY-eGFP mouse (RRID:IMSR_JAX:006417); Sst-ires-Flp (RRID:IMSR_JAX:028579); mCherry-Frt reporter (RRID:IMSR_JAX:029040); Sst-Cre (RRID:IMSR_JAX:013044); rosa26-tdTom (RRID:IMSR_JAX:007905); ChR2-tdTomato (RRID:IMSR_JAX:012567); and ArchT-EGFP (RRID:IMSR_JAX:021188). Based on the study in NPY or Sst neuron vagal gastric circuity, the driver and reporter strains were bred individually and were used to create a colony by mating differentially with each other to develop the distinct mouse lines used in the study. Mice of either sex 1–2 months old were anesthetized with isofluorane and killed to prepare brain slice by decapitation after cardiac perfusion following the National Institutes of Health guidelines, United Kingdom regulations for the ethical use of animals in research (Drummond, 2009), and the approval of the Georgetown University Animal Care and Use Committee. SST and NPY transgenic mice examined in this study have been described previously (Milstein et al., 2015; Lewin et al., 2016).
RNA scope for NPY, SLC7A6, GAD67, and tdTomato mRNA
Mice were anesthetized with isoflurane, followed by transcardiac perfusion with ice-cold PBS solution. Brainstems were extracted and fixed overnight in 10% neutral buffered formalin and transferred to 70% EtOH, sectioned at 5 µm on a rotary microtome and stored at room temperature until use in RNAScope assay. Slides were baked at 60°C for 1 h, deparaffinized in xylene twice for 5 minutes at room temperature and dehydrated with 100% EtOH twice for 2 minutes and dried for 5′ at 60°C. The RNAScope assay proceeded according to the protocol of the manufacturer, following the guidelines for standard tissue processing of formalin-fixed paraffin-embedded tissue. Sequential brainstem sections were hybridized with one of two three-plex probe mixtures, (1) NPY (catalog #313321, ACD Bio), tdTomato (catalog #317041-C2, ACD Bio), and Gad1 (catalog #400951-C3, ACD Bio), or (2) NPY, tdTomato, and SLC17A6 (catalog #319171-C3, ACD Bio). Signal amplification was achieved through multiple amplification steps using the RNAscope multiplex fluorescence kit (catalog #323100, ACD Bio) according to the protocol of the manufacturer. The following fluorophores were used: OPAL 570, 620, 690 (catalog #FP1488001KT, FP1495001KT, FP1497001KT, Akoya Biosciences). At the end of the procedure, slides were coverslipped using ProLong Gold Antifade Mountant (Thermo Fisher Scientific) and left to dry at RT before storing at 4°C and imaging.
Immunolocalization image acquisition and analysis
Brainstem sections were obtained from ∼3 week postnatal day transgenic mice. Following anesthesia with isoflurane, mice were initially perfused transcardially with PBS (0.1 m, pH 7.3), followed by 4% buffered paraformaldehyde fixative (PFA). The brains were removed and stored overnight in 4% PFA. Free-floating coronal brainstem sections (50 µm) were obtained using a vibratome (VT1000S, Leica). Sections were blocked with 4% donkey serum in PBS for 1 h at room temperature and washed three times for 10 min each in PBS containing 0.1% Triton X-100 (Tx). They were then incubated overnight (minimum 12 h in a cold room at 4°C). They were then incubated overnight (minimum 12 h in a cold room at 4°C) with a primary anti-NPY antibody (polyclonal rabbit, 1:400; catalog #ab30914, Abcam; RRID:AB_1566510) or anti-vGlut2 (polyclonal guinea pig, 1:500; catalog #AB2251-I, Millipore; RRID:AB_2665454) that were diluted in PBS/Tx/1% BSA. Following this treatment, the brainstem sections were rewashed three times for 10 min each in PBS/Tx and then further incubated (2–4 h) at room temperature with a secondary anti-rabbit antibody conjugated to Alexa Fluor 488 (catalog #A11094, Thermo Fisher Scientific; RRID:AB_221544) or Alexa Fluor 564 (catalog #A21312, Thermo Fisher Scientific; RRID:AB_221478) at 1:500 dilution, which were constituted in PBS/Tx/BSA. At the end of the incubation, the brainstem slices were washed three times for 10 min each with PBS/Tx. All slices were mounted in Vectashield Mounting Medium (Vector Laboratories).
Imaging
Slides were scanned at 10× magnification using the Vectra 3.0 Automated Quantitative Pathology Imaging System (PerkinElmer/Akoya Biosciences). Whole slide scans were viewed with Phenochart (PerkinElmer/Akoya Biosciences), which also allows for the selection of high-powered images at 40× (resolution of 0.25 µm per pixel) for multispectral image capture. Multispectral images were unmixed using inForm Advanced Image Analysis software (inForm 2.4.6; PerkinElmer/Akoya Biosciences) and exported as component image TIFFs for analysis in QuPath 0.3.0 (Bankhead et al., 2017). Cell segmentation was done using the StarDist extension (Schmidt et al., 2018).
Confocal microscope and colocalization analysis
To obtain the DVC images shown, brain slices were transferred into the viewing chamber of a resonant scanning confocal (Thorlabs) mounted on a Nikon Eclipse F1 microscope. Acquired z-stack images (0.5 µm) using 20× or 40× immersion objective were processed with ImageJ software (National Institutes of Health) to illustrate the distribution of cell bodies for green or red fluorescent neurons. For the analysis of colocalization z-stacks of images spanning the whole 50 µm thickness of stained brainstem slices were acquired using a 20× or 40× immersion objective and projected on a 2D image. Each channel was thresholded to generate a binary image following manual clearing of the background. The two colors images were merged, converted to 32-bit, and individual cells counted. One randomly chosen slice per animal was acquired.
Anatomical tracing
To record from identified stomach projecting DVC neurons, mice were labeled by monosynaptic tracer DiI18 or injected with the polysynaptic tracers pseudorabies virus (PRV)-152EFGP in the stomach (PRV-152EGFP was a gift from Lynn Enquist, Princeton University Center for Neuroanatomy with Neurotropic Viruses, National Institutes of Health Grant P40RR018604.). The tracer was applied to the gastric antrum of transgenic mice expressing -tdTomato in a manner previously described by us (Lewin et al., 2016). To retrograde uptake of DiI in the gastric-antrum-projecting DMV neurons, animals were allowed to recover for 7–10 d after surgery, whereas PRV-injected mice were killed 2–3 d after inoculation.
Brainstem slices
Slices were prepared from male and female mice at least 4 weeks of age. Briefly, after isoflurane anesthesia, followed by transcardiac perfusion with ice-cold N-Methyl-d-glucamine (NMDG) solution (Ting et al., 2018), brains were quickly removed into oxygenated solution (95% O2 plus 5% CO2, 4°C, pH 7.4, 296 mOsm) containing the following (in mm): 93 NMDG, 93 HCl, 2.5 KCL, 1.2 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, 10 MgSO4.7H2O, and 0.5 CaCl2.2H2O. Coronal brain sections (250 µm) containing the brainstem DVC were cut in NMDG solution using a vibratome (VT1000S, Leica) and incubated at 37°C in oxygenated HEPES holding artificial CSF (ACSF pH 7.4, 296 mOsm) with the following composition (in mM): 92 NaCl, 2.5 KCl, 1.2 NaH2PO4, 30 NaHCO3, 20 HEPES, 2.5 glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, and 2 MgSO4.7H2O. Slices were variably exposed to NaCl (Ting et al., 2018). They were then allowed to be equilibrated for an additional 3 h at room temperature (21°C). Subsequently, the slices were transferred to a recording chamber (500 µl volume) attached to a microscope stage (E600-FN, Nikon). There, they were continuously perfused with oxygenated ACSF (pH 7.4, 296 mOsm) composed of the following (in mm): 121 NaCl, 2.5 KCl, 26 NaHCO3, 1.25 NaH2PO4, 2 CaCl2, 2 MgSO4, 5 HEPES, and 2.5 glucose.
Electrophysiology
Neurons were identified visually by infrared-differential interference contrast or episcopic fluorescence optics and a CMOS camera (Thorlabs). All recordings from NPY and SST neurons were in the NTS. A 60× water immersion objective was used for identifying and approaching neurons. Recordings were made with patch electrodes (5–6 MΩ; Warner Instruments) with internal pipette solution (pH 7.2, 285 mOsm) that was composed of the following (in mm): 145 K-gluconate, 5 EGTA, 5 MgCl2, 10 HEPES, 5 ATP.Na, and 0.2 GTP-Na. In studies in which IPSCs were specifically studied, 145 mm KCl was substituted for potassium gluconate in the pipette solution. Cell-attached (loose seal <40 MΩ), whole-cell voltage-clamp mode at a holding potential (Vhold) = −60 mV or −30 mV) whole-cell voltage-clamp mode at a holding potential (Vhold) = −60mV or −30mV or current clamp or current-clamp recordings were performed using a MultiClamp 700B amplifier (Molecular Devices). (Note: Action potential firing frequency was not significantly different in parallel cell-attached recordings with ACSF pipette solution.) A 5 mV hyperpolarization pulse monitored input and access resistances; series resistance was typically <10 MΩ and was not compensated. Resting membrane potentials were corrected for liquid junction potential, which for the intracellular solutions, K-gluconate was −15 mV, and for the KCl was −3 mV. Signals were low-pass filtered at 2 kHz and acquired with a Digidata 1440A Digitizer (Molecular Devices).
In vitro optogenetic control
Light delivery in coronal brainstem slices was accomplished using filter cubes of the microscope [channelrhodopsin (ChR2), λ = 450–490 nm; ArchT-EGFP λ = 510–560 nm] via white light from an X-Cite 120LED (Excelitas Technologies). Slices were excited with a maximal light intensity adjusted to prevent loss of voltage clamp (Vhold = −70 mV). The diameter of the area exposed to optogenetic control under the 60× objective was <100 µm and encompassed the whole DVC area.
Drugs
Stock solutions of the following drugs were prepared in water: D(−)−2-Amino-5-phosphonopentanoic acid (catalog #3693, Tocris Bioscience), bicuculline methobromide (BMR; catalog #HB0894, Hello Bio), 2,3-Dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide disodium salt (NBQX, catalog #ab120046, Abcam), tetrodotoxin (TTX, catalog #HB1035, Hello Bio), and 4-Aminopyridine (catalog #A-0152, Sigma-Aldrich). Drug-containing stock solutions were diluted to desired concentrations in ACSF. All drugs were applied via the Y-tube application adapted to brain slices (Murase et al., 1989; Hevers and Lüddens, 2002).
In vitro data analysis
Electrophysiological data were analyzed off-line using pClamp 11 (Molecular Devices). Data were acquired from neurons that had a stable baseline membrane potential. Threshold, rheobase, and rate and characteristic of action potentials were assessed from a baseline membrane potential of −60 mV by injecting current. To analyze postsynaptic currents, semiautomated threshold pClamp 11 software was used with 5–10 times the baseline noise depending on the voltage clamp (Vhold = −60 mV or −30 mV) and the internal pipette solution (K-gluconate or KCl). The control and treatment data were acquired from two consecutive 500 ms segments. For studies where optogenetic activation was used to excite NPY-Cre;ChR2 or Sst-Cre;ChR2 neurons while recording from the NPY or Sst neuron or gastric-antrum-projecting DMV neurons, the voltage-clamp was set at −30 mV. This procedure allowed us to separate IPSCs and EPSCs, which were displayed as upward deflections (outward currents) and downward deflections (inward currents). Both IPSCs and EPSCs were measured together as there was minimal, if any, overlap between the two types of currents. Moreover, it enabled us to clearly visualize a one-to-one correlation between the onset of the optogenetic activation and its effect on the identity of the postsynaptic current (i.e., IPSC or EPSC).
Gastric physiology
Experiments were performed on NPY-Cre;ChR2 and Sst-Cre;ChR2/or Cre;ArchT-EGFP mice.
Surgical preparation
Before all experiments, food was withheld for 4 h, whereas water was provided ad libitum. Animals were anesthetized with an intraperitoneal injection containing a mixture of urethane (1000 mg/kg) and α-chloralose (60 mg/kg) dissolved in 0.9% saline. Body temperature was maintained at 37 ± 1°C with an infrared heat lamp.
After the depth of anesthesia was confirmed by lack of pedal and corneal reflexes, mice were intubated via the trachea following a tracheotomy to maintain an open airway and to institute artificial respiration when necessary. Both cervical vagi were carefully isolated from each carotid artery on either side and looped with a 5–0 silk thread for later avulsion during the experiment.
Subsequent, to the cervical vagi loop, a laparotomy was performed to expose the stomach. An intragastric balloon (made from the tip of a latex condom ∼0.3 cm long) was inserted into the stomach via the fundus and positioned in the distal region of the antrum, secured in place by a purse-string suture. The balloon was inflated (with warm water ∼100–150 µl) to produce a baseline pressure of 3–6 mmHg, which was measured by a pressure transducer (sensitivity 5 µV/V/mmHg) that was connected to it as described by (Richardson et al., 2013) for monitoring blood pressure in rats. At the end of the procedure, the abdominal cavity was closed using a gut suture.
In vivo optogenetic
To gain access to the dorsal medulla, mice were positioned in a stereotaxic apparatus (David Kopf Instruments). A limited dorsal craniotomy was performed to expose the medulla, and the underlying dura and pia were cut and reflected. The caudal tip of the area postrema, the calamus scriptorius (CS), was viewed as a reference point for determining the coordinates with optic fiber cannula (50–70 µm, Doric Lenses) directly connected to a laser output for blue light (∼473 nm, 1.20 µW) or green light (∼532 nm, 1.36 µW; Shanghai Laser & Optics Century). The laser output was controlled by a Master-8 stimulator (AMPI), and the intragastric pressure data were acquired using the PowerLab data acquisition system (ADInstruments).
Optogenetic control with the optic cannula was given either into the NTS or DMV at a 30° angle from the perpendicular. Responses evoked from the NTS were induced by a 90 or 60 s envelope of photograph excitation or inhibition by trains of light pulses (frequency 15 Hz, pulse width 10 or 40 ms, duty cycle of 15 or 60%). Duty cycles of 60% for a 90 s duration of light activation were used in early experiments, but a duty cycle of 15% for 60 s was more reliable and used for the rest of the study. These optogenetic control parameters are compatible with those used in respiration studies (Abbott et al., 2013). Stereotaxic coordinates for the optic fiber cannula into the NTS were area postrema (AP) = +0.3–0.5 mm rostral to CS, ML = 0.1–0.3 mm lateral to the midline, and DV = 0.0–0.1 mm ventral to the dorsal surface of the medulla. Coordinates for the DMV were +0.3–0.5 mm rostral to CS, 0.1–0.3 mm lateral to the midline, and 0.2–0.3 mm ventral to the dorsal surface of the medulla. Optogenetic activation tags were also acquired using the PowerLab data acquisition system.
Histologic verification of pipette tracks
At the end of each experiment, the mice were killed with an overdose of urethane. The brain was removed and placed in a fixative-cryoprotectant solution composed of 4% phosphate-buffered paraformaldehyde and 10% sucrose (0.1 m, pH 7.4) for at least 48 h. The brainstem was dissected and cut on a cryostat into 50 µm coronal sections, which were mounted serially onto gelatinized slides and stained with neutral red staining solution. The locations of the cannula tracks were identified using the mouse brain atlas (Paxinos and Franklin, 2007). To document cannula sites in the brain, microphotographs and camera lucida drawings were made of each pipette track.
Experimental design and statistical analysis
Analysis of the intragastric pressure recordings was completed off-line using Chart (ADInstruments) and Prism (GraphPad) software packages. All the experimental recordings were initially filtered using a root mean square (RMS; 2 s moving window) algorithm to account for respiration and various other signal artifacts. All peaks were selected within the 90 or 60 s of light activation and compared with the immediately prior peaks, selected for the same amount of time (baseline).
An algorithm for calculating average peak-to-peak values was used to determine a change in amplitude of gastric phasic contractions. This algorithm was used because gastric phasic contractions have a low frequency of occurrence. Because of the natural variance in the amplitude of gastric contractions, the average peak-to-peak values proved to be a uniform method to compare all datasets. However, following analysis of peak-to-peak parameters of baseline values and those measured during light activation, it was found that differences in intrinsic activity of an animal could introduce considerable variance in the results. An optogenetic-induced response in each animal was normalized to its baseline value to solve this problem, which was displayed as a percentage change in the gastric contractions. The noise level on all completed peak-to-peak analysis was set to zero as an RMS algorithm had already filtered the data.
Baseline and peak values during light activation were taken as a percentage difference using the following formula: ((light activation-baseline)/(baseline)) × 100). The same formula has been used to calculate the area under the trace (tone).
Quantification and statistical analysis
Statistical analyses were performed using Prism 9 (GraphPad). All data indicated in the text and figures were checked for normal distribution, and the results expressed as means ± SEM unless otherwise stated. The statistical analysis comparisons between two groups was performed by a paired Student's t test or unpaired t test with Welch's correction. Comparisons among three or more groups was performed with one-way ANOVA followed by the Holm–Šídák's post hoc test, Brown–Forsythe and Welch test followed by Dunnett's T3 multiple comparisons test, or mixed-effects model followed by Tukey's multiple comparisons test.
All percentage from in vivo data were analyzed for the significance of normality using the D'Agostino–Pearson omnibus and or Shapiro–Wilk normality test. A one-sample t test was performed on the data with the hypothetical value set at zero. For ipsilateral vagotomy data, a two-sample paired t test was performed.
Animal number and cell numbers are indicated in the text. The criterion used to determine statistical significance was p < 0.05.
Results
It is well documented that the DVC serves as an autonomic gateway that comprises many neuronal subtypes, which are essential to the homeostatic regulation of autonomic functions such as those related to the gastrointestinal, cardiovascular, and respiratory systems (Feldman and Ellenberger, 1988; Rinaman et al., 1989; Lettgen et al., 1994; Lawrence and Jarrott, 1996; Gillis et al., 2022). We focused on the NPY neurons in the DVC to delineate their role in gastric motility. To accomplish this, we examined their (1) distribution in the DVC, (2) presence in the vagovagal circuitry, (3) electrophysiological characteristics, (4) synaptic connectivity, and (5) influence on end-organ function (i.e., stomach).
Distribution of NPY neurons in the DVC and their phenotype
We first established distribution of the NPY neurons in the DVC of the npy-Cre knock-in mouse (catalog #027851, The Jackson Laboratory), crossed with the Rosa-tdTomato Flox reporter mouse (catalog #007905B6, The Jackson Laboratory) that expresses robust tdTomato fluorescence following Cre-mediated recombination directed by the mouse npy promoter (Milstein et al., 2015). This strain was also reported in a recent study (Chen et al., 2020). A prominent diagonal band of TdTomato-expressing neuron was observed in NTS (Fig. 1H), which extended mediolaterally from the central canal encompassing subnuclei, including the medial, central, dorsal (medial and lateral), with a few scattered neurons and their dendrites in the commissural and intermediate subnuclei. Distinct pockets of neurons were also observed in the ventrolateral NTS. In the underlying DMV nucleus, although relatively few neurons were scattered along the mediolateral axis, neurons favored a more lateral location and were interspersed with other DMV cells at its border with the NTS. This appearance was in contrast with NPY terminals in the DMV that showed distinct punctations that were more or less evenly spread throughout the nucleus as they were in the overlying NTS. Also, labeled neurons were present in the dorsally located AP (Fig. 1H).
A subset of NPY neurons in the NTS and DMV of a transgenic npy-Cre;tdTomato animal is glutamatergic. A, B, Representative micrograph of the DVC (A) that was used for RNAscope in situ hybridization probes for NPY, Slc17A6 (VGlut2), tdTomato mRNA (B), and GAD1 (GAD67 not shown) in distinct adjacent brainstem paraffin sections. C–G, Graph shows the percentage RNA target expression of GAD1, VGlut2, and NPY in the NTS (C). White box indicates region imaged at higher magnification in D and E; F and G are from adjacent 5 μm section. Colocalization with the NPY and Slc17A6 (E) or NPY and GAD1 (G) with tdTomato mRNA (TdTom, D, G). H, Distribution of td-Tomato neurons in the DVC of an npy-Cre;tdTomato mouse. I, J, Colocalization of td-Tomato neurons (I) with anti-NPY immunostaining in the NTS (J, NPY-Ab). K–M, Coexpression of transgenically td-Tomato labeled NPY neurons with YFP expression induced by unilateral virus injection in the DVC and their overlap in the NTS. N–P, Example of immunoreactivity to the VGlut2 transporter antibody as visualized by the secondary Alexa 647 antibody in the NTS of a npy-Cre;tdTomato mouse. (Note: Double arrowheads in all micrographs identify neurons that show respective colocalization.) 12N, Hypoglossal nucleus; CC, central canal. Calibration: A, B, 100 µm; D–G, 50 µm; H, 100 µm; I–M, 40 µm; N–O, 60 µm.
Using the Cre/Lox technology, transgenic breeding can potentially lead to recombination during development, inducing expression in off-target neurons (Taniguchi et al., 2011). Using in situ hybridization with RNAscope and probes targeting NPY and TdTomato (Fig. 1A–E) revealed that NPY mRNA expression is present in 84 ± 10% of a cell expressing mRNA for TdTomato (mean ± SD, three mice, 27 sections analyzed). We also performed immunohistochemistry for NPY (n = 2 mice, six slices) to visualize tdTomato neurons expressing NPY (Fig. 1I,J) using a polyclonal antibody, which was visualized with confocal microscopy using a far-red secondary Alexa 647 antibody. TdTomato neurons were partially labeled with anti-NPY antibodies (48 ± 8%; mean ± SD) whereas 74 ± 14% (mean ± SD) of NPY positive neurons expressed TdTomato. NPY antibody staining appeared to be restricted to the perikaryal cytoplasm and surrounding neuropil, as evident from the labeled fibers and puncta (Fig. 1J). The discrepancy between NPY mRNA and antibodies detection could be because, as in most peptides, NPY expression fluctuates because of alterations in physiological activity and development.
To further ascertain that cre-expression in DVC neurons was not because of developmental off-target recombination, unilateral microinjection of the rAAV5/EF1a-DIO-EYFP virus (University of North Carolina Virus Vector Core) was made in the dorsomedial NTS region to target neurons in npy-Cre;tdTomato mice (n = 2 mice). The virus-labeled NTS neurons also exhibited npy-Cre mediated recombinatorial tdTomato expression (Fig. 1K–M; 93 ± 4% colabeling, n = 5 slices). This indicates that the cre transcript is similar between neurons of adult mice. In general, when the transcriptional activity of NPY goes down, Cre expression should also decrease, although at a slower pace, which results in decreased reporter expression.
Because the regulation of gastric motility is intimately dependent on the nature of the ongoing neural activity in the DVC (Rogers et al., 1996), we next examined the potential neurotransmitter associated with the NPY neurons. We focused on glutamate as its VGlut2 transporter has been reported to be present, among others, in the medial and dorsomedial NTS subnuclei, which is central to the vagovagal circuitry of the DVC controlling gastric tone and motility (Lin et al., 2004, 2008; Okada et al., 2008).
To further characterize the phenotype of npy-cre neurons, RNAscope studies extended to probes targeting, in addition to NPY and TdTomato, VGlut2 (Slc17A6) or for comparison GAD67 (GAD1) in distinct adjacent brainstem paraffine sections (14 sections, three mice) analyzed at the intermediate rostrocaudal level, which is where the majority of subsequent studies were performed (Fig. 1B,C,F,G). Our experiments were similar to those performed to validate the Sst-Cre line in the DVC (Thek et al., 2019, their Fig. 1). These results revealed that Slc17A6 mRNA expression is present in 53 ± 24% (mean ± SD) of a cell expressing mRNA for TdTomato. In contrast, GAD1 mRNA expression was present in 22 ± 16% (mean ± SD) of a cell expressing mRNA for TdTomato.
We also assessed the VGlut2 transporter antibody with immunoreactivity as visualized by the secondary Alexa 647 antibody as shown in Figure 1N–P. In 19 hindbrain sections from three npy-Cre;tdTomato mice 52 ± 14% of TdTomato neurons were positive for the VGlut2 antibody.
Altogether, these studies demonstrate that NPY neurons in npy-Cre;tdTomato mice are distinctly distributed throughout the extent of the NTS and DMV, with their density especially notable in the NTS. Furthermore, a larger subset of the NPY neurons in the DVC is glutamatergic, whereas a smaller subset is GABAergic.
NPY neurons are part of the vagal reflex circuit that regulates gastric motility
To establish that the NPY neurons of the DVC are part of the autonomic circuitry that controls gastric motility, we trans-synaptically traced their connectivity from the stomach with PRV (Card et al., 1993; Smith et al., 2000; Lewin et al., 2016). The PRV isovariant 152 was injected into the gastric antrum of npy-Cre;tdTomato transgenic mice. Our reason for tracing NPY neurons from the antrum was the important role of this gastric region in gastric motility (el-Sharkawy and Szurszewski, 1978; Gillis et al., 1989; Lüdtke et al., 1991; Rogers et al., 1996; Gillis et al., 2022). PRV retrogradely labeled NPY neurons in the DMV and NTS (Fig. 2A–F). Notably, particularly in the DMV, premotor antrum-projecting neurons were observed to be surrounded by NPY terminals (Fig. 2E,F), suggesting that DMV neurons may be the recipients of regulatory modulation by these neurons. This observation led us to explore what other neurons may likewise be in close apposition with these terminals. Of particular interest were the SST neurons in the DVC, which we previously reported to have a strong influence on the activity of antrum-projecting DMV neurons and, by extension, on gastric motility (Lewin et al., 2016). To examine whether SST neurons received NPY terminals, as both SST and NPY transgenic mice are cre drivers, we have bred sst-Cre;tdTomato mice with npy-hrGFP-BAC (van den Pol et al., 2009; Fig. 2G–J), and their brains were processed for imaging. Inspection of the DVC under confocal microscopy showed that both NPY and SST neurons in this complex represented distinct populations of neurons with no overlap (n = 4 mice, eight slices) whose terminals were in apposition to each other (Fig. 2H–J). Moreover, each neuron population displayed terminals from the other population and suggested reciprocal connectivity within the same population (Fig. 2J, arrows).
NPY neurons in the DVC are part of the vagal circuit that regulates gastric motility and tone. A–C, Micrographs showing, after an ∼72 h survival period, the result of the injection of the PRV (isovariants 152) into the gastric antrum polysynaptically labels NPY neurons in DMV and NTS of a npy-Cre;tdTomato. D, Quantification of the percentage of neurons labeled with PRV-152 throughout the rostrocaudal extent of the DVC of npy-Cre;tdTomato mouse (n = 14 sections, >15 PRV+ cell count per section). AP, area postrema; 4V, 4th ventricle. E, F, Representative micrographs of PRV-614-labeled gastric-antrum-projecting neurons in the DMV displaying apposition with tdTomato positive terminals. (Note: Double arrowheads in all micrographs identify neurons showing colocalization.) G, Breeding strategy for transgenic mice labeling both NPY and SST neurons. H–I, Confocal micrographs showing digitally rendered images of NPY (H) and SST (J) neurons showing the apposition of NPY terminals onto SST neurons (white/green arrows) and SST terminal onto NPY neurons (white/red arrows). Calibration: A–C, 60 µm; E, 20 µm; F, 4 µm; H–J, 20 µm.
These results show that the NPY neurons are (1) part of the vagovagal circuitry that controls the stomach and (2) reciprocally connected to themselves and SST neurons in the DVC.
NPY neurons are functionally interconnected by an excitatory network
To study the functional characteristics of NPY neurons, we combined optogenetics with electrophysiology and pharmacological techniques in brainstem slices from transgenic mice expressing Chr2. Excitation of NPY neurons in the NTS of npy-Cre;ChR-EYFP mice produced consistent action potentials with a train or persistent (2 s) light activation in extracellular or intracellular recordings (Fig. 3A,B). Depolarization current steps produced action potentials that displayed little spike-frequency adaptation (Fig. 3C,D), whereas injection of hyperpolarizing currents produced a noticeable depolarization sag (2.1 ± 1.5 mV, n = 11 cells in six npy-Cre;tdTomato mice, mean ± SD). Cell-attached recordings (loose seal <40 MΩ) showed that NPY neurons were spontaneously active with a mean firing rate of 3.6 ± 3.6 Hz (n = 11 cells in five npy-Cre;tdTomato mice, mean ± SD; Fig. 3E). The intracellular current-clamp recording showed the mean peak amplitude, rise time, and half width of the action potential to be 76.5 ± 4.8 mV, 0.8 ± 0.04 ms, and 2.3 ± 0.2 ms in 11 neurons from 6 npy-Cre;tdTomato mice, respectively (Fig. 3E). The average resting membrane potential and input resistance were −61.4 ± 1.6 mV and 1.2 ± 0.2 GΩ in 11 neurons from 6 npy-Cre;tdTomato mice (Fig. 3E). Moreover, optogenetic inhibition of tonically active NPY neurons via activation of the proton-pump opsin ArchT in npy-Cre;ArchT-EGFP mice showed that they exhibit robust rebound activity on hyperpolarization (Fig. 3F) from a mean baseline of 3.3 ± 1.2 Hz before light activation to 5.9 ± 1.1 Hz afterward.
NPY neurons are functionally interconnected by an excitatory network. A, B, Response to light stimulation of NPY neurons in the cell-attached (A) and current-clamp modes (B) in a npy-Cre;ChR2-EYFP transgenic mouse. C, D, Representative voltage responses (C) and phase plot (D) of an NPY neuron to 10 pA current steps injections. [Note: hyperpolarization-induced depolarization sag in C (arrow)]. E, Graphs showing the spontaneous action potentials (AP) firing rate (5 npy-Cre;tdTomato mice), parameters of the action potential, resting potential, and input resistance (RM) in NPY neurons in the DVC (6 npy-Cre;tdTomato mice). F, Representative light-induced (train) suppression of action potentials in an NPY neuron in a npy-Cre;ArchT-EYFP mouse. [Note: the robust rebound response on cessation of the light train (red stippled arrows)]. G–I, Representative recordings of light-evoked excitatory currents in an NPY neuron before and after exposure to NBQX. J, K, Subtraction of the average trace of I from all traces in J exposure to NBQX. L, Graph of the peak of light-evoked EPSC from reciprocal connections between NPY neurons that were blocked by NBQX (EPSC) compared with peak of the direct-light-evoked ChR2 current (n = 9 cells in 9 mice).
We next examined the effect of the NPY neurons in response to short (5 ms) and long (2 s) pulses of light in neurons in the DVC of npy-Cre;ChR2-EYFP mice. Light excitation evoked large inward currents with superimposed postsynaptic currents (PSCs; Fig. 3G). Spontaneous and light-evoked PSCs but not the large inward currents were reversely blocked by local exposure, via a y-tube, to the glutamatergic AMPA receptor antagonist NBQX (5 µm; Fig. 3H,I). Subtraction of the average trace in Figure 3H from all traces in Figure 3G revealed light-evoked EPSC from reciprocal connections between NPY neurons that were blocked by NBQX (Fig. 3J; control −3K plus NBQX) as shown in the superimposed traces in Figure 3, J and K, and the summary of results in Figure 3L (n = 9 cells in nine mice) comparing the peak EPSCs with the peak of the direct ChR2 currents. These studies show that NPY neurons of the DVC of npy-Cre mice (1) are tonically active, (2) display action potentials with little spike-frequency adaptation in response to depolarizing current injection, (3) exhibit hyperpolarization-induced depolarization sags, and (4) are functionally connected via an excitatory synaptic network.
Gastric-projecting DMV neurons are excited by optogenetic activation of NPY neurons
The DMV forms the parasympathetic efferent arm of the vagovagal gastric reflex circuitry, which, among other gastric functions, controls motility of the upper gastrointestinal (GI) tract (Gillis et al., 1989; Ferreira et al., 2002; Niedringhaus et al., 2007; Cruz et al., 2007; Gillis et al., 2022). As gastric-projecting DMV neurons in this nucleus display close apposition of NPY terminals on their cell bodies (Fig. 2F), we wondered about the functional consequence of stimulating these terminals on the activity of the DMV neurons. To identify these gastric-projecting neurons, crystals of the cell-viable monosynaptic retrograde tracer DiI were applied to the gastric antrum and isolated from the surrounding tissue by a silicone glue as previously described by us (Lewin et al., 2016). Optogenetic excitation of NPY neurons increased the firing frequency of the antrum-projecting neurons from a mean baseline of 1.9 + 0.3–2.9 + 0.4 Hz (Fig. 4A–C; p = 0.0024, t = 3.300, two-tailed paired t test, n = 33). Similarly, baseline synaptic activity was also increased on long-lasting light exposure from a baseline of 4.4 + 0.7–7.9 + 1.2 Hz (Fig. 4D–F; p = 0.0132, t = 3.170, two-tailed paired t test, n = 9). Light excitation evoked synaptic currents with both excitatory and inhibitory components. This feature was particularly evident from the increase in the amplitude of the EPSCs as opposed to IPSCs, an observation that was made possible by voltage clamping each DMV neuron at −40 mV or by exposing the neurons to the GABAA or glutamate antagonists, respectively (Fig. 4G–J). Figure 4K summarized the percentage of neurons responsive to light on action potential firing and their synaptic activation. In cells displaying light-induced EPSCs or EPSC-IPSC sequence, we measured the latency of peak EPSC and IPSCs from the beginning of the light pulse (lag time) and their relative lag time as time difference between the peak EPSC minus peak IPSC (Fig. 4L). The EPSCs latency (11.2 ± 0.8 ms, n = 43) was shorter than that for IPSC (19.4 ± 2.8 ms, n = 17). The lag time between IPSC and EPSCs was −6.4 ± 2.8 ms (n = 17). These studies may indicate that the action of NPY neurons on gastric-antrum-projecting neurons is mainly excitation but can also trigger inhibition via the polysynaptic activation of GABAergic neurons. However, evoked PSC latency has been shown to be a poor predictor of monosynaptic versus polysynaptic connections in the DVC (Doyle and Andresen, 2001; Appleyard et al., 2007; Neyens et al., 2020), and light-evoked IPSCs could be because of the small portion of GAD1-expressing NPY neurons.
Gastric-projecting DMV neurons are excited by optogenetic stimulation of NPY neurons. A, B, Light-induced responses from retrogradely labeled DMV neurons (configuration as depicted in the diagram; star denotes the gastric site of the retrograde dye application). B, Consistent responses are elicited by a single pulse followed by a train of pulses (15 Hz; overlay of 5 sweeps shown). C, Graph showing light-evoked changes in the firing rate of action potentials in DMV neurons (n = 33 neurons from 9 mice). D–F, Change in excitatory synaptic currents in response to light stimulation in DMV neurons (n = 9 neurons from 4 mice). G, Overlay (top) and distribution (bottom) of traces illustrating light-induced EPSC-IPSC sequences at −40 mV holding voltage (Vh). H, Graph showing light-evoked changes in postsynaptic currents in DMV gastric-projecting neurons at −70 or −40 mV Vh. I, J, Representative recordings of light-evoked excitatory currents in at DMV neurons at Vh −70 mV and two additional DMV neurons at Vh −40 mV that are abolished by NBQX (5 μm) but not by BMR (25 μm). K, Summary graphs showing the percentage of cells displaying the light-evoked changes in action potentials (spike), EPSCs, or EPSC-IPSC sequence in DMV-gastric projecting neurons (n = 16 mice). L, Latency of peak EPSC and IPSCs from the beginning of the light pulse and the differences in their lag time as time difference between the peak EPSC minus peak IPSC (n = 43 and 17 neurons from 16 mice).
Synaptic interactions of NPY and SST neurons and their effect on premotor DMV neurons in the DVC
Previously, we reported that SST neurons are connected to premotor DMV gastric neurons, whose activity is robustly affected by their excitation (Lewin et al., 2016). Our present imaging data show that in addition to the DMV neurons, SST neurons appear to be anatomically connected to the NPY neurons (Fig. 2J). This observation compelled us to assess the interaction between these two neurons and gastric premotor neurons in the DVC. To accomplish this, a similar type of breeding strategy as mentioned before was used (i.e., crossing npy-Cre;Chr2-YFP transgenic mice with sst-Flp;;mCherry mice or crossing sst-Cre;Chr2-YFP mice with npy- hrGFP-BAC mice).
As with the DMV premotor gastric neurons (Fig. 4), we first determined whether the excitation of NPY neurons would influence the synaptic activity of the SST neurons. Optogenetic stimulation of NPY neurons in 4 npy-Cre;ChR2-YFP mice induced light-activated EPSC-IPSC timing sequences that varied between the different SST neurons; in some neurons, light-evoked EPSCs occurred before the IPSCs, whereas in others, the opposite was true (Fig. 5A,B). Overall, as seen in the percentage of light-responsive SST neurons, the EPSCs evoked were more than the IPSCs. The amplitudes and occurrence of both were enhanced by exposure to 4-AP (100 µm; Fig. 5C) to increase synaptic release, suggesting polysynaptic activation of inhibitory neurons.
Synaptic interactions and connectivity of neurons in the DVC. A, B, Optogenetic stimulation of NPY neurons evokes EPSCs and IPSCs in SST neurons that vary in their timing sequence. C, % neurons displaying light evoked EPSCs or IPSCs (top) and graphs showing differential changes in ChR2 light-evoked synaptic activity at −70 and −40 mV holding potential in two Sst neurons in a npy-Cre;ChR2-YFP mouse crossed with a sst-Flp;mCherry mouse in the absence (left) or the presence (right) of 4-AP. D, Representative waterfall plot displaying the robust suppression of action potentials in a NPY neuron in response to a short and long light-pulse stimulation in a sst-Cre;ChR2 neuron. E, Overlay (top) and distribution (bottom) of light-evoked IPSCs at −40 mV holding potential in an NPY-GFP neuron in a sst-Cre-Chr2-YFP; npy-hrGFP BAC mouse. F, Summary graph showing a side-by-side comparison of evoked postsynaptic currents by light in SST and NPY neurons in npy- or sst-Cre;ChR2-YFP mice crossed with the respective reporter n = 10 cells from 2 npy-Cre;ChR2-YFP mice; n = 12 cells from 5 sst-Cre;ChR2-YFP mice. G, Representative recording of light-induced suppression of action potentials in a NPY neuron in npy-Cre;ArchT-EGFP mouse is accompanied by an increase in the rebound firing of action potentials that is potentiated by closely spaced light pulses (green overlays). H, Dual recordings from SST and NPY neurons in response to light activation in an sst-Cre;ChR2-YFP mouse. I, J, Subcellular channelrhodopsin-assisted circuit mapping of light-evoked responses from DMV neurons in two sst-Cre;ChR2-YFP (IPSCs, I) and 3 npy-Cre;ChR2-YFP (EPSCs, J) mice. K, L, Summary of results. M, Micrograph showing red fluorescence expression in DVC neurons of sst-Cre;tdTomato mice (arrows, calibration: 20 µm). N, Light stimulation elicits postsynaptic currents in these SST neurons (left trace) that are inhibited by gabazine (10 μm), a GABAA antagonist, (right trace). (Note: The remaining light-induced current in the right trace is ChR2-induced direct current.)
Light activation of SST neurons, as was reported for the DMV, antrum projection neurons (Lewin et al., 2016) profoundly suppressed (88 ± 5%, 21 cells, six mice) the activity of NPY neurons from a mean baseline of 2.5 ± 0.5 Hz (Fig. 5D), which was associated with robust light-evoked IPSCs (Fig. 5E,F; n = 10 cells from 2 npy-Cre;ChR2-YFP mice; n = 12 cells from 5 sst-Cre;ChR2-YFP mice). Furthermore, as noted in our experiments with npy-Cre;ArchT-EGFP mice (Fig. 3F), recovery of NPY neurons from inhibition (in this case by SST neuron excitation) was accompanied by rebound spiking (Fig. 5G). Altogether, the striking inhibitory effect of SST stimulation on NPY neurons is best illustrated by the dual-cell attached recording example shown in Figure 5H (n = 5).
To determine whether the light-induced effects were direct from presynaptic neuron terminals, we used subcellular channel rhodopsin-assisted circuit mapping (Petreanu et al., 2009). We used TTX (1 µm and 4-AP, 50 µm) to prevent action potentials with ChR2 activation, thus demonstrating that the synaptic activity onto a neuron is because of the presence of synaptic terminal on the recorded neuron (Petreanu et al., 2009). Studies of light-evoked synaptic activity using this technique revealed that the DMV premotor gastric neurons received differential functional input from SST and NPY neurons. Whereas the SST neuron more reliably influenced DMV gastric neurons via a direct monosynaptic input (five of five cells responding), the NPY neuron was seen to elicit its effect with optogenetic excitation in the presence of both TTX and 4-AP in only two of six DMV neurons tested (Figure 5I–L; n = 5 cells from two sst-Cre;ChR2-YFP mice; n = 6 cells from three npy-Cre;ChR2-YFP mice).
In assessing the interaction between the different neurons in the DVC, we were also interested to know whether, similar to the NPY, light activation of SST neurons was capable of eliciting changes in the synaptic activity of other SST neurons in the DMV, as has been previously reported in the NTS (Thek et al., 2019). Light excitation of SST neurons evoked PSCs blocked by the GABAA antagonist gabazine (10 µm; n = 4 of 6 neurons, two mice; Fig. 5M,N). Altogether, these studies demonstrate that NPY neurons are connected to SST neurons in the DVC, which in turn have a powerful influence on their tonic activity. Furthermore, they show that whereas NPY neurons may indirectly influence DMV gastric neurons and SST neurons, they also directly influence premotor neurons.
In vivo optogenetic manipulation of NPY and SST neuronal activity in the DVC affects gastric motility and is site dependent
The DVC of the hindbrain is critical to the vagovagal control of gastric motility. In particular, its nuclei, NTS and DMV, differentially regulate the activity of the upper GI tract. Microinjection of l-glutamate in the NTS inhibits GI activity, whereas its application in the DMV excites it (Ferreira et al., 2002; Cruz et al., 2007; Niedringhaus et al., 2007; Herman et al., 2009, 2010; Richardson et al., 2013). As both NPY and SST neurons are found in these nuclei and strongly affect the activity of gastric neurons (Fig. 1; Gao et al., 2009; Lewin et al., 2016; Thek et al., 2019), optogenetic studies were undertaken to parse out the role of these neurons in regulating gastric motility. Responses evoked from the NTS were induced by a 60 s envelope of photograph excitation or inhibition using trains of light pulses (frequency 15 Hz, pulse width 10 ms, duty cycle of 15%). These optogenetic parameters are derived empirically from preliminary in vitro studies (Fig. 6A), which showed that they robustly activate SST or NPY neurons in the DVC. The parameters above were similar to those used by Guyenet and colleagues to study control of respiration (Abbott et al., 2013).
In vivo optogenetic stimulation of NPY and SST neurons in the DVC influences gastric motility and tone. A, Stimulation parameters used in vivo reliably elicit action potentials in vitro in both NPY and SST neurons in the DMV of npy- and sst-Cre;ChR2-YFP mice. B, A diagram of the ipsilateral vagotomy strategy used to separate light-evoked effects on gastric motility in the NTS from those in the underlying DMV (see text). C, D, Stereotaxic location and hindbrain approach of optogenetic stimulation in the DVC. E, F, Chart recordings demonstrating that optogenetic stimulation of the NTS in a npy-Cre;ChR2-YFP mouse robustly excites intragastric pressure (IGP) effect that is not blocked by ipsilateral vagotomy (E). G, H, Representative recordings showing that DMV stimulation leads to consistent inhibition of gastric motility (H) that is blocked by ipsilateral vagotomy. I, Graph displaying the extent of changes in percentage of motor index of the amplitude of phasic contractions (Amplitude) and gastric tone (Area Under Trace) evoked by light from baseline activity in npy-Cre;ChR2-YFP transgenic animals. **** p < 0.0001, * p = 0.0304, *** p = 0.0002. J, Representative recording showing inhibition of gastric tone and motility in the NTS in an npy-Cre;ArchT-EGFP. K, L, Two representative chart recordings showing how stimulation of the NTS in distinct sst-Cre;ChR2-YFP mice induce an initial excitatory phase followed by an inhibitory effect. M, Representative chart recording showing the light-induced inhibitory effect from the DMV of a sst-Cre;ChR2-YFP mouse that was blocked by ipsilateral vagotomy. N, Graph depicting the changes in the amplitude and gastric motility index evoked by light from baseline activity in sst-Cre;ChR2-YFP mice. O, Representative chart recording showing light-induced inhibition of gastric tone and motility in the DMV in a sst-Cre;ArchT-EGFP mouse that was blocked by ipsilateral vagotomy. *** p = 0.0008, * p = 0.0396(amplitude), * p = 0.0112 (AUT). P, Light (top) and phase-contrast (bottom) micrographs showing a representative optic cannula track in the DVC. Q, Camera lucida drawings of hindbrain sections denoting examples of the localization of cannula tracts (NTS, squares; DMV, circles) in npy- and sst-Cre;ChR2-YFP mice transgenic animals (blue and red circles, respectively) along the rostrocaudal axis of the DVC.
To differentiate responses induced from the NTS from those deriving from the underlying DMV, we used ipsilateral vagotomy, which we have shown to inhibit responses evoked only from the DMV as this nucleus projects ipsilaterally to postganglionic gastric neurons (Ferreira et al., 2002; Cruz et al., 2007; Herman et al., 2009; Richardson et al., 2013). This procedure is in contrast to the NTS, which projects bilaterally to the DMV (Norgren, 1978); hence, responses evoked from this nucleus are not blocked by ipsilateral vagotomy (Fig. 6B; Ferreira et al., 2002; Cruz et al., 2007; Herman et al., 2009; Richardson et al., 2013). To target the NTS or DMV, a direct visual approach was adopted in anesthetized mice (Fig. 6C,D) using calamus scriptorius as a reference point (Ferreira et al., 2002; Cruz et al., 2007).
In npy-Cre;ChR2-YFP mice, light activation of the NTS (n = 9) induced a robust increase in gastric motility and tone followed by a noticeable quiescent period marked by an initial drop in tone. (Fig. 6E). The increase in motility and tone to optogenetic activation was not affected by ipsilateral vagotomy (Fig. 6F; n = 7). This result is in contrast to light activation of Chr2 in the DMV (n = 5 mice), where suppression of gastric motility induced by light (Fig. 6G) was blocked by ipsilateral vagotomy (Fig. 6H), indicating an effective placement of the optic cannula in this nucleus.
During light activation of Chr2 neurons in the NTS, both amplitude of phasic contractions and gastric tone significantly increased compared with the baseline (44.2 + 5.8% and 54.3 + 8.5%, respectively; Fig. 6I, Table 1; p = 0.0001, t = 7.586 and p = 0.0002, t = 6.402, two-tailed, one sample t test). In contrast, light activation of ChR2 in the underlying DMV significantly decreased the amplitude of phasic contractions that averaged 26.5 + 8.1% relative to baseline, whereas gastric tone was not significantly affected (14.7 + 13.1% decrease from the baseline; Fig. 6I, Table 1; p = 0.0304, t = 3.283 and p = 0.3225, t = 1.128, two-tailed, one sample t test).
Optogenetic control of change of gastric motility (MMI%) and tone (AUT%) in the NTS and DMV nuclei
A group of npy-Cre;ArchT-EGFP mice that had been similarly tested showed opposite results on gastric activity compared with those described in npy-Cre;ChR2-YFP mice. Light activation of ArchT in the NTS in these mice (n = 7) produced inhibitory effects on gastric motility that exhibited a robust rebound activity at the end of the inhibition but not tone (Fig. 6J; average data are 27.6 ± 7.6 and 9.4 ± 5.8, respectively; p = 0.0112, t = 3.615 and p = 0.1561, t = 1.621, two-tailed, one sample t test; Table 1). Conversely, light activation of ArchT in the DMV (n = 5) in the of npy-Cre;ArchT-EGFP mice significantly increased gastric motility and tone (18.8 ± 4.2% and 17.5 ± 4.4%, respectively; p = 0.0116, t = 4.411 and p = 0.017, t = 3.939 and, two-tailed, one sample t test; Table 1).
In sst-Cre;ChR2-YFP transgenic mice, light-activated gastric motility was similar to that of npy-Cre;ChR2-YFP. However, despite the analogy with NPY-ChR2, the response elicited in the NTS produced short-duration contraction(s) that sometimes quickly proceeded to a longer duration of depression of the amplitude of motility (Fig. 6K,L, examples in different mice). In NTS of sst-Cre;ChR2-YFP transgenic mice (n = 9), light activation of ChR2 produced a significant increase in the amplitude of phasic contractions and gastric tone (35.6 + 6.8% and 26.2 + 7.7%, respectively; Fig. 6N, Table 1; p = 0.0008, t = 5.225 and 0.0112, t = 3.415, two-tailed, one sample t test), which was not affected by ipsilateral vagotomy (n = 7).
In contrast, in the DMV (n = 5) as shown in the example in Figure 6M, light activation of ChR2 suppressed the amplitude of phasic contractions by 25.4 + 8.4% (Fig. 6N, Table 1; p = 0.0396, t = 3.009, two-tailed, one sample t test), and this effect was blocked by ipsilateral vagotomy (Fig. 6M). Instead, the gastric tone was not significantly affected (16.0 + 11.5% decrease from the baseline, Fig. 6N, Table 1, p = 0.2365, t = 1.391, two-tailed, one sample t test). Similar to Sst-ChR2 mice, in sst-Cre;ArchT-EGFP transgenic mice, excitatory effects on gastric tone and motility increased during NTS light stimulation (n = 7). Gastric motility was significantly affected (27.3 ± 3.5%, p = 0.0002, t = 7.759, two-tailed, one sample t test), whereas tone was not (10.0 ± 5.6%, p = 0.1269, t = 1.772, two-tailed, one sample t test). Stimulation of the DMV (n = 6), conversely, induced significant inhibitory effects on the amplitude of phasic contractions but not on gastric tone (17.6 ± 4.5%, p = 0.0112 t = 3.921 and 6.9 ± 4.3%, p = 0.1724, t = 1.591, two-tailed, one sample t test, respectively), which were subsequently abolished by ipsilateral vagotomy (Fig. 6O, Table 1).
To document cannula placement sites in the hindbrain, photomicrographs and camera lucida drawings were made.
A representative photomicrograph illustrating an optic cannula track in the DVC (brightfield and darkfield images) is displayed in Figure 6P. The camera lucida drawings of the hindbrain sections denote the localization of cannula tracts in npy-Cre;ChR2-YFP and sst-Cre;ChR2-YFP transgenic animals (Fig. 6Q).
These in vivo studies (summarized in Table 1) show that in vivo optogenetic control in the DVC of NPY and SST transgenic mice affects gastric motility. Although light activation of ChR2 in NTS significantly increases the amplitude of phasic contractions and gastric tone in both mice, ChR2 activation in the DMV induces an opposite effect.
Altogether, the present study demonstrates that (1) NPY neurons in DVC, like SST neurons (i.e., Lewin et al., 2016), are important components of the gastric vagovagal reflex, are excitatory, and prevalently release glutamate, which is in contrast to other regions of the brain where NPY neurons are described as prevalently inhibitory (Pelkey et al., 2017); (2) NPY neurons may function in the gastric vagovagal reflex by mediating excitation from the NTS to the DMV and homeostatically regulate the vagal circuitry and gastric motility by the interaction of excitatory NPY neurons with inhibitory SST neurons; and (3) this is the first time to our knowledge that optogenetics has been used in vivo to excite (and inhibit) brainstem nuclei composing the main components of the gastric vagovagal reflex while monitoring this influence on gastric contractility and tone.
Discussion
This study observed the conspicuous presence of NPY neurons in the DVC that regulate vagal transmission to the stomach.
Distribution of NPY neurons and their connection to SST and gastric DMV neurons
Our results reveal abundant labeled cell distribution in the DVC of npy-Cre;tdTom mice. NPY mRNA expression in tdTomato neurons in the DVC was high, although the immunoreactivity for the neuropeptide was considerably lower. The discrepancy between NPY mRNA and antibodies detection could be because, as with most peptides, NPY expression fluctuates because of alterations in physiological activity and development. mRNA expression for both the VGlut2 and GAD67 transporter was detected in TdTomato neurons in the DVC of npy-Cre;tdTom mice. Excitatory Vglut2 mRNA-expressing neurons were more abundant than inhibitory GAD67 mRNA-expressing neurons, consistent with the distribution of light-evoked postsynaptic responses in target neurons of NPY neurons in npy-Cre;ChR2-EYFP mice. The immunoreactivity for Vglut2 is exceptional, as the NPY neurons are predominantly GABAergic in most brain areas (Horvath et al., 1997).
NPY neuron characteristics and functional interconnectivity in the DVC
In determining the synaptic interactions and connectivity of NPY neurons to SST neurons or DMV gastric output neurons, we initially optogenetically excited or inhibited these, profiling them in npy -Cre transgene crossed with ChR2 or ArchT reporter mice.
In NPY neurons of NPY-ChR2-eyfp mice, we observed evidence for reciprocal excitatory coupling to each other as seen using subtraction of direct ChR2 current recorded in NQBX. Silencing these neurons with ArchT opsin or with direct hyperpolarizations produced a large rebound increase in firing on recovery similar to that observed in cortical Camk2a kinase-expressing glutamatergic neurons (Han et al., 2011; Madisen et al., 2012). Our in vitro results with Vglut2 staining match the electrophysiological findings of light-induced EPSC in DMV neurons in NPY ChR2 mice. This excitatory action was observed between NPY and SST neurons, albeit much weaker. Thus, the most plausible hypothesis is that NPY neurons are prevalently excitatory. The coexpression between tdTomato and NPY mRNA argues against the possibility of misexpression in excitatory neurons using Cre-lox breeding (Madisen et al., 2012; Hu et al., 2013).
Light activation of an NPY neuron allows EPSC and IPSC sequences in the SST neuron, albeit with lower synaptic strength, and IPSCs probably because of the subpopulation of GAD67-NPY neurons or to excitatory polysynaptic activation of GABA neurons. Together, these results reveal a complexity of the interaction between the inhibitory SST neurons and the excitatory NPY neurons that is further increased by the reciprocal connectivity of both SST (Fig. 7N; Thek et al., 2019) and NPY neurons (Fig. 3J) and may play a substantial role as dynamic inhibitors of gastric vagal circuitry (Lewin et al., 2016).
A schematic illustration of a possible model to explain the results of the optogenetic stimulation in vivo of SST and NPY neurons in the DVC on gastric motility. A, In sst-Cre;ChR2-YFP mice light stimulation of the SST neurons of the NTS inhibited the GABA projection output neuron to the DMV, thus, exciting gastric motility. In the DMV, stimulation of SST neurons inhibits the acetylcholine (ACh) projection to the stomach, thereby, decreasing gastric motility. Light stimulation of NTS in SST ArchT mice increases contraction. SST neurons inhibit each other, and light decreases their firing. However, as their GABAergic synapses are not blocked, the remaining spontaneous activity of SST neurons increases inhibition of GABA output neurons. Light stimulation of DMV in SST-ArchT mice reduces contraction via a similar inhibitory action on the SST-SST reciprocal inhibition, which, however, has the end result of directly increasing inhibition of DMV neurons. NPY neurons are no longer inhibited by the SST neurons, thus allowing glutamate projection neurons on the NTS to become unopposed to enhance contraction motility. In the DMV, the ACh output projection neuron to the stomach has extensive GABA tone from local (SST) and extrinsic (GABA projection) neurons. Although ArchT proton pump activation silences SST neurons, extensive GABA tone still remains, which produces inhibition of the motility of the stomach. B, In npy-Cre;ChR2-YFP mice, light in the NTS excites NPY neurons, which in turn stimulate glutamate output neuron and ultimately activate ACh neuron in the DMV, increasing gastric motility. In the DMV, activation of the NPY neurons decreases motility by exciting SST neurons, which inhibit the ACh output projecting neuron to the stomach. In npy-Cre;ArchT-EGFP mice the light-induced gastric tone and motility in the two nuclei is opposite. In the NTS, silencing the NYP neurons via its proton pump stops the excitatory drive of the glutamatergic NTS-DMV projection neuron, inhibiting ACh neuron and in turn contractions.
In vivo optogenetic control of NPY and SST neurons in the DVC influences gastric motility and tone
After obtaining in vitro data on the synaptic interactions between NPY and SST neurons, we examined in vivo the end organ of the whole animal, namely, gastric stomach function, in NPY and SST mice (also, refer to a model of the signaling shown in Fig. 7).
Optogenetic stimulation of SST neurons may predominantly underlie the differential effects on gastric tone and motility seen with the blockade of the local GABA signaling in the DVC (Ferreira et al., 2002; Herman et al., 2009, 2010; Richardson et al., 2013). Moreover, the reciprocal inhibition between SST neurons is a vital synchronicity mechanism (Fig. 7A; Thek et al., 2019) that can be a target of suppression by both excitatory and inhibitory opsins. We suggest this as an explanation for the similar action of ChR2 activation and ArchT silencing, excitatory in the NTS and inhibitory in the DMV, and for the finding that light activation in the NTS and DMV of sst-Cre;ChR2-YFP transgenic mice elicited the same motility responses observed in npy-Cre;ChR2-YFP mice (see below).
Light activation of ArchT in the NTS of sst-Cre;ArchT-YFP mice increases contraction. SST neurons inhibit each other (Lewin et al., 2016; Thek et al., 2019), and light decreases their firing. However, as their GABAergic synapses are not blocked, the remaining spontaneous activity of SST neurons increases inhibition of GABA output neurons. Light activation of ArchT in the DMV of sst-Cre; ArchT-YFP mice reduces contraction via a similar inhibitory action on the SST-SST reciprocal inhibition that has the result to increase inhibition of DMV neurons directly. These results suggest that the SST-SST reciprocal inhibition acts as a brake of SST neurons and mediates inhibitory action on DMV neurons.
In the NTS, in vivo optogenetic control in NPY-ChR2-eyfp mice also showed activation of gastric motility, a robust effect that was not affected by ipsilateral vagotomy (Fig. 6). Previous reports showed that microinjection of glutamate, along with GABA blockade in the NTS, decreases gastric motility (Ferreira et al., 2002; Cruz et al., 2007; Herman et al., 2009; Richardson et al., 2013). Because NTS mainly inhibits DMV neurons projecting to the stomach, it is reasonable that glutamatergic and GABAergic exposure strikingly affects motility behaviors. We propose that NPY neurons may directly excite glutamatergic NTS-DMV-projecting neurons, overwhelming the dominant GABAergic-projecting neurons from the NTS to the DMV (Fig. 7B). In addition, the gastric motility induced by light activation of ChR2 was strong and durable during all the stimulation period, probably because of the monosynaptic excitatory network that NPY neurons create with each other (Fig. 7B). In fact, gastric motility is blocked in npy-Cre;ArchT-EGFP mice with light in the NTS (Fig. 6J). These results could also be explained with an activation of a direct NPY glutamatergic output to the DMV. However, this hypothesis is not consistent with the increased contractility produced by light in the DMV of ArchT mice and the unreliable light-evoked synaptic excitation seen with intracellular recording from DMV neurons in the presence of TTX and 4-AP when NPY-ChR2-YFP afferents were optogenetically activated. What should also be considered is the possibility of light-off rebound responses and the differential impact of ArchT activation at soma versus the synapse. (Mahn et al., 2016).
Light activation of ArchT in the DMV in npy-Cre;ArchT-EGFP mice increases baseline contraction because it decreases the ability of NPY neurons to activate the SST neurons. To explain the contradictory action of inhibition of contractility seen with the light on the DMV of sst-Cre;ChR2-YFP mice, we propose a significant role of activation of GABAB receptors that inhibits GABAergic terminals on DMV neurons and stimulates contraction (Cruz et al., 2019).
In the DVC, SST/GIN neurons of the vagovagal circuitry in the NTS receive direct vagus afferents (Glatzer et al., 2007; Lewin et al., 2016; Thek et al., 2019). In the DMV, these neurons inhibit projecting neurons to the stomach (Lewin et al., 2016). This inhibitory effect of SST is attenuated by α-melanocyte-stimulating hormone (a-MSH), which is a proopiomelanocortin (POMC)-derived peptide agonist melanocortin 4R, and DAMGO, µ-opioid agonist in the DMV (Lewin et al., 2016). However, the administration of these agonists in the DVC has contrasting effects on gastric motility. By suppressing local GABAergic signaling, α-MSH and DAMGO inhibit gastric motility in the NTS (Herman et al., 2010; Richardson et al., 2013) and excites it in the DMV (Richardson et al., 2013).
Like SST neurons, POMC neurons are also present in the DVC (Joseph et al., 1983; Appleyard et al., 2005) and may be directly activated by glutamatergic vagus afferents (Appleyard et al., 2005) and CCK neuron (Moran et al., 2001) via MC4-R signaling depending (Fan et al., 2004). These neurons are depressed by opioids (Appleyard et al., 2005; Glatzer et al., 2007). Furthermore, GLP-1 agonists in DVC (Ludwig et al., 2021; Zhang et al., 2021) that contribute to anorexic effects (Zhang et al., 2021) may do so via the POMC neurons releasing α-MSH in the NTS. This peptide could decouple GABA projection NTS-DMV neurons from local GABAergic SST neurons in the NTS. Although GABA projection NTS-DMV neurons receive vagus afferents, these neurons of gastric motility are subservient to local GABA signaling (Herman et al., 2009) mediated SST neurons (Lewin et al., 2016; Gillis et al., 2022). Altogether, we believe that the effects of SST and POMC neurons on gastric motility (among others) are influenced by vagus afferents and local neurons to regulate energy homeostasis. Of the local neurons, NPY neurons may play an essential role in this instance to stimulate SST neurons directly or indirectly. For example, orexigenic peptide NPY in NTS reduces the c-fos activation produced by peripheral CCK (McMinn et al., 2000) that positively influences POMC neurons. Activation of POMC neurons reduces feeding, partly by decreasing gastric motility by inhibiting local GABA SST neurons via releasing a-MSH. Moreover, microinjections of NPY in the NTS induced c-fos neuronal activity in DVC (Yang et al., 1995) and a significantly decreased gastric load-sensitive NTS cell response to distending gastric loads (Schwartz and Moran, 2002). An alternative explanation for our contrasting results may refer to noncanonical signaling (Chen et al., 2020) and the possibility of contrasting effects of secretion of NPY or either glutamate or both (Nectow et al., 2017). In summary, our finding show that the interaction of SST and NPY neurons in NTS has a potent effect on gastric contractions. This finding provides a model for studying other premotor systems in the nucleus, including cardiorespiratory systems, which among others, involve the monitoring of arterial baroreceptors and adapting pulmonary stretch receptors reflexes (Andresen and Kunze, 1994; Lawrence and Jarrott, 1996; Sapru, 1996; Kubin et al., 2006; Zoccal et al., 2014; Bai et al., 2019; Kupari et al., 2019).
Footnotes
This work was supported by the National Institutes of Health–National Institute of Diabetes and Digestive and Kidney Diseases Grant R01-DK117508. The Histopathology and Tissue Shared Resource at Georgetown University Medical Center is partially supported by National Institutes of Health–National Cancer Institute Grant P30 CA051008.
The authors declare no competing financial interests.
- Correspondence should be addressed to Stefano Vicini at svicin01{at}georgetown.edu or Niaz Sahibzada at sahibzan{at}georgetown.edu
