Figure 5. Insulin-induced AVP secretion is mediated by A1/C1 neurons.
(a) Upper: AAV-DIO-ChR2-mCherry was injected into the VLM of ThCre/+ mice or AvpGFP/+::ThCre/+ mice, targeting A1/C1 neurons. Lower: CRACM. Excitatory post-synaptic currents (EPSCs) were recorded in voltage-clamp mode in GFP+ (AVP) and GFP− neurons in the SON and PVH. The number (n) of neurons that responded to opto-activating A1/C1 terminals. Total of 36 neurons recorded from five mice. See Figure 5—figure supplement 1a. (b) Upper: Viral expression of ChR2-mCherry in A1/C1 neurons. Lower: A1/C1 neuron terminals co-localize with AVP-immunoreactive neurons. Representative from three mice. (c) Left: EPSCs evoked by opto-activation of A1/C1 terminals with 473 nm light pulses (arrows). Right: Light-evoked EPSCs following application of DNQX (20 µM). Representative of five recordings from three mice. (d) EPSC waveforms in a single GFP+ (AVP) neuron in response to repeated opto-activation of A1/C1 neuron terminals. Black line and shaded are=Mean ± SD of EPSCs. Light pulse=blue bar. Representative of three recordings from three mice. (e) EPSCs evoked by opto-activating A1/C1 terminals at baseline (left) and following addition of TTX (1 µM; middle) and 4-AP (1 mM; right). Representative of three recordings from three mice. (f) AAV-DIO-hM3Dq was injected into ThCre/+ mice, targeting A1/C1 neurons. CNO (1 mg/kg) was then injected (i.p.). Antagonists (or vehicle) for the V1bR (SSR149415, 30 mg/kg) or glucagon receptor (GCGR; LY2409021, 5 mg/kg) were injected 30 min prior to CNO. Plasma glucose and glucagon were then measured. See Figure 5—figure supplement 1b,c. (g) Plasma glucose in response to CNO and pretreatment with antagonists. n=8 mice. Two-way RM ANOVA (Sidak’s multiple comparison’s test). Time (p<0.0001), Treatment (p=0.03), and Interaction (p=0.0002). (h) Plasma glucagon at 30 min post-CNO (or vehicle) injection. Two-way RM ANOVA with Tukey’s (within treatment) and Sidak’s (between treatments) multiple comparisons. p<0.05=*, ns=not significant. Within treatment, CNO increased glucagon at 30 min versus 0 min (p=0.022). Saline did not (p=0.96). Between treatments, CNO increased glucagon at 30 min versus saline (p=0.001). n=6 mice. (i) In vivo fiber photometry measurements of population GCaMP6 activity in pituitary-projecting SON AVP neurons during A1/C1 neuron inhibition. AAV-DIO-GCaMP6s was injected into the SON and AAV-fDIO-hM4Di-mCherry into the VLM of Avpires-Cre/+::Dbhflp/+ mice. GCaMP6s was then imaged in response to an insulin tolerance test (ITT), following inhibition of the A1/C1 neuron (with CNO at 1 mg/kg), as indicated by the protocol in the lower horizontal bar. See Figure 5—figure supplement 2c. (j) Left: Example population activity in one mouse (as described in (i)) in response to an ITT, following saline or CNO treatment (on different trials). CNO strongly inhibited the response to insulin. Right: Average GCaMP6 signal ((F–F0)/F0) during response to insulin with either saline or CNO pretreatment (n=9 mice). CNO reduces the AVP GCaMP6 signal. t-test, p<0.01=**. (k) Plasma glucagon in response to an ITT in mice described in (i). 30 min before the insulin injection, either saline or CNO was given i.p. Glucagon is represented as fold of basal, where basal is 0 min (just prior to insulin) and the sample was taken at 30 min post-insulin. t-test, p=0.023 (*). n=8 mice. AVP, arginine-vasopressin; SON, supraoptic nucleus.
Figure 5—figure supplement 1. Viral tracing of A1/C1 terminals.
Figure 5—figure supplement 2. c-Fos expression in A1/C1 neurons during an ITT.
