Presynaptic sensor and silencer of peptidergic transmission reveal neuropeptides as primary transmitters in pontine fear circuit

Abstract

Neurons produce and release neuropeptides to communicate with one another. Despite their importance in brain function, circuit-based mechanisms of peptidergic transmission are poorly understood, primarily due to the lack of tools for monitoring and manipulating neuropeptide release in vivo. Here, we report the development of two genetically encoded tools for investigating peptidergic transmission in behaving mice: a genetically encoded large dense core vesicle (LDCV) sensor that detects presynaptic neuropeptide release and a genetically encoded silencer that specifically degrades neuropeptides inside LDCVs. Using these tools, we show that neuropeptides, not glutamate, encode the unconditioned stimulus in the parabrachial-to-amygdalar threat pathway during Pavlovian threat learning. We also show that neuropeptides play important roles in encoding positive valence and suppressing conditioned threat response in the amygdala-to-parabrachial endogenous opioidergic circuit. These results show that our sensor and silencer for presynaptic peptidergic transmission are reliable tools to investigate neuropeptidergic systems in awake, behaving animals.

Figure 1.. Design and characterization of CybSEP as an LDCV sensor

(A) Schematic of constructs and working principle. (B) Representative images of CybSEP expression in differentiated PC12 cells when perfused with acidic and NH₄Cl solutions, described in (C) (a, bath solution; b and c, acidic solution; d, NH₄Cl; Scale bar, 100 μm). (C) Traces of CybSEP and CybGam fluorescence change during application of various extracellular solutions. (D) Quantification of percent ΔF/F₀ peak intensity in CybSEP expressing PC12 cells (mean ± SEM; n= 9 over 3 experimental replicates; ***p < 0.001, ****p < 0.0001 via one-way ANOVA followed by Tukey’s multiple comparisons). BL indicates baseline. (E and F) Average trace of fluorescence change during 70 mM KCl treatment and quantification of percent ΔF/F₀ peak intensity in (E) (mean ± SEM.; n=8–10 over 2 experimental replicates; **p < 0.01 via two tailed-paired t test to the baseline). (G and H) Average traces of fluorescence change during electrical stimulation and quantification of ΔF/F₀ peak intensity in (G). CybGam response was measured at 100 Hz. For extracellular calcium removal, 5 mM EGTA was used instead of CaCl₂ (mean ± SEM; n=11–19, over 3 experimental replicates; **p < 0.001, ****p < 0.0001 via one-way ANOVA followed by Tukey’s multiple comparisons). (I) Representative images of CybSEP2 with mCherry (mCh) or TetTox-mCh expressed in PC12 cells (Scale bar, 100 μm). (J and K) Averaged fluorescence traces of CybSEP2 co-expressed with mCh or TetTox-mCh during electrical stimulation at 100 Hz and quantification of ΔF/F₀ peak intensity in (J) (mean ± SEM; n=19–23, 3 over 3 experimental replicates; ***p < 0.0001 via two-way ANOVA followed by Šidák multiple comparisons). See also Figure S1.

Figure 2.. Imaging LDCV release in brain slices

(A) Schematic diagram of CybSEP-miniSOG and brain targeting. (B) Representative electron microgram of LDCV-targeted CybSEP-miniSOG in the presynaptic terminal before (upper) and after (bottom) photostimulation. Pink arrow, labeled LDCVs; blue arrow, unlabeled LDCVs (Scale bar, 200 nm). (C) Quantification of labeled LDCVs before and after photostimulation (mean ± SEM; ****p < 0.0001 via Two-tailed unpaired t-test comparisons). (D) Schematic and representative images showing CybSEP2 targeting, projections to the CeAl and expression of CybSEP2 in the PBel and the CeAl (Scale bar, 100 μm). (E and F) Averaged fluorescence traces in response to various electrical stimulation and quantification of the data (E). For extracellular calcium removal, 5 mM EGTA was used instead of CaCl₂. Each trace is the average of 7–9 trials in 24 slice slices prepared from 4 mice (mean ± SEM; ****p < 0.0001 via one-way ANOVA followed by Šidák multiple comparisons). (G and H) Averaged fluorescence trace in response to repeated electrical stimulation at 100 Hz and quantification of ΔF/F₀ peak intensity in (G). Each trial was measured at 5 min interval between trials (mean ± SEM; n=6; ns, not significant via one-way ANOVA followed by Tukey’s multiple comparisons). (I and J) Averaged fluorescence traces in response to electrical stimulation at various frequency or plus numbers and quantification ΔF/F₀ peak intensity in (I) (mean ± SEM; n=4–5 slices for 3 mice). (K and L) Schematic for viral injection of CybSEP2 and TetTox-mCherry in the PBel of Calca^Cre/+ mice (Scale bar, 100 μm). (M and N) The trace of fluorescence changes in CybSEP2 with mCherry (n=10 slices from 3 mice) or TetTox-mCherry (n=10 slices from 3 mice) expressing neurons during electrical stimulation at 100 Hz and quantification of data in (M) (mean ± SEM; ***p < 0.0001 via two-way ANOVA followed by Šidák multiple comparisons). See also Figure S2.

Figure 3.. Monitoring LDCV release from synaptic terminals in behaving mice

(A) Schematic illustration of fiber photometry system used for CybSEP2 recording. (B) Histology of CybGam and CybSEP2 in the PBL and the CeA. Yellow dotted line represents the location of the optic fiber (Scale bar, 100 μm). (C) Heat map and averaged fluorescence traces elicited by footshocks (0.1, 0.3, and 0.5 mA) in CybGam (16 traces from 4 mice) and CybSEP2 (18 traces from 5 mice) expressing mice. (D) Quantification of data in (C) by area under curve (AUC) for 0–10 s (mean ± SEM; *p<0.05, **p < 0.001, ****p < 0.0001via two-way ANOVA followed by Tukey’s multiple comparisons). (E) Heat map and averaged fluorescence traces during thermal stimulus in CybGam (16 traces from 4 mice) and CybSEP2 (21 traces from 5 mice) expressing mice at 42 or 52°C hot plate with a cutoff time of 20 s. (F) Quantification of data in (E) by AUC for 0–10 s (mean ± SEM; ****p < 0.0001 via two-way ANOVA followed by Tukey’s multiple comparisons). (G) Heat map and averaged fluorescence traces during aversive taste stimulus in CybGam (16 traces from 4 mice) and CybSEP2 (17 traces from 5 mice) expressing mice elicited by 0.5 mM quinine solution or water. (H) Quantification of data in (G) by AUC for 0–10 s (mean ± SEM; ***p < 0.001 via two-way ANOVA followed by Tukey’s multiple comparisons to the CybGam or CybSEP2). (I) Averaged fluorescence traces elicited by pinching at 1- and 2- minute. Doted lines indicate time when pinch was given. (J) Quantification of data in (I) by AUC for 60–70 s and 120–130 s (mean ± SEM; *p<0.05 via two-tailed paired t-test comparisons). (K) Averaged fluorescence traces elicited by pinching at 1- and 3- minute. (L) Quantification of data in (K) by AUC for 60–70 s and 180–190 s (mean ± SEM; ns via two-tailed paired t-test comparisons). (M) Averaged fluorescence traces elicited by pinching at 1- and 4- minute. (N) Quantification of AUC in (M) between 60–70 s and 240–250 s (mean ± SEM; ns via two-tailed paired t-test comparisons). See also Figure S3.

Figure 4.. NEP_LDCV lowers neuropeptide release and attenuates threat learning

(A) Schematic illustrating working principle of LDCV targeted NEP. (B) Schematic of injection and representative electron microgram of LDCV-targeted NEP-miniSOG in the presynaptic terminal. Pink arrow, labeled LDCVs (scale bar, 200 nm). (C) Schematic of injection of NEP_LDCV and mCherry into the PBel of Calca^Cre/+ mice. (D) Representative images showing mCherry or mRuby3 co-labeling with CGRP (green) (Scale bar, 100 μm). (E) Mean fluorescence intensity of CGRP in mCherry or NEP expressing neurons (mean ± SEM; n=12–13 sections from three mice, ****p < 0.0001 via two-tailed unpaired t-test comparisons). (F) Schematic of viral injection and patch clamp recordings. (G) Example traces (Top, ChR2 + mCherry slice. Bottom, ChR2 + NEP_LDCV slice, 10 mW, 40 Hz, 5 ms pulse width). (H) Amplitude of oEPSCs in the CeA neurons elicited by photostimulation of ChR2-expressing CGRP^PBel axonal terminals (mean ± SEM; n=12 for mCherry, n=9 for NEP_LDCV; ns via two-tailed unpaired t-test comparisons). (I) Example traces of oEPSP (Top, ChR2 + mCherry slice. Bottom, ChR2 + NEP_LDCV slice). Photostimulation (10 mw, 40 hz, 5 ms pulse width). Onset is indicated by blue line. (J) Sustained oEPSP amplitude (right) after 40-Hz photostimulation (mean ± SEM; n=11 for mCherry, n=7 for NEP_LDCV; ns, *p < 0.05 via two-tailed unpaired t-test comparisons). (K) Initial membrane potential change (mean ± SEM; n=11 for mCherry, n=7 for NEP_LDCV; ns via two-tailed unpaired t-test comparisons). (L) Schematic of auditory threat conditioning experiment. (M) Freezing during threat conditioning in mice expressing mCherry (n = 8) and NEP_LDCV (n = 9) (mean ± SEM; *P<0.05, ***P<0.001 via repeated measures two-way ANOVA with Sidak’s multiple comparisons test). (N) Freezing at the same context 24 hours after threat conditioning (mean ± SEM; **p < 0.01 via two-tailed unpaired t-test comparisons). (O) Freezing to the tone 48 hours after threat conditioning (mean ± SEM; ***P<0.001 via two-tailed unpaired t-test comparisons). See also Figure S4.

Figure 5. Attenuating glutamatergic transmission in the CGRP^PBel neurons does not influence Pavlovian threat conditioning

(A) Schematic of injection into CGRP^PBel neurons. (B) Representative images of in situ hybridization detecting Calca (encodes CGRP) and Slc17a6 (encodes Vglut2) in PBL (Scale bar, 100 μm). (C) Mean fluorescence intensity of control and Slc17a6 against CGRP expression (mean ± SEM; 13 sections from 3 mice, ****p < 0.0001 via two-tailed unpaired t-test comparisons). (D) Example traces of evoked EPSCs in CeA neurons elicited by optogenetic stimulation of axonal terminals of ChR2-expressing CGRP^PBel neurons. Top, EPSC traces of control (sgRosa26) and sgSlc17a6 group. Bottom, EPSC traces with the bath application of TTX and 4-AP or DNQX in sgSlc17a6 group. (E) Amplitudes of EPSCs (mean ± SEM; n = 21 for control, n = 30 for sgSlc17a6; ***p < 0.0001 via two-tailed unpaired t-test comparisons). (F) Example traces of oEPSP (Top, ChR2 + mCherry slice. Bottom, ChR2 + sgRNA slice). Photostimulation (10 mW, 40 hz, 5 ms pulse width) onset is indicated by the blue line. (G and H) membrane potential at 40 hz and initial membrane potential change (mean ± SEM; n=8 for mCherry, n=8 for sgSlc17a6; ***P<0.001 via two-tailed unpaired t-test comparisons). (I) Schematic diagram of auditory threat conditioning experiment. (J) Freezing during threat conditioning in mice expressing sgRosa26 (n = 11) and sgSlc17a6 (n =12). (K and L) Percent of time to spend freezing at the same context 24 hr (K) and to the tone 48 hr (L) after learning in mice expressing sgRosa26 (n = 11) or sgSlc17a6 (n = 12) (mean ± SEM; ns via two-tailed unpaired t-test comparisons). See also Figure S5.

Figure 6. Monitoring endogenous opioid release from ENK^CeA→PBL terminals in behaving mice

(A) Schematic of injection and image showing Cre- and Flp-dependent expression of GCaMP6s in the CeA (Scale bars, 100 μm). (B) Averaged fluorescence trace during an acrophobic condition (8 traces from 4 mice). (C) Averaged fluorescence trace during a footshock (12 traces from 4 mice). (D) Averaged fluorescence trace aligned with onset of freezing during context tests (4 traces from 4 mice for 3 min). (E) Averaged fluorescence traces during cue tests (12 traces from 4 mice for 3 times cue). (F) Quantification of data in (C, D, and E) by AUC for 0–10 s (mean ± SEM; *p < 0.05 via one way ANOVA followed by Tukey’s multiple comparisons). (G) Schematic of injection and images showing expression of CybGam (Scale bar, 100 μm). (H) Averaged fluorescence trace during acrophobic conditions (12 traces from 4 mice). (I) Averaged fluorescence trace during a footshock (12 traces from 4 mice). (J and K) Averaged fluorescence trace during context tests (4 traces from 4 mice) and cue (12 traces from 4 mice for 3 times cue). (L) Quantification of AUC in (I, J, and K) between 0–10 s (mean ± SEM; ns via one way ANOVA followed by Tukey’s multiple comparisons). (M) Schematic of viral injection and images showing expression of CybSEP2 (Scale bar, 100 μm). (N) Averaged fluorescence trace during acrophobic conditions (12 traces from 4 mice). (O) Averaged fluorescence trace during a footshock (12 traces from 4 mice). (P and Q) Averaged fluorescence trace during context tests (4 traces from 4 mice) and cue (12 traces from 4 mice for 3 times cue). (R) Quantification of AUC in (O, P, and Q) between 0–20 s (mean ± SEM; ns, **p < 0.01, ***p < 0.001via two-way ANOVA followed by Tukey’s multiple comparisons). See also Figure S6.

Figure 7. Optogenetic stimulation of ENK^CeA→PBL terminals mediates positive valence and suppresses conditioned threat responses

(A) Schematic of injection and images showing expression of ChR2-eYFP in the CeA and PBL (Scale bar, 100 μm). (B) Representative heatmaps of time spent in RTPP for eYFP (n=6) and ChR2 (n=6) mice at 20 Hz photostimulation (10 mW, 5 ms pulse width) and quantification (mean ± SEM; **p < 0.01, ***p < 0.001 via two-way ANOVA followed by Šidák multiple comparisons). (C) Comparison of preference in ChR2 groups at 5 and 20 Hz (*p < 0.05 via two-tailed paired t test comparison). (D) Schematic of threat conditioning paired with photostimulation during retrievals. (E) Quantification of freezing response during context paired with 5 and 20 Hz photostimulation (mean ± SEM; *p < 0.05 via two-way ANOVA followed by Šidák multiple comparisons). (F) Quantification of freezing response during cue paired with 5 and 20 Hz photostimulation (mean ± SEM; *p < 0.05 via two-way ANOVA followed by Šidák multiple comparisons). (G) Schematic of injection and images showing expression of NEP_LDCV in the CeA (Scale bar: 100 μm). (H and I) Representative heatmaps of time spent on EPM and quantification of time spent in open arms and closed arms in the control (n=6) and NEP_LDCV expressing mice (n=6) (mean ± SEM; *p < 0.05 via two-tailed unpaired t test comparison). (J) Schematic of auditory threat conditioning experiment. (K) Freezing response during threat conditioning in control (n=6) and NEP_LDCV (n=6) expressing mice. (L and M) Freezing response during context and tone induced retrieval (mean ± SEM; ns, *p < 0.05 via two-tailed unpaired t test comparison). (N) Freezing response during extinction learning with 10 tone trials (mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 via two-way ANOVA followed by Šidák multiple comparisons). See also Figure S7.