Fig 1.
Involvement of the ACx and lPAG in noise-evoked defensive behaviors.
(A) Time line of the behavioral protocol. (B) Schematic showing the sound-evoked escape behavior. (C) Probability of noise-evoked escape behavior (t(6) = 6.874, P = 0.0005, paired t test, n = 7 mice). (D) Schematic showing how running speed is recorded in a head-fixed mouse. (E) Representative running speed traces. Arrowheads denote initiation of noise. (F) Running speed (t(6) = 8.282, P = 0.0002, paired t test, n = 7 mice). (G) c-Fos expression pattern. The silent group was not exposed to noise. Scale bar, 100 μm. (H) Density of c-Fos–positive cells (ACx, t(6) = 2.657, P = 0.0377; dlPAG, t(6) = 4.049, P = 0.0067; lPAG, t(6) = 4.793, P = 0.003. Unpaired t test, n = 4 slices from 3 mice/group). (I, J) Extracellular recordings for neuronal firing in the lPAG in response to 40-dB SPL (left panels) and 80-dB SPL (right panels) noise (I, top panel: raster plots; bottom panel: peristimulus time histograms; inset: spike waveforms) and summarized data of firing rates (J, t(119) = 5.865, P < 0.0001, unpaired t test, n = 60–61 units from 3 mice). The firing rates are normalized to the baseline level for each neuron. Scale bar, 50 μV, 1 millisecond. (K) Schematic showing viral injection and light stimulation. (L) Sample traces of action potentials evoked by 473-nm light (blue bars) recorded from lPAG ChR2-positive neurons in acute midbrain slices. (M, N) Probability of light-evoked escape behavior (M, U = 0, P = 0.0003, Mann-Whitney U test, n = 7 or 8 mice) and time spent in the opposite chamber (N, F(1, 13) = 5.94, P = 0.0299, two-way ANOVA, n = 7 or 8 mice). The underlying data for this figure can be found in S1 Data. Values are means ± SEM (*P < 0.05; **P < 0.01; ***P < 0.001). AAV, adeno-associated virus; ACx, auditory cortex; CaMKII, Ca2+/calmodulin-dependent protein kinase II; ChR2, channelrhodopsin-2; Cre, cyclization recombination; DIO, double-floxed inverted orientation; dlPAG, dorsolateral periaqueductal gray; FR, firing rate; lPAG, lateral periaqueductal gray; SPL, sound pressure level.
Fig 2.
Dissection of the GluACx→GlulPAG pathway.
(A) Schematic showing Cre-dependent retrograde Trans-monosynaptic RV tracing strategy. (B) Typical images of viral expression within the lPAG of CaMKII-Cre (top panels) and Gad2-Cre mice (bottom panels). Starter cells (yellow, arrowheads) co-expressing AAV-DIO-TVA-GFP, AAV-DIO-RVG (green), and RV-EnvA-ΔG-DsRed (red). Scale bars, 100 μm (left), 25 μm (right). (C) DsRed-labeled neurons in the ACx traced from GlulPAG. Scale bar, 100 μm. (D) Quantification of DsRed-labeled ACx neurons (t(6) = 4.21, P = 0.0056, unpaired t test, n = 4 slices from 4 mice/group). (E) DsRed signals co-localized with glutamate immunofluorescence in the ACx. Scale bar, 50 μm. (F) Percentage of DsRed-labeled neurons that contained glutamate in the ACx (n = 6 slices from 4 mice). (G) Schematic showing anterograde Trans-monosynaptic AAV tracing strategy. (H) Typical images of the ACx expressing AAV-Cre-GFP (left) and lPAG expressing AAV-DIO-mCherry (right) in wild-type mice. Scale bar, 100 μm. (I) mCherry signals indicative of Cre recombinase traced from the ACx co-localized with glutamate immunofluorescence in the lPAG. Scale bar, 50 μm. (J) Percentage of mCherry-labeled neurons that contained glutamate in the lPAG (n = 5 slices from 5 mice). (K) Schematic showing anterograde AAV tracing strategy for optogenetics. (L) Typical images of the ACx expressing AAV-DIO-ChR2-mCherry (left panel) and lPAG containing ChR2-expressing fibers from GluACx (right panel). Scale bars, 100 μm. (M) Schematic showing ACx injection of AAV-DIO-ChR2-mCherry in CaMKII-Cre mice and recording configuration in acute midbrain slices. (N) Sample traces of action potentials evoked by 473-nm light (blue bars) recorded from ACx mCherry-positive neurons in acute cortical slices. (O, P) Typical light-evoked EPSCs (O) recorded from lPAG neurons after photostimulation of GluACx terminals in the lPAG before and after bath application of 10 μM DNQX, and the summarized data (P) (t(9) = 9.2479; P < 0.0001, one-sample t test, n = 10 cells). The underlying data for this figure can be found in S1 Data. Values are means ± SEM (**P < 0.01). AAV, adeno-associated virus; ACx, auditory cortex; Aq, aqueduct; CaMKII, Ca2+/calmodulin-dependent protein kinase II; ChR2, channelrhodopsin-2; Cre, cyclization recombination; DIO, double-floxed inverted orientation; DNQX, 6,7-dinitroquinoxaline-2,3(1H,4H)-dione; EnvA, avian sarcoma leucosis virus envelope protein; EPSC, excitatory postsynaptic current; Gad2, glutamic acid decarboxylase 2; GFP, green fluorescent protein; Glu, glutamatergic; lPAG, lateral periaqueductal gray; RV, rabies virus; RVG, rabies virus glycoprotein; TVA, the subgroup A avian leukosis virus receptor.
Fig 3.
Optogenetic activation of the GluACx→GlulPAG pathway evoked defensive-like behaviors.
(A) Schematic showing protocols for viral injection and light stimulation. (B) Time line of optogenetic experiments. (C, D) Probability of light-evoked escape behavior (C, U = 49, P = 0.0006, Mann-Whitney U test, n = 7 mice/group) and time spent in the opposite chamber (D, F(1,12) = 7.22, P = 0.0198, n = 7 mice/group). (E, F) A representative trace (E) and speed of light-evoked running (F, F(1,10) = 54.9, P < 0.0001, n = 6 mice/group). (G, H) Quantification of wall rearing events (G, F(1,9) = 12.68, P = 0.0061, n = 5 or 6 mice) and wall rearing time (H, F(1,9) = 8.04, P = 0.0196, n = 5 or 6 mice) before (pre) and during (light) light stimulation. The underlying data for this figure can be found in S1 Data. Values are means ± SEM (*P < 0.05; **P < 0.01; ***P < 0.001). Two-way ANOVA with Bonferroni post hoc analysis for (D), (F), (G), and (H). AAV, adeno-associated virus; ACx, auditory cortex; CaMKII, Ca2+/calmodulin-dependent protein kinase II; ChR2, channelrhodopsin-2; Cre, cyclization recombination; DIO, double-floxed inverted orientation; Glu, glutamatergic; lPAG, lateral periaqueductal gray.
Fig 4.
Inhibition of the GluACx→GlulPAG pathway reduced noise-evoked defensive behaviors.
(A) Schematic showing protocols for optogenetic experiments. (B) Typical images of the ACx (left panel) expressing AAV-DIO-eNpHR3.0-eYFP and lPAG (right panel) containing eNpHR-expressing fibers from GluACx with a track of an optical fiber (arrowhead). Scale bar, 100 μm. (C) Yellow light (594 nm) hyperpolarized an ACx neuron expressing eNpHR. (D) Time line of optogenetic experiments. (E) Noise-evoked c-Fos expression. Scale bar, 100 μm. (F) Density of c-Fos–positive cells (t(6) = 10.58, P < 0.0001, unpaired t test, n = 4 slices from 4 mice/group). (G, H) Probability of noise-evoked escape behavior (G, F (1, 13) = 10.23, P = 0.007, n = 7 or 8 mice) and time spent in the opposite chamber (H, F (1, 13) = 16.74, P = 0.0013, n = 7 or 8 mice). (I, J) Representative recording traces (I) and summarized data (J, F(1,13) = 8.46, P = 0.0122, n = 7 or 8 mice) of the speed of noise-evoked running in the absence and presence of light stimulation. (K) Schematic for verifying chemogenetic protocols. (L, M) Typical light-evoked EPSCs (L) recorded from lPAG neurons after photostimulation of ACx terminals in the lPAG before and after bath application of 10 μM CNO, and summarized data (M) (t(4) = 5.76; P = 0.0045, paired t test, n = 5 cells). (N) Schematic showing protocols for chemogenetic experiments. (O, P) Probability of noise-evoked escape behavior (O, F(1, 11) = 25.54, P = 0.0004, n = 6 or 7 mice), and time spent in the opposite chamber (P, F(1,11) = 15.06, P = 0.0026, n = 6 or 7 mice). (Q) Summarized data (F(1, 12) = 4.66, P = 0.0518, n = 7 mice/group) of the speed of noise-evoked running in the absence or presence of CNO. The underlying data for this figure can be found in S1 Data. Values are means ± SEM (**P < 0.01; ***P < 0.001). Two-way ANOVA with Bonferroni post hoc analysis for (G), (H), (J), (O), (P), and (Q). AAV, adeno-associated virus; ACSF, artificial cerebrospinal fluid; ACx, auditory cortex; Aq, aqueduct; CaMKII, Ca2+/calmodulin-dependent protein kinase II; ChR2, channelrhodopsin-2; CNO, clozapine-N-oxide; Cre, cyclization recombination; DIO, double-floxed inverted orientation; eNpHR, enhanced natronomonas pharaonis halorhodopsin; EPSC, excitatory postsynaptic current; eYFP, enhanced yellow fluorescent protein; Glu, glutamatergic; hM4Di, human Gi-coupled M4 muscarinic receptor; lPAG, lateral periaqueductal gray.
Fig 5.
The GluACx→GlulPAG pathway mediated noise-evoked defensive behaviors independent of the ICx.
(A) Schematic showing protocols for injections of CTB-555 and CTB-488 into the lPAG and ICx, respectively. (B) A fluorescent image showing tracer injection sites (ICx, green; lPAG, red). Scale bar, 100 μm. (C) Typical images of traced ACx neurons positive for CTB-555 and CTB-488 and a magnified view of the boxed region. Scale bar, 50 μm. (D) Percentage of neurons traced from the ICx (ICx-positive) and lPAG (lPAG-positive) out of CTB-positive neurons in layer V of the ACx. n = 6 slices from 3 mice. (E) Schematic showing protocols for silencing the ICx and the contralateral ACx, and optogenetic stimulation of the ACx terminals in the lPAG. (F) A typical image of the ICx expressing AAV-DIO-hM4Di-mCherry. Scale bar, 100 μm. (G, H) Perfusion of CNO (10 μM) hyperpolarized ICx neurons expressing hM4Di in acute slices (G) and summarized data (H, t(3) = 7.96, P = 0.0041, one-sample t test, n = 4 cells). (I, J) Probability of light-evoked escape behavior (I, U = 0, P = 0.0007, Mann-Whitney U test, n = 6 or 8 mice) and time spent in the opposite chamber (J, F (1, 10) = 14.53, P = 0.0034, n = 6 mice/group). (K, L) A representative trace (K) and summarized data (L, F(1, 10) = 69.57, P < 0.0001, n = 6 mice/group) of the speed of light-evoked running before (pre) and during (light) light simulation. (M, N) Quantification of wall rearing events (M, F(1, 13) = 13.49, P = 0.0028, n = 7 or 8 mice) and wall rearing time (N, F(1, 13) = 11.58, P = 0.004, n = 7 or 8 mice). The underlying data for this figure can be found in S1 Data. Values are means ± SEM (*P < 0.05; **P < 0.01; ***P < 0.001). Two-way ANOVA with Bonferroni post hoc analysis for (J), (L), (M), and (N). AAV, adeno-associated virus; ACx, auditory cortex; CaMKII, Ca2+/calmodulin-dependent protein kinase II; ChR2, channelrhodopsin-2; CIC, central nucleus of the inferior colliculus; CNO, clozapine-N-oxide; Cre, cyclization recombination; CTB, cholera toxin subunit B; DCIC, dorsal cortex of the inferior colliculus; DIO, double-floxed inverted orientation; ECIC, external cortex of the inferior colliculus; Glu, glutamatergic; hM4Di, human Gi-coupled M4 muscarinic receptor; IC, inferior colliculus; ICx, cortex of the inferior colliculus; lPAG, lateral periaqueductal gray.
Fig 6.
The GluACx→GlulPAG pathway mediated noise-evoked defensive behaviors independent of the collateral pathway.
(A) Schematic showing a combinational viral strategy for mapping collateral pathways. (B) Typical images of the lPAG expressing AAV-Retro-Cre-GFP (left panel) and the ACx expressing AAV-DIO-ChR2-mCherry (right panel) in wild-type mice. Scale bar, 100 μm. (C) Typical images of collateral pathways of ACx→lPAG projection. Scale bar, 100 μm. (D) Schematic of viral injection, chemical inactivation, and light stimulation. (E, F) Probability of light-evoked escape behavior (E, U = 0, P = 0.0006, Mann-Whitney U test, n = 7 mice/group), and time spent in the opposite chamber (F, F (1, 12) = 13, P = 0.0036, n = 7 mice/group). (G, H) A representative trace (G) and summarized data (H, F(1, 12) = 32.03, P = 0.0001, n = 7 mice/group) of the speed of light-evoked running before (pre) and during (light) light simulation. The underlying data for this figure can be found in S1 Data. Values are means ± SEM (***P < 0.001). Two-way ANOVA with Bonferroni post hoc analysis for (F) and (H). AAV, adeno-associated virus; ACx, auditory cortex; Aq, aqueduct; BLA, basolateral amygdaloid nucleus; ChR2, channelrhodopsin-2; CIC, central nucleus of the inferior colliculus; Cre, cyclization recombination; DCIC, dorsal cortex of the inferior colliculus; DIO, double-floxed inverted orientation; ECIC, external cortex of the inferior colliculus; GFP, green fluorescent protein; Glu, glutamatergic; IC, inferior colliculus; LA, lateral amygdaloid nucleus; lPAG, lateral periaqueductal gray; SC, superior colliculus; SC-dg, deep gray layer of SC; SC-ig, intermediate gray layer of SC; SC-sg, superficial gray layer of SC.
Fig 7.
Escape mediated by the mSC→dlPAG pathway.
(A) Schematic showing protocols for viral injection and light stimulation. (B) Typical images of the mSC (left panel) expressing AAV-DIO-ChR2-mCherry and dlPAG (right panel) containing mCherry-expressing fibers from mSC with a track of an optical fiber. Scale bar, 100 μm. (C) Schematic showing ACx injection of AAV-DIO-ChR2-mCherry in CaMKII-Cre mice and recording configuration in acute midbrain slices. (D) Sample traces of changes in membrane voltage in response to step current injections (−280, 0, 280 pA) in an mCherry-positive mSC neuron. (E) Sample traces of action potentials evoked by 473-nm light (blue bars) recorded from mSC mCherry-positive neurons. (F, G) Typical light-evoked EPSCs (F) recorded from dlPAG neurons after photostimulation of mSC terminals in the dlPAG before and after bath application of 10 μM DNQX, and summarized data (G) (t(11) = 9.29; P < 0.001, one-sample t test, n = 12 cells). (H) Probability of light-evoked escape behavior by stimulating mSC→dlPAG and ACx→lPAG pathways, respectively (t(12) = 1.188, P = 0.2577, unpaired t test, n = 7 mice/group). (I–K) Representative traces (I) and summarized data of the maximum speed (J, t(12) = 3.085, P = 0.0094, unpaired t test,) and peak latency (K, t(12) = 2.514, P = 0.0272, unpaired t test, n = 7 mice/group) of light-evoked running. The underlying data for this figure can be found in S1 Data. Values are means ± SEM (*P < 0.05; **P < 0.01; ***P < 0.001; n.s., not significant). AAV, adeno-associated virus; ACx, auditory cortex; Aq, aqueduct; CaMKII, Ca2+/calmodulin-dependent protein kinase II; ChR2, channelrhodopsin-2; Cre, cyclization recombination; DIO, double-floxed inverted orientation; dlPAG, dorsolateral periaqueductal gray; DNQX, 6,7-dinitroquinoxaline-2,3(1H,4H)-dione; Glu, glutamatergic; EPSC, excitatory postsynaptic current; lPAG, lateral periaqueductal gray; mSC, medial superior colliculus.
Fig 8.
Proposed neural networks involved in noise-evoked defensive behaviors.
Frightening sound activates GluACx neurons, which send long-range excitatory projection onto GlulPAG neurons, the IC, and the SC. The activation of the lPAG (direct projection from ACx) or dlPAG (indirect projection from ACx with an intermediate of ICx or SC) leads to the generation of defensive behaviors. The possibility of the involvement of DNLL→ICx and IC→SC pathways also exists. ACx, auditory cortex; CIC, central nucleus of the inferior colliculus; dlPAG, dorsolateral periaqueductal gray; DNLL, the dorsal nucleus of the lateral lemniscus; Glu, glutamatergic; IC, inferior colliculus; ICx, dorsal and external cortex of the inferior colliculus; lPAG, lateral periaqueductal gray; MGB, medial geniculate body; SC, superior colliculus.