Hypothalamic Circuits for Predation and Evasion

•PAG-projecting LH neurons are activated during predatory attack•Activating PAG-projecting LH GABA neurons rapidly drives predatory attack•Inhibiting PAG-projecting LH GABA neurons reversibly blocks predatory attack•PAG-projecting LH glutamate neurons control predictive evasion The interactions between predator and prey represent some of the most dramatic events in nature and constitute a matter of life and death for both sides. The hypothalamus has been implicated in driving predation and evasion; however, the exact hypothalamic neural circuits underlying these behaviors remain poorly defined. Here, we demonstrate that inhibitory and excitatory projections from the mouse lateral hypothalamus (LH) to the periaqueductal gray (PAG) in the midbrain drive, respectively, predation and evasion. LH GABA neurons were activated during predation. Optogenetically stimulating PAG-projecting LH GABA neurons drove strong predatory attack, and inhibiting these cells reversibly blocked predation. In contrast, LH glutamate neurons were activated during evasion. Stimulating PAG-projecting LH glutamate neurons drove evasion and inhibiting them impeded predictive evasion. Therefore, the seemingly opposite behaviors of predation and evasion are tightly regulated by two dissociable modular command systems within a single neural projection from the LH to the PAG.Video AbstracteyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiI3NzFiNGJkMTI3MWFmMzBhOWI4ODc4OTFmNmY5MjJiMyIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNjc4NTE3MjQxfQ.QMpcTN2Y_QKnP1vutae42TD7uisLoI0eE6xUxcCoQeDzFXvSGr4hBhG3gATau6c0hTRgX_U4gGGUw8OuWTEVHjbqH-hT9cToGt9xpbvfik3OH469Fe7MTjKdbYQnyHVGARUfWV0_WMG66_1cxlcR8Cf8lJyIPFcbcLq1eOcUS-0nLRWu6zjTJFTP5guGjrIjUaaPJP8W7eFZWuFDMnlssqDITRzbE25OgjjuocuaaQDLVgTupOjqoyOFlaoKo-6K97uhsd99RDkD1r9gyM91UYddklLMPmUQE19Fog7M6hyFhCiTK1qhmseBKqcJaI_gXl5cr7AuWVcymEZ96k6mSg(mp4, (18.26 MB) Download video The interactions between predator and prey represent some of the most dramatic events in nature and constitute a matter of life and death for both sides. The hypothalamus has been implicated in driving predation and evasion; however, the exact hypothalamic neural circuits underlying these behaviors remain poorly defined. Here, we demonstrate that inhibitory and excitatory projections from the mouse lateral hypothalamus (LH) to the periaqueductal gray (PAG) in the midbrain drive, respectively, predation and evasion. LH GABA neurons were activated during predation. Optogenetically stimulating PAG-projecting LH GABA neurons drove strong predatory attack, and inhibiting these cells reversibly blocked predation. In contrast, LH glutamate neurons were activated during evasion. Stimulating PAG-projecting LH glutamate neurons drove evasion and inhibiting them impeded predictive evasion. Therefore, the seemingly opposite behaviors of predation and evasion are tightly regulated by two dissociable modular command systems within a single neural projection from the LH to the PAG. Animals in nature typically lack the luxury of eating readily available food at their pleasure. Before feeding, many predators perform a predatory action sequence that includes searching, pursuing, attacking, and consuming (Comoli et al., 2005Comoli E. Ribeiro-Barbosa E.R. Negrão N. Goto M. Canteras N.S. Functional mapping of the prosencephalic systems involved in organizing predatory behavior in rats.Neuroscience. 2005; 130: 1055-1067Crossref PubMed Scopus (49) Google Scholar). Conversely, the prey seeks to evade the predator. Although numerous studies have provided insights into the neural mechanisms underlying feeding behaviors (Morton et al., 2006Morton G.J. Cummings D.E. Baskin D.G. Barsh G.S. Schwartz M.W. Central nervous system control of food intake and body weight.Nature. 2006; 443: 289-295Crossref PubMed Scopus (1889) Google Scholar, Sternson, 2013Sternson S.M. Hypothalamic survival circuits: Blueprints for purposive behaviors.Neuron. 2013; 77: 810-824Abstract Full Text Full Text PDF PubMed Scopus (192) Google Scholar, Stuber and Wise, 2016Stuber G.D. Wise R.A. Lateral hypothalamic circuits for feeding and reward.Nat. Neurosci. 2016; 19: 198-205Crossref PubMed Scopus (275) Google Scholar), it has been challenging to define the neural circuits for predation. Classical electrical stimulations revealed various hotspots for “quiet biting attack,” including the thalamus, hypothalamus, cerebellum, periaqueductal gray (PAG), and the ventral tegmental area (VTA) (Bandler et al., 1972Bandler Jr., R.J. Chi C.C. Flynn J.P. Biting attack elicited by stimulation of the ventral midbrain tegmentum of cats.Science. 1972; 177: 364-366Crossref PubMed Scopus (60) Google Scholar, Brown et al., 1969Brown J.L. Hunsperger R.W. Rosvold H.E. Defence, attack, and flight elicited by electrical stimulation of the hypothalamus of the cat.Exp. Brain Res. 1969; 8: 113-129PubMed Google Scholar, Chi and Flynn, 1971Chi C.C. Flynn J.P. Neuroanatomic projections related to biting attack elicited from hypothalamus in cats.Brain Res. 1971; 35: 49-66Crossref PubMed Scopus (95) Google Scholar, Jansen et al., 1995Jansen A.S. Nguyen X.V. Karpitskiy V. Mettenleiter T.C. Loewy A.D. Central command neurons of the sympathetic nervous system: Basis of the fight-or-flight response.Science. 1995; 270: 644-646Crossref PubMed Scopus (546) Google Scholar, Siegel and Pott, 1988Siegel A. Pott C.B. Neural substrates of aggression and flight in the cat.Prog. Neurobiol. 1988; 31: 261-283Crossref PubMed Scopus (103) Google Scholar). Viewed together, these early studies suggested that there may be a “predatory attack center” in the brain that coordinates how these multiple brain areas drive predation behavior. The lateral hypothalamus (LH) is a possible candidate for this hypothetical predatory attack center. It connects the ventral tegmental area and the PAG via the medial forebrain bundle (Berthoud and Münzberg, 2011Berthoud H.-R. Münzberg H. The lateral hypothalamus as integrator of metabolic and environmental needs: From electrical self-stimulation to opto-genetics.Physiol. Behav. 2011; 104: 29-39Crossref PubMed Scopus (164) Google Scholar, Nieh et al., 2016Nieh E.H. Vander Weele C.M. Matthews G.A. Presbrey K.N. Wichmann R. Leppla C.A. Izadmehr E.M. Tye K.M. Inhibitory input from the lateral hypothalamus to the ventral tegmental area disinhibits dopamine neurons and promotes behavioral activation.Neuron. 2016; 90: 1286-1298Abstract Full Text Full Text PDF PubMed Scopus (204) Google Scholar, Siegel et al., 1999Siegel A. Roeling T.A. Gregg T.R. Kruk M.R. Neuropharmacology of brain-stimulation-evoked aggression.Neurosci. Biobehav. Rev. 1999; 23: 359-389Crossref PubMed Scopus (297) Google Scholar). The LH receives axon projections from multiple nuclei, including the feeding-related arcuate nucleus, the periventricular hypothalamic nucleus, and the recently identified predatory hotspot in the central amygdala (CeA) that is involved in processing sensorimotor information during predation (Berk and Finkelstein, 1981Berk M.L. Finkelstein J.A. Afferent projections to the preoptic area and hypothalamic regions in the rat brain.Neuroscience. 1981; 6: 1601-1624Crossref PubMed Scopus (266) Google Scholar, Berthoud and Münzberg, 2011Berthoud H.-R. Münzberg H. The lateral hypothalamus as integrator of metabolic and environmental needs: From electrical self-stimulation to opto-genetics.Physiol. Behav. 2011; 104: 29-39Crossref PubMed Scopus (164) Google Scholar, Han et al., 2017Han W. Tellez L.A. Rangel Jr., M.J. Motta S.C. Zhang X. Perez I.O. Canteras N.S. Shammah-Lagnado S.J. van den Pol A.N. de Araujo I.E. Integrated control of predatory hunting by the central nucleus of the amygdala.Cell. 2017; 168: 311-324Abstract Full Text Full Text PDF PubMed Scopus (138) Google Scholar, Reppucci and Petrovich, 2016Reppucci C.J. Petrovich G.D. Organization of connections between the amygdala, medial prefrontal cortex, and lateral hypothalamus: A single and double retrograde tracing study in rats.Brain Struct. Funct. 2016; 221: 2937-2962Crossref PubMed Scopus (90) Google Scholar). However, there is controversy about whether or not LH neurons do function in predation behaviors (Comoli et al., 2005Comoli E. Ribeiro-Barbosa E.R. Negrão N. Goto M. Canteras N.S. Functional mapping of the prosencephalic systems involved in organizing predatory behavior in rats.Neuroscience. 2005; 130: 1055-1067Crossref PubMed Scopus (49) Google Scholar). The LH has traditionally been considered to be important for promoting reward seeking, feeding, and arousal (Adamantidis et al., 2007Adamantidis A.R. Zhang F. Aravanis A.M. Deisseroth K. de Lecea L. Neural substrates of awakening probed with optogenetic control of hypocretin neurons.Nature. 2007; 450: 420-424Crossref PubMed Scopus (952) Google Scholar, Harris et al., 2005Harris G.C. Wimmer M. Aston-Jones G. A role for lateral hypothalamic orexin neurons in reward seeking.Nature. 2005; 437: 556-559Crossref PubMed Scopus (1043) Google Scholar, Jennings et al., 2013Jennings J.H. Rizzi G. Stamatakis A.M. Ung R.L. Stuber G.D. The inhibitory circuit architecture of the lateral hypothalamus orchestrates feeding.Science. 2013; 341: 1517-1521Crossref PubMed Scopus (314) Google Scholar, Jennings et al., 2015Jennings J.H. Ung R.L. Resendez S.L. Stamatakis A.M. Taylor J.G. Huang J. Veleta K. Kantak P.A. Aita M. Shilling-Scrivo K. et al.Visualizing hypothalamic network dynamics for appetitive and consummatory behaviors.Cell. 2015; 160: 516-527Abstract Full Text Full Text PDF PubMed Scopus (331) Google Scholar, Nieh et al., 2015Nieh E.H. Matthews G.A. Allsop S.A. Presbrey K.N. Leppla C.A. Wichmann R. Neve R. Wildes C.P. Tye K.M. Decoding neural circuits that control compulsive sucrose seeking.Cell. 2015; 160: 528-541Abstract Full Text Full Text PDF PubMed Scopus (240) Google Scholar, O’Connor et al., 2015O’Connor E.C. Kremer Y. Lefort S. Harada M. Pascoli V. Rohner C. Lüscher C. Accumbal D1R neurons projecting to lateral hypothalamus authorize feeding.Neuron. 2015; 88: 553-564Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar, Sakurai, 2007Sakurai T. The neural circuit of orexin (hypocretin): Maintaining sleep and wakefulness.Nat. Rev. Neurosci. 2007; 8: 171-181Crossref PubMed Scopus (927) Google Scholar, Stuber and Wise, 2016Stuber G.D. Wise R.A. Lateral hypothalamic circuits for feeding and reward.Nat. Neurosci. 2016; 19: 198-205Crossref PubMed Scopus (275) Google Scholar, Sutcliffe and de Lecea, 2002Sutcliffe J.G. de Lecea L. The hypocretins: Setting the arousal threshold.Nat. Rev. Neurosci. 2002; 3: 339-349Crossref PubMed Scopus (376) Google Scholar). Because electrical microstimulation can activate local neuron somata as well as the pass-by axons of neurons from other brain regions, the precise LH neural circuits underlying predation remain to be dissected precisely. Hypothalamic regions, especially those in the medial aspect, are also involved in evasion behavior (Lammers et al., 1988Lammers J.H. Kruk M.R. Meelis W. van der Poel A.M. Hypothalamic substrates for brain stimulation-induced patterns of locomotion and escape jumps in the rat.Brain Res. 1988; 449: 294-310Crossref PubMed Scopus (99) Google Scholar, Siegel and Pott, 1988Siegel A. Pott C.B. Neural substrates of aggression and flight in the cat.Prog. Neurobiol. 1988; 31: 261-283Crossref PubMed Scopus (103) Google Scholar), although the role of the LH in evasion remains largely unknown. The LH contains many genetically distinct cell populations (Berthoud and Münzberg, 2011Berthoud H.-R. Münzberg H. The lateral hypothalamus as integrator of metabolic and environmental needs: From electrical self-stimulation to opto-genetics.Physiol. Behav. 2011; 104: 29-39Crossref PubMed Scopus (164) Google Scholar, Nieh et al., 2016Nieh E.H. Vander Weele C.M. Matthews G.A. Presbrey K.N. Wichmann R. Leppla C.A. Izadmehr E.M. Tye K.M. Inhibitory input from the lateral hypothalamus to the ventral tegmental area disinhibits dopamine neurons and promotes behavioral activation.Neuron. 2016; 90: 1286-1298Abstract Full Text Full Text PDF PubMed Scopus (204) Google Scholar, Stuber and Wise, 2016Stuber G.D. Wise R.A. Lateral hypothalamic circuits for feeding and reward.Nat. Neurosci. 2016; 19: 198-205Crossref PubMed Scopus (275) Google Scholar), and recent advances in neural circuit probing technologies now enable examination of the discrete functions of particular hypothalamic circuits. Using both projection-specific and cell-type-specific optogenetic manipulation, as well as fiber photometry recording, we here show that PAG-projecting LH GABA neurons drive predatory attack upon natural and artificial prey and demonstrate that inhibiting these neurons blocks predatory behavior. Furthermore, we show that PAG-projecting LH glutamate neurons control evasion behavior and demonstrate that inhibiting these neurons impairs the ability to escape from danger. Using a dual-virus strategy that selectively expresses genetically encoded calcium indicator GCaMP6m in a projection-specific manner (Chen et al., 2013Chen T.W. Wardill T.J. Sun Y. Pulver S.R. Renninger S.L. Baohan A. Schreiter E.R. Kerr R.A. Orger M.B. Jayaraman V. et al.Ultrasensitive fluorescent proteins for imaging neuronal activity.Nature. 2013; 499: 295-300Crossref PubMed Scopus (3618) Google Scholar, Tervo et al., 2016Tervo D.G. Hwang B.Y. Viswanathan S. Gaj T. Lavzin M. Ritola K.D. Lindo S. Michael S. Kuleshova E. Ojala D. et al.A designer AAV variant permits efficient retrograde access to projection neurons.Neuron. 2016; 92: 372-382Abstract Full Text Full Text PDF PubMed Scopus (622) Google Scholar), we first tested whether or not PAG-projecting LH neurons are activated during predatory attack behavior. We injected a retrograde-transport-orientated adeno-associated viral (AAV) vector carrying Cre-recombinase (AAV-retro-Cre) into the PAG and then injected a Cre-dependent vector AAV-DIO-GCaMP6m into the LH to selectively express GCaMP6m in LH neurons projecting to the PAG (Figure 1A) (Atasoy et al., 2008Atasoy D. Aponte Y. Su H.H. Sternson S.M. A FLEX switch targets Channelrhodopsin-2 to multiple cell types for imaging and long-range circuit mapping.J. Neurosci. 2008; 28: 7025-7030Crossref PubMed Scopus (495) Google Scholar, Tervo et al., 2016Tervo D.G. Hwang B.Y. Viswanathan S. Gaj T. Lavzin M. Ritola K.D. Lindo S. Michael S. Kuleshova E. Ojala D. et al.A designer AAV variant permits efficient retrograde access to projection neurons.Neuron. 2016; 92: 372-382Abstract Full Text Full Text PDF PubMed Scopus (622) Google Scholar, Zhang et al., 2010Zhang F. Gradinaru V. Adamantidis A.R. Durand R. Airan R.D. de Lecea L. Deisseroth K. Optogenetic interrogation of neural circuits: Technology for probing mammalian brain structures.Nat. Protoc. 2010; 5: 439-456Crossref PubMed Scopus (548) Google Scholar). We implanted an optical fiber and used photometry to monitor changes in GCaMP fluorescence in the population of PAG-projecting LH neurons (Figure 1B) (Gunaydin et al., 2014Gunaydin L.A. Grosenick L. Finkelstein J.C. Kauvar I.V. Fenno L.E. Adhikari A. Lammel S. Mirzabekov J.J. Airan R.D. Zalocusky K.A. et al.Natural neural projection dynamics underlying social behavior.Cell. 2014; 157: 1535-1551Abstract Full Text Full Text PDF PubMed Scopus (753) Google Scholar, Li et al., 2016Li Y. Zhong W. Wang D. Feng Q. Liu Z. Zhou J. Jia C. Hu F. Zeng J. Guo Q. et al.Serotonin neurons in the dorsal raphe nucleus encode reward signals.Nat. Commun. 2016; 7: 10503Crossref PubMed Scopus (222) Google Scholar, Zhong et al., 2017Zhong W. Li Y. Feng Q. Luo M. Learning and stress shape the reward response patterns of serotonin neurons.J. Neurosci. 2017; 37: 8863-8875Crossref PubMed Scopus (46) Google Scholar). After we trained food-deprived mice to hunt crickets, they typically completed the searching, pursuit, and attack actions of the sequence in less than 2 s but spent a long time consuming the prey (∼60 s). Aligning the GCaMP signals with video-scored behavioral actions revealed an increase in the activity of PAG-projecting LH neurons starting when starving mice began to hunt crickets (Figure 1C). These results implicated PAG-projecting LH neurons in predatory behavior. To test whether or not activating PAG-projecting LH neurons would induce predatory attack behavior, we expressed ChR2-mCherry in the PAG-projecting LH neurons using the dual-virus strategy (Figures 1D, 1E, and S1A) (Boyden et al., 2005Boyden E.S. Zhang F. Bamberg E. Nagel G. Deisseroth K. Millisecond-timescale, genetically targeted optical control of neural activity.Nat. Neurosci. 2005; 8: 1263-1268Crossref PubMed Scopus (3373) Google Scholar, Tervo et al., 2016Tervo D.G. Hwang B.Y. Viswanathan S. Gaj T. Lavzin M. Ritola K.D. Lindo S. Michael S. Kuleshova E. Ojala D. et al.A designer AAV variant permits efficient retrograde access to projection neurons.Neuron. 2016; 92: 372-382Abstract Full Text Full Text PDF PubMed Scopus (622) Google Scholar). Whole-cell patch-clamp recordings in brain slices confirmed that brief light pulses reliably evoked spike firing from ChR2-mCherry-expressing neurons in the LH (Figure S1B). In order to assay stimulation-evoked predatory behavior, we used cricket-naive mice that had free access to food and water (Figure 1F). Light pulses were delivered through an optical fiber implanted into the LH (Figure S1C). Without optogenetic stimulation, the nonhungry cricket-naive mice rarely initiated predatory attack against crickets. Immediately following optogenetic stimulation, mice initiated predatory attack, sniffing vigorously around the test chamber, rushing toward a cricket and subduing it with their forepaws, and eventually delivering repeated killing bites with a 91.7% probability (Figures 1G–1I; Movie S1). The stimulated mice did not eat up these crickets within the 30-s stimulation period; rather, they attacked a cricket and left most of its dead body in the chamber (Figure 1G). As each new round of optogenetic stimulation was delivered, the stimulated mouse would initiate a new predatory attack sequence. The mean attack latency between stimulation and attack (the time from stimulation onset to mouse contact with a cricket) was less than 10 s (Figure S1D). The predatory attack behavior ceased immediately when the light pulses were stopped (Figure 1H; Movie S1, first part). Thus, PAG-projecting neurons in the LH are sufficient to rapidly orchestrate predatory attack without prior training. The LH neurons labeled with the two-virus strategy projected their axonal terminals in the lateral and ventral lateral region of the PAG (l/vlPAG) (Figure 1J). Photostimulating the axonal terminals in the PAG produced similarly strong predatory behavior with similarly short latency (Figures 1K and S1E). In patch-clamp recordings of brain slices, stimulating the axonal terminals from retrograde-labeled LH neurons evoked a mixture of GABAergic inhibitory currents and glutamatergic excitatory currents in PAG neurons, indicating that these PAG-projecting LH neurons can release the neurotransmitters GABA and/or glutamate (Figures 1L and 1M). Anatomical analysis of the PAG-projecting LH neurons of sacrificed mice revealed that about 61% were putatively GABAergic (vesicular GABA transporter-positive; Vgat+) and that about 33% are glutamatergic (vesicular glutamate transporter 2-positive; Vglut2+; Figures S1F–S1I). PAG-projecting LH neurons also project to other brain regions such as the thalamus and the lateral habenular (LHb), and project along the medial forebrain bundle (MFB) to the midbrain ventral tegmental area and the substantia nigra compacta (SNc) as well as the reticular nucleus posterior to the PAG (Figure S1J). Unilateral activating the thalamus projection, the LHb projection, and the reticular projection produced no obvious effects on predatory attack in the cricket-hunting task (Figures S1K, S1L, and S1N). Activating the ventral tegmental area/SNc projection located between the LH and the PAG induced attack behavior similar to what we observed with PAG activation (Figure S1M). However, it is possible that this predatory attack behavior could be caused by some coactivation of PAG-projecting pass-by fibers in the ventral tegmental area/SNc. Our results suggest that PAG-projecting LH neurons may coordinate neuronal activity in multiple brain regions as they orchestrate predatory behavior. Next, we specifically tested whether the activity of LH GABA neurons encodes particular aspects of the predatory attack behavior sequence. To monitor the activity of the inhibitory projections during predatory attack, we targeted GABA neurons by stereotaxic infusion of AAV-DIO-GCaMP6m vectors into the LH of Vgat-Cre mice (Figure 2A). Using fiber photometry, we measured changes in the GCaMP6 fluorescence of LH GABA neurons during the starvation-induced cricket-hunting task (Figure 2A). Aligning the GCaMP signals with video-scored behavioral actions revealed that neuronal activity increased at the moment that animals started hunting crickets; neuronal activity fell back to the baseline level as the mouse consumed the cricket (Figures 2B–2D and S2A–S2D; Movie S2). We did not detect significant changes in fluorescence signals in LH GABA neurons of GFP-expressing control mice during the cricket-hunting experiments, confirming that the GCaMP signals genuinely indicate neuronal activity and are not simply reflecting artifacts of motion (Figures S2E and S2F). In order to more precisely control the timing of predatory attack behaviors, we designed a computer-controlled food-chasing task. We used real-time monitoring and computer-controlled robotic arms to guide the movement of the food dish via remote magnetic force. After a mouse entered the specified trigger zone in an arena for 1 s, a moving dish that contained food pellets (artificial prey) was guided along one of two edges of the chamber (Figure 2E). When a mouse ran after the dish and came within 10 cm of the dish, the dish was stopped so that the mouse could retrieve one small food pellet. After training, mice showed predation-like behavior toward the moving food dish (Figures 2F and S2G). Following triggering of dish entry into the arena (“initiation” stage), animals quickly chased after the moving artificial prey dish (“chasing” stage), typically completing a “virtual predation sequence” within 2 s (Figures 2G, S2H, and S2I). Fiber photometry revealed that the activity of LH GABA neurons ramped up during pursuit and reached peak activity during pellet retrieval (“retrieval” stage) (Figures 2H–2J). It should be noted that our fiber photometry reflects the population-level activity of LH GABA neurons. Previous monitoring at single-cell resolution revealed that LH GABA neurons encode either the appetitive or consummatory aspects of reward-related behaviors (Jennings et al., 2015Jennings J.H. Ung R.L. Resendez S.L. Stamatakis A.M. Taylor J.G. Huang J. Veleta K. Kantak P.A. Aita M. Shilling-Scrivo K. et al.Visualizing hypothalamic network dynamics for appetitive and consummatory behaviors.Cell. 2015; 160: 516-527Abstract Full Text Full Text PDF PubMed Scopus (331) Google Scholar), but it had remained unclear how LH GABA neurons change their activity during hunting. To test whether the endogenous activity of LH GABA neurons is necessary for predation, we used viral-vector mediated, Cre-dependent expression of the Guillardia theta anion channel rhodopsin 1 protein (GtACR1) (Govorunova et al., 2015Govorunova E.G. Sineshchekov O.A. Janz R. Liu X. Spudich J.L. NEUROSCIENCE. Natural light-gated anion channels: A family of microbial rhodopsins for advanced optogenetics.Science. 2015; 349: 647-650Crossref PubMed Scopus (376) Google Scholar) to optically inhibit GABA neurons in the bilateral LH (Figures 3A, 3B, and S3A). In these cricket-hunting experiments, optogenetic inhibition abolished mouse predation on crickets (Figures 3C, 3D, and S3B). In separate food-chasing tasks (Figure S3C), optogenetic inhibition of LH GABA neurons during the initiation stage prevented mice from chasing the moving food dish (Figures 3E–3G and S3D–S3F; Movie S3). However, inhibition applied immediately after pellet retrieval did not disrupt an animal’s gnawing (putatively consummatory) behavior (Figures 3H, S3G, and S3H). Thus, the activity of LH GABA neurons is essential for initiating and actually conducting predatory attack, but not for the consumption of successfully captured food. To determine the role of the LH GABA neurons in predatory attack, we expressed ChR2 in LH GABA neurons by stereotaxic infusion of AAV-DIO-ChR2-mCherry vectors into the LH of Vgat-Cre mice (Figures 4A, 4B, and S4A). Following the expression of ChR2-mCherry in LH GABA neurons, optogenetic stimulation of the ChR2-expressing axonal terminals elicited GABAergic currents in PAG neurons in a brain slice preparation (Figures 4C and 4D). Applying light pulses into the LH rapidly drove strong predatory attack from the ChR2-expressing mice but not from the mGFP-expressing control mice (Figures 4E and S4B; Movie S1, latter part). Stimulation of LH GABA neurons also caused mice to initiate chasing, neck stretching, and biting upon another mouse, albeit with only about half of the attack percentage as for cricket attack (Figures S4C and S4D; Movie S4, first part). The attacks against other mice exhibited typical features of intraspecific predatory attack and differed from defensive attacks, which are characterized by piloerection and back arching (Lin et al., 2011Lin D. Boyle M.P. Dollar P. Lee H. Lein E.S. Perona P. Anderson D.J. Functional identification of an aggression locus in the mouse hypothalamus.Nature. 2011; 470: 221-226Crossref PubMed Scopus (583) Google Scholar, Siegel et al., 1999Siegel A. Roeling T.A. Gregg T.R. Kruk M.R. Neuropharmacology of brain-stimulation-evoked aggression.Neurosci. Biobehav. Rev. 1999; 23: 359-389Crossref PubMed Scopus (297) Google Scholar). Therefore, LH GABA neurons in mice can produce a sufficiently strong motivation to drive attack against both crickets and intraspecific targets. To test whether the promise of obtaining some caloric value is required for the observed light-evoked predatory attack, we designed an artificial prey object that offered no caloric content by embedding a small magnet in a wax disk, which was controlled by remote magnetic force (similar to the food dish, above). To mimic a natural prey item, the artificial prey was moved away as a mouse approached (Figure 4F). Optogenetic stimulation of LH GABA neurons caused the mice to pursue, attack, and voraciously bite the artificial prey disk (Figures 4G, S4E, and S4F; Movie S4, latter part). A substantial amount of wax had been bitten off the disk by the end of the trial sessions (Figure 4G), suggesting that activating LH GABA neurons can compel vigorous chasing and strong biting attack against a moving object without any caloric value. To determine whether LH GABA neurons have the potential of switching from evasion to predation, we conducted experiments in which we turned the wax disk into an attacker by guiding it to repetitively strike a mouse. Mice would turn and run away to escape from these strikes in typical evasion behavior (Figures 4H and S4G). However, immediately following stimulation of LH GABA neurons, a mouse would stop running away from disk strikes, instead turning its body around and starting to attack and bite the disk (Figures 4I–4K and S4G–S4K; Movie S5), demonstrating that stimulating LH GABA neurons can produce a sufficiently strong attack drive to switch behavior from evasion to predation. We carried out several additional experiments to study the role of LH GABA neurons that project specifically to the PAG. First, we carried out optogenetic inhibition specifically in the PAG-projecting LH GABA neurons and their terminals in the PAG during the cricket-hunting task and the food-chasing task. Using the dual-virus strategy, we injected retrogradely transported and Cre-recombinase-dependent AAV-retro-DIO-Flp into the PAG and injected Flp-dependent AAV-fDIO-GtACR1 into the LH of Vgat-Cre mice (Figure 5A). Optogenetic inhibition of both the PAG-projecting LH GABA neurons and their terminals in the PAG significantly suppressed predatory behavior in the cricket-hunting task (Figure 5B). Moreover, optogenetic inhibition of PAG-projecting LH GABA neurons starting from the initiation of food chasing decreased the successful hit rate during the food-chasing task (Figure 5C) but does not affect an animal’s gnawing behavior when such inhibition is started after the animal has begun to consume the food pellet (Figure 5D). These results indicate that the activity of PAG-projecting LH GABA neurons is necessary for predatory behavior. We then characterized the behavioral effect of activating LH GABA neurons that projected to the PAG. We expressed ChR2 in LH GABA neurons by single injection of AAV-DIO-ChR2 into the LH of Vgat-Cre mice. Stimulation of the axonal terminals in the l/vlPAG from LH GABA neurons also produced strong predatory attacks against crickets, intraspecific individuals, and artificial prey (Figures 5E, 5F, and S5A–S5E). We also expressed ChR2 in PAG-projecting LH GABA neurons using the dual-virus strategy of injecting AAV-retro-DIO-Flp vectors into the PAG and Flp-recombinase-dependent AAV-fDIO-ChR2-EYFP vectors into the LH of Vgat-Cre mice (Figure 5G). Optogenetic stimulation of the PAG-projecting LH GABA neurons resulted in attack against crickets, with an attack probability of over 92.5%, and latency of less than 10 s (Figures 5H and S5F–S5I). We observed no attack behavior while stimulating the LH of control mice that had received AAV-fDIO-ChR2-EYFP injection into the LH but no injection of AAV-retro-DIO-Flp into