Advertisement

Reduced Kv3.1 Activity in Dentate Gyrus Parvalbumin Cells Induces Vulnerability to Depression

Open AccessPublished:March 04, 2020DOI:https://doi.org/10.1016/j.biopsych.2020.02.1179

      Abstract

      Background

      Parvalbumin (PV)-expressing interneurons are important for cognitive and emotional behaviors. These neurons express high levels of p11, a protein associated with depression and action of antidepressants.

      Methods

      We characterized the behavioral response to subthreshold stress in mice with conditional deletion of p11 in PV cells. Using chemogenetics, viral-mediated gene delivery, and a specific ion channel agonist, we studied the role of dentate gyrus PV cells in regulating anxiety-like behavior and resilience to stress. We used electrophysiology, imaging, and biochemical studies in mice and cells to elucidate the function and mechanism of p11 in dentate gyrus PV cells.

      Results

      p11 regulates the subcellular localization and cellular level of the potassium channel Kv3.1 in cells. Deletion of p11 from PV cells resulted in reduced hippocampal level of Kv3.1, attenuated capacity of high-frequency firing in dentate gyrus PV cells, and altered short-term plasticity at synapses on granule cells, as well as anxiety-like behavior and a pattern separation deficit. Chemogenetic inhibition or deletion of p11 in these cells induced vulnerability to depressive behavior, whereas upregulation of Kv3.1 in dentate gyrus PV cells or acute activation of Kv3.1 using a specific agonist induced resilience to depression.

      Conclusions

      The activity of dentate gyrus PV cells plays a major role in the behavioral response to novelty and stress. Activation of the Kv3.1 channel in dentate gyrus PV cells may represent a target for the development of cell-type specific, fast-acting antidepressants.

      Keywords

      Along with impairments in glutamatergic signaling, recent imaging and biochemical studies suggest a reciprocal dysfunction of inhibitory signaling systems in the pathophysiology of stress-related neuropsychiatric disorders. These include reduced levels of GABA (gamma-aminobutyric acid) in the blood, cerebrospinal fluid, and cortex of depressed patients (
      • Petty F.
      • Sherman A.D.
      Plasma GABA levels in psychiatric illness.
      ,
      • Romeo B.
      • Choucha W.
      • Fossati P.
      • Rotge J.Y.
      Meta-analysis of central and peripheral gamma-aminobutyric acid levels in patients with unipolar and bipolar depression.
      ), as well as reduced levels of inhibitory neuronal markers including parvalbumin (PV), SNAP-25, and GABA receptor levels in the hippocampus postmortem from subjects with major depressive disorder and subjects with bipolar disorder (
      • Knable M.B.
      • Barci B.M.
      • Webster M.J.
      • Meador-Woodruff J.
      • Torrey E.F.
      • Stanley Neuropathology C.
      Molecular abnormalities of the hippocampus in severe psychiatric illness: postmortem findings from the Stanley Neuropathology Consortium.
      ,
      • Sequeira A.
      • Mamdani F.
      • Ernst C.
      • Vawter M.P.
      • Bunney W.E.
      • Lebel V.
      • et al.
      Global brain gene expression analysis links glutamatergic and GABAergic alterations to suicide and major depression.
      ). Preclinical studies support a link between stress and the dysfunction of the GABAergic signaling in the hippocampus. Early-life stress in rats led to an increase and a decrease in hippocampal glutamate and GABA release, respectively (
      • Martisova E.
      • Solas M.
      • Horrillo I.
      • Ortega J.E.
      • Meana J.J.
      • Tordera R.M.
      • et al.
      Long lasting effects of early-life stress on glutamatergic/GABAergic circuitry in the rat hippocampus.
      ), whereas α2 GABAA receptors in distinct hippocampal microcircuits are required for the anxiety-reducing actions of anxiolytics (
      • Engin E.
      • Benham R.S.
      • Rudolph U.
      An emerging circuit pharmacology of GABAA receptors.
      ).
      Among the inhibitory neuronal populations, the fast-spiking PV-expressing cells play a central role in hippocampal function (
      • Klausberger T.
      • Somogyi P.
      Neuronal diversity and temporal dynamics: The unity of hippocampal circuit operations.
      ). These cells accommodate their firing frequency according to the excitatory input to modulate hippocampal granule and pyramidal cell output via both feedforward and feedback inhibitory modalities (
      • Hu H.
      • Gan J.
      • Jonas P.
      Interneurons. Fast-spiking, parvalbumin(+) GABAergic interneurons: From cellular design to microcircuit function.
      ). Importantly, dysfunction of these cells is associated with neurological and psychiatric disorders, including schizophrenia (
      • Marin O.
      Interneuron dysfunction in psychiatric disorders.
      ,
      • Magloczky Z.
      • Freund T.F.
      Impaired and repaired inhibitory circuits in the epileptic human hippocampus.
      ).
      We recently identified an antidepressant role for p11, a member of the s100a calcium sensors (
      • Svenningsson P.
      • Chergui K.
      • Rachleff I.
      • Flajolet M.
      • Zhang X.
      • El Yacoubi M.
      • et al.
      Alterations in 5-HT1B receptor function by p11 in depression-like states.
      ). p11 in hippocampal cells is essential in mediating the response to antidepressants (
      • Egeland M.
      • Warner-Schmidt J.
      • Greengard P.
      • Svenningsson P.
      Neurogenic effects of fluoxetine are attenuated in p11 (S100A10) knockout mice.
      ). In this brain region, p11 is highly enriched in PV and cholecystokinin GABAergic interneurons (
      • Milosevic A.
      • Liebmann T.
      • Knudsen M.
      • Schintu N.
      • Svenningsson P.
      • Greengard P.
      Cell- and region-specific expression of depression-related protein p11 (S100a10) in the brain.
      ). In cholecystokinin cells, p11 regulates the surface levels of the 5-HT1B serotonergic receptor to initiate the response to antidepressants, an effect mediated by disinhibition of PV cells (
      • Medrihan L.
      • Sagi Y.
      • Inde Z.
      • Krupa O.
      • Daniels C.
      • Peyrache A.
      • et al.
      Initiation of behavioral response to antidepressants by cholecystokinin neurons of the dentate gyrus.
      ). Still, the mechanism by which p11 directly regulates PV cells remains unknown. The objectives of this study were to characterize the impact of p11 in PV neurons on cognitive and emotional behaviors and to identify a mechanism by which p11 regulates the function of PV cells of the dentate gyrus (DG).

      Methods and Materials

      Cell Culture

      For transient transfection studies, 80% confluent N2A cells (1 × 105 cells/well) were transfected with 3 μg of either mouse Kv3.1β or human influenza hemagglutinin-tagged Kv3.1α plasmid, and 3 μg of a plasmid expressing rat p11.

      Animals

      All procedures were approved by the Animal Care and Use Committee of the Rockefeller University. Mice were maintained on the C57BL/6N genetic background and were housed in a 12-hour light/dark interval with food and water ad libitum. p11 conditional knockout (cKO) mice were generated by crossing p11 floxed (
      • Medrihan L.
      • Sagi Y.
      • Inde Z.
      • Krupa O.
      • Daniels C.
      • Peyrache A.
      • et al.
      Initiation of behavioral response to antidepressants by cholecystokinin neurons of the dentate gyrus.
      ,
      • Warner-Schmidt J.L.
      • Flajolet M.
      • Maller A.
      • Chen E.Y.
      • Qi H.
      • Svenningsson P.
      • et al.
      Role of p11 in cellular and behavioral effects of 5-HT4 receptor stimulation.
      ,
      • Warner-Schmidt J.L.
      • Schmidt E.F.
      • Marshall J.J.
      • Rubin A.J.
      • Arango-Lievano M.
      • Kaplitt M.G.
      • et al.
      Cholinergic interneurons in the nucleus accumbens regulate depression-like behavior.
      ,
      • Virk M.S.
      • Sagi Y.
      • Medrihan L.
      • Leung J.
      • Kaplitt M.G.
      • Greengard P.
      Opposing roles for serotonin in cholinergic neurons of the ventral and dorsal striatum.
      ) with Pvalbtm1(cre)Arbr/J mice. Male mice were used for behavioral tests, whereas male and female mice were interchangeably used in all other assays.
      Synthesis of RE1, generation of cell line stably transfected with green fluorescent protein (GFP)-tagged Kv3.1β, biochemical and imaging studies in cell culture, generation and delivery of viral-mediated genes, chemogenetic activation, electrophysiology and optogenetic activation of DG PV cells, behavioral studies, molecular studies, and translating ribosome affinity purification are fully detailed in the Supplement.

      Results

      Deletion of p11 From PV Neurons Induces Susceptibility to Depression and Reduces Their Firing Frequency

      To examine if p11 in PV-expressing neurons plays a role in regulating emotional behaviors, mice with conditional deletion of p11 from PV cells (p11 cKO) were tested. In the open field (OF) test, p11 cKO spent 20% ± 1.7% and 16% ± 1.6% less time in the center of the arena relative to wild-type (WT) or PV-Cre mice, respectively, an indication of anxiety-like behavior (Figure 1A). In the elevated plus maze, another test for anxiety-like behavior, cKO mice spent 83% ± 6.3% less time exploring the open arm relative to WT mice (Figure S2). To test the possibility that p11 in PV cells mediates the adaptation in emotional behavior in response to novel stress, mice were subjected to subthreshold social defeat stress (Figure 1B). In the subsequent social interaction test, all p11 cKO mice manifested social avoidance (Figure 1C). This resulted in a 67% ± 9.0% reduction in the time spent interacting with an unfamiliar aggressor mouse, relative to that by WT mice (Figure 1D). Interestingly, p11 cKO mice did not manifest basal anhedonic or helplessness-like behaviors, as detected by the sucrose preference, the novelty suppressed feeding, the forced swim, and the tail suspension tests (Figure S2). However, the behavioral response to chronic antidepressant treatment was impaired in p11 cKO mice (Figure S2). Notably, recognition of novel objects was attenuated in cKO mice (Figure S2). To study whether p11 in PV cells might play a role in a cognitive behavior associated with the function of the DG, spatial pattern separation was tested. Small changes in object displacement resulted in increased exploration of the object located in unfamiliar positions in WT mice (Figure S2). In contrast, p11 cKO mice showed reduced exploration of the object in the unfamiliar position, an indication of a pattern separation deficit and impairments in the function of the DG.
      Figure thumbnail gr1
      Figure 1Deletion of p11 from PV neurons results in increased susceptibility to depression. (A) Thigmotaxis behavior in the OF in WT (black dots, n = 13 mice), p11 cKO (red dots, n = 13), and PV-Cre (purple dots, n = 7). One-way ANOVA. F2,30 = 13.76, p < .0001. ∗∗p = .005, ∗∗∗p < .001 by post hoc Bonferroni. (B) SSDS included 3 defeat sessions. The SI was conducted the next day. (C) Ratio between the times spent in the IZ in the presence and absence of an unfamiliar aggressor in WT (n = 6) and p11 cKO mice (n = 6). ∗∗p = .007 by unpaired t test. Dashed line represents social avoidance threshold. (D) Time spent in the IZ in the presence of an unfamiliar aggressor. ∗∗p = .001 by unpaired t test. (E) OF thigmotaxis 3 weeks after DIO-DREADD virus injection to the DG in mCherry (n = 8), Gi-DREADD (n = 9), and Gs-DREADD (n = 9) mice. One-way ANOVA. F2,23 = 15.07, p < .0001. ∗∗∗p < .0001 by post hoc Bonferroni. (F) Ratios of time spent in the IZ in mCherry (n = 8), Gi-DREADD (n = 8), and Gs-DREADD (n = 7) treated mice. One-way ANOVA. F2,20 = 6.43, p = .007. ∗∗p = .004 by post hoc Bonferroni. (G) Time spent in the IZ. One-way ANOVA. F2,20 = 9.21, p = .002. ∗p = .028 by post hoc Bonferroni. ANOVA, analysis of variance; cKO, conditional knockout; DG, dentate gyrus; IZ, interaction zone; OF, open field test; PV, parvalbumin; SI, social interaction test; SSDS, subthreshold social defeat stress; WT, wild-type.
      To test if an alteration in the activity of PV cells of the DG might affect emotional behavior, we utilized a chemogenetic approach and transfected Gi- or Gs-DREADD into DG PV cells (Figure S3). In the OF test, a single injection of clozapine N-oxide in Gs-DREADD-injected mice increased the time in the center of the arena by 22% ± 2.9% relative to that in mCherry controls (Figure 1E). In the subthreshold social defeat stress, acute clozapine N-oxide treatment induced social avoidance in 75% of the Gi-DREADD-treated mice, along with a 59% ± 6.5% reduction in the time they interacted with an unfamiliar aggressor mouse, relative to that by mCherry-transfected mice (Figure 1F, G). Taken together, these results suggest that p11 mediates resilience to depression by enabling the increase in the activity of DG PV cells in response to novel psychological stressors.
      To investigate the impact of p11 deletion on the physiology of DG PV neurons, we used patch clamp recordings in acute slices from the DG of WT and p11 cKO mice. The firing frequency of PV neurons from p11 cKO in response to injected current steps (100 pA) was lower than that in WT neurons (Figure 2A, B), with no changes observed between genotypes in either membrane potential value or action potential properties (Figure S4). Kv3.1, a potassium channel highly specific and abundant in PV neurons, has been shown to modulate the firing frequency of these cells (
      • Rudy B.
      • McBain C.J.
      Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing.
      ). The Kv-specific current was markedly reduced in p11 cKO mice in comparison with WT mice (Figure 2C, D). In contrast, the amplitude of the HCN (hyperpolarization-activated cyclic nucleotide–gated) channel, another channel associated with the modulation of cell activity by p11, was not different in cKO versus WT (Figure S5). Interestingly, although the density of the Kv current at 50 mV was reduced by 43% ± 5.4 % in p11 cKO mice (Figure 2E), the conductance of the channel was not different between genotypes (Figure 2F), suggesting that p11 regulates the membrane expression of these channels without interfering with their functional properties.
      Figure thumbnail gr2
      Figure 2Deletion of p11 from PV neurons results in decreased AP frequency and reduced Kv current amplitude. (A) Representative traces of the firing response to a 200-pA step injection of current in PV neurons from WT and cKO mice. (B) The firing frequency of PV neurons from WT (n = 16 neurons/10 mice, written as 16/10) and cKO (17/11) mice in response to 100-pA current steps injections. Two-way ANOVA. Genotype F4,134 = 73.43, p < .0001. ∗p < .05, ∗∗p < .01, ∗∗∗p < .001 by post hoc uncorrected Fisher’s LSD test. (C) Representative traces of Kv potassium currents evoked with 10 mV potential steps from −70 to +50 mV in WT and cKO mice. (D) I–V curves showing the Kv amplitude in PV neurons from WT (16/8) and cKO (15/8) mice in response to increasing 10-mV potential steps. Two-way ANOVA. Genotype F12,377 = 183.7, p < .0001. ∗p < .05, ∗∗p < .01, ∗∗∗p < .001 by post hoc uncorrected Fisher’s LSD test. (E) Histograms showing the maximal density of Kv currents at +50 mV. ∗∗p = .0011 by unpaired t test. (F) The relative conductance of Kv channels in PV neurons from WT (n = 13/7) and cKO (n = 12/7) mice in response to increasing 10-mV potential steps. Two-way ANOVA. Genotype F1,299 = 1.090, p = .2972, post hoc uncorrected Fisher’s LSD test. ANOVA, analysis of variance; AP, action potential; cKO, conditional knockout; LSD, least significant differences; PV, parvalbumin; WT, wild-type.

      Deletion of p11 Reduces Kv3.1 Levels and Abolishes the Capacity of PV Neurons to Adapt to High-Frequency Firing

      We next measured the protein level of Kv3.1 in hippocampal lysates from p11 WT and cKO mice. Western blot analysis confirmed a 41% ± 5.9% reduction in Kv3.1 protein level in p11 cKO mice, supporting the idea that p11 in PV cells modulates Kv3.1 function by regulating the expression level of the channel (Figure 3A, B). Moreover, inspection of PV cells throughout the hippocampus identified a difference in the distribution of Kv labeling in cells from cKO mice (Figure S6). In contrast to its reduction in the hippocampus, the protein level of the channel was not altered in other limbic areas including the nucleus accumbens, prefrontal cortex, or amygdala, suggesting that the regulation of Kv3.1 in PV cells by p11 is specific to the hippocampus (Figure S6).
      Figure thumbnail gr3
      Figure 3p11 regulates hippocampal Kv3.1 levels and mediates adaptation to high-frequency firing in DG PV neurons. (A) Representative immunoblot from hippocampal lysates of WT and cKO mice. Arrows and numbers represent protein weights in kilodaltons. (B) Densitometry of the data presented in (A) in n = 3 animals per group. ∗p = .027, ∗∗p = .006 by unpaired t test. (C) Messenger RNA expression levels of p11 (s100a10) and Kv3.1 isoforms in hippocampal PV cells from WT (n = 7 replicates) and cKO (n = 4) PVTRAP mice. ∗p = .042 by unpaired t test. (D) Representative traces of action potentials evoked by trains of injected currents (1–2 nA, 1 ms) at different frequencies in PV neurons from WT and cKO mice. (E) Success rate (measured as ratio) induced in PV neurons from WT (n = 7/4) and cKO (n = 6/4) mice. Two-way ANOVA. Genotype F10,121 = 20.28, p < .0001. ∗p < .05, ∗∗p < .01, ∗∗∗p < .001 by post hoc Bonferroni. (F) The ratio of the half-amplitude width between the last and the first action potential in the 100-Hz train in PV neurons from WT and cKO mice. ∗p = .013 by unpaired t test. ANOVA, analysis of variance; cKO, conditional knockout; DG, dentate gyrus; PV, parvalbumin; WT, wild-type.
      To study whether p11 in hippocampal PV cells regulates Kv3.1 synthesis, we measured the translated messenger RNA levels of the Kv3.1 α and β isoforms in hippocampal PV cells from WT and cKO mice, using translating ribosome affinity purification. Semiquantitative real-time polymerase chain reaction analysis confirmed that Kv3.1β (the protein product of the Kcnc1a transcript) is 3.3-fold more abundant than Kv3.1α (the product of Kcnc1b, Figure 3C). Importantly, while the translated messenger RNA level of s100a10 (p11) was downregulated in PV cells of cKO mice, no difference was detected in the ribosome-bound levels of either Kv3.1 transcript in hippocampal PV cells between WT and cKO mice, supporting the idea that p11 regulates the channel by inhibiting the degradation of the protein (Figure 3C).
      The most important physiological property of PV neurons is their ability to respond with high-frequency firing to inputs, and this function is dependent on Kv3 channels in these neurons (
      • Rudy B.
      • McBain C.J.
      Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing.
      ). We next measured the capacity of PV cells in WT and cKO mice to respond with action potentials induced by stimuli at increased frequencies (1 nA; 1 ms; 10–200 Hz). We noticed that PV neurons from cKO mice have a decrease in success rate above 100-Hz stimulations when compared with PV neurons from WT mice (Figure 3D, E). The decreased success rate can be explained by a 32% ± 11.1% increase in the width of the action potentials in a train in PV cKO (Figure 3F), which reflects the reduced function of Kv3.1 in these neurons (
      • Rudy B.
      • McBain C.J.
      Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing.
      ,
      • Atzori M.
      • Lau D.
      • Tansey E.P.
      • Chow A.
      • Ozaita A.
      • Rudy B.
      • et al.
      H2 histamine receptor-phosphorylation of Kv3.2 modulates interneuron fast spiking.
      ).

      Presynaptic Reduction of Kv3.1 Disrupts Short-term Plasticity at the PV-Granule Cells Synapse

      Anxiety-like behavior and vulnerability to stress are likely to reflect an impairment in the network of neurons rather than cellular impairment of a single neuronal type. We measured the inhibitory and excitatory synaptic input on PV neurons from WT and cKO mice but we did not notice any difference between genotypes (Figure S7). Several studies have shown that Kv channels are highly expressed in the axon terminal of PV neurons where they inhibit neurotransmitter release and the GABAergic output to the principal cells (
      • Kaczmarek L.K.
      • Zhang Y.
      Kv3 channels: enablers of rapid firing, neurotransmitter release, and neuronal endurance.
      ,
      • Hoppa M.B.
      • Gouzer G.
      • Armbruster M.
      • Ryan T.A.
      Control and plasticity of the presynaptic action potential waveform at small CNS nerve terminals.
      ,
      • Goldberg E.M.
      • Watanabe S.
      • Chang S.Y.
      • Joho R.H.
      • Huang Z.J.
      • Leonard C.S.
      • et al.
      Specific functions of synaptically localized potassium channels in synaptic transmission at the neocortical GABAergic fast-spiking cell synapse.
      ). Therefore, we infected PV neurons in the DG with the excitatory opsin ChETA, and recorded their monosynaptic output on granule cells (GCs) using optical stimulation and whole-cell patch clamp (Figure S8). The amplitude of the postsynaptic GABAergic response in GC evoked by light stimulation of PV neurons increased by 119% ± 35.3% in the p11 cKO mice compared to WT mice (Figure 4A, B). Bath addition of tetraethylammonium chloride (TEA) (1 mM), which blocks Kv channels, led to a 67% ± 18.8% increase in the amplitude of the evoked postsynaptic current in PV-GC synapses from WT mice, but not in synapses from p11 cKO (Figure 4C, D), suggesting that presynaptic Kv channels are not functional in p11 cKO mice. Paired-pulse experiments showed a reduction in the ratio between the second and the first response in p11 cKO mice at short time intervals (Figure 4E, F). The block of Kv channels with 1 mM TEA decreased the paired-pulse ratio of WT mice to p11 cKO levels but had no effect on the paired-pulse ratio in the p11 cKO mice (Figure 4G). Furthermore, a train of stimuli at 40 Hz that depletes the ready-releasable pool at the PV-GC synapses showed a drastic effect in p11 cKO mice, with more than 80% of the GABA vesicles being released after the first stimulation with respect to the ∼ 50% level noticed in WT synapses (Figure 4H, I). The block of Kv channels with 1 mM TEA did not have any effect on the depletion kinetics of p11 cKO mice, but it brought the WT depletion kinetics to p11 cKO levels (Figure 4I). Together, these data strongly suggest that presynaptic Kv channels are not functional in p11 cKO mice, and as a result, inhibition kinetics of GC by PV neurons are heavily altered.
      Figure thumbnail gr4
      Figure 4Reduction of Kv3.1 in DG PV neurons leads to loss of low-frequency synaptic filtering in PV-GC synapses. (A) Representative monosynaptic GABAergic responses (A) evoked in GC neurons from WT and cKO mice by light stimulation of ChETA-AAV-infected PV neurons. (B) Mean amplitude in WT (n = 21/5) and cKO (n = 7/4). ∗∗∗p < .001 by unpaired t test. (C) Amplitude of monosynaptic GABAergic responses evoked by light in GC neurons from WT (n = 3/3) and cKO (n = 3, 3) mice before and after the bath application of 1 mM TEA. Two-way ANOVA. Genotype F1,4 = 8.123, p = .046; treatment F1,4 = 16.88, p = .015. ∗p < .05 by post hoc Bonferroni. (D) Changes in the monosynaptic GABAergic responses evoked by light in GC neurons from WT and cKO mice by different presynaptic modulators like Kv channels (blocked by 1 mM TEA, n = 4/3 for WT and 3/3 for cKO), AMPA receptors (blocked by 10 mM CNQX, n = 4/3 for WT and 3/2 for cKO), and ERBB4 receptors (blocked by 10 μM PD158780, n = 4/2 for WT and 3/2 for cKO). One-way ANOVA. F5,14 = 14.76, p < .0001. ∗∗p = .0049 by post hoc Bonferroni. (E) Representative traces of paired-pulse responses at 100 ms evoked in GC neurons from WT and cKO mice by light stimulation of ChETA-AAV-infected PV neurons. (F) Paired-pulse ratio of GABAergic responses evoked in GC neurons from WT and cKO mice by light stimulation at different interstimulation intervals in WT (n = 5–13/5) and cKO (n = 3–7/4). Two-way ANOVA. Genotype F5,58 = 42.80, p < .0001. ∗∗p < .01 by post hoc Bonferroni. (G) Paired-pulse ratio at 100 ms evoked by light in GC neurons from WT (n = 3/3) and cKO (n = 3/3) mice before and after the bath application of 1 mM TEA. Two-way ANOVA. Genotype F1,4 = 11.27, p = .0284; treatment F1,4 = 84.63, p = .0008. ∗∗∗p < .001 by post hoc Bonferroni. (H) Representative traces of GABAergic responses evoked in GC neurons from WT and cKO mice by light stimulation at 40 Hz of ChETA-AAV-infected PV neurons. (I) Normalized amplitude of GABAergic responses evoked in GC neurons from WT (n = 5/3) and cKO (n = 4/2) mice before and after the bath addition of 1 mM TEA. Two-way ANOVA. Genotype F4,55 = 203.3, p < .0001. treatment F3,55 = 9.398, p < .0001. ∗p < .05, ∗∗∗p < .001 by post hoc Bonferroni. ANOVA, analysis of variance; cKO, conditional knockout; DG, dentate gyrus; GABAergic, gamma-aminobutyric acidergic; GC, granule cell; pp, paired-pulse; PV, parvalbumin; TEA, tetraethylammonium chloride; WT, wild-type.

      p11 Regulates the Protein Level and Intracellular Localization of Kv3.1

      Our studies indicate that p11 regulates the function of Kv3.1 in PV cells by regulating the cellular level of the protein. To study the detailed mechanisms by which p11 regulates Kv3.1, we measured changes in its level in N2A cells following transient cotransfection with either Kv3.1α or -β. Cotransfection with p11 increased the level of Kv3.1β and α proteins by 47% ± 13.0% and 431 ± 91.7% respectively (Figure 5A–C), supporting the idea that p11 enhances the stability of the channel. Next, we stably transfected N2A cells with GFP-tagged Kv3.1β and confirmed the molecular size of the tagged channel (Figure 5D). Downregulation of p11 resulted in 31% ± 9.6% reduction in the protein level of Kv3.1β-GFP (Figure 5E, F). Since p11 has been implicated in regulating the cell surface levels of several ion channels, it seemed likely that the regulation of Kv3.1 protein level by p11 is mediated via intracellular shuttling. To visualize the changes in the intracellular localization of Kv3.1β by p11, we next applied immunocytochemistry and visualized the colocalization of GFP with established markers of the Golgi, endoplasmic reticulum, and plasma membrane in fixed cells (Figure 5G, Figure S9). Importantly, downregulation of p11 resulted in reduced localization of Kv3.1β to the Golgi apparatus by 11% ± 1.6%, while increasing its localization to the cell membrane by 47% ± 4.1% (Figure 5H), supporting a role for endocytic shuttling by p11 in the regulation of Kv3.1.
      Figure thumbnail gr5
      Figure 5p11 regulates Kv3.1 shuttling. (A) Representative immunoblot from N2A cell lysates after 48 hours of transient cotransfection of Kv3.1β with p11 or without (control). Arrows and numbers represent protein weights in kilodaltons. (B) Immunoblot from N2A lysates after 48 hours of cotransfection of KV3.1α with p11 or without (control). (C) Densitometry of the protein levels in n = 5 wells per group. ∗p = .031, ∗∗p = .001 by unpaired t test. (D) Immunoblot from representative lysates of N2A cell lines: untransfected (control) or stably transfected with Kv3.1β-GFP. (E) Cells were transfected with siRNA for p11 or control (siRNA-scr) and lysed after 24 hours. (F) Densitometry of the protein levels in siRNA-p11 (n = 4 wells) and siRNA-scr (n = 4 wells). ∗p = .039, ∗∗∗p < .001 by unpaired t test. (G) Representative immunocytochemical images depicting co-localization between GFP and ATPase 1A1 (left), GM130 (right), and calnexin (bottom) in fixed Kv3.1β-GFP N2A cells transfected with either siRNA-scr or siRNA-p11 for 24 hours. Scale bar = 5 μm. (H) Colocalization ratio between GFP and the organelle markers. Dots represent individual cells and numbers inside the bars indicate the numbers of inspected cells. ∗∗∗p < .001 by unpaired t test. A.U., arbitrary units; GFP, green fluorescent protein; siRNA, small interfering RNA.

      Upregulation of Kv3.1 or Its Chemical Activation Induce Resilience After Chronic Stress

      To test if a direct upregulation of Kv3.1 in DG PV cells might induce anxiolytic response to novelty, we next generated an adeno-associated virus (AAV) bearing a Cre recombinase-dependent Kv3.1β and injected it to the DG of PV-Cre mice (Figure 6A). In the OF, mice overexpression (O/E) of Kv3.1β in DG PV cells spent 24% ± 3.4% more time in the center of the arena, relative to PV-Cre mice that were injected with the GFP control virus (Figure 6B). Interestingly, O/E of Kv3.1β in p11 cKO mice did not improve thigmotaxis (Figure 6B). This was in line with the fact that the O/E increased the protein level of Kv3.1 in the hippocampus of PV-Cre mice but not in that of p11 cKO (Figure S10), further supporting the idea that the stability of the Kv3.1 protein is regulated by p11 in DG PV neurons. We next subjected Kv3.1β O/E mice to 10 days of chronic social defeat stress. None of the Kv3.1β-O/E mice manifested social avoidance in the social interaction test (Figure 6C), and Kv3.1β-O/E mice spent 61% ± 21.5% more time interacting with an unfamiliar mouse, relative to that spent by the GFP-O/E mice (Figure 6D).
      Figure thumbnail gr6
      Figure 6Activation of Kv3.1β induces resilience response. (A) Representative immunohistochemical images showing Kv3.1β immunolabeling in DG PV cells from control mouse or in PV-Cre mouse after O/E of Kv3.1β. Scale bar = 50 μm. (B) OF thigmotaxis after O/E of either GFP or Kv3.1β in DG PV cells, in PV-Cre (control, n = 11 GFP, and n = 9 Kv.3.1β) and p11 cKO (n = 8 GFP and n = 8 Kv.3.1β). Two-way ANOVA. Genotype × AAV F1,32 = 11.21; p = .0021; genotype F = 11.78, p = .017; AAV F = 19.53, p = .001. ∗∗∗p < .001 by post hoc Bonferroni. (C) Ratios of time spent in the IZ after CSDS in PV-Cre mice with O/E of GFP (n = 10) or Kv3.1β (n = 8) in DG PV cells. ∗p= .034 by unpaired t test. (D) Time spent in the IZ. ∗p = .037 by unpaired t test. (E) WT mice were treated for 3 days with RE1 (0–100 μM in 100 μL intraperitoneally) or with vehicle (control, n = 5 mice per group). OF thigmotaxis was determined 30 minutes after the third injection. One-way ANOVA. F6,28 = 4.1, p = .043. ∗p < .05 vs. 0 μM by post hoc Bonferroni. (F–H) Stress sensitive mice (IZ ratio ≤ 1) were identified in SI test (SI1) after CSDS, and were treated for 3 days with vehicle (100 μL intraperitoneally) and for additional 3 days either with RE1 (500 nM, 100 μL intraperitoneally) or Veh. The second SI test (SI2) was conducted 30 minutes after the last injection. (G) Ratios of time spent in the IZ in n = 8 mice per group. ∗p = .049 by unpaired t test. (H) Time spent in the IZ in n = 8 mice per group. ∗p = .046 by unpaired t test. AAV, adeno-associated virus; ANOVA, analysis of variance; cKO, conditional knockout; CSDS, chronic social defeat stress; DG, dentate gyrus; GFP, green fluorescent protein; IZ, interaction zone; O/E, overexpression; OF, open field test; PV, parvalbumin; SI, social interaction test; SSDS, subthreshold social defeat stress; Veh, vehicle; WT, wild-type.
      We then tested the anxiolytic and the antidepressant effects by RE1, an activator of the Kv3.1 channel (
      • Boddum K.
      • Hougaard C.
      • Xiao-Ying Lin J.
      • von Schoubye N.L.
      • Jensen H.S.
      • Grunnet M.
      • et al.
      Kv3.1/Kv3.2 channel positive modulators enable faster activating kinetics and increase firing frequency in fast-spiking GABAergic interneurons.
      ). Patch clamp recordings in acute slices showed that concentrations of 0.5 or 1 μM of RE1 increased the activity of Kv3.1 channels (Figure S11). In the OF test, 3 days of treatment with 0.5 or 1 μM of RE1 resulted in 9% ± 1.5% and 10% ± 3.9% increase in the respective times spent in the center of the arena, relative to the vehicle-treated mice (Figure 6E). Notably, similar treatment with 0.5 μM RE1 did not improve thigmotaxis in p11 cKO (Figure S12). In the social interaction test, 3 days of 0.5 μM RE1 treatment resulted in a 450% ± 202.9% reduction in social avoidance in stress-sensitive mice, relative to that by the vehicle-treated mice (Figure 6F, G), along with 867% ± 431.3% increase in the time they spent interacting with an unfamiliar mouse (Figure 6H), supporting the idea that activation of Kv3.1 in DG PV cells may improve resilience to stress.

      Discussion

      p11 Regulates Diverse Ion Channels in Different Cell Types

      In PV neurons, Kv3.1 is localized to both dendritic spines and axonal terminals, where it respectively mediates cell firing and neurotransmission (
      • Rudy B.
      • McBain C.J.
      Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing.
      ). Notably, both of these functions were impaired in p11 cKO mice. Deletion of p11 resulted in reduced Kv3.1 protein levels in hippocampal lysates, and a similar effect was found in transfected cells following downregulation of p11, supporting the idea that the cellular level of the Kv3.1 channel is the target of p11. Previous studies identified a role for p11 in regulating the function of a variety of ion channels (
      • Okuse K.
      • Malik-Hall M.
      • Baker M.D.
      • Poon W.Y.
      • Kong H.
      • Chao M.V.
      • et al.
      Annexin II light chain regulates sensory neuron-specific sodium channel expression.
      ,
      • Renigunta V.
      • Yuan H.
      • Zuzarte M.
      • Rinne S.
      • Koch A.
      • Wischmeyer E.
      • et al.
      The retention factor p11 confers an endoplasmic reticulum-localization signal to the potassium channel TASK-1.
      ). Very recently we reported that p11 regulates the function of the HCN2 ion channels in cholinergic neurons from the nucleus accumbens (
      • Cheng J.
      • Umschweif G.
      • Leung J.
      • Sagi Y.
      • Greengard P.
      HCN2 channels in cholinergic interneurons of nucleus accumbens shell regulate depressive behaviors.
      ). The effect by p11 on HCN2 function was mediated via transcriptional regulation of its gene expression. In line with this, it was previously suggested that the behavioral response to antidepressant treatment is dependent on p11 via the regulation of transcription in hippocampal cells, an effect that was attributed to the interaction between the protein complex p11/annexin A2 and the chromatin remodeling factor SMARCA3 (
      • Oh Y.S.
      • Gao P.
      • Lee K.W.
      • Ceglia I.
      • Seo J.S.
      • Zhang X.
      • et al.
      SMARCA3, a chromatin-remodeling factor, is required for p11-dependent antidepressant action.
      ). Here we confirmed that p11 regulates the Kv3.1 protein level and that this effect is not transcriptionally mediated. Furthermore, the expression level of the HCN2 channel was not altered in hippocampal PV cells in the absence of p11, as suggested by the unaltered function of the channel in the cKO mice. Together, these studies strongly support the idea that diverse mechanisms by which p11 regulates ion channel function and suggest that the downstream channel targets of p11 are cell-type specific.
      The downregulation of p11 in cultured cells increased the localization of Kv3.1 to the cell membrane but reduced it from the Golgi, supporting the idea that p11 regulates the shuttling of the channel from the cell membrane. Furthermore, the accumulation of the channel in the plasma membrane following transfection suggests that its removal from the membrane was impaired in these cells. We previously showed that deletion of p11 resulted in downregulation of its binding partner, annexin A2 (
      • Oh Y.S.
      • Gao P.
      • Lee K.W.
      • Ceglia I.
      • Seo J.S.
      • Zhang X.
      • et al.
      SMARCA3, a chromatin-remodeling factor, is required for p11-dependent antidepressant action.
      ). Since annexin A2 is involved in endocytic shuttling to the early endosome (
      • Morel E.
      • Gruenberg J.
      The p11/S100A10 light chain of annexin A2 is dispensable for annexin A2 association to endosomes and functions in endosomal transport.
      ), the deletion of p11 could indirectly lead to impairment in Kv shuttling and increased degradation. This is in line with the fact that viral-mediated delivery of Kv to DG PV cells failed to increase the level of the channel in p11 cKO mice. Furthermore, p11 is essential for the trafficking of diverse ion channels, including ASIC1a and NaV1.8, among others (
      • Okuse K.
      • Malik-Hall M.
      • Baker M.D.
      • Poon W.Y.
      • Kong H.
      • Chao M.V.
      • et al.
      Annexin II light chain regulates sensory neuron-specific sodium channel expression.
      ,
      • Donier E.
      • Rugiero F.
      • Okuse K.
      • Wood J.N.
      Annexin II light chain p11 promotes functional expression of acid-sensing ion channel ASIC1a.
      ). It was suggested that association of p11 with ion channels masks their endoplasmic reticulum retention signal (
      • Renigunta V.
      • Yuan H.
      • Zuzarte M.
      • Rinne S.
      • Koch A.
      • Wischmeyer E.
      • et al.
      The retention factor p11 confers an endoplasmic reticulum-localization signal to the potassium channel TASK-1.
      ). Together, our data support the idea that impaired protein shuttling leads to reduced stability of the Kv3.1 ion channel in vivo and that this role of p11 could involve endocytosis. Future studies should determine the exact mechanism by which p11 regulates the cellular localization of Kv3.1 and whether this effect is dependent on interaction with annexin A2 or another binding partner of p11.

      Multiple Roles for DG PV Cells in Mood-Related Behaviors

      p11 in PV cells was essential for mediating resilience and anxiolytic behavior in response to stress and novelty. The behavioral deficits in p11 cKO were mimicked by chemogenetic inhibition of DG PV cells, suggesting that p11 activates PV neurons to mediate emotional modalities of DG functions. Chemogenetic inhibition of DG PV cells mimicked the impairments in anxiolytic response in p11 cKO mice but did not induce a depressive-like behavior. In line with our results, Zou et al. reported that chemogenetic activation of DG PV using Gq-DREADD induced anxiolytic response but did not induce changes in the tail suspension test (
      • Zou D.
      • Chen L.
      • Deng D.
      • Jiang D.
      • Dong F.
      • McSweeney C.
      • et al.
      DREADD in parvalbumin interneurons of the dentate gyrus modulates anxiety, social interaction and memory extinction.
      ). Moreover, the deletion of p11 from PV cells did not induce anhedonia or helplessness behaviors. This is in line with previous reports showing that deficits in reward and motivation-related behaviors are found in mice with constitutive deletion of p11 or in those with conditional deletion in the cholinergic neurons of the nucleus accumbens (
      • Warner-Schmidt J.L.
      • Schmidt E.F.
      • Marshall J.J.
      • Rubin A.J.
      • Arango-Lievano M.
      • Kaplitt M.G.
      • et al.
      Cholinergic interneurons in the nucleus accumbens regulate depression-like behavior.
      ). The idea that differences in mood-related behaviors are regulated by p11 in different cell types and brain regions is supported by the fact that the behavioral response to selective serotonin reuptake inhibitors (SSRIs) is impaired in mice with deletion of p11 from hippocampal cells and cortical cells, but not cholinergic accumbal cells (
      • Egeland M.
      • Warner-Schmidt J.
      • Greengard P.
      • Svenningsson P.
      Neurogenic effects of fluoxetine are attenuated in p11 (S100A10) knockout mice.
      ,
      • Warner-Schmidt J.L.
      • Schmidt E.F.
      • Marshall J.J.
      • Rubin A.J.
      • Arango-Lievano M.
      • Kaplitt M.G.
      • et al.
      Cholinergic interneurons in the nucleus accumbens regulate depression-like behavior.
      ,
      • Schmidt E.F.
      • Warner-Schmidt J.L.
      • Otopalik B.G.
      • Pickett S.B.
      • Greengard P.
      • Heintz N.
      Identification of the cortical neurons that mediate antidepressant responses.
      ). Taken together, the current study identifies a unique role for p11 in DG PV cells in mediating anxiolytic response and resilience for depression, as well as in regulating the behavioral response to chronic antidepressant treatment.

      Opposing Regulation of Kv3.1 Activity Is Required Before and After Chronic SSRIs

      PV interneurons play a critical role in neuronal networks for complex processes, such as learning and memory, cognition, and emotional behavior, whereas disruption of their function has been associated with several mental illnesses and most consistently with epilepsy and schizophrenia (
      • Hu H.
      • Gan J.
      • Jonas P.
      Interneurons. Fast-spiking, parvalbumin(+) GABAergic interneurons: From cellular design to microcircuit function.
      ,
      • Marin O.
      Interneuron dysfunction in psychiatric disorders.
      ). In PV neurons, Kv3 channels provide the rapid repolarization of action potentials during a very brief interspike interval, allowing high-frequency firing rates (
      • Rudy B.
      • McBain C.J.
      Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing.
      ,
      • Hu H.
      • Roth F.C.
      • Vandael D.
      • Jonas P.
      Complementary tuning of Na(+) and K(+) channel gating underlies fast and energy-efficient action potentials in GABAergic interneuron axons.
      ). Moreover, in terminals of presynaptic neurons, Kv channels contribute to neurotransmitter release evoked by a presynaptic action potential (
      • Kaczmarek L.K.
      • Zhang Y.
      Kv3 channels: enablers of rapid firing, neurotransmitter release, and neuronal endurance.
      ,
      • Hoppa M.B.
      • Gouzer G.
      • Armbruster M.
      • Ryan T.A.
      Control and plasticity of the presynaptic action potential waveform at small CNS nerve terminals.
      ,
      • Goldberg E.M.
      • Watanabe S.
      • Chang S.Y.
      • Joho R.H.
      • Huang Z.J.
      • Leonard C.S.
      • et al.
      Specific functions of synaptically localized potassium channels in synaptic transmission at the neocortical GABAergic fast-spiking cell synapse.
      ). We show here that alterations in both of these functions of Kv channels in PV neurons lead to anxiety-like behavior in response to novelty and susceptibility to depressive-like behavior in response to stress. These findings are in line with recent data suggesting that the pathophysiological effects of PV neurons in epilepsy and schizophrenia may be related also to the dysfunction of Kv3.1. For example, a recurrent de novo mutation in Kv3.1 that suppresses the current amplitude when assembled into heteromers with WT Kv3.1 results in progressive myoclonus epilepsy, an inherited disorder that causes tonic-clonic seizures (
      • Muona M.
      • Berkovic S.F.
      • Dibbens L.M.
      • Oliver K.L.
      • Maljevic S.
      • Bayly M.A.
      • et al.
      A recurrent de novo mutation in KCNC1 causes progressive myoclonus epilepsy.
      ). In schizophrenia, the levels of Kv3.1β are significantly reduced in the prefrontal and parietal cortex of untreated patients with schizophrenia and these levels were normalized by antipsychotic medication, suggesting that these agents restore Kv3.1β levels to those in healthy control subjects (
      • Yanagi M.
      • Joho R.H.
      • Southcott S.A.
      • Shukla A.A.
      • Ghose S.
      • Tamminga C.A.
      Kv3.1-containing K(+) channels are reduced in untreated schizophrenia and normalized with antipsychotic drugs.
      ). Moreover, it was recently shown that mice expressing truncated Disc1, which mirrors a high-risk gene for psychiatric disorders including schizophrenia and depression, show depressive-like behavior (
      • Sauer J.F.
      • Struber M.
      • Bartos M.
      Impaired fast-spiking interneuron function in a genetic mouse model of depression.
      ,
      • Shen S.
      • Lang B.
      • Nakamoto C.
      • Zhang F.
      • Pu J.
      • Kuan S.L.
      • et al.
      Schizophrenia-related neural and behavioral phenotypes in transgenic mice expressing truncated Disc1.
      ). This behavioral deficit was correlated with reduced number of PV interneurons, disrupted synaptic input and output, and abnormal limbic network oscillations in the low-gamma range, further supporting a major role for these neurons in pathological emotional behavior (
      • Sauer J.F.
      • Struber M.
      • Bartos M.
      Impaired fast-spiking interneuron function in a genetic mouse model of depression.
      ). Interestingly, in contrast to the downregulation of Kv3.1 in the hippocampus in p11 cKO, we did not identify changes in Kv levels outside the hippocampus in these mice, supporting the idea that Kv3.1 is differentially regulated in different neuronal circuits. This is supported by the fact that while antipsychotic treatment elevates Kv3.1 level in the cortex, chronic antidepressant drug use resulted in reduced activity of this channel in the hippocampus. We recently found that the behavioral response to chronic SSRIs is dependent on the serotonergic 5-HT5A receptor signaling in DG PV cells and that this signaling pathway mediates delayed inhibition of Kv3.1 channel function. This inhibitory effect is detected only after chronic SSRI treatment and was mediated by an upregulation of the phosphorylation level of Ser-503 Kv3.1β (
      • Sagi Y.
      • Medrihan L.
      • George K.
      • Barney M.
      • McCabe K.A.
      • Greengard P.
      Emergence of 5-HT5A signaling in parvalbumin neurons mediates delayed antidepressant action.
      ). Indeed, our results strongly support the idea that the activity of Kv3.1 in DG PV cells is highly modulated during the course of chronic SSRI treatment, with maximal activity during the initiation of the treatment and subsequent reduction that requires delayed activation of the 5-HT5A receptor (
      • Sagi Y.
      • Medrihan L.
      • George K.
      • Barney M.
      • McCabe K.A.
      • Greengard P.
      Emergence of 5-HT5A signaling in parvalbumin neurons mediates delayed antidepressant action.
      ). Two recent findings support the idea that DG PV cell activity is dynamically changed during the course of the SSRI treatment. First, the initial SSRI treatment is associated with increased DG PV cell activity. This induction of activity was found to be mediated by 5-HT1B heteroreceptors on DG cholecystokinin cells, and their activation led to reduced disinhibition of DG PV cells (
      • Oh Y.S.
      • Gao P.
      • Lee K.W.
      • Ceglia I.
      • Seo J.S.
      • Zhang X.
      • et al.
      SMARCA3, a chromatin-remodeling factor, is required for p11-dependent antidepressant action.
      ). Second, we recently showed that activation of DG PV cells after chronic SSRI treatment attenuated the behavioral response to SSRIs (
      • Sagi Y.
      • Medrihan L.
      • George K.
      • Barney M.
      • McCabe K.A.
      • Greengard P.
      Emergence of 5-HT5A signaling in parvalbumin neurons mediates delayed antidepressant action.
      ). Taken together, our studies suggest that changes in DG PV cell activity during chronic SSRI treatment correlates with the changes in the function of the Kv3.1 channel. The opposing regulation of Kv3.1 activity before and after chronic SSRIs is initially mediated by p11, which is required for the initial demand in Kv activity and later by the 5-HT5A receptor and its signaling, which mediate the subsequent downregulation of the channel activity. Our current and previous findings suggest that the regulation of the ionic mechanisms that allow PV neurons to accommodate high-firing frequencies is essential for the DG microcircuit to cope with stress and novelty.

      Acknowledgments and Disclosures

      This work was supported by the Fisher Center for Alzheimer’s Research Foundation, The JPB foundation (PG and SS), The Leon Black Family Foundation (PG), and the United States Army Medical Research Acquisition Activity Grant No. W81XWH-14-0390 (to YS).
      LM, GU, SCS, PG, and YS designed experiments and discussed results. LM performed and analyzed electrophysiology recordings. GU designed, performed, and analyzed social defeat studies. SR performed and analyzed biochemical and imaging studies. JL performed behavioral studies. AS synthetized RE1 with the assistance of KG. YS designed and analyzed the molecular and behavioral studies. LM and YS wrote the manuscript.
      We thank Jodi Gresack for valuable suggestions on behavioral studies, Jerry Cheng for suggestions on imaging studies, Elisabeth Griggs for the artwork, Debra Poulter for proofreading, and Zintis Inde, Chelsea Daniels, and Katia George for technical assistance.
      The authors report no biomedical financial interests or potential conflict of interests.

      Supplementary Material

      References

        • Petty F.
        • Sherman A.D.
        Plasma GABA levels in psychiatric illness.
        J Affect Disord. 1984; 6: 131-138
        • Romeo B.
        • Choucha W.
        • Fossati P.
        • Rotge J.Y.
        Meta-analysis of central and peripheral gamma-aminobutyric acid levels in patients with unipolar and bipolar depression.
        J Psychiatry Neurosci. 2017; 42: 160228
        • Knable M.B.
        • Barci B.M.
        • Webster M.J.
        • Meador-Woodruff J.
        • Torrey E.F.
        • Stanley Neuropathology C.
        Molecular abnormalities of the hippocampus in severe psychiatric illness: postmortem findings from the Stanley Neuropathology Consortium.
        Mol Psychiatry. 2004; 9 (544): 609-620
        • Sequeira A.
        • Mamdani F.
        • Ernst C.
        • Vawter M.P.
        • Bunney W.E.
        • Lebel V.
        • et al.
        Global brain gene expression analysis links glutamatergic and GABAergic alterations to suicide and major depression.
        PLoS One. 2009; 4: e6585
        • Martisova E.
        • Solas M.
        • Horrillo I.
        • Ortega J.E.
        • Meana J.J.
        • Tordera R.M.
        • et al.
        Long lasting effects of early-life stress on glutamatergic/GABAergic circuitry in the rat hippocampus.
        Neuropharmacology. 2012; 62: 1944-1953
        • Engin E.
        • Benham R.S.
        • Rudolph U.
        An emerging circuit pharmacology of GABAA receptors.
        Trends Pharmacol Sci. 2018; 39: 710-732
        • Klausberger T.
        • Somogyi P.
        Neuronal diversity and temporal dynamics: The unity of hippocampal circuit operations.
        Science. 2008; 321: 53-57
        • Hu H.
        • Gan J.
        • Jonas P.
        Interneurons. Fast-spiking, parvalbumin(+) GABAergic interneurons: From cellular design to microcircuit function.
        Science. 2014; 345: 1255263
        • Marin O.
        Interneuron dysfunction in psychiatric disorders.
        Nat Rev Neurosci. 2012; 13: 107-120
        • Magloczky Z.
        • Freund T.F.
        Impaired and repaired inhibitory circuits in the epileptic human hippocampus.
        Trends Neurosci. 2005; 28: 334-340
        • Svenningsson P.
        • Chergui K.
        • Rachleff I.
        • Flajolet M.
        • Zhang X.
        • El Yacoubi M.
        • et al.
        Alterations in 5-HT1B receptor function by p11 in depression-like states.
        Science. 2006; 311: 77-80
        • Egeland M.
        • Warner-Schmidt J.
        • Greengard P.
        • Svenningsson P.
        Neurogenic effects of fluoxetine are attenuated in p11 (S100A10) knockout mice.
        Biol Psychiatry. 2010; 67: 1048-1056
        • Milosevic A.
        • Liebmann T.
        • Knudsen M.
        • Schintu N.
        • Svenningsson P.
        • Greengard P.
        Cell- and region-specific expression of depression-related protein p11 (S100a10) in the brain.
        J Comp Neurol. 2017; 525: 955-975
        • Medrihan L.
        • Sagi Y.
        • Inde Z.
        • Krupa O.
        • Daniels C.
        • Peyrache A.
        • et al.
        Initiation of behavioral response to antidepressants by cholecystokinin neurons of the dentate gyrus.
        Neuron. 2017; 95: 564-576 e564
        • Warner-Schmidt J.L.
        • Flajolet M.
        • Maller A.
        • Chen E.Y.
        • Qi H.
        • Svenningsson P.
        • et al.
        Role of p11 in cellular and behavioral effects of 5-HT4 receptor stimulation.
        J Neurosci. 2009; 29: 1937-1946
        • Warner-Schmidt J.L.
        • Schmidt E.F.
        • Marshall J.J.
        • Rubin A.J.
        • Arango-Lievano M.
        • Kaplitt M.G.
        • et al.
        Cholinergic interneurons in the nucleus accumbens regulate depression-like behavior.
        Proc Natl Acad Sci U S A. 2012; 109: 11360-11365
        • Virk M.S.
        • Sagi Y.
        • Medrihan L.
        • Leung J.
        • Kaplitt M.G.
        • Greengard P.
        Opposing roles for serotonin in cholinergic neurons of the ventral and dorsal striatum.
        Proc Natl Acad Sci U S A. 2016; 113: 734-739
        • Rudy B.
        • McBain C.J.
        Kv3 channels: voltage-gated K+ channels designed for high-frequency repetitive firing.
        Trends Neurosci. 2001; 24: 517-526
        • Atzori M.
        • Lau D.
        • Tansey E.P.
        • Chow A.
        • Ozaita A.
        • Rudy B.
        • et al.
        H2 histamine receptor-phosphorylation of Kv3.2 modulates interneuron fast spiking.
        Nat Neurosci. 2000; 3: 791-798
        • Kaczmarek L.K.
        • Zhang Y.
        Kv3 channels: enablers of rapid firing, neurotransmitter release, and neuronal endurance.
        Physiol Rev. 2017; 97: 1431-1468
        • Hoppa M.B.
        • Gouzer G.
        • Armbruster M.
        • Ryan T.A.
        Control and plasticity of the presynaptic action potential waveform at small CNS nerve terminals.
        Neuron. 2014; 84: 778-789
        • Goldberg E.M.
        • Watanabe S.
        • Chang S.Y.
        • Joho R.H.
        • Huang Z.J.
        • Leonard C.S.
        • et al.
        Specific functions of synaptically localized potassium channels in synaptic transmission at the neocortical GABAergic fast-spiking cell synapse.
        J Neurosci. 2005; 25: 5230-5235
        • Boddum K.
        • Hougaard C.
        • Xiao-Ying Lin J.
        • von Schoubye N.L.
        • Jensen H.S.
        • Grunnet M.
        • et al.
        Kv3.1/Kv3.2 channel positive modulators enable faster activating kinetics and increase firing frequency in fast-spiking GABAergic interneurons.
        Neuropharmacology. 2017; 118: 102-112
        • Okuse K.
        • Malik-Hall M.
        • Baker M.D.
        • Poon W.Y.
        • Kong H.
        • Chao M.V.
        • et al.
        Annexin II light chain regulates sensory neuron-specific sodium channel expression.
        Nature. 2002; 417: 653-656
        • Renigunta V.
        • Yuan H.
        • Zuzarte M.
        • Rinne S.
        • Koch A.
        • Wischmeyer E.
        • et al.
        The retention factor p11 confers an endoplasmic reticulum-localization signal to the potassium channel TASK-1.
        Traffic. 2006; 7: 168-181
        • Cheng J.
        • Umschweif G.
        • Leung J.
        • Sagi Y.
        • Greengard P.
        HCN2 channels in cholinergic interneurons of nucleus accumbens shell regulate depressive behaviors.
        Neuron. 2019; 101: 662-672 e665
        • Oh Y.S.
        • Gao P.
        • Lee K.W.
        • Ceglia I.
        • Seo J.S.
        • Zhang X.
        • et al.
        SMARCA3, a chromatin-remodeling factor, is required for p11-dependent antidepressant action.
        Cell. 2013; 152: 831-843
        • Morel E.
        • Gruenberg J.
        The p11/S100A10 light chain of annexin A2 is dispensable for annexin A2 association to endosomes and functions in endosomal transport.
        PLoS One. 2007; 2: e1118
        • Donier E.
        • Rugiero F.
        • Okuse K.
        • Wood J.N.
        Annexin II light chain p11 promotes functional expression of acid-sensing ion channel ASIC1a.
        J Biol Chem. 2005; 280: 38666-38672
        • Zou D.
        • Chen L.
        • Deng D.
        • Jiang D.
        • Dong F.
        • McSweeney C.
        • et al.
        DREADD in parvalbumin interneurons of the dentate gyrus modulates anxiety, social interaction and memory extinction.
        Curr Mol Med. 2016; 16: 91-102
        • Schmidt E.F.
        • Warner-Schmidt J.L.
        • Otopalik B.G.
        • Pickett S.B.
        • Greengard P.
        • Heintz N.
        Identification of the cortical neurons that mediate antidepressant responses.
        Cell. 2012; 149: 1152-1163
        • Hu H.
        • Roth F.C.
        • Vandael D.
        • Jonas P.
        Complementary tuning of Na(+) and K(+) channel gating underlies fast and energy-efficient action potentials in GABAergic interneuron axons.
        Neuron. 2018; 98: 156-165 e156
        • Muona M.
        • Berkovic S.F.
        • Dibbens L.M.
        • Oliver K.L.
        • Maljevic S.
        • Bayly M.A.
        • et al.
        A recurrent de novo mutation in KCNC1 causes progressive myoclonus epilepsy.
        Nat Genet. 2015; 47: 39-46
        • Yanagi M.
        • Joho R.H.
        • Southcott S.A.
        • Shukla A.A.
        • Ghose S.
        • Tamminga C.A.
        Kv3.1-containing K(+) channels are reduced in untreated schizophrenia and normalized with antipsychotic drugs.
        Mol Psychiatry. 2014; 19: 573-579
        • Sauer J.F.
        • Struber M.
        • Bartos M.
        Impaired fast-spiking interneuron function in a genetic mouse model of depression.
        eLife. 2015; 4
        • Shen S.
        • Lang B.
        • Nakamoto C.
        • Zhang F.
        • Pu J.
        • Kuan S.L.
        • et al.
        Schizophrenia-related neural and behavioral phenotypes in transgenic mice expressing truncated Disc1.
        J Meurosci. 2008; 28: 10893-10904
        • Sagi Y.
        • Medrihan L.
        • George K.
        • Barney M.
        • McCabe K.A.
        • Greengard P.
        Emergence of 5-HT5A signaling in parvalbumin neurons mediates delayed antidepressant action.
        Mol Psychiatry. 2020; 25: 1191-1201