(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Brn3b regulates the formation of fear-related midbrain circuits and defensive responses to visual threat [1]

['Hyoseo Lee', 'Department Of Ophthalmology', 'Visual Science', 'Yale University School Of Medicine', 'New Haven', 'Connecticut', 'United States Of America', 'Hannah Weinberg-Wolf', 'Hae-Lim Lee', 'Department Of Cellular']

Date: 2023-12

Defensive responses to visually threatening stimuli represent an essential fear-related survival instinct, widely detected across species. The neural circuitry mediating visually triggered defensive responses has been delineated in the midbrain. However, the molecular mechanisms regulating the development and function of these circuits remain unresolved. Here, we show that midbrain-specific deletion of the transcription factor Brn3b causes a loss of neurons projecting to the lateral posterior nucleus of the thalamus. Brn3b deletion also down-regulates the expression of the neuropeptide tachykinin 2 (Tac2). Furthermore, Brn3b mutant mice display impaired defensive freezing responses to visual threat precipitated by social isolation. This behavioral phenotype could be ameliorated by overexpressing Tac2, suggesting that Tac2 acts downstream of Brn3b in regulating defensive responses to threat. Together, our experiments identify specific genetic components critical for the functional organization of midbrain fear-related visual circuits. Similar mechanisms may contribute to the development and function of additional long-range brain circuits underlying fear-associated behavior.

Funding: This research was funded by the National Institutes of Health ( https://nei.nih.gov ) (EY031751, EY031512, EY029820 to IJK; EY 014454, EY029323 to JBD; EY026878, EY022312 to Yale School of Medicine) and by Whitehall Foundation ( http://whitehall.org )(2017-08-39 to AR). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2023 Lee et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Previous studies demonstrated a critical role for Brn3b in retina development [ 16 , 17 ], but little is known about Brn3b functions in other parts of the nervous system. Here, we report multifaceted roles of Brn3b in the organization and function of visual threat-related circuits, involving both superficial SC and deep SC/PAG. Conditional deletion of Brn3b in the midbrain decreased axonal projections to the LP due to neuronal loss in the superficial SC and also decreased expression of the neuropeptide tachykinin 2 (Tac2) in the deep SC/PAG. Brn3b mutants exhibited diminished freezing responses to a threatening looming stimulus [ 18 ] following single housing-based social isolation. This behavioral impairment was rescued by overexpression of Tac2 in the deep SC/PAG. These findings suggest that Tac2 acts downstream of Brn3b and that a change of Tac2 expression plays an important role in behavioral alterations in Brn3b mutants. Altogether, our findings define, for the first time, the molecular mechanism that regulates development and function of midbrain circuits conveying defensive responses to visual threat.

To examine molecular mechanisms that control development of the circuits mediating visually triggered fear responses, we sought genes expressed by subsets of midbrain neurons. We focused on the superficial SC, which conveys visual information from the retina and visual cortex to other subcortical areas [ 15 ], and discovered the expression of the transcription factor Brn3b within the bottom layer of superficial SC. Neurons in this layer innervate the lateral posterior nucleus (LP) of the thalamus, and SC-LP connections apparently mediate innate fear responses to visual threat manifested by freezing behavior [ 9 , 10 , 14 ]. We also discovered Brn3b expression confined to the deep SC and PAG. The overall expression pattern of Brn3b led us to hypothesize that this transcription factor may play important roles in the development of subcortical circuits critical for fear-related defensive behaviors.

Behavioral responses to threat are crucial for both animal and human survival. In humans, the inaccurate interpretation of threat- and fear-related information can lead to devastating psychiatric conditions [ 1 – 4 ]. Threatening stimuli are detected by sensory systems, including the visual system. Visually triggered defensive fear responses depend on midbrain structures, including the superior colliculus (SC) and periaqueductal gray (PAG) [ 5 – 7 ]. Threatening visual signals are conveyed to the superficial SC and subsequently delivered to the PAG either directly via the deep SC or indirectly within circuits connecting SC to other subcortical areas that ultimately project to the PAG [ 8 – 10 ]. The PAG integrates threat-related information and executes relevant defensive reactions. Specific midbrain circuits mediating defensive behaviors to visual threat have been identified using electrical and neurochemical stimulation as well as optogenetic and chemogenetic tools [ 7 – 14 ]. However, these studies evaluated rather broadly defined cell populations (e.g., based on the expression of CaMKIIa, parvalbumin, and VGluT2) leaving a gap in our understanding of the molecular mechanisms underlying the organization and functional specificity of the identified circuits.

Results

Layer-restricted Brn3b expression in the dorsal midbrain Our initial screening for SC neuronal markers in the superficial SC showed that Brn3b is expressed at the bottom layer of the superficial SC [15]. The organization of superficial SC can be defined relative to the axonal projections of retinal ganglion cells (RGCs). To visualize RGC axons, cholera toxin B subunit (CTB)-conjugated fluorescent dyes were delivered into 1 eye (Fig 1A–1C). CTB labeling divides superficial SC into 2 layers: stratum griseum superficiale (SGS), a densely labeled top layer, and stratum opticum (SO), a weakly labeled bottom layer. CTB labeling verified that Brn3b+ neurons, identified by immunostaining are localized in SO (Fig 1D). Furthermore, Brn3b is expressed sparsely in the intermediate SC and abundantly in the deep SC/PAG. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Layer-restricted Brn3b expression in the dorsal midbrain. (A, B) Schematic diagrams of the midbrain after CTB injection into the contralateral eye (A); detailed layer distribution in the coronal section of the dorsal midbrain (top) and a sagittal image showing the location (dotted line) of the imaged coronal section (bottom) (B). (C, D) Brn3b expression (green), visualized by anti-Brn3b antibody is detected in the sSC, iSC, dSC, and PAG at P11 (C). Magnified view of the boxed area (D) shows Brn3b expression confined to the bottom layer of the sSC (SO). CTB labeling (red) divides the superficial SC into the SGS and SO. The aqueduct (Aq) is also shown. DAPI (blue). (E) Schematic diagram of the conditional Brn3b allele. Conditional deletion (by En1-Cre) removes an open reading frame of Brn3b and places a human placental AP under the control of the Brn3b promoter. (F) Dorsal midbrain section from Brn3bflox/+:: En1-Cre mouse histochemically processed for AP signals at P24. Brackets represent the layer segregation. (G-K) Axonal projections from the dorsal midbrain to adPBG and Pn (G), CnF (H), central lateral nucleus of the thalamus (CL) (I), lateral posterior nucleus (LP) and hypothalamus (HY) (J), vLGN and PRC (K). Each area was identified by its anatomical position (see Methods). (L) Schematic diagram of the midbrain Brn3b+ neuronal projections. Scale bars: 250 μm. adPBG, adjacent parabigeminal nucleus; AP, alkaline phosphatase; CnF, cuneiform nucleus; CTB, cholera toxin B; dSC, deep SC; iSC, intermediate SC; PAG, periaqueductal gray; Pn, pons; PRC, precommisural nucleus; SC, superior colliculus; SGS, stratum griseum superficiale; SO, stratum opticum; sSC, superficial SC; vLGN, ventral lateral geniculate nucleus. https://doi.org/10.1371/journal.pbio.3002386.g001 Neurons in specific layers of the dorsal midbrain selectively innervate different subcortical areas [19,20]. Restricted expression of Brn3b, particularly in the bottom layer of superficial SC (SO), prompted us to examine axonal projections of Brn3b+ neurons to other brain areas. To trace Brn3b+ axonal projections, we used a conditional Brn3b mouse carrying the Brn3b gene flanked by loxP sites and a human placental alkaline phosphatase (AP) coding region inserted immediately downstream of 3′ loxP (Brn3bflox/+, Fig 1E) [21]. Following Cre-mediated recombination, a copy of Brn3b is deleted and AP is expressed under the control of the Brn3b promoter. To visualize Brn3b+ neurons, we crossed Brn3bflox/+ to the En1-Cre line, in which Cre expression is confined to the midbrain [22,23]. The layer-restricted distribution of AP staining in the SC matched the labeling from the Brn3b antibody (S1A–S1G Fig): both AP and immunostaining signals in the superficial SC were detected close to the pia at P1 when SGS and SO are barely distinguishable, and then became restricted to the SO at P10/P12 when superficial SC development was completed [24]. Both methods revealed abundant labeling of somas/nuclei in the deep SC/PAG at P1. These observations confirmed that AP signals report Brn3b expression. We next examined axonal projections to other brain areas, based on AP signals, and found that Brn3b+ neurons clearly innervate several subcortical areas, including LP, ventral lateral geniculate nucleus (vLGN), precommisural nucleus, hypothalamus, central lateral nucleus of the thalamus, adjacent parabigeminal nucleus, pons, and cuneiform nucleus (Fig 1F–1L). To rule out the possibility that observed axonal projections originate from areas outside the midbrain, we first conducted immunostaining of the thalamus including LP, and found no Brn3b staining, excluding the contribution of local thalamic neurons to the detected AP signals (S1H and S1I Fig). Considering the broad expression of Brn3b in the retina, we also examined Brn3bflox/+:: En1-Cre mice to ensure that labeling was not present in the retina’s output neurons, RGCs. Indeed, no AP signals were detected in the retina, confirming the midbrain-restricted Cre expression in the En1-Cre mouse (S1J Fig). Together, these results demonstrate that the observed axonal labeling represents Brn3b+ neuronal projections from the dorsal midbrain. Next, we examined whether Brn3b+ neurons are glutamatergic or GABAergic, considering that SC projection neurons comprise both populations [9,25]. We initially attempted to mark glutamatergic and GABAergic neurons by either immunostaining or crossing Slc17a6-Cre (VGluT2-ires-Cre) or Gad2-Cre to a Cre-dependent td-Tomato reporter line (Ai14). However, very dense immunolabeling of synaptic terminals and overly abundant filling of neuronal processes by td-Tomato signals precluded visualizing individual somas. Instead, we performed in situ hybridization using Slc17a6 and Gad1 probes. The brain tissue for the hybridization was obtained from a newly generated Brn3bGFP/+ mouse, in which an open reading frame of Brn3b was replaced with GFP by CRISPR-based genome editing (S2A–S2C Fig). First, we confirmed that GFP signals faithfully represent Brn3b expression by double immunostaining, which showed that approximately 95% of Brn3b+ cells express GFP and approximately 93% of GFP+ cells express Brn3b. Subsequent double labeling with anti-GFP combined with in situ probes to either Slc17a6 or Gad1 revealed that all Brn3b+ neurons are glutamatergic and not GABAergic (S2D–S2L Fig).

Brn3b is required for survival of neurons projecting to LP To assess the role of Brn3b in circuit assembly of the midbrain, we crossed Brn3bflox/flox to Brn3b+/-:: En1-Cre mice to generate Brn3bflox/-:: En1-Cre progeny (Brn3b cKO) and confirmed Brn3b loss in midbrain by immunostaining (S3A–S3E Fig). The conditional approach avoids confounding effects of retinal degeneration and, consequently, the compromised development of superficial SC in global Brn3b mutants [17,26]. Therefore, we proceeded to analyze and compare the phenotypes of control animals carrying conditional Brn3b and WT alleles (Brn3bflox/+:: En1-Cre) and mutants carrying conditional and null Brn3b alleles (Brn3bflox/-:: En1-Cre, cKO); in both cases, mice carry 1 copy of En1-Cre and 1 copy of Brn3bflox, producing AP signals. The AP signals faithfully represent Brn3b expression in both control and cKO mice. These signals disappeared in the superficial SC of the mutants, with no obvious changes in other layers. Thus, Brn3b deletion induces a selective loss of Brn3b+ neurons in the superficial SC (Fig 2A). We also examined projections of Brn3b+ neurons to other brain areas and found that axonal projections to LP and ventral LGN were decreased in cKO mice, most significantly in the rostral LP and ventral LGN (Fig 2B–2D). Quantification of AP signals in the LP revealed an approximately 48% decrease. Axonal projections to other subcortical areas exhibited no obvious differences (Fig 2B and 2C and Fig 2E–2G). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Brn3b deletion reduces the number of Brn3b+ neurons in the superficial SC and decreases projections to LP. Brn3b+ neurons visualized by AP signals. (A) (Left) Midbrain sections from control and conditional Brn3b knockout (cKO) at P24. (Right) Magnified view of the boxed areas (i, ii) showing the loss of Brn3b+ neurons in the superficial SC (arrow) of the cKOs. Brackets represent the layer distribution. (B–D) Decreased projections to the LP and vLGN and no difference in projections to the hypothalamus (HY) and PRC in cKO brains. Magnified view of the boxed images (iii, vi) showing the LP areas used for quantification (D; n = 5 mice/group). The AP-labeled LP area, normalized to the average control value, was reduced in cKO brains. Unpaired two-tailed Student’s t test (mean ± SEM, p < 0.0001 [****]). (E–G) No obvious differences in axonal projections to the central lateral nucleus of the thalamus (CL), adPBG, Pn, and CnF between control and cKO at P24 (n = 5 mice/group). Scale bars: 250 μm. The data underlying this figure can be found in S1 Data. adPBG, adjacent parabigeminal nucleus; AP, alkaline phosphatase; CnF, cuneiform nucleus; LP, lateral posterior nucleus; Pn, pons; PRC, precommisural nucleus; SC, superior colliculus; vLGN, ventral lateral geniculate nucleus. https://doi.org/10.1371/journal.pbio.3002386.g002 The cKO animals lacked Brn3b+ neurons at the bottom layer of the superficial SC (SO), where LP-projecting neurons localize, which apparently explains the altered projections to LP. However, because some axonal projections to LP were still detectable in these cKO mice, we considered the possibility that Brn3b+ neurons in other layers of SC may also innervate LP. To test this idea, we first evaluated whether superficial SC Brn3b+ neurons project to LP using a second Cre line (S4A–S4D Fig). We crossed Brn3bflox/+ to Ntsr1-GN209-Cre mice, which express Cre in the LP-projecting superficial SC neurons [27,28], but we unexpectedly found AP signals also in the retina. However, projections of Brn3b+ neurons to LP persisted after enucleating both eyes, suggesting that labeled axons in LP originated from the SC and not from the retina. We also quantified the number of Brn3b+ neurons in the offspring of Ntsr1-GN209-Cre crossed to the Ai14 line expressing Cre-dependent td-Tomato. We found that approximately 23% of td-Tomato+ cells were Brn3b+. Interestingly, Brn3b mutants (Brn3bflox/-:: Ntsr1-GN209-Cre) showed no changes in the survival of the superficial SC Brn3b+ neurons and projections to the LP, likely because Cre expression occurs too late during development in this Cre line (see Discussion). Therefore, for further analysis, we utilized En1-Cre-based mutants only. To confirm that Brn3b+ neurons in superficial SC project axons to LP, we labeled these neurons using 2 constructs delivered by adeno-associated virus (AAV): GFP-dependent FLP recombinase (FLP-DOG) and FLP-dependent mCherry (fDIO-mCherry). The FLP-DOG (DOG: Dependent On GFP) is genetically engineered to be unstable and degrade, whereas binding to GFP prevents its degradation [29]. Inside GFP+ cells, FLP-DOG becomes stabilized and converts fDIO-mCherry to an active configuration, allowing mCherry expression. We injected AAV-FLP-DOG and AAV-fDIO-mCherry into the superficial SC of Brn3bGFP/+ mouse to selectively label Brn3b+ neurons (S4E–S4I Fig). Double immunostaining revealed that approximately 98% of mCherry+ cells express GFP, suggesting high fidelity of the FLP-DOG method. Subsequent analysis revealed axons labeled with mCherry in LP, confirming that Brn3b+ neurons in superficial SC indeed innervate LP and that they are not local neurons. Lastly, we injected AAV-mCherry into the deep SC/PAG of wild-type mice and found axonal labeling in LP (S5A–S5D Fig). Together, our results indicate that LP is innervated by neurons of the superficial SC as well as deep SC/PAG and that Brn3b deletion by En1-Cre causes a loss of the superficial SC Brn3b+ neurons specifically, resulting in the decreased projection to the LP. In addition to affecting the SC-LP connections, Brn3b deletion by En1-Cre completely eliminated innervation of the ventral LGN (Fig 2B and 2C). Given that the superficial SC Brn3b+ neurons project only to the LP and that delivering AAV-mCherry into the deep SC/PAG did not show projections to the ventral LGN, we speculated that the ventral LGN-projecting Brn3b+ neurons are located in the intermediate SC layer. To test this possibility, we delivered AAV-Cre into the intermediate SC of Brn3bflox/+ mice, which successfully labeled Brn3b+ neurons in the intermediate SC and confirmed their axonal projections to the ventral LGN (S5E–S5G Fig).

Brn3b deletion-induced cell death occurs around birth We next analyzed development of Brn3b+ neurons to define specific time points of neuronal loss and altered projections to LP. During development, Brn3b expression in the dorsal midbrain becomes detectable as early as E13 (Fig 3A and 3B). Therefore, we examined the possibility that Brn3b could regulate early development including neurogenesis. However, we found that Brn3b expression is barely detectable in the layer defined by a marker of proliferating cells, Ki67, suggesting that Brn3b is expressed post-mitotically and regulates later aspects of neuronal development and differentiation. We subsequently examined control mouse brains at later developmental time points, E16, E18, P1, and P4 (Fig 3C–3E). Based on the AP signals, segregation of Brn3b+ neurons into the superficial SC was first detectable at E18 and became obvious at P1. In cKO mice, the loss of superficial SC Brn3b+ neurons was already evident at E18 and persisted until adult stages (Fig 2A). It is known that Brn3b deletion causes apoptotic death of RGCs during retinal development [17]. To examine whether Brn3b deletion likewise causes apoptotic death of Brn3b+ neurons in the superficial SC, we performed immunostaining with anti-cleaved caspase-3 (Fig 3F–3J). Indeed, Brn3b deletion increased the number of cleaved caspase-3+ cells at E18 (approximately 59% increase in cKO) and this labeling was mainly detected in superficial and intermediate SC, suggesting that Brn3b expression is necessary for neuronal survival in these layers. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Loss of the superficial SC neurons in Brn3b mutants occurs around birth. (A) Schematic diagram illustrating a specific localization of the imaged sagittal section according to a lateral distance from the midsagittal line at E13. (B) (Left) Example of sagittal section. (Right) Magnified view of the boxed area. Brn3b+ neurons (red) were missing in the layer labeled by Ki67, a maker of proliferating cells (cyan) at E13 (n = 3 mice). (C) Schematic diagrams showing the brain area (boxed) of a coronal image used for analysis (top) and a sagittal image depicting the level (dashed line) where each coronal section was obtained (bottom) at E16-E18 (i) and at P1-P5 (ii). (D, E) Brn3b+ neurons, visualized by AP signals, were weakly detectable at E18 and become obvious at P1 and P4 (arrows) in the superficial SC of control (D). However, these neurons were missing at E18, P1 and P4 in the superficial SC of cKO (E) (n = 3 mice/group/development stage). (F) Schematic diagram depicting the brain areas (boxed) analyzed to examine the number of cleaved caspase-3+ cells in the sagittal section. The localization of imaged sections was shown as a lateral distance from the midsagittal line. (G, H) Cleaved caspase-3 immunoreactivity (red) revealed increased cell death at E18 in cKO (H), compared to control (G). DAPI (blue). (I, J) Quantification revealed increased cell death in cKO (35.7 ± 4.9/mm2 for control, 56.6 ± 3.3/mm2 for cKO; n = 5 mice/group), primarily occurring in the superficial and intermediate layers; unpaired two-tailed Student’s t test (mean ± SEM, p = 0.008 [**] in I and p = 0.03 [*] and 0.02 [*] in J). Brackets represent the layer segregation and dashed lines indicate the pial surface. Scale bars: 250 μm (B) and 150 μm (D–E and G–H). The data underlying this figure can be found in S1 Data. AP, alkaline phosphatase. https://doi.org/10.1371/journal.pbio.3002386.g003 The lack of AP signals in the superficial SC of Brn3b mutants indicates the death of Brn3b+ neurons. However, extensive AP signals in other layers made it difficult to distinguish somas from processes, preventing visualization of individual Brn3b+ cells. To assess possible changes of Brn3b+ neurons in other layers, we utilized the molecular markers ETV1 and Brn3a. ETV1 is expressed in Brn3b+ neurons occupying the intermediate SC layer (S6A–S6G Fig). Double immunostaining showed that approximately 50% of Brn3b+ neurons were ETV1+ and approximately 49% of ETV1+ neurons were Brn3b+ at P2. Quantification of ETV1+ neurons at E16 and P2 revealed no difference between control and cKO mice, suggesting that Brn3b deletion has no effect on the development of Brn3b+/ETV1+ cells. We also examined Brn3b+ neurons expressing Brn3a (S6H–S6M Fig). Double immunostaining showed that approximately 60% of Brn3b+ neurons were Brn3a+ and approximately 45% of Brn3a+ neurons were Brn3b+ at P1 and that Brn3b+/Brn3a+ neurons are mainly located in deep SC/PAG. Quantification of Brn3a+ neurons at E15 and P1 also revealed no difference in their number and distribution in control and cKO mice, indicating that Brn3b deletion has no substantial effects on the development of Brn3b+/Brn3a+ neurons. Considering that ETV1 and Brn3a are expressed in only about half of Brn3b+ neurons in their respective areas of co-expression (intermediate SC and deep SC/PAG), we also examined survival of Brn3b+ neurons using the Brn3bGFP/+ mouse, in which GFP expression faithfully reports Brn3b expression (S2 Fig). We crossed Brn3bflox/flox to Brn3bGFP/+:: En1-Cre mice and confirmed Brn3b loss in Brn3bGFP/ flox:: En1-Cre offspring (Brn3b-GFP cKO) by immunostaining (S7 Fig). Quantification of GFP+ signals at E16 revealed no difference between Brn3bGFP/+ (control) and Brn3b-GFP cKO. Also, no changes in GFP+ signals were detected in deep SC/PAG at P1/P2. However, we found approximately 30% decrease of GFP+ cell number in superficial and intermediate SC, confirming that Brn3b deletion affects neuronal survival in these layers specifically. Collectively, these results suggest that the loss of Brn3b has no effect on neuronal survival in the layers of dorsal midbrain outside the superficial and intermediate SC, including the deep SC/PAG. We also examined axonal development during several embryonic and postnatal stages. We observed no obvious differences between control and Brn3b mutants at E13, suggesting that Brn3b deletion had no effects on initial axonal growth (S8 Fig). Consistent with our observations at P24, no obvious changes of AP signals were detected in other brain areas at E13, E16, and P1/P2. Projections to LP became visible at P4/P5 in control and cKO mice (S9 Fig), indicating that Brn3b deletion causes no significant delay in axonal growth. However, despite seemingly unaffected developmental timing, projections to LP were clearly decreased in the mutants (approximately 80% decrease at P4/P5, approximately 60% at P8/P9, approximately 55% at P12/P13). Altogether, our findings suggest that Brn3b deletion in the midbrain causes decreased projections to the LP due to the loss of superficial SC Brn3b+ neurons occurring around birth.

Brn3b regulates expression of Tac2 in the deep SC/PAG The deep SC/PAG highly expresses Brn3b but did not show obvious morphological alterations in Brn3b mutants. Considering the diverse roles of Brn3b in retinal development [21,30,31], we speculated that apart from specific structural changes, Brn3b loss could cause various transcriptional alterations, including dysregulation of the genes involved in visually triggered behavioral responses. To test this hypothesis, we conducted transcriptomic profiling of the dorsal midbrain and identified 1,992 genes that were significantly differentially expressed in Brn3b cKOs compared to controls (padj < 0.05; Fig 4A and 4B and S1–S3 Tables). We discovered that besides Brn3b itself, neuropeptide tachykinin 2 (Tac2) showed the highest level of down-regulation in the Brn3b mutants. Interestingly, Tac2 was previously implicated in fear learning [32,33]. Additionally, Tac2 expression was up-regulated by single housing-based social isolation, which caused enhancement of freezing responses to a visually threatening looming stimulus [34]. Therefore, we hypothesized that decreased Tac2 expression due to the loss of Brn3b could affect neural circuits mediating fear-related behavior. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Brn3b regulates expression of Tac2 in the deep SC/PAG. (A, B) Transcriptomic profiling (RNA-seq) using dorsal midbrain at P2-4 (n = 3 for control and 4 for cKO). (A) Volcano plot of the RNA-seq results. Brn3b and Tac2 were down-regulated in cKO (arrows). The vertical dashed lines indicate the lowest log2-fold change value of significantly down- or up-regulated genes. The horizontal dashed line indicates adjusted p value (p adj ) at 0.05. (B) Violin plots showing TPM values of Brn3b and Tac2 for each animal. The log2-fold change was 2.01 for Brn3b and 3.10 for Tac2. (C) Assessment of Tac1, Tac2, and Crh levels at P3-P5 showed decreased expression of Tac2 in cKO mice. The data for transcripts were normalized to the average control value. Quantification (Tac1: 1.01 ± 0.07 for control, 1.22 ± 0.12 for cKO; Tac2: 1.07 ± 0.16 for control, 0.54 ± 0.12 for cKO; Crh: 1.02 ± 0.11 for control, 1.37 ± 0.18 for cKO; Brn3b: 1.05 ± 0.17 for control, 0.01 ± 0.00 for cKO; n = 5 mice/group/gene). Unpaired two-tailed Student’s t test (mean ± SEM, p = 0.177 for Tac1, p = 0.030 [*] for Tac2, p = 0.136 for Crh and p = 0.0003 [***] for Brn3b). (D) Schematic diagram of the brain area analyzed for Tac2 expression in a coronal section (top) and a sagittal image depicting the level (dashed line) where the coronal section was acquired (bottom). (E, F) Brn3b deletion decreased Tac2 expression (red) in the deep SC/PAG, as detected by in situ hybridization. DAPI (blue). (G) Quantification of Tac2+ cell number (76.7 ± 6.5/mm2 for control, 48.2 ± 3.3/mm2 for cKO; n = 3 mice/group). Unpaired two-tailed Student’s t test (mean ± SEM; p = 0.017 [*]). Scale bars: 250 μm. The data underlying this figure can be found in S1 Data and S1–S3 Tables. PAG, periaqueductal gray; SC, superior colliculus; TPM, transcripts per million. https://doi.org/10.1371/journal.pbio.3002386.g004 To further assess whether Tac2 may act downstream of Brn3b, we performed reverse transcription-quantitative PCR (RT-qPCR) on independent biological samples and confirmed that Brn3b deletion indeed significantly reduces Tac2 expression (approximately 50%, Fig 4C). Two related neuropeptides, tachykinin1 (Tac1) and corticotropin-releasing hormone (Crh), known to mediate effects of stress [35,36], showed no difference between controls and mutants. To determine whether Brn3b+ neurons express Tac2, we performed in situ hybridization with Tac2 probes on brain tissues of the Brn3bGFP/+ mouse (S10 Fig). Tac2+ cells were mainly located in the deep SC/PAG and approximately 57% of Tac2+ cells were Brn3b+. No Tac2+ cells were found in the superficial and intermediate SC. Furthermore, Brn3b deletion reduced Tac2+ cell number (approximately 37% decrease, Fig 4D–4G), consistent with the finding that Brn3b regulates Tac2 expression. Unlike the superficial SC, altered Tac2 expression in the deep SC/PAG of the mutants very unlikely resulted from neuronal death, as neither an obvious loss of AP signals (Fig 2A) nor a change in the number of Brn3a+ cells and GFP+ cells was detected in that area (S6 and S7 Figs).

Brn3b-dependent circuits are required for behavioral responses to visual threat Upon discovering that Brn3b loss disrupts SC-LP connections and down-regulates Tac2 expression, we proceeded to evaluate whether Brn3b deletion affects visually triggered fear responses using the looming stimulus assay (Fig 5A). This assay examines the mouse’s reaction to a dark, overhead expanding disk (a looming stimulus) that mimics an approaching predator (see Methods) [18]. The stimulus causes mice to either freeze or escape to a shelter. Here, we omitted the shelter to focus on freezing responses that are thought to depend on the SC-LP pathway [9,10,14]. The behavioral examination was performed using group-housed as well as single-housed mice to investigate potential enhancement of the freezing responses precipitated by social isolation and linked to Tac2 activity [34]. Basic visual abilities were examined using the light/dark exploration test [37,38]. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Brn3b-dependent circuits are necessary for visually triggered freezing responses. Fear-related defensive behavior was analyzed by examining freezing responses to visual threat in a looming stimulus assay. (A) Schematic diagram of the behavioral paradigms. The looming stimulus was an expanding dark disk on a gray background presented overhead (0.25 s expansion; 0.25 s at largest disk size; 0.5 s interstimulus interval; 10 repeats). Following the looming stimulus assay, the mice were tested on light/dark discrimination. For the experiments involving single housing, the mice were housed individually for 2 weeks prior to the behavioral tests. (B) Quantification of freezing responses for group-housed (32.4 ± 6.5 s for control, 22.6 ± 6.1 s for cKO; n = 9 for control and 8 for cKO) and single-housed (78.1 ± 12.3 s for control, 35.3 ± 7.4 s for cKO; n = 12 for control and cKO) mice. A significant difference in total freezing time was observed between control and cKO in single-housed but not in group-housed mice. Tukey post hoc analysis (mean ± SEM, p = 0.903 for group-housed and p = 0.006 [**] for single-housed animals). (C) Quantification of the time spent in the brightly lit area for group-housed (132.5 ± 14.0 s for control, 137.5 ± 21.9 s for cKO; n = 9 for control and 8 for cKO) and single-housed (108.7 ± 15.6 s for control, 136.1 ± 12.9 s for cKO; n = 9 for control and cKO) mice. No significant difference was detected in the light/dark exploration test. Tukey post hoc analysis (mean ± SEM, p = 0.996 for group-housed, p = 0.618 for single-housed animals). The data underlying this figure can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3002386.g005 Our looming stimulus assay evoked prolonged freezing responses (Fig 5B). Indeed, nearly all mice froze within a few seconds of the onset of the 10 s looming stimulus and remained frozen for up to 3 min in some cases. A 2 × 2 ANOVA (genotype × housing condition) showed a main effect of genotype (F = 7.866, p = 0.008) and a main effect of housing condition (F = 9.727, p = 0.004) with no interaction (F = 3.099, p = 0.087). Specifically, the Brn3b cKO group froze less than the control group, and both groups froze longer following single housing. Follow-up post hoc analysis showed a significant difference between groups only in the single housing condition (t = 5.018, p = 0.006; group housing: t = 0.965, p = 0.903). All mouse groups behaved similarly in the light/dark exploration test (Fig 5C). A 2 × 2 ANOVA showed no statistical differences between genotypes (F = 1.011, p = 0.323) and housing conditions (F = 0.6073; p = 0.442). Overall, these results suggest that Brn3b cKO mice have overall normal vision; however, they exhibit impaired visually evoked freezing responses following single housing.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002386

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.