1. Neuroscience
Download icon

Roundabout receptor 2 maintains inhibitory control of the adult midbrain

  1. Bryan B Gore
  2. Samara M Miller
  3. Yong Sang Jo
  4. Madison A Baird
  5. Mrinalini Hoon
  6. Christina A Sanford
  7. Avery Hunker
  8. Weining Lu
  9. Rachel O Wong
  10. Larry S Zweifel  Is a corresponding author
  1. University of Washington, United States
  2. Boston University Medical Center, United States
Research Article
  • Cited 9
  • Views 970
  • Annotations
Cite this article as: eLife 2017;6:e23858 doi: 10.7554/eLife.23858

Abstract

The maintenance of excitatory and inhibitory balance in the brain is essential for its function. Here we find that the developmental axon guidance receptor Roundabout 2 (Robo2) is critical for the maintenance of inhibitory synapses in the adult ventral tegmental area (VTA), a brain region important for the production of the neurotransmitter dopamine. Following selective genetic inactivation of Robo2 in the adult VTA of mice, reduced inhibitory control results in altered neural activity patterns, enhanced phasic dopamine release, behavioral hyperactivity, associative learning deficits, and a paradoxical inversion of psychostimulant responses. These behavioral phenotypes could be phenocopied by selective inactivation of synaptic transmission from local GABAergic neurons of the VTA, demonstrating an important function for Robo2 in regulating the excitatory and inhibitory balance of the adult brain.

https://doi.org/10.7554/eLife.23858.001

eLife digest

Although no two people are alike, we all share the same basic brain structure. This similarity arises because the same developmental program takes place in every human embryo. Specific genes are activated in a designated sequence to generate the structure of a typical human brain. But what happens to these genes when development is complete – do they remain active in the adult brain?

A gene known as Robo2 encodes a protein that helps neurons find their way through the developing brain. Many of these neurons will ultimately form part of the brain’s reward system. This is a network of brain regions that communicate with one another using a chemical called dopamine. The reward system contributes to motivation, learning and memory, and also underlies drug addiction. In certain mental illnesses such as Parkinson’s disease and schizophrenia, the dopamine-producing neurons in the reward system work incorrectly or die.

To find out whether Robo2 is active in the mature nervous system, Gore et al. used genetic techniques to selectively remove the gene from the reward system of adult mice. Doing so reduced the ability of the dopamine neurons within the reward system to inhibit one another, which in turn increased their activity. This changed the behavior of the mice, making them hyperactive and less able to learn and remember. Cocaine makes normal mice more active; however, mice that lacked the Robo2 gene became less active when given cocaine.

Overall, the work of Gore et al. suggests that developmental axon guidance genes remain important in the adult brain. Studying developmental genes such as Robo2 may therefore open up new treatment possibilities for a number of mental illnesses and brain disorders.

https://doi.org/10.7554/eLife.23858.002

Introduction

The midbrain dopamine system, consisting of the VTA and substantia nigra pars compacta (SNc), is essential for motor function, motivation, reward, learning, and memory. Alterations in the activity patterns of dopamine neurons have been proposed as a key contributor to mental illness (Grace, 1991). In addition to dopamine, accumulating evidence points to an essential balance of excitatory and inhibitory neurotransmitter systems in the brain to allow for proper function. Shifts in this balance are increasingly linked to a variety of mental disorders including schizophrenia (Eichler and Meier, 2008) and autism (Nelson and Valakh, 2015). The VTA dopamine system is also broadly implicated in addiction where changes in inhibitory and excitatory strength are proposed to underlie drug-seeking behavior (Chen et al., 2010). The cellular mechanisms that maintain the excitatory and inhibitory synaptic control of the adult VTA are not well known, but resolving these processes has important implications for resolving the molecular regulation of circuit connectivity.

To identify additional genes that may regulate synaptic connectivity in the midbrain dopamine system, we surveyed the Allen Institute mouse brain expression atlas (Lein et al., 2007) for genes with enriched or partially enriched expression in the VTA/SNc that had previously defined roles in axonal pathfinding, synaptogenesis, or plasticity, and have linkage to mental illness. One of the genes that we identified encodes for the axon guidance receptor Robo2. Robo2 was first discovered as one of the four mammalian homologs to the Drosophila Roundabout gene (Robo), which was named because of the unique axonal recrossing pattern generated at the CNS midline (Seeger et al., 1993; Kidd et al., 1998; Brose et al., 1999; Simpson et al., 2000). Further studies found that Robo family members perform additional axon guidance functions to regulate brain wiring broadly throughout the nervous system (Ypsilanti and Chedotal, 2014), including midbrain dopamine neurons (Dugan et al., 2011). Additional studies expanded the essential functions of Robo2 to include cell migration, synaptogenesis, synaptic plasticity, neuronal survival, and dendritic patterning during prenatal and early postnatal development (Shen and Cowan, 2010; Gibson et al., 2014; Koropouli and Kolodkin, 2014; Ypsilanti and Chedotal, 2014).

Previous studies demonstrate that the neuronal axon guidance receptor deleted in colorectal cancer (DCC) is also highly expressed in the adult dopamine neurons, is dynamically regulated by amphetamine (Yetnikoff et al., 2007), and mice with haploinsufficiency for DCC demonstrate blunted responses to the psychomotor activating effects of amphetamine (Flores et al., 2005). Consistent with the continued function of DCC in the adult nervous system, it has recently been demonstrated that DCC regulates excitatory synaptic transmission in the adult hippocampus (Horn et al., 2013). These findings are consistent with a role for adult expression of neuronal pathfinding genes in the regulation of the adult dopamine system.

During early development of the spinal cord, Robo and DCC signaling have opposing roles in commissural axon guidance, with DCC activation promoting axon growth toward the midline and Robo serving as a chemorepellant to prevent recrossing (Tessier-Lavigne and Goodman, 1996). Robo receptors are activated upon binding chemorepellants of the Slit family, Slit1, Slit2, and Slit3 (Tessier-Lavigne and Goodman, 1996), which cause growth cone collapse through cytoskeletal remodeling via Rho-family GTPase signaling (Wong et al., 2001). All three genes for the Slit ligands and the receptors Robo1 and Robo2 (Marillat et al., 2002) are expressed in the adult midbrain, including the VTA and SNc; however, the function of these ligands and receptors in this context and their cell-specific requirements has yet to be established. Of further interest, genes involved in Robo signaling have been linked to schizophrenia, autism, conduct disorder, intellectual disability, language impairment, and bipolar depression (Anitha et al., 2008; Potkin et al., 2009, 2010; Viding et al., 2010; Suda et al., 2011; Dadds et al., 2013; St Pourcain et al., 2014; Wang et al., 2014; Okbay et al., 2016).

Based on the expression of Robo2 in the adult limbic system of the brain, its role in establishing the early wiring of the midbrain dopamine system (Dugan et al., 2011), and its opposing function related to netrin/DCC signaling, we hypothesized that this gene may play an important role in maintaining or modulating synaptic connectivity of the adult dopamine system. To test this hypothesis, we conditionally inactivated the Robo2 gene exclusively in the adult VTA of mice. We find that Robo2 regulates the inhibitory synaptic connectivity of the VTA, and that disruption of this connectivity has profound behavioral effects including altered psychomotor activity and impaired learning. Most intriguing was a paradoxical inversion of behavioral responses to the psychostimulants cocaine and amphetamine, whereby these drugs calmed hyperactivity in mice lacking Robo2 in the adult midbrain. These findings strongly implicate Robo2 signaling in maintaining balanced control of the VTA that is highly relevant for behavioral disorders linked to altered psychomotor control and cognition such as autism, ADHD, addiction, and schizophrenia.

Results

Robo2 in the adult VTA regulates inhibitory synaptic transmission

Analysis of the Allen Institute mouse brain expression atlas (Lein et al., 2007) reveals that Robo2 is broadly expressed in the adult brain with the highest expression in the ventral midbrain, hippocampus and cerebellum, similar to previous reports (Marillat et al., 2002) (Figure 1—figure supplement 1A). To confirm the expression of Robo2 in the VTA, we performed immunostaining for Robo2 and tyrosine hydroxylase (TH, a dopamine neuron marker) on brain sections from adult mice (>8 weeks old). Robo2 expression was observed throughout the VTA, largely localized to TH-positive neurons, though several TH-negative neurons were observed to also express Robo2 (TH+Robo2+ = 75.28 ± 0.56%, TH+Robo2− = 7.81 ± 1.46%, and TH-Robo2+ = 16.90 ± 1.77%, n = 4) (Figure 1A).

Figure 1 with 3 supplements see all
Robo2 VTA mutants have a reduction in the frequency of sIPSCs.

(A) Robo2 is expressed in the VTA, predominantly in dopamine neurons (TH+ cells, arrows), but also non-dopamine neurons (TH- cells, arrowheads). Scale bar = 100 µm (Robo2), 10 µm (merge/zoom). (B) The genomic structure of the Robo2lox/lox allele before and after Cre recombination, adapted from Lu et al. (2007). (C) Schematic of strategy to virally inactivate Robo2 in the VTA (left). Representative image (right) showing expression of Cre-GFP and TH in the VTA. Scale bar = 100 µm. (D) sIPSC representative traces for control and mutant Ih positive cell. (E) sIPSC frequency (left) and cumulative distribution plot (right) for control (n = 13) and mutant (n = 16) Ih positive cells. Unpaired t-test, t(27) = 3.944, ***p<0.001. Two-way repeated measures ANOVA, genotype x time interaction, F(27,100) = 17.30, ****p<0.0001. (F) sIPSC amplitude (left) and cumulative distribution plot (right) for control and mutant Ih positive cells. (G) sIPSC representative trace for control and mutant Ih negative cell. (H) sIPSC frequency (left) and cumulative distribution plot (right) for control (n = 10) and mutant (n = 14) Ih negative cells. Unpaired t-test, t(22) = 3.459, **p<0.01. Two-way repeated measures ANOVA, genotype x time interaction, F(22,100)=11.40, ****p<0.0001. (I) sIPSC amplitude (left) and cumulative distribution plot (right) for control and mutant Ih negative cells. Bars represent mean ± SEM.

https://doi.org/10.7554/eLife.23858.003

To confirm Robo2 expression in the midbrain, we performed Ribotag, a method which allows for cell-type specific enrichment of actively translating mRNAs (Sanz et al., 2009). Mice containing the floxed HA-tagged Ribosomal Protein L22 (Rpl22) gene were crossed to mice containing Cre recombinase expression under the control of the endogenous dopamine transporter locus (Slc6a3Cre/+), or the vesicular GABA transporter (Slc32a1Cre/+), allowing for isolation of dopamine and GABA neuron translatomes, respectively (Figure 1—figure supplement 1D–E). Tissue punches of the adult midbrain from double heterozygous animals (Slc6a3Cre/+::Rpl22lox-HA/+, n = 3; Slc32a1Cre/+::Rpl22lox-HA/+, n = 3) were immunoprecipitated (IP) to isolate Rpl22-HA tagged polyribosomes. TaqMan real time PCR was used to test for the relative enrichment of Robo2 in the IP fraction relative to total mRNA (input). We found Robo2 to have enriched expression in the IP fraction relative to the input in both mouse lines (Figure 1—figure supplement 1F) confirming Robo2 expression in both dopamine and GABA neurons of the midbrain neurons.

Based on previous findings that DCC regulates excitatory synaptic strength in the adult hippocampus (Horn et al., 2013), we hypothesized that Robo2 may also play an important role in this process. To establish whether Robo2 in the adult VTA plays a functional role in synaptic function, we measured spontaneous excitatory and inhibitory postsynaptic currents (sEPSCs and sIPSCs, respectively) in putative dopamine and non-dopamine neurons following Robo2 inactivation. To inactivate Robo2, an adeno-associated viral vector (AAV) containing an expression cassette for a GFP-tagged Cre recombinase (AAV1-Cre-GFP) was injected into the VTA of mice with a Cre-conditional allele for Robo2 (‘mutant mice’) (Figure 1B–C and Figure 1—figure supplement 1B–C). The generation of Robo2lox/lox mice were described previously (Lu et al., 2007). In this line, loxP recognition sites are inserted to flank exon 5 of Robo2, and following Cre-mediated recombination, a frameshift occurs resulting in a premature stop codon in exon 6 (Figure 1B). We confirmed a specific reduction in Robo2 protein in cells expressing Cre-GFP but not in cells expressing the enzymatically dead version of Cre, ∆Cre-GFP (Figure 1—figure supplement 1G–H).

Putative dopamine neurons in an acute slice preparation from AAV1-Cre-GFP::Robo2lox/lox (mutant) and AAV1-∆Cre-GFP::Robo2lox/lox (control) were identified by the presence of an afterhyperpolarization induced current (Ih) as described (Johnson and North, 1992) (Figure 1—figure supplement 2A), though this is not a definitive feature of all dopamine producing neurons (Margolis et al., 2006). Surprisingly, we did not observe significant changes in either sEPSC frequency or amplitude (Figure 1—figure supplement 2B-DA-C). In contrast, we did observe a significant reduction in sIPSC frequency in mutant mice relative to controls (Figure 1D–E), with no change in sIPSC amplitude (Figure 1D and F).

Due to expression in putative non-dopamine neurons, as well as putative dopamine neurons, Robo2 signaling might also alter the connectivity and/or physiology of these cells. To test this, we recorded sIPSCs and sEPSCs on putative non-dopamine neurons in the VTA (Ih-negative; Figure 1—figure supplement 2E). Similar to putative dopamine neurons, sIPSC frequency was reduced in Ih-negative neurons and sIPSC amplitude was not altered (Figure 1G–I). We also did not observe differences in sEPSCs in these neurons (Figure 1—figure supplement 2F-HD-F).

A reduction in sIPSC frequency following Robo2 inactivation is consistent with a reduction in inhibitory synaptic input to midbrain neurons. Ultrastructural analysis of synaptic contacts using transmission electron microscopy did not reveal gross morphological differences in either symmetrical (inhibitory) or asymmetrical (excitatory) synapses (Figure 1—figure supplement 3A), nor the number of observable synapses (Figure 1—figure supplement 3B); indicating that the reduction in sIPSC frequency is not due to a gross loss of inhibitory synaptic connectivity. Additionally, we did not observe differences in the inactivation kinetics of sIPSCs suggesting that the subunit composition of the postsynaptic GABAA receptors is not altered (Figure 1—figure supplement 3C–F).

Robo2 inactivation in the adult VTA enhances phasic dopamine

GABAergic synaptic transmission potently regulates dopamine neuron activity patterns (Paladini and Tepper, 1999). To determine whether altered inhibitory synaptic inputs to dopamine neurons alters their activity in vivo, we recorded single-unit activity from the VTA of control and mutant mice (Figure 2—figure supplement 1A). Putative dopamine neurons were identified by end of session sensitivity to the D2 agonist quinpirole that could be reversed with the D2 antagonist eticlopride (Figure 2A–B). No differences in quinpirole sensitivity were observed between the two groups (Figure 2—figure supplement 1D–F). We observed no difference between controls and Robo2 VTA mutants in the basal firing rate (Figure 2C), nor did we detect significant differences in the number of burst firing, or phasic events (Figure 2D). In contrast, we did observe a significant increase in the duration of burst events (Figure 2E) and the number of spikes per burst (Figure 2F), consistent with a reduced inhibitory termination of burst epochs.

Figure 2 with 1 supplement see all
Robo2 VTA mutants show increased burst activity.

(A,B) Firing rate histogram of a representative cell displaying inhibition to quinpirole that is reversed with eticlopride in control (A) and mutant animal (B). (Left inset) Waveform of represented cell on each wire of the tetrode. (Right inset) Cumulative spikes of represented cell across time (min). Scale bar = 0.5 ms. (C) Firing rate of dopamine cells in controls (n = 42) and mutants (n = 50) is unchanged. (D) Burst rate of dopamine cells. (E) Burst duration of dopamine cells is longer in mutant cells. Unpaired t-test, t(90) = 2.17, *p<0.05. (F) Spikes per burst in dopamine cells are increased in mutants. Unpaired t-test, t(90) = 2.33, *p<0.05. (G) Firing rate of non-dopamine cells during in vivo electrophysiological recordings in controls (n = 61) and mutants (n = 71). (H) Total normalized distribution of the inter-spike interval (ISI) in controls and mutants. (I) Skew of the ISI in non-dopamine cells is reduced in Robo2 VTA mutants. Unpaired t-test, t(130) = 2.42, *p<0.05. (J) Coefficient of variation of the ISI in non-dopamine cells is reduced in mutants. Unpaired t-test, t(130) = 2.66, **p<0.01. Bar represent mean ± SEM.

https://doi.org/10.7554/eLife.23858.007

To determine whether activity patterns of putative non-dopamine neurons are altered in vivo, we analyzed firing of quinpirole insensitive neurons recorded on the same tetrodes as quinpirole sensitive neurons. Similar to putative dopamine neurons, quinpirole insensitive neurons did not show changes in overall firing rate (Figure 2G). However, these cells did show a significant change in the skew of their interspike-interval distribution (ISI, Figure 2H–I). Consistent with this change in ISI distribution, quinpirole insensitive neurons showed a reduction in the coefficient of variation of the ISI (CV-ISI, Figure 2J), indicative of a reduced number of inhibitory pauses.

We next sought to determine whether changes in dopamine activity patterns result in altered dopamine release. To test this, we performed fast scan cyclic voltammetry (FSCV) to monitor dopamine release in the nucleus accumbens (NAc) in Robo2 mutant mice. Dopamine release was measured in response to increasing stimulus intensity delivered to an excitatory afferent input, the pedunculopontine tegmental area (PPTg) in urethane-anesthetized animals (Figure 3A–B; Figure 3—figure supplement 1A) as previously described (Soden et al., 2013). At lower stimulus intensities, dopamine release was significantly elevated in Robo2 mutant mice compared to controls (Figure 3C–H). We did not observe significant differences in dopamine release evoked by direct stimulation of the medial forebrain bundle (MFB; Figure 3—figure supplement 1E–J), suggesting that dopamine transporter function is not altered in Robo2 mutant mice. Consistent with a lack of observed differences in the overall firing rate of dopamine neurons (Figure 2C), we did not detect differences in tonic dopamine release as measured by microdialysis (Figure 3—figure supplement 1K–L).

Figure 3 with 1 supplement see all
Robo2 VTA mutants have increased phasic dopamine release.

(A) Schematic of experimental design showing stimulation of the PPTg and FSCV recording in the NAc. (B) Representative voltammogram of dopamine recorded in the NAc in control (left) and mutant (right). (C–G) Average dopamine oxidation current recorded in NAc after PPTg stimulation is increased in mutants (n = 6) compared to controls (n = 5). (C) 50 μA: Two-way repeated measures ANOVA, effect of genotype, F(1,9) = 12.75, **p<0.01. (D) 100 μA: Two-way repeated measures ANOVA, effect of genotype, F(1,9) = 5.84, *p<0.05. (E) 150 μA: Two-way repeated measures ANOVA, effect of genotype, F(1,9) = 8.65, *p<0.05. (F) 200 μA. (G) 250 μA. (H) Area under the curve of the dopamine current across all stimulation intensities. Two-way repeated measures ANOVA, effect of genotype, F(1,9) = 5.99, *p<0.05. Bar represent mean ± SEM.

https://doi.org/10.7554/eLife.23858.009

Inactivation of Robo2 in the adult VTA causes psychomotor dysfunction

To test whether inactivation of Robo2 in the adult VTA has a functional impact on dopamine-dependent behavioral regulation, we monitored Robo2 mutant and control mice in a series of behavioral tests. Both AAV1-∆Cre-GFP injected Robo2lox/lox mice and AAV1-Cre-GFP injected Robo2+/+ mice were used as controls. We first tested locomotor activity across the day-night cycle. Robo2 mutants were significantly hyperactive compared to controls (Figure 4A), particularly during the night (Figure 4B). Consistent with altered locomotion in mutant mice, we observed reduced forelimb separation and greater print separation (Figure 4—figure supplement 1A–F).

Figure 4 with 2 supplements see all
Behavioral characterization of Robo2 VTA mutants show phenotypes in locomotion and cocaine sensitization.

(A) Locomotion measured across three consecutive days and nights showing mutants (n = 31) are hyperactive relative to controls (n = 36). Two-way repeated measures ANOVA, genotype x time interaction, F(252,16380) = 2.53, ****p<0.0001. (B) Total distance traveled across three consecutive nights. Two-way repeated ANOVA, effect of genotype, F(1,65) = 12.49, ***p<0.001, Bonferroni’s multiple comparison test, ***p<0.001, **p<0.01. (C–D) Locomotor response to cocaine (20 mg/kg) in controls (n = 20) and mutants (n = 11) on day 1 (C), Two-way repeated measures ANOVA, genotype x time interaction, F(35,1015) = 7.10, ****p<0.0001 and day 5 (D), Two-way repeated measures ANOVA, genotype x time interaction, F(35,1015) = 8.23, ****p<0.0001. (E) Normalized locomotor response to cocaine by subtracting 60 min pre-cocaine from 60 min post-cocaine. Two-way repeated measures ANOVA, effect of genotype, F (1,29) = 18.26, ***p<0.001, Bonferroni’s multiple comparison test, ****p<0.0001, **p<0.01. (F) Locomotor response to amphetamine (2.5 mg/kg) in controls (n = 19) and mutants (n = 17). Two-way repeated measures ANOVA, genotype x time interaction, F(35,1190) = 6.43, ****p<0.0001. (G) Locomotor response to the DAT blocker, GBR-12909 (10 mg/kg) in controls (n = 19) and mutants (n = 17). Two-way repeated measures ANOVA, genotype x time interaction, F(35,1190) = 2.94, ****p<0.0001. (H) Normalized locomotor response to amphetamine and GBR-12909 by subtracting 60 min pre-drug from 60 min post-drug. AMP: Unpaired t-test, t(34) = 2.968 **p<0.01,. GBR: Unpaired t-test, t(34) = 2.061, *p<0.05. (I) Discrimination between a CSp and CSm is impaired in AAV1-Cre-GFP::Robo2lox/lox animals (n = 16) compared to controls (n = 22). Two-way repeated measures ANOVA, genotype x time interaction, F(6,216) = 2.40, *p<0.05, Bonferroni’s multiple comparison test, **p<0.01 and *p<0.05. Bars represent mean ± SEM.

https://doi.org/10.7554/eLife.23858.011

Increased locomotor activity in Robo2 mutant mice suggests these animals may have enhanced psychomotor activation. To test this, control and Robo2 mutant mice were subjected to daily injections of cocaine (20 mg/kg) for 5 days. In control mice, we observed a characteristic locomotor sensitization to cocaine (Figure 4C–E). In contrast, Robo2 mutant mice showed an inversion in the normal locomotor response to cocaine (Figure 4C–E), decreasing rather than increasing their activity; this pattern persisted across days. To determine whether the inverted psychomotor response to cocaine in Robo2 mutant mice leads to an anhedonic as opposed to a hedonic state, we measured conditioned place preference (CPP) to cocaine. Although cocaine CPP was not as robust in mutant mice compared to controls (Figure 4—figure supplement 2A), it did not lead to an avoidance of the cocaine-paired side, suggesting that although their psychomotor response is altered, the hedonic aspect of cocaine is not significantly changed.

An inverted psychostimulant response to cocaine is remarkably similar to that reported following genetic inactivation of the dopamine transporter (Slc6a3-KO mice) (Gainetdinov et al., 1999). Analysis of Slc6a3-KO mice demonstrated that these animals not only have inverted responses to psychostimulants, but also a variety of monoamine transport blockers (Gainetdinov et al., 1999). Similar to their response to cocaine, Robo2 mutant mice had inverted locomotor responses to amphetamine (AMP) and the DAT blocker GBR12909 (GBR; Figure 4F–H). Robo2 mutant mice also had a differential response to the serotonin transport blocker fluoxetine with the mutants showing greater inhibition, but the norepinephrine transport blocker nisoxetine reduced activity equally in control and mutant mice (Figure 4—figure supplement 2B–E).

In addition to regulating psychomotor activation, VTA dopamine neurons play an essential role in regulating reward-related learning. To establish whether alterations in the dopamine system of Robo2 mutant mice affect this process, we monitored their performance in a Pavlovian associative learning paradigm. Robo2 mutant mice showed significant deficits in this task compared to control mice, with mutant mice failing to discriminate between predictive (CSplus) and non-predictive (CSminus) cues (Figure 4I). This impairment was not associated with a reduction in reward seeking as total head entries into the reward receptacle did not differ (Figure 4—figure supplement 2F).

Hyperactivity and cognitive or learning impairments associated with enhanced phasic dopamine release may be related to a psychosis-like state in mutant mice (Grace, 1991). To address this, we monitored pre-pulse inhibition of the acoustic startle reflex and enhanced sensitivity to the psychomimetic NMDA receptor antagonist MK-801, two metrics used to assess psychosis-related phenotypes in animal models (Geyer, 2008). Robo2 mutant mice did not differ from control mice in PPI, and although they are hyperactive at baseline, they do not show increased sensitivity to MK-801. (Figure 4—figure supplement 2G–H).

Behavioral dysfunction requires Robo2 inactivation globally in the VTA

Alterations in inhibitory control of dopamine neurons, enhanced phasic dopamine release, behavioral responses to psychostimulant drugs, and impaired reward learning all point to dopaminergic phenotypes following Robo2 inactivation in the VTA. To establish whether the observed behavioral phenotypes are the result of inactivation in dopamine producing neurons of adult mice, we crossed Robo2lox/lox mice to mice containing an inducible Cre coupled to the estrogen receptor (iCreER) under control of the dopamine transporter promotor (Slc6A3-iCre/ERT2). Treatment of adult mice with the estrogen receptor agonist tamoxifen (75 mg/kg x 5 days) resulted in Cre-mediated recombination in dopamine neurons of the VTA and SNc (Figure 5A). Behavioral analysis of tamoxifen treated DAT-iCreER::Robo2lox/lox mice did not result in hyperactivity, altered psychostimulant response to cocaine, or learning deficits relative to sham injected Slc6A3-iCre/ERT2::Robo2lox/lox mice (Figure 5B–C; Figure 5—figure supplement 1A–D).

Figure 5 with 1 supplement see all
Adult or embryonic inactivation of Robo2 in dopamine neurons does not recapitulate the Robo2 VTA mutant phenotypes.

(A) Schematic showing expression of tamoxifen-inducible Cre specifically in adult dopamine neurons. Representative images from Slc6a3-iCreERT2 crossed to a Cre-dependent reporter line, Gt(ROSA)26Sortm(CAG-tdTomato)Hze, shows induction of the tdTomato reporter is specific to animals receiving tamoxifen injections. Scale bar = 100 µm. (B) Locomotion measured across three consecutive days and nights showing mutants (n = 12) are normal compared to control animals (n = 12). (C) Discrimination between CSp and CSm (left) is normal in Slc6a3-iCreERT2::Robo2lox/lox animals (n = 12) compared to controls (n = 12). (D) Schematic showing expression of Cre in dopamine neurons starting in embryonic development and continuing into adulthood. (E) Locomotion measured across three consecutive days and nights showing mutants (n = 12) are normal compared to control animals (n = 11). (F) Discrimination between CSp and CSm (left) is normal in Slc6a3-Cre::Robo2∆/lox animals (n = 11) compared to controls (n = 12). (G) Schematic of strategy for injecting AAV1-Cre-GFP into the midbrain of Slc6a3-Cre; Robo2∆/lox animals. (H) Locomotion measured across three consecutive days and nights showing mutants (n = 13) are hyperactive relative to controls (n = 15). Two-way repeated measures ANOVA, genotype x time interaction, F(252,6552) = 3.17, ****p<0.0001. (I) Discrimination between a CSp and CSm is impaired in AAV1-CreGFP::Slc6a3-Cre::Robo2∆/lox animals (n = 6) compared to controls (n = 8). Two-way repeated measures ANOVA, genotype x time interaction, F(6,72) = 2.68, *p<0.05, Bonferroni’s multiple comparison test, **p<0.01. (J, K) Locomotor response to cocaine (20 mg/kg) in controls (n = 15) and mutants (n = 13) on day 1 (d), Two-way repeated measures ANOVA, genotype x time interaction, F (35,910)=6.57, ****p<0.0001 and day 5 (e), Two-way repeated measures ANOVA, genotype x time interaction, F (35,910)= 17.03, ****p<0.0001. (L) Normalized locomotor response to cocaine by subtracting 60 min pre-cocaine from 60 min post-cocaine on day 1 and day 5, Two-way repeated measures ANOVA, genotype x time interaction, F (1,26)= 10.97, **p<0.01, Bonferroni’s multiple comparison test, ****p<0.0001, **p<0.01. Bars represent mean ± SEM.

https://doi.org/10.7554/eLife.23858.014

The observed lack of behavioral effects in Slc6A3-iCre/ERT2::Robo2lox/lox mice was unexpected, but could reflect an inefficiency in Robo2 inactivation. The conventional Cre-driver mouse line (Slc6A3-Cre) has been previously shown to effectively inactivate genes in virtually all dopamine producing neurons (Engblom et al., 2008; Zweifel et al., 2008). One caveat to this mouse line is that the early expression of Cre may result in developmental compensations (Engblom et al., 2008). Nonetheless, we generated conditional, non-inducible knockouts of Robo2 in dopamine neurons (Slc6A3-Cre::Robo2∆/lox mice, Figure 5D). Consistent with inactivation of Robo2, Ribotag analysis following injection of an AAV containing Cre-dependent Rpl22HA (AAV1-FLEX-Rpl22HA) demonstrated a significant reduction in Robo2 mRNA in dopamine neurons (Figure 5—figure supplement 1E); consistent with nonsense-mediated mRNA decay (Gibson et al., 2014). Similar to Slc6A3-iCre/ERT2::Robo2lox/lox mice, Slc6A3-Cre::Robo2∆/lox mice did not show hyperactivity, altered psychostimulant response to cocaine, or learning deficits relative to control mice (DAT-iCreER::Robo2∆/+) (Figure 5E–F; Figure 5—figure supplement 1F–I).

The lack of observed effects in Slc6A3::Robo2∆/lox mice could be the result of developmental compensation as discussed above, and if this were true then injection of AAV1-Cre-GFP into the VTA of Slc6A3::Robo2∆/lox mice should not result in behavioral alterations due to an occlusion-like effect. To test this, adult Slc6A3::Robo2∆/lox mice were injected with AAV1-Cre-GFP into the VTA (AAV1-Cre-GFP::Slc6A3-Cre::Robo2∆/lox mice, Figure 5G). Similar to injection of AAV1-Cre-GFP into the VTA of Robo2lox/lox mice, AAV1-Cre-GFP::Slc6A3-Cre::Robo2∆/lox mice were hyperactive, had inverted cocaine responses, and deficits in Pavlovian associative learning (Figure 5H–L; Figure 5—figure supplement 1J–K). These results indicate that Robo2 inactivation is required in non-dopamine producing neurons to induce the observed behavioral phenotypes.

Loss of local GABAergic control of the VTA evokes psychomotor dysfunction

Changes in the inhibitory control of both dopamine and non-dopamine neurons in the VTA following Robo2 inactivation suggests that local GABAergic control of the VTA is responsible for the observed behavioral alterations. To test this, we selectively silenced GABAergic neurons in the VTA through conditional expression of the light-chain of tetanus toxin fused to GFP (GFP-TetX) to block synaptic transmission (Kim et al., 2009). An AAV1 vector containing a Cre-dependent expression cassette for GFP-TetX (AAV1-FLEX-GFP-TetX, [Han et al., 2015]) was injected into the VTA of mice expressing Cre under the control of the endogenous VGAT locus, Slc6a32A1-Cre mice (Figure 6A). Unexpectedly, inactivation of GABAergic transmission in the VTA of Slc6a32A1-Cre::AAV1-FLEX-GFP-TetX mice resulted in rapid weight loss and compromised health in a number of animals (11 out of 20 fell below 80% of their pre-surgery bodyweight and were removed from the study) that could not be further assessed for behavioral changes. Of the remaining Slc6a32A1-Cre::AAV1-FLEX-GFP-TetX mice (9 out of 20) that did not show dramatic weight loss, possibly due to a reduced efficiency of GABA inactivation, we observed hyperactivity and inverted locomotor responses to cocaine (Figure 6B–E). Post-hoc histological analysis showed robust viral expression within the VTA, so it is unclear why the surviving animals were less affected; however, behaviorally testing the survivors possibly caused a selection bias that underestimated the severity of the phenotype. These findings suggest that reduced GABAergic control of the VTA following Robo2 inactivation is likely the cause of hyperactivity and altered behavioral responses to psychostimulants.

Midbrain GABA neuron inactivation recapitulates Robo2 VTA mutant phenotypes.

(A) Schematic of strategy for injecting AAV1-GFP-TetX into the midbrain of Vgat-Cre animals (left). Schematic of Cre-dependent AAV1-GFP-TetX (top). Representative image (bottom) showing viral transduction is specific to non-dopamine neurons (TH- neurons). Scale bar = 100 µm (25 µm in zoom). (B) Locomotion measured across three consecutive days and nights showing mutants (n = 9) are hyperactive relative to controls (n = 6). Two-way repeated measures ANOVA, genotype x time interaction, F(252,3276) = 1.35, ***p<0.001. (C,D) Locomotor response to cocaine (20 mg/kg) in controls (n = 6) and mutants (n = 9) on day 1 (C), Two-way repeated measures ANOVA, genotype x time interaction, F(35, 455) = 1.34, p=0.0987, and on day 5 (D), Two-way repeated measures ANOVA, genotype x time interaction, F(35, 455) = 2.86, p<0.0001. (E) Normalized locomotor response to cocaine by subtracting 60 min pre-cocaine from 60 min post-cocaine on day 1 and day 5, Two-way repeated measures ANOVA, genotype x time interaction, F(1,13) = 5.35, *p<0.05. Bars represent mean ± SEM.

https://doi.org/10.7554/eLife.23858.016

Discussion

Our results provide the first demonstration that Robo2 has a critical function beyond early development of the nervous system by potently regulating GABAergic synapses. The observed behavioral phenotype of inverted psychostimulant responses is a rare genetic finding. Numerous other genetic models have been described that cause hyperactivity, but few have inverted psychostimulant responses (Leo and Gainetdinov, 2013). Given the link between hyperactivity and paradoxical calming by psychostimulants in ADHD, it is attractive to think that Robo2 regulation of inhibitory control of the midbrain may play an important role in this disorder. Although selective inactivation of Robo2 exclusively in the adult VTA is not a true genetic model of ADHD, our results strongly implicate balanced excitatory and inhibitory control of the midbrain as a likely source of ADHD-like phenotypes. Similar to inactivation of the dopamine transporter (Gainetdinov et al., 1999), we observed increased sensitivity to the serotonin transport blocker fluoxetine, suggesting the hyperactivity associated with loss of Robo2 is due to a potential enhancement of the serotonin system.

We find that Robo2 is highly expressed in the VTA in both dopamine and non-dopamine producing neurons. Given the dopamine-related nature of the phenotypes we observed, we were surprised to find that inactivation of Robo2 exclusively in dopamine neurons was not sufficient to invoke these phenotypes, particularly given the partial enrichment of Robo2 mRNA in these cells. Instead our data suggests that Robo2 is required in multiple cell types for the observed behavioral changes. There are numerous cell types within the VTA, including dopamine neurons, GABAergic interneurons, GABAergic projection neurons, and glutamatergic neurons (Barker et al., 2016). Because the phenotype emerged only when dopamine and non-dopamine neurons were genetically targeted, it is possible that Robo2 is required globally within the VTA. Unfortunately, Cre driver lines that allow for the inducible expression of Cre recombinase only in GABA or glutamate neurons of the VTA do not currently exist that would allow us to test this in greater detail. However, given that many of the observed behavioral deficits are dopamine-dependent and the genetic manipulation alters synaptic physiology in both dopamine and non-dopamine neurons, it is highly likely that the observed deficits are the results of a global change in GABAergic connectivity in the VTA. Consistent with the function of GABA in this process, reducing synaptic release from midbrain GABA neurons recapitulated the adult VTA Robo2 mutant phenotype, though the effects were much more severe overall. These findings point to a key role of GABAergic transmission in the VTA for behavioral hyperactivity and learning deficits. Given the high degree of specificity for partial GABA agonists in targeting GABA receptors with specific subunit compositions (Hoestgaard-Jensen et al., 2014), drugs targeting specific GABAA receptors in the VTA may be a promising therapeutic avenue for multiple disorders.

Additional studies are required to understand the molecular mechanisms whereby Robo2 signaling controls inhibitory synapses in the VTA. The reduced frequency of sIPSCs following Robo2 inactivation is consistent with a requirement for Robo2 in presynaptic GABA release. It was recently shown that Robo receptors interact with A-kinase anchoring proteins (AKAPs) (Samelson et al., 2015), placing Robo receptors in key scaffolding complexes. It has also been shown that AKAP 79/150 is expressed in dopamine and non-dopamine neurons in the VTA where it regulates GABAergic synaptic transmission. Intriguingly, disruption of AKAP function induces a long-term depression like effect on GABAergic synapse (Dacher et al., 2013). Thus, an intriguing possibility is that Robo2 receptors interact with AKAPs to stabilize inhibitory synaptic connectivity. Another, non-mutually exclusive possibility is that Robo2 interacts with Robo1 either in cis or in trans to stabilize synapses. Both Robo1 and Robo2 are expressed in the VTA, and it has been demonstrated in Drosophila that during commissural axon pathfinding, Robo2 interacts with Robo1 in trans to suppress the repulsive properties of Robo1 (Evans et al., 2015). Future exploration into these potential mechanisms will provide important insight into this process.

In addition to expression in the ventral midbrain, Robo2 is highly expressed in other brain regions, most notably the limbic system including the prefrontal cortex, striatum, and hippocampus. Whether adult Robo2 signaling functions in a similar fashion in these regions will be important to elucidate, particularly given the linkage between Slit/Robo signaling and mental illness. Our results also suggest that Robo2 receptors may play an important role in regulating synaptic changes associated with drugs of abuse. Interestingly, during cocaine sensitization, changes in spine density in the striatum results from inhibition of the small Rho-family GTPase, Rac1, which is a known effector molecule for Robo signaling, and is a potent regulator of spine turnover during development (Dietz et al., 2012). In addition, cocaine sensitization reduces inhibitory synaptic transmission in the VTA (Liu et al., 2005) that may be mediated in part by Robo2 signaling. Determining how Slit/Robo signaling regulates synaptic processes in other cell types will be an important next step towards further resolving its function in the adult nervous system.

Materials and methods

Animals

Animals were group housed (max 5 animals/cage), maintained on a 12:12 light:dark cycle, and given ad libitum access to food and water except during food restriction when they were restricted to 85% of their ad libitum bodyweight. All behavioral experiments were performed during the light cycle. Equal numbers of C57Bl/6 male and female mice (8–12 weeks old) were used for all experiments. Animals were assigned to experimental groups to randomize for sex, age, and genotype. The experimenter was blinded to the genotype during data collection, and all data was generated from at least two independent experiments. The sample sizes were chosen based upon the predicted variability of the assay. Two to six independent experiments were performed for each behavioral test. In at least one of these experiments, the animals underwent a single behavioral test. When multiple tests were performed with the same group of animals, an additional cohort of animals had the behavioral tests reversed to counterbalance any interaction between tests; however, all pharmacology studies occurred subsequent to non-invasive behavioral tests (ie Pavlovian conditioning). The Robo2lox/lox mouse line was generated as described (Lu et al., 2007). The delta allele of Robo2∆/lox was generated by crossing the Robo+/lox to the Meox2-Cre mouse line to recombine the Robo2 lox allele. This approach was taken with the Slc6a3-Cre line to prevent mosaicism caused by potential transient Cre expression in the gametes of the Slc6a3-Cre line. The Slc6a3-iCreERT2, Slc6a32A1-Cre, and Gt(ROSA)26Sortm(CAG-tdTomato)Hze mouse lines were obtained from Jackson Laboratory. To induce expression of Slc6a3-iCreERT2, daily injections of tamoxifen at 75 mg/kg were performed for five days (Sigma-Aldrich). All experiments were done in accordance with protocols approved by the University of Washington Animal Care and Use Committee.

Stereotaxic viral injections

Request a detailed protocol

AAV1 virus was prepared as described (Gore et al., 2013). For stereotaxic injections into the brain, either AAV1-Cre-GFP, AAV1-∆Cre-GFP, or AAV1-FLEX-GFP-TetX was injected in the VTA (Relative to bregma: M-L = 0.5, A-P = −3.25*F, D-V = −4.5, F= (Bregma-Lambda)/4.21). For the generation of control mice, either a virus containing a partially deleted, nonfunctional version of Cre (AAV1-∆Cre-GFP) was injected into Robo2lox/lox animals (control for Robo2lox/lox genotype) or AAV-Cre-GFP was injected into wild type mice (control for viral Cre expression). Equal numbers of these two control groups were generated and showed no statistical difference, so they are presented as a single combined ‘control’ group. Animals were allowed to recover for 2 weeks prior to testing to allow for expression of the virus. Exclusion criteria was based upon histological examination that verified the expression of the virus in the VTA.

Ribotag

Request a detailed protocol

For Slc6a3Cre;Rpl22HA+/lox and Slc32a1Cre;Rpl22HA+/lox experiments, tissue was collected from mice 8–10 weeks of age. For AAV1-FLEX-Rpl22HA injected Slc6a3Cre::Robo2∆/lox injected mice, tissue was collected 3 weeks following injection to allow for sufficient HA-tagged Rpl22 incorporation. Tissue punches of the ventral midbrain were homogenized as described previously (Sanz et al., 2009). Homogenized tissue from individual mice were incubated with 5 µl of anti-HA antibody (Covance) coupled to 200 µl of magnetic beads (Pierce) overnight at 4°C. Following elution from magnetic beads, RNA from both immunoprecipitated (IP) samples and non-HA tagged (input) samples was obtained using the RNeasy micro kit (Qiagen) according to manufacturer's directions. Total RNA was quantified using a Ribogreen RNA assay kit (Invitrogen). cDNA was generated with Superscript IV (Invitrogen) using oligo dT primers from equal amounts of starting RNA, according to the manufacturer protocol. IP purified mRNA was converted to cDNA and analyzed via qRT-PCR. TaqMan (Applied Biosystems) primers against Robo2 and Actb were used to detect expression levels. Relative expression values were obtained using the comparative CT method and normalized to Actb levels. Fold enrichment was calculated as the IP versus input ratio and represented the amount of the transcript in the targeted cell type (IP) when compared to equal amounts of RNA from the input.

Behavior

Pavlovian conditioning

Request a detailed protocol

Mice underwent 7 days of conditioning, receiving 25 conditioned stimulus plus (CSp) and 25 conditioned stimulus minus (CSm) each day. The CSp and CSp were 10 s tones that were either 10 kHz or 20 kHz with a two minute intertrial interval, and animals groups receiving each tone were counterbalanced. The CSp auditory tone co-terminated with the delivery of food reward pellet 100%, whereas the CSm auditory tone co-terminated with the delivery of a food rewards pellet only 12.5%.

Basal locomotion and pharmacology studies

Request a detailed protocol

For all locomotor studies, standard Allentown cages were inserted into the locomotion chamber, and activity was measured by infrared beam breaks (Columbus Instruments). For Day/Night locomotion, mice were provided ad libitum access to food and water. For pharmacological studies, animals were removed from their home cage and placed into the locomotion chamber. Locomotor activity was measured for 90 min before injections with a single compound. The compounds tested were quinpirole (0.2 mg/kg; Sigma-Aldrich), eticlopride (0.05 mg/kg; Sigma-Aldrich), nisoxetine (10 mg/kg; Tocris Bioscience), fluoxetine (20 mg/kg; Tocris Bioscience), AMPH (2.5 mg/kg; Sigma-Aldrich), cocaine (20 mg/kg; Sigma-Aldrich), GBR-12909 (10 mg/kg; Tocris Bioscience), and MK-801 (0.2 mg/kg; Sigma-Aldrich).

Prepulse inhibition

Request a detailed protocol

Mice were tested in an acoustic startle chamber as described (Soden et al., 2013). Briefly, mice underwent 10 trials of different prepulse amplitudes, including 0 dB, 70 dB, 75 dB, and 80 dB, which preceded a 120 dB startle amplitude. The prepulse inhibition was calculated as the % reduction in startle at each prepulse intensity compared to no prepulse.

Cocaine conditioned place preference

Request a detailed protocol

A three chamber CPP arena was used in an unbiased cocaine CPP paradigm (Hnasko et al., 2007). In brief, one side had vertical stripes and white alpha-dri bedding, and the other side had horizontal stripes and regular bedding. On Day 1, animals were placed in the center chamber and given 30 min to explore all three chambers. On the drug pairing days (Day 2 and Day 3), animals were injected with saline in the morning and confined for 30 min to one side. In the afternoon, cocaine was injected at 15 mg/kg and the animals were confined to the drug-paired side for 30 min. On the fourth day, the animals were allowed to freely explore all chambers. The % CPP was calculated as the % non-center time spent in the drug-paired side.

Footprint analysis

Request a detailed protocol

The footprint analysis was performed as described (Crawley, 2008). In brief, forelimbs were dipped in green paint (Crayola) and the hindlimbs were dipped in red paint. The mice ran on a 6 × 55 cm piece of white paper elevated to a 10° incline. For each animal, three independent measures were made for each parameter comparing the centroid of the left and right forelimb or hindlimb.

In vitro slice electrophysiology

Request a detailed protocol

Whole-cell recordings were made using an Axopatch 700B amplifier (Molecular Devices) with filtering at 1 KHz using 4–6 MΩ electrodes. Horizontal brain slices (200 μm) were prepared from 6 week old mice in an ice slush solution containing (in mM): 92 NMDG, 2.5 KCl, 1.25 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 0.5 CaCl2, 10 MgSO4, pH 7.3–7.4 (Ting et al., 2014). Slices recovered for ~12 min in the same solution at 32°C and then were transferred to a room temperature solution including (in mM): 92 NaCl, 2.5 KCl, 1.25 NaH2PO4, 30 NaHCO3, 20 HEPES, 25 glucose, 2 thiouria, 5 Na-ascorbate, 3 Na-pyruvate, 2 CaCl2, 2 MgSO4. Slices recovered for an additional 60 min. All solutions were continually bubbled with O2/CO2, and all recordings were made in ACSF containing (in mM): 126 NaCl, 2.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2 11 D-glucose, 18 NaHCO3, 2.4 CaCl2 at 32°C continually perfused over slices at a rate of ~2 ml/min. Ih currents were induced by 2 s hyperpolarizing voltage steps from −70 mV to −120 mV. All recorded cells were medial and proximal to the MT, where the presence of an Ih current correlates highly with dopamine neuron identity (Margolis et al., 2006).

For recording spontaneous EPSCs, electrodes were filled with an internal solution containing (in mM): 130 K-gluconate, 10 HEPES, 5 NaCl, 1 EGTA, 5 Mg-ATP, 0.5 Na-GTP, pH 7.3, 280 mOsm. 200 mM picrotoxin was included in the ACSF to inhibit GABAA receptor-mediated events. For recording spontaneous IPSCs, electrodes were filled with an internal solution containing (in mM): 135 KCl, 12 NaCl, 0.05 EGTA, 100 HEPES, 0.2 Mg-ATP, 0.02 Na-GTP, pH 7.3, 280 mOsm. 2 mM kynurenic acid was included in the ACSF to inhibit AMPA and NMDA receptor-mediated events. Cells were held at −60 mV, and access resistance was monitored throughout all experiments.

To calculate the Tau (τw) for the sIPSCs, we calculated the weighted time constant, τw, for 10–90% of the rise time by generating an average waveform for each cell using the equation, τw =(A1 τ1)+(A2 τ2), where τ1 and τ2 are the time constants of the first and second exponential functions, respectively, and A1 and A2 are the proportion of the amplitude of the IPSC of the first and second exponential functions, respectively as described (Eyre et al., 2012).

In vivo electrophysiology

Request a detailed protocol

Four recording tetrodes were made from tungsten wires (25 µm in diameter; California Fine Wire Company) and mounted on a custom-built Microdrive (Jo et al., 2013, Jo and Mizumori, 2016). Each wire was gold-plated to adjust its final impedance to 200–400 KΩ (tested at 1 kHz). The microdrive was surgically implanted in one hemisphere of the VTA after bilateral virus infusion of AAV-Cre-GFP or AAV- ∆Cre-GFP into the VTA. Two weeks after surgery, each mouse was placed in a rectangular box (21.6 cm × 17.8 cm) and single-unit activity was daily monitored as follows: the microdrive was connected to a preamplifier and its outputs were transferred to a RZ5 BioAmp processor (Tucker-Davis Technologies). Unit signals were filtered between 0.3 and 5 kHz, and digitized at 24 kHz. Tetrodes were advanced in 40 µm increments, up to 160 µm per day. Once stable and isolated units were found, their sensitivity to quinpirole and eticlopride was tested to identify putative dopamine cells (Zweifel et al., 2009). A VTA cell was classified as putative dopaminergic if it exhibited a low basal firing rate (less than 12 Hz) and showed severe inhibition by quinpirole (0.2 mg/ml, i.p.; ≥70% inhibition of the basal frequency) and subsequent recovery by eticlopride (0.5 mg/ml, i.p.; ≥70% the basal frequency). The other neurons that did not meet the criteria were considered as putative non-dopaminergic cells. VTA neurons were recorded for 20 min to determine baseline firing properties. Spikes from these single neurons were isolated by cluster analysis using Offline Sorter (Plexon). Subsequent data analyses, such as comparison of average firing rates and burst rates were performed with Matlab software (Mathworks, Zweifel, 2017 https://github.com/zweifellab/ephys with a copy archived at https://github.com/elifesciences-publications/ephy). The onset of a burst was identified as two consecutive spikes with an interspike interval of <80 ms and its termination was defined as the an interspike interval of >160 ms (Grace and Bunney, 1984).

Fast scan cyclic voltammetry

Request a detailed protocol

FSCV was performed using carbon fiber microelectrodes encased in fused-silica capillary tubing (Polymicro Technologies) (Clark et al., 2010). The carbon fiber microelectrode was positioned in the dorsomedial striatum at stereotaxic coordinates: A-P = 1.2 mm, M-L = 1.2 mm, D-V = 3.5 below the dura and a waveform was applied at a frequency of 60 Hz. After 40 min, the waveform frequency was reduced to 10 Hz for 10 min or until baseline recordings stabilized. Next, the carbon fiber microelectrode was lowered into the nucleus accumbens in 0.05 mm increments until maximum dopamine release was observed (D-V coordinates ranged from 3.60–3.98 below bregma, average 3.80). An Ag/AgCl reference electrode was placed in the contralateral hemisphere. The stimulation electrode (Plastics One) was placed above the PPTg at stereotaxic coordinates: A-P: −4.6 mm, M-L: 0.8 mm and lowered in 0.1 mm increments from 2.5 mm below the dura until optimal dopamine release was observed after 1 s stimulations of 60 Hz (average D-V coordinate: 2.7 mm below the dura). For the MFB stimulations, the stereotaxic coordinates were A-P: 2.4 mm, M-L: 1.1 mm, DV: 0.1 increments from 4 mm.

Microdialysis-UPLC

Request a detailed protocol

Microdialysis-UPLC experiments were modified for mouse tissue from a protocol previously described (Schindler et al., 2016). In brief, a 1 mm long microdialysis probe (BASi instruments) was equilibrated in artificial cerebrospinal fluid (ACSF) containing 154.7 mM Na+, 0.82 mM Mg+2, 2.9 mM K+, 132.49 mM Cl, 1.1 mM Ca+2, and 5.9 mM D-glucose, and pH to 7.4 with CO2. The probe was inserted unilaterally into the nucleus accumbens (M-L = +1.0, A-P = +1.3*F, D-V = 4.0, F=(distance between lambda and bregma)/4.21). The probe was flushed with ACSF at a rate of 2 µl/min for 90 min, followed by 1 µl/min for 30 min. Four 20 min fractions were collected. An equal volume of dialysate was mixed with an internal standard solution containing 100 pg/µl of deuterated D4 (2-(3,4-dihydroxyphenyl)ethyl-1,1,2,2-d4-amine HCL) 0.01 N perchloric acid and frozen on dry ice until run on the UPLC. For the detection of dopamine on the UPLC, serial dilutions of purified D4 dopamine samples of known concentrations were first analyzed to detect the retention time for dopamine on a Xevo TQ-S mass spectrometer. Samples for controls and Robo2 VTA mutants were then ran in an interleaved manner to control for any column artifact. The dopamine and D4 peaks were identified and the area under the curve was calculated for each sample.

Histology

Request a detailed protocol

Animals were deeply anesthetized with beuthanasia and transcardially perfused with 4% paraformaldehyde (PFA). The brain was removed and fixed overnight in 4% PFA before being cryoprotected in 30% sucrose. 30 μm frozen sections were made on a Leica cryostat and fluorescence imaging was performed on a Nikon upright microscope. For immunohistochemistry, the following antibodies were used: Tyrosine hydroxylase (rabbit polyclonal, 1:2000, Millipore, or mouse monoclonal, 1:2000, Millipore), GFP (mouse monoclonal, 1:1000, Invitrogen or chicken polyclonal, 1:1000, Abcam), Robo2 (rabbit polyclonal, 1:1000, Abcam (discontinued), and dsRed (rabbit, 1:1000, Clontech). For transmission electron microscopy, tissue was fixed overnight in 4% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, and then postfixed in 1% osmium tetroxide, washed, en bloc stained in 1% uranyl acetate, embedded in epoxy resin, and then sectioned to 80 nm.

Data analysis

Request a detailed protocol

Statistical analyses were performed in Prism (GraphPad), Excel (Microsoft), and Matlab (Mathworks). For data that was normally distributed with equal variance between groups, an unpaired t-test or two-way repeated measures ANOVA was performed. For data that was not normally distributed (Figure 1—figure supplement 2E), a non-parametric Mann-Whitney test was performed. The sample sizes were chosen based upon previously established variability of the assay. The images were processed using either Photoshop (Adobe).

References

    1. Grace AA
    2. Bunney BS
    (1984)
    The control of firing pattern in nigral dopamine neurons: burst firing
    Journal of Neuroscience 4:2877–2890.
    1. Okbay A
    2. Beauchamp JP
    3. Fontana MA
    4. Lee JJ
    5. Pers TH
    6. Rietveld CA
    7. Turley P
    8. Chen GB
    9. Emilsson V
    10. Meddens SF
    11. Oskarsson S
    12. Pickrell JK
    13. Thom K
    14. Timshel P
    15. de Vlaming R
    16. Abdellaoui A
    17. Ahluwalia TS
    18. Bacelis J
    19. Baumbach C
    20. Bjornsdottir G
    21. Brandsma JH
    22. Pina Concas M
    23. Derringer J
    24. Furlotte NA
    25. Galesloot TE
    26. Girotto G
    27. Gupta R
    28. Hall LM
    29. Harris SE
    30. Hofer E
    31. Horikoshi M
    32. Huffman JE
    33. Kaasik K
    34. Kalafati IP
    35. Karlsson R
    36. Kong A
    37. Lahti J
    38. van der Lee SJ
    39. deLeeuw C
    40. Lind PA
    41. Lindgren KO
    42. Liu T
    43. Mangino M
    44. Marten J
    45. Mihailov E
    46. Miller MB
    47. van der Most PJ
    48. Oldmeadow C
    49. Payton A
    50. Pervjakova N
    51. Peyrot WJ
    52. Qian Y
    53. Raitakari O
    54. Rueedi R
    55. Salvi E
    56. Schmidt B
    57. Schraut KE
    58. Shi J
    59. Smith AV
    60. Poot RA
    61. St Pourcain B
    62. Teumer A
    63. Thorleifsson G
    64. Verweij N
    65. Vuckovic D
    66. Wellmann J
    67. Westra HJ
    68. Yang J
    69. Zhao W
    70. Zhu Z
    71. Alizadeh BZ
    72. Amin N
    73. Bakshi A
    74. Baumeister SE
    75. Biino G
    76. Bønnelykke K
    77. Boyle PA
    78. Campbell H
    79. Cappuccio FP
    80. Davies G
    81. De Neve JE
    82. Deloukas P
    83. Demuth I
    84. Ding J
    85. Eibich P
    86. Eisele L
    87. Eklund N
    88. Evans DM
    89. Faul JD
    90. Feitosa MF
    91. Forstner AJ
    92. Gandin I
    93. Gunnarsson B
    94. Halldórsson BV
    95. Harris TB
    96. Heath AC
    97. Hocking LJ
    98. Holliday EG
    99. Homuth G
    100. Horan MA
    101. Hottenga JJ
    102. de Jager PL
    103. Joshi PK
    104. Jugessur A
    105. Kaakinen MA
    106. Kähönen M
    107. Kanoni S
    108. Keltigangas-Järvinen L
    109. Kiemeney LA
    110. Kolcic I
    111. Koskinen S
    112. Kraja AT
    113. Kroh M
    114. Kutalik Z
    115. Latvala A
    116. Launer LJ
    117. Lebreton MP
    118. Levinson DF
    119. Lichtenstein P
    120. Lichtner P
    121. Liewald DC
    122. Loukola A
    123. Madden PA
    124. Mägi R
    125. Mäki-Opas T
    126. Marioni RE
    127. Marques-Vidal P
    128. Meddens GA
    129. McMahon G
    130. Meisinger C
    131. Meitinger T
    132. Milaneschi Y
    133. Milani L
    134. Montgomery GW
    135. Myhre R
    136. Nelson CP
    137. Nyholt DR
    138. Ollier WE
    139. Palotie A
    140. Paternoster L
    141. Pedersen NL
    142. Petrovic KE
    143. Porteous DJ
    144. Räikkönen K
    145. Ring SM
    146. Robino A
    147. Rostapshova O
    148. Rudan I
    149. Rustichini A
    150. Salomaa V
    151. Sanders AR
    152. Sarin AP
    153. Schmidt H
    154. Scott RJ
    155. Smith BH
    156. Smith JA
    157. Staessen JA
    158. Steinhagen-Thiessen E
    159. Strauch K
    160. Terracciano A
    161. Tobin MD
    162. Ulivi S
    163. Vaccargiu S
    164. Quaye L
    165. van Rooij FJ
    166. Venturini C
    167. Vinkhuyzen AA
    168. Völker U
    169. Völzke H
    170. Vonk JM
    171. Vozzi D
    172. Waage J
    173. Ware EB
    174. Willemsen G
    175. Attia JR
    176. Bennett DA
    177. Berger K
    178. Bertram L
    179. Bisgaard H
    180. Boomsma DI
    181. Borecki IB
    182. Bültmann U
    183. Chabris CF
    184. Cucca F
    185. Cusi D
    186. Deary IJ
    187. Dedoussis GV
    188. van Duijn CM
    189. Eriksson JG
    190. Franke B
    191. Franke L
    192. Gasparini P
    193. Gejman PV
    194. Gieger C
    195. Grabe HJ
    196. Gratten J
    197. Groenen PJ
    198. Gudnason V
    199. van der Harst P
    200. Hayward C
    201. Hinds DA
    202. Hoffmann W
    203. Hyppönen E
    204. Iacono WG
    205. Jacobsson B
    206. Järvelin MR
    207. Jöckel KH
    208. Kaprio J
    209. Kardia SL
    210. Lehtimäki T
    211. Lehrer SF
    212. Magnusson PK
    213. Martin NG
    214. McGue M
    215. Metspalu A
    216. Pendleton N
    217. Penninx BW
    218. Perola M
    219. Pirastu N
    220. Pirastu M
    221. Polasek O
    222. Posthuma D
    223. Power C
    224. Province MA
    225. Samani NJ
    226. Schlessinger D
    227. Schmidt R
    228. Sørensen TI
    229. Spector TD
    230. Stefansson K
    231. Thorsteinsdottir U
    232. Thurik AR
    233. Timpson NJ
    234. Tiemeier H
    235. Tung JY
    236. Uitterlinden AG
    237. Vitart V
    238. Vollenweider P
    239. Weir DR
    240. Wilson JF
    241. Wright AF
    242. Conley DC
    243. Krueger RF
    244. Davey Smith G
    245. Hofman A
    246. Laibson DI
    247. Medland SE
    248. Meyer MN
    249. Yang J
    250. Johannesson M
    251. Visscher PM
    252. Esko T
    253. Koellinger PD
    254. Cesarini D
    255. Benjamin DJ
    256. LifeLines Cohort Study
    (2016) Genome-wide association study identifies 74 loci associated with educational attainment
    Nature 533:539–542.
    https://doi.org/10.1038/nature17671

Decision letter

  1. Eunjoon Kim
    Reviewing Editor; Institute for Basic Science, Korea Advanced Institute of Science and Technology, Korea (South), Republic of

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Roundabout receptor 2 maintains inhibitory control of the adult midbrain" for consideration by eLife. Your article has been favorably evaluated by a Senior Editor and three reviewers, one of whom is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal his identity: Garret D. Stuber (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This is a very thorough paper by Gore et al., where they examine the role of Robo2 in the VTA of mice in regulating the activity of dopamine neurons and a variety of behavioral phenotypes. The main findings are that loss of Robo2 in the VTA reduces inhibitory synaptic drive in the VTA, results in enhanced phasic activity of VTA dopamine neurons (and dopamine release in the NAc), and is associated with enhanced locomotion. These mice show alterations in reward learning and also show a paradoxical reduction in locomotion in response to psychomotor stimulants. The authors also delete Robo2 selectively in dopamine neurons and show that this alone cannot recapitulate the behavioral phenotypes seen when the gene is removed non-specifically from VTA neurons. Thus, they conclude that Robo2 is essential for inhibitory synapses maintenance in the VTA in adult mice. Overall, the experiments are well performed and the paper is clearly written. The findings are also quite novel as to my knowledge previous work has not investigated how Robo2 alters the function of VTA neurons, despite the enrichment of the expression of this gene within the VTA.

Essential revisions:

1) The authors showed the difference of the phenotypes between AAV-mediated k/o in VTA versus DAT cre mediated k/o. Given that it is almost impossible to make AAV injections just within VTA, could it be possible that the unintended virus spread infected some cells that are right next to VTA (some interneurons right next to VTA, but outside of it)? The authors did state that exclusion criteria was based on histological examination. But it is unclear whether injections that have 'tails', which is almost impossible to avoid with VTA injections, are included. It is also unclear what are the criteria and how robust they are. The authors' explanation is also a plausible one: the k/o with DATcre is incomplete or there is compensation for the protein and therefore they generated δ allele of Robo. When using DAT-cre::Robo2Δ/lox mice, there is still one allele that suffers the same problems. Therefore, quantifying the expression of Robo is important, to actually show the incomplete of k/o and improvement with DAT-cre::Robo2Δ/lox (e.g. by IHC).

2) Given the above mentioned phenotype differences, Figure 6 is the key experiment of the paper. To the authors credit, they found that using VGAT-cre mediated k/o, there are phenotypes. However, this part of the experiments seemed done in a rush (only one sentence about the key results) and was complicated by the fact that 11/20 mice had weight loss phenotypes. What does it mean 'dramatic weight loss' and what is 'not dramatic weight loss'? How objective is this? With complications, the only convincing way is the rescue experiment: expression Robo in VGAT+ cells, or stimulation of VGAT+ cells.

3) Images in Figure 1A is the foundation of the work. The authors stated that 'Robo2 is expressed in the VTA, predominantly in dopamine neurons (TH+ cells), but also non-dopamine neurons.' This was referred to as 'yellow cells' and 'red cells' respectively. But it looks to me that there are also a lot of TH-positive neurons that do not express Robo (green cells). Given the possibility of non-dopamine cells in this area might be an important component of the circuits based on functional data, it is worth quantifying the three categories of the cells. Similarly, in Figure 1—figure supplement 1A, if I understand correctly, top panels are supposedly to show that there are yellow signals around the green cells, and bottom panels are supposedly to show that there is no red around green cells. But from what I can see, there are definitely green cells that are not surrounded with red signals on the top panel. This needs to be quantified.

4) The authors present some example images of Robo2 localization with TH+ and TH- negative neurons. However, additional quantification of these data would be helpful. That is, is the proportion of TH+ and TH- neurons that express Robo2 similar or different.

5) Although the data implicate that Robo2 is expressed in GABAergic neurons in the VTA, there is no direct evidence to support this. The authors should try direct immunohistochemistry experiments to support this hypothesis. Ideally, the authors need to use several GABAergic neuronal markers. It would also be ideal if the authors could test whether other cell types in the VTA such as glutamatergic and GABAergic projection express Robo2 to exclude potential contribution of these neurons to the observed phenotype.

https://doi.org/10.7554/eLife.23858.017

Author response

Essential revisions:

1) The authors showed the difference of the phenotypes between AAV-mediated k/o in VTA versus DAT cre mediated k/o. Given that it is almost impossible to make AAV injections just within VTA, could it be possible that the unintended virus spread infected some cells that are right next to VTA (some interneurons right next to VTA, but outside of it)? The authors did state that exclusion criteria was based on histological examination. But it is unclear whether injections that have 'tails', which is almost impossible to avoid with VTA injections, are included. It is also unclear what are the criteria and how robust they are.

We apologize for the confusion on the language for the exclusion criteria. The criteria that was used excluded any mice that did not have direct viral infusion into the VTA. No animals met these criteria, so no animals were excluded. For all animals in this study, the peak viral expression occurred within the VTA. A small volume of diluted virus was injected into the VTA to minimize spread outside the VTA. A minority of animals exhibited weak “tail” expression outside the VTA. We have now performed a quantitative analysis (n=10 animals) of the spread of the AAV1-Cre-GFP virus within the adult midbrain. We traced the viral spread throughout the rostral, medial, and caudal VTA. We also performed cell counts within these brain regions. We observed good coverage of the virus throughout the VTA with minimal spread outside the VTA.

In the instances of minor spillover outside the VTA, it spread to adjacent midbrain regions that do not express Robo2. As can be seen in the Robo2 in situ data (Figure 1—figure supplement 1A), Robo2 expression in the adult midbrain is largely confined to the VTA and SNc, so cre-mediated recombination of the floxed Robo2 locus in non-Robo2 expressing cells outside the VTA and SNc should presumably have no effect. In the subset of the animals with minor viral spillover expression into the SNc, inactivating Robo2 in the SNc might have contributed to the phenotypes. However, we observed the same phenotypes in animals that selectively expressed the virus in the VTA with no expression in the SNc, which demonstrates the VTA is the key site of Robo2 expression in the adult responsible for the observed phenotypes.

The authors' explanation is also a plausible one: the k/o with DATcre is incomplete or there is compensation for the protein and therefore they generated δ allele of Robo. When using DAT-cre::Robo2Δ/lox mice, there is still one allele that suffers the same problems. Therefore, quantifying the expression of Robo is important, to actually show the incomplete of k/o and improvement with DAT-cre::Robo2Δ/lox (e.g. by IHC).

We apologize for the lack clarity regarding the use of the δ allele in the DAT-Cre::Robo2△/lox mice. We commonly use the δ allele for cell-specific inactivation in non-inducible cre-driver lines. This is based on the observation that many genes are transiently or ectopically expressed in germ cells, but not with 100% penetrance. Thus, in a mouse breeder that is DAT-Cre::Robo2lox/+ for example, some germ cells may have ectopically expressed Cre from the DAT locus, but this may not occur in all germ cells. Thus, some gametes will be Robo2△/+ and some will be Robo2lox/+ yielding offspring from the same litter with potentially different genotypes. To circumvent this, we generate a δ allele and only breed mice that are DAT-Cre:: Robo2△/+ with mice that are Robo2lox/lox. In this way we can avoid mosaicism within a given litter. To control for potential effects of heterozygosity, control mice are DAT-Cre::Robo2lox/+, Robo2 △/lox (Cre-negative), and DAT-Cre::Robo2+/lox. We did not observe any behavioral differences in these genotypes and collapsed these mice into a single control group. This has been more accurately described in the Methods section.

With regard to the IHC for Robo2, that relates specifically to this point as well as points 3, 4, and 5 below, the Robo2 rabbit polyclonal antibody we used in this study from Abcam that was rigorously validated by us and other investigators (for example see Gibson et al., 2014), has been discontinued. We have made a note of this discontinuation in our Methods section. This prevented us from performing additional experiments to complete some of the detailed analysis requested. We acquired multiple additional Robo2 antibodies, however these antibodies did not prove to be selective for Robo2 (likely cross-reacting with other Robo family members). In some cases, such as in relation to specific points 3 and 4, we had a sufficient amount of immunostaining from previous analyses to complete quantitative comparisons of Robo2 overlap with TH+ and TH- cells. To address outstanding questions relating to cell-specific expression of Robo2, we performed cell-specific mRNA analysis utilizing the Ribotag strategy (Sanz et al., 2009). Using this approach we were able to isolate the translatomes from DAT-Cre control and DAT-cre::Robo2Δ/lox mice and demonstrate a de-enrichment of Robo2 in dopamine producing neurons consistent with genetic inactivation and nonsense-mediate mRNA decay. See response to specific point 5 for additional details.

2) Given the above mentioned phenotype differences, Figure 6 is the key experiment of the paper. To the authors credit, they found that using VGAT-cre mediated k/o, there are phenotypes. However, this part of the experiments seemed done in a rush (only one sentence about the key results) and was complicated by the fact that 11/20 mice had weight loss phenotypes. What does it mean 'dramatic weight loss' and what is 'not dramatic weight loss'? How objective is this? With complications, the only convincing way is the rescue experiment: expression Robo in VGAT+ cells, or stimulation of VGAT+ cells.

We apologize for the confusion in our previous version of the manuscript; we have expanded our description of this section. This experiment was performed on three separate cohorts of 6-8 mice per group per cohort. Numerous mice with inactivation of the GABA neurons had to be euthanized because they exhibited “dramatic weight loss”, defined as animals dropping below 80% of their pre-surgery bodyweight. The surviving animals also lost bodyweight, but stabilized two weeks post-surgery, allowing for their behavioral characterization. All of the surviving animal had robust viral expression within the VTA, so they did not survive due to being partial or unilateral viral hits. Thus, we believe that our analysis of these mice is valid, caveats notwithstanding, since they did show robust viral expression. It is unclear why some animals were more affected than others, but possibly a compensatory mechanism occurred in the survivors. As noted above, we have added to the text additional information regarding this caveat. Importantly, this selection bias of only be able to analyze the survivors might have caused an underestimation of the phenotype.

3) Images in Figure 1A is the foundation of the work. The authors stated that 'Robo2 is expressed in the VTA, predominantly in dopamine neurons (TH+ cells), but also non-dopamine neurons.' This was referred to as 'yellow cells' and 'red cells' respectively. But it looks to me that there are also a lot of TH-positive neurons that do not express Robo (green cells). Given the possibility of non-dopamine cells in this area might be an important component of the circuits based on functional data, it is worth quantifying the three categories of the cells. Similarly, in Figure 1—figure supplement 1A, if I understand correctly, top panels are supposedly to show that there are yellow signals around the green cells, and bottom panels are supposedly to show that there is no red around green cells. But from what I can see, there are definitely green cells that are not surrounded with red signals on the top panel. This needs to be quantified.

We appreciated the reviewers’ comments; fortunately we had a sufficient number of mice in which this staining was performed previously to do this quantitative analysis. We now have quantified the three categories of cells that include TH+Robo2+, TH+Robo2-, and TH-Robo2+ cells (Figure 1A). In addition, we have performed Ribotag from dopamine and GABA neurons and demonstrate that Robo2 has enriched expression in both populations (Figure 1—figure supplement 1D-F).

Regarding Figure 1—figure supplement 1A (now Figure 1—figure supplement 1G), your understanding of the figure is correct; namely, the inactive Cre (top panel) has no effect on Robo2 expression, but in the active Cre (bottom panel), Cre-GFP cells are never found co-expressed with Robo2. However, since the AAV-Cre-GFP and AAV-∆Cre-GFP viruses are non-conditional, it will be expressed in a variety of cell types that include midbrain neurons that do not express Robo2; hence, some cells will be GFP positive but Robo2 negative. In both the top and bottom panel, these cells would appear as cells that are green but not red. We have now quantified the distribution of the three categories of cells, GFP+Robo2+, GFP-Robo2+, and GFP+Robo2- in animals receiving either the AAV1-Cre-GFP or the AAV1-∆Cre-GFP virus (Figure 1—figure supplement 1H).

4) The authors present some example images of Robo2 localization with TH+ and TH- negative neurons. However, additional quantification of these data would be helpful. That is, is the proportion of TH+ and TH- neurons that express Robo2 similar or different.

As noted above for specific point #3, we have quantified the three categories of cells that include TH+Robo2+, TH+Robo2-, and TH-Robo2+ cells.

5) Although the data implicate that Robo2 is expressed in GABAergic neurons in the VTA, there is no direct evidence to support this. The authors should try direct immunohistochemistry experiments to support this hypothesis. Ideally, the authors need to use several GABAergic neuronal markers. It would also be ideal if the authors could test whether other cell types in the VTA such as glutamatergic and GABAergic projection express Robo2 to exclude potential contribution of these neurons to the observed phenotype.

As mentioned in response to specific points 1 and 3, we were not able to perform detailed immunohistochemical analyses due to the discontinuation of the validated antibody to Robo2. However, our Ribotag experiment that we have added to the manuscript (Figure 1—figure supplement 1D-F), demonstrate Robo2 to be enriched ~2 fold in purified translatomes from VGAT-Cre GABA neurons compared to the ventral midbrain tissue punch input mRNA. These findings, together with our Ribotag analysis of dopamine neurons correlate with the immunohistochemistry showing expression of Robo2 in both midbrain dopamine (TH+) and non-dopamine (TH-) neurons.

https://doi.org/10.7554/eLife.23858.018

Article and author information

Author details

  1. Bryan B Gore

    1. Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, United States
    2. Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    BBG, Conceptualization, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  2. Samara M Miller

    Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    SMM, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Contributed equally with
    Yong Sang Jo and Madison A Baird
    Competing interests
    The authors declare that no competing interests exist.
  3. Yong Sang Jo

    1. Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, United States
    2. Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    YSJ, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Contributed equally with
    Samara M Miller and Madison A Baird
    Competing interests
    The authors declare that no competing interests exist.
  4. Madison A Baird

    Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    MAB, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Contributed equally with
    Samara M Miller and Yong Sang Jo
    Competing interests
    The authors declare that no competing interests exist.
  5. Mrinalini Hoon

    Department of Biological Structure, University of Washington, Seattle, United States
    Contribution
    MH, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  6. Christina A Sanford

    Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    CAS, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  7. Avery Hunker

    Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    AH, Data curation, Formal analysis, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  8. Weining Lu

    Department of Medicine, Renal Section, Boston University Medical Center, Boston, United States
    Contribution
    WL, Data curation, Formal analysis, Supervision, Funding acquisition, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  9. Rachel O Wong

    Department of Biological Structure, University of Washington, Seattle, United States
    Contribution
    ROW, Data curation, Formal analysis, Supervision, Funding acquisition, Methodology, Writing—original draft, Writing—review and editing
    Competing interests
    The authors declare that no competing interests exist.
  10. Larry S Zweifel

    1. Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, United States
    2. Department of Pharmacology, University of Washington, Seattle, United States
    Contribution
    LSZ, Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Methodology, Writing—original draft, Writing—review and editing
    For correspondence
    larryz@uw.edu
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3465-5331

Funding

National Institutes of Health (R01-MH094536)

  • Larry S Zweifel

National Institute for Health Research (EY10699)

  • Rachel O Wong

National Institute for Health Research (R01-DK078226)

  • Weining Lu

National Institutes of Health (R01-MH104450)

  • Larry S Zweifel

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to thank the entire Zweifel lab for their critical reading of the manuscript. This work was supported by the NIH (R01-MH094536 and R01-MH104450 to LZ, EY10699 to RW, and R01DK078226 to WL).

Ethics

Animal experimentation: All experiments were done in accordance with a protocol (4249-01) approved by the University of Washington Animal Care and Use Committee.

Reviewing Editor

  1. Eunjoon Kim, Institute for Basic Science, Korea Advanced Institute of Science and Technology, Korea (South), Republic of

Publication history

  1. Received: December 2, 2016
  2. Accepted: April 9, 2017
  3. Accepted Manuscript published: April 10, 2017 (version 1)
  4. Version of Record published: May 5, 2017 (version 2)

Copyright

© 2017, Gore et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 970
    Page views
  • 220
    Downloads
  • 9
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Immunology and Inflammation
    2. Neuroscience
    Ibrahim T Mughrabi et al.
    Tools and Resources Updated

    Vagus nerve stimulation (VNS) suppresses inflammation and autoimmune diseases in preclinical and clinical studies. The underlying molecular, neurological, and anatomical mechanisms have been well characterized using acute electrophysiological stimulation of the vagus. However, there are several unanswered mechanistic questions about the effects of chronic VNS, which require solving numerous technical challenges for a long-term interface with the vagus in mice. Here, we describe a scalable model for long-term VNS in mice developed and validated in four research laboratories. We observed significant heart rate responses for at least 4 weeks in 60–90% of animals. Device implantation did not impair vagus-mediated reflexes. VNS using this implant significantly suppressed TNF levels in endotoxemia. Histological examination of implanted nerves revealed fibrotic encapsulation without axonal pathology. This model may be useful to study the physiology of the vagus and provides a tool to systematically investigate long-term VNS as therapy for chronic diseases modeled in mice.

    1. Neuroscience
    Shinya Ohara et al.
    Research Article Updated

    The entorhinal cortex, in particular neurons in layer V, allegedly mediate transfer of information from the hippocampus to the neocortex, underlying long-term memory. Recently, this circuit has been shown to comprise a hippocampal output recipient layer Vb and a cortical projecting layer Va. With the use of in vitro electrophysiology in transgenic mice specific for layer Vb, we assessed the presence of the thus necessary connection from layer Vb-to-Va in the functionally distinct medial (MEC) and lateral (LEC) subdivisions; MEC, particularly its dorsal part, processes allocentric spatial information, whereas the corresponding part of LEC processes information representing elements of episodes. Using identical experimental approaches, we show that connections from layer Vb-to-Va neurons are stronger in dorsal LEC compared with dorsal MEC, suggesting different operating principles in these two regions. Although further in vivo experiments are needed, our findings imply a potential difference in how LEC and MEC mediate episodic systems consolidation.