Increased GABAB receptor signaling in a rat model for schizophrenia

Schizophrenia is a complex disorder that affects cognitive function and has been linked, both in patients and animal models, to dysfunction of the GABAergic system. However, the pathophysiological consequences of this dysfunction are not well understood. Here, we examined the GABAergic system in an animal model displaying schizophrenia-relevant features, the apomorphine-susceptible (APO-SUS) rat and its phenotypic counterpart, the apomorphine-unsusceptible (APO-UNSUS) rat at postnatal day 20–22. We found changes in the expression of the GABA-synthesizing enzyme GAD67 specifically in the prelimbic- but not the infralimbic region of the medial prefrontal cortex (mPFC), indicative of reduced inhibitory function in this region in APO-SUS rats. While we did not observe changes in basal synaptic transmission onto LII/III pyramidal cells in the mPFC of APO-SUS compared to APO-UNSUS rats, we report reduced paired-pulse ratios at longer inter-stimulus intervals. The GABAB receptor antagonist CGP 55845 abolished this reduction, indicating that the decreased paired-pulse ratio was caused by increased GABAB signaling. Consistently, we find an increased expression of the GABAB1 receptor subunit in APO-SUS rats. Our data provide physiological evidence for increased presynaptic GABAB signaling in the mPFC of APO-SUS rats, further supporting an important role for the GABAergic system in the pathophysiology of schizophrenia.

To explore the pathophysiological mechanisms underlying schizophrenia, we used the apomorphine-susceptible (APO-SUS) rat model and its phenotypic counterpart, the apomorphine-unsusceptible (APO-UNSUS) rat. These two rat lines have been pharmacogenetically selected and bred from an outbred Wistar population based on their stereotypical gnawing responses to the dopamine D1/D2 agonist apomorphine 21 . The APO-SUS rat is a well-characterized animal model that displays schizophrenia-relevant features during adulthood, such as altered density of central dopamine receptors 22 , high sensitivity to dopaminergic drugs (i.e., apomorphine and amphetamine) 21,23,24 , decreased prepulse inhibition and diminished latent inhibition 24,25 , increased novelty-induced exploration and accumbal dopamine response 21,23,26 , and an increased HPA-axis response to stress 22 , as well as learning and memory deficits 27,28 . However, the pathophysiology underlying the phenotypes observed in this model remains unknown.
In the present study, we compared the inhibitory system in the medial PFC (mPFC) of APO-SUS and APO-UNSUS rats at postnatal day (PND) 20-22. We found a decrease in the protein level of GAD67 as well as a reduced cell-count for GAD67 + cells specifically in the prelimbic region (PRL) of the mPFC. While basal synaptic transmission onto LII/III pyramidal cells was not different between the two rat lines, we report a decrease in the paired-pulse ratio in APO-SUS rats at longer (≥ 150 ms) intervals compared to APO-UNSUS rats. Importantly, this decrease could be abolished by application of the GABA B antagonist CGP 55845, indicating that the decreased paired-pulse ratio was caused by an increased activity in GABA B signaling. We indeed observed an increase in GABA B receptor expression in the mPFC of APO-SUS rats compared to APO-UNSUS rats. Collectively, our data identify GABA B receptor signaling as a possible player in the etiology of schizophrenia.

Results
Reduced levels of GAD67 in the mPFC of APO-SUS rats. Previous studies have reported altered levels of specific interneuron markers in schizophrenia patients [13][14][15][16][17]29 . In order to determine whether the number of GABAergic interneurons is altered in APO-SUS compared to APO-UNSUS rats, we performed western blot analysis to measure the protein expression levels of GAD67, CB and CR in punches from the mPFC of PND 20-22 rats (Fig. 1). We found protein levels of GAD67 to be reduced by ~35% in APO-SUS rats (Fig. 1a,d; p = 0.04). Conversely, we found no change in the levels of CB (Fig. 1b,d; p = 0.33) and CR (Fig. 1c,d; p = 0.99) between the two rat lines.
Reduction of GAD67 + cells in the prelimbic region, but not the infralimbic region, of the mPFC of APO-SUS rats. Next, we wanted to assess whether the decrease in GAD67 protein levels in the mPFC was due to a decreased amount of GAD67 per cell or to a decrease in the total number of interneurons. To this end we performed immunohistochemistry to assess the number of interneurons positive for GAD67. We also assessed the number of CB + , CR + and PV + cells, which together label ~90% of GABAergic interneurons 30 . We used coronal sections of the mPFC and analyzed the number of positive cells separately for the prelimbic (PRL)-and infralimbic (IL) region, two subregions of the mPFC.
Unaltered basal synaptic input on LII/III pyramidal neurons in the mPFC. Since reduced levels of GAD67 are indicative of reduced interneuronal activity 31,32 , we subsequently tested if the observed changes in GAD67 protein levels would be accompanied by changes at the electrophysiological level. In order to assess basal synaptic input we used whole-cell voltage clamp to record miniature inhibitory postsynaptic currents (mIPSCs) Previous studies have shown that changes in postsynaptic GABA receptor subunit composition are reflected in the decay time of mIPSC 33,34 . We found no changes in the mIPSC decay time in APO-SUS rats (APO-UNSUS 15.58 ± 0.52 ms; APO-SUS 13.46 ± 0.93 ms, p = 0.06), indicating the postsynaptic receptor content of inhibitory synapses was unaltered (Fig. 3g).
Next, we measured the maximal response to bath application of 20 μ M GABA (Fig. 3h). No difference was observed between APO-UNSUS and APO-SUS rats (APO-UNSUS 154.1 ± 48.3 pA; APO-SUS 163.9 ± 51.2 pA, p = 0.78), indicating that the total number of GABA A receptors was identical. Together, these data show that there is no difference in basal synaptic inhibitory and excitatory input connectivity onto LII/III pyramidal cells in the mPFC between APO-UNSUS and APO-SUS rats.
We then analyzed pyramidal cell morphology of cells that were filled with biocytin during the recordings. All cells had a typical pyramidal morphology, hallmarked by a skirt of basal dendrites radiating ~200 μ m from the soma, and a single long apical dendrite that reached into upper LI (Fig. 3i). The apical dendrite mostly branched close to the soma, and had an apical tuft close to the pia mater, mainly within LI, consistent with previous reports 35 . The dendritic morphology of both apical and basal dendrites did not differ between pyramidal cells from APO-UNSUS and APO-SUS rats (Table 1).
Increased expression of GABA B receptors, but not GAT1, in the mPFC of APO-SUS rats. The increase in GABA B receptor signaling could be caused either by an increase of GABA B receptor activation, through increased levels of extracellular GABA, or by an increase in the number of GABA B receptors. Increased levels of extracellular GABA could be caused by a reduction in the expression of GAT1, the main GABA transporter in the central nervous system 41 . GAT1 reduces the amount of extracellular GABA in the synapse, which reduces spillover 41 . GABA B receptors are located perisynaptically, and are activated by GABA spillover. In addition, reduced GAT1 expression has been found in schizophrenia patients 42 and animal models 12 of schizophrenia. We therefore first assessed the levels of GAT1 in punches from the mPFC of APO-SUS and APO-UNSUS rats by performing western blot analysis. However, we found no significant difference between APO-SUS and APO-UNSUS rats (p = 0.73; Fig. 5a,b). We next studied the protein expression levels of GABA B receptors. Western blot analysis using an antibody against the GABA B1 subunit, which together with the GABA B2 receptor forms an active GABA B receptor 43 , revealed that the levels of GABA B1 were increased by ~110% in the mPFC of APO-SUS rats compared to APO-UNSUS mPFC (p < 0.05; Fig. 5a,b). Together, our data suggest that increased GABA B receptor signaling in APO-SUS rats is accompanied by increased expression of the GABA B1 subunit, and underlies the decreased paired-pulse ratio in APO-SUS rats.

Discussion
The GABAergic system plays an important role in the pathophysiology of schizophrenia 44,45 . Here, using the APO-SUS/APO-UNSUS rat model at PND 20-22 we further highlight the importance of dysfunctional GABA signaling in schizophrenia-related traits. We provide several lines of evidence for deficits in the inhibitory  circuitry within the PRL of APO-SUS rats: (i) a reduced level of GAD67 in interneurons in the mPFC of APO-SUS rats, specifically in the PRL; (ii) a reduced paired-pulse ratio at longer (≥ 150 ms) ISIs in APO-SUS rats; and (iii) increased GABA B signaling due to an increased expression of the GABA B1 subunit of the GABA B receptor. Finally, we were able to reverse the increased GABA B signaling in APO-SUS rats by the application of the GABA B antagonist CGP 55845. In our present study on the GABAergic system in the APO-SUS rats, we observed a decreased protein level of GAD67 in the PRL. Also, we found a reduced number of GAD67 + cells in this region. However, the number of CB + , CR + and PV + neurons, which together account for ~90% of interneurons 30 , remained unchanged in the APO-SUS rat, indicating that the total number of interneurons is unaltered. The reduced number of GAD67 + cells could result from a reduction of GAD67 levels per cell, causing these interneurons to be under the detection threshold of our analysis. Indeed, reduced levels of GAD67 have been reported before in schizophrenia both in animal models and post-mortem patient studies 17,46,47 .
A number of studies has shown that reduced prefrontal inhibitory transmission induces various cognitive, emotional and dopaminergic abnormalities that resemble aspects of schizophrenia 48 . More specifically, reduced GABAergic activity in the PRL of the rat PFC results in deficits in speed of processing information, cognitive flexibility, recall of relevant information and enhanced dopamine activity 49,50 . In this respect, it is of interest to note that GABAergic interneurons in the mPFC receive direct input from mesocortical dopaminergic fibers, and this control matures during adolescence and appears crucial for cortical network activity at adulthood 51 . Moreover, together with the genetic background and environmental stress the developmental ingrowth of dopaminergic fibers may lead to an abnormal functioning of the GABAergic interneurons at adulthood, especially under stressful conditions that also alter the dopamine system 52 . A functional dysregulation of the mesostriatal dopaminergic neurons in schizophrenia has indeed been well established 8,53,54 .
Given that we found altered levels of GAD67 in the mPFC of APO-SUS rats, indicative of a reduced inhibitory drive, we studied changes in the basal synaptic input connectivity of LII/III pyramidal neurons. Surprisingly, mIPSC amplitude and frequency on these neurons were unaltered, as was the presynaptic release probability for GABA, suggesting that the changes in interneuron composition were not accompanied by changes in basal synaptic input. Additionally, we did not observe any changes in the response to bath application of GABA, indicating that the total number of GABA A receptors was unaltered. Together, these results suggest that the total number and strength of inhibitory synapses are not different between APO-SUS and APO-UNSUS rats. However, this study does not include a measure of excitability of the individual classes of interneurons within the mPFC and it is conceivable that the reduced inhibitory activity is due to a reduced excitability of one or more subclasses of interneurons. The use of reporter rodent lines will allow the targeted exploration of the excitability of specific interneuron subclasses in future studies.
While we did not observe differences at the level of synaptic transmission and morphology of LII/III pyramidal cells, we found a clear reduction in the paired-pulse ratio in the APO-SUS compared to APO-UNSUS rats. This reduction was specific for ISIs ≥150 ms, and was not observed in ≤ 100 ms ISIs. The reduced paired-pulse ratio at ≥ 150 ms ISIs indicates a reduced release of neurotransmitter at the second pulse. We show that enhanced GABA B receptor signaling, at least in part, underlies this reduction. Indeed, the paired-pulse ratio in the APO-SUS rats could be restored to the levels observed in the APO-UNSUS rats by application of the GABA B receptor antagonist CGP 55845. GABA B receptors are metabotropic G-protein-coupled GABA receptors, and are located perisynaptically at both the pre-and postsynapse 55 . Presynaptic GABA B receptors reduce calcium inflow mainly by inhibiting N-type and P/Q type voltage-gated calcium channels and to a lesser extent by activating inward-rectifying potassium channels such as Kir3, causing a hyperpolarization 55 . Both pathways reduce the inflow of calcium, and occur with a delay of ~150 ms and slow decay of ~1 s 40 consistent with the maximal effect that we observed at the 150 ms ISI. Postsynaptic GABA B receptors mediate opening of potassium channels, but both the absence of an outward current in our recordings as well as the use of a cesium-based intracellular recording solution exclude the possibility of a postsynaptic contribution of GABA B receptors to our recordings and point to presynaptic GABA B receptor activity underlying the observed changes. Activation of presynaptic GABA B receptors reduces calcium conductance and subsequent GABA release, and therefore provides a negative feedback to the GABAergic system 56,57 . Interestingly, a reduced inhibitory input onto excitatory pyramidal cells has been found in models for schizophrenia 20 and autism 58 , and underlines the notion that a proper tuning of excitation and inhibition is required for proper brain function 11 . Importantly, a number of studies have highlighted a crucial link between GABA transmission and cognitive dysfunction in schizophrenia, indicating that reduced prefrontal inhibitory transmission induces various cognitive, emotional and dopaminergic abnormalities that resemble aspects of this disorder 48 .
Consistent with our finding of increased GABA B signaling, we show that the expression of GABA B1 is increased in the mPFC of APO-SUS compared to APO-UNSUS rats. However, our results do not show if the increase in GABA B1 is due to an increase in the number of GABA B1 -expressing cells, or an increase in the amount of GABA B1 per cell. Thus far, no comparative studies are available describing the levels of GABA B receptors in individual interneuron subtypes. Interestingly, a reduction in GABA B subunit expression has been observed in various brain regions of schizophrenia patients [59][60][61] . In humans, stimulation of cortical GABA B receptors in the fronto-parietal network has led to better attentional allocation in reinforcement learning 62 . In addition, GABA B receptor manipulation has been shown to reverse behavioral changes related to psychosis 63 , improve pre-pulse inhibition deficits and ameliorate sensorimotor gating in rodent models 64 . Combining the results of the present animal study with the results of the previously reported patient studies suggests that both reduced as well as increased GABA B signaling may underlie some of the aspects of schizophrenia. Of note, most schizophrenia research on patients is hampered by the use of medication, whereas the rat model used in this study is drug naive. Additional preclinical studies are warranted to further evaluate the hypothesis that the GABA B receptor Scientific RepoRts | 6:34240 | DOI: 10.1038/srep34240 represents a promising pharmacological target for treating appropriately stratified subsets of individuals with schizophrenia.
It is important to notice that schizophrenia is a disorder with an onset during adolescence or early adulthood 65 . The PND 20-22 rats used for this study are in their early adolescence, and therefore do not necessarily display the schizophrenia-relevant features that have been described in adult rats [21][22][23][24][25][26] . Even though the exact mechanisms underlying neurodevelopmental disorders, including schizophrenia, are unknown, many theories exist about a distorted balance between neuronal excitation and inhibition during development 51,66,67 . While it is difficult to distinguish primary effects from homeostatic-and compensatory mechanisms, a substantial amount of evidence points in the direction of disrupted inhibitory signaling as an important factor in the etiology of schizophrenia 51,68,69 . Our data suggest a possible role for GABA B receptor signaling in the development of the schizophrenia-relevant features observed in adult APO-SUS rats.
In conclusion, our findings highlight the importance of GABAergic signaling for inducing schizophrenia-related phenotypes, and identify GABA B receptors as potential new key players in the distorted network functioning in this disorder.

Animals.
All experiments were performed with male Wistar (PND20-22) rats pharmacogenetically selected for high susceptibility (APO-SUS) or low susceptibility (APO-UNSUS) to apomorphine. The generation of APO-SUS and APO-UNSUS rat lines has been described previously 21 . In short, upon injection of apomorphine, a bimodal distribution of the gnawing response was found, i.e. approximately 40% of the Wistar rats showed a weak gnawing response (< 10 counts/45 min) and a similar percentage showed an intense gnawing response (> 500 counts/45 min). Following this initial selection, the nine males and females with the highest scores, and the nine males and females with the lowest scores were paired and their offspring was again tested for their gnawing response. For each new generation, nine new pairs of rats were selected out of the four male and female litters showing the highest (APO-SUS) and the lowest (APO-UNSUS) mean gnawing response per gender, with the condition that brother/sister pairing was not allowed. APO-SUS rats are defined as animals born from an APO-SUS mother and father; APO-UNSUS rats are likewise defined as animals born from an APO-UNSUS mother and father. The present experiments were performed with naive male APO-SUS and APO-UNSUS rats belonging to the 38 th and 40 th generation, i.e. apomorphine was used only during the procedure to select and generate the APO-SUS and APO-UNSUS lines.
All rats were bred and reared in the Central Animal Facility of the Radboud University Nijmegen. The animals were reared and housed in macrolon cages (42 × 26 × 15 cm) in a controlled (20 ± 2 °C) and enriched environment under a 12 h light/dark cycle (lights on at 7:00 A.M.). Food pellets and water were provided ad libitum. Experimental procedures were performed between 9:00 A.M. and 5:00 P.M. The experimental procedures were approved by the Animal Ethics Committee of the Radboud University Nijmegen (Nijmegen, the Netherlands) and conducted in accordance with the Dutch legislation. Every effort was made to minimize the number of animals used and their suffering.
Western blotting. For the western blot experiments the animals were decapitated and the brains were collected quickly, frozen on dry ice, and kept at − 80 °C. The brains were sliced into 200 μ m coronal sections using a cryostat (Microm HM 500OM) at − 15 °C and mounted on glass slides. Punches from 6 sections of the mPFC (prelimbic and infralimbic region) were taken with a 2 mm diameter micropunch needle (Harris Inc.) between the first appearance of the corpus callosum and the start of the caudate putamen. The coordinates were determined according to the atlas of Paxinos and Watson 70 . Samples from each hemisphere were pooled. All the samples were stored at − 80 °C before protein extraction took place.

Morphological reconstructions.
For morphological reconstructions, the internal solution was supplemented with 0.4% biocytin (Sigma-Aldrich). Following single-cell electrophysiology, the slices with biocytin-filled neurons were fixated in 4% paraformaldehyde in 0.1 M PBS overnight at 4 °C, and subsequently processed following a modified staining protocol based on Marx et al. 72 . In brief: after fixation, slices were rinsed in 0.1 M PB, incubated in 3% H2O2 in 0.1 M PB for 30 minutes at room temperature to quench endogenous peroxidases, rinsed in 0.1 M PBS, then incubated in Avidin-Biotin-Peroxidase solution (Vectastain Elite, with 1% v/v Triton-100) overnight on a shaking platform at 4 °C. The next day, slices were washed with 0.1 M PBS and pre-incubated with Di-Aminobenzidine (DAB) solution with Nickel enhancer (Vector Peroxidase substrate kit, SK-4100) for 30 min. Then, the DAB solution was replaced with the same solution plus H2O2 and incubated for ca. 30 seconds. Slices were then rinsed in 0.1 M PB, mounted on gelatinized coverslips, and dried for 3-6 h in a custom-made moist chamber at room temperature. Slices were dehydrated in an ethanol series and Xylene, put on coverslips and sealed with Eukitt (Sigma). Slices were imaged on a Zeiss Axioskop 1 upright brightfield microscope with 20x and 40x objectives (Zeiss EC Plan-Neofluar, NA 0.5 and 0.75, respectively) and motorized stage (MicroBrightField). The camera and motorized stage were connected to a Neurolucida workstation (MicroBrightField). Cells were selected based on position in the cortical layer II/III of the prefrontal cortex, pyramidal morphology, and staining intensity. Somata, apical and basal dendrites were reconstructed in Neurolucida as three-dimensional drawings. Reconstructions were analysed in NeuroExplorer (MicroBrightField) for intrinsic parameters and Sholl analysis. Statistical analysis. Statistical analysis was performed using Prism (Graphpad). Significance was tested with a two-sided Student's t-test. Correction for multiple comparisons was performed using the Holm-Sidak method where indicated. Data is expressed as mean ± SEM. Significance was defined as p < 0.05 (*) or p < 0.01 (**).