GluN2D-mediated excitatory drive onto medial prefrontal cortical PV+ fast-spiking inhibitory interneurons
Loading metrics
Open Access
Peer-reviewed
Research Article
Jonathan Garst-Orozco ,
Ruchi Malik,
Thomas A. Lanz,
Mark L. Weber,
Hualin Xi,
Dominique Arion,
John F. Enwright III,
David A. Lewis,
Patricio O’Donnell,
Vikaas S. Sohal,
Derek L. Buhl
Jonathan Garst-Orozco,
Ruchi Malik,
Thomas A. Lanz,
Mark L. Weber,
Hualin Xi,
Dominique Arion,
John F. Enwright III,
David A. Lewis,
Patricio O’Donnell,
Vikaas S. Sohal
x
Published: June 4, 2020
https://doi.org/10.1371/journal.pone.0233895
AbstractDeficits in fast-spiking inhibitory interneurons (FSINs) within the dorsolateral prefrontal cortex (dlPFC) are hypothesized to underlie cognitive impairment associated with schizophrenia. Though representing a minority of interneurons, this key cell type coordinates broad neural network gamma-frequency oscillations, associated with cognition and cognitive flexibility. Here we report expression of GluN2D mRNA selectively in parvalbumin positive cells of human postmortem dlPFC tissue, but not pyramidal neurons, with little to no GluN2C expression in either cell type. In acute murine mPFC slices the GluN2C/D selective positive allosteric modulator (PAM), CIQ(+), increased the intrinsic excitability as well as enhanced NMDAR-mediated EPSCs onto FSINs. This increase in intrinsic excitability with GluN2C/D PAM was also observed in the Dlx 5/6+/- FSIN developmental deficit model with reported FSIN hypoexcitability. Together these data speak to selective modulation of FSINs by a GluN2D PAM, providing a potential mechanism to counter the FSIN-deficit seen in schizophrenia.
Citation: Garst-Orozco J, Malik R, Lanz TA, Weber ML, Xi H, Arion D, et al. (2020) GluN2D-mediated excitatory drive onto medial prefrontal cortical PV+ fast-spiking inhibitory interneurons. PLoS ONE 15(6):
e0233895.
https://doi.org/10.1371/journal.pone.0233895Editor: Pavel I. Ortinski, University of Kentucky, UNITED STATESReceived: February 24, 2020; Accepted: May 14, 2020; Published: June 4, 2020Copyright: © 2020 Garst-Orozco et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.Data Availability: RNAseq data have been uploaded to GEO (GSE149154) https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE149154 Electrophysiology data are within the manuscript and its Supporting Information files.Funding: At the time of this study, Pfizer, Inc. provided support in the form of salaries for JGO, TAL, MLW, HX, POD, and DLB, and provided funding grants to the laboratories of DAL and VSS. The funder did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. CDMRP (TS150059) to VSS and RM. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. NIMH R01 (MH106507) to VSS. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. NARSAD Young Investigator Award (Leichtung Family Investigator, BBRF) to RM. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are listed in the ‘author contribution’ section.Competing interests: Numerous contributing authors had at the time of the study and continue to have commercial affiliations. However the funds from Pfizer Neuroscience covered salaries and materials, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript beyond supporting the listed authors’ salaries (Funding Statement), materials, and the dissemination costs (commitment to cover submission fees). Commercial affiliations in no way have impacted our adherence to PLOS ONE policies on sharing data and materials.
IntroductionStudies of the underlying pathophysiology of schizophrenia have led to the NMDA, GABA and dopaminergic hypotheses, backed by pharmacological manipulations that replicate key aspects of the pathology. The dopamine (DA) hypothesis, which has remained the most prevalent in the field, is substantiated by the ability of DA agonists to replicate, and DA antagonists to address, positive symptoms. Though positive symptoms are perhaps the most striking aspect of the disease, negative and cognitive symptoms have been shown to be more predictive of patient outcome and remain unaddressed by current pharmacology [1–5]. These deficits include working memory, attention, cognitive flexibility, semantic processing and verbal learning. In addition to the cognitive impairment associated with schizophrenia (CIAS), these subjects also exhibit an associated marked decrease in evoked cortical gamma band frequency (30–80 Hz) amplitude, which is thought to be indicative of fast-spiking interneuron (FSIN) dysfunction in synchronizing the firing of broad neural networks [6–9]. The GABA hypothesis is also supported by human post-mortem tissue from individuals with schizophrenia showing markedly reduced GAD67 expression [10,11], an enzyme that synthesizes GABA, in addition to reduced GAT1 [12], the membrane transporter of GABA, somatostatin (SOM) [13] and parvalbumin (PV) [14], a calcium buffer expressed within FSINs in the dorsolateral prefrontal cortex (dlPFC). The NMDA hypothesis stems from the ability of NMDAR antagonists phencyclidine (PCP) and ketamine to induce both positive and cognitive deficits in healthy subjects that mirror schizophrenia and reinstate symptoms in stabilized patients with schizophrenia [15]. Accompanying the induction of cognitive deficits is a marked increase in neuronal firing and decreased coordinated bursting within the PFC [16], suggesting NMDAR antagonists act with greatest efficacy at inhibitory interneurons [17].
One possible site of convergence of the glutamatergic, GABAergic and dopaminergic hypotheses is at the level of inhibitory interneurons, which are driven by NMDA-mediated excitatory inputs and control inhibition of the excitatory pyramidal network throughout the brain, including in the hippocampus, which drives downstream DA neurons [18]. The nexus of the hypotheses at inhibitory interneurons offers a target for therapeutic intervention in selectively boosting inhibition [19]. Previous data point to the selective expression of GluN2C and GluN2D receptors at key classes of interneurons in the adult rodent hippocampus [20]. Though initially expressed in both pyramidal and inhibitory interneurons at birth, GluN2D expression decreases with development, becoming selectively enriched within PV- and somatostatin (SOM)-expressing fast-spiking interneurons in the adult hippocampus [21]. GluN2D expression was confirmed functionally by the modulation of NMDAR-mediated current using CIQ(+), a GluN2C/D specific PAM [22], in hippocampal interneurons of young mice, but not in pyramidal cells [23], nor in the GluN2D -/- animals [24].
Here we report that patterns of GluN2D selective expression described in the young murine hippocampus are conserved in PV+ interneurons of the dlPFC in postmortem adult human tissue. Consistent with previous findings, we show CIQ(+) increases the FSIN intrinsic excitability in addition to potentiating NMDAR-mediated excitatory currents onto FSINs in the adult mPFC. Together these results point to the viability of GluN2D-selective pharmacology in the remediation of NMDAR- and GABAergic hypofunction in schizophrenia.
Materials and methods
Human transcriptomics
Human laser-capture microdissection.Post-mortem brain samples were obtained from human subjects with a relatively small spread in age, low post-mortem interval (PMI) and limited agonal state to ensure high quality RNA could be collected. Subjects had no known neuropsychiatric or neurodegenerative disorders. No statistically significant differences between the two groups were found for age, PMI, brain pH, RIN or storage time (summarized in Table 1). Laser-capture methodology for RNA has been described previously [25,26]. From each sample, 12 mm sections of dlPFC (Brodmann area 9) were cut and stained with thionin for pyramidal neurons, or aggrecan for parvalbumin interneurons. For each cell type, 200 cells were cut from deep layer 3 and collected into RLT buffer with β-mercaptoethanol, then stored at -80°C. Adjacent stained and unstained slides were collected in the same manner to serve as controls. Informed consent for brain tissue donation was obtained from the next-of-kin via a recorded and witnessed telephone call with a licensed clinician using procedures approved by the University of Pittsburgh Committee on Research and Clinical Training Involving Decedents.
Human transcriptional profiling.Generation of microarray data from pyramidal and PV neurons was previously described [25,26]. For RNAseq, RNA was isolated using RNeasy micro kits (Qiagen), and cDNA libraries were prepared using the pico input SMARTer stranded total RNA-seq kit (Clontech). Library size was measured by high sensitivity DNA kit (Agilent), and concentration was quantified by Qubit (Life Technologies). Libraries were sequenced on a NextSeq500 (Illumina), and reads were aligned to the human genome using SALMON [27], and differential expression of genes between cell types was quantified using DESEQ2. On average, 5 million reads mapped to annotated gene regions for each sample. The number of genes with TPM (transcripts per million)>1 was 12,000–13,000 for the isolated cell types, and over 16,000 for slide controls. All data have been deposited to GEO (GSE149154).
Murine electrophysiology
All procedures were approved by the Institutional Animal Care and Use Committees of the Pfizer Neuroscience Research Unit and the University of California, San Francisco.
Adult (> 8-week old) male GAD67-GFP C57BL/6 mice (G42 line, Jackson Labs) were anesthetized with inhaled isoflurane and transcardially perfused with oxygenated ice-cold aCSF cutting solution containing (in mM): 130 NaCl, 26 NaHCO3, 2.5 KCl, 1.25 NaH2PO4, 0.5 CaCl2, 3 MgSO4 and 10 glucose. Adult (P40-P60) Dlx 5/6+/ mice (see [28]) were anesthetized with an intraperitoneal injection of euthasol and transcardially perfused with oxygenated ice-cold cutting solution containing (in mM): 210 sucrose, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 0.5 CaCl2, 7 MgCl2 and 7 dextrose. Following rapid decapitation, brains were removed and sectioned coronally on a vibrating microtome (VT1200S, Leica Biosystems). 250-μm slices containing the mPFC were transferred to warm (35°C) oxygenated aCSF containing (in mM) 130 NaCl, 26 NaHCO3, 2.5 KCl, 1.25 NaH2PO4, 2.5 CaCl2, 1 MgSO4 and 10 glucose and allowed to recover for> 1hr. Slices were then transferred to a recording chamber with aCSF containing 50 μM picrotoxin and 10 μM CNQX (Sigma Aldrich) heated to 35°C. Whole-cell recordings were performed on layer 5/6 neurons using (4–8 MOhm) electrodes containing (in mM): 150 K-Gluconate, 10 HEPES, 5 NaCl, 1 MgCl, 0.2 EGTA, 2 Mg-ATP and 0.5 Na-GTP. Cells were characterized in current clamp in response to 500 ms DC injections from –100 to +300 pA in 50 pA steps. Inhibitory (GFP+) FSINs were identified by rapid (>200 Hz) sustained firing [29]. Pyramidal cells with large (GFP-) somata showed rapid spike-frequency adaptation. For EPSC measures neurons were held at –50 mV to characterize weakly rectifying GluN2D-mediated evoked currents in response to bipolar stimulation of layer 2/3. Stimulation intensity was set to consistently evoke an EPSC with stable current amplitude. CIQ was dissolved in DMSO stock prior to dilution 1/1000. Baseline EPSCs were recorded for 10 min in solution containing DMSO vehicle prior to application of solution containing 10 μM CIQ(+) for 15 min. Average current amplitude for the final 5 min period was compared to that of the 5 minutes prior to application of CIQ(+). During current clamp recordings, series resistance and pipette capacitance were appropriately compensated. Series resistance was usually 10–20 MΩ, and experiments were terminated if series resistance exceeded 25 MΩ.
Data analysis
Data were analyzed using custom routines written in IGOR Pro (Wavemetrics). Input resistance was calculated from the steady-state current in response to a 10 mV step for voltage-clamp experiments and from the slope of the linear fit of the voltage–current plot generated from a family of hyperpolarizing and depolarizing current injections (-50 to +20 pA, steps of 10 pA) for current-clamp experiments. The membrane time constant (tau) was calculated as the slow component of a double‐exponential fit of the average voltage decay in response to a hyperpolarizing current injection (−400 pA, 1 ms). Firing output was calculated as the number of action potentials (APs) fired in response to 800 ms long depolarizing current injections (25–300 pA). Firing frequency was calculated as the number of APs fired per second. Sigmoid fits of firing frequency–current curves were used to obtain xhalf and linear rising rates. Rheobase was measured as the minimum current injection that elicited spiking. Firing traces in response to 50 pA current above the rheobase were used for analysis of single AP properties–AP threshold, maximum dV/dt (rate of rise of AP), AP amplitude, AP half-width and fast after hyperpolarization (fAHP) amplitude. Threshold was defined as the voltage at which the value of third derivative of voltage with time is maximal. Action potential amplitude was measured from threshold to peak, with the half-width measured at half this distance. Fast after hyperpolarization (fAHP) was measured from the threshold to the negative voltage peak after the AP.
Results
GluN2D transcription is selectively enriched in PV+ interneurons of the human dlPFC
Laser-capture microdissection was performed on thionin-stained pyramidal neurons and aggrecan-stained parvalbumin neurons in dlPFC. For each cell type, 100–200 microdissected cells were collected and processed for RNAseq alongside stained adjacent gray matter sections. Fig 1 demonstrates selectivity of each cell population for markers characteristic of each neuron type, such as VGAT (SLC32A1) and VGLUT2 (SLC17A6). Expression of the GluN2 subunits is shown in Fig 2 for the RNAseq samples, along with a second set of samples processed for microarray. GluN2A and B were expressed in both pyramidal neurons and PV interneurons, but also showed expression in the adjacent sections, suggesting that expression is not enriched in pyramidal or PV neurons relative to other cell types present in cortical gray matter. GluN2C showed very low expression in pyramidal neurons and parvalbumin interneurons, and higher expression in the section, suggesting that the majority of GluN2C expression in these regions comes from another cell type. GluN2D showed a clear enrichment in microdissected parvalbumin interneurons versus pyramidal neurons in both microarray and RNAseq data. The relatively low level of GluN2D in full sections relative to parvalbumin interneurons suggests that the latter cell type is the predominant cell expressing this receptor subtype in dlPFC. The interneurons stained with aggrecan expressed other markers of PV+ interneurons, such as KCNS3, but also some markers common to both PV+ and SST, such as LHX6 and SST, thus expression should be interneuron-specific, but not completely restricted to PV+ interneurons. The full TPM table for all samples can be found in S1 Table.
Fig 1. Interneuron versus pyramidal neuron enriched genes.Expression of cell-type specific or enriched genes in parvalbumin interneurons (black circles) versus pyramidal neurons (red triangles). Each point represents 200 laser-captured cells pooled from a single brain section. The y-axis represents TPM expression of each gene on a log scale. A vertical line separates interneuron markers (left) from pyramidal neuron markers (right).
https://doi.org/10.1371/journal.pone.0233895.g001Fig 2. GluN2D transcripts are enriched in PV interneurons relative to pyramidal neurons in the dlPFC.(A) Gene expression data as measured by microarray is plotted for GluN2A, B, C and D, and enrichment in one cell type versus the other is listed. Triangles represent pyramidal neurons, circles represent parvalbumin interneurons; intensity is expressed on a log2 scale. *p
Read More