Transcriptional Dysregulation of γ-Aminobutyric Acid Transporter in Parvalbumin-Containing Inhibitory Neurons in the Prefrontal Cortex in Schizophrenia

Abstract

Parvalbumin (PV)-containing neurons are functionally compromised in schizophrenia. Using double in situ hybridization in postmortem human prefrontal cortex, we found that the messenger RNA (mRNA) for the γ-aminobutyric acid transporter GAT-1 was undetectable in 22-41% of PV neurons in layers 3-4 in schizophrenia. In the remaining PV neurons with detectable GAT-1 mRNA, transcript expression was decreased by 26% in layer 3. Hence, the dysfunction of PV neurons involves the molecular dysregulation of presynaptic GABA reuptake.

1.INTRODUCTION

Schizophrenia is a neurodevelopmental disorder (Lewis and Murray, 1987; Weinberger, 1987) characterized by deficits in cognitive processes mediated by the circuitry of the dorsolateral prefrontal cortex (DLPFC) (Lewis and Levitt, 2002). Although no single pathological mechanism exists to explain the evolution of the symptoms associated with schizophrenia, it is considered a disease of aberrant synaptic plasticity and dysconnection stemming from deficits in key processes and structures including myelogenesis (Benes, 1989; Bartzokis et al., 2003), synaptic pruning (Feinberg, 1982; Keshavan et al., 1994; McGlashan and Hoffman 2000), perineuronal nets (Bitanihirwe and Woo, 2014) in addition to the active maturation of cortical inhibitory circuits (Berretta and Benes, 2001, Lewis et al., 2012).

Inhibitory neurons in the cerebral cortex are morphologically, neurochemically, physiologically and thereby functionally diverse, with different subsets of these neurons regulating distinct aspects of information processing (Buzsaki, 2006; Markram et al., 2004; Soltesz, 2005; Wang et al., 2004). Recent evidence suggests that disturbances of the sculpting of pyramidal circuit network activities by inhibitory neurons represent a core pathophysiologic feature of schizophrenia (Benes and Berretta, 2001; Lewis et al., 2005). Thus, the expression of the mRNA for the 67 kD isoform of the synthesizing enzyme of γ-aminobutyric acid (GABA), glutamic acid decarboxylase (GAD)₆₇, has been consistently shown to be decreased in the cerebral cortex in schizophrenia (Akbarian et al., 1995; Costa et al., 2004; Volk et al., 2000; Woo et al., 2008; Woo et al., 2004). Furthermore, this reduction appears to occur preferentially, albeit not exclusively, in the subset of inhibitory neurons that contain parvalbumin (PV) (Hashimoto et al., 2003).

PV-containing neurons comprise two connectionally distinct populations of cells that synaptically target distinct compartments of pyramidal neurons: the perisomatically projecting basket neurons and the axoaxonally projecting chandelier neurons (DeFelipe and Farinas, 1992; Freund and Katona, 2007; Howard et al., 2005). The axon terminals of chandelier neurons form characteristic candlestick-like profiles that are termed “cartridges”. The density of these cartridges, which can be visualized immunohistochemically using an antibody against the GABA transporter GAT-1 (DeFelipe and Gonzalez-Albo, 1998), has been found to be decreased by ~40% in the DLPFC in schizophrenia (Konopaske et al., 2006; Pierri et al., 1999; Woo et al., 1998). Furthermore, the expression of the GABA_A μ2 subunit protein, which is preferentially localized to synapses formed by chandelier neurons, has also been shown to be increased (Volk et al., 2002). Together these findings suggest that, in schizophrenia, chandelier neuron-mediated inhibitory neurotransmission at the axon initial segment of pyramidal neurons is disturbed at both the pre- and postsynaptic sites. In this regard, it has also been shown that the number of the GABA_A μ1 receptor subunit, which is enriched in synapses formed by basket neurons, is significantly decreased in subjects with schizophrenia, suggesting that, like chandelier neurons, inhibitory regulation of pyramidal neurons by basket neurons is also aberrant (Curley et al., 2011; Curley and Lewis, 2012). Consistent with this, a recent study by Glausier and colleagues reported that while the number of PV basket cell inputs to pyramidal neurons was unaltered in schizophrenia, the amount of PV protein within PV basket cell axon terminals was decreased by 23% (Glausier et al., 2014). Likewise, the amount of GAD₆₇ protein in PV basket cell terminals has also been shown to be decreased by as much as 50% (Curley et al., 2011). In the context of these observations, the present study aimed to evaluate whether pre-synaptic GABA reuptake in PV neurons is altered in subjects with schizophrenia.

2. METHODS and MATERIALS

2.1. Subjects

Postmortem human brains from 20 schizophrenia and 20 normal control subjects, matched for age, postmortem interval (PMI), freezer storage time, pH and sex were obtained from the Harvard Brain Tissue Resource Center (HBTRC), at McLean Hospital in Belmont, Massachusetts (Table 1). Brain collection procedures of the HBTRC, including the informed consent process, were approved by the Partners Human Research Committee. Written informed consent for use of each of the brains for research was obtained by the legal next of kin. The diagnosis of schizophrenia was made by two psychiatrists by reviewing medical records and an extensive family questionnaire that included medical, psychiatric and social history. All of the brains were also examined by a neuropathologist to rule out any neurological conditions, such as the various forms of dementia. Toxicological analysis revealed that no substances of abuse were detected in any of the schizophrenia or normal control subjects at the time of death.

Table 1.

Demographic information on subjects included in the present study

Case	Diagnosis		Age	Sex	Race	pH	PMI	CED	Cause of Death
	Psychotropics received at the time of death
1	CON		49	M	W	6.76	24.6		Myocardial
infarction	None
2	CON		37	M	W	6.68	18.8		Electrocution
	None
3	CON		54	M	W	6.53	24.2		Cardiopulmonary
arrest	None
4	CON		78	F	W	6.22	14.1		Myocardial
infarction		None
5	CON		53	M	W	U	20.2		Cardiopulmonary
arrest	None
6	CON		65	F	W	6.40	24.3		Lung cancer
	None
7	CON		89	M	W	6.39	7.42		Cancer
		None
8	CON		69	M	W	6.88	15.3		Respiratory failure
		None
9	CON		74	F	W	U	12.5		U
	None
10	CON		66	F	W	6.03	7.4		Cancer
		None
11	CON		42	M	W	6.78	18.3		Myocardial
infarction		None
12	CON		78	F	W	6.67	23.9		Breast cancer
	None
13	CON		40	M	W	6.24	16.6		Myocardial
infarction		None
14	CON		67	M	W	6.42	22.3		Cardiopulmonary
arrest	None
15	CON		70	F	W	6.26	22.5		Liver cancer
	None
16	CON		66	M	W	6.76	18.7		Myocardial
infarction		None
17	CON		79	M	W	6.74	20.9		Cancer
		None
18	CON		38	M	W	6.53	28.8		Myocardial
infarction		None
19	CON		70	F	W	6.59	15		Cardiac arrest
		None
20	CON		29	M	W	U	19		U
		None
Mean (±SD)			60.4±17.3		6.58±0.25	18.7±5.5
21	SZ		85	F	W	U	15.7	150	Sepsis
		Risperidone, lorazepam
22	SZ		48	F	W	6.63	33.8	450	Cardiac arrest
		Risperidone, divalproex
23	SZ		44	M	W	6.20	19	266	Pneumonia
		Clozapine
24	SZ		89	F	W	U	13.5	20	Pneumonia
		Trifluoperazine
25	SZ		78	F	W	6.81	13.4	750	Sinus node disease
		Haloperidol, lithium, benztropine
26	SZ		61	M	W	6.68	19.9	300	Sepsis
		Clozapine
27	SZ		61	F	W	6.14	11	150	Myocardial
infarction		Paroxetine, clonazepam, clozapine
28	SZ		84	F	W	6.14	25.8	U	Cardiac arrest
	None
29	SZ		26	M	W	6.75	16	357	Suicide by hanging
		Fluphenazine,decanoate
30	SZ		55	F	W	6.52	18	U	Lung cancer
	None
31	SZ		47	M	W	6.57	19.2	U	Lung cancer
	Clonazepam, hydroxyzine
32	SZ		73	F	W	6.08	24	600	Lung cancer
	Risperidone, fluoxetine, clorazepate, midazolam
33	SZ		49	M	W	6.60	19	500	Suicide by hanging
		Haloperidol decanoate, lorazepam
34	SZ		63	M	W	6.55	22.3	500	Cardiac arrest
		Cloazapine, haloperidol, lorazepam, trazodone
35	SZ		72	F	W	6.65	21.7	400	Ovarian cancer
		Risperidone, paroxetine
36	SZ		66	M	W	6.43	22.1	1000	Emphysema
	Haloperidol
37	SZ		83	F	W	6.91	23.2	2000	Gastrointestinal
bleed		Haloperidol decanoate
38	SZ		46	F	W	6.31	18.5	200	Sepsis
		Olanzapine, divalproex
39	SZ		42	M	W	6.64	27.1	U	Leukemia
	None
40	SZ		31	M	W	6.46	14	600	U
	Risperidone, olanzapine, buproprion
Mean (±SD)			60.2±16.7		6.65±0.28	19.8±5.4	433±468.0

Abbreviations are as follows: PMI = postmortem interval, CED = chlorpromazine equivalent dose, U = unknown or unavailable, CON = normal control, SZ = schizophrenia, M = male, F = female, W = white

2.2. Tissue preparation and processing

Two sections (10 μm) per subject were taken from Broadmann area 9 of the DLPFC for in situ hybridization analysis. Laminae from the DLPFC were delineated and identified in adjacent sections by Nissl-counterstaining.

All of the experimental methods, including hybridization conditions (Bitanihirwe et al., 2009; Bitanihirwe et al., 2010; Woo et al., 2008; Woo et al., 2004), visualization of digoxigenin (DIG) and radiolabled probes (Bitanihirwe et al., 2009; Bitanihirwe et al., 2010; Woo et al., 2008; Woo et al., 2004) in addition to quantification procedures (Bitanihirwe et al., 2009; Bitanihirwe et al., 2010; Woo et al., 2008; Woo et al., 2004) have been described extensively in previous publications.

2.3. Riboprobe preparation

The complementary RNA (cRNA) probe for GAT-1, which was radiolabeled with Sulfur-35, was transcribed in vitro from a cDNA spanning nucleotides 1185 to 2154 within the coding region of the human SLC6A1 gene (Genbank Accession No. NM_003042). A corresponding sense probe was also generated. Hybridization of the sense probe resulted in no specific labeling. The DIG-labeled PV mRNA probe utilized here has been used in previously published studies (Bitanihirwe et al., 2009; Bitanihirwe et al., 2010).

2.4. Statistical Analysis

The densities of single (PV+) and double-labeled (PV+/GAT-1+) neurons and the amount of mRNA for GAT-1 in PV+ neurons were compared between the schizophrenia and normal control groups across layers 2 through 6 using repeated-measures analysis of variance (ANOVA) with diagnosis and layer as main effects, as previously described (Bitanihirwe et al., 2009; Bitanihirwe et al., 2010). We evaluated the effects of all of the confounding variables, such as age, PMI, pH, freezer storage time and exposure to antipsychotic medication (expressed as chloropromazine equivalent dose or CED) using analysis of covariance (ANCOVA). Because ANCOVA did not detect any statistically significant effect of any of the covariates on the dependent variables, only results from repeated-measures ANOVA are reported. In addition, Pearson's correlation was used to assess if there was any linear relationship between cell or grain densities and any of the continuous variables. All statistical tests were calculated with α < .05, and computed using JMP (SAS Institute Inc, Cary, NC).

3. RESULTS

Consistent with previous observations (Hashimoto et al., 2003; Woo et al., 1997), we found that the density of cells that expressed a detectable level of PV mRNA was not altered in subjects with schizophrenia. The use of DIG, however, precludes us from addressing any possible alteration in transcript expression level per neuron. For the PV neurons that expressed GAT-1 mRNA, the effect of diagnosis was highly significant (F=20.95, P<0.0001) and this effect was layer-specific (F=5.54, P=0.02). Thus, the density of these neurons was decreased by 22% and 41% in layers 3 and 4, respectively (Figure 1). We observed a significant effect of diagnosis on the density of silver grains per neuron (F=4.49, P<0.04; Figure 1), which reflected a 26% decrease in grain density in layer 3 in the schizophrenia subjects. Hence, it appears that in the PV neurons with detectable GAT-1 mRNA, the amount of this transcript in layer 3 was decreased in the subjects with schizophrenia. Finally, we detected no statistically significant correlation between any of the potential confounding covariates and cell or grain densities. ANOVA also revealed no statistically significant interaction between sex and diagnosis.

Figure 1.

Upper panel: Mean (±SEM) density of PV+/GAT-1+ neurons is significantly decreased in layers 3 and 4 in the PFC in schizophrenia. Photomicrograph shows a double-labeled (PV+/GAT-1+) neurons. Scale bar = 10 μm. Lower panel: Mean (±SEM) density of silver grains over PV+ neurons is significantly decreased in layer 3 in the PFC in schizophrenia. Layer 1 was not included in the analyses because no PV+ neurons were found in this layer. *p<0.05, based on post-hoc group comparisons.

4. DISCUSSION

In the present study, we found that the expression of the mRNA for GAT-1 was undetectable in 22-41% of PV neurons and, among the remaining PV neurons with detectable GAT-1 mRNA, the amount of this transcript was decreased by 26%. Because basket neurons constitute the majority, possibly as much as 80-90%, of all PV neurons (Kawaguchi, 1995; Krimer et al., 2005; Markram et al., 2004; Zaitsev et al., 2004), GAT-1 mRNA reduction must occur in basket neurons in order to account for the magnitude of the reduction in neuronal densities and GAT-1 expression observed in this study. Together with the recent findings of decreased PV and GAD₆₇ protein levels within PV basket cell axon terminals in schizophrenia (Glausier et al, 2014; Curley et al., 2011), the findings of the current study strengthen the notion that presynaptic PV basket cell neurotransmission is altered in this illness and suggest that this alteration involves the molecular dysregulation of GABA reuptake. However, the extent to which the observed decrease in mRNA expression can be translated into altered protein content is unclear.

In concert with the well-documented observation of decreased expression of the mRNA for GAD₆₇, the finding of decreased GAT-1 mRNA expression in this study may represent a downstream consequence of some yet-to-be-defined pathophysiologic process (Curley et al., 2013). Interestingly, recent evidence suggests that oxidative stress may be a key mechanism that could contribute to PV neuronal dysfunction in schizophrenia, stemming from a combination of predisposing gene variations and environmental factors that include immune activation and stress (Behrens and Sejnowski, 2009; Bitanihirwe and Woo, 2011; Do et al., 2009; O'Donnell, 2012; O'Donnell et al., 2014). In this context, it would be of interest to investigate if the downstream consequences of oxidative stress may include decreased GAD₆₇ and GAT-1 expression by utilizing animal models.

• - Expression of GAT-1 mRNA is reduced in the DLPFC of schizophrenia subjects.

• - Pre-synaptic PV neurotransmission is altered in schizophrenia.

• - These findings suggest that GAT-1 reduction occurs in PV-expressing basket cells.