Molecular and functional characterization of GAD67-expressing, newborn granule cells in mouse dentate gyrus

Abstract

Dentate gyrus granule cells (GCs) have been suggested to synthesize both GABA and glutamate immediately after birth and under pathological conditions in the adult. Expression of the GABA synthesizing enzyme GAD67 by GCs during the first few weeks of postnatal development may then allow for transient GABA synthesis and synaptic release from these cells. Here, using the GAD67-EGFP transgenic strain G42, we explored the phenotype of GAD67-expressing GCs in the mouse dentate gyrus. We report a transient, GAD67-driven EGFP expression in differentiating GCs throughout ontogenesis. EGFP expression correlates with the expression of GAD and molecular markers of GABA release and uptake in 2-4 weeks post-mitotic GCs. These rather immature cells are able to fire action potentials (APs) and are synaptically integrated in the hippocampal network. Yet they show physiological properties that differentiate them from mature GCs. Finally, GAD67-expressing GCs express a specific complement of GABAA receptor subunits as well as distinctive features of synaptic and tonic GABA signaling. Our results reveal that GAD67 expression in dentate gyrus GCs is a transient marker of late differentiation that persists throughout life and the G42 strain may be used to visualize newborn GCs at a specific, well-defined differentiation stage.

Keywords: GABA; GAD; adult neurogenesis; dentate gyrus; granule cells.

Figure 1

Dentate gyrus EGFP-positive neurons in G42 mice are mostly granule cells. (A) Widefield fluorescence images of the isocortex and hippocampus of G42 mice at P30. EGFP (green) colocalizes with parvalbumin (PV, red) in neocortical interneurons. However, in the hippocampus, EGFP expression is restricted mostly to a subpopulation of dentate gyrus neurons and does not colocalize with PV staining. Scale: right, 500 μm; left, 50 μm. (B) Confocal maximum projection images showing EGFP expression in the dentate gyrus of P7, P21, and 6 months-old mice. Dentate gyrus granule cells (GCs) are stained with Prox-1, a specific marker of GCs, showing colocalization with EGFP at all post-natal ages. Note the mossy fiber track revealed by EGFP immunoreactivity (arrow). EGFP expression persists in the adult, although in fewer cells (arrowheads), mostly confined near the subgranular zone. Boxed areas in right panels are shown enlarged in left panels. Scale: left, 200 μm; right, 50 μm. (C) Estimation of the proportion of GCs expressing EGFP showing a progressive decrease with age, reaching a steady-state in the adult. n ≥ 3 slices/animal and 3 animals/age.

Figure 2

GAD67-driven EGFP expression is restricted to immature granule cells in G42 dentate gyrus. (A) Maximum intensity projections of confocal images of hippocampus from P20–30 G42 mice, showing EGFP-positive GCs colocalize primarily with DCx as well as NeuN. Arrowheads indicate co-localization with EGFP whereas arrows indicate lack of co-localization. (B) Most EGFP-positive GCs express calretinin (Cr), another marker of immature GCs. Arrowheads indicate co-localization of EGFP with calretinin whereas arrows indicate lack of calretinin immunoreactivity in EGFP+ cells. (C) In contrast, no expression of calbindin (CB) was detected in EGFP-positive GCs (arrows). Scale: 20 μm. (D), Summary data of developmental marker quantification in juvenile (P14, P20–30) and adults animals (2–6 months) respectively. n ≥ 6 slices/animal; n ≥ 3 animals/age.

Figure 3

Timing of GAD67-driven EGFP expression in differentiating granule cells. P15 and P70 G42 mice received intraperitoneal injection of BrdU and were sacrificed at increasing times after injection. After fixation, brain slices were processed for BrdU and EGFP immunostaining. (A) Colocalization of BrdU and EGFP immunoreactivity in maximum intensity projections of confocal stacks reveals EGFP+ neurons born at the time of injection. Boxed areas in left panels are shown enlarged in right panels. Scale 50 μm and 20 μm (insets). (B) Top, scheme of BrdU treatment and time course of experiment; Bottom, quantification of data revealing expression of EGFP in approximately 14–28 days postmitotic GCs, both in young and adult mice. n ≥ 2 animals/age/day post-injection.

Figure 4

Cluster analysis of dentate gyrus GCs in G42 mice based on their physiological properties and transcriptome. X-axis, individual cells; Y-axis, average within-cluster linkage distance. Two main groups can be distinguished: group 1, including immature GCs, is composed of EGFP+ cells from both juvenile and adult mice (49.2%) and EGFP− cells from juvenile animals (50.8%). Group 2, includes EGFP− GCs from adults. Dotted line indicates the limits between clusters suggested by the Thorndike procedure.

Figure 5

Electrophysiological properties of GAD67/EGFP expressing and non-expressing GCs. (A) Compared firing of each GC group identified in the unsupervised cluster analysis: group 1 EGPF+, group 1 EGFP−, and group 2. Response to a +150 pA current step. Insets show the first action potential of the train. Dotted line: 0 mV. Scale: 15 mV, 300 ms (20 mV, 10 ms for insets). (B) Summary data of input resistance for all groups (n = 32 for group 1 EGFP+, n = 33 for group 1 EGFP−, and n = 22 for group 2). (C,D) Summary data of action potential (AP) amplitudes and frequency (number during a 500 ms pulse) for all groups (n = 22 for group 1 EGFP+, n = 29 for group 1 EGFP−, and n = 22 for group 2). (E) Recordings of spontaneous PSCs (spPSCs) at −70 mV in voltage clamp mode, in each subtype of GCs. Scale; 15 pA, 100 ms. (F,G) Summary data of spPSC amplitude and frequency, respectively (n = 25 for group 1 EGFP+, n = 24 for group 1 EGFP−, and n = 20 for group 2). Mann–Whitney rank sum test significance ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, for comparisons with group 1 EGFP+ GCs and ^##p < 0.01, ^###p < 0.001 for comparison with group 1 EGFP− cells.

Figure 6

Molecular characterization of GAD67/EGFP expressing and non-expressing GCs. (A) Agarose gel analysis of the scPCR products from a representative cell of each GC group (group 1 EGFP+, group 1 EGFP− and group 2). (B–D) Quantification of developmental markers (DCX, doublecortin; CB, calbindin; Cr, calretinin), glutamate/GABA synthesis, vesicular release and uptake and subunits of GABAA and glutamate receptors. n = 32, 33, and 22 for group 1 EGPF+, group 1 EGFP− and group 2 GCs. Significance ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, for comparisons with group 1 EGFP+ GCs and ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 for comparison with group 1 EGFP− cells.

Figure 7

Differential GABA signaling in GAD67/EGFP expressing vs. non-expressing GCs. (A) Top spIPSC recordings from Group2/EGFP− and Group1/EGFP+ GCs. Bottom, 150–200 spIPSCs from each recording in top panels were averaged (left) and scaled in peak amplitude (right) to show similar decay kinetics. (B) Summary plots of mean IPSC frequency and amplitude from 11 EGFP− and 9 EGFP+ GCs. ^**p < 0.005. (C) Top, recording of a representative EGFP− mature GC from young adult (P35) G42 mouse. 2 μM gaboxadol (GBX) was applied after a 5 min control before bicuculline (40 μM) was added to the recording solution. Bottom, same as above in a representative EGPF+ GC. Note the small effect of GBX in the EGFP+ as compared to that observed in the EGFP− GC, suggesting a lower contribution of δ subunit-containing GABAA receptors. (D) Summary data of GBX-induced currents in EGFP+ and EGFP− GCs (E) Summary data of GBX effect on the standard deviation of holding current in EGFP− (n = 11) and EGFP+ (n = 9) GCs. Mann–Whitney rank sum test significance: ^**p < 0.01, ^***p < 0.001.