Pioneer GABA cells comprise a subpopulation of hub neurons in the developing hippocampus

Abstract

Connectivity in the developing hippocampus displays a functional organization particularly effective in supporting network synchronization, as it includes superconnected hub neurons. We have previously shown that hub network function is supported by a subpopulation of GABA neurons. However, it is unclear whether hub cells are only transiently present or later develop into distinctive subclasses of interneurons. These questions are difficult to assess given the heterogeneity of the GABA neurons and the poor early expression of markers. To circumvent this conundrum, we used "genetic fate mapping" that allows for the selective labeling of GABA neurons based on their place and time of origin. We show that early-generated GABA cells form a subpopulation of hub neurons, characterized by an exceptionally widespread axonal arborization and the ability to single-handedly impact network dynamics when stimulated. Pioneer hub neurons remain into adulthood, when they acquire the classical markers of long-range projecting GABA neurons.

Figure 1. The distribution of EGins in horizontal sections of the hippocampus in P7 and adult mice

A–B. Fate-mapped sparse neurons (arrows) are visible all over the hippocampus after GFP immunoperoxidase detection in a P30 tamoxifen-treated at E7.5 Dlx1/2^CreERTM;RCE:LoxP mouse. B1–3 are enlargements of the areas indicated in A showing the multipolar feature of labelled somata and their relative extensive dendritic arborisation. Axonal processes (double arrows) can be observed in all layers of the Dentate Gyrus (DG) and Hilus (B1), CA3 (B2) and CA1 (B3) areas. C. Histogram reporting the fraction of EGins present in different hippocampal layers of P7 (410 cells, n = 7 pups) and P30 (963 cells, n = 5 mice) tamoxifen-treated Dlx1/2^CreERTM;RCE:LoxP animals. Alv: alveus; gl: granular layer; H: hilus; ml: molecular layer; slm, sr, sl, sp, so, stratum laculosum moleculare, radiatum, lucidum, pyramidale and oriens, respectively. Scale: 100 μm.

Figure 2. The morphological and neurochemical properties of EGins in horizontal sections of the hippocampus at P7

A–B. Distribution of GFP-expressing early generated neurons (EGins, induction: E7.5) and late generated neurons (LGins) in horizontal sections of the hippocampus of P7 mice. EGins (arrows in A) represent a sparse population compared to the numerous LGins (arrows and double arrows in B). LGins are present in all hippocampal layers but are predominantly located in stratum pyramidale (sp) and stratum lacunosum-moleculare (slm). Several LGins exhibit morphological features resembling neurogliaform cells (double arrows in B, B1). C. GFP positive axons are running along the fimbria (arrows) in EGins-GFP mice. D–F. Examples of immunofluorescence for SOM, mGluR1α and M2R in EGins within CA3 stratum oriens (so). Abbrevations: alv, alveus; DG, dentate gyrus; gl, granular layer; H, hilus; ml, molecular layer; sl, sr, stratum lucidum and radiatum. Scale is 100 μm except for B1 where it is 20 μm.

Figure 3. SOM and mGluR1α are preferentially expressed in EGins in the adult hippocampus

All images are from the CA3 area. All GFP positive neurons correspond to EGins labelled using tamoxifen-treated Dlx1/2^CreERTM;RCE:LoxP mice (induction: E7.5). A. A GFP positive EGin (arrow) co-expressing mGluR1α and SOM in the CA3 stratum oriens. B–E. Generally, EGins do not express PV (arrows in B, D, E) even though very occasional PV positive cells can be seen (arrow in C). GFP positive cells never co-express PV and SOM (arrows in C, D, E) but can be observed next to cells that are positive for PV or SOM (single or double arrowheads). Note that the GFP cell in D (arrow) is co-labelled for SOM. F. M2R labelling in a GFP positive cell (arrows). G–K. In the vast majority of fate-mapped GFP positive cells (arrows), neither immunoreactivity (arrowheads) for VIP, nor for NPY, CB, CR, NOS is detected. L. Histogram showing the fraction of EGins immunoreactive for the various tested molecular markers. Note that two thirds and almost half of the EGins contain mGLuR1α and SOM, respectively. The n numbers in brackets represent the number of tested mice. M. In triple labelled sections, most of the GFP and SOM positive cells are also labelled for mGluR1α. Few EGins are mGluR1α positive only, and even fewer only SOM positive. A few EGins were immunonegative for both markers. Data were obtained from 4 tamoxifen-treated Dlx1/2^CreERTM;RCE:LoxP mice. Abbrevations: sl, sp, so, stratum lucidum, pyramidale and oriens, respectively. Scale: 100 μm.

Figure 4. Morphological features of early- and late-generated hippocampal GABA interneurons

A. 1. Distribution of the somatic location of all morphologically recovered EGins (n=21) on a schematic representation of the hippocampus. All these cells correspond to GFP positive neurons in tamoxifen-treated Dlx1/2^CreERTM;RCE:LoxP mice that were filled with neurobiotin and processed post hoc. Scale: 200μm. 2. Photomicrograph at low (scale: 200μm) and high magnification (inset, scale: 50μm) of the confocal image of a neurobiotin filled EGin. Neuronal morphology was analyzed between P5 and 7. Tamoxifen was provided at E7.5 except for cell#23 for which induction occurred at E9.5. The Neurolucida reconstruction of the cell is superimposed and shows a dense and widespread axonal arborisation (orange). Soma and dendritic arborisation are white. 3. Neurolucida reconstructions of six representative EGins. Note that despite their variable morphological features, all these cells display a long axonal arborisation (orange) spanning across hippocampal layers. Axons may present a dense arborization (e.g. #35, 25 and 23) and/or a long branch often running toward the fimbria (arrow in cells # 36, 37, 24 and 35). Scale: 200μm. B. Same as (A) but for LGins (n=9) corresponding to neurobiotin-filled GFP positive neurons from P5–7 Mash1^CreERTM;RCE:LoxP mice tamoxifen-treated at E18.5 (see Experimental Procedures). Note that the axonal arborisation (green) of these cells is more confined and less ramified than that of early generated neurons. Scale: 200μm.

Figure 5. EGins display remarkable morphometric features similar to those of hub neurons

A. Bar graphs comparing the averaged axonal (top) and dendritic (bottom) lengths (left) and surfaces (right) obtained in Neurolucida reconstructed High Connectivity (HC from (Bonifazi et al., 2009), red, n=10), early generated (EGin, orange, n= 20), Low Connectivity (LC from (Bonifazi et al., 2009), dark green, n=10) and late generated (LGin, light green, n=10) interneurons. *: p<0.05, **: p<0.001. B. Normalized distribution graphs of the fraction of axonal intersections with concentric circles of increasing radius (20μm steps) centred at the soma for the cell populations described in A (same color code). Distributions obtained for HC and EGins are best fitted by a log normal function (black) indicating heavy tailed distributions whereas LGins are best fitted by two Gaussian functions (black). C. Cluster analysis tree of the morphological variables describing the same cells as in A (Ward’s method, Dlink: Euclidian distances, see Experimental Procedures). Distances were normalized. Most HC and EGins segregated in the same group whereas LC and LGins segregated in another.

Figure 6. EGins cells display a high effective connectivity

A. Neurolucida reconstruction of a CA3c EGin (induction: E9.5, analysis: P6) on a schematic representation of the hippocampus reveals an extended axonal arborisation (orange, dendrites and soma in black) Black rectangle marks the imaged region and includes the contour map of imaged cells (gray). Image below is a color coded representation of the effective connectivity map (see Experimental Procedures). The asterisk indicates the cell body position. Red color represents high density of effective targets (A.U.: arbitrary units). This EGin therefore displays a high effective connectivity. B. Same as (A), but for a representative LGin (induction: E18.5, analysis: P6). Note that stimulation of that cell does not induce any significant calcium response in other imaged neurons, even though the axonal arborisation (green) of that LGin was fairly developed. Scale: 200 μm.

Figure 7. Stimulation of early- but not late- generated interneurons significantly perturbs network dynamics

A. Summary table of the effects on network dynamics after stimulating EGins.. Significant results are displayed in orange whereas gray boxes indicate no effect. Columns indicate the three possible types of effects (illustrated in B): 1. Increase (+) or decrease (−) of GDP frequency (GDPf); 2. GDP triggered after stimulation (p value obtained after statistical analysis of the PeriStimulusTimeHistogram (PSTH) is indicated; 3. Forward (1) or backward (−1) phase (Φ) shift. B. Three types of effects on network dynamics observed in response to repetitive phasic stimulation of EGins. Tamoxifen induction: E9.5 (cells #22-21), E7.5 (#34). Analysis: P6–7. 1. GDP Triggering: GDP occurrence (indicated by *) as a function of time following repetitive phasic stimulation of a representative EGin (22 consecutive trials). 6 out of 22 triggered a GDP. Reconstruction of the corresponding EGin is illustrated below (axon: orange; dendrites and soma: black). Scale bar: 200μm. 2. GDP frequency: graph plotting the time interval between GDPs (inter GDP interval) as a function of time in a representative EGin (illustrated below same as (1)) significantly decreasing the GDP occurrence when stimulated (orange). 3. Backward phase shift: Same stimulation as (1) but for a representative EGin inducing a backward GDP phase shift when stimulated. Phase shift is illustrated in graph plotting the number of GDP cycles skipped during phasic stimulation (orange) as a function of time. The number of expected GDPs was calculated during resting conditions (white) based on the average interval between GDPs. Recorded cell is illustrated below (same as 1). C. Same as (B), but stimulating LGins. Stimulation of none of the tested LGins perturbed network dynamics significantly despite their fairly developed axonal morphology (green). Tamoxifen induction: E18.5; Analysis: P6–7.

Figure 8. Electrophysiological characteristics of EGins

A. Current clamp traces of EGins recorded at resting membrane potential during stimulation (orange) corresponding to the same cells as illustrated on Figure 7B. 1. Three trials showing a stimulation (square pulse below) followed by a polysynaptic membrane potential depolarization characteristic of GDPs (marked by*). 2. The frequency of GDPs (marked by*) is significantly reduced during stimulation (indicated by square pulses below). 3. Current-clamp recordings from 4 consecutive stimulation trials (square pulse) show the progressive delay in the occurrence of a GDP (*) following stimulation (orange). B. Table comparing basic electrophysiological measurements obtained in EGins, LGins and previously sampled High Connectivity hub cells (HC). P values obtained when comparing EGins with LGins or HC cells are indicated (Student or Mann-Whitney tests). Vrest: resting membrane potential; Rinput: input resistance; Vthreshold: action potential threshold; AP width: action potential width measured at half maximal amplitude; AP amplitude: action potential amplitude; sEPSP: spontaneous Excitatory PostSynaptic Potentials.