Neuroanatomic relationships between the GABAergic and serotonergic systems in the developing human medulla

Abstract

gamma-Amino butyric (GABA) critically influences serotonergic (5-HT) neurons in the raphé and extra-raphé of the medulla oblongata. In this study we hypothesize that there are marked changes in the developmental profile of markers of the human medullary GABAergic system relative to the 5-HT system in early life. We used single- and double-label immunocytochemistry and tissue receptor autoradiography in 15 human medullae from fetal and infant cases ranging from 15 gestational weeks to 10 postnatal months, and compared our findings with an extensive 5-HT-related database in our laboratory. In the raphé obscurus, we identified two subsets of GABAergic neurons using glutamic acid decarboxylase (GAD65/67) immunostaining: one comprised of small, round neurons; the other, medium, spindle-shaped neurons. In three term medullae cases, positive immunofluorescent neurons for both tryptophan hydroxylase and GAD65/67 were counted within the raphé obscurus. This revealed that approximately 6% of the total neurons counted in this nucleus expressed both GAD65/67 and TPOH suggesting co-production of GABA by a subset of 5-HT neurons. The distribution of GABA(A) binding was ubiquitous across medullary nuclei, with highest binding in the raphé obscurus. GABA(A) receptor subtypes alpha1 and alpha3 were expressed by 5-HT neurons, indicating the site of interaction of GABA with 5-HT neurons. These receptor subtypes and KCC2, a major chloride transporter, were differentially expressed across early development, from midgestation (20 weeks) and thereafter. The developmental profile of GABAergic markers changed dramatically relative to the 5-HT markers. These data provide baseline information for medullary studies of human pediatric disorders, such as sudden infant death syndrome.

Figure 1

Semi-quantitative grading system to compare immunostaining intensity within and across different cases and at different ages. (A). Schematic showing 5-HT nuclei. We first identified upon survey of all cases a site as an internal standard which was immunostained with the same intensity irrespective of age; this site was the (B) ependymal cells in the lining of the fourth ventricle. We designated the intensity of this “standard” as grade 2 and visually scored neurons in the selected nuclei relative to this standard as: C. Grade 0 = no neuronal immunostaining; D. Grade 1 = less intense immunostaining than standard, E. Grade 2 = equal immunostaining to standard; F. Grade 3 = more intense immunostaining than standard, as illustrated for the GABA_Aα3 antibody (Fig. 1). Scale bars = 100μm.

Figure 2

GAD65/67 immunoreactivity in neurons in the hypoglossal and in the raphé obscurus and pre-absorbtion assay with GAD65/67 peptide which blocks antibody staining. (A) Immunostaining for GAD65/67 localized to the raphé obscurus (ROb); at 41 postconceptional weeks (term). (B) Immunostaining for GAD65/67 in raphé obscurus blocked when incubated with GAD65/67 peptide. (C) Immunostaining for GAD65/67 localized to the neurons in the hypoglossal nucleus at 41 postconceptional weeks (term). (D) Immunostaining for GAD65/67 in hypoglossal nucleus blocked when incubated with GAD65/67 peptide. Scale bars = 100μm.

Figure 3

Western blot analysis using infant medulla lysate of GAD65/67, with band seen at 65kDa. We incubated with 2 concentrations of immunizing peptide 7×, and 20×; protein was completely blocked as seen in the immunohistochemistry when incubated with peptide also.

Figure 4

The co-expression of GAD65/67 and TPOH in neurons in the raphé obscurus (ROb), with the site indicated in blue in the diagram at 41 weeks postconceptional weeks (term) ×40. (A) TPOH positive neuron; (B) GAD65/67 positive; (C) Merged images. Scale bars = 100 μm.

Figure 5

GAD65/67 immunoreactivity in the hypoglossal nucleus at 41 postconceptional weeks (term) with the site demonstrated schematically (×40). (A) Punctate GAD65/67 immunostaining localizes around the edge of the somata consistent with GABAergic terminals projecting upon the somata. (B) Positively-stained alpha motor neurons for choline acetyl transferase (ChAT) indicative of the cholinergic phenotype of these motorneurons; (C) alpha motor neurons with only surface punctate immunostaining around the edge of the somata. Scale bars = 100 μm.

Figure 6

³H-GABA binding (fmol/mg tissue) to GABA_A receptors with tissue autoradiography, at the (A) mid (level of hypoglossal nucleus) level section of the medulla (B), rostral (level of prepositus nucleus) level section of the medulla; (C) Non-specific section (incubated with high concentration of ligand displacer) of medulla. The distribution of GABA_A binding is ubiquitous across nuclei (see text).

Figure 7

GABA_A receptor immunoreactivity in 5-HR Medullary Nuclei at 41 postconceptional weeks. The regional distribution of the GABA_A receptor subtypes studied is consistent with that found for the GABA_A receptor binding. Immunostaining is punctate and localized to neuronal cell bodies and processes, consistent with postsynaptic localization of the receptors. Scale bars = 100 μm.

Figure 8

Developmental profile of GABA_A receptor subtypes in 5-HT medullary nuclei in the medulla of the human fetus and infant. From 15 gestational weeks to the end of the first postnatal year, there is a significant increase in GABA_Aα3 expression in the raphé obscurus (p<0.001), gigantocellularis (p=0.005) and the paragigantocellularis lateralis (p=0.003). The GABA_Aα1 receptor expression changes with age with a significant decrease in the raphé obscurus (p=0.02) and non-significant trends to decrease in the gigantocellularis and paragigantocellularis.

Figure 9

Double-label immunofluorescent images showing the co-localization of the GABA_Aα3, with 5-HT neurons in the raphé obscurus and paragigantocellularis lateralis. (A) 5-HT positive neurons in raphé obscurus, B. GABA_Aα3 positive neurons in raphé obscurus; C. Merged images, D. 5-HT positive neurons in paragigantocellularis lateralis; E. GABA_Aα3 positive neuron in paragigantocellularis; F. Merged images. Scale bars = 100 μm.

Figure 10

Developmental profile of KCC2 in 5-HT medullary nuclei in the medulla of the human fetus and infant from cases of different ages 22 and 34 gestational weeks, 41 postconceptional weeks (term) and 80 postconceptional weeks. KCC2 immunostaining is punctate and localized to neuronal cell bodies and processes in early fetal development, but through development KCC2, immunoreactivity increases to the neuropil. Scale bars = 100 μm.

Figure 11

Double-label immunofluorescent images showing KCC2-positive fibers are not axons, as KCC2 immunoreactivity did not colocalize with fibers positive for axonal neurofilament (NFT) and myelin basic protein (MBP) in the raphé obscurus and paragigantocellularis lateralis, (×40) A. KCC2 positive fibers in the PGCL; B. NFT positive fibers in PGCL; C. Merged images. D. KCC2 positive fibers in the PGCL); E. MBP positive fibers in PGCL; F. Merged images. Scale bars = 100 μm.

Figure 12

Double-label Immunofluorescent images showing the co-expression of KCC2 and MAP2 in neurons in 5-HT nuclei raphé obscurus (×10) and paragigantocellularis lateralis, (×40) A. KCC2 positive neurons in the ROb; B. MAP2 positive neurons in ROb; C. Merged images. Scale bars= 500μm; D. KCC2 positive neuron in PGCL; E. MAP2 positive neuron in PGCL; F. Merged images. Scale bars = 100 μm.

Figure 13

Summary diagram of the inter-relationships of the GABAergic and 5-HT systems during early development of the human medulla. Profiles for the GABAergic system (GABA_A receptor subtypes studied, transporters) and 5HT system in the (A) raphé; (B) paragigantocellularis lateralis; (C) gigantocellularis; (D) arcuate showing the dynamic changes that occur in different nuclei across development for the two systems. The diagram highlights the complexity of the developmental changes in the GABAergic and 5-HT systems in the human medulla relative to one another.