Subcellular localization of intercellular adhesion molecule-5 (telencephalin) in the visual cortex is not developmentally regulated in the absence of matrix metalloproteinase-9

Abstract

The telencephalon-associated intercellular adhesion molecule-5 (telencephalin; ICAM-5) regulates dendritic morphology in the developing brain. In vitro studies have shown that ICAM-5 is found predominantly within dendrites and immature dendritic protrusions, with reduced expression in mushroom spines, suggesting that ICAM-5 downregulation is critical for the maturation of synaptic structures. However, developmental expression of ICAM-5 has not been explored in depth at the ultrastructural level in intact brain tissue. To investigate the ultrastructural localization of ICAM-5 with transmission electron microscopy, we performed immunoperoxidase histochemistry for ICAM-5 in mouse visual cortex at postnatal day (P)14, a period of intense synaptogenesis, and at P28, when synapses mature. We observed the expected ICAM-5 expression in dendritic protrusions and shafts at both P14 and P28. ICAM-5 expression in these dendritic protrusions decreased in prevalence with developmental age to become localized predominantly to dendritic shafts by P28. To understand better the relationship between ICAM-5 and the endopeptidase metalloproteinase-9 (MMP-9), which mediates ICAM-5 cleavage following glutamate activation during postnatal development, we also explored ICAM-5 expression in MMP-9 null animals. This analysis revealed a similar expression of ICAM-5 in dendritic elements at P14 and P28; however, an increased prevalence of ICAM-5 was noted in dendritic protrusions at P28 in the MMP-9 null animals, indicating that, in the absence of MMP-9, there is no developmental shift in ICAM-5 subcellular localization. Our ultrastructural observations shed light on possible functions mediated by ICAM-5 and their regulation by extracellular proteases.

Keywords: ICAM-5; MMP-9; dendrite; electron microscopy; plasticity; telencephalin.

FIGURE 1. Immunocytochemical analysis of anti-ICAM-5 and anti-Iba-1 antibodies in mouse brain

(A) WT cortex shows ICAM-5 immunoreactivity in telencephalic structures (including hippocampus) but not in diencephalon (thalamus). (B) Anti-ICAM-5 immunohistochemistry in ICAM-5 KO cortex shows no immunoreactivity in telencephalic or diencephalic structures. (C) Anti-Iba-1 immunoreactivity in WT cortex shows specific labeling in microglial cell bodies and processes. (D) Immunofluorescent confirmation of anti-Iba-1 reactivity (first panel) with the astrocyte-specific marker, GFAP (second panel). The third panel is a merged image showing a lack of colocalization. Cell nuclei (DAPI counterstain) are denoted in blue. (E) Immunofluorescent confirmation of anti-Iba-1 immunoreactivity with the neuronal nuclei marker, NeuN. The third panel is a merged image showing a lack of colocalization. Cell nuclei (DAPI counterstain) are denoted in blue. Scale bar= (A–B) 1 mm; (C) 100 microns; (D,E) 25 microns. Abbreviations: ctx, cortex; hpc, hippocampus; th, thalamus; cc, corpus callosum.

FIGURE 2. ICAM-5 Labeling in WT P14 and P28 mouse visual cortex

ICAM-5 labeled dendritic elements at P14 (A–B) and P28 (C–D). (E) ICAM-5 is found predominantly in dendritic protrusions during early visual development (P14) and in dendritic shafts during later development (P28). (F) At P14, the density of immunoreactive ICAM-5 dendritic protrusion is higher than at P28. *= p<0.05, ***= p<0.001; n=3 animals; 1,000 um² of neuropil analyzed for each animal. Labeling index= axon terminals (blue), dendrites (purple), dendritic protrusions (yellow), glia (green). Scale bars= 200 nm. Abbreviations: ?, unknown; d, dendrite; dp, dendritic protrusions; t, terminal; g, glia.

FIGURE 3. EM analysis of pre-embedding anti-Iba-1 and anti-ICAM-5 reactivity in P28 mouse visual cortex

(A) Low-magnification image of a characteristic microglial cell body (note condensed chromatin in the cell soma, vacuoles and other types of cytoplasmic inclusions) with anti-Iba-1 (visualized with a DAB precipitate) distributed throughout the cytoplasm. Anti-ICAM-5 (visualized with nanogold particles) is found primarily along the microglial plasma membrane. It should be noted that the surrounding elements of neuropil are devoid of non-specific immunogold staining (B) Larger magnification image of region in A showing a characteristic microglial inclusion (inc). (C–D) Magnified view of regions in B showing Iba-1 immunoreactivity in the microglial cytoplasm with ICAM-5 immunogold particles along the plasma membrane. (E) ICAM-5 immunogold reactivity in P28 mouse visual cortex. Arrows point to immunogold particles localized to dendritic membranes. Labeling index= axon terminals (blue), dendrites (purple), dendritic protrusions (yellow). Scale bars= (A) 1 mm; (B–D) 200 nm; (E) 500 nm. Abbreviations: inc= inclusion.

FIGURE 4. EM analysis of pre-embedding anti-ICAM-5 immunoperoxidase reactivity in P14 and P28 MMP-9 KO mouse visual cortex

(A–B) ICAM-5 labeled dendritic protrusions at P14 and P28. (C) The percent of ICAM-5 labeled elements in MMP-9 KO mice at P14 and P28 showed no significant differences in distribution across age. Immunoreactivity was found predominantly in dendritic protrusions at both ages. (D) A high density of immunoreactive dendritic protrusions was observed at P14 and P28. Labeling index= axon terminals (blue), dendritic protrusions (yellow), glia (green). Scale bar= 200 nm. Abbreviations: ?, unknown; d, dendrite; dp, dendritic protrusions; t, terminal; g, glia.

FIGURE 5. Localization of ICAM-5 is compromised in the absence of MMP-9

(A) ICAM-5 cleavage by MMP-9 regulates dendritic spine maturation. ICAM-5 is expressed in immature dendritic protrusions during early development (A, left side). Cleavage by MMP-9 results in the loss of ICAM-5 from putative spines, an increase in ICAM-5 in dendritic shafts and a resultant shift to a mature phenotype (A, right side). (B) In the absence of MMP-9 (KO), ICAM-5 is expressed initially in immature dendritic protrusions (B, bottom panel left side), and is retained in these structures as development proceeds (B, bottom panel, ‘nascent spine’). Interestingly, we further found ICAM-5 retention in maturing dendritic protrusions by P28 (B, bottom panel, right side). This suggests a more complex relationship between ICAM-5 localization and synaptic maturation than previously envisioned. A possibility is that maturation effects may be mediated by soluble portions of ICAM-5 feeding back onto integrins in the dendritic spine.