CLINICAL NEUROSCIENCE AND NEUROPATHOLOGY NEUROREPORT 0959-4965 & Lippincott Williams & Wilkins Vol 12 No 12 28 August 2001 2767 Cajal-Retzius cell density as marker of type of focal cortical dysplasia Rita Garbelli, C. Frassoni, A. Ferrario, L. Tassi, 1 M. Bramerio 2 and R. Sprea®co CA Department of Experimental Neurophysiology, Istituto Nazionale Neurologico C. Besta, Via Celoria 11, 20133 Milano; Departments of 1 Epileptic Surgery and 2 Pathology, Niguarda Hospital, Milano, Italy CA Corresponding Author Received 6 June 2001; accepted 27 June 2001 Cajal-Retzius cells, identi®ed using calretinin antiserum, were studied in layer I (LI) of adult human temporal cortex from epileptic patients with Taylor's focal cortical dysplasia and architectural dysplasia, in comparison with normal cortex. Both types of dysplasia showed LI hypercellularity, but only in architectural dysplasia was the density of Cajal-Retzius cells signi®cantly increased. A subset of Cajal-Retzius cells were reelin immunoreactive, but none were GABA positive. These ®ndings suggest that differences in the persistence of Cajal- Retzius cells, which probably re¯ect different types of altera- tion during brain development, can assist in characterizing different forms of cortical dysplasia. NeuroReport 12:2767±2771 & 2001 Lippincott Williams & Wilkins. Key words: Calretinin; Cajal-Retzius cells; GABA; Immunocytochemistry; Molecular layer; Reelin INTRODUCTION The most distinctive cellular feature of the outermost lamina of the cerebral cortex (LI) just below the pial surface, is the presence of Cajal-Retzius cells (CjR), the earliest neurons generated in human corticogenesis [1]. Because of their strategic localization in the marginal zone (MZ) above the developing cortical plate, CjR may be important for neocortex maturation [2]. This is supported by the ®nding that in developing cortex, including the human cortex [3] CjR express reelin, the extracellular matrix protein responsible for the cortical development defect in reeler mutant mice [4]. The calretinin calcium binding protein is considered a marker of early cortical cell subsets and labels CjR in developing and adult mammalian LI [4±6]. The postnatal fate and function of CjR are still debated: they may gradually degenerate during cortical development, or be- come progressively dispersed within the growing LI, or transform into non-pyramidal neurons [7]. Histochemical and immunocytochemical studies have demonstrated the existence of three morphological types of CjR in adult human LI similar to those present in developing brain [5,6,8]. Persistence of neurons, including CjR, in L1 has been reported in human focal cortical dysplasia (CD) associated with epilepsy [9]; however CD is not a homogenous pathological entity and recent data suggest the existence of several morphofunctional subgroups that presumably sub- serve diverse epileptogenic mechanisms [10]. The aim of this immunocytochemical study was to assess CjR in the temporal cortex of adult patients who had undergone surgery for intractable epilepsy associated with CD, and to compare the ®ndings with those in normal cortex. MATERIALS AND METHODS Temporal lobe specimens were obtained from seven adult patients operated on for intractable epilepsy secondary to focal CD. Three patients had Taylor's focal cortical dyspla- sia (TFCD) characterized by laminar disruption of the cortex, giant dysmorphic neurons and balloon cells [11]; the remaining four had architectural dysplasia (AD) char- acterized by disruption of the cortical lamination only. Specimens of temporal neocortex from three patients operated on for deep glial lesions and without cortical migration defects served as controls. In all cases, the surgery was performed for strictly therapeutic reasons after informed consent. Immunostaining: Specimens were ®xed in 4% buffered paraformaldheyde for 24±48 h at 48C then cut into 60 ìm serial sections. Some sections were processed using single immunoperoxidase procedures; others were processed for double immuno¯uorescence labeling and examined by confocal microscopy. The immunoperoxidase procedure was performed on free-¯oating sections as described elsewhere [10], using polyclonal anti-calretinin serum (1:5000, Swant, Bellinzona, Switzerland) or monoclonal anti-glial ®brillary acidic pro- tein (GFAP) antibody (1:500, Boehringer Mannheim, Man- nheim, Germany). Adjacent sections were counterstained with thionin (0.1%) to reveal the cytoarchitecture of the cortex, including the borders between LI and LII. For double immuno¯uorescence, sections were incu-