Temporal lobe sclerosis associated with hippocampal sclerosis in temporal lobe epilepsy: neuropathological features

Abstract

Widespread changes involving neocortical and mesial temporal lobe structures can be present in patients with temporal lobe epilepsy and hippocampal sclerosis. The incidence, pathology, and clinical significance of neocortical temporal lobe sclerosis (TLS) are not well characterized. We identified TLS in 30 of 272 surgically treated cases of hippocampal sclerosis. Temporal lobe sclerosis was defined by variable reduction of neurons from cortical layers II/III and laminar gliosis; it was typically accompanied by additional architectural abnormalities of layer II, that is, abnormal neuronal orientation and aggregation. Quantitative analysis including tessellation methods for the distribution of layer II neurons supported these observations. In 40% of cases, there was a gradient of TLS with more severe involvement toward the temporal pole, possibly signifying involvement of hippocampal projection pathways. There was a history of a febrile seizure as an initial precipitating injury in 73% of patients with TLS compared with 36% without TLS; no other clinical differences between TLS and non-TLS cases were identified. Temporal lobe sclerosis was not evident preoperatively by neuroimaging. No obvious effect of TLS on seizure outcome was noted after temporal lobe resection; 73% became seizure-free at 2-year follow-up. In conclusion, approximately 11% of surgically treated hippocampal sclerosis is accompanied by TLS. Temporal lobe sclerosis is likely an acquired process with accompanying reorganizational dysplasia and an extension of mesial temporal sclerosis rather than a separate pathological entity.

Figure 1

Quantitative measurements in evaluation of TLS and controls. (a) Sequential 5-mm coronal slices of a fixed temporal lobe specimen with the pole to the left. “A” indicates the anterior block, “P” the posterior and “I” the inferior resection margin. The region of interest used for quantitative analysis is highlighted in red on the mid-temporal gyrus cortex at 1.5 cm from the pole. (b) NeuN immunostained section of cortical layers III to V from a case with hippocampal sclerosis (HS) but without temporal lobe sclerosis (TLS) as viewed on the image analyzer. (c) The same field as shown in (b) using field fraction analysis to detect the percentage of NeuN immunostaining highlighted in green. (d) NeuN stained coronal section taken 1.5 cm from the pole and showing TLS with severe neuronal loss from the superficial cortex, visible at this magnification affecting all gyri, including the superior and inferior temporal gyri (arrows).

Figure 2

Distribution and severity of neuronal loss in temporal lobe sclerosis (TLS). (a) Gyrus from the crown to the depth in a case with hippocampal sclerosis (HS) but without TLS showing preservation of normal hexalaminar architecture. (b) A gyrus from a case with severe TLS showing a laminar band of neuronal loss in layer II and III (arrowhead) and apparent hypercellularity and crowding of neurons in the outer part of layer II at the interface with layer I which forms a dense cell band (arrow). (c) TLS case showing gyral to sulcal variation of neuron loss in layer II. In addition to the overlying cell crowding of layer II (arrow) in this region there is more apparent loss in the depth of the gyrus (arrowhead); the loss extends half way up the gyrus. (d) Mild TLS case showing crowding and clustering of cells in layer II through the gyral depth (arrow); neuronal loss is not perceivable at this magnification. (e) Control normal section from an adult male patient with no history of temporal lobe epilepsy for comparison to epilepsy cases. The temporal lobe was removed for the management of acutely increased intracranial pressure following head trauma. The laminar architecture in layers I-III is normal with no evidence of neuronal loss or dysplasia. (f) Cortical layers I to IV from a patient with HS and temporal lobe epilepsy but without evidence of TLS. (g) Severe TLS case with marked reduction in neurons from layers II and III compared to (e) and (f). (h) Case with less marked TLS compared to (g) Patchy and incomplete depletion of neurons from layers II and III is appreciated in a NeuN-stained section. (i) Same region as (g) on corresponding H&E-stained section demonstrates than neuronal loss is much less easily detectable in comparison to NeuN staining. All cases are stained for NeuN except h; Bars: a-d = 1000 μm; e-i = 100 μm.

Figure 3

Layer II dysplasia in temporal lobe sclerosis (TLS) cases. (a) A horizontal alignment, clustering and malorientation of residual neurones at the interface of layer I and layer II is appreciated in a NeuN immunostained section. (b) Glial fibrillary acidic protein (GFAP) immunostained section highlights a ‘tramline’ pattern of gliosis at low power involving mainly layers II and III. (c) GFAP immunostaining in layer I-III of a non-TLS case showing gliosis mainly in layer I and Chaslin's subpial band with delicate radial fibres extending toward the deeper cortical layers. (d) A TLS case shows a dense band of cellular and fibrillary gliosis in layers II and III in addition to superficial subpial gliosis. (e) TLS case with a region of marked neuron loss from deeper layer II extending into layer III with some clustering of remaining neurons in outer layer II at the interface with layer I as seen on NeuN immunostaining. (f) The corresponding section to (e) stained with calretinin immunohistochemistry shows a normal distribution and preservation of immunopositive interneurons. (g) A calretinin-stained section of postmortem tissue from the inferior temporal lobe gyrus of a normal adult control shows comparable distribution, morphology and number of calretinin-immunostained cells in the superficial cortical layers as in the TLS case in (f). (h) Focal cortical dysplasia (FCD) IIB case for comparison to TLS/layer II dysplasia cases demonstrating hypercellularity of layer II but with enlarged hypertrophic and dysmorphic neurons through the cortex, including layer II. Bar in a = 20 μm; in b = 1000 μm; in all other figures, 100 um.

Figure 4

Voronoi tessellation measurements. NeuN fields from cortical layer II in temporal lobes (a, c, e) with corresponding Voronoi polygon diagrams for the 3 fields (b, d, f). (a, b) is a non-temporal lobe sclerosis (TLS) case with a relatively even distribution of polygon sizes. (c, d) and (e, f) are TLS cases that demonstrate clustering on NeuN staining and correspondingly greater variability in polygon sizes. Bar = 10 μm.