- Email: [email protected]
Remodeling dendritic spines in the rat pilocarpine model of temporal lobe epilepsy
Neuroscience Letters 258 (1998) 73–76
Remodeling dendritic spines in the rat pilocarpine model of temporal lobe epilepsy Masako Isokawa* Brain Research Institute, Center for Health Sciences, University of California, Los Angeles, CA 90024-1761, USA Received 27 August 1998; accepted 13 October 1998
Abstract Dendritic degeneration is a common pathology in temporal lobe epilepsy and its animal models. However, little is known when and how the degeneration occurs. In the present study of the rat pilocarpine model, visualization of dendrites of the hippocampal dentate granule cells (DGCs) by biocytin revealed a generalized spine loss immediately after the acute seizure induced by pilocarpine. However, this generalized damage was followed by recovery and plastic changes in spine shape and density, which occurred 15–35 days after the initial acute seizure, i.e., during the period of establishing a chronic phase of this model with the induction of spontaneous seizures. The present finding suggests that initial acute seizures do not cause permanent damages in dendrites and spines of DGCs; instead, dendritic spines are dynamically maintained in the course of the establishment and maintenance of spontaneous seizures. Local dendritic spine degeneration, detected later in the chronic phase of epilepsy, is likely to have a separate cause from initial acute insults. 1998 Elsevier Science Ireland Ltd. All rights reserved
Keywords: Dendrite; Spine; Seizure; Degeneration; Plasticity; Pilocarpine
It has long been known that hippocampal hyperexcitability and cell degeneration (hippocampal sclerosis) co-exist in the temporal lobe epilepsy [2]. Recent studies on human epileptic brain specimens not only provided a quantitative estimate of neuronal loss specific to neurotransmitters and peptides [6,20], but also elucidated epileptic pathologies of surviving neurons, i.e., aberrant reorganization of the mossy fiber collaterals [10] and local dendritic spine degeneration [1,8]. These pathological changes were also confirmed to be present in animal models. Neuronal losses were reported in the kainate model [14], the electrical stimulation model [16], and the pilocarpine model [13]. Aberrant mossy fiber reorganization was reported in the kainate model [3], the kindling model [22] and the pilocarpine model [13]. Local dendritic spine degeneration was reported in the alumina cream model [5], the iron model [19], the electrical stimulation model [16], the kainate model [17,18], the pilocarpine * Corresponding author. Reed Neurological Research Center, UCLA, 710 Westwood Plaza, Los Angeles, CA 90095-1769, USA. Tel.: +1 310 2061361; fax: +1 310 2068461; e-mail: [email protected]
0304-3940/98/$ - see front matter PII S0304- 3940(98) 00848- 9
model [9,11] and the kindling model [7,15] although the evidence is controversial in the last model. In the pilocarpine model, glutamate-mediated epileptic hyperexcitation was recorded from hippocampal neurons that exhibited local dendritic spine degeneration, suggesting that epilepsy-specific dendritic alteration does not necessarily accompany a decline or cessation of synaptic transmission. Although there is ample evidence to suggest that (1) acute seizures play a major role as a cause of neuronal death and (2) a loss of specific groups of neurons could be a critical step in the gradual development of aberrant axonal reorganization and spontaneous seizures, there is little literature to indicate when and how local dendritic spine degeneration occurs in ‘surviving’ neurons in chronic epilepsy. Acute seizures produce a marked elevation of extracellular glutamate [21], which may strongly activate both AMPA/KA and NMDA receptors and trigger a cascade of excitotoxic processes. Local dendritic spine degeneration, observed in chronically seizure-exposed hippocampi of human epileptic patients and the animal models, could be the pathologies resulting from initial acute insults in the ‘surviving’ neurons. The present study was conducted to
1998 Elsevier Science Ireland Ltd. All rights reserved
74
M. Isokawa / Neuroscience Letters 258 (1998) 73–76
test this hypothesis. My results provide evidence that dendritic degeneration, identified later in the chronic phase of epilepsy, is not the direct outcome of initial acute seizures. Instead, following initial acute seizures, ‘surviving’ neurons undergo substantial changes in the morphology and density of dendrites and spines in the chronic phase, during which gradual development of spontaneous seizures is established. Male Sprague–Dawley rats (100 g: n = 50) were used for (1) control (n = 10) in which rats were injected with saline and (2) pilocarpine injection (n = 40). In the experimental group, rats received a 350 mg/kg i.p. pilocarpine and experienced acute status epilepticus for 6 h (see Ref. [9] for details of methods). The pilocarpine-injected animals were divided
Fig. 1. (a)Generalized spine loss in DGCs right after the acute status epilepticus induced by pilocarpine. (b) Recovery of spines in 15–35 days after the initial status. (c) Local spine degeneration observed 12 months later in the chronic phase [9,11]. There are areas that show high spine-density (arrowheads 1 and 2), low spine density (arrowheads 3 and 4), and no spines (arrowhead 5) within a given dendritic branch. Calibration: 20 mm.
Fig. 2. Spine densities in control and pilocarpine-treated rat DGCs. Spine densities are shown as mean (no. of spines/mm) and SEM. (A) In control rat DGCs, spine density was 1.462 ± 0.269 spines/mm (±SEM). Spine density declined to 0.042 ± 0.038 spines/mm 1 day after the status (P , 0.0001) and 0.084 ± 0.008 spines/mm 3 days after the status (P , 0.0001). However, spines started to recover in 15–35 days after the status: 0.716 ± 0.078 spines/mm (15 days) and 0.838 ± 0.120 spines/mm (35 days). These densities were still low in comparison with control DGCs (P , 0.0005 for 15 days, P , 0.03 for 35 days); however, these numbers were already significantly higher than the densities observed 1–3 days after the status (P , 0.0001). (B) Density of dendritic spines that showed mushroom-type morphology without an obvious neck. In control rat DGCs, the ‘mushroom spine’ density was 0.671 ± 0.168 spines/ mm. In pilocarpine-treated rat DGCs, few ‘mushroom spines’ were identified 1–3 days after the status. However, ‘mushroom spines’ started to recover to 0.222 ± 0.026 spines/mm in 15 days, and 0.265 ± 0.077 spines/mm in 35 days. This recovery was still significantly low compared with the density in control DGCs (P , 0.0001 for 15 days, P , 0.03 for 35 days). (C) Density of dendritic spines whose morphology showed a clear spine neck. In control DGCs, ‘neck-bearing thin spine’ density was 0.781 ± 0.137 spines/mm. In pilocarpine-treated rat DGCs, few neck-bearing thin spines were observed 1–3 days after the status. However, ‘neck-bearing thin spine’ density recovered in 15–35 days: 0.523 ± 0.057 spines/mm in 15 days (P = 0.069), and 0.578 ± 0.109 spines/mm in 35 days (P = 0.252), both of which were not significantly different from control. These results suggested that pilocarpine-induced status epilepticus severely degenerated dendritic spines in DGCs. However, this degeneration was transient, and the spines did recover in the 15–35 days after the initial status epilepticus. The recovery rate depended on spine morphology. ‘Mushroom-shaped spines’ recovered slower and less completely than thin spines bearing a clear neck in the time period examined.
into four groups depending on the survival periods, i.e., 1 day (n = 10), 3 days (n = 10), 15 days (n = 10) and 35 days (n = 10). The rats who survived for 15–35 days all developed spontaneous seizures. Biocytin (5%; Sigma) was dissolved in a Tris–HCl buffered saline (0.05 M), and iontophoresed into the dentate gyrus with the use of stereotaxic coordinates by applying negative current pulses of 0.1 Hz (50% duty cycle) for 10 min. After the injection, rats were kept alive overnight to allow time for the dye to be taken up by the neurons and distributed to their dendrites. Subsequently, the animals received cardiac perfusion with 4% paraformaldehyde. The brains were removed, cut coronally into 30-mM thick sections, and processed for the biocytin visualization using avidin-biotin complex reagent (Vector). Neuroanatomical data were determined satisfactory when dendritic arbors were observed in the full length with distal dendritic tips
M. Isokawa / Neuroscience Letters 258 (1998) 73–76
Fig. 3. (a)Dendritic neck became longer after spine-recovery in the 35 days after the initial status epilepticus. Arrowheads 1 and 4 indicate the distal end of the proximal dendrite. Arrowheads 2 and 3 indicate lacuna formation in a dendritic shaft. (b) A giant mushroom spine with a gigantic head (arrowhead 1) was only identified in recovered dendrites. (c) Recovered ‘neck-bearing thin spines’ of different sizes and forms: large round spine heads (arrowheads 1 and 2), a flattened spine head with a long spine neck (arrowhead 3) and the aggregation of mushroom spines (arrowheads 4 and 5). Calibration: 20 mm.
75
reaching to the hippocampal fissure and individual spines were clearly seen. When cells satisfied these criteria, they also showed secondary and tertiary axonal branches clearly visible in addition to the discrete visualization of somata and primary axons. In each group of ten rats, 150 areas/rat (500 mm2/area) were selected from a uniform location in the dendritic arbor in the dentate gyrus. Spines were counted, dendritic lengths were measured and spine density was calculated. Spine counts were obtained ‘blind’ as to experimental conditions. Dendritic spines were severely damaged or lost in DGCs, right after the initial acute seizures (status) induced by pilocarpine. Fig. 1a shows an example of a DGC dendrite that lost most of its spines 1 day after the status epilepticus. The pattern of spine loss was generalized and was different from the pattern of spine loss that is typically observed in the chronic phase of this model (see Fig. 1c for a reference– comparison; [9,11]). The generalized spine loss after the status epilepticus was not permanent. Instead, spines recovered in 15 days (Fig. 1b). By this time, a chronic phase with spontaneous seizures had almost been established. Time course of changes in spine density is summarized in Fig. 2A. Interestingly, the recovery of spines was morphologyspecific: mushroom-shaped spines with large heads recovered slower compared with neck-bearing thin spines (Fig. 2B,C; statistical figures are discussed in the legend). The spine density of control DGCs was 1.46 spines/mm, which is similar to those reported in Golgi material from rat CA1 pyramidal cells (0.55–1.2 spines/mm) [12]. As these measurements were not corrected for three-dimensional structures [23], the actual spine density is estimated to be higher in both control and epileptic DGCs than what I report here. The recovered spines have different shape and distribution compared to those observed in control DGCs. One clear difference was the length of the proximal dendrite, which became longer after the recovery of spines in epileptic DGCs (mean ± SEM, 28.4 ± 2.85 mm; P , 0.01) compared with the control DGCs (14.5 ± 2.5 mm) (arrowheads 1 and 4 in (Fig. 3a) show the distal end of proximal dendrite for the neurons whose somata were dark stained in the center of the figure). Size and shape of individual spines were also changed as shown in (Fig. 3b,c). In addition, lacunae formation was identified in dendritic shafts (arrowheads 2 and 3 in Fig. 3a). However, lacunae formation was low in density and restricted, if at all, to local spots in a few proximal dendrites and was not a generalized change in the entire dendritic trees. It is concluded that (1) dendritic degeneration, which occurs in the pilocarpine model of temporal lobe epilepsy, is not consolidated during initial acute seizures and (2) morphology and density of dendritic spines are dynamically modified and maintained in the course of the development and maintenance of spontaneous seizures during the chronic phase. These findings are supported by others who reported that long-term application of convulsants, such as a GABAA receptor antagonist bicuculline or picrotoxin, to hippocam-
76
M. Isokawa / Neuroscience Letters 258 (1998) 73–76
pal slice cultures generated pathological changes in dendrites leading to spine loss and vacuolation [4]. However, when the culture was returned to normal medium, the neurons recovered healthy morphology and exhibited a normal complement of dendritic spines. This observation supports my current findings that dendritic spines can recover from acute damage in vivo in epileptic animals once the initial status epilepticus is terminated. As discussed in the introduction, local dendritic spine loss is prominent in the DGCs of chronically epileptic hippocampus. Thus, spine degeneration detected later in the chronic phase is likely to have a separate cause from the acute degeneration. A possible candidate may be a brief spontaneous seizure, which recurs on a steady basis in the chronic phase of the pilocarpine model [9,11]. Indeed, the neuronal time constant, which can assess a cell’s total surface area and geographic extent of dendritic branches, was reported to be significantly reduced in rats that experienced many spontaneous seizures in the chronic phase (mean ± SEM, 14.2 ± 2.1 ms; P , 0.01) compared with control rat DGCs (21.2 ± 3.7 ms) [11]. This suggests that the higher the frequency of spontaneous seizures, the more severe the local dendritic spine degeneration. Although the NMDA receptor-mediated toxic process has been implicated in this model to be a potential cause of spine alteration [9], elucidation of its mechanisms awaits future investigation. Lastly, when an epilepsy model does not involve either a prolonged initial status epilepticus or the later appearance of spontaneous seizures, there is no spine degeneration. Thus, local dendritic spine alteration, observed in the chronic phase of epilepsy, may not only be a simple correlate of brief seizures, but it could also indicate a mechanistic relationship with the mode of generating seizures, i.e., spontaneity. [1] Belichenko, P.V., Sourander, P., Malmgren, K., Nordborg, C., von Essen, C., Rydenhag, B., Lindstrom, S., Hedstrom, A., Uvebrant, P. and Dahlstrom, A., Dendritic morphology in epileptogenic cortex from TRPE patients, revealed by intracellular Lucifer Yellow microinjection and confocal laser scanning microscopy, Epilepsy Res., 18 (1994) 233–247. [2] Brown, W.J., Structural substrates of seizure foci in the human temporal lobe. In M.A. Brazier (Ed.), Epilepsy, its Phenomena in Man, Academic Press, NY, 1973, pp. 339–374. [3] Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E., Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers, Brain Res., 573 (1992) 305–310. [4] Drakew, A., Muller, M., Gahwiler, H.H., Thompson, S.M. and Frotscher, M., Spine loss in experimental epilepsy: quantitative light and electron microscopic analysis of intracellularly stained CA3 pyramidal cells in hippocampal slice cultures, Neuroscience, 70 (1995) 31–45. [5] Franceschetti, S., Bugiani, O., Panzica, F., Tagliavini, F. and Avanzini, G., Changes in excitability of CA1 pyramidal neurons in slices prepared from AlCl3-treated rabbits, Epilepsy Res., 6 (1990) 39–48.
[6] de Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer, D.D., Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy, Brain Res., 495 (1989) 387–395. [7] Geinisman, Y., Morrell, F. and de Toledo-Morrell, L., Increase in the number of axospinous synapses with segmented postsynaptic densities following hippocampal kindling, Brain Res., 569 (1992) 341–347. [8] Isokawa, M. and Levesque, M.F., Increased NMDA responses and dendritic degeneration in human epileptic hippocampal neurons in slices, Neurosci. Lett., 132 (1991) 212–216. [9] Isokawa, M. and Mello, L.E.A.M., NMDA receptor-mediated excitability in dendritically deformed dentate granule cells in pilocarpine-treated rats, Neurosci. Lett., 129 (1991) 609–613. [10] Isokawa, M., Levesque, M., Babb, T.L. and Engel, J. Jr., Single mossy fiber axonal systems of human dentate granule cells studied in hippocampal slices from patients with temporal lobe epilepsy, J. Neurosci., 13 (1993) 1511–1522. [11] Isokawa, M., Decreased time constant in hippocampal dentate granule cells in pilocarpine-treated rats with progressive seizure frequencies, Brain Res., 718 (1996) 1901–1908. [12] Lacey, D.J., Normalization of dendritic spine numbers in rat hippocampus after termination of phenylacetate injections (PKU model), Brain Res., 329 (1985) 354–355. [13] Mello, L.A.M., Cavalheiro, E.A., Tan, A.M., Kupfer, W.R., Pretorius, J.K., Babb, T.L. and Finch, D.M., Circuit mechanisms of seizures in the pilocarpine model of chronic epilepsy: cell loss and mossy fiber sprouting, Epilepsy, 34 (1993) 985–995. [14] Nadler, J.V., Perry, B.W. and Cotman, C.W., Intracentricular kainic acid preferentially destroys hippocampal pyramidal cells, Nature, 271 (1978) 676–677. [15] Nishizuka, M., Okada, R., Seki, K., Arai, Y. and Iizuka, R., Loss of dendritic synapses in the medial amygdala associated with kindling, Brain Res., 552 (1991) 351–355. [16] Olney, J.W., de Gubareff, T. and Sloviter, R.S., Epileptic brain damage in rats induced by sustained electrical stimulation of the perforant path. II. Ultrastructural analysis of acute hippocampal pathology, Brain Res. Bull., 10 (1983) 699–712. [17] Phelps, S., Mitchell, J. and Wheal, H.V., Changes to synaptic ultrastructure in field CA1 of the rat hippocampus following intracerebroventricular injection of kainic acid, Neuroscience, 40 (1991) 687–699. [18] Pyapali, G.K. and Turner, D.A., Denervation-induced dendritic alterations in CA1 pyramidal cells following kainic acid hippocampal lesions in rats, Brain Res., 652 (1994) 279–290. [19] Reid, S.A., Sypert, G.W., Boggs, W.M. and Willmore, L.J., Histopathology of the ferric-induced chronic epileptic focus in cat: a Golgi study, Exp. Neurol., 66 (1979) 205–219. [20] Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D., Calcium-binding protein (Calbindin-D28K) and parvalbumin immunoicytochemistry in the normal and epileptic human hippocampus, J. Comp. Neurol., 308 (1991) 381–396. [21] Smolders, I., van Belle, K., Ebinger, G. and Michotte, Y., Hippocampal and cerebellar extracellular amino acids during pilocarpine-induced seizures in freely moving rats, Eur. J. Pharmacol., 319 (1977) 21–29. [22] Sutula, T., Koch, J., Golarai, G., Watanabe, Y. and McNamara, J.O., NMDA receptor dependence of kindling and mossy fiber sprouting: evidence that the NMDA receptor regulates patterning of hippocampal circuits in the adult brain, J. Neurosci., 16 (1996) 7398–7406. [23] Trommald, M., Jensen, V. and Andersen, P., Analysis of dendritic spines in rat CA1 pyramidal cells intracellularly filled with a fluorescent dye, J. Comp. Neurol., 353 (1995) 260–274.
Remodeling dendritic spines in the rat pilocarpine model of temporal lobe epilepsy Masako Isokawa* Brain Research Institute, Center for Health Sciences, University of California, Los Angeles, CA 90024-1761, USA Received 27 August 1998; accepted 13 October 1998
Abstract Dendritic degeneration is a common pathology in temporal lobe epilepsy and its animal models. However, little is known when and how the degeneration occurs. In the present study of the rat pilocarpine model, visualization of dendrites of the hippocampal dentate granule cells (DGCs) by biocytin revealed a generalized spine loss immediately after the acute seizure induced by pilocarpine. However, this generalized damage was followed by recovery and plastic changes in spine shape and density, which occurred 15–35 days after the initial acute seizure, i.e., during the period of establishing a chronic phase of this model with the induction of spontaneous seizures. The present finding suggests that initial acute seizures do not cause permanent damages in dendrites and spines of DGCs; instead, dendritic spines are dynamically maintained in the course of the establishment and maintenance of spontaneous seizures. Local dendritic spine degeneration, detected later in the chronic phase of epilepsy, is likely to have a separate cause from initial acute insults. 1998 Elsevier Science Ireland Ltd. All rights reserved
Keywords: Dendrite; Spine; Seizure; Degeneration; Plasticity; Pilocarpine
It has long been known that hippocampal hyperexcitability and cell degeneration (hippocampal sclerosis) co-exist in the temporal lobe epilepsy [2]. Recent studies on human epileptic brain specimens not only provided a quantitative estimate of neuronal loss specific to neurotransmitters and peptides [6,20], but also elucidated epileptic pathologies of surviving neurons, i.e., aberrant reorganization of the mossy fiber collaterals [10] and local dendritic spine degeneration [1,8]. These pathological changes were also confirmed to be present in animal models. Neuronal losses were reported in the kainate model [14], the electrical stimulation model [16], and the pilocarpine model [13]. Aberrant mossy fiber reorganization was reported in the kainate model [3], the kindling model [22] and the pilocarpine model [13]. Local dendritic spine degeneration was reported in the alumina cream model [5], the iron model [19], the electrical stimulation model [16], the kainate model [17,18], the pilocarpine * Corresponding author. Reed Neurological Research Center, UCLA, 710 Westwood Plaza, Los Angeles, CA 90095-1769, USA. Tel.: +1 310 2061361; fax: +1 310 2068461; e-mail: [email protected]
0304-3940/98/$ - see front matter PII S0304- 3940(98) 00848- 9
model [9,11] and the kindling model [7,15] although the evidence is controversial in the last model. In the pilocarpine model, glutamate-mediated epileptic hyperexcitation was recorded from hippocampal neurons that exhibited local dendritic spine degeneration, suggesting that epilepsy-specific dendritic alteration does not necessarily accompany a decline or cessation of synaptic transmission. Although there is ample evidence to suggest that (1) acute seizures play a major role as a cause of neuronal death and (2) a loss of specific groups of neurons could be a critical step in the gradual development of aberrant axonal reorganization and spontaneous seizures, there is little literature to indicate when and how local dendritic spine degeneration occurs in ‘surviving’ neurons in chronic epilepsy. Acute seizures produce a marked elevation of extracellular glutamate [21], which may strongly activate both AMPA/KA and NMDA receptors and trigger a cascade of excitotoxic processes. Local dendritic spine degeneration, observed in chronically seizure-exposed hippocampi of human epileptic patients and the animal models, could be the pathologies resulting from initial acute insults in the ‘surviving’ neurons. The present study was conducted to
1998 Elsevier Science Ireland Ltd. All rights reserved
74
M. Isokawa / Neuroscience Letters 258 (1998) 73–76
test this hypothesis. My results provide evidence that dendritic degeneration, identified later in the chronic phase of epilepsy, is not the direct outcome of initial acute seizures. Instead, following initial acute seizures, ‘surviving’ neurons undergo substantial changes in the morphology and density of dendrites and spines in the chronic phase, during which gradual development of spontaneous seizures is established. Male Sprague–Dawley rats (100 g: n = 50) were used for (1) control (n = 10) in which rats were injected with saline and (2) pilocarpine injection (n = 40). In the experimental group, rats received a 350 mg/kg i.p. pilocarpine and experienced acute status epilepticus for 6 h (see Ref. [9] for details of methods). The pilocarpine-injected animals were divided
Fig. 1. (a)Generalized spine loss in DGCs right after the acute status epilepticus induced by pilocarpine. (b) Recovery of spines in 15–35 days after the initial status. (c) Local spine degeneration observed 12 months later in the chronic phase [9,11]. There are areas that show high spine-density (arrowheads 1 and 2), low spine density (arrowheads 3 and 4), and no spines (arrowhead 5) within a given dendritic branch. Calibration: 20 mm.
Fig. 2. Spine densities in control and pilocarpine-treated rat DGCs. Spine densities are shown as mean (no. of spines/mm) and SEM. (A) In control rat DGCs, spine density was 1.462 ± 0.269 spines/mm (±SEM). Spine density declined to 0.042 ± 0.038 spines/mm 1 day after the status (P , 0.0001) and 0.084 ± 0.008 spines/mm 3 days after the status (P , 0.0001). However, spines started to recover in 15–35 days after the status: 0.716 ± 0.078 spines/mm (15 days) and 0.838 ± 0.120 spines/mm (35 days). These densities were still low in comparison with control DGCs (P , 0.0005 for 15 days, P , 0.03 for 35 days); however, these numbers were already significantly higher than the densities observed 1–3 days after the status (P , 0.0001). (B) Density of dendritic spines that showed mushroom-type morphology without an obvious neck. In control rat DGCs, the ‘mushroom spine’ density was 0.671 ± 0.168 spines/ mm. In pilocarpine-treated rat DGCs, few ‘mushroom spines’ were identified 1–3 days after the status. However, ‘mushroom spines’ started to recover to 0.222 ± 0.026 spines/mm in 15 days, and 0.265 ± 0.077 spines/mm in 35 days. This recovery was still significantly low compared with the density in control DGCs (P , 0.0001 for 15 days, P , 0.03 for 35 days). (C) Density of dendritic spines whose morphology showed a clear spine neck. In control DGCs, ‘neck-bearing thin spine’ density was 0.781 ± 0.137 spines/mm. In pilocarpine-treated rat DGCs, few neck-bearing thin spines were observed 1–3 days after the status. However, ‘neck-bearing thin spine’ density recovered in 15–35 days: 0.523 ± 0.057 spines/mm in 15 days (P = 0.069), and 0.578 ± 0.109 spines/mm in 35 days (P = 0.252), both of which were not significantly different from control. These results suggested that pilocarpine-induced status epilepticus severely degenerated dendritic spines in DGCs. However, this degeneration was transient, and the spines did recover in the 15–35 days after the initial status epilepticus. The recovery rate depended on spine morphology. ‘Mushroom-shaped spines’ recovered slower and less completely than thin spines bearing a clear neck in the time period examined.
into four groups depending on the survival periods, i.e., 1 day (n = 10), 3 days (n = 10), 15 days (n = 10) and 35 days (n = 10). The rats who survived for 15–35 days all developed spontaneous seizures. Biocytin (5%; Sigma) was dissolved in a Tris–HCl buffered saline (0.05 M), and iontophoresed into the dentate gyrus with the use of stereotaxic coordinates by applying negative current pulses of 0.1 Hz (50% duty cycle) for 10 min. After the injection, rats were kept alive overnight to allow time for the dye to be taken up by the neurons and distributed to their dendrites. Subsequently, the animals received cardiac perfusion with 4% paraformaldehyde. The brains were removed, cut coronally into 30-mM thick sections, and processed for the biocytin visualization using avidin-biotin complex reagent (Vector). Neuroanatomical data were determined satisfactory when dendritic arbors were observed in the full length with distal dendritic tips
M. Isokawa / Neuroscience Letters 258 (1998) 73–76
Fig. 3. (a)Dendritic neck became longer after spine-recovery in the 35 days after the initial status epilepticus. Arrowheads 1 and 4 indicate the distal end of the proximal dendrite. Arrowheads 2 and 3 indicate lacuna formation in a dendritic shaft. (b) A giant mushroom spine with a gigantic head (arrowhead 1) was only identified in recovered dendrites. (c) Recovered ‘neck-bearing thin spines’ of different sizes and forms: large round spine heads (arrowheads 1 and 2), a flattened spine head with a long spine neck (arrowhead 3) and the aggregation of mushroom spines (arrowheads 4 and 5). Calibration: 20 mm.
75
reaching to the hippocampal fissure and individual spines were clearly seen. When cells satisfied these criteria, they also showed secondary and tertiary axonal branches clearly visible in addition to the discrete visualization of somata and primary axons. In each group of ten rats, 150 areas/rat (500 mm2/area) were selected from a uniform location in the dendritic arbor in the dentate gyrus. Spines were counted, dendritic lengths were measured and spine density was calculated. Spine counts were obtained ‘blind’ as to experimental conditions. Dendritic spines were severely damaged or lost in DGCs, right after the initial acute seizures (status) induced by pilocarpine. Fig. 1a shows an example of a DGC dendrite that lost most of its spines 1 day after the status epilepticus. The pattern of spine loss was generalized and was different from the pattern of spine loss that is typically observed in the chronic phase of this model (see Fig. 1c for a reference– comparison; [9,11]). The generalized spine loss after the status epilepticus was not permanent. Instead, spines recovered in 15 days (Fig. 1b). By this time, a chronic phase with spontaneous seizures had almost been established. Time course of changes in spine density is summarized in Fig. 2A. Interestingly, the recovery of spines was morphologyspecific: mushroom-shaped spines with large heads recovered slower compared with neck-bearing thin spines (Fig. 2B,C; statistical figures are discussed in the legend). The spine density of control DGCs was 1.46 spines/mm, which is similar to those reported in Golgi material from rat CA1 pyramidal cells (0.55–1.2 spines/mm) [12]. As these measurements were not corrected for three-dimensional structures [23], the actual spine density is estimated to be higher in both control and epileptic DGCs than what I report here. The recovered spines have different shape and distribution compared to those observed in control DGCs. One clear difference was the length of the proximal dendrite, which became longer after the recovery of spines in epileptic DGCs (mean ± SEM, 28.4 ± 2.85 mm; P , 0.01) compared with the control DGCs (14.5 ± 2.5 mm) (arrowheads 1 and 4 in (Fig. 3a) show the distal end of proximal dendrite for the neurons whose somata were dark stained in the center of the figure). Size and shape of individual spines were also changed as shown in (Fig. 3b,c). In addition, lacunae formation was identified in dendritic shafts (arrowheads 2 and 3 in Fig. 3a). However, lacunae formation was low in density and restricted, if at all, to local spots in a few proximal dendrites and was not a generalized change in the entire dendritic trees. It is concluded that (1) dendritic degeneration, which occurs in the pilocarpine model of temporal lobe epilepsy, is not consolidated during initial acute seizures and (2) morphology and density of dendritic spines are dynamically modified and maintained in the course of the development and maintenance of spontaneous seizures during the chronic phase. These findings are supported by others who reported that long-term application of convulsants, such as a GABAA receptor antagonist bicuculline or picrotoxin, to hippocam-
76
M. Isokawa / Neuroscience Letters 258 (1998) 73–76
pal slice cultures generated pathological changes in dendrites leading to spine loss and vacuolation [4]. However, when the culture was returned to normal medium, the neurons recovered healthy morphology and exhibited a normal complement of dendritic spines. This observation supports my current findings that dendritic spines can recover from acute damage in vivo in epileptic animals once the initial status epilepticus is terminated. As discussed in the introduction, local dendritic spine loss is prominent in the DGCs of chronically epileptic hippocampus. Thus, spine degeneration detected later in the chronic phase is likely to have a separate cause from the acute degeneration. A possible candidate may be a brief spontaneous seizure, which recurs on a steady basis in the chronic phase of the pilocarpine model [9,11]. Indeed, the neuronal time constant, which can assess a cell’s total surface area and geographic extent of dendritic branches, was reported to be significantly reduced in rats that experienced many spontaneous seizures in the chronic phase (mean ± SEM, 14.2 ± 2.1 ms; P , 0.01) compared with control rat DGCs (21.2 ± 3.7 ms) [11]. This suggests that the higher the frequency of spontaneous seizures, the more severe the local dendritic spine degeneration. Although the NMDA receptor-mediated toxic process has been implicated in this model to be a potential cause of spine alteration [9], elucidation of its mechanisms awaits future investigation. Lastly, when an epilepsy model does not involve either a prolonged initial status epilepticus or the later appearance of spontaneous seizures, there is no spine degeneration. Thus, local dendritic spine alteration, observed in the chronic phase of epilepsy, may not only be a simple correlate of brief seizures, but it could also indicate a mechanistic relationship with the mode of generating seizures, i.e., spontaneity. [1] Belichenko, P.V., Sourander, P., Malmgren, K., Nordborg, C., von Essen, C., Rydenhag, B., Lindstrom, S., Hedstrom, A., Uvebrant, P. and Dahlstrom, A., Dendritic morphology in epileptogenic cortex from TRPE patients, revealed by intracellular Lucifer Yellow microinjection and confocal laser scanning microscopy, Epilepsy Res., 18 (1994) 233–247. [2] Brown, W.J., Structural substrates of seizure foci in the human temporal lobe. In M.A. Brazier (Ed.), Epilepsy, its Phenomena in Man, Academic Press, NY, 1973, pp. 339–374. [3] Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E., Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers, Brain Res., 573 (1992) 305–310. [4] Drakew, A., Muller, M., Gahwiler, H.H., Thompson, S.M. and Frotscher, M., Spine loss in experimental epilepsy: quantitative light and electron microscopic analysis of intracellularly stained CA3 pyramidal cells in hippocampal slice cultures, Neuroscience, 70 (1995) 31–45. [5] Franceschetti, S., Bugiani, O., Panzica, F., Tagliavini, F. and Avanzini, G., Changes in excitability of CA1 pyramidal neurons in slices prepared from AlCl3-treated rabbits, Epilepsy Res., 6 (1990) 39–48.
[6] de Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer, D.D., Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy, Brain Res., 495 (1989) 387–395. [7] Geinisman, Y., Morrell, F. and de Toledo-Morrell, L., Increase in the number of axospinous synapses with segmented postsynaptic densities following hippocampal kindling, Brain Res., 569 (1992) 341–347. [8] Isokawa, M. and Levesque, M.F., Increased NMDA responses and dendritic degeneration in human epileptic hippocampal neurons in slices, Neurosci. Lett., 132 (1991) 212–216. [9] Isokawa, M. and Mello, L.E.A.M., NMDA receptor-mediated excitability in dendritically deformed dentate granule cells in pilocarpine-treated rats, Neurosci. Lett., 129 (1991) 609–613. [10] Isokawa, M., Levesque, M., Babb, T.L. and Engel, J. Jr., Single mossy fiber axonal systems of human dentate granule cells studied in hippocampal slices from patients with temporal lobe epilepsy, J. Neurosci., 13 (1993) 1511–1522. [11] Isokawa, M., Decreased time constant in hippocampal dentate granule cells in pilocarpine-treated rats with progressive seizure frequencies, Brain Res., 718 (1996) 1901–1908. [12] Lacey, D.J., Normalization of dendritic spine numbers in rat hippocampus after termination of phenylacetate injections (PKU model), Brain Res., 329 (1985) 354–355. [13] Mello, L.A.M., Cavalheiro, E.A., Tan, A.M., Kupfer, W.R., Pretorius, J.K., Babb, T.L. and Finch, D.M., Circuit mechanisms of seizures in the pilocarpine model of chronic epilepsy: cell loss and mossy fiber sprouting, Epilepsy, 34 (1993) 985–995. [14] Nadler, J.V., Perry, B.W. and Cotman, C.W., Intracentricular kainic acid preferentially destroys hippocampal pyramidal cells, Nature, 271 (1978) 676–677. [15] Nishizuka, M., Okada, R., Seki, K., Arai, Y. and Iizuka, R., Loss of dendritic synapses in the medial amygdala associated with kindling, Brain Res., 552 (1991) 351–355. [16] Olney, J.W., de Gubareff, T. and Sloviter, R.S., Epileptic brain damage in rats induced by sustained electrical stimulation of the perforant path. II. Ultrastructural analysis of acute hippocampal pathology, Brain Res. Bull., 10 (1983) 699–712. [17] Phelps, S., Mitchell, J. and Wheal, H.V., Changes to synaptic ultrastructure in field CA1 of the rat hippocampus following intracerebroventricular injection of kainic acid, Neuroscience, 40 (1991) 687–699. [18] Pyapali, G.K. and Turner, D.A., Denervation-induced dendritic alterations in CA1 pyramidal cells following kainic acid hippocampal lesions in rats, Brain Res., 652 (1994) 279–290. [19] Reid, S.A., Sypert, G.W., Boggs, W.M. and Willmore, L.J., Histopathology of the ferric-induced chronic epileptic focus in cat: a Golgi study, Exp. Neurol., 66 (1979) 205–219. [20] Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D., Calcium-binding protein (Calbindin-D28K) and parvalbumin immunoicytochemistry in the normal and epileptic human hippocampus, J. Comp. Neurol., 308 (1991) 381–396. [21] Smolders, I., van Belle, K., Ebinger, G. and Michotte, Y., Hippocampal and cerebellar extracellular amino acids during pilocarpine-induced seizures in freely moving rats, Eur. J. Pharmacol., 319 (1977) 21–29. [22] Sutula, T., Koch, J., Golarai, G., Watanabe, Y. and McNamara, J.O., NMDA receptor dependence of kindling and mossy fiber sprouting: evidence that the NMDA receptor regulates patterning of hippocampal circuits in the adult brain, J. Neurosci., 16 (1996) 7398–7406. [23] Trommald, M., Jensen, V. and Andersen, P., Analysis of dendritic spines in rat CA1 pyramidal cells intracellularly filled with a fluorescent dye, J. Comp. Neurol., 353 (1995) 260–274.