Amygdala sclerosis in sudden and unexpected death in epilepsy

Epilepsy Research 37 (1999) 53 – 62 www.elsevier.com/locate/epilepsyres

Amygdala sclerosis in sudden and unexpected death in epilepsy M. Thom a,*, B. Griffin a, J.W. Sander b, F. Scaravilli a a

Department of Neuropathology, National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3 BG, UK b Department of Neurology, National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3 BG, UK Received 8 November 1998; received in revised form 13 April 1999; accepted 13 April 1999

Abstract Sclerosis of the amygdala is a not uncommon finding in patients with chronic epilepsy. The amygdala has efferent connections, via the central nuclei, to cardioregulatory centres in the medulla. Experimental studies have suggested that damage to the central nucleus may be of functional significance in patients with sudden and unexpected death in epilepsy (SUDEP) in particular with regard to their susceptibility to cardiac arrhythmias. We investigated this possibility by carrying out a quantitative immunohistochemical analysis of the patterns of neuronal loss and gliosis in three amygdala subnuclei (central, basal and lateral) in post mortem material from 15 SUDEP cases and seven normal controls. We identified significant neuronal loss in the medial division of the lateral amygdaloid nucleus in SUDEP cases but not in central or basal nuclei. These patterns of cell loss in the amygdala do not differ from previous studies in both humans and animal models of chronic epilepsy suggesting that there is not a specific pattern of amygdaloid sclerosis in SUDEP patients which could implicate a functional role for this nucleus in the mechanism of the sudden death. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Amygdala; Epilepsy; Sudden death

1. Introduction In sudden unexpected deaths occurring in epilepsy (SUDEP), by definition no anatomical or toxicological cause for the death is identified at post mortem examination (Nashef, 1997). Accidental death occurring during a seizure and * Corresponding author. Present address: Department of Neuropathology, Institute of Neurology, University College London, London WC1N 3BG, UK. Fax: +44-171-9169546. E-mail address: [email protected] (M. Thom)

deaths from status epilepticus are excluded from this group and although most SUDEP cases are unwitnessed, many are considered likely to be peri-ictal events (Shorvon, 1997; Nashef et al., 1998). The underlying mechanisms of SUDEP have been the subject of much debate; central apnoea, asphyxia and pulmonary oedema occurring during a seizure have been suggested as possible causes (Nashef et al., 1996). In addition, life threatening cardiac arrhythmias including bradyarrhythmias have been recorded during seizures (Earnest et al., 1992; Nashef et al., 1996; Reeves

0920-1211/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 9 9 ) 0 0 0 3 1 - 5

54

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

et al., 1996; Jallon, 1997; Saussu et al., 1998) and may represent another possible mechanism. The major output nuclei of the amygdala are the central nuclei (CN) with efferent projections to the hypothalamus and cardio-regulatory nuclei of the dorsal medulla. The CN receive convergent information from the adjacent lateral, basal and accessory basal nuclei of the amygdala (Pitkanen et al., 1997). Experimental evidence has shown that stimulation of the CN increases vagal output to the heart inducing bradycardia whereas lesioning this nucleus abolishes conditioned bradycardia and it has been suggested that this nucleus may be of functional significance in SUDEP (Healy and Peck, 1997). Sclerosis of the amygdala is recognised as a possible sequel of chronic epilepsy, occurring with or without adjacent hippocampal damage (Bruton, 1988; Hudson et al., 1993; Honavar and Meldrum, 1997; Wolf et al., 1997) and it has also been reproduced in animal models of epilepsy (Tuunanen et al., 1996). We aimed to investigate whether specific patterns of neuronal loss and gliosis were present within the amygdala nuclei of patients with SUDEP. We carried this out using a quantitative and immunohistochemical analysis of the three major nuclei within the amygdala. We hypothesised that specific structural alterations if present, could in theory effect seizure spread through the amygdala, and output via the CN with a potential effect on autonomic cardio-regulatory control, and thus offer a possible explanation for SUDEP.

2. Methods Neuropathological tissue from 15 patients with known epilepsy and SUDEP was examined at the National Hospital for Neurology and Neurosurgery, London, and included in the study. In all cases a general post mortem examination had not revealed an accidental, traumatic or other cause to explain the sudden death. Clinical data were retrieved from the hospital notes and routine neuropathological investigation was carried out; the data and findings are presented in Tables 1 and 2. The mean age of the patients was 34.5 years (range 14–69), 11 patients were female and four

male. All had histories of generalised seizures, six had complex partial seizures and in two patients, episodes of status epilepticus had been recorded in the past. In the neuropathological examinations, three of the SUDEP patients showed hippocampal sclerosis and six showed an end-folial pattern of sclerosis. A total of seven control cases were studied (mean age 50.7 years, range 32–68 years) from four female and three male neurologically normal patients. All the cases and controls were collected in the department between the years 1991 and 1998. The brains were formalin fixed and sections of the caudal part of both the left and right amygdaloid nuclear complex were taken; fixation times varied between cases from 1 month to 4 years. The tissue was routinely processed and serial sections of 20-mm thickness were alternatively stained with Luxol fast blue/cresyl violet for cytoarchitectural analysis and immunohistochemically for GFAP (Dako UK, 1:400) with cresyl violet counterstain, using routine Avidin-Biotin techniques. Based on nomenclature proposed by Amaral et al. (1992), the CN was identified at the level of the ‘dog leg’ of the magnocellular division of the basal nucleus, where it is partly surrounded by myelinated bundles, lying medial to the external capsule and lateral to the medial nucleus (Fig. 1). The basal nucleus, magnocellular division (Bmc), was identified by its comprising the largest and most darkly stained neurones in the amygdala complex and at its dorsal most tail it bends over the dorsal part of the lateral nucleus forming the ‘dog leg’ (Fig. 1). The lateral nucleus, the largest and most lateral nucleus of the amygdala, was identified and its medial division (Lm), lying lateral to the basal nucleus, was distinguished from the less densely packed neurones of its lateral division. The three nuclei (CN, Bmc, Lm) were identified as above on the serial LFB/N stained sections and borders outlined on adjacent GFAP/ Nissl stained sections with permanent ink. For quantitative analysis, one GFAP/Nissl section from both the left and right amygdala of each case which clearly demonstrated all three nuclei was selected. Quantification of neuronal and astrocyte volumetric densities within each nucleus was carried

Table 1 Clinical findings in patents with SUDEPa Age (years) at death and sex

Age (years) at onset of epilepsy

Seizure type

Treatment

Circumstances of death

Seizure history prior to death

1

19F

14 months

GTCS, CPS

Unwitnessed

Two seizures in previous 48 h

2

69F

49

PB, CM, NV, Lam. PHT, PB

3

17F

5

4

33F

5

GTCS

CM

19

CPS, GTCS

CM, PHT

45F

14

GTCS, TS, PS

PB, PHT, Cloz.

6 7

14F 34F

9 months 32

GS GTCS

NIA CM

8

27F

13

GTCS, CPS

9

29F

16

GTCS

PB, PHT, NV, Lam. CM, NV

10

47M

18 months

PHT, PB

11

42F

GTCS, drop attacks GTCS

12 13 14

31M 53F 38M

18 months 12 months 8

GTCS, SE CPS, SGS CPS, SGS, SE

NIA PHT, PB, CM CBZ, primidone

15

20F

17

CPS, SGS

Vagal stimulator

8

PHT, PB, CBZ

Unwitnessed; found on floor Unwitnessed; found in bed Unwitnessed; found collapsed Unwitnessed

Two seizures in previous 24 h

Unwitnessed Unwitnessed; found in bed Unwitnessed; found in bed Unwitnessed; found in bath Witnessed; ‘cardiac arrest’ Unwitnessed; found in bed Witnessed; ‘cardiac arrest’ Witnessed; ‘collapsed’ Unwitnessed; found in bed Unwitnessed; found collapsed

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

Case

Series of fits in last 48 h

a

Cloz, clonazepam; CM, carbamezapine; CPS, complex partial seizures; GTCS, generalised tonic clonic seizures; Lam, lamotrigine; NV, sodium valproate; PB, phenobarbitone; PS, simple partial seizures; SE, episodes of status epilepticus; SGS, secondary generalised seizures; TS, tonic seizures.

55

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

56

Table 2 Neuropathological findings in patients with SUDEPa Case

Brain weight at postmortem

Macroscopic neuropathological findings

Additional histological neuropathological findings

1 2 3 4 5 6

1460 1260 1360 1440 1500 1055

Old frontal lobe contusions No lesions Congested white matter No lesions Cerebellar atrophy No lesions

7

1245 g

No lesions

8

1760 g

No lesions

9 10 11 12

1160 1100 1030 1065

13 14 15

1270 g 1395 g 1390 g

No lesions Old right frontal contusion Ulegyria of frontal and occipital lobes Ulegyria in left and right MCA territory, right hippocampal atrophy Bilateral hippocampal atrophy No lesions Left hippocampal atrophy

Bilateral EFS Bilateral EFS, acute neuronal changes in CA1 Microglial nodules in amygdala Left EFS Left and right EFS Right EFS and dispersion of dentate granule cells Hamartia in hippocampus, bilateral acute neuronal cells in CA1 and subiculum Acute neuronal changes in basal ganglia and subiculum Acute neuronal changes in CA1 and subiculum Capillary telangiectasia in left amygdala Bilateral EFS, cerebellar atrophy Right HCS

a

g g g g g g

g g g g

Bilateral HCS Acute neuronal changes in CA1 Bilateral HCS

EFS, end folial sclerosis; HCS, hippocampal sclerosis; MCA, middle cerebral artery.

out using an optical dissector and a stereological three-dimensional cell counting method as described by Williams and Rakic (1988). We used a counting box of area 80 ×80 mm and depth 10 mm and counted with an oil immersion objective lens (aperture 1.3) fitted to a Zeiss research microscope. The microscope was also fitted with a digital length gauge to monitor movement of the stage in the z (focusing) axis. The bottom plane of the counting box was set at 5 mm above the bottom plane of the 20-mm tissue section and moved in a systematic fashion to cover the whole area of the nucleus within the defined boundaries. Cells which were fully inside the box or touching the right, bottom or upper planes (non-forbidden planes) were counted, including all nerve cells (large and small) identified on Nissl stain and all GFAP positive cells with astrocytic morphology.

All counts were carried out by one person (MT) and average times for counting cells in one nucleus was 120 min, and around 300 counting boxes per nucleus were analysed. Differences in mean neuronal and glial densities in each nucleus between the groups of patients studied were analysed using the independent t-test.

3. Results There was no significant difference between the mean neuronal densities (ND) in the CN or Bmc of SUDEP cases and controls (Table 3) and separate analysis of left and right nuclei gave similar results. Overall, slightly higher mean ND were observed in the CN of SUDEP cases (Fig. 3) and the degree of gliosis, as measured by astrocytic

Fig. 1. (a) Section through the caudal amygdala illustrating the location of the central (C), basal (B) and lateral nuclei (L). The position of the ‘dog leg’ of the magnocellular division of the basal nucleus is indicated with an arrowhead (Luxol fast blue/cresyl violet ×4.4). (b) Higher magnification of amygdala nucleus in a control case indicating the ‘dog leg’ of the magnocellular basal nucleus (M). Arrows indicate the boundaries of the central nucleus (Luxol fast blue/cresyl violet × 28).

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

Fig. 1.

57

58

Central nucleus

SUDEP, n= 15 Controls (normal), n=7 P-values

Basal nucleus (magnocellular division)

Lateral nucleus (medial division)

Mean ND/ mm3 (S.D.)

Mean ND/ mm3 left

Mean ND/ mm3 right

Mean AD/ mm3

Mean ND/ mm3

Mean ND/ mm3 left

Mean ND/ mm3 right

Mean AD/ mm3

Mean ND/ mm3

Mean ND/ mm3 left

Mean ND/ mm3 right

Mean AD/ mm3

16 860 (3185)

16 471 (2665)

17 198 (3635)

4194 (3374)

9361 (2460)

8530 (1909)

10 081 (2712)

2824 (2903)

11 778 (3740)

11 158 (3575)

12 655 (3522)

5030 (4034)

15 303 (5325)

15 105 (6808)

15 561 (2971)

2553 (2241)

9525 (2525)

9667 (2915)

9023 (1944)

2026 (1387)

15 911 (4372)

16 239 (3618)

15 468 (5211)

2389 (2129)

P=0.5

P= 0.6

P =0.34

P =0.84

P =0.29

P= 0.004*

P=0.005*

a AD, astrocytic density; ND, neuronal density. * Values significant at PB0.05.

P =0.3

P= 0.4

P= 0.3

P =0.166

P =0.009*

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

Table 3 Results of neuronal densities and astrocytic densities in amygdalaa

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

density (AD), was also higher although not significantly so. Significant reduction in mean ND was noted in the Lm nuclei of SUDEP cases compared to controls (Table 3). This significance was maintained when left Lm nuclei alone were compared but not for the right side. Mean ND in the Lm between SUDEP cases with or without a history of complex partial seizures (10 920 and 12 800/ mm3, respectively) were not significantly different (P =0.18) and similarly between SUDEP patients with or without hippocampal sclerosis (11 104 and 12 150/mm3, respectively), analysing the Lm ipsilateral to HCS (P=0.26). There was also a significant gliosis noted in the Lm in all SUDEP cases compared to control cases with higher AD observed (Fig. 2, Table 3). Other significant neuropathological findings included a small telangiectasia in one SUDEP case (case 10) involving the lateral and basal nuclei of the left amygdaloid complex with surrounding gliosis and haemosiderin impregnation of nerve cells. In a further case (case 3) prominent microglial nodules with neuronophagia were present in the lateral nucleus as evidence of ongoing nerve cell damage, possibly related to seizures. In five SUDEP cases eosinophilic neuronal changes were

59

identified in the hippocampus and subiculum (Table 2) as evidence of acute neuronal injury, again possibly secondary to seizures occurring in the more immediate pre-mortem period.

4. Discussion The finding of significant neuronal loss in the lateral nuclei but not in basal or central subnuclei of SUDEP cases compared to controls is evidence of a regional pattern of damage to the amygdala in this group of patients with epilepsy. The mean age of the control group was slightly older than that of the SUDEP group but, as we might expect a reduction of neuronal density with age, this is unlikely to have influenced this result. The age differences between the groups may explain the finding of slightly higher neuronal densities in the central nuclei of the SUDEP group versus controls although the increased astrocytosis also observed in the SUDEP group, which is likely to have reduced the volume of the nucleus, may also have contributed to this finding. This regional pattern of amygdala pathology is similar to observations in animal models of status epilepticus where the medial division of the lateral nucleus

Fig. 2. Gliosis within the lateral nucleus of the amygdala in SUDEP with an arrow indicating the reactive astrocytes (Cresyl violet/GFAP stain ×420).

60

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

Fig. 3. Histogram illustrating the mean neuronal densities/mm3 from both left and right amygdala subnuclei in SUDEP cases and controls. Significantly lower neuronal densities were present in the lateral nuclei of SUDEP cases (P =0.004).

showed nerve cell loss but the central and basal magnocellular nuclei appeared relatively spared (Tuunanen et al., 1996). In the present study however, we note that previous episodes of status epilepticus were documented in only two of the 15 SUDEP cases. Although sclerosis of the amygdala is well recognised to occur in patients with epilepsy (Bruton, 1988), in contrast to hippocampal sclerosis there is relatively little information available regarding the anatomical distribution of cellular loss. This may be a consequence of the complex anatomy of the amygdala and problems of orientation and identification of subnuclei particularly in partial surgical resections. However, two previous quantitative studies of amygdaloid sclerosis in lobectomies from patients with temporal lobe epilepsy have been published (Hudson et al., 1993; Wolf et al., 1997). In both, neuronal loss and

gliosis of the lateral nuclei was also documented and, although the central and basal nuclei were not analysed, these findings suggest that damage to the lateral nuclei is not a specific finding in patients with epilepsy who die suddenly and unexpectedly. Whether the degree of sclerosis and neuronal loss in the lateral nucleus is greater in SUDEP patients will be addressed in a future study by comparison with non-SUDEP epileptic controls which are acknowledged to be a more relevant control group in sudden death studies (Nilsson et al., 1999). We are not able to ascertain this by direct comparison of our data with the previous two studies referred to above as only the lateral (and not medial) divisions of the lateral nucleus were assessed previously and in addition, the different quantitative methodologies employed (two dimensional as opposed to three dimensional cell counting) may result in differences in estima-

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

tions of neuronal densities. We did, however, observe a trend of greater lateral nuclear cell loss in SUDEP patients with complex partial seizures or hippocampal sclerosis compared to those with end-folial sclerosis or no hippocampal cell loss. This is similar to the findings of Hudson et al. (1993), suggesting that more severe amygdaloid sclerosis occurs in association with gliosis of other mesial temporal structures. The lateral nucleus of the mammalian amygdala has a large afferent input from cortical areas (Pitkanen et al., 1997) and gives rise to heavy projections to most other amygdaloid nuclei. We would logically expect nerve cell loss and gliosis involving the lateral nucleus to have an effect on intra-amygdaloid circuitry and possibly on output from the central nucleus. However we have not identified a pattern of cell loss exclusive to SUDEP victims which could implicate a ‘malfunctioning’ of the amygdala to explain their sudden death. We suggest that it is more likely that observed neuronal loss from the lateral nucleus reflects a selective vulnerability of these cells to seizure-induced damage. One possible explanation for this would be the larger cortical afferent projection to this nucleus compared with other amygdala subnuclei (Pitkanen et al., 1997). The observation of neuronophagia and microglial nodules within the lateral nucleus of one SUDEP case are suggestive of a more recent pathophysiological cellular insult, possibly seizure-induced. We have also noted more significant neuronal loss in the left than right Lm nucleus of SUDEP cases; this asymmetry of cellular loss is unexplained but again is unlikely to be relevant to the sudden death. Animal studies involving stimulation of amygdala nuclei have shown similar changes in the heart rate following stimulation of either the left or right hemisphere (Healy and Peck, 1997) and additionally, in a study of the ictal bradycardia syndrome in patients with temporal lobe epilepsy, no clear lateralising predominance on EEG was demonstrated (Reeves et al., 1996). In summary, we have identified focal neuronal loss and gliosis within amygdaloid subnuclei in SUDEP patients. By comparison with information from previous human and animal studies of epilepsy, we cannot identify any specific patterns

61

of amygdaloid damage that distinguish the SUDEP group and would thus implicate a role for this nucleus in the mechanism of sudden death via autonomic deregulation. However, this study has provided further quantitative information regarding the anatomical localisation of amygdaloid sclerosis in patients with chronic epilepsy. Acknowledgements We acknowledge the help of Dr L. Nashef of the Kent and Canterbury Hospital for her help with this project. References Amaral, D.G., Price, J.L., Pitkanen, A., Carmichael, S.T., 1992. Anatomical organization of the primate amygdaloid complex. In: Aggleton, J. (Ed.), The Amygdala: Neurobiological Aspects of Emotion, Memory and Mental Dysfunction. Wiley-Liss, New York, pp. 1 – 66. Bruton, C.J., 1988. The Neuropathology of Temporal Lobe Epilepsy. Maudsley Monographs, Oxford University Press, New York. Earnest, M.P., Thomas, G.E., Eden, R.A., Hossack, K.F., 1992. The sudden unexplained death syndrome in epilepsy: demographic, clinical and post mortem features. Epilepsia 33, 310 – 316. Healy, B., Peck, J., 1997. Bradycardia induced from stimulation of the left versus the right central nucleus of the amygdala. Epilepsy Res. 28, 101 – 104. Honavar, M., Meldrum, B.S., 1997. Epilepsy. In: Graham, D.I., Lantos, P.L. (Eds.), Greenfield’s Neuropathology, 6th ed. Arnold, London, pp. 931 – 962. Hudson, L.P., Munoz, D.G., Miller, L., McLachlan, R.S., Girvin, J.P., Blume, W.T., 1993. Amygdaloid sclerosis in temporal lobe epilepsy. Ann. Neurol. 33, 622 – 631. Jallon, P., 1997. Arrhythmogenic seizures. Epilepsia 38 (Suppl. 11), 43 – 47. Nashef, L., 1997. Sudden unexpected death in epilepsy: terminology and definitions. Epilepsia 38 (Suppl. 11), 6 – 8. Nashef, L., Walker, F., Allen, P., Sander, J.W., Shorvon, S.D., Fish, D.R., 1996. Apnoea and bradycardia during epileptic seizures: relation to sudden death in epilepsy. J. Neurol. Neurosurg. Psychiatr. 60, 297 – 300. Nashef, L., Garner, S., Sander, J.W., Fish, D., Shorvon, S., 1998. Circumstances of death in sudden death in epilepsy: interviews of bereaved relatives. J. Neurol. Neurosurg. Psychiatr. 64, 349 – 352. Nilsson, L., Farahmand, B.Y., Persson, P.G., Thiblin, I., Thomson, T., 1999. Risk factors for sudden unexpected death in epilepsy: a case control study. Lancet 353, 888 – 893.

62

M. Thom et al. / Epilepsy Research 37 (1999) 53–62

Pitkanen, A., Savander, V., LeDoux, J.E., 1997. Organization of intra-amygdaloid circuitries in the rat: an emerging framework for understanding functions of the amygdala. Trends Neurosci. 20, 517–523. Reeves, A.L., Nollet, K.E., Klass, D.W., Sharbrough, F.W., So, E.L., 1996. The ictal bradycardia syndrome. Epilepsia 37, 983 – 987. Saussu, F., van-Rijckevorsel, K., de-Barsy, T., 1998. Bradycardia: an unrecognised complication of some epileptic crises. Rev. Neurol. Paris 154, 250–252. Shorvon, S., 1997. Risk factors for sudden unexpected death in epilepsy. Epilepsia 38 (Suppl. 11), S20–S22.

Tuunanen, J., Halonen, T., Pitkanen, A., 1996. Status epilepticus causes selective regional damage and loss of GABAergic neurons in rat amygdaloid complex. Eur. J. Neurosci. 8, 2711 – 2725. Williams, R.W., Rakic, P., 1988. Three dimensional counting: an accurate and direct method to estimate numbers of cells in sectioned material. J. Comp. Neurol. 278, 344 – 352. Wolf, H.K., Aliashkevich, A.F., Blumcke, I., Wiestler, O.D., Zentner, J., 1997. Neuronal loss and gliosis of the amygdaloid nucleus in temporal lobe epilepsy. Acta Neuropathol. (Berl.) 93, 606 – 610.

.