Methodical approach to brain hypoxia/ischemia as a fundamental problem in forensic neuropathology

Legal Medicine 5 (2003) 190–201 www.elsevier.com/locate/legalmed

Review article

Methodical approach to brain hypoxia/ischemia as a fundamental problem in forensic neuropathologyq Manfred Oehmichen*, Christoph Meissner, Nicole von Wurmb-Schwark, Thorsten Schwark Institute of Forensic Medicine, University Hospital of Schleswig-Holstein, Campus Lu¨beck, Kahlhorststraße 31-35, D-23562 Lu¨beck, Germany Received 12 May 2003; accepted 9 June 2003

Abstract A review is given summarizing different methods that have been applied to the specific forensic neuropathological question of brain hypoxia/ischemia. On the microscopic level the authors applied routine stains and immunohistochemistry (MAP2, ALZ 50, GFAP, CD68, b-APP) for characterization of the functional activity of neurons as well as of different cell types in various brain areas. Moreover, using molecular techniques for evaluation of the mitochondrial 4977-bp deletion in correlation to hypoxia and to age brain tissue and single cell analyses are described. The demonstrated scope of methods and results give evidence of the wide spectrum of possibilities to visualize hypoxic brain injuries for determining the cause (and matter) of death and for reconstructing the time-dependent process. q 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Brain hypoxia; Ischemia; Immunohistochemistry; Molecular techniques

1. Introduction More than 50% of all forensic autopsies give evidence of brain induced functional arrest of the organ system, i.e. on brain caused death, which can be partly the result of mechanical injury, of chemical influences or of hypoxic/ischemic events. The hypoxic/ischemic event may be caused by passing q In part presented on occasion of the International Symposium on Mechanisms of Neuron Survival and Death, Shiga University of Medical Sciences, October 7, 2002. * Corresponding author. Tel.: þ451-500-2750; fax: þ451-5002760. E-mail address: [email protected] (M. Oehmichen).

cardiac or respiratory arrest, by asphyxia, by vessels’ obstruction and by traumatic or chemical mechanisms as well. In this respect it will be obvious to review the neuropathological phenomena of brain hypoxia/ ischemia. The authors will limit the review to findings of their own work group to address questions confronting every forensic neuropathologist, namely recognition of tissue and cell destruction and recovery, especially of the cause that is suspected to be induced by hypoxic/ischemic influences. The red thread will be the subject of forensic interest, which is associated by different microscopic and molecular methods. Partly the described investigations are still published, others are still in progress. The review will be

1344-6223/03/$ - see front matter q 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S1344-6223(03)00077-4

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

concentrated on the phenomenological level only, especially on neuronal and reactive alterations. For aspects of pathophysiologic principles we refer to additional reviews [1,2].

2. Irreversible intracranial circulatory arrest The permanent intracranial circulatory arrest is termed ‘brain death’. Brain death under conditions of an intact cardiac function will be termed ‘respirator brain’. The morphologic features of brain death and respirator brain include cerebral edema and absence of reactive changes. After an artificial respiration of 15 –36 h the morphologic hallmarks of the respirator brain develop: global softening of the brain, dusky discoloration of the gray matter, and often necrotic and sloughing tonsillar herniations. The diagnostic criteria of the respirator brain are summarized by Oehmichen [3]. In contrast, the diagnostic features of brain death are to be discussed. The irreversible damage of the function of all neurons of the cerebrum including cerebellum and nearly the total brain stem will be the precondition of the diagnosis. We have tried to demonstrate irreversibility of neuronal damage using different immunohistochemical markers of neuronal function:

191

microtubule-associated protein 2 (MAP2), a-tubulin, somatostatin (comp. [4 –6]). We could demonstrate a reactive expression in single cortical and subcortical neurons as an indication of a lack of irreversibility of hypoxic injury of all the neurons. We have to suppose that the hypoxic/ischemic insult is a time dependent process, which does not involve all neurons at the same time. This hypothesis is supported by the signs of reactive alterations within the perivascular neuropil of the brain parenchyma [3]: aggregations of neutrophils are seen in single brains and different brain regions as an indication of focal blood supply, which may take place in different parts of the brain while the circulation of most of the brain vessels stopped. Summarizing these findings, we have to state (Fig. 1) that morphological findings of about 75% of cases will undoubtedly give evidence of the total necrotic alteration of brain and cerebellum. However, about 25% of cases expressed cytological criteria that do not concur with clinical findings: neurons express specific proteins as an indication of a still existing functional state, and emigrated leukocytes give evidence of blood flow within the brain. These findings will be the result of a temporary (and still past) state but not the final state. Obviously, these findings do not allow the conclusion that the diagnosis

Fig. 1. Morphological diagnosis of respirator brain by macroscopic and microscopic criteria.

192

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

‘respirator brain’ can be exclusively based on cytological criteria.

3. Cytological criteria of hypoxia/ischemia Under conditions of passing hypoxia/ischemia, the irreversible neuronal changes are characterized by an eosinophilic alteration of the cytoplasm using hematoxylin and eosin stain (H and E stain) as well as a (reduction or) loss of MAP2 expression. Under normal (non-hypoxic) circumstances nearly all cortical neurons express MAP2 (Fig. 2a). Blood reflow of about 3– 5 h leads to a loss of expression. The number of cortical neurons expressing a MAP loss depends on the time of circulatory arrest. Eosinophilic cell changes and loss of MAP2 expression are especially observed after cardiac arrest at the borders of the blood supply, i.e. in the depth of the sulcus between the first and second frontal gyrus as well as in the globus pallidum, the CA1 region of the hippocampus and in the Purkinje cell layer. Moreover, we have to point to two additional pathological events regarding hypoxic changes in the CNS: 1. Under conditions of cortical hemorrhages after mechanical impact, the very early loss of neuronal expression of MAP gives evidence of additional hypoxic/ischemic processes around the contusional cortical hemorrhages (Fig. 2b). 2. Brain hypoxia has been discussed to be the cause of sudden infant death syndrome (SIDS), but this hypothesis still is not proven [7 – 9]. After application of a distinct antibody to ALZ 50 antigen [10] we could observe in about 100% an expression by neurons in the hippocampal cortex (Fig. 3a) as well as in seven hypoxia sensitive areas of the brain including the hippocampal area. We compared the SIDS group with two other groups: cerebral hypoxia or ischemia without and with perfusion (Fig. 3b). As ALZ 50 is associated with the tauantigen, we suppose that the expression gives evidence of an acute hypoxic event in SIDS—to compare with victims of cerebral hypoxia without reperfusion, though this consideration has to be proven by further investigations.

4. Axonal injury as an indication of hypoxia/ischemia Axonal injury has been known since the first observations by Strich in 1956 [11] to be one of the phenomena of degeneration of the white matter— especially—following non-missile head injury [12] and hypoxia/ischemia [2,13]. Axonal injury can be demonstrated by using an antibody against b-amyloid precursor protein (Fig. 4). A wave-like type of axonal injury (Fig. 4a) must be distinguished from an aggregated type (Fig. 4b). While the aggregated type may result from hypoxic insult the wave-like type may be mainly produced as an acceleration/deceleration response to impulsive head rotation or other type of mechanical impact. Three questions have been of interest: 1. Is axonal injury specific for mechanical forces? 2. Can it be demonstrated in brain hypoxia? 3. Is axonal injury accompanied by a microglial reaction? We could demonstrate [14,15] that axonal injury in the pons is associated with fatal subdural hemorrhages without cortical hemorrhages as a result of an acceleration/deceleration mechanism (Fig. 5a) as well as with fatal brain hypoxia without mechanical impact, but at a lesser degree (Fig. 5b). The last question (question 3) will be answered below in Section 7.

5. Hypoxic/ischemic event and edema A fundamental problem in neuropathology is the visualization of edema. It is known that hypoxic injury to the brain primarily causes a cytotoxic edema, which is characterized by a decrease in intracellular potassium. By occluding the main vessels supplying one brain hemisphere in gerbils, we could observe (and measure by densitometry) the loss of potassium within the hypoxic hemisphere [16] (Fig. 6). To get information on vasogenic edemas further animal experiments were performed. We intravenously applied Evan’s blue [17], a fluorescent marker which immediately binds to albumin in serum and functions as an albumin-marker (Fig. 7a). After disturbance of the blood –brain barrier (BBB) by

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

193

Fig. 2. MAP2 expression in the human brain. (a) Normal and intact cortex as demonstrated by neuronal expression of MAP2 (magnification: 100 £ ). (b) Contusional hemorrhage (left part of the figure) and multiple neurons negative for MAP2 expression 12 h after trauma (magnification: 500 £ ).

applying a cold lesion to the brain surface, the exsudation of albumin could be observed (Fig. 7b). The barrier disturbance was also demonstrable by using an antibody against albumin in human

formalin fixed brain. In most cases, however, the fluid was diffusely spread during the postmortem interval. Therefore, commonly no focal disturbances of BBB are demonstrable in human routine material.

194

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

Fig. 3. ALZ 50 expression in victims of SIDS. (a) ALZ 50 expression of neurons in sudden infant death syndrome demonstrated in the hippocampus (magnification: 100 £ ). (b) Semiquantitative analysis of ALZ 50 expression in victims of SIDS (upper part), in victims of hypoxia/ischemia without perfusion (middle part) and with perfusion (lower part).

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

195

Fig. 4. Different expression of types of axonal injury as demonstrated by b-APP-. (a) Mechanic induced wave-like pattern and (b) hypoxic/edematous, irregularly-shaped aggregates of injured axons (magnification, a and b: 500 £ ).

6. 4977 bp Deletion of mitochondrial DNA Molecular methods were used to investigate the occurrence of the 4977-bp deletion of mitochondrial DNA (mtDNA) in the brain. Mitochondria are thought to be one element in the aging process due to their central role in cellular energy generation. Moreover, the mitochondrial genome is particularly prone to oxidative damage from harmful free radicals continuously produced as side products of oxidative

phosphorylation. Characteristic to this type of lesion is the ‘common’ 4977-bp deletion, which has been shown to accumulate with age in postmitotic tissues [18 –20]. Total DNA was extracted from five different brain regions from brains obtained at autopsies (n ¼ 26). Fragments specific for total and 4977-bp mtDNA were amplified, detected and quantified after capillary electrophoresis in an ABI Prism 310 (Applied Biosystems). The amount of the 4977-bp deletion in

196

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

Fig. 5. Axonal injury in pons. The number of investigated cases (upper part of both the diagrams) and the percentage of cases with axonal injury (AI) expression in the pons are demonstrated (lower part of both the diagrams): In cases with (a) traumatic subdural hemorrhages and (b) fatal brain hypoxia/ischemia—each with different survival times.

the caudate nucleus increases with advancing age ([20], see Fig. 8). In the caudate nucleus no deletionspecific signal was detected in children 6 years old and younger but the signal increased in this assay to more than fivefold of the mtDNA-specific fragment in 73- and 84-year-old persons. In cerebellar tissue, however, the deletion was not detectable in 20 out of 26 persons independent of their age. These six cases, in which the deletion was observed, were characterized by clinical and morphological criteria of chronic cerebral and cerebellar hypoxia. To investigate whether certain cell characteristics underlie this mosaic-like distribution pattern of the 4977-bp deletion and to look for possible age-related changes, two cell types—GFAP expressing astrocytes and MAP expressing neurons—were visualized and isolated in the caudate nucleus from five young and five old subjects [21]. The cells were microdissected

and analyzed by single cell PCR. For each of 10 subjects, at least 30 cells of each cell type were collected and subjected to PCR individually. Screening for the presence of the common deletion (Fig. 8) yielded no significant differences in relative distribution, neither between astrocytes and neurons, nor between healthy young and old humans. These findings show that cellular susceptibility to copy errors during mtDNA replication obviously does not change as a function of age and that the mere passage of time is crucial for intracellular fixation and expansion of the deletion. Therefore, we have to state that we neither know the cellular origin of the aging process of mtDNA in the brain nor the cellular specificity of hypoxic changes of the level of mtDNA. Moreover, a deletion could be demonstrated in mitotic cells like astrocytes, i.e. not only in postmitotic cells. This work is still in progress.

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

197

Fig. 6. Potassium concentrations in human brain after different types of intoxication. (a) Relative potassium concentrations in triethyltin intoxication (loss of intracellular potassium) and alcohol intoxication (accumulation of intracellular potassium). (b) Densitometric analysis for quantification of potassium concentration in brains after different types of intoxication: In white matter (black columns) different intoxications show low potassium concentrations (left part), others high potassium concentrations (right part).

198

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

Fig. 7. Albumin as a marker of extravasation in the brain. (a) Experimental edema using Evan’s blue as a marker of albumin extravasation: fluorescence around vessels as an indication of a break-down of the blood–brain barrier (magnification: 100 £ ). (b) The expression of albumin in human brain can also demonstrate the extravasation of proteins in specific cases (magnification: 300 £ ).

7. Posthypoxic reactions The hypoxia induces a focal chemokine cascade, which results in an accumulation and proliferation

induction of local cells (astrocytes) and immigration of different blood cell types: neutrophils and macrophages. The damaged neurons are phagocytized and therefore are lost. Additionally, the reduction of nerve

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

199

Fig. 8. 4977-bp deletion in human brain. Increase of deleted mtDNA 4977 with age in caudate nucleus and cerebellum. Only in six cases the deletion is detectable in the cerebellar tissue (black columns).

cells as the reaction of local and immigrating cells are scar markers, i.e. indications of a past hypoxic/ischemic event. A double-labeling technique for simultaneous detection of macrophages and axonal injury failed to detect macrophage increase in areas of axonal injury [22]. We found an increase in activated microglia in edematous brain tissue, which was independent of the presence of axonal injury. The absence of a microglial reaction can be attributed primarily to poor chemotactic attractivity of injured axons, which prevents macrophages from recognizing the b-APP-positive axons as ‘foreign’ or ‘necrotic’ and therefore as requiring scavenging. Neuropathological studies were carried out on the brains of 162 HIV-seronegative intravenous drug addicts [23]. No (acutely) ischemically injured neurons were found in any case, but there was a loss of neurons in a quarter of the cases, mainly in the Purkinje cell layer (Fig. 9a). It is generally accepted that the hypoxic lesion of brain parenchyma—white and gray matter—also initiates a microglial reaction. This leads within 12– 24 h to an increase in the number of activated microglia/macrophages in the injured white matter as well as to neuronophagia in gray matter as demonstrated by the expression of CD68 epitope. A third phase is characterized by an increase in the number of activated astrocytes with upregulation of glial fibrillary acidic protein (GFAP).

Investigation of microglial and astrocytic reactions (Fig. 9b) in the brain of drug addicts, especially in the hippocampal area of further 46 victims revealed reactive changes in about 90% of the cases. These results indicate the existence of recurrent hypoxic injury in nearly all long-term drug addicts (victims of chronic drug abuse). In 17 cases of shaken baby syndrome, autopsy was followed—as usual—by macroscopic and microscopic examination of the brain, cervical spine, cervical cord and the eyes (unpublished data, see also [24]). In addition to acute fatal subdural hematoma, the macroscopic and microscopic investigations revealed the following phenomena: † recurrent subdural hematomas, as demonstrated by a positive Prussian blue reaction; † retinal bleeding; † epidural bleeding of the cervical spinal cord; † b-APP-positive axonal injury in the corpus callosum, pons, cervical part of the medulla indicating DAI in single cases; † microglial increase in gray and white matter indicating old injuries; and finally † activated glial cells indicative of recurrent edema. Not every case gave evidence of all these phenomena but the investigation pattern and

200

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201

Fig. 9. Morphological alterations of different cell types in cases of chronic drug addicts. (a) Neuronal loss in 162 cases of chronic drug addicts. (b) Reactive changes in the hippocampal area (n ¼ 46).

the applied methods are recommended in all cases of infant autopsies.

8. Conclusion We have described a number of special neuropathological findings and methods, which are preponderately used, in our routine laboratory for neuropathological diagnosis. The applied methods are selected especially for the demonstration of acute or repeated hypoxia/ischemia in the brain. The hypoxic changes in neurons are demonstrable within a specific time span of about 7 h. But there is a lack of methods for

demonstrating acute hypoxic changes in the brain without reflow which are to be differentiated from agonal hypoxia. We are still not able to demonstrate specific neuropathological findings in cases of acute cause of death as strangulation or drowning. We hope that the ongoing development of histological, histochemical and DNA or RNA-based methods will aid to solve these problems in the near future.

References [1] Auer RN, Sutherland GR. Hypoxia and related conditions. In: Graham DJ, Lantos PL,, editors, 7th ed. Greenfield’s neuropathology, 1. London: Arnold; 2002. p. 233 –80.

M. Oehmichen et al. / Legal Medicine 5 (2003) 190–201 [2] Oehmichen M, Meissner C. Forensic neuropathological aspects of cerebral anoxia/ischemia and hypoxia/hypoxemia. In: Oehmichen M, editor. Brain hypoxia and ischemia. Research in legal medicine, 24. Lu¨beck: Schmidt-Ro¨mhild; 2000. p. 13–25. [3] Oehmichen M. Brain death: neuropathological findings and forensic implications. Forensic Sci Int 1994;69:205 –19. [4] Yanagihara T. Immunohistochemical assessment of cerebral ischemia. In: Schurr A, Rigor BM, editors. Cerebral ischemia and resuscitation. Boca Raton: CRC Press; 1990. p. 389– 412. [5] Yanagihara T, Brengman JM, Mushynski WE. Differential vulnerability of microtubule components in cerebral ischemia. Acta Neuropathol (Berl) 1990;80:499–505. [6] Tomimoto H, Yanagihara T. Vulnerability of dendrites and nerve cell bodies in cerebral ischemia. In: Oehmichen M, editor. Research in legal medicine, 24. Lu¨beck: SchmidtRo¨mhild; 2000. p. 65 –84. [7] Oehmichen M. Recent neuropathologic research in sudden infant death syndrome. Forens Sci Progr 1990;4:128–39. [8] Oehmichen M, Linke P, Zilles K, Saternus K-S. Reactive astrocytes and macrophages in the brain stem of SIDS victims. Eleven age- and sex-matched SIDS and control cases. Clin Neuropathol 1989;8:276–83. [9] Oehmichen M, Wullen B, Zilles K, Saternus K-S. Cytologic investigations on the cerebellar cortex of sudden infant death victims. Acta Neuropathol (Berl. 1989;78:404–9. [10] Oehmichen M, Theuerkauf I, Bajanowski T, Merz H, Meissner C. Enhanced reactivity of Alz-50 antibody in brains of sudden infant death syndrome victims versus brains with lethal hypoxic/ischemic inquiry. Diagnostic significance after application of the ImmunoMax technique on routine paraffin material. Acta Neuropathol (Berl) 1998; 95:280–6. [11] Strich SJ. Diffuse degeneration of the cerebral white matter in severe dementia following head injury. J. Neurol. Neurosurg. Psychiatry. 1956;19:163 –85. [12] Oehmichen M, Meissner C, Schmidt V, Pedal I, Ko¨nig HG, Saternus K-S. Axonal injury – a diagnostic tool in forensic neuropathology? A review. Forens Sci Int 1998;95:67– 83. [13] Oehmichen M, Ko¨nig HG, editors. Neurotraumatology: Biomechanic aspects, cytologic and molecular mechanisms. Lu¨beck: Schmidt-Ro¨mhild; 1997. [14] Oehmichen M, Meissner C, Schmidt V, Pedal I, Ko¨nig HG. Axonal injury (AI) in a forensic neuropathological material. In: Oehmichen M, Ko¨nig HG, editors. Neurotraumatology:

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

201

Biomechanic aspects, cytologic and molecular mechanisms. Lu¨beck: Schmidt-Ro¨mhild; 1997. p. 203–23. Oehmichen M, Meissner C, Schmidt V, Pedal I, Ko¨nig HG. Pontine axonal injury after brain trauma and nontraumatic hypoxic-ischemic brain damage. Int J Leg Med 1999;112: 261 –7. Oehmichen M, Ochs U, Meissner C. Regional potassium distribution in the brain in forensic relevant types of intoxication preliminary morphometric evaluation using a histochemical method. Neurotoxicology 2001;22: 99– 107. Oehmichen M, Gencic M, Gru¨ninger H. Prae- und postmortale intracerebral Plasma diffusion. Lichtmikroskopische Untersuchungen am Hirnoedem. Beitr gerichtl Med 1979;37: 271 –5. Meissner C, Mohamed SA, von Wurmb N, Oehmichen M. Das mitochondriale Genom und Altern. Z Gerontol Geriatr 2001; 34:447–51. Meissner C, Mohamed SA, von Wurmb N, Oehmichen M. Aging and mitochondria. In: Oehmichen M, Ritz-Timme S, Meissner C, editors. Aging – morphological, biochemical, molecular and social aspects. Lu¨beck: Schmidt-Ro¨mhild; 2002. p. 325– 40. Meissner C, Schwark T, von Wurmb N, Oehmichen M. Cerebellar mitochondrial DNA deletion associated with survival of prolonged ischemia. In: Oehmichen M, editor. Brain hypoxia and ischemia. Research in legal medicine, 24. Lu¨beck: Schmidt-Ro¨mhild; 2000. p. 179–92. Storm T, Rath S, Mohamed SA, Bruse P, Kowald A, Oehmichen M, Meissner C. Mitotic brain cells are just as prone to mitochondrial deletions as neurons: a large-scale single-cell PCR study of the human caudate nucleus. Exp. Gerontol. 2002;37:1389–400. Oehmichen M, Theuerkauf I, Meissner C. Is traumatic injury (AI) associated with an early microglial activation? Application of a double-labeling technique for simultaneous detection of microglia and AI. Acta Neuropathol (Berl) 1999;97:491–4. Oehmichen M, Meissner C, Reiter A, Birkholz M. Neuropathology in non-human immunodeficiency virus infected drug addicts: Hypoxic brain damage after chronic intravenous drug abuse. Acta Neuropathol (Berl) 1996;)91: 642–6. Saternus K-S, Kernbach-Wighton G, Oehmichen M. The shaking trauma infants – kinetic chains. Forensic Sci Int 2000; 109:203–13.