Clinical Study Stereotact Funct Neurosurg 2013;91:45–55 DOI: 10.1159/000341076
Received: March 14, 2012 Accepted after revision: June 11, 2012 Published online: November 29, 2012
Hypothalamic Hamartomas: Neuropathological Features with and without Prior Gamma Knife Radiosurgery John F. Kerrigan a, b Angela Parsons b Stephen G. Rice b Kristina Simeone b, f Andrew G. Shetter c Adib A. Abla c Erin Prenger d Stephen W. Coons e a
Hypothalamic Hamartoma Program, Phoenix Children’s Hospital, and b Division of Pediatric Neurology and Comprehensive Epilepsy Center, Divisions of c Neurological Surgery, d Neuroradiology and e Neuropathology, Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Ariz., and f Department of Pharmacology, Creighton University School of Medicine, Omaha, Nebr., USA
Abstract Background: The neuropathological consequences of Gamma Knife radiosurgery (GK) on hypothalamic hamartoma (HH) are unknown. Objective: In a cohort of patients undergoing surgery for treatment-resistant epilepsy, we compared surgically resected HH tissue from patients without (group I; n = 19) and with (group II; n = 10) a history of GK (median dose 16 Gy to the 50% isodose margin). Methods: Techniques included thick-section stereology for total nucleated and total neuron cell counts, and thin-section immunohistochemistry. Normal human hypothalamus derived from age-matched autopsy material was used as control tissue for CD68 immunohistochemistry. Qualitative scoring of tissue sections was performed by a neuropathologist who was blind to the GK treatment history. Results: GK is associated with decreased total cell density (p ! 0.02). A dose-dependent association of GK with decreased total neuron density approached significance (p = 0.06). Group II HH tissue had significantly more (1) reactive gliosis, (2) thickened capil-
© 2012 S. Karger AG, Basel 1011–6125/13/0911–0045$38.00/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/sfn
lary endothelium and (3) microglial activation. Degenerative features, including karyorrhexis and pyknotic nuclei, were infrequent in group II and absent in group I HH tissue. Conclusions: Nonnecrotizing doses of GK radiosurgery decrease cell density in human HH tissue. Cell loss resulting from GK may contribute to decreased excitation in the neuronal networks responsible for seizure onset in HH tissue. Copyright © 2012 S. Karger AG, Basel
Introduction
Gamma Knife radiosurgery (GK) is effective for treating several conditions responsible for epilepsy in humans, including arteriovenous malformations [1], mesial temporal lobe epilepsy [2–4] and hypothalamic hamartomas (HH) [5–7]. However, the basic cellular mechanisms by which radiosurgery exerts its therapeutic effect are poorly understood. Early treatment paradigms with GK delivered high doses of energy that resulted in cavitating lesions in the therapeutic target. High doses are
J.F.K. and A.P. contributed equally to this work.
John F. Kerrigan, MD c/o Neuroscience Publications, Division of Pediatric Neurology Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center 350 West Thomas Road, Phoenix, AZ 85013 (USA) Tel. +1 602 406 3593, E-Mail neuropub @ chw.edu
Downloaded by: Creighton University Health Sciences Libr. 198.143.32.33 - 7/8/2015 6:10:23 PM
Key Words Hypothalamic hamartoma ⴢ Gamma Knife ⴢ Radiosurgery ⴢ Epilepsy ⴢ Neuropathology
46
Stereotact Funct Neurosurg 2013;91:45–55
Materials and Methods Study Subjects Patients underwent surgical resection of HH at the Barrow Neurological Institute between April 2003 and February 2010. Inclusion criteria were treatment-resistant epilepsy associated with HH. Ten patients had previously been treated with GK, and 19 patients had no history of GK treatment. Patients were excluded from the study if they had any history of surgical intervention (biopsy, subtotal resection or stereotactic thermoablation) directed at the HH lesion. HH neuroradiological types were classified preoperatively according to the method of Delalande and Fohlen [25]. Imaging and surgical techniques have been previously described [26, 27]. Permission for enrollment into a prospective proprietary database and the use of resected HH tissue for research purposes was obtained after informed consent under protocols approved by the Institutional Review Board for Human Research at the Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Ariz., USA. Clinical features of the study cohorts are noted in table 1. Age-matched human hypothalamic tissue was used as control tissue for CD68 immunohistochemistry. This material was obtained postmortem from donors to the Brain and Tissue Bank for Developmental Disorders at the University of Maryland (Baltimore, Md., USA; see Acknowledgment). Tissue Preparation for Thick-Section Stereology Freshly resected HH tissue was placed directly into 4% paraformaldehyde for at least 24 h, and then stored at –80 ° C. Tissue was cut with a cryostat into sections with 50 m thickness to provide the appropriate depth for staining and stereological counting. Tissue slices were stored in 0.1 M phosphate buffer until staining in wells for cell counting. All 29 cases were stained with hematoxylin and eosin (HE) to quantify total nucleated cells. Neuronal nucleus immunohistochemistry (NeuN; mouse monoclonal antibody, 1:500 dilution in phosphate-buffered saline; Millipore, Billerica, Mass., USA; table 2) was used to quantify total neurons. Total neuron counts were available from only 17 cases (59% of total cohort) due to limited tissue availability.
Stereology A mean of 4.3 tissue sections (range 2–10 sections) obtained randomly from surgically resected HH tissue was counted for each stain. The number of sections varied due to differences in the amount of tissue available from surgery. A minimum of 2 sections with 100–200 dissectors were counted for each stain on each case. Before a count was begun, the microscope (Zeiss Axioskop 2 plus; Carl Zeiss Inc., Oberkochen, Germany) was aligned and calibrated to ensure consistency within the counting frames. After calibration, the counting frame and dissectors were placed randomly according to standard specifications (Bioquant Image Analysis Corp., Nashville, Tenn., USA). Each dissector was 50 ! 50 ! 20 m (dissector volume 5 ! 104 m3) with a 2.5-m guard height on each side. There were approximately 25 dissectors per tissue section due to the small size of the resected material. The software was enabled with all appropriate parameters to perform an unbiased stereology study. The results are reported as cell density, rather than absolute cell counts, because the total volume of the region of interest (including microscopically defined margins) was not possible with surgical resection within the hypothalamus.
Kerrigan /Parsons /Rice /Simeone /Shetter / Abla /Prenger /Coons
Downloaded by: Creighton University Health Sciences Libr. 198.143.32.33 - 7/8/2015 6:10:23 PM
still indicated in functional radiosurgery for movement disorders [8]. For patients with epilepsy, however, radiosurgery is effective at dose levels that usually do not result in necrosis, defined here as an absence of encephalomalacia on long-term follow-up magnetic resonance imaging. Long-term follow-up magnetic resonance imaging of successfully treated mesial temporal lobe epilepsy shows no change from baseline or mild volume loss in the targeted region [4, 9]. Studies in animal models have suggested a number of possible mechanisms of action that may contribute to seizure control after radiosurgery with dosimetry that results in subnecrotizing changes, including apoptosis [10], remodeling of neuronal microanatomy and functional connectivity [11, 12], differential changes in excitatory and inhibitory neurotransmitter levels [13], and decreased neuronal excitability [14]. Data from chronic epilepsy models in rats suggest that there may be a positive treatment response in the absence of histological changes in GK-treated tissue [14–16]. Current hypotheses regarding the mechanisms of GK on chronic epileptic tissue in humans emphasize either tissue injury [4, 9] or modulatory changes [17]. Published research on the neuropathological consequences of GK on human epileptic tissue is limited. A small number of studies have described the findings of human mesial temporal lobe tissue after GK, collectively reporting a total of 12 patients [18–22]. Necrosis, here identified by histological criteria, was identified in 9 of 12 cases (75%). Additional pathological features in this cohort included reactive gliosis [18, 22], microvascular sclerosis [18–21], perivascular inflammation [21, 22], and microglial activation [21, 22]. It is important to make the point that post-GK tissue findings are usually derived from patients who failed to respond adequately to radiosurgical treatment, as patients responding successfully to GK do not undergo subsequent surgical resection. In an earlier report on the neuropathology of HH (cohort of 57 patients undergoing surgical resection, including 3 with a history of GK), we did not observe obvious histological changes in the GK-treated tissue [23]. Likewise, in reporting a single HH case with prior GK, Akai et al. [24] did not make note of postirradiation changes. In an effort to examine the histological consequences of GK more rigorously, we compared surgically resected HH tissue from 10 patients with a history of GK therapy to that from 19 HH patients with no history of radiosurgery. Our null hypothesis is that no differences would be observed between HH tissue with and without a history of GK.
Table 1. Clinical summary of HH patients with and without a history of GK
Clinical feature
Group I: no prior radiosurgery (n = 19) Group II: prior radiosurgery (n = 10)
Mean age at surgery, years Female gender Mean lifetime duration of epilepsy, years History of gelastic seizures Seizure frequency at time of surgery (>1 seizure/day) Mean number of AEDs at time of surgery History of central precocious puberty Mental or developmental retardation (IQ/DQ
