Five clinically normal, neurologically intact, 2-year-old, purpose-bred Beagle crossbred dogs (Covance, USA) weighing 11~14 kg were used for this study. The N-TIRE ablation procedure we performed has been described elsewhere [9]. Briefly, general anesthesia was induced with intravenous (IV) diazepam (1 mg/kg; Abbott Laboratories, USA), propofol (4~6 mg/kg; Abbott Laboratories), and fentanyl (10 µg/kg/hr; Abbott Laboratories). Magnetic resonance imaging (MRI) examination of the brain was then performed with a low-field magnet (0.2 T Vet-MRI; Esaote S.p.A., Italy) using standard T1-weighted (T1W), T2-weighted (T2W), fluid attenuated inversion recovery (FLAIR), and T1W-post gadolinium sequences in three planes. A routine right lateral rostrotentorial approach to the canine skull was performed. Right parietotemporal craniectomy defects were created with a high-speed pneumatic drill (Hall Surgairtome Two; ConMed, USA), and regional durectomies were performed along the limits of the craniectomy. N-TIRE lesions were created in the right parietal lobe of the cerebrum in four dogs (Table 1) using a NanoKnife pulse generator and electrodes (AngioDynamics, USA). An additional dog underwent a sham control procedure (Table 1). The lesions were created by administering nine sets of ten electrical pulses at a repetition rate of 4 Hz. The pulses were each 50 µseconds in duration and configured with alternating polarity between each set to minimize charge build-up on the electrode surface. The sets were delivered every 3.5 seconds. The strength of the electric field applied in this configuration was calculated based on the voltage-to-distance ratio between the energized and grounded electrodes (Table 1) [9]. For the sham procedure, the electrodes were placed in the cerebral cortex for 2 min without delivering electrical pulses. The left hemisphere served as an internal (untreated) control for each dog.

Additional MRI data were obtained 1 h after N-TIRE. The dogs were then allowed to recover from anesthesia and received subcutaneous buprenorphine (0.05 mg/kg; Abbott Laboratories) every 8 h for 48 h. During the recovery period, physical and neurological examinations were performed every 6 h for 72 h, and the dogs were then humanely sacrificed by intravenous barbiturate overdose (35 mg/kg, Beuthanasia-D; Abbott Laboratories). Immediately following euthanasia, the dogs were infused with physiologic saline for 15 min to maintain a hydrostatic pressure of 90 mm Hg.

Formalin-fixed brains were cut into 5-mm sections (5 mm Porcine Matrix Slicer; Zivic Instruments, USA) in the transverse plane from the olfactory bulb through the cervicomedullary junction and embedded in paraffin. The paraffin-embedded sections were then cut into 5-µm sections, mounted on positively charged slides (Fisher Scientific, USA) and stained with hematoxylin and eosin (H&E; Fisher Scientific), luxol fast blue and Periodic Acid Schiff (LFB-PAS; Hitobiotech, USA), modified trichrome (TRIC; Fisher Scientific), Bielschowsky silver (BS; Abcam, USA), and glial fibrillary acidic protein (GFAP; Millipore, USA) according to standard procedures [10].

Brain tissues were fixed in buffered 2.5% glutaraldehyde (Fisher Scientific), post-fixed in buffered 1% osmium tetraoxide (Fisher Scientific), dehydrated in graded series of 35~100% ethanol and propylene oxide (Sigma Aldrich, USA), trimmed, and then embedded in epoxy. One-µm sections were stained with 1% toludine blue (Fisher Scientific) and 2% borate in distilled water, air dried, and examined by light microscopy (Eclipse; Nikon, Japan) to select representative fields for electron microscopy. The tissue samples were then cut into 50-nm sections (EM UC7; Lecia Microsystems, Germany), stained with uranyl acetate (Fisher Scientific) for 2 hours and lead citrate (Fisher Scientific) for 5 minutes, and examined using a transmission electron microscope (EMC-10 CA; Carl Zeiss AG, Germany).

Formalin-fixed, paraffin-embedded, transversely oriented, brain tissue samples were cut into 5-µm sections for IHC studies. The sections were stained using an automated system (Dako Autostainer; Dako, USA). Antigen retrieval was performed in 10 mM citrate buffer (pH 6.0) with two 5-min cycles of microwaving at 600 W and 2.5 GHz (Bcl-2) or by incubating for 30 min at 37℃ in a proteinase K solution (CD31; IHC-Tek, USA). Antigen retrieval for caspase-3 and caspase-9 was performed with antigen retrieval citrate buffer (citrate pH 6.0; caspase-3, EDTA pH 9.0; caspase-9, Dako 3-in-1; Dako) in a DAKO PT Link (Dako) by boiling for 10 (caspase-3) or 15 (caspase-9) min. The sections were washed in neutral PBS for 5 min following antigen retrieval. Endogenous peroxidase activity was quenched by exposing the sections to 0.3% hydrogen peroxide in methanol for 30 min at room temperature. After blocking with 10% normal goat serum (Genway Biotech, USA) for 30 min at 37℃, the sections were incubated with rabbit polyclonal antibodies against Bcl-2 (1 : 100; Abcam) caspase-3 (1 : 60; Abcam), caspase-9 (1 : 100; Abcam), and CD31/PECAM-1 (1 : 200, Bioss, USA) in 0.01 M PBS containing 1% normal goat serum and 0.3% Triton X-100 for 30 min at 20℃. The sections were then rinsed with Tris buffered saline and incubated for 40 min (CD31) or 1 h (Bcl-2, caspase-3, and caspase-9) at 37℃ with anti-mouse or biotinylated anti-rabbit IgG (1 : 200, ABC kit; Vector Laboratories, USA) in 0.01 M PBS containing 1% normal goat serum and 0.3% Triton X-100 [3]. Subsequently, the sections were rinsed with Tris buffered saline band incubated with an avidin-biotin-horseradish peroxidase complex (1 : 200, ABC kit; Vector Laboratories) for 1 h at 37℃. The slides were chromogenically developed with 3,3'-diaminobenzidine tetrahydrochloride (Dako Cytomation, USA) for 5 min at room temperature and then counterstained with Fast Red (Ventana Medical Systems, USA). Positive controls for IHC included canine tonsil (caspase-3), normal canine brain tissue (CD31), canine mammary adenocarcinoma (caspase-9), and murine hippocampus (Bcl-2). Positive control tissues were obtained from the Necropsy and Veterinary Pathology Service of the Virginia-Maryland Regional College of Veterinary Medicine, VA, USA. Negative control sections received identical treatment except that they were not exposed to the primary antibodies.

Apoptotic indices were determined to quantitate the degree of apoptosis in the histological sections stained for Bcl-2, caspase-3, and caspase-9. This was done using an equation to calculate the percentage of positively labeled cells per the total number of cells in a section:

Images of the microscopic fields of interest in the treated and homologous contralateral control cerebral cortices were captured with a charge-coupled device digital camera (DS-Fi2; Nikon). In addition, the right parietal cerebral cortices from three normal aged-matched, purpose-bred Beagle dogs (Covance) that were euthanized for reasons unrelated to this study were examined as untreated controls. The digitized images were optically segmented using density and size thresholding techniques in order to discriminate negative from positive (Image J version 1.46r, National Institute of Health, USA). This segmentation process allowed the generation of binary images from which the number of positively stained objects, as indicated by a chromogen intensity greater or equal to that of postive control tissue, in each region of interest could be determined. To calculate the labeling indices for each immunohistochemical stain, five random 200× fields were selected from each slide and at least 1,000 total cells were counted.

Differences in apoptosis LI values between N-TIRE-treated and control brain sections for each staining procedure (Bcl-2, caspase-3, and caspase-9) were compared using an unpaired Student's t-test (SAS version 9.0; SAS Institute, USA). The p values < 0.05 were considered statistically significant.

Dogs 1~4 recovered from the sham or N-TIRE procedures without complications. Results of the post-operative neurological examinations were normal within 12 h of surgery. Dog 5 (Table 1) developed signs consistent with transtentorial brain herniation (unilateral unresponsive mydriasis and decerebrate rigidity) ipsilateral to the area of N-TIRE treatment 14 h after the procedure. The animal was euthanized.

No abnormalities were detected by pre-treatment MRI in any dog (Figs. 1A and B). N-TIRE resulted in clearly marginalized, elliptical to ovoid defects of various size in the parietal lobe. The lesions were T1-iso- to hypointense (Figs. 1C and D), and predominantly T2 (Figs. 1E and F) and FLAIR hyperintense with central and peripheral regions of hypointensity. Peripheral contrast enhancement was observed following gadolinium administration (Figs. 1G and H).

Focal areas of hemorrhage and malacia in the right parietal region clearly defined the areas ablated by N-TIRE (Figs. 1I~L). There was excellent morphologic correlation between the MRI findings and pathologic features of the N-TIRE lesions (Figs. 1I~P). Data obtained from the gross brain sections indicated that the volumes of N-TIRE-affected tissue correlated positively with the applied electrical charges (Table 1, Fig. 1). In addition to the N-TIRE-induced lesion, right unilateral transtentorial herniation (Fig. 2A), marked edema of the white matter in the right cerebral hemisphere, and infarction of the right caudate nuclear area secondary to thrombosis of a penetrating artery were observed in Dog 5 (Fig. 2B). No lesions were found in the contralateral internal control hemispheres. In the sham-operated dog (Dog 1), gross examination revealed two punctate, 1-mm superficial cortical depressions consistent with the insertion points of the electrodes.

No significant lesions were noted during light microscopy evaluation of the control brain tissues (Fig. 3A). In brain sections from the sham-operated control dog, injury was limited to physical disruption of the neuropil in the electrode trajectory without any collateral damage (Fig. 3B). N-TIRE resulted in focal regions of severely disrupted gray (Fig. 3C) and white (Fig. 4) matter architecture that were sharply demarcated from the adjacent untreated brain tissue (Figs. 1 and 5A). Changes in the white matter consisted of diffuse edema (Figs. 4A and B), disrupted axonal projections, (Fig. 4B), and axonal spheroid formation (Fig. 4C).

N-TIRE gray matter lesions included a vacuolated amorphous coagulum and multifocal areas of intraparenchymal hemorrhage. The intraparenchymal hemorrhage was most pronounced in the gray matter in the electrode insertion trajectory (Fig. 3C), and the capillary beds in the peripheral junctional region between the treated and untreated areas of the brain (Fig. 5A). Remnant neurons within the treated regions were shrunken with hypereosinophilic and angular cytoplasm as well as nuclear pyknosis or chromatolysis (Fig. 5B). Multifocal aggregates of inflammatory cells including neutrophils, lymphocytes, macrophages, and gitter cells were also found within the treated regions (Figs. 5B and C). Free glial nuclei in various states of degeneration (Fig. 5B) along with neuronal satellitosis and neuronophagia were also observed.

Within the treated regions of Dogs 2-4, blood vessels with luminal diameters exceeding 30 µm (Figs. 5C, 6A and B) appeared intact. In occasional endothelial cells within the unaffected vasculature, pyknosis with cytoplasmic eosinophilia and shrinkage were found. Separation of the tunica muscularis layers of the arterioles and diapedesis was also apparent in some sections. Collagenous scaffold morphology of large vessels within the N-TIRE-affected regions was preserved (Fig. 6C) and the endothelium of intact vessels in the treated regions retained immunoreactivity to CD31 (Fig. 6D). Capillary endothelial cells demonstrated frequent pyknosis (Fig. 6F, inset) and were disrupted within regions of intraparenchymal hemorrhage (Fig. 6F). In contrast, there was complete disruption of tissue architecture and vasculature in the N-TIRE-induced lesions of Dog 5 (Fig. 2B).

Adjacent to the N-TIRE lesion interspersed between normal cerebrocortical tissue or underlying subcortical white matter was a transitional zone (TZ) characterized by vacuolization of the white and gray matter (Figs. 7A and B) and perivascular inflammatory cuffs (Fig. 7B). A rim of astrogliosis was also observed in the TZ. Severity of the glial reaction was proportional to the applied energy (Figs. 7B and C).

No significant differences in Bcl-2, caspase-3, or caspase-9 LI values were noted between the N-TIRE-induced lesions, internal controls, or untreated control samples (Fig. 8). Occasional Bcl-2 and caspase-3 immunoreactivity was found in vascular endothelial cells in all groups of dogs. Rarely, moderately intense immunostaining specific for Bcl-2 and caspase-9 was noted within random individual neurons in all layers of the cerebral cortex and hippocampus of all dogs.