Do different anesthesia regimes affect hippocampal apoptosis and neurologic deficits in a rodent cardiac arrest model?

Of the total of 130 rats, 57 were used to test different cardiac arrest times of 6, 7, 7.5, and 8 minutes by either induction of ventricular fibrillation or asystole. In these pilot series we could not establish sustained ROSC after induction of ventricular fibrillation, so we used the potassium/esmolol-induced asystolic CA as described below. After establishing 8 minutes as the target, the remaining 73 rats were included in the study to evaluate the damage and recovery after cardiac arrest.

The study protocol consisted of induction of anesthesia, surgical preparation, adjustment of mechanical ventilation based on post-surgery blood gas analysis, induction of cardiac arrest for 8 minutes, resuscitation, and behavioral follow-up after successful resuscitation for 1, 2, or 5 days. At the end of follow up, the animals were euthanized and brain tissue was harvested for analyses from animals followed up for 2 and 5 days.

The animals were anesthetized either with sevoflurane/fentanyl (sev/fent) or medetomidine/ketamine (med/ket). Assignment to the different anesthesia groups was not randomized for logistic reasons: the vaporizer was only available at certain days. At these days, the ventilator stayed configured for the vaporizer and no experiments of the other group were conducted and vice versa. Assignment to duration of follow up was decided simultaneously with the anesthesia regime before the experiment, and an even distribution between groups was pursued. No cross-over after assignment was allowed. However, because low overall survival rates, we decided to abandon the day one follow up and to redistribute the remaining animals into the 2 and 5 days follow-up.

Animals in the sevoflurane/fentanyl group (sev/fnt) were placed in an induction chamber with 6% sevoflurane (Sevorane, Abbott Switzerland), with an additional dose of fentanyl (Fentanyl, Janssen-Cilag, Switzerland) 20 μg/kg intraperitoneally (IP). The animals in the medetomidine/ketamine (med/ket) group received IP a mixture of 0.25 mg/kg Medetomidine (Medetor, Virbac Switzerland) and 60 mg/kg ketamine (Ketalar, Pfizer Switzerland). Orotracheal intubation was performed with an angiocath 2.0 mm diameter (Venflon, BD Germany) and the position was confirmed by capnography (Datex S/5 Anaesthesia Monitor, GE, Helsinki, Finland). Animals were ventilated with a pressure-controlled, time-cycled ventilator (KTR-5, Hugo-Sachs Elektronik, March, Germany). The initial ventilator settings were 50 breaths per minutes, 30% oxygen/air mix, an inspiratory pressure of 12 cm H2O, and a PEEP of 3 cm H20. After stabilization (10 minutes), blood gases were analyzed and the inspiratory pressure of the ventilator adapted to achieve normoventilation. To spare blood, no further blood gas analysis was performed before cardiac arrest. The animals were fixed on a surgical heating pad, a thermometer was placed into the esophagus, and the animals were warmed to 36°. Subcutaneous ECG electrodes were placed on standard positions. The right femoral artery was exposed and a PE 50 catheter for blood pressure monitoring and blood sampling was inserted. Another PE 50 catheter for drug administration was inserted into the right jugular vein. All operations were done by either MH or DS. Time from anesthesia induction to start surgery took about 15 minutes in the med/ket group and 10 minutes in the sev/fnt group. The surgical procedure was about 20 minutes in length. Recordings of ECG, arterial blood pressure and temperature were performed with a standard anesthesia monitor (Datex S/5 Anaesthesia Monitor, GE, Helsinki, Finland) and the Wincollect data logger (GE, Helsinki, Finland) at a rate of 100 Hz.

The rats were paralyzed with 2 mg/kg body weight vecuronium (Norcuron, MSD, Switzerland) via the jugular vein catheter to suppress agonal breathing during the experiment. Cardiac arrest was induced with a mixture of potassium chloride and esmolol (Esmolol-Orpha, OrPha Swiss, Switzerland). The rational for adding esmolol was to lower the total amount and the concentration of the KCl solution, as done for cardioplegia in human cardiac surgery. The final solution had a concentration of 0.2 mmol/ml KCl and 9 mg/ml esmolol. Initially, a dose of 2 ml/kg body weight (BW) was given until cardiac arrest followed, indicated by a drop of the arterial blood pressure below 15 mmHg, which occurred in all animals within 5 seconds. If electrical cardiac activity in the ECG was present, immediately an additional dose of the mixture was added (up to totally 1 ml). The catheter was flushed with 0.2 ml of a 0.02 mmol/ml KCl solution. This lower concentrated KCL flush- solution was used to avoid “auto-resuscitation” which occurs when some spontaneous very weak contractions of the heart, sometimes ensuing about 1–2 minutes after cardiac arrest, increase in strength and frequency. Ventilation was disconnected, and the heating lamp turned off. After 7.5 minutes, the heating lamp was turned on, the ventilator was reconnected, a recruitment maneuver performed (sustained inspiratory pressure of 20 mmH2O), and ventilation resumed with 55 breaths per minute and 100% oxygen. At 8 minutes, manual metronome guided chest compression with 2 fingers at a rate of 220/min were started. The chest compression were made always by the same two operators, compression depth was about 25 percent of the anterior–posterior diameter of the chest with complete recoil, and targeted to achieve a mean artery pressure (together with adrenaline) of 50 – 55 mmHg. Diluted adrenaline 15 μg/kg (Adrenalin Bichsel, Switzerland) and 0.2 ml of a diluted calcium solution containing 0.1125 mmol Ca/ml (Calcium-Sandoz, Sandoz, Switzerland) were injected at the beginning of resuscitation maneuvers, separated by 0.1 ml of normal saline. If ROSC was not achieved within 60 seconds, adrenaline was repeated with 5 μg/kg every 30 seconds until ROSC. Calcium was repeated after 2 minutes of resuscitation. ROSC was defined as regular cardiac activity > 200 beats per minute with a mean arterial pressure of ≥ 60 mmHg sustaining for at least 20 seconds (one monitor length). If ROSC was not achieved within 4 minutes, no further resuscitation attempts were made.

After successful resuscitation, ventilation was increased to 70 breaths per minute, with inspiratory pressures of 18–20 mmH2O, and the rats received 50 mg/kg BW ampicillin (Clamoxyl, GlaxoSmithKline, Switzerland) for perioperative antibiotic prophylaxis and 20 μg/kg BW buprenorphine (Temgesic, Reckitt Benckiser, Switzerland) as pain relieve intramuscularly. The animals in the sev/fnt group received 0.6 to 0.8% sevoflurane once hemodynamics were stable, the concentration was targeted accordingly to avoid excessive hypertension and ventilator dyssynchronicity. Blood for arterial blood gases were drawn at ROSC +5 min and ROSC + 15 min. The catheters were withdrawn immediately afterwards, the wounds closed and the mechanical ventilation weaned within the following 15 minutes. The still comatose rats were placed on a heating mattress, without additional oxygen. After 4 hours, the rats were returned into a single cage with access to water and food ad libitum.

Animals were followed according the Animal Care and Experimentation Committee of the Canton of Bern for general health, including daily weight measurement with an electronic balance.

Neurological functions were assessed daily from day 1 to day 5 with an established wide-ranging neuroscore (NDS) [14]. The NDS evaluates general behavior and respiration (maximum, 40 points), cranial nerves function (maximum, 20 points), sensitivity to tactile stimuli (maximum, 10 points), motor function (maximum, 10 points), and coordination (maximum, 20 points).

To assess gross locomotor abilities, a modified simple vertical pole test [15] was used. The test scores the capability of the animals to move along a horizontal wooden spar of 1.5 cm width, followed by stepwise elevation of the bar to 45° to 90°, with the animal still on it. Inability to move along the horizontal spar (0°) scored 3, keeping at 90° scored 0.

To evaluate sensorimotor integration, the Tape Removal Test was used [16]. This test measures the time needed to successfully remove a standard (10×12 mm) adhesive tape attached to the front paw of the animals (maximum 120 seconds).

To determine the general activity level, the Open Field task was used [17]. Assessment took place in a square (0.75 × 0.75 m), wooden box, with the floor divided by marks in 25 squares. The animals were placed in the arena and allowed to freely move for 4 minutes while being recorded by an overhead camera. Normal behavior of rats consists of moving around and exploring the arena, including doing rears to investigate the surroundings. Accordingly, we analyzed footage by hand counting the squared crossed (with at least three paws) and rears.

In order to estimate neuronal damage consecutive to cardiac arrest and resuscitation, animals were euthanized at day 1, day 2 or day 5 with an overdose of pentobarbital intraperitoneally. Then the animal was transcardially perfused via the left ventricle with 150 ml of buffered formalin 4% in PBS (Hospital pharmacy Inselspital Bern, Switzerland). A vertical midline incision extending from the level of the cervical-cranial junction to the tip of the nose was performed with scissors. The skull edges were retracted with forceps so that the exposed brain could be removed intact. Brains were post-fixed in buffered formalin 4% in PBS for 24 hours at 4°C. The brains were then transferred to an 18% sucrose solution in PBS at 4°C until further processing.

Histological cryo-sections were obtained by cutting the brain (45 μm thick coronal sections) using a Leica CM 1850 cryostat, after snap freezing of the brain in methylbutane pre-cooled at -80°C. Injury to the hippocampus was determined by histomorphometry on 4 sections of different hippocampal regions (from rostral to caudal, therefor examining mostly the dorsal part of the cornus ammoni 1 (CA1 segment) stained with cresyl violet [18]. Histological sections were scanned at a resolution of 2700 dpi using a dia scanner (Polaroid SprintScan 35 Plus). Atrophy of the cornus ammoni 1 cell layer, resulting from neuronal loss was determined using the software ImageJ 1.45 l (National Institutes of Health, USA, http://​imagej.​nih.​gov/​ij). For this aim, the area of the CA1 subfield stained by cresyl violet was determined. To correct for differences in the length of the CA1 between sections, normalization was performed by dividing the calculated surface by the length of the structure, determined at its border proximal to the dentate gyrus, and expressed as cell layer ratio (area/length). Furthermore, the percentage of cells with apoptotic morphology (pyknotic nucleus, shrunken appearance) was estimated in the CA1 region of the same 4 sections by visual observation under a bright-field microscope, and reported as % pyknotic cells. Because the rostral/septal portion of the hippocampus is especially vulnerable, the two rostral/septal and the two caudal/temporal segments were analyzed separately. The correlation between atrophy and apoptosis was estimated with the Pearson’s correlation coefficient (see below).

All histopathological evaluations were performed by the same investigator blinded to the clinical and treatment data of the respective animal.

Cerebrospinal fluid was obtained by puncture of the cisterna magna with a 26 gauge injection needle via the suboccipital access after the pentobarbital injection was given. Between 50 and 80 μl of CSF could be harvested, immediately centrifuged and then stored at -80°C until further examination. Measurement of cytokines/chemokines concentrations in the undiluted CSF samples was performed using the suspension array technology (http://​www.​luminexcorp.​com), using a Millipore MILLIPLEX MAP Rat Cytokine/Chemokine Magnetic Bead Panel (Millipore AG, Zug, Switzerland). Quantification was based on the use of standard curves for each analyte. Concentrations of cytokines were calculated using Bio-Plex Manager Software 4.1 (Bio-Rad Laboratories) with a 5-parameter logistic curve-fitting method. The cytokines/chemokines measured were: interleukin 6 (IL-6, lower limit of detection 30.7 pg/ml); TNF-α (lower limit of detection 1.9 pg/ml); monocyte chemoattractant protein-1 (MCP-1, lower limit of detection 9.0 pg/ml) and macrophage inflammatory protein-1α (MIP-1α, lower limit of detection 0.8 pg/ml).