The effects of anaesthesia on cell death in a porcine model of neonatal hypoxic-ischaemic brain injury

Background Hypothermia is neuroprotective after neonatal hypoxic-ischaemic brain injury. However, systemic cooling to hypothermic temperatures is a stressor and may reduce neuroprotection in awake pigs. We compared two experiments of global hypoxic-ischaemic injury in newborn pigs, in which one group received propofol–remifentanil and the other remained awake during post-insult hypothermia treatment. Methods In both studies, newborn pigs were anaesthetised using halothane during a 45-min global hypoxic-ischaemic insult induced by reducing Fio2 and graded hypotension until a low-voltage <7 μV electroencephalogram was achieved. On reoxygenation, the pigs were randomly allocated to receive 24 h of normothermia or hypothermia. In the first study (n=18) anaesthesia was discontinued and the pigs' tracheas were extubated. In the second study (n=14) anaesthesia was continued using propofol and remifentanil. Brain injury was assessed after 72 h by classical global histopathology, Purkinje cell count, and apoptotic cell counts in the hippocampus and cerebellum. Results Global injury was nearly 10-fold greater in the awake group compared with the anaesthetised group (P=0.021). Hypothermia was neuroprotective in the anaesthetised pigs but not the awake pigs. In the hippocampus, the density of cleaved caspase-3-positive cells was increased in awake compared with anaesthetised pigs in normothermia. In the cerebellum, Purkinje cell density was reduced in the awake pigs irrespective of treatment, and the number of cleaved caspase-3-positive Purkinje cells was greatly increased in hypothermic awake pigs. We detected no difference in cleaved caspase-3 in the granular cell layer or microglial reactivity across the groups. Conclusions Our study provides novel insights into the significance of anaesthesia/sedation during hypothermia for achieving optimal neuroprotection.

Therapeutic hypothermia is neuroprotective after neonatal hypoxic-ischaemic encephalopathy, as demonstrated in various animal models 1 and clinical trials. 2 The effects of hypothermia have been thoroughly explored in large animal models of hypoxic-ischaemic encephalopathy regarding optimal temperature 3 and duration 4 of cooling, the effect of delayed onset, 5e7 concurrent inflammation 8 and stress. 9mong large animal models, the newborn pig is an excellent species for studying hypoxic-ischaemic injury and hypothermia as it has similar brain morphology to humans, including gyrencephaly, comparable brain growth spurt, 10 similar white matter distribution, 1 and comparable cardiovascular responses. 11espite the successful translation of hypothermia into clinical practice, there are significant differences in the treatment benefit from hypothermia.The number needed to treat to reduce death or major neurodevelopmental disability is estimated to be 7. 2 Concurrent drug use in the neonatal intensive care unit, including sedation and analgesia, is believed to be among the many factors thought to influence neuroprotection. 12Cooling to below normothermia is a stressor for most species, resulting in increased plasma cortisol concentration in unanaesthetised newborn pigs 9 and prolonged cortisol release in unanaesthetised fetal lambs. 13horesen and colleagues 9 reported neuroprotection to be abolished after 24 h of immediate hypothermia in unsedated pigs, whereas consistent neuroprotection was demonstrated in anaesthetised newborn pigs after hypothermia for 3 h, 14 12 h, 15 and 24 h. 16Moreover, we recently reported neuroprotection to be abolished in immature rats subjected to restraint stress and discomfort from a rectal thermometer probe. 17In preterm infants, the number of painful procedures is associated with decreased cerebellar volume 18 and cortical thickness 19 at school age.This evidence raises concern for infants with hypoxic-ischaemic encephalopathy undergoing hypothermia, likely prompting clinicians to administer increasing doses of opioids during cooling. 12he anaesthetic and sedative agents administered during hypothermia in the newborn pig may independently influence injury susceptibility and neuroprotection. 20Propofol and remifentanil, the anaesthesia cocktail commonly used in these models, are suggested to exert antiapoptotic and neuroprotective effects.21e23 Fentanyl, a different opioid, was reported to induce apoptosis in the cerebellar granular cell layer, 24 but not the cerebrum 25 in newborn pigs.Overall, the available evidence suggests anaesthetic and sedative agents to have differing effects on cell death in the newborn brain.
With this study, we aimed to conduct a comparative analysis of two historical experiments of global hypoxic-ischaemic injury in newborn pigs.Our surviving pig model integrates clinically relevant brain damage, a therapeutic standard of clinical care including post-insult seizures, multiorgan injury, and cell resolution neocortical neuropathology. 26We compare two studies in which all pigs received halothane-induced anaesthesia during the hypoxic-ischaemic insult.In one study, the anaesthesia was switched to propofoleremifentanil during the 24 h of normothermia or whole-body hypothermia 15 ; whilst in the other study, the pigs remained awake throughout the normothermia or the cooling period. 9Based on previously published findings, 9 we hypothesised that the administration of anaesthesia would enhance hypothermia neuroprotection in this model.

Experimental design
The study in which pigs were anaesthetised throughout the hypoxic-ischaemic insult and treatment was conducted in Bristol (UK) and approved by the University of Bristol Ethical Review Panel. 15The study in which pigs were awake during subsequent treatment was conducted in Oslo (Norway) and approved by the Norwegian Experimental Animal Board. 9Both projects applied the same experimental hypoxic-ischaemicneuroprotection protocol, were led by the same scientist, and used large white pigs.
Representative pigs from each study were selected based on comparable insult severity, as assessed by duration of lowamplitude electroencephalogram (EEG) <7 microvolts (mV) to match injury severity, with a minimum duration of 1200 s (20 min).Animals with premature death (survival <72 h) were excluded from the analysis.As halothane anaesthesia was administered in both studies during the hypoxic-ischaemic insult, the terms 'awake' and 'anaesthetised pigs' refer merely to anaesthesia administered during the post-insult treatment.

Animal preparation
Supplementary Figure S1 shows a flowchart of the experimental design for this study.The pigs were kept at a physiological temperature for this species (38.5e39.0C) 27 and bottlefed with pig formula ad libitum after transportation.The pigs were administered anaesthesia gas mixture in a closed box before tracheal intubation followed by umbilical vessel cannulation for fluid (dextrose 5% in NaCl 0.45% [5 ml kg À1 h À2 ]), drug administration and continuous mean arterial blood pressure (MAP) recording.Transcutaneous arterial O 2 saturation (TcSaO 2 ) was measured by a pulse oximeter.Thirty minutes before hypoxia-ischaemia, in both studies, halothane concentration was adjusted to maintain end-tidal halothane concentration at 1% and end-tidal CO 2 4.5e5.0kPa during baseline before the insult.Core temperature was recorded with a rectal thermometer at 6 cm depth and maintained at 38.5e39.5 C with an overhead heater Two-channel EEG signals were continuously recorded from subcutaneous needle electrodes over each hemisphere (interelectrode distance 3 cm).The baseline EEG was recorded for 60 min before, during, and after the hypoxic-ischaemic insult.

Acute global hypoxic-ischaemic insult
Acute hypoxia was induced by reducing fractional inspired O 2 (FiO 2 ) to 4e7% pending the EEG voltage response.The inspired halothane concentration was reduced to 0.6e0.8%pending the MAP response to achieve controlled hypotension <40 mm Hg.The FiO 2 and inspired halothane concentration were continuously titrated to maintain a low-amplitude EEG (peak-to-peak amplitude <7 mV) whilst avoiding severe hypotension (MAP <30 mm Hg) and bradycardia (<100 beats min À1 ).When hypotension or bradycardia occurred, the FiO 2 was transiently increased, halothane decreased, or both until the recovery of blood pressure and HR.Severe bradycardia that was unresponsive to increased FiO 2 or cardiac arrest were treated with 100% FiO 2 and cardiac compressions.The hypoxic-ischaemic insult was terminated after 45 min and pigs were reoxygenated with O 2 100% until the TcSaO 2 reached !95%, 9 which was the clinical practice at the time.The anaesthetised pigs were first reoxygenated in air and then with the minimal FiO 2 required to achieve !95% TcSaO 2.

Post-insult treatment
In the awake pigs, anaesthesia was stopped upon reoxygenation.Mechanical ventilation was terminated when the pigs were able to breathe independently.In the anaesthetised pigs, inhalation anaesthetics were replaced by i.v.anaesthetics (propofol 4 mg kg À1 bolus followed by propofol 4e12 mg kg À1 h À2 and remifentanil 20e80 mg kg À1 h À2 ), with the dose adjusted according to clinical need (HR, MAP, responsiveness).
Immediately after reoxygenation, the pigs were randomly allocated to treatment with normothermia or hypothermia for 24 h.During treatment, the awake pigs were kept unrestrained in a closed neonatal temperature-controlled incubator.All intravascular lines and the deep rectal temperature probe were taped along the flank and onto the back.
In the awake pigs, the rectal temperature was maintained at 39.0 C during normothermia or 35 C during hypothermia.In the anaesthetised pigs, the rectal temperature was maintained at 38.5 C during normothermia or 33.5 C during hypothermia.After 24 h of hypothermia, pigs were rewarmed to 38.5e39 C rectal temperature over a period of 6e10 h.

Histological processing and classical histopathology
After survival for 72 h, the pigs were euthanised under anaesthesia with halothane (awake pigs) or isoflurane (anaesthetised pigs).The brain was perfusion-fixed with phosphate 4%-buffered paraldehyde, paraffin-embedded, and sectioned coronally fronto-caudally into 15 blocks.Histological sections of 5 mm thickness were cut from each block and stained with haematoxylin and eosin (H&E) and immunohistochemistry (as described below).
The injury severity and distribution in the cerebral cortex, hippocampus, thalamus, basal ganglia, and cerebellum were evaluated on H&E-stained sections by pathologists EML and HP who were blinded to treatment and clinical severity.Each region was graded with a pathology score ranging from 0.0 to 4.0 27 with intervals of 0.5.A score of 0 indicates no visible injury; a score of 1 indicates the presence of individual dead neurones or patchy areas of infarcts; a score of 2 indicates partly confluent infarctions; a score of 3 indicates large, complete infarcts; a score of 4 indicates complete disintegration of the tissue. 27The assessment was based on the presence of cell death, gliosis, and injury size on H&E-stained sections.The global pathology score is the average score of all assessed regions.

Immunohistochemistry
Histological sections of 5 mm thickness from blocks 6e7 containing the hippocampus and blocks 11e12 containing the cerebellum were analysed with brightfield immunohistochemistry.Sections were deparaffinised at 60 C for 1 h, incubated for 15 min in xylene 100%, followed by graded rehydration in 5-min steps in ethanol and distilled water.Sections were washed in a solution of Tween-20-phosphate buffer 0.1% (0.1% PBS-T), followed by antigen retrieval by boiling at 96 C for 25 min in citrate buffer 10 mM (pH¼6.0).Dual enzyme block (Dako, S2003, Agilent, Santa Clara, California, USA) was applied for 10 min to block endogenous peroxidases and phosphatases.All sections were incubated for 1 h in horse serum 10% in PBS-T 0.2%, after which primary antibodies were applied.
After 48 h, the sections were washed in PBS-T0.1% and secondary antibodies were applied (horse-anti-rabbit HRP [brown] þ horse-anti-rabbit AP [magenta]) with the Immpress Duet Double Staining Polymer kit (MP-7724, Vector Laboratories, Burlingame, California, USA).DAB and AP chromogens were applied sequentially.Haematoxylin was used for counterstaining.Lastly, stepwise dehydration followed before mounting with permanent mounting medium (DPX).After the sections had dried, they were scanned using a Zeiss Axioscan Z1 slide scanner at pixel resolution 0.220 mmÂ0.220mm, then pre-processed in Zen-Lite and Fiji for image analysis as below.

Cell analysis
For the analysis of cerebellar Purkinje cells and granular cells, 8e10 representative areas (counting frame area¼1500Â1500 .Likewise, 4e5 representative hippocampal areas (counting frame area-¼600Â600 mm À2 ) were selected from CA1eCA4.Manual counting of cleaved caspase-3-positive pyramidal cells was performed.
Microglial reactivity was assessed in the cerebellar cortex, cerebral cortex, and hippocampus using morphological grading.Microglia were graded based on their somal size, ramifications, and density using a 6-point score ranging from 0.0 to 2.0, with resting microglia exhibiting a small soma, thin ramified processes, and sparse density, whereas reactive microglia were rounder with thicker and shorter processes or an amoeboid larger cell body with very few ramifications. 28,29lobal microglial reactivity was calculated by averaging over all regions assessed.

Results
A total of 18 awake (normothermia n¼11, hypothermia n¼7) and 14 anaesthetised (normothermia n¼7, hypothermia n¼7) pigs were included in this study.The median (IQR) weight was 1.68 kg (1.49e1.81kg), with no differences between the groups.The median (IQR) age was 30 h (21e36 h) in the awake study and 23 h (12e31 h) in the anaesthetised study, with no difference between the treatment groups.

Greater cell death in CA1eCA4 of the hippocampus in awake pigs
We further investigated the hippocampus as it is a region with a high susceptibility to hypoxic-ischaemic injury.We did not observe a difference in microglial reactivity between the groups (Fig. 4).Interestingly, we detected increased numbers of cleaved caspase-3-positive pyramidal cells (indicating increased cell death) in CA1eCA4 of awake normothermia pigs compared with anaesthetised normothermia pigs (P¼0.031,95% CI [5.46e135.11])but no significant difference between the hypothermia groups (P¼0.062,95% CI [À0.00008 to 167.4], Fig. 4).

Discussion
We have explored the effects of anaesthesia on brain injury, neuroinflammation, and apoptosis in two distinct studies of global hypoxic-ischaemic injury in newborn pigs randomised to treatment of normothermia or whole-body hypothermia.
Our study demonstrates increased injury in the pigs that remained awake throughout the 24-h hypothermia treatment period, with a loss of neuroprotection from hypothermia.This finding is specific to the newborn pig compared with other term-equivalent animal models of hypoxic-ischaemic encepaholpathy. 10Although the translatability of hypoxicischaemic hypothermia experiments in rodents is limited, it is noteworthy that loss of neuroprotection has not been reported in awake and unsedated postnatal day 7 rats or postnatal day 17 ferrets. 31,32This disparity may be attributed to physiological differences in heat-conserving thermoregulation, 33 experimental differences in insult protocol, or both.Based on our data, it is challenging to determine whether anaesthesia provides neuroprotection directly, through reduction of stress, or both.From previous work we know that pigs remaining awake and unsedated during hypothermia develop elevated plasma cortisol concentrations, reflective of an ongoing stress response during cooling. 9Anaesthesia mitigates this cortisol surge. 16During hypothermia, the HR is expected to decrease by 10 beats min À1 C À1 . 34In our study, the HR did not decrease during hypothermia in the awake pigs, being significantly higher than in cooled anaesthetised pigs (Table 1, Fig. 1) hence indicating a stress response.It is possible that the haemodynamic changes associated with stress, such as increased HR, contribute to the difference in injury susceptibility and neuroprotection.
There is evidence supporting direct neuroprotection from anaesthetics, such as propofol and remifentanil.Propofol has demonstrated neuroprotective effects in vivo through reduced The two studies differenced in their protocol for reoxygenation.In the awake study, pigs were reoxygenated in O 2 100% as was the clinical practice at the time.By the time of the anaesthetised study, the clinical practice had moved towards resuscitation in air.Despite this difference, the level of FiO 2 at 10 and 20 min post-insult was comparable between the groups, indicating that that time of reoxygenation (and thereby difference in hyperoxia exposure) likely is negligible.
Table 1 Physiological recordings.Physiological data of rectal temperature, cardiovascular and metabolic changes at different timepoint during hypoxia, treatment, and recovery.The awake pigs undergoing hypothermia had significantly greater heart rate throughout the cooling period, compared with pigs remaining anaesthetised during hypothermia.HT, hypothermia; MAP, mean arterial blood pressure; NT, normothermia.
Anaesthetised study Awake study KruskaleWallis NT, n¼7 HT, n¼7 NT, n¼11 HT, n¼7 Ca 2þ -induced mitochondrial swelling 35 and suppression of excessive reactive oxygen species (ROS), 36 which are important mechanisms involved in delayed cell death. 37In adult rats, propofol infusion attenuated infarct volume 22 and improved Montoya staircase test performance after acute cerebral ischaemia. 38Remifentanil was found to reduce apoptosis in the gyrus dentatus and improve latency in the step-down avoidance task after transient global cerebral ischaemia in adult gerbils, 39 and exert antiapoptotic and protective effects against ROS in postnatal day 2 mice. 23e two studies assessed here are suitable for comparative analysis because they share a similar protocol in terms of injury induction, duration, and mode of hypothermia (i.e.immediate onset, whole body, 24 h duration), post-insult management, and survival.Notably, clinical practice in neonatology with respect to reoxygenation and resuscitation changed between the earlier (awake) and later (anaesthetised) study.The awake pigs were resuscitated in 100% O 2 , which has been shown to induce excessive ROS formation, exacerbate hypoxic-ischaemic injury, 40 and abolish hypothermia

NT anaesthesia n=7
HT anaesthesia n=7 NT awake n=11 HT awake n=7 Effect of anaesthesia on cell death in hypoxic-ischaemic brain injury -7 neuroprotection. 41The anaesthetised pigs were initially resuscitated in air followed by increasing FiO 2 as clinically indicated.This difference could potentially alter ROS burden and thereby explain some of the difference in injury between the awake and anaesthetised groups.However, in our experiments the duration of reoxygenation lasted only a few minutes, and FiO 2 stabilised at comparable levels between the two studies after only 10 min (Fig. 2).Therefore, we are confident our results primarily reflect the difference arising from 24 h of anaesthesia rather than the brief difference in FiO 2 during reoxygenation.Another noteworthy difference between the studies is the discrepancy in target rectal temperature in the hypothermia groups.The awake pigs were cooled 4 C from 39.0 to 35.0 C core temperature, which yields a similar relative temperature reduction as used clinically (À3.5 C reduction from physiological temperature).The anaesthetised pigs were cooled from 38.5 to 33.5 C core temperature, which is the target temperature used clinically, however yielding a greater relative temperature reduction of À5 C. Whether it is best to induce the target temperature or the temperature reduction for translational studies remains unclear. 42However, cooling to either 35 C or 33.5 C after this global insult has previously shown similar neuroprotection and systemic effects in pigs. 42,43e were especially interested in studying the effect of anaesthesia on cell death, as our group previously demonstrated increased apoptosis in the granular cell layer after continuous fentanyl infusion in healthy newborn pigs. 24We used cleaved caspase-3 as a marker of cell death on the Effect of anaesthesia on cell death in hypoxic-ischaemic brain injury -9 apoptosisenecrosis continuum. 44,45We unfortunately did not investigate the role of caspase-independent apoptosis and necroptosis 37 specifically.We found no evidence to suggest that propofoleremifentanil induces cell death in our model.Interestingly, we observed an increase in cleaved caspase-3 cells in the awake hypothermia pigs.These findings could provide an explanation for the difference in injury and neuroprotection observed in this group.

Conclusion
Our study provides novel insights into the significance of anaesthesia/sedation during hypothermia for achieving optimal neuroprotection.While the direct transferability of anaesthesia during hypothermia to sedation in clinical practice may be limited, our findings highlight the potential impact of sedation on mitigating stress-related cell death.These findings emphasise the need for further studies on analgesia/ sedation/anaesthesia in the clinical setting to improve outcomes after hypoxic-ischaemic encephalopathy.

mm 2 )
within the cerebellar cortex were selected from the vermis and lateral cerebellar hemispheres.Separate counting of calbindin-positive cells, cleaved caspase-3 positive Purkinje cells, and cleaved caspase-3-positive granular cells were performed manually in ImageJ facilitated by the 'cell counter'plugin.The cell density (cells mm À2 ) was calculated by the following formula: density¼ ðaverage cell countÞ* b 106 ð1500*1500Þ

Fig 1 .
Fig 1. Physiological recordings.(a) Duration of low-amplitude EEG <7 mV was used to match injury severity across the two studies (anaesthesia vs awake).(b)e(d) Show recordings of rectal temperature (b) for anaesthetised (triangles) or awake (circles) pigs, heart rate for normothermia (NT), (c) and hypothermia (HT) (d).During hypothermia, heart rate was significantly greater in pigs that were awake during cooling compared with being anaesthetised.

Fig 2 .
Fig 2. Reoxygenation fractional inspired O 2 .Distribution of FiO 2 at 10 and 20 min after insult.The two studies differenced in their protocol for reoxygenation.In the awake study, pigs were reoxygenated in O 2 100% as was the clinical practice at the time.By the time of the anaesthetised study, the clinical practice had moved towards resuscitation in air.Despite this difference, the level of FiO 2 at 10 and 20 min post-insult was comparable between the groups, indicating that that time of reoxygenation (and thereby difference in hyperoxia exposure) likely is negligible.

Fig 4 .Fig 5 .
Fig 4. Hippocampus.(a) Microglial reactivity (image not shown) graded by morphological characteristics, assessed in the cornu ammoni of the hippocampus.(b) Density of pyramidal cells in CA1eCA4 positive for cleaved caspase-3.Wilcoxon ManneWhitney statistics are presented.(c) Correlation between pyramidal CA1eCA4 cleaved caspase-3 density and hippocampal pathology score.(d)e(g) Show representative selection of pyramidal cells positive for cleaved caspase-3 in each group.HT, hypothermia; NT, normothermia.