Neuroprotective properties of xenon in different types of CNS injury

Abstract

The noble gas xenon produces neuroprotective and anaesthetic effects via different pathways including activation of background potassium channels, stimulation of the hypoxia-inducible factor 1 alpha pathway, and competition at the glycine co-activation site on the N-methyl-D-aspartate (NMDA) receptor.1Alam A. Suen K.C. Hana Z. Sanders R.D. Maze M. Ma D. Neuroprotection and neurotoxicity in the developing brain: an update on the effects of dexmedetomidine and xenon.Neurotoxicol Teratol. 2017; 60: 102-116Crossref PubMed Scopus (82) Google Scholar Overstimulation of the NMDA receptor plays a key role in neuronal death resulting from hypoxia, ischaemia, or both, but seems to be of less importance in the pathophysiological mechanisms of other types of neuronal injury.2Cerejeira J. Firmino H. Vaz-Serra A. Mukaetova-Ladinska E.B. The neuroinflammatory hypothesis of delirium.Acta Neuropathol. 2010; 119: 737-754Crossref PubMed Scopus (269) Google Scholar We systematically reviewed and quantified in vivo evidence on xenon-mediated neuroprotection in all investigated settings of neuronal injury, including trials in humans. We relied on the Stroke Therapy Academic Industry Roundtable (STAIR) recommendations to consider study methodology and to investigate whether xenon meets the rigorous standards for preclinical testing.3Lapchak P.A. Zhang J.H. Noble-Haeusslein L.J. RIGOR guidelines: escalating STAIR and STEPS for effective translational research.Transl Stroke Res. 2013; 4: 279-285Crossref PubMed Scopus (193) Google Scholar After a systematic literature search of the electronic databases PubMed, EMBASE, and the Cochrane Library, 50 articles (detailed reference list available on request) met our inclusion criteria, including 38 animal studies and 12 clinical trials. For the meta-analyses, quantitative data could be extracted from all articles, resulting in 422 effect sizes. Effect sizes were grouped by species (rodent, pig, human), administered concentration of xenon (arbitrarily chosen cut off of 50 vol% which is the mean of frequently studied xenon concentrations between 30% and 70%), window of protection (pre-, re- and post-conditioning) and settings of CNS injury (hypoxia/asphyxia, ischaemia, cardiac arrest, anaesthetic neurotoxicity, traumatic brain injury, cardiopulmonary bypass, excitotoxicity, postoperative delirium, postoperative cognitive dysfunction). Quality assessment was performed using the Cochrane Collaboration tool for human trials, and the SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE) risk tool for animal studies. The majority of studies were performed in rodents. In all species, meta-analyses showed a statistically significant neuroprotective effect of xenon (Fig 1). Neuroprotection was mainly observed in cerebral hypoxia/ischaemia, could be achieved by pre- and post-conditioning, and occurred in doses below and higher than xenon 50%. Why xenon is effective in animal models and much less so in humans remains unclear (Fig 1). As suggested by STAIR, such a failure in translation from preclinical to clinical efficacy can be attributed to either the properties of the drug itself or the preclinical models used to assess them. Small animal models are often too simplistic and frequently unable to mirror the complex interplay of events causing injury in man. Most preclinical experiments are performed in healthy young male animals free from co-morbidities and co-medications known to counteract cardioprotection.4Ludman A.J. Yellon D.M. Hausenloy D.J. Cardiac preconditioning for ischaemia: lost in translation.Dis Model Mech. 2010; 3: 35-38Crossref PubMed Scopus (90) Google Scholar,5Akhtar S. Pharmacological considerations in the elderly.Curr Opin Anaesthesiol. 2018; 31: 11-18Crossref PubMed Scopus (18) Google Scholar Note that in most preclinical experiments, neuroprotection was observed with sub-anaesthetic concentrations of xenon [because of the low potency of xenon with MAC values frequently >100% (piglets: 120%,6Liu X. Dingley J. Elstad M. Scull-Brown E. Steen P.A. Thoresen M. Minimum alveolar concentration (MAC) for sevoflurane and xenon at normothermia and hypothermia in newborn pigs.Acta Anaesthesiol Scand. 2013; 57: 646-653Crossref PubMed Scopus (14) Google Scholar pigs: 119%,7Hecker K.E. Horn N. Baumert J.H. Reyle-Hahn S.M. Heussen N. Rossaint R. Minimum alveolar concentration (MAC) of xenon in intubated swine.Br J Anaesth. 2004; 92: 421-424Abstract Full Text Full Text PDF PubMed Scopus (15) Google Scholar rats: 161%, mice: 95%8Koblin D.D. Fang Z. Eger E. et al.Minimum alveolar concentrations of noble gases, nitrogen, and sulfur hexafluoride in rats: helium and neon as nonimmobilizers (nonanesthetics].Anesth Analg. 1998; 87: 419-424PubMed Google Scholar), it is impossible to administer xenon mono-anaesthesia to most animal species unless using hyperbaric conditions)]. In most clinical trials, however, xenon was used in concentrations below or close to 1 MAC. Notably, administering 1 and 2 MAC of xenon in a hyperbaric chamber to hippocampus slice cultures from rats was associated with dose-dependent neurotoxicity, while 0.75 MAC did not increase hippocampal cell death.9Brosnan H. Bickler P.E. Xenon neurotoxicity in rat hippocampal slice cultures is similar to isoflurane and sevoflurane.Anesthesiology. 2013; 119: 335-344Crossref PubMed Scopus (42) Google Scholar Another plausible explanation for the reduced neuroprotective efficacy of xenon observed in humans might be the concomitant use of propofol in virtually all clinical trials. Propofol seems to interfere with organ protection elicited by either ischaemic or pharmacological pre-conditioning. In fact, the use of propofol has been demonstrated to abrogate the cardioprotective effects of ischaemic conditioning10Kottenberg E. Thielmann M. Bergmann L. et al.Protection by remote ischemic preconditioning during coronary artery bypass graft surgery with isoflurane but not propofol - a clinical trial.Acta Anaesthesiol Scand. 2012; 56: 30-38Crossref PubMed Scopus (300) Google Scholar and to aggravate sevoflurane-induced neuroapoptosis in neonatal mouse brain.11Tagawa T. Sakuraba S. Kimura K. Mizoguchi A. Sevoflurane in combination with propofol, not thiopental, induces a more robust neuroapoptosis than sevoflurane alone in the neonatal mouse brain.J Anesth. 2014; 28: 815-820Crossref PubMed Scopus (37) Google Scholar Propofol has been postulated to be the common denominator in two phase III trials that failed to demonstrate efficacy for remote ischaemic pre-conditioning.12Hausenloy D.J. Candilio L. Evans R. et al.Remote ischemic preconditioning and outcomes of cardiac surgery.N Engl J Med. 2015; 373: 1408-1417Crossref PubMed Scopus (518) Google Scholar,13Meybohm P. Bein B. Brosteanu O. et al.A multicenter trial of remote ischemic preconditioning for heart surgery.N Engl J Med. 2015; 373: 1397-1407Crossref PubMed Scopus (466) Google Scholar In clinical contexts, neuroprotective treatments can usually only be started with a certain delay after neuronal damage, as the onset of brain injury is nearly always unpredictable. Hence, protective strategies relying on pre-conditioning are frequently not applicable (with the exception of few selected intraoperative situations e.g. deep hypothermic circulatory arrest or clamping of the carotid artery).14Chakkarapani E. Dingley J. Liu X. et al.Xenon enhances hypothermic neuroprotection in asphyxiated newborn pigs.Ann Neurol. 2010; 68: 330-341Crossref PubMed Scopus (124) Google Scholar Hence, our findings that xenon is not only protective when used for pre-conditioning, but also when administered as post-conditioning (up to 2 h after injury), renders xenon an interesting candidate for future clinical trials (Fig 1). A limit to this conclusion is that only a few studies have analysed the comparative efficacy of xenon pre-conditioning vs post-conditioning in the same model.15Zhao H. Yoshida A. Xiao W. et al.Xenon treatment attenuates early renal allograft injury associated with prolonged hypothermic storage in rats.FASEB J. 2013; 27: 4076-4088Crossref PubMed Scopus (29) Google Scholar Our meta-analysis demonstrated protective effects for xenon even when used in subanaesthetic concentrations, suggesting that neuroprotection can be attained independently from anaesthesia (Fig 1). A protective effect was seen for non-anaesthetic doses of xenon 50% in rats corresponding to a dose of 19% in humans, although exact dose conversions for neuroprotection are currently unknown. This is particularly interesting as significant cost reductions can be achieved by using lower doses of this expensive noble gas. Using lower concentrations of xenon potentially helps avoid sedation/anaesthesia in patients with brain injury, allows higher doses of inspired oxygen in patients with concomitant hypoxic respiratory failure, or both. Further clinical trials should therefore probably avoid anaesthetic concentrations of xenon and rather systematically explore the effectiveness of lower xenon doses. Our study is subject to some methodological weaknesses that have to be acknowledged when interpreting the results. First, the funnel plots for both the experimental and the human studies might be suggestive of publication bias. Given the obvious lack of reported small negative effect sizes, a negative relation between effect size and sample size was observed. Second, the preclinical studies are subject to many methodological discrepancies and use a wide spectrum of experimental set-ups. While this diversity allows a comprehensive overview, the effect sizes are derived from widely differing data leading to significant heterogeneity (I2Cerejeira J. Firmino H. Vaz-Serra A. Mukaetova-Ladinska E.B. The neuroinflammatory hypothesis of delirium.Acta Neuropathol. 2010; 119: 737-754Crossref PubMed Scopus (269) Google Scholar) in the results. In contrast, this heterogeneity is virtually absent in the included clinical trials. The results of this meta-analysis remain inconclusive. Neuroprotective effects of xenon were primarily observed in animal models investigating cerebral ischaemia. However, these results are affected by possible publication bias and by significant heterogeneity. In humans, the effects of xenon were small. Clinical effectiveness should be confirmed in adequately powered randomised clinical trials, particularly focusing on patients suffering from post-ischaemic/post-hypoxic brain injury.16Gardner A.J. Menon D.K. Moving to human trials for argon neuroprotection in neurological injury: a narrative review.Br J Anaesth. 2018; 120: 453-468Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar Comply with ICMJE recommendations: all authors. Assessment of articles: L.V.H., S.R. Data extraction: L.V.H. Analysis of data: S.F. Interpretation of data: L.V.H., S.R. Primary manuscript preparation: L.V.H. Manuscript editing: S.R., S.D., L.A., S.F., R.D.S. Approval of the final version and accountable for all aspects of the work: L.V.H., L.A., S.D., S.F., R.D.S, S.R. S.R. was supported by an unrestricted research grant from AirLiquide Belgium, a research grant of AirLiquide France and the Foundation Annie Planckaert-Dewaele (Biomedical Sciences Group, KU Leuven). L.V.H., R.D.S, L.A., S.D., and S.F. have no conflicts of interest to declare. We would like to thank the librarians of the KU Leuven for sharing their expertise in searching the electronic databases and C. Annys for valuable help in the search and screening of articles.