Presented at the Neonatal Society 2005 Spring Meeting.

Carmichael DW, Wang X, Bainbridge A, Thornton JS, Cady EB, Raivich G, Ordidge RJ, Peebles DM

Centre for Perinatal Brain Protection and Repair, University College London, London, UK

Introduction: Perinatal brain injury is commonly multifactorial in origin. We have previously reported that in the chick embryo in-ovo histological evidence of neuronal cell death is more severe following a combination of hypoxia and exposure to bacterial endotoxin (a model of infection) than for each factor alone (1). The hypothesis to be addressed in this study was that endotoxin pre-treatment increases hypoxia-associated neural cell death by impairing the normal metabolic and haemodynamic responses to acute hypoxia. The two main aims were 1) to measure, by magnetic resonance, changes in cerebral metabolism, water diffusion, blood volume and oxygenation during and after acute hypoxia 2) to determine if the acute response was altered by pre-treatment with bacterial endotoxin.

Methods: Nineteen white Leghorn chicken eggs (55-65g) were incubated at 38ºC and 55-70 % humidity. On incubation day 19, chick embryos received 0.4 ml of either saline (n=10) or lipopolysaccharide (LPS, 3mg, n=9) from Salmonella Typhimurium 4 hours prior to transient hypoxia (FiO2=0.04, 60 minutes) which was initiated whilst in the bore of a 7T Bruker Spectrometer. All the embryos received 0.5mg of tubocurarine dropped onto the chorioamniotic membrane for immobilization during data acquisition. The periovo environment was controlled to maintain gas flow, oxygenation and temperature (38ºC). The following MRI/MRS measurements were repeated approximately every 15 minutes before, during and post hypoxia: proton spectroscopy, quantitative diffusion weighted imaging (an indicator of metabolic energy status), and quantitative T2 and T2* imaging (reflects changes in blood oxygenation and volume). MR parameters are displayed in Table 1. Metabolite peak-area ratios were calculated from the spectra using LCModel (2). Apparent water diffusion coefficient (ADC) maps were calculated for each diffusion direction (x/y/z) and from these 1/3 Trace images (Dav) were obtained. Additionally, both T2 and T2* maps were calculated. For ADC, T2 and T2* maps a region was drawn covering the whole brain for each image and the mean value and approximate time of scan recorded. The average responses of each group (hypoxia only and hypoxia + LPS) were obtained for each time interval. Additionally, the measured values were tested for intergroup (hypoxia / hypoxia + LPS) and interperiod (baseline / hypoxia / post-hypoxia) differences using a multifactor analysis of variance (ANOVA).

Results: Results Compared to baseline values lactate (Lac) / total creatine (Cr) had increased significantly by the end of hypoxia (60 mins) (see Table ) and then returned close to baseline by 120-minute post hypoxia. Similar findings were observed in both groups. N-acetylaspartate (Naa) peak-area ratios did not change significantly with respect to hypoxia or experimental group (values not given). Global ADC fell significantly (typically by 10-15% baseline values) during hypoxia and then returned to baseline values within 30 minutes of the end of hypoxia. There were no significant intergroup differences between baseline ADC or size of ADC reduction during hypoxia. T2 and T2* also fell during hypoxia and then returned to baseline. There was no significant T2 difference between the experimental groups. However, compared to the hypoxia only group T2* was significantly reduced (p<0.05) in the LPS + hypoxia group before, during and after hypoxia. In addition, the T2* decrease during hypoxia was significantly attenuated (p<0.05) in the LPS + hypoxia group.

Conclusions: The observed decrease in global ADC suggests that despite anaerobic glycolysis, indicated by the large increase in Lac peak-area ratios, this hypoxic insult results in impaired cellular energy generation. However, both the size of ADC fall (only 10-15%) and lack of histological cell death (1) suggest that the extent of energy failure following hypoxia alone is limited. Pretreatment with LPS did not appear to increase the effect of hypoxia on cellular energetics (the magnitude of ADC and lactate/Cr change was similar in both groups). The decrease in T2* during hypoxia is likely to be due to a reduction in blood oxygenation as well as a compensatory increase in cerebral blood volume. The lower baseline values for T2* observed in the endotoxin group are consistent with LPS increasing cerebral blood volume. Although it is possible that these haemodynamic perturbations may contribute to the increased cell death observed with LPS pre-treatment, the absence of any effect of LPS on the metabolic response to hypoxia suggests that this synergistic effect is more likely to be caused by an endotoxin related upregulation of cellular inflammatory mechanisms.

Acknowledgements: The Wellcome Trust supported this work

References
1. Wang et al; Society for Neuroscience, 2004.
2. Provencher SW; NMR Biomed. 2001 Jun;14(4):260-4.