Pretreatment prediction of the chemotherapeutic response of human glioma cell cultures using nuclear magnetic resonance spectroscopy and artificial neural networks.

Both tumor metabolism and its response to cytotoxic drugs are intrinsic properties of tumor cells. It is therefore likely that there is a relationship between the two properties, however subtle and complex, wherein the metabolic characteristics of tumor cells can reflect the inherent response (resistance or sensitivity) of these cells to cytotoxic drugs. We used artificial neural network analysis to show that it is possible to distinguish, prior to treatment, between drug-resistant and drug-sensitive human glioma cell cultures from their metabolic profiles, as given by high-resolution proton nuclear magnetic resonance spectra of the cell extracts, and to predict their cellular response to the chemotherapeutic drug 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea in vitro. The results suggest that neural network analysis of tumor nuclear magnetic resonance spectra has potential as a prognostic tool for determining treatment of gliomas, ultimately noninvasively, and may be used to provide information about the metabolic pathways involved in drug response that may be helpful in developing novel treatments for these tumors.

Contribution of cerebrovascular parasympathetic and sensory innervation to the short-term control of blood flow in rat cerebral cortex.

In two groups of normotensive rats anaesthetised with halothane, either the nasociliary nerve (NCN) or the NCN and parasympathetic (PS) fibres together (NCN-PS) were functionally blocked at the right ethmoidal foramen. Blocking was achieved reversibly and repeatedly using a cooling probe. Cortical regional CBF (rCBF) was measured bilaterally using laser-Doppler probes. In Group 1, bilateral common carotid occlusion (CCO) was applied for 1 min both with and without block. In Group 2, CCO was applied permanently followed by stages of controlled haemorrhagic hypotension to deepen the ischaemia and the block applied at each stage. In Group 1, during CCO, rCBF was unaffected by blocking NCN-PS. However, during the transient postocclusive hyperaemia, blocking NCN-PS, but not NCN alone, significantly increased right side rCBF. In Group 2 and in Group 1 in the absence of CCO (normotension), rCBF was unaffected by blocking either set of fibres. We conclude that neither NCN nor PS fibres contribute to the tonic level of rCBF or to its autoregulatory control, but PS fibres conduct impulses tending to resolve postischaemic hyperaemia. We suggest that a subpopulation of PS fibres containing neuropeptide Y is activated under conditions of supernormal rCBF.

Comparison of the effects of ischaemia on early components of the somatosensory evoked potential in brainstem, thalamus, and cerebral cortex.

In 14 ventilated, normocapnic baboons anaesthetised with alpha-chloralose, local CBF (hydrogen clearance) and the amplitude and latency of local components of the somatosensory evoked potential (SEP, median nerve stimulation) were measured bilaterally in ventrobasal thalamus (VPL), medial lemniscus (ML), and cerebral cortex before and during progressive ischaemia, produced by occlusion of the right middle cerebral artery and subsequent controlled reductions in mean systemic blood pressure (MSBP). The first significant reduction from control of the left cortical SEP amplitude occurred in the range of 30-40 mm Hg MSBP, but those of the VPL and ML responses only below 30 mm Hg; in the range of 20-30 mm Hg, the average SEP amplitudes in cortex, VPL, and ML were 8.6, 72.6, and 90.7% of control, respectively. In terms of local CBF, the cortical SEP threshold was in the range of 15-20 ml/100 g/min (as in previous work), that of VPL in the range of 10-15 ml/100 g/min, but the ML response was only markedly reduced below 10 ml/100 g/min. Thus, the differential ischaemic sensitivity of the SEP between the three regions was clearly demonstrated. These results indicate that as one descends the neuraxis, there is an increasing resistance of electrophysiological function to systemic hypotension, together with a decreasing threshold for local ischaemia.

Changes in extracellular calcium activity in cerebral ischaemia.

Changes in extracellular ion activities were measured during partial ischaemia of the cerebral cortex of primates anaesthetized with alpha-chloralose. Triple-barrelled, double-ion-sensitive microelectrodes were used to measure the extracellular potassium (Kc) and calcium (Cac) activity at the same point simultaneously. The ion changes were related to local cerebral blood flow, and it was shown that at a blood flow of approximately 10 ml 100 g-1 min-1, there is a threshold below which ion homeostasis is disturbed. This is associated with a dramatic rise in Ke and fall in Cae. Cae falls from a normal value of 1.31 +/- 0.1 mM to approximately 0.28 mM in densely ischaemic tissue. In ischaemia, Ke reaches 13.4 +/- 3.8 mM before Cae begins to fall. The fall in Cae, although related to reduced blood flow, is closely associated with and follows the rise in Ke. The change in Cae is probably due to an increase in membrane permeability, as a result of either depolarisation or a critical lowering of cellular energy reserves.

Firing patterns of pallidal cells in parkinsonian patients correlate with their pre-pallidotomy clinical scores.

It is unclear how the disordered activity of cells in the basal ganglia contributes to the symptoms of Parkinson's disease (PD). We recorded from single neurons extracellularly in 3 regions of the globus pallidus (GPe, GPie and GPii) in patients undergoing pallidotomy for PD. Movement-related cell firing patterns, analysed using hidden Markov models, were significantly correlated with patients' preoperative clinical scores (off drugs). Responses of cells in GPii correlated best with the scores for specific motor tasks, rather than general ones related to activities of daily living, but the reverse was true for responses from GPe. In both GPii and GPe, a higher score (i.e. greater parkinsonian severity) was associated with greater variability in cell firing rather than an increase in firing rate itself.

Cerebrovascular parasympathetic innervation contributes to coupling of neuronal activation and blood flow in rat somatosensory cortex.

We measured the increase of regional cerebral blood flow (rCBF) in the somatosensory cerebral cortex occurring in response to a standard stimulation of the L. side mystacial vibrissae (facial whiskers) in rats anaesthetised with halothane, in conjunction with blocking of activity in the R. side parasympathetic (PS) and sensory fibres innervating the cerebral vessels. Blocking was achieved reversibly and repeatedly by means of a cooling probe. When the PS fibres and the nasociliary nerve (NCN) were blocked together, but not when the NCN was blocked alone, the R. side rCBF increase occurring with whisker stimulation was significantly reduced. Our results indicate that, in addition to the intrinsic cortical factors demonstrated in earlier studies, the cerebrovascular PS innervation, but not the NCN, contributes to the increase in cortical rCBF associated with somatosensory cortical neuronal activation.

Extracellular potassium activity, evoked potential and tissue blood flow. Relationships during progressive ischaemia in baboon cerebral cortex.

Extracellular K+ activity (Ke), local tissue blood flow and the cortical evoked potential (EP) were measured concurrently in the cerebral cortex of baboons anaesthetised with a-chloralose. Flows were progressively reduced from normal by occlusion of the middle cerebral artery and controlled steps of exsanguination. Our data suggest that 3 stages may be identified in the disturbance of K+ homeostasis produced by progressive ischaemia. In the first stage, at flow levels similar to those sufficient to abolish the EP (12-16 ml/100 g/min), small, self-limiting increases in Ke occur, probably reflecting K+ efflux into the extracellular space (ECS) with partial impairment of K+ clearance from the ECS. The second stage occurs at distinctly lower (P less than 0.01) levels of flow (8-11 ml/100 g/min), and is characterized by a massive (30-80 mM) increase in Ke, which we attribute to an increase in ionic permeability of cell membranes with further impairment or overloading of K+ clearance mechanisms. In the third stage, at flows below about 6-8 ml/100 g/min, the data indicate an inverse relationship between flow and Ke with persisting high Ke levels, suggesting complete loss of K+ clearance. Transient increases of Ke also occur in the flow range 4-13 ml/100 g/min, the rate of recovery of Ke in their decay phase being positively corelated with flow (P less than 0.005).

Reversibility of ischaemically induced changes in extracellular potassium in primate cortex.

Following the massive increase in extracellular potassium activity that occurs in cerebral cortex when local blood flow falls below 8--11 ml/100 g/min, recovery of potassium toward normal levels might be expected when flow is restored. This study assessed the reversibility of such potassium increases, produced by middle cerebral artery occlusion in 13 baboons anaesthetised with alpha-chloralose, in relation to a wide range of ischaemic duration and density and post-occlusion flow. Potassium was measured with ion-exchanger microelectrodes and flow by hydrogen clearance. The artery was occluded for 136 +/- 63 min (mean +/- SD) and measurements were continued thereafter for 93 +/- 57 min without systemic hypertension. Upon reperfusion, partial or complete recovery (i.e., to within control confidence limits) of potassium was seen in all animals, but the rate of recovery varied widely and potassium clearance showed bi-compartmental characteristics in 7 animals. The fast component (or initial slope) rate constant was significantly correlated with post-occlusion flow and (inversely) with the duration of occlusion for which flow fell below the arbitrary threshold of 10 ml/100 g/min (the flow deficit). The slow component was unrelated to these quantities. Complete recovery was associated with a significantly higher post-occlusion flow, and smaller flow deficit, than was partial recovery. Secondary increases in potassium, associated with relatively high flow deficits and post-occlusion flows, were seen in 5 animals. These results are discussed in terms of factors that may determine potassium clearance and the possibility that elevated levels of potassium (demonstrated here to be prolonged well into the post-occlusion phase) might influence the evolution of a cortical infarct.

Changes in regional cortical tissue oxygen tension and cerebral blood flow during temporary middle cerebral artery occlusion in baboons.

Cortical tissue oxygen measured by a platinum cathode, and cerebral blood flow recorded by a hydrogen clearance technique, were measured in 13 baboons before, during and after temporary occlusion of the middle cerebral artery. Mean control pO2 was 23.8 +/- 14 mm Hg and mean flow 51.3 +/- 12 ml/100g/min. During the occlusion, there was a gradation in pO2 from values in the opercular area of 3.6 +/- 5.9 mm Hg, to values in the high parietal area of 11.9 +/- 11.7, these being statistically different (P less than 0.05) from each other. The corresponding flow values were 5.5 +/- 7.5 (opercular) and 22.3 +/- 21.7 ml/100 g/min parietal (P less than 0.01). Following removal of the MCA clip, between 20% and 30% of the electrodes registered an early hyperoxia and hyperaemia, which lasted up to 5 min. A late and prolonged hyperoxia, with less evidence of hyperaemia, was also noted in about 20%. The mean tissue pO2, however, at 5-min intervals up to 40 min following the removal of the clip only reached 60-80% of control values in the most ischaemic areas. Only the parietal region showed a mean pO2 above control levels. The mean flow data were uniformly reduced in all regions to about 80% of control values. During and after a second occlusion in 6 animals, similar changes were noted but with even fewer instances of hyperoxia. The mean oxygen and flow results were lower than with the first occlusion, but the reduction was not significant. There was no overall effect of hypercapnia on cortical tissue pO2 during the control period, but there was a significant (P less than 0.05) reduction during the same procedure after the period of ischaemia. An increase in pO2 during hypercapnia could be observed if there were arousal responses of blood pressure "spikes".

Measurements of oxygen tension in the cerebral cortex of baboons.

A polystyrene-covered platinum electrode (100-150 mum diameter) has been used to measure cortical tissue oxygen tension in baboon brains. The method of preparation, calibration, and the importance of small residual current (less than 40 nA) as an attribute of a reliable electrode, are described. With electrodes of this size, there was a large (16 +/- 12nA/torr) and linear current output with pO2 changes. The effect of avrious gases in addition to oxygen is described; halothane inhalation increases the apparent pO2 and hydrogen, used for blood flow estimations, reduces the recorded pO2. In 48 separate electrode placements in 13 baboons, the mean cortical qo2 was 23.8 +/- 12 mm Hg, with a range from 1-79 mm Hg; following occlusion of one middle cerebral artery, 37 electrodes recorded a pO2 of less then 5 mm Hg pO2 Oscillations were invariably noted in control conditions, independent of blood pressure; these waves disappeared during MCA occlusion and appeared to be augmented following release of the clip. Blood pressure "spikes" produce immediate and synchronous changes in all electrodes entirely different from the spontaneous waves. Such blood pressure changes may mask the true effect of hypercapnia on tissue pO2 and, if ignored, may lead to erroneous assumptions regarding local neural control of the circulation, the increased pO2 secondary to hypertension being regarded as evidence of regional vasodilation. A SUdden change in inspired pO2-the "air test"-was performed in control conditions and following the ischaemic insult, and the rate of change of cortical pO2 compared. The gradient was significantly greater (P less than 0.05) following ischaemia, suggesting a changed ratio in the tissue's flow to oxygen requirements and/or a persisting vasodilatation.

CBF and time thresholds for the formation of ischemic cerebral edema, and effect of reperfusion in baboons.

Ischemic cerebral edema has been studied in 41 baboons, with regional cerebral blood flow (CBF) determined by hydrogen clearance, and edema measured by microgravimetry. A threshold of ischemia has been identified for baboon cortex and subcortical white matter, which has to be crossed before edema formation begins. This threshold is 40.5% of normal CBF in cortex, and 34.4% of normal flow in subcortical white matter. A time threshold has also been determined, and the baboon brain can withstand 30 minutes of ischemia of the middle cerebral artery without significant edema formation. Reperfusion of ischemic brain has no effect on tissue water if the ischemic flow and time thresholds have not been crossed. Reperfusion of cortex, where water has begun to accumulate, exacerbates the water accumulation in proportion to the extent of the reperfusion. If these results are applicable to man, restoration of flow should not be attempted after an ischemic insult that reduces flow to less than 40% of normal unless it can be accomplished within 30 minutes of the insult. Provided CBF can be restored to above the 40% threshold within 30 minutes, reversal of the neurological deficit and prevention of ischemic edema can be expected.

Kinetics of resolution of transient increases in extracellular potassium activity: relationships to regional blood flow in primate cerebral cortex.

Previous studies have established that in cerebral cortex subjected to progressive reduction in blood flow, two distinct thresholds of flow may be identified below which cellular function is impaired: the cortical evoked response loses amplitude when local flow falls below 18ml/100gm/min, and below 11ml/100gm/min a major increase in extracellular K+ activity (Ke) occurs. However, further evidence suggests that even at higher flows the capacity of the tissue to handle induced ionic changes may be impaired. To investigate this point, we studied the kinetics of resolution of Ke following a transient increase produced by local electrical stimulation, in relation to the local pre-stimulus flow (reduced by acute middle cerebral artery occlusion) in baboons. Flow was measured by the hydrogen clearance method and Ke by ion-exchanger micro-electrodes, in the same cortical regions. In primary induced transients (those increases in Ke elicited by cortical stimulation, and reported previously,) Ke attained a maximum value of 8-10 mM and then decayed towards the 4-mM baseline. The half-time of this decay was significantly increased from normal in the flow range 20-40 ml/100 gm/min, and increased further at lower flows until, below 11ml/100gm/min, Ke clearance was undetectable. Thus, cortical ion homeostasis appeared impaired at flows substantially closer to normal than those thresholds mentioned above, a result discussed in terms of impairment of active Ke clearance mechanisms. Secondary induced transients arose during a primary induced transient, reaching considerably higher peak values (8-30 mM) of Ke (indicating temporary clearance loss) and with slower decay rate than the primary. Spontaneous transients, not associated with any stimulus, were also observed; like secondary transients, they occurred only at flows below 20ml/100gm/min and showed a reduction in clearance rate with progressive ischemia. They resemble spreading depression and their generation is discussed in terms of the ionic and metabolic conditions at their time of origin.

Intracranial pressure changes following primate subarachnoid haemorrhage.

In 29 anaesthetized baboons avulsion of a small intracranial artery was used to produce a subarachnoid haemorrhage, in a closed-skull situation. Intracranial pressure was measured by extradural transducers, and arterial pressure was also measured continuously, with periodic measurements of cerebral blood flow. After haemorrhage there was an immediate fall in cerebral perfusion pressure in nearly all cases, reaching zero in 9 animals. In 18 there was a significant pressor response in the systemic circulation, but perfusion pressure usually remained low in spite of this response. Perfusion pressure recovered after a few minutes in most cases. In the 19 cases where intracranial pressure was measured on both sides, differences occurred in 11, with the higher pressure always on the same side as the haemorrhage. The difference was evident very soon after haemorrhage in 9 cases, and lasted over half an hour in 5 of them. The mechanism of arrest of bleeding was, in most of this series, not that of a zero perfusion pressure. Explanations for this and for the occurrence of differential pressures are discussed.