1.

Introduction

Previous studies have demonstrated that there is a decrease in cerebral blood flow (CBF) in animals with acute or chronic hyperglycemia.¹ Hyperglycemia might have a longterm effect on CBF, especially under the condition of cerebral ischemia. Lanier² reported that within $24\mathrm{h}$ of the onset of cerebral ischemia, even mild hyperglycemia is associated with exacerbated cerebral damage and increased morbidity and mortality; Bruno³ also described patients with nonlacunar stroke, in which high blood-glucose levels on admission were associated with a worse outcome at $3\text{months}$ . It has been reported that 20–50% of strokes are accompanied with concomitant hyperglycemia regardless of the presence of pre-existing diabetes.⁴ This hyperglycemia occurring during experimental and clinical strokes has been associated with increased cerebral damage.⁴

Cortical spreading depression (CSD) is characterized by transitory, spreading neuronal depolarization that propagates like a wave on the cortical surface. Many studies have shown that normal brain tissue invaded by CSD presents reversible changes in CBF, glucose metabolism, transmembrane ion transport and so on.^{5, 6, 7, 8} There is evidence that CSD plays an important role in ischemic stroke, and it has been commonly accepted that CSD increases the ischemic volume and aggravates ischemic injury.⁸ This deterioration of brain tissue resulting from CSD waves may be caused by its reducing further the already decreased blood flow in the penumbral zone of the ischemic cortex.⁸

Hyperglycemia and CSD can worsen the outcome of ischemic stroke and also influence CBF.^{1, 8} Therefore, it is possible that both of them have a prolonged cooperative effect on CBF in ischemic stroke. Research on the changes in CBF following CSD during acute hyperglycemia for a long period might provide evidence for this possibility.

To shed light on the spatiotemporal changes in CBF for the different vascular compartments in the cortex, we selectively used a technique of laser speckle imaging to provide a directly spatial pattern of changes in blood flow, avoiding the sampling restriction of the previous techniques. Laser speckle imaging,⁹ as a means of enabling CBF imaging at high spatiotemporal resolution, has recently attracted extensive attention, and it has been successfully applied to explore relative changes in CBF induced by CSD,¹⁰ cerebral ischemia,^{11, 12} and sensory stimulation.¹³ It has been demonstrated that not only laser speckle imaging and laser Doppler imaging provide equivalent assessments of blood flow,¹⁴ but also the relative measurement of flow changes by laser speckle imaging can be accurately calibrated by the absolute flow imaging provided by spectral-domain Doppler optical coherence tomography.¹⁵

Here, we adopt an acute rat hyperglycemic model, in conjunction with laser speckle imaging, to monitor the long-term effect of acute hyperglycemia and the additional effect of CSD during acute hyperglycemia on CBF for different vascular compartments.

2.

Materials and Methods

3.

Results

All physiological parameters were maintained within normal limits throughout the experiments.

3.1.

The Effect of Acute Hyperglycemia on CBF in Normal Rat Cortex

Figures 1a and 1b illustrate the status of CBF before or after the administration of agents (saline or D-glucose) in a representative subject. In these ICT images with pseudo color, the regions with a higher level of blood flow are reddish, whereas the regions with lower velocity of blood flow are shown by blue. By comparing the images of two panels, it was apparent that as a result of the glucose administration, the CBF in the rat with acute hyperglycemia exhibited a distinct increase, peaking about $90–120\min$ after glucose injection, and finally recovered toward the baseline level [Fig. 1b].

Fig. 1

The effect of acute hyperglycemia on CBF in a normal rat cortex. (a) The distribution of CBF perfusion was measured by laser speckle imaging before or after saline administration in a representative subject. (b) The distribution of CBF perfusion was measured by laser speckle imaging before or after glucose administration in a representative subject. Three vascular compartments (parenchyma, arteriole, and venule) (a)- $0\min$ or (b)- $0\min$ are indicated as three solid squares numbered 1, 2, and 3. (c) The time courses of saline-induced relative changes in CBF at the three vascular compartments are shown. (d) The time courses of glucose-induced changes in CBF at the three vascular compartments are denoted. Direction: A, anterior; M, median. Scale bar, $1\mathrm{mm}$ . All of the ICT images share the same direction and color bar. (Color online only.)

The time courses of the relative changes in CBF for three vascular compartments [as drawn in Fig. 1a- $0\min$ or 1b- $0\min$ ] are shown in Fig. 1c and 1d, correspondingly. These curves in Fig. 1c show that the relative changes in CBF after saline administration at different vascular compartments had slight fluctuation over $180\min$ periods, ranging from 88 to 108% baseline value. Generally, we believe that normal CBF levels always vary from 80 to 120% baseline. However, the relative changes of CBF during acute hyperglycemia had a significant difference among the various vascular compartments as shown by Fig. 1d. First, the relative changes of CBF at the arteriole and parenchyma showed significant increase, but the amplitudes of the arterial response were larger than that of parenchyma. Second, the relative changes of CBF in the arteriole rise from $30\min$ (111% baseline), peaking at $150\min$ (183% baseline value). Increases in CBF at the parenchyma begin at $60\min$ (108% baseline value), peaking at $120\min$ (162% baseline value) after the glucose injection. Whereas no significant changes in CBF were detected at the venule, the relative changes of CBF varied from 89 to 108% baseline value [Fig. 1d].

3.2.

The Effect of Acute Hyperglycemia on CBF Following CSD in Rat Cortex

The length of time for a wave to first appear in the imaged cortex was usually $\sim 2\min$ after delivering electrical stimulation into the frontal cortex. CBF $30\min$ after each CSD wave was also monitored using laser speckle imaging. Figures 2a and 2b illustrate spatial and temporal changes in CBF following each CSD in the rats of the control group and the hyperglycemia group, respectively. Two series of ICT images in Fig. 2a and 2b were acquired at a $30\text{-}\min$ interval. In the control group, CBF following each CSD was usually close to the baseline level during the experiment, as shown in Fig. 2a; however, CBF following each CSD in the hyperglycemia rats had a stepwise reduction within $180\min$ after glucose injection [Fig. 2b].

Fig. 2

The effect of acute hyperglycemia on CBF after CSD in rats. (a) The distribution of CBF perfusion was measured by laser speckle imaging before or after saline administration in a representative subject. (b) The distribution of CBF perfusion was measured by laser speckle imaging before or after glucose administration in a representative subject. Three vascular compartments (parenchyma, arteriole, and venule) in (a)- $0\min$ or (b)- $0\min$ are indicated as three solid squares with numbered 1, 2, and 3. (c) The time courses of saline-induced relative changes in CBF after CSD monitored at the selected region are shown. (d) The time courses of glucose-induced changes in CBF after CSD monitored at the region [(b)- $0\min$ ] are denoted. Direction: A, anterior; M, median. Scale bar, $1\mathrm{mm}$ . All of the ICT images share the same direction and color bar. (Color online only.)

Figures 2c and 2d show the time courses of the saline (glucose)-induced changes in CBF following CSD for all vascular components, as identified using the procedure illustrated in Fig. 1c and 1d. From Fig. 2c, we obtained that the relative changes of CBF in the three vascular compartments (parenchyma, arteriole, and venule), respectively, were 100–124, 96–124, and 95–108% of the baseline value, which were all close to the baseline level. There were no significant differences between them.

The relative changes of CBF following each CSD gradually fell off in the vascular compartments during acute hyperglycemia. These decreases in CBF for all compartments first appeared at $30\min$ after glucose administration and reached their minimums at $150\min$ . The minimums of relative changes were, respectively, 49, 56, and 60% of the baseline value [Fig. 2d]. There were not significant differences in relative changes of CBF among the various vascular compartments.

4.

Discussion

It has been previously shown that CBF decreases due to acute or chronic hyperglycemia.¹ This phenomenon may provide useful information for the pathophysiology of ischemic stroke, suggesting possible mechanisms for the deleterious effects of hyperglycemia on cerebral injury due to ischemia. However, the methods used in the previous researches might have some limitations in spatial or temporal dimension, and these limitations might result in some details being neglected in this phenomenon. Laser speckle imaging based on laser speckle temporal contrast analysis, as an advanced technique, can monitor the changes of CBF with high spatiotemporal resolutions, and this technique has been demonstrated in our previous works.^{17, 18} Here, we also used the technique for tracing the long-lasting effect of acute hyperglycemia on the CBF in normal cortexes or those with CSD, and thus to further explore the possible location of the vasculature where CBF is regulated.

We found that CBF over the normal rat cortex during acute hyperglycemia shows an extensive increase in this study. These results were also consistent with previous studies from systemic arteries which demonstrated that basal blood flow increased after $6\text{-}\mathrm{h}$ hyperglycemia.¹⁹ It is likely that the increase in CBF due to hyperglycemia was largely an effect of a tight coupling between cerebral metabolic rates of oxygen/glucose and CBF. The tight coupling was reportedly present in the resting brain, where high-level metabolic activity is accompanied with a high perfusion level of blood flow.²⁰ In general, full oxidation of glucose that produces carbon dioxide and water always occurs in the resting brain. During acute hyperglycemia, high plasma glucose could lead to an increase in cerebral metabolic rates of oxygen due to increases in consumption of oxygen to perform the full oxidation of glucose.²¹ Thus, the raise in cerebral metabolic rates of oxygen could result in a distinct increase in CBF according to the tight coupling as described above, that is, increased blood flow in response to increased metabolism.

We also found that this increase of CBF related to acute hyperglycemia, as mentioned above, began at the arteriolar level and then was observed at the capillaries (parenchyma). Arterioles had the largest response, whereas the venous response was less evident. These results are in agreement with the activity-dependent changes of cortical blood volume in cats reported by Vanzetta ²² Changes in cerebral blood volume displayed an apparent similarity with the CBF.²³ While the mechanism linked to the different changes in CBF at the all vascular components is still unknown, one potential mechanism may be related to the orientation of blood flow. Augmentation of CBF at the arteriole could lead to a passive increase in CBF at the downstream of the involved arteriole.

Our results also showed that in the normal rat, the relative changes in CBF following CSD at all vascular components were close to the baseline level. Previous studies have shown that a single episode of CSD is accompanied by reversible changes in CBF; it is characterized by a temporary CBF increase, followed by a long-lasting hypoperfusion in the cortex, terminated by a brief, small hyperperfusion,⁸ and the duration of this reversible changes in CBF is within $15\pm 3\min$ .²⁴ In our experiments, the interval of $30\min$ between two CSD waves is sufficient for recovering the disturbance of CBF due to CSD.

Finally, we found that acute hyperglycemia causes simultaneous and gradual decreases in CBF following CSD throughout the vascular compartments. First, there have been reports asserting that most of the additional energy needed by activated brain area during CSD is provided by nonoxidative metabolism (glycolysis), although there is sufficient oxygen in the brain tissue. Meanwhile, hyperglycemia can also inhibit oxygen consumption in the active brain.¹ The two factors, hyperglycemia and CSD, have cooperatively aggravated the fall of the oxidative metabolism rates, finally resulting in the decline of blood flow in the brain. Second, CSD is consistently associated with changes in the caliber of surface blood vessels, which was first reported by Leao.²⁵ Recently, Brennan, ²⁶ also reported there is significant dilatation of cortical surface arteriolar associated with CSD. Many different mechanisms of CSD-mediated arteriolar changes have been proposed.²⁶ Possible dilators include nitric oxide,²⁷ and parasympathetic innervation by acetylcholine,²⁸ and so on. Nakahata noted that acute exposure toward hyperglycemia reduces cerebral vasodilatation produced by acetylcholine and related to nitric oxide synthases.²⁹ The long-lasting hypoperfusion among the disturbance of CBF due to CSD in the acute hyperglycemia group would be more difficult to recover than that in the control group because of a lack of effective cerebral vasodilatation. When a rat experienced acute hyperglycemia, its CBF presented gradual decreases after disconnected stimulation that evoked repetitive CSD.

These results also raise a question of why the vascular compartments showed stepwise reductions of CBF under this situation. Possible mechanisms could include impaired cerebrovascular autoregulation. It has been accepted that cerebrovascular autoregulation related to flow-pressure relationships play an important role in regulation of CBF. The autoregulation of CBF in the acute-hyperglycemia rats without CSD invasion was very good within a certain arterial pressure range, just as in normal rats.³⁰ However, cerebral vascular autoregulation of blood flow might be compromised as a result of hyperglycemia³¹ accompanied with CSD invasion. This hypothesis is still awaiting future experimental confirmation.

The ability of hyperglycemia to aggravate cerebral ischemic damage has been well accepted,³² whereas the underlying mechanisms are still not completely understood. Our results show that there are significant differences in the time course of changes in CBF during hyperglycemia between normal rats and those with CSD, suggesting that there may be a coeffective action between hyperglycemia and CSD on CBF, leading to the exacerbated outcome of ischemic stroke.