In Vivo Uptake of [ 3 H]Nimodipine in Focal Cerebral Ischemia: Modulation by Hyperglycemia

Focal cerebral ischemia and hyperglycemia

Under anesthesia femoral arterial and venous cannulae were placed to measure mean arterial blood pressure, obtain blood samples and administer drugs, glucose and radiotracers. Arterial blood gases were monitored and ventilator settings adjusted accordingly. The left middle cerebral artery (MCA) was exposed (Tamura et al., 1981) and occluded by two microaneurysm clips (Sundt #1, Codman, Peterborough, Ontario, Canada) placed in tandem on the vessel proximal and distal to the olfactory tract. The ipsilateral common carotid artery (CCA) was simultaneously occluded with an aneurysm clip (George Tiemann and Co., Hauppauge, NY, U.S.A.) to produce a combined MCA + CCA occlusion. The clips were kept in place for 30 minutes, at which time the study was terminated by decapitation. Brain temperature was monitored using a 300-μm-diameter thermocouple probe placed into exposed cortex adjacent to the MCA occlusion site and a warming blanket and radiant heat were used to maintain this temperature between 36° and 37°C.

Hyperglycemia was established before MCA + CCA occlusion by continuous intravenous infusion of a 50% glucose solution in normal saline at an initial rate of 1.2 mL/h. The infusion was started 54 ± 9 minutes before MCA + CCA occlusion and the rate was adjusted to maintain blood glucose concentration between 2 and 3 times control levels. Normoglycemic rats were infused with normal saline intravenously at the same rate.

Cerebral blood flow measurements

Regional CBF was measured using the method of Sakurada et al. (1978). An intravenous ramp infusion of 30 μCi of [¹⁴C]iodoantipyrine ([¹⁴C]IAP) (57 mCi/mmol, Amersham International PLC, Little Chalfont, England) in 1.6 mL of normal saline was started 29 minutes after MCA + CCA occlusion and continued for 1 minute. The rats were then decapitated and the brains rapidly removed and frozen. Both normoglycemic (n = 8) and hyperglycemic (n = 9) rats were studied.

Simultaneous tissue pH and [3H]nimodipine uptake measurements

A dual-label autoradiographic technique was developed to simultaneously determine regional tissue pH and [³H]nimodipine uptake during focal cerebral ischemia. In vivo uptake of [³H]nimodipine was measured as described by Hakim and Hogan (1991). Net tissue pH was calculated from the in vivo uptake of the volatile radiotracer [¹⁴C]DMO (Kobatake et al., 1984), which has been previously used in dual-label studies with a second nonvolatile radiotracer (Hakim and Arrieta, 1986).

Preliminary studies. To determine the optimal dosage and sublimation period for [¹⁴C]DMO a normal rat was infused intravenously with either 2 or 20 μCi of [¹⁴C]DMO (51 mCi/mmol, Amersham) 1 hour before decapitation or with 250 μCi of [³H]nimodipine (130 Ci/mmol, NEN-DuPont, Boston, MA, U.S.A.) 30 minutes before decapitation. The brains were rapidly removed and frozen, and 20-μm cryostat sections were apposed to both Kodak SB5 film (Picker International, Brampton, Ontario, Canada) and tritium-sensitive Hyperfilm (Amersham, Etobicoke, Ontario, Canada). To assess for potential loss of [¹⁴C]DMO during the prolonged exposure to the Kodak SB5 film 85 μCi of [¹⁴C]DMO was infused intravenously and allowed to circulate for 1 hour in a normoglycemic and a hyperglycemic rat undergoing 30 minutes of focal ischemia. Twenty-micrometer brain sections were obtained and repeatedly apposed to a new Kodak SB5 film at 4°C every 5 to 6 days to obtain nine sequential image sets of [¹⁴C]DMO distribution for 47 days.

[³H]Nimodipine uptake and tissue pH. An intravenous bolus of 4 μCi of [¹⁴C]DMO in 500 μL of normal saline was infused 30 minutes before MCA + CCA occlusion. Plasma [¹⁴C]DMO concentration was measured 30 minutes and 1 minute before study termination. Immediately after MCA + CCA occlusion, 200 μCi [³H]nimodipine in 600 μL of carrier (BAY e9736 placebo, Bayer, Etobicoke, Ontario, Canada) was infused intravenously for 3 minutes and allowed to circulate for 30 minutes so that an equilibrium of distribution between plasma and brain was approached (Hogan et al., 1991). Plasma [³H]nimodipine concentration was measured 1 minute before study termination. All rats were decapitated 30 minutes after MCA + CCA occlusion and the brains were rapidly removed and frozen. Activity of ¹⁴C and ³H in plasma was determined using dual-label liquid scintillation counting with appropriate quench correction verified with internal standards. Both normoglycemic (n = 12) and hyperglycemic (n = 12) rats were studied.

Autoradiography. The brains were warmed to −20°C and 20-μm cryostat sections were obtained every 140 μm. Sections containing [¹⁴C]IAP were apposed to Kodak SB5 film for 10 days. Sections containing [³H]nimodipine and [¹⁴C]DMO were apposed to Kodak SB5 film for 40 days at 4°C. [¹⁴C]DMO was allowed to sublimate from the sections for 50 days under a stream of air at room temperature and the sections were then apposed to Hyperfilm for 60 to 90 days. Brain tissue sections (20- and 40-μm thick) from a rat infused with 2 μCi [¹⁴C]DMO were used to ensure adequate sublimation of the ¹⁴C signal. Commercial ¹⁴C standards (Amersham, Oakville, Canada) were used to calibrate the ¹⁴C autoradiographs. Tritium autoradiographs were calibrated using 20-μm tissue sections of brain paste standards prepared from rat forebrain containing known concentrations of [³H]nimodipine. In a subset of normoglycemic (n = 6) and hyperglycemic (n = 5) studies [³H]nimodipine brain paste standards were also apposed to SB5 film to assess for possible ³H signal contributing to the [¹⁴C]DMO image. In this subset of studies paired measurements of ¹⁴C and ³H activity in a coronal section at the level of the MCA occlusion were obtained using images from both SB5 and Hyperfilm. These measurements were made in the region of ischemic cortex showing the greatest increase in [³H]nimodipine content and in the contralateral nonischemic cortex. The data were used to determine the contribution of ³H signal to the total signal on the [¹⁴C]DMO autoradiographs.

Image analysis

Autoradiographic images at four coronal levels in the brain were digitized using a microcomputer-based image analysis system (Imaging Research Inc., St. Catharines, Ontario, Canada) and regional radiotracer content was measured using a standardized region of interest (ROI) template (Fig. 1). At each coronal level ¹⁴C activity was averaged over three adjacent aligned images and tritium activity was measured on one aligned image. Regional CBF and net tissue pH were calculated from the autoradiographs on SB5 film using the measured activity within each ROI along with plasma radiotracer and pH measurements. To normalize measurements of brain tissue [³H]nimodipine, regional volumes of distribution were calculated as the ratio of tissue [³H]nimodipine content in each ROI to the plasma [³H]nimodipine concentration at study termination. (Gjedde and Wong, 1990) Previously determined [³H]nimodipine metabolite correction factors for plasma and brain were applied (Hogan et al., 1991).

OPEN IN VIEWER

Fig. 1.

The region of interest (ROI) and volume of interest (VOI) templates at four coronal levels in the rat brain are shown. The ROI template is on the left side of the sections with arbitrary numbering used to identify each ROI within a section. The ROI groupings used to define the six volumes of interest (VOI) are on the right side of the sections and are labeled at the bottom of the figure.

A K-means cluster analysis was performed on the average of the logarithm of the ratio of ischemic to nonischemic hemisphere CBF values for all ROI on two variables representing the normoglycemic and hyperglycemic groups. The clusters of ROI obtained were further subdivided to separate cortical and striatal ROI. The resulting set of ROI clusters, termed volumes of interest (VOI), represented a fixed grouping of ROI for all studies over which CBF, pH, and [³H]nimodipine volume of distribution ROI measurements were averaged. Differences between groups were analyzed by repeated measures analysis of variance (ANOVA) (BMDP Statistical Software, Inc, Los Angeles, CA, U.S.A.). To control the level of type 1 error during multiple within-group comparisons, the Greenhouse-Geisser adjusted degrees of freedom and Neumann-Keuls procedures were used. For between-group comparisons, we used the pooled ANOVA error and adjustment to the degrees of freedom (Howell, 1992).

To more closely examine the relationship between tissue pH and [³H]nimodipine tissue content in ischemic cortex, paired pH and [³H]nimodipine volume of distribution measurements were obtained using a template consisting of a grid of contiguous 0.09-mm² sampling regions. To obtain data over a wide range of severity of ischemia this template was applied within ROI 1, 2, and 3 within the ischemic cortex at coronal level 2. For each rat the slope of the linear fit to the paired [³H]nimodipine volume of distribution–pH measurements was determined. These were averaged for the normoglycemic and hyperglycemic groups and compared using a Student's t-test. To further assess differences in the relationship between [³H]nimodipine volume of distribution and tissue pH a subset of 10 measurement pairs was randomly selected from the set of paired measurements for each rat and analyzed using repeated measures multivariate ANOVA (SPSS Inc., Chicago, IL, USA).

All results are expressed as mean ± standard deviation.

Physiologic measurements

In the dual-label [¹⁴C]DMO – [³H]nimodipine studies average values for arterial blood gases in the normoglycemic group (pH = 7.40 ± 0.03, P_aco₂ = 39.5 ± 3.8 mm Hg; n = 12) did not differ from the hyperglycemic group (pH = 7.38 ± 0.03, P_aco₂ = 40.7 ± 2.5 mm Hg; n = 12). Average mean arterial blood pressure in the normoglycemic group was 84 ± 10 mm Hg and did not differ from 84 ± 6 mm Hg observed in the hyperglycemic group. Comparison of average arterial blood gases in the normoglycemic group of the [¹⁴C]IAP CBF studies (pH = 7.39 ± 0.03, P_aco₂ = 37.2 ± 4.0 mm Hg; n = 8) showed no difference from the hyperglycemic group (pH = 7.40 ± 0.03, P_aco₂ = 37.7 ± 2.5 mm Hg; n = 9). Similarly in these studies average mean arterial blood pressure in the normoglycemic group (94 ± 8 mm Hg) did not differ from the hyperglycemic group (94 ± 5 mm Hg). In all studies average P_ao₂ levels were greater than 100 mm Hg and did not differ between groups.

Plasma glucose levels before glucose infusion did not differ between normoglycemic and hyperglycemic groups in the dual-label [¹⁴C]DMO – [³H]nimodipine studies (6.2 ± 1.2 mmol/L; n = 24) or [¹⁴C]IAP CBF studies (6.8 ± 1.7 mmol/L; n = 17). Plasma glucose remained constant in the sham-perfused normoglycemic groups but increased to 17 ± 2.8 mmol/L (n = 12) in the dual-label studies and 16.4 ± 1.6 mmol/L (n = 9) in the CBF studies after glucose infusion (P < 0.01).

Cerebral blood flow measurements

Cluster analysis on the mean of the logarithm of the CBF ratio in the normoglycemic and hyperglycemic studies showed the within-cluster sum of squares to decline most rapidly up to four clusters. These four initial ROI clusters were further subdivided to separate cortical from striatal structures, resulting in six final VOI; cortex 1, cortex 2, cortex 3, cortex 4, striatum 1, and striatum 2, as shown in Fig. 1. Table 1 summarizes the CBF data from each VOI. Cerebral blood flow did not differ in corresponding VOI between normoglycemic and hyperglycemic rats. No significant regional differences in CBF were observed in the nonischemic hemisphere in either group. Compared with the contralateral hemisphere, CBF was significantly depressed in cortex 2, cortex 3, cortex 4, striatum 1, and striatum 2 in the ischemic hemisphere in both groups. Within the ischemic hemispheres CBF was significantly depressed in all VOI compared with cortex 1 (corresponding to the cingulate cortex). The lowest CBF values were observed in ischemic cortex 4, corresponding to the cortex just distal to the MCA occlusion site.

OPEN IN VIEWER

Table 1.

Average cerebral blood flow (mL/100g/min) after 30 minutes of focal cerebral ischemia for each volume of interest in the normoglycemic and hyperglycemic rats

Volume of interest	Normoglycemic group (n = 8)	Hyperglycemic group (n = 9)
Ischemic hemisphere
Cortex 1	146 ± 58	180 ± 67
Cortex 2	67 ± 23†‡	75 ± 30†§
Cortex 3	31 ± 22†‡	33 ± 21†§
Cortex 4	15 ± 16†‡	17 ± 15†§
Striatum 1	68 ± 46*‡	62 ± 55†§
Striatum 2	38 ± 49†‡	51 ± 52†§
Nonischemic hemisphere
All VOI	200 ± 55	214 ± 50

VOI, volume of interest. Values are mean ± standard deviation.

Differs from contralateral hemisphere volume of interest at *P < 0.05, †P <0.01.

Differs from ipsilateral cortex-1 VOI at ‡P < 0.05, §P < 0.01.

Dual-label methodology

Autoradiographs obtained by apposing 20-μm brain sections from a normal rat infused with 250 μCi [³H]nimodipine to SB5 film up to 60 days did not show any visible tritium signal. Sections obtained from a normal rat infused with either 2 or 20 μCi [¹⁴C]DMO that were exposed to the same film showed ¹⁴C activity; however, the signal was weak in sections from the rat infused with 2 μCi. After a 40- to 60-day sublimation period autoradiographs of 20- and 40-μm-thick sections obtained from a rat infused with 2 μCi [¹⁴C]DMO produced no visible signal after a 60- to 90-day exposure to Hyperfilm. However, similarly treated 20-μm sections from a rat infused with 20 μCi [¹⁴C]DMO resulted in a visible signal on Hyperfilm that averaged 14% of the nonischemic hemisphere [³H]nimodipine signal in two normoglycemic rats. Subsequent studies were performed using 4 μCi [¹⁴C]DMO to produce an adequate ¹⁴C signal for image analysis and negligible ¹⁴C activity on the tritium images after sublimation of [¹⁴C]DMO. Images calibrated using ¹⁴C standards in the studies in which both ¹⁴C and ³H standards were apposed to SB5 film demonstrated a tritium signal of 0.0015 ± 0.0002 (n = 11) nCi per actual nCi of ³H present. Thus, in the nonischemic cortex the percentage of activity measured in the SB5 image as a result of ³H was 3.6% ± 0.8% (n = 6) in the normoglycemic studies and 4.8% ± 1.2% (n = 5) in the hyperglycemic studies, resulting in an overestimation of tissue pH by 0.019 ± 0.005 and 0.027 ± 0.01 pH units in these groups. In the region of greatest [³H]nimodipine uptake in the ischemic hemisphere the activity resulting from ³H was 11.0% ± 2.7% in the normoglycemic group and 16.1% ± 5.3% in the hyperglycemic group. Net tissue pH in this region was overestimated by 0.065 ± 0.020 and 0.122 ± 0.048 pH units in these groups (P < 0.05). The two studies in which repeated measurements of [¹⁴C]DMO activity for 47 days were obtained demonstrated the average decrease in calculated tissue pH as a result of [¹⁴C]DMO loss to be 0.013 pH units in the normoglycemic study and 0.005 pH units in the hyperglycemic study.

[3H]Nimodipine–[14C]DMO uptake studies

Autoradiographs of [³H]nimodipine and [¹⁴C]DMO uptake at coronal levels 2 and 3 from a selected normoglycemic study are shown in Fig. 2. [³H]Nimodipine uptake was increased in the cortex and striatum of the ischemic hemisphere whereas [¹⁴C]DMO uptake was depressed in these regions indicating increased tissue acidosis.

OPEN IN VIEWER

Fig. 2.

Histologic and autoradiographic sections at coronal level 2 (left column) and coronal level 3 (right column) from a selected normoglycemic rat. The images in the second row are averages of three adjacent [¹⁴C]DMO autoradiographs at the two coronal levels. The [³H]nimodipine autoradiographs shown in the third row were obtained after sublimation of [¹⁴C]DMO. The top row shows images of cresyl violet staining on nearby adjacent tissue sections and serves as an anatomic guide.

pH measurements. Figure 3A summarizes regional pH measurements obtained from the normoglycemic group of studies. Similar data from the hyperglycemic group are shown in Fig. 3B. Within VOI in the nonischemic contralateral hemisphere pH values ranged from 6.91 ± 0.13 to 6.97 ± 0.10 in the normoglycemic group and from 6.83 ± 0.13 to 6.93 ± 0.08 in the hyperglycemic group, with no significant differences found either within or between groups. In the ischemic cortex of the normoglycemic group, net tissue pH was most depressed in cortex 3 (6.76 ± 0.11) and cortex 4 (6.80 ± 0.14) and was significantly lower than that observed in cortex 1 of the ischemic hemisphere (P < 0.05). In the hyperglycemic group pH was significantly depressed in ischemic hemisphere cortex 3 (6.57 ± 0.13), cortex 4 (6.57 ± 0.15), and striatum 2 (6.73 ± 0.14) compared with the corresponding contralateral hemisphere VOI (P < 0.01). Within the ischemic hemisphere, cortex 3 and cortex 4 pH levels were significantly lower than pH in cortex 1 (P < 0.01). Comparison of tissue pH between the normoglycemic and hyperglycemic groups showed significantly greater acidosis in the ischemic hemisphere VOI of cortex 2 (P < 0.05) and cortex 3 and cortex 4 (P < 0.001) of the hyperglycemic group.

OPEN IN VIEWER

Fig. 3.

Net tissue pH calculated from [¹⁴C]DMO uptake into each volume of interest in (A) normoglycemic (n = 12) and (B) hyperglycemic (n = 12) studies (mean ± standard deviation). pH differs from the ischemic hemisphere cortex 1 pH at ⁺P < 0.05 and ⁺⁺P < 0.01 and from the corresponding contralateral hemisphere VOI pH at ^++,××P < 0.01. pH differs between the normoglycemic and hyperglycemic groups at *P < 0.05 and **P < 0.001.

[ ³ H]Nimodipine uptake. In normoglycemic rats the volume of distribution of [³H]nimodipine ranged from 0.98 ± 0.16 to 1.14 ± 0.19 mL/g in nonischemic hemisphere VOI and increased to 1.74 ± 0.29 mL/g in ischemic hemisphere cortex 3 (P < 0.05). In hyperglycemic rats the volume of distribution ranged from 1.03 ± 0.15 to 1.20 ± 0.17 mL/g in the nonischemic hemisphere and was 1.42 ± 0.45 mL/g in cortex 3 of the ischemic hemisphere. To reduce between-subject variation in nimodipine volume of distribution data the ratio of [³H]nimodipine volume of distribution in the ischemic hemisphere relative to that in the nonischemic hemisphere ([³H]nimodipine uptake ratio) was calculated for each VOI and is summarized in Fig. 4 for both the normoglycemic and hyperglycemic groups. In the normoglycemic rats this ratio was increased in cortex 3 (P < 0.01) and cortex 4 (P < 0.05) compared with cortex 1 and striatum 2. In the hyperglycemic rats the ratio was highest in cortex 3 but this was not significantly different from the values observed in the other VOI. Comparing corresponding VOI in the two study groups demonstrated the uptake ratio to be greater in normoglycemic rats (1.83 ± 0.45) compared with hyperglycemic rats (1.40 ± 0.50) in cortex 3 (P < 0.05).

OPEN IN VIEWER

Fig. 4.

Ratio of [³H]nimodipine uptake in ischemic hemisphere volumes of interest (VOI) over corresponding nonischemic hemisphere VOI in normoglycemic (n = 12) and hyperglycemic (n = 12) studies (mean ± standard deviation). Within the normoglycemic studies this ratio differs from cortex 1 and striatum 2 values at ⁺P < 0.05 and ⁺⁺P < 0.01. Between the normoglycemic and hyperglycemic groups the ratio for cortex 3 differs at *P < 0.05.

An average of 138 ± 13 paired pH–[³H]nimodipine volume of distribution measurements were obtained from 0.09-mm² sampling areas in ROI 1, 2, and 3 (coronal level 2) in the ischemic hemisphere cortex in each rat and these data are shown in Fig. 5. The volume of distribution of [³H]nimodipine was averaged over 0.1 pH units from pH 6.3 to 7.1 and was observed to increase with decreasing pH in the normoglycemic group. The average slope of the linear fit between [³H]nimodipine volume of distribution and pH differed significantly between the two groups and was −3.5 ± 1.5 mL/g per pH unit in the normoglycemic rats and −0.9 ± 1.3 mL/g per pH unit in the hyperglycemic rats (P < 0.01). Multivariate repeated measures ANOVA on the randomly selected subset of paired dependent measures (volume of distribution and tissue pH) demonstrated a significant difference (P = 0.006) between the normoglycemic and hyperglycemic rats. Univariate ANOVA showed a significant effect for both pH (P = 0.033) and nimodipine volume of distribution (P = 0.008) between these two groups. A stepdown analysis using pH as a covariate (Tabachnick and Fidell, 1989) demonstrated a significant effect for [³H]nimodipine volume of distribution between the groups independent of pH change (P = 0.014). Thus the relationship between nimodipine uptake and net tissue pH in ischemic brain, as estimated by the slope of a linear fit, differs between normoglycemia and hyperglycemia. Furthermore, changes in nimodipine uptake and tissue pH contribute independently to the observed difference between these two groups.

OPEN IN VIEWER

Fig. 5.

[³H]Nimodipine volume of distribution (mL/g) plotted against net tissue pH data from normoglycemic (n = 12) and hyperglycemic (n = 12) rats. Data were obtained from multiple paired measurements of [³H]nimodipine and [¹⁴C]DMO content in contiguous 0.09-mm² sample regions in the cortex of the ischemic hemisphere at the level of the middle cerebral artery occlusion (coronal level 2). Within each study [³H]nimodipine volumes of distribution measurements are grouped in 0.1-pH unit bins and the mean value of each group then averaged across all studies to obtain the means and standard deviations shown.