Altered hippocampal arteriole structure and function in a rat model of preeclampsia: Potential role in impaired seizure-induced hyperemia

Abstract

We investigated the effect of experimental preeclampsia on hyperemia during seizure in the hippocampus and vascular function and structure of hippocampal arterioles using Sprague Dawley rats (n = 14/group) that were nonpregnant, pregnant (d20), or had experimental preeclampsia (induced by a high cholesterol diet d7–20). Hyperemia was measured via hydrogen clearance basally and during pentylenetetrazol-induced seizure (40–130 mg/kg i.v.). Reactivity of isolated and pressurized hippocampal arterioles to KCl, nitric oxide synthase inhibition with NG-nitro-L-arginine methyl ester and the nitric oxide donor sodium nitroprusside were investigated. Capillary density was quantified via immunohistochemistry. Cerebral blood flow increased during seizure vs. baseline in pregnant (118 ± 14 vs. 87 ± 9 mL/100 g/min; p < 0.05) and nonpregnant rats (106 ± 9 vs. 82 ± 9 mL/100 g/min; p < 0.05) but was unchanged in preeclamptic rats (79 ± 16 vs. 91 ± 4 mL/100 g/min; p > 0.05), suggesting impaired seizure-induced hyperemia in preeclampsia. Hippocampal arterioles from preeclamptic rats had less basal tone, and dilated less to 15 mM KCl (9 ± 8%) vs. pregnant (61 ± 27%) and nonpregnant rats (20 ± 11%). L-NAME had no effect on hippocampal arterioles in any group, but dilation to sodium nitroprusside was similar. Structurally, hippocampal arterioles from preeclamptic rats underwent inward hypotrophic remodeling and capillary rarefaction. Impaired seizure-induced hyperemia, vascular dysfunction, and limited vasodilatory reserve of hippocampal arterioles could potentiate hippocampal injury in preeclampsia especially during eclampsia.

Introduction

Preeclampsia (PE) is a hypertensive complication of pregnancy that involves widespread maternal endothelial dysfunction that can manifest with neurologic symptoms including persistent headache, cortical blindness, and de novo seizure (eclampsia).^1,2 PE and eclampsia are leading causes of maternal and fetal morbidity and mortality worldwide, with 40% of maternal deaths involving complications of the cerebral circulation.³ Eclamptic seizures usually remain isolated events, i.e. women who have eclampsia do not tend to develop a long-lasting seizure disorder such as epilepsy.⁴ However, acute memory deficits are common in women with eclampsia, with either retrograde or anterograde amnesia lasting hours to days after seizure.⁵ Additionally, formerly eclamptic women self-report cognitive impairment later in life and have increased incidence of cerebral white matter lesions.^6–9 Thus, eclamptic seizures appear to have both acute and long-lasting effects on the maternal brain. More specifically, that eclamptic seizures affect memory and cognition suggests that the hippocampus may be affected by seizure in PE.¹⁰

The hippocampus is a brain region susceptible to pathological insults including seizure and ischemia.^11,12 During seizure, cerebral blood flow (CBF) increases in response to increased neuronal metabolism.¹³ However, this hyperemic response has limits, and severe and prolonged seizures often result in uncoupling of CBF and metabolism such that the seizure focus is no longer supplied with sufficient oxygen and glucose, resulting in neuronal injury.^13–15 Interestingly, studies in rats using autoradiography and MRI report the hippocampus may be more susceptible to seizure-induced ischemic injury than regions of the cerebral cortex due to less effective hyperemia during seizure.^13,15–17 The hyperemic response to seizure depends upon the capacity of the cerebral vasculature to dilate, decrease vascular resistance, and increase CBF to match metabolic demands. Thus, pathological conditions in which the cerebral vasculature is affected could further impair the hyperemic response to seizure and potentiate ischemic injury.¹⁸

The CBF response to changes in neuronal activity involves dynamic communication between neurons, astrocytes, and the cerebral vasculature.¹⁹ Local control of blood flow appears to occur at the level of pre-capillary arterioles and upstream penetrating arterioles, making these vessels an intricate player in hyperemia during seizure.²⁰ However, if the function of cerebral arterioles is compromised under pathological conditions, such as hypertension, the hyperemic response to seizure could be compromised.²¹ In fact, chronic hypertension has been shown to impair functional hyperemia due to hypertension-induced endothelial dysfunction and vascular remodeling.^21,22

PE is a unique hypertensive state, and there is evidence that PE impairs functional hyperemia as well. A study using transcranial Doppler to measure CBF in the posterior cerebral artery during visual stimulation reported abnormal functional hyperemia in women with pregnancies complicated by PE.²³ One mechanism by which functional hyperemia may be adversely affected in PE is if the ability of arterioles to appropriately dilate in response to increased neuronal metabolic demand was impaired. How PE affects the hippocampal vasculature and hyperemia during seizure in the hippocampus remains unknown. However, if vascular dysfunction is present and the hyperemic response to seizure is impaired in PE, the hippocampus may be at even greater risk of seizure-induced injury.

In the current study, we measured local CBF in the CA3 region of the hippocampus before and during pentylenetetrazol (PTZ)-induced seizure in nonpregnant and pregnant rats and rats with experimental PE. We also isolated hippocampal arterioles (HippAs) that supply the CA3 region and studied them pressurized in vitro.²⁴ The response of HippAs to mediators of dilation involved in the hyperemic response were investigated, including extracellular K⁺ and nitric oxide (NO).²¹ In addition, the structure and biomechanical properties of HippAs were compared, as well as quantification of capillary density in the hippocampus. We hypothesized that CBF would increase during seizure in all groups, but that it would increase to a lesser degree in PE rats due to impaired dilation and structural remodeling of HippAs.

Materials and methods

Animals

All experiments were conducted using virgin, nonpregnant (Nonpreg) or timed-pregnant (Preg) Sprague Dawley rats between 14 and 16 weeks of age (Charles River, Canada). Pregnant rats were used late in gestation (day 20 of a 22-day gestation), a time point when eclampsia occurs most often.²⁵ Rats were housed singly with environmental enrichment in the University of Vermont Animal Care Facility, an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) accredited facility. Rats were maintained on a 12-h light/dark cycle and allowed access to food and water ad libitum. All procedures were approved by the Institutional Animal Care and Use Committee and conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. The investigator was not blinded to animal group during experiments, as this was not possible due to the presence of pregnancy. However, analyses were completed by an investigator that was blinded to group. All euthanasia was under isoflurane anesthesia according to NIH guidelines, and experiments were conducted and are reported in compliance with the Animal Research: Reporting in Vivo Experiments (ARRIVE) guidelines.

Rat model of PE and assessment of pregnancy outcome

Pregnancy is a state of hyperlipidemia that is augmented in PE and thought to contribute to maternal endothelial dysfunction and the pathogenesis of PE.²⁶ To model PE, we used an established model that involves maintaining pregnant rats on a high cholesterol diet. This model has previously shown to cause hyperlipidemia by increasing total plasma cholesterol, maternal endothelial dysfunction, increased blood pressure, and fetal growth restriction.^27,28 Briefly, pregnant rats were fed a high cholesterol diet (Prolab 3000 rat chow with 2% cholesterol and 0.5% sodium cholate; Scotts Distributing Inc., Hudson, NH, USA) on days 7–20 of gestation (PE). After euthanization, uterine horns were examined for total number of pups and any reabsorbed fetuses. To assess fetal growth restriction and placental disease, pups and placenta from a separate group of PE rats (n = 3; n = 41 pups and placentae) were removed, weighed individually and compared with control rats (n = 3; n = 44 pups and placentae).

Measurement of CBF via hydrogen clearance in the hippocampus

CBF in the hippocampus was determined under baseline conditions and during seizure using hydrogen clearance. Hydrogen clearance was chosen for CBF measurements because it allows for repeated measures of absolute local CBF.^29,30 Rats that were Nonpreg, Preg, or with PE (n = 6/group) were anesthetized initially with isoflurane (1–3% in oxygen) for intubation and instrumentation. Rats were mechanically ventilated to maintain blood gases and pH within normal physiological ranges (Supplementary Table 1). Body temperatures were monitored with a rectal thermometer and maintained with a heating pad at ∼37℃ throughout the experiment (Supplementary Table 1). Blood glucose levels were sampled using a FreeStyle Lite glucometer (Abbott Diabetes Care, Inc., Alameda, CA, USA). While in supine position, femoral arteries were cannulated to obtain blood samples for blood gas measurements and continuous arterial blood pressure measurements via a pressure transducer (BIOPAC Systems Inc., Goleta, CA, USA; Supplementary Table 1). Femoral veins were cannulated for administration of the anesthetic chloral hydrate, infusion of the paralytic vecuronium, and the chemoconvulsant PTZ.

After instrumentation, rats were placed in prone position, secured in a stereotaxic apparatus, and the scalp retraced and the skull exposed. Rats were tapered off isoflurane and anesthesia maintained by continuous intravenous infusion of chloral hydrate (50 mg/mL; 30 µL/min). Choral hydrate was used because it has less depression of neuronal function, and has previously been used in studies measuring seizure threshold of rats.^31–33 Through a burr hole, a 50 µm tip glass hydrogen microsensor (Unisense, Aarhus, Denmark) containing both the sensing anode and reference electrode was inserted into the CA3 region of the hippocampus (−4.0 mm posterior, 3.0 mm lateral, 3.5 mm ventral to bregma).³⁴ In preliminary studies, these coordinates were confirmed by coating a wire electrode with DiI to visualize placement into the CA3 region of the hippocampus. The hydrogen microsensor was calibrated daily, and the hydrogen current was sampled at 20 Hz and recorded using a Multimeter (Unisense, Aarhus, Denmark). After placement of the microsensor, rats inhaled 4% hydrogen gas until the hydrogen current reached a steady state, at which point tissue saturation was achieved. Hydrogen gas was turned off and tissue desaturation was recorded under baseline conditions. Hydrogen gas was then inhaled again until tissue saturation was once again achieved. Vecuronium was infused intravenously (0.05 mg/kg) to initiate paralysis, followed by intravenous infusion of PTZ to induce seizure. Hydrogen gas was then turned off and tissue desaturation was recorded during seizure. From tissue desaturation measurements, the half-life of hydrogen was calculated.^29,30 Rats were then euthanized under chloral hydrate anesthesia by decapitation. Vecuronium induces paralysis by blocking neuromuscular transmission, and in preliminary studies, we found vecuronium dramatically decreased blood pressure. Thus, to avoid the confounding hypotensive effect of vecuronium on CBF, dosing was carefully titrated such that there was no effect of vecuronium administration on blood pressure. Previously, we used EEG and determined PTZ-induced seizure threshold in Nonpreg and Preg rats to be 65 mg/kg and 37 mg/kg of PTZ, respectively.³³ In preliminary studies, we determined seizure threshold of PE rats to be 20 mg/kg of PTZ using the same methodology. Thus, to control for the differences in threshold to PTZ-induced seizure between groups, each group was administered a dose of PTZ twice that of their seizure threshold that was confirmed to elicit seizure detected via intrahippocampal EEG in preliminary studies. EEG was not used during CBF measurements in this study due to the use of a hydrogen microsensor that precluded the use of EEG electrodes simultaneously. Despite this limitation, we assumed that PTZ initiated seizure in all rats.

Isolation of hippocampal arterioles

Separate groups of rats (n = 8/group) were decapitated under deep isoflurane anesthesia (3% oxygen) and brains immediately removed and placed in cold, oxygenated artificial cerebrospinal fluid (aCSF). HippAs branching off the internal transverse arteries of the dorsal hippocampus were carefully dissected and studied, as they supply the CA3 region of the hippocampus (Figure 1(a)).²⁴ HippAs were isolated and pressurized in an arteriograph chamber (Living Systems Instrumentation, Burlington, VT, USA). For more information on the isolation of these arterioles, please see Supplementary Material. Figure 1(b) shows a representative wide-field image of an internal transverse artery that is cannulated and pressurized and the HippA that branches from it at 20× magnification. The boxed inset is the portion of the HippA where lumen diameter and wall thickness measurements were made via video microscopy.

Figure 1.

Anatomic location and photomicrograph of hippocampal arterioles (HippAs). (a) Illustration of the hippocampal vasculature showing the branches of the longitudinal hippocampal artery (Long hipp a.) that penetrate into the hippocampal cleft. Internal transverse hippocampal arteries (Int trans hipp a.) were dissected to access the hippocampal arterioles (boxed inset) supplying the CA3 region of the dorsal hippocampus. Reprinted from P. Coyle, Vascular Patterns of the Rat Hippocampal Formation, Exp Neurol, 1976; 52(3): 450, 1976, with permission from Elsevier. (b) Wide field image (20×) of an Int trans hipp a. from a nonpregnant rat secured to a glass cannula with a HippA branching off at a 90° angle. HippA was pressurized to 60 mmHg. The region of the HippA where lumen diameter and wall thickness were measured is delineated by the boxed inset.

Experimental protocol for isolated arterioles

HippAs were equilibrated at 10 mmHg for 1 h, after which intravascular pressure was increased to 120 mmHg in a stepwise manner to determine whether vessels developed spontaneous myogenic tone and to measure myogenic reactivity. Lumen diameter and wall thickness were recorded at each intravascular pressure. Pressure was then returned to 60 mmHg for the remainder of the experiment. To investigate the response of HippAs to some mediators of dilation, reactivity to various pharmacological agents was measured: extracellular KCl (3–60 mM); NG-nitro-L-arginine methyl ester (L-NAME, 10⁻³M), a non-specific nitric oxide synthase (NOS) inhibitor, sodium nitroprusside (SNP), a nitric oxide donor (10⁻⁸ to 10⁻⁵M), and NS309, a SK/IK channel agonist (10⁻⁸ to 10⁻⁵M). At the end of each experiment, aCSF was replaced with aCSF containing zero calcium, papaverine (10⁻⁴M) and diltiazem (10⁻⁵M) to fully relax the vascular smooth muscle, and passive structural measurements made within the pressure range of 5–120 mmHg.

Quantification of hippocampal capillary density

Separate groups of Nonpreg, Preg, and PE rats (n =6/group) were euthanized under isoflurane anesthesia (3% in O₂) and brains immediately removed. A 3-mm section (4–7 mm posterior Bregma) was taken of the posterior cerebral cortex and fixed in 10% buffered formalin at 4℃ overnight, then transferred to 0.1 M PBS and slices paraffin embedded. Capillary density of the hippocampus was determined by immunohistochemical staining for Collagen IV, a structural component of basement membrane,³⁵ and morphometric analysis done using standard procedures. For detailed information regarding the immunohistochemical protocol, please see Supplementary Material. For each brain section, six micrographs (two per region of the hippocampus) were captured using a Zeiss LSM 510 META confocal microscope at 25× magnification with oil immersion. Images were digitized and imported into image analysis software (Metamorph, Sunnyvale, CA, USA) for quantification of Collagen IV staining. Vessels that stained positive with Collagen IV that were less than 10 µm in diameter were considered capillaries. Capillaries were counted per square millimeter and averaged per group for the entire hippocampus, as well as for each hippocampal region. Images were randomized and a reviewer that was blinded to group quantified capillary density.

Drugs and solutions

Chloral hydrate, PTZ, KCl, L-NAME, SNP, NS309, and papaverine were purchased from Sigma Aldrich (St. Louis, MO, USA). Chloral hydrate and PTZ were made daily in sterile lactated Ringer’s solution. Vecuronium was purchased from the University of Vermont Medical Center Pharmacy Services (Burlington, VT, USA) and diluted in sterile lactated Ringer’s solution and used for 5 days. Diltiazem was purchased from MP Biomedicals (Santa Ana, CA, USA). Stock solutions of L-NAME, SNP, papaverine, and diltiazem were made weekly and stored at 4℃ until use. NS309 stock solution was aliquoted and stored at −20℃ until use. Please see Supplementary Material for details on buffers used in isolated arteriole experiments.

Data calculations and statistical analyses

The number of animals used in each experiment was justified by statistical power calculation based on our previous studies using similar methodology.^30,36 Results are presented as mean ± SEM. Differences between three groups were determined by one-way ANOVA with a post hoc Bonferroni’s test to correct for multiple comparisons, and CBF during seizure was compared with baseline within group using a parametric paired t-test. Differences were considered significant at p < 0.05. Please see Supplementary Material for all data calculations.

Results

Physiological parameters of Preg and PE rats

High cholesterol diet did not significantly increase maternal body weight compared with pregnant controls (459 ± 15 g vs. 439 ± 13 g, p > 0.05), however, both groups of pregnant rats weighed more than the Nonpreg group (391 ± 17 g), as expected. In addition, blood glucose levels were similar between groups (Supplementary Table 1). Blood pressures were similar between groups under chloral hydrate anesthesia (Supplementary Table 1); however, in our previous studies, we have reported increased conscious blood pressures in this rat model of experimental PE.^27,28 There was no difference in pregnancy outcome between Preg and PE rats, as the number of pups (13 ± 1 pups vs. 15 ± 1 pups; p > 0.05) and reabsorbed fetuses (2 ± 1 fetuses vs. 1 ± 1 fetuses; p > 0.05) were similar between groups. However, rats with PE had smaller pups compared with Preg rats (2.41 ± 0.03 g vs. 2.51 ± 0.03 g; p < 0.05) and smaller placentae (0.427 ± 0.010 g vs. 0.461 ± 0.007 g; p < 0.05), indicating fetal and placental growth restriction, as previously reported.²⁸

Seizure-induced changes in CBF in the CA3 region of the hippocampus

Figure 2(a) shows a representative original tracing of hydrogen desaturation during baseline and PTZ-induced seizure of a Nonpreg rat. During seizure, the rate of hydrogen clearance was faster than during the baseline measurement, and the time required to reach the half-life (t_1/2) was used to calculate CBF. Absolute CBF values during baseline and seizure of Nonpreg, Preg, and PE rats is presented in Figure 2(b). Baseline CBF in the CA3 region of the hippocampus was similar between groups. During seizure, CBF significantly increased in the hippocampus compared with baseline in both Nonpreg and Preg rats. Importantly, the hyperemic response to seizure was absent in PE rats (Figure 2(b)). The percent change of the seizure-induced increase in CBF was not different between Nonpreg and Preg rats (Figure 2(c)). However, there was a decrease in CBF in PE rats during seizure (Figure 2(c)). Two Nonpreg, three Preg, and three PE rats were excluded due to technical difficulties. A power calculation revealed 1 -β = 0.95, indicating the percent change in blood flow during seizure was sufficiently powered despite the decrease in sample size due to technical difficulties.

Figure 2.

Changes in cerebral blood flow (CBF) in the CA3 region of the hippocampus in response to seizure in nonpregnant (Nonpreg), pregnant (Preg), and preeclamptic (PE) rats. (a) Representative hydrogen (H₂) desaturation curves under baseline conditions and during seizure in a Nonpreg rat from which CBF was determined using the half-life of desaturation (t_1/2). (b) Graph showing baseline CBF and CBF during seizure in Nonpreg, Preg, and PE rats. CBF significantly increased during seizure compared with baseline in Nonpreg and Preg rats, but was unchanged during seizure in PE rats. (c) Percent change in CBF during seizure was positive and similar between Nonpreg and Preg rats; however, the change in CBF was negative in PE rats. *p < 0.05 vs. baseline by paired t-test; ^p < 0.05 by one-way ANOVA with post-hoc Bonferroni test.

Reactivity of isolated and pressurized HippAs

To investigate whether vascular dysfunction may have contributed to the decrease in seizure-induced hyperemia in PE, we studied the function of isolated HippAs that supply the CA3 region. Figure 3 shows active and passive diameters over the pressure range from 10 to 120 mmHg. HippAs from all groups developed spontaneous myogenic tone at pressures >40 mmHg. However, the difference between active and passive diameters at 60 mmHg was significantly less in HippAs from PE rats compared with Preg and Nonpreg rats (7.8 ± 1.5 µm vs. 18.3 ± 2.8 µm vs. 19.2 ± 1.5 µm; p < 0.001), mostly because of their structurally smaller lumens. Thus, the dilatory capacity of HippAs from PE rats was significantly less compared with the other groups. HippAs from two Nonpreg, one Preg, and two PE rats were excluded due to technical difficulties.

Figure 3.

Myogenic reactivity of hippocampal arterioles (HippAs) from nonpregnant (Nonpreg), pregnant (Preg), and preeclamptic (PE) rats. Pressure-diameter curves of lumen diameters in both the active (circles) and passive (fully relaxed; squares) states of HippAs from Nonpreg (a), Preg (b), and PE (c) rats. Active diameters of HippAs from all groups were smaller than passive diameters. However, the difference between active and passive diameters was considerably less in PE rats. *p < 0.05, **p < 0.01 vs. passive lumen diameter by t-test.

Small- and intermediate-conductance calcium-activated potassium (SK/IK) channels are present in the endothelium of pial and cerebral parenchymal arterioles.³⁷ Activation of SK/IK channels causes hyperpolarization and subsequent vasodilation.³⁷ However, whether these channels are present in HippAs is not known. Supplementary Figure 1 shows the dose response of HippAs to the SK/IK channel agonist NS309. HippAs from all groups dilated with activation of SK/IK channels in a dose-dependent manner, however, HippAs from PE rats dilated less than those from Nonpreg and Preg rats. Thus, SK/IK channels appear to be present in HippAs, however, activation with NS309 had a diminished dilatory effect in PE.

The response of HippAs to increased extracellular KCl

Figure 4(a) shows representative raw diameter tracings from one Nonpreg, Preg, and PE rat during elevations in extracellular KCl from 5 to 10 mM. HippAs dilated within the KCl range 5–15 mM in Preg and Nonpreg rats, however, this dilation was nearly absent in HippAs from PE rats (Figure 4(b)). When extracellular KCl was increased > 20 mM, vessels decreased in diameter, and constricted in a dose-dependent manner. HippAs from PE rats constricted more than arterioles from Nonpreg and Preg rats at 40 and 60 mM KCl (Figure 4(b)).

Figure 4.

Reactivity of hippocampal arterioles (HippAs) to elevated extracellular KCl. (a) Representative inner diameter traces from nonpregnant (Nonpreg; top panel), pregnant (Preg; middle panel), and preeclamptic (PE; bottom panel) rats in response to increased extracellular KCl concentrations from 5 to 10 mM. Diameters increased with increasing KCl concentration in HippAs from Nonpreg and Preg rats, but remained unchanged in the HippA from the PE rat. (b) Graph showing the percent change in vessel diameter from baseline of HippAs from Nonpreg, Preg, and PE rats in response to elevated extracellular KCl. HippAs dilated in a dose dependent manner to elevations in KCl up to 15 mM. There was a minimal dilation of HippAs from PE rats. At KCl concentrations greater than 20 mM, arterioles constricted in a dose-dependent manner with the greatest constriction occurring in HippAs from PE rats at 40 and 60 mM KCl. ^p < 0.05 vs. Preg; *p < 0.05 vs. Nonpreg and Preg by one-way ANOVA with a post-hoc Bonferroni test.

The effect of NO in HippAs

Figure 5(a) shows HippA diameters at baseline and after treatment with a single high dose of the NOS inhibitor L-NAME. Diameters remained unchanged with NOS inhibition in HippAs from all groups. However, all HippAs dilated to SNP, a NO donor, in a dose-dependent manner, as seen in Figure 5(b).

Figure 5.

Reactivity of hippocampal arterioles (HippAs) to nitric oxide (NO). (a) Lumen diameters of HippAs from nonpregnant (Nonpreg), pregnant (Preg), and preeclamptic (PE) rats at baseline and after 20-min incubation with the NOS inhibitor L-NAME (10⁻³M). Diameters remained unchanged after NOS inhibition in all groups. (b) Reactivity of HippAs to the NO-donor SNP. All HippAs dilated in a dose-dependent manner.

Structural and biomechanical properties of HippAs

Passive inner diameters (Figure 6(a)) and outer diameters (Figure 6(b)) of HippAs from PE rats were significantly smaller than HippAs from Preg and Nonpreg rats. Outer diameters of HippAs from Preg rats were also smaller than Nonpreg rats, although this did not reach statistical significance (Figure 6(b)). In addition, the vascular wall of HippAs from both Preg and PE rats were significantly thinner than those from Nonpreg rats (Figure 6(c)). Cross-sectional area and wall tension of HippAs from PE rats were significantly less than Preg and Nonpreg rats at 5 and 60 mmHg, but wall stress was similar between groups (Supplementary Table 2). Wall stress–strain curves were calculated to investigate changes in vessel stiffness in pregnancy and PE and are shown in Figure 6(d). Compared with Preg and Nonpreg controls, HippAs from PE rats were stiffer, demonstrated by the leftward shift in the stress–strain curve. HippAs from Preg rats were also stiffer than those from Nonpreg rats, indicated by an intermediary leftward shift in the stress–strain curve (Figure 6(d)). In addition, arterioles from PE rats were less distensible at 60 mmHg than HippAs from Preg and Nonpreg rats (49.6 ± 5.8% vs. 80.6 ± 8.6% vs. 81.6 ± 8.2%; p < 0.05). Passive structural measurements were unable to be obtained from one arteriole from a Preg rat due to technical difficulties.

Figure 6.

Structural and biomechanical properties of hippocampal arterioles (HippAs). (a) Inner diameters of fully relaxed HippAs from Nonpregnant (Nonpreg), pregnant (Preg), and preeclamptic (PE) rats. Diameters were smaller in HippAs from PE rats compared with Nonpreg rats. (b) Outer diameters of fully relaxed HippAs were smaller in PE rats than Nonpreg rats. (c) Wall thickness was decreased in HippAs from both PE and Preg rats. (d) Stress–strain curves of HippAs show a leftward shift in the stress–strain curve of HippAs from PE rats, with an intermediary left shift in HippAs from Preg rats. *p < 0.05, **p < 0.01 vs. Nonpreg; ^^p < 0.01 vs. Nonpreg and Preg by one-way ANOVA with a post-hoc Bonferroni test.

Capillary density in the hippocampus

Figure 7(a) shows representative images of the Collagen IV⁺ vessels in the CA3 region of the hippocampus from a Nonpreg, Preg, and PE rat. There was a significant decrease in capillary density in CA3 of the hippocampus in PE rats compared with Nonpreg rats (Figure 7(b)), with no differences in capillary density between groups in the dentate gyrus (Figure 7(c)) or CA1 region of the hippocampus (Figure 7(d)).

Figure 7.

Capillary density in the hippocampus of nonpregnant (Nonpreg), pregnant (Preg), and preeclamptic (PE) rats. (a) Representative photomicrographs of collagen IV⁺ vessels within the CA3 region of the hippocampus of a Nonpreg (left), Preg (middle), and PE (right) rat. (b) Graph showing capillary density in the CA3 region of the hippocampus was decreased in PE rats compared with Nonpreg rats. Capillary density was similar between groups in the dentate gyrus (c) and CA1 region (d). *p < 0.05 vs. Nonpreg by one-way ANOVA with a post-hoc Bonferroni test.

Discussion

The main finding of the current study was the lack of seizure-induced increase in CBF in the CA3 region of the hippocampus in a model of PE, suggesting that CBF and neuronal metabolism were uncoupled during seizure in experimental PE. In addition, we showed isolated and pressurized HippAs from PE rats dilated minimally to increased extracellular KCl compared with arterioles from Preg and Nonpreg rats that could contribute to the lack of seizure-induced increase in CBF. Additionally, HippAs from PE rats underwent inward, hypotrophic remodeling and were stiffer than the nonpregnant and pregnant states that significantly limited the vasodilatory reserve of these arterioles. Finally, decreased capillary density in CA3 of PE rats suggests capillary rarefaction occurred in this region of the hippocampus in experimental PE. Collectively, these vascular changes at the arteriolar and capillary levels may contribute to the lack of increase in CBF during seizure in PE.

Functional hyperemia, the increase in CBF evoked by neuronal activity, has been shown to be impaired in several pathological conditions including PE.^18,23 However, whether seizure-induced hyperemia is impaired in PE is not known. The current study is the first we are aware of to report impaired hyperemia during seizure in the hippocampus in experimental PE. The findings of the current study suggest impaired dilation and remodeling of the arterioles supplying the hippocampus in PE rats could limit the hyperemic response and predispose the hippocampus to ischemic injury during seizure. Previous studies have assessed functional hyperemia in women within 6 years of a pregnancy complicated by PE and/or eclampsia. A study using transcranial Doppler ultrasonography of the posterior cerebral artery during a visual stimulation task reported that abnormal functional hyperemia occurred 3 times more often in formerly PE and eclamptic patients than in women with uncomplicated pregnancies.²³ Abnormal functional hyperemia was not predictive of PE, as there was no association between changes in neurovascular coupling at 25–28 weeks of gestation and the occurrence of PE.³⁸ This finding suggests a potential causative role of PE in impaired functional hyperemia. It is possible that similar mechanisms underlie the impaired seizure-induced hyperemic response measured in the current study in experimental PE. In addition, both impaired functional- and seizure-induced hyperemia could increase the risk of ischemic brain injury in PE.

The hippocampus is particularly susceptible to ischemic injury including during seizure.^11,12 In fact, it has been suggested that injury to the hippocampus during seizure is due to less effective hyperemia compared with other brain regions.^15,16 In a study measuring regional CBF using autoradiography and glucose utilization in the whole brain of rats during kainic acid-induced limbic seizures, the most dramatic uncoupling of CBF and metabolism occurred in the CA3 region of the hippocampus that was associated with greater neuronal injury.¹⁶ Another study utilized MRI with arterial spin labeling to measure CBF in the hippocampus and parietal cortex of rats during pilocarpine-induced status epilepticus.¹⁵ In that study, there was less of an increase in CBF in the hippocampus in the early stages of seizure onset that, remarkably, was followed by a decrease in CBF below baseline levels at later stages of seizure that did not occur in the cortex.¹⁵ Further, studies using blood oxygen level-dependent (BOLD) functional MRI during bicuculline-induced seizures in rats report positive BOLD signal in the cerebral cortex and thalamus, but negative BOLD signal in the hippocampus, suggesting a mismatch in neuronal activity and CBF in the hippocampus during seizure.^39,40 In the current study, there were no changes in basal blood flow between groups, despite vascular remodeling of HippAs from PE rats. It is likely that this remodeling occurred in response to elevated blood pressure in these rats, thereby maintaining CBF in the face of higher blood pressure. However, we showed that seizure-induced hyperemia was absent in the hippocampus of PE rats, suggesting that similar to the previous studies, the hippocampus may be at even greater risk of seizure-induced injury in PE.

Pial arteries and cerebral parenchymal arterioles dilate to KCl concentrations up to ∼15 mM due to activation of inward rectifier K⁺ (K_IR) channels and subsequent vascular smooth muscle hyperpolarization.^41,42 At the immediate onset of seizure, extracellular K⁺ concentrations in the seizure domain rise from 3 mM to 8–12 mM, which is within the range that activates K_IR channels.⁴³ K_IR channels have been suggested to be a mechanism by which increased neuronal activity is coupled to changes in local blood flow by dilating arterioles in response to extracellular K⁺ accumulation.⁴⁴ In parenchymal arterioles within the cerebral cortex, 15 mM extracellular KCl caused near maximal dilation;^42,45 however, in the current study, HippAs from PE rats responded very little to extracellular KCl within the range that activates K_IR channels. The mechanism by which HippAs have impaired dilation to increased extracellular K⁺ is not clear. However, the K_IR-dependent dilation of cerebral parenchymal arterioles and pial arteries has been reported to be impaired in hypertensive rats.^45,46 Therefore, the presence of hypertension in PE and/or circulating factors that promote endothelial dysfunction in PE may underlie the decreased K_IR channel response. Further, the structural remodeling of HippAs in PE rats significantly decreased their dilatory capacity that likely resulted in sustained high vascular resistance during seizure, limiting the capacity of CBF to increase to the hippocampus.⁴⁷ Thus, the decreased dilation to K⁺ in conjunction with vascular remodeling may impair the hyperemic response to seizure in PE.

In the current study, HippAs did not constrict to a high concentration of L-NAME, suggesting they did not have basal production of NO to inhibit tone. This finding was surprising and appears to be a unique property of this vascular segment compared with parenchymal arterioles supplying the cerebral cortex. Parenchymal arterioles in the cerebral cortex have a substantial contribution of basal NO production that opposes tone.^48,49 In addition, administration of acetylcholine to HippAs to activate eNOS and stimulate NO production⁵⁰ had no effect on vessel diameter (data not shown). However, while HippAs dilated to SNP, the dilation was less than in reports of other cerebrovascular segments.⁵¹ The finding that NOS inhibition appears to have no effect on arteriolar diameter in the current study is in contrast to studies measuring the effect of NOS inhibition on hippocampal microvessel diameter in hippocampal slices. Fergus et al. report administration of the NOS inhibitor L-NNA constricted microvessels ∼60% of their resting diameter (9–23 µm); however, the source of NO was not identified due to the use of a nonspecific NOS inhibitor.⁵² In fact, studies using fluorescent imaging of NO report nitrergic perivascular nerves in close proximity to the smooth muscle of intraparenchymal arterioles in the hippocampus that exert a tonic dilator influence on arterioles.^53,54 Thus, vessel diameter may be preferentially regulated by nNOS in the hippocampus, rather than eNOS as in the cerebral cortex. The lack of basal NO production by eNOS in HippAs may be a mechanism to maintain tight control of NO-mediated dilatory responses in the hippocampus, a concept that is supported by many in vivo studies that have shown functional- and seizure-induced hyperemia in the hippocampus are dependent upon nNOS-derived NO.^55–57

This is the first investigation that we are aware of to study HippAs isolated and pressurized. Previously, studies have measured the dilatory response of hippocampal microvessels to numerous neurotransmitters and endogenous signaling molecules including NO, calcitonin gene-related peptide, N-methyl-D-aspartate (NMDA), and gamma-aminobutyric acid (GABA) using a hippocampal slice model.^52,53,55 In the current study, we investigated PE-induced vascular changes in the hippocampus that could contribute to impaired hyperemia during seizure by isolating HippAs and studying their function at physiological pressures, without neuronal and glial influences that would be present in a slice preparation. While this in vitro methodology allows study of the arteriole-specific response to mediators of dilation involved in coupling blood flow and metabolism, we recognize the limitation of this approach, as neurons, glia, and microvessels are intimately associated and form the neurovascular unit in vivo.²¹ Additionally, while we did not measure neuronal activation during seizure that may be different in PE, we did measure reactivity to K⁺ and NO that are thought to be major dilatory signals that increase CBF during seizure.^19,21,55 Further, we controlled for differences in seizure severity that could confound the CBF response to seizure by accounting for pregnancy and PE-induced changes in seizure threshold in dosing of PTZ. Future studies are required to determine changes in neuronal or glial contributions to impaired hyperemia during seizure in PE, in which a slice model may be useful.

In conclusion, PE appears to have a detrimental effect on seizure-induced hyperemia in the hippocampus that may contribute to both acute and long-lasting cognitive impairment reported to occur in formerly eclamptic women.^5,6 Uncoupling of CBF and neuronal metabolism can result from changes in vascular dynamics to biochemical messengers coordinating blood flow and energy demand.¹⁸ Increased extracellular K⁺ and neuronal-derived NO production occurs during neuronal activation and to a greater extent during seizures that are considered critical components of dilating nearby arterioles to match local blood flow to neuronal metabolism.^19,55 Interestingly, cerebral white matter lesions are more abundant in formerly PE and eclamptic women.⁸ Thus, PE-induced cerebrovascular dysfunction and impaired functional hyperemia detected 6 years post-pregnancy suggests the vascular impairment is persistent.²³ This could explain why white matter lesions are present in PE women, even in the absence eclamptic seizure.⁸ The hippocampus is central to higher-order cognitive function, and understanding the function of HippAs under physiological and pathophysiological conditions may be important as several debilitating neurological diseases involve vascular-mediated hippocampal injury, including stroke, vascular dementia, and PE/eclampsia.^6,58,59

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: National Institutes of Health (NIH) National Institute of Neurological Disorders and Stroke (NINDS) R01 NS045940, the Preeclampsia Foundation, the Cardiovascular Research Institute of Vermont, and the Totman Medical Research Trust.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

ACJ and MJC made a substantial contribution to the concept and design, acquisition of data, or analysis and interpretation of data. ACJ and MJC drafted the article or revised it critically for important intellectual content, and approved the version to be published.

Supplementary material

Supplementary material for this paper can be found at http://jcbfm.sagepub.com/content/by/supplemental-data