Memory impairment in spontaneously hypertensive rats is associated with hippocampal hypoperfusion and hippocampal vascular dysfunction

Abstract

We investigated the effect of chronic hypertension on hippocampal arterioles (HippAs) and hippocampal perfusion as underlying mechanisms of memory impairment, and how large artery stiffness relates to HippA remodeling. Using male spontaneously hypertensive rats (SHR) and normotensive Wistar rats (n = 12/group), long-term (LTM) and spatial memory were tested using object recognition and spontaneous alternation tasks. Hippocampal blood flow was measured via hydrogen clearance basally and during hypercapnia. Reactivity of isolated and pressurized HippAs to pressure and pharmacological activators and inhibitors was investigated. To determine large artery stiffness, distensibility and elastin content were measured in thoracic aorta. SHR had impaired LTM and spatial memory associated with decreased basal blood flow (68 ± 12 mL/100 g/min) vs. Wistar (111 ± 28 mL/100 g/min, p < 0.01) that increased during hypercapnia similarly between groups. Compared to Wistar, HippAs from SHR had increased tone at 60 mmHg (58 ± 9% vs. 37 ± 7%, p < 0.01), and decreased reactivity to small- and intermediate-conductance calcium-activated potassium (SK/IK) channel activation. HippAs in both groups were unaffected by NOS inhibition. Decreased elastin content correlated with increased stiffness in aorta of SHR that was associated with increased stiffness and hypertrophic remodeling of HippAs. Hippocampal vascular dysfunction during hypertension could potentiate memory deficits and may provide a therapeutic target to limit vascular cognitive impairment.

Introduction

The American Heart Association currently estimates there are >100 million adults in the United States that have hypertension.¹ Hypertension is a major risk factor for cerebrovascular disease, stroke, and vascular cognitive impairment.^2,3 Although vascular cognitive impairment and dementia are a result of specific vascular events such as ischemic stroke, hypertensive patients also exhibit cognitive impairment prior to and independently of stroke.^3,4 A prospective, longitudinal study across the adult age span recently reported that middle-aged participants with hypertension demonstrated greater cognitive decline than normotensive participants, including a faster rate of deterioration in memory function.⁵ Thus, chronic hypertension may affect brain regions involved in cognition and memory such as the hippocampus in a progressive manner that contributes to cognitive decline.

The hippocampus is a cognition-centric brain region particularly susceptible to pathological insults such as ischemia.^6,7 Hippocampal neurons have high metabolic demands that require very tightly regulated delivery of glucose and oxygen, making local cerebral blood flow (CBF) critical to neuronal health.^8,9 Interestingly, the hippocampus is less effective than the cerebral cortex at maintaining CBF during sustained metabolic demand such as seizure, an effect that could potentiate ischemic injury.^10–12 We recently reported in a model of preeclampsia – a hypertensive disorder of pregnancy – that an impaired hyperemic response to seizure occurred in the hippocampus that was associated with smaller, stiffer hippocampal arterioles (HippAs).¹³ HippAs are critical for maintenance of hippocampal neuronal homeostasis that if affected in other hypertensive disorders could lead to decreased perfusion. Importantly, the hippocampal vasculature appears to be structurally and functionally distinct from the vasculature supplying the cerebral cortex and may respond to chronic hypertension in a different manner. While there is a breadth of knowledge regarding hippocampal neuronal network function as it relates to learning and memory formation, relatively little is known about the arterioles supplying this brain region that is critical to higher order cognitive function and involved in several neurological diseases.

In the current study, we determined hippocampal-dependent memory function in adult spontaneously hypertensive rats (SHR) and normotensive Wistar rats and assessed HippA function in vivo and in vitro. We measured hippocampal CBF and studied HippAs isolated and pressurized in vitro to investigate potential mechanisms by which hypertension may cause hippocampal vascular dysfunction and ultimately affect hippocampal perfusion and memory function. Vascular responses to mediators of neurovascular coupling were investigated, including activation of small- and intermediate-conductance calcium-activated (SK/IK) channels and nitric oxide (NO). We further compared the response of HippAs to that of the aorta that undergoes adaptive increases in aortic stiffness and is clinically associated with increased risk of stroke and cognitive decline.^14–16 Hypertension-induced aortic stiffness increases transmission of pulse-wave velocity (PWV) more deeply into the brain causing microvascular injury.¹⁴ However, the effect of chronic hypertension-induced aortic stiffness on hippocampal perfusion and the hippocampal microvasculature that could directly affect memory function is largely unknown. We hypothesized that SHR would demonstrate impaired hippocampal-dependent memory function that would be associated with reduced hippocampal blood flow, HippA dysfunction and remodeling, and increased aortic stiffness.

Materials and methods

Animals

All experiments were conducted using male Wistar rats (14–16 weeks old) or male SHR (26–30 weeks old) that were purchased from Charles River, Canada. Male rats were used because the incidence for vascular dementia has been reported to be higher in men than women.^17–19 SHR are a widely used animal model to investigate the effect of chronic hypertension on cognitive decline, as they share similar neuropathology of vascular cognitive impairment and dementia that occurs in humans.²⁰ Thus, this rodent model is useful in the investigation of the effect of chronic hypertension on hippocampal blood flow and vascular function. Wistar rats were used as normotensive controls rather than Wistar-Kyoto (WKY) rats that are the genetic controls for SHR because WKY rats are an established model of depression and anxiety-like behavior.^21–23 Thus, investigation of hippocampal vascular function and perfusion as it relates to memory could be confounded by such depressive and anxiety-like behavior demonstrated by WKY rats. A pilot study was conducted to determine the optimal time to investigate hippocampal perfusion and vascular function as it may relate to cognition in SHR. Long-term memory and spatial memory were tested at 16, 20, and 26–28 weeks' age in separate groups of SHR. SHR at 26–28 weeks' age demonstrated deficits in both memory types. Therefore, hippocampal perfusion and vascular function were investigated in SHR ≥ 26 weeks of age when both long-term memory and spatial memory were known to be impaired. As Wistar rats do not demonstrate age-related cognitive decline until > 52 weeks of age or older,²⁴ the groups were not specifically age-matched, but rather both studied in adulthood of middle age.

Rats were housed in pairs 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 allowed to acclimate to the animal facility for at least five days prior to handling, and were maintained on a 12-h light/dark cycle and allowed access to food and water ad libitum. All procedures occurred at the same time of day, and 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. Investigators were blinded to animal group during behavioral and blood flow analyses. All euthanasia was under either isoflurane or chloral hydrate anesthesia according to NIH guidelines, and experiments were conducted and are reported in compliance with the Animal Research: Reporting in Vivo Experiments (ARRIVE) guidelines.

Behavioral tests of memory function

To determine the effect of chronic hypertension on hippocampal-dependent cognitive function, long-term memory and spontaneous alternation as a test of spatial memory were tested using a novel object recognition (NOR) task and continuous Y maze task, respectively.^25,26 Groups of SHR and Wistar rats (n = 6/group) acclimated to the room in which behavioral testing took place for 1 h prior to behavioral testing. For the NOR task, each rat was allowed to habituate to the open field arena for 5 min. The acquisition phase consisted of each rat being placed in the arena with two identical objects and allowed to explore for 10 min, an exposure time of sufficient duration to induce, and therefore test, long-term memory.²⁷ After the acquisition phase, rats were returned to their home cages. Forty-eight hours later, rats were placed in the same arena containing one familiar and one novel object. Automated live tracking software (ANY-maze, Stoelting Co., Wood Dale, IL, USA) was used to quantify baseline locomotion and exploratory behavior during the habituation phase, as well as the time that each rat spent investigating each object during acquisition and testing phases. The time spent investigating both the novel and familiar objects was used to calculate a recognition index as a measure of long-term memory function. A preliminary study was conducted using separate Wistar rats (n = 4) that, after habituation, were placed in the arena with the four objects used in this study and the time spent investigating each object was quantified to confirm no innate object preference (Supplementary Figure 1).

For the continuous Y maze task, rats were allowed to freely explore a Y maze for 8 min. Y maze tasks were video recorded to allow for quantification of spontaneous alternation behavior (SAB) and total arm entries that were used to calculate an alternation index as a measure of spatial memory.²⁶ All videos were analyzed by a reviewer that was blinded to group. For more details on objects, apparatus, and methods to avoid innate bias, please see Supplementary Material.

Measurement of CBF via hydrogen clearance in the hippocampus

CBF in the hippocampus was determined under baseline conditions and during hypercapnia using hydrogen (H₂) clearance as previously described with modifications.¹³ H₂ clearance was chosen for CBF measurements because it allows for repeated measures of absolute local CBF.²⁸ Rats that were Wistar or SHR (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 and to induce hypercapnia. Body temperature was maintained at 37.0 ± 0.2℃ with a heating pad throughout the experiment. 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). Femoral veins were cannulated for administration of the anesthetic chloral hydrate and infusion of the paralytic vecuronium.

After instrumentation, rats were placed in prone position, secured in a stereotaxic apparatus, and the scalp retracted and the skull exposed. Rats were tapered off isoflurane and anesthesia maintained by continuous infusion of chloral hydrate (50 mg/mL; 30 µL/min i.v.). Choral hydrate was used because it is less vasoactive than isoflurane and has previously been used in studies measuring CBF via H₂ clearance in rats during normo- and hypercapnia.^13,29–31 Through a 2 mm burr hole, a 50 µm glass tip H₂ microsensor (Unisense, Aarhus, Denmark) containing both the sensing anode and reference electrode was inserted into the CA1 region of the dorsal hippocampus (−3.5 mm posterior, 3.0 mm lateral, 2.0 mm ventral to Bregma).³² In preliminary studies, these coordinates were confirmed visually by assessing the electrode tract running to the CA1 region of the hippocampus. A limitation of the H₂ clearance method is its invasiveness that can cause local tissue damage that could confound CBF measurements if care not taken to accurately place the microsensor upon first insertion. Thus, microsensor placement was done with precision to avoid additional local tissue damage. If placement was inaccurate and multiple probe insertions were required to measure CBF, the animal was excluded from analysis. The CA1 region was chosen because of its high susceptibility to ischemic injury and its prominent role in learning and memory.^7,33–35 The H₂ microsensor was calibrated daily, and the H₂ current was sampled at 5 Hz and recorded using a Multimeter (Unisense, Aarhus, Denmark). After placement of the microsensor, rats inhaled 4% H₂ gas until the H₂ current reached a steady state, at which point tissue saturation was achieved. H₂ gas was turned off and tissue desaturation was recorded under baseline conditions.

To induce hypercapnia, the ventilation of rats was adjusted to increase pCO₂ by controlled hypoventilation. The paralytic vecuronium (0.05 mg/kg) was infused i.v. to initiate diaphragmatic paralysis, then the breathing rate decreased to 30–35 breaths/min and inhalation of air increased from 1% to 4%. Oxygen inhalation was also increased from 0.1% to 0.4% to maintain pO₂ levels within the physiological range. Hypercapnic conditions were achieved when pCO₂ was 60–65 mmHg for 10 min. H₂ clearance was then measured during hypercapnia as described above, and the half-life of H₂ was obtained from tissue desaturation measurements and used to calculate absolute CBF.²⁸

Experimental protocol for isolated HippAs

To understand the potential role of the hippocampal vasculature in hypertension-induced hippocampal dysfunction, HippAs supplying CA1 in the dorsal hippocampus were isolated and studied in a pressurized arteriograph system, as previously described.¹³ Briefly, rats that underwent behavioral testing (n = 6/group) were decapitated under deep isoflurane anesthesia (3% in 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, mounted on glass cannulas and pressurized in an arteriograph chamber (Living Systems Instrumentation, Burlington, VT, USA).¹³

HippAs were equilibrated at 20 mmHg for 1 h, after which intravascular pressure was increased to 120 mmHg in a stepwise manner to determine if 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 40 mmHg for HippAs from Wistar rats and 60 mmHg for HippAs from SHR for the remainder of the experiment. To investigate the response of HippAs to some mediators of cerebrovascular function, reactivity to various pharmacological agents was measured: NS309 (10⁻⁸–10⁻⁵ M), a SK/IK channel agonist; TM5441 (10⁻⁸–10⁻⁵ M), a plasminogen activator inhibitor-1 (PAI-1) inhibitor; 1-(2-trifluoromethylphenyl) imidazole (TRIM, 3 × 10⁻⁴ M), a neuronal NO synthase (nNOS) inhibitor; NG-nitro-L-arginine methyl ester (L-NAME, 10⁻³ M), a non-specific NOS inhibitor; sodium nitroprusside (SNP, 10⁻⁸–10⁻⁵ M), a NO donor. At the end of each experiment, aCSF was replaced with aCSF containing zero calcium, papaverine (10⁻⁴ M) and diltiazem (10⁻⁵ M) to fully relax the vascular smooth muscle (VSM), and passive structural measurements made within the pressure range of 5–120 mmHg.

Experimental protocol for isolated thoracic aorta

To understand the relationship between hypertension-induced aortic stiffness and hippocampal vascular dysfunction, thoracic aortic segments were isolated from the same animals as HippAs and studied pressurized in an arteriograph chamber, modified for large arteries. After decapitation under isoflurane anesthesia, a thoracotomy was performed and entire thoracic aorta quickly dissected and placed in ice-cold physiological saline solution (PSS). Adipose tissue and nerve plexus were carefully dissected from the aortic wall. The aortic arch was removed and discarded. The remaining aortic segments spanning intercostal arteries one to five were mounted on grooved, blunt-ended 14 gauge metal cannulas in an arteriograph chamber (Living Systems Instrumentation, Burlington, VT, USA) modified for study of large conduit arteries. The proximal and distal ends were secured on cannulas by 4–0 braided silk sutures and intercostal arteries tied off such that aortic segments could be studied pressurized. Aortas were equilibrated at 120 mmHg for 30 min in PSS containing zero calcium, papaverine (10⁻⁴ M) and diltiazem (10⁻⁵ M) to fully relax aortic segments. Due to the robust vascular wall of aorta, segments were not able to be trans-illuminated and therefore lumen diameter and wall thickness were not distinguishable. However, outer diameters were measured under fully relaxed conditions within the intravascular pressure range of 5–200 mmHg.

Histological assessment of elastin content of thoracic aorta

The effect of chronic hypertension on elastin content of thoracic aorta was investigated histologically using Verhoeff-Van Gieson (VVG) staining of cross-sections of aortic rings adjacent to segments used for isolated vessel experiments. Rings were immersion fixed in 10% buffered formalin for 24 h and paraffin-embedded along the longitudinal axis. VVG staining was performed using standard techniques on paraffin sections. Two sections that were 5 µm thick were taken 200 µm apart from each aortic ring and stained for elastin. Four images were captured at 40 × magnification from each ring using an Olympus BX50 microscope. As VVG stains elastin fibers black, elastin content was quantified by calculating the mean grey scale value of the medial layer of the aortic wall using ImageJ (NIH, Bethesda, MD, USA). Mean grey scale values were inverted and averaged within animal and group, then multiplied by 10³ power. Images were analyzed by a Reviewer blinded to group.

Drugs and solutions

Chloral hydrate, TRIM, L-NAME, SNP, NS309, and papaverine were purchased from Sigma Aldrich (St. Louis, MO, USA). Chloral hydrate was 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 five days. Diltiazem was purchased from MP Biomedicals (Santa Ana, CA, USA) and TM5441 purchased from R&D Systems (Minneapolis, MN, USA). Stock solutions of L-NAME, SNP, papaverine and diltiazem were made weekly and stored at 4℃ until use. TRIM, TM5441, and NS309 stock solutions were aliquoted and stored at −20℃ until use. Please see Supplementary Material for details on buffers used in isolated vessel experiments.

Data calculations and statistical analyses

The number of animals used in each experiment was justified by a statistical power calculation based on our previous study comparing structure of HippAs between pregnant and preeclamptic rats using similar methodology.¹³ The measured variable was lumen diameter (µm) being (mean ± SD) 42.8 ± 6.7 for pregnant and 31.2 ± 6.1 for preeclamptic rats with n = 6/group. We therefore know that with a sample size of n = 6, a two sided 95% confidence interval for a single mean and 1–β of 0.80 is sufficient to detect statistical differences between groups. Results are presented as mean ± SD. Differences between Wistar rats and SHR were determined using a Student's t-test. To determine if there was a relationship between elastin content and distensibility of thoracic aorta, a linear regression analysis and Pearson's correlation was performed. To better understand the relationship between memory function, arteriolar function and aortic stiffness, a multiple linear regression analysis was performed. All statistical analyses were completed using GraphPad 8 (GraphPad Software Inc., La Jolla, CA). Differences were considered significant at p < 0.05. Please see Supplementary Material for all data calculations used in behavioral, blood flow and isolated vessel experiments.

Results

Hippocampal-dependent memory function was reduced in SHR

To investigate hippocampal function in SHR, long-term memory and spatial memory were assessed using behavioral tasks. Baseline locomotion was reduced in SHR, with SHR traveling less distance and having a slower average speed during the habituation phase of NOR compared to Wistar rats (Figure 1(a) and (b)). However, there was no difference in exploratory behavior between groups, with the latency to and number of entries into the center of the arena being similar between groups (Supplementary Figure 2(a) and (b)). Wistar and SHR spent a similar amount of time investigating the identical objects during the acquisition phase of the NOR task (82 ± 22 s vs. 102 ± 40 s; p > 0.05). However, when long-term memory was tested 48 h later, SHR had a significantly reduced recognition index compared to Wistar rats (Figure 1(c)). In the continuous Y maze assessing SAB, both groups entered the maze arms a similar number of times (24 ± 5 entries in Wistar vs. 23 ± 2 entries in SHR; p > 0.05), indicating similar exploratory behavior. However, SHR had more reversals than Wistar, resulting in a significantly lower alternation index (Figure 1(d)). Thus, hippocampal-dependent long-term memory and spatial memory were impaired in SHR.

Figure 1.

Behavioral assessments of long-term memory via a novel object recognition (NOR) task and spatial memory via a continuous Y maze task in Wistar rats and spontaneously hypertensive rats (SHR). During the habituation phase of the NOR task SHR (a) traveled a shorter distance (b) at a slower speed than Wistar rats. (c) The recognition index during NOR was decreased in SHR compared to Wistar rats. (d) The Alternation Index during the Y maze assessment of spontaneous alternation behavior (SAB) was lower in SHR compared to Wistar rats. *p < 0.05 vs. Wistar by Student's t-test.

CBF in CA1 of the hippocampus was reduced in SHR

To understand the mechanisms of hypertension-induced impairment of hippocampal-dependent memory function that may be vascular in nature, we measured absolute hippocampal blood flow using the H₂ clearance method. Under chloral hydrate anesthesia, average blood pressures were higher in SHR compared to Wistar rats as expected (130 ± 8 mmHg vs. 87 ± 4 mmHg; p < 0.05). Table 1 shows physiological parameters during H₂ clearance. Blood pressure decreased during hydrogen inhalation in SHR; however, BP remained higher than in Wistar rats. Arterial pO₂, pCO₂ and pH levels were within the physiological range in both groups during baseline CBF measurements (Table 1). One Wistar rat was excluded from baseline CBF measurements due to technical difficulties resulting in repeated microsensor placement. Figure 2(a) shows representative tracings of H₂ clearance in the hippocampus of a Wistar rat and SHR, from which the half-life (t_1/2) derived and used to calculate CBF. H₂ cleared at a faster rate in Wistar compared to SHR, demonstrating higher CBF. Basal hippocampal CBF was significantly lower in SHR compared to Wistar rats (Figure 2(b)).

Table 1.

Physiological parameters during CBF measurements.

Baseline	Average BP (mmHg)	Arterial pO2 (mmHg)	Arterial pCO2 (mmHg)	Arterial pH
Wistar (n = 5)	87.6 ± 2.5	103.5 ± 11.0	42.3 ± 1.9	7.44 ± 0.02
SHR (n = 6)	97.5 ± 11.7	118.8 ± 18.6	39.0 ± 4.9	7.46 ± 0.03
Hypercapnia
Wistar (n = 3)	84.3 ± 4.0	99.0 ± 13.0	63.3 ± 5.2	7.33 ± 0.01
SHR (n = 4)	95.8 ± 13.1	93.8 ± 5.7	65.4 ± 2.8	7.30 ± 0.01

SHR: spontaneously hypertensive rats; BP: blood pressure; CBF: cerebral blood flow.

Note: Data are mean ±  SD.

Figure 2.

Cerebral blood flow (CBF) in the CA1 region of the hippocampus in Wistar rats and spontaneously hypertensive rats (SHR). (a) Representative hydrogen (H₂) desaturation curves from Wistar rats (black) and SHR (blue) under baseline conditions (normocapnia) from which CBF was determined using the half-life of desaturation (t_1/2). (b) Basal CBF was significantly decreased in SHR compared to Wistar rats. (c) Representative H₂ desaturation curves during baseline (normocapnia) and hypercapnia from which CBF was determined using t_1/2. (d) Paired CBF measurements during baseline and hypercapnia of Wistar and SHR rats demonstrating that hypercapnia increased CBF compared to baseline conditions in both groups. **p < 0.01 vs. Wistar by Student's t-test.

To assess cerebrovascular reactivity in-vivo, hippocampal CBF was also measured during hypercapnia. Hypercapnia was induced without affecting blood pressure or causing hypoxia (Table 1). Figure 2(c) shows representative H₂ clearance traces during baseline (normocapnic) and hypercapnic conditions. As expected, hypercapnia increased blood flow, indicated by the leftward shift in H₂ clearance. Hypercapnia could not be induced in two rats from each group that were therefore excluded from hypercapnic blood flow analyses. Figure 2(d) shows paired measurements of Wistar rats and SHR under baseline and hypercapnic conditions. Despite reduced basal blood flow in SHR, CBF increased in response to hypercapnia to a similar level as in Wistar rats (118 ± 24 mL/100 g/min and 114 ± 11 mL/100 g/min; p > 0.05). However, the percent increase in blood flow was greater and more variable in SHR compared to Wistar rats (84 ± 66 % vs. 20 ± 8 %; p > 0.05).

Reactivity of isolated HippAs to pressure and SK/IK channel activation

To investigate underlying mechanisms by which chronic hypertension may reduce hippocampal perfusion, arterioles that supply CA1 were studied isolated and pressurized. Figure 3(a) shows a representative photomicrograph of a HippA segment pressurized to 40 mmHg. Figure 3(b) shows the lumen diameter over the intravascular pressure range 20–100 mmHg. In HippAs from Wistar rats, lumen diameter decreased as pressure increased above 20 mmHg demonstrating myogenic reactivity. Lumen diameters of HippAs from SHR remained constant across the entire pressure range (Figure 3(b)). HippAs from SHR had smaller diameters than those of Wistar rats (Figure 3(b)) that was partly due to increased myogenic tone at each intravascular pressure (Figure 3(c)).

Figure 3.

Myogenic reactivity of hippocampal arterioles (HippAs) from Wistar rats and spontaneously hypertensive rats (SHR). (a) Wide field image (20 × magnification) of a HippA from a Wistar rat mounted on a glass cannula and pressurized to 40 mmHg. The boxed inset represents the region of interest studied. (b) Pressure–diameter curves of HippAs from Wistar rats and SHR across the pressure range of 20–100 mmHg. (c) Calculated percent myogenic tone was higher in HippAs from SHR compared to Wistar rats at each intravascular pressure studied. **p < 0.01 vs. Wistar by Student's t-test.

Activation of SK/IK channels in the endothelium of cerebral penetrating arterioles causes hyperpolarization and subsequent vasodilation that contributes to conducted vasodilation and neurovascular coupling.^36,37 Whether SK/IK channels are present in HippAs in males and whether the SK/IK activation-induced vasodilation is intact in SHR is unknown. Supplementary Figure 3 shows the concentration-response curve of HippAs to the SK/IK channel agonist NS309. Both groups dilated to activation of SK/IK channels in a dose-dependent manner, although HippAs from SHR had smaller lumen diameters than those from Wistar (Supplementary Figure 3(a)). However, reactivity to NS309 was decreased in HippAs from SHR (Supplementary Figure 3(b)). Thus, SK/IK channels appear to be present in HippAs from male rats; however, activation with NS309 had a diminished dilatory effect in SHR.

The effect of PAI-1 inhibition and NO in HippAs

PAI-1 is elevated during chronic hypertension that contributes to vascular stiffening and dysfunction, and inhibition of PAI-1 caused vasodilation of cerebral arterioles that was NO-dependent that reduced ischemic injury.³⁸ To determine the vasodilatory potential of PAI-1 inhibition in the hippocampus during chronic hypertension, HippAs were exposed to a PAI-1 inhibitor TM5441. Arterioles from SHR were more constricted, with lumen diameters being smaller in SHR compared to Wistar rats (Figure 4(a)). However, PAI-1 inhibition caused ∼25% vasodilation in both groups at the highest concentration of TM5441. Reactivity to TM5441 was blunted in arterioles from SHR (Figure 4(b)) that may be due to arterioles from SHR being more constricted.

Figure 4.

Reactivity of hippocampal arterioles (HippAs) to plasminogen activator inhibitor-1 (PAI-1) inhibition with TM5441 and nitric oxide (NO). (a) Lumen diameters of HippAs were smaller in spontaneously hypertensive rats (SHR) than Wistar rats despite arterioles dilating to TM5441 in a concentration-dependent manner. (b) Reactivity to PAI-1 inhibition was blunted in HippAs from SHR compared to Wistar rats. (c) Lumen diameters of HippAs from Wistar and SHR at baseline, after a 15 min incubation with the nNOS inhibitor TRIM (3 × 10⁻⁴ M), and subsequent 15 min incubation with the NOS inhibitor L-NAME (10⁻³ M). Diameters remained unchanged after NOS inhibition in both groups. (d) HippAs dilated in a dose-dependent manner to the NO-donor SNP. **p < 0.01 vs. Wistar by Student's t-test.

In the hippocampus, NO derived from nNOS is a primary mediator of neurovascular coupling, making the vasodilatory response of vascular smooth muscle to NO particularly important in this vascular bed.³⁹ Further, NO derived from endothelial (e)NOS is basally produced to inhibit myogenic tone in other cerebrovascular territories that is impaired during chronic hypertension.^40,41 To determine the potential contribution of nNOS and eNOS to inhibition of myogenic tone in HippAs and how this may be affected in chronic hypertension, NOS inhibitors were used. Figure 4(c) shows HippA diameters at baseline and after treatment with a single high dose of the selective nNOS inhibitor TRIM, followed by a single high dose of the nonselective NOS inhibitor L-NAME to inhibit remaining eNOS. Lumen diameters were consistently smaller in arterioles from SHR compared to Wistar rats (Figure 4(c)). Interestingly, lumen diameters remained unchanged in response to both nNOS and eNOS inhibition in HippAs from both groups. However, HippAs in both groups dilated to the NO donor SNP in a concentration-dependent manner, although this dilation trended towards being blunted in HippAs from SHR (Figure 4(d)).

Structural and biomechanical properties of HippAs

To determine if HippAs underwent hypertension-induced remodeling, we investigated structural and biomechanical properties of arterioles under fully relaxed (passive) conditions. Passive lumen diameters (Figure 5(a)) of HippAs were similar between groups. However, arteriolar walls were thicker (Figure 5(b)), outer diameters larger (Supplementary Figure 4), and cross-sectional areas larger (Figure 5(c)) in HippAs from SHR compared to Wistar rats. Thicker walls in HippAs of SHR resulted in normalized wall tension (Figure 5(d)) and decreased wall stress (Figure 5(e)). Distensibility curves were calculated to investigate changes in vessel stiffness in chronic hypertension and are shown in Figure 5(f). HippAs from SHR were less distensible than those from Wistar rats at higher pressures, suggesting increased stiffness in HippAs from SHR (Figure 5(f)).

Figure 5.

Structural and biomechanical properties of hippocampal arterioles (HippAs). (a) Lumen diameters of fully relaxed HippAs from Wistar rats and spontaneously hypertensive rats (SHR) were of similar size. (b) Wall thickness and (c) cross-sectional area were increased in HippAs from SHR compared to Wistar rats. (d) Wall tension was similar between groups, and (e) wall stress was decreased in HippAs from SHR compared to Wistar rats. (f) Distensibility of HippAs from SHR was decreased compared to arterioles from Wistar rats at higher pressures. **p < 0.01, *p < 0.05 vs. Wistar by Student's t-test.

Biomechanical properties of thoracic aorta

Segments of thoracic aorta were cannulated and pressurized under fully relaxed (passive) conditions to understand the relationship between hypertension-induced aortic stiffness and HippA function and structure. Figure 6(a) shows a photomicrograph of a cannulated and pressurized aortic segment. In the relatively unpressurized state (5 mmHg), there was no difference in vessel diameter between groups (Figure 6(b)), suggesting aorta are of similar size structurally. However, at intravascular pressures > 60 mmHg, diameters of aorta from SHR were smaller than those from Wistar rats (Figure 6(c)) that was likely due to decreased distensibility in SHR (Figure 6(d)). Figure 6(e) shows elastin content in thoracic aorta was decreased in SHR compared to Wistar rats. Further, elastin content significantly correlated with distensibility at 100 mmHg (Figure 6(f)).

Figure 6.

Structural and biomechanical properties of thoracic aorta from Wistar rats and spontaneously hypertensive rats (SHR). (a) A representative wide field image (15 × magnification) of a cannulated thoracic aorta segment from an SHR pressurized to 120 mmHg. (b) Aortic diameters were of similar size between groups at 5 mmHg. (c) Diameters of pressurized thoracic aorta from SHR were smaller than those from Wistar rats that was due to (d) decreased distensibility of aorta from SHR. (e) Aorta from SHR had less elastin content compared to Wistar rats. (f) Elastin content of thoracic aorta significantly correlated with distensibility at 100 mmHg. **p < 0.01 vs. Wistar by Student's t-test; ^#p < 0.01 by linear regression analysis and Pearson's correlation.

Relationship between memory, HippA function and aortic stiffness

To better understand the relationship between memory function and hippocampal vascular function and large artery stiffness, a multiple regression analysis was performed. Multiple linear regression was calculated to predict long-term memory based on percent tone of HippAs at 60 mmHg, aortic distensibility at 100 mmHg, and spatial memory function. A significant regression equation was found (F(3.8) = 7.28, p = 0.011), with an R² = 0.73.

Discussion

Understanding pathological mechanisms that contribute to vascular cognitive impairment is important as incidence and progression of cerebrovascular disease continue to increase in an ever-aging population. Interestingly, very little is known about the vasculature supplying the hippocampus, a brain region intimately involved in cognitive function that may be adversely affected in pathological states associated with cerebrovascular disease such as chronic hypertension. The primary findings of the current study were that hippocampal-dependent memory was impaired in a model of chronic hypertension that was associated with hyperconstriction of HippAs and hippocampal hypoperfusion. However, the hyperemic response to hypercapnia was similar between groups that were likely due to vessels being of similar size structurally. Finally, hypertension-induced increased aortic stiffness was associated with increased stiffness and microvascular dysfunction in the hippocampus. Together, these findings provide novel insight into the detrimental effects of chronic hypertension on hippocampal vascular function that may contribute to impaired memory function.

Hypoperfusion is considered a primary mechanism by which neuropathological changes occur during chronic hypertension and predicts cognitive decline.^42–44 However, few studies have investigated hippocampal CBF during chronic hypertension as it may relate to cognitive decline. In the current study, hippocampal CBF was decreased ∼40% in SHR compared to normotensive rats that may contribute to impaired hippocampal-dependent memory function. This finding is in partial support of clinical studies reporting that hypertension was associated with decreased hippocampal CBF measured via arterial spin labeling (ASL) MRI; however, in those studies, all subjects were cognitively intact (e.g. non-demented).^45,46 In contrast to our current findings, previous studies have reported baseline hippocampal CBF was similar in SHR and normotensive rats, as measured via ASL MRI and autoradiography.^47,48 The difference between past studies and the current findings may be due to the use of anesthetics including isoflurane and nitrous oxide that are potent cerebral vasodilators that could mask potential differences in CBF.^31,47–50 In the current study, hypertension-induced vasoconstriction of HippAs likely contributes to hippocampal hypoperfusion in SHR, a response that may represent a vascular mechanism by which memory function was impaired in SHR. However, the exact causal relationship between hippocampal vascular dysfunction and impaired memory is not clear from the current study and requires further investigation. Regardless, the hyperconstriction of HippAs during chronic hypertension may represent a therapeutic target such that ameliorating the increased tone and vasoconstriction pharmacologically could potentially restore hippocampal perfusion and rescue memory function.

The vasculature of the hippocampus appears to be unique regarding the role of NO in vascular function. To our surprise, arterioles from neither normotensive nor hypertensive rats responded to NOS inhibition in the current study. While this finding is in agreement with our previous study of HippAs,¹³ it is in striking contrast to the substantial contribution of eNOS-derived NO to inhibition of myogenic tone reported in other cerebrovascular territories.⁴⁰ Further, increased oxidative stress and decreased bioavailability of NO result in increased myogenic tone of cerebral arterioles during chronic hypertension that is a well-established mechanism of hypertension-induced endothelial dysfunction.⁴¹ However, that does not appear to be present in HippAs. Vascular NO plays an important role not only in regulating vascular tone, but also exerts anti-inflammatory and anti-thrombotic actions through its ability to inhibit platelet aggregation, adhesion molecules and lipid oxidation.⁵¹ Further, NO suppresses the expression of PAI-1 in VSM cells, contributing to its anti-thrombotic actions.⁵² Thus, a vascular territory without basal NO production could be at greater risk for pathological processes including thrombosis and ischemic stroke.

In the current study, all HippAs dilated ∼25% to PAI-1 inhibition with TM5441 that has previously been shown to be NO-dependent,³⁸ suggesting eNOS was expressed in HippAs and could produce NO similarly upon stimulation. Further, the hypercapnia-induced vasodilation of the cerebral vasculature is largely dependent upon NO production from eNOS at the hypercapnic status used in the current study (pCO₂ ∼65 mmHg).^53–55 That hypercapnia caused an increase in hippocampal CBF to similar levels in hypertensive and normotensive rats supports a vasoactive role for eNOS under stimulated conditions. Importantly, HippAs vasodilated similarly to the NO-donor SNP in both groups, suggesting the dilatory response of VSM to NO was intact in chronic hypertension. Further, the lumen size of HippAs under fully relaxed conditions was similar, suggesting that the vasodilatory reserve was similar between normotensive and hypertensive rats. Thus, stimulation of eNOS may provide a therapeutic target to stimulate NO production, decrease arteriolar resistance and restore hippocampal perfusion.

The hyperconstriction of arterioles supplying the hippocampus during chronic hypertension in the current study likely contributes to hippocampal hypoperfusion due to increased segmental vascular resistance. Unlike other vascular territories, this does not appear to involve NO. Alternatively, increased tone of HippAs during chronic hypertension may be due to impaired endothelium-derived hyperpolarization (EDH). SK/IK channels are involved in EDH and HippAs from SHR had impaired dilation in response to SK/IK channel activation; however, it is unclear from the current study whether SK and/or IK channels were basally active in HippAs to inhibit myogenic tone. Additionally, SK/IK channels are involved in the conducted vasodilation to increase local CBF that contributes to neurovascular coupling that may be impaired during chronic hypertension.³⁷

Neurovascular coupling is the dynamic communication between neurons, glia, and vasculature that matches neuronal metabolic demands with appropriate blood flow.⁵⁶ During neuronal activity, vasoactive mediators are generated that act locally to vasodilate the microvasculature. That vasodilation is conducted proximally to dilate upstream arterioles, decrease segmental vascular resistance and increased local blood flow. Impaired dilation of HippAs to SK/IK channel activation in SHR in the current study suggests that conducted vasodilation may also be impaired, resulting in uncoupling of neuronal activity and local CBF. Such uncoupling could result in neuronal injury in the hippocampus that is already particularly susceptible to ischemia, thereby directly contributing to memory impairment. Although chronic hypertension impairs neurovascular coupling in the somatosensory cortex of rodents,^57,58 whether this occurs in the hippocampus requires further investigation.

Large conduit arteries such as the aorta play a critical role in maintaining relatively constant pressure in arteries despite the pulsatile nature of blood flow. Under healthy conditions, the elasticity or distensibility of the aorta functions to cushion pulse pressure and minimize pulsatility.⁵⁹ However, chronic hypertension induces structural changes in the aortic wall that increase aortic stiffness and PWV.^16,59,60 Increased PWV results in greater pulsatility being transmitted into the microcirculation of the brain that increases sheer stress and causes endothelial damage.¹⁴ To compensate for the excessive pressure and flow pulsatility, cerebral arterioles remodel and constrict to protect microvessels from damage.⁶¹ However, increased vasoconstriction can cause hypoperfusion and neuronal injury that, if occurring in the hippocampus, could potentiate memory and cognitive dysfunction.

In the current study, we found that during chronic hypertension, decreased elastin content correlated with increased stiffness of aorta that was associated with increased stiffness and remodeling of HippAs, hippocampal hypoperfusion and memory deficit. Clinical studies suggest that aortic stiffening directly influences cognition, including hippocampal-dependent memory function via cerebrovascular dysfunction.^62–64 Muela et al.⁶² reported that patients with severe hypertension had increased aortic PWV that was associated with decreased cerebral vasoreactivity to hypercapnia and impaired memory function. However, assessment of cerebrovascular reactivity was completed using laser Doppler of middle cerebral arteries. Importantly, the hippocampal vasculature primarily stems from the posterior cerebral artery, and laser Doppler of the hippocampal arteries in patients cannot safely be assessed due to the hippocampus being a small, deep brain structure.⁶⁵ Regardless, the current study is the first investigation that we are aware of in to understanding how large artery stiffness may negatively affect the hippocampal microvasculature and subsequent cognitive performance. Overall, hypertension-induced aortic stiffening induces what appears to be widespread cerebrovascular dysfunction including within the hippocampus that may represent a therapeutic target to minimize cognitive burden associated with cardiovascular disease and hypertension.¹⁵

In the current study, SHR demonstrated impaired hippocampal-dependent long-term memory and spatial memory compared to normotensive Wistar rats. SHR are widely used as models of several neuropsychological conditions including attention-deficit/hyperactivity disorder (ADHD) and vascular cognitive impairment and dementia. Importantly, SHR demonstrate ADHD-like behavior that is exclusively studied during the juvenile phase, during which the young rats show no impairment in memory, and even enhanced performance in novel object recognition.^66–68 However, SHR are used as a model of vascular cognitive impairment and dementia in adulthood. In the current study, adult SHR were used (∼ 5–6.5 months of age) to better model the patient population and target investigation of chronic hypertension-induced hippocampal dysfunction. Further, in an open field arena, SHR demonstrated reduced activity (Figure 1), not hyperactivity, confirming adult SHR did not demonstrate ADHD-like behavior. However, a limitation of the current study was that Wistar rats were used at three to four months of age, a time point earlier in adulthood than SHR. Although hippocampal-dependent tasks (e.g. spatial learning and memory) decline with age in both SHR and Wistar rats, changes in normotensive controls do not occur until 12–24 months of age.^24,69–71 In fact, object recognition was unaffected by age in Wistar rats, with there being no difference in performance at 4, 8, 18, or 24 months of age.²⁴ Thus, despite this limitation, the impaired memory function measured in SHR was likely due to a detrimental effect of chronic hypertension on hippocampal function.

In conclusion, the findings of the current study provide novel insight into functional aspects of the hippocampal vasculature under physiological conditions that seem unique compared to other cerebrovascular territories. Further, chronic hypertension disrupts hippocampal memory function potentially via increased aortic stiffness-induced dysfunction of the hippocampal microvasculature. These findings suggest that the hippocampal vasculature may hold a currently untapped therapeutic potential in alleviating at least some of the cognitive burden associated of cerebrovascular disease. The hippocampus is critically involved in higher order cognitive function. Understanding the function of the hippocampal vasculature under physiological and pathophysiological conditions is important, particularly as several neurodegenerative diseases involve vascular-mediated hippocampal injury, including stroke, epilepsy, Alzheimer's disease, vascular cognitive impairment and dementia.^3,4,72–74

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institutes of Health (NIH) National Institute of Neurological Disorders and Stroke (NINDS) R01 NS093289 and R01 NS108455, 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, JEM and MJC made a substantial contribution to the concept and design, acquisition of data or analysis and interpretation of data. ACJ, JEM and MJC drafted the article or revised it critically for important intellectual content, and approved the version to be published. Data will be provided upon request.

Supplementary material

Supplementary material for this paper can be found at the journal website: http://journals.sagepub.com/home/jcb