Autonomic Dysfunction Impairs Baroreflex Function in an Alzheimer’s Disease Animal Model

Abstract

Background:

Alzheimer’s disease (AD) patients frequently present with orthostatic hypotension. This inability to reflexively increase blood pressure on standing is a serious health concern and increases the risk of stroke and cardiovascular diseases.

Objective:

Since there are no clear mechanisms for orthostatic hypotension in human AD, the present study assessed the autonomic changes that could explain this comorbidity in an AD animal model.

Methods:

We used the established streptozotocin-induced rat model of AD (STZ-AD), which mimics many hallmark symptoms of sporadic AD in humans. Baroreflex responses were analyzed in anesthetized STZ-AD rats using femoral catheterization for blood pressure and heart rate, and autonomic activity was assessed using specific blockers and splanchnic sympathetic nerve recordings. Expression levels of autonomic receptors at the heart were examined using the western blot technique.

Results:

Baroreflex function in STZ-AD showed a blunted heart rate (HR) response to low blood pressure challenges, and the maximal sympathetic nerve activity was reduced. Conversely, HR responses to high blood pressure were similar to control, indicating no change in parasympathetic nerve activity. Under resting conditions, autonomic blockade demonstrated a baseline shift to increased sympathetic tone in STZ-AD. Protein expression levels of beta-1 adrenergic receptor and muscarinic acetylcholine receptor M2 in the heart were unchanged.

Conclusion:

Our study provides the first data on the pathological influence of AD on baroreflex function, which primarily affected the sympathetic nervous system in STZ-AD. These results represent the first mechanisms that may correlate with the orthostatic hypotension in human AD.

INTRODUCTION

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder often characterized by impairments in memory and cognition [1]. A frequent comorbidity in AD is orthostatic hypotension which affects about 35% of cases [2, 3]. Orthostatic hypotension manifests as a fall in blood pressure after postural changes to the standing position, and the adverse health outcomes include falls, syncope, and cardiovascular events [4, 5].

Disturbances in baroreflex control have been previously correlated with a potential autonomic dysfunction in AD patients and were assessed using indirect cardiovascular reflex testing and measurement of heart rate (HR) [6, 7]. For example, patients with mild cognitive impairments had significantly lower HR during deep breathing tests, and the drop in systolic blood pressure on standing was significantly higher than in age-matched controls [8]. Further, some studies found altered HR variability (beat-to-beat alterations of HR) in AD, and close analysis of the time and frequency domain gave indications of the probable underlying autonomic division involved [9, 10]. However, other studies found no change in HR variability [11], and the overall role of the sympathetic and parasympathetic nervous system for orthostatic hypotension in AD remains poorly understood.

The streptozotocin-induced AD rat model (STZ-AD) mimics many of the symptoms found in AD patients, including memory deficits, respiratory dysfunction, brain atrophy, amyloid-β deposition, tau tangle formation, oxidative stress, and inflammation [12–17]. Unlike genetically-induced AD models, we previously demonstrated significant astrogliosis in a brainstem nucleus of the STZ-AD model, namely the nucleus tractus solitarii [16], which is important for normal baroreflex function. Disturbances of neuronal activity in the nucleus tractus solitarii or any subsequent nucleus involved in baroreflex function will therefore affect heart function and blood pressure control [18, 19]. The combined pathophysiological changes make the STZ-AD model ideally suited for studying the central mechanisms behind baroreflex dysfunction known from human AD. Thus, the aim of the current study was to determine whether the STZ-AD rat model also mimics the baroreflex dysfunction known from human AD and whether alterations in sympathetic and parasympathetic nerve activity explain changes in baroreflex function.

MATERIALS AND METHODS

Animals

Male Sprague-Dawley rats (8–11 weeks, n = 67 rats) were kept in an AAALAC-accredited facility at A.T. Still University’s Kirksville College of Osteopathic Medicine (ATSU-KCOM). Animals were bred in-house from rats previously purchased from Hilltop Lab Animals or Envigo. Cages were in a temperature-controlled room (23°C, 46% humidity) under a 12 : 12 h light/dark cycle, and rats had ad libitum access to commercial rat food and water. A subset of experiments was conducted at the University of Missouri and rats (purchased from Envigo) were housed under the same conditions in the AAALAC-accredited vivarium of the Dalton Cardiovascular Research Center. All protocols and procedures were approved by the Animal Care and Use Committee of ATSU-KCOM and the University of Missouri, and were in accordance with NIH guidelines (“Guide for the Care and Use of Laboratory Animals”).

Alzheimer’s disease model

Experimental rats were randomly assigned to either a control (CTL) or STZ-AD group, ensuring that individual rat litters contributed equally to each group to minimize potential skewing of data. Similar to before [15, 16], isoflurane-anesthetized rats (5% induction, 2–2.5% maintenance, Patterson Veterinary) were positioned in a stereotaxic frame (Model 68005, RWD Life Science), and core body temperature was maintained during the surgery using a heating pump (HTP-1500, Adroit Medical Systems) connected to a heating pad (HP 3275, Hallowell EMC). Dexamethasone (2 mg/kg, Vet One) was administered subcutaneously as an immunosuppressant to reduce cerebral inflammation. After sterilization of the crown of the head with iodine and alcohol, a small medial incision (~10–15 mm) was made to expose the cranial suture lines of the skull. Bregma was visually identified and small bilateral holes were drilled (Dremel 7300 with engraving cutter 105) to provide access to the lateral ventricles according to the following stereotaxic coordinates [20]: AP, −0.9 mm; ML,±1.5 mm; DV, 3.6 mm. Borosilicate glass micropipettes (type 1B120F-4, World Precision Instruments) were pulled using a micropipette puller (P-97, Sutter Instruments) and filled with either vehicle (0.9 mM citrate buffer saline, pH 4.5) or 2 mg/kg STZ (Alfa Aesar) in vehicle. We have previously shown that this STZ dosage is subdiabetogenic [16, 17]. A fluid volume of 5 μL per ventricle was slowly injected over 1 min to allow pressure equilibration in the ventricle. After injection, the surgical site was cleaned with sterile saline (0.9%, Baxter Healthcare) and sutured shut (MV-J397-V, Oasis). All animals were postsurgically injected with buprenorphine hydrochloride (0.05 mg/kg, s.q., Par Pharmaceutical), Baytril (7 mg/kg, i.m., Vet One), and 3 mL of normal saline (0.9%, s.q.). After the surgical procedure, the animals were closely monitored for signs of pain and given caloric supplementation (Froot Loops, Kellogg’s) to ensure weight recovery back to presurgical weight.

Morris water maze

Rats were tested in the Morris water maze for 4 days to assess spatial memory (12–14 days after induction of the AD model) [21, 22]. The maze consisted of a round metal tub (D = 178 cm) filled with darkened water (~730 L, pH = 7.2, 22°C, Black Dye-Mond Pond dye, Pond Logic) to conceal a circular escape platform (D = 10 cm) that was placed in one of four hypothetical pool quadrants of equal dimension (see Fig. 1C). Each quadrant was marked with a visual cue (i.e., yellow star, blue square, red circle, or green triangle) to allow for spatial orientation. Generally, animals were tested for 3 days. On day 1 only, each rat did 5 pretrials with a visible escape platform (1 cm above the water surface) to familiarize the animals with the experimental setup. For the acquisition trials, animals were randomly placed in a quadrant that did not contain the platform, and the time it took to locate and climb onto the escape platform (1 cm beneath the water surface) was recorded. If a rat was unable to locate the platform within 60 s (cutoff time), the animal was gently placed on the platform and allowed to rest and survey its surroundings for 60 s until the next trial. This acquisition period lasted for 3 days with 15 consecutive trials/day, and each set of trials was 24 h apart. A subset of rats was exposed to a probe trial on day 4, where the platform was removed and the rat was allowed to swim for 1 min. A video recording was taken to calculate the average time spent in the quadrant that previously contained the escape platform using Fiji (ImageJ version 1.52 g, NIH) and the Animal Tracker plugin [23].

Fig. 1.

Memory impairment in STZ-AD. A) Group averages of escape latency to find the hidden platform (HP) for CTL and STZ-AD groups over a 3-day (d1-d3) acquisition period (CTL, n = 10 versus STZ-AD, n = 11). VP, visible platform. B) Average group swim velocity demonstrating similar motor function (CTL, n = 10 versus STZ-AD, n = 11). C) Representative swim pattern of CTL (left) and STZ-AD (right) rats during the probe trial on day 4. Quadrant D was the previous position of the HP. D) Group data comparing the time animals spent in quadrant D during the probe trial (n = 6/group). **p ≤ 0.01.

Femoral catheterization and drug administration

Similar to before [24] and 2 weeks after STZ-AD model induction, rats underwent femoral catheterization for baroreflex assessment. Rats were anesthetized with isoflurane (5% induction, 2–2.5% maintenance) and placed on a heating pad. The femoral artery and vein were accessed via the groin using blunt dissection. Once isolated, a catheter (PE-30, Braintree) was inserted medially into the femoral artery. This catheter was connected to a pressure transducer (MLT844, ADInstruments), bridge amplifier (FE221, ADInstruments), and PowerLab acquisition system (PL3508, ADInstruments). The pressure transducer was calibrated before each experiment using a portable manometer (8205, AZ Instrument Corp.). Three separate cannulas (PE-10 fused to PE-50, Intramedic) were inserted into the femoral vein, also extending medially. Two venous catheters were connected to a syringe pump (Model A-99, Razel) that delivered sodium nitroprusside (SNP, 1 mg/mL in 0.9% saline, Fisher) or phenylephrine (PE, 1 mg/mL in 0.9% saline, Sigma) at manually controlled speed. Mean arterial pressure (MAP) was altered within 1–2 min from baseline to 50 mmHg using SNP or to 150 mmHg using PE. The third venous catheter was used for bolus injections of several drugs (1 mg/kg atropine, Alfa Aesar; 1 mg/kg atenolol, Acros Organics; 30 mg/kg hexamethonium, Calbiochem) diluted in heparinized 0.9% saline. Efficacy of autonomic blockade at the heart was verified in a subset of rats by administration of a second dose of the autonomic blocker or by bolus injection of the agonist (SNP and PE) to exclude additional changes in HR. After catheter placement, anesthesia was reduced to ≤2% isoflurane to keep the experimental animal just at surgical plane and minimize potential cardiovascular effects of the anesthetic (see Discussion). Cardiovascular parameters were allowed to stabilize at least 1 h after catheter placement, and baseline MAP and HR were held constant for at least 30 min before starting experimental protocols. The computer software LabChart (ADInstruments) was then used to calculate MAP and HR from the raw arterial pressure trace (1 kHz sample rate).

Heart rate variability (HRV) and spectral analysis

Intervals between heart beats were calculated from a 5-min recording at baseline condition using LabChart software. Intervals were imported into Kubios HRV Standard software (version 3.5.0) to calculate HRV in the time and frequency domain. The RR time series rate was interpolated at 10 Hz and the Welch’s periodogram for fast Fourier transformation was set at 512 s window width with 50% overlap [25]. Frequency power bands were set as follows [26]: VLF: 0 – 0.2 Hz, LF: 0.2 – 0.75 Hz, and HF: 0.75 – 3 Hz. Group comparison was done in Excel.

Sympathetic nerve recordings

In a subset of animals, a splanchnic sympathetic nerve was isolated in deeply anesthetized rats using the retroperitoneal approach [24]. The nerve was recorded with bipolar Teflon-coated silver electrodes (0.005–0.007 inch, A-M Systems). To isolate electric signals from the surrounding noise, the silver electrodes and nerve were covered in silicone elastomer (Kwik-Cast, WPI). Nerve activity was amplified (1000×) and filtered (30–3000 Hz) using a P511 Grass amplifier (Grass Technologies). The signal was then rectified and integrated using a root mean square converter (time constant 28 ms). The splanchnic sympathetic nerve activity (sSNA) was electronically averaged, and the signal between bursts of activity was defined as the background noise and subtracted from the nerve recording. To allow comparison between groups, sSNA was normalized to the individual resting activity at set point ( = 100%) [27].

Baroreflex curves and autonomic balance

Baroreflex curves were generated in SigmaPlot 14.0 (Systat Software) from blood pressure ramps and HR or sSNA responses using the following sigmoid regression equation:

$$\text{y}={\text{y}}_0+\frac{\text{a}}{1+e^{-\left(\frac{x-x_0}{b}\right)}}$$

The midpoint of the curve is represented by y₀, whereas a and x₀ refer to the maximum and minimum HR or sSNA values, respectively. Curves depicting the gain were calculated from the first derivative of the baroreflex curve. All gains curves were aligned at the calculated midpoint and averaged for each group.

Baseline autonomic balance was calculated from HR responses to atropine and atenolol from rest condition according to the following equation:

$$\begin{array}{l}\text{Balance}\left[\%\right]\\ =100\times \frac{\Delta \text{HR}\hspace{0.17em}\text{atenolol}-\Delta \text{HR}\hspace{0.17em}\text{atropine}}{\Delta \hspace{0.17em}\text{HR}\hspace{0.17em}\text{atenolol}+\Delta \hspace{0.17em}\text{HR}\hspace{0.17em}\text{atropine}}\end{array}$$

The sign of the balance describes dominance of either sympathetic (positive) or parasympathetic (negative) input to the heart.

Frequency and power density analysis of sSNA

LabChart software was used to measure the frequency spectrogram between 0–10 Hz and the respective power density from the integrated sSNA recording [28]. Fast Fourier transformation size was set to 128,000 with a Hann (cosine-bell) data window and 93.75% window overlap. Power densities for each individual recording were averaged over a 40-s window at baseline and low blood pressure.

Western blot

Tissue of frozen heart atria was homogenized with 300 μL RIPA buffer (1% NP-40, 0.5% deoxycholate, 0.1% SDS, 0.15 M NaCl, 50 mM Tris-HCl, and 2.5 mM EDTA), incubated on ice for 45 min, and centrifuged for 10 min at 4°C (13,000 rpm). The protein concentration of each sample was determined using the Bradford assay [29] by adding 1 or 2 μL sample to a total of 400 μL molecular biology grade H₂O and 100 μL Protein Assay dye (#5000006, Bio-Rad). The absorbance was measured in triplicate at 595 nm against bovine serum albumin standards. Next, 30 μg of total protein for each sample was diluted in loading buffer (4×Laemmli SDS sample buffer with 5% mercaptoethanol), heated for 10 min at 95°C, and then loaded on a 10% hand-casted polyacrylamide gel to separate proteins using a vertical electrophoresis cell (Mini-PROTEAN, Bio-Rad). Next, proteins were transferred onto a PVDF membrane (IPFL00005, Immobilon-FL), and nonspecific binding was blocked with Intercept Protein-Free Blocking Buffer (927–90001, LI-COR) for 1 h. The membrane was then washed in Tris-buffered saline containing 0.1% Tween-20 and incubated with primary antibodies against either β1 adrenergic receptor (1 : 1000, rabbit, #AAR-023, Alomone labs) or muscarinic acetylcholine receptor M2 (1 : 600, rabbit, #AMR-002, Alomone labs) and β-actin (1 : 5000, mouse, #926–42212, LI-COR) overnight at 4°C. After washing (5×5 min with TBST), the membrane was incubated for 1 h in secondary antibody solution (anti-mouse: #926–68072, 1 : 15,000 and anti-rabbit: #926–32213, 1 : 10,000, LI-COR). The membrane was allowed to dry before imaging using the LI-COR Odyssey imaging system. The Chameleon Duo protein ladder (#928–60000, LI-COR) was used to determine correct protein size. Densitometric analysis of the Western blots was done using ImageJ software (v1.53r, NIH) and protein concentrations were normalized to the loading control (β-actin).

Statistical analysis

SigmaPlot 14.0 and Microsoft Excel were used for all statistical analyses. Most data sets and the individual points on the baroreflex curve were compared between groups using t-tests. Other group comparisons with multiple conditions were analyzed using either 1-way or 2-way repeated ANOVA with Student-Newman-Keuls post hoc test where appropriate. If normality was not given, the nonparametric-equivalent test was used. Variance between 2 data sets was compared using the F-Test. Results are reported as mean±standard error of the mean (SEM), and p values ≤ 0.05 were considered significantly different.

RESULTS

Impaired short-term and long-term memory with STZ-AD

To verify AD model induction, we tested spatial memory of CTL and STZ-AD rats using the Morris water maze paradigm (Fig. 1). During our 3-day acquisition period, STZ-AD rats took significantly more time to learn the location of the submerged escape platform than CTL rats (Fig. 1A), which indicated reduced short-term spatial memory. Longer escape latencies in the STZ-AD group were not caused by motor impairments since both groups had similar swim velocities (Fig. 1B). To test long-term memory on day 4 (probe trial), we removed the submerged platform and measured the time rats spent in the quadrant that previously had the platform (quadrant D) (Fig. 1C, D). STZ-AD rats spent significantly less time in the target quadrant than CTL rats, verifying a reduced long-term memory in this AD model.

Blunted baroreflex function of STZ-AD rats

We previously found STZ-induced changes of neuronal activation in a brainstem nucleus important for blood pressure control [17]. To identify whether the STZ-AD rat model also mimics baroreflex dysfunction of AD patients, we accessed blood pressure and heart rate (HR) through cannulation of the femoral artery of our experimental rats. We initially compared resting values between our groups. At baseline condition (also defined as set point [SP]), there was no difference between groups for resting mean arterial pressure (MAP; CTL, 96±3 mmHg, n = 18 versus STZ-AD, 95±2 mmHg, n = 16, p = 0.84) and HR (CTL, 347±5 bpm, n = 18 versus STZ-AD, 332±7 bpm, n = 16, p = 0.12).

Next, we challenged the baroreflex with injections of vasoactive drugs. In response to decreasing MAP with sodium nitroprusside (SNP) ramp infusion (to 50 mmHg), CTL rats had an expected baroreflex-mediated increase in HR (Fig. 2A₁, left). A similar SNP-induced decrease in MAP was found in STZ-AD rats, but the increase in HR was lacking (Fig. 2A₁, right). When increasing MAP with phenylephrine (PE, to 150 mmHg), CTL rats had an expected compensatory decrease in HR (Fig. 2A₂, left). A comparable decrease in HR to high MAP was observed in the STZ-AD group (Fig. 2A₂, right).

Fig. 2.

Blunted baroreflex function in STZ-AD rats. A) Typical responses of arterial pressure (AP, red trace), mean arterial pressure (MAP, white trace), inter-beat interval (IBI, purple trace), and heart rate (HR, blue trace) to sodium nitroprusside (SNP, A₁) and phenylephrine (PE, A₂) ramp infusions for CTL (left) and STZ-AD (right) animals. B) Group data comparing the HR response to changes in MAP (baroreflex curve) between CTL and STZ-AD groups. Points depict maximal and minimal HR, set point (SP, blue dots; baseline, no prior intervention), and midpoint (MP, green dots) of the baroreflex curve. C) Comparison of average baroreflex curve slopes (gain) between groups. CTL, n = 18 and STZ-AD, n = 16. **p ≤ 0.01.

Figure 2B shows the sigmoidal baroreflex functions calculated from averaged blood pressure ramps of CTL and STZ-AD groups. Overall, the STZ-AD curve shifted toward lower HR values. Although this downward shift was nonsignificant at SP (at baseline before any intervention), the calculated midpoint (MP) of the baroreflex curve had a significantly reduced HR for STZ-AD with no obvious shift in MAP. In response to low MAP (with infusion of SNP), STZ-AD rats did not increase HR, and the maximum HR was significantly lower than in the CTL group. Furthermore, the difference between maximum HR and SP was significantly reduced in STZ-AD (∆ from SP to max. HR: CTL,+18.1±5.8 bpm, n = 18 versus STZ-AD, −0.4±4.9 bpm, n = 16, p = 0.021), demonstrating a missing reserve capacity to further increase HR (from SP) and thus the inability to compensate for low blood pressures. Conversely, at high MAP (with infusion of PE), the magnitude of the compensatory HR reduction was similar between STZ-AD and CTL groups (∆ from SP to min. HR: CTL, −22.0±4.3 bpm, n = 18 versus STZ-AD, −34.2±7.2 bpm, n = 16, p = 0.15). Minimal HR values (Fig. 2B) tended to be lower in STZ-AD, resembling the general shift of the baroreflex curve to a lower HR. Nevertheless, SP was similar to control at the expense of reserve capacity not allowing compensation for lower blood pressure.

To identify baroreflex sensitivity at MP of the baroreflex curves, we compared the slope (∆bpm/mmHg) ±30 mmHg from MP between groups (Fig. 2C). The gain was similar between groups, indicating a similar sensitivity of the baroreflex system across its dynamic range.

Similar pulse pressures between CTL and STZ-AD groups

Pulse pressure (PP, difference between systolic and diastolic blood pressure) is used as a clinical indicator (among many others) for heart problems and offers insight into arterial compliance, contractile force of the heart, and autonomic input to the heart [30]. To identify whether PP changed in STZ-AD, we analyzed raw PP at baseline and during stress conditions with low and high blood pressure. Figure 2A shows exemplary PP traces (red traces) for CTL (left) and STZ-AD (right). At baseline, STZ-AD rats had a similar PP to CTL (PP at SP: CTL, 31±3 mmHg, n = 18 versus STZ-AD, 30±4 mmHg, n = 16, p = 0.79). The PP were also similar between groups at the nadir of low MAP (at ~50 mmHg: CTL, 23±2 mmHg, n = 18 versus STZ-AD, 21±3 mmHg, n = 16, p = 0.58) and at the peak of high MAP (at ~150 mmHg: CTL, 54±4 mmHg, n = 18 versus STZ-AD, 46±5 mmHg, n = 16, p = 0.23). While the absent change in PP does not exclude other potential changes at the heart and arteries with STZ-AD, it may be an indicator for a more centrally mediated (i.e., autonomic outflow) baroreflex dysfunction in this model of AD.

Time and frequency domain analysis of heart rate variability (HRV) in STZ-AD

We analyzed HRV as an initial indicator for potential changes in autonomic balance as underlying cause for the baroreflex dysfunction in STZ-AD rats [31]. Group analysis of the interbeat intervals at baseline showed no difference between groups (SDNN: CTL, 2.9±0.2 ms, n = 18 versus STZ-AD, 3.4±0.2 ms, n = 16, p = 0.10). Next, we did a Fourier transformation of HR intervals to reveal its spectral components. The high frequency power band, which indicates changes in parasympathetic nerve activity, was similar between groups (HF: CTL, 5.80±0.93 ms², n = 18 versus STZ-AD, 7.58±0.91 ms², n = 16, p = 0.182). The low frequency component, typically representing a mix between parasympathetic and sympathetic activity, showed a weak trend towards an increased power in STZ-AD (LF: CTL, 0.58±0.16 ms², n = 18 versus STZ-AD, 1.25±0.36 ms², n = 16, p = 0.087). Increased LF power can indicate a shift towards higher sympathetic control over HR [31]. However, the trend in increased LF did not translate into an altered ratio of low and high frequency, which depicts the global sympathetic/parasympathetic balance (LF/HF: CTL, 0.11±0.03, n = 18 versus STZ-AD, 0.15±0.03, n = 16, p = 0.44). Lastly, the very low frequency band, which is associated with sympathetic activity influenced by the renin-angiotensin system, was not different between groups (VLF: CTL, 0.61 ± 0.15, n = 18 versus STZ-AD, 0.49 ± 0.08, n = 16, p = 0.51).

Increased resting sympathetic tone in STZ-AD

To better understand the underlying mechanisms of baroreflex dysfunction in the STZ-AD rat model, we assessed the influence of central autonomic output on HR by sequentially injecting atropine (parasympathetic blocker, leading to increased HR) (Fig. 3A₁) and atenolol (sympathetic blocker, leading to decreased HR) (Fig. 3A₂) in a subgroup of animals under resting conditions. The magnitude of HR change to each blocker will reveal the relative balance of the sympathetic and parasympathetic nervous system at rest and their influence on the heart. Again, an overall slight shift toward lower HR was observed in STZ-AD when compared with CTL (Fig. 3B). After atropine injections, STZ-AD rats had a significantly blunted increase in HR (Fig. 3C) and absolute HR was significantly lower than in CTL (Fig. 3B). Subsequent bolus injection of atenolol decreased HR in both groups and the magnitude of response was similar (Fig. 3D). We verified that these effects were not dependent on the sequence of blocker injections in another subset of animals (data not shown). A reduced parasympathetic influence on HR at rest indicates a shift in relative autonomic balance towards a greater sympathetic control over the heart and is consistent with the significant leftward shift of SP towards maximal HR on the baroreflex curve of STZ-AD (Fig. 2B). To further examine the relative autonomic balance and its influence on the heart, we first identified the total autonomic input to the heart by summating the absolute change of HR to each autonomic blocker. Although not significant, the total autonomic input to the heart was smaller in the STZ-AD group when compared to control (CTL, 79.1±10.1 bpm, n = 8 versus STZ-AD, 67.1±4.0 bpm, n = 10, p = 0.25). Next, we calculated the relative contribution of each autonomic branch to the total autonomic input in each group (see methods; e.g., no change in HR to a particular blocker would shift autonomic balance 100% to the opposite autonomic branch). Figure 3E shows the overall autonomic balance over HR at rest for our experimental groups. In STZ-AD, autonomic balance at SP is significantly shifted toward an increased sympathetic tone, which is consistent with the leftward shift of SP towards maximal HR in the baroreflex curve and a loss in sympathetic reserve capacity when stressed with low blood pressure.

Fig. 3.

Increased resting sympathetic activity in STZ-AD. A) Typical mean arterial pressure (MAP) and heart rate (HR) tracings during successive bolus injection of atropine (A₁) and atenolol (A₂) in a control animal. B) Group averages of HR to injections of atropine and atenolol from set point (SP, baseline). C) Group data comparing ∆HR from SP to parasympathetic blockade with atropine. D) Group data comparing ∆HR to sympathetic blockade with atenolol. E) Overall resting autonomic influence on HR. CTL, n = 8 and STZ-AD, n = 10. *p ≤ 0.05; **p ≤ 0.01.

Blunted sympathetic activity to low MAP challenges in STZ-AD

Next, we looked at HR responses under stress conditions using low and high MAP to induce maximal possible outflow of the sympathetic and parasympathetic systems, respectively. Figure 4 shows typical raw traces for MAP and HR to an atenolol bolus during SNP infusion (Fig. 4A) and an atropine bolus during PE infusion (Fig. 4C). At maximal HR and maximal sympathetic activation in response to low MAP, the drop in HR to sympathetic blockade was significantly reduced in STZ-AD when compared with CTL rats (Fig. 4B). This result suggests that STZ-AD rats have an overall reduced maximal sympathetic output to the heart and is consistent with our blunted increase in HR to low MAP (Fig. 2). Conversely, under maximal parasympathetic outflow, the STZ-AD group had a similar change in HR to atropine when compared with CTL animals (Fig. 4D), suggesting no change of parasympathetic output between groups under high MAP stress conditions.

Fig. 4.

STZ-AD induces blunted sympathetic outflow at low MAP. A, B) Representative traces (A) and group data (B) for heart rate (HR) responses to atenolol during low mean arterial pressure (MAP, under sodium nitroprusside [SNP] infusion) and maximal sympathetic activity for CTL (left) and STZ-AD (right) rats. C, D) Representative traces (C) and group data (D) for HR responses to atropine during high MAP (under phenylephrine [PE] infusion) and maximal parasympathetic activity for CTL (left) and STZ-AD (right) rats. n = 7/group. *p ≤ 0.05.

Blunted sympathetic nerve activity in STZ-AD rats

To directly analyze the dynamics of sympathetic outflow in STZ-AD, we recorded splanchnic sympathetic nerve activity (sSNA) in a subset of rats. All individual recordings from STZ-AD animals had a lower resting sSNA when compared to CTL (noise-corrected sSNA at SP: CTL, 0.12±0.01 mV.s, n = 3 versus STZ-AD, 0.07±0.01 mV.s, n = 3, p = 0.03). Despite this decrease in resting sSNA, resting HR was similar between groups (SP of HR: CTL, 311 ± 19 bpm, n = 3 versus STZ-AD, 315±19 bpm, n = 3, p = 0.87; also see Fig. 2B). This outcome indicates a compensatory change in STZ-AD, such as a shift in autonomic balance (Fig. 3E) to maintain a normal HR.

Next, we challenged sSNA with blood pressure ramps (Fig. 5). Similar to before, a reduction of MAP with SNP infusion induced a compensatory increase of HR in the CTL group (Fig. 5A, left). Simultaneously, sSNA increased to low MAP, validating that the observed increase in HR was mediated by increasing sympathetic outflow. Conversely, there were no changes in HR and sSNA in STZ-AD (Fig. 5A, right). This inability to increase sSNA to low MAP in STZ-AD was also seen in the group data in Fig. 5B (normalized sSNA to its baseline activity at SP). While CTL rats were able to increase sSNA to low MAP by 56%, STZ-AD rats increased by only 8%. Accordingly, the difference between maximum HR and SP tended to be reduced in STZ-AD (∆ from SP to max. HR: CTL,+58.8±19.5 bpm, n = 3 versus STZ-AD,+11.3±6.3 bpm, n = 3, p = 0.081). With infusion of PE, there was no difference of sSNA at MP (despite a slight shift toward higher MAP pressures) and at high MAP between our experimental groups (Fig. 5B).

Fig. 5.

Blunted sympathetic outflow in STZ-AD rats. A) Typical responses of arterial pressure (AP, red trace), mean arterial pressure (MAP, white trace), heart rate (HR, blue trace), integrated splanchnic sympathetic nerve activity (sSNA, green trace), and mean sSNA (black trace) to sodium nitroprusside (SNP) ramp infusions for CTL (left) and STZ-AD (right) animals. B) Group data for a subset of rats comparing the % sSNA response (normalized to baseline [set point, SP]) to changes in MAP (baroreflex curve) between CTL and STZ-AD. Points depict maximal and minimal sSNA, SP (blue dots), and midpoint (MP, green dots) of the baroreflex curve. sSNA between groups: *p ≤ 0.05. MAP between groups: #, p ≤ 0.05. C) Time course of frequency spectra (top) for the corresponding examples shown in panel A. Power densities are a 40-s average at baseline (Bsl) and during low blood pressure (LBP). D) Group data for peak power densities at 5–7 Hz (HR modulation, closed arrow head in C), 1–2 Hz (respiratory modulation, open arrow head in C), and 0.5–1 Hz (low-frequency HR modulation) at Bsl and LBP. versus CTL: ***p ≤ 0.001. versus Bsl: ^‡‡p ≤ 0.01; ^‡p ≤ 0.05. CTL, n = 3 and STZ-AD, n = 3.

Figure 5C (top) shows typical frequency spectra of the integrated sSNA traces from CTL and STZ-AD (time frames correspond to those given in Fig. 5A). We found a strong coupling of sSNA to HR at baseline/resting condition in the CTL group (5–7 Hz peak measured at ‘Bsl’ of the trace in Fig. 5C, closed arrow head). Additionally, and as previously reported [32, 33], we identified main frequency bands between 1–2 Hz and 0.5–1 Hz, resembling coupling of sSNA to respiratory rate (open arrow head in Fig. 5C) and low-frequency HR modulation, respectively. Other smaller oscillations were found between 2–6 Hz and likely represent centrally-generated rhythms [34]. The STZ-AD group, on the other hand, had an overall lower sSNA activity and the coupling to HR had a significantly lower power under baseline condition when compared to CTL (Fig. 5C, D). The power of the frequency bands between 1–2 Hz and 0.5–1 Hz, however, were similar to CTL (Fig. 5D).

With infusion of SNP and resultant low blood pressure, the pattern of frequency bands in sSNA changed only in the CTL group (measured at ‘LBP’ of the trace in Fig. 5C). CTL rats lost the strong coupling to HR (5–7 Hz peak in Fig. 5D), which is consistent with an overall increase of sSNA and desynchronized firing of sympathetic fibers in the splanchnic nerve. Similarly, there was a significantly reduced coupling to low-frequency HR modulation (0.5–1 Hz). On the other hand, CTL rats trended towards stronger sSNA-coupling with respiration when compared to baseline (t-test of 1–2 Hz peak power, p = 0.081). This coupling became significantly higher in CTL than in STZ-AD (∆ from Bsl to LBP: CTL,+164±23%, n = 3 versus STZ-AD,+21±46%, n = 3, p = 0.049). Power densities in the STZ-AD group did not change in all 3 frequency bands during the low blood pressure condition. Frequency peaks for respiratory and low-frequency HR modulation of sSNA shifted to a slightly higher frequency (by 0.1–0.2 Hz) in both groups, which was a significant change in the CTL group only. This shift in peak frequency was due to the baroreflex-mediated increase in HR and respiration [35]. The combined data of the sSNA frequency spectra again demonstrate the inability of STZ-AD rats to adequately respond to a low blood pressure challenge.

In summary, the blunted baroreflex function in STZ-AD is primarily due to an overall reduction of central sympathetic output (lower ceiling response), and a resetting of SP on the baroreflex curve to maximal sympathetic output (in its reduced limits), which maintains a seemingly normal HR at resting conditions.

Unaltered autonomic receptor expression at the heart of STZ-AD

To examine whether alterations of beta-1 adrenergic receptors (β1AR) at the heart have a role in the blunting of the baroreflex response, we compared their expression at the right atrium between our groups using the western blot technique. Figure 6A shows representative β1AR expression from STZ-AD and CTL groups. Antibodies against β1AR recognize a prominent band at 75 kDa and two additional faint bands at 80 and 50 kDa. The predicted molecular mass of the rat β1AR is about 50 kDa [36]; however, posttranslational modification and SDS-resistant dimerization can affect the size of the mature receptor in cardiomyocytes, as previously described for β1AR [37]. The normalized group data indicated that the expression of β1AR was not altered between groups. We further assessed whether an altered muscarinic acetylcholine receptor M2 (mAChR-M2) expression at the heart contributed to the observed change in autonomic balance under resting conditions (Fig. 6B). Antibodies against mAChR-M2 detected a single band at almost 50 kDa, as previously shown [38]. Although the group data of mAChR-M2 expression did not show a significant difference between groups, we found a high variance in the STZ-AD group data when compared with CTL (CTL, σ² = 0.03, n = 6 versus STZ-AD, σ² = 0.88, n = 6, p = 0.003). Changes in expression for mAchR-M2 may indicate a compensatory response to the shift in autonomic input.

Fig. 6.

No change in autonomic receptor expression at the heart. A, B) Western blot analysis of beta-1 adrenergic receptor (β1AR, n = 7/group) (A) and muscarinic acetylcholine receptor M2 (mAChR-M2, n = 6/group) (B) protein levels in right atria between CTL and STZ-AD. β-actin served as loading control.

Unaltered sympathetic influence on the vasculature in STZ rats

Lastly, we wanted to elucidate the sympathetic influence on peripheral vascular resistance by delivering a bolus of the preganglionic blocker hexamethonium (subsequent to atropine and atenolol infusion). Figure 7A depicts the typical raw tracings of MAP and HR in response to hexamethonium. After delivery of a hexamethonium bolus, MAP decreased strongly in both groups (CTL, 46±2 mmHg, n = 8 versus STZ-AD, 52±3 mmHg, n = 10, p = 0.14). However, there was no difference in the magnitude of MAP change between groups (Fig. 7B), suggesting no change of autonomic output to the vasculature in STZ-AD rats. Furthermore, there was a negligible small change in HR between groups (CTL, 3.6±1.0%, n = 8 versus STZ-AD, 3.0±1.0%, n = 10, p = 0.47), suggesting that our previous autonomic blockade (with atropine and atenolol) to the heart was sufficient (Fig. 7C). After complete blockade, there was a trend for a lower intrinsic HR in the STZ-AD group (by ~13%) than in the CTL group (Fig. 7D, p = 0.069). The slight reduction in intrinsic HR may indicate altered activity of the pacemakers in the sinoatrial node.

Fig. 7.

Influence of STZ-AD on the vasculature and intrinsic HR. A) Examples of typical mean arterial pressure (MAP) and heart rate (HR) responses to ganglionic blockade using hexamethonium (following prior boluses of atropine and atenolol) in CTL and STZ-AD groups. B-D) Group data showing comparisons of ∆MAP (B), ∆HR (C), and intrinsic HR (D) to ganglionic blockade between groups. CTL, n = 8 and STZ, n = 10.

DISCUSSION

In the present study, we present novel data for a profoundly blunted baroreflex function in the STZ-AD animal model. STZ-AD rats were unable to increase HR in response to low blood pressure. The absent HR response was most likely due to an overall decrease of central sympathetic outflow (lower ceiling response), since peripheral expression of β1AR and mAChR-M2 at the heart was unchanged. The balance of autonomic outflow to the heart was strongly shifted toward increased sympathetic tone in the STZ-AD model, which kept a seemingly normal HR at rest. Our results provide the first insights into the autonomic role behind baroreflex dysfunction in an AD animal model, elucidating possible mechanisms for the orthostatic hypotension seen in human AD.

The human autonomic nervous system is complex, and there is no single test that completely describes its overall functional state [39]. Therefore, most studies include a series of indirect tests (e.g., skin response, deep breathing, Valsalva maneuver, head-up tilt, HR variability) for a more complete picture of the sympathetic and parasympathetic branches of the autonomic nervous system. Although clinical dysautonomia in AD occurs at various degrees, a rather consistent finding in human AD is the orthostatic hypotension experienced during transition from supine to a standing position [6, 40, 41]. This finding is in agreement with a blunted sympathetic response to low blood pressure on standing [42] and corroborates other data showing impaired sympathetic outflow when challenged during autonomic function tests [2, 7]. Similar to the orthostatic hypotension in human AD, we found a severe blunting of sympathetic nerve activity in response to low MAP in the STZ-AD animal model, as shown by a significantly reduced ceiling response of HR and sSNA. STZ-AD rats already operated at maximal HR and sSNA under resting condition (at SP), disabling a potential compensation for lower MAP (i.e., loss of reserve capacity).

Under resting condition, studies indicate a shift towards greater sympathetic tone with either normal or suppressed parasympathetic activity in AD patients [6, 41, 43]. Baseline HR of AD patients may be normal [44] or slightly lower [7] than healthy age-matched controls. We also saw a shift towards higher sympathetic tone (within its lower limits) and away from parasympathetic modulation of the heart at resting condition in the STZ-AD model. This outcome was corroborated by our spectral analysis of HR intervals, autonomic blocker experiments, and baroreflex curves for HR and directly for sSNA. Altogether, the symptoms of the STZ-AD animal model shown here mimic the findings of autonomic dysfunction in human AD studies and may thus constitute a suitable model to further identify the neuronal mechanisms behind orthostatic hypotension in AD.

The neurophysiological mechanisms behind orthostatic hypotension in AD are currently unknown. It has been previously shown that brainstem nuclei involved in the classic baroreflex, namely the nucleus tractus solitarii, nucleus ambiguous, caudal ventrolateral medulla, and rostral ventrolateral medulla [45], are affected by amyloid-β and tau tangles in patients with AD [46]. Although these pathologies are seen at various times in the progression of AD, they are associated with, and likely preceded by inflammation and increased reactive oxygen species, as shown in brain tissue from human AD [47, 48] and the STZ-AD animal model [13, 49]. Our study specifically chose to analyze baroreflex function in the STZ-AD model two weeks after model induction in order to assess the impact of these early symptoms (inflammation and oxidative stress) before amyloid beta and tau tangles have established. It has been previously shown that persistent inflammation and increased reactive oxygen species are already able to induce neuronal hyperactivity in nuclei important for the classical baroreflex pathway, like the nucleus tractus solitarii [50, 51], and thus are able to profoundly alter baroreflex function [52, 53]. Potential hyperexcitability in cardiovascular nuclei may limit the overall responsiveness to cardiovascular perturbations (ceiling effect), which may have a role in the baroreflex dysfunction observed in our study. Neuronal hyperexcitability has been shown in cortical regions of human AD patients [54], and mechanisms may be similar for the brainstem. Whether this is the case in the STZ-AD animal model needs to be examined in future studies.

Changes of sympathetic and parasympathetic input to the heart can trigger cardiac remodeling that increases the risk of heart failure [55–57]. In the present study, we analyzed cardiac remodeling on the level of the autonomic receptors that are primarily found at the heart, namely β1AR and mAChR-M2 [58, 59]. Expression of both receptor types was similar between groups, which may indicate that the baroreflex dysfunction in STZ-AD was to a large part caused by central changes in the brain. However, we noticed significant variance in mAChR-M2 expression. This variance may indicate the beginning of a compensatory response of the heart to the altered autonomic input. Since our experiments occurred early (2 weeks) after STZ-AD model induction, the compensation may be further developed at a later time. Similarly, a trend for a lower intrinsic HR in STZ-AD may indicate a possible adaptive response. In humans, a reduced intrinsic HR with aging is coupled with a decreased maximal HR [60]. Thus, a possible reduced intrinsic HR in the STZ-AD animal model may contribute to the overall shift of the baroreflex curve to a lower HR. This contribution, however, is likely minimal, and the drastic drop of sympathetic nerve activity is probably responsible for the shift of the baroreflex curve. In human AD, it is currently unknown whether the altered autonomic balance influences the heart, but heart failure is one of the causes of death in human AD [61]. However, to date, cardiac remodeling in AD has only been studied in regard to amyloid-β aggregation at the heart [62, 63].

Isoflurane anesthesia has been shown to alter HR, MAP, and baroreflex function [64–66] through changes of baroreceptor discharge, postganglionic nerve activity, calcium currents in the sinoatrial node, and responses of small resistance arteries [64, 67, 68]. These changes were more pronounced with depth of anesthesia. In the present study, after initiation of anesthesia, we reduced isoflurane delivery to a concentration less than or equal to 2%, which was just at the surgical plane. Although depression of MAP and maximal HR responses cannot be fully excluded with this isoflurane dosage [65], significant changes of baroreflex function should not be evident until 2.6% isoflurane [64]. Furthermore, CTL and STZ-AD animals received the same isoflurane dosage and any depression of reflexes should have been equal between groups.

In conclusion, the present study provides the first data on the contribution of the autonomic nervous system to baroreflex dysfunction in the STZ-AD animal model. Blunted HR responses to low blood pressure in STZ-AD was caused by a reduced sympathetic outflow (ceiling response) and is consistent with orthostatic hypotension frequently found in human AD. Our current data further expand on the versatility of the STZ-AD rat model to mimic symptoms of AD patients and may serve as an important tool to test future therapeutic approaches to alleviate comorbidities with AD.