Blunted Respiratory Responses in the Streptozotocin-Induced Alzheimer’s Disease Rat Model

Animals

Male Sprague-Dawley rats (Harlan; n = 28, 272±8 g) were maintained in the AAALAC accredited vivarium of the ATSU Kirksville College of Osteopathic Medicine. Rats were held on a 12-h day/night cycle at 24°C and 46% humidity with water and food available ad libitum. A subset of experiments was conducted at the University of Missouri and rats were held under the same conditions in the AAALAC accredited vivarium of the Dalton Cardiovascular Research Center. All experimental protocols are in accordance with NIH guidelines (“Guide for the Care and Use of Laboratory Animals”) and approved by the Animal Care and Use Committee of A.T. Still University and University of Missouri.

Intracerebroventricular injection of streptozotocin

Similar to previous studies [20, 23], STZ was injected into both lateral ventricles of the brain. We used a sub-diabetogenic dose of STZ [24, 25] as indicated by peripheral blood glucose levels that were comparable to control rats (CTL) injected with artificial cerebrospinal fluid (aCSF/vehicle, n = 6, 85.5±5.3 mg/dL blood versus STZ, n = 5, 98.8±10.7 mg/dL blood; p = 0.27). Blood glucose levels were tested immediately before euthanasia using a commercially available blood glucose tester (TRUEresult, Nipro diagnostics). For ICV injections, rats were anesthetized using isoflurane (Piramal; 5% induction, 2% maintenance) and positioned in a stereotaxic frame (David Kopf Instruments). Before injections, rats received 2 mg/kg Dexamethasone (Med-Pharmex; immunosuppressant) to prevent possible brain swelling. A midline sagittal incision was made in the scalp to expose the coronal and sagittal suture of the skull. Two burr holes were drilled into the skull with a rotary tool (Dremel 7300 with engraving cutter 105) to allow access to the lateral ventricles. Injection glass capillaries (World Precision Instruments) were pulled using a P-97 micropipette puller (Sutter Instruments) and used for ICV-injections according to the following stereotaxic coordinates: –0.9 mm AP,±1.5 mm ML, and 3.6 mm DV [26]. Figure 1 shows the typical injection site and injection canal through the cortex marked with 2% methylene blue. 25 mg/mL STZ was dissolved in aCSF (in mM: 123 NaCl, 1.1 CaCl₂, 3 KCl, 1.9 MgCl₂, 25 NaHCO₃, 0.5 NaH₂PO₄, and 0.25 Na₂HPO₄, pH 7.3) and pressure-injected at a variable volume (<10 μL per side) to achieve a dose of 1.5 mg STZ per kg rat. Control animals received vehicle only. The wound was sutured shut using absorbable suture (4-0, Oasis, MV-J397-V) and rats were postoperatively treated with 50 μg/kg Buprenorphine hydrochloride (Sigma) for pain management, 7 mg/kg Baytril (Bayer) for antibiotic treatment, and 3 mL 0.9% sodium chloride solution (Hospira) for fluid reconstitution. The surgical procedure for a second ICV injection was repeated 2 days later. Within the first week following surgery, rats were supplemented with sweetened cereal (Froot loops, Kellogg’s) for weight recovery.

Passive avoidance memory test

Similar to previous studies [27, 28], we validated our model to confirm memory decline in a subset of rats using the passive avoidance step-through paradigm. The avoidance system (GEMINI avoidance system, San Diego Instruments) consisted of a light and a dark chamber separated by a guillotine door. In the acquisition trial 14 days following ICV injections, rats were placed in the light chamber. Following 2 min of habituation, the guillotine door was raised and the time recorded until the rat entered the dark chamber. There was no difference in initial crossing latency between CTL and STZ rats (CTL, n = 5, 62±21 s versus STZ, n = 3, 34±10 s; p = 0.786). Immediately after entering the dark chamber, rats received a 3-s electric foot shock (0.4 mA) though the metal floor grids and were then transferred back to their cage. In the retention trial 24 h later, rats were subjected to the same procedure (without the electric foot shock). All CTL rats remained in the light chamber for more than 300 s (cut-off time). On the other hand, all STZ rats showed impaired recall latency as indicated by a significant shorter crossing latency into the dark chamber (CTL, n = 5, 300±0 s versus STZ, n = 3, 126±77 s; p = 0.036).

Immunohistochemistry

Similar to our previous manuscripts [29, 30], immunohistochemistry was performed in a subset of rats two weeks following the second ICV injection. Rats were deeply anaesthetized with isoflurane and transcardially perfused with heparinized 0.01 M phosphate-buffered saline (PBS; in mM: 2.7 NaH₂PO₄, 7.7 Na₂HPO₄, and 154 NaCl, pH 7.4) that was followed by 4% paraformaldehyde (PFA, Acros organics) in PBS. The brain was removed, post-fixed overnight, cut into 30 μm thick coronal sections using a vibrating microtome (7000smz-2, Campden Instruments), and collected either in a 1/6 series (brainstem; 24 well-tissue culture plate) or 1/4 series (forebrain; 12 well-plates). Sections were washed in PBS and blocked for 30 min in 10% normal donkey serum (NDS, Millipore) with 0.3% Triton-PBS. Subsequently, brain sections were incubated in 1% NDS with 0.3% Triton-PBS and the primary antibody against S100 calcium-binding protein B (S100B; astroglial marker; rabbit, 1:200, ab41548, Abcam). The following day, sections were rinsed in PBS and incubated for 2 h in 1% NDS with 0.3% Triton-PBS and the secondary antibody Alexa Fluor 488-conjugated donkey anti-rabbit IgG (1:200; 711-545-152, Jackson Immuno). Brain sections were mounted on gelatin-coated glass slides, air dried for 1 h, and cover-slipped using ProLong Diamond with 4’,6-diamidino-2-phenylindole (ThermoFisher Scientific). One section was incubated in the absence of the primary antibody and showed no fluorescent staining (negative control). While all hippocampal sections were simultaneously labeled against S100B using the same solutions, tissue sections from the brainstem were processed in pairs consisting of one aCSF control and one STZ rat. Specificity of the S100B antibody was previously confirmed [31].

Immunoreactivity was generally visualized using a conventional epifluorescent microscope (Eclipse 80i, Nikon), a digital monochrome camera (DS-Qi1Mc, Nikon), and appropriate filter sets and excitation wavelength for the fluorophore. Figure 2D is a maximum projection of 8 images taken in a z-stack (0.5 μm) with a 63x (oil) objective on a spinning disc confocal microscope (DMI6000 B, Leica) with CCD camera (ImagEM Enhanced, Hamamatsu). Exposure times were kept identical for all hippocampal section or tissue that was processed in pairs to allow comparison between both treatment groups. Images were post-processed using ImageJ (Version 1.49t, NIH) by rescaling pixel size (bicubically) and adjusting contrast and brightness (identical histogram range for all slice pair comparisons). The clarity of Fig. 2D was enhanced by deconvolution (10 iterations, blind PSF, medium noise) using AutoQuant X3 (Media Cybernetics). According to the fluorophore of the secondary antibody images were given the pseudo-color green. Templates shown in Figs. 1, 2, and 3 were drawn according to coronal brain slices using the freeware program Inkscape (version 0.91; http://www.inkscape.org/). The position from Bregma and location of some brain structures were determined using the rat atlas from Paxinos & Watson [26]. In each animal, one corresponding brain section per brain region (hippocampus, nTS, pre-Bötzinger complex) was used to analyze astrocyte number and process density. Astrocyte numbers were manually counted by two blinded individuals in a 330×330 μm box for the CA1 region of the hippocampus, a 100×100 μm box for the commissural nTS, and a 150×150 μm box for the pre-Bötzinger complex using ImageJ. Box sizes were chosen to include most of the target area while excluding potential differences in nucleus size. S100B staining of astrocytic processes was quantified as mean staining intensity in ImageJ using the original grey-scale images. Measurements were taken adjacent to the cell bodies (see Fig. 2A & D) and represent an average of multiple measurements within the nucleus of interest for each animal. In the hippocampus, each average consists of measurements using 30×30 μm boxes at three different locations. Five measurements using 10×10 μm boxes were used for the other nuclei. Staining intensity from tissue that was processed in pairs was normalized to aCSF control to allow group comparison.

Plethysmography

Prior to any surgical procedures (pre-aCSF/STZ) and 14 days following the second ICV-injection (post-aCSF/STZ), respiratory parameters were recorded in conscious unrestrained rats using a whole-body plethysmography chamber (Data Sciences International). Pressure changes from respiration of the rat inside the test chamber were compared to those in a reference chamber and measured with a differential pressure transducer (DP45, Validyne). The test chamber was continuously flushed with bias gas flow of 3 L/min, making the recorded pressure signal directly proportional to changes in air flow inside the chamber. The pressure signal was demodulated (Sine wave carrier demodulator CD15, Validyne), digitized (PowerLab 8/35, AD Instruments), and recorded via LabChart Software (AD Instruments). Rats were exposed to different gas mixtures by blending pure gases using two mass flow controllers (MC-5SLPM-D) and control software (Flow Vision MX) with hardware interface (RS-232 Multi-Drop, Model BB9) from Alicat Scientific. For the experiments conducted at the University of Missouri pure gases were mixed using the gas blender Hypoxydial (Starr Life Sciences). The other devices utilized for plethysmography were similar at both institutes. Rats were acclimated to the plethysmography chamber for 2 days (1 h each) prior experimental protocols. On the day of the experiment, gas mixtures were given (following stabilization of breathing at 21% O₂) for 7 to 10 min to allow adequate time for gas exchange within the chamber and chemoreflex responses. To stimulate the peripheral or central chemoreflex, rats were exposed to hypoxia (14, 12, 10, and 8% O₂, equilibrated with N₂; following 21% O₂) or hypercapnia (5% CO₂, equilibrated with O₂; subsequent to 20 minutes of 100% O₂ to eliminate influence of the peripheral chemoreflex), respectively. Data were analyzed using LabChart and Microsoft Excel Software. Respiratory rate (breath per minute) and tidal volume (normalized to 100 g rat) were measured during steady and uninterrupted breathing (in the absence of movement, sniffing, and sighs) within the last 30 s of a gas mixture. Minute ventilation (per 100 g rat) was calculated as the product of respiratory rate and tidal volume/100 g. Rate and volume of sighs (i.e., augmented breaths; per 100 g rat) were analyzed within the last 3 min of a given gas mixture. Sigh and tidal volume were calculated as the area under the inspiratory pressure curve and by using a formula previously described to correct for barometric pressure, temperature, and water vapor pressure inside the animal and plethysmography chamber [32]. Volume measurements were calibrated to a pressure change of a known volume of air injected into the chamber using a rodent respirator (Harvard Apparatus, Model 680).

Statistical analysis

Statistical analyses were performed using SigmaPlot 13.0 (Systat Software). Immunohistochemical data and percent changes during the central chemoreflex were analyzed using paired t-tests. Data from the passive avoidance test were compared using the Mann-Whitney rank sum test. All other data were compared by two-way repeated measures ANOVA (with Student–Newman–Keuls post hoc test). Results are considered significantly different at p values≤0.05. Group data are presented as means±SEM.

STZ-injections induce astrogliosis in the hippocampus and commissural nTS

Reactive astrocytes (astrogliosis) are highly associated with clinical AD [33]. An increase in astroglia has also been previously demonstrated in the hippocampus of the STZ-induced AD rat model using immunoblots [20]. To verify that similar alterations occur in this study, we immunohistochemically labeled astrocytes in the hippocampus using S100B antibody. Figure 2A shows a typical staining pattern of astrocytes in the CA1 region of the hippocampus (box in Fig. 2B) from aCSF-injected control (CTL) rats and STZ-injected rats. There was no statistical difference in the number of astrocytes between the CTL and STZ rat group (Fig. 2C). However, since astrogliosis is mainly characterized by an increased thickening and number of cellular processes [34, 35], we measured S100B staining intensity adjacent to the cell bodies. Typically, the network of astrocytic processes in the CA1 region had a higher density in STZ rats when compared to CTL (Fig. 2D). This was supported by a significant increase of S100B staining intensity in the group data (Fig. 2E).

To identify if astrogliosis also affects respiratory centers in the brainstem, we analyzed S100B immunoreactivity in the commissural nTS and pre-Bötzinger complex representatively for the dorsal and ventral respiratory group (Fig. 3). Similar to the hippocampus, the commissural nTS of STZ rats exhibited a consistently higher degree of S100B staining intensity when compared to CTL rats, while the number of astrocytes was unchanged (Fig. 3A-C).The pre-Bötzinger complex (Fig. 3D, E), again, showed no change in astrocyte number (Fig. 3F left). However, and in contrast to the commissural nTS and hippocampus, S100B staining intensity was not altered (Fig. 3F, right).

Increased resting respiration with ICV STZ

Astrogliosis within the commissural nTS may be associated with alteration of respiratory parameters. Thus, we analyzed respiratory rate, tidal volume, and minute ventilation using whole-body plethysmography in conscious rats. Baseline values (at 21% O₂) before (pre-) and two weeks following (post-) ICV injections of aCSF or STZ are given in Table 1. While most parameters remained constant within the two treatment groups, resting minute ventilation fell significantly in CTL rats over time. This decrease is similar to the reduction of ventilation with rat age when adjusted for body weight [36–38]. STZ rats, however, did not show the time-dependent decrease in ventilation. These results indicate increased respiration at baseline in response to STZ treatment.

Peripheral chemoreflex responses are bluntedin STZ rats

To evaluate possible alterations in the peripheral chemoreflex, we analyzed respiratory parameters in response to hypoxia (14–8% O₂). Reducing environmental O₂ content within the plethysmography chamber reliably elevated respiratory rate in CTL and STZ rats (p≤0.001, n = 8 for each group). The representative example shown in Fig. 4A exhibited a maximum respiratory rate at 10% O₂ that slightly dropped with reduction of O₂ to 8%. These changes in respiratory rate are similar in both rat groups prior to any intervention (open circles/squares Fig. 4B, top). Following ICV injections, post hoc analysis revealed a significant reduction of respiratory rate in the STZ rat, whereas CTL rats did not change (closed circles/squares Fig. 4B, top). Tidal volume, on the other hand, did not change over most hypoxic severities tested (Fig. 4B, middle). At 8% O₂, however, both rat groups had significantly increased tidal volume before treatment as shown by post hoc analysis (p≤0.01, n = 8 for each group, against all other hypoxic severities). Two weeks after ICV injections this significant increase in response to hypoxia was absent in the STZ-treated group but not the control group. Minute ventilation was steadily rising with increasing hypoxia (Fig. 4B, bottom; p≤0.001, n = 8 for each group). In the STZ rat group this rise in minute ventilation was significantly blunted after STZ treatment. CTL rats were similar pre- and post-aCSF treatment.

Normalizing the data sets from Fig. 4B to their baseline values at 21% O₂ allows assessment of chemoreflex gain independent of the changes in resting ventilation over time that were observed in CTL rats. Fig. 5 shows the difference between pre- and post-injection of normalized respiratory rate, tidal volume, and minute ventilation during hypoxia for CTL and STZ rats. Even though there were no statistically significant changes between rat groups and hypoxic severities for respiratory rate and tidal volume (Fig. 5, top and middle), respiratory rate in CTL rats was generally increased at post-injection. The response of AD rats, on the other hand, was reduced by more than 20% at 14% O₂. This decrease gradually disappeared with increasing severity of hypoxia. Tidal volume remained stable during 14–10% O₂ in both rat groups. At 8% O₂, however, the response of STZ rats showed a decrease of about 30% in comparison to the pre-injection data. Minute ventilation (Fig. 5, bottom) of CTL rats was increased to each hypoxia tested. Responses of STZ rats, in contrast, were consistently decreased by more than 20% for each hypoxic severity tested. At 10% and 8% O₂ this change in minute ventilation was significantly different from control, indicating a progressive blunting of chemoreflex gain with increased severity of hypoxia.

The central chemoreflex shows little impairment in STZ rats

With the STZ-induced blunting of ventilatory response to hypoxia, central chemoreflex responses to hypercapnia may also be altered. This was tested by increasing the CO₂ concentration (5% CO₂ equilibrated with O₂) subsequent to 20 min of 100% O₂.The high concentration of O₂ was given to minimize any influence of the peripheral chemoreflex. Prior to aCSF or STZ injections, hypercapnia induced a robust increase in breathing rate, tidal volume, and minute ventilation (p≤0.01, n = 5 for each group) as shown by the raw values in Fig. 6A (open circles/squares). These responses did not change following aCSF injection. STZ rats showed a significantly reduced respiratory rate during 5% CO₂ when compared to its baseline value (i.e., before STZ injection). This decrease in respiratory rate in combination with unchanged tidal volume, however, did not alter minute ventilation. Normalizing these data to 100% O₂ (Fig. 6B) shows minimal changes in both rat groups for respiratory rate, tidal volume, and minute ventilation.

STZ rats demonstrate augmented sigh amplitude with hypoxia

Sighs are augmented breaths that occur intermittently during normal breathing to improve lung oxygenation by reopening collapsed alveoli [39]. Sigh frequency generally increases with hypoxia [40] and thus is indicative of the hypoxic state of the animal. We analyzed sigh frequency and volume in the CTL and STZ rat. Fig. 7A shows a typical example of intermittent sighs during the breathing rhythm when exposed to hypoxia. Similar to previous reports [41], sighs can be distinguished from regular breaths by its augmented amplitude and its biphasic shape (Fig. 7A, bottom). The initial phase (I) appears similar to the inspiratory part of a regular breath. In phase II, the amplitude of inspiration is greatly extended. The mean data shows that sigh frequency increases with the severity of hypoxia (Fig. 7B, top open circles/squares; p≤0.001, n = 8 for each group). Neither treatment with aCSF nor STZ changed the magnitude of this relationship (Fig. 7B, top closed circles/squares). Sigh volume, in contrast, generally decreased with hypoxic severity (Fig. 7B, bottom; p≤0.001, n = 8 for each group). Following ICV injection, the fall of sigh amplitude to hypoxia was significantly attenuated in STZ rats.