Synaptic Loss and Gliosis in the Nucleus Tractus Solitarii with Streptozotocin-Induced Alzheimer’s Disease

Abstract

Obstructive sleep apnea is highly prevalent in Alzheimer’s disease (AD). However, brainstem centers controlling respiration have received little attention in AD research, and mechanisms behind respiratory dysfunction in AD are not understood. The nucleus tractus solitarii (nTS) is an important brainstem center for respiratory control and chemoreflex function. Alterations of nTS integrity, like those shown in AD patients, likely affect neuronal processing and adequate control of breathing. We used the streptozotocin-induced rat model of AD (STZ-AD) to analyze cellular changes in the nTS that corroborate previously documented respiratory dysfunction. We used 2 common dosages of STZ (2 and 3 mg/kg STZ) for model induction and evaluated the early impact on cell populations in the nTS. The hippocampus served as control region to identify site-specific effects of STZ. There was significant atrophy in the caudal nTS of the 3 mg/kg STZ-AD group only, an area known to integrate chemoafferent information. Also, the hippocampus had significant atrophy with the highest STZ dosage tested. Both STZ-AD groups showed respiratory dysfunction along with multiple indices for astroglial and microglial activation. These changes were primarily located in the caudal and intermediate nTS. While there was no change of astrocytes in the hippocampus, microglial activation was accompanied by a reduction in synaptic density. Together, our data demonstrate that STZ-AD induces site-specific effects on all major cell types, primarily in the caudal/intermediate nTS. Both STZ dosages used in this study produced a similar outcome and can be used for future studies examining the initial symptoms of STZ-AD.

1. Introduction

Alzheimer’s disease (AD) is a neurological pathology commonly associated with memory and cognitive decline (Alzheimer’s Association, 2020; Folstein and Whitehouse, 1983). Sleep-disordered breathing is another frequent symptom of AD (Hoch et al., 1986), and moderate to severe obstructive sleep apnea (OSA) is found in about 50% of patients with AD (Gaig and Iranzo, 2012). The high prevalence of OSA in AD leads to poor quality of life and increased mortality (Khattak et al., 2018; Partinen et al., 1988). Although respiratory problems are common with AD, the mechanisms behind this impairment are unknown.

Brain atrophy, amyloid beta plaques, and hyperphosphorylated tau protein are linked to AD and contribute to many of the neurological deficits seen in AD (Ingelsson et al., 2004; Kesslak et al., 1991). These pathological changes are also found in brainstem structures that house critical neuronal networks for the control of breathing (Grinberg et al., 2011; Lee et al., 2015; Parvizi et al., 2001; Simic et al., 2010). The nucleus tractus solitarii (nTS) in the dorsomedial medulla is an important upstream center for respiratory control and essential for integration of chemoafferent information during hypoxemia (e.g., from OSA) (Cutsforth-Gregory and Benarroch, 2017; Zoccal et al., 2014). Any disturbance in nTS processing will alter respiratory responses to hypoxia (Favero et al., 2011), including the AD-related disorders found in the nTS of brains from patients with AD (Parvizi et al., 2001).

The streptozotocin-induced rat model of AD (STZ-AD) mimics many symptoms from human AD (Salkovic-Petrisic et al., 2013). We have previously shown that the STZ-AD model also mimics respiratory dysfunction when exposed to hypoxia (Ebel et al., 2017). Concomitantly, there is a lower neuronal activation to hypoxia in the caudal/intermediate aspect of the nTS (Brown et al., 2019), which is the primary integration site of peripheral chemoafferent information (Housley et al., 1987). For memory-related forebrain structures (e.g., hippocampus and cortex), cellular changes, including cell loss, have been shown to account for poor memory performance in the STZ-AD model (Knezovic et al., 2015; Shoham et al., 2003). Whether morphological and cellular decline in the nTS corroborates the respiratory dysfunction in STZ-AD is currently unknown.

The goal of the current study was two-fold, to determine early histopathological alterations in the nTS of the STZ-AD model and to compare the degree of disease expression with 2 commonly used streptozotocin (STZ) concentrations (2 and 3 mg/kg) for AD model induction.

2. Results

2.1. Impact of STZ on forebrain morphology

We first analyzed dose-dependent changes of intraventricular STZ injections (icv-STZ) on forebrain morphology 2–3 weeks following model induction. Forebrain regions have been previously described in the STZ-AD model and serve as an important control for the data obtained in the brainstem. We chose 2 STZ dosages typically used for induction of this model and compared the effect of 2 mg/kg STZ with that of 3 mg/kg STZ (two 1.5 mg/kg doses given 2 days apart). Figure 1a shows the area of the lateral ventricles in the STZ-treated groups when compared with vehicle control. Similar to patients with AD (Ertekin et al., 2016; Ferrarini et al., 2006), icv-STZ tended to enlarge the ventricle space for both concentrations tested. The 3 mg/kg STZ-AD group had the strongest increase of ventricle area and was significant different when directly compared with the control group (CTL, 5.2 ± 0.8 mm² vs. 3 mg/kg STZ-AD, 8.4 ± 1.0 mm²; n = 8/group; p = 0.03, t-test). Enlargement of the ventricles can be explained by the significantly reduced size of the septum (Table 1). Similarly, the hippocampal area had a dose-dependent reduction in size (Figure 1b, significant change for the 3 mg/kg STZ-AD group only). This early decrease in hippocampal volume (2–3 weeks following AD model induction) likely contributes to the memory loss previously shown by us and others using these STZ concentrations (Sachdeva et al., 2015; Vicente et al., 2020). The size of other structures in the ventricle’s vicinity were unaffected (Table 1), indicating a site-specific effect of STZ on the septum and hippocampus.

Figure 1:

Gross morphological changes in the forebrain of the STZ-AD rat model. a-b) Effect of intracerebroventricular injections of 2 and 3 mg/kg STZ on the lateral ventricles (a, left, red shading) and hippocampal area (b, left, red shading). Group data (right) depicts bilateral measurements. Ct, cortex; Cp, striatum; Sp, septum; Th, thalamus. * = p ≤ 0.05 vs. CTL, 1-way ANOVA (n = 8/group).

Table 1:

Differential effect of STZ-AD on paraventricular brain areas. Group data depicts bilateral measurements (n = 8/group).

Group	Septum [mm2]	Striatum [mm2]	Cortex [mm]	Thalamus [mm2]
CTL	5.15 ± 0.31	19.12 ± 0.47	1.33 ± 0.02	20.77 ± 0.24
2 mg	3.68 ± 0.51 *	17.80 ± 0.32	1.38 ± 0.02	21.20 ± 0.63
3 mg	2.77 ± 0.31 ***	18.98 ± 0.48	1.33 ± 0.02	21.27 ± 0.68

^{* =}p ≤ 0.05,

^{*** =}p ≤ 0.001, vs. CTL, 1-way ANOVA

2.2. General morphology of the nTS and breathing dysfunction in the STZ-AD model

To identify possible mechanisms for the respiratory problems frequently experienced by patients with AD, we wanted to elucidate the effects of 2 and 3 mg/kg STZ on the general morphology of the nTS in the brainstem. Figure 2a shows examples of coronal sections that illustrates the nTS (highlighted in red) in relation to its caudal-rostral extent within the brainstem. Bilateral measurements of nTS area in our 3 groups are summarized in Figure 2b. Although the nTS area of the 2 mg/kg STZ-AD group did not differ from CTL, there was a significant reduction in the caudal aspect of the nTS with 3 mg/kg STZ. This reduction was significantly different from CTL and the 2 mg/kg STZ-AD group, indicating a stronger effect on nTS morphology with 3 mg/kg STZ. Morphology of the intermediate and rostral nTS was unaffected.

Figure 2:

Gross morphological changes of the nucleus tractus solitarii (nTS) in STZ-AD. a) Schematic representation of the nTS within the brainstem (left, red shaded area) and coronal example sections depicting sub-regions of the nTS (right, unilateral red shading). b) Group data for the nTS area (bilateral measurements) of vehicle control (CTL, n = 7), 2 mg/kg STZ-AD (n = 8), and 3 mg/kg STZ-AD (n = 7) groups. Note, a significant decrease of nTS size was seen in the caudal area after injections of 3 mg/kg STZ. AP, area postrema; CS, calamus scriptorius; V, ventricle. * = p ≤ 0.05 vs. CTL; ^# = p ≤ 0.05 vs. 2 mg/kg STZ, 2-way repeated ANOVA.

The caudal nTS is important for processing information from the peripheral chemoreflex (Favero et al., 2011; Zoccal et al., 2014), and decreasing brain mass in this area correlates with previously documented chemoreflex dysfunction in STZ-AD animals (Brown et al., 2019; Ebel et al., 2017). Figure 3 shows comparisons of the respiratory responses of the 2 mg/kg and 3 mg/kg STZ-AD groups during 10% hypoxia. Exposing the rats to 30 minutes of hypoxia in a plethysmography chamber reliably increased respiratory rate (RR; Fig. 3a). The group data shows that this increase in RR was significantly blunted with both STZ concentrations when compared to their controls (Fig. 3b). The %-increase of RR was similar between STZ-AD groups (2 mg/kg STZ-AD, 43.4 ± 7.6% vs. 3 mg/kg STZ-AD, 43.8 ± 3.4%; n = 6/group; p = 0.97, t-test). Although tidal volume (TV) did not appreciably change in any experimental group (Fig. 3c), the 2 mg/kg STZ-AD group had a slight, non-significant increase that may compensate for the blunted RR. This partial compensation rendered minute ventilation (MV, Fig. 3d) similar to CTL. On the other hand, the 3 mg/kg STZ-AD group had a significant reduction of MV under hypoxia. Comparing the %-increase of MV between STZ-AD groups revealed a trend for greater blunting of MV in the 3 mg/kg STZ-AD group (2 mg/kg STZ-AD, 68.5 ± 13.5% vs. 3 mg/kg STZ-AD, 37.4 ± 8.7%; n = 6/group; p = 0.081, t-test).

Figure 3:

STZ-AD induces chemoreflex dysfunction during hypoxia. a) Example respiratory traces of baseline breathing (21% O₂, top trace) and increased breathing frequency during exposure to hypoxia (10% O₂, bottom trace) of control (CTL) and STZ groups. b-d) Group data for respiratory rate (b), tidal volume (c), and minute ventilation (d) of CTL (white squares), 2 mg/kg STZ-AD (blue squares), and 3 mg/kg STZ-AD (red squares) groups. Although the hypoxia-induced increase in respiratory rate was blunted for both STZ-AD groups, only the 3 mg/kg STZ dosage had an effect on the minute ventilatory response. * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001 vs. CTL, 2-way repeated ANOVA (n = 6/group).

2.3. Reduced synapse density with STZ-AD

Next, we used an immunohistochemical approach to examine the effect of 2 and 3 mg/kg STZ on different cell populations and morphology in the hippocampus and 3 representative sections of the nTS (caudal = CS −0.36, intermediate = CS +0.54, and rostral = CS +1.62). Our analysis of the neuronal cell population is shown in Figure 4. Although a decline in neuron number was previously shown in the STZ-AD model (Rostami et al., 2017), the density of NeuN-identified neurons did not change between STZ-AD groups and CTL in the hippocampus CA1 region or throughout the caudal-rostral extent of the nTS (Figs. 4a and 4b). Conversely, the density of synaptophysin-identified neuronal synapses decreased significantly in the hippocampus for both STZ concentrations (Fig. 4c). Interestingly, although the 3 mg/kg STZ-AD group had more severe damage to the hippocampal area (Fig. 1b), the 2 mg/kg STZ-AD group had a stronger decline in synaptic density than the 3 mg/kg group. Similarly, the caudal aspect of the nTS in the 2 mg/kg STZ-AD group had a strong trend for lower synaptic density when individually compared with control (p = 0.054, t-test). All other nTS regions showed a synaptophysin intensity similar to our CTL group (Fig. 4d). Collectively, these results indicate that a reduction of synaptic density (significant reduction in the hippocampus, trending in the caudal nTS) may occur earlier than a decline in neuron number in the STZ-AD rat model, and the decline in hippocampal area may be partially explained by a reduction in synaptic density.

Figure 4:

Reduced synaptic density with STZ-AD. a-b) Density of NeuN-identified neurons in the hippocampus (a) and 3 representative sections throughout the rostral-caudal extent of the nucleus tractus solitarii (b; nTS) for CTL (n = 7), 2 mg/kg STZ-AD (n = 8), and 3 mg/kg STZ-AD (n = 7) groups. c-d) Synaptophysin-identified synaptic density is shown for the hippocampus (c) and sections of the nTS (d). ** = p ≤ 0.01, *** = p ≤ 0.001 vs. CTL; ^# = p ≤ 0.05 vs. 2 mg/kg STZ, 1-way ANOVA.

2.4. Region-specific impact of STZ-AD on astrocytes in the nTS

Next, we analyzed the astrocyte population in the hippocampus and nTS and whether STZ dosage had an effect on their density and branching pattern. Using S100B immunostaining, we quantified astrocyte density as shown in Figure 5. Although there were no differences between STZ-AD groups and CTL in the hippocampus (Fig. 5a), both STZ concentrations induced an increased astrocyte density in the intermediate and rostral nTS (Fig. 5b). There was no change in the caudal nTS when compared to CTL. Note, because of the low density of astrocytes in the hippocampus, the size of the analysis window was different between the hippocampus and nTS region; astrocyte densities are therefore not directly comparable between these brain regions as presented in Figure 5.

Figure 5:

Increased astrocyte density in nTS subregions. a) S100B-identified astrocyte density is similar between CTL and both STZ-AD groups in the CA1 region of the hippocampus (n = 7/group). b) Astrocyte density in caudal, intermediate, and rostral nTS sections. * = p ≤ 0.05, ** = p ≤ 0.01 vs. CTL, 1-way ANOVA. n = 7 for CTL, 8 for 2 mg STZ, 7 for 3 mg STZ.

To analyze the morphology of astrocytes, we used GFAP immunoreactivity and analyzed branch length, number of branches, and branch thickness in the hippocampus and sections of the nTS (Fig. 6). Although these parameters were similar among groups in the CA1 region of the hippocampus (Fig. 6a₂), there were significant changes in astrocyte morphology in the nTS. Astrocyte branch length was affected in both STZ-AD groups and primarily decreased in length in intermediate (direct comparison for CTL vs. 2 mg/kg STZ-AD: p = 0.029, t-test) and rostral nTS sections (Fig. 6b₂). This decrease in branch length resembles an atrophic phenotype in the astrocyte population as previously shown during early disease stages (Zhou et al., 2019). Interestingly, caudal nTS sections did not change (2 mg/kg STZ-AD group) or had an increase in branch length (3 mg/kg STZ-AD group) when compared with control (Fig. 6b₂). The number of astrocyte branches also showed a significant increase in the caudal nTS only for the 3 mg/kg STZ-AD group (Fig. 6b₃). The morphological changes in the caudal nTS indicates a hypertrophic and activated state of astrocytes and is associated with later stages of the disease (Zhou et al., 2019).

Figure 6:

Region-specific morphology of astrocytes in STZ-AD. a-b) Typical examples and group data for astrocyte branch length, number, and thickness in the hippocampus (a, n = 7/group) and 3 representative sections throughout the rostral-caudal extent of the nTS (b; n = 7 for CTL, 8 for 2 mg STZ, 7 for 3 mg STZ) for CTL and STZ-AD groups. Yellow squares indicate the region in the magnified image below. Green lines depict anatomical tracings for cell morphology analysis performed using Fiji software (see Methods). * = p ≤ 0.05, ** = p ≤ 0.01 vs. CTL; ^# = p ≤ 0.05 vs. 2 mg/kg STZ, 1-way ANOVA.

2.5. Activation of microglia with STZ-AD

Activation of microglia is often observed in AD patients (Davies et al., 2017). Figure 7 shows our analysis of microglia density in the hippocampus and nTS of our experimental groups. In the hippocampus, only the 3 mg/kg STZ-AD group had a significant increase in microglia number. However, the variability within the group was high. Microglia numbers tended to decrease in the nTS (Fig. 7b), with a slight reduction in cell number for the individual comparison between the 2 mg/kg STZ-AD group and control in intermediate (p = 0.028, t-test) and rostral (p = 0.028, t-test) sections.

Figure 7:

Microglia density in the hippocampus and nTS of CTL and STZ-AD groups. a-b) IBA1-identified microglia density in the CA1 region of the hippocampus (a, n = 7/group) and in caudal, intermediate, and rostral nTS sections (b; n = 7 for CTL, 8 for 2 mg STZ, 7 for 3 mg STZ). Yellow squares indicate the region in the magnified images in Figure 8 for the corresponding group. * = p ≤ 0.05 vs. CTL, 1-way ANOVA.

There were strong morphological alterations of microglia in both STZ-AD groups (Fig. 8). Although average branch length of hippocampal microglia was unaltered, there was a strong reduction in the number of branches that was accompanied by a significant increase in branch thickness for both STZ-AD groups (Fig. 8a₂). Such phenotypic changes are expected from microglia in an activated state (Fernández-Arjona et al., 2017). A similar pattern was observed primarily in the caudal and intermediate nTS of both STZ-AD groups (Fig. 8b). The number of microglia branches again showed a significant decline (Fig. 8b₃). Conversely, branch thickness only tended to increase in the group comparison (significant for the intermediate nTS) (Fig. 8b₄). When directly compared with the control group, significant branch thickening was also found in caudal nTS sections of the 2 mg/kg STZ-AD group (p = 0.016, t-test). Branch length of nTS microglia tended to increase in overall group comparisons (Fig. 8b₂). When individually compared with control, there was a significant lengthening in the caudal (CTL vs. 2 mg/kg STZ-AD: p = 0.022, t-test) and rostral sections (CTL vs. 2 mg/kg STZ-AD: p = 0.021, t-test; CTL vs. 3 mg/kg STZ-AD: p = 0.006, t-test).

Figure 8:

Activated microglia morphology with STZ-AD. a-b) Typical examples and group data for microglia branch length, number, and thickness in the hippocampus (a, n = 7/group) and 3 representative sections throughout the rostral-caudal extent of the nTS (b, n = 7 for CTL, 8 for 2 mg STZ, 7 for 3 mg STZ) for CTL and STZ-AD groups. Magnified images are from the regions outlined with yellow squares in Figure 7 for the corresponding group. Cyan lines depict anatomical tracings for cell morphology analysis performed using Fiji software (see Methods). * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001 vs. CTL; ^# = p ≤ 0.05 vs. 2 mg/kg STZ-AD, 1-way ANOVA.

3. Discussion

The STZ-induced AD rat model displays the respiratory dysfunction that has also been reported in humans with AD (Gaig and Iranzo, 2012). The current study analyzed morphological and cellular alterations in this model that may explain this malfunction in breathing. Data from the nTS are presented in comparison to the memory-related CA1 region of the hippocampus using 2 STZ concentrations (2 and 3 mg/kg) frequently used for induction of this model. The main outcome was a gross morphological atrophy in the hippocampus of both STZ-AD groups, but only the 3 mg/kg STZ-AD group had a significant decline in the caudal nTS area. Further, both STZ-AD groups had multiple indices of microglial activation in the caudal and intermediate nTS. Microgliosis was also evident in the hippocampus and accompanied by a significant reduction in synaptic density. The morphology of nTS astrocytes were mainly affected in the 3 mg/kg STZ-AD group in a region-specific manner throughout the caudal-rostral extent of the nTS, indicating hypertrophic (caudal nTS) and atrophic (intermediate and rostral nTS) morphologies.

3.1. Effect of STZ dosage on AD phenotype

A range of STZ concentrations are used in various studies to induce the STZ-AD rat model. Although early memory dysfunction can be produced with STZ concentrations as low as 0.3 mg/kg (Knezovic et al., 2015), an effective range for measurable histopathological changes in brain tissue seems to be 1–6 mg/kg STZ (with higher dosages split into multiple injections on separate days) (Sharma and Gupta, 2001; Sharma and Garabadu, 2020; Shoham et al., 2003; Sofic et al., 2014; Yin et al., 2016). However, studies directly comparing responses to different STZ concentrations showed an expected dose-dependent effect on the time course and severity of pathological changes (Knezovic et al., 2015; Mehla et al., 2013; Rostami et al., 2017; Zappa Villar et al., 2018). This dose-dependent effect of STZ may resemble different stages in the progression of human AD, where lower STZ dosages have a slower time course developing the same pathologies as those seen earlier with high STZ dosages (Kraska et al., 2012; Rostami et al., 2017). However, there are inconsistent data between different studies and outcomes seem to strongly dependent on the observation period (Knezovic et al., 2015; Mehla et al., 2013). Here, we evaluated the impact of 2 mg/kg (1 injection) and 3 mg/kg (2 injections of 1.5 mg/kg) STZ on respiration and histopathology in the hippocampus and nTS area. From our data, we can conclude that both dosages induced very similar effects, except for absent nTS atrophy and a milder change of astrocyte morphology in the 2 mg/kg STZ-AD group. Our data are consistent with a previous study that tested comparable STZ dosages on memory dysfunction and histopathological markers in the forebrain (Dehghan-Shasaltaneh et al., 2016). They found that both STZ dosages produced effects that were similar in magnitude, but the 3 mg/kg dosage also negatively affected locomotor activity, which somewhat limited evaluation of behavioral data for memory assessment (e.g., Morris water maze). In the current study, 2 mg/kg STZ was the most suitable concentration to induce respiratory dysfunction and ‘mild’ histopathological changes in the absence of nonreversible atrophy in the nTS. Such mild changes may also provide a critical window where intervention may still be possible. Furthermore, a single icv injection provides an important advantage to minimize the surgical stress for the animal.

3.2. Brain atrophy in STZ-AD

Gross morphological changes in human AD brains typically include atrophy of the hippocampus, cortex, and brainstem; disruption of white matter tracts; and ventricular enlargement (Ferreira et al., 2017; Lee et al., 2015; Villain et al., 2010). In particular, medial temporal lobe atrophy and ventricular enlargement seem to be among the earliest changes in AD (Nestor et al., 2008; Risacher et al., 2009). The current study was able to reproduce the findings from previous studies (Kraska et al., 2012; Shoham et al., 2003; Voronkov et al., 2019) showing hippocampal atrophy and increased ventricular space in the STZ-AD rat model. These histopathological changes were more severe with higher STZ concentration. Although we did not see a reduction of cortical thickness, it has been shown to occur at a later time point in the STZ-AD model (Knezovic et al., 2015). Our data on dose-dependent forebrain atrophy serve as an important internal control to evaluate the severity of changes in the nTS from our 2 STZ concentrations. Even though hippocampal atrophy was evident in both STZ-AD groups, only the 3 mg/kg STZ-AD group had a decreased caudal nTS volume. Thus, tissue loss in the nTS seems less severe than in the hippocampus in this model. Most likely, reduced nTS atrophy is not entirely caused by the distance from the forebrain injection site and spread of STZ via the ventricular system. Other forebrain structures (e.g., striatum, thalamus), which are much closer to the injection site than the nTS, showed no atrophy. A similar lack of effect of STZ-AD on striatal volume was previously shown by others (Kraska et al., 2012). Furthermore, a dopaminergic cell group situated directly adjacent to a strongly enlarged third ventricle was ‘spared’ by STZ (Shoham et al., 2003). In line with these results is our previous study that showed a specific and localized effect of STZ on neuronal activation in respiratory and cardiovascular regions of the brainstem (Brown et al., 2019). A possible explanation for STZ-induced pathologies in selected brain regions may be related to the distribution and quantity of glucose transporter 2 (GLUT2), which allows the specific uptake of STZ into cells (Schnedl et al., 1994). GLUT2 is found in the hippocampus and nTS (Arluison et al., 2004) and may contribute to the effects of STZ seen here.

3.3. STZ-induced changes of cell morphology

3.3.1. Neurons.

The caudal/intermediate nTS is known to processes information relevant for chemoreflex control (Favero et al., 2011; Zoccal et al., 2014), and atrophy in this area for the 3 mg/kg STZ-AD group corroborates our findings of respiratory dysfunction. However, the 2 mg/kg AD-STZ group also had respiratory dysfunction in the absence of apparent nTS atrophy. Thus, other factors impact cellular functions and their role in breathing. Although nTS neuron density was unchanged, we observed a strong trend for decreased synaptic density in the caudal nTS. This decrease of synaptic density was significant in the hippocampus region, which was also observed by others in this model (Moreira-Silva et al., 2018). However, data for individual synaptic markers seem inconsistent and vary between brain region (Moreira-Silva et al., 2018; Pilipenko et al., 2020; Xu et al., 2014). A reduction of synaptic contacts indicates breakdown of neural circuits (Jackson et al., 2019) and serves as an early marker in AD patients when found in memory-related brain regions (de Wilde et al., 2016; Ingelsson et al., 2004). Similarly, reduced synaptic density may be an early event in the STZ-AD rat model that is already visible 2–3 weeks after AD induction. Additional reduction of neuronal number may occur at a later time point as shown by others using this model (Rostami et al., 2017).

3.3.2. Astrocytes.

We found an increased number of astrocytes in the intermediate and rostral extent of the nTS. This finding indicates likely effects of STZ-AD on other reflexes due to the semi-viscerotopic organization of the nTS (Cutsforth-Gregory and Benarroch, 2017; Loewy, 1990). For example, the dorsomedial nTS (caudal to intermediate nTS) is also known for its importance in baroreflex control (Andresen and Kunze, 1994) and alteration in astrocyte number may interfere with normal processing. In fact, we have recently shown that STZ-AD induces significant baroreflex dysfunction that is similar to orthostatic hypotension frequently seen in AD patients (Ehlen et al., 2022). On the other hand, the intermediate to rostral nTS is implicated in the control of gastrointestinal, gustatory, and swallowing function (Lang, 2009; Roussin et al., 2012; Travagli et al., 2006). Patients with AD show symptoms associated with these functions, including constipation, altered taste identification, and dysphagia (Affoo et al., 2013; Naudin et al., 2015; Zakrzewska-Pniewska et al., 2012). However, the exact origin of these impairments and whether they occur in the STZ-AD model is currently unknown. Whatsoever, astrocyte number has previously been shown as a rather unreliable marker for astrocyte activation in the STZ-AD model because of its highly variable effect in response to STZ dosage and time course (Rostami et al., 2017). A more reliable marker for gliosis is the morphological appearance of astrocytes. Close analysis of astrocyte morphology indicated a region-specific effect of STZ in the nTS. Although intermediate and rostral astrocytes were similar to control (2 mg/kg STZ-AD) or resembled a morphology of an early atrophic phenotype (3 mg/kg STZ-AD), caudal nTS astrocytes were hypertrophic with the strongest STZ dosage. This hypertrophic state shows activation of astrocytes and is typically seen later in disease progression (Zhou et al., 2019). These area-specific morphologies of astrocytes are indicative of a differential impact of STZ on nTS subregions. STZ may affect caudal nTS regions earlier and astrocytes may have already surpassed the atrophic state at the time of our analysis. This assumption agrees with our gross morphology data, which showed atrophy only in the caudal nTS. It also agrees with our previous study showing activated astrocytes in the caudal nTS (Ebel et al., 2017). For the hippocampus CA1 region, we found no indices of astrocyte activation. This result was surprising because of the marked atrophy of the hippocampus. We have previously shown increased thickening of astrocyte branches in the CA1 region using 3 mg/kg STZ (Ebel et al., 2017). Although our current results suggest a similar outcome, these data were considered non-significant because of variability between animals. However, activation of astrocytes in the hippocampus has been previously documented at a later time point in the STZ-AD rat model (Rai et al., 2014; Yang et al., 2020; Zappa Villar et al., 2018) and in human AD patients (Verkhratsky et al., 2010). This activated phenotype of astrocytes has a profound impact on neuronal function and may lead to excitotoxicity and subsequent neuron death from malfunctioning glutamate handling (Acosta et al., 2017; Conway, 2020).

3.3.3. Microglia.

The primary effect of STZ-AD on microglia in the hippocampus and nTS was by decreasing the number of branches and increasing their thickness. This result resembles a morphology of activated/reactive microglia (Fernández-Arjona et al., 2017) and our findings in the hippocampal area are consistent with previous studies showing activated microglia in various forebrain regions of the STZ-AD rat model (Bassani et al., 2018; Shoham et al., 2003) and in humans with AD (Davies et al., 2017). In fact, reactive microglia as part of the inflammatory process are considered one of the key events in AD progression (Leng and Edison, 2021). In the current study, the effect of STZ-AD on microglia was strong and consistent for the different STZ dosages and brain regions tested, indicating that microglia activation is among the earliest events in this model. It has been previously shown that reactive microglia in AD are able to profoundly alter neuronal processing (Piccioni et al., 2021). Mechanisms for how reactive microglia affect neuronal signaling include synaptic pruning (a lower synapse density was also evident in the current study), impaired glutamate handling by astrocytes at the synaptic cleft, and neuronal death (Allen and Attwell, 2001; Lyman et al., 2014; Piccioni et al., 2021). These pathological changes in the STZ-AD rat model and human AD were associated with overproduction of proinflammatory cytokines and reactive oxygen species derived from lasting microglia activation of the neurotoxic M1 phenotype (and not from the M2 phenotype that is neuroprotective, releases anti-inflammatory cytokines, and is morphologically indistinguishable from the M1 type) (Lu et al., 2017; Piccioni et al., 2021). An early functional consequence from increased proinflammatory cytokines in the nTS is neuronal hyperexcitability, which was shown by extracellular recordings of nTS neurons and their increased action potential discharge to perfusion with TNF-α (Emch et al., 2000). An increase of reactive oxygen species also induces prolonged hyperactivity of neurons in the nTS (Ostrowski et al., 2017, 2014). Thus, activated microglia in STZ-AD may lead to hyperactivity of nTS neurons, affecting proper function of the chemoreflex, and result in the observed breathing dysfunction in this model when exposed to hypoxia.

3.4. Effects of STZ-AD on respiratory centers in the brainstem

Human patients with progressed AD show amyloid beta plaques and hyperphosphorylated tau in various brainstem nuclei, including those important for chemoreflex control (Parvizi et al., 2001). Thus, respiratory dysfunction in human AD likely involves pathological changes throughout the chemoreflex axis. Similar alterations are expected for the STZ-AD animal model. In deed, we have previously demonstrated effects of STZ-AD in multiple nuclei within the chemoreflex axis. For example, besides a blunted activation of nTS neurons in response to hypoxia, we also found significantly reduced activation in select nuclei within the ventral respiratory group (Brown et al., 2019). Namely, the rostral ventral respiratory group and Bötzinger complex had reduced responses when compared to control. On the other hand, the pre-Bötzinger complex and caudal ventral respiratory group were unchanged. These results indicate a rather differential impact of STZ-AD on components within the chemoreflex loop. In regards to the central chemoreflex, we have previously shown that STZ-AD increases the CO₂-sensitivity of locus coeruleus neurons (Vicente et al., 2020). Other centers, such as the retrotrapezoid nucleus, have not been analyzed. However, it is very likely that the immunohistochemical targets chosen for the current study are also altered in downstream nuclei of the reflex axis and play an important role for altered respiratory responses in STZ-AD.

3.5. Intersection of sleep-disordered breathing and AD

Sleep-disordered breathing, including OSA, is highly prevalent in AD (Gaig and Iranzo, 2012; Hoch et al., 1986), and current evidence suggests a tight interrelation between OSA and AD. Although the likelihood of AD patients acquiring OSA as part of the progression of AD is much higher than in age-matched controls (Emamian et al., 2016), OSA is also known as a major risk factor leading to the development of AD (Osorio et al., 2015; Tsai et al., 2020). Furthermore, treatment strategies using positive airway pressure to target OSA have been shown to reduce the risk of developing AD (Tsai et al., 2020) and decelerate memory decline in patients already diagnosed with AD (Troussière et al., 2014). Thus, these data from the bidirectional relationship of OSA and AD suggest shared mechanisms for both conditions. Along these same lines, it has been shown that pathologies associated with AD affect the entire brain, including the brainstem, which is an important area for a multitude of noncognitive functions (Lee et al., 2015; Parvizi et al., 2001). It has been suggested that the noncognitive symptoms in AD are due to pathological changes in the brainstem and occur earlier than the pathology seen in forebrain regions (Simic et al., 2010). However, there is still a large lack of studies focusing on the brainstem in AD, and identification of mechanisms related to the life-threatening breathing dysfunction may expose avenues to also address the prototypical memory decline in AD.

4. Experimental Procedure

4.1. Animals

Male Sprague-Dawley rats (6–8 weeks, total of 42 animals) were bred in the AAALAC-accredited vivarium of A.T. Still University’s Kirksville College of Osteopathic Medicine. Breeders were previously purchased from Hilltop Lab Animals Inc. or Envigo. Animals were housed on a 12-hour day/night cycle at 23 °C with food and water available ad libitum. Animals were randomly assigned to 3 experimental groups: a control group (CTL), 2 mg/kg STZ-AD group, or 3 mg/kg STZ-AD group. All experimental procedures were in accordance with NIH guidelines (“Guide for the Care and Use of Laboratory Animals”) and approved by A.T. Still University’s Animal Care and Use Committee.

4.2. Alzheimer’s disease model

Similar to our previous studies (Brown et al., 2019; Ebel et al., 2017; Vicente et al., 2020), the head of isoflurane-anesthetized rats (Piramal, 5% induction, 2% maintenance) was positioned in a stereotaxic frame (Model 68005, RWD Life Science), and the cranial sutures were exposed with a 1-cm incision. Using a rotary tool (Dremel 7300 with engraving cutter 105), 2 small burr holes were drilled into the skull to access the lateral ventricles using the following stereotaxic coordinates (Paxinos and Watson, 2007): −0.9 mm AP, ± 1.5 mm ML, and 3.6 mm DV. Glass capillaries (type 1B120F-4, World Precision Instruments) were pulled using a P-97 micropipette puller (Sutter Instruments) and filled with 0.9 mM citrate-buffered saline solution (pH = 4.5, vehicle CTL group) or with dissolved STZ (Alfa Aesar, 2 mg/kg STZ-AD group or 3 mg/kg STZ-AD group). The 3 mg/kg STZ dosage was split and injected on day 1 and day 3 to circumvent mortality when given as a single dose. Intracerebroventricular (icv) injections were made bilaterally (5 μL per side) and surgical wounds were closed using absorbable suture (MV-J397-V, Oasis). Experimental animals were subdiabetogenic (CTL, 105.0 ± 4.0 mg/dL, n = 7; 2 mg/kg STZ-AD, 134.0 ± 7.0 mg/dL, n = 8; 3 mg/kg STZ, 122.0 ± 8.0 mg/dL, n = 7), as shown for a subgroup using a blood glucose tester (TRUEresult, Nipro Diagnostics), and had similar weights (CTL, 287 ± 23 g, n = 7; 2 mg/kg STZ-AD, 249 ± 18 g, n = 8; 3 mg/kg STZ-AD, 276 ± 21 g, n = 7). Before injections, animals received 2 mg/kg dexamethasone (VetOne) as an immunosuppressant. Postoperative care included 50 μg/kg buprenorphine hydrochloride (s.q., Reckitt Benckiser), 7 mg/kg Baytril (i.m., VetOne), and 3 mL normal saline solution (s.q., Hospira). To facilitate postsurgical weight recovery, the normal rat diet was supplemented with high caloric supplement until presurgical weight was reached.

4.3. Plethysmography

Similar to before (Brown et al., 2019; Ebel et al., 2017), chemoreflex assessment was done using a whole-body flow-through plethysmography chamber (Data Sciences International) in separate animals (n = 6/group) 2–3 weeks after model induction. The respiratory responses were measured in conscious and unrestrained rats exposed to different O₂ concentrations (10% or 21% O₂, equilibrated with N₂) at 3 L/min. Gas concentrations were established using 2 mass flow controllers (MC-5SLPM-D, Alicat Scientific), and fluctuations in chamber pressure from the animal’s respiration were measured with a DP45 low differential pressure transducer (Validyne). Respiratory data were extracted from the carrier wave using a CD15 sine wave carrier demodulator (Validyne) and digitized with PowerLab 8/35 (AD Instruments). LabChart software (AD instruments) was used to record and analyze the raw data before final processing in Excel (Microsoft). A 30-second segment of uninterrupted breathing (without movements, sniffs, or sighs) was used to measure respiration rate (RR, breaths/minute), tidal volume (TV, area under the inspiratory pressure curve), and minute ventilation (MV, product of RR and TV) immediately before (21% O₂, baseline) and 30 minutes into hypoxic exposure (10% O₂). Changes in chamber pressure (ΔP) were calibrated to pressure changes (ΔPcal) from a known volume of air (Vcal) using a 1-mL syringe. TV was corrected for barometric pressure (PB), rat body temperature (Tb), chamber pressure (Tc), and water vapor pressure in the rat (Pw,b) and chamber (Pw,c) according to the following formula (Drorbaugh and Fenn, 1955):

$$\text{TV}=\text{Vcal}\times \frac{\Delta \text{P}}{\Delta \text{Pcal}}\times \frac{1}{1-\frac{\text{PB}-\text{Pw},\text{b}}{\text{PB}-\text{Pw},\text{c}}\times \frac{\text{Tc}}{\text{Tb}}}$$

Barometric pressure, temperature, and humidity was measured inside the plethysmography chamber using a wireless sensor (model 02038W, AcuRite). Core body temperature and water vapor pressure inside the rat was assumed at 310.15 K and 47.1 mmHg, respectively.

4.4. Brain tissue processing

Brain tissue was collected 2–3 weeks following model induction. Rats were deeply anesthetized with 5% isoflurane and transcardially perfused with heparinized 10 mM phosphate-buffered saline (PBS, in mM: 2.7 NaH₂PO₄, 7.7 Na₂HPO₄, and 154 NaCl at pH 7.4). Brains were subsequently fixed with 4% paraformaldehyde (PFA, Acros Organics) in PBS, extracted, post-fixed in PFA overnight at 4 °C, and stored in PBS until further processing. Coronal brain slices of the hippocampus and nTS were obtained using a vibratome (7000smz-2, Campden Instruments) and collected into 12- and 24-well tissue culture plates, respectively. Culture plates were filled with cryoprotectant (50% PBS, 30% ethylene glycol, and 20% glycerol) for long-term storage at −20 °C. Hippocampal sections were collected as follows: 1) at bregma −0.96 mm as identified by the extent and relative location of the stria medullaris and fornix fiber bundles and 2) at bregma −3.60 mm as identified by the initial appearance (from rostral) of a medial anatomical connection (retrouniens area) between each ventral posteromedial thalamic nucleus. The nTS region was collected between Bregma −14.94 and −12.42 mm by first identifying calamus scriptorius (CS, caudal tip of the area postrema) and then collecting multiple equidistant (180 μm apart) brain slices caudal and rostral from CS.

4.5. Gross morphological analysis

Hippocampal sections (150 μm thick) were photographed using AmScope software (version 3.7) and a trinocular stereoscope (SM-7TX-FRL, AmScope) with digital camera (MU300, AmScope). Images from 30-μm thick brainstem sections were obtained using a 4× objective on a conventional microscope (BX51WI, Olympus) with a digital camera (optiMOS, QImaging). Images of the nTS were stitched using the MosaicJ plugin (Thévenaz and Unser, 2007) of Fiji software (version 1.52i, NIH). All images were postprocessed in Fiji by outlining and measuring the area of the ventricle space, septum, striatum, hippocampus, and thalamus, and by measuring the thickness of the cortex.

4.6. Immunohistochemistry

Similar to our previous studies (Brown et al., 2019; Ebel et al., 2017; Vicente et al., 2020), 30-μm thick free-floating coronal sections of the hippocampus or nTS were washed in PBS and blocked to prevent nonspecific antibody binding using 10% normal donkey serum (NDS, Millipore) with 0.3% Triton-PBS. Brain sections were then incubated with primary antibodies in 0.3% Triton-PBS with 1% NDS at room temperature overnight. The following primary antibodies were used: NeuN (rabbit, 1:400, MABN140, Millipore), synaptophysin 1 (guinea pig, 1:1000, 101 004, SYSY), S100B (rabbit, 1:1000, ab41548, Abcam), GFAP (chicken, 1:3000, GFAP, Aves Labs), and IBA1 (guinea pig, 1:300, 234 004, Synaptic Systems). The next day, sections were washed in PBS and incubated for 2 hours with the appropriate secondary antibodies: (Jackson ImmunoResearch): Alexa Fluor 488 (donkey anti-chicken, 1:200, 703-545-155), Alexa Fluor 594 (donkey anti-rabbit, 1:200, 711-586-152), and Alexa Fluor 647 (donkey anti-guinea pig, 1:200, 706-605-148). Next, tissue sections were positioned on gelatin-coated glass slides, air dried, and coverslipped using ProLong Diamond (P36962, ThermoFisher Scientific). One section was incubated without primary antibody to ensure antibody validity by the absence of fluorescent staining (negative control). Immunohistochemical images were obtained with a digital monochrome camera (DS-QiMc, Nikon) mounted on an epifluorescent microscope (Eclipse 80i, Nikon). Exposure times remained constant to analyze staining intensity among groups. All analyses of immunohistochemical images were done in Fiji software. Images were postprocessed uniformly for contrast and brightness to enhance clarity. Manual cell counts using the multi-point tool (for anti-NeuN, anti-S100B, and anti-IBA1 positive cells) and staining intensity using mean gray values (for anti-synaptophysin 1) were completed in the CA1 region of the hippocampus or in the middle of a hemi-nTS section using a 100 × 100 or 200 × 200 μm analysis window (size of the window was constrained by the respective nucleus area and density of cells for adequate analysis). Areas devoid of cells (e.g., blood vessels) or nucleus borders were avoided for cell density analyses. Images were independently and blindly evaluated by 3 persons using the same counting criteria. Final results of counters were averaged and compared among experimental groups. Fiji and Inkscape software (version 0.91) were used to create representative brain sections for the current publication. For group analysis (n = 7–8 animals/group), each animal contributed 1 hippocampal CA1 region (bregma −3.60 mm) and 3 nTS sections (representative for the rostral-caudal nTS extent: caudal nTS, CS −0.36 mm; intermediate nTS, CS +0.54 mm; rostral nTS, CS +1.62 mm). Sections shown in the figures were from the hippocampal CA1 region and caudal nTS.

4.7. Cell morphology analysis

For assessment of astrocyte and microglia morphology, images were obtained from a confocal microscope (DMI6000 B, Leica) with laser excitation (Argon, HeNe 633) and appropriate filter sets to visualize fluorescence. Confocal z-stacks were taken at 10× and 0.5-μm intervals between images. Maximum intensity z projections were done using Fiji software. To ensure visibility of fine processes, each image was de-speckled in Fiji and adjusted for brightness and contrast allowing for 1% of pixels to be fully saturated. The Fiji neuroanatomy plugin Simple Neurite Tracer (SNT, v3.2.14) was then used to trace all visible cell processes and calculate average branch number and length. Branch width was determined using the ‘fill out’ feature in SNT’s path manager (fill threshold of 0.3). Cell processes were traced blindly by 2 persons. Each data point given in figure 6 and 8 represents the average cell morphology of one animal in the respective brain area (hippocampal CA1 region and the 3 subregions of the nTS). Each of these averages consists of 3 representative cells that were analyzed in that section. Cells were only analyzed when the morphology allowed a clear separation from neighboring cells.

4.8. Statistical analysis

Statistical analyses were performed using SigmaPlot 14.0 (Systat Software). Gross-morphological data from the hippocampus and all immunohistochemical data assessing cell/synapse density and process morphology were analyzed using a t-test (2 groups) or 1-way ANOVA (multiple groups per brain area) followed by the Newman-Keuls post-hoc test. Chemoreflex and gross-morphological data from nTS subregions were analyzed using two-way repeated ANOVA with Newman-Keuls post hoc test. If data were non-normal, the corresponding nonparametric test was used. Statistical significance was defined as p ≤ 0.05. Group data are presented as mean ± standard error of the mean (SEM).

Highlights.

• Alzheimer’s disease (AD) is highly associated with respiratory dysfunction

• Brainstem pathology corroborates respiratory problems in the STZ model of AD

• Regional atrophy is found in a respiratory brainstem area

• Reduced synapses and gliosis may initiate respiratory dysfunction in STZ-AD