Morphological changes of the digestive canal organs in animals with a model of Alzheimer's disease

Alzheimer's disease (AD) is a neurodegenerative disorder characterized by the progressive development of morphological changes leading to severe dementia. A key molecular event in its pathogenesis is impaired proteolysis of amyloid precursor protein (APP) [1]. This event initiates the formation of aggregation-prone amyloid beta (Aβ) peptides. Owing to its conformational properties, Aβ undergoes irreversible polymerization and is deposited in the neuropil of various regions of the central nervous system (CNS), where it forms deposits. This leads to neuroinflammation accompanied by astrogliosis, microgliosis, and neuronal death [2]. Aβ forms amyloid plaques, accumulates in perivascular and intracellular compartments, and enters the cerebrospinal fluid and bloodstream.

For a long time, AD has been regarded predominantly as a disorder of the CNS. However, evidence of peripheral tissue involvement indicates a potential role of the gastrointestinal (GI) tract and alterations in the enteric nervous system in disease development. In particular, the expression of APP and APP-like proteins has been demonstrated in the human GI tract, including structures of the enteric nervous system, suggesting that Aβ deposits may also form there [3]. Nevertheless, direct evidence of specific Aβ accumulation in enteric neurons in patients with AD remains limited. Earlier studies identified Aβ deposits in nonneural tissues and blood vessels, including the intestine; however, their cellular and anatomical localization within the intestinal wall requires cautious interpretation [4].

Evidence linking Aβ to the gastrointestinal tract was obtained in a study that detected elevated levels of Aβ aggregates in the faeces of patients with AD [5]. However, the source of these aggregates remains unclear: they may originate from the enteric nervous system or reach the intestine through other routes, including hepatobiliary excretion. In addition, signs of intestinal barrier dysfunction have been described in patients with dementia, including increased diamine oxidase and sCD14 levels, which may reflect increased intestinal permeability and endotoxin burden [6]. Similarly, elevated serum zonulin levels in patients with AD indicate possible dysregulation of the epithelial barrier, although these findings do not constitute direct morphological evidence of damage to intestinal epithelial tight junctions [7].

In addition to barrier dysfunction, intestinal inflammation has been implicated in AD. Elevated fecal calprotectin levels are associated with age and AD pathology, indicating a relationship between intestinal inflammation and the neurodegenerative process [8]. Functional bowel disturbances may also be clinically relevant: constipation in patients with AD is associated with a more rapid decline in cognitive function and the progression of white matter changes [9]. Nevertheless, most of these findings are associative and do not establish a causal role of intestinal alterations in the pathogenesis of AD.

Numerous transgenic animal lines have been developed to investigate AD in greater detail, most of which are mouse models carrying mutant human APP and PS1 (presenilin 1) genes. Several lines exhibit pronounced cerebral amyloidosis; however, functional GI abnormalities have been described only in some of them, including Tg2576 (transgenic line 2576) [10] and 5xFAD (five familial Alzheimer's disease mutations) [11], whereas combined functional and morphological changes have been reported in APP/PS1 [12] and APP23 mice [13].

One of the most promising models is the APP/PS1 – APPswe/PS1dE9 (amyloid precursor protein with the Swedish mutation / presenilin 1 deleted exon 9 line) line, in which mice older than 1 year exhibit amyloidosis of the colonic muscularis, accelerated intestinal transit, and a reduced population of nitrergic neurons [12].

Moreover, data on intracellular Aβ accumulation in neurons in the intestinal plexuses remain fragmentary. Information is also limited regarding the response of enteric glia, including glial fibrillary acidic protein (GFAP) positive and S100 calcium-binding protein B (S100B) positive cells, which play key roles in the pathogenesis of neuroinflammation in the CNS.

MATERIALS AND METHODS

Animals

The experiments included male wild-type C57BL/6J mice and male APPswe/PS1dE9 mice on the same genetic background (n = 5 per group). Only animals without concomitant disorders were included. The mice were obtained from the animal facility at Sechenov University. The study was conducted without blinding.

After a two-week quarantine period and veterinary examination, wild-type and transgenic mice were maintained in the animal facility in accordance with all ethical standards and housing requirements. At the age of 2 months, the animals were genotyped by polymerase chain reaction followed by gel electrophoresis. Based on the genotyping results, the animals were assigned to two groups: an intact group (without pathology), which included wild-type animals, and an experimental group, which included transgenic mice with an Alzheimer's disease model. Animals were selected using a random number method. No animals were excluded from the experimental groups. Until 1 year of age, the mice were housed at 22 °C and 55–60% relative humidity with ad libitum access to clean water and pelleted feed.

Behavioral tests

At 1 year of age, the animals underwent behavioral tests commonly used in AD research for additional model validation: the open-field test, elevated plus maze, and Morris water maze. The tests were performed sequentially, with sufficient intervals and a mandatory habituation stage before each test. Horizontal locomotor activity was assessed using PoKi v0.0.1 software, which was used for tracking mice, generating heatmaps, and recording the time spent in predefined test zones [14].

Histological examination

After completion of all the behavioral tests, the animals were anesthetized with xylazine at 5 mg/kg (Interchemie, Castenray, the Netherlands) and tiletamine/zolazepam at 40 mg/kg (Zoletil 100; Virbac, Carros, France). They were then euthanized via perfusion with a 0.002% heparin solution, followed by in situ fixation with Carnoy's fluid.

The scalp was first removed, the skull was exposed, and the brain was collected. The abdominal cavity was then opened, and the stomach and intestine were isolated. Autopsy samples from the stomach, small intestine, and colon were separated into the corresponding regions, opened longitudinally, rinsed with fixative, and rolled using the Swiss-roll technique. All samples were subsequently fixed in Carnoy's mixture for an additional 24 hours, thoroughly washed, processed using isopropyl alcohol, and embedded in paraffin.

The stomach and intestinal samples were sectioned on an HM325 rotary microtome (Thermo Fisher Scientific, Waltham, MA, USA) perpendicular to the coils of the Swiss roll, whereas the brain samples were cut into coronal sections. All sections were 3 μm thick. To improve tissue adhesion, silane-coated adhesive slides were used and dried at 56 °C.

Extracellular amyloid deposits in the brain were identified using general histological staining with sulfated Alcian blue (SAB) followed by Nissl counterstaining with neutral red.

Some sections were used for immunohistochemical analysis. Antigen retrieval was performed in one-step dewaxing/antigen retrieval buffer, pH 9 (item number E-IR-R220A, lot XF05RT4N9592; Elabscience, Wuhan, China), for 20 min at 98 °C and 105 kPa. After cooling, the sections were washed three times in 10% phosphate-buffered saline (PBS; Eco-Service LLC, Saint Petersburg, Russian Federation) and incubated with hydrogen peroxide for 30 min at 37 °C. The sections were then washed three times in PBS and incubated with 1% bovine serum albumin (lot RM-T1725/1000; Bi-osera, Nuaillé, France) for 30 min at 37 °C. Primary antibodies were subsequently applied, and the sections were incubated for 18 hours at room temperature.

Aβ deposits were detected using rabbit polyclonal anti-Aβ antibodies (1:200; item number E-AB-15509, lot DM04J4P29901; Elabscience, Wuhan, China). The peripheral nervous system was visualized using chicken polyclonal anti-peripherin antibodies (1:1000; item number GTX85464, lot 822503406; GeneTex, Irvine, CA, USA). Synaptic innervation was assessed using rabbit polyclonal anti-synaptophysin antibodies (SYP, synaptophysin; 1:200; item number O407-2; lot HG0614; Huabio, Hangzhou, China). Enteric glia and cerebral astrocytes were evaluated using rabbit monoclonal anti-GFAP antibodies (glial fibrillary acidic protein; 1:500; clone SA03-04, item number ET1601-23, lot HO0913; Huabio, Hangzhou, China) and rabbit monoclonal anti-S100B antibodies (1:1000; clone SC57-02, item number ET-1610-3, lot H661380007; Huabio, Hangzhou, China).

After incubation, the stomach and intestinal sections were washed three times with PBS and incubated with secondary goat anti-rabbit IgG antibodies conjugated with the fluorophore TRITC (tetramethylrhodamine isothiocyanate) (Goat Anti-Rabbit-TRITC, 1:100, catalog No. E-AB-1053, lot No. 22038, Elabscience, Wuhan, China) and goat anti-chicken IgY antibodies conjugated with FITC (fluorescein isothiocyanate) (Goat Anti-Chicken-IgY-FITC, 1:1000, catalog No. GTX26873, lot No. 822503077, GeneTex, Irvine, USA) for 2 hours at 37 °C. The sections were then washed three times with PBS and stained with the nuclear dye DAPI (4',6-diamidino-2-phenylindole) (catalog No. E-IR-R103, Elabscience, Wuhan, China) for 15 minutes at 37 °C. Afterward, the sections were washed, dehydrated through a graded series of alcohols, cleared in xylene, and mounted in the xylene-containing mounting medium Vitrogel.

After incubation with the primary antibodies, the brain sections were also washed in PBS and incubated with ready-to-use horseradish peroxidase-conjugated goat anti-rabbit antibodies (goat anti-rabbit-HRP; item number HA1119, lot M05-22-P2; Huabio, Hangzhou, China). After washing in PBS, 3,3'-diaminobenzidine (CDH, New Delhi, India) was used as the chromogen for 1–2 min. The sections were then rinsed in distilled water, counterstained with Carazzi hematoxylin, and mounted.

Morphometric analysis

Imaging and histological analysis were performed using an Axio Imager. A1 microscope, an Axiocam 305 color camera, and ZEN 3.10 software (Zeiss, Jena, Germany) were used. Immunofluorescence images were acquired at wavelengths ranging from 359–461 nm for DAPI, 495–519 nm for FITC, and 557–576 nm for TRITC.

The area of the SYP⁺ plexuses was measured using the standard area-measurement tool in Fiji 2.17.0. The structure of GFAP⁺ enteric glia was evaluated using open-source CellSkeletDetector v0.0.2 software [15]. The numbers of GFAP⁺ glial cells and Aβ deposits in the brain, Aβ⁺ neurons in the enteric plexuses, and S100B⁺ cells within and outside the myenteric plexus were quantified.

Statistical analysis

A pilot experiment was conducted to determine the minimum acceptable sample size required to detect potentially significant differences between groups. Normality of data distribution was assessed using the Shapiro–Wilk test. When normality was confirmed (p > 0.05), differences between two independent groups were evaluated using Student's t-test. One-way ANOVA followed by Tukey's test was used to compare Aβ, GFAP, and S100B levels among the GI regions. In-between-group differences were considered significant at p < 0.05 and a statistical power greater than 0.8. Statistical analyses were performed using GraphPad Prism 10.4.0 (GraphPad Software, San Diego, CA, USA).

RESULTS

Behavioral assessment of cognitive impairment

Heat maps of animal movement trajectories during the behavioral tests are shown in Fig. 1A. In the open-field test, APPswe/PS1dE9 mice traveled a greater total distance (Fig. 1B) and moved at a greater mean speed (Fig. 1C) than wild-type mice did, while making fewer stops (Fig. 1D). APPswe/PS1dE9 mice also entered the central zone more frequently (Fig. 1E). These findings indicate a pronounced increase in locomotor activity in APPswe/PS1dE9 mice.

In the elevated plus maze, APPswe/PS1dE9 mice similarly showed an increase in the total distance traveled (Fig. 1F). The transgenic animals made more entries into the closed arms, further indicating increased locomotor activity (Fig. 1H). No differences in entries into the open arms were observed between wild-type and APPswe/PS1dE9 mice (Fig. 1G), suggesting the absence of overtly increased anxiety-like behavior. The longer time spent in the central zone may reflect more frequent transitions between maze compartments and increased exploratory behavior (Fig. 1I).

After five days of training with the platform in the Morris water maze, swimming ability was assessed during a probe trial without the platform. Compared with wild-type mice, APPswe/PS1dE9 mice spent less time in the target quadrant containing the former platform location (Fig. 1J). In addition, APPswe/PS1dE9 mice crossed the former platform zone fewer times (Fig. 1K) and showed a longer latency to reach the platform zone (Fig. 1L). The swimming speed did not differ significantly between the wild-type and APPswe/PS1dE9 mice (Fig. 1M). These changes indicate impaired spatial memory retention and reduced accuracy in searching for the former platform location in APPswe/PS1dE9 mice.

Morphological features of the neuropathological process in the central nervous system

Histochemical analysis using a modified SAB method revealed deposits that were unevenly distributed throughout the brain (Fig. 2A). Numerous SAB⁺ deposits were detected in the CA1 and CA3 regions and dentate gyrus of the hippocampus, as well as in the cerebral cortex, of APPswe/PS1dE9 mice.

Immunohistochemical analysis using anti-Aβ antibodies confirmed that the SAB⁺ structures were amyloid plaques (Fig. 2F). Aβ staining also revealed intracellular deposits, particularly in cells within the plaque microenvironment (Fig. 2F). Among the regions examined, the largest numbers of deposits were found in the hippocampal CA1 region and primary auditory cortex (Figs. 2B and 2C). No Aβ immunoreactivity was detected in wild-type animals.

GFAP immunohistochemistry revealed markedly increased activation in the cerebral cortex and hippocampus of APPswe/PS1dE9 mice (Fig. 2G). In contrast, only isolated GFAP⁺ cells per field of view were observed in the corresponding brain regions of the wild-type animals (Figs. 2D and 2E).

Amyloid beta accumulation in the intramural ganglia of the gastrointestinal tract

Immunohistochemical analysis of Aβ revealed cytoplasmic staining in neurons of the myenteric plexus in APPswe/PS1dE9 mice (Fig. 3A). No Aβ staining was observed in wild-type animals. The number of positively stained neurons in the myenteric plexus did not differ between the stomach and small intestine (Fig. 3B). In the colon, however, the number of Aβ⁺ neurons per field of view was significantly greater than that in the stomach and small intestine.

Changes in synaptophysin immunoreactivity in enteric neurons

SYP immunohistochemistry was used to evaluate vesicular transport in neurons of the myenteric plexus. APPswe/PS1dE9 mice presented an increased area of SYP⁺ enteric neurons in the small intestine and colon (Figs. 4A and 4B). In contrast, SYP staining occupied a smaller mean area in the gastric myenteric plexus of APPswe/PS1dE9 mice than in that of wild-type animals.

Enteric glial response

Immunohistochemical analysis of the GFAP and S100B markers enabled the assessment of two enteric glial populations (Fig. 5). Increased GFAP⁺ glial reactivity was observed in the intestinal myenteric plexus (Fig. 5A). Skeletonization analysis revealed increases in the number and branching complexity of glial processes (Figs. 5B and 5C). The opposite effects were observed in the gastric myenteric plexus: the branching of GFAP⁺ glia was markedly reduced, whereas the number of nodes remained at the wild-type level.

In the stomach and intestine, S100B is expressed by a population of Schwann cell-like glia. In mice, S100B⁺ cells are located in both the myenteric and submucosal plexuses and along nerve fibers in the submucosal layer (Fig. 6A). Immunohistochemical analysis of S100B in the myenteric plexus of the stomach and small intestine revealed no significant differences (Fig. 6B). In contrast, the number of S100B⁺ cells in the colonic myenteric plexus was reduced APPswe/PS1dE9 mice. Assessment of S100B immunoreactivity outside the plexuses, in the submucosa and muscularis, revealed an increased response in all the GI regions of APPswe/PS1dE9 mice (Fig. 6C).

DISCUSSION

In this study, APPswe/PS1dE9 mice exhibited increased locomotor activity in the open-field test and elevated plus maze, characterized by greater total distance traveled and higher mean speed. The animals also made fewer stops and entered the central areas of both tests more frequently. These findings are consistent with previously reported alterations in behavioral activity in transgenic mouse models of R. Lalonde et al. [16] reported that 12-month-old APP(695)swe/co+PS1dE9 mice presented altered exploratory behavior in the open-field test and elevated plus maze, including increased time spent in the open arms of the elevated plus maze and greater distance traveled in the central zone of the open field.

In the Morris water maze, APPswe/PS1dE9 mice spent less time in the quadrant containing the former platform location, crossed the former platform zone less frequently, and took longer to reach this zone than wild-type animals did, indicating impaired spatial memory in the transgenic mice. Similar findings were reported by L. Fu et al. [17] in APPswe/PS1dE9 mice. Compared with wild-type animals, transgenic animals presented progressive spatial memory impairment, characterized by fewer crossings of the platform zone and reduced time and distance in the target quadrant.

APPswe/PS1dE9 mice exhibited excessive accumulation of beta-amyloid plaques in the brain, including in the hippocampal CA1 region and primary auditory cortex. According to the literature, these brain regions are among the principal sites in which beta-amyloid aggregates are detected [18][19]. Plaque deposition was accompanied by pronounced astrogliosis. These changes are characteristic of this transgenic line and confirm the reproduction of the major features of the model [20].

Our study revealed pronounced morphological changes in the stomach, small intestine, and colon of 1-year-old APPswe/PS1dE9 mice. The pathological process prominently involves neurons and enteric glia of the myenteric plexus. The severity of alterations across the tissue layers followed a craniocaudal gradient. This finding is supported by recent studies identifying the distal colon as the principal site of involvement in APPswe/PS1dE9 mice [12]. The literature describes an absence of intestinal manifestations in 5xFAD mice, despite their more aggressive CNS amyloidosis associated with a greater number of APP and PS1 mutations. It may therefore be assumed that GI alterations are related to the specific APPswe/PS1dE9 genetic cassette, which is controlled by a prion promoter. Similarly, other lines expressing AD-associated mutant genes under a prion promoter may exhibit intestinal manifestations. This assumption is supported by enteric nervous system alterations in TgCRND8 mice, which also carry a prion promoter [21].

With respect to amyloid pathology, the key observation was pronounced cytoplasmic alterations in enteric neurons. These effects involve not only intracellular beta-amyloid accumulation but also apparent impairment of secretory function, as indirectly indicated by increased SYP levels. Aβ accumulation was most pronounced in the cytoplasm of neurons in the colonic myenteric plexus. Our findings are consistent with ELISA-based studies of Aβ immunoreactivity, which revealed increased Aβ1-40 and Aβ1-42 in the colons of APPswe/PS1dE9 mice and the humanized AppNL-G-F line at 3, 6, and 12 months of age [22]. Increased immunoreactivity measured by ELISA has also been reported in segments of the small intestine in transgenic animals overexpressing mutant forms of Aβ [23]. The literature indicates that Aβ accumulation in the enteric nervous system of APPswe/PS1dE9 mice leads to the loss of enteric nNOS⁺ neurons [24].

Our findings suggest that neuronal dysfunction may be associated with impaired vesicular transport resulting from altered SYP activity. A marked increase in SYP immunoreactivity in myenteric neurons was demonstrated in the intestines of APPswe/PS1dE9 mice. This condition may be associated with neuronal stress and occurs in several inflammatory disorders of the GI tract.

Notably, enteric glia respond to amyloid deposition through a mechanism similar to the astrocytic response in the CNS in AD. The morphological signs of glial activation observed in this study indicate the initiation of a neuroinflammatory cascade in the intestinal wall analogous to central gliosis. GFAP is considered a marker of astrocyte-like enteric glia, which correspond functionally and in their gene expression signature to CNS astrocytes [25]. Immunohistochemical analysis revealed increased GFAP immunoreactivity in the intestine, most prominently in the colon. This observation provides further evidence that the pathological process progresses in a craniocaudal direction.

Increased S100B immunoreactivity, which marks a population of non-myelinating Schwann cells (neurolemmal enteric glia), indicates enhanced trophic and modulatory activity outside the myenteric plexus in APPswe/PS1dE9 mice [26]. Under conditions of beta-amyloid accumulation outside the myenteric plexus, excessive S100B release may be assumed to alter calcium homeostasis and enteric neuronal excitability, thereby contributing to the development of colonic motor disturbances.

Study limitations

The main limitation of the study is that changes in the enteric plexuses in humans with Alzheimer’s disease may be considerably more pronounced than those reproduced in animal models. In laboratory rodents, the corresponding phenotype generally develops in the presence of mutations in the APP and PS1 genes; however, such models do not reproduce neurofibrillary pathology, which plays an important role in the development of Alzheimer’s disease in humans. In contrast, the predominant form of Alzheimer’s disease in humans (approximately 95% of cases) is sporadic, which makes it difficult to reproduce in rodents. In addition, the expression profiles of most promoters used in mouse lines often do not correspond to the expression profiles of the APP and PS1 genes in humans. An additional limitation is that the study was conducted without blinding.

Future directions

The most promising direction is to investigate pharmacological and biological factors that may influence the progression of intestinal manifestations of AD. Of particular interest is the study of factors capable of slowing the development of pathological changes in the CNS by reducing the effects of Aβ in intestinal tissue. Particular attention should be given to studies using recently developed mouse lines carrying mutations in the APOE4 and TREM2 genes, which are associated with sporadic forms of AD.

CONCLUSION

This study examined qualitative and quantitative morphological changes in the stomach, small intestine, and colon of APPswe/PS1dE9 mice used as a model of Alzheimer's disease. The alterations became more pronounced from the proximal to distal regions of the GI tract and involved enteric neurons and glia.

AUTHOR CONTRIBUTIONS

Alexander N. Yatskovskiy: study concept and design, critical revision of the manuscript. Ksenia S. Pokidova: experimentation, preparation of histological specimens, statistical analysis, literature review, manuscript preparation, preparation of illustrations. Egor A. Kuzmin: histological and immunohistochemical studies, literature review, manuscript preparation, preparation of illustrations. All authors approved the final version of the manuscript.

Ethics statements. The study was conducted in accordance with the provisions of the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes. All procedures involving animals were performed in accordance with the approval of the Local Ethics Committee of Sechenov First Moscow State Medical University (Sechenov University) (Protocol No. 04-23 dated March 2, 2023).

Data availability. The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Conflict of interest. The authors declare that there is no conflict of interests.

Financing. The study had no sponsorship (own resources).

Use of artificial intelligence. No artificial intelligence tools were used in the preparation of this manuscript.