Caffeoylquinic Acid Mitigates Neuronal Loss and Cognitive Decline in 5XFAD Mice Without Reducing the Amyloid-β Plaque Burden

Abstract

Background:

Caffeoylquinic acid (CQA), which is abundant in coffee beans and Centella asiatica, reportedly improves cognitive function in Alzheimer’s disease (AD) model mice, but its effects on neuroinflammation, neuronal loss, and the amyloid-β (Aβ) plaque burden have remained unclear.

Objective:

To assess the effects of a 16-week treatment with CQA on recognition memory, working memory, Aβ levels, neuronal loss, neuroinflammation, and gene expression in the brains of 5XFAD mice, a commonly used mouse model of familial AD.

Methods:

5XFAD mice at 7 weeks of age were fed a 0.8% CQA-containing diet for 4 months and then underwent novel object recognition (NOR) and Y-maze tests. The Aβ levels and plaque burden were analyzed by enzyme-linked immunosorbent assay and immunofluorescent staining, respectively. Immunostaining of markers of mature neurons, synapses, and glial cells was analyzed. AmpliSeq transcriptome analysis and quantitative reverse-transcription-polymerase chain reaction were performed to assess the effect of CQA on gene expression levels in the cerebral cortex of the 5XFAD mice.

Results:

CQA treatment for 4 months improved recognition memory and ameliorated the reduction of mature neurons and synaptic function-related gene mRNAs. The Aβ levels, plaque burden, and glial markers of neuroinflammation seemed unaffected.

Conclusions:

These findings suggest that CQA treatment mitigates neuronal loss and improves cognitive function without reducing Aβ levels or neuroinflammation. Thus, CQA is a potential therapeutic compound for AD, improving cognitive function via as-yet unknown mechanisms independent of reductions in Aβ or neuroinflammation.

Keywords

Alzheimer’s disease amyloid-β caffeoylquinic acid chlorogenic acid coffee cognition DeepLabCut

Introduction

Alzheimer’s disease (AD) is an irreversible neurodegenerative disorder accounting for 60%–80% of dementias, in which cognitive dysfunction, including memory impairment, is a core manifestation [1]. In the United States (US), 6 million people aged 65 years and older are living with AD, and the number is estimated to reach 13.8 million by 2060 [2]. The mechanisms underlying the onset of AD are not well understood. Amyloid-β (Aβ) plaques in the brain, derived from sequential cleavage of the Aβ protein precursor (AβPP), are considered hallmarks of AD pathology [2]. Several anti-Aβ antibody immunotherapies and anti-Aβ drugs aimed at reducing the Aβ levels in the brain are in development [3, 4]. Anti-amyloid agents have demonstrated either limited or no cognitive benefits in AD clinical trials, however, and clinical development of most agents has been discontinued due to insufficient clinical efficacy for cognitive improvement despite significant effects on biomarkers [4 –6]. At present, the US Food and Drug Administration (FDA) has approved only 2 drugs for AD, including mild cognitive impairment [7, 8], and the development of more effective drugs is highly desired.

Epidemiologic investigations suggest that coffee consumption has a neuroprotective effect [9] and that higher coffee consumption is associated with slower cognitive decline and reduced AD pathology [10, 11]. Caffeoylquinic acid (CQA) or chlorogenic acid, a major compound abundant in coffee beans [12] and the plant Centella asiatica [13], improves components of cognitive function such as attention, psychomotor speed, and executive function in healthy human subjects [14]. Moreover, long-term CQA treatment improves cognitive function assessed by recognition memory tasks in 2 AD mouse models, namely PS2APP transgenic (Tg) mice [15] and 5XFAD Tg mice [16], both of which express mutant forms of human APP and presenilin genes [17, 18]. Interestingly, CQA with coffee bean extract and Centella asiatica extract have distinct effects on the pathologies observed in AD model mice. CQA treatment reduced the pathologic Aβ plaque burden in both the cerebral cortex and hippocampus of PS2APP mice [15], whereas treatment with coffee bean extract or Centella asiatica extract, the major constituent of which is CQA, had no significant effect on the cortical Aβ plaque burden in PS2APP and 5XFAD mice [19, 20]. Despite the lack of a significant effect on cortical Aβ plaque deposition, both coffee bean extract and Centella asiatica extract ameliorated cognitive or working memory impairments in PS2APP and 5XFAD mice [19, 20]. Therefore, how CQA exerts its effects on Aβ pathology and cognitive function remains unknown. CQA treatment has a neuroprotective effect, inhibiting Aβ-induced cell death in the SH-SY5Y neuroblastoma cell line [21], which is a putative mechanism underlying the improvement of cognitive function. Whether CQA treatment provides neuroprotective benefits in vivo, however, has not been elucidated.

Here, we sought to reveal how long-term CQA treatment provides neuronal benefits in AD. First, we investigated the effect of CQA on cognitive function and the Aβ levels and plaque burden in 5XFAD model mice. In addition, we evaluated the effect of CQA on neuronal loss, synaptic density, and neuroinflammation in the brains of 5XFAD mice by immunostaining. Furthermore, we conducted AmpliSeq and quantitative reverse transcription-polymerase chain reaction (qRT-PCR) to address the effects of CQA treatment on global gene expression in 5XFAD mice.

MATERIALS AND METHODS

Animals

All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Tsukuba. Transgenic (APPSwFlLon, PSEN1*M146L*L286V) 6799Vas mice, also known as 5XFAD mice, with a congenic C57BL/6J genetic background from The Jackson Laboratory (Bar Harbor, ME, USA) were used in this study. Wild-type (WT) littermates were used as controls. Genotyping of mice was performed using the following primers: 5XFAD_Mutant_R 5’-CGGGCCTCTTCGCTATTAC-3’ (Jackson primer# 27367), 5XFAD_Common 5’-CGG GCCTCTTCGCTATTAC-3’ (Jackson primer# 37598), and 5XFAD_WT_R 5’-TATACAACCTTGGGGGATGG-3’ (Jackson primer# 37599). Male animals were used for all the experiments and were group-housed (3–5 mice/cage). Mice were maintained under a 12-h light/dark cycle. Food and water were available ad libitum.

Diet

Mice were fed either a normal MF diet (MF, Oriental Yeast, Japan) or a diet supplemented with 0.8% CQA (#70930; Cayman Chemical, Ann Arbor, USA) as previously reported [15]. The normal MF diet served as the control diet, and the only difference between the 2 diets was the addition of 0.8% CQA in the supplemented diet. When 5XFAD mice were 6 weeks old, they were evenly divided based on their body weights to either the control diet group or the CQA diet group. The feeding started when the mice were 7 weeks old. The diet was exchanged to a fresh one every week. The amounts of the added diet and remaining diet were weighed every week for estimation of food intake per each cage for both control and CQA diet groups throughout the experiment. Estimated food intake for control diet and CQA diet was 4.3±0.3 and 4.4±0.4 g/day/mouse, respectively, and there was no significant difference (p = 0.75, Welch’s test). Behavioral tests and brain sampling were conducted at 16 weeks and 21–23 weeks after starting the CQA treatment, respectively.

General behavioral analysis

Behavioral tests were performed during the light phase (zeitgeber time 0 to 4) when mice were 23 weeks old. The same groups of mice were tested in both the NOR test and the Y-maze test. Mice first underwent the NOR test, and after a 24 h interval, they underwent the Y-maze test. Prior to the tests, mice were gently handled by the same experimenter for 5 consecutive days. During the behavioral test, the room temperature was maintained at 23.5°C and illumination in the room was kept at 40 lux. Behavior was recorded with a digital video camera (HDR-CX675, SONY, Japan) at a resolution of 1440×1080 pixels and a frame rate of 30 frames/s. The video was then down-sampled and analyzed using the DeepLabCut (DLC) posture-estimation package [22 –24], which allowed us to track certain body parts (keypoints) of the mice. To analyze the data obtained from the NOR test using DLC, network training was conducted using 284 image frames and the ResNet-50 neural network with 781,000 training iterations. For analyzing data from the Y-maze test, network training was separately conducted using 55 image frames and the ResNet-50 neural network with 140,000 training iterations.

For the analysis of behavioral data in the NOR test based on keypoint-tracking using DLC, the coordinates of 5 keypoints of the mouse (snout, center back, tail base, right ear, left ear) and the 2 objects were labeled using the graphical user interface provided by the DLC package (Supplementary Figure 1). For analyses of behavioral data in the Y-maze test, the coordinates of 3 keypoints of the mouse (snout, center back, tail base) and the arms of the Y-maze apparatus were labeled. Exploration of objects in the NOR test and entrance into the arms in the Y-maze test were quantified using a custom Python code available on Github (https://github.com/hayashi-laboratory/deeplabcut-analysis).

NOR test

For acclimation, mice were placed in an open field box (40 cm×40 cm×40 cm) and given 10 min to explore the arena. Mice were then allowed to rest in their home cage for 24 h. During the training session, mice were again placed in the open field box and presented with 2 identical objects (either 4×3 cm rubber cylinders covered with blue plastic tape or triangular pyramids of a similar size assembled with glass beads and covered with red plastic tape) for 10 min. After a 2-h retention period in the home cage, the mice were returned to the open field box in which 1 of the objects was replaced with the alternative object and again given 10 min to explore. The location and combination of objects were counterbalanced to avoid bias. The discrimination index was calculated as follows: $Discrimination index = \frac{({Time}_{novel} - {Time}_{familiar})}{({Time}_{novel} + {Time}_{familiar})}$

Time_novel is the time spent exploring the novel object and Time_familiar is the time spent exploring the familiar object. The time spent exploring each object was determined by automatically counting frames in which the mouse’s snout touched or sniffed around the object without climbing on it using DLC-labeled keypoint analysis as shown in Supplementary Figure 1.

Y-maze test

A Y-shaped maze consisting of 3 plastic arms (40 cm×12 cm) positioned 120° apart from one another was used. The mice were placed at the end of 1 arm and allowed to freely explore the maze for 8 min. The number of arm entries was analyzed based on the DLC-labeled keypoints. Mice were considered to have entered an arm when both their center back and snout were within the arm area. The spontaneous alternation rate was calculated as follows:

$\begin{matrix} Spontaneous alternation rate (%) \\ = \frac{Number of 3 consecutive entries to different arms}{Total number of arm entries - 2} \times 100 \end{matrix}$

Immunofluorescence

Following transcardial perfusion with phosphate-buffered saline (PBS), dissected brains were divided at the midline and 1 hemisphere was used for immunostaining and the other for enzyme-linked immunosorbent assay (ELISA). Hemispheres used for immunostaining were post-fixed overnight with 4% paraformaldehyde in PBS, equilibrated with 30% sucrose in PBS, frozen, and sectioned at 40μm using a sliding microtome (Yamato Kohki, Japan). The sections were washed with PBS and placed in 0.1% H₂O₂/methanol for 3 min to inactivate endogenous peroxidase activity and washed again with PBS. For Aβ staining, the sections were incubated with a primary antibody for human anti-Aβ, which recognizes the N-terminal region of Aβ42 (#82E1, 1 : 1000 dilution, IBL, Japan). For double staining of NeuN and PSD95, mouse anti-NeuN (#MAB377, 1 : 1000 dilution, Merck Millipore, Burlington, MA, USA) and rabbit anti-PSD95 (#ab238135, 1 : 1000 dilution, Abcam, Cambridge, MA, USA) antibodies were used. For double staining of ionized calcium-binding adapter molecule 1 (Iba1) and glial fibrillary acidic protein (GFAP), rabbit anti-Iba1 (#019-19741, 1 : 1000 dilution, FUJIFILM Wako Pure Chemical, Japan) or chicken anti-GFAP (#ab4674, 1 : 1000 dilution, Abcam) antibody were used. The sections were incubated with primary antibodies at 4°C overnight and washed twice with Tris-buffered saline with 0.05% Tween 20 (TBST; #T9142, Takara Bio, Japan). The sections were incubated with secondary antibodies, i.e., horseradish peroxidase-conjugated donkey anti-mouse IgG (#ab7061, 1 : 200 dilution, Abcam) for Aβ, Alexa488-conjugated donkey anti-mouse IgG (#ab150109, 1 : 200 dilution, Abcam) for NeuN, CF647-conjugated donkey anti-rabbit IgG (#20047, 1;200 dilution, Biotium, Fremont, CA, USA) for PSD95, horseradish peroxidase-conjugated donkey anti-rabbit IgG (#ab7083, 1 : 200 dilution, Abcam) for Iba1, and CF594-conjugated donkey anti-chicken IgY (#SAB4600031, 1 : 200 dilution, MilliporeSigma, St. Louis, MO, USA) for GFAP, at room temperature for 60 min. For Aβ and Iba1 staining, the signals were amplified with the TSA fluorescein system (NEL701A001KT, 1 : 50 dilution, Akoya Biosciences, Brooklyn, NY, USA), i.e., sections were washed twice with TBST, subjected to a 10-min TSA reaction at room temperature and then washed twice with TBST following the manufacturer’s instructions. For Iba1 and GFAP double staining following the reactions with secondary antibodies, the sections were additionally incubated with 1μg/mL of methoxy-X04 (#4920, Tocris Bioscience, UK), which reacts with Aβ and generates a fluorescent signal [25], and then washed with 70% ethanol for 3 min. The sections were cover-slipped using an antifade-aqueous mountant (#TA-006-FM, Thermo Fisher Scientific, Waltham, MA, USA) and imaged using a laser-confocal scanning microscope (LSM800, Carl Zeiss, Germany) through a 10× or 40× objective lens. Fluorescent signals were analyzed using Fiji software [26] (available on https://github.com/hayashi-laboratory/ImmunofluorescenceAnalysis).

Aβ₄₂ quantification by ELISA

Soluble and insoluble fractions were obtained from the cerebral cortex and hippocampus as previously described [27]. Briefly, the dissected hippocampus and cerebral cortex were homogenized in 50 mM Tris-HCl, pH 7.6, containing 150 mM NaCl and protease inhibitor cocktail (cOmplete, #4693159001, Roche, Germany) using a disruptor with a zirconia bead (φ7) shaking at 1600 rpm for 60 s. Then, the homogenate was ultracentrifuged at 200,000×g for 20 min at 4°C and the supernatant was collected as the soluble fraction. Guanidine-HCl (6M; #17356-24, Nacalai Tesque, Japan) containing the protease inhibitor cocktail was added to the pellet and sonicated to dissolve it. The solution was centrifuged again at 200,000×g for 20 min at 4°C and the supernatant was collected as the insoluble fraction. The soluble and insoluble fractions were diluted 40-fold and 12,000–24,000-fold, respectively. Sandwich ELISA for Aβ_X–42 was conducted with Human/Rat β Amyloid (42) ELISA Kit Wako (#290-62601, FUJIFILM Wako Pure Chemical) in accordance with the manufacturer’s instructions using a microplate washer machine (#5165000, Thermo Fisher Scientific) and measuring the absorbance at 405 nm with a microplate reader (Multiskan™ FC, Thermo Fisher Scientific). Finally, the Aβ₄₂ concentration was calculated from the standard curve.

RNA extraction

The cerebral cortex was dissected and homogenized with TissueLyser LT (#85600, Qiagen, Germany), and total RNA was extracted with the RNeasy Plus Universal Mini Kit (#73404, Qiagen) according to the manufacturer’s instructions. Subsequently, RNA samples were treated with the RNase-free DNase kit (#79254, Qiagen) and further cleaned with the RNeasy MinElute Cleanup kit (#74204, Qiagen) according to the manufacturer’s instructions. The quantity of the RNA was measured using the Qubit Fluorometer (#Q33216, Thermo Fisher Scientific) and the Qubit RNA Broad Assay Kit (#Q10210, Thermo Fisher Scientific).

Ion AmpliSeq

The Ion AmpliSeq Transcriptome mouse Gene Expression kit (#A36554, Thermo Fisher Scientific) was used to prepare the sequencing library. Briefly, 20 ng of RNA was reverse transcribed with VILO Reaction Mix and SuperScript III Enzyme (included in the above kit) at 25°C for 10 min and 42°C for 90 min, and then incubated at 85°C for 5 min to generate cDNA. The target cDNA was amplified by mixing the cDNA solution, Ion AmpliSeq HiFi Mix (included in the above kit), and Ion AmpliSeq Transcriptome Mouse Gene Expression Core Panel (included in the above kit) under the following conditions: 99°C for 15 s and 62°C for 16 min for 12 cycles. After digesting the primer sequence with FuPa reagent (included in the above kit), Ion Xpress Barcode adapters (#4471250, Thermo Fisher Scientific) were ligated. The sequencing library was purified by mixing with AMPure XP beads (#A63880, Beckman Coulter, USA) according to the manufacturer’s instructions. A quality check of the sequencing library was performed using the High Sensitivity D1000 ScreenTape (#5067-5582, Agilent Technologies, Santa Clara, CA, USA) on the Agilent 4200 TapeStation. The sequencing library was quantified using the Ion Library TaqManTM Quantitation Kit (#4468802, Thermo Fisher Scientific). The templates for sequencing were prepared using the Ion Chef System (#A30011, Thermo Fisher Scientific) followed by sequencing with the Ion GeneStudio S5 prime sequencer system (#A37260, Thermo Fisher Scientific). As for the AmpliSeq read counts analysis, differentially expressed genes were detected using DESeq2 of iDEP94 (http://bioinformatics.sdstate.edu/idep94/) and enrichment pathway analysis was performed [28].

qRT-PCR

Total RNA (200 ng) extracted from the cerebral cortex was subjected to reverse transcription with SuperScript IV VILO Master Mix (#11756050, Thermo Fisher Scientific) to generate cDNA in accordance with the manufacturer’s instructions. cDNA was mixed with a commercial quantitative (q)PCR reagent (TaqMan Fast Universal PCR Master Mix, #4352042, Thermo Fisher Scientific) and either of the following TaqMan probes (TaqMan Gene Expression Assay FAM, #4331182, ThermoFisher Scientific): NeuN (Rbfox3), Mm01248771_m1; Psd95 (Dlg4), Mm00492193_m1; Homer1, Mm00516275_m1; Apoe, Mm01307193_g1; Synaptophysin, Mm00436850_m1; Gapdh, Mm99999915_g1; Map2, Mm00485230_m1; cFos, Mm00487425_m1; Fosb, Mm00500401_m1; Arc, Mm01204954_g1; Iba1 (Aif1), Mm00479862_g1; Tmem119, Mm00525305_m1; P2ry12, Mm01950543_s1; Cd68, Mm03047343_m1; Cd11b (Itgam), Mm00434455_m1; Gfap, Mm01253033_m1. The PCR reaction and analysis were conducted using the QuantStudio 3 System (Thermo Fisher Scientific), in which cDNA was quantified by fitting Ct values to the standard curve. The relative expression level of each gene was calculated by normalization with Gapdh.

Statistical analyses

Data analyses and graph generation were conducted using Python3.9 unless otherwise stated. Line plots with error bars were used to represent the mean±standard error. Box plots were used to represent the median±interquartile range (IQR), with whiskers indicating 1.5 times the IQR; individual sample points are also shown. Comparisons between diets in 5XFAD mice were tested with Welch’s t-test. Comparisons between diets over time or between diets among multiple brain regions were conducted using a 2-way analysis of variance (ANOVA) for mixed design. Statistical tests were performed using the Pingouin package in Python [29]. Statistical significance was set at p < 0.05. Where applicable, all tests were 2-tailed.

Results

CQA treatment ameliorates cognitive memory decline in 5XFAD mice

To assess the effect of CQA treatment on cognitive impairment in 5XFAD mice, we evaluated cognitive memory and spatial working memory using the NOR test and Y-maze test, respectively, following 16 weeks of feeding with the CQA-containing diet. In advance, we measured their body weights, starting from when mice were 6 weeks old (before the start of CQA treatment) until they reached 23 weeks old (the age at which behavioral tests were conducted). No significant difference in body weight was detected between CQA-treated and control 5XFAD mice (Fig. 1A, B), suggesting that CQA-treated mice consumed the same amount of diet as the control mice. In the NOR test, mice were placed into an arena and presented with 2 identical objects during the training session. After a 2-h retention period spent in the home cage, the mice were placed back into the arena, in which 1 of the objects was replaced with a novel object. If an animal retains memory of exploring an object during the training session, it will spend more time exploring the novel object. We adopted DLC, which allows semi-automatic tracking of various mouse body parts (keypoints) based on transfer learning with deep neural networks [20, 21]. Using DLC, we quantified the time each mouse spent in each part of the arena in an unbiased manner (Fig. 1C). The discrimination index of the control 5XFAD mice was nearly 0, suggesting that these mice did not retain memory of exploring the objects. The discrimination index was significantly higher in the CQA-treated 5XFAD mice than in the control 5XFAD mice (Fig. 1D, p = 0.025, Welch’s test).

Fig. 1

CQA treatment mitigates cognitive memory decline in 5XFAD mice. A, B) Comparison of the temporal change in body weights (A) and body weights when mice were 23 weeks old (B) between control (Cont) and CQA-treated (CQA) 5XFAD mice. P-values in the 2-way ANOVA for mixed design (A) and the Welch’s test (B) are shown. Control 5XFAD, n = 14 mice; CQA-treated 5XFAD, n = 13 mice. C) Representative examples of heatmaps depicting the duration of time the mouse’s snout was in each position within the arena during the NOR test. The heatmap color represents the time fraction. White circles indicate the position of either the novel or familiar object. Examples of WT mice and control or CQA-treated 5XFAD mice are shown. D) Comparison of discrimination index scores in the NOR test between control and CQA-treated 5XFAD mice. Results of WT mice are also shown. WT, n = 19 mice; Control 5XFAD, n = 12 mice; CQA-treated 5XFAD, n = 12 mice. E, F) Comparison of the Y-maze test alternation rate (E) and the number of arm entries (F). WT, n = 20 mice; Control 5XFAD, n = 14 mice; CQA-treated 5XFAD, n = 13 mice. p-values in the Welch’s test are shown (*p < 0.05).

In the Y-maze test, mice are placed into an arena with 3 arms of identical length at 120° angles and are allowed to freely explore the arena. If the animal has intact spatial working memory, it will remember the arm most recently visited and show a tendency toward entering the other arm, i.e., exhibit a high spontaneous alternation rate. CQA-treated 5XFAD mice tended to have a higher spontaneous alternation rate than control 5XFAD mice (Fig. 1E, p = 0.07, Welch’s test).

CQA treatment does not affect the Aβ plaque burden or glial activation in 5XFAD mice

To assess the effect of CQA treatment on Aβ plaque formation in 5XFAD mice, we performed immunostaining using an antibody against human Aβ, focusing on the somatosensory cortex layer 5, subiculum, entorhinal cortex, dentate gyrus, cornu ammonis, and amygdala, areas with high levels of Aβ accumulation in 5XFAD mice. No significant difference in the Aβ plaque area or size was observed between the CQA-treated and control 5XFAD mice (Fig. 2A–D). We also conducted immunostaining for microglia and astrocytes using an Iba1 or GFAP antibody, respectively. No significant differences in the glia-positive area and size were detected in the somatosensory cortex layer 5, subiculum, or dentate gyrus between the CQA-treated and control 5XFAD mice (Fig. 2E, F). In addition, we quantified the area that overlapped between microglia and Aβ plaques by simultaneously immunostaining for IBA1 and staining with methoxy X-04, which reacts with Aβ and generates fluorescent signals [25]. No significant difference was detected between the CQA-treated and control 5XFAD mice (Fig. 2G, H).

Fig. 2

Lack of effect of CQA treatment on the Aβ plaque burden and neuroinflammation in the cerebral cortex and hippocampus of 5XFAD mice. A, B) Representative immunofluorescence images of anti-Aβ antibody staining in the left brain hemisphere from control (Cont) (A) and CQA-treated (CQA) (B) 5XFAD mice. Scale bar, 500μm. C, D) Comparison of the Aβ plaque area fraction (%) (C) and average size (μm²) (D) in the indicated brain regions. Control 5XFAD, n = 10 mice; CQA-treated 5XFAD, n = 10 mice. E, F) Comparison of anti-Iba1 (E) or anti-GFAP (F) immunofluorescence signals. Representative images of cortical layer 5 (left) and quantitative analyses of immunofluorescence signals (area (%) and size (μm²)) in the indicated brain regions (right) are shown. Scale bar, 50μm. Control 5XFAD, n = 5-6 mice; CQA-treated 5XFAD, n = 6-7 mice. G) Representative immunofluorescence image of cortical layer 5 following double fluorescent staining with anti-Iba1 antibody (green) and the Aβ-fluorescent marker methoxy X04 (magenta). Scale bar, 100μm. H) Comparison of the ratio of the area that overlapped between anti-Iba1 and methoxy X04 to the Iba1-positive area in the indicated brain regions. Control 5XFAD, n = 5-6 mice; CQA-treated 5XFAD, n = 6 mice. p-values in the 2-way ANOVA for mixed design are shown. L5, layer 5; Sub, subiculum; Ent, Entorhinal cortex; DG, Dentate gyrus; CA, Cornu Ammonis; Amy, Amygdala.

CQA treatment does not affect Aβ peptide levels in 5XFAD mice

We next compared the Aβ peptide levels by ELISA in the cerebral cortex and hippocampus between CQA-treated and control 5XFAD mice. No significant difference in the concentration of soluble or insoluble Aβ was detected in the cerebral cortex and hippocampus between CQA-treated and control mice (Fig. 3), indicating that CQA treatment does not affect Aβ levels in the brains of 5XFAD mice.

Fig. 3

Lack of effect of CQA treatment on Aβ_1 - 42 levels in the cerebral cortex and hippocampus of 5XFAD mice. A-D) Comparison of Aβ_1 - 42 levels measured by ELISA in the soluble fraction (A and C) or the insoluble fraction (B and D) derived from the cortex (A and B) or hippocampus (C and D) between control (Cont) and CQA-treated (CQA) 5XFAD mice. Results of WT mice are also shown. WT, n = 3 mice; Control 5XFAD, n = 14 mice; CQA-treated 5XFAD, n = 13 mice. p-values in the Welch’s test are shown. n.d., not detected.

CQA treatment reduces neuronal loss and increases PSD95 intensity in 5XFAD mice

In 5XFAD mice, neuronal loss occurs in cortical layer 5 and in the subiculum, areas that show high levels of plaque formation [18, 30]. To test whether CQA treatment affects the loss of mature neurons in 5XFAD mice, we performed immunostaining with anti-NeuN antibodies, which label mature neurons (Fig. 4A, C). Consistent with previous reports [18], our immunofluorescence analyses revealed that the average numbers of mature neurons in cortical layer 5 and the subiculum in the 5XFAD mice were approximately 70% and 50%, respectively, of those in WT mice (Fig. 4C). We found that CQA treatment increased the number of mature neurons in both cortical layer 5 and the subiculum (Fig. 4A, C). Reduction of PSD95, a major scaffold protein that maintains the structural and functional integrity of excitatory synapses, may mark postsynaptic degeneration and contribute to functional deficits in 5XFAD mice [31]. Thus, we next conducted immunostaining with anti-PSD95 antibody to assess the effect of CQA on the abundance of PSD95. In 5XFAD mice, the average intensity of PSD95 was approximately 80% of that in WT mice in both the cortical layer 5 and the subiculum (Fig. 4B, D), consistent with the findings of a previous study [31]. CQA treatment increased the PSD95 fluorescence intensity in cortical layer 5 (Fig. 4B, D). Whether this is due to an elevated expression level of PSD95, an increased number of functional synapses, or an increased number of mature neurons remains unclear.

Fig. 4

CQA treatment mitigates the loss of mature neurons in the cerebral cortex and subiculum of 5XFAD mice. A, B) Representative immunofluorescence images of the cortex (top) or the subiculum (bottom) stained with anti-NeuN (A) or anti-PSD95 (B) antibody. Scale bar, 100μm. L1, layer1; L2-3, layer 2 and 3; L4, layer 4; L5, layer 5; L6, layer 6. C, D) Comparison of the density of cells labeled with anti-NeuN antibodies (C) or the average intensity of immunofluorescent signals labeled with anti-PSD95 antibodies (D) in cortical layer 5 (left) or the subiculum (right) between control (Cont) and CQA-treated (CQA) 5XFAD mice. Results of WT mice are also shown. WT, n = 6 mice; Control 5XFAD, n = 7 mice; CQA-treated 5XFAD, n = 8 mice. p-values in the Welch’s test are shown (*p < 0.05).

Transcriptome analyses and qRT-PCR revealed that CQA treatment leads to increased levels of mRNA of genes related to synaptic function in 5XFAD mice

To further investigate the effect of CQA treatment on the brains of 5XFAD mice, we attempted

to explore cellular or molecular pathways that were altered by CQA treatment. We performed AmpliSeq transcriptome analysis on RNA extracted from the cerebral cortex. The mRNA levels of 252 genes were increased and the mRNA levels of 13 genes were decreased in CQA-treated mice compared with control mice (Fig. 5A). Pathway analysis, using PGSEA, suggested that CQA affects synaptic and mitochondrial functions, and neural components as defined by Gene Ontology terms (Fig. 5B). Significant changes in the mRNA levels of genes well known to be associated with these pathways could not be detected from the AmpliSeq read counts, however, perhaps due to the small sample size (n = 5). We thus performed qRT-PCR analysis on the cerebral cortex and found that CQA treatment significantly increased the mRNA levels of mature neuronal markers (NeuN and Map2), genes involved in synaptic function (Synaptophysin, Psd95, and Homer1), and neuronal activity makers (cFos, Fosb, and Arc) (Fig. 5C–J). In contrast, the mRNA levels of genes expressed in glial cells, including microglial marker genes (Tmem119, P2ry12, and Iba1), activated microglial marker genes (Cd68 and Cd11b), astrocyte marker gene (Gfap), or Apoe, which is expressed in both microglia and astrocytes, were not altered by CQA treatment (Fig. 5K–Q), consistent with findings from the immunostaining analysis.

Fig. 5

CQA treatment increases mRNA levels of genes related to synaptic function in 5XFAD mice. A) Volcano plot showing results of AmpliSeq applied to bulk RNA obtained from the cerebral cortex of control (Cont) or CQA-treated (CQA) 5XFAD mice (n = 5 mice each). FC, fold-change; FDR, false discovery rate. B) List of Gene Ontology terms (molecular functions) annotated by the gene sets whose mRNA expression levels were altered by CQA treatment in 5XFAD mice based on AmpliSeq data. C–Q) Comparison of mRNA levels of genes related to synaptic function, neuronal markers, and glial markers quantified by qRT-PCR. Control 5XFAD, n = 13 mice; CQA-treated 5XFAD, n = 10 mice. p-values in the Welch’s test are shown (*p < 0.05).

Fig. 6

Graphical abstract of the present study.

Discussion

The findings of the present study indicate that CQA treatment in 5XFAD model mice mitigates neuronal loss, increases mRNA levels of genes associated with synaptic function in the cerebral cortex, and may improve memory deficits compared with control 5XFAD mice. Importantly, Aβ levels and neuroinflammation were not affected by CQA treatment, suggesting that CQA treatment potentially provides an alternative therapeutic approach for AD from those targeting Aβ or neuroinflammation.

To our knowledge, besides the current study, only 2 other studies using rodent AD models have investigated the effects of CQA treatment on cognitive function. One study demonstrated that CQA treatment in PS2APP mice improved cognitive memory functions assessed by the NOR test and reduced the Aβ plaque burden in the brain [15]. Another study revealed that CQA treatment improves memory function in 5XFAD mice as assessed by the fear conditioning paradigm [16]. Plant extracts derived from green coffee beans or Centella asiatica, both of which contain CQA, are also reported to mitigate cognitive decline in AD mouse models [19, 32]. The current study further supports the positive effects of CQA treatment in mitigating memory decline associated with AD.

In this study, the impact of CQA treatment on Aβ levels was evaluated by immunostaining and ELISA. Surprisingly, CQA treatment had no effect on Aβ levels in the cortex and hippocampus, in contrast to previous reports demonstrating that CQA treatment reduced the Aβ plaque burden in the cortex and hippocampus of PS2APP model mice [15]. Other reports described that treatment with extracts from coffee beans or Centella asiatica reduced the Aβ plaque burden in the hippocampus but not in the cerebral cortex of PS2APP and 5XFAD model mice [19, 20]. Importantly, compared with the above studies that assessed the Aβ plaque burden, the current study used either a different AD mouse model (PS2APP versus 5XFAD) or a different form of CQA (crude extract containing CQA versus purified CQA). Thus, the lack of an ameliorative effect on the Aβ plaque burden in our study may be attributed to these differences and the results should thus be interpreted cautiously. The 5XFAD mouse model used in the present study is designed to overexpress humanized APP with 3 AD-linked mutations and humanized presenilin PS1 with 2 AD-linked mutations, resulting in the rapid accumulation of Aβ₄₂ in the cerebrum starting as early as 2 months of age [18], which is much earlier than that observed in PS2APP mice (9 months) [33]. While a direct comparison may not be feasible due to the different conditions, our investigation revealed a remarkable concentration of insoluble Aβ₄₂ levels in the cerebral cortex of 5XFAD mice at 7 months (∼30,000 pmol/g or ∼135 ng/mg), a level much higher than that in the PS2APP mice measured in a previous study (<10 ng/mg at 9 months) [33]. Perhaps the high Aβ₄₂ levels and plaque burden in 5XFAD mice cannot be reduced by CQA in our experimental conditions. If CQA treatment were initiated earlier, for example from prenatal periods when accumulation of Aβ is still not very high, it might affect the Aβ plaque burden even in 5XFAD mice. Additionally, although the effect of 0.8% CQA diet has been established, different amount of CQA might result in different outcomes. Moreover, in the 5XFAD model, a previous study demonstrated that there are sex differences in amyloid pathology [34]. A limitation of our study is that only male mice were evaluated, and further investigation is required to evaluate differences in effects between males and females. Nevertheless, our study highlights the potential for CQA to ameliorate cognitive decline in an Aβ-independent manner. Several clinical trials of AD therapeutics focusing on the removal of Aβ have been conducted targeting early AD patients, and several reagents were revealed to effectively reduce the Aβ plaque burden as measured by positron-emission tomography [4 , 36]. With the exception of lecanemab (BAN2401) and donanemab (LY3002813), however, most interventions did not significantly improve clinical cognition scales [35, 36]. Even for lecanemab, which has demonstrated efficacy in early-stage AD patients and has been approved by the US Food and Drug Administration, the effectiveness for late-stage AD patients remains unclear. Our findings may shed light on novel Aβ-independent therapeutic approaches for AD, which could be particularly crucial for patients in which a reduction of Aβ alone is insufficient for cognitive improvement.

5XFAD mice exhibit a reduction of synaptic marker proteins and Aβ accumulation at 4 months, and neuronal loss occurs in layer 5 cortical pyramidal neurons by 9 months [18]. The present study showed that layer 5 cortical neurons of 5XFAD mice are reduced to approximately 70% of those in WT mice at 7 months. Moreover, we found that 5XFAD mice that underwent CQA treatment exhibited higher numbers of mature neurons in cerebral cortex layer 5 and the subiculum compared with control 5XFAD mice. Thus, CQA treatment may reduce the neuronal loss that accompanies AD via unknown mechanisms. Considering that the loss of PSD95 induces neuronal cell death [37] and synaptic dysfunction likely promotes neuronal death, the protective effect of CQA treatment on neuronal loss may be partly attributed to its ability to alleviate synaptic dysfunction. Indeed, the histologic results showed that CQA treatment in 5XFAD mice leads to higher PSD95 levels in the cortex. Whether this is due to an elevated expression level of PSD95, an increased number of functional synapses, or an increased number of mature neurons was not determined in the present study. Findings from the cortical bulk AmpliSeq transcriptome analysis and subsequent pathway analysis further supported an association between CQA treatment and increased intensities of pathways involved in synaptic function. Indeed, the mRNA levels of Synaptophysin, Homer1a, Psd95, and Map2 were all elevated by CQA treatment. Moreover, mRNA levels of the immediate early genes cFos, Fosb, and Arc were also elevated by CQA treatment. Thus, there is a possibility that CQA treatment ameliorates cognitive decline by potentiating synaptic functions and increasing neuronal activity in 5XFAD mice. Based on the present findings, however, we cannot conclude whether the increased mRNA levels of these genes by CQA treatment are due to a direct upregulation of gene expression or are a secondary effect of increased neuronal survival. Further studies are needed to verify the effects of CQA on the function of these proteins. Additionally, we are not sure whether the changes in neuronal numbers and synaptic intensities detected in the cerebral cortex layer 5 or the subiculum by immunostaining and the differences in mRNA expression across the entire cerebral cortex detected by qRT-PCR represent the same or distinct aspects of the CQA effect. The current study does not provide comprehensive analyses regarding neuronal loss or synaptic intensities across broad brain areas, which should be further investigated in the future.

Additionally, CQA is an electron-rich compound containing a phenol hydroxyl group that acts as a scavenger against free radicals, thereby exhibiting antioxidant effects [38], which may prevent oxidative stress-induced cell death [39, 40]. CQA can reduce NMDA-dependent neuronal death in primary cultures associated with high levels of glutamate [41]. Furthermore, extracts from Centella asiatica, which contains CQA, can induce Nrf2 signaling and reduce oxidative stress associated with the pathology of AD [32]. In our AmpliSeq transcriptome analysis and subsequent pathway analysis, however, we detected no enrichment of pathways related to oxidative stress.

The beneficial effects of CQA treatment on neuronal and/or synaptic loss may underlie the improved performance of 5XFAD mice in the NOR test. The cerebral cortex, especially visual cortical areas V1 and V2, and the hippocampus are involved in cognitive memory based on visual and spatial information [42 –45], and synaptic plasticity of the cortex contributes to memory storage [46]. Moreover, efferents from the visual cortex projecting through the subiculum to the hippocampal CA1 and the entorhinal cortex are involved in memory [47,48, 47,48]. Thus, improved maintenance of these circuits by CQA treatment may have contributed to the improved performance of the 5XFAD mice in the NOR test.

Neuroinflammation, in which microglia and astrocytes play key roles, is a major contributor to AD pathogenesis [49]. In our study, we found no effect of CQA treatment on the area and size of Iba1-positive microglial cells and GFAP-positive astrocytes, the density of microglia around Aβ plaques, or the mRNA levels of Iba1, Cd68, Cd11b, P2ry12, Tmem119, Gfap, and Apoe, suggesting that CQA has no effect on neuroinflammation. These results further support the notion that CQA could exert its beneficial effects on AD by directly acting on neurons.

Currently, whether CQA or its derivative exerts its effects by passing the blood-brain barrier and directly acting on neurons remains unknown. Following ingestion, CQA undergoes metabolism via gut microbial conversion, generating caffeic acid and its methylated metabolite, ferulic acid [38]. Caffeic acid, which is solely derived from the diet, improves memory in rats [50] and is detected in the cerebrospinal fluid in humans [51]. Thus, brain-permeable forms of CQA metabolites might be involved.

In conclusion, the findings of this study suggest that CQA, which is contained in coffee beans and Centella asiatica, may provide therapeutic or preventive effects for AD via novel mechanisms related to an attenuation of neuronal loss and enhanced synaptic function independent of a reduction in Aβ or neuroinflammation. Considering these findings, CQA treatment might also be effective for cognitive decline unrelated to AD, such as natural ageing or other neurodegenerative diseases, which should be tested in the future.

AUTHOR CONTRIBUTIONS

Takaya Suganuma (Conceptualization; Formal analysis; Investigation; Methodology; Software; Visualization; Writing – original draft); Sena Hatori (Investigation; Visualization; Writing – review & editing); Chung-Kuan Chen (Data curation; Formal analysis; Methodology; Software; Validation; Writing – review & editing); Satoshi Hori (Data curation; Investigation); Mika Kanuka (Methodology; Resources); Chih-Yao Liu (Resources; Validation); Chika Tatsuzawa (Investigation); Masashi Yanagisawa (Project administration; Supervision); Yu Hayashi (Conceptualization; Funding acquisition; Project administration; Resources; Supervision; Writing – review & editing)

Footnotes

ACKNOWLEDGMENTS

We thank Shinnosuke Yasugaki for technical advice and Naoki Yamamoto, Takatoshi Murase, and Tadashi Hase for discussion and providing reagents. English-language editing of the manuscript was provided by SciTechEdit International LLC (Highlands Ranch, CO, USA).

FUNDING

This work was supported by AMED under Grant Numbers JP19gm1110008, JP21wm0425018, and JP21zf0175005; JSPS KAKENHI under Grant Numbers JP22K19354, JP23H04668, and JP23H04210; the Asahi Glass Foundation; the Kao Foundation for Research on Health Science; and financial support from Kao Corporation (to YH).

CONFLICT OF INTEREST

T. S. and S. H. are employees of Kao Corporation. This study received financial support from Kao Corporation (to YH).

DATA AVAILABILITY

All data are available upon reasonable request to the corresponding author.

The supplementary material is available in the electronic version of this article: .

References

Alzheimer’s Association (2021) 2021 Alzheimer’s disease facts and figures. Alzheimers Dement 17, 327–406.

Abubakar

, Sanusi

, Ugusman

, Mohamed

, Kamal

, Ibrahim

, Khoo

, Kumar

(2022) Alzheimer’s disease: An update and insights into pathophysiology. Front Aging Neurosci 14, 742408.

Chen

, Fu

AKY

, Ip

(2019) Synaptic dysfunction in Alzheimer’s disease: Mechanisms and therapeutic strategies. Pharmacol Ther 195, 186–198.

Tolar

, Abushakra

, Hey

, Porsteinsson

, Sabbagh

(2020) Aducanumab, gantenerumab, BAN2401, and ALZ-801-the first wave of amyloid-targeting drugs for Alzheimer’s disease with potential for near term approval. Alzheimers Res Ther 12, 95.

Sevigny

, Chiao

, Bussière

, Weinreb

, Williams

, Maier

, Dunstan

, Salloway

, Chen

, Ling

, O’Gorman

, Qian

, Arastu

, Li

, Chollate

, Brennan

, Quintero-Monzon

, Scannevin

, Arnold

, Engber

, Rhodes

, Ferrero

, Hang

, Mikulskis

, Grimm

, Hock

, Nitsch

, Sandrock

(2016) The antibody aducanumab reduces Aβ plaques in Alzheimer’s disease. Nature 537, 50–56.

Salloway

, Sperling

, Fox

, Blennow

, Klunk

, Raskind

, Sabbagh

, Honig

, Porsteinsson

, Ferris

, Reichert

, Ketter

, Nejadnik

, Guenzler

, Miloslavsky

, Wang

, Lu

, Lull

, Tudor

, Liu

, Grundman

, Yuen

, Black

, Brashear

(2014) Two phase 3 trials of bapineuzumab in mild-to-moderate Alzheimer’s disease. N Engl J Med 370, 322–333.

Cummings

, Rabinovici

, Atri

, Aisen

, Apostolova

, Hendrix

, Sabbagh

, Selkoe

, Weiner

, Salloway

(2022) Aducanumab: Appropriate use recommendations update. J Prev Alzheimers Dis 9, 221–230.

Mahase

(2023) Alzheimer’s disease: Lecanemab gets full FDA approval and black box safety warning. BMJ 382, p1580.

Carneiro

, Oliveira

MBPP

, Alves

(2021) Neuroprotective properties of coffee: An update. Trends Food Sci Technol 113, 167–179.

10.

Kim

, Byun

, Yi

, Lee

, Jeon

, Jung

, Lee

, Sohn

, Lee

J-Y

, Kim

, Shin

, Sohn

C-H

, Lee

(2019) Coffee intake and decreased amyloid pathology in human brain. Transl Psychiatry 9, 1–10.

11.

Gardener

, Rainey-Smith

, Villemagne

, Fripp

, Doré

, Bourgeat

, Taddei

, Fowler

, Masters

, Maruff

, Rowe

, Ames

, Martins

, the AIBL Investigators (2021) Higher coffee consumption is associated with slower cognitive decline and less cerebral a β-amyloid accumulation over 126 months: Data from the Australian Imaging, Biomarkers, and Lifestyle Study. Front Aging Neurosci 13, 681.

12.

Monteiro

, Farah

, Perrone

, Trugo

, Donangelo

(2007) Chlorogenic acid compounds from coffee are differentially absorbed and metabolized in humans. J Nutr 137, 2196–2201.

13.

Gray

, Zweig

, Murchison

, Caruso

, Matthews

, Kawamoto

, Harris

, Quinn

, Soumyanath

(2017) Centella asiatica attenuates Aβ – induced neurodegenerative spine loss and dendritic simplification. Neurosci Lett 646, 24–29.

14.

Saitou

, Ochiai

, Kozuma

, Sato

, Koikeda

, Osaki

, Katsuragi

(2018) Effect of chlorogenic acids on cognitive function: A randomized, double-blind, placebo-controlled trial. Nutrients 10, 1337.

15.

Ishida

, Misawa

, Nishimura

, Hirata

, Yamamoto

, Ota

(2020) 5-caffeoylquinic acid ameliorates cognitive decline and reduces Aβ deposition by modulating Aβ clearance pathways in APP/PS2 transgenic mice. Nutrients 12, E494.

16.

Matthews

, Caruso

, Alcazar Magana

, Wright

, Maier

, Stevens

, Gray

, Quinn

, Soumyanath

(2020) Caffeoylquinic acids in centella asiatica reverse cognitive deficits in male 5XFAD Alzheimer’s disease model mice. Nutrients 12, 3488.

17.

Toda

, Noda

, Ito

, Maeda

, Shimizu

(2010) Presenilin-2 mutation causes early amyloid accumulation and memory impairment in a transgenic mouse model of Alzheimer’s disease. Biomed Res Int 2011, e617974.

18.

Oakley

, Cole

, Logan

, Maus

, Shao

, Craft

, Guillozet-Bongaarts

, Ohno

, Disterhoft

, Eldik

, Berry

, Vassar

(2006) Intraneuronal β-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: Potential factors in amyloid plaque formation. J Neurosci 26, 10129–10140.

19.

Ishida

, Yamamoto

, Misawa

, Nishimura

, Misawa

, Ota

, Shimotoyodome

(2020) Coffee polyphenols prevent cognitive dysfunction and suppress amyloid β plaques in APP/PS2 transgenic mouse. Neurosci Res 154, 35–44.

20.

Gray

, Zweig

, Caruso

, Zhu

, Wright

, Quinn

, Soumyanath

(2018) Centella asiatica attenuates hippocampal mitochondrial dysfunction and improves memory and executive function in β-amyloid overexpressing mice. Mol Cell Neurosci 93, 1–9.

21.

Gray

, Morré

, Kelley

, Maier

, Stevens

, Quinn

, Soumyanath

(2014) Caffeoylquinic acids in Centella asiatica protect against β-amyloid toxicity. J Alzheimers Dis 40, 359–373.

22.

Mathis

, Mamidanna

, Cury

, Abe

, Murthy

, Mathis

, Bethge

(2018) DeepLabCut: Markerless pose estimation of user-defined body parts with deep learning. Nat Neurosci 21, 1281–1289.

23.

Nath

, Mathis

, Chen

, Patel

, Bethge

, Mathis

(2019) Using DeepLabCut for 3D markerless pose estimation across species and behaviors. Nat Protoc 14, 2152–2176.

24.

, Kejzar

, Sasaguri

, Saito

, Saido

, De Strooper

, Bauza

, Krupic

(2023) A fully automated home cage for long-term continuous phenotyping of mouse cognition and behavior. Cell Rep Methods 3, 100532.

25.

Klunk

, Bacskai

, Mathis

, Kajdasz

, McLellan

, Frosch

, Debnath

, Holt

, Wang

, Hyman

(2002) Imaging Abeta plaques in living transgenic mice with multiphoton microscopy and methoxy-X04, a systemically administered Congo red derivative. J Neuropathol Exp Neurol 61, 797–805.

26.

Schindelin

, Arganda-Carreras

, Frise

, Kaynig

, Longair

, Pietzsch

, Preibisch

, Rueden

, Saalfeld

, Schmid

, Tinevez

J-Y

, White

, Hartenstein

, Eliceiri

, Tomancak

, Cardona

(2012) Fiji: An open-source platform for biological-image analysis. Nat Methods 9, 676–682.

27.

Iwata

, Mizukami

, Shirotani

, Takaki

, Muramatsu

, Lu

, Gerard

, Ozawa

, Saido

(2004) Presynaptic localization of neprilysin contributes to efficient clearance of amyloid-β peptide in mouse brain. J Neurosci 24, 991–998.

28.

, Son

, Yao

(2018) iDEP: An integrated web application for differential expression and pathway analysis of RNA-Seq data. BMC Bioinformatics 19, 534.

29.

Vallat

(2018) Pingouin: Statistics in Python. J Open Source Softw 3, 1026.

30.

Eimer

, Vassar

(2013) Neuron loss in the 5XFAD mouse model of Alzheimer’s disease correlates with intraneuronal Aβ42 accumulation and Caspase-3 activation. Mol Neurodegener 8, 2.

31.

Shao

, Mirra

, Sait

HBR

, Sacktor

, Sigurdsson

(2011) Postsynaptic degeneration as revealed by PSD-95 reduction occurs after advanced Aβ and tau pathology in transgenic mouse models of Alzheimer’s disease. Acta Neuropathol (Berl) 122, 285–292.

32.

Zweig

, Brandes

, Brumbach

, Caruso

, Wright

, Quinn

, Soumyanath

, Gray

(2021) Prolonged treatment with Centella asiatica improves memory, reduces amyloid-β pathology, and activates NRF2-regulated antioxidant response pathway in 5xFAD mice. J Alzheimers Dis 81, 1453–1468.

33.

Richards

, Higgins

, Ouagazzal

A-M

, Ozmen

, Kew

JNC

, Bohrmann

, Malherbe

, Brockhaus

, Loetscher

, Czech

, Huber

, Bluethmann

, Jacobsen

, Kemp

(2003) PS2APP transgenic mice, coexpressing hPS2mut and hAPPswe, show age-related cognitive deficits associated with discrete brain amyloid deposition and inflammation. J Neurosci 23, 8989–9003.

34.

Poon

, Wong

STN

, Roy

, Wang

, Chan

HWH

, Steinbusch

, Blokland

, Temel

, Aquili

, Lim

(2023) Sex differences between neuronal loss and the early onset of amyloid deposits and behavioral consequences in 5xFAD transgenic mouse as a model for Alzheimer’s disease. Cells 12, 780.

35.

van Dyck

, Swanson

, Aisen

, Bateman

, Chen

, Gee

, Kanekiyo

, Li

, Reyderman

, Cohen

, Froelich

, Katayama

, Sabbagh

, Vellas

, Watson

, Dhadda

, Irizarry

, Kramer

, Iwatsubo

(2023) Lecanemab in early Alzheimer’s disease. N Engl J Med 388, 9–21.

36.

Shcherbinin

, Evans

, Lu

, Andersen

, Pontecorvo

, Willis

, Gueorguieva

, Hauck

, Brooks

, Mintun

, Sims

(2022) Association of amyloid reduction after donanemab treatment with tau pathology and clinical outcomes. JAMA Neurol 79, 1015–1024.

37.

Gardoni

, Bellone

, Viviani

, Marinovich

, Meli

, Pellegrini-Giampietro

, Cattabeni

, Di Luca

(2002) Lack of PSD-95 drives hippocampal neuronal cell death through activation of an αCaMKII transduction pathway. Eur J Neurosci 16, 777–786.

38.

Alcázar Magaña

, Kamimura

, Soumyanath

, Stevens

, Maier

(2021) Caffeoylquinic acids: Chemistry, biosynthesis, occurrence, analytical challenges, and bioactivity. Plant J 107, 1299–1319.

39.

Schroeter

, Williams

, Matin

, Iversen

, Rice-Evans

(2000) Phenolic antioxidants attenuate neuronal cell death following uptake of oxidized low-density lipoprotein. Free Radic Biol Med 29, 1222–1233.

40.

Kwon

S-H

, Lee

H-K

, Kim

J-A

, Hong

S-I

, Kim

H-C

, Jo

T-H

, Park

Y-I

, Lee

C-K

, Kim

Y-B

, Lee

S-Y

, Jang

C-G

(2010) Neuroprotective effects of chlorogenic acid on scopolamine-induced amnesia via anti-acetylcholinesterase and anti-oxidative activities in mice. Eur J Pharmacol 649, 210–217.

41.

Mikami

, Yamazawa

(2015) Chlorogenic acid, a polyphenol in coffee, protects neurons against glutamate neurotoxicity. Life Sci 139, 69–74.

42.

Cooke

, Bear

(2015) Visual recognition memory: A view from V1. Curr Opin Neurobiol 35, 57–65.

43.

López-Aranda

, López-Téllez

, Navarro-Lobato

, Masmudi-Martín

, Gutiérrez

, Khan

(2009) Role of layer 6 of V2 visual cortex in object-recognition memory. Science 325, 87–89.

44.

Weible

, Rowland

, Pang

, Kentros

(2009) Neural correlates of novel object and novel location recognition behavior in the mouse anterior cingulate cortex. J Neurophysiol 102, 2055–2068.

45.

Wang

, Furlong

, Stratton

, Lee

CCY

, Xu

, Merlin

, Nolan

, Arabzadeh

, Marek

, Sah

(2021) Hippocampus–prefrontal coupling regulates recognition memory for novelty discrimination. J Neurosci 41, 9617–9632.

46.

Bear

(1996) A synaptic basis for memory storage in the cerebral cortex. Proc Natl Acad Sci U S A 93, 13453–13459.

47.

Sun

, Jin

, Lin

, Chen

, Qiao

, Jiang

, Zhou

, Johnston

, Golshani

, Nie

, Holmes

, Nitz

, Xu

(2019) CA1-projecting subiculum neurons facilitate object-place learning. Nat Neurosci 22, 1857–1870.

48.

Roy

, Kitamura

, Okuyama

, Ogawa

, Sun

, Obata

, Yoshiki

, Tonegawa

(2017) Distinct neural circuits for the formation and retrieval of episodic memories. Cell 170, 1000–1012.e19.

49.

Heneka

, Carson

, Khoury

, Landreth

, Brosseron

, Feinstein

, Jacobs

, Wyss-Coray

, Vitorica

, Ransohoff

, Herrup

, Frautschy

, Finsen

, Brown

, Verkhratsky

, Yamanaka

, Koistinaho

, Latz

, Halle

, Petzold

, Town

, Morgan

, Shinohara

, Perry

, Holmes

, Bazan

, Brooks

, Hunot

, Joseph

, Deigendesch

, Garaschuk

, Boddeke

, Dinarello

, Breitner

, Cole

, Golenbock

, Kummer

(2015) Neuroinflammation in Alzheimer’s disease. Lancet Neurol 14, 388–405.

50.

Deshmukh

, Kaundal

, Bansal

, Samardeep (2016) Caffeic acid attenuates oxidative stress, learning and memory deficit in intra-cerebroventricular streptozotocin induced experimental dementia in rats. Biomed Pharmacother 81, 56–62.

51.

Grabska-Kobylecka

, Kaczmarek-Bak

, Figlus

, Prymont-Przyminska

, Zwolinska

, Sarniak

, Wlodarczyk

, Glabinski

, Nowak

(2020) The presence of caffeic acid in cerebrospinal fluid: Evidence that dietary polyphenols can cross the blood-brain barrier in humans. Nutrients 12, 1531.

Caffeoylquinic Acid Mitigates Neuronal Loss and Cognitive Decline in 5XFAD Mice Without Reducing the Amyloid-β Plaque Burden

Abstract

Background:

Objective:

Methods:

Results:

Conclusions:

Keywords

Introduction

MATERIALS AND METHODS

Animals

Diet

General behavioral analysis

NOR test

Y-maze test

Immunofluorescence

Aβ42 quantification by ELISA

RNA extraction

Ion AmpliSeq

qRT-PCR

Statistical analyses

Results

CQA treatment ameliorates cognitive memory decline in 5XFAD mice

CQA treatment does not affect the Aβ plaque burden or glial activation in 5XFAD mice

CQA treatment does not affect Aβ peptide levels in 5XFAD mice

CQA treatment reduces neuronal loss and increases PSD95 intensity in 5XFAD mice

Transcriptome analyses and qRT-PCR revealed that CQA treatment leads to increased levels of mRNA of genes related to synaptic function in 5XFAD mice

Discussion

AUTHOR CONTRIBUTIONS

Footnotes

ACKNOWLEDGMENTS

FUNDING

CONFLICT OF INTEREST

DATA AVAILABILITY

References

Aβ₄₂ quantification by ELISA