The Institutional Animal Care and Use Committee of the Research Institute for Future Medicine (RIFM) at the Samsung Medical Center approved this animal study on June 29, 2021 (approval number: 20210629002). The Institute is committed to adhering to the guidelines set forth by the Institute for Laboratory Animal Resources (ILAR) and is a facility fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). All experiments were conducted in accordance with the ARRIVE guidelines 2.0. Genotyping was conducted using genomic DNA extracted from the tail snips of littermates born by mating PS19 (P301S) transgenic and B6C3 wild type mice originally procured from the Jackson Laboratory (Bar Harbor, ME, USA). This transgenic mouse model expresses a mutant human microtubule-associated protein tau. The expression of the human tau protein is approximately fivefold higher than that of the endogenous mouse tau protein [7]. Experiments were performed using transgenic mice that were 3–4 months old (n = 32). Mice were surgically implanted with ameroid constrictors (MC-0.75-TI, inner diameter: 0.75 mm, Research Instruments SW, Escondido, CA, USA) at an early age to allow for the gradual occlusion of the arteries to be observed over an extended period of time. Four separate PS19 mice and four separate B6C3 mice were used as positive and negative controls, respectively, for the tau enzyme-linked immunosorbent assay (ELISA). Mice were fed ad libitum and maintained on a 12-hour light/12-hour dark cycle.

PS19 mice were placed in the supine position and surgical procedures were performed under 2% isoflurane (1015, Hana Pharmaceuticals, Seoul, Republic of Korea) as general anesthesia. CCH was induced according to the methodology described in a previously reported procedure [10]. After removal of the fur and povidone iodine sterilization, a midline incision was made over the thyroid bone. The carotid arteries were exposed by retracting the omohyoid and sternomastoid muscles and by meticulously dissecting surrounding fascia and sheath, including the vagus nerve. A 6-0 silk thread (Ethicon, Cincinnati, OH, USA) was used to wrap the ameroid constrictors (inner diameter: 0.75 mm; Research Instruments SW, Escondido, CA, USA) around each CCA. An opening slit on one side of the ameroid constrictor allowed it to be wrapped around the vessel. In the Sham group, the same procedure was performed without the use of the ameroid constrictors after exposure of both CCAs. After surgery, the midline incision was closed with a 5-0 silk suture (Ethicon, Cincinnati, OH, USA). Prior to inducing CCH for 6 months, a preliminary study was performed to determine the optimal duration of CCH induction in PS19 mice. PS19 mice were sacrificed (Sham: n = 3, CCH: n = 5) at either 1 or 3 months (Fig. 1). PS19 mice were subjected to CCH for up to 6 months. Following this period, the mice were deeply anesthetized with 1.5–2.0 isoflurane (1015, Hana Pharmaceuticals, Seoul, Republic of Korea) and sacrificed via transcardial perfusion (Sham: n = 6, CCH: n = 10) at 9–10 months of age (Fig. 1A).

Fig. 1.

Schematic diagram illustrating overall study design. (A) In the preliminary experiments, PS19 mice were allocated to one of two experimental groups: the Sham group or the CCH group. Mice were subsequently sacrificed either 1 or 3 months following surgery. Subsequently, PS19 mice were subject to a 6-month period of chronic cerebral hypoperfusion and subsequently sacrificed. (B) An initial experiment performed using positive (PS19) and negative (B6C3) controls (each group: n = 4) validated the experimental procedures for pT231, pT181, and total tau ELISA. Statistical significance was defined at ^*⁣**p 0.001, ^*⁣*⁣**p 0.0001, and ^**p 0.01 vs Sham at each respective time point (Welch’s t-test, unpaired, two-tailed). Values are shown as mean $\pm$ standard deviation (SD). CCH, chronic cerebral hypoperfusion; ELISA, enzyme-linked immunosorbent assay.

Peak systolic velocity (PSV; mm/s) was measured for both the Sham and CCH groups before surgery (pre) and at 1, 3, and 6 months post-surgery. PSV was measured for both the left and right CCAs. Mice from the Sham and CCH groups were anesthetized with 1.5–2.0% isoflurane (1015, Hana Pharmaceuticals, Seoul, Republic of Korea) during the acquisition of the ultrasound image. During image acquisition mice were restrained in a supine position on a pad equipped with a heater, temperature sensor, and electrocardiogram electrodes (serial number: 2100-0286, Visual Sonics Inc., Toronto, ON, Canada). Ultrasound images were acquired using a Vevo® 2100 Imaging System and an MS-550D (22–55 MHz) transducer probe (serial number: 2100-0286, Visual Sonics Inc., Toronto, ON, Canada). Image acquisition was by pulsed wave Doppler and Color Doppler modes. PSV was calculated using Vevo® 2100 1.6.0 software (serial number: 2100-0286, Visual Sonics Inc., Toronto, ON, Canada).

To assess changes in motor coordination, rotarod tests were conducted [28] prior to animal sacrifice. Over the course of 3 consecutive days, three trials were conducted each day in which mice were placed on a revolving cylindrical rod. Prior to each trial, the mice underwent a 1-minute training session on the rotating rod at a speed of 4 revolutions per minute (RPM). In the actual trial, the speed of the revolving rod was increased from 4 to 40 rpm, and the trial lasted for a total of 4 minutes. The latency to fall was recorded during the aforementioned time period. Additionally, the mice were permitted a 10-minute recovery interval between each trial.

To validate the ELISA procedures, ELISA (pT181, KHO0631; pT231, KHB8051, Invitrogen, Carlsbad, CA, USA) was first performed to quantify pTau (pT181, pT231) and total tau protein levels using homogenized brain samples from PS19 (positive control; n = 4) and B6C3 (negative control, n = 4) mice (Fig. 1B).

After a period of 1, 3, or 6 months of inducing CCH, PS19 mice were sacrificed and samples of brain tissue were harvested. On the day of sacrifice mice were euthanized by cardiac perfusion (only PBS) under deep anesthesia (2% isoflurane, 1015, Hana Pharmaceuticals, Seoul, Republic of Korea). The hippocampus was excised from the cortex of the left hemisphere of the Sham and CCH groups and the resulting tissue samples were stored separately in a freezer (TFF900S-UULTF, Thermo Fisher Scientific, Waltham, MA, USA) at –80 °C. For the samples harvested after 1 and 3 months, pTau (pT181, pT231) and total tau protein levels were quantified via ELISA. For the samples harvested at post 6 months, half of the Sham (n = 3) and CCH (n = 5) samples (both hippocampus and cortex) were used for quantification of pTau (pT181, pT231) and total tau protein levels via ELISA. Phospho-tau 217 (pT217, 59672, Cell Signaling Technology; Danvers, MA, USA) protein levels were also quantified via ELISA. The manufacturer’s instructions were followed in the use of each ELISA kit, with each sample tested in duplicate wells. For each of the tau ELISAs, each sample was measured three times and results were averaged.

Whole-transcriptome sequencing was conducted on the remaining samples (Sham: n = 3, CCH: n = 5). The total RNA concentration was extracted and quantified using the Quant-IT RiboGreen assay. To evaluate the quality of total RNA, samples were analyzed using a TapeStation RNA screen tape (5067-5576, Agilent Technologies, Santa Clara, CA, USA). Only high-quality RNA (RNA integrity number greater than 7) was used for library construction. An independent preparation of the library was conducted for each sample using the Illumina TrueSeq Stranded mRNA Sample Prep Kit (20020595, Illumina, San Diego, CA, USA). The poly-A-containing mRNA molecules were purified using poly-T-attached magnetic beads (61002, Thermo Fisher Scientific, Waltham, MA, USA), followed by fragmentation and complementary DNA (cDNA) synthesis using SuperScript II reverse transcriptase (18064014, Thermo Fisher Scientific, Waltham, MA, USA) and random primers. Subsequently, the products underwent further purification and enrichment through PCR and the libraries were quantified using the KAPA library quantification kit (07960140001, Roche, Pleasanton, CA, USA) for Illumina sequencing platforms. Alignment of the whole-transcriptome sequencing reads to the reference human genome (GRCh37, hg19) was performed using the STAR 2.7.11b alignment tool (https://github.com/alexdobin/STAR). Gene expression levels were quantified and determined based on fragments per kilobase of transcript per million mapped reads. These values were subsequently log2-transformed and used in differentially expressed gene (DEG) analysis using the DEseq2 1.46.0 software (http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html). Gene set enrichment analysis was performed to acquire pathway-level activities. The mRNA expression levels of genes associated with AD were evaluated and the gene expression of immune cell types was also identified via immune deconvolution analysis using CIBERSORTx software (http://cibersortx.stanford.edu).

To validate mRNA sequencing results, the gene expressions of Glutathione s-transferase pi 1 (Gstp1) and Fibroblast growth factor receptor 1 (Fgfr1) were validated by carrying out real-time quantitative polymerase chain reaction (RT-qPCR). With the remaining RNA samples following mRNA sequencing, cDNA was synthesized using the AccuPower® RocketScript™ Cycle RT PreMix kit (K-2241, Bioneer, Daejeon, Republic of Korea). RT-qPCR was subsequently performed by mixing the respective cDNA samples (1 µg per sample) with the SYBR green master mix (4367659, Applied Biosystems, Foster City, CA, USA). All samples were subjected to a 40-cycle amplification at 95 °C for initial holding (10 min), 95 °C for denaturation (15 sec), and 57 °C for annealing (1 min).

All cDNA samples were assayed in triplicate. Results were normalized against Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) levels and quantified using the 2^{-Δ⁢Δ⁢Ct} method proposed by Livak and Schmittgen [29]. The primer sequences were: Fgfr1, forward, GCCTCACATTCAGTGGCTGAAG; reverse, AGCACCTCCATTTCCTTGTCGG (Origene, Rockville, MD, USA), Gstp1, forward, TGGAAGGAGGAGGTGGTTACCA; reverse, GGTAAAGGGTGAGGTCTCCATC (Origene, Rockville, MD, USA), and Gapdh, forward, CAGTATGACTCCACTCACGG; reverse, GTGAAGACACCAGTAGACTCC.

The right hemispheres of the Sham and CCH groups were fixed in 4% paraformaldehyde (Biosesang, Seongnam, Republic of Korea) then paraffin blocked. Serial sections measuring 4 µm were made for immunohistochemical (IHC) staining. Previously reported methods were used to deparaffinize the slides and perform heat antigen retrieval [30]. The phospho-tau (pT231) primary antibody (1:400, MN1040, ThermoFisher Scientific, Waltham, MA, USA) was used for IHC and visualized using 3,3^′-diaminobenzidine (DAB, K3468, Dako, Santa Clara, CA, USA). CD68 macrophage expression (1:200, ab125212, Abcam, Cambridge, UK) was observed by immunofluorescence using Alexa Fluor 594 (A-11012, Thermo Fisher Scientific, Waltham, MA, USA). Upon completion of IHC, a Scanscope AT scanner (Leica Biosystems, Wetzlar, Germany) was employed to scan the slides, then pTau expression in the cortex, hippocampus, and thalamus of the Sham and CCH groups was quantified using the Vectra® Automated Imaging System (PerkinElmer Applied Biosystems, Waltham, MA, USA). Two adjacent stained and imaged paraffin slides were used. Phenochart (Vectra version 2.6.0) software (PerkinElmer Applied Biosystems, Waltham, MA, USA) was used to delineate rectangular regions of interest (ROIs) for each brain region. Subsequently, the DAB optical density threshold was calibrated to quantify pTau expression for the aforementioned ROIs.

To evaluate the histologic changes caused by CCH, a hematoxylin and eosin (H&E) stain was first performed [10]. Luxol fast blue (IW-3005, IHC World, Woodstock, MD, USA) was used for myelin staining and to specifically determine white matter changes in the corpus callosum following CCH. The Scanscope AT scanner was used for the scanning of the H&E and Luxol fast blue-stained slides. The corpus callosum area for both the Sham and CCH groups was assessed using the Otsu threshold algorithm [31] in the ImageJ analysis software (https://imagej.nih.gov/ij/, National Institutes of Health (NIH), Bethesda, MD, USA).

GraphPad Prism 8.0 software (GraphPad Software, Inc., San Diego, CA, USA) was used for the statistical analysis. Values are given as mean $\pm$ standard deviation (SD). To assess significance, either a Welch’s t-test or a Mann-Whitney U test was employed and a p-value $\le$ 0.05 was considered to be statistically significant.

A preliminary study was conducted to ascertain the optimal time point for induction of CCH in PS19 mice. Sham and CCH mice were sacrificed at either 1 or 3 months (Fig. 1A). Prior to assessing the 1- and 3-month samples from the Sham and CCH groups, the positive and negative control samples were examined using ELISA. Statistical analysis revealed significant differences in pT181 (p 0.001), pT231 (p 0.0001), and total tau (p 0. 01) levels between the positive (PS19) and negative (B6C3) controls (pT181: p = 0.00036, pT231: p = 0.000019, Total Tau: p = 0.0027), as shown in Fig. 1B. These findings validated the ELISA procedures employed in this study. There was no discernible change in the levels of pTau and total tau between the Sham and CCH groups at either the 1 or 3-month time points for both the hippocampal and cortical lysates. Similarly at both 1 and 3 months following CCH induction, no significant increase in either pTau (pT181, pT231) (Fig. 2A,B) or total tau (Fig. 2C) was observed in the CCH group (both cortex and hippocampus) when compared with that of the Sham group.

Fig. 2.

Pathological alterations of the tau protein in PS19 mice induced with chronic cerebral hypoperfusion for up to 1 or 3 months. PS19 mice were sacrificed 6 months after sham or CCH surgery. Phosphorylated tau (A) pT181, (B) pT231, and (C) total tau levels were determined via ELISA using brain homogenates from the right cortex and hippocampus of the Sham (n = 3) and CCH (n = 5) groups at 1 and 3 months (Mann-Whitney U test). Values are shown as a percentage of Sham, mean $\pm$ SD.

A comprehensive analysis of the mice in the CCH group 3 months after CCH induction revealed that two of the five mice examined exhibited histopathological alterations such as loss of hippocampal neurons and early signs of infarction (foamy macrophages), with alterations in hippocampal structure after H&E staining noted in one mouse. The presence of lipid-laden macrophages or foamy macrophages was identified in the corpus callosum and cortex of the CCH-induced PS19 mouse. To ascertain the presence of macrophages, IHC was employed. Based on IHC staining, CD68-positive cells were observed to be strongly visible at the sites where foamy macrophages were identified by H&E staining (Supplementary Fig. 1).

The outcomes observed following induction at 1 and 3 months indicate a gradual accumulation of tau. To ascertain the chronic effects of hypoperfusion on tau phosphorylation, it was necessary to extend the duration of CCH induction for a period exceeding 3 months. Consequently, the final endpoint for the remaining set of experiments was set to occur 6 months post-induction.

For the Sham group, eight mice were observed over the course of the study (6 months). Unexpectedly, two mice died before the final endpoint of the study was reached. A total of five (out of 15) CCH mice died within 3 months of the surgical procedure. In comparison with the Sham group, a consistent decline in peak systolic velocity (mm/s) was observed in the CCH group across all time points (1, 3, and 6 months post-surgery) in both the left (1 month: p = 0.0001, 3 months: p = 0.000046, 6 months: p = 0.00016) and right (1 month: p = 0.000099, 3 months: p = 0.000046, 6 months: p = 0.00016) CCAs (Fig. 3A,B).

Fig. 3.

Induction of chronic cerebral hypoperfusion up to 6 months. (A) At various time points (pre-, 1, 3, and 6 months) following surgery, the peak systolic velocity (PSV; mm/s) of both the left and right CCAs of both the Sham and CCH groups was examined. A statistically significant and consistent decline in PSV was observed over time for both bilateral PSVs. Statistical significance is defined as ^*⁣*⁣**p 0.0001 and ^*⁣**p 0.001 vs Sham at each respective time point (Mann-Whitney U test). Values are shown as mean $\pm$ SD. (B) The motor function of the Sham and CCH groups was assessed by placing the mice on a rotating rod. The latency to fall was measured on three separate days, with no significant difference observed between the two groups (Mann-Whitney U test). CCA, common carotid artery.

The rotarod test was repeated to examine changes in motor function. The latency to falls from the revolving rod was measured for each experimental animal. For all three test days, the CCH group demonstrated a tendency to remain on the rotarod for longer durations than the Sham group, although the difference was not statistically significant (Fig. 3B).

Using the separate cortical and hippocampal lysates of both the Sham and CCH groups, the pTau (pT181, pT231, pT217) and total tau protein levels were quantitated using ELISA. A non-significant increasing trend in pT181 (Fig. 4A) and pT231 (Fig. 4B) levels was observed for both the cortical (pT181, pT231) and hippocampal (pT181, pT231) lysates of the CCH group in comparison with those of the Sham group. No significant difference in total tau levels was observed for the cortical or hippocampal lysates of the CCH group (Fig. 4C), although there was a non-significant trend toward elevated levels of pT217 in the cortex and hippocampus of the CCH group compared with the Sham group (Fig. 4D).

Fig. 4.

Quantitative analysis of phosphorylated tau and total tau following induction of chronic cerebral hypoperfusion up to 6 months. PS19 mice were sacrificed 6 months after Sham or CCH surgery. ELISA was used to measure phosphorylated and total tau: (A) pT181, (B) pT231, (C) total tau, and (D) pT217, in the right hemisphere (cortex and hippocampus separately) brain homogenate from the Sham (n = 3) and CCH (n = 5) groups. All four tau ELISAs showed no significant differences between the groups (Mann-Whitney U test). Values are shown as a percentage of Sham, mean $\pm$ SD.

Following ELISA, accumulation of pTau (pT231) was histologically evaluated across various regions in the PS19 mouse parenchyma (cortex, hippocampus, and thalamus). Significant elevation of pTau accumulation was not seen in the CCH group across all three brain regions in comparison with the Sham group, although there was a discernible tendency towards moderately elevated cortical pT231 levels, with a slight non-significant attenuation in the thalamus of the CCH group relative to the Sham group (Fig. 5). The progression of the disease was evaluated based on changes in either the white matter tract or the corpus callosum, which were assessed using Luxol fast blue. However, no significant differences were observed between the Sham and CCH groups in the corpus callosum (Supplementary Fig. 2).

Fig. 5.

Detection of alterations in phosphorylated tau in the cortex, hippocampus, and thalamus. Representative 3,3-diaminobenzidine immunohistochemical stained images showing the expression of phosphorylated tau (pT231, dark brown) for both the Sham and CCH groups. Nuclei were counterstained with hematoxylin. pT231 expression was quantified for both groups in the cortex, hippocampus, and thalamus (% of Sham, Mann-Whitney U test). Scale bar: right hemisphere: 2 mm, hippocampus: 400 µm. Values are shown as a percentage of Sham, mean $\pm$ SD.

To identify genes that are uniquely enriched in the CCH and Sham groups based on anatomical locations, genome-wide DEG analysis was initially performed. Notably, functional analysis using the Gene Ontology database revealed that the Sham group exhibited an upregulation of genes related to the immune system and synaptic activity, whereas the CCH group displayed an upregulation of genes related to cell cycle and DNA repair mechanisms (Fig. 6A). Moreover, the Fgfr1 pathway was collectively upregulated in the CCH group, while the immune system was upregulated in the Sham group (Fig. 6B). Both the cortex and hippocampus of the CCH group exhibited higher Gstp1 mRNA expression levels compared with the matched Sham group for both brain regions (Fig. 6C). The immune deconvolution analysis of the hippocampal samples indicated a statistically significant reduction (p = 0.029; solid red arrow) in the abundance of B memory cells in the CCH group in comparison with the Sham group (Fig. 6D). To validate the mRNA sequencing results, RT-qPCR was conducted using the respectively remaining RNA samples. A noticeable, but non-significant, increase in Fgfr1 levels was apparent in both the cortex and hippocampus of the CCH group when compared with the Sham group (Fig. 6E). No statistically significant difference was observed in Gstp1 expression between the two groups (Fig. 6F).

Fig. 6.

Whole-transcriptome sequencing analysis of hippocampal and cortical lysates and validation of results using RT-qPCR. At the 6-month sacrifice point, the Sham and CCH group brain tissues were harvested and the cortex and hippocampus (Hippo) from the left hemisphere were homogenized for transcriptome sequencing analysis. (A) DEG and Gene Ontology analyses. The colors represent genes or pathways significantly enriched in the corresponding groups based on the experimental models. Blue: CCH-Hippo; Red: CCH-Cortex & Hippo; Green: CCH-Cortex; Orange: Sham-Hippo; Light Blue: Sham-Cortex & Hippo; Purple: Sham-Cortex. (B) DEG and single sample gene set enrichment analyses. (C) Gstp1 mRNA expression. The box plot encompasses the first and third quartiles, with the median indicated by a line extending from the box. The whiskers extend to 1.5 times the interquartile range. (D) Immune deconvolution analyses using CIBERSORTx. (E) qPCR validation of Fgfr1 mRNA expression. (F) qPCR validation of Gstp1 mRNA expression. Values are expressed as mean $\pm$ SD. Statistical significance is defined as ^*p 0.05 vs Sham (Mann-Whitney U test). DEGs, differentially expressed genes; RT-qPCR, quantitative reverse transcription polymerase chain reaction.