Increased Expression of C/EBP Homologous Protein, a Marker of Endoplasmic Reticulum Stress, in the Brains of App^{NL-G-F/NL-G-F} Knock-in Alzheimer’s Disease Model Mice

Abstract

Alzheimer’s disease (AD) is one of the most common types of progressive dementia. Recently, endoplasmic reticulum (ER) stress was suggested as a potential event involved in AD development. Thus, targeting ER stress may be an effective AD treatment. The involvement of ER stress in the brains of amyloid precursor protein knock-in AD model mice (App^{NL-G-F/NL-G-F}) with Swedish/Iberian/Arctic mutations found in human familial AD remains unclear. This study aimed to clarify whether the expression of ER stress marker C/EBP homologous protein (CHOP) was enhanced in the brains of App^{NL-G-F/NL-G-F} AD model mice. Our immunofluorescence staining results showed that similar to the expression pattern of amyloid-β (Aβ), CHOP demonstrated an age-dependent increase in the numbers and sizes of spotted signals in the cerebral cortex and hippocampus of these 4- to 10-month-old AD model mice but not in their age-matched controls. These findings suggested that CHOP expression was upregulated in close association with Aβ expression, and that CHOP was involved in neuropathy caused by Aβ accumulation. Future investigations of the localization and variations in expression levels of other ER stress-related proteins in this mouse model using immunofluorescence staining will lead to a more detailed estimation of the relationship between ER stress and AD pathogenesis.

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disease that severely affects cognitive abilities, leading to memory loss. It is the prevalent cause of dementia in late adult life and is associated with high morbidity and mortality rates in older adults.¹⁾

The neuropathological hallmarks of AD are amyloid-β (Aβ) plaques and neurofibrillary tangles formed by the intracellular accumulation of hyperphosphorylated tau protein.²⁾ To date, only a few approved drugs have been used for the clinical treatment of AD, including cholinesterase inhibitors and NMDA receptor antagonists, which only relieve the symptoms or slow down disease progression and do not provide a cure.³⁾ Based on the amyloid cascade hypothesis,⁴⁾ the clearance of Aβ plaques from the brain may treat AD and halt disease progression, and this has promoted the development of innovative antibody drugs for preventing Aβ aggregation in the brain, such as lecanemab.⁵⁾ However, these antibody drugs are indicated for patients with early AD⁶⁾ and are unable to cure advanced cases.³⁾ Therefore, a thorough investigation of the changes in the brain following Aβ accumulation is necessary to identify novel targets and subsequently develop effective AD treatments.

Endoplasmic reticulum (ER) stress was recently reported as a potential event in AD development, and several studies have suggested that Aβ accumulation-induced ER stress is involved in neurodegeneration in AD.^7,8) The accumulation of unfolded/misfolded proteins within the ER lumen leads to ER dysfunction (ER stress).⁹⁾ The unfolded protein response (UPR) activates and promotes the degradation of misfolded proteins and increases the ability to correct protein folding. However, under chronic or excessive ER stress, the UPR fails to maintain ER homeostasis, resulting in activation of the apoptosis signaling pathway, leading to the development of various diseases.¹⁰⁾ ER-induced apoptosis occurs via three primary pathways: IRE1/ASK1/JNK, caspase-12 kinase, and C/EBP homologous protein (CHOP)/GADD153. The CHOP pathway plays a critical role in ER stress-induced apoptosis caused by neurological diseases.^11,12) Therefore, ER stress is considered to act in a mediator-like role in AD pathology, and elucidating the relationship between ER stress and AD development could enable the identification of new therapeutic targets in AD treatment.

Mice with mutations inserted into the Aβ and tau genes, which are considered responsible for AD onset, have been used as models of memory impairment. Transgenic mouse models of AD include the Tg2576 model, in which amyloid precursor protein (APP) is overexpressed; the 5XFAD transgenic model, characterized by the overexpression of APP and presenilin 1 and the presence of mutations observed in familial AD; and the 3xTg model, with mutations leading to the overexpression of presenilin, APP, and tau. However, the AD pathology in these mouse models differs from the AD pathology in humans due to the overproduction of APP. Recently, Saito et al. generated a single APP knock-in mouse model, App^{NL-G-F/NL-G-F}.¹³⁾ In this model, the mutations observed in familial AD (Swedish, Beyreuther/Iberian, and Arctic mutations) were introduced into the endogenous mouse APP gene, leading to the development of progressive AD-related pathologies, including cognitive decline, Aβ deposition, and neuroinflammation. This model is considered to reflect the human condition without the artificial effects of APP overexpression.

The involvement of ER stress in AD development requires careful evaluation to enable the identification of effective prevention and treatment strategies. The detection of ER stress markers, including CHOP, in the brain of several different AD mouse models was the subject of a review by Hashimoto and Saido.¹⁴⁾ Although an enhanced ER stress response was observed in 5XFAD and 3xTg mouse models,^15,16) no ER stress response was observed in APP-single-Tg or APP knock-in mouse models, including App^{NL-G-F/NL-G-F}.¹⁷⁾ Because most of these reports were based on results derived from western blot analysis of total proteins extracted from the hippocampus and cerebral cortex, the target proteins might have been diluted, making it difficult to accurately detect increases or decreases in their expression. Thus, it is imperative that conclusions are not only based on the results of western blot analysis but also on the results of careful examination of the localization of Aβ and ER stress markers, such as CHOP, in the brains of AD model mice. Therefore, in this study, we used immunofluorescence staining to analyze CHOP expression in the brains of App^{NL-G-F/NL-G-F} mice. Additionally, we investigated whether CHOP expression was associated with the increased amyloid levels that have been observed with age.

MATERIALS AND METHODS

Animals

The original lines of APP knock-in (App^{NL-G-F/NL-G-F} and App^NL/NL) mice with a C57BL/6J genetic background were obtained from the RIKEN Center for Brain Science (Wako, Japan) and maintained by brother–sister mating at the Institute for Animal Experimentation, Toho University. In addition to the humanized mouse Aβ sequence, the Swedish familial AD mutation is inserted in App^NL/NL knock-in mice and a triple mutation (Swedish, Beyreuther/Iberian, and Arctic familial AD mutations) is inserted in App^{NL-G-F/NL-G-F} knock-in mice. The App^NL/NL knock-in mice were used as negative controls. Following weaning at postnatal days 28–35, all mice were housed socially in same-sex groups in a temperature-controlled environment (24 ± 2°C) under a 12-h light/dark cycle (lights on at 8:00, lights off at 20:00), with access to food and water ad libitum. The mice were tested at approximately 4–10 months of age.

All experiments were performed in accordance with the Toho University guidelines for animal care, handling, and termination, which comply with the international and Japanese guidelines for animal care and welfare (approval number: 21-31-476).

Brain Fixation

While under anesthesia with a cocktail combination (medetomidine [0.3 mg/kg] + midazolam [4 mg/kg] + butorphanol [5 mg/kg]), the mice were intracardially perfused with phosphate-buffered saline (PBS, Wako Pure Chemical Industries, Osaka, Japan), followed by 4% paraformaldehyde solution (Nacalai Tesque, Inc., Kyoto, Japan). Then, the brains were collected and postfixed in 4% paraformaldehyde solution at 4°C overnight before storage in FD Tissue Storage Solution (FD NeuroTechnologies, Inc., MD, USA) at −20°C until tissue sectioning.

Immunofluorescence Staining of Aβ and CHOP

Coronal sections of the brain (thickness, 50 μm) were prepared using a vibratome (Neo LinearSlicer MT, Dosaka EM Co., Ltd., Kyoto, Japan). Antigens were retrieved by soaking the sections in HistoVT One solution (Nacalai Tesque, Inc.) for 30 min at 70°C. After washing with PBS containing 0.3% Triton X-100, sections for Aβ staining were further incubated with 70% formic acid for 15 min. Then, all sections were blocked for 1 h at room temperature using BlockAid Blocking Solution (Thermo Fisher Scientific, Inc., MA, USA) containing 0.3% Triton X-100, followed by overnight incubation at 4°C with primary antibodies in the blocking buffer. The primary antibody used for Aβ staining was mouse anti-Aβ (specific for human Aβ1–16; 1:500; BAN50; Wako Pure Chemical Industries, Osaka, Japan), and mouse anti-DDIT3 (1:100; ab11419; Abcam plc., Cambridge, UK) and mouse anti-CHOP (1:100; L63F7; Cell Signaling Technology, Inc., MA, USA) were used to stain CHOP. DDIT3 is the gene name for CHOP and refers to the same.

After three washes with PBS, the sections were incubated for 1 h at room temperature in blocking buffer with the following secondary antibodies: Alexa Fluor 488 Goat anti-Mouse IgG1 (1:1000; A-21121; Thermo Fisher Scientific, Inc.) for anti-Aβ antibody, Alexa Fluor 488 Goat anti-Mouse IgG2b (1:1000; A-21141; Thermo Fisher Scientific, Inc.) for anti-DDIT3 antibody, and Alexa Fluor 594 Goat anti-Mouse IgG2a (1:1000; A-21135; Thermo Fisher Scientific, Inc.) for anti-CHOP antibody. After three washes with PBS, the immunostained slides were mounted with Prolong Glass Antifade Mountant with NucBlue Stain (Thermo Fisher Scientific, Inc.) and cured for 48 h. Immunofluorescence images of the sections were captured using a BZ-X700 fluorescence microscope (Keyence, Osaka, Japan). Multiple images were taken at 20× magnification and then joined together for quantification of the areas of Aβ plaques using WinROOF Education Version 2.4.0 (Mitani Corp., Tokyo, Japan).

RESULTS AND DISCUSSION

Changes in Aβ Plaque Staining in App^{NL-G-F/NL-G-F} Mice with Age

App^NL/NL mice are recombinants in which the mouse Aβ sequence is humanized and a Swedish familial AD mutation is inserted. In contrast, App^{NL-G-F/NL-G-F} mice have a humanized Aβ sequence and Swedish, Beyreuther/Iberian, and Arctic familial AD mutations, and Aβ deposition in the cortex of these model mice has been reported to begin by 2 months old and show saturation by the age of 7 months. Thus, we used App^NL/NL mice as negative controls in our study.¹³⁾

Figure 1 shows images of Aβ deposition in App^NL/NL and App^{NL-G-F/NL-G-F} brains. Figure 1A shows the immunofluorescence staining results for Aβ in the hippocampus and cerebral cortex of 8-month-old mice. Specific fluorescence derived from Aβ antibodies (arrows) was observed in App^{NL-G-F/NL-G-F} brains. Figure 1B shows the immunofluorescence staining results for Aβ at 4, 6, 8, and 10 months of age. Anti-Aβ antibody-derived fluorescence was observed in 4-month-old App^{NL-G-F/NL-G-F} mice, with particularly high expression in the cortex. Consistent with the literature¹³⁾, age-dependent progressive Aβ amyloidosis was observed in App^{NL-G-F/NL-G-F} mice, with increased deposition throughout the cortex at 8 months of age. The hippocampus region of the brain is involved in memory, and the slight accumulation of Aβ in the hippocampus was observed at 4 months of age, with increased deposition throughout the hippocampus at 10 months of age. Figure 1C shows the percentage of anti-Aβ antibody-derived fluorescent areas in the brains of App^{NL-G-F/NL-G-F} mice, including the hippocampus and cerebral cortex. Aβ aggregation markedly increased from the age of 6–8 months and continued to increase with age until the mice were 10 months old.

Fig. 1

Aβ Deposition in App^NL/NL and App^{NL-G-F/NL-G-F} Mice Brains

(A) Immunofluorescence staining images of the hippocampus and cortex areas in the brains of 8-month-old App^NL/NL and App^{NL-G-F/NL-G-F} mice using anti-Aβ antibody. The same magnification was used for each pair of photos. Scale bars represent 1,000 μm in the overall image and 200 μm in the images of the hippocampus and cortex.

(B) Immunofluorescence staining images of App^NL/NL and App^{NL-G-F/NL-G-F} mice brains at 4, 6, 8, and 10 months old using anti-Aβ antibody. All photos have the same magnification. Scale bars represent 1,000 μm.

(C) Quantification of the ratio of the Aβ-positive areas. Data are presented as means ± standard deviations for each group (n = 3). *p < 0.05 versus 4 months, #p < 0.05 versus 6 months (Tukey multiple comparison test).

CHOP Identification in App^{NL-G-F/NL-G-F} Mice Brains

We compared immunofluorescence staining images between 8-month-old App^NL/NL and App^{NL-G-F/NL-G-F} mice to determine the expression of ER stress markers in mouse brains (Figs. 2A and 2B). First, we compared the immunofluorescence staining results obtained using two commercially available antibodies against CHOP. CHOP-derived luminescence was observed in App^{NL-G-F/NL-G-F} brains but not in App^NL/NL brains. Furthermore, the staining patterns in the cerebral cortex and hippocampus were similar to those obtained using the Aβ antibody. Next, we compared the staining patterns between the two types of antibodies, revealing that CHOP antibody L63F7 was more sensitive compared with ab11419.

Fig. 2

Immunofluorescence Staining with Two Different Antibodies Against CHOP

(A) Immunofluorescence staining images of the hippocampus and cortex areas of 8-month-old App^NL/NL and App^{NL-G-F/NL-G-F} mice using anti-DDIT3 antibody Ab11419. The same magnification was used for each pair of photos. Scale bars represent 1,000 μm in the overall image and 200 μm in the images of the hippocampus and cortex.

(B) Immunofluorescence staining images of the hippocampus and cortex areas of 8-month-old App^NL/NL and App^{NL-G-F/NL-G-F} mice using anti-CHOP antibody L63F7. The same magnification was used for each pair of photos. Scale bars represent 1,000 μm in the overall images and 200 μm in the images of the hippocampus and cortex.

The CHOP antibody L63F7 used in this study was the same as that used in previous studies¹⁷⁾ examining CHOP expression by western blot analysis in App^{NL-G-F/NL-G-F} brains. Although these studies did not show increased CHOP expression in the brain total proteins of App^{NL-G-F/NL-G-F} knock-in AD model mice,¹⁷⁾ our study, which used an immunofluorescence staining method, showed a clear increase in CHOP expression in App^{NL-G-F/NL-G-F} knock-in AD model mice. As mentioned in the review by Hashimoto and Saido, the relationship between AD pathology and ER stress has been investigated using various AD mouse models, but no unified conclusion has been reached.¹⁴⁾ In our study, immunofluorescence staining revealed that CHOP expression patterns in the brains of App^{NL-G-F/NL-G-F} mice were similar to those of Aβ plaques.

Changes in CHOP Expression in App^{NL-G-F/NL-G-F} Mice with Age

We observed changes in CHOP expression with age using the L63F7 antibody. Figure 3 shows the immunofluorescence staining results for CHOP in brains of 4- to 10-month-old App^{NL-G-F/NL-G-F} mice. In 4-month-old mice brains, CHOP expression was mainly observed in the cerebral cortex. Additionally, CHOP expression in the cerebral cortex increased until 8 months of age. In the hippocampus, clear CHOP expression was observed in 6-month-old mice brains, with a marked increase in expression until 8 months of age. Similar to the increase in Aβ that was observed with age, CHOP expression in App^{NL-G-F/NL-G-F} mice brains was also increased with age. Collectively, our immunofluorescence staining results suggested that CHOP expression was associated with an increase in age-related Aβ expression in the brains of App^{NL-G-F/NL-G-F} knock-in AD model mice.

Fig. 3

Immunofluorescence Staining of CHOP in App^NL/NL and App^{NL-G-F/NL-G-F} Mice Brains

Brain sections of 4- to 10-month-old mice were immunostained using mouse anti-CHOP antibody L63F7. All photos have the same magnification. Scale bars represent 1,000 μm.

Various in vivo and in vitro studies have suggested that targeting ER stress and ER stress-mediated apoptosis contributes to AD recovery.^18,19) Nevertheless, owing to species differences between humans and mice, existing AD mouse models, such as 3xTg or 5XFAD, are unable to fully simulate the pathologic and clinical features of human AD. The App^{NL-G-F/NL-G-F} knock-in AD mouse model was developed to more closely resemble human pathology. In this study, CHOP expression in App^{NL-G-F/NL-G-F} brains was confirmed using immunostaining, supporting the possibility that ER stress plays an important role in AD development due to Aβ accumulation. Besides CHOP, numerous other factors (ER stress marker proteins) are associated with ER stress, which form a series of cascades, and the detection of these marker proteins in brain sections should be undertaken in future studies. Our study findings strongly suggest that evaluation of the localization and expression variation of a variety of ER stress markers, including CHOP, in brain tissue sections, rather than analysis using brain total proteins, will enable clarification of the true role of ER stress in AD pathogenesis. Thus, the App^{NL-G-F/NL-G-F} knock-in AD mouse model is useful for assessing the significance of ER stress in human AD pathogenesis, therapy, and prevention.

Acknowledgments

This work was partially supported by JSPS KAKENHI Grant Numbers JP16H05096 (KN) and JP23K16806 (HT).

Conflict of interest

The authors declare no conflict of interest.

REFERENCES

• 1)    Cass  SP. Alzheimer’s disease and exercise: a literature review. Curr. Sports Med. Rep., 16, 19–22 (2017).
• 2)    Shamsi  TN,  Athar  T,  Parveen  R,  Fatima  S. A review on protein misfolding, aggregation and strategies to prevent related ailments. Int. J. Biol. Macromol., 105, 993–1000 (2017).
• 3)    Morris  G,  Puri  BK,  Walder  K,  Berk  M,  Stubbs  B,  Maes  M,  Carvalho  AF. The Endoplasmic Reticulum Stress Response in Neuroprogressive Diseases: emerging Pathophysiological Role and Translational Implications. Mol. Neurobiol., 55, 8765–8787 (2018).
• 4)    Hardy  JA,  Higgins  GA. Alzheimer’s disease: the amyloid cascade hypothesis. Science, 256, 184–185 (1992).
• 5)    Logovinsky  V,  Satlin  A,  Lai  R,  Swanson  C,  Kaplow  J,  Osswald  G,  Basun  H,  Lannfelt  L. Safety and tolerability of BAN2401--a clinical study in Alzheimer’s disease with a protofibril selective Aβ antibody. Alzheimers Res. Ther., 8, (2016).
• 6)    van Dyck  CH,  Swanson  CJ,  Aisen  P,  Bateman  RJ,  Chen  C,  Gee  M,  Kanekiyo  M,  Li  D,  Reyderman  L,  Cohen  S,  Froelich  L,  Katayama  S,  Sabbagh  M,  Vellas  B,  Watson  D,  Dhadda  S,  Irizarry  M,  Kramer  LD,  Iwatsubo  T. Lecanemab in Early Alzheimer’s Disease. N. Engl. J. Med., 388, 9–21 (2023).
• 7)    Costa  RO,  Ferreiro  E,  Oliveira  CR,  Pereira  CM. Pereira Inhibition of mitochondrial cytochrome c oxidase potentiates Aβ-induced ER stress and cell death in cortical neurons. Mol. Cell. Neurosci., 52, 1–8 (2013).
• 8)    Scheper  W,  Hoozemans  JJ. The unfolded protein response in neurodegenerative diseases: a neuropathological perspective. Acta Neuropathol., 130, 315–331 (2015).
• 9)    Liu  MQ,  Chen  Z,  Chen  LX. Endoplasmic reticulum stress: a novel mechanism and therapeutic target for cardiovascular diseases. Acta Pharmacol. Sin., 37, 425–443 (2016).
• 10)    Mattson  MP,  Guo  Q,  Furukawa  K,  Pedersen  WA. Presenilins, the endoplasmic reticulum, and neuronal apoptosis in Alzheimer’s disease. J. Neurochem., 70, 1–14 (1998).
• 11)    Oyadomari  S,  Mori  M. Roles of CHOP/GADD153 in endoplasmic reticulum stress. Cell Death Differ., 11, 381–389 (2004).
• 12)    Hetz  C. The unfolded protein response: controlling cell fate decisions under ER stress and beyond. Nat. Rev. Mol. Cell Biol., 13, 89–102 (2012).
• 13)    Saito  T,  Matsuba  Y,  Mihira  N,  Takano  J,  Nilsson  P,  Itohara  S,  Iwata  N,  Saido  TC. Single App knock-in mouse models of Alzheimer’s disease. Nat. Neurosci., 17, 661–663 (2014).
• 14)    Hashimoto  S,  Saido  TC. Critical review: involvement of endoplasmic reticulum stress in the aetiology of Alzheimer’s disease. Open Biol., 8, 180024 (2018).
• 15)    O’Connor  T,  Sadleir  KR,  Maus  E,  Velliquette  RA,  Zhao  J,  Cole  SL,  Eimer  WA,  Hitt  B,  Bembinster  LA,  Lammich  S,  Lichtenthaler  SF,  Hébert  SS,  De Strooper  B,  Haass  C,  Bennett  DA,  Vassar  R. Phosphorylation of the translation initiation factor eIF2alpha increases BACE1 levels and promotes amyloidogenesis. Neuron, 60, 988–1009 (2008).
• 16)    Soejima  N,  Ohyagi  Y,  Nakamura  N,  Himeno  E,  Iinuma  KM,  Sakae  N,  Yamasaki  R,  Tabira  T,  Murakami  K,  Irie  K,  Kinoshita  N,  LaFerla  FM,  Kiyohara  Y,  Iwaki  T,  Kira  J. Intracellular accumulation of toxic turn amyloid-β is associated with endoplasmic reticulum stress in Alzheimer’s disease. Curr. Alzheimer Res., 10, 11–20 (2013).
• 17)    Hashimoto  S,  Ishii  A,  Kamano  N,  Watamura  N,  Saito  T,  Ohshima  T,  Yokosuka  M,  Saido  TC. Endoplasmic reticulum stress responses in mouse models of Alzheimer’s disease: overexpression paradigm versus knockin paradigm. J. Biol. Chem., 293, 3118–3125 (2018).
• 18)    Oliveira  MM,  Lourenco  MV,  Longo  F,  Kasica  NP,  Yang  W,  Ureta  G,  Ferreira  DDP,  Mendonça  PHJ,  Bernales  S,  Ma  T,  De Felice  FG,  Klann  E,  Ferreira  ST. Correction of eIF2-dependent defects in brain protein synthesis, synaptic plasticity, and memory in mouse models of Alzheimer’s disease. Sci. Signal., 14, eabc5429 (2021).
• 19)    Kou  JJ,  Shi  JZ,  He  YY,  Hao  JJ,  Zhang  HY,  Luo  DM,  Song  JK,  Yan  Y,  Xie  XM,  Du  GH,  Pang  XB. Luteolin alleviates cognitive impairment in Alzheimer’s disease mouse model via inhibiting endoplasmic reticulum stress-dependent neuroinflammation. Acta Pharmacol. Sin., 43, 840–849 (2022).