Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 11 (1), 507

YAP-dependent Necrosis Occurs in Early Stages of Alzheimer's Disease and Regulates Mouse Model Pathology

Affiliations

YAP-dependent Necrosis Occurs in Early Stages of Alzheimer's Disease and Regulates Mouse Model Pathology

Hikari Tanaka et al. Nat Commun.

Abstract

The timing and characteristics of neuronal death in Alzheimer's disease (AD) remain largely unknown. Here we examine AD mouse models with an original marker, myristoylated alanine-rich C-kinase substrate phosphorylated at serine 46 (pSer46-MARCKS), and reveal an increase of neuronal necrosis during pre-symptomatic phase and a subsequent decrease during symptomatic phase. Postmortem brains of mild cognitive impairment (MCI) rather than symptomatic AD patients reveal a remarkable increase of necrosis. In vivo imaging reveals instability of endoplasmic reticulum (ER) in mouse AD models and genome-edited human AD iPS cell-derived neurons. The level of nuclear Yes-associated protein (YAP) is remarkably decreased in such neurons under AD pathology due to the sequestration into cytoplasmic amyloid beta (Aβ) aggregates, supporting the feature of YAP-dependent necrosis. Suppression of early-stage neuronal death by AAV-YAPdeltaC reduces the later-stage extracellular Aβ burden and cognitive impairment, suggesting that preclinical/prodromal YAP-dependent neuronal necrosis represents a target for AD therapeutics.

Conflict of interest statement

Shingo Yamada is an employee of Shino-Test Corporation. Hitoshi Okazawa and all other authors have nothing to report.

Figures

Fig. 1
Fig. 1. HMGB1 levels are elevated in the CSF of MCI and AD patients.
a CSF-HMGB1 levels in the normal control (nc) (N = 19 persons), disease control (dc) (N = 11 persons), MCI (N = 21 persons), and AD (N = 56 persons) groups were evaluated by high-sensitivity ELISA. The box plot shows the median and quartiles. Statistical differences among groups were evaluated using the Wilcoxon rank–sum test with post-hoc Bonferroni correction. b Receiver operating characteristic (ROC) curves for the MCI or AD group versus the normal control (nc) and disease control (dc) groups. Area under the ROC curve (AUC) values are shown in the graphs. Source data are provided as a “Source Data file”.
Fig. 2
Fig. 2. Necrosis occurs most actively at the preclinical stage in mouse and human AD brains.
a Morphological definition of active necrosis, secondary necrosis and ghost of cell death. Active necrosis is specified by a single degraded nucleus detected by DAPI surrounded by pSer46-MARCKS-positive degenerative neurites. Secondary necrosis is a cluster of multiple dying cells with residual Aβ in extracellular space and surrounding pSer46-MARCKS stains. Ghost of cell death is an extension of secondary necrosis in which DAPI and pSer46-MARCKS stains have faded out. b Upper graphs show time course of active necrosis, secondary necrosis and ghost of cell death in 5xFAD mice, in retrosplenial dysgranular cortex. Representative images from each time point are shown below the graph. Yellow arrow indicates a single degraded nucleus surrounded by reactive pSer46-MARCKS stains. N = 3 mice, n = 30 fields. c Time course of active necrosis in human mutant APP knock-in mice. Time course of necrosis (N = 3 mice, n = 30 fields) and representative images are shown. d Representative images of active necrosis in human MCI (MCI) and non-neurological disease control (NC). Rupturing or deformed nucleus undergoing necrosis is surrounded by Aβ and pSer46-MARCKS-positive degenerative neurites (white arrow). e The box plot shows the number of active necrosis per visual field in the median, quartiles and whiskers that represent 1.5× the interquartile range. **p < 0.01, Tukey’s HSD test (n = 10 images/person). f Simulation of active necrosis. A formula was generated by assuming that cell death occurs at a constant rate in the residual neurons and in regular time interval (top). Modulation of each parameter changed simulation curves (graphs). g Numerical simulation program generated the optimized curve (red line) based on observed values of active necrosis in occipital cortex of 5xFAD mice (black line) and predicted parameter values and active necrosis at an unmeasured time point (2 months). h The number of active necrosis observed afterwards with samples at 2 months (60 days) matched exactly with the predicted number. Values in each group are summarized by mean ± S.E.M. Source data are provided as a “Source Data file”.
Fig. 3
Fig. 3. Extreme instability of ER in AD model mice revealed by in vivo ER imaging.
a In vivo ER and Aβ images were acquired by two-photon microscopy from 1-month-old 5xFAD mice into which ER-tracker and BTA1 had been injected in one shot 4 h before observation. ER and Aβ image sets were taken in tandem every 20 min. 3D images of ER and Aβ stains were merged by IMARIS (Bitplane, Zurich, Switzerland). Dot-line indicates a single neuron. b Total volumes of ER puncta belonging to a single cell were quantified by IMARIS (Bitplane, Zurich, Switzerland), and time courses are shown in the graph. Changes were more pronounced in 5xFAD mice than in non-transgenic sibling mice (Non-Tg sibling). N = 3 mice, n = 9 cells. c To verify the finding in b, standard deviation (SD) and quartile deviation of ER volumes from a single cell at multiple time points were compared between groups of 5xFAD and non-transgenic sibling mice. Box plots show the median, quartiles and whiskers that represent 1.5× the interquartile range. P-values were determined by Welch’s test, **p < 0.01 (N = 3 mice, n = 9 cells). Source data are provided as a “Source Data file”.
Fig. 4
Fig. 4. ER enlargement of neurons in 5xFAD mice and human MCI/AD patients.
a Electron microscopy of neurons in control (B6/SJL) and 5xFAD mice. Marked ER enlargement was observed in 5xFAD mice at 5 months of age before the onset. b Electron microscopy of neurons in non-neurological disease control, MCI (Braak stage III by Gallyas-Braak staining) and AD (Braak stage V) patients. Higher magnification in two subcellular fields reveal ribosomes attached to vacuoles (arrow). Amyloid plaque in AD patient is shown (asterisk). Quantitative analysis of three groups is shown in the graph. The bar graph indicates average and mean ± S.E.M., together with the corresponding data points. P-values were determined by Tukey’s HSD test, **p < 0.01 (N = 3 persons, n = 100 cells). c A high magnification image of ballooned ER. Ribosomes on ER membrane help identification of the origin of the medium-size ERs (black ER), while ribosomes were detached when ER lumen was further enlarged and only a few ribosomes were remained (white ER). Mt: mitochondria. Nuc: nucleus. d Calnexin and MAP2 or KDEL and MAP2 co-staining of human neurons of non-neurological disease (control), MCI and AD patients. Remarkable enlargement of ER was detected (arrow) in MCI more frequently than in AD. e Calnexin and MAP2 or KDEL and MAP2 co-staining of mouse neurons of 5xFAD mice at 3 and 6 months of age. Remarkable enlargement of ER was detected (arrow) at 3 months more frequently than at 6 months. Source data are provided as a “Source Data file”.
Fig. 5
Fig. 5. Aβ sequesters YAP from the nucleus to intracellular aggregates.
a Immunohistochemistry of YAP and Aβ of human postmortem brains revealed sequestration of YAP into cytoplasmic Aβ aggregates and the resultant decrease of nuclear YAP in MCI and AD patients. b Quantification of signal intensities of nuclear YAP staining in neurons confirmed similar findings in three MCI patients and three symptomatic AD patients. A total of 160 neurons (N = 8 persons) in the occipital lobe were selected at random from each patient, and nuclear YAP signal intensities were quantified by confocal microscopy (FV1200IXGP44, Olympus, Tokyo, Japan). Signal intensity of DAPI per a nucleus was quantified in cortical neurons of occipital lobe of non-neurological disease controls and MCI or AD patients. The DAPI signals were compared in MCI between cytoplasmic YAP-positive and Aβ-positive neurons (n = 60) and cytoplasmic YAP-negative and Aβ-negative neurons (n = 60). In addition, signal intensity of DAPI per a nucleus of cytoplasmic was compared between YAP-positive and Aβ-positive neurons (n = 180) of MCI or AD patients (N = 3) and normal neurons (n = 180) of non-neurological disease controls (N = 3). Box plots show the median, quartiles and whiskers that represent 1.5× the interquartile range. The bar graph indicates average and mean ± S.E.M., together with the corresponding data points. P-values were determined by Tukey’s HSD test or student’s t-test, **p < 0.01 c Immunoprecipitation reveals the interaction between YAP and Aβ in human postmortem brain. Upper panels: anti-Aβ antibody (82E1) co-precipitated YAP from cerebral cortex tissues of AD patients, but not from non-neurological disease control. Lower panels: reverse co-precipitation of Aβ protofibrils with YAP. d Western and dot blots of Aβ and YAP. Temporal tip and occipital tip tissues from pathologically diagnosed AD patients or controls were immunoblotted with anti-Aβ antibody (82E1, dot blot) or anti-YAP antibody (sc-15407). e Inverse correlation between Aβ burden and YAP in human patient/control brains. P-values were determined by Pearson’s correlation coefficient (AD: N = 8 persons, Control: N = 6 persons). Source data are provided as a “Source Data file”.
Fig. 6
Fig. 6. Time lapse imaging of multiple neurons suggests pathological cascade.
a Experimental protocol to evaluate the relationship among intracellular Aβ, TEAD/YAP transcriptional activity and ER ballooning. b High magnification revealed that the membrane of the cytoplasmic balloon was reactive to ER-tracker in iPSC-derived neurons carrying APP mutations (left panels). ER ballooning occurs frequently in iPSC-derived neurons carrying APP mutations (heterozygous and homozygous mutants carrying APP KM670/671NL) (right graph). P-values were determined by Wilcoxon’s rank sum test with post-hoc Bonferroni correction, **p < 0.01 (N = 3 wells, n = 10 visual fields). c Accumulation of BTA1-stained intracellular Aβ occurs before ER ballooning in iPSC-derived neurons carrying APP mutations (left panels). Alignment of BTA1 signals changes in multiple neurons to the time point of ER ballooning initiation revealed intracellular Aβ began to accumulate 10 h before ER ballooning (right graphs). d Construction of a plasmid vector to monitor TEAD/YAP-transcriptional activity. e TEAD/YAP transcriptional activity was decreased in iPSC-derived neurons carrying APP mutations in comparison to normal iPSC-derived neurons (left panels). Alignment to ER ballooning time point revealed TEAD/YAP transcriptional activity started to decrease 8 h before ER ballooning. These results suggest a pathological cascade, increase of intracellular Aβ → decrease of TEAD/YAP transcriptional activity → ER ballooning. f Protocol of YAP-knockdown in normal human iPSC-derived neurons. g Time lapse imaging of siRNA-transfected neurons revealed the increase of ER ballooning by YAP-siRNA. Right graph shows quantitative analysis. P-values were determined by Wilcoxon’s rank sum test, **p < 0.01 (N = 3 wells, n = 10 visual fields). h Conformation of siRNA-mediated YAP-knockdown by immunohistochemistry. Green and white arrows indicate siRNA-mediated and non-transfected cells, respectively. i Conformation of siRNA-mediated YAP-knockdown by western blot. j HMGB1 concentration in culture medium was quantified by using ELISA, and the increase of HMGB1 from the initial concentration after siRNA transfection was compared between scrambled control siRNA and YAP-siRNA. P-values were determined by Student’s t-test (N = 3 wells). The box plot shows the median, quartiles and whiskers that represent 1.5× the interquartile range. The bar graph indicates average and mean ± S.E.M., together with the corresponding data points. Source data are provided as a “Source Data file”.
Fig. 7
Fig. 7. YAP knockdown induces ER ballooning in normal mice.
a Protocol of YAP knockdown in vivo. Labeled YAP-siRNA or scrambled control RNA was transfected into normal mice (B6/SJL), and ER images of siRNA-positive neurons were obtained by two-photon microscopy from 18 h later for 4 h. b In vivo time-lapse imaging of normal mice after siRNA transfection by two-photon microscopy. Right graph shows quantitative analysis of ER volume of siRNA-positive neurons. c Knockdown of YAP was confirmed by immunohitochemistry after in vivo imaging of two-photon microscopy. Green arrow indicates siRNA-transfected cell, and white arrow indicates non-transfected cell. d Knockdown of YAP was confirmed by western blot of mouse cortex tissues dissected after observation by two-photon microscopy. e Immunohistochemistry of pSer46-MARCKS after siRNA transfection. Patty immunostains were observed by transfection of YAP-siRNA but not scrambled control siRNA (low magnification). High magnification revealed DAPI-negative cell death surrounded by pSer46-MARCKS signals. Quantitative analysis confirmed the increase of pSer46-MARCKS immunostain signals by YAP-siRNA (right graph). The bar graph indicates average and mean ± S.E.M., together with the corresponding data points. P-values were determined by Student’s t-test, **p < 0.01 (N = 3 mice, n = 30 fields). f Quantitative analysis of nuclear volume after YAP-knockdown supported TRIAD necrosis. Left panels: representative nuclei in 3D imaging, right graph: quantitative analysis of a cell nucleus. The box plot shows the median, quartiles and whiskers that represent 1.5× the interquartile range. P-values were determined by Wilcoxon’s rank sum test, **p < 0.01 (N = 3 mice, n = 50 cells). Source data are provided as a “Source Data file”.
Fig. 8
Fig. 8. Time lapse imaging of a single neuron elucidates pathological cascade.
a Time lapse imaging protocol of a single neuron to analyze relationship between intracellular localization of YAP, intracellular Aβ and ER ballooning. Representative images of the same cells (similar to the cells in b) are shown in right panels. b Time lapse imaging of an iPSC-derived neurons carrying APP mutations (homozygous mutants carrying APP KM670/671NL). Nuclear YAP was shifted to intracellular Aβ in the cytoplasm (magenta arrow) and further to ballooned ER (green arrow). YAP was released to extracellular space via leakage of ballooned ER (white arrow) while intracellular Aβ remains as aggregates (blue arrow). The details were described in the text. c Chronological change of nuclear YAPdeltaC intensity and cytoplasmic BTA intensity in three iPSC-derived neurons carrying APP mutations. d Immunohistochemistry of human cerebral cortex with anti-calnexin, an ER membrane marker, and anti-YAP antibodies. Abnormal localizations of YAP in the cytoplasm or ballooned ER were observed frequently observed in MCI patients and at a low frequency in AD patients (white arrow), consistently with the findings in iPSC-derived neurons carrying APP mutations. Source data are provided as a “Source Data file”.
Fig. 9
Fig. 9. S1P and YAPdeltaC rescue ER instability and cognitive impairment in AD model mice.
a Experimental protocol for the rescue effect of S1P or YAPdeltaC (YAPdC) on ER instability and cell death in 5xFAD mice. Two protocols, administration before symptomatic onset (1 month) or administration just after onset (5 months), were used. AAV-NINS: AAV-CMV-no insert. b Alteration rate in Y-maze test of 6-month-old 5xFAD mice that had been treated with S1P from 1 month (upper left panel) or 5 months (lower left panel) or injected with AAV-YAPdeltaC at 1 month (upper right panel) or 5 months (lower right panel). P-values were determined by Tukey’s HSD test, *p < 0.05, **p < 0.01. N: shown below graphs. c ER instability was rescued at 6 months following treatment from 1 month with S1P or YAPdeltaC. P-values were determined by Welch’s test, **p < 0.01 (N = 3 mice, n = 9 cells). d Pathological examination at 6 months following treatment from 1 month with S1P or YAPdeltaC. Staining of YAP with the sc-15407 antibody detecting YAP-FL and YAPdeltaC, intracellular/extracellular Aβ, and pSer46MARCKS is shown. Right graphs show quantification of the four stains before and after the treatment. P-values were determined by Welch’s test, **p < 0.01 (N = 3 mice, n = 30 or 60). e Western blot with anti-Aβ antibody (82E1) confirmed the effect of S1P and AAV-YAPdC on Aβ burden in 5xFAD mice. f ELISA of Aβ1-40 and Aβ1–42 consistently showed the effect of S1P and AAV-YAPdC reducing Aβ burden in 5xFAD mice. P-values were determined by Tukey’s HSD test, *p < 0.05, **p < 0.01 (n = 4 mice). g Western blot confirming the increase of YAPdeltaC protein in cerebral cortex tissue of 5xFAD mice after AAV-YAPdeltaC infection. S1P also increased YAPdeltaC. P-values were determined by Tukey’s HSD test, #p < 0.05, ##p < 0.01 (n = 3 tests). h Immunohistochemistry of cerebral cortex tissue of 5xFAD mice supported the increase of total YAP after S1P and AAV-YAPdeltaC treatments. P-values were determined by Tukey’s HSD test, ##p < 0.01 (n = 50 cells). Box plots show the median, quartiles and whiskers that represent 1.5× the interquartile range. Bar graphs indicate average and mean S.E.M., together with the corresponding data points. Source data are provided as a “Source Data file”.
Fig. 10
Fig. 10. S1P and YAPdeltaC rescue ER instability and necrosis in genome-edited iPSC-derived AD neurons.
a Protocol for the rescue effect of S1P on ER instability and cell death. b Time-lapse images of non-treated iPSC-derived AD neurons exhibited ER ballooning (white arrows) and rupture. S1P treatment stabilized the ER and decreased necrosis. Aβ accumulation in ER (yellow area) and the leak to cytoplasm (green arrow). Representative images are shown in Supplementary Movie 7-12. c S1P did not change %BTA1-positive iPSC-derived AD neurons. d, e Suppressive effect of S1P on ER ballooning and necrosis of iPSC-derived AD neurons d and BTA1-positive neurons e. f Protocol for the rescue effect of YAPdeltaC on ER instability and cell death. AAV-NINS: AAV-CMV-no inset. g Time-lapse images showed ER ballooning (white arrows) and rupture of iPSC-derived AD neurons, which were suppressed by YAPdeltaC expression. Aβ was mainly accumulated in ER (yellow area), but a portion leaked into the cytoplasm (green arrow). Representative images are shown in Supplementary Movie 13-18. h YAPdeltaC did not change %BTA1-positive iPSC-derived AD neurons. i Suppressive effect of YAPdeltaC on ER ballooning and necrosis in iPSC-derived AD neurons i and BTA1-positive neurons j. k, l. Rescue of transcriptional function, as determined using TEAD-responsive element reporter plasmid in iPSC-derived neurons. S1P k and YAPdeltaC l restored TEAD-YAP/YAPdeltaC-dependent transcription, which was suppressed in iPSC-derived AD neurons. **p < 0.01 (N = 6 wells), Tukey’s HSD test. m Time-lapse images of shrinkage apoptosis and ballooning necrosis (arrow). n The ratio of each type of cell death (shrinkage apoptosis, rupture necrosis and ballooning necrosis) to total neurons that occurred naturally during 24 h of time-lapse imaging in normal and iPSC-derived AD neurons. ##p < 0.01 (N = 3 wells, n = 10 visual fields), Tukey’s HSD test. o. The ratio of each type of cell death to total neurons that occurred during 24 h after transfection of YAP-siRNA or scrambled-siRNA. **p < 0.01 (N = 3 wells, n = 10 visual fields), Wilcoxon’s rank sum test. c, d, e, h, i, j **p < 0.01 (N = 3 wells, n = 10 visual fields) by Wilcoxon’s rank–sum test with post-hoc Bonferroni correction. Box plots show the median, quartiles and whiskers that represent 1.5× the interquartile range. Source data are provided as a “Source Data file”.

Similar articles

See all similar articles

References

    1. Doody RS, Farlow M, Aisen PS, Alzheimer’s Disease Cooperative Study Data A, Publication C. Phase 3 trials of solanezumab and bapineuzumab for Alzheimer’s disease. N. Engl. J. Med. 2014;370:1459–1460. doi: 10.1056/NEJMoa1312889. - DOI - PubMed
    1. Honig LS, et al. Trial of Solanezumab for Mild Dementia Due to Alzheimer’s Disease. N. Engl. J. Med. 2018;378:321–330. doi: 10.1056/NEJMoa1705971. - DOI - PubMed
    1. Doody RS, et al. A phase 3 trial of semagacestat for treatment of Alzheimer’s disease. N. Engl. J. Med. 2013;369:341–350. doi: 10.1056/NEJMoa1210951. - DOI - PubMed
    1. Mullard A. BACE failures lower AD expectations, again. Nat. Rev. Drug Discov. 2018;17:385. - PubMed
    1. Sepulcre J, et al. Neurogenetic contributions to amyloid beta and tau spreading in the human cortex. Nat. Med. 2018;4:1910–1918. doi: 10.1038/s41591-018-0206-4. - DOI - PMC - PubMed

Publication types

Feedback