Sertad1 is elevated and plays a necessary role in neuron loss and cognitive impairment in a mouse model of Alzheimer's disease

Abstract

Aberrant activation of autophagy contributes to neuronal cell death and plays a pivotal role in the pathogenesis of Alzheimer's disease (AD). To study this further, we assessed autophagy-related (ATG) proteins in 5xFAD mice at different ages and found a progressive inappropriate elevation of autophagic proteins in these mice. We identified a transcriptional coregulator, Sertad1 (SERTA domain–containing protein 1), which plays a necessary role in neuron death, as a key regulator of aberrant autophagy in AD. We found a progressive elevation in Sertad1 levels in 5xFAD mice with age compared with WT mice. Sertad1 knockdown in 5xFAD mice brains lowered levels of ATG proteins and lysosomal proteins, suggesting its role in the regulation of the autophagy–lysosomal pathway. We found that Sertad1 knockdown restored Akt activity, which is inhibited in AD, and blocked the activation of its target, FoxO3a, which is translocated to the nucleus in the absence of active Akt and mediates neuron death by apoptosis and autophagy. Furthermore, we showed that lentivirus-mediated RNAi targeting of Sertad1 in 5xFAD mice led to better performance in behavioral experiments compared with 5xFAD mice treated with nontargeting shRNA, accompanied by significant restoration of synaptic integrity. Overall, our results demonstrated that autophagy is robustly induced with disease progression, but ATG proteins accumulate in the brain because of their impaired clearance; Sertad1 knockdown restored synaptic failure and improved cognition in 5xFAD mice by enhancing clearance of ATG proteins and neuronal survival. Hence, Sertad1 could be an excellent target for therapeutic intervention to combat the multifaceted pathologies of AD.

Alzheimer's disease (AD) is an age-associated progressive neurodegenerative disorder marked by the accumulation of aberrantly processed amyloid-β (Aβ) and neurofibrillary tangles (1). Autophagy is an evolutionarily conserved homeostatic pathway by which long-lived misfolded protein aggregates are cleared and nutrients are recycled. Although autophagy is highly efficient and constitutively active at basal levels in neurons (2, 3, 4), excessive autophagy can be detrimental and can lead to cell death (5, 6, 7, 8). In AD, increasing proteinopathic stress is accompanied by failing lysosomal machinery that overburdens the system, resulting in the accumulation of undigested cellular waste causing neurotoxicity (9). Autophagy–lysosomal pathway (ALP) is a key modulator of the cellular phase of AD that proceeds to the clinical phase, where all cellular mechanisms fail to maintain homeostasis and cognitive decline ensues (10). Mounting evidence suggests that dysfunctional autophagosomes and lysosomes accumulate in the AD brain, and these structures bear resemblance to the dense autophagic vacuoles seen when the last stages of autophagy are impeded by chemical modulators (9, 11). A currently unresolved question is how impaired autophagy is regulated in AD and at what stage does autophagy switches from its homeostatic to detrimental role.

Biswas et al. (12), in a microarray screening of genes, identified SERTA domain–containing protein 1 (Sertad1), also known as Trip-Br1, Sei1, or p34 to be rapidly upregulated 46-fold by DNA-damaging agent in cortical neurons. This protein was found to be necessary for neuron death in response to trophic factor withdrawal, Aβ_(1–42) treatment, and DNA damage (12). Sertad1 was found to bind and activate CDK4 (13), triggering a pathogenic cascade that involves the derepression of E2F transcription factor 1 (E2F1) complex, retinoblastoma hyperphosphorylation, and subsequent activation of Myb and Bim, all of which lead to neuronal apoptosis in these models (12, 14, 15). Much of the available literature on Sertad1 stems from non-neuronal contexts, especially cancer. For example, Sertad1 has been reported to act as a transcriptional coregulator of E2F1–DP1 in SAOS2 cells (16). In another report, expression of Sertad1, Sertad2, or Sertad3 activated the transcriptional activity of p53 in HeLa and U2OS cells (17). A yeast two-hybrid library screening of mouse embryonic heart revealed that Sertad1 interacts with SMAD1 and acts as its transcriptional coregulator during cardiogenesis (18). In cancer models, Sertad1 exhibits an antiapoptotic role by binding and stabilizing the X-linked inhibitor of apoptosis (19). Sertad1 plays an important role in suppressing multiple cell death pathways, including apoptosis, autophagy, and necrosis upon nutrient or serum deprivation (20, 21). These reports reflect cell type–specific functions of Sertad1 that differ from its prodeath role in postmitotic neurons. In addition, Sertad1 has been shown to increase inflammatory activity by specifically binding to NLRP3, where it plays an adaptor function in immune cells (22). Overall, these findings establish Sertad1 as a multifunctional signaling molecule with context-dependent roles, making it a relevant target for therapeutic intervention.

AD is a complex multifactorial disorder, where using a single approach for disease intervention might not be effective. It is imperative to understand the molecular mechanism underlying the pathogenic events in AD to design an effective combinatorial therapeutic strategy. Dysregulated autophagy is one of the primary reasons why accumulated amyloid plaques and neurofibrillary tangles fail to get cleared. Sertad1 is known for regulating cell cycle, apoptosis, and DNA damage responses, conditions significantly relevant in neurodegeneration in AD. Moreover, Sertad1 is responsive to nutrient and growth factor withdrawal, both of which are potent autophagy inducers. These associations suggest that downregulation of Sertad1 might facilitate autophagic degradation of amyloid plaques and prevent neuron loss. In the present study, we found that knockdown of Sertad1 using RNAi conferred cognitive benefits and restored synaptic health in a transgenic 5xFAD mouse model, unraveling its excellent potential as a therapeutic target for AD.

Results

Autophagosome and lysosome markers accumulate with age in 5xFAD brain

For temporal profiling of autophagy in AD, we checked autophagy markers of different stages with disease progression from cortical and hippocampal protein lysates of 5xFAD and WT mice brains, as these regions are the most affected areas during disease (23). Cytosolic LC3I is lipidated to form LC3II (24) and is used as a marker for autophagosome number (25). We found that both LC3II:LC3I and LC3II:actin ratios are significantly increased with age in the hippocampus (Fig. 1, A–C) and cortex (Fig. S1, A–C) in 5xFAD mice brains. Defects in autophagy flux result in p62 accumulation because of impaired clearance (26, 27). Our immunoblots showed that p62 levels increased significantly at 6 and 12 months of age in 5xFAD mice (Fig. 1, A and D and Fig. S1, A and D). We also checked for other autophagy-related (ATG) proteins, including ATG5 and ATG7, and found a similar trend where both these proteins show elevated levels with age progression in 5xFAD mice relative to WT mice (Fig. 1, A and E, F and Fig. S1, A, E and F). Consistent with previous observations, we found that lysosomal membrane protein, lysosomal-associated membrane protein 1 (LAMP1), is significantly elevated as early as 1 month of age compared with WT and accumulates with increasing age in 5xFAD brain relative to age-matched WT brain (Fig. 1, A and G and Fig. S1, A and G). Our data reveal that autophagy–lysosome impairment occurs early in AD pathogenesis.

Figure 1.

Autophagosome and lysosome proteins accumulate in the 5xFAD brain.A–G, hippocampal tissues were isolated from the brains of 5xFAD or WT mice of 1 month, 3 months, 6 months, and 12 months of age, protein lysates were prepared and subjected to Western blotting. A, representative immunoblots of autophagosome and lysosome proteins show their accumulation with increasing age in the 5xFAD hippocampal lysates when compared with their age-matched WT mice. Actin was used as a loading control. It is to be noted that representative images of LC3I and LC3II (18 kDa and 16 kDa), p62 (62 kDa), and actin were taken from the same immunoblot shown in the second and fourth panels of the 1-month column and 12-month column, respectively; thus, the same actin image is depicted in both the columns. Similarly, representative images of ATG5 (56 kDa), ATG7 (78 kDa), and actin were taken from the same immunoblot shown in the sixth and eighth panels of the 1-month column, thus the same actin image is depicted in the column. Graph of relative expression of LC3II:LC3I (B), LC3II (C), p62 (D), ATG5 (E), ATG7 (F), and LAMP1 (G) normalized with actin quantified using NIH-ImageJ software. Individual data points on the graph represent data from each mouse. Information on the number of mice used for the analysis is given in Table S2. Data represent mean ± SEM. Statistical analysis was done using a two-tailed unpaired t test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. ATG, autophagy-related; FAD, familial Alzheimer's disease; LC3, microtubule-associated protein 1A/1B-light chain 3; NIH, National Institutes of Health.

SERTAD1 is elevated in AD human brains, primary rat cortical neurons upon Aβ_(1–42) treatment, and 5xFAD mice brains

We performed bioinformatic analysis on gene expression datasets available in the public domain (Gene Expression Omnibus [GEO] accession ID: GSE5281) from previously published research on brain regions isolated from postmortem brains (age range, 63–102 years) (28, 29, 30, 31). Brain areas considered for gene expression analysis include the entorhinal cortex, hippocampus, medial temporal gyrus, posterior cingulate, superior frontal gyrus, and primary visual cortex. Among the differentially altered genes in these six regions, we found that SERTAD1 gene transcripts were upregulated ∼1.4-fold (p = 0.0009) in AD-affected samples (N = 87) as compared with the control samples from normal-aged brains (N = 74), suggesting that it plays a crucial role in AD pathogenesis (Fig. 2A). Previously, Biswas et al. (12) found an early and progressive induction of Sertad1 transcript levels in response to DNA-damaging agent and Aβ_(1–42) treatment in rat primary cortical neurons. Similar results were observed in response to trophic factor deprivation in primed PC12 cells and primary superior cervical ganglion sympathetic neurons from neonatal rats, where Sertad1 showed an early rise both at the transcript and protein levels. To determine the involvement of Sertad1 in AD progression, we assessed Sertad1 expression in primary rat cortical neurons upon Aβ_(1–42) treatment. We observed that Sertad1 increased significantly at 16 h and 24 h of Aβ_(1–42) exposure relative to control neurons (Fig. 2, B and C). We found significant upregulation of Sertad1 in the hippocampus in 1-month-old 5xFAD mice and its significant upregulation in both the hippocampus and cortex from 3 months onward compared with WT mice (Fig. 2, D–F). Immunostaining with Sertad1 and MAP2 (microtubule-associated protein 2) antibodies showed that Sertad1 is expressed in neurons and is upregulated in the hippocampus of 6-month-old 5xFAD mice as compared with WT (Fig. 2, G and H). Therefore, Sertad1 is elevated at the transcriptional level in human neurons from AD-affected brains and at the protein level in rat cortical neurons treated with Aβ_(1–42) oligomers and 5xFAD brains, respectively.

Figure 2.

SERTAD1 is elevated in human AD patients' brains and in vitro and in vivo models of AD.A, bioinformatic analysis was done based on a reference dataset of genes altered in AD pathogenesis. Violin plot shows the fold change differences in SERTAD1 gene expression in postmortem brain samples from six brain regions (control: N = 74; AD: N = 87). Data represent mean ± SEM. Statistical analysis was done using a two-tailed unpaired t test with Welch's correction (p = 0.0009). B and C, primary rat cortical neurons were treated with 1.5 μM Aβ_(1–42) for the indicated time points, and total cell protein lysates were prepared and subjected to Western blotting. B, representative immunoblot showing Sertad1 and actin at the indicated time points after Aβ_(1–42) exposure in primary cortical neurons. C, graph of Sertad1 relative to actin normalized with control is depicted. Individual data points represent normalized fold change of Sertad1 from each biological experiment. D–F, cortical and hippocampal tissues were isolated from the brains of 5xFAD or WT mice of 1 month, 3 months, 6 months, and 12 months of age, and protein lysates were prepared and subjected to Western blotting. D, representative immunoblots showing Sertad1 and actin expression in the cortex and hippocampus of 5xFAD or WT mice at the indicated ages. It is to be noted that the actin images of the bottom panels of the columns of 3 months, 6 months, and 12 months from Fig. 1A have been reused in this figure in the bottom panels of the columns of 3 months, 6 months, and 12 months, respectively, as LAMP1 (90 kDa), Sertad1 (34 kDa), and actin (42 kDa) were visualized on the same immunoblot. Similarly, LC3I and LC3II (18 kDa and 16 kDa), Sertad1 (34 kDa), and actin were visualized on the same immunoblot shown in the second panel of the 3-month column of this figure and the second panel of 3-month column of Fig. S1A, thus the image of actin has been depicted in both the figures. Graph of Sertad1 normalized with actin in the cortex (E) and hippocampus (F) of 5xFAD mice or WT mice as indicated was quantified using NIH-ImageJ software. Individual data points represent data from each mouse. N = 7 to 8 mice/group; details about the number of male and female mice used are given in Table S2. G and H, transcardially perfused brains were isolated and stained to check for Sertad1 expression in 6-month-old 5xFAD or WT mice. F, immunofluorescence image showing Sertad1 (green), MAP2 (red), and Hoechst (blue) staining shows elevated levels of Sertad1 in the CA1 region of the hippocampus of 5xFAD brain. H, graph of integrated density of Sertad1 relative to WT mice brain. Individual data points represent the Integrated (Int.) density of Sertad1 from brain tissue slices obtained from each mouse (N = 4 mice/group). Data represent mean ± SEM. Statistical analysis was done using a two-tailed unpaired t test. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001. Aβ, amyloid-β; AD, Alzheimer's disease; LAMP1, lysosomal-associated membrane protein 1; LC3, microtubule-associated protein 1A/1B-light chain 3; MAP2, microtubule-associated protein 2; NIH, National Institutes of Health; SERTAD1, SERTA domain–containing protein 1.

Knockdown of Sertad1 using lentiviral delivery of shSertad1 in mice brains

To study the role of Sertad1 in AD pathogenesis, we established a Sertad1 knockdown model using lentiviral delivery of shRNA. After validation of the efficacy of shRNA against Sertad1 that has been previously described (12) (Fig. S2, A–D), we injected the viral particles in the mouse hippocampus (Fig. S3A; see the Experimental procedures section) and confirmed GFP expression in the hippocampus and cortex after 14 days of surgery, confirming virus expression (Fig. S3, B and C). To assess for downregulation, we coimmunostained tissue sections with Sertad1 and GFP 14 days postsurgery and found significant downregulation of Sertad1 levels along with GFP expression in mouse brain after 14 days in the hippocampus (Fig. 3, A and B) and cortex (Fig. S3, D and E). Behavioral analysis was done 14 days postsurgery, and mice were sacrificed 21 days postsurgery for biochemical experiments (Fig. S3A). To account for any nonspecific RNAi effects, we infused scrambled shRNA-expressing lentiviral particles (scLV) into the hippocampus of WT and 5xFAD mice. We then stained brain sections with GFP to validate the virus expression in scLV infused, empty vector (EV)-lentivirus infused, and shSertad1-LV infused to WT and 5xFAD mice, respectively, after 2 weeks of surgery and compared it with uninfused WT and 5xFAD mice (Fig. S3C). No GFP staining was observed in the uninfused WT and 5xFAD mouse brain. We performed the locomotion test, which assesses anxiety in rodents (32, 33). We found no significant difference in the distance traveled and time spent in the center between WT and WT+scLV mice and between 5xFAD and 5xFAD+scLV mice. Whereas, 5xFAD and 5xFAD+scLV showed significant hyperlocomotor activity and spent more time in the center because of their anxious behavior as compared with the WT and WT+scLV (Fig. S4, A–C). In addition, we performed the contextual and cue-dependent fear conditioning experiments to assess the associative memory in rodents. We found that WT and WT+scLV showed a higher %freezing associated with better memory retention in both contextual (Fig. S4D) and cue-dependent fear conditioning tasks as compared with 5xFAD and 5xFAD+scLV mice (Fig. S4E).

Figure 3.

Sertad1 knockdown confers neuroprotection in in vivo AD models. shSertad1- or EV-expressing lentiviral particles were stereotactically infused in 5.5-month-old 5xFAD or WT mice brains. Transcardially perfused mice brains were isolated, tissue sections (30 μm) were collected, and staining was performed. A, coimmunostaining with anti-Sertad1 (red), anti-EGFP (green) antibodies, and Hoechst (blue) 14 days postsurgery shows reduced levels of Sertad1 upon knockdown in the hippocampal region. B, graph showing integrated (Int.) density of Sertad1 quantified using NIH-ImageJ software. Individual data points represent Int. density of Sertad1 from the brain tissue slices of each mouse (N = 3–4 mice/group). C and D, representative image showing TUNEL (red) staining in the CA1 region of the hippocampus of brain sections from shSertad1- or EV-infused 5xFAD or WT mice as indicated. Nuclei were stained with Hoechst (blue). D, graph showing the number of TUNEL-positive cells relative to total cells in the field expressed as a percentage. Individual data points represent percent of TUNEL-positive cells from brain tissue slices of each mouse (N = 3 mice/group). E–H, immunofluorescence image of 6E10 (red) in shSertad1- or EV-infused mice of the indicated groups in the CA1 region of the hippocampus. Graph of plaque number (F), plaque burden (G), and Int. density have been depicted (H). Individual data points represent the Int. density values of each mouse. Details about the number of mice used for the study are given in Table S4. Data represent mean ± SEM. Statistical analysis was done using one-way ANOVA, Tukey's post hoc analysis. ∗∗p < 0.01 and ∗∗∗p < 0.001. AD, Alzheimer's disease; EGFP, enhanced GFP; EV, empty vector; FAD, familial Alzheimer's disease; NIH, National Institutes of Health; Sertad1, SERTA domain–containing protein 1.

Sertad1 knockdown reduces apoptosis levels and amyloid plaque burden in 5xFAD mice

Since Sertad1 is elevated in 5xFAD brain, we checked if its ablation can ameliorate apoptosis and plaque pathology in 5xFAD mice. First, we transfected primary neurons with shSertad1 and checked the morphology and % survival of neurons upon Aβ_(1–42) exposure. Our results indicated that Sertad1 knockdown showed higher % survival as compared with shRand-transfected neurons and retention in morphology up to 48 h as compared with shRand-transfected neurons (Fig. S2, C and D). Next, we assessed cell death by TUNEL assay in 5xFAD brain and found that the percent of TUNEL-positive cells was significantly reduced upon Sertad1 knockdown in 5xFAD brain (Fig. 3, C and D). We observed that the intensity of TUNEL staining was considerably higher in EV-infused 5xFAD mice as compared with shSertad1-infused 5xFAD mice. This can be attributed to the high level of apoptosis and DNA damage in 5xFAD mice. Similar results for the TUNEL assay were also observed by Zhang et al. (34), where the intensity of TUNEL staining appears to be higher in 5xFAD mice as compared with other groups. Amyloid plaque deposition is the primary hallmark of AD pathology and is observed in 5xFAD mice at early stages (35). We assessed for amyloid plaque burden, number of plaques, and amyloid integrated density and found it to be significantly reduced in shSertad1-injected 5xFAD mice as compared with 5xFAD infused with EV (Fig. 3, E–H) or nontargeting shRNA (Fig. S4, F–H). These results show that Sertad1 is associated with AD pathogenesis, and its knockdown increases neuron survival, reduces the percent of TUNEL-positive cells, and ameliorates amyloid pathology in 5xFAD mice.

Sertad1 plays a role in dysfunctional autophagy in 5xFAD mice

To determine the effect of Sertad1 on autophagy modulation, we assessed the protein levels of autophagy markers in the shSertad1-infused 5xFAD mice. We found that LC3II, p62, ATG5, and lysosomal marker, LAMP1, which was increased in 5xFAD mice brain, were reduced upon Sertad1 knockdown (Figs. 4, A–E and S5, A and B). To distinguish if reduced LC3II:LC3I and p62 levels upon Sertad1 knockdown were due to less autophagy induction or enhanced clearance of autophagosomes, we measured ALP activity biochemically based on previous reports (25, 36, 37). The schema for the turnover of LC3II and p62 has been depicted in Figure 4F and gives the measure of autophagic degradation or autophagy flux. Bafilomycin A1 (BafA1) inhibits autophagy by blocking vATPase-dependent lysosomal acidification and prevents the fusion of autophagosomes and lysosomes (38). First, we standardized the dose of BafA1 at which the autophagy markers LC3II and p62 accumulate in primed PC12 cells. We found that both LC3II and p62 accumulate at significantly high levels compared with the control cells at 15 nM concentration and reach a saturation point (Fig. S6, A–C). Therefore, we selected a 15 nM dose of BafA1 for the turnover assay. We treated primed PC12 cells in the presence and absence of Aβ_(1–42) and/or BafA1 and measured LC3II/LC3I and p62/actin ratios in the indicated conditions. Even though the LC3II band in the presence of both Aβ_(1–42) plus BafA1 seems to be of high intensity, the value of turnover of LC3II was reduced significantly when we quantified the ratio of LC3II/LC3I bands and compared it with or without BafA1. Our results clearly indicated that both LC3II and p62 turnover were significantly reduced upon Aβ_(1–42) treatment and were restored upon Sertad1 knockdown in primed PC12 cells (Fig. 4, G–I). Furthermore, we measured the total protein content of LC3II/LC3I and p62 in the absence of BafA1. Consistent with our in vivo results, we found that LC3II/LC3I and p62 levels were significantly upregulated upon Aβ_(1–42) treatment and reduced upon Sertad1 knockdown in primed PC12 cells in the absence of BafA1 (Fig. 4, G and J–K). We also found that shSertad1-transfected cells show low expression levels of LC3 and p62 in immunofluorescence studies in primed PC12 cells treated with oligomeric Aβ_(1–42) and reduced the number of puncta of LC3 and p62, respectively, as compared with control cells, suggesting that Sertad1 plays a role in autophagy modulation (Fig. S7, A–H). Furthermore, we checked for lysosomal proteases, cathepsin B (CTSB) and cathepsin D (CTSD) levels, and found that pro and mature forms of CTSB and CTSD were elevated in the hippocampus and cortex of 5xFAD mice but were significantly reduced upon Sertad1 knockdown (Fig. S5, A and C–F). These results indicate that Sertad1 knockdown enhances the clearance of autophagosomes by increasing ALP activity in AD models.

Figure 4.

Sertad1 knockdown ameliorates autophagy dysfunction in models of AD. shSertad1- or EV-expressing lentiviral particles were stereotactically infused in 5.5-month-old 5xFAD or WT mice brains, and protein lysates from the cortex and hippocampus were prepared 21 days postsurgery and subjected to Western blotting. A–E, immunoblots showing autophagosome markers: LC3, p62, lysosomal marker: LAMP1 and actin in hippocampal and cortical lysates of shSertad1- or EV-infused 5xFAD and WT mice as indicated. Graph of fold change of LC3II:LC3I (B), LC3II (C), p62 (D), LAMP1 (E), represented as normalized fold change (N = 4–6 mice/group). Data represent mean ± SEM. Individual data points represent relative expression from each mouse; details about the number of mice used are given in Table S4. F–K, PC12 cells were transfected with shSertad1 or shRand, primed, and treated with 3 μM Aβ_(1–42) and/or 15 nM bafilomycin A1 (BafA1) for 16 h as indicated. F, schematic diagram of the autophagic turnover assay. LC3II turnover was calculated by comparing LC3II:LC3I ratio in the presence and absence of BafA1. Similarly, p62 turnover was calculated by comparing normalized p62:actin in the presence and absence of BafA1. G, immunoblots showing LC3II, LC3I, p62, and actin expression in protein lysates from PC12 cells treated with Aβ_(1–42) and/or 15 nM BafA1, as indicated. Graph indicates LC3II turnover obtained by comparing LC3II:LC3I (+BafA1/-BafA1) (H) and p62 turnover (+BafA1/-BafA1) (I) represented as fold change with shRand-transfected cells. Graph of protein content of LC3II/LC3I (J) and p62 (K) normalized with actin represented as fold change with control cells. Individual data points represent each independent experiment. Statistical analysis was done using one-way ANOVA, Tukey's post hoc analysis. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Aβ, amyloid-β; AD, Alzheimer's disease; EV, empty vector; FAD, familial Alzheimer's disease; LAMP1, lysosomal-associated membrane protein 1; LC3, microtubule-associated protein 1A/1B-light chain 3; Sertad1, SERTA domain–containing protein 1.

Sertad1 knockdown blocks FoxO3a translocation to the nucleus and restores Akt activity

FoxO3a has been previously reported to modulate cell death modalities like autophagy and apoptosis (39, 40). FoxO3a is dephosphorylated at serine 253 and translocates to the nucleus and executes neuron death in AD models (41). We assessed if Sertad1 is required for FoxO3a localization and activation in primed PC12 cells treated with oligomeric Aβ_(1–42) as well as in our in vivo AD model. Our results revealed that in shSertad1-transfected cells, FoxO3a was retained exclusively in the cytosol as compared with shRand-transfected cells, where FoxO3a translocated to the nucleus in response to Aβ_(1–42) exposure (Fig. 5, A–C). Figure 5B depicts quantification by visual counting of primed PC12 cells, where FoxO3a is present in the nucleus and cytoplasm. In cells where FoxO3a staining overlapped with nuclear staining using Hoechst, we considered that FoxO3a was nuclear, whereas in cells where FoxO3a in the nucleus was absent, we considered FoxO3a to be cytoplasmic. We observed that in shRand-transfected cells treated with Aβ_(1–42), FoxO3a was present in the nucleus in a significantly higher number of cells as compared with shSertad1-transfected cells treated with Aβ_(1–42). To further validate these results, we measured the intensity of FoxO3a in the nucleus and cytoplasm using ImageJ (National Institutes of Health [NIH]) and measured their ratio. We observed that upon Aβ_(1–42) treatment, FoxO3a translocated to the nucleus, as evident from a significantly higher nuclear:cytoplasmic FoxO3a ratio, and this was reduced in Aβ_(1–42)-treated cells transfected with shSertad1 (Fig. 5, A and C). We next assessed phosphorylation of FoxO3a at Ser253 and found that phosphorylation was restored in shSertad1-5xFAD mice (Fig. 5, D and E). Interestingly, our study revealed that Sertad1 knockdown not only blocked FoxO3a nuclear translocation but also reduced total FoxO3a level (Fig. 5, D–F). Akt remains inactive in AD (42) and is an upstream negative regulator of FoxO3a (43). We observed reduced Akt phosphorylation in 5xFAD mice as compared with shSertad1-injected 5xFAD mice. Our results show that Sertad1 knockdown not only inhibits FoxO3a translocation to the nucleus but also reduces its protein levels (Fig. 5, G and H). Furthermore, Akt, a survival kinase whose activity is lost in 5xFAD mice, is restored in shSertad1-infused mice, suggesting its possible role in autophagy modulation.

Figure 5.

Sertad1 aids in FoxO3a nuclear translocation and inhibits its upstream kinase, Akt activity.A–C, PC12 cells were transfected with shSertad1 or shRand, primed and treated with 3 μM Aβ for 16 h. A, immunofluorescence images showing FoxO3a (red) staining in shSertad1- or shRand-transfected PC12 cells treated as indicated. Percentage-stained cells indicate the proportion of transfected cells where FoxO3a is present in the cytosol or nucleus. Graph of the percent of stained cells with FoxO3a in the cytoplasm or nucleus is depicted in (B). The nuclear FoxO3a: cytosolic FoxO3a ratio in transfected PC12 cells treated with Aβ is shown in (C). Individual data points represent nuclear:cytosolic FoxO3a levels for each individual cell. Approximately 15 to 20 cells were analyzed per group per experiment, and three independent experiments were performed. shSertad1- or EV-expressing lentiviral particles were stereotactically infused in 5.5-month-old 5xFAD or WT mice brains, protein lysates from cortex and hippocampus were prepared 21 days postsurgery and subjected to Western blotting. D–F, immunoblots showing P-FoxO3a, FoxO3a and actin in mice of the indicated groups. Graph of fold change of P-FoxO3a:T-FoxO3a (E) and FoxO3a fold change relative to actin. F, in the cortex and hippocampus of mice of the indicated groups quantified using NIH-ImageJ (N = 3–4 mice/group). G and H, representative immunoblots showing P-Akt (Thr308):T-Akt and actin in mice of the indicated group. H, graph showing fold change of P-Akt:T-Akt in the hippocampus and cortex of mice of the indicated groups quantified using NIH-ImageJ. Individual data points represent data from each mouse. N = 4 mice/group; details about number of mice used are given in Table S4. Data represent mean ± SEM. Statistical analysis was done using one-way ANOVA, Tukey's post hoc analysis. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Aβ, amyloid-β; FAD, familial Alzheimer's disease; NIH, National Institutes of Health; Sertad1, SERTA domain–containing protein 1.

Sertad1 knockdown ameliorates cognitive deficits that are observed in the 5xFAD mice

We assessed whether Sertad1 knockdown in 5xFAD mice can ameliorate behavioral deficits in AD mice. Hyperlocomotion is a measure of anxiety levels in mice. Our data indicated that 5xFAD mice show hyperlocomotor activity, total distance traveled, and reduced thigmotaxis behavior, time spent in the outer zone, significantly attenuated in shSertad1-5xFAD mice (Fig. 6, A–C). Figure 6, D and F shows the schema for contextual- and cue-dependent fear conditioning. Contextual fear conditioning test showed significant restoration in percent of freezing in shSertad1-infused 5xFAD mice relative to 5xFAD mice infused with EV (Fig. 6, D and E). In the cue-dependent fear conditioning test, we observed that 5xFAD mice show a lower percentage of freezing, which is significantly restored in shSertad1-injected 5xFAD mice, indicating Sertad1 knockdown restored cue-dependent fear-based memory in 5xFAD mice (Fig. 6, F and G). Our results indicate that regions of the brain associated with fear-based associative learning, that is, amygdala and hippocampal coordination, are significantly improved in shSertad1-injected 5xFAD mice as compared with EV-injected 5xFAD mice. Next, we performed the novel object recognition task that assesses long-term recognition memory. Figure 6H shows the schema for the experiment on the test day. We exposed the mice to an open arena and allowed them to explore two similarly shaped objects and recorded the time spent near each object. On the next day, we replaced one of the objects with a differently shaped novel object and recorded the time spent by mice near each of the objects. Our results reveal that shSertad1-5xFAD mice spend more time near the novel object compared with the familiar object (Fig. 6I) and display higher discrimination index (DI) and preference index (PI) as compared with 5xFAD mice, indicating that Sertad1 knockdown restores recognition memory (Fig. 6, J and K). We also performed the Morris water maze experiment, which tests hippocampus-dependent spatial navigation and memory (44). We trained the animals to find the hidden platform and recorded escape latency. Figure 6L depicts the path traversed by mice on the probe day of one mouse from each group. 5xFAD mice take a longer time to find the hidden platform as compared with shSertad1-5xFAD mice, which take less time to find the platform (Fig. 6M). On the fourth day, the time spent by mice in the target quadrant and the number of target quadrant entries were recorded in the absence of the platform. We observed that shSertad1-5xFAD mice spent significantly more time in the target quadrant relative to 5xFAD mice (Fig. 6, N and O). Our results show that Sertad1 knockdown ameliorates impaired cognitive functions in 5xFAD mice.

Figure 6.

Sertad1 knockdown ameliorates behavioral deficits in 5xFAD mice. shSertad1 or EV lentiviral particles were injected bilaterally in 5.5-month-old 5xFAD or WT mice, and they were subjected to cognitive tests after 14 days. A–C, representative images of the path traversed by mice of the indicated groups in the open field test. Graph of total distance traveled by mice (cm) (B) and time spent in the center by animals of the indicated groups (seconds) (C). D and E, diagram of contextual fear conditioning apparatus. E, graph of percentage freezing of mice in the indicated groups recorded 1 day after training for contextual fear conditioning test. F and G, diagram of cue-dependent fear conditioning apparatus. G, graph of percentage freezing of mice in the indicated groups recorded 1 day after training for cue-dependent fear conditioning test. H–K, diagram of experimental setup for novel object recognition (NOR). Graph of time spent near familiar and novel object, respectively, in seconds by mice of the indicated groups on the probe day recorded for 5 min in the NOR task (I). Graph depicting discrimination index (DI) (J) and preference index (PI) (K) of mice of the indicated groups. L, diagram of path traced by mice of the indicated groups in Morris water maze experiment on test day. Graph of time taken in seconds to find the hidden platform (escape latency) (M), time spent in target quadrant on test day recorded for 90 s (N), and number of target quadrant entries on the test day recorded for 90 s (O) by mice of the indicated groups. Information on the number of mice used for each behavioral test is given in Table S3. Individual data points represent data for each mouse. Data represent mean ± SEM. Statistical analysis was done using one-way ANOVA, Tukey's post hoc analysis. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. EV, empty vector; FAD, familial Alzheimer's disease; Sertad1, SERTA domain–containing protein 1.

Sertad1 knockdown ameliorates synaptic deficits and restores dendritic spine density in 5xFAD mice

We then investigated the protein levels of presynaptic and postsynaptic markers, SNAP25 and PSD95, respectively, which are the most important determinants of synaptic health in Sertad1 knockdown 5xFAD mice. We found that PSD95 and SNAP25 protein levels, which were significantly reduced in 5xFAD mice, were restored in shSertad1-5xFAD whole tissue lysates (Fig. 7, A–D). Dendritic spines constitute the majority of excitatory synapses and correlate with cognitive functions (45). We studied the cytoarchitecture of the brain by Golgi–Cox staining to quantify dendritic spine density. Figure 7E depicts an image of the brain area showing the region of interest. We found that shSertad1-injected 5xFAD mice show significant restoration in spine density as compared with 5xFAD mice injected with EV, particularly in the hippocampus and cortex (Fig. 7, F and G). We propose that Sertad1 knockdown improves synapse health substantially and thus improves cognitive function in 5xFAD mice.

Figure 7.

Sertad1 knockdown restores synapse health in 5xFAD brain. shSertad1- or EV-expressing lentiviral particles were stereotactically infused in 5.5-month-old 5xFAD or WT mice brain, and protein lysates from the cortex and hippocampus were prepared 21 days postsurgery and subjected to Western blotting. A and B, representative immunoblots of PSD95 and actin and graph showing PSD95 fold change relative to WT. C and D, representative immunoblots of SNAP25 and actin and quantification of SNAP25 fold change relative to WT mice. shSertad1- or EV-expressing lentiviral particles were stereotactically infused in 5xFAD or WT mice brains, and the whole brain was isolated 21 days postsurgery and Golgi staining was performed. Individual data points represent relative expression of synaptic proteins from the cortex and hippocampus, respectively, from each mouse. E, representative image of one Golgi-stained section from mice at 10X to show the region of interest. F and G, representative image and quantification of dendritic spines in hippocampus and cortex of tissue sections from mice of the indicated groups. The scale bar represents 5 μm. Individual data points represent the average number of spines per 10 μm from brain tissue slices of each mouse, N = 4 to 5 mice/group; detailed information about the number of mice used is given in Table S4. Data represent mean ± SEM. Statistical analysis was done using one-way ANOVA, Tukey's post hoc analysis. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. EV, empty vector; FAD, familial Alzheimer's disease; Sertad1, SERTA domain–containing protein 1.

Discussion

An overwhelmed and dysfunctional autophagy machinery in AD brains results in failed clearance of toxic misfolded protein aggregates within swollen dystrophic neurites, resulting in neurodegeneration (9, 46). Electron microscopic analysis of AD patients' brains displayed robust accumulation of dense double-membraned autophagic vacuoles within dystrophic neurites, where failing lysosomes deficient in proteases are unable to clear their substrates (11, 47). Our data show a progressive increase in LC3II and p62 levels in 5xFAD mice compared with WT brains, indicating that these markers accumulate with age in AD because of impairment in the autophagy machinery. Whereas ATG5 has been reported to be elevated in the plasma of AD patients (48), reports on ATG7 levels in AD are conflicting (49). Supporting our observations, bioinformatic and biochemical analysis reveals that autophagy proteins in AD patients are elevated with an impairment in autophagy flux (46, 50). It appears that early autophagic induction occurs to cope with the aggravating burden of plaques and misfolded protein aggregates to degrade them. Contrarily, there are some reports that suggest that autophagy markers are downregulated in the AD brain and serum of patients (51, 52). Autophagy failure in AD is reported to be due to defects in lysosomal acidification, a detailed mechanism of which is still elusive (53, 54). By temporal profiling of autophagy markers, we wanted to correlate autophagic dysregulation with the onset of other already known pathologic changes that take place in the AD brain. Based on characterization studies in 5xFAD mice, it was reported that hippocampal long-term potentiation decline and synaptic impairments begin at 10 weeks of age in 5xFAD mice, preceding plaque accumulation (55). These changes are accompanied by microgliosis, astrocyte reactivity, impaired synaptic transmission, and long-term potentiation because of major genetic alterations (35, 56). Extracellular plaque deposition becomes significantly apparent at 5 months of age and shows a progressive increase in the hippocampus, cortex, and subiculum regions of the 5xFAD brain (35, 56, 57). Our data show that autophagy dysfunction starts at 3 months, and it coincides with the onset of other cellular alterations that lead to the manifestation of cognitive decline. The cellular phase of AD comprises multiple dysregulated pathways occurring simultaneously that negatively influence each other, which manifests later as the clinical phase of AD (10).

Our experimental observations suggest that Sertad1 is involved in autophagy dysfunction and cognitive deficiencies in 5xFAD mice, suggesting a strong mechanistic link between autophagy impairment and memory function. Bioinformatic analysis of gene expression from publicly available datasets revealed that Sertad1 is elevated significantly in six brain regions isolated from AD-affected individuals as compared with the corresponding brain regions isolated from healthy aged individuals. Supporting this, we found a significant increase in Sertad1 protein levels in AD transgenic mice brains with increasing age as compared with their WT counterparts. A similar increase in Sertad1 levels was also observed in primary rat cortical neurons treated with oligomeric Aβ_(1–42) at increasing time points. Analysis of the percent of surviving neurons in response to Aβ_(1–42) treatment revealed preservation of neuronal morphology and an increased percent of survival upon Sertad1 knockdown. This correlated with reduced accumulation of autophagosome markers like LC3 and p62 and lysosomal proteins LAMP1, CTSB, and CTSD in Sertad1-knockdown 5xFAD brains. Furthermore, we investigated for ALP activity biochemically using ALP inhibitor BafA1. Our results indicated that reduced levels of autophagy and lysosomal proteins in Sertad1 knockdown cells were due to increased autophagy flux, rather than decreased induction of the autophagy pathway. While β- and γ-secretases are canonical enzymes responsible for the initiation of amyloidogenic processing of amyloid-beta precursor protein (APP), they are not traditional lysosomal enzymes. It has been reported that lysosomal hydrolases are active and favor amyloidogenic processing of APP, resulting in increased deposits of amyloid (58, 59). A recent study by Hook et al. demonstrated that CTSB, a lysosomal cysteine protease, functions as an alternative β-secretase, and downregulating CTSB improves cognitive deficiencies in AD (60). Therefore, we investigated the protein levels of lysosomal enzymes, CTSB and CTSD, and found that their pro and mature forms were elevated in 5xFAD brain, but their levels were reduced upon Sertad1 knockdown. Interestingly, we found that the levels of the active forms of these enzymes were significantly reduced in Sertad1-knockdown 5xFAD mice, suggesting that Sertad1 depletion activates an alternate pathway for autophagic degradation, independent of the cathepsins. Further studies are warranted here to understand how Sertad1 influences lysosomal degradation defects in a disease context.

A pertinent question that arises here is how Sertad1 regulates autophagy in AD. We investigated for FoxO3a activity and localization, which has been reported to play an important role in autophagy regulation (39, 61, 62). Numerous reports suggest that FoxO group of transcription factors, especially FoxO3, helps in autophagy induction and is essential for flux maintenance in different model systems (36, 40, 62, 63, 64). Interestingly, FoxOs can regulate autophagy from the nucleus as well as the cytoplasm. Within the nucleus, FoxO3 possesses transactivation activity and can directly regulate the autophagy network in adult neural stem cells (39) and cardiomyocytes (64). On the other hand, FoxO also exhibits indirect transcription-independent functions from the cytoplasm. It has been reported that FoxO1 is acetylated in the cytoplasm in response to oxidative stress or serum deprivation and physically binds with Atg7 to induce autophagy by interacting with other important proteins involved in autophagosome biogenesis (65). We have previously reported that FoxO3a translocates to the nucleus to induce apoptotic genes (41, 66). Nucleus-to-cytoplasmic shuttling of FoxO3a is tightly regulated through dynamic phosphorylation–dephosphorylation events by kinases and phosphatases like Akt and protein phosphatase 2A (43, 67). We and others have shown that Akt activity is lost during AD pathogenesis (41, 42). We found that phosphorylation of FoxO3a at serine 253, a site phosphorylated by Akt, is restored upon Sertad1 knockdown, and this is accompanied by restoration in Akt activity. While a report by Degtyarev et al. (68) showed that Akt inhibition promotes autophagy in a human tumor xenograft model, there are other studies that show the important role of Akt signaling in autophagy activation (69, 70). Thus, Akt modulation of autophagy displays an intricate bidirectional relationship, which is largely context dependent. Our observations revealed that cells with Sertad1 knockdown treated with Aβ_(1–42) showed exclusive retention of FoxO3a in the cytoplasm as opposed to cells without Sertad1 knockdown. Our observations led to the speculation that Sertad1 aids in FoxO3a translocation to the nucleus by blocking its inhibitory phosphorylation by Akt. Interestingly, we also observed reduced levels of FoxO3a protein upon Sertad1 knockdown. Studies have shown that E2F1, which is downstream of Sertad1, regulates the expression of FoxO3a (71). Furthermore, FoxO family also regulates the specificity of E2F1 transcription, thus controlling apoptosis (72). We predict that Sertad1 acts proximal to the interplay between these two transcription factors and plays an essential role in orchestrating neuron death. Based on our results, we speculate that increased Akt activity upon Sertad1 knockdown co-occurred with increased ALP activity in our AD models. However, in the present study, the mechanistic link between Sertad1, Akt, and FoxO3a is correlative. These results open up new avenues for investigation to dissect the precise molecular mechanisms in further detail.

AD is considered a synaptopathy, where impairments in synaptic transmission commence during early stages of AD and lead to memory loss (73, 74). Maintenance of continuous and efficient autophagy flux is essential for long-term memory consolidation and synaptic homeostasis (75, 76). Our observations revealed that Sertad1 knockdown in 5xFAD mice brains could ameliorate the cognitive impairments associated with AD. This was accompanied by a remarkable restoration in presynaptic and postsynaptic proteins and dendritic spine density in Sertad-knockdown 5xFAD mice. Our results demonstrated that elevated levels of Sertad1 in 5xFAD mice brains play a necessary role in neuron death, and its intervention can ameliorate AD pathogenesis.

This study is based on the foundational observations laid by Biswas et al. (12), who first identified the role of Sertad1 as an apoptosis mediator in neurons. While the earlier work established the essential role of Sertad1 in neuronal apoptosis, the present study posits Sertad1 as a vital mediator of both impaired autophagy and apoptosis, directly impacting cognitive learning and synapse health in vivo. The previous report places Sertad1 proximal to retinoblastoma hyperphosphorylation → E2F1 complex derepression → Myb/Bim activation → apoptosis. While supporting this conclusion, we have highlighted the role of Sertad1 in apoptosis alongside autophagy, taking into consideration a broader cell death paradigm. Using the 5xFAD mouse model, we have performed a comprehensive dissection of the role of Sertad1 in autophagy impairment, FoxO3a translocation, and Akt activity. Thus, this study not only corroborates but also expands on the original findings, providing novel mechanistic aspects and preclinical validation through behavioral and pathological evidence of Sertad1 as a potential target for therapeutic intervention.

In conclusion, we proposed a model (Fig. 8), where Sertad1 plays multifunctional roles in cell death regulation and synaptic plasticity, which explains why knockdown of Sertad1 could ameliorate cognitive deficiencies and other pathogenic features like plaque deposition in AD. Autophagy regulation in AD is complex and multidimensional, whose mechanism remains to be fully elucidated (77). Sertad1 has been associated with neurodegenerative paradigms like Parkinsons's disease (78), DNA damage, Aβ insult, and nerve growth factor (NGF) deprivation (12). Gene-based therapeutic approaches hold tremendous potential in understanding the molecular mechanism and designing treatment strategies for combating neurodegenerative disorders. This study sheds light on the efficiency of using viral vectors to target Sertad1 for amelioration of neurodegenerative disorders like AD.

Figure 8.

Schematic representation depicting that Sertad1 interference can ameliorate cognitive and synaptic deficiencies in 5xFAD mice. Graphical abstract shows that autophagosome and lysosome markers accumulate in the AD brain because of impaired clearance. Sertad1 is upregulated in 5xFAD mice in an age-dependent manner. Knockdown of Sertad1 ameliorated autophagy dysfunction, synaptic impairment, and cognitive deficiencies in the 5xFAD mice. AD, Alzheimer's disease; FAD, familial Alzheimer's disease; Sertad1, SERTA domain–containing protein 1.

Experimental procedures

Materials

Dulbecco's modified Eagle's medium (DMEM)–F12, DMEM high glucose medium, Neurobasal, OPTI-MEM, B27 supplement, fetal bovine serum (FBS), horse serum (HS), normal goat serum, Penstrep antibiotic, Lipofectamine, Hoechst 33342, and Prolong Gold antifade were purchased from Thermo Fisher Scientific. Human recombinant NGF, insulin, progesterone, putrescine, selenium, apotransferrin, poly-d-lysine, radioimmunoprecipitation assay (buffer), 1,1,1,3,3,3-hexafluoro-2-propanol, paraformaldehyde, dimethyl sulfoxide, PhosSTOP and bovine serum albumin were purchased from Sigma. Protease inhibitor cocktail, Lenti-X concentrator, enhanced chemiluminescence, Western blot stripping buffer, and protein ladder were purchased from Takara Bio. Polyvinylidene fluoride (0.45 μm) membrane was purchased from Merck Millipore. Other fine chemicals and reagents were procured from standard local suppliers. A detailed list of antibodies used in this study is given in Table S1.

Animals

5xFAD (AD transgenic) mice and their nontransgenic WT littermates, C57BL6 mice were procured from The Jackson Laboratory. 5xFAD mice harbor five mutations as follows: Swedish (K670N, M671L), Florida (I716V), and London (V717I) in APP, and M146L, and L286V in presenilin 1 regulated by mouse Thy1 promoter to drive overexpression in the brain, resulting in recapitulation of major pathologic features of familial AD.

Animal housing and care

Male and female 5xFAD and WT mice were housed separately in pathogen-free environments in individually ventilated cages with sufficient bedding material. At most, five mice were kept in individual cages in a controlled environment with temperatures of 22 to 25 °C, 12 to 12 h light–dark cycle, 60% to 70% humidity, and were given access to food and water ad libitum in the animal house of CSIR-Indian Institute of Chemical Biology, Kolkata. Details of the number of mice used for each experiment are given in the supporting information section and in the figure legends. All experimental procedures were carried out in strict compliance with the guidelines formulated for the care and use of animals by the Institutional Animal Ethics Committee, CSIR-Indian Institute of Chemical Biology, Kolkata.

Plasmid constructs

shRNA specific for Sertad1 with sequence 5′-CCAGACCTCCGACACCTGG-3′ validated in a previous report was used for transfection and lentivirus preparation (12). A scrambled version of this sequence, shRand, was used as a control plasmid. PLVTHM1 lentiviral vector–expressing shSertad1 from the H1 promoter and GFP was used for generation of lentiviral particles. PLVTHM1 EV was used as a control. The nontargeting scrambled shRNA-expressing GFP was purchased from Origene (catalog no.: TR30021V).

Preparation of lentiviral particles

Lentiviral particles were produced as described previously (79). Briefly, human embryonic kidney 293T (HEK293T) cells were cotransfected with shSertad1 or EV, along with the packaging plasmid and envelope plasmid. Lentiviral supernatant was collected after 48 and 72 h. The solution was concentrated using Lenti-X concentrator for infusion as per the manufacturer's protocol and stored at −80 °C until further use.

Stereotaxic procedure for lentiviral infusion in mouse brain

Male and female 5xFAD mice and WT mice (5.5 months old) were used for the experiments. Mice were anesthetized using a mixture of ketamine (60 mg/kg i.p.) and xylazine (8 mg/kg i.p.) intraperitoneally for a short time to carry out the stereotaxic injection procedures. Each mouse was placed on a stereotaxic frame (Stoelting). Five microliters of viral suspension containing 0.8 × 10¹⁰ lentiviral particles were injected bilaterally in the CA1 region of the hippocampus (coordinates: anterior–posterior axis: −0.2 cm, lateral axis: ±0.18 cm and dorsal–ventral axis: −0.15 cm from the bregma point) using a controlled syringe pump (BAS).

Cell culture

Rat pheochromocytoma (PC12) cells were cultured in DMEM containing 10% HI-HS and 5% HI-FBS. Naïve PC12 cells were differentiated using NGF at 100 ng/ml concentration with 1% HI-HS for 3 days prior to treatment with 3 μM Aβ_(1–42). HEK293T cells were cultured in DMEM with 10% HI-FBS. Primary rat cortical neurons were cultured as described previously (41). In brief, cells from the neocortex of an E18-day pregnant Sprague–Dawley rat fetus were dissected under a dissection microscope. Isolated tissue was triturated to form a single-cell suspension and plated on poly-d-lysine-coated plates. Cultured cells were maintained in vitro in DMEM–F12 medium supplemented with d-glucose (6 mg/ml), apotransferrin (0.1 mg/ml), insulin (25 μg/ml), progesterone (20 ng/ml), putrescine (60 μg/ml), and selenium (30 ng/ml). The next day, the cortical medium was replaced partially with neurobasal medium containing B-27. Cortical neurons were treated with 1.5 μM Aβ_(1–42) on the sixth day of plating in neurobasal medium containing B27.

Preparation of oligomeric Aβ_(1–42)

Lyophilized synthetic Aβ_(1–42) peptide with sequence DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVVIA (catalog no.: AB-250-10) was purchased from AlexoTech (Vasterbottens Ian). Lyophilized peptide was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol solvent to a final concentration of 1 mM and centrifuged in a speed-vac (Eppendorf). The peptide pellet was then resuspended in dimethyl sulfoxide to a final concentration of 5 mM followed by sonication at 37 °C for 10 min. The obtained solution was then stored in small aliquots at −80 °C. Prior to use, aliquots were diluted in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4) and 0.2% SDS to a final concentration of 400 μM and incubated at 37 °C for 6 to 18 h. The solution was further diluted with PBS to a final concentration of 100 μM and incubated at 37 °C for 18 to 24 h.

Transfection

DNA preparation was done using the HiPure Plasmid Midiprep kit purchased from Invitrogen, Thermo Fisher Scientific. For the survival assay, 0.5 μg of pSiren-shSertad1-zsgreen and pSiren-shRand-zsgreen plasmids were transfected on day 3 of plating primary cortical neurons and maintained for 48 h prior to treatment.

Transduction

HEK293T cells were cultured in DMEM with 10% HI-FBS. Lentiviral supernatant was collected after 48 h and 72 h, fed to cells, and samples were harvested 72 h after transduction.

Western blot assay

Samples were lysed in radioimmunoprecipitation assay buffer with phosphatase inhibitor (Roche) and protease inhibitor. Proteins were resolved on SDS-PAGE and transferred to polyvinylidene fluoride membrane (Merck) followed by blocking and primary antibody incubation overnight. After washing, the membrane was incubated with a secondary antibody. Protein bands were visualized on iBright Imaging System (Thermo Fisher Scientific) and analyzed using ImageJ software. Raw images of immunoblots are incorporated as a separate file in the supporting information.

Immunohistochemistry

Brain tissue sections of 30 μm thickness were permeabilized, washed, and incubated with primary antibody overnight, followed by incubation with Alexa Fluor–tagged secondary antibody and nuclear staining with Hoechst. TUNEL assay was performed using the In-Situ Cell Death Detection Kit, TMR red purchased from Roche, Merck as per the manufacturer's protocol. After mounting, images were taken in a Leica STED confocal microscope. Image analysis was done using ImageJ software. 3,3'-Diaminobenzidine staining was done using VECTASTAIN ABC Kit, peroxidase purchased from Vector Laboratories, as per the manufacturer's protocol. Sections were then mounted and imaged under a bright-field microscope.

Golgi–Cox staining

Golgi staining was performed according to a procedure described previously (80). Briefly, isolated mice brains were immersed in impregnation solution in the dark for 2 to 3 weeks, after which they were stored in 30% sucrose solution at 4 °C. Tissue sections of 75 μm thickness were treated with ammonia solution in the dark, dehydrated using ethanol, delipidated using xylene, mounted, and imaged using a Leica STED confocal microscope for analysis of dendritic spine density.

Immunocytochemistry

Fixed cells were blocked and incubated with primary antibody at appropriate dilution overnight at 4 °C. Cells were then incubated with appropriate Alexa Fluor–tagged secondary antibodies for 1 h, followed by nuclear staining with Hoechst and then mounted. Microscopic images were taken in a Leica STED confocal microscope. Image analysis was done using ImageJ software.

Analysis of nuclear translocation

Primed PC12 treated with oligomeric Aβ_(1–42) was immune stained with anti-FoxO3a antibody, and Hoechst was used for nuclear staining as mentioned previously. Images were obtained using a Leica STED confocal microscope. Visual counting was used to count the number of cells where FoxO3a was present in the cytoplasm and the nucleus. FoxO3a was considered to be localized in the nucleus if the FoxO3a staining colocalized with nuclear staining with Hoechst. Counting was done in a blinded manner by a lab member who was unaware of the experimental conditions. The percentage of cells where FoxO3a was present in the nucleus was calculated with respect to the total number of transfected cells in shRand- or shSertad1-transfected cells. ImageJ software was used to measure the intensity of FoxO3a in the whole cell and nucleus. Briefly, the intensity of the whole cell was calculated using ImageJ by specifying the area of the whole cell. The intensity of FoxO3a in the nucleus was calculated by specifying the region overlapping with the nuclear staining using Hoechst. Then, the intensity of cytosolic FoxO3a was calculated by subtracting the intensity of FoxO3a in the nucleus from FoxO3a intensity present in the whole cell. The ratio of intensity of FoxO3a in the cytoplasm and nucleus was calculated and plotted on the graph for comparison.

Analysis of plaque burden

Plaque pathology was analyzed as described previously (81). For assessment of amyloid plaque pathology, brain tissue sections were stained with antibody for 6E10, and ImageJ software was used for analysis. For each image, background intensity was subtracted by setting the rolling ball radius at 25. Thresholding was done using values that were set at the default threshold settings. Particle size of 5 μm² was counted as plaque and was plotted on the graph as plaque number. For plaque burden calculation, the total area occupied by Aβ plaques relative to the area of the selection or field (% area) was analyzed in the hippocampal region and plotted on the graph. For amyloid integrated density, the mean fluorescence intensity after background correction was plotted.

Behavioral experiments

Mice were divided into four groups: (1) EV-WT, (2) EV-5xFAD, (3) shSertad1-WT, and (4) shSertad1-5xFAD. Behavioral investigations were done 14 days after surgery.

Open field test

The experiment was carried out as described previously (32, 33). ActiTrac software (Panlab, Harvard apparatus) was used for tracking of mouse trajectories. An inner square area encompassing one-third of the total arena was considered as the inner zone, and the remaining peripheral area was considered as the outer zone. The total distance traveled (cm), time spent in the outer zone (seconds), and time spent in the inner central zone (seconds) were recorded by the software.

Fear conditioning test

Cue- and context-dependent fear conditioning experiments were conducted as described previously (32). A sound-proof chamber with a steel grid to deliver footshock is used for the experiment. The chamber is connected to a computer that has PACKWIN 2.0 software (Panlab) for recording freezing behavior in mice.

Context-dependent fear conditioning

On the acquisition day, each mouse was placed in the chamber and given 2 min for acclimatization to the environment. Following this, mice were subjected to four trials of an unconditional stimulus (89 s of intertrial interval followed by 1 s 0.4 mA footshock) and then a rest time of 2 min. The nex t day, mice were returned to the same chamber, but no footshock was given. The PACKWIN 2.0 software records percentage freezing, indicative of retention of context-dependent fear memory.

Cue-dependent fear conditioning

Similar to context-dependent fear conditioning, on the acquisition day, each mouse was placed in the chamber and given 2 min for acclimatization. The mouse was then subjected to four trials; 29 s of light and sound (80 dB and 2500 Hz) terminating in 1 s footshock, followed by 2 min of rest. The next day, the context was changed by attaching white sheets on the walls of the chamber. This ensures that context has no role to play in this test and that mice only respond to light and sound cues. The protocol was run without any footshock. The PACKWIN 2.0 software is used to record the percentage of freezing that reflects retention in cue-dependent fear memory.

Novel object recognition

The experiment was conducted as described previously (32, 33). Briefly, each mouse was allowed to explore a rectangular box of dimensions 50 cm × 30 cm with two objects of the same color and shape for 5 min on the acquisition day. On the test day, one of the similar objects was replaced with a differently shaped object, and each mouse was allowed to explore the objects for a total time of 5 min. Time spent near both the objects on the acquisition day and test day was recorded and tabulated. DI and PI (%) were calculated as follows: DI = (TN-TF)/(TN+TF); PI = TN/(TN+TF) ∗100 on day 2, where TN is the time spent near the novel object and TF is the time spent near the familiar object.

Morris water maze

The experiment was carried out as described previously with slight modifications (82). Briefly, in a circular pool, mice were trained to find the hidden platform by giving 12 trials over 3 consecutive days. Mice were removed from the maze after 90 s. Time taken to find the platform was recorded using the ANY-MAZE software (Stoelting Co) on each day and tabulated as escape latency in seconds. On the fourth day, the platform was removed, and time spent in the target quadrant and the number of times mice entered the target quadrant were recorded for 90 s using ANY-MAZE software.

Bioinformatic analysis

Differential gene expression analysis was performed on AD-affected human postmortem brain samples with transcriptomic dataset available in the public domain (GEO accession ID: GSE5281) using GEO2R tool for bioinformatic analysis (R script available with the dataset). Benjamini and Hochberg (false discovery rate) test was performed to obtain statistical significance in differential gene expression analysis. In particular, the gene expression of SERTAD1 was compared between unaffected control and AD-affected samples. Six different brain regions considered for the gene expression analysis include the entorhinal cortex, hippocampus, medial temporal gyrus, posterior cingulate, superior frontal gyrus, and primary visual cortex. Details of the SERTAD1 gene expression values and sample characteristics have been provided in the supporting information Excel sheet.

Statistical analysis

Student's unpaired t test was used for comparison between two groups. The ifference between more than two groups was calculated by one-way or two-way ANOVA with Tukey's multiple comparisons test using GraphPad Prism, version 9.0.0 (GraphPad Software, Inc). Data represent mean ± SEM. p < 0.05 was considered statistically significant.

Data availability

All data generated and used in this study have been included in this article and supporting information.

Supporting information

This article contains supporting information.

Ethical approval

All animal studies were carried out in accordance with the guidelines formulated by the Committee for Control and Supervision of Experiments on Animals (Ministry of Fisheries, Animal Husbandry and Dairying, Department of Animal Husbandry and Dairying, Government of India) with approval from the Institutional Animal Ethics Committee. We have consulted the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines for the relevant aspects of animal studies.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.

Reviewed by members of the JBC Editorial Board. Edited by Elizabeth J. Coulson