Targeting the overexpressed mitochondrial protein VDAC1 in a mouse model of Alzheimer’s disease protects against mitochondrial dysfunction and mitigates brain pathology

Background Alzheimer's disease (AD) exhibits mitochondrial dysfunctions associated with dysregulated metabolism, brain inflammation, synaptic loss, and neuronal cell death. As a key protein serving as the mitochondrial gatekeeper, the voltage-dependent anion channel-1 (VDAC1) that controls metabolism and Ca2+ homeostasis is positioned at a convergence point for various cell survival and death signals. Here, we targeted VDAC1 with VBIT-4, a newly developed inhibitor of VDAC1 that prevents its pro-apoptotic activity, and mitochondria dysfunction. Methods To address the multiple pathways involved in AD, neuronal cultures and a 5 × FAD mouse model of AD were treated with VBIT-4. We addressed multiple topics related to the disease and its molecular mechanisms using immunoblotting, immunofluorescence, q-RT-PCR, 3-D structural analysis and several behavioral tests. Results In neuronal cultures, amyloid-beta (Aβ)-induced VDAC1 and p53 overexpression and apoptotic cell death were prevented by VBIT-4. Using an AD-like 5 × FAD mouse model, we showed that VDAC1 was overexpressed in neurons surrounding Aβ plaques, but not in astrocytes and microglia, and this was associated with neuronal cell death. VBIT-4 prevented the associated pathophysiological changes including neuronal cell death, neuroinflammation, and neuro-metabolic dysfunctions. VBIT-4 also switched astrocytes and microglia from being pro-inflammatory/neurotoxic to neuroprotective phenotype. Moreover, VBIT-4 prevented cognitive decline in the 5 × FAD mice as evaluated using several behavioral assessments of cognitive function. Interestingly, VBIT-4 protected against AD pathology, with no significant change in phosphorylated Tau and only a slight decrease in Aβ-plaque load. Conclusions The study suggests that mitochondrial dysfunction with its gatekeeper VDAC1 is a promising target for AD therapeutic intervention, and VBIT-4 is a promising drug candidate for AD treatment. Supplementary Information The online version contains supplementary material available at 10.1186/s40035-022-00329-7.

protein for measuring VDAC1 expression levels and 60-80 µg protein for visualizing VDAC1 oligomerization) were resolved by SDS-PAGE and electro-transferred to nitrocellulose membranes that were subsequently blocked with 5% non-fat dry milk and 0.1% Tween-20 in Tris-buffered saline (TBS). Samples were then incubated in primary antibodies (sources and dilutions listed in Table S1), and then with the appropriate secondary HRP-conjugated anti-mouse or anti-rabbit antibody (Table  S1). Enhanced chemiluminescence substrate EZ-ECL (Cat. No. 205001000;Biological Industries, Israel) was used for detection of HRP activity. Analyses of immuno-reactive VDAC1 monomers, dimers, and multimer bands was performed using FUSION-FX (Vilber Lourmat, France).

TUNEL assay
Paraffin-embedded fixed brain sections were processed for a TUNEL assay using the DeadEnd Fluorometric TUNEL assay (Promega, Madison, WI), according to the manufacturer's instructions. Sections were deparaffinized, equilibrated in PBS, permeabilized with proteinase K (20 μg/ml in PBS), post-fixed in 4% paraformaldehyde, and incubated in TdT reaction mix for 1 h at 37°C in the dark. The slides were then washed in saline-sodium citrate buffer and counter-stained with propidium iodine (1µg/ml). After mounting with Vectashield mounting medium (Vector Laboratories, Burlingame, CA), images were collected using a confocal microscope (Olympus IX81).

RNA isolation from fixed samples and quantitative real-time PCR (RT-PCR) analysis
Total RNA was isolated from 20-µm sections cut from formaldehyde-fixed brains using the Recover All™ Total Nucleic Acid Isolation Kit for FFPE (Cat. No. AB-AM1975; Ambion™, Austin, Texas, US), according to the manufacturer's protocol. Total RNA quality was analyzed using the Agilent RNA 6000 nano kit. The RNA integrity values obtained for total RNA extracted were 8-10. Complementary DNA (cDNA) was synthesized using a qPCRBIO cDNA synthesis kit PCR Biosystems,London,UK,) according to the manufacturer's protocol. Quantitative-PCR was performed with specific primers (Table S2) in triplicates using SYBR Green master mix (Cat. No. 4367659;Thermo Fisher,Waltham,MS,US). Next, samples were amplified by the 7300 Real-Time PCR System (Applied Biosystems). Following this, target gene levels were normalized with α-actin mRNA as an internal control, and the mean fold change (± SEM) of the three replicates of 3-4 mice per group were calculated.  The genes examined and the forward and reverse sequences of the primers used are indicated.

Table S3. VDAC1 promoter sites matching sequence profiles generated from AβID decamers identified by Maloney and Lahiri
Profiles for the Aβ-Interacting Domain (AβID) were generated from two sets of sequences: (A) six EMSAvalidated AβID decamers from APOE, APP, BACE1, and TP53, presented in Table 2 in (110); and (B) 35 computationally-predicted AβID decamers from the promoters of 10 different genes, presented in Table 3 in (110). The profiles were constructed using MEME (111), while requesting that for each sequence set, only one profile (a.k.a. "motif") of 10 bp would be generated. The profiles were then searched with FIMO (111) against a 997-bp region surrounding the VDAC1 mRNA (RefSeq NM_003374.3) transcription start site (TSS, chr5:134,004,568-134,005,564 on the Human GRCh38/hg38 Assembly), consisting of 589 bp upstream and 408 bp downstream of the TSS. The results show profile occurrences with a p-value less than 0.0001. The pvalue of a profile occurrence is defined as the probability of a random sequence of the same length as the profile matching that position of the sequence with as good or a better score. The score for the match of a position in a sequence to a profile is computed by summing the appropriate entries from each column of the positiondependent scoring matrix that represents the profile. The q-value of a profile occurrence is defined as the false discovery rate if the occurrence is accepted as significant.
The most significant site on the VDAC1 promoter was located 309 bp upstream of the TSS. This site contained the four 3' Gs which appeared in all EMSA-validated sequences, including the G marked with an asterisk in Fig. 4 in (2), whose alteration to A was shown to significantly diminish Aβ binding (2). Moreover, this site was nearly identical (9 out of 10 bases) to the decamer predicted by dynamic weight analysis (2) to have the maximum possible score (GGGGTTGGGG) as AβID.  (c-e) Cells were incubated with cisplatin in the absence and presence of the indicated concentrations of VBIT-4, then VDAC1 oligomerization, as revealed using EGSbased cross-linking and quantitative analysis of VDAC1 dimers (c,d), and apoptotic cell death was assayed using propidine iodin and FACS analysis (e), as previously described (19). The positions of VDAC1 monomers and oligomers and of the molecular weight standards are indicated. The cells were also subjected to a survival assay using XTT (f). (g) Primary neural culture was infected with APP swe/lnd-EGFP Sindbis virus for 16 h to overexpress and secret A into the medium (conditioned medium). Immunostaining of APP-infected cells for VDAC1 and the expressed GFP is seen. (h) Apoptosis stimuli, pathological or stress conditions, and Ab enhance VDAC1 expression via activating the VDAC1 promoter by transcription factors (TFs) and/or Ca 2+ , leading to VDAC1 overexpression, and oligomerization of the overexpressed VDAC1, allowing cytochrome c or/and mtDNA release, and thereby apoptosis and inflammation.

Fig. S2. VBIT-4 improves cognition, learning, and memory performance in the 5XFAD mouse model
and had no effect on WT mice. (a) Y-Maze test was performed on WT (n=8), and untreated and 5-monthtreated 5XFAD mice (n=9) with VBIT-4 (n=9). Performance in the Y-maze allowing mice to explore all three arms of the maze. An ANOVA yielded a significant difference among the groups [f(2,33)=4.46, p=0.019].
Tukey post hoc analysis revealed that 5XFAD mice performed more poorly than WT mice (p=0.007), and VBIT-4-treated 5XFAD mice performed better than untreated mice (p=0.041). (b-e) Confocal images of cortical and hippocampal sections from WT, untreated, and VBIT-4-treated mice co-immunostained for VDAC1 and TUBB3 (B), or GFAP or IBA-1 (d) and their quantification (c,e). Results show means ± SEM (n=3), NS, not significant, showing no effect of VBIT-4 on the expression of the tested proteins in WT mice.      Cortical (a) and hippocampal (b) sections co-immunostained for citrate synthase (CS) and VDAC1. In WT, CS was strongly expressed in the cell cytosol (arrows in A). This expression was dramatically decreased in untreated, and largely restored in the VBIT-4-treated 5XFAD mice. CS was not localized with VDAC1 in the neuropils surrounding the A plaques.   ) and hippocampal (c,d) sections co-immunostained for Na,K-ATPase and VDAC1. No expression of Na,K-ATPase was observed in the A plaques surrounding neuropils in the 5XFAD mice, but Na,K-ATPase was expressed and co-localized with VDAC1 in the A plaques surrounding the neuropils (circled) in VBIT-4-treated 5XFAD mice, is further shown in the enlarged image (a(i)). Quantitative analysis of total Na,K-ATPase expression levels in the cortical neuropils surrounding the A plaques of VBIT-4-treated and untreated 5XFAD mice (b,d). Results show means ± SEM (n=3), **p <0.01 ***p <0.001. Analysis of the 3D-reconstructed images of astrocytes in 50m cortical sections from untreated and VBIT-4treated 5XFAD mice, immunostained for GFAP, imaged using Spinning disk microscopy, and analyzed using Imaris software. Total area of all processes (d) and total process area as a function of branch order (e) are larger in astrocytes of treated mice. (f) Illustration of IBA-1 staining and its quantification in the A plaques (a) and outside of the plaques (b) of black and white converted IBA1-staning. (g,h) Analysis for total process areas for microglia (as described above for astrocytes). Results show means ± SEM (n=3), p-values are presented, ***p <0.001. (i) IHC staining of brain sections from WT and 5XFAD mice immunostained for TSPO, showing its increased expression in the form of plaques in the 5XFAD mice.  I (a,b), or CS (c,d), or COX-IV (e,f). The morphology improved astrocytes in VBIT-4-treated 5XFAD mice, showing increased expression of the mitochondrial enzymes and their co-localization with GFAP, as is further shown in the enlarged images (a(i), c(ii), e(iii)). Results show means ± SEM (n=3). Results show means±SEM (n=3), ****p <0.0001. p-value in blue color represents the significance of VBIT-4-treated relative to untreated mice. NS, not significant.   . (a,b) Confocal images of hippocampal sections from WT, and untreated and VBIT-4-treated 5XFAD mice co-immunostained for p-NF-kB-p65 and TNF- (a) and their relative expression levels (b) (i) shows an enlargement of a section to show co-staining of p-NF-kB and TNF- (arrows). Results show means ± SEM (n=3), **p <0.01 ***p <0.001. p-value in blue represents the significance of VBIT-4-treated mice relative to untreated mice. NS, not significant. (c,d) Co-immunostaining for p-NF-kB-p65 with GFAP (c) or with TUBB3,b, arrows) (d), showing p-NF-kB expression in astrocytes and neurons in cortical and hippocampal sections from 5XFAD mice. Enlargement are shown in b(ii,iii), arrows point to co-staining.   S16. VBIT-4 reduces NRLP3 and caspase-1 expression in 5XFAD mice. Confocal images of cortical and hippocampal sections from WT, and untreated and VBIT-4-treated 5XFAD mice co-immunostained for caspase-1 and NRLP3 (a,b), or GFAP (c,d), or TUBB3 (e) to visualize their expression in astrocytes or/and neurons. (f) Quantitative analysis of NRLP3 and caspase-1 in the hippocampus. Results show means ± SEM (n=3), **p <0.01 ***p <0.001. p-value in blue color represents the significance of VBIT-4-treated relative to untreated mice. NS, not significant.    Cortical (a,c) and hippocampal (b,d) confocal images from WT, and untreated and VBIT-4-treated 5XFAD mice coimmunostained for GFAP and IL-4 (a,b) or TGF- (c,d).