Blocking STAU1 may stop p53-related cell death in brain diseases

Beginning with transcriptomic analysis in vitro, we examined the p53 pathway directly under conditions of STAU depletion in non-neuronal cells. This was followed by functional analysis in human neurons and finally in vivo using genetic interaction of ALS and STAU1-knockout (KO) mice.

Interrogation of canonical pathways using the Ingenuity Pathway Analysis (IPA) software revealed 241 significant pathways (with p ≤ 0.05) (Fig. 1C for the top 10, Supplemental Table 1 for the complete list). In our previous studies we described how STAU1 KD decreases the activity of eIF2 and mTOR, as well as their corresponding downstream targets [17 –19]. In agreement with this, the top 3 IPA canonical pathways in the current analysis were eIF2 signaling, regulation of eIF4 and p70S6K signaling and mTOR signaling (Fig. 1C). IPA analysis of the DEGs in the context of their pathway assigns a positive z-score to the pathways that are predicted to be activated and a negative z-score to pathways predicted to be inhibited [27]. Out of 241 significant canonical pathways, IPA assigned significant z-scores (z-score ≥ 2 or ≤ -2) [27] to 26 pathways (Fig. 1D). The top 3 canonical pathways eIF2 signaling, regulation of eIF4 and p70S6K signaling and mTOR signaling had significant or strongly trending negative z-scores, mirroring our previous findings and indicating an accurate IPA prediction of their functional inhibition (−5.859, −1.886 and −1.732 correspondingly, Fig. 1D, Fig. S3, SI Table 1 and 2) [17 –19].

IPA Upstream Regulator analysis identifies and predicts a positive or negative activation state for upstream transcription factors, microRNAs, kinases, compounds and drugs based on prior knowledge of their effects on the target genes in our dataset. Upstream analysis predicted an inhibition of the effects of multiple cytotoxic drugs with antineoplastic activity (camptothecin, prexasertib, etoposide and topotecan), along with a decrease in immune responses (TCR, Salmonella enterica LPS, poly rl:rc-rna, IFN-γ, diclofenac) (Fig. 2C). Several activated upstream regulators further support attenuation of cell death and immune responses by STAU1 KD (ESR1, Rictor, sirolimus), however some can promote cell death and positively modulate immune responses, depending on the context (CD24, PD98059, Mknk1, Tgfbr2, DSCAM) (Fig. 2C). Activity plots for predicted responses to etoposide and camptothecin show the z-scores for these compounds in the context of >47,000 datasets from the IPA database, indicating STAU1 KD has a profound effect (Fig. 2D).

The Upstream analysis of transcription and translation factors shows the top five factors with most significant p-values are MYC, MYCN and MLXIPL (belonging to the Myc/Max/Mad superfamily) and TP53 and LARP1(Fig. 2E, F, Fig. S4). All these factors can be regulated by mTORC1, and STAU1 binds to the mTOR 5'-UTR and controls its abundance and activity [19]. In addition, multiple other factors with significant p-values and z-scores were identified as upstream regulators (Fig. 2F, SI Table 2).

Altogether, STAU1 KD cells demonstrated extensive transcriptomic changes that suggest a cellular state of suppressed immune and apoptotic responses.

Because the transcriptomic analysis indicated STAU1 KD could result in apoptotic resistance, we tested this hypothesis by treating cells with stressors that affect different cellular processes but converge at the activation of the intrinsic apoptosis pathway. We quantified cleaved Caspase-3 (cCaspase 3) and cleaved PARP (cPARP), as molecular effectors and markers of apoptosis [28, 29]. We utilized siRNA to KD STAU1 in HEK293 cells and an AAV coding for a miRNA targeting STAU1 in SH-SY5Y cells (AAV-PhP.eB-miSTAU1-45, due to low transfection efficiency), and additionally tested primary fibroblasts derived from Stau1-KO mice.

We also carried out a Comet assay, which detects DNA damage at the single cell level. Due to the assay's high sensitivity, our cells' baseline measurements after transfection left us with a narrow dynamic range and we were not able to meaningfully quantify DNA damage changes with this technique (Fig.). S6

To further support our findings, we studied primary skin fibroblast cultures from WT and Stau1 KO mice [33]. In WT fibroblasts, etoposide caused p53 stabilization and increased PUMA, associated with increased cCaspase 3, whereas in Stau1^−/− fibroblasts their levels were indistinguishable from untreated cells (Fig. 4C).

In contrast with SH-SY5Y cells and fibroblasts, siSTAU1 did not reduce p53 activation in neurons; etoposide and Nutlin-3 induced p53 to the same level or higher (correspondingly) in control and STAU1 KD neurons. At the same time, induction of p21 in response to etoposide and Nutlin-3 was higher in the STAU1 KD neurons. The protein p21 is a key target of p53, and its activation promotes cell cycle arrest, DNA repair and inhibition of apoptosis [37 –39]. In addition to iNeurons, we studied primary mouse cortical neurons and found they mirrored the iNeuron results, with an increase in p21 after STAU1 KD (Fig. 5B). These results demonstrate that STAU1 KD can prevent DNA damage and p53-dependent apoptosis by harnessing the signaling pathway in a cell type and context-dependent manner.

To investigate whether STAU1 can modulate the p53 pathway to prevent apoptosis in vivo we utilized the C9orf72-500 mouse model of ALS/FTD [43]. We found that in spinal cords, STAU1, PUMA and cCaspase-3 were elevated (Fig. 6B). When crossed with Stau1^−/− mice to produce C9orf72-500 mice haploinsufficient for Stau1, both PUMA and cCaspase-3 were reduced to levels comparable to WT mice (Fig. 6B). Thus, genetic reduction of Stau1 was sufficient to prevent p53-mediated apoptotic signaling in a mouse model of neurodegeneration caused by C9orf72-exp. Altogether, our results indicate reducing STAU1 abundance can effectively prevent p53 apoptotic signaling to protect from apoptosis in human and mouse cellular models and in vivo.

Cell culture media and reagents were purchased from Thermo Fisher Scientific unless otherwise specified. HEK293 cells, mouse skin fibroblasts and human skin fibroblasts were grown in DMEM supplemented with 10% fetal bovine serum and 1X penicillin and streptomycin.

Mouse primary fibroblasts were obtained by culturing ~3 mm skin explants attached to a cell culture dish for 2–3 weeks. Explants were then removed and fibroblasts were passaged 2–3 times before experimentation.

Human fibroblasts ND38530 (normal) and ND42506 (GGGGCC repeat expansion>24; in 1st intron – individual at risk for FTD) were obtained from the Coriell Cell Repositories (Camden, NJ, USA).

Identity authentication of HEK293, SH-SY5Y cells and human fibroblasts was carried out by short tandem repeat (STR) analysis with the GenePrint 24 System (Promega, USA) and PCR mycoplasma testing was carried out regularly.

Lipofectamine 2000 was used for all siRNA and plasmid transfections, according to the manufacturer's instructions. STAU1 siRNA was used at 100 μM, and subsequent pharmacological treatments were added 48 h after transfection in fresh media.

Mice were bred and handled according to the protocol for ethical use of animals for research approved by the University of Utah IACUC (protocol number 00002040). Primary cortical neuron cultures were prepared from neonatal mice [70, 71]. Brain cortices from 6–7 animals were isolated and incubated with 50 units of papain (Worthington) in Earle's balanced salt solution (EBSS) with 1mM L-cysteine and 0.5 mM EDTA for 15 min at 37 °C, followed by washing in EBSS and mechanical trituration. The remaining tissue was removed by filtration through a 50 µm strainer (Falcon). Neurons were seeded at 50,000 per cm² on plates coated with poly-L-ornithine and laminin in Neurobasal Plus medium containing 2% B27 Plus supplement (Life technologies) and glutamax 0.5 mM. On day 2 10 µM cytosine arabinoside was added for 24 h to prevent proliferation of glial cells and 75% of culture medium volume was replenished every 2–3 days from there on. After 2 weeks in culture, we added ASO against STAU1 or a non-targeting control ASO for overnight gymnotic delivery. Neurons were treated 48 h later.

The iPSC line KOLF2.1 J AAVS-TO-NGN2 was a gift from Michael Ward (NIH, Maryland) and Bill Skarnes (Jackson laboratories, Connecticut) as part of the iPSC Neurodegenerative Disease Initiative (iNDI) and has been previously validated [72]. iPSCs were cultured on Geltrex substrate with StemFlex cell culture media. For differentiation, iPSCs were passaged using Accutase (Sigma) and plated at high density on Geltrex substate. For the first 3 days cells were grown in Induction Media composed of DMEM/F12 supplemented with 1% N2, 1% NEAA and 1% GlutaMax, with the addition of 2 μg/mL doxycycline (Sigma), with daily media changes. On day 3 cells were dissociated with Accutase and plated on Geltrex substrate in 12 well plates at a density of 5 × 10⁴ cells per well. The next day, media was changed to Maturation Media, composed of Neurobasal Plus supplemented with 1% B27 Plus, 1% N2, 10 ng/mL BDNF (StemCell) and 2 μg/mL doxycycline. Doxycycline was discontinued on day 8. 50% of the media was replaced with fresh media 2-3 times per week. iN from at least two different batches of differentiation were used for experiments. iN were transfected between days 15 and 20 after initial induction with Lipofectamine RNAiMAX and 400 nM siSTAU1 or siControl, overnight, for 2 consecutive days, and experiments were carried out 48 h after the last transfection.

For immunofluorescence analysis iNeurons were plated on coverslips coated with geltrex on day 3 of differentiation, transfected on days 4 and 5 and on day 7 treated with etoposide for 2 h. iNeurons were pre-fixed by adding 4% paraformaldehyde into the cell culture media and incubating for 15 min, then fixed for further 15 minutes by replacing the entire volume of media with paraformaldehyde 4%. iNeurons were permeabilized and blocked for 1 h with 5% goat serum and 0.1 Triton X-100 in PBS, then incubated with γH2AX antibody at a 1:500 dilution overnight. The secondary antibody was an Alexa Fluor Plus 594 used at a 1:1000 dilution for 1 h. Images were acquired with a Nikon widefield microscope or a Leica SP8 at the Cell imaging core of the University of Utah.

The plasmids used to generate adeno associate virus (AAV) include the single-strand AAV expression plasmid, pAAV-U6-sgRNA-CMV-GFP, (gift from Hetian Lei; Addgene plasmid # 85451), and pUCmini-iCAP-PHP.eB, (gift from Viviana Gradinaru; Addgene, plasmid # 103005), and pHelper (Stratagene, La Jolla, CA, USA). The STAU1 miRNA DNA insert "CGAGTGAGCGCTAACTGCCATGATAGCCCGAGCTGTAAAGCCACAGATGGGCTCGGGCTATCATGGCAGTTACCGCCTACTATTTTTTA" was cloned downstream of the U6 promoter of sgRNA pre-depleted AAV-U6-CMV-GFP plasmid using Sac I and Spe I restriction sites and designated as pAAV-U6-iSTAU1-45F(h + m)-CMV-GFP. The plasmid constructs were verified by sequencing. Recombinant AAV particles were generated as previously described [73] with minor modifications. HEK293T cells were co-transfected with three plasmids: either pAAV-U6-CMV-GFP (control) or pAAV-U6-iSTAU1-45F(h + m)-CMV-GFP along with pUCmini-iCAP-PHP.eB and pHelper, at a ratio of 1:4:2 (based on micrograms of DNA; ~ 30 μg of total DNA per 10 cm dish) using lipofectamine 2000 transfection reagent (ThermoFisher Scientific) according to the manufacturer's protocol. After 12–18 h media was changed and incubated for further 72 h. Cells were then harvested and lysed in 25 mM Tris, pH 8.5, 150 mM NaCl with repeated freezing and thawing cycles using liquid nitrogen and a 37 °C water bath, followed by centrifugation at 4 °C for 30 min at 14,000 rpm. The resultant supernatants containing AAV-PhP.eB-Control or AAV-PhP.eB-miSTAU1-45 were filtered and stored at −80 C. SH-SY5Y cells were incubated overnight with these AAVs in a 1:4 dilution with fresh media, and experiments were conducted 48 h later.

For quantification of cell viability, we used the Cell-Titer Fluor Viability Assay (Promega), a glycyl-phenylalanyl-aminofluorocoumarin based-assay. SH-SY5Y cells were plated in black 96 well plates with optical bottoms, and the assay was carried out according to the manufacturer's protocol.

All mice were housed and bred in standard vivarium conditions and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Utah. The Stau1^{tm1Apa(−/−)} (Stau1^−/−) mouse was a generous gift from Prof. Michael A. Kiebler, Ludwig Maximilian University of Munich, Germany and genotyped according to published protocols [33]. Stau1^−/− mice were maintained in a C57BL/6 J background.

FVB/NJ-Tg(C9orf72)500Lpwr/J mice [43] were purchased from JAX Laboratories and backcrossed to C57BL/6 J. C9-500 mice were first genotyped with a multiplex PCR amplifying mouse actin with JAX primers 21238 F and 21239 R, and human C9orf72 primers amplifying across exon 10 (C9Ex10F) and exon 11 (C9Ex11R) producing a 415 bp and 320 bp amplicon, respectively. PCR was performed with an initial denaturation at 95 °C for 5', 30 cycles of 95 °C for 1', 60 °C for 30" and 72 °C 1' with a final extension of 72 °C for 5'. The amplicon was run on a 3% agarose gel and detected with ethidium bromide. The G4C2 repeat expansion was then verified utilizing previously published Repeat Prime PCR protocol [74]. Only animals verified to be non-mosaic and with a repeat expansion greater than 200 bp were utilized. C9-500 mice were crossed with Stau1^−/− mouse to generate C9-500/Tg Stau1^(+/−), C9-500/Wt Stau1^(+/−), C9-500/Tg Stau1^(+/+) and C9-500/Wt Stau1^(+/+) in a mixed background of FVB/NJ and C57BL/6 J.

Total RNA was extracted from tissues using the RNeasy Mini-Kit (Qiagen) according to the manufacturer's protocol. RNA quality was determined using the Agilent ScreenTape Assay. Library preparation was performed using the Illumina TruSeq Stranded Total RNA library prep Ribo-Zero gold. Paired-end 150 bp reads were generated on a Novaseq 6000 S2 cell sequencing instrument at the High-Throughput Genomics and Bioinformatic Analysis Shared Resource at Huntsman Cancer Institute (University of Utah). The human GRCh38 genome and gene annotation files were downloaded from Ensembl release 100 and a reference database was created using STAR version 2.7.3a with splice junctions optimized for 150 base pair reads [75]. Optical duplicates were removed from the paired end FASTQ files using clumpify v38.34 [76] and reads were trimmed of adapters using cutadapt 1.16 [77]. The trimmed reads were aligned to the reference database using STAR in two pass mode to output a BAM file sorted by coordinates. Mapped reads were assigned to annotated genes using featureCounts version 1.6.3 [78]. The output files from cutadapt, FastQC, FastQ Screen, Picard CollectRnaSeqMetrics, STAR and featureCounts were summarized using MultiQC to check for any sample outliers [79]. Differentially expressed genes were identified for siSTAU1 vs siControl using a 5% false discovery rate with DESeq2 version 1.26.0 [80]. Pathways were analyzed using the fast gene set enrichment package with a 10% false discovery rate [81]. For IPA, we used a 5% false discovery rate and a log₂FC cutoff of -0.1 and 0.1.

Protein homogenates from cultured cells were prepared by scraping cells in phosphate buffered saline and lysing the pellets in Laemmli sample buffer (Bio-Rad), followed by boiling for 5 min [82]. HEK293 protein extracts for detection of p53 were diluted 1:10. Mouse spinal cords were hydraulically ejected and protein extracts were prepared by homogenization in extraction buffer (25 mM Tris-HCl pH 7.6, 300 mM NaCl, 0.5% Nonidet P-40, 2 mM EDTA, 2 mM MgCl2, 0.5 M urea and protease inhibitors) followed by centrifugation at 4 °C for 20 min at 14,000 RPM. Supernatants were then diluted in laemmli sample buffer and boiled.

Proteins were resolved by SDS-PAGE and transferred to Hybond P membrane (Amersham Bioscience), blocked in Tris-buffered saline 0.1% Tween-20 with 5% skim milk and primary antibody was incubated overnight in this same solution. After incubation with the corresponding secondary antibody signal was detected using Immobilon Western Chemiluminescent HRP Substrate (EMD Millipore) or SuperSignal™ West Pico PLUS Chemiluminescent Substrate (ThermoFisher Scientific) and photographed with a Bio-Rad ChemiDoc. Analysis and quantification were performed with Image Lab software (Bio-Rad). Relative protein abundance was first normalized against actin band intensity and then expressed as the ratio to the normalized control.

We used the CometAssay Single Cell Gel Electrophoresis Assay kit from BioTechne, according to the manufacturer's instructions. The percentage of DNA in tail was quantified in ImageJ with the OpenComet plug-in [83].

All results are presented as mean ± standard error of the mean (SEM) unless noted otherwise. For western blot quantifications 3 independent experiments were carried out and the means for each group were compared with an Ordinary one-way ANOVA followed by Sidak's multiple comparison test in the GraphPad Prism software.

All mice were housed and bred in standard vivarium conditions and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Utah under protocol number 16-09007.

The KOLF2.1 J AAVS-TO-NGN2 cell line was obtained from Dr. Michael Ward and Dr. William Skarnes. The parental KOLF2.1 J iPSC line was produced according to all regulations and informed consent of the NRES Committee Yorkshire & The Humber Leeds West, approval number 15/YH/0391.