Spermidine/spermine-N¹-acetyltransferase ablation impacts tauopathy-induced polyamine stress response

Mice

All animal procedures were performed in accordance with the Institutional Animal Care and Use Committee (IACUC) at University of South Florida Health Byrd Alzheimer’s Institute. Male and female rTg4510 mice [43] were used for western blot (aged 12 months) and polyamine quantification (aged 8 months) experiments and have been previously described [44]. For biochemical and polyamine quantification experiments group sample sizes and sex distribution were as follows: nTg: n = 7 (1M/6F); rTg4510 n = 8 (4M/4F). For biochemical analyses, a random selection of n = 5 per group were used. For immunohistochemical experiment group sample sizes and sex distribution were as follows: nTg: n = 2 (0M/2F); rTg4510 n = 2 (1M/1F). For all experiments, statistical outliers were defined as falling more than two standard deviations from the mean and were removed prior to analyses (see “Statistical analyses” section).

Male and female spermidine/spermine N¹-acetyltransferase (SSAT) null mice (SSAT-/-) on the C57BL/6 background were used for all other experiments (aged 11 months at time of surgery) and were provided as a generous gift from collaborator M. Soleimani at University of Cincinnati College of Medicine, Cincinnati, OH, USA. SSAT-/- mice comprise of a conventional disruption (neo cassette insertion in the SAT1 gene), thereby affecting all cell types that govern SSAT regulation and show altered polyamine homeostasis [45, 46]. SSAT transcript expression was confirmed by RT-PCR with primers targeting exon 1-3 (Fig. 7g). For immunohistochemical, biochemical, polyamine quantification, and behavioral experiments, group sample sizes and sex distribution were as follows: nTg AAV9 empty capsid: n = 10 (3M/7F); nTg AAV9 Tau ΔD421: n = 11 (8M/3F); SSAT-/- AAV9 empty capsid: n = 11 (7M/4F); SSAT-/- AAV9 Tau ΔD421: n = 10 (7M/3F). For biochemical analyses, a random selection of n = 5 per group were used. For electrophysiology analyses, group sample sizes and sex distribution were as follows [1]: nTg: n = 6 (6M/0F) [2]; SSAT-/-: n = 6 (6M/0F). For all experiments, statistical outliers were defined as falling more than two standard deviations from the mean and were removed prior to analyses (see “Statistical analyses” section).

Thioflavin T assay

Recombinant 4R0N (P301L) tau was purified as described [47]. For the thioflavin T (ThT) assay, 10 μM tau was combined with 10 μM ThT (Sigma-Aldrich, St. Louis, MO, USA) in 10-mM sodium phosphate, pH 7.4 buffer. Compounds at 300 μM, putrescine (Sigma-Aldrich, St. Louis, MO, USA), spermidine (Sigma-Aldrich, St. Louis, MO, USA), spermine (Sigma-Aldrich, St. Louis, MO, USA), acetylputrescine (Sigma-Aldrich, St. Louis, MO, USA), acetylspermidine (Sigma-Aldrich, St. Louis, MO, USA), acetylspermine (Sigma-Aldrich, St. Louis, MO, USA), or vehicle control were added to as indicated in a black, clear bottom 96-well plate (Nunc). Compounds and vehicle controls were kept strictly to 5% of the total sample volume. Aggregation was initiated with 4 μM heparin, less than 2% of total well volume, and samples were incubated in a BioTek Synergy H1 plate reader for 10-min readings for a 72-h period at excitation wavelength 442 nm and emission wavelength 482 nm. Each compound was assessed in triplicate wells and compared to tau + vehicle. Average Buffer background was subtracted and accounted for at each time point. Lowest tau at 0 h time point was created, as it was lowest tau relative fluorescence unit (RFU). That RFU value of tau was used and subtracted from each compound at each time point (RFU value of tau used to subtract was from the 0 h.). Hours were calculated by doing a 1/6, 2/6… 433/6 to calculate hours from 10-min readings that were taken.

Split GFP-Tau oligomerization assay

The split GFP-Tau plasmids pmGFP10C-Tau (#71433) and pmGFP11C-Tau (#71434) were purchased from Addgene and used together to visualize tau dimerization [48, 49]. These two plasmids were desired for bimolecular fluorescence complementation (BiFC) experiment based on the complementary superfolder GFP split into β-strands 1–10 (1–212 aa) and the 11th β-strand (213–228 aa). The GFP fragments were fused to the C-terminus of tau (0N4R). Both plasmids were sequenced to confirm their correct identity. Neuro-2a cells (N2a, mouse neuroblastoma, ATCC® CCL131™) were cultured in DMEM supplemented with 10% fetal bovine serum, 1% (v/v) GlutaMAX (2 mM), penicillin (100 IU/ml), streptomycin (100 μg/ml), and sodium pyruvate (1 mM) at 37 °C in a 5% CO₂ humidified incubator. For functional validation of the two split GFP-Tau plasmids, both were individually and co-transfected into N2a cells using Lipofectamine 2000 (Life Technologies) according to manufacturer’s protocols. Twenty-four hours post transfection, cells were imaged at 4× and 20× under bright field and GFP fluorescence by using Cytation™ 3 Cell Imaging Multi-Mode Reader (BioTek™). 72-hours post transfection, cells co-transfected with both plasmids were selected with zeocin (125 μg/ml, R25001, Life Technologies) and geneticin (400 μg/ml, 30234CI, Corning) in N2a complete medium for obtaining stable transfected clones. After a total of 6-month selection process, several single stable clones were picked and eventually proliferated into stable monoclonal cell lines. The novel cell line named N2a split superfolder GFP-Tau (N2a-ssGT) was established and characterized by detecting high percentage of GFP using Accuri® C6 flow cytometer (BD Biosciences) under FL1 detector (Laser 488). Thus, N2a-ssGT cells were used as a model for visualizing tau oligomerization. For split GFP-Tau oligomerization assay, N2a-ssGT cells were plated on 96-well assay plate (3603, Corning), then treated with N⁸-acetylspermidine dihydrochloride (A3658, Sigma-Aldrich) at a final concentration of 30 μM on the following day. The medium for diluting the drug was supplemented with 1 mM aminoguanidine hydrochloride (396494-25G, Sigma-Aldrich) to inhibit bovine amine oxidases activity in the fetal bovine serum. Twenty-four, 48, and 72 h post treatment, the 96-well plate was automatically read to collect GFP fluorescence arbitrary units (A.U.) and imaged simultaneously under an optimized preset protocol on the Cytation™ 3 reader.

Viral production

Truncated human tau 1-421 amino acids (tau ΔD421) were cloned into the recombinant adeno-associated virus (rAAV) vector pTR2-MCS. This vector expresses the tau with the chicken beta-actin CMV hybrid promoter (CBA) and contains the AAV2 terminal repeats. The tau ΔD421 was also appended on the N-terminus with a hemagglutinin (HA) tag for easier detection of the expressed protein. rAAV serotype 9 viral particles were generated from a triple transfection into HEK293 cells, followed by purification as described previously [50]. Viral titer was performed via dot blot as described previously [51].

Surgical procedures

At 11 months of age, SSAT null mice and SSAT sufficient littermates were anesthetized by isoflurane, and a volume of 2 μL of rAAV particles containing rAVV9-empty capsid or C-terminally truncated tau (tau ΔD421) (3 × 10¹² vg/ml) were stereotaxically injected (David Kopf Instruments, Tujunga, CA, USA) bilaterally into the CA3 of the hippocampus and the anterior cortex. Stereotaxic coordinates from bregma were AP = − 2.5 mm, ML = ± 2.9 mm, DV = − 3.0 mm, for the hippocampus, and AP = 2.2 mm, ML = ± 1.7 mm, and DV = − 3.0 for the cortex. Virus was administered using convection-enhanced delivery at a constant rate of 1.5 μL/min [52]. Viral incubation was 3 months at time of behavioral testing and 4 months at time of euthanasia and tissue collection.

Behavioral testing

Three months after viral incubation, mice performed a battery of behavioral tasks to assess alterations in cognitive performance, motor performance, and anxiety. Behavioral tasks were run in order of increasing stress to minimize stress-effects on behavioral performance. White noise (55 dB) was present during all testing.

Immunohistochemistry and Staining (IHC)

Following the completion of the one-month long behavioral testing battery, mice were euthanized with SomnaSol and transcardially perfused with 0.9% saline at 15 months of age, with a total of 4 months of viral incubation. One hemisphere of the brain was dissected, frozen, and stored at – 80 °C for biochemical analysis; the opposite hemisphere was fixed in 4% paraformaldehyde in 100 mM phosphate buffer (pH 7.4) for 24 h. Tissue was cryoprotected by sequential immersion in 10%, 20%, and 30% sucrose solutions. Brains were sectioned at 25 μm using a sliding microtome and stored in 4 °C in Dulbecco’s phosphate-buffered saline (DPBS) containing 100 mM sodium azide until staining.

Immunohistochemistry was performed on free-floating sections (6–8 per mouse) as previously described [54]. Antibodies were as follows: rabbit anti-Iba1 (1:3000; Wako Chemicals USA, Inc.), mouse anti-NeuN biotin (1:30,000; EMD Millipore), mouse anti-HT7 biotin (1:5000, Thermo Fisher Scientific), rabbit anti-Tau pS199/202 (1:5000; AnaSpec), and mouse paired helical filament tau biotin (AT8, 1:5000, Thermo Fisher Scientific).

Biochemical analysis/western blotting

For the in vivo samples, dissected hippocampal tissue was used for western blot (WB) analysis and prepared as previously described [44] or using an AllPrep DNA/RNA/Protein Mini Kit (Qiagen, Germantown, MD), according to the manufacturer’s protocol. The detergent-soluble fraction (S1/SF) was subjected to western blot analysis. Protein concentration was determined with the Pierce BCA protein assay. Protein was loaded onto a 4–20% gradient gel (20- to 26-well Midi SureLock System; Life Technologies). Antibodies were as follows: mouse anti-Tau C3 (tau ΔD421)(1:1000; Sigma Aldrich), mouse anti-HT7 (1:1000, Thermo Fisher Scientific), mouse anti-Tau-5 (1:1000, Thermo Fisher Scientific), rabbit pS199/202 (1:1000; AnaSpec), mouse anti-Tau PHF (AT8; 1:1000; Thermo Fisher Scientific), rabbit anti-Tau pS396 (1:1000; AnaSpec), rabbit anti-ODC (1:1000; Epitomics), rabbit anti-OAZ1 (1:1000; Pierce), rabbit anti-SRM (1:1000; Epitomics), rabbit anti-SMS (1:1000; Proteintech), rabbit anti-SMOX (1:1000; Proteintech Group), rabbit anti-PMF1 (1:1000; Proteintech Group), rabbit anti-PMFBP1 (1:1000; Proteintech Group), and mouse anti-GAPDH (1:800,000; Meridian).

Polyamine quantification

Polyamine quantification in brain homogenate was performed in collaboration with Sanford Burnham Prebys (Orlando, FL, USA) by liquid chromatography tandem mass spectrometry (LC-MS/MS) using a standard curve of each polyamine and acetylated polyamine (Sigma Aldrich).

qRT-PCR

Total RNA from each mouse hippocampus was extracted using Qiagen’s All Prep DNA/RNA/Protein kit and used for RT-qPCR. A standard curve spanning two logs of a dynamic range was generated from a pool of naïve mice and experimental groups. Fifty nanograms of RNA from each mouse were used to generate cDNA using Superscript II First Strand Synthesis System (Life Technologies). Integrated DNA Technologies (IDT) PrimeTime® qPCR gene primers were used (SAT1: Mm.PT.58.45831345; GAPDH: Mm.PT.39a.1) with SYBR green-based real-time qPCR (Sigma-Aldrich). Quantitative PCR reaction proceeded by the following protocol: incubation at 95 °C/5 min, followed by 39 cycles of incubation at 95 °C/30 s, amplification at 60 °C/1 min, and a plate read. Primers were validated after each cycle by performing a dissociation curve analysis. The DNA Engine Opticon 2 TM Real-Time PCR System (Version 4.3; Bio-Rad) was used to detect the amplicon.

Electrophysiology

Hippocampal slice electrophysiology was performed on 7-month-old SSAT-/- and non-transgenic littermates as previously described [54–56]. Briefly, baseline stimulus intensity was set at the level that elicited 40–50% of the maximum fEPSP response as determined from the input-output curve. Input-output relationships were determined by stimulating slices from 1–15 at 0.5 mV increments. Short-term plasticity was measured via paired-pulse facilitation, which was induced by stimulating slices at half-max intensity with sequential pulses spaced at 20 ms intervals from 20–300 ms. Long-term potentiation (LTP) was induced by a theta-burst protocol, which consisted of five bursts at 200 Hz separated by 200 ms, repeated six times with an inter-train interval of 10 s as previously described [56–58].

siRNA knockdown of SAT1 in C3 HeLa cell culture

HeLa cells stably overexpressing 4R0N wild-type human tau (C3 HeLa) were used to model tau biology. Cells were maintained in 75 cm² flasks at 37 °C in a 5% CO₂ humidified chamber in OPTIMEM containing 10% fetal bovine serum, 100 U/ml penicillin and 100 ug/ml streptomycin. C3 HeLa cells were transfected with scramble negative siRNA control or SAT1 (FlexiTube siRNA; negative control siRNA, Hs_SAT_5 FlexiTube siRNA, Qiagen) using DhamaFECT1 (Horizon, Dharmacon) with or without 10 uM N1, N11-diethylnorspermine (DENSPM/DSPM; Tocris Cat#0468 batch NO. 4, Minneapolis, MN) (used to induce SAT1) for 72 h. Cells were harvested 72 h post transfection and whole cell pellets were lysed using M-PER (Thermo Scientific). Protein concentration was determined using Pierce BCA protein Assay (Thermo Scientific) and subjected to western blot analysis. Antibodies were as follows: rabbit anti-Tau pSer356 (1:1000; AnaSpec), rabbit anti-Tau pSer396 (1:1000, AnaSpec), mouse total tau (H150 (discontinuted); 1:1000; SantaCruz), rabbit anti-SAT1 (SSAT; 1:1000, Abcam), and mouse anti-GAPDH (1:800,000; Meridian).

Statistical analyses

All statistical analyses were performed using SPSS Version 24. An independent samples t test was used for all non-transgenic (nTg) and rTg4510 comparisons, with a Levene’s test for equality of variances correction when appropriate. Data is represented by means ± S.E.M. *p < .05. Area under the curve (AUC) was calculated using GraphPad Prism, and data is represented as averages of triplicate wells for thioflavin T assays. A 2 × 3 factorial analysis of variance (ANOVA) (treatment × time) was used to determine the simple main effects of drug (vehicle vs 30 μM acetylspermidine) and time (24 h vs 48 h vs 72 h) as well as interaction of variables, followed by post hoc comparisons to further investigate specific effects of time within each drug on tau oligomerization. A 2 × 2 factorial analysis of variance (ANOVA) (genotype × viral treatment) was used to determine simple main effects of genotype (nTg vs SSAT-/-) and treatment (AAV9 empty capsid vs AAV9 Tau ΔD421) as well as interaction of variables, followed by pair-wise comparisons to further investigate specific effects of treatment within each genotype. A repeated measures mixed ANOVA was performed to determine simple main effects of genotype (nTg vs SSAT-/-) and treatment (AAV9 empty capsid vs AAV9 Tau ΔD421) as well as interaction of variables on performance across blocks in the radial arm water maze and trials in the rotarod task. A nonlinear regression was performed, and the line fits were compared using the extra sum-of-squares F test for electrophysiology experiments. A one-way analysis of variance (ANOVA) followed by Fisher’s PLSD post hoc comparisons was used to determine the effect of treatment in the siRNA cell culture experiments. For all analyses, statistical outliers were defined as falling more than two standard deviations from the mean. Data is represented by means ± S.E.M., p < .05; asterisks indicate the main effect of treatment and main effect of genotype, ampersands indicate the interaction of genotype and treatment, and number signs indicate the pairwise comparison within genotype.

Tau rTg4510 mice show significantly disrupted polyamine homeostasis

Tau rTg4510 harbor the MAPT P301L mutation on a tetracycline response element regulated by the calcium/calmodulin-dependent protein kinase type II alpha chain (CaMKII-α) promoter. Mice deposit tau, phospho-tau (Fig. 1b, c), and tangles (not shown) in the forebrain and hippocampus but also develop many components of tauopathies including inflammation and atrophy [43, 44, 59]. One prominent post translational modification that incites further tau deposition and neuropathology includes cleavage by caspase-3 at aspartate 421 (tau ΔD421) which associates with Pick’s disease (PiD) [15], corticobasal degeneration (CBD) [60], progressive supranuclear palsy (PSP) [14], and AD [9–12]. Further, tau ΔD421 has been shown to increase tau spreading [61] and produce cognitive impairment [62].

We show tau ΔD421 in rTg4510 mice (Fig. 1d–h), which was associated with polyamine dysregulation at the level of enzymatic and protein control (Fig. 1i, j). Specifically, pro-synthetic enzymes ornithine decarboxylase (ODC), and spermine synthase (SMS) significantly decreased, whereas catabolic enzymes, spermidine/spermine-N¹-acetyltransferase (SSAT) and spermine oxidase (SMOX) significantly increased. Spermidine synthase (SRM) also significantly increased, possibly as a compensatory mechanism in response to reduced ODC. Furthermore, polyamine-modulated factor1 (PMF1; a transcription factor that binds NF-E2 released factor 2 (Nrf-2) to increase SSAT expression [63]) also moderately increased (Fig. 1i, j), relative to non-transgenic controls. Importantly, the reduction in ODC and increase in SSAT were most notable, as they alone are thought to be the key regulatory enzymes to control polyamine homeostasis in the brain and provide evidence for a unique tau-PSR [30]. Although the role of polyamine-modulated factor 1 binding protein 1 (PMFBP1) is unknown, expression significantly decreased in tau transgenic mice. These data indicate that the tau phenotype renders dysregulation of anabolic and catabolic enzymes throughout the polyamine axis.

To determine the extent of end-product polyamine levels, we measured polyamines and their acetylated products in rTg4510 mice (Fig. 2a, b). Although many of the polyamine enzyme levels were altered in some way, only putrescine and acetylspermidine were significantly increased in the brains of rTg4510 mice (Fig. 2a, b), compared to non-transgenic littermates. An increase in both products indicates evidence of the SSAT-induced polyamine back conversion revealing a signature of the tau-PSR. Notably, increased putrescine in rTg4510 does not account from increased metabolism of ornithine (the precursor substrate for putrescine), because ODC decreased in tau transgenic mice. This suggests sustained SSAT-dependent polyamine back conversion as one potential mechanism for the tau-PSR. To date, acetylation of spermidine (and other polyamines) has only been shown to occur under the control of SSAT, signifying the impact of SSAT dysregulation.

Polyamines and acetylated polyamines differentially impact tau fibrillization

To determine whether polyamines and acetylated polyamines directly impact tau fibrillization, we co-incubated recombinant 4R0N (P301L) tau with polyamines and acetylated polyamines (300 μM) or vehicle control in solution using the ThT assay (Fig. 3a–c). Polyamines prevented tau fibrillization, as measured by area under the curve (AUC), with higher-order polyamines (longer chain length) inhibiting fibrillization to greater extent and with greater potency. Strikingly, both spermidine and spermine appear capable of completely blocking tau fibrillization in this system. Interestingly, acetylated polyamines displayed a different profile for tau fibrillization. Specifically, acetylspermidine failed to inhibit fibrillization of tau and actually increased fibrillization. In addition, acetylputrescine and acetylspermine failed to reduce tau fibrillization to the same extent as putrescine and spermine respectively, suggesting different impacts of polyamines versus acetylated polyamines on tau biology.

Acetylspermidine increases tau oligomerization

To further determine whether increases in acetylspermidine could be contributing to disease progression, we developed a monoclonal superfolder split GFP-tau cell line (murine neuroblastoma; N2a-ssGT) to measure tau oligomerization. Characterization and validation of individual plasmids were performed. Subsequently, characterization and validation of stably expressed tau in a monoclonal N2a-ssGT cell line was conducted by measurement of GFP fluorescence by Cytation™ 3 Cell Imaging Multi-Mode Reader (BioTek™) and Accuri® C6 flow cytometer (BD Biosciences) (Fig. 3d, e, respectively). N2a-ssGT cells were incubated with acetylspermidine for 24, 48, and 72 h, and tau oligomerization was measured by GFP fluorescence. In this model, the SSAT-dependent byproduct acetylspermidine increased tau oligomerization (Fig. 3f, g). Although it is difficult to predict the various endogenous concentrations of polyamines and acetylated polyamines within subcellular compartments and the impact on tau biology, these data signify a detrimental role for accumulation of acetylated products, namely acetylspermidine, on tau fate.

AAV9 Tau ΔD421 overexpression induces a tau phenotype and tau-PSR: SSAT disruption prevents the tau-PSR and partially reduces phospho-tau but produces an independent behavioral phenotype

To determine the impact of SSAT disruption on the tau-PSR, we overexpressed tau ΔD421 (AAV9 Tau ΔD421) in SSAT null mice (SSAT-/-). This post-translational modification was chosen based on its observed expression in rTg4510 mice (Fig. 1d–h) as well as its ability to enhance the spreading of tau pathology [61], assemble more rapidly and more extensively into tau filaments than wild-type tau in vitro [10], and produce cognitive impairment [62].

Immunohistochemical analyses revealed AAV9 Tau ΔD421 increased hippocampal, but not cortical, microglial activation (Iba1); however, SSAT disruption did not impact any of these measurements (Fig. 4a–d). Interestingly, AAV9 Tau ΔD421 decreased hippocampal NeuN expression in both genotypes compared to AAV9 empty capsid; however, SSAT disruption significantly increased hippocampal and cortical NeuN expression (Fig. 4a–d).

Relative to injection sites, AAV9 Tau ΔD421 increased total tau (HT7) in the CA3 of the hippocampus and anterior cortex (Fig. 4e–h) and importantly was not significantly different between nTg and SSAT-/- mice, ensuring equal tau expression across genotypes. Examining phospho-tau epitopes within these regions, analyses revealed that targeted SSAT disruption reduced both hippocampal and cortical tau pSer199/202, but not PHF tau (AT8) (Fig. 4e–h), indicating a discrete reduction in tau phospho epitopes by targeting the tau-PSR through SSAT.

Western blot analyses confirmed the level of tau ΔD421 (Fig. 5a, b) and total tau (HT7; Fig. 5a, c) were not significantly different between nTg and SSAT-/- mice, confirming that human tau expression was equal across genotypes. Further, analyses revealed SSAT disruption reduced the aggregation of monomeric and high-molecular weight phospho-tau ser396 (tau pSer396) (Fig. 5a, h, i), suggesting that SSAT disruption discretely impacts the level of phospho-tau species and certain forms of tau associated with the Tau-5 epitope (Fig. 5a, d, h, i).

Furthermore, AAV9 Tau ΔD421 produced a unique tau-PSR (Fig. 6a–h). Similar to that of rTg4510 mice, AAV9 Tau ΔD421 decreased polyamine anabolic enzymes ODC and SMS (Fig. 6a, e). One enzyme capable of binding ODC to render its degradation by the proteasome is ornithine decarboxylase antizyme1 (OAZ1 or AZ). Although the mean average was higher in AAV9 Tau ΔD421 mice, OAZ1 failed to reach statistical significance (Fig. 6c). Additionally, SRM and SMOX show no change between groups (Fig. 6d, f), suggesting that tau induces a response specific to ODC and SMS. Interestingly, PMF1 significantly increased in SSAT-/- mice (Fig. 6g), possibly as a compensatory mechanism for SSAT deficiency, whereas PMFPBP1 was only significantly induced in SSAT-/- mice that received AAV9 Tau ΔD421 (Fig. 6h) suggesting a unique interaction of tau pathology and SSAT regulation. These interactions between the tau phenotype and enzymatic signatures warrant further investigation.

Regarding polyamines and their metabolites, SSAT-/- mice in general displayed reduced putrescine, acetylputrescine, spermidine, and acetylspermidine (Fig. 7a–d; Additional file 1: Figure S1 a-g). Importantly, AAV9 Tau ΔD421 increased acetylspermidine and showed a trend toward increased putrescine whereas SSAT deficiency prevented these effects (Fig. 7a, d). Conversely, spermine showed no change between groups whereas acetylspermine was slightly but still significantly elevated (Fig. 7e–f). These data suggest that SSAT disruption does not completely eliminate all acetylated products but primarily impacts polyamine retro-conversion, recycling, and homeostasis of polyamine levels. It also suggests that unknown and alternative acetylation processes may exist, specifically in regard to spermine. Furthermore, tau seemingly promotes SSAT-induced back-conversion that can be measured at the level of acetylated polyamines, namely acetylspermidine (Fig. 7d). Of note, spermidine shows the largest abundance of all three polyamines in the brain and likely outcompetes for SSAT as a primary substrate; therefore, acetylated spermidine increases in response to SSAT induction elicited by tau. Curiously, while qRT-PCR confirmed SAT1 disruption in SSAT-/- mice, AAV9 Tau ΔD421 did not significantly increase SAT1 expression (Fig. 7g), as the elevations in back-converted putrescine and acetylspermidine would predict (Fig. 7a, d).

AAV9 Tau ΔD421 also produced cognitive impairment in the radial arm water maze task (Fig. 8a); however, despite SSAT disruption’s ability to partially reduce specific phospho-tau epitopes (Fig. 4e–h, Fig. 5a, d, h, i) and prevent the tau-induced increases in putrescine and acetylspermidine (Fig. 7a, d), cognitive impairment induced by AAV9 Tau ΔD421 was not improved by SSAT disruption (Fig. 8a). Interestingly, a behavioral phenotype emerged as a product of SSAT disruption, characterized by increased hypervigilance in the rotarod task (Fig. 8b), decreased inhibition in the elevated plus maze (Fig. 8f, g), and increased compulsion in the marble burying task (Fig. 8h), compared to SSAT sufficient littermates, which may have influenced performance in the RAWM task. These data are highly intriguing given the role of SSAT in affective disorders [64–69]. Interestingly, AAV9 Tau ΔD421 significantly decreased time in the center zone of the open field task, but only in SSAT-/- mice (Fig. 8e), suggesting a modest interaction of SSAT disruption and AAV9 Tau ΔD421. To confirm that SSAT disruption alone did not influence learning and memory, we performed electrophysiology on hippocampal slices. The behavioral phenotype was not associated with any deficiencies in synaptic transmission or pre-synaptic short-term plasticity (Fig. 8i–k), solidifying behavioral changes in SSAT null mice as being rooted in affect rather than cognitive processing.

siRNA SAT1 repression reduces tau aggregation

Lastly, in an effort to better understand the impact of SSAT disruption on tau biology in vivo, we repressed SSAT using siRNA in C3 HeLa cells stably overexpressing tau (4R0N). Here, we show that SSAT repression also decreased monomeric total and phospho-tau levels compared to scrambled siRNA controls (Fig. 9a–d), further suggesting that simply modifying SSAT levels impacts tau biology.