All experiments and procedures were approved by the University of Tasmania Animal Ethics Committee (A12780 and A15121) and were in accordance with the Australian Guidelines for the Care and Use of Animals for Scientific Purposes. C57Bl/6 mice were housed in optiMICE cages on a 12-h light/dark cycle with free access to food and water.

Mouse cortical neuron cultures were prepared as previously described (King et al., 2013). Cortical neurons were prepared from embryonic day 15.5 (E15.5) C57Bl/6 mice. Females were sacrificed using CO₂ and embryos were removed and decapitated. Heads were kept on ice during dissection to prevent tissue degradation. Cortices were removed using a dissecting microscope (Leica), dissociated from the meningeal layers and transferred to 5 ml Hanks Balanced Salt Solution (HBSS, Gibco). Trypsin (0.0125%) was added for 4 min at 37°C, followed by removal of HBSS and replaced with 1 ml of initial plating media [Neurobasal media (Gibco) containing 2% B27 supplement, 0.5 mM glutamine, 25 μM glutamate, 10% fetal calf serum and 1% antibiotic/antimiotic] and mechanically triturated with a 1 ml pipette. Cell density was assessed using a trypan blue exclusion assay and 200,000 cells were plated either into the soma compartment of microfluidic chambers (450 μm long, 10 μm wide, and 3 μm high barrier grooves, Xona Microfluidics), which allows fluidic isolation of cells from the axon (Taylor et al., 2006), or directly into the wells of poly-L-lysine coated culture plates. Cells were allowed to adhere to the coverslip (Livingstone) at 37°C for 30 min prior to adding initial plating media to the chambers and were incubated at 37°C overnight before initial media was replacement with subsequent plating media with no fetal calf serum and glutamate. Cells were grown to 10 days in vitro (DIV) at 37°C in 5% CO₂.

Cells were treated with 0, 10, or 25 μM of kainic acid (Sigma, lot #SLBD1491V) in DMSO (Sigma, Lot # RNBF1056) at 10 DIV, a timepoint which corresponds to dense axonal growth in the axon compartment of microfluidic chambers (Hosie et al., 2012; Millet and Gillette, 2012). Cells in 12-well plates or the soma compartment of microfluidic chambers were exposed to kainic acid for 1, 6, and 18 h. For a set of experiments cells were treated with trichostatin A, an inhibitor of the deacetylating enzyme HDAC6, which promotes microtubule acetylation. Cells at 10 DIV in 12-well plates or axonal compartment of microfluidic chambers were treated with 10 or 100 nM trichostatin A (Sigma, Lot # 026M4036V) for 2 h before 6 h of kainic acid exposure prior to ELISA analysis for acetylated tubulin. The health of cells following TSA treatment was not significantly different from vehicle control, as determined using an alamarBlue assay (Supplementary Figure 1). For all pharmacological manipulations vehicle controls were performed and are reported. No significant difference was found in any measure for vehicles to untreated cells.

Microfluidic cultures were imaged on a Nikon TiE live cell microscope, with chambers maintained at 37°C. Imaging was performed prior to treatment and 18 h following treatment. Axonal side of the microfluidic chambers were imaged using 40 × objective lens to measure axon fragmentation and degeneration. A quantitative measure of axonal degeneration was obtained from five random regions of interest (200 × 200 μm) in the axonal chamber, at least 40 μm away from the microgrooves. Identical regions were imaged before and after treatment. Axon degeneration was calculated by counting degenerated axons. A degeneration index (DI) was determined by the following equation.

The percentage fragmentation refers to the percentage of axons the have fragmented after treatment, compared to the total number of axons before treatment. In our experiments we also analyzed 35, 50, and 100 μM of kainic acid in the cell culture using live imaging, however the axons were too degenerated to be able to target therapeutically.

Plating media was removed from cells plated in 12-well trays, and cultures were rinsed with HBSS. Cells were harvested with RIPA buffer with protease (Complete^TM Mini Protease Inhibitor Cocktail tablets, Roche) and phosphatase inhibitors (Phosphatase Inhibitor Cocktail, A.G. Scientific). Samples were centrifuged at 13,000 rpm for 1 min and pellets were discarded. Samples were diluted at 1:300 in 50 μl bicarbonate/carbonate coating buffer (AbCam), added to 96-well plate and incubated overnight at 4°C. For the standard curve, protein samples were serially diluted at 1:100, 1:200, 1:400, and 1:800. A blank and no primary control were included to correct for ELISA results. Plate was washed with washing buffer (0.01M PBS with 0.05% tween-20) prior to blocking with blocking buffer (0.01M PBS with 5% fetal bovine serum) and incubated at 37°C for 30 min. Plates were washed prior to incubation with detecting antibodies (acetylated tubulin 1:500 mouse, Sigma; tyrosinated tubulin 1:500 rabbit, Millipore; tau 1:1,000 rabbit, DAKO; CRMP2 1:1,000 rabbit, Sigma; MAP1B 1:500 mouse, AbCam) diluted in blocking buffer for 1 h at room temperature. Plates were washed and incubated with species-specific HRP secondary antibody (DAKO) diluted in blocking buffer at 1:2,000 and incubated at room temperature for 45 min. Plates were washed and incubated with room temperature tetramethylbenzidine (TMB) substrate (Thermo Scientific, Lot # SA2328991) for 15 min. The reaction was stopped 0.1M H₂SO₄. Plate was read using plate reader at 450 nm.

Plating media was removed from microfluidic chambers and culture was rinsed with HBSS. Axons were harvested with RIPA buffer with protease (Complete^TM Mini Protease Inhibitor Cocktail tablets, Roche) and phosphatase inhibitors (Phosphatase Inhibitor Cocktail, A.G. Scientific), with protein pooled from 2 to 4 chambers per treatment group. Denatured protein samples (20 μg) were electrophoresed into Bolt^®; 4–12% Bis-Tris gels (Invitrogen), transferred to PVDF membrane (Bio-Rad) and incubated overnight with primary antibodies acetylated tubulin (1:1,000 mouse, Sigma), tyrosinated tubulin (1:1,000 rabbit, Millipore), MAP1B (1:1,000 mouse, AbCam). The corresponding anti-rabbit or anti-mouse horseradish peroxidase conjugated secondary antibody (1:7,000, DAKO) was used as previously described (King et al., 2018). GAPDH (1:5,000 mouse, Millipore) was used as a loading control. Bands were visualized with enhanced chemiluminescence (ECL) solution-Luminata Forte Western horseradish peroxidase (HRP) substrate (Millipore) and images acquired with a Chemi-Smart 5000 Imaging System (Vilber Lourmat) equipped with Chemi-Capt 5000 software. Band intensity was measured as the integrated intensity using Fiji software. After standardizing to GAPDH, each value was calculated as a percentage of control samples (See Supplementary Figure 1).

Values were reported as means ± standard error of the mean (SEM), with differences considered significant at p < 0.05. Differences for ELISA and axon degeneration counts were evaluated using one-way ANOVA, with Tukey's post-hoc test for multiple comparisons between groups. For kainic acid live imaging experiments two coverslips on five separate culture days were analyzed. For trichostatin A live imaging experiments, 1–2 coverslips from four separate culture days were analyzed. For whole cell kainic acid ELISAs, three coverslips per treatment from four separate culture days were used. For whole cell trichostatin A ELISAs, three coverslips per treatment from five separate culture days were used. For axon-only kainic acid ELISAs, cells were pooled from 2 to 3 chambers per treatment from four separate culture days. For axon-only trichostatin A ELISAs, cells were pooled from 2 to 3 chambers per treatment from five separate culture days. All statistical analysis and graphs were prepared in GraphPad Prism (v6.1).

Microfluidic chambers allow compartmentalized separation of axons from the somatodendritic compartment, which permits treatment to be applied exclusive to either compartment of the chamber. To establish an acute model of excitotoxin-induced axon degeneration, mouse cortical neurons at 10 DIV were exposed to 10 and 25 μM kainic acid in the somatodendritic compartment for 18 h. Live imaging of axons was performed immediately prior to and post-treatment to determine a degeneration index. Axon loss was significantly increased (p < 0.05) at 25 μM kainic acid compared to control (Figure 1A). Axonal degeneration at both 10 and 25 μM kainic acid was significantly increased (p < 0.05) compared to control (Figure 1B).

We next determined whether there were changes in microtubule PTMs and the microtubule-associated protein, tau, following 25 μM kainic acid exposure at 1 and 6 h, and whether these changes were related to downstream degeneration (Figure 2). Neuronal cells grown on 12-well trays and cells grown in the somatodendritic compartment of microfluidic chambers were exposed to 25 μM kainic acid for 6 h. ELISA analysis of neurons demonstrated that, at 1 h after 25 μM kainic acid treatment, acetylated and tyrosinated tubulin levels were unchanged, however the microtubule associated protein tau was significantly (p < 0.05) increased, compared to control (Figures 2A–C). At 6 h after 25 μM kainic acid treatment in both neurons and axons, acetylated tubulin levels were significantly (p < 0.05) decreased (Figure 2D). Tyrosinated tubulin levels were unchanged (Figure 2E), whereas, tau levels were also significantly (p < 0.05) increased (Figure 2F). Western blot analysis of acetylated tubulin levels in axons confirmed these were also significantly (p < 0.05) decreased (Figure 2G).

Since 25 μM kainic acid treatment affected levels of the microtubule associated protein tau, we examined whether other microtubule associated proteins were altered at 6 h post treatment. Neuronal cells grown on 12-well trays and cells grown in somatodendritic compartment of microfluidic chambers were exposed to 25 μM kainic acid for 6 h. CRMP2 and MAP1B levels in neurons and axons were examined using ELISA. Like tau, CRMP2 was significantly (p < 0.05) increased following excitotoxicity (Figures 3A,C) and MAP1B was unchanged in both neurons and axons (Figures 3B,D).

Since our data showed significantly reduced acetylated tubulin levels in the axon after 25 μM kainic acid treatment (Figure 2), we hypothesized that promoting microtubule acetylation could rescue these changes and prevent axon degeneration. To promote microtubule acetylation, trichostatin A was used to inhibit the de-acetylating enzyme HDAC6.

We first determined the concentrations of trichostatin A required to promote acetylation in our primary mouse cortical cultures. ELISA analysis demonstrated that acetylated tubulin levels after 10 and 100 nM trichostatin A were significantly (p < 0.05) increased compared to control at 2 h post-treatment (Figure 4A).

We next determined whether HDAC6 inhibition with trichostatin A could rescue acetylated tubulin following excitotoxin exposure. Neuronal cells grown on 12-well trays and the axonal compartment of microfluidic chambers were pre-treated with low (10 nM) trichostatin A for 2 h prior to exposure to 25 μM kainic acid. ELISA analysis demonstrated that 10 nM trichostatin A restored acetylated tubulin levels after kainic acid exposure, as compared to control, in both neurons and axons (Figures 4B,C). Western blot analysis also demonstrated that 10 nM trichostatin A restored acetylated tubulin levels after excitotoxicity in axons (Figure 4D).

Since our results had shown that tau and CRMP2 levels were increased by kainic acid, we examined the effect of trichostatin A on these MAPs. ELISA analysis of neurons demonstrated that 10 nM trichostatin A treatment restored tau to control levels in the presence of kainic acid (Figure 5B). However, CRMP2 levels remained significantly (p < 0.05) higher following kainic acid in the presence of 10 nM trichostatin A (Figure 5D). As we had previously shown, tyrosinated tubulin and MAP1B levels were unaffected by kainic acid and these were also unchanged at 10 nM trichostatin A treatment (Figures 5A,C). ELISA analysis of axons also demonstrated that tau levels were restored after 10 nM trichostatin A, and tyrosinated tubulin levels were unchanged (Figures 5E,F).

We next determined if the rescue of acetylation by trichostatin A treatment following excitotoxin exposure was also protective against axonal degeneration (Figure 6). Neurons were grown in microfluidic chambers and axons exposed to low (10 nM) and high (100 nM) trichostatin A for 2 h prior to somatodendritic treatment with 25 μM kainic acid. Axon degeneration was quantitated from images acquired before and 18 h after kainic acid treatment. In the presence of 10 and 100 nM trichostatin A, axon degeneration (Figure 6B) and axon loss (Figure 6A) were significantly reduced (p < 0.05) after exposure to 25 μM kainic acid.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.