HDAC1 nuclear export induced by pathological conditions is essential for the onset of axonal damage

Abstract

Histone deacetylase 1 (HDAC1) is a nuclear enzyme involved in transcriptional repression. We detected cytosolic HDAC1 in damaged axons in brains of humans with multiple sclerosis and of mice with cuprizone-induced demyelination, in ex vivo models of demyelination and in cultured neurons exposed to glutamate and tumor necrosis factor-alpha. Nuclear export of HDAC1 was mediated by the interaction with the nuclear receptor CRM-1 and led to impaired mitochondrial transport. The formation of complexes between exported HDAC1 and members of the kinesin family of motor proteins hindered the interaction with cargo molecules, thereby inhibiting mitochondrial movement and inducing localized beading. This effect was prevented by inhibiting HDAC1 nuclear export with leptomycin B, treating neurons with pharmacological inhibitors of HDAC activity or silencing HDAC1 but not other HDAC isoforms. Together these data identify nuclear export of HDAC1 as a critical event for impaired mitochondrial transport in damaged neurons.

Figures

Figure 1. Cytosolic HDAC1 is detected in animal models of demyelination, in the brain of MS patients and in demyelinated slice cultures

(a) Confocal micrographs of the corpus callosum of control and 4 week cuprizone-fed mice (Cupri4w), stained with antibodies specific for class I (HDAC1, HDAC2, HDAC3, HDAC8) and class II HDAC isoforms (HDAC4, HDAC5, HDAC6 and HDAC7), indicated on the left (green) and with anti-neurofilament medium chain antibody (NFs, red). Scale bar 10 μm; 63× objective. (b) Confocal image of the co-localization (white arrows) of hypophosphorylated neurofilament heavy chain (SMI32, red) and HDAC1 (green) in damaged axons in the corpus callosum of cuprizone fed mice. Scale bar 5 μm. (c) Luxol fast blue/periodic acid Schiff staining of human brain sections identified areas of white matter demyelination (arrows, scale bar 200 μm). (d) Serial sections of the same lesions were used to show the presence of HDAC1+ end bulbs (green) in neurofilament positive (red) axons. Scale bar 10 μm. (e) Adjacent sections were processed for immunohistochemistry with antibodies for SMI32 (red) and HDAC1 (green) in white matter lesions. DAPI (blue) was used as nuclear counterstain. Scale bar 10 μm. (f) Confocal images of cerebellar slice cultures either untreated (Control) or treated with lysolecithin and LPS to induce demyelination (Lyso/LPS17h+LPS24h). Cultures were stained with antibodies specific for myelin basic protein (MBP, red) and NFs (green) antibodies. (g) The same cultures were then stained for HDAC1 (green) and SMI32 (red).

Figure 2. Glutamate and TNF-α treatment of primary neuronal cultures induces neuritic beading followed by transections

(a) Confocal image of cultured neurons either untreated (Control) or treated with 50μM glutamate/200ng/ml TNF-α for 2 or 24 hours. Cells were stained with antibodies for NFM (green) to visualize neurites and SMI32 (red) to identify damaged neurites. Scale bar 10μm in low and 2 μm in high magnification. (b) Bar graphs show the percentage of SMI32+ beaded neurites relative to the total number of NFM+ neurites (mean ± SD, n = 83±12 neurites counted in each condition. * P < 0.05, *** P < 0.001 two-tailed Student’s t-test). (c) Time-lapse video microscopy of cultured neurons exposed to 50μM glutamate/200ng/ml TNF-α. Images were captured every 30 minutes for 12 hours. Scale bar 5μm. (d) The graph indicates the percentage of normal appearing, beaded and transected neurites, relative to the total, at each time point. Error bars represent standard deviation.

Figure 3. Cytosolic localization of HDAC1 precedes the onset of localized beadings and impaired mitochondrial transport

(a) Confocal images of neurons before and after exposure to 50μM glutamate/200ng/ml TNF-α for the indicated time periods. Control and treated cultures were co-stained with antibodies specific for HDAC1 (green) and for neurofilament medium and light chains (NFs, red). Upper panel, scale bar 5μm or 2 μm; lower panels 10 μm. (b) Bar graphs indicate the percentage of beaded neurites relative to the total NFs+ population at the indicated time points; error bars represent standard deviation. *P < 0.05. (c) Live cell images of moving mitochondria, labelled with MitoTracker in neurons exposed to 50μM glutamate/200ng/ml TNF-α. The position of each mitochondrion was recorded every 5 seconds for 5 minutes at the indicated time points and the tracks of mitochondrial movement were pseudocolored. Scale bar 5μm. (d) Bar graphs indicate the average speed of moving mitochondria with standard deviation. **P < 0.01.

Figure 4. Calcium-depletion prevents HDAC1 nuclear export and the onset of neuritic beading induced by glutamate and TNF-α treatment

(a) Confocal image of primary cultures transfected with FLAG-tag-HDAC1 on day 10 and then analyzed three days later in the absence (Control) or presence of 50μM glutamate/200ng/ml TNF-α for 2 hours. Transfected cells were detected using antibodies against FLAG-tag (green) and NFM (red). Scale bar 2 μm for high magnification and 10 μm for low magnification. (b) Confocal images of cultured neurons pre-treated with increasing concentrations of the calcium chelator EDTA and stained with antibodies specific for HDAC1 (green) and NFs (red). Note that EDTA prevents neurite beading induced by treatment with 50μM glutamate/200ng/ml TNF-α. Scale bar 10μm for low magnification and 2 μm for high magnification (High mag.). (c) Bar graphs show the quantification of the results shown in (b) (mean ± SD; **P < 0.01, ***P < 0.001).

Figure 5. Silencing HDAC1, but not other isoforms prevents neurite beading

(a) Primary neurons were infected with Hdac1-shRNA lentiviral particles and then exposed to 50μM glutamate/200ng/ml TNF-α for 2 hours and processed for immunocytochemistry with anti-HDAC1 (green) and NFs (red) antibodies. DAPI (blue) was used as nuclear counterstain. Scale bar 10μm for low magnification and 2 μm for high magnification (High mag.). (b) Cultured neurons were infected with class I (Hdac2, Hdac3 and Hdac8) and class II (Hdac4 and Hdac6) Hdac-shRNA lentiviral particles. 72 hours after infection, cultures were exposed to 50μM glutamate/200ng/ml TNF-α for 2 hours and then processed for immunocytochemistry with antibodies specific for HDAC2, HDAC3, HDAC4, HDAC6, or HDAC8 (green) and NFs (red). Scale bar 10 μm for low and 2 μm for high magnification.

Figure 6. CRM1 dependent nuclear export of HDAC1 is essential for the induction of damage induced by glutamate and TNF-α

(a) Schematic diagram of HDAC1 protein sequence, with potential nuclear export sequences (NESs, red), and nuclear localization domains (NLS, blue). (b) Protein extracts from cultured neurons either untreated or treated with toxic stimuli for 5 or 20 minutes were immunoprecipitated with antibodies specific for HDAC1 (IP:HDAC1) and processed for western blot analysis using CRM1 or HDAC1 antibodies. Whole cell lysates (WCL) were used as controls. (c) Wild type (Wt) and NES mutant HDAC1 (Mut) over-expression, followed by immunoprecipitation with anti-FLAG antibody and western blot analysis with anti-CRM1 and FLAG antibodies. (d) Immunocytochemistry of primary neurons pre-treated with increasing concentrations of leptomycin B (LMB) for 30min, then exposed to 50μM glutamate/200ng/ml TNF-α (Gluta/TNF-α) and stained with antibodies for HDAC1 (green) and NFs (red). Scale bar 10μm for low and 2 μm for high magnification. (e) Bar graphs indicate the percentage of beaded neurites relative to the total population (mean ± SD; *P < 0.05). (f) Pseudo-colored image of mitochondrial movement, using time-lapse video microscopy of primary neurons treated with LMB and exposed to glutamate and TNF-α. Untreated cultures were used as controls. Cultures were labeled with Mitotracker GreenFM and photograms were taken every 5 seconds within a 5 minute time period. Note that LMB treatment ameliorated mitochondrial movement. Scale bar 5μm. (g) Bar graphs indicate the average speed of moving mitochondria in each condition and error bars represent standard deviation *P < 0.05.

Figure 7. HDAC1 binding partners in demyelinating region of cuprizone-treated mice

(a) Validation of the MALDI-TOF results in vivo. Total proteins from demyelinated (Corpus callosum) and (b) non-demyelinated (Spinal cord) regions of cuprizone-treated (Cupri4w) or control mice were immunoprecipitated with anti-HDAC1, HDAC4, or HDAC6 antibody and probed with the indicated antibodies. Shown are representative results of 3 independent experiments, each comparing 3 mice for each condition. Note that HDAC1 forms protein complexes with proteins involved in axonal transport only in demyelinated regions. The expression levels of CNPase, an enzyme critical for myelination process, was used to verify the occurrence of demyelination, while expression levels of HDAC1, HDAC4 and HDAC6 were used to confirm the efficiency of the immunoprecipitation.

Figure 8. Activity-dependent interaction of HDAC1 with motor proteins impairs mitochondrial transport

(a) Pseudo-colored image of time-lapse video-microscopy of neuronal cultures treated with 50μM glutamate/200ng/ml TNF-α in the presence of HDAC1 inhibitor MS-275, or HDAC6 inhibitor tubacin. Mitochondrial movement was examined using the lipophilic mitochondrial dye MitoTracker. Scale bar 5μm. (b) Bar graphs indicate the average speed of moving mitochondria (mean ± SD). *P < 0.05, **P < 0.01 (c) Confocal image of neurites in cultures treated with 50μM glutamate/200ng/ml TNF-α for 2 hours (Glu/TNF-α), or in the presence of MS-275 (Glu/TNF-α+MS-275) or tubacin (Glu/TNF-α+tubacin). Immunoreactivity for HDAC1 (green) and NFs (red); DAPI (blue) as nuclear counterstain. (d) Bar graphs represent the percentage of beaded neurites in (c) (Mean ± SD; *P < 0.05). (e) Protein complexes containing HDAC1 (left) or HDAC6 (right) and motor proteins were identified by immunoprecipitation and western blot analysis. Proteins were extracted from cultured neurons treated as indicated in (a). (f) Co-immunoprecipitation with anti-dynamin antibody of protein extracts from either untreated or MS-275 pre-treated neurons exposed to 50μM glutamate/200ng/ml TNF-α for 20min or 2 hours. Note that the association of dynamin from motors was partially restored by treatment with MS-275 or silencing HDAC1 with shRNAs. Silencing efficiency was confirmed by western blot analysis using HDAC1 antibody. (g) Western blot analysis of whole cell lysates (WCL) from cultures treated as described in (e) and (f).