Background
Glaucoma is a leading cause of blindness worldwide and is characterized by damage to the optic nerve [
1,
2]. Increased intraocular pressure results in an increase in the strain surrounding the optic nerve head [
3,
4], which is believed to precipitate focal damage to retinal ganglion cell (RGC) axons as they pass into the optic nerve [
5]. Axogenic neurodegeneration precedes somatogenic neurodegeneration in the predicted pathophysiology of a majority of optic neuropathies, such as glaucoma [
6]. Optic nerve crush (ONC) mimics these molecular events by inducing partial RGC axonal damage [
7], and it is a widely accepted model of acute RGC injury that has been used to study intrinsic and extrinsic apoptotic mechanisms of RGC death [
7‐
12].
The intrinsic mechanism of apoptosis involves activation and translocation of the pro-apoptotic protein BAX to the mitochondria following cellular injury. Oligomerization of BAX at the mitochondrial outer membrane releases cytochrome c from the mitochondria [
13], which leads to activation of caspases and subsequent cell death [
14]. The steps involving BAX mark the committed step of intrinsic apoptosis [
15]. In a previous study,
Bax-deficient RGCs remained viable up to at least 72 weeks post ONC, however, these cells exhibited many early stages of atrophy typical of wild type cells undergoing cell death [
8]. This was particularly evident in structural and functional changes in RGC nuclei. The RGC nuclei were found to exhibit atrophic characteristics including nuclear shrinkage, histone H4 deacetylation, heterochromatin formation, and RGC-specific gene silencing soon after ONC [
8,
10,
16].
Changes in the transcriptional profile of damaged neurons have been described in several models of neurodegeneration, including the down regulation (silencing) of normal gene expression and the increase in expression of stress-response and pro-apoptotic genes [
10,
17‐
26]. Transcriptional downregulation and initiation of the cell death mechanism in several cases of neuronal injury, including RGC death, were linked with epigenetic processes such as histone deacetylase (HDAC) activity [
10,
18,
19,
27‐
29]. Although most HDACs are found ubiquitously in tissues, class I HDAC isoforms 1, 2, 3, and 6 are found primarily in the cells of the inner nuclear layer and the ganglion cell layer (GCL) of the murine retina [
10,
30]. A previous study showed that RGC gene silencing and RGC death were attenuated following ONC as a result of pretreatment with the broad spectrum HDAC inhibitor Trichostatin A (TSA) [
10]. The same study demonstrated that HDAC3 translocated to the nucleus in concert with H4 deacetylation during RGC death [
10]. These results suggested a potential role for HDAC3 in early RGC gene downregulation and global deacetylation events in RGC death following axonal injury. Other studies have also reported that HDAC3 is toxic to differentiated neurons, indicating an important role for HDAC3 molecular events in neurodegeneration [
28,
31,
32].
Here we show that conditional knock out of Hdac3 in RGCs ameliorates global deacetylation and heterochromatin formation, while improving nuclear integrity and RGC viability following ONC. Interestingly, conditional knockout of Hdac3 does not prevent the downregulation of RGC-specific gene expression, even though TSA does. We interpret these data as indicating that a different class I HDAC may be responsible for global transcriptional regulation in the early stage of nuclear atrophy. Overall, the results indicate an important role for HDAC3 in the early events of neuronal intrinsic apoptosis and provide direction for dissecting the roles of other class I HDACs in the process of early transcriptional silencing during the RGC apoptotic program.
Discussion
The time course of RGC apoptosis can be temporally divided into phases of early cell and nuclear atrophy, initiation of BAX oligomerization, and late caspase and endonuclease activation leading to nuclear fragmentation. The main characteristics of the events of nuclear atrophy include gene silencing, global deacetylation, heterochromatin formation, and a decrease in nuclear structural integrity. Based on previous work showing neuronal toxicity of HDAC3, amelioration of RGC death with HDAC1 and HDAC3 inhibitors, and HDAC3 localization to the nuclei of RGCs prior to maximal histone H4 deacetylation during apoptosis [
10,
28,
39], we sought to determine the role of HDAC3 in early RGC nuclear atrophy by analyzing each of the atrophic characteristics following ONC.
Here, we demonstrated that H4 deacetylation and heterochromatin formation were prevented in RGCs of
Hdac3 cKO retinas at 5 days following ONC. Although nucleoli and nuclear pores remained normal appearing, we found that the RGCs of
Hdac3 cKO retinas had wavy-appearing nuclear envelopes, which may reflect an intermediate stage of perinuclear swelling that was present in the nuclei of cells of crushed control mice. Alternatively, HDAC3 may play a role in maintaining the latticework that supports the inner nuclear membrane [
43], by associating with lamin A/C (see below).
HDAC3 enzymatic activity, resulting in histone deacetylation and heterochromatin formation, are temporally situated early in the apoptotic program, because they still occur in
Bax- deficient mice, where completion of apoptosis is effectively permanently blocked [
8,
44]. It is not known, however, if these HDAC3-mediated changes are essential to activate further downstream events in the apoptotic pathway, such as caspase activation and nuclear fragmentation. Our results showing a correlation between HDAC3 activity and CASPASE-3 activation support a cause and effect relationship between these early events and later stages in the apoptotic pathway. Whether or not this relationship is linked to chromatin remodeling is not known at this time. HDAC3 may also play a role in modifying non-histone targets early in the apoptotic process. One such target may be the transcription factor p53. Acetylation and phosphorylation play a critical role in regulating p53 activity by altering cellular localization or binding affinity to specific targets [
45]. Specifically, Chao and colleagues observed that HDACs could activate p53 by deacetylating lysine residue 317, leading to activation of pro-apoptotic gene expression such as
Bbc3 (PUMA) and
Bim[
46]
. Both of these BH3-only containing proteins of the
Bcl2 family of genes function by modulating the activity of BAX to promote mitochondrial permeability leading to the release of cytochrome-c and the activation of the caspase cascade. Importantly, both of these proteins have been strongly implicated in regulating RGC apoptosis after optic nerve damage [
42]. Separately, Brochier et al. [
47] observed that inhibition of p53 DNA-binding and transcriptional activity in neurons was obtained by acetylation on lysines 381 and 382. The same post-translational modifications enhance p53 activity in cancer cells, however. If HDAC3 plays such a role in regulating the acetylation status of p53, this would provide a mechanism for the selective toxicity of this HDAC in differentiated neurons [
28,
40,
48].
One of the events in nuclear atrophy not affected by
Hdac3 deletion is the silencing of normal RGC gene expression. Previously, we showed that injection of broad-spectrum HDAC inhibitor, TSA, prior to ONC prevented silencing of the RGC specific gene promoter,
Fem1c, in both acute and chronic (glaucomatous) models of axonal injury [
10,
16]. We hypothesized that amelioration of gene silencing would occur with the conditional knockout of
Hdac3 in RGCs of mice that later underwent ONC. However, knockout of
Hdac3 did not lead to protection from gene silencing at 5 days following ONC, rather RGC specific gene transcript abundance also decreased similar to the change exhibited by wild type retinas. Conversely, TSA administration prior to ONC led to significant amelioration of RGC specific gene silencing in RGCs. These results, taken together with evidence that HDACs 1, 2, 3, and 6 localized in the retina [
30], indicated a potential role for these other HDACs in the early event of gene silencing during RGC atrophy.
Several class I HDACs have been implicated in regulating gene transcription. An appropriate candidate for investigation may be HDAC2 due to it’s increased mRNA abundance in retinas at 1 day following ONC [
10]. Previous work demonstrated that HDAC2 comprised approximately 35% of the total HDAC activity in the mouse retina and that retinas lacking HDAC2 underwent less retinal degeneration following ischemic insult [
30]. The HDAC1/HDAC2 corepressor complex can be targeted to genes by transcription factors such as Sp1 and Sp3, playing a role in regulation of gene expression via chromatin remodeling [
49].
A different mechanism of gene silencing, involving lamin-associated domains, was recently demonstrated in
Drosophila S2 somatic cells, which exhibit silencing of a multi-genic testes gene cluster. This gene cluster is sequestered at the nuclear envelope as heterochromatin that interacts with the lamin protein complex. Domains, such as this one, are maintained by the activity of the
Drosophila HDAC1 homolog, with the HDAC3 homolog playing an apparent auxiliary role [
43]. Lamin-associated domains are increasingly recognized as a mechanism for repressing gene activity during development [
50]. This phenomenon offers an interesting mechanism for gene silencing during cell death, and is consistent with one of the hallmarks of apoptosis, which is an initial accumulation of heterochromatin along the inner surface of the nuclear envelope [
51,
52]. It is unknown if this accumulation of heterochromatin associates with lamin domains, or selectively involves aggregation of genomic DNA with actively transcribed genes that are targeted for silencing. This potential model warrants further investigation.
Although HDAC3, by itself, evidently does not seem to play a large role in early gene silencing, it may be a valuable molecular target for drug inhibitor therapy since knockout of Hdac3 expression has been shown to halt the progression of histone deacetylation and attenuate subsequent RGC death after axonal injury. An important consideration resulting from these experiments is that HDAC-mediated changes in RGCs can be segregated into very early events, likely associated with gene-silencing, and later events, associated with global chromatin remodeling. It is still relevant to address the mechanism of gene silencing, since RGC function is expected to rely on these cells having a normal profile of gene expression that defines cell identity and function. We anticipate that selective targeting of HDAC3 for protective therapy may yield living, but non-functional RGCs, because of their inability to express a genetic profile that identifies them as ganglion cells. Further investigation into functions and timing of different HDAC activities during the process of chromatin remodeling will provide insight into early epigenetic events that play a role in RGC apoptosis associated with optic neuropathies.
Methods
Experimental animals, AAV2-Cre/GFP injection, and ONC
All mice were handled in accordance with the Association for Research in Vision and Ophthalmology statement for the use of animals for research, and experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin. A random mixture of male and female C57BL/6 mice between the ages of 4–6 months, were used for experiments. The
Rosa26-LacZ
fl/fl
mice were generously provided by Dr. Jing Zhang from the University of Wisconsin-Madison McArdle Laboratory for Cancer Research (Madison, Wisconsin).
Rosa26-Tomato
fl/fl
mice were obtained from the Jackson Laboratory (Bar Harbor, Maine), and
Hdac3
fl/fl
mice [
53] were obtained from Dr. Scott Hiebert at Vanderbilt University (Nashville, Tennessee, USA). The
Rosa26 mice contain the
loxP flanked PGK-Neo stop transcript found upstream of either
Tomato or
LacZ (βGEO) reporter transcript and were used to monitor for RGC transduction of the AAV2-Cre/GFP virus as well as
Hdac3+/+ controls for viral injections (see below). Each treatment group contained at least 4 mice for all experiments except for heterochromatin formation analysis (n = 3). The C57BL/6
Hdac3
fl/fl
mice contain loxP sites flanking exon 7 of the
Hdac3 gene transcript [
53].
Injections of 1 μL AAV2-Cre/GFP (equaling a total of 10
9 gc) were administered intravitreally 4 weeks prior to ONC. The 4-week time point was chosen after testing transduction efficiency at 1, 2, 4, and 8 weeks post AAV2-Cre/GFP injection showed optimal expression of
Tomato starting at 4 weeks (Additional file
1: Figure S1). AAV2-Cre/GFP was obtained from Vector Biolabs (Philadelphia, PA, USA), and 1:10 dilution of stock AAV2-Cre/GFP was made in 5% glycerol in sterile phosphate buffered saline (PBS, 137 mM NaCl, 1.8 mM KH
2PO
4, and 10 mM Na
2HPO
4, pH 7.5) prior to injection. Intravitreal injections were conducted using a 10 μL Hamilton syringe with a 35G needle attached. A volume of 1 μL of diluted AAV2-Cre/GFP was injected over a period of 40 seconds, and the needle was held in the eye for at least 30 seconds before it was retracted. ONC was performed unilaterally using self-closing forceps to initiate degeneration of RGCs. Retinas were harvested at 5, 14, 28, and 56 days following ONC for analysis. Previously, we observed peak histone H4 deacetylation in the GCL by 5 days post-ONC [
10]. Therefore, retinas were harvested at 5 days to assess mRNA abundance, histone H4 deacetylation, and heterochromatin formation. Retinas were collected at 14, 28, and 56 days post ONC for cell counts. Cell counts were obtained using a modification of the method described in Li et al. (2007) [
54]. Briefly, digital images collected at 400× magnification were taken of each lobe of a retinal whole mount stained with DAPI. Cell numbers were determined in 24 separate 100 μm
2 fields for each retina. Change in cell number for each experimental eye was calculated as a percentage of cell numbers in the corresponding control eye of each mouse.
β-galactosidase staining and bright field microscopy
β-galactosidase reporter expression was identified histochemically in retinal sections and whole mounts by X-Gal assay. X-gal staining solution was prepared ahead of time in the dark by adding N, N dimethylformamide to X-gal (0.02% Igepal, 0.01% sodium deoxycholate, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 2 mM MgCl2 diluted in 0.1 M PBS [pH 7.3]). Mice were euthanized and eyes were collected and fixed in 4% paraformaldehyde in PBS at room temperature for 50 minutes. For whole mounts, the eye was then rinsed with PBS, and the anterior portion of the eye was removed, leaving an eyecup. The eyecup was then washed in PBS containing 2 mM MgCl2 and 2 μM CaCl2 and stained by incubation in staining solution containing 1 mg/mL X-gal at 37°C for 18 hours. After staining, the retina was dissected from the eyecup and whole mounted on glass slides. For sections, the eye was rinsed with PBS, and the anterior portion of the eye was removed, leaving an eyecup. The eyecup was then equilibrated in 30% sucrose overnight at 4°C before mounting in blocks of Tissue-Tek O.C.T. Compound from Fisher Scientific (Pittsburgh, PA) for cryosectioning (5 μm thick). Sections were then rinsed in PBS containing 2 mM MgCl2 and 2 μM CaCl2 and stained by incubation in staining solution containing 1 mg/mL X-gal at 37°C for 18 hours. Retinas were then washed in PBS and stained with nuclear fast red stain for 5 minutes. The slides were examined and photographed using an Olympus BX40 light microscope (Olympus America Inc., Center Valley, PA) and digital camera attachment.
Immunofluorescence
Indirect immunofluorescence on 5 μm thick frozen retinal sections and whole mounts was done as described previously [
10]. Cryosections and whole mounts were mounted on Superfrost Plus microscope slides (Fisher Scientific) and rinsed in PBS. The sections were then blocked in 5% bovine serum albumin (BSA) in PBS for 3 hours at room temperature and later rinsed in PBS. Primary antibodies including polyclonal rabbit antibody to human HDAC3 (#sc-11417) and polyclonal rabbit antibody HDAC2 (#sc-7899) (both from Santa Cruz, Dallas, TX), polyclonal rabbit antibody to human AcH4 (#06-866) and monoclonal mouse primary to human BRN3A (#MAB1585) (both from EMD Millipore Inc., Billerica, MA), monoclonal mouse primary to human TUJ-1 (#ab14545) (AbCam, Cambridge, MA), and polyclonal rabbit primary antibody to human CASPASE-3 (#AF835) (R&D Systems, Minneapolis, MN) were used at 1:100 dilutions. Sections and whole mounts were incubated in primary antibody for 24–48 hours at 4°C and washed in PBS afterwards. Secondary antibodies used included goat anti-rabbit TEXAS RED (1:1,000) and goat anti-mouse FITC (1:1,000) (Jackson ImmunoResearch Laboratories, West Grove, PA). Sections and whole mounts were incubated in secondary antibody at room temperature in the dark for 2 hours and washed in PBS. All sections and whole mounts were counter-stained for 10 minutes with 4’, 6-diamidino-2-phenylindole (DAPI) and were washed in PBS. Finally, sections and whole mounts were mounted using Immumount mounting medium (Fisher Scientific) and coverslipped. Fluorescent images were obtained using a Zeiss Axioplan 2 Imaging microscope with Axiovision 4.6.3.0 software (Carl Zeiss MicroImaging Inc., Thornwood, NY).
Transmission Electron Microscopy (TEM)
Rosa26-Tomato
fl/fl
and Hdac3
fl/fl
mouse eyes were injected with AAV2-Cre/GFP and after 4 weeks were subjected to ONC surgery. Animals were analyzed 5 days after ONC surgery. Enucleated eyes were immersed in 4% paraformaldehyde in 0.1 M Phosphate buffer (PB) for 5 minutes, after which the anterior chambers and lenses were dissected away from each eyecup. A small region of the superior eyecup was then removed and placed in 2.5% glutaraldehyde, 2% paraformaldehyde in PB overnight at 4°C. Tissues were postfixed in 1% osmium tetroxide in PB, dehydrated in ethanol, and embedded in Epon epoxy. Sections (60–90 nm) were cut, stained with 50% ethanoic uranyl acetate and Reynold’s lead citrate, and viewed using a Phillips CM120 transmission electron microscope (FEI Company, Hillsboro, OR).
Heterochromatin scoring analysis
Tissues processed for TEM were also sectioned for bright field microscopy. Thick (1 μm) sections were cut from epoxy embedded samples and stained with Richardson’s stain (methylene blue and azure blue). Sections were imaged using an Olympus BX40 light microscope and a digital camera attachment. Nuclear morphology of cells in the GCL was scored by 3 masked observers. A score of 1 indicated cells that exhibited healthy euchromatic nuclei with well-formed nucleoli; a score of 2 indicated cells that were partially heterochromatic; and a score of 3 indicated cells that had completely condensed pyknotic chromatin or fragmented nuclei.
Evaluation of transcript abundance in the retina by qPCR
Total retinal RNA was isolated from 5 pooled retinas at 5 days post ONC by acid-phenol extraction, and RNA was then DNase I treated (Promega, Madison, WI). First strand cDNA using reverse transcriptase and oligo (dT) was synthesized from 2 μg of isolated and purified total RNA [
55]. The resulting cDNA was diluted 10-fold and 5 μl of cDNA was used for each qPCR reaction with SYBR Green PCR master mix (Applied Biosystems, Foster City, CA) and the appropriate RGC gene-specific primers as listed in the table of primer sequences (Table
1). Quantitative PCR was conducted on triplicate samples in each run using ABI 7300 Real Time PCR system (Applied Biosystems). Data were obtained from triplicate samples for each target cDNA. Absolute transcript quantification was based on a standard
S16 curve run during the same reaction and the copy number was normalized to
S16 ribosomal protein mRNA. The mRNA transcript values are expressed as the percent change from contralateral control eye to treatment eye. Data were reported as the mean ± SD of these differences.
Table 1
Primers for qPCR analysis
Thy1
| 5’-CTTGCAGGTGTCCCGAGGGC-3’ | 379 |
5’-CTGAACCAGCAGGCTTATGC-3’ |
Sncg
| 5’-GACCAAGCAGGGAGTAACGG-3’ | 240 |
5’-TCCAAGTCCTCCTTGCGCAC-3’ |
Nrn1
| 5’-TTCACTGATCCTCGCGGTGC-3’ | 238 |
5’-TACTTTCGCCCCTTCCTGGC-3’ |
Nfl
| 5’-AGCACGAAGAGCGAGATGGC-3’ | 173 |
5’-TGCGAGCTCTGAGAGTAGCC-3’ |
S16
| 5’-CACTGCAAACGGGGAAATGG-3’ | 198 |
5’-TGAGATGGACTGTCGGATGG-3’ |
Western blot analysis
Western blot analysis was conducted on 5 pooled Tomato
fl/fl
and Hdac3
fl/fl
mouse retinas from each treatment group described. Individual retinas harvested from Hdac3 cKO and control eyes were also analyzed. Retinal protein was loaded in triplicate with 50 μg per lane on 12% polyacrylamide gels and transblotted onto Immobilon P (Millipore, Inc., Billerica, MA). Membranes were probed for HDAC3, HDAC2, and ACTIN. Rabbit polyclonal antibodies were used at 1:1,000 for HDAC3 and HDAC2 and a goat monoclonal antibody was used at 1:250 for ACTIN (I-19) (cat# sc-1616) (Santa Cruz, CA). Firstly, the blots were incubated in donkey anti-goat secondary (1:10,000) conjugated to IRDye 800CW (cat# 926–32214), and after washing in PBS, incubated in goat anti-rabbit secondary (1:10,000) conjugated to IRDye 680RD (cat# 926–68071) (LICOR, Lincoln, NE). Images were scanned and analyzed using the Odyssey Clx (LICOR). Band fluorescence was quantified using Image Studio software, and data were normalized to the ACTIN loading control on each blot.
Statistical analysis
Data were collected from a minimum of 4 independent samples in all experiments except for analysis of heterochromatin formation (n = 3), and shown as the mean ± standard deviation in all experiments except for cell counts, where data was shown as the mean ± standard error. All statistical analyses were performed using either the Student’s t-test with statistical significance set at P ≤ 0.05 for comparison of two groups or ANOVA with Bonferroni adjustments with statistical significance set at P ≤ 0.05 for comparison of multiple groups.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they no competing interests.
Authors’ contributions
HMS performed viral and HDAC inhibitor injections, ONC, sample harvesting, RNA isolations, qPCR analysis, Western blotting, bright field and fluorescence microscopy, fluorescent antibody labeled cell counts, sample processing for TEM analysis, and writing of all sections of the manuscript. HRP performed some of the HDAC inhibitor injections and qPCR analysis, sample processing, and drafting the manuscript. CLS processed image data and performed some data analysis and project design. RWN conceived the study, actively participated in the design and coordination of the study, reviewed all data, and helped draft the manuscript. All authors read and approved the final manuscript.