The expression of dynein light chain DYNLL1 (LC8-1) is persistently downregulated in glaucomatous rat retinal ganglion cells

Experimental Eye Research 92 (2011) 138e146

Contents lists available at ScienceDirect

Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer

The expression of dynein light chain DYNLL1 (LC8-1) is persistently downregulated in glaucomatous rat retinal ganglion cells Christian van Oterendorp a, b, Barbara Lorber a, Zorica Jovanovic c, Giles Yeo c, Wolf A. Lagrèze b, Keith R. Martin a, d, * a

Centre for Brain Repair, University of Cambridge, The E.D. Adrian Bldg., Forvie Site, Robinson Way, Cambridge, CB2 2PY, United Kingdom University Eye Hospital Freiburg, Germany Institute for Metabolic Science, University of Cambridge, United Kingdom d Cambridge NIHR Biomedical Research Centre, Cambridge, United Kingdom b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 August 2010 Accepted in revised form 30 November 2010 Available online 7 December 2010

High intraocular pressure induces glaucomatous degeneration of retinal ganglion cells (RGCs). The cellular mechanisms leading to activation of the apoptosis cascade are multidimensional and only partially understood. A small dynein subunit, the light chain DYNLL1 (synonym LC8-1, PIN) has recently been shown to be an important regulator of neuron proteins known to be involved in glaucomatous RGC death including NO synthases, the pro-apoptotic protein Bim and the dynein intermediate chain. Also, DYNLL1 is a regulator of mitochondria anchorage in axons, which is impaired in glaucoma. We investigated expression of DYNLL1 and 2 and its dynein binding partner dynein intermediate chain in a rat model of chronic glaucoma. Laser capture microdissection (LCM) allowed us to collect distinct cell layers and cell bodies from the retina to gain data highly specific for retinal ganglion cells. Glaucoma was induced in 23 rats by laser treatment to the aqueous outflow tract. RNA was extracted from LCM dissected ganglion cell layers (GCL) and 100 pooled RGCs per retina. Expression levels for 1, 2 and 4 week timepoints were analysed by quantitative PCR for DYNLL1 and 2, dynein intermediate chain and GFAP. DYNLL protein abundance in RGCs was quantified in immunostained retina sections. DYNLL gene 1 but not 2 was expressed in RGCs. In the glaucoma model DYNLL1 was strongly and persistently downregulated at all timepoints. DYNLL protein was significantly less abundant at the 4 week timepoint. In contrast, the motorprotein binding partner dynein intermediate chain 1 was more stably expressed. DYNLL2 was upregulated in glia cells at 2 weeks. Expression of DYNLL1, the only form of the dynein light chain expressed in RGCs, is downregulated persistently in glaucoma, while its binding partner dynein IC-1 is unchanged. The specific lack of DYNLL1 could have an impact on the function of their regulatory binding partners and contribute in several ways to neuron dysfunction and apoptosis. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: glaucoma axonal transport retinal ganglion cell apoptosis dynein dynein light chain DYNLL

1. Introduction Glaucoma is characterised by progressive loss of retinal ganglion cells (RGCs), which die mainly by apoptosis (Quigley et al., 1995; Libby et al., 2005). The mechanisms leading to activation of the apoptotic cascade are complex and only partially understood. Neurotrophin deprivation and oxidative stress are considered to be

* Corresponding author. Centre for Brain Repair, University of Cambridge, The E.D. Adrian Bldg., Forvie Site, Robinson Way, Cambridge, CB2 2PY, United Kingdom. Tel.: þ44 1223216427; fax: þ44 1223331174. E-mail addresses: [email protected] (C. van Oterendorp), [email protected] (B. Lorber), [email protected] (G. Yeo), [email protected] (W.A. Lagrèze), [email protected] (K.R. Martin). 0014-4835/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2010.11.013

important factors (Quigley, 2005; Tezel, 2006). The constant supply of neurotrophic signals originating from pre- and postsynaptic neurons or glia relies on an intact endosome transport system from the ganglion cell periphery to the nucleus. The motorprotein mainly executing this retrograde transport is the multisubunit molecule dynein (Heerssen et al., 2004). Not all subunits of the dynein complex are entirely dedicated to transport, with the dynein light chain DYNLL (syn. LC8, PIN) consisting of the subgroups DYNLL1 and 2, known to have other regulatory functions in the cell (Barbar, 2008). DYNLL1 is an evolutionarily highly conserved 10kD protein, with an identical sequence throughout mammals and a 94% similarity between human and Drosophila. Null mutations in Drosophila are lethal (Phillis et al., 1996). Although first described as part of the

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

dynein complex, DYNLL1 binds to a variety of other proteins, modulating their function by alteration of the secondary structure (Fan et al., 2001; Nyarko et al., 2004; Benison et al., 2006; Barbar, 2008). Among the target proteins are the nitric oxide synthases (NOS), the pro-apoptotic factor Bim, the NFkB-inhibitor IkBa, the axonal mitochondria anchorage protein syntaphilin and the dynein intermediate chain. DYNLL2, which is encoded by a different gene but shares a 93% sequence homology and presumably functional similarity with DYNLL1 is less well known. It specifically binds to the pro-apoptotic BH3-only protein Bmf (Day et al., 2004). The list of DYNLL target proteins shows a remarkable overlap with the proteins considered to be involved in glaucomatous RGC death. It includes the dynein intermediate chain (axonal transport), NO synthase (oxidative stress) and Bim (apoptosis). Other DYNLL target proteins are related to the induction of RGC death by their function (syntaphilin: axonal energy supply and Ca2þ homeostasis; IkBa: NFkB signalling). This led us to investigate the expression of DNYLL in RGCs under physiological conditions and in a rat model of chronic glaucoma. A methodological obstacle in glaucoma research is that RGCs represent only a minor population in the retina. When analysing whole retinal lysates, changes in RGC gene expression may be ‘diluted’ or are prone to misleading results caused by the simultaneous glial cell activation. Laser capture microdissection (LCM) allows dissection of distinct cell layers or single cells from tissue sections, allowing the collection of samples that are more specific compared to whole retina analyses, as recently demonstrated for advanced RGC injury in a glaucoma model (Guo et al., 2010). We therefore used this technique in the current project. 2. Material and methods 2.1. Animals used A total of 23 male Wistar rats weighing 250e400 g were used. All animals were treated in accordance to the NIH Guide for the Care and Use of Laboratory Animals, the EC Directive 86/609/EEC for animal experiments and the UK Home Office regulations. The animals were housed with a 14 h light/10 h dark cycle and with chaw and water ad libidum. 2.2. Intraocular pressure elevation All procedures were carried out under general anaesthesia. A mixture of ketamine (50 mg/kg body weight) and xylazine (5 mg/kg body weight) was administered intraperitoneally and 1% topical tetracaine was applied to the eye. In each animal only the left eye was treated. The right eye served as matched control. Chronic IOP elevation was obtained by external 532 nm diode laser treatment to the aqueous outflow tract as previously described (Levkovitch-Verbin et al., 2002). Briefly, laser energy was delivered to the trabecular meshwork at the slit lamp without the use of additional lenses. Initial treatment was 40e60 spots of 50e75 mm size, 0.6 Watt power and 0.6 s duration. Treatment was repeated at 1 week if the difference in IOP between the two eyes was less than 6 mmHg. IOP was measured under general anaesthesia using the TonoLab device (Tiolat Oy, Helsinki, Finland). A minimum of 3 sets of readings, each consisting of 6 measurements, were used. Measurements were taken immediately before, one day and one week after each treatment, then weekly for the duration of the experiment. The last IOP value was obtained immediately before killing the animal. For gene expression studies three groups with different duration of IOP elevation were studied: 1 week (n ¼ 4), 2 weeks (n ¼ 4) and 4

139

weeks (n ¼ 3). For immunostaining a control group (n ¼ 3) was compared to groups with 1 week (n ¼ 4) or 4 weeks (n ¼ 4) of IOP elevation. 2.3. Retina explantation and cryosectioning for gene expression studies Animals were killed by increasing CO2 concentrations followed by neck dislocation. The eyes were enucleated, the anterior segment removed and the retina dissected in ice cold DEPC-treated PBS. The whole retina was placed in OCT embedding medium (Raymond A. Lamb UK, Eastbourne, UK) and immediately frozen in liquid nitrogen. Samples were stored at 80  C. Prior to cryosectioning, RNAse-free polyethylene naphthalate PALM membrane slides (P.A.L.M. Microlaser Systems, Bernried, Germany) were UV treated (1200 mJ for 20 min) and precooled on dry ice. Approximately 40 sections of 18 mm thickness were collected while cutting through the whole tissue block in order to obtain samples from all parts of the retina. Slides were stored at 80  C. 2.4. Fixation and rapid immunostaining To identify RGCs for laser capture, cryosections were immunolabelled for beta-III-tubulin, a selective marker for RGCs in the ganglion cell layer (Fournier and McKerracher, 1997; Cui et al., 2003). We chose not to use retrograde labelling of RGCs for identification because the labelling process itself causes neuronal damage in the superior colliculus which could potentially alter RGC gene expression. All solutions for fixation and washing were made up with DEPCtreated water and fresh batches of ethanol. For fixation the frozen sections on the membrane slides were fixed in 70% ethanol for 5 min at room temperature. The primary antibody for immunostaining (monoclonal anti-beta-III-tubulin, Promega UK, Southampton, UK) was diluted 1:100 in 20% BSA and 500U/ml SuperaseIn RNAse inhibitor (Ambion/Applied Biosystems, Warrington, UK) in PBS and incubated for 5 min 2nd antibody (Anti-mouse biotin conjugated (Vector laboratories, Peterborough, UK)) was diluted 1:200 in PBS and 200 U/ml SuperaseIn and incubated for 3 min 4 standard concentration of avidin-HRP (ABC kit, Vector laboratories) with 200 U/ml SuperaseIn was incubated for 3 min. Colour precipitation was obtained using standard DAB solution (Sigmafast tablets, SigmaeAldrich, Gillingham, UK). Subsequently, sections were counterstained with hematoxylin (Gill No 1, SigmaeAldrich) containing 200 U/ml SuperaseIn and dehydrated by increasing ethanol concentrations followed by immersion in xylene for 5 min and air drying for 5 min. 2.5. Laser capture microdissection and total RNA extraction Laser capture microdissection was performed on a PALM LCM device (P.A.L.M. Microlaser Systems). First, 100 individual retinal ganglion cells positive for beta-III-tubulin were dissected and directly collected in lysis buffer (RNAqueous micro kit, Ambion). Care was taken to capture cells from different sections and different areas within one section to minimise sampling error as glaucoma damage is known to be focal rather than diffuse. After collection and pooling of 100 single RGCs the remaining sections from the same eye were used to dissect out 100 roughly equally sized pieces of the ganglion cell layer. These samples were again captured from several different sections and collected in a cap with fresh lysis buffer. The resulting harvest from each retina was one sample of 100 pooled RGCs and another sample of 100 pooled pieces of the GCL.

140

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

Total RNA extraction and DNAse treatment were carried out following the manufacturer’s instructions (RNAqueous micro kit, Ambion). RNA concentration was determined using a Nanodrop ND-1000 spectrometer (Fisher Scientific, Loughborough, UK). 2.6. RNA amplification As the amount of RNA obtained from 100 pooled laser captured RGCs was too low for immediate reverse transcription the mRNA was amplified in two cycles. The amplification and purification were performed as previously described (Tung et al., 2008). 2.7. Reverse transcription of RNA cDNA synthesis was performed using Superscript III reverse transcriptase (Invitrogen) with a larger reaction volume (100 ml) than the manufacturer’s standard recommendation. 3 mg of total RNA or 0.1 mg of amplified RNA were mixed with 750 ng random hexamers (Invitrogen) plus 5 ml 40 mM (10 mM each) dNTP mix (Bioline) in 40 ml water and incubated at 65  C for 5 min. Then, 5 first strand buffer (Invitrogen), DTT (Invitrogen; final concentration 10 mM) Superscript III RT (250U) and RNAse Out (40U, Invitrogen) were added and the reaction incubated for 10 min at 25  C, 60 min at 50  C and 15 min at 70  C. 2.8. PCR Each PCR reaction was performed in duplicate. For a reaction volume of 12.5 ml, 1 ml of cDNA was mixed with the following reagents (final concentration in brackets): DNA polymerase buffer (1), dNTP (0.3 mM each), MgCl2 (1.5 mM) and DNA polymerase (1.25U; all from Bioline). Primers (6.5pM each primer per reaction) were: DYNC1-I1 (syn. IC-1) forward: GAAGCTGGAAGCCAAGACGATC, reverse: ACCTGGGTA ACCTTTGACACGC; DYNC2-I2 (syn. IC-2) forward: CGAAGCCTTGCTT CAGAGCATG, reverse: TCTTCCTCCTCCTCATCTTCTTTGG or alternative reverse: TGGCTTCCAGCTTCACTTGGC (all following Pfister et al., 1996); GFAP forward: AGAAAACCGCATCACCATTC, reverse: TCCTTA ATGACCTCGCCATC; CD11b forward: CGCAAGAACACCAAGGAC, reverse: AGAAGGCTCGGACAACTG. PCR reactions were mixed 1:6 with Orange G loading buffer (0.35% Orange G (Sigma), 3.33% Glycerol in TBE buffer) and separated on a 2.8% agarose gel. Documentation with the GelDoc imager and Quantity One software (both Bio-Rad, Hemel Hempstead, UK). For quantitative expression analyses background subtracted quantification of the band intensity was performed with ImageJ software (National Institutes of Health, Bethesda, Maryland, USA). The values for the duplicates were averaged.

5-fold dilution series (depending on the abundance of the gene of interest) of rat brain cDNA were used. The relative quantity (RQ) of expression was calculated as the ratio between the expression level of the gene of interest in glaucoma versus control eyes (DDCt method). The expression levels in glaucoma and control tissue were normalised to beta-actin, which was previously used as housekeeping gene in the same glaucoma model (Levkovitch-Verbin et al., 2006). PCR efficiency of each primer pair as derived from standard curves was used as a correction factor in DCt-calculation. 2.10. Immunostaining Animals whose eyes were used for immunostaining were deeply anaesthetised by a mixture of ketamine (50 mg/kg body weight) and xylazine (5 mg/kg body weight) injected intraperitoneally followed by exsanguination and transcardial perfused with 4% paraformaldehyde in PBS. The eyes were removed, postfixed in 4% paraformaldehyde, cryoprotected in 30% sucrose in PBS and cryosectioned to 14 mm. Immunostainings were performed as previously described (Bull et al., 2008) using the following antibodies: DYNLL mouse polyclonal antibody, recognising both DYNLL1 and 2, diluted 1:200 (Abnova, Heidelberg, Germany), GFAP rabbit polyclonal 1:500 (DAKO, Ely, UK) and beta-III-tubulin rabbit polyclonal 1:2000 (Covance, UK) all in PBS 4% goat serum, 0.3% Triton X-100. Secondary antibodies were Alexa Fluor 488 and 568 (Invitrogen) diluted 1:500 in PBS 4% goat serum. Stained sections were scanned using a fluorescence microscope with fixed settings for all images (Leica FW4000, Leica Inc., Wetzlar, Germany). Quantification of immunofluorescence intensity was performed on 20 images per eye (4 eyes per experimental condition) taken from different locations within the retina. By using ImageJ software between 234 and 275 individual RGCs (as determined by beta-IIItubulin staining) per experimental condition were manually outlined and mean DYNLL-stain pixel intensity was measured. For correction of background fluorescence variability a correction factor calculated from the mean background fluorescence intensity of all images and the background intensity of the individual image was used. 2.11. Statistical analyses Relative quantity of expression in glaucoma versus control eyes was analysed using the one sample t-test. As three timepoints were measured for each gene, a Bonferroni correction was performed. For immunostaining quantification a t-test was used. 3. Results

2.9. Quantitative real-time PCR 3.1. Intraocular pressure exposure QPCR was performed on an iCycler iQ with software version 3.1 (Bio-Rad). iQ Sybr green super mix (Bio-Rad) was used for all reactions. Primers were designed with Beacon Designer 7.0 software (Premier Biosoft, Palo Alto, USA) and manufactured by Sigma. Primer sequences were: Dynein intermediate chain IC-1 all isoforms: forward GCAGTGTTTGGTGTCCTTTC, reverse GGTGGATTT CGGGTGAGTAG; DYNLL1: forward AGAAGGAGTTTGACAAGAAG, reverse AGATTTGAACAGGAGAATGG; DYNLL2: forward TGCGAG GACAGGGTTAGGG, reverse AGCCATAGCGTCACAGTTCAG; betaactin: forward GTCCACCTTCCAGCAGATG, reverse CTCAGTAACA GTCCGCCTAG. Final volume was 25 ml and each cDNA sample was analysed in triplicate. For standard curves to calculate the PCR efficiency 10- or

IOP elevation was achieved in all laser treated eyes (Table 1). As is characteristic for the glaucoma model we used, there was a biphasic pressure rise within 2 weeks followed by a decrease in pressure back to normal values within 3e4 weeks. This method of laser treatment has been shown to result in progressive RGC death (Levkovitch-Verbin et al., 2002; Bull et al., 2009; Johnson et al., 2010). 3.2. Purity of the laser captured retinal ganglion cell samples To identify RGCs for collection with the laser capture microscope we performed a rapid immunostaining for beta-III-tubulin immediately before the LCM dissection, which is a specific marker for

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

141

Table 1 IOP exposure of glaucoma and control eyes (mean and SEM). Mean peak IOP is the mean of the IOP peaks that occurred after each of the two laser treatments. Time (weeks)

1 2 4

n

8 5 7

Mean IOP (mmHg)

Mean peak IOP (mmHg)

Glaucoma

Control

Difference

Glaucoma

Control

Difference

25.2  7.0 19.9  2.2 16.3  2.7

10.7  0.5 11.2  0.9 11.4  0.9

14.5  6.7 8.7  1.7 4.9  2.7

31.2  8.5 29.7  4.9 27.3  6.5

10.6  0.6 11.8  0.6 11.2  1.7

20.5  8.2 17.9  4.5 16.1  6.0

RGCs (Fournier and McKerracher, 1997; Cui et al., 2003). We verified the purity of our samples by performing PCRs for GFAP (astrocytes, Müller cells) and CD11b (microglia) on the cDNA derived from pooled single RGCs. All samples were negative for these genes (data not shown). 3.3. Dynein light chain expression in the ganglion cell layer The two members of the LC8 family of dynein light chains, DYNLL1 and 2, are expressed by two different genes. In the laser captured ganglion cell layer of untreated control eyes both genes were expressed at almost equal levels (Fig. 1a). However, analysis of pure RGC-cDNA derived from laser capture microdissection revealed that ganglion cells express only the DYNLL1 but not the DYNLL2 light chain (Fig. 1b, showing an exemplary amplification curve). Glaucomatous stress did not induce a gain of DYNLL2 expression, as in cDNA from laser captured RGCs derived from 1 and 4 week glaucoma eyes no expression of DYNLL2 could be detected (data not shown). 3.4. DYNLL1 expression in glaucoma eyes The stability of expression of the housekeeping gene beta-actin was tested on cDNA derived from laser captured RGCs. There was no statistically significant difference in the expression levels between the glaucoma and control eyes used (RQ 0.92  0.053; p ¼ 0.24; n ¼ 4). Induction of glaucoma led to an early downregulation of DYNLL1 expression in the ganglion cell layer (Fig. 2a). The amount of DYNLL1 as measured by quantitative real-time PCR dropped significantly to an RQ of 0.27 (p ¼ 0.0003) at one week post-glaucoma compared to control GCL and beta-actin expression. With persisting ocular hypertension after two weeks the relative DYNLL1 expression remained significantly lower compared to beta-actin (RQ 0.53 (p ¼ 0.012)). After 4 weeks, when the intraocular pressure had

returned to control eye values in all laser treated eyes, the DYNLL1 RQ was still lower compared to beta-actin although this value was not statistically significant (0.57; p ¼ 0.063). To clarify whether the lower DYNLL1 expression observed at the 4 week timepoint is due to persistent downregulation or simply reflects loss of RGCs in the laser captured ganglion cell layer, we analysed pure RGC-cDNA from two 4 week animals. Similar to results obtained for the ganglion cell layer, we observed DYNLL1 expression in purified control RGCs, while at 4 weeks after glaucoma, DYNLL1 was below detection limits (see Supplemental Figure 1 for real-time PCR amplification curves of DYNLL and beta-actin). The difference between low but detectable DYNLL1 expression levels in the GCL-derived cDNA and no detectable expression in the RGC-derived cDNA may arise from the fact that RGC-derived cDNA is transcribed from amplified RNA, as the amount of RNA harvested from 100 laser captured RGCs was relatively low. The amplification procedure magnifies differences in expression levels, resulting in the case of DYNLL1 in a ‘negative’ magnification of the DYNLL1 downregulation in RGC-derived cDNA compared to GCL-derived cDNA. Also, DYNLL1 may be co-expressed with DYNLL2 in glia cells which could maintain a certain level of DYNLL1 expression in the GCL. 3.5. DYNLL2 expression in glaucoma eyes 3.5.1. DYNLL2 expression in the GCL and DYNLL protein abundance in non-RGC cells As for DYNLL1, we measured DYNLL2 expression in the ganglion cell layer of glaucoma eyes. There was a significant upregulation of DYNLL2 in the GCL after two weeks of glaucoma (relative quantity 1.56, p ¼ 0.038; Fig. 2b). As RGCs remain DYNLL2 negative even under glaucomatous stress (see above) the upregulation must occur in a different cell population within the ganglion cell layer. Immunolabelling of control and glaucoma retinae using an antiDYNLL-antibody revealed that the only cell population in the

Fig. 1. a) Expression of the DYNLL1 and 2 genes in the ganglion cell layer laser captured from control retinae of nine rats. The mean threshold cycle number and SEM of the real-time PCR amplification is plotted for both genes. The difference is not significantly different. b) Expression of DYNLL1 and 2 in retinal ganglion cells of control eyes. Exemplary amplification curves of real-time PCR for both genes from one control eye are shown. Each sequence is amplified in triplicate. There is good amplification of DYNLL1 but not DYNLL2.

142

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

Fig. 2. aee) Expression of different genes in the glaucomatous ganglion cell layer. The DYNLL1, DYNLL2 and Dynein IC-1 data was obtained by real-time PCR, GFAP and Dynein IC-2B expression was quantified by densitometry of PCR gel electrophoresis bands and normalised to beta-actin. The charts show mean and SEM of the quantity of expression of the gene of interest in individual glaucoma eyes relative to the matched control eye and normalised to beta-actin. a) DYNLL1 expression is lower than in control eyes at all timepoints. The difference is statistically significant after 1 week (RQ 0.27; p ¼ 0.0003) and 2 weeks (RQ 0.53; p ¼ 0.012), but not after 4 weeks (RQ 0.57; p ¼ 0.063). b) DYNLL2 expression in the ganglion cell layer rises at 2 weeks (RQ 1.56; p ¼ 0.038). The 1 and 4 week timepoints are not significantly different (1 week RQ 0.94; p ¼ 0.90; 4 weeks RQ 1.29; p ¼ 0.24). c) GFAP expression in the ganglion cell layer. The upregulation starts after 1 week and diminishes after 4 weeks (1 week RQ 4.44; p ¼ 0.15; 2 weeks RQ 3.31; p ¼ 0.017; 4 weeks RQ 2.18; p ¼ 0.08). d) Dynein intermediate chain (IC-) 1 gene (all 3 isoforms) expression did not change significantly over time. At 1 week after glaucoma induction the expression appears reduced but the change did not reach significance (RQ 0.71; p ¼ 0.15). e) The dynein intermediate chain (IC-) 2B isoform, the dominant gene 2 isoform in RGCs (see Fig. 2f) was significantly downregulated at both timepoints analysed: 1 and 4 weeks (1 week RQ 0.14; p ¼ 0.016; 4 weeks RQ 0.22; p ¼ 0.015). f) Expression of the different dynein intermediate chain isoforms in control brain and different control retina compartments. The neuronal isoform 1B, responsible for TrkB receptor transport is the main IC-1 isoform in RGCs. The neuronal isoform 2B is the dominant gene 2 isoform in RGCs. The isoform 2C, known to be expressed in neurons and glia in hippocampal tissue (Pfister et al., 1996) was amplified much less in the pure RGC samples than in whole retina and GCL.

ganglion cell layer exhibiting elevated DYNLL protein abundance are glia cells (Fig. 3).

DYNLL2 expression was delayed, relative to the onset of glial reactivity.

3.5.2. Timecourse of astroglial response in the GCL Quantitative PCR for GFAP was carried out to determine onset and magnitude of the astroglial response in our glaucoma model. There was a marked increase in GFAP expression already at the 1 week timepoint and also after 2 weeks (Fig. 2c). The increase in

3.6. DYNLL1 protein abundance in RGCs of glaucoma eyes By immunostaining RGCs in retina sections of glaucoma and control eyes with an anti-DYNLL-antibody combined with quantification of the staining intensity of individual RGCs we could

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

143

Fig. 3. Immunohistochemistry for DYNLL(1 and 2) and GFAP demonstrating the increase in DYNLL abundance in the glaucomatous nerve fibre layer (2 weeks glaucoma) due to glia activation. Marked with an arrow are putative astrocyte cell bodies, as identified by GFAP and DAPI staining. Müller cell end feet also exhibit higher DYNLL abundance in glaucoma compared to control tissue. Scalebar: 25 mm. No nonlinear adjustments were made to the images.

demonstrate that the impaired gene expression was followed by a reduction of DYNLL protein abundance even in neurons that strongly stained for beta-III-tubulin (Fig. 4). This reduction was statistically significant after 4 weeks (p ¼ 0.039; t-test). As only DYNLL1 is expressed in RGCs the quantification was specific for this DYNLL form although an antibody recognising both DYNLL1 and 2 was used.

3.7. Expression of the TrkB-transporting dynein IC-1 isoforms in the GCL The dynein intermediate chains (CD-IC) are the central subunit of the dynein complex and binding partners of DYNLL1. Six different splice variants are produced from two different genes (IC-1 and -2). The neuronal IC-1 gene products have recently been shown to

144

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

Fig. 4. DYNLL immunostaining in RGCs 1 and 4 weeks after glaucoma induction versus control eyes. aei) Representative images of the inner retina. Columns are showing DYNLL (left), the RGC marker beta-III-tubulin (centre), DYNLL and beta-III-tubulin merged (right). Rows are showing different durations of glaucoma: untreated control eyes (top), 1 week of glaucoma (centre) and 4 weeks of glaucoma (bottom). j) Quantification of DYNLL immunostaining intensity at the above stated timepoints. The total number of cells analysed in 4 eyes per timepoint was: control ¼ 271, 1 week ¼ 275, 4 weeks ¼ 234. DYNLL expression was significantly reduced in RGCs at 4 weeks post-glaucoma versus control (p ¼ 0.039). Scalebar: 25 mm. No nonlinear adjustments were made to the images.

preferentially execute retrograde transport of TrkB-neurotrophin signalling endosomes (Ha et al., 2008). The IC-2C isoform is ubiquitously expressed and much less involved in neurotrophic signalling. We investigated whether both isoforms are present in RGCs and whether CD-IC expression is downregulated in our glaucoma model similarly to DYNLL1. As previously described for brain tissue (Pfister et al., 1996) we detected all isoforms in the retina. While the expression ratio between the IC-1 isoforms was similar to brain tissue the isoforms 2B and 2C, which are equally expressed in brain, were differently expressed in the retina with 2B being the dominant isoform. In pure RGCs cDNA the ubiquitous 2C isoform was expressed at very low levels (Fig. 2f). Conventional PCR analysis showed that the more abundant IC-2 isoform 2B was downregulated at 1 and 4 week

timepoint (Fig. 2e). No compensatory upregulation of the very weakly expressed ubiquitous 2C isoform, which is potentially able to execute neurotrophin signalling, was observed (data not shown). Quantitative real-time PCR for dynein IC-1 in the ganglion cell layer (Fig. 2d) showed a small, but not significant, decrease relative to control eyes after 1 week of experimental glaucoma (mean RQ: 0.71; p ¼ 0.15). At later timepoints the expression levels in glaucoma eyes returned to control levels. Thus, the TrkB-transporting dynein intermediate chain 1 is more stably expressed compared to DYNLL1. 4. Discussion By quantitative analysis of gene expression in the ganglion cell layer and in pure RGC samples we found that moderate chronic

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

glaucomatous stress induces a strong and persistent downregulation of the dynein light chain DYNLL1, which is the only DYNLL- (LC8-) type light chain expressed in RGCs. The reduction in DYNLL1 expression occurred early after IOP elevation and persisted for a remarkably long time. It even remained depressed at 4 weeks, a timepoint when the IOP had returned to normal levels in all animals in the glaucoma model used. The onset of DYNLL1 downregulation in our model occurs very early as after 1 week the number of DYNLL1 transcripts was already significantly lower (RQ 0.27) than in control GCL. The reduction in DYNLL1 protein abundance occurs with a delay and was significantly lower after 4 weeks. Levkovitch-Verbin et al. (2006) have investigated the timecourse of cleaved capsase-3 abundance, an early event in apoptosis, in the same laser glaucoma model. Onset of caspase-3 cleavage was not detected before 2 weeks after glaucoma induction and, like DYNLL1, remained in the altered state at 4 weeks when IOP has returned to normal. Thus, there is an overlap in the timecourse of DYNLL1 reduction and the onset of apoptosis in the same glaucoma model. Whether the diminution of DYNLL1 alone is capable of inducing apoptosis in RGCs remains unknown. There is evidence that a reduction of DYNLL1 causes neuronal dysfunction and can trigger neuron death. According to the relatively broad role of DYNLL1 as a “universal protein hub” (Barbar, 2008), and as a stabiliser and regulator of different proteins there is not one but several pathways that link DYNLL1 reduction and neuron dysfunction. Studies in PC12 neurons (Chang et al., 2000) showed a neuroprotective effect of DYNLL1 for neurons starved from neurotrophins. This effect was mediated by inhibition of NO-synthases, which is in conjunction with other publications on NO and RGC death (Klocker et al., 1998; Manabe et al., 2005; Tezel, 2006). Recent work from Chen et al. (2009) revealed a role for DYNLL1 in regulating axonal immobilisation of mitochondria. Using primary cultured hippocampal neurons and COS7 cells they demonstrated that DYNLL1 regulates syntaphilin, a specific anchor for mitochondria on axonal microtubuli. Dysfunctional DYNLL1 reduced syntaphilin mediated immobilisation of mitochondria. Thus, the physiological ability to recruit mitochondria to zones of increased energy demand and Ca2þ levels within the axon is perturbed. Interestingly, the healthy human optic nerve exhibits a marked heterogeneity in mitochondria distribution (accumulation in the unmyelinated proximal nerve) which is lost in glaucoma (Barron et al., 2004). The proapoptotic factor Bim, which also occurs in RGCs (Guo et al., 2009), is also a DYNLL1 binding partner. Experiments in mammary gland cells with dysfunctional DYNLL1 (Song et al., 2008) showed increased intracellular levels of Bim and higher apoptosis rates. DYNLL1 binds to the dynein intermediate chain (IC) within the dynein complex. This interaction has been characterised in great detail and is likely to be representative also for the other binding partners (Nyarko et al., 2004; Benison et al., 2006), leading to an increase in dynein-IC secondary structure that amplifies dimerisation. As the IC chains occur as a dimer (Myers et al., 2007) reduced DYNLL1 abundance is likely to result in an increased number of dysfunctional dynein complexes. Unlike DYNLL1, the expression of the dynein intermediate chains did not change dramatically. There was a moderate downregulation of the IC-1 chains, which are responsible for neurotrophin signalling, in all 4 samples after 1 week but not at later timepoints. Thus, we conclude that functional impairment of the dynein-IC-1 protein (Martin et al., 2006) rather than reduced expression levels of dyneinIC-1 may contribute to RGC death in our glaucoma model. There is no published data on DYNLL1 in glaucoma and only two publications on DYNLL in neurodegenerative diseases. Gillardon et al. (1998) reported an upregulation of DYNLL1 in ischemiaresistant brain areas but unchanged levels in vulnerable areas after

145

transient global ischemia in the rat, suggesting a neuroprotective effect of relatively higher DYNLL1 levels. Dietz et al. (2005) observed no change in gene expression and protein levels 6 days after optic nerve transection in rat. Thus, our work demonstrates for the first time a downregulation of DYNLL1 in the context of a neurodegenerative disease. However, whether DYNLL1 diminution triggers neuron death in an early phase of stress or only accelerates cell death in already lethally injured neurons remains to be clarified by future studies including selective silencing of DYNLL1 in cultured retinal ganglion cells. Our data indicates that DYNLL2 is expressed in glia cells. Whether DYNLL1 is co-expressed or not was not determined. Although the function of DYNLL2 is much less well characterised than DYNLL1, the large sequence similarity (both 89 amino acids; 93% sequence similarity) suggests a very similar role in the cell. While DYNLL1 binds to the pro-apoptotic Bcl-2 family protein Bim, DYNLL2 binds to Bmf, also a pro-apoptotic Bcl-2 protein (Day et al., 2004). Analogous to the connection between low DYNLL1 levels and the occurrence of apoptosis, higher levels of DYNLL2 may coincide with anti-apoptotic conditions. In addition glial cells, the only proliferating retinal cell population in our glaucoma model, upregulate DYNLL2 after 2 weeks of glaucoma. GFAP expression indicates an onset of glia activation already after 1 week, thus DYNLL2 occurs with a delay when glia proliferation is already initiated. The fact that two DYNLL genes exist which are differently expressed in a mixed cell compound and whose gene products are very similar and thus difficult to distinguish by antibodies makes the expression analysis technically demanding. In this case laser capture microdissection (LCM) has a great advantage over the analysis of whole retina lysate. The use of in situ hybridisation would have allowed a more precise determination of the retinal cell types expressing one or the other DYNLL form but quantification of expression levels would have been imprecise. However, LCM is a challenging technique, highly labour intensive and relatively expensive. These factors, combined with strict limits on animal use in the UK, meant that the number of animals used in our study was relatively small. Nevertheless, it is clear that the study design was adequate to detect large differences in gene expression that would be predicted to be biologically important. With ongoing glaucomatous neurodegeneration the changes in gene expression observed is restricted to the surviving cells. In order to reliably analyse changes in gene expression the choice of a relatively stably expressed housekeeping gene is critical. On the basis of sample tests, we chose beta-actin as the most stably expressed from three housekeeping candidate genes (beta-actin, HPRT and GAPDH). Analysing multiple housekeeping genes for each sample could potentially have increased the reliability of the normalisation process (Vandesompele et al., 2002) but was impossible due to the very limited amount of cDNA available from laser captured tissue. The fact that we have observed a gene upregulation for DYNLL2 provides further reassurance that the gene downregulation observed for DYNLL1 was not just an effect of global reduction of expression in dying RGCs. The strong and persistent downregulation of DYNLL1 suggests that it may have detrimental effects on neuron function in our glaucoma model. To what extent and at which time point of the neuronal death process the diminution of DYNLL1 has its main effect will be subject of future studies.

Acknowledgments The authors thank Aviva M. Tolkovsky for critically reading of the manuscript.

146

C. van Oterendorp et al. / Experimental Eye Research 92 (2011) 138e146

This work was funded by Deutsche Forschungsgemeinschaft e DFG (CvO) and the GSK Clinician Scientist Program (KRM). Appendix. Supplementary material Supplementary information associated with this article can be found in the online version, at doi:10.1016/j.exer.2010.11.013. References Barbar, E., 2008. Dynein light chain LC8 is a dimerization hub essential in diverse protein networks. Biochemistry 47, 503e508. Barron, M.J., Griffiths, P., Turnbull, D.M., Bates, D., Nichols, P., 2004. The distributions of mitochondria and sodium channels reflect the specific energy requirements and conduction properties of the human optic nerve head. Br. J. Ophthalmol. 88, 286e290. Benison, G., Nyarko, A., Barbar, E., 2006. Heteronuclear NMR identifies a nascent helix in intrinsically disordered dynein intermediate chain: implications for folding and dimerization. J. Mol. Biol. 362, 1082e1093. Bull, N.D., Irvine, K.A., Franklin, R.J., Martin, K.R., 2009. Transplanted oligodendrocyte precursor cells reduce neurodegeneration in a model of glaucoma. Invest. Ophthalmol. Vis. Sci. 50, 4244e4253. Bull, N.D., Limb, G.A., Martin, K., 2008. Human Muller stem cell (MIO-M1) transplantation in a rat model of glaucoma: survival, differentiation and integration. Invest. Ophthalmol. Vis. Sci.. Chang, Y.W., Jakobi, R., McGinty, A., Foschi, M., Dunn, M.J., Sorokin, A., 2000. Cyclooxygenase 2 promotes cell survival by stimulation of dynein light chain expression and inhibition of neuronal nitric oxide synthase activity. Mol. Cell Biol. 20, 8571e8579. Chen, Y.M., Gerwin, C., Sheng, Z.H., 2009. Dynein light chain LC8 regulates syntaphilin-mediated mitochondrial docking in axons. J. Neurosci. 29, 9429e9438. Cui, Q., Yip, H.K., Zhao, R.C., So, K.F., Harvey, A.R., 2003. Intraocular elevation of cyclic AMP potentiates ciliary neurotrophic factor-induced regeneration of adult rat retinal ganglion cell axons. Mol. Cell. Neurosci. 22, 49e61. Day, C.L., Puthalakath, H., Skea, G., Strasser, A., Barsukov, I., Lian, L.Y., Huang, D.C., Hinds, M.G., 2004. Localization of dynein light chains 1 and 2 and their proapoptotic ligands. Biochem. J. 377, 597e605. Dietz, G.P., Schott, M., Labes, M., Bahr, M., 2005. Expression of the protein inhibitor of nitric oxide synthase in the adult rat retina before and after optic nerve lesion. Brain Res. Mol. Brain Res. 136, 118e124. Fan, J., Zhang, Q., Tochio, H., Li, M., Zhang, M., 2001. Structural basis of diverse sequence-dependent target recognition by the 8 kDa dynein light chain. J. Mol. Biol. 306, 97e108. Fournier, A.E., McKerracher, L., 1997. Expression of specific tubulin isotypes increases during regeneration of injured CNS neurons, but not after the application of brain-derived neurotrophic factor (BDNF). J. Neurosci. 17, 4623e4632. Gillardon, F., Krep, H., Brinker, G., Lenz, C., Bottiger, B., Hossmann, K.A., 1998. Induction of protein inhibitor of neuronal nitric oxide synthase/cytoplasmic dynein light chain following cerebral ischemia. Neuroscience 84, 81e88. Guo, Y., Cepurna, W.O., Dyck, J.A., Doser, T.A., Johnson, E.C., Morrison, J.C., 2010. Retinal cell responses to elevated intraocular pressure: a gene array comparison between the whole retina and retinal ganglion cell layer. Invest. Ophthalmol. Vis. Sci. 51, 3003e3018. Guo, Y., Johnson, E., Cepurna, W., Jia, L., Dyck, J., Morrison, J.C., 2009. Does elevated intraocular pressure reduce retinal TRKB-mediated survival signaling in experimental glaucoma? Exp. Eye Res. 89, 921e933.

Ha, J., Lo, K.W., Myers, K.R., Carr, T.M., Humsi, M.K., Rasoul, B.A., Segal, R.A., Pfister, K.K., 2008. A neuron-specific cytoplasmic dynein isoform preferentially transports TrkB signaling endosomes. J. Cell. Biol. 181, 1027e1039. Heerssen, H.M., Pazyra, M.F., Segal, R.A., 2004. Dynein motors transport activated Trks to promote survival of target-dependent neurons. Nat. Neurosci. 7, 596e604. Johnson, T.V., Bull, N.D., Hunt, D.P., Marina, N., Tomarev, S.I., Martin, K.R., 2010. Neuroprotective effects of intravitreal mesenchymal stem cell transplantation in experimental glaucoma. Invest. Ophthalmol. Vis. Sci. 51, 2051e2059. Klocker, N., Cellerino, A., Bahr, M., 1998. Free radical scavenging and inhibition of nitric oxide synthase potentiates the neurotrophic effects of brain-derived neurotrophic factor on axotomized retinal ganglion cells in vivo. J. Neurosci. 18, 1038e1046. Levkovitch-Verbin, H., Dardik, R., Vander, S., Nisgav, Y., Kalev-Landoy, M., Melamed, S., 2006. Experimental glaucoma and optic nerve transection induce simultaneous upregulation of proapoptotic and prosurvival genes. Invest. Ophthalmol. Vis. Sci. 47, 2491e2497. Levkovitch-Verbin, H., Quigley, H.A., Martin, K.R., Valenta, D., Baumrind, L.A., Pease, M.E., 2002. Translimbal laser photocoagulation to the trabecular meshwork as a model of glaucoma in rats. Invest. Ophthalmol. Vis. Sci. 43, 402e410. Libby, R.T., Li, Y., Savinova, O.V., Barter, J., Smith, R.S., Nickells, R.W., John, S.W., 2005. Susceptibility to neurodegeneration in a glaucoma is modified by Bax gene dosage. PLoS Genet. 1, 17e26. Manabe, S., Gu, Z., Lipton, S.A., 2005. Activation of matrix metalloproteinase-9 via neuronal nitric oxide synthase contributes to NMDA-induced retinal ganglion cell death. Invest. Ophthalmol. Vis. Sci. 46, 4747e4753. Martin, K.R., Quigley, H.A., Valenta, D., Kielczewski, J., Pease, M.E., 2006. Optic nerve dynein motor protein distribution changes with intraocular pressure elevation in a rat model of glaucoma. Exp. Eye Res. 83, 255e262. Myers, K.R., Lo, K.W., Lye, R.J., Kogoy, J.M., Soura, V., Hafezparast, M., Pfister, K.K., 2007. Intermediate chain subunit as a probe for cytoplasmic dynein function: biochemical analyses and live cell imaging in PC12 cells. J. Neurosci. Res. 85, 2640e2647. Nyarko, A., Hare, M., Hays, T.S., Barbar, E., 2004. The intermediate chain of cytoplasmic dynein is partially disordered and gains structure upon binding to light-chain LC8. Biochemistry 43, 15595e15603. Pfister, K.K., Salata, M.W., Dillman 3rd, J.F., Vaughan, K.T., Vallee, R.B., Torre, E., Lye, R.J., 1996. Differential expression and phosphorylation of the 74-kDa intermediate chains of cytoplasmic dynein in cultured neurons and glia. J. Biol. Chem. 271, 1687e1694. Phillis, R., Statton, D., Caruccio, P., Murphey, R.K., 1996. Mutations in the 8 kDa dynein light chain gene disrupt sensory axon projections in the Drosophila imaginal CNS. Development 122, 2955e2963. Quigley, H.A., 2005. New paradigms in the mechanisms and management of glaucoma. Eye 19, 1241e1248. Quigley, H.A., Nickells, R.W., Kerrigan, L.A., Pease, M.E., Thibault, D.J., Zack, D.J., 1995. Retinal ganglion cell death in experimental glaucoma and after axotomy occurs by apoptosis. Invest. Ophthalmol. Vis. Sci. 36, 774e786. Song, C., Wen, W., Rayala, S.K., Chen, M., Ma, J., Zhang, M., Kumar, R., 2008. Serine 88 phosphorylation of the 8-kDa dynein light chain 1 is a molecular switch for its dimerization status and functions. J. Biol. Chem. 283, 4004e4013. Tezel, G., 2006. Oxidative stress in glaucomatous neurodegeneration: mechanisms and consequences. Prog. Retin. Eye Res. 25, 490e513. Tung, Y.C., Ma, M., Piper, S., Coll, A., O’Rahilly, S., Yeo, G.S., 2008. Novel leptinregulated genes revealed by transcriptional profiling of the hypothalamic paraventricular nucleus. J. Neurosci. 28, 12419e12426. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., Speleman, F., 2002. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 3 (7) RESEARCH0034.