Retinal degeneration modulates intracellular localization of CDC42 in photoreceptors

Animals and light exposure

All procedures were conducted in accordance with the guidelines published by the Institute for Laboratory Animal Research and with the regulations of the Veterinary Authority of Zurich. Cdc42 floxed mice (Cdc42^flox/flox) [23] were a generous gift from Dr. Joao Relvas (Institute of Cell Biology, ETH Zurich, Zurich, Switzerland). The 129S6/SvEvTac mice were purchased from Taconics (Eiby, Denmark). Mice with Cre recombinase expression controlled by the rod opsin promoter (LMOPC1, which will be called opsin-Cre from now on) [24] were provided by Dr. Yun Le (University of Oklahoma, Oklahoma City, OK). The mouse model for autosomal dominant RP (called VPP from now on) was kindly provided by Dr. Muna Naash (University of Oklahoma). Pde6brd10 [4] (called Rd10 from now on) and Pde6brd1 [5] (called Rd1 from now on) were purchased from Jackson laboratories (Bar Harbor, MA) and Harlan (Horst, Netherlands), respectively. A summary of mouse models for inherited retinal degeneration used in this study is given in Table 1. Rd1 and rd10 mice were homozygous for the Rpe65_450Met variant, whereas all other strains were homozygous for Rpe65_450Leu [25]. All mice were kept as breeding colonies at the animal facility of the University Hospital Zurich in a 12 h:12 h light-dark cycle with 60 lux at cage level. All experiments were performed with 12-week-old mice unless stated otherwise.

For light exposure experiments, mice were dark adapted overnight and pupils were dilated with 1% cyclopentolate (Cyclogyl; Alcon, Cham, Switzerland) and 5% phenylephrine (Ciba Vision, Niederwangen, Switzerland) 30 min before exposure to 13,000 lux of white light for 2 h, as described previously [26].

Generation of rod photoreceptor–specific Cdc42 knockdown mice and genotyping

For the experiments, Cdc42^flox/flox;opsin-Cre mice were crossed with Cdc42^flox/flox mice to obtain litters with Cdc42^flox/flox;opsin-Cre (rod photoreceptor–specific Cdc42 knockdown) and Cdc42^flox/flox (control mice). The late onset of Cre recombinase expression postnatally allowed a normal retinal development, permitting the use of the Cdc42^flox/flox;opsin-Cre to study the role of CDC42 in adult rod photoreceptor cells.

Genotypes were determined using genomic DNA from ear biopsies, and the following conventional PCR conditions: initial denaturation (95 °C, 5 min); 35 cycles of denaturation (95 °C, 45 s), annealing (60 °C, 45 s), and elongation (72 °C, 45 s); and final extension (72 °C, 10 min). For the detection of the floxed Cdc42 allele, primers (forward: 5′-TTC TTC CTC CAA CCT CCT GAT GGG-3′, reverse 5′-TGC TGT GTG TGG CAT TTG CTG C-3′) spanning both loxP sites were used. The amplicon of the floxed allele was 1.5 kb long, whereas amplification of the wild-type allele yielded a 1.3 kb fragment. Excision of the floxed sequence by the Cre recombinase resulted in an amplification product of 0.9 kb. For the detection of the Cre transgene, primers (forward: 5′-AGG TGT AGA GAA GGC ACT TAG C-3′, reverse 5′-CTA ATC GCC ATC TTC CAG CAG G-3′) specific for Cre were used. PCR products were run on a 1% agarose gel for size detection.

RNA preparation and semiquantitative RT–PCR

Retinas were removed through a slit in the cornea and immediately frozen in liquid nitrogen. Total RNA was prepared with an RNA isolation kit (RNeasy; Qiagen, Hilden, Germany) including a DNase treatment to remove residual genomic DNA. Identical amounts of RNA were used for reverse transcription using oligo(dT) and M-MLV reverse transcriptase (Promega, Madison, WI). Gene expression was analyzed by real-time PCR using specific primer pairs (Table 2) spanning an intronic region of the respective gene, a polymerase ready mix (LightCycler 480 SYBR Green I Master Mix; Roche Diagnostics, Indianapolis, IN), and a thermocycler (LightCycler Roche Diagnostics). Signals were normalized to Actb and relative expression was calculated by the comparative threshold cycle (ΔΔCT) method using a control sample for calibration [27].

Laser capture microdissection

Mouse eyes were enucleated and immediately frozen in tissue-freezing medium using a methylbutanol bath cooled by liquid nitrogen. Retinal sections (20 µm) were cut, fixed (5 min acetone), air dried (5 min), and dehydrated (30 s in 100% ethanol, 5 min in xylol). Microdissection of the retinal nuclear layers was performed with an Arcturus XT Lasercapture device (Molecular Devices, Silicon Valley, CA). RNA was isolated using the Arcturus RNA isolation kit (Molecular Devices), and residual genomic DNA removed by a DNase treatment. cDNA synthesis was performed as described in the previous section. Gene expression was determined by conventional PCR (40 cycles) or semiquantitative PCR using the primer pairs listed in Table 2. Amplified fragments were run on a nondenaturing polyacrylamide gel and detected after staining with ethidium bromide.

Light microscopy and spider diagram

Enucleated eyes were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) overnight at 4 °C. The cornea and lens were removed and the superior and inferior retina of each eye was prepared, washed in cacodylate buffer, incubated in osmium tetroxide for 1 h, dehydrated in a series of increasing ethanol concentrations, and embedded in Epon 812. Semithin cross-sections (500 nm) were prepared from the ventral central retina, the most affected region in our light-damage model [3]. Sections were counterstained with toluidine blue and analyzed by light microscopy.

The thickness of the photoreceptor nuclear layer was measured using the Adobe Photoshop CS3 ruler tool at 250, 500, 1,000, 1,500, 1,750, 2,000, and 2,250 μm distances from the optic nerve head in both the dorsal and ventral directions. Results of n=3 retinas were plotted as a spider diagram.

Immunofluorescence

Eyes were enucleated and fixed in 4% (wt/vol) paraformaldehyde in 0.1 M phosphate buffered saline (PBS; pH 7.4) overnight. After removing the cornea and lens, eyecups were postfixed in 4% paraformaldehyde for an additional 2 h before being immersed in 30% sucrose in PBS at 4 °C overnight. The eyes were embedded in tissue-freezing medium (Leica Microsystems Nussloch GmbH, Nussloch, Germany) and frozen in a 2-methylbutane bath cooled by liquid nitrogen. Retinal sections (12 µm) were cut, placed on slides and incubated with a blocking solution (3% normal goat serum, 0.3% Triton X-100 in 0.1M PBS, or 3% horse serum, 0.3% Triton X-100 in 0.1M PBS) for 1 h at room temperature. For protein detection, sections were incubated with primary antibodies (Table 3) diluted in blocking solution at 4 °C overnight. After three washes with PBS, slides were incubated with the appropriate secondary antibody coupled to Cy3 or Cy2 for 1 h at room temperature, washed, counterstained with 4',6-diamidino-2-phenylindole (DAPI), and mounted with antifade medium (10% Mowiol 4–88; vol/vol; Calbiochem, San Diego, CA), in 100 mM Tris (pH 8.5), 25% glycerol (wt/vol), and 0.1% 1,4-diazabicyclo (2.2.2) octane.

TUNEL assay and cell death detection

Detection of photoreceptor cell death in retinal sections was done using the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) method according to the manufacturer’s recommendations (Roche Diagnostics, Rotkreuz, Switzerland). Immunofluorescent signals were analyzed with a digital microscope (Axiovision; Carl Zeiss Meditec, Inc., Dublin, CA).

Apoptotic cell death was quantified 36 h after light exposure by measuring the release of free nucleosomes in isolated retinas using the cell death detection kit (Roche Diagnostic, Basel, Switzerland) according to the manufacturer’s instructions. Briefly, retinas were homogenized and the free nucleosomal content was quantified photometrically by a sandwich-enzyme-immunoassay principle using a spectrophotometer.

Western blot analysis

Retinas were homogenized in 0.1 M Tris/HCl (pH 8.0) and protein content was analyzed using Bradford reagent. Equivalent amounts of proteins were resolved by electrophoresis on sodium dodecyl sulfate–polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were blocked in 5% milk (Bio-Rad, Hercules, CA) in TBST (10 mM Tris/HCl [pH 8.0], 150 mM NaCl, and 0.05% Tween-20) for 1 h at room temperature before they were incubated overnight at 4 °C in 5% milk (in TBST) containing the respective primary antibody (Table 3). Detection was with horseradish peroxidase (HRP)-conjugated secondary antibodies and proteins were visualized using the Renaissance Western Blot Detection Kit (PerkinElmer Life Sciences, Boston, MA).

Electroretinography

Electroretinograms (ERGs) were recorded from both eyes simultaneously, as previously described [28,29]. Briefly, mice were dark-adapted overnight and anesthetized the next day with ketamine (66.7 mg/kg) and xylazine (11.7 mg/kg). Pupils were dilated before performing single flash ERG recordings under dark-adapted (scotopic) followed by light-adapted (photopic) conditions. Light adaptation was accomplished with a background illumination of 30 cd/m² starting 10 min before photopic recording. Single white-flash stimulus intensity ranged from –4 to 1.5 log cd*s/m² under scotopic and from –2 to 1.5 log cd*s/m² under photopic conditions, divided into 10 and 8 steps, respectively. Ten responses were averaged with an interstimulus interval of either 5 s or 17 s (for 0, 0.5, 1, and 1.5 log cd*s/m²).

Statistical analysis

Statistical analyses were performed using Prism4 software. All data are the mean ± standard deviation (SD) of three animals per group. Statistical differences of means were calculated using ANOVA (ANOVA) followed by a Bonferroni post-hoc test. A p value of less than 0.05 was considered significant.

Accumulation of CDC42 in the perinuclear region of TUNEL–negative photoreceptors during retinal degeneration

Immunofluorescent stainings of CDC42 on retinal sections showed that CDC42 was ubiquitously expressed in the retina. This was supported by analysis of Cdc42 mRNA (mRNA) expression in individual retinal layers separated by laser capture microdissection (Figure 1E). Localization of CDC42 in the inner retinal layers remained unchanged after light exposure (Figure 1A) and in the genetic models of photoreceptor degeneration studied (not shown). However, after induction of photoreceptor degeneration by light, CDC42 was additionally detected in distinct cell bodies in the outer nuclear layer (ONL; Figure 1A). Closer inspection of CDC42 localization by confocal microscopy revealed that CDC42 accumulated in the perinuclear region of photoreceptors (Figure 1B). Perinuclear accumulation of CDC42 in the ONL of degenerating retinas was an early and transient event, peaking around 14 h after light exposure. At this time point, Cdc42 showed a tendency of increased gene expression (Figure 1D). A similar CDC42 staining pattern was observed in all tested inherited mouse models (rd1, rd10, and VPP) of retinal degeneration (Figure 1C). This suggested that the perinuclear accumulation of CDC42 in individual cells of the ONL is a common mechanism in retinas with degenerating photoreceptors, and is independent of the toxic stimulus (light, mutation) that induced cell death.

Costaining for CDC42 and apoptotic cells using the TUNEL assay showed that perinuclear CDC42 staining appeared around 6 h after light exposure and was strongest around 14 h (Figure 2). It decreased rapidly thereafter as the number of TUNEL-positive cells rose (Figure 2). At the peak of apoptosis, around 36 h after light exposure [3], perinuclear CDC42 staining was no longer detectable in photoreceptors (data not shown). Importantly, CDC42 accumulated in the perinuclear region exclusively in TUNEL-negative photoreceptors, suggesting that CDC42 may support survival of cells or, alternatively, may prepare photoreceptors to die.

Rod photoreceptor–specific knockdown of Cdc42

To address the role of CDC42 in photoreceptor survival or death directly, we generated Cdc42^flox/flox;opsin-Cre mice that lack functional CDC42 specifically in rods. Although the expression of Cre recombinase has been described as being somewhat patchy in opsin-cre mice, expression is specific to rods and affects about 75% of photoreceptors [24,30]. By 12 weeks of age, Cre-mediated excision of Cdc42 genomic sequences was analyzed by amplifying genomic DNA isolated from retinal tissue. A 0.9 kb fragment was detected in Cdc42 knockdown (Cdc42^flox/flox;opsin-Cre) but not in control Cdc42^flox/flox mice, indicating successful excision of the floxed exon 2 of Cdc42 in the presence of Cre recombinase (Figure 3A). Since Cre recombinase was specifically expressed in rods, other retinal cells still carried the full-length floxed Cdc42 gene. Thus, PCR amplification of total retinal genomic DNA resulted in two bands representing the knockdown (0.9 kb) and floxed exon 2 (1.5 kb) alleles of Cdc42, respectively. A successful knockdown of Cdc42 in the ONL was also determined by the relative quantification of Cdc42 mRNA levels in the individual retinal layers after laser capture microdissection. Expression of Cdc42 was reduced by 60% in the ONL of Cdc42 knockdown mice relative to control mice (Figure 3B). Thus, we estimate that about 60% of the floxed Cdc42 alleles have been deleted in the ONL. Given that only about roughly 75% of rods express the CRE recombinase [24] we expect that 25% to 30% of rods are wild type for Cdc42 (Cdc42^+/+). Of the remaining rods, 40% to 50% have to be full knockouts (Cdc42^−/−), whereas the remaining 20% to 30% may be Cdc42^+/−. The reason for the slightly elevated levels of Cdc42 mRNA in the inner nuclear layer (INL) and the ganglion cell layer (GCL) of knockdown mice is not known, but may potentially include some reactions compensating for the absence of Cdc42 in the ONL. The reduced expression of Cdc42 in the ONL of knockdown mice was also reflected in a reduced number of photoreceptors with a perinuclear localization of CDC42 in the degenerating retina (Figure 3C). Together, these results suggested an effective CDC42 knockdown at 12 weeks of age. All subsequent experiments were performed on animals at this age, unless otherwise stated.

Rod-specific ablation of CDC42 does not affect retinal morphology and function

Recent studies have shown that CDC42 is involved in photoreceptor morphogenesis and polarity in Drosophila [31,32]. To test whether rod-specific ablation of CDC42 has an effect in the mouse retina, we analyzed tissue morphology, the expression of photoreceptor markers, and the retinal function in knockdown animals. The overall retinal morphology and photoreceptor structure of Cdc42 knockdown mice was maintained relative to controls (Figure 4A). Similarly, ablating CDC42 did not change expression levels of markers for rod and cone photoreceptors (Figure 4B). Furthermore, the intracellular localization pattern of rhodopsin, rod transducin, and short-wavelength (SWL) cone opsin was normal in Cdc42 knockdowns and indistinguishable from control littermates (Figure 4C). In particular, no mislocalization of these proteins was observed, suggesting that protein transport through the cilium was not affected by the lack of CDC42 (Figure 4C). Normal rod (and retinal) physiology in Cdc42 knockdown mice was also supported by normal scotopic and photopic ERG responses to light stimuli (Figure 4D). Together, these results suggest that deficiency of CDC42 in adult mouse rods does not influence photoreceptor physiology or function.

Rod-specific knock-down of CDC42 does not affect progression of photoreceptor degeneration

RAC1, another classical member of the small Rho GTPases, has recently been shown to have a pro-apoptotic role in light-induced retinal degeneration [21]. The perinuclear localization of CDC42 in TUNEL-negative photoreceptor cells after excessive light exposure (Figure 1 and Figure 2) suggested that CDC42 might have a similar role to RAC1 in degeneration. However, CDC42 ablation did not result in any detectable differences in the progression of induced (light damage) or inherited (VPP) photoreceptor apoptosis (Figure 5). At 24 h after light exposure, retinal morphology showed severely disturbed rod inner and outer segments, as well as a large number of photoreceptor nuclei with condensed chromatin in both control and knockdown mice (Figure 5A). At 10 days after exposure, the ONL was reduced to 2–3 rows of mainly pyknotic photoreceptor nuclei. Biochemical quantification of photoreceptor apoptosis at 36 h after light exposure confirmed that Cdc42 knockdown mice were susceptible to light injury in a similarly fashion to controls (Figure 5B). Similarly, VPP mice showed a comparable progression of photoreceptor loss independently of the presence or absence of CDC42. At 20 weeks of age, both control and knockdown mice possessed one row of photoreceptor nuclei, most of which contained condensed chromatin (Figure 5C). A similar progression of the degeneration was further supported by the comparable reduction of the ONL thickness at 4, 12, and 20 weeks of age in both control and Cdc42 knockdown mice (Figure 5D). Therefore, the presence or absence of CDC42 did not influence the susceptibility of rods to degeneration or the progression toward cell death.

Rod-specific ablation of CDC42 does not affect the induction of endogenous survival pathways after light exposure

Degeneration of photoreceptors activates a LIF-controlled protective signaling cascade in the retina [33]. This signaling involves induced expression of endothelin 2 (Edn2) and fibroblast growth factor 2 (Fgf2), as well as activation of JAK2, STAT1, and STAT3 proteins through phosphorylation [33-36]. Since it has been shown in other systems that CDC42 can participate in the endothelin [37] and LIF signaling pathways [38], we tested whether the ablation of CDC42 in rods might affect the molecular response during retinal degeneration. Semiquantitative analysis of gene expression after light exposure showed a similar fold induction over the respective basal levels in control and Cdc42 knockdown mice (Figure 6A). However, basal expression of Edn2 and Gfap was elevated in retinas of untreated knockdown mice, as well as in the early degeneration phase after light exposure (Figure 6A). Activation of JAK2, STAT1, STAT3, and AKT by phosphorylation was not grossly affected by the lack of CDC42 (Figure 6B). Together, these results suggest that lack of CDC42 in photoreceptors does not strongly affect the regulation of the LIF-controlled survival pathway and corroborates our conclusion of a nonessential role of CDC42 in photoreceptor degeneration.