Capn5 Expression in the Healthy and Regenerating Zebrafish Retina

Purpose Autosomal dominant neovascular inflammatory vitreoretinopathy (ADNIV) is a devastating inherited autoimmune disease of the eye that displays features commonly seen in other eye diseases, such as retinitis pigmentosa and diabetic retinopathy. ADNIV is caused by a gain-of-function mutation in Calpain-5 (CAPN5), a calcium-dependent cysteine protease. Very little is known about the normal function of CAPN5 in the adult retina, and there are conflicting results regarding its role during mammalian embryonic development. The zebrafish (Danio rerio) is an excellent animal model for studying vertebrate development and tissue regeneration, and represents a novel model to explore the function of Capn5 in the eye. Methods We characterized the expression of Capn5 in the developing zebrafish central nervous system (CNS) and retina, in the adult zebrafish retina, and in response to photoreceptor degeneration and regeneration using whole-mount in situ hybridization, FISH, and immunohistochemistry. Results In zebrafish, capn5 is strongly expressed in the developing embryonic brain, early optic vesicles, and in newly differentiated retinal photoreceptors. We found that expression of capn5 colocalized with cone-specific markers in the adult zebrafish retina. We observed an increase in expression of Capn5 in a zebrafish model of chronic rod photoreceptor degeneration and regeneration. Acute light damage to the zebrafish retina was accompanied by an increase in expression of Capn5 in the surviving cones and in a subset of Müller glia. Conclusions These studies suggest that Capn5 may play a role in CNS development, photoreceptor maintenance, and photoreceptor regeneration.


Introduction
Calpains are a family of calcium-dependent, non-lysosomal cysteine proteases that include at least 15 family members in humans. [1][2] Calpains share similarity in their protease and calcium binding domains, and are unique to other proteases because they are localized to the cytosol or nucleus instead of the lysosome. Activated by influxes of calcium, calpains do not degrade their protein substrates, but rather modify the activity of their targets through proteolytic processing. Calpains have been implicated in numerous cellular processes, including cell death, signal transduction, intracellular signaling, and sex determination. 3 Hyperactivation of calpains is associated with the pathogenesis of several diseases, including Alzheimer's, Parkinson's, cardiovascular, and autoimmune diseases, as well as neurodegeneration following traumatic brain injury (TBI). 4,5 Calpains have also been implicated in many eye and retinal diseases, such as retinitis pigmentosa, retinal detachment, and glaucoma. 6,7,8,9 Although calpains have been the subject of extensive research, their functions and substrates are not fully defined. Moreover, while the majority of studies have focused on the ubiquitously expressed classical calpains CAPN1 (-calpain) and CAPN2 (m-calpain), the functions of the non-classical calpains in health and disease are much less well understood.
Capn5 is grouped with the non-classical calpains because it contains a C2-like domain instead of a penta-EF hand domain (domain 4) at its C-terminus. Capn5 is considered the vertebrate homolog of C. elegans TRA3, which plays a role in sex determination and mediates a necrotic pathway in neurons. 10,11 Capn5 has been shown to be the second most abundantly expressed calpain in the mammalian central nervous system (CNS). 12 Expression of Capn5 has also been demonstrated in the mammalian retina, where it is found in the outer plexiform layer (OPL) and outer nuclear layer (ONL), specifically the inner and outer segments and synapses of the rod and cone photoreceptors, some ganglion cells and the inner plexiform layer. 13 Within cells, Capn5 has been shown to be associated with promyelocytic leukemia protein bodies (PML) in the nucleus, which have been implicated in cellular stress response, apoptosis, cellular senescence, and protein degradation. [12][13][14] Mutations in CAPN5 are associated with the devastating retinal degenerative disease autosomal dominant neovascular inflammatory vitreoretinopathy (ADNIV). 15,16,17 ADNIV is a hereditary autoimmune disease of the eye that is characterized by abnormal retinal pigmentation, retinal neovascularization, photoreceptor degeneration, vitreous hemorrhage, intraocular fibrosis, and tractional retinal detachment. As the disease progresses it phenocopies more commonly known ocular diseases such as non-infectious uveitis, glaucoma, diabetic retinopathy and retinitis pigmentosa. 15,18 The three point mutations that have been identified in ADNIV patients (p.Arg243Leu, p.Leu244Pro, and p.Lys250Asn) are located in the calcium-sensitive domain 2 near the active site and are thought to cause the mislocalization of CAPN5 from the cell membrane to the cytosol. 16,17 Thus, ADNIV is thought to result from gain-of-function mutations in CAPN5 that lower its threshold for activation by calcium. 15,19 However, the precise mechanism whereby mutant CAPN5 causes ADNIV is not well understood.
Elucidating the role of CAPN5 in the retina could reveal the underlying pathogenetic mechanisms of ADNIV as well as other retinal degenerative diseases that display similarities to ADNIV.
The normal function of Capn5 during development and in the adult retina is not well understood. Previous studies using two different Capn5 mutant mouse models yielded conflicting results. In one study, Capn5 null mice (Capn5 trn1Nde) were born viable and fertile, but some mutant offspring were runted and died two months after birth. 20 In another study, the Capn5 -/null mutation (Capn5 trn1Dgen) was preimplantation embryonic lethal. 20 These conflicting results highlight a need for additional animal models to study Capn5 function. Zebrafish are an ideal model for developmental studies due to their strong genetic homology to humans, completely sequenced genome, and rapid, transparent, external development. Furthermore, unlike mammals, the zebrafish is capable of regenerating all classes of retinal neurons in response to injury or genetic insult. This regenerative response relies on the retinal Müller glia, which are stimulated to dedifferentiate, re-enter the cell cycle, and generate retinal progenitor cells to replace the cells that have been lost. [21][22] In this study, we examined the expression profile of capn5 during embryonic development and in the adult retina of the zebrafish. The zebrafish has two orthologs of Capn5 (capn5a and capn5b; http://zfin.org/ ) and we demonstrate that both are expressed in the CNS and the retina during zebrafish development. We also demonstrate that Capn5 is expressed in the adult zebrafish retina in a similar pattern to that described for the mammalian retina, however we find that expression of Capn5 is specific to cone, but not rod, photoreceptors. Finally, we show that Capn5 is upregulated in response to photoreceptor degeneration and is expressed in the Müller glia following acute light exposure, suggesting that Capn5 could be playing a role in the regenerative process to retinal damage.

Developmental expression of zebrafish capn5 orthologs
While some studies have examined the expression and function of Capn1 and Capn2 orthologs in zebrafish, 23 the developmental expression pattern of Capn5 has not previously been reported. Through BLAST searches of zebrafish genome databases we identified two, full-length cDNA sequences with significant sequence similarity to human CAPN5. Due to an ancient genome duplication that occurred during the evolution of teleost fishes, 24-25 zebrafish often possess two orthologs of single-copy mammalian genes. The two zebrafish orthologs of CAPN5, hereafter referred to as capn5a and capn5b, are located on zebrafish chromosomes 18 and 21, respectively. The predicted zebrafish Capn5a protein is 68% identical and 81% similar to human CAPN5, whereas zebrafish Capn5b is 71% identical and 83% similar. We identified unique regions of each cDNA sequence and designed PCR primers. We extracted RNA from embryos at selected developmental time points (4,18,24,48,72, and 120 hours post fertilization), and performed RT-PCR followed by agarose gel electrophoresis to detect expression of capn5a and capn5b. Expression of both genes was detectable starting at 4 hours post fertilization (hpf), indicating capn5 is maternally deposited, and both capn5a and capn5b were expressed at every subsequent developmental timepoint tested ( Figure 1A). To quantify the expression of capn5a and capn5b, we performed quantitative real-time RT-PCR (qPCR). For both capn5a and capn5b, we observed a gradual increase in expression as development progressed ( Figure 1B). With the exception of 4 hpf, expression of capn5a was consistently higher than capn5b ( Figure 1B). From these data, we conclude that capn5a and capn5b are expressed during zebrafish embryonic development and that capn5a is expressed more strongly during development than capn5b.
In the adult mouse, Capn5 expression has been identified in many tissues including the brain, eye, uterus, and prostate; 12, 26-29 however, less is known about the expression pattern of Capn5 during vertebrate embryonic development. To address this question, we performed whole-mount in situ hybridization (WISH) with unique probes for capn5a and capn5b at selected developmental timepoints in the zebrafish. At 18 hpf, capn5a expression was detected in the optic vesicles, developing diencephalon, mesencephalon and hindbrain ( Figure 1C-C'). Capn5b expression was present in the diencephalon, mesencephalon, and hindbrain, but was not detected in the optic vesicles at 18 hpf ( Figure 1E). At 24 hpf, expression of capn5a and capn5b is was observed in the developing zebrafish brain, more specifically the tectum, hindbrain, cerebellum and floor-plate. Expression of capn5a was also observed in the hatching gland ( Figure 1D, 1F). We did not detect expression of capn5a or capn5b in the developing eye at 24, 36 or 48 hpf (data not shown). We conclude that capn5a/b are expressed in the developing zebrafish brain and in the optic vesicle prior to its invagination to form the bi-layered optic cup. Furthermore, the lack of expression of capn5a/b in the retina prior to 3 dpf suggests that Capn5 plays a role in photoreceptor cell maintenance rather than development or specification.

Capn5 expression is cone-specific in the adult zebrafish retina
To further analyze the expression of capn5a/b in the adult zebrafish retina we performed FISH with capn5a/b probes on sections of adult wild type zebrafish retina.
Expression of capn5a and capn5b was again observed in the photoreceptor inner segments ( Figure 3). The rod and cone photoreceptors in the adult zebrafish retina are highly tiered and morphologically distinguishable, such that (moving from the vitreal to scleral direction), the round rod nuclei are located most proximal to the OPL, followed by the UV and blue cones, then the ellipsoid red/green double cone nuclei interspersed with rod inner segments, the red/green cone outer segments, and finally the rod outer segments located most proximal to the RPE. 30 Interestingly, the FISH expression pattern for capn5a/b appeared to localize primarily to the cone, but not rod inner segments. To determine whether this was the case, we used the TC-GFP transgenic line, in which GFP is expressed specifically in the cones, 31,32 and performed FISH for capn5a in combination with IHC for GFP. We observed a perfect co-localization of GFP and capn5a expression ( Figure 3D-D''). We then performed the same experiment using the XOPS:GFP transgenic line, which expresses GFP specifically in the rods. 30 We did not observe any co-localization of the rod GFP signal and capn5a expression ( Figure   3E-E''), which we confirmed using confocal microscopy ( Figure 3F-F"). Capn5b displayed the same co-localization pattern as capn5a (not shown). We confirmed the expression of Capn5 protein in the adult retina using the CAPN5 antibody ( Figure 3C-C"). Capn5 expression was observed in the cone inner segments, outer segments and in the synaptic terminals of the adult retina. Taken together, these results suggest that, similar to mammals, zebrafish Capn5 is expressed in retinal photoreceptors. However, zebrafish Capn5 expression appears to be cone-specific, which has not been reported for Capn5 in the mammalian retina.

Capn5a expression increases in response to rod photoreceptor degeneration
As described above, gain of function mutations in CAPN5 have been associated with ADNIV, an autosomal dominant disease that eventually leads to the photoreceptor degeneration and retinal detachment. 12,16,17,19,33 Moreover, other members of the calpain family have been implicated in apoptosis, necrosis, and photoreceptor degeneration. 6,7,34 To determine whether Capn5 expression is altered in response to photoreceptor degeneration in zebrafish, we evaluated the expression of capn5a and capn5b in the XOPs:mCFP transgenic line, which displays continual degeneration and regeneration of rod (but not cone) photoreceptor cells. 35 RT-PCR and qPCR were performed on mRNA prepared from dissected wild type (WT) and XOPS:mCFP retinas.
In both experiments, we observed elevated expression of capn5a in the XOP:mCFP retinas compared to WT, whereas capn5b expression levels remained unchanged ( Figure 4A-B). Next, we performed FISH on retinal cryosections from WT and XOPs:mCFP zebrafish with capn5a/b probes. We found that the expression patterns for both capn5a and capn5b were similar in the WT and the XOPs:mCFP retina; however, the signal for capn5a was stronger in XOPS:mCFP cones than in WT, whereas capn5b expression levels did not change ( Figure 4C-H'). These results indicate that the cone-specific expression of capn5a is upregulated in response to rod photoreceptor degeneration in zebrafish.

Capn5 expression increases in response to acute light damage
Our results indicate that cone-specific expression of Capn5 is induced in the XOPS:mCFP transgenic line. However, in the XOPs:mCFP retina, rod photoreceptor degeneration does not result in any secondary degeneration of the cones. 35,36 This led us to ask whether upregulation of Capn5 in cones serves a cell-autonomous protective function, or whether it plays a non-cell autonomous role in promoting rod degeneration or regeneration. To begin to address this question, we used an acute light damage approach, which allowed us to temporally separate photoreceptor degeneration and regeneration. We adopted the light damage protocol described by Vithelic and Hyde, 37 in which dark adapted albino zebrafish are exposed to 20,000 lux of constant light for  39 Moreover, we observed strong co-localization of the Capn5 signal with Zrf-1 ( Figure 7D-F'). These results demonstrate that Capn5 is induced in Müller glia in response to photoreceptor degeneration.

Discussion
Although mutations in CAPN5 are associated with a severe retinal degenerative disease, very few studies have described the developmental expression pattern of Capn5. In the mouse, developmental expression of Capn5 was suggested to be relatively low, 27 with specific sites of expression noted in the developing thymus and the neurons of the sympathetic and dorsal root ganglia. 27, 40 Further in situ hyrbridization studies on embryonic tissue sections revealed Capn5 expression in the developing mouse brain and in epithelial cells surrounding several tissues. 20,28 In our study, we found that Capn5 is expressed in the developing brain at 18 hpf, and its expression increases throughout embryogenesis. We detected expression of Capn5 in the mesencephalon, telencephalon and optic vesicles at 18 hpf, and the hindbrain and cerebellum at later stages. This expression pattern supports a role for Capn5 during vertebrate brain development.
Cell death is a common and often essential process throughout embryonic development, 41 and calpains have been implicated in mediating cell death pathways.
Calpain 1 has been shown to cleave procaspase-12, activating caspase-12, and initiating cell death. 42 In another study, calpain 2 was shown to mediate the cleavage of Atg5 switching the cell death pathway from autophagy to apoptosis. 4,43 Given that calpains are thought to modify a wide variety of target substrates, we hypothesize that is not needed for the development or specification of photoreceptors, but rather that Capn5 is important for mature photoreceptor function or maintenance.
Previous studies have analyzed the expression of Capn5 in the adult human and rat retina, identifying protein expression in the outer nuclear layer, the outer plexiform layer, and the inner segments of the photoreceptors, as well as lower expression in some ganglion cells and the inner plexiform layer. 12 Co-localization of Capn5 in the outer plexiform layer (OPL) with PSD95 expression, a marker for neural synaptic densities, coupled with the detection of Capn5 in the synaptic membrane by Western blot, indicated that Capn5 is expressed in the photoreceptor synapses. More specifically, expression was identified in both the outer segment and synaptic fractions of rod photoreceptors from mouse retina. 13 In our study, we found that Capn5 is also expressed in mature photoreceptors in the wild type adult zebrafish retina, however its expression appears to be cone specific. We observed capn5 mRNA expression in the inner segments of the cone photoreceptors, and analysis of protein expression by immunohistochemistry identified expression in the inner segments and the photoreceptor synaptic terminals, which is consistent with the mammalian expression pattern retina. 13 Previous studies of Capn5 expression in the mouse and human retina did not report whether Capn5 expression was associated with a specific photoreceptor subtype.
The zebrafish possesses a cone-rich retina with each of the four spectral subtypes having distinct morphologies, making them easily identifiable in tissue sections. We found, using both FISH and immunohistochemical analyses, that Capn5 is not expressed in the rod photoreceptors of the zebrafish. It will be interesting to determine whether mammalian Capn5 is also expressed primarily in cones, or whether there are species-specific differences in its localization. Using a genetic model of rod-specific degeneration and regeneration, 35 we found that the cone-specific expression of Capn5 increases in response to rod degeneration. This is intriguing because the cone photoreceptors in the XOPs-mCFP zebrafish do not degenerate secondary to the rod degeneration, as is typically observed in other photoreceptor degeneration and disease models. 35 This raises the question: Is Capn5 playing an anti-apoptotic role in the cone photoreceptors and protecting them from the residual effects of the rod degeneration?
Using an acute light-damage paradigm, which results in almost total rod photoreceptor loss and some cone photoreceptor damage, we found that there is a large increase in Capn5 expression during the acute light damage phase that correlates with the amount of photoreceptor damage produced. In the zebrafish retina, the cones are much more resistant to the toxic effects of acute light exposure. Therefore, induction of Capn5 in cones in response to acute light damage could indicate a protective role for this protein during photoreceptor degeneration.
How do our data fit into the context of previously described functions of calpains?
The role of calpains in regulating cell death pathways such as apoptosis, programmed cell death and necrosis has been extensively explored, 44 with a focus on their function as either pro-or anti-apoptotic proteases. It has been shown that calpains play a proapoptotic role in the presence of a wide variety of stimuli. For example, in fibroblasts overexpression of Calpastatin, the endogenous calpain inhibitor, protected cells from okadaic acid induced apoptosis. 45 Depletion of calpastatin in neutrophils promoted cell death during cycloheximide induced apoptosis. 46 Evidence has also been presented that calpains play a pro-apoptotic role during cell death induced by hydrogen peroxide, ultra-violet light, and serum starvation through the PI3-kinase/Akt survival pathway. 1,26 Overexpression of Capn2 in CHO cells resulted in sensitivity to ER stress induced cell death. 44 In contrast, anti-apoptotic roles for calpains have also been demonstrated in the presence of some stimuli. For example, Capn1 cleaves p53 which protects the cell from DNA damage induced apoptosis. 47,48 Given that no Capn5-specific substrates have been identified yet, it is possible that Capn5 plays an apoptotic or anti-apoptotic role, depending on the context. As mentioned above, our data indicate a possible protective role for Capn5 in the cones in response to rod photoreceptor degeneration. Finally, one of the most intriguing results from our study is that, in addition to upregulation of Capn5 expression in the cones in response to acute light damage, we also observed induction of Capn5 in a subset of Müller glia. It should be noted that this Müller glia expression was not observed in the wild type or XOPs-mCFP zebrafish models, and has not been observed in the wild type mammalian retina. In teleosts, Müller glia are the source of retinal stem and progenitor cells for injury-induced regeneration, and they also phagocytose cell debris to clean up the retina in the initial response to damage. 49,50 Therefore, the upregulation of Capn5 in Müller glia in response to acute light damage suggests a potential role for this calpain in photoreceptor regeneration as well.
In summary, this body of work lays the foundation for understanding the physiological function of Capn5 in the developing retina and in response to photoreceptor degeneration. As one of the few calpains that is not tightly associated with its inhibitor, understanding the function of Capn5 provides an opportunity to investigate calpain protease function in the absence of endogenous inhibition. Further investigation has the potential to shed light on the importance of proteolytic proteases in degeneration and regeneration, and possibly unlock the underlying mechanism associated with ADNIV.

Zebrafish lines and maintenance
All zebrafish lines were bred, housed, and maintained at 28.5°C on a 14 hour light:10 hour dark cycle, except where indicated for the light damage experiments. The Tg(XRho:gap43-mCFP) q13 transgenic line (hereafter called XOPS:mCFP) has been previously described. 35,36,51 The Tg(3.2TαC:EGFP) transgenic line (TαC:EGFP) has been previously described, 31 and was generously provided by Susan Brockerhoff (University of Washington, Seattle, WA). The Tg(XlRho:EGFP) transgenic line (hereafter called XOPS:GFP) has been previously described, 30 and was obtained from James Fadool (Florida State University, Tallahassee, FL). Zebrafish were bred, raised and maintained in accordance with established protocols for zebrafish husbandry. 52 Embryos were anaesthetized with Ethyl 3-aminobenzoate methanesulfonate salt (MS-222, Tricaine, Sigma-Aldrich, St. Louis, MO) and adults were euthanized by rapid cooling as previously described. 53 All animal procedures were carried out in accordance with guidelines established by the University of Kentucky Institutional Animal Care and Use Committee (IACUC) and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

RNA extraction, RT-PCR and Real-time quantitative RT-PCR (qPCR)
Total RNA was extracted from whole embryos at selected developmental time points or from the dissected retinas of adult zebrafish using TRIzol reagent (Invitrogen, Grand Island, NY) according to the manufacturer's protocol. The samples were treated with RNAse-free DNAse I (Roche, Indianapolis, IN) and purified using a chloroform/phenol extraction. The GoScript Reverse Transcriptase System (Promega, Madison, WI) was used to synthesize first strand cDNA from 1g of the extracted RNA. PCR primers were designed to amplify unique regions of the capn5a, capn5b, and atp5h cDNAs (Eurofins Genomics; www.eurofinsgenomics.com; Table 1). Faststart Essential DNA Green Master mix (Roche) was used to perform qPCR on a Lightcycler 96 Real-Time PCR System (Roche). The relative transcript abundance was normalized to atp5h expression as the housekeeping gene control, 54 and was calculated as fold-change relative to 4 hours post fertilization (hpf) for developmental expression, and fold-change relative to wild type, untreated adult fish (WT) for the XOPs:mCFP and light damage experiments.
RT-PCR and qPCR experiments were performed with three biological replicates and three technical replicates. RT-PCR was performed on a Mastercycler Pro thermocycler (Eppendorf, Westbury, NY). PCR products were visualized on a 1% agarose gel. The sequences for the primers used to produce the PCR products are listed in Table 1.

Tissue sectioning
Whole embryos and adult retinas were collected as described above and fixed in 4% paraformaldehyde (PFA) at 4°C overnight. Fixed embryos or retinas were cryoprotected in 10% sucrose for a minimum of 8 hours, followed by 30% sucrose overnight at 4°C.
Samples were placed into OTC medium (Ted Pella, Redding, CA) and frozen at -80°C for 2 hours. Ten micron-thick tissue sections were cut on a cryostat (Leica CM 1850, Leica Biosystems, Buffalo Grove, IL) and the sections were mounted on gelatin-coated or Superfrost Plus slides (VWR, Randor, PA) and air-dried overnight at room temperature.

Riboprobe synthesis
PCR products from the unique regions of capn5a and capn5b were cloned into the pGEMT-easy vector (Promega, Madison, WI). Plasmids were linearized using either SpeI or SacII restriction enzymes (NEB, Ipswich, MA). Riboprobes were synthesized from the plasmids by in vitro transcription using either T7 or Sp6 polymerase and a digoxigenin (DIG) labeling kit (Roche Applied Science, Indianapolis, IN). The sequences for the primers used to produce the PCR products are listed in Table 1.

Fluorescent in situ hybridization (FISH)
FISH was performed essentially as previously described. 55

Whole-mount in situ hybridization (WISH)
Embryos were manually dechorionated, collected at selected developmental time points (18, 24, 48 hpf, 72 hpf and 120 hpf) and fixed as described above. WISH was performed as previously described. 55

Immunohistochemistry
Immunohistochemistry was performed as previously described. 55 Primary antibodies used in this paper are described in Table 2

Light damage experiment
Acute light damage (LD) was performed essentially as previously described. 37,53 Briefly, 18 month old, albino zebrafish were dark adapted for 14 days. Fish were then placed in a 2.8 liter clear plastic tank surrounded by four 250W halogen bulbs placed 7 inches away, which collectively produced 20,000 lux of light. A bubbler, cooling fan, and water circulatory system were used to maintain the water level and keep the temperature under 32°C. The fish were maintained in constant light for 3 days, at which point they were returned to normal lighting conditions. Fish were collected at various time points during and after light damage (LD), and the left eyes were dissected for cryosectioning and in situ hybridization, whereas the right eyes were dissected and processed for RNA extraction followed by RT-PCR and qPCR. The fish were collected at 3 days LD, and at 2 and 7 days post LD. The light damage experiment was repeated three times, and three fish were collected for each time point.

Quantification and Statistical analysis
For comparisons between groups, statistical significance was determined using the Student's t-test with p<0.05 considered as significant. For all graphs, data are represented as the mean +/-the standard deviation (s.d.).       Table 1. Primer sequences used for RT-PCR and qPCR. RT-PCR primers were also used to design WISH and FISH probes.