Background
Axonal injuries occur in a wide range of CNS disorders. They contribute to neuronal damage in diseases and trauma of the human brain and spinal cord. Similarly, injury to the optic nerve (ON), an extension of the CNS, is associated with various ocular diseases and abnormalities, such as glaucoma, traumatic optic neuropathy, ischemic optic neuropathy, and compressive optic neuropathy. In humans, there are approximately 1.2 million axons in the ON. They originate from retinal ganglion cells (RGCs) in the inner retina and are responsible in transmitting image-forming visual and other signals from the retina to the visual centers and circadian pacemakers in the brain [
1]. Damage to these axons leads to RGC apoptotic death, and consequently, loss of vision and disturbance of the circadian rhythm [
2].
Apoptosis of RGC involves caspases. Caspases are a family of cysteine proteases that activate apoptotic pathways. They serve as either ‘initiators’ or ‘effectors’ based on their molecular functions. The initiator caspases, such as caspases-2, −8, −9, and −10, cleave and activate effector caspases, such as, caspases-3, −6, and −7, which, in turn, hydrolyze target proteins and initiate apoptosis [
3]. The involvement of specific caspases in optic neuropathy, glaucoma, and RGC death has been previously implicated. For example, in the ON of glaucoma patients, more axons were found to express caspase-3 [
4]. Similarly, activation of specific caspases occurs in retinas of animal models of ON injury, glaucoma, and RGC degeneration: (a) ON crush activates caspase-2 [
5]; (b) ON axotomy activates caspases-3, −6, −8, and −9 in RGCs [
6‐
11]; (c) higher levels of active caspases-3, −8, and −9 were shown in the rat retina subjected to chronic ocular hypertension [
12‐
17]; (d) retinal ischemia/reperfusion activates both caspases-2 and −3 [
18,
19] with a reported reduction in caspase-7 expression [
19]; and (e) intravitreal injection of kainate cleaved caspase-3 with an increase in effector caspase activity [
20]. Interestingly, other than caspase-2 [
5,
21], inhibition or down-regulation of other initiator caspases or caspase-3 only provides partial protection against injury-induced RGC death [
7‐
9,
12,
16].
Despite the studies described above, very little is known about the involvement of a unique caspase, caspase-7, in RGC apoptosis. Until recently, contribution of caspase-7 to neuronal apoptosis was controversial. Previously, it was generally believed that caspase-7 was not present in the CNS, not activated if present, or even if activated, ineffective in inducing apoptosis [
22]. Furthermore, because caspase-7 exhibits a high degree of homology with caspase-3, it was speculated that caspase-7 simply served as a redundant version of caspase-3, and thus did not play a critical role in the apoptosis cascade [
22]. However, it was later determined that caspase-7 cleaves exclusive substrates different from those of caspase-3, such as kinectin and co-chaperone P23 [
23‐
25]. Several other non-ocular studies strongly suggest that caspase-7 has a critical, non-redundant role in apoptotic cell death and in normal development [
26,
27]. For example, Larner et al. demonstrated that caspase-7 is upregulated and activated in the rat brain after traumatic brain injury [
28]. Unlike other caspases, only caspase-7 can be activated by calpain-1 [
29], indicating that caspase-7 may be a unique concomitant sensor of Ca
2+ dysregulation, via changes in calpain-1 activity, and the classical caspase activation, two important death cascades in RGCs.
Thus, the objective of this study is to evaluate the involvement of caspase-7 in ON injury-induced RGC apoptosis in mice. In this study, injury to the ON was produced by intraorbital ON crush, which initiates rapid onset and progression of RGC apoptotic death [
30,
31]. We showed that caspase-7, together with its upstream activator calpain-1, were activated and that the selective caspase-7 downstream substrates, P23 and kinectin, were hydrolyzed. We further showed that knockout of caspase-7 significantly both morphologically and functionally protected from the ON crush-induced loss of RGCs.
Discussion
In many neurodegenerative diseases including glaucoma, axonal injury is a critical insult leading to apoptotic death of RGCs [
39‐
41]. Although previous research has highlighted a central role of caspases as mediators of RGC apoptosis, the specific role of the distinct effector caspase, caspase-7, was not clear. We used ON crush as an experimental model of CNS injury to delineate this relationship. Similar to previous findings [
31,
42,
43], we observed a ∼ 60 % decrease in RGC density at 28 d post injury in adult mice.
We found that caspase-7 was activated in RGCs following ON crush starting at 12 h post injury, and activation reached a plateau at 3 d. Consistent with this finding, we also showed the activation of calpain-1, an upstream activator of caspase-7, in the ON crush-injured retina. Since the enzymatic site of caspase-7 differs from that of caspase-3, substrates hydrolyzed by caspase-7 differ from those of caspase-3. In our study, we observed the hydrolysis of the caspase-7-selective substrates kinectin and co-chaperone P23. Together, our results demonstrate that ON crush activates caspase-7 in the mouse retina.
The finding that caspase-7 was activated in a majority of RGCs as early as 12 h after injury is intriguing. In comparison, activation of caspase-3 begins at 3 d [
6] or later [
32] after ON injury and is typically only observed in only a fraction of RGCs, likely those actively going through apoptosis. Since RGC death starts at or after 3 d post ON crush and not all RGCs immediately become apoptotic [
31], our findings suggest that caspase-7 activation is involved in an earlier step in the apoptotic signaling cascade. Compared to other caspases, caspase-7 is unusual in that it can be activated by both the traditional initiator caspase pathway as well as a novel calpain-1 pathway. It is interesting to note that activation of calpain in our study was observed only at 12 and 24 h after ON crush injury, whereas, activation of caspase-7 was seen up to 7 days post injury. Together these data suggest that caspase-7 activation may be initiated by the calcium-calpain pathway and sustained by the traditional caspase activation cascade. The functional implication of this unique temporal profile and the exact role that caspase-7 plays require further investigation.
Although the exact role that caspase-7 activation in ON damage plays remains unclear, biological functions of substrates selectively hydrolyzed by caspase-7 may provide some clues. Co-chaperone P23, which was hydrolyzed in the injured retina of only WT but not
Casp7
−/−
mice, is a ubiquitous protein found in most eukaryote tissues. It binds and stabilizes Heat-shock protein 90 kDa (Hsp90), serves as an essential component of the Hsp90 molecular chaperone complex [
44], and interacts with a variety of Hsp90 clients, including most steroid receptors, active telomerase, tyrosine kinases, and transcription factors [
45]. In addition, genomic and proteomic studies further revealed an extensive P23 molecular network independent of Hsp90 clients [
46]. Most intriguingly, this protein is recognized as a pro-survival factor in cancer cells [
25] and perinatal mice [
47]. Thus, caspase-7 mediated hydrolysis of P23 will disrupt these intricate biological interactions and may synergize with the canonical caspase-mediated cell death pathway. Kinectin, another caspase-7 selective substrate, is found on the cytoplasmic face of endoplasmic reticulum (ER) membranes [
48], as well as the cargo-binding site of conventional kinesins. It has been implicated in the formation, stabilization, shaping, and cytoskeletal interactions of sheet-like cisterns of the ER [
49]. As a motor-binding protein, kinectin is also associated with active intracellular organelle movements powered by kinesin, thus controlling intracellular cargo transport [
50]. These vital functions of kinectin are expected to be compromised by caspase-7 activation.
To investigate the importance of caspase-7 in ON crush-induced RGC loss, we assessed the potential RGC protective effects of Casp7 knockout. Our results showed significant protection of RGCs in the ON crushed eyes of Casp7
−/−
mice when compared to WT mice 7 to 28 d post crush. We demonstrated this protection using both morphological and functional evaluations. Morphologically, we analyzed post mortem RGC density and measured in vivo thickness of retinal layers. RGC counts of flat-mounted retinas indicated that ON crush dramatically reduced the density of these neurons in WT mice, with only 38 % remaining at 28 d post injury. In contrast, 76 % of the RGCs survived this insult in Casp7
−/−
mice. An interesting observation is that RGC density of these animals appeared to stabilize from 7 d to 28 d compared to the continuous decline in the WT animals. This suggests that the protective effect of Casp7 knockout is not a transient delay of neuronal death, and instead it may provide a long-term protection.
SD-OCT analysis allows non-invasive,
in vivo monitoring of morphological changes of the retina. We showed that GCC thickness significantly decreased in WT mice from 7 to 28 d post ON crush. GCC is comprised of three retinal layers: the NFL, containing RGC axons; the GCL, containing RGC soma; and the IPL, containing the RGC dendrites. ON injury causes loss of RGCs and thinning of the GCC [
51‐
53]. In comparison, ON crush-induced thinning of the GCC was significantly and stably ameliorated in
Casp7
−/−
mice from 7 to 28 d after injury. Together, the
in vivo and
post mortem results indicate that absence of caspase-7 protects against morphological deterioration of RGCs caused by ON injury.
Most importantly, knockout of
Casp7 not only preserved RGC density and structures, but also significantly protected against loss of RGC function subsequent to ON crush. The STR response is derived mainly from RGCs [
54]. In WT mice, we found an approximately 35 % decrease in pSTR responses at 7 d and approximately 65 % at 28 d after ON crush. This coincides with a time-dependent RGC loss in WT mice after ON crush. In contrast, the pSTR response in
Casp7
−/−
mice showed only a 15 % reduction at 7 d and a 25 % decrease at 28 d after ON crush, clearly showing protection of RGC function.
Our discovery that caspase-7 plays a critical role in ON crush-induced RGC loss is important for the following reasons. Being a concomitant sensor of both the calpain and traditional caspase pathways, caspase-7, through hydrolysis of its specific downstream substrates may regulate unique and yet-to-be-revealed mechanisms in controlling cell death. In addition, caspase-7 is a p53 responsive gene; p53 binds to the first intron of the human and mouse caspase-7 gene [
55]. p53 is a transcription factor, involved in the regulation of neuronal viability after injury [
56,
57]. Activated p53 can modulate mitochondrial membrane permeability and the release of mitochondrial proteins during apoptosis. p53 is phosphorylated and its pro-apoptotic targets PUMA and Fas/CD95 are upregulated in ON axotomy-induced RGC death [
58]. It is interesting to speculate that caspase-7 may be involved in this apoptosis signaling mechanism.
Based on morphological and functional criteria, we showed that
Casp7 knockout was protective against ON crush-induced loss of RGCs. However, this protection was not complete (i.e., it did not reach the uninjured levels), suggesting that caspase-7 is not the sole mediator of this ON crush pathology and that other cellular pathways may be involved. It is likely that caspase-3 also contributes to this injury. This is consistent with the findings that inhibition of caspase-2, an upstream initiator caspase for both caspases-3 and −7, via pharmacological inhibition or siRNA provided almost complete protection of RGC against ON crush-induced death for at least 30 days [
5,
21]. Furthermore, caspase-independent pathways may also be involved in neuronal cell death associated with ON injury, such as mitochondria-mediated death. Mitochondria are essential in controlling cell death. Exposure of RGCs to TNF-α or hypoxia leads to mitochondrial alterations, including the release of cytochrome c and apoptosis inducing factor, leading to RGC death through caspase-independent events [
59]. After ON injury, amplification of TNF-α activates JNK, which causes RGC death [
60]. Fernandes et al. have shown that JNK signaling is a major early pathway triggering RGC death after axonal injury [
61]. Moreover, reactive oxygen species are associated with mitochondrial dysfunction and may be involved in signaling RGC apoptosis after axonal injury [
62]. Therefore, it is clearly important to delineate the relative roles each of these signaling pathways plays and their interactions in ON crush-induced RGC death and other forms of CNS injury.
Methods
Animals
Female C57BL/6 J mice, serving as the WT control, and caspase-7 knockout (Casp7
−/−
) (B6.129S6-Casp7tm1Flv/J) breeders were purchased from the Jackson Laboratory (Bar Harbor, ME). The Casp7
−/−
mice were backcrossed with the C57BL/6 J strain for at least 10 generations, making the C57BL/6 J strain an appropriate WT control. All mice were maintained in 12:12 light/dark cycle and supplied with food and water ad libitum. All experiments were conducted in accordance with the ARVO Statement of the Use of Animals in Ophthalmic and Vision Research and the University of North Texas Health Science Center’s Institutional Animal Care and Use Committee guidelines.
Optic nerve crush
ON crush was performed on the left eyes of C57BL/6 J and
Casp7
−/−
mice at 8–10 weeks of age according to a previously described procedure [
31]. Briefly, mice were anesthetized by intraperitoneal injection of ketamine and xylazine, (100 and 10 mg/kg body weight, respectively). The cornea was anesthetized with a topical drop of 0.5 % proparacaine HCl (Alcon Laboratories, Fort Worth, TX). The ON of the left eye was exposed intraorbitally through a small opening made among the surrounding extraocular muscles. A pair of self-closing forceps was used to crush the ON approximately 1 mm behind the globe for 4 s. Care was taken not to block the blood flow to the retina through the central retinal artery. The eye was covered with antibiotic ointment (polymyxin B sulfate and bacitracin; Bausch & Lomb, Tampa, FL). All animals undergoing surgery were given buprenorphine (0.05 mg/kg, subcutaneously) immediately after surgery.
Western blot analysis
Retinal protein extracts were obtained from dissected retinas by sonication in mammalian protein extraction buffer (Thermo Scientific, Rockford, IL) containing protease inhibitor and phosphatase inhibitor cocktails (Thermo Scientific). Protein concentrations were determined using the Bio-Rad Dc protein assay according to the manufacturer’s instructions (Bio-Rad Laboratories, Richmond, CA). A total of 60 μg protein from each sample was loaded and electrophoresed on 15 % denaturing polyacrylamide gels and then electrophoretically transferred to polyvinylidene fluoride membranes. Blots were incubated in 5 % milk in Tris Buffered Saline Tween solution (TBST; 20 mM Tris, 0.5 M NaCl, and 0.05 % Tween 20, pH 7.4) for 1 h to block nonspecific binding. Membranes were incubated overnight at 4 °C with primary antibodies diluted in Super Block T20 (Thermo Scientific). The primary antibodies used were caspase-7 (1:1000, Cell Signaling, Danvers, MA, Cat # 9492), calpain-1 (1:1000, Cell Signaling, Cat # 2556), kinectin (1:1000, Novus Biologicals, Littleton, CO, Cat # NBP1-91592), P23 (1:1000, GeneTex, Irvine, CA, Cat # GTX112655), PARP (1:1000, Cell Signaling, Cat # 9542), and mouse-anti-β-actin (1:5000; clone C4, Millipore, Billerica, MA, Cat # MAB1501). The following day, membranes were washed with TBST 3X for a total of 30 min and were incubated for 1 h with corresponding secondary antibodies: goat-anti-rabbit-horseradish peroxidase (1:10,000; Thermo Scientific, Cat # 32460) or goat-anti-mouse-horseradish peroxidase conjugated antibodies (1:10,000; Thermo Scientific, Cat # 32430) diluted in TBST. Membranes were washed 3X with TBST for 10 min each at room temperature. The proteins were then visualized in a FluorChem 8900 imager (Alpha Innotech, San Leandro, CA) using ECL detection reagent Super Signal West Femto Maximum Sensitivity Substrate (Thermo Scientific). The relative expression of proteins was determined by densitometry analysis and then normalized to β-actin.
Immunohistochemistry
Eyes from WT and Casp7
−/−
mice were enucleated and fixed overnight in freshly prepared 4 % paraformaldehyde (PFA) in phosphate-buffered saline (PBS). Afterwards, eyes were submerged sequentially in solutions of 10 %, 20 % and 30 % sucrose for at least 1 h each. Eyes were embedded in cryostat compound (Tissue TEK OCT, Sakura Finetek, Torrance, CA) and frozen at −80 °C. Sections (10 μm thick) were cut using a cryostat (Leica Microsystem, Buffalo Grove, IL) and transferred to glass slides. Retinal cryosections were soaked in PBS for 10 min and blocked in 2 % goat serum, 0.3 % Triton X-100 in PBS for 1 h at room temperature. The sections were double immunolabeled with mouse monoclonal caspase-7 antibody (1:50, Santacruz Biotechnology, Dallas, TX, Cat # SC-28295) and rabbit polyclonal RBPMS antibody (1:200, Genetex, Irvine, CA Cat # GTX118619) overnight at 4 °C, washed three times with PBS and incubated for 1 h with the respective secondary antibodies at room temperature. Secondary antibodies used were AlexaFluor488 goat-anti-rabbit (1:1000; Life Technologies, Grand Island, NY, Cat # A11008) and tetramethylrhodamine goat-anti-mouse IgG (1:1000; Life Technologies, cat # T2762). Sections were mounted with Vectashield Mounting Medium containing DAPI (Vector Laboratories, Burlingame, CA). Images were acquired using a Carl Zeiss inverted microscope Axiom vision series (Carl Zeiss Microscopy, Thornwood NY).
Retinal flat mounts and cell counts
Following fixation in 4 % PFA, the anterior segment of each eye was removed and the posterior eyecup was processed for whole mount immunostaining to quantitate RGCs. Retinas from fixed eyes were carefully dissected, flat-mounted on glass slides, and blocked in 0.3 % Triton X-100 in PBS containing 2 % goat serum for 2 h. Retinas were then incubated in rabbit polyclonal RBPMS antibody (1:200, diluted in 0.3 % Triton X-100 in PBS) overnight at 4 °C. Following washes in PBS, the retinas were incubated with AlexaFluor488 goat-anti-rabbit (1:1000, diluted in PBST) for 24 h at 4 °C and then mounted with Vectashield Mounting Medium containing DAPI (Vector Laboratories). Eight 40x images were taken from peripheral and mid-peripheral regions of the four quadrants of each retina. The number of labeled cells from each image was counted. For each individual retina, the RGC count was obtained by averaging the eight counts for each retina.
Spectral domain-optical coherence tomography
SD-OCT imaging was performed using the Bioptigen Envisu™ Ophthalmic Imaging System (Bioptigen, Durham, NC) to measure GCC thickness (combination of NFL, GCL, and IPL layers). Mice were anesthetized using a ketamine/xylazine cocktail (100/10 mg/kg), and their pupils were dilated with a drop of 2.5 % phenylephrine hydrochloride (Bausch & Lomb). To prevent corneal desiccation during the procedure, topical Systane Ultra® lubricant eye drop (Alcon) was applied frequently and bilaterally. The retina was scanned with the rectangular scanning mode consisting of 100 B-scan tomograms, each with 1000 A-scans, by using InVivoVue software (Bioptigen Inc. Durham, NC). Horizontal volume scans of retina at 100, 200, 300 and 400 μm through the area dorso-temporal of the ON head (superior retina) and the area ventro-temporal of the ON head (inferior retina) were used to evaluate the thickness of the GCC. For measuring the GCC thickness at peripheral retina, two calibrated calipers were placed at 400 μm and 500 μm on both sides from the center of each of the eight optical slices. The peripheral GCC thickness of each retina was determined by averaging the 32 measurements. To determine GCC thickness at central retina, two calibrated calipers were placed at 100 μm and 200 μm on both sides of the center of optical slices that were 200 and 300 μm superior or inferior to the ON head. The central thickness of the GCC of each retina was determined by averaging the 16 measurements.
Scotopic threshold response electroretinography
The STR of ERG was performed with the Handheld Multi-species-ElectroRetinoGraph unit (Ocuscience, Kansas City, MO) to assess the functional analysis of RGCs [
31]. Mice were dark-adapted overnight prior to ERG recordings. During the whole procedure, mice were anesthetized using 1.5 % isoflurane. The mouse pupils were dilated with a drop of 2.5 % phenylephrine, and the cornea was anesthetized with a drop of 0.5 % proparacaine. A ground electrode was inserted subcutaneously in the tail and two reference electrodes were inserted subcutaneously underneath both eyes. Recording electrodes were placed on the cornea. To maintain moisture, a drop of 2.5 % methylcellulose was added to the concave side of a contact lens, which was placed on the electrodes. Light stimuli (0.03 mcd · s/m
2) were generated in a Ganzfeld dome with light emitting diodes. Responses from 10 flashes with 2 s intervals were averaged. The positive component of the STR amplitude was quantified and analyzed by using ERGVIEW software (Ocuscience).
Statistical analysis
Results between two groups were analyzed for differences using the unpaired Student’s t-test (GraphPad Prism 6.0; GraphPad Software, La Jolla, CA). Results among three or more groups were compared using one-way ANOVA followed by Tukey–Kramer post hoc test. Statistical significance was determined at p values < 0.05.
Competing interests
The authors declare that they have no competing interest.
Authors’ contributions
SC performed ON crush, sample harvesting, Western blotting, Immunohistochemistry, fluorescence microscopy, retinal flat mounts, Optical coherence tomography, Scotopic Threshold Response-Electroretinogram and the data analysis. YL also performed some of the ON crush, sample processing, and actively participated in the design and coordination of the study. IHP participated in the design of the study, statistical analysis as well as reviewed the data. AFC reviewed all data, participated in the design and coordination of the study as well as revising the manuscript. Manuscript was written by SC and IHP with revisions from AFC and YL. All authors read and approved the final manuscript.