Abstract
Glaucoma is characterized by programmed cell death of retinal ganglion cells (RGCs) after axonal injury. Several studies have shown the cell-intrinsic drivers of RGC degeneration act in a compartment-specific manor. Recently, the transcription factors JUN and DDIT3 were identified as critical hubs regulating RGC somal loss after mechanical axonal injury. It is possible somal DDIT3/JUN activity initiates axonal degeneration mechanisms in glaucoma. Alternatively, DDIT3/JUN may act downstream of inciting degenerative mechanisms and only drive RGC somal loss. The MAP2Ks MKK4 and MKK7 control all JNK/JUN activity and can indirectly activate DDIT3. Furthermore, MKK4/7 have been shown to drive RGC axonal degeneration after mechanical axonal injury. The present work investigated whether JUN and DDIT3, or their upstream activators MKK4 and MKK7, control degeneration of RGC axons and somas after glaucoma-relevant injury. Ddit3/Jun deletion did not prevent axonal degeneration in ocular hypertensive DBA/2J mice but prevented nearly all RGC somal loss. Despite robust somal survival, Ddit3/Jun deletion did not preserve RGC somal viability (as assessed by PERG decline and soma shrinkage) in DBA/2J mice or after glaucoma-relevant mechanical axonal injury. In contrast, Mkk4/7 deletion significantly lessened degeneration of RGC somas and axons, and preserved somal function and size after axonal injury. In summary, activation of MKK4 and MKK7 appears to be the inciting mechanism governing death of the entire RGC after glaucoma-relevant injury; driving death of the RGC soma (likely through activation of DDIT3 and JUN), decline in somal viability, and axonal degeneration via DDIT3/JUN-independent mechanisms.
Introduction
Vision loss in glaucoma is caused by progressive loss of retinal ganglion cells (RGCs). The most important risk factors for developing glaucoma are age and elevated intraocular pressure (IOP). After prolonged elevated IOP, RGC axons are injured at the optic nerve head—ultimately leading to programmed cell death. Much progress has been made in dissecting the mechanisms of RGC death after glaucoma-relevant injury. The pro-apoptotic molecule BAX was shown to be required for RGC somal death after chronic ocular hypertension (the DBA/2J mouse model of glaucoma). However, BAX was not required for axonal degeneration (1). Thus, RGC degeneration is compartmentalized after axonal injury; RGC somal and axonal degeneration are governed by distinct molecular mechanisms. Furthermore, the inciting mechanism triggered by ocular hypertension that leads to both axonal and somal degeneration must ultimately cause somal BAX activation, making upstream activators of BAX attractive targets of investigation.
Recently, the BAX-inducing transcription factors JUN and DNA-damage inducible transcript 3 (DDIT3, also known as CHOP) were identified as the critical regulators of RGC somal loss after glaucoma-relevant injury (2–4). Ddit3 and Jun deletion provided near-complete protection to RGC somas after controlled optic nerve crush (CONC) (2), an acute model of mechanical axonal injury. Like Bax deletion, Ddit3/Jun deletion did not prevent RGC axonal degeneration after CONC (2), suggesting degenerative DDIT3/JUN activity is restricted to the RGC soma. Several lines of evidence have indicated axonal degeneration mechanisms are critical to initiating pathways driving both somal and axonal degeneration in glaucoma (5–8). However, it has also been suggested that degenerative mechanisms initiated in the soma are essential for driving axonal degeneration (9, 10). Somal JUN activity has been suggested to play a role in anterograde axonal injury signaling (9), and DDIT3 has been suggested to promote axonal degeneration, presumably as a result of its somal function (11). Individual deletions of Ddit3 (4) or Jun (3) were not sufficient to prevent axonal degeneration in DBA/2J mice. However, it is possible DDIT3 and JUN play redundant and compensatory roles in initiating axonal degeneration. Given the near-complete somal protection conferred by combined Jun/Ddit3 deletion after CONC (2), it is feasible somal DDIT3 and JUN act together to integrate somal and axonal degeneration cascades and thus govern death of the entire RGC in the context of ocular hypertension. Alternatively, it is also possible that DDIT3 and JUN act as critical downstream hubs regulating only RGC somal loss. If this is the case, the inciting mechanism governing degeneration of the entire RGC lies upstream of somal DDIT3 and JUN activation.
MAP2Ks 4 and 7 (MKK4/7) are known to act upstream of both DDIT3 and JUN activation. In fact, MKK4/7 are the only kinases known to activate the JUN-N-terminal kinases (JNKs 1, 2, and 3, also known as MAPKs 8, 9, and 10). Thus, MKK4/7 activation acts as a bottleneck for JNK and subsequent JUN activation (12, 13). As upstream effectors, MKK4/7 control several molecular mechanisms in addition to DDIT3/JUN. Through the JNKs, MKK4/7 can control JUN/DDIT3-independent cell death pathways, including NMNAT2 degradation (14, 15), SARM1 activation (16), and BH3-only protein activation (17). MKK4/7 also control JNK-independent degenerative pathways, including p38 (MAPK11-14) signaling (18). Individually, Mkk4 or Mkk7 deletion afforded significant but incomplete protection to RGC somas after CONC (42% and 17% protection, respectively) (19). Neither Mkk4 nor Mkk7 deletion alone was sufficient to prevent all JNK and JUN activation in RGC axons and somas (19), suggesting at least partial compensatory activity. Excitingly, blocking both MKK4 and MKK7 activity preserved morphological integrity of RGC axons after axonal injury (16). Therefore, MKK4/7 signaling may be a critical early mechanism governing RGC somal death via DDIT3/JUN activation and may also regulate axonal degeneration. Here, we assessed the importance of DDIT3 and JUN, along with their upstream activators MKK4 and MKK7, in RGC degeneration after glaucoma-relevant injury. Importantly, we investigated the role of these molecules in mitigating not only morphologically observable cell loss, but also deterioration of viability and gross function.
Methods
Mice
Ddit3 null alleles (20) (Jackson Laboratory, Stock# 005530), floxed alleles of Jun (21) (Junfl), and the Six3-cre transgene (22) (Jackson Laboratory, Stock# 019755) were backcrossed >10 times to both the C57BL/6J genetic background (>99% C57BL/6J) and the DBA/2J background (>99% DBA/2J). Junfl alleles were recombined in the optic cup using Six3-cre. Floxed alleles of Mkk4 (23) and Mkk7 (24) were backcrossed >10 times to the C57BL/6J background and were recombined from retinal neurons via bilateral intravitreal delivery of AAV2.2-Cmv-cre-Gfp or AAV2.2-Cmv-Gfp with no cre (UNC vector core). Mice were fed chow and water ad libitum and were housed on a 12-hour light-to-dark cycle. All experiments were conducted in adherence with the Association for Research in Vision and Ophthalmology’s statement on the use of animals in ophthalmic and vision research and were approved by the University of Rochester’s University Committee on Animal Resources.
Experimental rigor and statistical analysis
For all procedures, the experimenter was masked to genotype and condition. Roughly equal numbers of male and female mice were used for each experimental group. Animals were randomly assigned to experimental groups. Before experiments were performed, it was established that animals with pre-existing abnormal eye phenotypes (e.g. displaced pupil, cataracts) would be excluded from the study.
Data were analyzed using GraphPad Prism9 software. Power calculations were performed before experiments were conducted to determine appropriate sample size. The comparison of the percent of optic nerves at each grade between genotypes was analyzed using a Chi-square test. Data from experiments designed to test differences between two groups were subjected to an F test to compare variance and a Shapiro-Wilk test to test normality to ensure appropriate statistical tests were utilized. For normally distributed data with equal variance, a two-tailed independent samples t test was utilized. Data from experiments designed to test differences among more than two groups across one condition were subjected to a Brown-Forsythe test to compare variance and a Shapiro-Wilk test to test normality to ensure an appropriate statistical test was utilized. Normally distributed data with equal variance were analyzed using a one-way ANOVA followed by Holm-Sidak’s post-hoc test. Non-normally distributed data were analyzed using a Kruskal-Wallis test with Dunn’s post-hoc test. Data from experiments designed to detect differences among multiple groups and across two conditions were analyzed using a two-way ANOVA followed by Holm-Sidak’s post-hoc test. Data from experiments designed to test differences among multiple groups across more than two conditions were analyzed using a three-way ANOVA followed by Holm-Sidak’s post-hoc test. For these statistical tests, every possible comparison was made when relevant, and multiplicity adjusted P values are reported. In all cases, data met the assumptions of the statistical test used. P values <0.05 were considered statistically significant. Throughout the manuscript, results are reported as mean± standard error of the mean (SEM).
Full field and pattern electroretinograms
Pattern and full-field electroretinography (PERGs and ERGs, respectively) was conducted using Diagnosys LLC’s Celeris rodent ERG system according to manufacturer’s instructions. Briefly, mice were dark-adapted for 60 minutes. Mice were anaesthetized with an intraperitoneal injection of 0.05 ml/10 g solution containing 20 mg/mL ketamine and 2 mg/mL xylazine. Hypromellose GenTeal (0.3%, Novartis Pharmaceuticals Corporation, NDC 0078-0429-47) was applied to the eyes before placement of electrodes. PERGs were obtained using 50cd/m2 mean luminance with spatial frequency 0.155 cycles/degree with 100% contrast. A total of 600 sweeps were recorded and averaged per eye. ERGs were obtained with 1 cd s/m2 luminance.
Other surgeries and procedures
Controlled optic nerve crush (CONC) (1, 19, 25), intraocular pressure measurement (3, 25, 26), and intravitreal injections (27, 28) were performed as previously described. CONC was performed at least 28 days after intravitreal delivery of AAV2.2-Cmv-cre-gfp or AAV2.2-Cmv-gfp to allow for recombination and endogenous protein degradation.
Compound action potentials
Compound action potentials were recorded as previously described (2, 29, 30). 5 days following CONC, animals were euthanized with CO2 asphyxiation and optic nerves were dissected free. Fresh optic nerves were transferred to a chamber of artificial cerebral spinal fluid (ACSF) aerated with 95% O2/5%CO2 for at least 60 minutes before recording. Optic nerves were transferred to a temperature-controlled chamber perfused with ACSF bubbled with 95% O2 / 5% CO2. Nerves were drawn into glass pipet suction electrodes (filled with ACSF) at each end for stimulation and recording. The recording pipet resistance was measured before (17-20 KΩ) and after (29-34 KΩ) insertion of the nerve and monitored continuously during the experiment. The ratio of this resistance during each sweep divided by the resistance of the pipet alone allowed a normalization of the amplitude of the compound action potential (CAP) to our standard ratio of 1.7 (30, 31). This corrects for any drift in the seal resistance during an experiment. Signals were fed to one input of an AC differential amplifier of our design. The second input came from a pipet electrode placed near the recording electrode. This served to subtract much of the stimulus artifact. Stimuli of 50μs duration were delivered by an optically isolated constant current unit (WPI, Sarasota FL) driven by the computer. Stimulus currents were monitored by a linear optically coupled amplifier of our design. All signals were electronically low pass filtered with a cutoff of 10 kHz. All records were taken at 37 ± 0.5 °C.
Histology, nerve grading, and immunofluorescence
Optic nerve processing for plastic sectioning and optic nerve severity grading (3, 4) were performed as previously described. Immunofluorescence of whole-mounted retinas (4, 19, 25) and cryosectioned optic nerves (29, 30) were performed as previously described using the following antibodies: rabbit anti-cCASP3 (AF835, R&D, 1:1000), rabbit anti-RBPMS (GTX118619, GeneTex, 1:250), chicken anti-GFP (ab13970-100, Abcam, 1:1000), rabbit anti-pJNK (4668S , Cell Signaling, 1:250), rabbit anti-pJUN (3270S, Cell Signaling, 1:250), donkey anti-rabbit (A31572 and A-21206, ThermoFisher, 1:1000), donkey anti-mouse (A31570, ThermoFisher, 1:1000), and donkey anti-chicken (703-545-155, Jackson ImmunoResearch, 1:1000). RGC soma sizes (measured using images assessed for RGC soma survival) were quantified using ImageJ by using a Gaussian blur filter with a sigma of 4, converting images to binary with an automatic triangle thresholding setting allowing detection of RGC somas. Somas were separated using the watershed function. Somas were defined as an area ≥30µm2 and ≥.2 circularity. RGCs cut off at the boarder of the image were excluded from analysis. Average soma area per image was measured, and 8 images were averaged per retina.
Results
Ddit3 and Jun control RGC somal loss after ocular hypertension
DDIT3 and JUN together have been shown to regulate death of RGC somas after CONC, potentially additively (2). It has yet to be determined whether somal DDIT3 and JUN activity initiates axonal degeneration mechanisms. To determine whether both JUN and DDIT3 act in tandem to elicit RGC axonal degeneration in the context of glaucoma, Ddit3 and/or Jun were deleted from the full body and neural retina, respectively, from ocular hypertensive DBA/2J mice.
Neurodegeneration in the DBA/2J model of glaucoma is dependent upon elevated IOP (32–37). Some genetic manipulations have been reported to lower IOP in DBA/2J mice (25). However, with Six3-cre, Jun is not deleted in the structures in the anterior segment of the eye that control IOP, and D2.Six3-cre+Junfl/flmice did not have altered IOP compared to WT controls (3). Furthermore, full-body deletion of Ddit3 did not alter the IOP profile of the DBA/2J mouse model (4). To ensure combined Ddit3/Jun deletion did not lessen IOP elevation, IOPs were measured from D2.Ddit3+/?Jun+/?(WT) and D2.Six3-cre+Ddit3-/-Junfl/fl (Ddit3/Jun-/-) mice at 5M, 9M, 10.5M, and 12M of age. Mice of both genotypes had elevated IOP at each timepoint compared to 5M, and IOP was not lowered compared to WT at each timepoint measured. Therefore, Ddit3/Jun deletion did not substantially alter the profile of ocular hypertension typical of the DBA/2J model (Fig. 1A).
To determine whether DDIT3 and JUN control death of the entire RGC after ocular hypertension, Ddit3/Jun-/-and WT control optic nerves were assessed for axonal degeneration at 12M—a timepoint at which roughly 50% of DBA/2J optic nerves will have severe optic nerve damage (26). Ddit3/Jun deletion did not lessen instances of severe glaucomatous neurodegeneration. In fact, Ddit3/Jun-/- mice had slightly worse outcomes relative to WT controls (Fig. 1B). Therefore, somal DDIT3 and JUN did not act in tandem to perpetuate axonal degeneration after chronic ocular hypertension.
Several molecules contribute to degeneration of the soma, but not the axon, after glaucoma-relevant injury (1, 3, 4). To determine whether DDIT3 and JUN play an important role in RGC somal degeneration after severe axonal injury, 12M Ddit3/Jun-/- and WT retinas with corresponding severe optic nerves were assessed for RGC somal survival. Ddit3/Jun deletion conferred robust (77%) protection to RGC somas in retinas with severe optic nerve degeneration (Fig. 1C). These data suggest DDIT3 and JUN are the critical regulators of somal death but do not contribute to axonal degeneration in glaucoma.
Despite somal survival, Ddit3/Jun deletion did not prevent PERG amplitude decline or somal shrinkage
Ocular hypertension is known to cause impaired RGC somal gross potentials (5–7, 38, 39), even before the onset of detectable axonal damage (39, 40). Others have shown neuroprotective treatment or genetic manipulation also preserved physiological activity in glaucoma-relevant models (5–7, 38). Given Ddit3/Jun deletion protected most RGC somas after chronic ocular hypertension (Fig. 1C), it remained important to determine whether surviving Ddit3/Jun-/-RGC somas retained physiological function. To test this, pattern electroretinograms (PERGs) were longitudinally recorded from WT, D2.Ddit3-/-Jun+/? (Ddit3-/-), D2.Six3-cre+ Ddit3+/?Junfl/fl (Jun-/-), and Ddit3/Jun-/- animals at 5M (before the onset of IOP elevation and RGC degeneration (26)), 9M (when IOP is elevated, but morphologically detectable RGC degeneration has not yet occurred (26)), and 12M of age (when roughly 50% of eyes have severe glaucomatous neurodegeneration (26), Fig. 1B). D2.Gpnmb+ (Gpnmb+) animals were assessed as an age- and background-matched normotensive control (35).
Compared to non-glaucomatous Gpnmb+ age-matched controls, PERG amplitude significantly declined in all genotype groups over time (Fig. 2A, B). As previously observed, PERG amplitude significantly declined by 9M for DBA/2J mice compared to Gpnmb+ age-matched controls (39), and Ddit3/Jun deficiency did not prevent this decline. Similar results were observed at 12M. Thus, despite conferring robust protection to RGC somas, Ddit3/Jun deletion did not prevent loss of RGC somal function. Notably, PERG amplitude decline did not appear to be caused by photoreceptor or bipolar cell dysfunction or improper light penetration. ERG a- and b-wave amplitudes declined slightly with age for all genotype groups, including Gpnmb+ eyes, but not nearly to the same extent as PERG amplitude decline (Fig. 2C, D). %PERG and ERG a- and b-wave amplitudes relative to respective 5M controls are listed in Table 1.
Notably, PERG amplitude decline did not depend on morphologically observed RGC loss. Regardless of genotype, 12M eyes with no or early and severe glaucomatous damage had similarly reduced PERG amplitudes relative to age-matched Gpnmb+ controls (Fig. 2E). Interestingly, Ddit3/Jun deletion did not prevent ocular hypertension-induced shrinkage of the RGC soma (Fig. 2F), suggesting surviving RGCs are likely injured and/or undergoing metabolic stress (41, 42). Thus, despite conferring protection from somal loss, Ddit3/Jun deletion did not preserve RGC somal viability, at least as measured by gross potentials and soma size. These data suggest the mechanism(s) driving somal shrinkage and loss of gross potentials must act either upstream or independently of somal DDIT3/JUN activation.
Mkk4 and Mkk7 deletion prevented somal and axonal degeneration after axonal injury
MKK4 and MKK7 (MAP2Ks 4 and 7) are known to control DDIT3 signaling and all JUN activation after injury, and thus may drive RGC somal death after axonal injury. In addition, MKK4/7 activate a variety of downstream targets independently of DDIT3/JUN activation, which may contribute to decline of somal viability and axonal degeneration. In fact, recent work has suggested the importance of both MKK4 and MKK7 in driving axonal degeneration (16, 43). Therefore, MKK4/7 activation may drive degeneration of the entire RGC. To test this possibility, Mkk4fl and Mkk7fl alleles were recombined from RGCs using intravitreal AAV2.2-delivered Cmv-cre (AAV2.2-Cmv-cre-Gfp) to generate animals with Mkk4/7-/- RGCs. AAV2.2 with no cre vector (AAV2.2-Cmv-Gfp) was intravitreally injected into Mkk4?Mkk7? mice to generate Mkk4/7+/+ controls.
To ensure sufficient recombination of Mkk4/7 floxed alleles, activation of MKK4/7’s downstream targets was evaluated after CONC. Robust JNK activation in the optic nerve head and JUN activation in RGC somas occurs early after CONC (2, 44). Compared to Mkk4/7+/+controls, Mkk4/7-/- eyes had little appreciable JNK activation in the optic nerve head (Fig. 3A) and had a 91% reduction of RGCs with JUN activation (Fig. 3B) after CONC. Therefore, AAV2.2-delivered Cmv-cre effectively recombined Mkk4/7 floxed alleles in RGCs. To determine whether MKK4 and MKK7 together play an important role in RGC somal loss after glaucoma-relevant injury, RGC soma survival was assessed for Mkk4/7+/+ and Mkk4/7-/- mice after CONC. Mkk4/7 deletion provided robust and sustained long-term protection to RGC somas after CONC—Mkk4/7 deletion prevented the vast majority of caspase 3 activation 5 days post-CONC (Fig. 4A) and preserved ∼90% of RGC somas at both 14 days and 2 months post-CONC (Fig. 4B). Therefore, MKK4 and MKK7 controlled RGC somal death after glaucoma-relevant injury.
Recent evidence has suggested a role for MKK4/7 in driving Wallerian degeneration cascades after glaucoma-relevant injury (16, 43). To clarify the role of MKK4/7 in axonal degeneration, RGC axonal integrity was evaluated for Mkk4/7+/+ and Mkk4/7-/- optic nerves after glaucoma-relevant injury. Consistent with previous reports (16, 43), Mkk4/7-/- RGC axons had substantially fewer histological indications of degeneration (Fig. 5A). Importantly, Mkk4/7 deletion preserved axonal physiological function as assessed by compound action potentials (CAPs, Fig. 5B, C). These data show that together, MKK4 and MKK7 control degeneration of both the RGC soma and axon after axonal injury.
Given PERG amplitude decline and soma shrinkage were not prevented with Ddit3/Jun deletion despite robust somal survival, it remained important to assess RGC somal and function size in Mkk4/7-/-eyes after glaucoma relevant injury. In striking contrast to Jun/Ddit3 deletion, Mkk4/7 deletion significantly attenuated PERG amplitude decline (Fig. 6A) and soma shrinkage (Fig. 6B) after CONC. These data suggest MKK4/7 govern not only somal and axonal survival, but also drive loss of viability and gross function. Thus, activation of MKK4/7 is likely a critical inciting event integrating mechanisms controlling death and degeneration of somal and axonal RGC compartments in the context of axonal injury.
Discussion
In glaucoma, injury to RGC axons drives degenerative signaling cascades that are critical for eliciting degeneration of the RGC soma and axon. The mechanisms governing RGC degeneration act in a compartment-specific manor (1, 2, 4, 29, 30). BAX-dependent apoptosis governed degeneration of the RGC proximal segment, but did not contribute to axonal degeneration (1). Several lines of evidence have suggested axonal degeneration mechanisms are critical drivers initiating cell death pathways ultimately driving degeneration of all RGC compartments (5, 7). Therefore, mechanisms important in axonal degeneration that can lead to downstream somal BAX activation have been recent targets of investigation.
The present study investigated the degenerative mechanisms by which RGCs die after glaucoma-relevant injury. Specifically, we examined the dual role of DDIT3 and JUN and their upstream regulators MKK4/7 in controlling RGC death after glaucoma-relevant injury. DDIT3/JUN controlled the majority of RGC somal death in ocular hypertensive DBA/2J mice but did not control axonal degeneration. Despite playing a critical role in RGC somal loss, DDIT3 and JUN did not contribute to decline in somal viability as measured by RGC somal shrinkage and loss of PERG amplitudes in DBA/2J mice or after CONC. In contrast, Mkk4/7 deficiency significantly lessened not only somal apoptosis but also PERG amplitude decline, somal shrinkage, and the rate of axonal degeneration after glaucoma-relevant axonal injury. Together, these data suggest activation of MKK4 and MKK7 is a critical early event after glaucoma-relevant injury which activates pathways governing degeneration of RGC axons and somas. Future studies should assess the importance of MKK4/7 in driving RGC degeneration after chronic ocular hypertension. Furthermore, identifying the upstream activators of MKK4/7 in the context of axonal injury and elucidating the DDIT3/JUN-independent downstream effectors of MKK4/7 driving loss of somal viability and axonal degeneration will be important next lines of investigation.
Upstream regulators of MKK4 and MKK7 have previously been implicated in driving axonal and somal degenerative cascades after glaucoma relevant injury. Several studies have suggested the importance of Dual leucine kinase (DLK, MAP3K12), an upstream activator of MKK4 and MKK7, in initiating cell death pathways important in somal and axonal degeneration (30, 45, 46). Inhibition of DLK lessened Wallerian degeneration after axonal injury in vitro (15, 16, 46) and in vivo (16, 46). DLK inhibition modestly protected RGC somas and axons in a model of inducible ocular hypertension (47). However, deletion of Dlk did not phenocopy Mkk4/7 deletion’s protection to RGC axons after CONC (30). Importantly, Dlk deletion was not sufficient to prevent axonal JNK activation after CONC (30), suggesting Dlk is not the sole activator of MKK4 and MKK7 after mechanical optic nerve injury. For example, MKK4/7 are also known to be activated by the MAP3K LZK (48, 49). Thus, it is possible that multiple MAP3Ks contribute to MKK4/7 activation in the context of glaucoma-relevant injury. Regardless of their upstream activators, MKK4/7 together controlled death of somal and axonal compartments after axonal injury. Our findings suggest MKK4/7 drive somal loss via activation of DDIT3 and JUN. However, axonal degeneration, somal functional decline, and soma shrinkage must be driven by MKK4/7 via a mechanism independent of DDIT3/JUN activation.
Much research has begun to uncover the mechanisms important in governing Wallerian axonal degeneration—many of which may be controlled by MKK4/7 after glaucoma-relevant injury. For example, recent work has suggested the importance of JNK signaling in axonal degeneration. Loss of JNK1, 2, and 3 activation prevented axonal degeneration in vitro (16, 43, 46) and in vivo after axonal injury (46, 50), including after CONC (16), suggesting MKK4/7 mediate axonal degeneration via activation of the JNKs. Importantly, although JNK2/3 are the JNKs known to be expressed in the central nervous system, Jnk2/3 deletion did not prevent axonal degeneration after CONC (30) or in DBA/2J mice (51). These data either indicate a critical role for JNK1 in mediating axonal degeneration or suggest MKK4/7 mediate axonal degeneration through mechanisms in addition to or independently of JNK activation.
MKK4/7 (potentially via JNK activation) may drive axonal degeneration by contributing to NMNAT2 depletion after axonal injury. Walker et al. showed Mkk4/7 silencing prevented NMNAT2 degradation and subsequent axonal degeneration in cultured dorsal root ganglion cell after axotomy. Furthermore, this study showed both MKK4/7 activation and consequential NMNAT2 degradation acted upstream of pro-degenerative SARM1 activation to drive axonal degeneration (43) (NMNAT2 degradation has been shown to act upstream of SARM1 activation in several models of axon injury (15, 43, 52–55)). It has been suggested members of the MAPK cascade, including DLK, directly target NMNAT2 for degradation (14, 15). It is possible MKK4/7 and/or JNK1/2/3 directly target NMNAT2 for degradation, ultimately triggering SARM1 activation and allowing axonal degeneration after glaucoma-relevant injury. Alternatively, work done by Yang et al. suggested SARM1 activity and NMNAT2 depletion acts upstream of MKK4/7-JNK1/2/3 activation to drive Wallerian degeneration after axonal injury (16). However, recent work has shown Sarm1 deletion does not phenocopy Mkk4/7 deletion after CONC. While Sarm1 deletion protected RGC axons, it did not prevent somal JUN activation and subsequent somal death (29). Thus, it remains more likely MKK4/7 drive axonal degeneration by facilitating NMNAT2 degradation and allowing pro-degenerative SARM1 activation. Future work should elucidate the mechanisms downstream of MKK4/7 activation that ultimately drive Wallerian degeneration in the context of glaucomatous injury.
An intriguing result of our study indicated glaucoma-relevant loss of RGC gross potentials and soma shrinkage are not merely consequences of somal loss. To date, few studies have investigated the compartment-specific RGC-intrinsic mechanisms by which RGCs undergo shrinkage and lose the ability to fire in response to light-evoked inner retinal neuronal signals. Interestingly, previous work suggested Bax deletion, despite providing protection against neuronal loss, did not prevent somal shrinkage after neurotrophic deprivation. In this study, soma shrinkage corresponded with indicators of metabolic stress such as loss of glucose uptake and decreased rate of protein synthesis—which were not prevented by Bax deletion (42). In contrast, inhibition of mixed lineage kinases (and downstream MKK4/7 and JNK signaling) preserved soma size and metabolic integrity (41). Consistent with our reported results, these data indicate MKK4/7 promote deterioration of somal health independently of JUN/DDIT3-induced BAX activation.
It is possible MKK4/7 mediate somal and axonal decline via DDIT3/JUN-independent NMNAT2 degradation. WLDS and Nmnat gene therapy’s preservation of RGC electrical function in ocular hypertensive mice, suggesting a role for NMNAT2 in maintaining RGC somal viability (5, 7, 8, 56). It is also important to consider manipulations like WLDS and Nmnat1 gene therapy preserved both RGC somas and axons in DBA/2J mice (5–7, 57). This could suggest protection of both the RGC soma and axon is required for protection of somal viability after glaucoma-relevant injury. Work by Chou et al. showed temporarily blocking RGC retrograde axonal transport without injuring RGCs caused a significant and reversable reduction PERG amplitude (58). Thus, loss of axonal transport itself could lead to subsequent loss of somal gross potentials. This suggests preservation of axons (and therefore preservation of axon transport) by Mkk4/7 deletion could also preserve RGC somal gross potentials.
MKK4/7 could also drive RGC degeneration via activation of DDIT3/JUN-independent mechanisms in the soma. For example, MKK4/7 may drive somal and axonal degeneration via activation of pro-apoptotic BH3 only proteins. JNK is known to phosphorylate and activate pro-apoptotic BH3 only proteins, thereby inhibiting pro-survival Bcl-2 family proteins and allowing BAX activation (17). Overexpression of Bcl2l1 (Bclxl) prevented axonal degeneration in DBA/2J mice (28) and reduced RGC soma loss but not axonal degeneration after axonal injury (27, 59). Notably, Bclxl overexpression did not prevent somal shrinkage or PERG amplitude decline after axonal injury, suggesting BCLXL loss alone is not the sole driver of RGC somal degeneration in the context of glaucoma. Bcl2 overexpression also protected RGC somas (60–62) but not distal axons (61, 62) after CONC. Interestingly, Bcl2 overexpression prevented PERG amplitude decline even up to 2 months post-CONC (38), suggesting a potential role of Bcl2 family proteins in maintaining loss of somal viability after injury. Therefore, BCL2 or other Bcl-2 family proteins may be critical in maintaining both somal survival, somal health, and axonal degeneration, and MKK4/7-mediated inhibition of Bcl-2 family proteins may contribute to RGC degeneration in glaucoma.
In conclusion, we demonstrate MKK4/7 controlled loss of the RGC soma after glaucoma-relevant injury—likely via somal DDIT3/JUN activation. MKK4/7 also controlled RGC axonal degeneration and decline in somal viability via DDIT3/JUN-independent mechanisms. These data suggest MKK4/7 activation may be an early inciting mechanism initiating degenerative cascades in glaucoma. The DDIT3/JUN independent downstream mechanisms by which MKK4/7 drive axonal degeneration cascades, and deterioration of somal health remain unidentified. Future work should elucidate the downstream effectors of MKK4/7 in each RGC compartment and should also identify upstream activators of MKK4/7 in the context of glaucoma to further identify neurotheraputic targets.
Acknowledgments
The authors would like to thank Drs. C. Tournier and R.J. Davis for generously providing the Mkk4 and Mkk7 floxed alleles. Thank you to Alyssa Parker and Thurma McDaniel for excellent technical assistance.
List of abbreviations
- ACSF
- artificial cerebrospinal fluid
- CAP
- compound action potential
- cCASP3
- cleaved caspase 3
- CONC
- controlled optic nerve crush
- ERG
- electroretinography
- IOP
- intraocular pressure
- noe
- no or early
- PERG
- pattern electroretinography
- RGC
- retinal ganglion cell
- SEM
- standard error of the mean
- sev
- severe
- WT
- wild type