Age-related deficits in retinal autophagy following intraocular pressure elevation in autophagy reporter mouse model

This study quantified age-related changes to retinal autophagy using the CAG-RFP-EGFP-LC3 autophagy reporter mice and considered how aging impacts autophagic responses to acute intraocular pressure (IOP) stress. IOP was elevated to 50 mm Hg for 30 minutes in 3-month-old and 12-month-old CAG-RFP-EGFP-LC3 (n = 7 per age group) and Thy1-YFPh transgenic mice (n = 3 per age group). Compared with younger eyes, older eyes showed diminished basal autophagy in the outer retina, while the inner retina was unaffected. Autophagic flux (red:yellow puncta ratio) was elevated in the inner plexiform layer. Three days following IOP elevation, older eyes showed poorer functional recovery, most notably in ganglion cell responses compared to younger eyes (12 months old: -33.4 ± 5.3% vs. 3 months mice: -13.4 ± 4.5%). This paralleled a reduced capacity to upregulate autophagic puncta volume in the inner retina in older eyes, a response that was seen in younger eyes. Age-related decline in basal and stress-induced autophagy in the retina is associated with greater retinal ganglion cells' susceptibility to IOP elevation.


Introduction
Glaucoma is a multifactorial optic neuropathy characterized by progressive retinal ganglion cell (RGC) degeneration (Quigley et al., 1981) and is a leading cause of irreversible blindness (2021; Keel et al., 2019;Tham et al., 2014).Although aging has been identified as the strongest risk factor for glaucoma, the precise age-related mechanisms involved in causing RGC degeneration are still poorly understood.
Autophagy regulates cellular health and survival (Adornetto et al., 2020).Autophagy encompasses 3 major pathways (macroautophagy [also including mitophagy], microautophagy, and chaperone-mediated autophagy), facilitating the breakdown and recycling of defective or aged cellular components for energy production (Adornetto et al., 2020;Scrivo et al., 2018).The pathway of interest, macroautophagy/autophagy, is the most well-studied process, involving the formation of a double-membraned autophagosome, engulfing cellular debris, fusion with the lysosome to form the autolysosome, and, finally, breakdown of the contents.Normal levels of autophagy promote cell survival, while autophagic overactivation can initiate type 2 cell death (Ou et al., 2017).Impairment of autophagy is thought to be involved in aging and in age-related neurodegeneration (Kaden and Li, 2013;Nixon et al., 2005;Pan et al., 2008), including glaucoma (Kaden and Li, 2013).
The retina is highly metabolically demanding in order to sustain second messenger systems, synaptic transmission, and ion pumps (Ames et al., 1992).Not surprisingly, key autophagic protein markers, including microtubule-associated protein 1 light chain 3 (LC3), are expressed throughout the retina (Mitter et al., 2012).However, a detailed layer-by-layer quantification of autophagy in the aging retina has yet to be undertaken.
Even though autophagy has been linked to glaucoma, its role is not well understood.Outcomes in severe and chronic models of optic nerve injury have been divided, with some suggesting autophagy activation corresponded with RGC death (Knoferle et al., 2010;Park et al., 2012Park et al., , 2018;;Piras et al., 2011;Produit-Zengaffinen et al., 2014;Wei et al., 2015;Zhang et al., 2021), while others suggested that autophagy upregulation promoted RGC survival (Kim et al., 2008;Kitaoka et al., 2013;Rodríguez-Muela et al., 2012;Russo et al., 2011Russo et al., , 2018;;Wei et al., 2015).Few studies have considered the relationship between autophagy and the capacity of the retina to cope with mild levels of IOP elevation.Studying mild recoverable stress (He et al., 2006) can provide insight into the relationship between autophagy and age-related neurodegeneration.Therefore, this study considers the following hypotheses.Firstly, aging alters autophagy in specific retina layers, and older eyes show different autophagy responses to mild recoverable IOP elevation.

Animals
All experimental procedures have been approved by the Florey Institute of Neuroscience and Mental Health Animal Ethics Committee (18-111-UM) and performed in compliance with the National Health and Medical Research Council of Australia guidelines.All mice were housed at the Melbourne Brain Centre animal facility (Parkville, Victoria, Australia).Facility was kept at constant room temperature (20°C) with a 12-hour diurnal light-dark cycle (< 50 lux inside the cage).Standard chow (Barastoc, Ridley Corporation, Melbourne, VIC, Australia) and water were provided ad libitum.

Mouse strains and age
Autophagy quantification is possible using the dual reporter LC3 mouse (CAG-RFP-EGFP-LC3), in which the LC3 protein is tagged with both a red fluorescent protein (RFP) and an enhanced green fluorescent protein (EGFP).Colocalization of RFP and EGFP (yellow fluorophores) mark the presence of autophagosomes.During fusion, EGFP is quenched by the low pH environment of the lysosome, leaving only RFP (red fluorophores) that indicate autolysosomes, allowing quantification of autophagic flux (Li et al., 2014;Yang et al., 2021), including in the retina (Ramachandra Rao and Fliesler, 2021).For the purposes of this study, autophagic flux is defined as the ratio of all puncta (autophagosome and autolysosomes) to autophagosomes, with a higher ratio indicating greater relative number of autolysosomes.CAG-RFP-EGFP-LC3 reporter mice show a normal phenotype (Li et al., 2014).
It would also be important to quantify autophagy distribution in RGC soma and dendrites.The Thy1-YFPh mouse strain expresses intrinsic fluorescence in approximately 0.2%-10% of total RGCs (with minimal overlap between cells), making this model useful for LC3 protein immunostaining in RGC soma and dendrites (Barnstable and Dräger, 1984;Feng et al., 2000).
Based on previous studies (Kong et al., 2012b;Lee et al., 2022), it is known that by 12 months of age, mice already show greater functional susceptibility to mild IOP elevation.While 12 months is by no means an old animal, it is roughly equivalent to middle age in humans where the prevalence of glaucoma increases.In this study, we chose to compare 3-and 12-month-old mice.

Experimental protocol
Twenty-six CAG-RFP-EGFP-LC3 mice (LC3) of 2 ages were examined in the current study; 7 young (3-4 months old, hereby referred as 3-month-old) and 9 old (12-14 months old, hereby referred as 12-month-old) animals underwent IOP elevation, while 5 young and 5 old mice underwent a sham treatment (cannulation without IOP elevation).At baseline, animals were subject to acute IOP elevation (IOP; 50 mm Hg for 30 minutes) in 1 randomly chosen eye, while the contralateral eye served as an internal untreated control.After 3 days of post-IOP recovery, all animals underwent end point ERG assessment for quantification of retinal function.Animals were then immediately euthanized (overdose of pentobarbital 100 mg/kg followed by cervical dislocation), and tissue was harvested.Dissected retinas were then treated with Hoechst nuclei stain (Sigma-Aldrich, Castle Hill, New South Wales, Australia) and flat mounted for confocal imaging.

Acute intraocular pressure elevation
The acute IOP elevation model (Kong et al., 2009b) produces a recoverable injury that targets RGCs, thus providing a means to probe age-related susceptibility to "glaucomatous" stress (Kong et al., 2012b;Lee et al., 2022).Prior to cannulation, general anesthesia (ketamine and xylazine) was administered alongside topical corneal anesthetic (0.5% proxymetacaine hydrochloride; Alcon Laboratories Pty Ltd, 0998-0016-15) on the treated eye.The contralateral eye was kept hydrated with gel lubricant (GenTeal Gel; Alcon Laboratories Pty Ltd, 0065-0426-36).Under an operating microscope, 1 randomly chosen eye was cannulated parallel to and 0.5 mm anterior to the limbus using a glass micropipette (pulled from 1.0 mm diameter, 6-inch Borosilicate Glass; World Precision Instruments, Sarasota, FL, USA) connected to a micromanipulator (World Precision Instruments).Pressure was elevated to 50 mm Hg for 30 minutes by adjusting the height of a reservoir filled with Hanks balanced salt solution (Sigma-Aldrich, H6648-500ML), which had been precalibrated using a manometer (Livingstone, Sydney, New South Wales, Australia).
Following 30 minutes, the micropipette was removed from the anterior chamber, and the animal was administered atipamezole hydrochloride (1:20, Antipam injection; Jurox Pty Ltd, 504280) diluted in saline (0.9% sodium chloride; Baxter Sodium Chloride IV Infusion; Baxter Healthcare Pty Ltd, AHB1364) to reverse the effect of anesthetic and assist recovery.Animals were monitored daily until experimental end point at 3 days post-IOP elevation where retinal function was assessed via electroretinography.

Electroretinography
Scotopic full-field ERG, as previously described (Lee et al., 2022), was performed to examine retinal function at 3 days postcannulation.Prior to recording, animals were dark-adapted overnight (12 hours minimum) to optimize RGC sensitivity.General anesthesia (ketamine and xylazine) was used alongside topical corneal anesthetic (1% proxymetacaine; Alcon Laboratories Pty Ltd) and mydriasis (0.5% tropicamide; Alcon Laboratories Pty Ltd).Animals were placed on a warmed pad in the plane of a Ganzfeld bowl (Photometric Solutions International, Huntingdale, Victoria, Australia) for electrode placement.ERG electrodes (99.9% silver; A&E Metal Merchants, Sydney, New South Wales, Australia) were handmade before each recording session.Each eye was recorded by 2 electrodes, an L- shaped active placed on the corneal apex and a ring inactive electrode placed around the equator of the eye, all referenced to a ground electrode (F-E2-30; Grass Telefactor, West Warwick, RI, USA) inserted into the mouse tail.
Scope software (AD Instruments Pty Ltd., Bella Vista, New South Wales, Australia) was used to record the ERGs.Luminous energies ranging from −5.53 to 2.07 log cd.s/m 2 were delivered through a customized LED array from dimmest to brightest to extract the responses from different retinal cell classes.At the brightest light level, a twin-flash (with 0.5-second interstimulus interval) was used to isolate rod-and cone-mediated photoreceptor responses.The corneal negative a-wave photoreceptor response was modeled using a delayed Gaussian function to return the maximum amplitude (RmP3, µV), sensitivity (S, m 2 .cd−1 .s−3 ), and time delay (td, milliseconds).The positive ON-bipolar cell response was obtained by subtracting the P3 model at each luminous exposure from their respective raw ERG waveforms, then applying a low-pass filter (−3 dB at 50 Hz) and a band pass filter (−3 dB at 50 and 280 Hz) to isolate the P2 and the oscillatory potentials (OPs), respectively.The P2 amplitude as a function of luminous energy was fit with a using saturating hyperbolic function to return the maximum amplitude (V max , µV) and sensitivity (K −1 , log cd.s/m 2 ).RGC function was measured as the peak positive amplitude and its peak time at the dimmest light exposures (−4.90 and 5.01 log cd.s/m 2 ).
Thy1-YFPh retinae: Tissue fixation and dissection were as described above.Retinae were incubated for 30 minutes (room temperature, on a gentle agitator) in 100 μL of PBS diluent consisting of 0.5% Triton X-100 (MP Biomedicals, ICN807423) for permeabilization, 1% donkey serum (Sigma-Aldrich, D9663) for blocking, and 0.05% sodium azide solution (Sigma-Aldrich, 26628-22-8) as an antibiotic agent.Retinae were then incubated in 1:200 LC3-5F10 mouse monoclonal primary antibody (Nanotools Antikörpertechnik GmbH & Co.KG, 0231-100/LC3-5F10) diluted in PBS for 1 hour at room temperature, followed by 3 consecutive overnights at 4°C.After removing the primary antibody, retinae were washed thrice with 0.01 M PBS in 5-minute intervals, then incubated in 1:400 donkey anti-mouse IgG secondary antibody, Alexa Fluor 568 conjugate (Thermo Fisher Scientific, A10037) diluted in PBS for 50 minutes at room temperature.Tissues were washed again twice at 5-minute intervals with PBS, then incubated with 1:600 Hoechst nuclei stain diluted with PBS for 15 minutes at room temperature on a gentle agitator before washing with PBS 3 more times in 5-minute intervals.Retinae were then flat mounted as described above.

Confocal microscopy
All retinal tissues were imaged using the Zeiss LSM880 Airyscan Fast confocal microscope (Biological Optical Microscopy Platform, the University of Melbourne, Parkville, Victoria, Australia) paired with the Zen 2.3 SP1 recording system (Zen black software; Zeiss, Jena, Germany).This microscope was chosen for its capacity to take super-resolution images, sampled at 2.0× the Nyquist.This function allowed z-stack images to be captured at a resolution and surface area high enough to detect the small, fluorescent, intracellular LC3 autophagic puncta (0.100 µm minimum radius to calculate minimum puncta volume).
CAG-RFP-EGFP-LC3 retinae: A 7×7 tile scan map (10% overlap) was first taken with the 20× objective lens on the Hoechst channel to define central and peripheral regions.One central and one peripheral region of interest (ROI) from each retina was randomly selected and imaged with the 40× oil objective using the Airyscan function at super resolution.The 3 channels (RFP, EGFP, and Hoechst) were imaged consecutively after each z-stack to minimize movement of the imaging plane.Twelve (2 retina × 2 ROIs × 3 depths) z-stack 3dimensional (3D) images (z-stack interval = 0.169 µm; pixel size [x, y] = 0.0380 µm, 0.0380 µm) containing the 3 channels were captured per animal for both 3-and 12-month age groups.These images include 3 retinal depth layers (ganglion cell layer to inner plexiform layer [GCLIPL], inner nuclear layer to outer plexiform layer [INLOPL], and outer nuclear to photoreceptors [ONL]) at the 2 ROIs each for both control and IOP-treated eyes as shown in Supplementary Fig. S7.
Thy1-YFPh retinae: A 7×7 tile scan map (10% overlap) was first taken with the 10× objective lens on the GFP channel to define central and peripheral RGCs.Two central and 2 peripheral RGCs were randomly selected for imaging for each retina with the 63× oil objective using the Airyscan function at super resolution.At each RGC, a randomly selected quadrant was used for z-stack analysis.Eight (2 retina × 4 ROIs) z-stack 3D images (z-stack interval = 0.159 µm; pixel size [x, y] = 0.0353 µm, 0.0353 µm) containing all 3 channels (YFP, RFP, and Hoechst) were acquired per animal for both 3-and 12-month age groups.These images include 4 ROIs each in both control and IOP-treated eyes.
All confocal images were then processed via Zen Black Airyscan processing (Zen Black software) and converted to Imaris 3D image file format (Imaris File Converter).It is important to note that the sampling of the retina (~0.2%) is limited to resolve individual puncta across the entire retinal depth.This approach may miss regional differences in autophagy.

Image analysis
Analysis of LC3 puncta location and depth was undertaken in a masked fashion using Imaris 3D Imaging software (Version 9.6; Bitplane AG, Oxford Instruments).
CAG-RFP-EGFP-LC3 retinae: All images were analyzed across the 3 retinal depths (GCLIPL, INLOPL, and ONL), which were demarcated by Hoechst nuclei position and EGFP background fluorescence (Supplementary Fig. S7).The Imaris surface generation tool was used to quantify the number and volume of LC3 fluorophores in both the early and late stages (colocalized and RFP channels, respectively) using automated thresholding to detect RFP puncta by outlining the maximum puncta size and average fluorescence intensity.A mask using the RFP (red) puncta was then used to isolate the colocalized (yellow) puncta from the EGFP (green) channel.Please note that for a z-stack, the same threshold was applied to all slices.Nuclei surfaces were created from the Hoechst channel to demarcate the retinal layers.
Thy1-YFPh retinae: All images were analyzed for Thy1-YFPh RGC volume and the associated endogenous LC3 puncta volume, distance, and depth parameters.The Imaris surface generation tool was first used to map and mask out RFP, which was highly saturated along the blood vessels.A clean surface map of the RGC soma and dendritic was used to mask and detect RFP puncta within the RGC soma and dendrites and for total LC3 expression surrounding the RGC.Due to residual saturation around blood vessels affecting automatic thresholding, manual thresholding was optimized and kept consistent between images.Nuclei surfaces were also created from the Hoechst channel.

Group data analysis
Volume data and associated x, y, and z coordinates of all surfaces created were extracted from all 3-month-old and 12-month-old eyes.Maximal z-depth for each image ROI was used to normalize all puncta.Central and peripheral regions were averaged per retina in each animal.
CAG-RFP-EGFP-LC3 retinae: RFP puncta, colocalized puncta, and nuclei surface volumes were averaged within their age group to produce a single value per parameter.The yellow colocalized puncta represented autophagosome volume prior to fusion with the lysosomes (i.e., before EGFP quenching).Meanwhile, total RFP volume represents the volume of all autophagic bodies, and thus, autophagic flux is given by the red:yellow volume ratio (%).
Thy1-YFPh retinae: Volume data from RGC soma and/or dendrites and LC3 puncta within the RGC was returned.In addition, LC3 puncta in the entire ROI surrounding the RGC.Hoechst nuclei surfaces were used to express LC3 expression as a function of RGC dendritic depth in the inner plexiform layer (IPL).RGC images were averaged within each age group to produce a single value per parameter.

Statistical analysis
Group data are presented as mean ± standard error of the mean unless stated otherwise.Statistical analyses were performed in GraphPad Prism (version 9.2.0;San Diego, CA, USA).To examine effects across one factor (e.g., age), a nonpaired t-test was used as appropriate.To examine effects across 2 factors (e.g., age × treatment and treatment × layer), a 2-way analysis of variance (ANOVA) or repeated measures ANOVA was used.In cases where missing data were present, a restricted maximum likelihood analysis was conducted, which facilitates "repeated measures" analysis with missing data points.Post hoc comparisons were performed with Šidák's correction factor.A ROUT outlier test was used to identify and remove outliers in retinal depth assessment.

Age-related decline in function with acute intraocular pressure injury
At 3 days after mild IOP elevation, retinal function was impaired in both 3-month-old and 12-month-old eyes (Fig. 1A).Attenuation of responses was more prominent at lower light levels, indicating that RGC-mediated responses were most affected.Importantly, functional deficits were greater in older eyes compared with younger eyes (Fig. 1B) at both dim and brighter light levels.
This age-related functional susceptibility was confirmed using 2way ANOVA, which identified a significant interaction between age and IOP for photoreceptor amplitude (p = 0.034).Post hoc comparison identified that a reduction in the photoreceptor amplitude between IOP-treated eyes and contralateral controls was only present in older mice (IOP, −402 ± 71 μV vs. control, −640 ± 32 μV, p = 0.012; Fig. 1C).
Like the photoreceptor response, there was a significant age and IOP interaction (p = 0.035, Fig. 1D) for bipolar cell amplitude (VmaxP2).Post hoc analysis showed significantly smaller amplitudes in older IOP-treated eye (IOP, 485 ± 83 μV vs. control, 892 ± 106 μV, p = 0.0047) and not younger eyes.Bipolar cell sensitivity (K) showed no IOP treatment effect, but there was a significant increase in sensitivity with age (p < 0.001, Fig. 1E).
The ganglion cell-mediated response (pSTR) showed a significant age and IOP treatment interaction (p = 0.017).Specifically, there was a greater decline in older eyes (IOP, 9.1 ± 2.3 μV vs. control, 42.4 ± 6.3 μV, p = 0.0002) compared to younger eyes (IOP, 23.3 ± 3.5 μV vs. control, 36.7 ± 3.8 μV, p = 0.028, Fig. 1F).Finally, OPs, which provide information about inner retinal feedback circuits involving amacrine cells, showed significant age and IOP treatment effects (p = 0.001 and p = 0.017, respectively, Fig. 1G); however, there was no significant interaction effect (p = 0.39).There was a significant decline in OPs with IOP treatment that was similar in both young and older eyes.The effect of sham treatment is summarized in Supplementary Fig. S1, with older eyes also showing some evidence of greater susceptibility to injury, although not to the same magnitude as with IOP elevation.

Characterization of autophagic puncta across retinal layers
Fig. 2 shows representative retinal layers captured from a single young CAG-RFP-EGFP-LC3 contralateral control eye from the IOPtreated group to illustrate changes in autophagy with retinal depth.The full extent of retinal depths was analyzed in 3 segments according to retinal layers demarcated by the Hoechst nuclei staining: (1) ganglion cell layer (GCL) to IPL, (2) inner nuclear layer (INL) to outer plexiform layer (OPL), and (3) outer nuclear layer (ONL).In young contralateral control eyes, colocalized (autophagosome, yellow fluorophores) and RFP-tagged (all autophagic puncta, red fluorophores) LC3 expression was most pronounced in the GCL, OPL, and ONL (Fig. 2).There were fewer autophagic puncta in the IPL.This is summarized in Fig. 3, with RFP puncta volume for both young and older eyes plotted as a function of retinal depth.The 3 retinal segments are expressed in terms of relative depth to account for variation in total retinal depth between eyes.The autophagosome puncta showed similar trends to RFP throughout all retinal layers for both age and IOP effect (Supplementary Fig. S2).
It is worth noting that the retinal layer designations depended on the quality of the Hoechst penetration on retinal flat mounts.It was of interest that nuclei staining in the INL and ONL showed significant age and depth effects, with significantly more nuclei highlighted in the older retinae (Fig. 3D-F).A possible explanation for this is that age-related changes to tissue make it more permeable, allowing the Hoechst stain to penetrate deeper into the retinae (Fig. 3E and F).As the GCL is superficial, stain penetration would be less of an issue (Fig. 3D).

Autophagic flux changes with age
Fig. 3 shows that in 3-month-old eyes, RFP puncta volume had a peak at and around the RGC nuclei (~30 µm 3 ) as well as in the IPL (Fig. 3D and G).In the middle retina, particularly in the OPL, RFP puncta volume peaked at ~80 µm 3 (Fig. 3E and H).In the outer retina, near the outer edge of the ONL, RFP puncta volume was the highest, with a peak of ~200 µm 3 (Fig. 3F and I).
In comparison to younger eyes, 12-month-old control eyes showed decreased autophagic puncta volume in the IPL, OPL, and ONL (Fig. 3G-I).Significant retinal depth and interaction (depth × age) effects were found across all retinal layers (2-way RM ANOVA [p < 0.0001]).Post hoc analysis revealed significant reduction in total RFP puncta volume in what could be considered the OFF sublayer of the IPL (~75%-100% of GCLIPL relative depth), as well as the OPL and ONL.
Autophagic flux was quantified as the ratio of red to colocalized yellow puncta.Interestingly, 3-month-old eyes display a relatively constant autophagic flux across all retinal layers (Fig. 3J-L).Comparatively, 12-month-old eyes showed a significant increase in autophagic flux in the OFF region of the IPL (p = 0.014) and the INL (p = 0.019) despite an overall reduction in RFP puncta volume in these areas (Fig. 3G and H).Age-related changes were evident in both ON and OFF regions of the IPL (Supplementary Fig. S3).A sharp rise in flux in the outer margins of the ONL was also observed; however, care should be taken with the interpretation of data in this (G) Peak oscillatory potential amplitude (OP, µV).A 2-way analysis of variance was performed comparing age (3 mo vs. 12 mo) and treatment (control vs. IOP-treated eye), significant effects outlined in the results.Post hoc analysis was performed using Šidák's multiple comparison tests, with significant effects denoted as * (p < 0.05).IOP, intraocular pressure.region particularly, given the possibility of damage to photoreceptors during tissue processing (Fig. 3J-L).

Acute intraocular pressure elevation activates inner retinal autophagy in young eyes but not older eyes
RFP puncta volume across relative retinal depth is compared between young control and IOP-treated eyes in Fig. 4. Two-way repeated measures ANOVA revealed a significant increase in RFP puncta volume in the inner retina 3 days following acute IOP elevation (Fig. 4A, p = 0.001).This effect was not observed in the outer retina (Fig. 4B and C).Šidák's post hoc analysis showed that the GCL showed significant IOP effects (Supplementary Figs.S4 and S5).Interestingly, autophagic flux remained constant between control and treated eyes throughout the retina 3 days following IOP elevation (Fig. 4G-I).It is worth noting that autophagic flux in control eyes showed greater variability in the outer retina (Fig. 4I, similar to that shown in Fig. 3L).
In contrast to the increase in RFP volume seen in younger eyes, IOP elevation did not appear to affect RFP puncta volume or autophagic flux in any retinal layers in 12-month-old eyes.There was an interaction effect in the ONL for autophagic flux; however, as previously noted (Fig. 4D-F), the outer retina shows high variability in autophagic flux, particularly at the border of the ONL (Fig. 4L).
As summarized in Fig. 5, there was an overall age-related reduction in the production of colocalized (Fig. 5A and B) and RFP (Fig. 5C and D) autophagic bodies throughout inner and outer layers of the retina, alongside a significant increase in autophagic flux in older eyes (Fig. 5E and F).Acute IOP elevation stress evoked a significant increase in autophagic bodies in the GCL in young eyes but not older eyes.

Acute intraocular pressure elevation leads to increased autophagy around RGC soma in young eyes
In addition to the CAG-EGFP-RFP-LC3 mouse strain, autophagy responses to IOP elevation in young and older RGCs were compared by examining autophagic bodies (LC3-II) in intrinsically fluorescent cells from Thy1-YFPh mouse retinae (Fig. 6 as well as Supplementary Fig. S6).
LC3 puncta volume was also quantified in the region surrounding the RGC (Fig. 7A and B) as a function of depth, with 100% indicating the maximum extension of the respective RGC dendrites.A 2-way repeated measures ANOVA showed that IOP elevation in 3-monthold eyes resulted in a significant increase in LC3 puncta near the RGC soma (Fig. 7A, 0%-25% of the relative depth, p = 0.0015).In contrast, IOP elevation did not induce a change in RGC LC3 puncta in older eyes (Fig. 7B, p = 0.83).
When LC3 puncta located in Thy1-labeled RGCs were analyzed as a function of depth, the same effect was evident.There was a significant interaction between depth and IOP treatment in young eyes (Fig. 7C, p = 0.0049) but not older eyes (Fig. 7D, p = 0.47).Analysis of LC3 puncta as a function of their distance from the RGC soma shows that IOP elevation in 3-month-old eyes caused an upregulation in the number of autophagic bodies located close to the RGC soma (Fig. 7E).In contrast, 12-month-old eyes showed a significant decrease in LC3 puncta near the cell body 3 days following IOP stress (Fig. 7F).These data support those for CAG-RFP-EGFP-LC3, showing that young and older eyes demonstrate different autophagic responses to IOP-related injury.

Discussion
This study showed for the first time that autophagy differs in different layers of the retina.With age, autophagic puncta volume is reduced throughout the retinal layers except for the GCL, but autophagic flux was increased in the inner retina.In younger eyes, acute IOP elevation increased the total volume of autophagic puncta, an effect not seen in older eyes.This upregulation of autophagic puncta in the GCL of younger eyes paralleled better ganglion cell functional recovery from stress.This may suggest that older eyes Fig. 3. Aging effect on autophagic activity throughout retina.Average distribution of RFP (all LC3) puncta volume and autophagic flux (mean ± standard error of the mean) through normalized tissue depth in the GCLIPL, INLOPL, and ONL retinal layers of 3-mo-old (blue circles, n = 7 animals, 2 regions from each eye) and 12-mo-old (orange squares, n = 7 animals, 2 regions from each eye) contralateral control eyes.(A-C) Representative confocal images at retinal depths showing largest difference in LC3 puncta volume between 3-mo-old and 12-mo-old eyes.All images were taken on a super-resolution confocal laser scanning microscope using a 40× objective lens, then processed, resized, and brightness adjusted for viewing purposes.Scale bar = 10 µm.(D-F) Hoechst nuclei volume (µm 3 ).(G-I) RFP puncta volume (µm 3 ).(J-K) LC3 flux calculated as the red:yellow puncta ratio.A 2-way repeated measures analysis of variance was performed comparing age (3 mo vs. 12 mo) and relative depth (from 0% to 100% in increments of 2.5%), with significant effects denoted as # (p < 0.05) in the top right corner.Post hoc analysis was performed using Šidák's multiple comparisons test, with significant effects denoted as * (p < 0.05).GCLIPL, ganglion cell layer to inner plexiform layer; INLOPL, inner nuclear layer to outer plexiform layer; LC3, microtubule-associated protein 1 light chain 3; ONL, outer nuclear to photoreceptors; RFP, red fluorescent protein.
become more vulnerable to injury because of age-related dampening of autophagy responses, as has been observed in other neurodegenerative diseases (Barbosa et al., 2018).
Studies using immunohistochemical staining of cryosections identify LC3-II fluorescence peaks in the GCL (Kim et al., 2008) and ONL (Kim et al., 2008;McWilliams et al., 2019).Using the CAG-RFP-EGFP-LC3 reporter line, the current data show that in young, healthy eyes, autophagic puncta volume peaks in the GCL, IPL, OPL, and ONL.The volume of autophagic bodies is some 2-3 times higher in the outer retina compared with the inner retina.
The observed autophagic peaks align with regions of high energy production and demand in human (Andrews et al., 1999), nonhuman primate (Wong-Riley et al., 1998), and rodent retina (Rueda et al., 2016).Specifically, intensity profiles for cytochrome-c and Na + /K + -ATPase immunostaining (Wong-Riley et al., 1998) are consistent with areas of high metabolic activity in the retina (Country, 2017;Liu and Prokosch, 2021), specifically at the level of photoreceptor inner segments (Hagins et al., 1970;Yao et al., 2014;Zhou et al., 2015), the OPL and the retinal nerve fiber layer.Areas of high metabolic demand can also be seen in oxygen distribution profiles, indicating high consumption at the level of photoreceptor inner segments and IPL in rodent retina (Cringle et al., 2002).In addition, autophagy peaks (Fig. 3) were also found in the OPL and GCL; these peaks are consistent with the location of retinal capillary beds (Kornfield and Newman, 2014), which are also regions of high Na + / K + -ATPase expression (Wong-Riley, 2010).
In the GCL and retinal nerve fiber layer, RGCs are largely reliant on oxidative phosphorylation for energy, as indicated by the high number of mitochondria and the high energy demands for mitochondrial trafficking in axons (Ito and Di Polo, 2017).Autophagy is thought to have a cytoprotective effect on RGCs by recycling aged or damaged mitochondria and thus reducing the risk of oxidative damage (Boya et al., 2016).
The current study highlights the effect of aging on basal autophagy in different retinal layers.Specifically, in older eyes, LC3 puncta volume was preserved in the GCL; elsewhere in the retina, basal autophagy was reduced (Fig. 3G-I).Studies in other tissues, including the rodent retina (Rodríguez-Muela et al., 2013), brain (Kaushik et al., 2012), heart (Hua et al., 2011), and kidney (Kume et al., 2010) indicate that autophagic bodies are reduced with age.Age-related LC3-II reductions are on the order of ~25% (Hua et al., 2011;Kaushik et al., 2012;Rodríguez-Muela et al., 2013), consistent with the changes observed in our study.Few studies have reported an increase in autophagic flux that was observed in the IPL and INL in the current study (Fig. 3J-L).
In the retina, Nettesheim et al. (2020) provide some evidence that intermediate LC3 bodies, as well as P62 and LAMP1 staining are reduced in 18-month-old C57Bl6 mice compared with 4-month-old animals.Kozhevnikova et al. (2019) utilizing RNA-Seq have shown in 18-month-old Wistar rats that gene regulation of autophagy is significantly altered.A genome-wide screen in human cells also showed similar reductions in autophagy markers with normal aging and demonstrated that age-related change in reactive oxygen species modifies autophagy (Lipinski et al., 2010).Age-related oxidative stress associated with prolonged mTORC1 activation and autophagy inhibition is also known to lead to accumulated cell damage and reduced tissue repair in the brain (Laplante and Sabatini, 2012).This has yet to be investigated in the eye; however, these mechanisms may be important as impaired autophagy increases the risk of agerelated diseases, including age-related macular degeneration (Hyttinen et al., 2021).While we show here that macroautophagy declines with age, there is also emerging evidence of a compensatory relationship between macroautophagy, chaperone-mediated autophagy activation (Cuervo and Wong, 2014;Kaushik and Cuervo, 2018), and the ubiquitin-proteasome system (Du et al., 2009) in the brain and eye with aging.Thus, a more comprehensive study of autophagy processes would help us better understand changes to self-repair processes in aging.
The current results also show that autophagic flux is increased in aging, but only in the IPL.Why only the IPL should show an age- ).(E and F) LC3 flux (%), calculated as the red:yellow puncta ratio.A 2-way analysis of variance was performed comparing age (3 mo vs. 12 mo) and treatment (control vs. IOP), with significant effects denoted as # (p < 0.05) in the top right corner.Post hoc analysis was performed using Šidák's multiple comparisons test, with significant effects denoted as *(p < 0.05).related increase in autophagic flux is unclear.The IPL is where RGC dendrites stratify and receive abundant synaptic inputs from bipolar cells and amacrine cells (Liu and Prokosch, 2021), and this layer is particularly vulnerable to injury (Yu et al., 2013).This age-related increase in flux, together with the finding that autophagic bodies are not reduced in the inner retina with age, might reflect a relative preservation of the RGCs and its synapses to maintain retinal output.This may be consistent with previous studies showing an independent preservation of ganglion cell function, despite declines in outer retinal photoreceptor and bipolar cell responses (Charng et al., 2011;Terzibasi et al., 2009).
A key finding of the current study was that in response to IOP stress, young eyes showed upregulated autophagy in the inner retina, a response not found in older eyes.This increase in LC3 puncta in RGCs after injury in young eyes was generally consistent with findings from other experimental models of glaucoma in rats, including retinal ischemia-reperfusion (110 mm Hg for 60 minutes) (Piras et al., 2011;Wei et al., 2015) and 3 days after optic nerve transection (Kim et al., 2008;Rodríguez-Muela et al., 2012), and 7 days after episcleral vein cauterization (Park et al., 2018;Park et al., 2012).A model of IOP elevation using intracameral injection of magnetic microbeads shows that 4 weeks of ocular hypertension led to hypoxic RGCs, muller cells, microglia and astrocytes in the retina, ON degeneration, and decreased RGC density accompanied by increased autophagic induction (LC3-II: LC3-I via Western blot analysis) (Jassim and Inman, 2019).
The absence of a significant change in autophagy after IOP injury in older eyes was unexpected.The results from Nettesheim et al (2020) suggest that LC3 staining in the GCL increased following 6 weeks of chronic IOP elevation induced by hypertonic saline injection in 18-month-old but not 4-month-old mice.This discrepancy raises the possibility that autophagic responses in older eyes may show different responses to acute (current) compared with chronic injury (Morrison et al., 1997).Contrasting the effects of acute and chronic IOP elevation in young and older eyes would provide useful insights.
Why IOP-related stress does not induce the same upregulation of autophagy in older eyes is unclear.The age-related preservation of autophagy in the inner retina may mean that any capacity for stressrelated upregulation is used up, which may account for poorer functional recovery in older eyes (Charng et al., 2013;Kong et al., 2012b;Lee et al., 2022).
In young eyes, LC3 increased in both RGC soma and their surround, whereas in older eyes, LC3 decreased more in RGCs somas compared with the surround (Fig. 7Bi inset).This outcome suggests that there may be other cells near the RGCs, such as Müller glial cells, that respond differently to stress with age.Several studies have shown that neurons and glia are differentially regulated in response to stress (Kulkarni et al., 2020;Motori et al., 2013).
It is important to acknowledge the potential for IOP effect on the contralateral "control" eye.To address this issue, we have analyzed the autophagy response (LC3 puncta volume distribution) in the GCLIPL layer of young (3-month-old) naïve animals, as this layer showed a significant change in young eyes after IOP elevation.These naïve versus contralateral control data are presented as Supplementary Fig. S8.We find that in young animals, RFP volume is significantly higher in contralateral control eyes compared with .(E and F) Number of RGC-specific LC3 puncta at vector distances from the soma (µm).A 2-way repeated-measures ANOVA was performed comparing treatment (control vs. IOP-treated) and relative depth (from 0% to 100% in increments of 2.5%) or distance from soma (from 0 to 150 µm in increments of 5 µm), with significant effects denoted as # in the top right corner (p < 0.05).(Bi) Percentage change of mean LC3 puncta volume (IOP/control at 7.5%-12.5% relative depth) in RGC compared to surrounding quadrant.A 2-way repeated measures ANOVA was performed comparing location (RGC vs. surround) and age (3 vs 12 months old).Post hoc analyses were performed using Šidák's multiple comparisons test, with significant effects denoted as *(p < 0.05).
naïve eyes.In contrast, in older eyes, contralateral controls eyes showed significantly lower RFP volume compared to naïve eyes.Given the direction of these contralateral eye effects by comparing the IOP-treated eyes to their own contralateral control eyes we would have underestimated our IOP effects.As such, the contralateral eye effect does not change our overall conclusions.

Conclusions
The current study utilizes the CAG-RFP-EGFP-LC3 mouse to show that in older eyes, basal autophagy is reduced in the outer retina but preserved in the inner retina.In response to IOP elevation younger eyes mount an autophagy response, whereas older eyes do not.These layer-specific age-related changes to autophagy regulation in the retina may help explain RGC susceptibility to IOP elevation in aging.

Fig. 2 .
Fig. 2. Microtubule-associated protein 1 light chain 3 (LC3) puncta distribution throughout the retina in a young contralateral control eye from the intraocular pressure (IOP) treatment group.Representative confocal z-stack slices (0.169 µm thickness) showing differences in the basal expression of LC3 fluorescent markers in the inner (ganglion cell layer [GCL], inner plexiform layer [IPL]-OFF) and outer (outer plexiform layer [OPL], outer nuclear layer [ONL]) retinal layers of a 3-mo-old contralateral control eye.(A-D) All 3 (red, green, and blue) imaging channels displayed.Hoechst nuclei stain (blue) used to identify the various retinal layers, and enhanced green fluorescent protein (EGFP) and red fluorescent protein (RFP) stain were shown in green and red, respectively.(E-H) EGFP channel.(I-L) RFP channel indicating the presence of autolysosomes during late-stage autophagy.(M-P) Colocalized EGFP and RFP channels (yellow puncta) indicating the presence of autophagosomes during early-stage autophagy and illustrating autophagic flux when compared with red puncta alone.Close up of colocalized puncta within the upper left quadrant of panels M-P are displayed in magnified inset window (10 µm width).All images were taken on a super-zresolution confocal laser scanning microscope using a 40× objective lens, then processed, resized, and brightness adjusted for viewing purposes.Scale bar = 10 µm.

Fig. 6 .
Fig. 6.Overview of retinal ganglion cell (RGC)-specific LC3 puncta quantification analysis in young and older Thy1-YFPh mouse retinae.(A) Map of a 3-mo-old contralateral control Thy1-YFPh mouse retina displaying all intrinsically green fluorescent RGCs.Confocal image taken using a 10× objective lens, with the white box indicating a centrally located RGC selected for analysis.(B) Close up of the representative centrally located RGC.Image has been cropped and magnified from panel B retinal map, with the white box indicating the specific RGC quadrant analyzed.(C) RGC quadrant showing intrinsically fluorescent RGC (green) and the LC3 puncta stain (red; Nanotools, mouse anti-LC3 antibody clone 5F10).Image taken using 63× oil objective lens, and overlaying axons from neighboring cells have been digitally removed.(D) Masked red channel showing LC3 puncta (red) localized specifically within the RGC, with the white box indicating RGC soma.(E) Depth profile of RGC quadrant, showing GCL and IPL segmentation and the distribution of LC3 puncta (red) along the Z-axis.(F) Close up of the LC3 puncta (red) localized in the RGC soma.Image has been cropped and magnified from panel E and brightness adjusted for viewing purposes.Scale bars are shown in lower right corner.GCL, ganglion cell layer; IPL, inner plexiform layer; LC3, microtubule-associated protein 1 light chain 3;

Fig. 7 .
Fig. 7. Aging and IOP injury effect on autophagy throughout RGC profile.Average distribution of LC3 puncta count (mean ± standard error of the mean) in analyzed RGC quadrants from 3-mo-old (left column, n = 3 animals) comparing contralateral control (blue with white-filled circles, n = 9 cells) and IOP-treated (blue-filled circles, n = 8 cells) eyes, and 12-mo-old (right column, n = 3 animals) contralateral control (orange with white-filled squares, n = 11 cells) and IOP-treated (orange-filled squares, n = 11 cells) eyes.(A and B) LC3 puncta in quadrant of interest, excluding the RGC.Data are shown as a function of tissue depth normalized to the RGC depth (%).(C and D) Number of RGC-specific LC3 puncta normalized to the RGC depth (%).(E and F) Number of RGC-specific LC3 puncta at vector distances from the soma (µm).A 2-way repeated-measures ANOVA was performed comparing treatment (control vs. IOP-treated) and relative depth (from 0% to 100% in increments of 2.5%) or distance from soma (from 0 to 150 µm in increments of 5 µm), with significant effects denoted as # in the top right corner (p < 0.05).(Bi) Percentage change of mean LC3 puncta volume (IOP/control at 7.5%-12.5% relative depth) in RGC compared to surrounding quadrant.A 2-way repeated measures ANOVA was performed comparing location (RGC vs. surround) and age (3 vs 12 months old).Post hoc analyses were performed using Šidák's multiple comparisons test, with significant effects denoted as *(p < 0.05).