Treatment regimen.
To assess 8-AG's effect on age-related retinal degeneration, we administered 5 mg/kg/day of 8-AG to 22-month-old F344 rats via drinking water over 8 weeks. We observed that 8-AG remains stable in water at room temperature for up to 3 days (Fig. 1B), allowing for consistent daily dosing through daily water replacement with fresh 8-AG solution. We evaluated the efficacy, safety, and mechanism of action of 8-AG through ERG and endpoint assessments as illustrated in Fig. 1C.
8-AG improves the function of aged rat retinae.
We recorded scotopic and photopic ERGs at baseline and after 8 weeks of treatment in the aged rats. Aged rats showed substantially lower scotopic and photopic ERG responses than young ones, indicating diminished rod and cone functions (Supplementary Figure S1). From 22 to 24 months of age, control rats showed stable scotopic a- and b-waves (Fig. 2A&B) but decreased photopic responses (Fig. 2C). In contrast, 8-AG treated animals showed a left-shift of scotopic a-wave response curve (P < 0.0001, Fig. 2D) and increased scotopic b-wave responses (P < 0.0001, Fig. 2E). Scotopic a- and b-wave responses to 0.1 cd.s/m2 flashes increased from 30 and 164 µV to 57 and 274 µV, respectively (P < 0.0001 and < 0.01, Fig. 2D&E), indicating improved rod photosensitivity and function during the treatment period. Photopic b-wave responses of the 8-AG treated animals decreased similarly to controls (P < 0.001, Fig. 2C&F). After 8 weeks, 8-AG treated animals had significantly higher scotopic a- and b-wave responses than the water-treated control group (P = 0.0002 and < 0.0001; Fig. 2G and 2H, respectively), but similar photopic responses (Fig. 2I). Thus, 8-AG treatment for eight weeks enhanced, rather than merely preserved, the rod functions in aged F344 rats, without notably affecting cones.
8-AG improves the structure of aged rat retinae.
After an 8-week 8-AG regimen, we conducted histological analysis on the retinae (Figs. 3A and S2). The untreated aged rats, as compared to young rats, showed notable age-related thinning, reduced nuclei in the ONL, diminished outer and inner segment (OS + IS) layers, and fewer nuclei in RGC layer (Fig. 3A-E). The most severe degeneration occurred in the superior peripheral region, with a complete loss of photoreceptor layers in some of the aged retinae (Fig. 3A-D). Conversely, the 8-AG treated animals showed significantly thicker retinae (P < 0.01), higher ONL cell count (P < 0.0001), thicker OS + IS layer (P < 0.001), and increased cells count in the RGC layer (P < 0.001), compared to untreated controls (Fig. 3A-E). Particularly, 8-AG preserved the ONL at the peripheral ends, which was absent in some control retinae (Fig. 3A). The histological data indicate that 8 weeks of oral 8-AG initiated at 22 months of age provides highly effective protection against age-related retinal degeneration in F344 rats.
8-AG reduces cell death and apoptosis in aged rat retinae.
We assessed retinal cell death by the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL assay) (Figs. 4A&B and S3) and caspase-3 cleavage via immunoblotting (Figs. 4C&D and S4). Even healthy young retinae had some TUNEL+ cells in the INL and RGC layer. Aged retinae had over double the TUNEL+ cells compared to young ones (P < 0.001), but 8-AG treatment reduced TUNEL+ cells to levels close to young retinae (Fig. 4A&B, P < 0.01 comparing 8-AG vs. aged). Caspase-3 cleavage, a key apoptotic event, initiates a caspase cascade leading to programmed cell death (43). Immunoblotting revealed aging doubled the cleaved-to-intact caspase-3 ratio; 8-AG treatment reduced this ratio by 1.5-fold (Figs. 4C&D and S4), indicating it lowers age-related retinal apoptosis. These data indicates that oral 8-AG is safe and decreases cell death in aged retinae.
8-AG rescues rhodopsin level in rod photoreceptors and preserves cones.
As we observed that 8-AG increased the rod function (Fig. 2D,E,G,and H) and rescued the length of OS + IS layers (Figs. 3A&D and S2) in the aged F344 rat retinae, we then investigated if 8-AG affects the level of rhodopsin (RHO), the most abundant phototransduction components in rods ( Figs. 5, S5 and S6). RHO levels decreased considerably in the aged retinae compared to young controls, shown by both immunohistochemistry (IHC) (Fig. 5A&B, P < 0.0001) and immunoblots (Fig. 5C&D, over 3-fold decrease, P < 0.01). 8-AG treatment markedly increased RHO levels, over 2-fold compared to the control (P < 0.01 and < 0.05 in Fig. 5B and D, respectively). However, 8-AG treatment did not significantly increase other phototransduction components, such as ARRESTIN1 and PDE6B, in the aged rat retinae (Supplementary Figures S7 and S8). These findings suggest that 8-AG reverts the rods from a senescence-like state to a healthier and functional state, characterized by the restored OS + IS morphology, RHO levels, and rod function.
While 8-AG didn’t improve cone function within the 8 weeks treatment period (Fig. 2), a significant higher number of cones in the inferior retinae of 8-AG-treated aged rats were observed (Figs. 5E&F and S9), indicating some degree of cone protection.
Oral 8-AG reduces oxidative damage in Fischer 344 rats.
To investigate if 8-AG reverses aging-associated oxidative damage, we used IHC to assess malondialdehyde (MDA) level for lipid oxidation, 8-hydroxy-2'-deoxyguanosine (8-OHdG) for DNA oxidation (44, 45), and a mitochondria marker, translocase of outer mitochondrial membrane 20 (TOMM20), in young (3 m), aged (24 m) and 8-AG treated aged Fischer 344 rat retinae (Fig. 6). Aged retinae exhibited increased MDA level, especially in the RGC layer (Figs. 6A-C, S10, P < 0.0001). 8-AG did not alter overall retinal MDA levels but significantly decreased MDA staining in the RGC layer (Figs. 6A-C, S10, P < 0.05), suggesting 8-AG protects RGCs specifically from lipid peroxidation. No obvious difference was observed in MDA staining in photoreceptor OS, ONL, and INL layers. Perhaps, the potential reason that the lipid-enriched photoreceptors OS layer is more resistant to lipid oxidation may be due to its continuous renewal and phagocytosis of OS tips by the RPE cells.
Moreover, while we noted marginal increases in 8-OHdG in aged retinae, implying minimal DNA oxidative damage, 8-AG significantly reduced 8-OHdG in photoreceptor IS, INL, and RGC layers (Figs. 6D-G and S11), indicating mitigated DNA oxidation. No significant change in TOMM20 was observed with 8-AG treatment (Figs. 6D&H and S11). Thus, 8-AG exhibits potent antioxidant effects in the neural retina, reducing RGC lipid oxidation and DNA oxidation in photoreceptor mitochondria and nuclei of inner retinal neurons. We noticed that the ONL is free of 8-OHdG staining, suggesting little oxidative damage to the chromosomal DNA of photoreceptor cells, potentially due to the condensed form of chromatin.
8-AG treatment reduces the number of injury-induced Müller glia and microglia in rat retina.
Glial cells play a supportive role in maintaining the structural and functional stability of the central nervous system (CNS) (46). Müller glia and microglia are activated in response to retinal neuron stress. We immunostained F344 rat retinae for activated Müller glia with glial fibrillary acidic protein (GFAP, Figs. 7A&B, S12) and for phagocytic macrophage/microglia with cluster of differentiation 68 (CD68) and ionized calcium binding adaptor molecule 1 (IBA1, Figs. 7C-F, S13, and S14). Aged retinae had increased number of GFAP+ filaments throughout, which 8-AG reduced in central and equatorial regions (Figures. 7A&B, S12, P < 0.01), but not in severely degenerated peripheries. Similarly, 8-AG reduced CD68+ and IBA1+ cell numbers that were elevated in the untreated aged rat retinae, indicating its anti-inflammatory effects (Figures. 7C-F, S13 and S14, P < 0.0001). Notably, all CD68+ cells were IBA1+, but not vice versa, suggesting distinct macrophage subpopulations. Collectively, we show that age-related increases of activated Müller glial filaments and macrophage/microglia were effectively reduced by 8-AG treatment, suggesting its potential in mitigating age-related retinal inflammation.
Age-related accumulation of autophagosomes is reduced by 8-AG treatment.
Transmission electron microscopy (TEM) of retinal cross-sections taken at the central and peripheral areas of the retinae of young, aged and 8-AG-treated aged rats (Fig. 8 and Supplementary Data File 1) revealed more severe structural damage in the peripheral retina of aged rats. These damages include swollen mitochondria (marked with “*”), abundant electron-dense phagosomes (marked with “➜”), disorganized photoreceptor OS membranes, and fragmented inner segment (IS) mitochondria, as compared to the central region (Fig. 8A-R). Notably, IS diameter increased with age (Fig. 8G,H,P,Q), whereas the OS diameter was stable (Fig. 8D,E,M,N). 8-AG significantly reduced the age-related accumulation of the phagosomes in RPE (Fig. 8S), with no effects on the number and morphology of mitochondria in the RPE and IS (Fig. 8T). This result suggests that 8-AG treatment restored RPE phagocytosis flux which was compromised by aging.
Transcriptome of 8-AG treated retina shows downregulation of immune and stress responses.
To explore the molecular changes caused by aging and 8-AG treatment, we isolated retinae and RPE/choroids from young, aged and 8-AG treated aged rats (8 weeks treatment) and performed bulk RNA-Seq (Figure S15, Supplementary Data Files 2–5, and GEO accession number GSE254123). A total of 293 upregulated and 814 downregulated differentially expressed genes (DEGs) were identified, comparing the aged vs young retina, with over 1.5-fold difference and P < 0.05 (Supplementary Data File 2). Gene ontology (GO) pathway analysis showed that upregulated genes in aged retinae were associated with stress responses like JAK-STAT, MAPK, and ERK cascades, axon injury response, and defense mechanisms (Fig. 9A). Retina cell remodeling and stress in the aged retina is also reflected by the activation of pathways including actin filament polymerization, response to tumor necrosis factor, cellular response to amino acid starvation, and cellular oxidation detoxification (Fig. 9A). Importantly, aged retinae showed upregulation of genes involved in the pro-inflammatory signaling and production of pro-inflammatory cytokines (Fig. 9A, complement activation, response to interferons-alpha, -beta, and -gamma, NF-kappaB signaling, response to cytokines, positive regulation of chemokines, interleukin-1 beta (IL1b), IL-6 and IL10). Consequentially, biological pathways (BPs) related to the activation of immune cells were observed to be significantly upregulated (Fig. 9A, the activation of microglial cell, T cell, neutrophil, leukocyte, macrophage, B cells and upregulation of phagocytosis and autophagy). Finally, the upregulation of innate and adaptive immune responses were identified supporting the inflammatory environment in the aged retinae. The upregulation of MHC class II antigens, expressed only on the antigen-presenting cells (APCs), suggests the infiltration of immune cells in the aged retinae. Among the DEGs, upregulation of complement factors including C1s, C2, C3, and C4a, C4b, C1r, Cfh, Cfi, C1rl, was seen in the aged retinae (Supplementary Data File 2). Activation of the complement system is a significant factor contributing to age-related macular degeneration. In agreement with the immunostaining data, we observed upregulated expression of the markers activated microglia/macrophages, Cd68 and Aif1 (encoding IBA1), as well as the marker of activated Müllar glia, Gfap, in the aged retina (Supplementary Data File 2). The downregulated DEGs in the aged vs. young retinae affected BPs related to the retina or eye function and development, as well as those in response to cell stress and cell adhesion (Fig. 9B), suggesting a compromised retinal blood barrier and reduced retinal function.
8-AG treatment led to detecting 80 downregulated and 87 upregulated DEGs compared to the untreated control (Supplementary Data File 3). 8-AG reversed aging-related pathways linked to cell stress, such as ERK1 & ERK2 cascade and MAPK activity, and reduced pro-inflammatory signaling, evidenced by lower levels of interferon-gamma, NF-KappaB, and chemokine production (Fig. 9C). Consequentially, the activation of microglia, astrocytes, T cells, neutrophils, leukocytes, and B cells was downregulated by 8-AG, along with downregulation of phagocytosis and angiogenesis, showing a potent anti-inflammatory effect. Upregulated DEGs were involved in oxygen transport, neural processing, and anti-inflammatory pathways (Fig. 9D), with notable transcripts including growth hormone-releasing hormone receptor (Ghrhr) (47, 48), Gpr171 (49), and Mir-124 (50, 51), known for neuroprotective and anti-inflammatory properties (Supplementary Data File 3). Collectively, these changes suggest that 8-AG alleviates cell stress and exhibits a broad-spectrum anti-inflammatory effect and these effects could all contribute to retinal protection.
The transcriptome of RPE/choroids suggests an age-related weakening of tight junctions and upregulated inflammatory signals, partially reversed by 8-AG. RPE/choroid tissue identity was confirmed by the enrichment of RPE-specific mRNAs, including Cst3, Efemp1, Itgav, Crispld1, Itgb8, Gulp1, Rpe65, Best1, Rbp1, Rlbp1, Rgr, Lrat, Pmel, Tttr, Tyr, Tyrp1, and Ptgds (Supplementary Data Files 4 and 5) (52). Comparing aged vs young rat RPE transcriptomes, we identified 166 upregulated and 118 downregulated DEGs (Supplementary Data File 4). Three known AMD-associated genes, C3, Cfi, and Mbp (52), were upregulated in the aged rat RPE/choroids. GO pathway analysis of the upregulated DEGs points to immune responses, including complement activation (C3, C4b, and Cfi), antimicrobial humoral immune response mediated by antimicrobial peptide (Camp, Ccl27, Cxcl1, Cxcl6, Reg3g, S100a9), response to tumor necrosis factor (Ccl2; Chi3l1; Mbp; Ubd), astrocyte development (Gfap; S100a8; S100a9), cellular response to transforming growth factor beta stimulus (Cdkn2b; Ovol2; Postn; Wnt10a; Wnt2), and so forth (Fig. 10A and Supplementary Data File 4). Aged downregulated DEGs suggested down-regulation of hydrogen peroxide catabolic process (Hba-a2; Hbb; Pxdn), angiogenesis (Angpt2; Aplnr; Col15a1; Enpep; Tek), negative chemotaxis (Flrt2; Itgb3; Sema3g), regulation of cell cycle (Ccnd2; Dusp1; Mecom; Plcb1; Skil; Xiap), cell adhesion mediated by integrin (Adam17; Itgb3), response to hypoxia (Adam17; Adipoq; Alas2; Angpt2; Loxl2; Tek), as shown in Fig. 10B and Supplementary Data File 4. These changes suggest decreased tight junction and decreased anti-inflammatory responses in the aged RPE/choroid.
8-AG treatment downregulated 140 genes and upregulated 519 genes in aged RPE relative to controls (Supplementary Data File 5). The downregulated DEGs suggest the down-regulation of pathways including cell signaling and metabolism, ion/water balance, cell differentiation, and immune responses. In contrast to the aging-associated upregulation of pathways involved in immune responses, 8-AG treatment led to downregulated immune responses including responses to IL-7, macrophage activation and antimicrobial humoral immune response (Fig. 10C). The 8-AG downregulated BPs include the ion homeostasis which potentially affects water balance and blood pressure (Fig. 10C), and 8-AG is known to modulate blood pressure and water-ion balance (53).. 8-AG upregulated genes were involved in metabolism, cell adhesion and differentiation, including genes encoding tight junction proteins like claudins (Cldn2, Cldn3, Cldn4, Cldn7, Cldn19 and Cldn23, Supplementary Data File 5). Cldn-19 is dominantly expressed in RPEs (54, 55). These changes suggest 8-AG may restore RPE and choroid capillary junction integrity, thereby improving the retina-blood barrier.
Age-related accumulation of hypoxanthine and xanthine is accompanied by the age-related decline of guanine and 3’5’-cGMP.
To assess the effects of aging and 8-AG treatment to the retinal purine metabolites, for the first time in record, we quantitatively profile the purine metabolome of the retina using UPLC-MS/MS (Fig. 11). We found the abundant purine metabolites in the rat retina are 5’-AMP, 5-’GMP, inosine and adenosine (Fig. 11D,E,L,N). Compared to the young retinae, the aged rat retinae showed a substantial reduction in guanine (> 3 fold), and 3’5’-cGMP levels (> 3 fold, Fig. 11G&K), with a rise in inosine (> 2 fold), hypoxanthine (~ 2 fold) and xanthine levels (~ 3 fold, Fig. 11E,I,J), whereas the levels of remaining purines were not affected significantly. As hypoxanthine is the product of inosine, the age-related accumulation of hypoxanthine is due to the age-related accumulation of inosine (Fig. 11A). Interestingly, the RNA-seq data showed that while genes encoding the adenosine deaminases (Adar, Adarb1, Adarb2, Adat1, Adat2), and the PNPase (Pnp) are not affected by aging, the expression of guanine deaminase (Gda) and xanthine dehydrogenase (Xdh/Xo) were 5-fold and 1.7 fold in the aged retinae, respectively, compared to the young retinae (Fig. 11O and Supplementary Data File 2). Higher expression of Gda can lead to higher consumption of guanine and a higher production of xanthine, while a higher level of Xdh can lead to a higher rate of production of xanthine from hypoxanthine. These results suggest age-related accumulation of hypoxanthine and xanthine is due to the age-related accumulation of inosine and elevated Gda and Xdh levels, which results in guanine drop and xanthine accumulation. 8-AG treatment increased its own retinal levels, suggesting the retinal bioavailability of 8-AG (Fig. 11B). 8-AG didn’t reverse the decline of guanine and 3’5’-cGMP in the aged rat retinae (Fig. 11G&K). Although 8-AG slightly reduced PNPase products, hypoxanthine, and xanthine, (Fig. 11I&J), it did not increase PNPase substrates inosine and guanosine (Fig. 11C&E), suggesting the retinal protection by 8-AG include other mechanisms in addition to direct PNPase inhibition within the retina, possibly including peripheral inhibition. Indeed, peripheral inhibition of PNPase, for example in the erythrocytes which are a rich source of PNPase, by 8-AG could be involved in the mechanism of action of 8-AG in the retina.
Long-term efficacy of 8-AG in Fischer 344 rats.
Following the beneficial effects observed after 8 weeks of treatment in 24-month-old rats, we extended the study to assess if 8-AG maintains effectiveness through the rats' lifespan. Specifically, we examined the F344 rats’ retinal structure and function after 17 weeks of treatment starting at 23 months (Fig. 12). Due to high mortality past 24 months, only one untreated (n = 2 eyes) and two treated (n = 4 eyes) rats survived for final analyses. Spectral domain-optic coherence tomography (SD-OCT) revealed maintained retinal structure in 8-AG-treated rats, while untreated ones showed significant degeneration (Fig. 12A&B). Scotopic and photopic ERG responses in 8-AG-treated rats indicated preserved rod and cone functions, contrasting with the near-complete loss of responses in untreated rats (Fig. 12C-E). IHC at 27 months showed abnormal localization of rhodopsin, loss of OS, and zero to one row of ONL nuclei in untreated rats, whereas 8-AG-treated retinae retained rhodopsin localization in the OS and 6–7 rows of ONL nuclei, despite no rescue in peripheral areas (Fig. 12F-H). Collectively, the long-term treatment with 8-AG showed even higher efficacy in preserving the structure and function of the aged retinae.
8-AG confers temporary retinal protection in the RhoP23H/+ knock-in mouse model of retinitis pigmentosa.
We then tested 8-AG’s efficacy in the RhoP23H/+ knock-in mouse model of retinitis pigmentosa (RP), a different retinal degeneration model caused by rhodopsin misfolding (Fig. 13 and S16-S18) (56). Daily intraperitoneal (i.p) injections from PND10 to 28, followed by oral administration of 8-AG in drinking water, until PND38 or PND53. 8-AG treatment led to enhanced scotopic a- and b-waves at PND 36 (P < 0.0001), but only scotopic b-wave improvement persisted at PND 50 (Fig. 13B,C,K,L), suggesting a short-term rod function enhancement by 8-AG. No impact on photopic b-waves was seen, suggesting no effects of 8-AG on cone function in this model, similarly as observed in F344 rats (Fig. 13D&M). Retinal histology showed significantly increased OS + IS thickness at PND 38 and 53 (Figs. 13E,F,N,O, S16 and S17). The ONL cell count was slightly increased by 8-AG at PND 53, but not at PND 38 (Fig. 13E,G,N,P). Interestingly, 8-AG also led to a higher number of neurons in the RGC layer at PND 38 (P < 0.001) and PND53 (not significant, Fig. 13E,H,N,Q). Immunostaining showed that RHO level in the OS + IS increased with 8-AG treatment at PND 38 (Figs. 13I&J and S18, P < 0.0001), but not at PND 53 (Fig. 13R&S and S18). Interestingly, the time point when RHO level was increased by 8-AG can be directly correlated with the time when scotopic ERG responses were increased by the treatment (Fig. 13B&C). Collectively, 8-AG’s retinal protection appears limited to short-term, in the RhoP23H/+ knock-in mouse model of RP.
8-AG treatment reduces activated microglia/macrophages in RhoP23H/+ mice.
We then asked whether 8-AG has any effects on the microglia/macrophages in this animal model (Figs. 14 and S19). Similar to our previous report (57), RhoP23H/+ mouse retinal flat mounts exhibited over 10-fold increase in CD68+ cells and IBA1+ cells compared to normal Rho+/+ mice, with IBA1+ cells outnumbering CD68+ cells (Figs. 12A-H,M,N and S19). As CD68 and IBA1 are markers of activated microglia/macrophages, this result suggests the activated microglia/macrophages increased 10-fold in the RhoP23H/+ mouse retinae. Treatment with 8-AG roughly halved the population of both cell types (Fig. 12E-N), mirroring the anti-inflammatory effects observed in F344 rat retinae and suggesting that 8-AG's anti-inflammatory action may be effective across different models of retinal degeneration.