The AKT2/SIRT5/TFEB pathway as a potential therapeutic target in non-neovascular AMD

Abstract

Non-neovascular or dry age-related macular degeneration (AMD) is a multi-factorial disease with degeneration of the aging retinal-pigmented epithelium (RPE). Lysosomes play a crucial role in RPE health via phagocytosis and autophagy, which are regulated by transcription factor EB/E3 (TFEB/E3). Here, we find that increased AKT2 inhibits PGC-1α to downregulate SIRT5, which we identify as an AKT2 binding partner. Crosstalk between SIRT5 and AKT2 facilitates TFEB-dependent lysosomal function in the RPE. AKT2/SIRT5/TFEB pathway inhibition in the RPE induced lysosome/autophagy signaling abnormalities, disrupted mitochondrial function and induced release of debris contributing to drusen. Accordingly, AKT2 overexpression in the RPE caused a dry AMD-like phenotype in aging Akt2 KI mice, as evident from decline in retinal function. Importantly, we show that induced pluripotent stem cell-derived RPE encoding the major risk variant associated with AMD (complement factor H; CFH Y402H) express increased AKT2, impairing TFEB/TFE3-dependent lysosomal function. Collectively, these findings suggest that targeting the AKT2/SIRT5/TFEB pathway may be an effective therapy to delay the progression of dry AMD.

Fig. 1. Akt2 upregulation triggers lysosomal dysfunction and an AMD-like phenotype in mice.

a Quantification of ZO-1 immunostaining on RPE flatmounts from 10-month-old Akt2 KI mice shows an increased number of dysmorphic RPE cells in the peripheral regions (arrows), not seen in age-matched WT RPE. n (biological replicate)=5. Scale bar = 50 μm. b Immunofluorescence studies revealed a noticeable decrease in EBP50 (magenta) and rhodopsin (red) staining in the RPE, as well as decreased microvilli length (green and graph) in retinal sections, from12-month-old Akt2 KI mice, compared to WT. n (biological replicate) = 4. Datapoints show individual microvili lengths at different regions in the retina sections of 4 different biological replicates. Scale bar = 50 μm. c Transmission electron micrographs showing accumulation of basal laminar deposits in the RPE cells (asterisks) above Bruch’s membrane (BrM) in 15-month-old Akt2 KI mice, but not in age-matched WT RPE cells. M =mitochondria. n = 5. Scale bar 600 nm. d, e Statistically significant decline in expression of CLEAR network genes Ctsd, Ctsb, Lamp1, and Atp6v0a1 in Akt2 KI, but not in Akt2 cKO RPE cells. n (biological replicate)=4. f Western blot analysis showing elevated p-Akt2:Akt2 ratio in iPSC-derived RPE cells from CFH Y402H risk allele [homozygous; CFH (H/H)] containing donors (with no AMD), relative to controls [CFH (Y/Y)] (n; biological replicate = 3), and g reduction of the p-Akt2:Akt2 ratio when CFH I62V mutation is also present in CFH (H/H) cells (n; biological replicate = 6) or h by complete knockout of CFH (CFH null) in iPSC-derived RPE cells. n (biological replicate) = 3. All values are Mean ± S.D. *P < 0.05, **P < 0.01, ****P < 0.0001. The statistical test used in (a, b, d–h) is Student’s t-test (Two-sided). The exact p-values are a P = 0.040 (Akt2 KI vs WT); b P = 2.201E-22 (Akt2 KI vs WT); d Ctsd: P = 0.0257, Ctsb: P = 0.0079, Lamp1: 0.0118, Atp6v0a1: P = 0.0177 (Akt2 KI vs WT); e Ctsd: P = 0.0770, Ctsb: P = 0.8538, Lamp1: 0.3277, Atp6v0a1: P = 0.1918 (Akt2 cKO vs Akt2^fl/fl). f P = 0.000092 (CFH (H/H) vs CFH (Y/Y)); g P = 0.0469 (CFH (H/H) + I62V vs CFH (H/H)); h P = 0.0052 (Cfh null vs Isogenic control). Source Data is provided in the Source Data file.

Fig. 2. Akt2 upregulation in the RPE cells is associated with loss of lysosomal function in AMD.

a Western blot showing elevated p-TFEB (S211) and p-TFE3 (S321) levels along with decreased levels of lysosomal hydrolases CTSD and CTSL in CFH(H/H) cells relative to CFH (Y/Y) cells. n (biological replicate) = 3. b, c Colorimetric analysis revealed significant downregulation of activities of both CTSD and CTSL in iPSC-derived RPE cells from CFH Y402H risk allele [CFH (H/H)] containing donors, relative to controls. n (biological replicate) = 3. d Western blot analysis showing elevated levels of p-Akt2 (S474), p-TFE3 (S321), and p-TFEB (S211) and e downregulation of CTSD and CTSL in RPE lysates from human AMD donors, compared to age-matched controls. n (biological replicate) = 3. f, g RPE lysates from human AMD donors also showed significant downregulation of both CTSD and CTSL activities, compared to controls. n (biological replicate) = 3. All values are Mean ± S.D. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. The statistical test used in (a–g) is Student’s t-test (Two-sided). The exact p-values are a CTSL: P = 0.0422, CTSD: P = 0.00028274, p-TFEB (S211): P = 0.0498, p-TFE3 (S321): P = 0.0142 (CFH (H/H) vs CFH (Y/Y)); b P = 0.0419 (CFH (H/H) vs CFH (Y/Y)); c P = 8.585E-05 (CFH (H/H) vs CFH (Y/Y)); d p-TFE3 (S321): P = 0.0488, p-TFEB (S211): P = 0.0041, p-AKT2 (S474): AKT2: P = 0.0491 (AMD vs Control); e CTSL: P = 0.0175, CTSD: P = 0.0173 (AMD vs Control); f P = 0.042 (AMD vs Control); g P = 7.824E-05 (AMD vs Control). Source Data is provided in the Source Data file.

Fig. 3. AKT2/SIRT5 signaling axis regulates TFEB activity.

a A hypothetical heterodimer of human AKT2/SIRT5 suggests a potential protein–protein interaction between these proteins. The ribbon structures of AKT2 and SIRT5 are colored orange and cyan, respectively. Residues related to the AKT2 and SIRT5 active sites are shown in yellow. The model suggests that both proteins have active sites with potential exposure to substrates and ligands, indicating that the AKT2/SIRT5 complex is catalytically active. b Co-immunoprecipitation study showing that SIRT5 and AKT2 are binding partners in SIRT5-HA-pcDNA and AKT2-GFP-pcDNA overexpressing ARPE19 cells, compared to untransfected controls upon pulldown with anti-IgG or anti-HA antibodies, respectively. n = 3. c Pulldown of AKT2 with anti-AKT2 antibody and subsequent incubation of the pulldown complex with TFEB (recombinant; rec) in vitro to evaluate phosphorylation of TFEB using the phospho-tag gel staining, showed increased p-TFEB/total TFEB in AKT2 overexpressing ARPE19 cells, compared to control. This TFEB phosphorylation was reduced when cells were overexpressing both AKT2 and SIRT5 or the AKT2 pull-down complex was incubated with an AKT2 inhibitor, indicating SIRT5 can modulate AKT2 activity and that AKT2 activity is critical in regulating TFEB phosphorylation. n (biological replicate) = 3. d Western blot showing that simultaneous transfection of AKT2 and SIRT5 pcDNAs could rescue the levels of p-TFEB (S211) in ARPE19 cells, relative to cells overexpressing only AKT2. AKT2 and SIRT5 overexpression in the ARPE19 cells was also confirmed by western blot. n (biological replicate) = 3. Western blot analysis showing reduced expression of SIRT5 and PGC-1α in RPE cells from e Akt2 KI, f PGC-1α KO, and g Sirt5 KO mice, as well as h iPSC-derived RPE cells from CFH (H/H) donors, compared to controls. n = 4 (mice) and n = 3 (iPSC cells). All values are Mean ± S.D. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. The statistical tests used in (c and d) is One-way ANOVA followed by Tukey’s post-hoc test whereas in (e–h) is Student’s t-test (Two-sided). The exact p-values are c P = 0.00002 (AKT2-pcDNA vs IgG), P = 0.0001 (AKT2-pcDNA+SIRT5-HA-pcDNA vs AKT2-pcDNA), P = 0.0462 (AKT2-pcDNA+ AKT2 inhibitor vs AKT2-pcDNA); d P = 0.00062 (AKT2-GFP-pcDNA vs untransfected), P = 0.00051 (AKT2-GFP-pcDNA+SIRT5-HA-pcDNA vs AKT2-pcDNA); e SIRT5: P = 0.0021 and PGC-1α: P = 0.008 (Akt2 KI vs WT); f p-AKT2 (S474): AKT2: P = 0.000832 and SIRT5: P = 0.03 (PGC-1α KO vs WT); g P = 0.0079 (Sirt5 KO vs WT); h PGC-1α: P = 0.0252 and SIRT5: P = 0.0057 (CFH (H/H) vs CFH (Y/Y)). Source Data is provided in the Source Data file.

Fig. 4. Akt2 upregulation inhibits macroautophagy and triggers secretory autophagy in RPE cells.

a Western blot showing elevated levels of FKBP51 and Snap23 in AKT2 overexpressing ARPE19 cells, which were reduced upon simultaneous upregulation of SIRT5 or overexpression of an AKT2 inactive mutant (K14A/R25E). n (biological replicate) = 4. b Immunofluorescence assay showing association of LC3-positive autophagosomes (green) with lysosomes (Lamp1-positive; red) in control ARPE19 cells (arrows in b), which was significantly reduced upon AKT2 overexpression. Scale bar = 50 μm. n (biological replicate) = 8. c RNAseq analysis from RPE cells of 4-month-old WT and Akt2 KI mice showing differential expression of several autophagy-related genes. n = 3. d Western blot showing reduced expression of autophagy mediators ATG9B and ULK1, as well as upregulation of the autophagosome marker p62/SQSTM1 in Akt2 KI RPE cells, relative to WT. n (biological replicate) = 4. e Chromatin immunoprecipitation showing diminished binding of TFEB on Atg9b promoter region in Akt2 KI RPE cells, compared to WT. n (biological replicate) = 4. f Western blot showing reduced autophagy flux (Ratio of LC3-II/ Vinculin in BafA1 treated vs untreated) in Akt2 KI RPE explants compared to WT when treated with Bafilomycin A1 (BafA1; 1 μm) for 4 h. n (biological replicate) = 4. g ARPE19 cells were transfected with AKT2 construct for 48 h or left untreated (control), followed by an overnight infection with an Adenovirus-GFP-RFP-LC3B construct to label the autophagosomes (yellow) and autolysosomes (red). The number of autolysosomes (red puncta) was significantly decreased in AKT2 overexpressing cells (arrows in c) when compared to controls, suggesting a decline in autophagy flux. Scale bar = 50 μm. n (biological replicate) = 4. h Transmission electron micrographs showing double membranous autophagosomes in the RPE cells of 15-month-old Akt2 KI mice, but not in age-matched WT. Scale bar = 600 nm. n = 5. All values are Mean ± S.D. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. The statistical tests used in (a, c, and f) is One-way ANOVA followed by Tukey’s post-hoc test whereas in (b, d, g) is Student’s t-test (Two-sided). The exact p-values are a FKBP51: P = 0.0007 (AKT2-GFP-pcDNA vs untransfected), P = 0.0327 (AKT2-GFP-pcDNA+SIRT-HA-pcDNA vs AKT2-GFP-pcDNA), P = 0.0081 (AKT2(K14A/R25E)-GFP-pcDNA vs AKT2-GFP-pcDNA); b P = 1.598E-10 (AKT2-pcDNA vs Control); d ATG9B: P = 0.0139, p62: P = 0.0013, ULK1: P = 0.017; e P = 0.001 (Akt2 KI vs WT); f P = 9.45202E-05 (Akt2 KI vs WT); g Autophagosome: P = 0.0012, Autolysosome: P = 0.000082 (AKT2-pcDNA vs Control). Source Data is provided in the Source Data file.

Fig. 5. Targeting TFEB independent of mTOR can rejuvenate lysosomal function in CFH Y402H iPSC-derived RPE cells.

a Western blot analysis showing that in iPSC-derived RPE cells from CFH (H/H) risk allele-containing donors, overexpression of mutant TFEB-S467A construct (10⁶ vg/ml for 48 h) or b correction with CFH I62V mutation or c treatment with trehalose (100 mM for 20 h) rescued the levels of the lysosomal hydrolases Cathepsin D, Cathepsin L and the autophagosome mediator ATG9B, compared to untreated CFH (H/H) cells. n (biological replicate) = 3 (TFEB-S467A and trehalose groups), n (biological replicate) = 6 (CFH I62V rescue). d, e Cathepsin D and cathepsin L activities also showed significant rescue upon TFEB-S467A infection, I62V correction, or trehalose treatment to CFH (H/H) cells. n (biological replicate) = 3. f Trehalose (100 mM for 48 h) or AKT2 inhibitor (10 nM for 48 h) treatment to iPSC-derived RPE cells from CFH (H/H) donors incubated with CCHS (5% for 48 h) or g, h iPSC-derived RPE cells from human AMD donors with CFH (H/H) genotype, rescued the levels of Lamp1 and cathepsin D, compared to untreated CFH (H/H)+CCHS or CFH (H/H; from AMD donors) cells. n (biological replicate) = 3. All values are Mean ± S.D. *P < 0.05, ****P < 0.0001, ***P < 0.008, **P < 0.01, *P < 0.05. The statistical tests used in (a, c, d–h) is One-way ANOVA followed by Tukey’s post-hoc test, and in (b) is Student’s t-test (Two-sided). The exact p-values are a ATG9B: P = 0.0167, CTSD: P = 0.0234, CTSL: P = 0.0006 (CFH (H/H)+WT TFEB vs CFH (Y/Y)), ATG9B: P = 0.0085, CTSD: P = 0.0235, CTSL: 0.0012 (CFH (H/H)+TFEB-S467A vs CFH (H/H)+WT TFEB); b CTSD: P = 0.0067, CTSL: P = 0.007 (CFH (H/H)+I62V vs CFH (H/H)); c ATG9B: P = 0.00002, CTSD: P = 0.00001 (CFH (H/H) vs CFH (Y/Y)) and ATG9B: P = 0.0020, CTSD: 0.00003 (CFH (H/H)+Tre vs CFH (H/H)). d P = 0.00001 (CFH (H/H) vs CFH (Y/Y)), P = 0.00003 (CFH (H/H)+AAV2-WT TFEB vs CFH (Y/Y)), P = 0.00225 (CFH (H/H)+AAV2-TFEB-S467A vs CFH (H/H)+AAV2-WT TFEB), P = 0.00037 (CFH (H/H)+I62V vs CFH (H/H)), P = 0.00191 (CFH (H/H)+Tre vs CFH (H/H)); e P = 0.00001 (CFH (H/H) vs CFH (Y/Y)), P = 0.000001 (CFH (H/H)+AAV2-WT TFEB vs CFH (Y/Y)), P = 0.00002 (CFH (H/H)+AAV2-TFEB-S467A vs CFH (H/H)+AAV2-WT TFEB), P = 0.00008 (CFH (H/H)+I62V vs CFH (H/H)), P = 0.00001 (CFH (H/H)+Tre vs CFH (H/H)). f CTSD: P = 0.0401 (CFH (Y/Y)+CCHS vs CFH (Y/Y)+CIHS), P = 0.0478 (CFH (H/H)+CCHS vs CFH (Y/Y)+CIHS), P = 0.0011 (CFH (H/H)+CCHS+Tre vs CFH (H/H)+CCHS), LAMP1: P = 0.049 (CFH (Y/Y)+CCHS vs CFH (Y/Y)+CIHS), P = 0.0003 (CFH (H/H)+CCHS+Tre vs CFH (H/H)+CCHS). g CTSD: P = 0.0031, Lamp1: P = 0.0452 (CFH (H/H)+Tre vs CFH (H/H)); h CTSD: P = 0.0263, Lamp1: P = 0.0491 (CFH (H/H)+Akt2 inhibitor vs CFH (H/H)). Source Data is provided in the Source Data file.

Fig. 6. Trehalose rejuvenates lysosomal function and rescues autophagy in mouse models in vivo.

RPE cells from 6-month-old Cryba1 cKO mice treated with trehalose for 3 consecutive months show significant rescue in the levels of a CLEAR network genes like Ctsd, Lamp1 and Atp6v0a1, b cathepsin D (CTSD) activity and c protein levels of the autophagosome marker p62/SQSTM1, relative to RPE cells from vehicle (water) treated Cryba1 cKO mice. n (biological replicate) = 4. RPE cells from 6-month-old Akt2 KI mice treated with trehalose for 3 consecutive months showed statistically significant rescue of d protein levels of autophagosome marker p62/SQSTM1 and lysosomal hydrolase, CTSD as well as e CLEAR network gene (Ctsd, Ctsb, Lamp1, Atp6v0a1) expression, compared to vehicle-treated Akt2 KI mice. n (biological replicate) = 4. Trehalose treatment to Akt2 KI mice showed noticeable rescue in f CTSD and g CTSL activities relative to untreated Akt2 KI mice. n (biological replicate) = 4. Hematoxylin-eosin stained sections and quantification showing a decrease in ONL thickness in h Akt2 KI and i Cryba1 cKO mice, which was rescued upon trehalose treatment. Prevalence of a patchy RPE layer with the accumulation of drusen-like deposits (asterisks in i and inset) in Cryba1 cKO retina (vehicle control), with rescue by trehalose treatment. Scale bar = 100 μm (inset = 20 μm). n (biological replicate) = 5. j Created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en. Cartoon showing that AKT2 upregulation in RPE cells from Akt2 KI mice and CFH (H/H) iPSC-derived cells trigger deregulation of SIRT5/PGC-1α dependent lysosomal function. Trehalose and S467A TFEB mutant construct rescued the lysosomal abnormalities and delayed the progression of early RPE changes in a mouse model. All values are Mean ± S.D. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. The statistical test used in (a–i) is One-way ANOVA followed by Tukey’s post-hoc test. The exact p-values are a Ctsd: P = 0.0042, Lamp1: P = 0.0010, Atp6v0a1: P = 0.0062 (Cryba1 cKO vs Cryba1^fl/fl), Ctsd: P = 0.0129, Lamp1: P = 0.0055, Atp6v0a1: P = 0.0489 (Cryba1 cKO+Trehalose vs Cryba1 cKO); b P = 0.00006 (Cryba1 cKO vs Cryba1^fl/fl), P = 0.0013 (Cryba1 cKO+Trehalose vs Cryba1 cKO); c P = 0.0013 (Cryba1 cKO vs Cryba1^fl/fl), P = 0.0258 (Cryba1 cKO+Trehalose vs Cryba1 cKO); d p62: P = 0.0037 and CTSD: P = 0.0461 (Akt2 KI vs WT), p62: P = 0.0072 and CTSD: P = 0.0092 (Akt2 KI+Trehalose vs Akt2 KI); e Ctsd: P = 0.00001, Ctsb: P = 0.00002, Lamp1: P = 0.000035, Atp6v0a1: P = 0.000029 (Akt2 KI vs WT), Ctsd: P = 0.0020, Ctsb: P = 0.0005, Lamp1: P = 0.0017, Atp6v0a1: P = 0.0056 (Akt2 KI+Trehalose vs Akt2 KI); f P = 0.00002 (Akt2 KI vs WT), P = 0.000042 (Akt2 KI+Trehalose vs Akt2 KI); g P = 0.000049 (Akt2 KI vs WT), P = 0.00006 (Akt2 KI+Trehalose vs Akt2 KI). h P = 0.00001 (Akt2 KI vs WT), P = 0.0001 (Akt2 KI+Trehalose vs Akt2 KI); i ONL thickness: P = 0.000027 (Cryba1 cKO vs Cryba1^fl/fl), P = 0.000019 (Cryba1 cKO +Trehalose vs Cryba1 cKO), drusen-like deposits: P = 0.0005 (Cryba1 cKO+Trehalose vs Cryba1 cKO; Student’s t-test). Source Data is provided in the Source Data file.