ATF4 leads to glaucoma by promoting protein synthesis and ER client protein load

Abstract

The underlying pathological mechanisms of glaucomatous trabecular meshwork (TM) damage and elevation of intraocular pressure (IOP) are poorly understood. Here, we report that the chronic endoplasmic reticulum (ER) stress-induced ATF4-CHOP-GADD34 pathway is activated in TM of human and mouse glaucoma. Expression of ATF4 in TM promotes aberrant protein synthesis and ER client protein load, leading to TM dysfunction and cell death. These events lead to IOP elevation and glaucomatous neurodegeneration. ATF4 interacts with CHOP and this interaction is essential for IOP elevation. Notably, genetic depletion or pharmacological inhibition of ATF4-CHOP-GADD34 pathway prevents TM cell death and rescues mouse models of glaucoma by reducing protein synthesis and ER client protein load in TM cells. Importantly, glaucomatous TM cells exhibit significantly increased protein synthesis along with induction of ATF4-CHOP-GADD34 pathway. These studies indicate a pathological role of ATF4-CHOP-GADD34 pathway in glaucoma and provide a possible treatment for glaucoma by targeting this pathway.

Fig. 1. ATF4–CHOP–GADD34 is induced in TM cells/tissues of human and mouse glaucoma.

a and b Cellular lysates from age-matched normal and glaucomatous primary human TM cells were examined for ER stress markers using Western blotting a and analyzed by densitometric analysis b. The average from two independent experiments is shown graphically (n = 3 cell strains for ATF4 and CHOP; n = 6 cell strains for GADD34; n = 4 cell strains for XBP-1; data are presented as mean ± SEM, 2-tailed unpaired t-test). c and d Representative immunostaining for ATF4 and GADD34 c and its intensity measurements d in age-matched normal and glaucoma donor eyes shows significantly increased ATF4 and GADD34 in glaucomatous TM tissues. Arrow shows TM. SC Schlemm’s canal. Scale bar is 100 µm (n = 9 normal and 11 glaucoma eyes, data are presented as mean ± SEM, two-tailed unpaired t-test). *Indicates fold in section (artifact). e Periocular Dex injections elevate IOP significantly in C57BL/6J mice. 3 months old C57BL/6J mice received vehicle or Dex (200 µg/eye) via periocular conjunctival fornix injections bilaterally every week up to 5 weeks and night IOPs were monitored weekly using rebound tonometry (n = 8 in each group; data are presented as mean ± SEM, two-way ANOVA, ***P < 0.001). f and g Western blot for ER stress markers f and its densitometric analyses g of anterior segment tissues from vehicle and Dex-treated mice demonstrated that induction of chronic ER stress markers is associated with Dex-induced ocular hypertension (n = 3 in each group, data are presented as mean ± SEM, two-tailed unpaired t-test).

Fig. 2. ATF4 induces ocular hypertension and reduces aqueous humor outflow facility in mice.

a 3 months old C57BL/6J mice received bilateral intravitreal injections of Ad5.Empty or Ad5.ATF4 (2 × 10⁷ pfu/eye). IOPs were monitored using rebound tonometry on conscious mice. ATF4-injected mice showed significant IOP elevation starting from 3 weeks injections compared to Empty-injected mice (n = 10 each group, data are presented as mean ± SEM, two-way ANOVA, Bonferroni’s multiple comparisons test *P < 0.05, ***P < 0.001). b Conventional outflow facility was significantly reduced in Ad5.ATF4-injected mice (0.022 µL/min/mmHg; n = 6) compared to Ad5.Empty injected mice (0.042 µL/min/mmHg; n = 6) (data are presented as mean ± SEM, two-tailed unpaired t-test).

Fig. 3. Lentiviral expression of ATF4 but not CHOP induces glaucoma in mice.

a Night-time IOP measurements in mice injected intravitreally with lentiviral particles expressing GFP (control), ATF4, and CHOP (2 × 10⁶ TU/eye). A significant IOP elevation was observed in ATF4-injected mice compared to GFP or CHOP-injected mice (n = 7 in GFP group, n = 8 in ATF4 and CHOP groups; data are presented as mean ± SEM, two-way ANOVA, Tukey’s multiple comparison test, ****P < 0.0001). b RGC functional loss in ATF4-injected mice. A decrease in pattern ERG amplitudes was observed in 8 weeks ATF4-injected mice (n = 7) compared to GFP-injected mice (n = 5; data are presented as mean ± SEM, two-tailed unpaired t-test. c Loss of RGCs in 10 weeks ATF4-injected mice. Representative images showing peripheral, mid-peripheral, and central regions of whole mount retinas stained with RBPMS (RGC marker) in GFP (n = 7) and ATF4 (n = 8) injected groups. d RGC counts per 0.393 mm² of total retina in peripheral, mid-peripheral, and central regions of ATF4 and GFP-injected mice demonstrated significant loss of RGCs in periphery and mid-periphery of ATF4-injected retinas. Scale bar is 50 µm. Data are presented as mean ± SEM, one-way ANOVA; n = 7 for GFP and n = 8 for ATF4. e Representative images of optic nerve sections stained with PPD showing loss of axons in 10 weeks ATF4-injected mice. Scale bar is 10 µm (n = 5 in GFP and n = 5 in ATF4 injected mice). f PPD stained optic nerve axons were counted and total number of axons per optic nerve were shown graphically in GFP (n = 5) vs. ATF4 (n = 5) injected mice groups (data are presented as mean ± SEM, two-tailed unpaired t-test).

Fig. 4. CHOP is required for ATF4-induced IOP elevation.

a GTM3 cells were transduced with Ad5.Empty, ATF4, and CHOP for 24 h. Cellular lysates were subjected to Western blot analysis of ATF4 and CHOP. Increased CHOP expression levels were observed in ATF4-transduced TM cells whereas CHOP transduction did not alter ATF4 levels in TM cells (n = 3 replicates). b GTM3 cells transduced with Ad5.Empty, ATF4, and CHOP for 24 h. Cellular lysates were immunoprecipitated using CHOP antibody and immunoblotted for ATF4. IgG antibody was used as a negative control (n = 3 replicates). Experiments were replicated three times independently and similar results were observed in NTM-5 cell line. c 3-month-old WT (CHOP^+/+) and CHOP knockout (CHOP^−/−) mice were intravitreally injected with Ad5.Empty or ATF4. A significant IOP elevation was observed in CHOP^+/+ mice injected with Ad5.ATF4 (n = 8) compared to CHOP^+/+ mice injected with Ad5.Empty (n = 6). However, ATF4 did not significantly elevate IOP in CHOP^−/− mice (n = 8) compared to Ad5. Empty injected WT mice (data are presented as mean ± SEM, two-way ANOVA with Bonferroni’s multiple comparison test, *P < 0.05, ***P < 0.001).

Fig. 5. ATF4 increases protein synthesis and ER client protein load in TM cells and tissues.

a GTM3 cells were transduced with Ad5.Empty, ATF4, and CHOP for 36 h. and cellular lysates were examined for markers of chronic ER stress by Western blot analysis. Expression of ATF4 increased XBP-1 (s) and GADD34 while ATF4 reduced p-eIF2α (n = 3 replicates). b GTM3 cells transfected with ATF4 or CHOP were incubated with cycloheximide (CHX) (10 μg/ml) for 16 h. Puromycin (10 µg/ml) was added to cells for 30 min before harvesting cell lysates. Total cellular lysates were subjected to Western blot analysis using anti-puromycin and GAPDH antibodies. Increased puromycin incorporation observed in the total ER fractions of ATF4 transfected cells (n = 3 replicates), which was blocked by CHX signifying a higher rate of de novo protein synthesis. c and d GTM3 cells stably expressing DsRed-tagged WT c or mutant d MYOC were transfected with plasmids expressing Empty and ATF4 for 48 h. Cells were fixed and stained with calreticulin (ER marker). ATF4 increased intracellular WT and mutant myocilin accumulation in the ER as evident from a strong colocalization of myocilin with calreticulin in ATF4-transfected cells (n = 3 replicates). Scale bar is 25 µm. e 3-month-old C57BL/6J mice were intravitreally injected with Ad5.Empty, Ad5.ATF4, and Ad5.CHOP (2 × 10⁷ pfu/eye). After 3 weeks of injections, anterior segment lysates were subjected to Western blot (Supplementary Fig. 15) and densitometric analyses e of ECM (fibronectin, collagen-1) and ER stress (Grp78, ATF4, and CHOP) markers (n = 4 mice; data are presented as mean ± SEM, one-way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). f Increased ATF4 and KDEL staining in TM of Ad5.ATF4-injected mice compared to Ad5.Empty-injected mice (n = 3 mice). Scale bar is 100 µm.

Fig. 6. ATF4 leads to TM cell death in a CHOP and GADD34-dependent manner.

GTM3 cells were transduced with Ad5.Empty, ATF4, or CHOP for 36 h. a Cellular lysates were subjected to Western blot analysis for apoptotic markers (n = 3 independent experiments). b Fixed GTM3 cells transduced with Ad5.Empty, ATF4, or CHOP were analyzed by TUNEL assay. Increased number of TUNEL-positive cells (green) were observed in Ad5.ATF4-transduced TM cells (n = 3 independent experiments). Scale bar is 50 µm. c Increased number of TUNEL-positive cells (green) were observed in the TM region of 2 weeks Ad5.ATF4-injected mice compared to Ad5.Empty or Ad5.CHOP (n = 3 in each groups). Bright field images (bottom panel) merged with DAPI shows TM orientation and the TM area is represented in a white box (scale bar = 50 µm). d and e GTM3 cells were transduced with Ad5.Empty, ATF4, CHOP, or ATF4 + CHOP KO (CRISPR-Cas9 vector targeting CHOP) for 36 h. Western blot d and densitometric analyses e demonstrated that depletion of CHOP significantly reduced ATF4-induced cleaved PARP suggesting that CHOP is required for ATF4-mediated cell death. n = 3 independent experiments, data are presented as mean ± SEM, *p < 0.05, one-way-ANOVA. f and g Western blot f and densitometric analyses g of cleaved PARP in GTM3 cells transduced with ATF4 and treated with mild dose of CHX (2.5 μg/ml) demonstrated that reduction of protein synthesis prevents ATF4-induced TM cell death. n = 3 independent experiments, data are presented as mean ± SEM, one-way-ANOVA. h GTM3 cells expressing ATF4 were transfected with plasmid expressing CRISPR-Cas9 targeting GADD34. Puromycin (10 µg/ml) was added to cells for 30 min before harvesting cell lysates. Total cellular lysates were subjected to Western blot analysis using anti-puromycin and anti-GAPDH antibodies (shown in Supplementary Fig. 17). Densitometric analysis demonstrated that GADD34 knock down significantly reduces ATF4-induced protein synthesis. n = 4 independent experiments, data are presented as mean ± SEM, two-tailed unpaired t-test. i and j Western blot i and densitometric analyses j of cleaved PARP and GADD34 of GTM3 cells expressing ATF4 and plasmid expressing CRISPR-Cas9 targeting GADD34 demonstrated that knockdown of GADD34 significantly reduced ATF4-induced cleaved PARP and GADD34 protein levels. n = 3 independent experiments, data are presented as mean ± SEM, **p < 0.01, two-way-ANOVA.

Fig. 7. Pharmacological inhibition of ATF4 rescues mouse models of glaucoma.

a C57 mice were injected with vehicle (n = 4) and Dex (n = 4) weekly via periocular conjunctival fornix injections. After 2 weeks, left eyes received 5 µl of 2 mM ISRIB topical eye drops while the contralateral right eyes received vehicle eye drops (DMSO) twice daily. One-week ISRIB treatment significantly lowered IOPs in Dex-injected left eyes compared to the contralateral right eyes (n = 4 biologically independent samples; data are presented as mean ± SEM, one-way ANOVA, Tukey’s multiple comparison test, **P < 0.01). b Western blot (Supplementary Fig. 21) and densitometric analysis b of mouse anterior segment lysates injected with vehicle or Dex for 2 weeks and treated with 5 µl of 2 mM ISRIB topical eye drops or DMSO control. ISRIB treatment prevented Dex-induced FN and ER stress markers (n = 3 biologically independent samples; data are presented as mean ± SEM, one-way ANOVA, **P < 0.01, ***P < 0.001, ****P < 0.0001). c The ocular hypertensive Tg.MYOC^Y437H mice received ISRIB eye drops in left eyes whereas the contralateral right eyes received vehicle (DMSO) eye drops twice daily. IOPs were recorded after one-week treatment (n = 5 biologically independent samples; data are presented as mean ± SEM, two-tailed paired t-test). d C57 mice were treated with Dex for 3 weeks via periocular route. IOPs were measured after 3 weeks. After ocular hypertension was confirmed, Ad5.ATF4ΔRK (2 × 10⁷ pfu/eye) was injected intravitreally in the left eyes while the contralateral right eyes were injected with Ad5.Empty. Both eyes were treated with Dex for another week and IOPs were measured every week. Ad5.ATF4ΔRK reduced elevated IOP compared to contralateral eyes injected with Ad5. empty (n = 4 biologically independent samples; data are presented as mean ± SEM, two-tailed paired t-test). e–h Western blot e and its densitometric analyses f–h of mouse anterior segment lysates show that ATF4ΔRK prevented Dex-induced fibronectin and CHOP. n = 3 biologically independent samples; data are presented as mean ± SEM, two-tailed unpaired t-test.

Fig. 8. Increased protein synthesis is associated glaucomatous primary human TM cells.

a and b Protein synthesis was examined in human primary TM cells from normal (n = 4 biologically independent cells) and POAG donor eyes (n = 4 biologically independent cells) using SUnSET assay. Western blot a and its densitometric analysis b demonstrated significantly increased protein synthesis in glaucomatous TM cells. n = 4 biologically independent samples, data are presented as mean ± SEM, two-tailed unpaired t-test. c and d GTM3 cells were treated with vehicle and Dex along with DMSO (control) or ISRIB for 2 h. Protein synthesis was examined using the SUnSET assay. Western blot c and its densitometric analysis d demonstrated ISRIB significantly reduced Dex-induced protein synthesis. n = 3 independent experiments, data are presented as mean ± SEM, **p < 0.01; *p < 0.05; 1-way ANOVA. e Western blot analysis demonstrated that ISRIB reduced Dex-induced GADD34 levels while ISRIB effects on p-eIF2α were minimal. n = 3 independent experiments. f Western blot analysis of total ER fractions with puromycin antibody isolated from GTM3 cells treated with vehicle, Dex, and Dex plus ATF4ΔRK for 36 h. Cells were incubated with puromycin antibody for 30 min before harvesting lysates (n = 3 independent experiments). Genetic inhibition of ATF4 prevented Dex-induced ER client protein load in GTM3 cells. g Western blot analysis of GTM3 cells expressing mutant myocilin treated with ISRIB showing that ISRIB reduces intracellular accumulation of mutant myocilin along with ATF4 and CHOP protein levels. n = 3 independent experiments.