Fig 2: HDAC6-mediated deacetylation enhances PRMT1 stability.(A) Immunoprecipitation and immunoblotting to examine PRMT1 acetylation in cells transfected with HA vector, HA-HDAC6 or its inactive mutants, in the presence of GFP-PRMT1. (B, C) The acetylation of PRMT1 was analyzed by immunoprecipitation with GFP antibody followed by immunoblotting in cells transfected with GFP-PRMT1 and treated with or without tubastatin A (TubA, B), and in cells transfected with control or HDAC6 siRNAs in presence of GFP-PRMT1 (C). (D) Analysis of endogenous PRMT1 acetylation in cells transfected with HA vector or HA-HDAC6. (E) Analysis of endogenous PRMT1 acetylation in retinas from wild-type or Hdac6 knockout mice. (F, G) Analysis of endogenous PRMT1 acetylation in retinas from control and diabetic mice, and the relative PRMT1 acetylation level was determined by densitometry. (H) Schematic diagram of PRMT1 wild-type and lysine-to-arginine (KR) mutants. (I, J) Identification of the site of PRMT1 deacetylation by HDAC6 by immunoprecipitation with the anti-GFP antibody followed by immunoblotting with the acetylated-lysine (Ace-lysine) antibody in cells transfected with GFP-PRMT1 wild-type or its KR mutants in the presence of HA-HDAC6 or HA vector. (K, L) Analysis of the stability of PRMT1 wild-type or its K128R mutant. (M–P) The protein stability of PRMT1 was examined in cells transfected with control or HDAC6 siRNAs (M), and in cells transfected with HA vector, HA-HDAC6 wild-type or its inactive mutants (O), and the relative PRMT1 level was determined by densitometry (N, P). Data are presented as mean ± SEM. ****p < 0.0001. The data underlying this figure are available in S1 Data and S1 Raw Images.

Fig 3: PRMT1-mediated asymmetric arginine dimethylation promotes HDAC6 stability.(A) The asymmetric demethylation of HDAC6 was analyzed by immunoprecipitation with the anti-HA antibody followed by immunoblotting in cells transfected with GFP-PRMT1, GFP-PRMT1 mutant (VLD-AAA, catalytically inactive mutant) or GFP vector, in the presence of HA-HDAC6. (B, C) Immunoprecipitation and immunoblotting analysis of asymmetric dimethylation of HDAC6 in cells transfected with control or PRMT1 siRNAs in the presence of HA-HDAC6 (B), and in cells transfected with HA-HDAC6 with or without C-7280948 (C). (D) In vitro HDAC6 dimethylation assay using purified GST-HDAC6 and various concentrations of Flag-PRMT1 as indicated. (E) Immunoprecipitation and immunoblotting analysis asymmetric dimethylation of endogenous HDAC6 in mouse retinas. (F, G) Analysis of the asymmetric dimethylation of endogenous HDAC6 in retinas from wild-type and Prmt1 conditional knockout mice (F), or from control and diabetic mice (G). (H) Identification of the site of HDAC6 asymmetric dimethylation by immunoprecipitation with the anti-HA antibody followed by immunoblotting with the anti-ADMA antibody in cells transfected with GFP-PRMT1 or GFP vector in the presence of HDAC6 wild-type or arginine-to-lysine mutants. (I, J) The protein stability of wild-type or R16K mutant HDAC6 was examined by immunoblotting in cells transfected with HA-HDAC6 wild-type or the R16K mutant and treated with 20 μg/mL cycloheximide (CHX) for the indicated time. (K) Quantification of HDAC6 stability in cells transfected with GFP vector, GFP-PRMT1 or its mutant. The data underlying this figure are available in S1 Data and S1 Raw Images.

Fig 4: Inhibiting the PRMT1-HDAC6 feedback loop ameliorates pathological alterations in DR and DN.(A) Schematic illustration of intravitreal or intraperitoneal injection of C-7280948 and tubastatin A to inhibit PRMT1 an HDAC6 in diabetic mice, respectively. (B–D) ERG recordings (B) and measurement of retinal a-wave (C) and b-wave (D) amplitudes in diabetic mice following intravitreal injection with C-7280948, tubastatin A, or a combination of both inhibitors, vs. vehicle control (n = 20 mice from three independent experiments). (E, F) Photomicrographs (E) and quantification (F) of the retinal histology assessed by H&E staining in diabetic mice intravitreally injected with C-7280948, tubastatin A, or a combination of both inhibitors, vs. vehicle control (n = 3 independent experiments using a total of 24 mice). Scale bar, 30 μm. (G–I) Immunofluorescence images (G) and quantifications of the thickness of outer segment membranous discs of rods (H) and cones (I) from diabetic mice intravitreally injected with C-7280948, tubastatin A, or a combination of both inhibitors, vs. vehicle control (n = 50 fields from three independent experiments). Scale bar, 10 μm. (J–L) Immunofluorescence images (J) and quantifications of the length (K, n = 100 fields from three independent experiments) and density (L, n = 3 independent experiments) of ciliary axonemes in retinas from diabetic mice intravitreally injected with C-7280948, tubastatin A, or a combination of both inhibitors, vs. vehicle control. Scale bar, 2 μm. (M–O) Immunofluorescence images (M) and quantification of ciliary length (N) and density (O) in LTL+ proximal tubules from diabetic mice intraperitoneally injected with C-7280948, tubastatin A, or a combination of both inhibitors, vs. vehicle control (n = 50 renal tubules from three independent experiments). Scale bar, 6 μm. Data are presented as mean ± SEM. ****p < 0.0001. The data underlying this figure are available in S1 Data.

Fig 5: PRMT1 and HDAC6 overexpression exacerbates the progression of DR and DN.(A) Differentially expressed proteins in retinas and kidneys from control and diabetic mice. (B) Heat maps of representative cilium-associated proteins significantly altered in both retinas and kidneys from control and diabetic mice, filter criteria used DESeq P-values < 0.05, fold-change > 1.5. (C, D) Immunoblot analysis of HDAC6, PRMT1, and β-actin in the retina and kidney from control and diabetic mice. (E–G) Immunofluorescence images (E) and quantifications of the length (F, n = 100 fields from three independent experiments) and density (G, n = 3 independent experiments) of ciliary axonemes in retinas from Hdac6 knockout or Prmt1 conditional knockout mice and their respective controls under diabetic conditions. Scale bar, 2 μm. (H–J) Immunofluorescence images (H) and quantifications of the thickness of outer segment membranous disc in rods (I) and cones (J) from Hdac6 knockout or Prmt1 conditional knockout mice and their respective controls under diabetic conditions (n = 50 fields from three independent experiments). Scale bar, 10 μm. (K) Photomicrographs of the retinal histology assessed by H&E staining in Hdac6 knockout or Prmt1 conditional knockout mice and their respective controls under diabetic conditions. Scale bar, 30 μm. (L–N) ERG recordings (L) and measurements of retinal a-wave (M) and b-wave (N) amplitudes were performed for the Hdac6 knockout or Prmt1 conditional knockout mice and their respective controls under diabetic conditions (n = 24 mice from three independent experiments). (O–Q) Immunofluorescence images (O) and quantifications of ciliary length (P) and density in LTL+ proximal tubules from Hdac6 knockout or Prmt1 conditional knockout mice and their respective controls under diabetic conditions (n = 50 renal tubules from three independent experiments). Scale bar, 6 μm. Data are presented as mean ± SEM. **p < 0.01, ****p < 0.0001. The data underlying this figure are available in S1 Data and S1 Raw Images.