Phenylbutyrate modulates polyamine acetylase and ameliorates Snyder-Robinson syndrome in a Drosophila model and patient cells

Abstract

Polyamine dysregulation plays key roles in a broad range of human diseases from cancer to neurodegeneration. Snyder-Robinson syndrome (SRS) is the first known genetic disorder of the polyamine pathway, caused by X-linked recessive loss-of-function mutations in spermine synthase. In the Drosophila SRS model, altered spermidine/spermine balance has been associated with increased generation of ROS and aldehydes, consistent with elevated spermidine catabolism. These toxic byproducts cause mitochondrial and lysosomal dysfunction, which are also observed in cells from SRS patients. No efficient therapy is available. We explored the biochemical mechanism and discovered acetyl-CoA reduction and altered protein acetylation as potentially novel pathomechanisms of SRS. We repurposed the FDA-approved drug phenylbutyrate (PBA) to treat SRS using an in vivo Drosophila model and patient fibroblast cell models. PBA treatment significantly restored the function of mitochondria and autolysosomes and extended life span in vivo in the Drosophila SRS model. Treating fibroblasts of patients with SRS with PBA ameliorated autolysosome dysfunction. We further explored the mechanism of drug action and found that PBA downregulates the first and rate-limiting spermidine catabolic enzyme spermidine/spermine N1-acetyltransferase 1 (SAT1), reduces the production of toxic metabolites, and inhibits the reduction of the substrate acetyl-CoA. Taken together, we revealed PBA as a potential modulator of SAT1 and acetyl-CoA levels and propose PBA as a therapy for SRS and potentially other polyamine dysregulation-related diseases.

Keywords: Genetic diseases; Genetics; Lysosomes; Polyamines; Therapeutics.

Figure 1. PBA attenuates SAT1.

(A) Diagram of the hypothesis of PBA-derived phenylacetyl-CoA competing with acetyl-CoA to interfere with SAT1-mediated polyamine catabolism and acetyl-CoA consumption. SPDSY, spermidine synthase; PAO: peroxisomal N¹-acetyl-spermine/spermidine oxidase; SMOX, spermine oxidase. (B) Acetyl-CoA level in HEK293T cells with SAT1 overexpression with or without 2 mM PBA treatment. Cells from 4 separate experiments were stored at –80°C and then tested in a single plate. n = 4; 1-way ANOVA multiple comparisons (matched). (C) Dot blot of acetyl-lysine in HEK293T cells with SAT1 overexpression with or without 2 mM PBA treatment. The values of the quantification were normalized with the control samples. (D) Western blot of the samples in C. Acetylated histones and nonhistone proteins in the bracket areas were quantified separately and normalized with β-Actin level. All values were further normalized by the control samples. (E) Western blot of HA-tagged SAT1 overexpressed in HEK293T cells with PBA, MG132, or the combination treatment. The HA-SAT1 level was normalized with the β-Actin level. The values of the cells without HA-SAT1 plasmid transfection were set as background. All the values were further normalized by that of the cells with indicated treatment. (F) Western blot of SAT1 induced by DENSPM in patient fibroblasts (CMS1849a) with indicated treatment. The SAT1 level was normalized with the β-Actin level. Values of the cells without DENSPM treatment were set as background. All values were further normalized by that of cells with DENSPM and without PBA or MG132 treatment. Images in C–F are representative of 3 separate experiments. n =3; *P < 0.05, **P < 0.01, ***P < 0.001; 1-way ANOVA multiple comparisons (matched). Data represent mean ± SEM.

Figure 2. PBA treatment extends life span and reduces ROS and aldehydes in a Drosophila SRS model.

(A) Life span of female SRS flies fed with indicated concentration of PBA. n = 76 at 0 mM, 77 at 2 mM, 53 at 5 mM, and 53 at 10 mM; log-rank (Mantel-Cox) test, Bonferroni-corrected α = 0.0167. (B) Life span of female SRS flies fed with lower concentration of PBA. n = 60, 60, and 71; log-rank (Mantel-Cox) test, Bonferroni-corrected α = 0.025. (C) ROS staining of brains of 10 days after eclosion (DAE) flies with or without PBA feed. Scale bar: 100 μm. The image is a representative of multiple brains in each group. (D) Quantification of the brain size in C. n = 4, 4, and 3. (E) Quantification of the relative ROS level in C. The ROS signal was normalized with the brain size. All the values were further normalized by that of the first WT brain. n = 4, 4, and 3. (F) Aldehyde level measurement of 10 DAE flies with or without PBA feed. Each dot indicates a sample of homogenized mixture of 10 flies. n = 6, 6, and 6; *P < 0.05, **P < 0.01, ***P < 0.001; ordinary 1-way ANOVA multiple comparisons in D–F. Data represent mean ± SEM.

Figure 3. PBA treatment partially restores mitochondria in a Drosophila SRS model.

(A) Mitochondrial membrane protein ATP5α and contractile filament component F-actin staining of flight muscle of 10 DAE flies with or without PBA feed. The third row shows recognized mitochondria by automatic segregation of ATP5α staining signal using ImageJ H-watershed (NIH). The images are representatives of 5 flies in each group. Scale bar: 10 μm. The diagram of muscle fiber structure on the top right is adapted from (97). (B) Quantification of mitochondrial size indicated by segregated ATP5α signal. Each mitochondrion is marked as a dot. Multiple dots of mitochondria with the same size aggregate into a black line. The blue dash lines indicate the average values. The red bars indicate SD. (C) Quantification of mitochondrial shape indicated by circularity of segregated ATP5α signal. Each mitochondrion is marked as a dot. Multiple dots of mitochondria with the same circularity aggregate into a black line. The blue dash lines indicate the average values. The red bars indicate SD. (D) COX activity staining of flight muscle of 10 DAE flies with or without PBA feed. The image is a representative of 10 samples in each group. Scale bar: 50 μm. (E) Quantification of COX activity in D. n = 10; **P < 0.01, ***P < 0.001; ordinary 1-way ANOVA multiple comparisons in B, C, and E. Data represent mean ± SEM.

Figure 4. PBA treatment improves lysosomal function in a Drosophila SRS model.

(A) Lysosomal membrane protein LAMP1 and CtsL staining of lamina of 10 DAE flies with or without PBA feed. The diagram of lamina structure is on the top right. Scale bar: 10 μm. (B) Quantification of LAMP1 or (C) CtsL intensity in A. n = 6, 6, and 8; ordinary 1-way ANOVA multiple comparisons. (D) Western blot of autophagy cargo recruiter Ref(2)p in heads of 10 DAE flies with or without PBA feed. The image is a representative of 4 separate experiments. (E) Quantification of the Ref(2)p level in D. The Ref(2)p level was normalized with the β-Actin level. All the values were further normalized by that of WT flies. n = 4; *P < 0.05, **P < 0.01, ***P < 0.001; 1-way ANOVA multiple comparisons (matched). Data represent mean ± SEM.

Figure 5. PBA treatment improves lysosomal function in fibroblasts of patients with SRS.

(A) BMV109(Cy5) labeling of active lysosomal Cts in patient fibroblasts (Ctr, GM09503; M35R, CMS25081a; and I150T, CMS1849b) with indicated SMS mutation with or without PBA treatment. The image is a representative of 3 separate experiments. (B) Quantification of the BMV109(Cy5) intensity of the strongest band in the bracket area, normalized with Coomassie staining signal in the bracket area in A. All the values were further normalized by that of the control cells. (C) Western blot of CtsD and autophagy cargo recruiter p62/SQSTM1 in patient fibroblasts (Ctr, CMS24833a; Q148R, CMS4627; and I150T, CMS1849a) with indicated SMS mutation with or without PBA treatment. The image is a representative of 3 separate experiments. (D) Quantification of the ratio of matured CtsD to pro CtsD in C. (E) Quantification of the p62/SQSTM1 level in C. The p62/SQSTM1 level was normalized with the β-Actin level. All the values were further normalized by that of the control cells. In B, D, and E, n = 3; *P < 0.05, **P < 0.01; 1-way ANOVA multiple comparisons (matched). Data represent mean ± SEM.

Figure 6. PBA treatment restores acetyl-CoA and protein acetylation in a Drosophila SRS model and fibroblasts of patients with SRS.

(A) Acetyl-CoA level in 10 DAE flies with or without PBA feed. Each dot indicates a sample of homogenized mixture of 10 flies. n = 8; ordinary 1-way ANOVA multiple comparisons. (B) Dot blot of acetyl-lysine in 10 DAE flies with or without PBA feed. The values of the quantification were normalized with the WT fly samples. Each dot indicates a sample of homogenized mixture of 10 flies. n = 6. (C) Western blot of acetyl-lysine in 10 DAE flies with or without PBA feed. Acetylated histones and nonhistone proteins in the bracket areas were quantified separately and normalized with the β-Actin level. All the values were further normalized by that of the WT flies. n = 3. (D) Acetyl-CoA level in patient fibroblasts (Ctrl, CMS24833a; I150T, CMS1849a) with or without PBA treatment. Cells from 4 separate experiments were stored at –80°C and then tested in a single plate. n = 4. (E) Dot blot of acetyl-lysine in patient fibroblasts with or without PBA treatment. The values of the quantification were normalized with the control cells. n = 5. (F) Western blot of acetyl-lysine in patient fibroblasts with or without PBA treatment. Acetylated histones and nonhistone proteins in the bracket areas were quantified separately and normalized with the β-Actin level. All the values were further normalized by that of the control cells. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001; 1-way ANOVA multiple comparisons (matched) in B–F. Data represent mean ± SEM.

Figure 7. PBA heterogeneously regulates polyamine level.

(A) PBA treatment does not restore the polyamine levels in SRS flies. No acetyl-spermidine was detected in any of the samples. Each dot indicates a sample of homogenized mixture of 10 flies. n = 3; ordinary 1-way ANOVA multiple comparisons. (B) PBA treatment regulates spermidine level in fibroblasts of patients with SRS. No putrescine was detected in any of the samples. Each dot indicates an individual drug treatment experiment. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001; 1-way ANOVA multiple comparisons (matched). Data represent mean ± SEM. Put, putrescine; Spd, spermidine; Spm, spermine.