4-Phenylbutyrate restores localization and membrane repair to human dysferlin mutations

Abstract

Dysferlinopathies are muscular dystrophies caused by recessive loss-of-function mutations in dysferlin (DYSF), a membrane protein involved in skeletal muscle membrane repair. We describe a cell-based assay in which human DYSF proteins bearing missense mutations are quantitatively assayed for membrane localization by flow cytometry and identified 64 localization-defective DYSF mutations. Using this platform, we show that the clinically approved drug 4-phenylbutryric acid (4-PBA) partially restores membrane localization to 25 mutations, as well as membrane repair to cultured myotubes expressing 2 different mutations. Two-day oral administration of 4-PBA to mice homozygous for one of these mutations restored myofiber membrane repair. 4-PBA may hold therapeutic potential for treating a subset of humans with muscular dystrophy due to dysferlin deficiency.

Keywords: Drugs; Musculoskeletal medicine.

Graphical abstract

Figure 1

Determination of DYSF^PMMs PM localization: The 2A-assay (A) Schematic outline of the 2A-assay. Human DYSF isoform 8 cDNA was inserted into a bicistronic lentiviral vector expression vector regulated by a minimal cytomegalovirus (CMV) promoter and tetracycline response element (TRE). Transfection of LV-TRE-DYSF-T2A-DsRed into HEK293T cells results in equimolar expression of DYSF-2A, a fusion protein with a C-terminal 2A-peptide, and DsRed protein, which are separated by cleavage at the T2A peptide sequence during translation. A mouse α-2A peptide antibody recognizes the extracellular region of DYSF-T2A protein in live cells; subsequent binding of Alexa 647 α-mouse IgG secondary antibody enables quantitation of PM-localized DYSF intensity on the surface of live cells that express cytoplasmic DsRed by flow cytometry. The amount of DsRed translationally expressed is equimolar to the amount of DYSF-T2A. Quantification of these 2 signals allows us to calculate a “2-A assay value,” which represents the amount of PM-localized Dysferlin relative to DsRed for any HEK cell population expressing a mutant DYSF relative to the 2-A assay value in a similar population of cells expressing wild-type DYSF. (B) Validation of the HEK cell-based 2A-assay was performed using vector constructs expressing DYSF^WT and 1 of the following 3 known pathogenic DYSF^PMMs: DYSF^V67D, DYSF^R555W and DYSF^L1341P, and DYSF^A170E, a commonly occurring DYSF SNP that is predicted to be non-pathogenic. HEK cells were transiently transfected, cultured, and processed for flow cytometry as described in materials and methods. PM localized expression of DYSF^PMMs is determined relative to DYSF^WT. Data are represented as mean (n = 3) ± S.D., ∗∗∗p < 0.001 versus DYSF^WT, by Student’s t test. (C) Immunofluorescence images localizing DYSF^PMMs expressed in HEK cells. Forty-eight hours post transfection HEK cells expressing the listed DYSF^PMMs were FACS sorted for DsRed, cultured, fixed, and stained with the Hamlet α-DYSF 1˚Ab (green) and the α-Na/K ATPase-Ab (red) to identify the plasma membrane (PM). Scale bar: 25 μm.

Figure 2

Determination of PM-localized expression of 113 DYSF^PMMs by 2A assay and ICC HEK cells were transiently transfected with LV-TRE-DYSF-T2A-DsRed vectors bearing DYSF^WT or 1 of 113 various DYSF^PMMs. The amount of PM-localized DYSF was determined by 2A-assay; expression of PMMs is reported relative to DYSF^WT (n = 6. Data are means ± S.D.). DsRed-positive cells were sorted and cultured on coverslips and subject to ICC to visually determine DYSF localization; red bars indicate no observable PM localization, whereas blue bars indicate that PM localization was observed for a given DYSF^PMM.

Figure 3

Restoration of DYSF^L1341P PM localization by 4-PBA and corr-2b (A) Results of 2A-assays (n = 3, data are means ± S.D.) on 64 DYSF^PMMs that have less than 25% of DYSF^WT PM localization (Figure 2) following treatment with DMSO (0.1%), 4-PBA (1 mmol/L), or corr-2b (25 μmol/L) for 24 h. Twenty-one mutants (highlighted in light gray) significantly respond to 4-PBA only and 4 mutants (highlighted in dark gray) responded to both 4-PBA and corr-2b, boosting DYSF^PMMs PM localization above the 25% threshold (dashed line). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 by Student t-test for all figures. (B) ICC DYSF localization in transfected (unsorted) HEK cells expressing DsRed and either DYSF^WT or DYSF^L1341P treated with DMSO (0.1%), 4-PBA (1 mM), or corr-2b (25 μM) for 24 h. Live cells were stained with α-2A-Ab (green) to identify PM-localized DYSF protein. DAPI staining and α-Na/K ATPase-Ab (red) hybridization was used to identify the nuclei and PM in all cells, respectively. Scale bar: 50 μm.

Figure 4

Treatment with 4-PBA restores membrane repair in GREG (DYSF^L1341P) myotubes and MMex38 (DYSF^L1360P) mouse myofibers (A–D) Selected image frames from beginning, middle, and end of membrane repair assays in the presence of FM1-43 dye following laser irradiation of myotube (A) or myofiber (C and F) membranes; white arrowheads show sites of membrane wounding by laser. Quantification of the change in fluorescent intensity (ΔF) caused by intracellular FM1-43 dye infiltration after membrane injury (B, D, and G); n-values are the total number of fibers tested in 2 independent experiments. (A and B) DYSF^WT or DYSF^L1341P transfected GREG myotubes treated with listed compounds; DYSF^WT (DMSO) n = 15, DYSF^WT (4-PBA) n = 19, DYSF^L1341P (DMSO) n = 16, DYSF^L1341P (4-PBA) n = 26. (C and D) C57BL6/NJ (+/+) or MMex38 (L3160P) mouse myofibers from explanted EDL muscles treated with DMSO or 4-PBA (1 mM) in vitro for 24 h; fiber number; +/+ (DMSO) n = 6, +/+ (4-PBA) n = 16, L1360P (DMSO) n = 15, L1360P (4-PBA) n = 18. (E–G) (E) Immunofluorescent staining of fresh frozen histological cross sections of EDL muscle isolated from 3-month-old male +/+ and MMex38 mice treated with vehicle or 4-PBA (2 mg/mL) in drinking water for 48 h. DAPI staining was performed for nuclear localization, and DYSF staining was done using Romeo α-DYSF-1 Ab and Alexa 647 α-mouse IgG Ab fluorescent secondary Ab. Images were all taken at the same exposure time and magnification; scale bar, 50 μm. Representative assay images (F) and membrane repair kinetics (G) of EDL muscle myofibers isolated from +/+ and MMex38 (L1360P) mice treated with vehicle or 4-PBA in drinking water for 48 h. Number and genotype of mice treated; +/+ (water) n = 2, +/+ (4-PBA) n = 2, L1360P (water) n = 3, L1360P (4-PBA) n = 3; Number and genotype of fibers used for membrane repair assay; +/+ (water) n = 9, +/+ (4-PBA) n = 5, L1360P (water) n = 17, L1360P (4-PBA) n = 22. All plotted data are means ± S.D; p values calculated by Student’s t test (∗∗∗p < 0.001).