Fig 2: CARM1 is required for the regulation of lipid metabolism by C9orf72. (A) Wild-type V5-CARM1 and mutant V5-CARM1(?347–380) were expressed in wild-type and C9KO MEFs under glucose starvation conditions for 6 h. Eleven to 12 fields of view containing 39–47 cells from three independent experiments were statistically analyzed. (B) Wild-type and C9KO MEFs were cultured in CM or starved with glucose-free medium for 6 h followed by immunostaining of endogenous CARM1. Fifteen to 18 fields of view in each group from three independent experiments were statistically analyzed for the nuclear enrichment of CARM1. (C) CARM1 protein levels were determined by immunoblotting of the nuclear fractions of wild-type and C9KO MEFs grown with or without glucose starvation conditions for 6 h. The ratio of CARM1 to PARP was calculated and statistically analyzed. n = 3. (D,E) LDs were stained with BODIPY493/503 after wild-type, C9KO, and C9orf72/CARM1 double-knockout MEFs were grown with glucose-free medium for 16 h. The loss of C9orf72 led to an accumulation of LDs, which was reversed by the deletion of CARM1. The number of LDs per cell was counted, and 40 cells in each group from three independent experiments were statistically analyzed. (F,G) Wild-type, C9KO, and C9orf72/CARM1 double-knockout MEFs were grown with glucose-free medium for 16 h followed by quantitation of lipid amounts and FFA levels in each group. n = 3. (H) Wild-type, C9KO, and C9orf72/CARM1 double-knockout MEFs were starved for 16 h. LDs were isolated and immunoblotted against anti-LC3. ADRP served as a loading control. (I) Wild-type, C9KO, and C9orf72/CARM1 double-knockout MEFs were cultured in glucose-free medium with [1,2-14C] acetate at the final concentration of 0.3 µCi for 6 h. The incorporation of [1,2-14C] in total lipids was measured by a scintillation counter. Counts per minute (CPM) were normalized to the total amount of protein. n = 3. (J) After starving wild-type, C9KO, and C9orf72/CARM1 double-knockout MEFs with glucose-free medium for 6 h, the levels of ACC in each group were determined by immunoblotting. (K) ACC mRNA level was determined by quantitative PCR (qPCR) after MEFs were grown with glucose-free medium for 6 h. n = 3. (L–N) Wild-type and C9KO MEFs were starved for 6 h. The enrichment of CARM1 and H3R17me2 on the promoter of Atg14, Map1lc3b, or ACC in each group was measured by ChIP-qPCR (chromatin immunoprecipitation [ChIP] combined with qPCR) using anti-CARM1 and anti-H3R17me2 antibodies, respectively. n = 3. (GS) Glucose starvation. Data are presented as mean ± SEM. (*) P < 0.05; (**) P < 0.01; (***) P < 0.001; (****) P < 0.0001.

Fig 3: NDST3 is downregulated in C9orf72-linked ALS patient cells and tissues A, BNDST3 protein levels relative to ß-tubulin in B lymphocytes derived from healthy controls and C9orf72-linked ALS patients, as examined by immunoblotting (n = 3 biologically independent samples, **P = 0.0026).C–FThe protein levels of NDST3 relative to GAPDH (D) (**P = 0.0013), Ac-a-tubulin relative to total a-tubulin (E) (*P = 0.0379), and total a-tubulin relative to GAPDH (F) (P = 0.5700), as examined by immunoblotting, in spinal cord tissues from healthy controls and C9orf72-linked ALS patients (n = 4 biologically independent samples in the control group and n = 5 biologically independent samples in the ALS patient group).G–IImmunoblotting assays showing downregulation of NDST3 in C9orf72 shRNA (shC9orf72)-treated RPE1 cells, as compared to control shRNA (shCTRL)-treated cells. Quantification of C9orf72 and NDST3 expression relative to GAPDH is shown in (H) (n = 4 independent experiments, ***P = 0.0004) and (I) (n = 4 independent experiments, ***P = 0.0001), respectively.J–LImmunoblotting assays showing up-regulation of NDST3 in RPE1 cells over-expressing Flag-C9orf72 compared to those expressing Flag-EGFP as control. Quantification of C9orf72 and NDST3 expression relative to GAPDH is shown in (K) (n = 3 independent experiments, **P = 0.0011) and (L) (n = 3 independent experiments, **P = 0.0033), respectively.M, NBaf A1 sensitivity assay for WT, NDST3 KO, and Noco-treated NDST3 KO RPE1 cells expressing proline-arginine poly-dipeptide (PR82). 24 h after the transfection of PR82, WT and NDST3 KO cells were treated with vehicle or 10 µg/ml Noco overnight. The cell survival measured using calcein AM staining at 48 h post-transfection was quantified as the percentage of the fluorescence intensity for treated cells relative to that of untreated cells (n = 3 independent cultures, *P = 0.0369, ***P = 0.0007). Scale bar, 200 µm.O, PImmunoblotting analysis of PR82 levels in WT, NDST3 KO, and Noco-treated NDST3 KO RPE1 cells from (M, N) after the drug treatment. Quantification of the levels of PR82 relative to ß-tubulin is shown in (P) (n = 4 independent experiments, *P = 0.0141, **P = 0.0045). Data information: Error bars represent ± standard deviation. Source data are available online for this figure.

Fig 4: Loss of C9orf72 increases FA flux and biogenesis. (A,B) Wild-type and C9KO MEFs were starved with glucose-free medium for 16 h, and then the association of autophagy with LDs was monitored by immunostaining of BODIPY and lysosomal marker LAMP1. The number of BODIPY and LAMP1 double-positive puncta per cell was counted, and 30–40 cells in each group from three independent experiments were statistically analyzed. (C) Wild-type and C9KO MEFs were grown with or without glucose starvation for 16 h, and then LDs were isolated and subjected to SDS-PAGE and immunoblot against anti-LC3 and P62. ADRP served as a loading control. The ratios of LC3II/I and the amounts of P62 were calculated and statistically analyzed. n = 3. (D–F) Wild-type, C9KO, and C9/Atg5 double-knockout MEFs were starved with glucose-free medium for 16 h. The LDs were labeled with BODIPY and counted, and the association of autophagy and LDs was monitored as in A. Twenty-eight to 30 cells in each group from three independent experiments were statistically analyzed. (G) Wild-type and C9KO MEFs were cultured in CM or glucose-free medium with [1,2-14C] acetate at a final concentration of 0.3 µCi for 6 h. The incorporation of [1,2-14C] in total lipids was measured by a scintillation counter. Counts per minute (CPM) were normalized to the total amount of protein. n = 6. (H–J) Wild-type and C9KO MEFs were grown in glucose-free medium, and the protein levels of acetyl-CoA carboxylase (ACC) and NOX2 were determined at the indicated time points by immunoblotting. n = 3. (GS) Glucose starvation. Data are presented as mean ± SEM. (*) P < 0.05; (**) P < 0.01.

Fig 5: Loss of C9orf72 alters lipid metabolism under glucose starvation. (A) Wild-type and C9KO MEFs were starved with glucose-free medium for the indicated hours, and the lipid amount at each time point was detected by a fluorometric lipid-binding assay as described in the Materials and Methods. n = 3. (B,C) Wild-type and C9KO MEFs were cultured with CM or starved with glucose-free medium for 16 h, and LDs were stained by BODIPY493/503. The number of LDs per cell was counted, and 50 cells per group from three independent experiments were analyzed. (D,E) Wild-type and C9KO MEFs cultured with CM or starved with glucose-free medium (16 h) were triple-labeled with LDs by BODIPY493/503 and ER marker calnexin for confocal immunofluorescence microscopy analysis. The colocalizations of BODIPY493/503 and calnexin per cell were analyzed by Pearson's R-value. Thirty cells in each group from three independent experiments were statistically analyzed. (F) Wild-type and C9KO MEFs were starved with glucose-free medium for 0, 6, and 16 h. The FFA amounts were determined by an enzyme-coupled fluorometric assay as described in the Materials and Methods. n = 3. (GS) Glucose starvation. Data are presented as mean ± SEM. (*) P < 0.05; (**) P < 0.01; (***) P < 0.001.