Fig 2: GFI1-dependent survival of MM cells is mediated by intracellular S1P levels. MM.1S, OPM-2, and KMS-11 GFI1 o/e and empty vector (EV) control cells were treated with different doses of SK1-I (5, 7.5 and 10 µM) or vehicle control (H2O) (0 µM) for 48 h. Viability was measured by AlamarBlue assay and reported as percent vs. untreated EV control (MM.1S left panel) or percent vs. vehicle control for each cell type to better appreciate the difference between EV and GFI1 o/e, (OPM-2 middle panel and KMS-11 right panel) (* p < 0.05 vs. untreated control and # p < 0.005 vs. untreated control) (A); Mouse pre-osteoblastic MC4 cell line (left panel) and mouse pre-osteocytes MLO-A5 cell line were co-cultured in a 3D-model with MM.1S-gLuc cells in a 1:5 ratio and treated or not with SK1-I (10 µM). Relative secreted Guassia luciferase activity in the supernatant was measured after 24, 48, and 72 h by a luminometer (* p < 0.05, ** p < 0.01 and *** p < 0.0001) (B); SPNS2 specific S1P transporter was inhibited by either pharmacological treatment with Calcitonin (CT) (0, 25, 50, and 100 nM) of KMS-11 cells (left panel) or transcriptional repression (transduction with doxycycline-inducible shSPNS2 plasmid, 30% inhibition of SPNS2 gene expression in dox-treated vs. untreated controls after three days) in MM.1S cells. Viability was measured by AlamarBlue assay and reported as percent vs. untreated control (C); GFI1 knockdown was induced in MM.1S ishGFI1 and OPM-2 ishGFI1 cells by dox treatment (1 µg/mL) for three days. MM.1S ishGFI1 dox treated (40% GFI1 KD) and OPM-2 ishGFI1 dox treated (30% GFI1 KD) were then treated with THI (10 µM), K6PC5 (5 µM), or vehicle (DMSO) for an additional 48 h (MM.1S) or 24 h (OPM-2) and viability was measured by AlamarBlue assay and reported as percent vs. each cell type vehicle control (D).

Fig 3: Microenvironmental factors activate the GFI1-SphK1-c-Myc axis for survival. GFI1, SphK1, and c-Myc mRNA expression detected by qPCR and expressed as fold change versus untreated control in MM cell lines H929 (p53 WT) (A) and U266 (p53 mut) (B) treated or not with IL-6 (5 ng/mL; 4 h), indicated proteins expression levels by WB (using total ERK1/2 (T-ERK1/2) and ß actin [see Figure S6] as loading control) in H929 cells transiently knocked down using shGFI1 and Sh-vector control (C), and in KMS-11 (p53 null) cells stably expressing GFI1 o/e and their EV controls (using ß actin as loading control) (D) treated or not with IL-6 (5 ng/mL; 4 h); MM cells H929 (p53 WT) (left panel) and JJN3 (p53+/-) (right panel) were directly co-cultured for 24 h with bone marrow stromal cells (BMSC) obtained from MM donors (MM 3, 6, and 9 in Table S1). MM cells were harvested separately and SphK1 and c-Myc mRNA levels were measured by qPCR and compared with levels in MM cells alone (n = 3) (E); GFI1, P-SphK1, and c-Myc protein expression by WB using VCP as loading control in H929 cells transiently knocked down by shGFI1 and sh-vector control and co-cultured or not with MM patient derived BMSC (MM3 in Table S1) (F); GFI1, P-SphK1, and c-Myc protein expression by WB in MM.1S (p53 WT) (left panel) and OPM-2 (p53 mut) (right panel) treated for 15 min with the specified concentrations of S1P or vehicle control (ethanol) and normalized by VCP (G). Original WB data in Figure S6 (C,D,F,G).

Fig 4: Intracellular S1P levels modulate c-Myc protein stability in a protein phosphatase 2A (PP2A)-dependent manner. Cell lysates were analyzed by WB (original data in Figure S6) using specific antibodies as indicated and representative immunoblots are shown (A,B,D,F,H). Immunoblots are arranged with the internal control used for the same samples and, therefore, are not in the same order for the different cell lines in (D). DMSO is the vehicle control (also shown as 0) for all treatments. MM.1S ishGFI1 and OPM-2 ishGFI1 were treated or not with dox (1 µg/mL, 3 days) to induce GFI1 KD (E, F, H-I). MM.1S, OPM-2 and KMS-11 cells were treated with SK1-I for 24 h. A schematic of S1P synthesis and degradation (main enzymes and their specific inhibitors) is shown below (A); MM.1S and OPM-2 cells were treated with THI (0, 1, 10 µM, 24 h) (B); Spearman correlations (r= -0.1934) between c-Myc and SGPP1 mRNAs (p < 0.0001 by two-tailed test) in MM patient samples (n = 414) of GSE4581 database (C); MM.1S and OPM-2 cells pre-exposed to OA (2.5 nM, 2 h) were treated with SK1-I for 24 h (D); OPM-2 GFI1 KD and control cells were exposed to OA for 2 h. PP2A activity was measured and results (n = 3) were graphed as percentage activity vs. vehicle control (E); MM.1S GFI1 o/e cells and EV treated with either THI (10 µM, 2 h) or SK1-I (5 µM, 2 h) and as well as OPM-2 GFI1 KD and Control cells were IP with anti-PP2Ac antibody and probed for I2PP2A and PP2Ac by WB (F); MM.1S and OPM-2 GFI1 o/e and EV control cells pre-exposed to OA (2.5 nM, 2 h) were then treated with SK1-I (3 µM and 5 µM respectively). Viability after 48 h was reported as percent vs. vehicle treated control for each cell type (G); MM.1S and OPM-2 GFI1 KD and Control (H,I) were treated with either K6PC5 (5 µM), THI (10 µM), OA (2.5 nM) or DMSO for 4 h for WB (H), or were exposed to OA (0, 0.1, 1, 2.5 nM) and viability after 48 h was reported as percent vs. vehicle treated control for each cell type (I).

Fig 5: Insufficient intracellular S1P in MM cells induces cell death via autophagy. MM.1S and OPM-2 cells were treated with SK1-I (5 and 10 µM) or vehicle control (DMSO shown as 0) for 24 h (A,B) or 48 h (C) and pre-treated or not with Baf (50 nM, 2 h) (B,C). Whole cell lysates were probed by WB with indicated specific antibodies and presented by representative immunoblots (A,B) (original data in Figure S6). Viability was measured by AlamarBlue assay and reported as percent vs. EV vehicle treated control (p < 0.0005 in both MM.1S and OPM-2 panels: * vs. vehicle control; ** vs. respective SK1-I treatments alone; # vs. GFI1 o/e vehicle control and ## vs. EV respective treatments) (C).