Runx1 Orchestrates Sphingolipid Metabolism and Glucocorticoid Resistance in Lymphomagenesis

Abstract

The three-membered RUNX gene family includes RUNX1, a major mutational target in human leukemias, and displays hallmarks of both tumor suppressors and oncogenes. In mouse models, the Runx genes appear to act as conditional oncogenes, as ectopic expression is growth suppressive in normal cells but drives lymphoma development potently when combined with over-expressed Myc or loss of p53. Clues to underlying mechanisms emerged previously from murine fibroblasts where ectopic expression of any of the Runx genes promotes survival through direct and indirect regulation of key enzymes in sphingolipid metabolism associated with a shift in the "sphingolipid rheostat" from ceramide to sphingosine-1-phosphate (S1P). Testing of this relationship in lymphoma cells was therefore a high priority. We find that ectopic expression of Runx1 in lymphoma cells consistently perturbs the sphingolipid rheostat, whereas an essential physiological role for Runx1 is revealed by reduced S1P levels in normal spleen after partial Cre-mediated excision. Furthermore, we show that ectopic Runx1 expression confers increased resistance of lymphoma cells to glucocorticoid-mediated apoptosis, and elucidate the mechanism of cross-talk between glucocorticoid and sphingolipid metabolism through Sgpp1. Dexamethasone potently induces expression of Sgpp1 in T-lymphoma cells and drives cell death which is reduced by partial knockdown of Sgpp1 with shRNA or direct transcriptional repression of Sgpp1 by ectopic Runx1. Together these data show that Runx1 plays a role in regulating the sphingolipid rheostat in normal development and that perturbation of this cell fate regulator contributes to Runx-driven lymphomagenesis. J. Cell. Biochem. 118: 1432-1441, 2017. © 2016 Wiley Periodicals, Inc.

Keywords: GLUCOCORTICOID; ONCOGENE; Runx1; SPHINGOLIPID; SURVIVAL.

Figure 1

Ectopic Runx1 promotes S1P release from T‐lymphoma cells. (A) Total protein was extracted from p/m97 or p53/184 thymic lymphoma cells transduced with the pBabeRunx1 retroviral vector or the pBabePuro vector control and probed against antibodies to Runx1 (Cell Signalling #8229) or actin (Santa Cruz sc‐1616) as a loading control. Lymphoma cells over‐expressing Runx1 (9) were included as a positive control. (B) The blot was quantified using image J software and the Runx1 fold change indicated below the histogram. (C) Long chain ceramides were extracted from cell pellets from pBabePuro vector control and pBabeRunx1‐expressing T lymphoma cells (p/m97 shown), and separated, identified, and semi‐quantitated by HPLC mass spectrometry. The data are means ± SD where n = 4 from one experiment typical of two. Solid bars represent combined levels of 16.0, 24.1, and 24.0‐Cer (P = <0.01). A significant difference between ectopic Runx1 expressing and non‐expressing cells was preserved when the remaining ceramides (14.0, 16.1, 18.0, 20.0, 22.1, 24.2‐Cer—displayed in hashed bars) were included in the total (P = <0.01). (D) C18‐S1P was analyzed from conditioned media collected from pBabePuro vector control and pBabeRunx1‐expressing T lymphoma cells by HPLC mass spectrometry (p/m97 shown). FCS was included to activate Sphk1. The data are means ± SD where n = 4 from one experiment typical of two.

Figure 2

Enforced deletion of Runx1 impairs S1P release in vivo. (A) PCR genotyping of 60‐day‐old splenic tissue DNA from Runx1^fl/flMx1Cre⁺ mice treated with vehicle control (PBS) for background excision or pIpC to excise Runx1. Samples: A–C, 60‐day spleen tissue samples from two groups of three Runx1^fl/flMx1Cre⁺ mice treated with either PBS or pIpC; D–F DNA controls, non‐excised Runx1^fl/flMx1Cre⁻ (D), a mixture of partially excised Runx1^fl/flMx1Cre⁺ and Balb/c normal kidney to show all three possible PCR products (E) and Balb/c normal kidney (F). Positions of Floxed (Runx1Fl), deleted (ΔRunx1Fl), and wildtype (WT) Runx1 alleles are as shown. Quantitation to estimate relative excision was performed using image J software. Estimated background and pIpC excision rates are labeled under the blots. (B) qt‐RT‐PCR analysis of Runx1 expression in 60‐day splenic tissue RNA samples from Runx1^fl/flMx1Cre⁺ mice treated with vehicle control or pIpC (dark and light green bars, respectively). Parallel samples were analyzed from Runx1^fl/fl mice lacking the Mx1Cre gene to control for PBS and pIpC treatment (black and gray bars, respectively). The data are means ± SD where n = 9 representing three technical replicates of each biological replicate (3). A significant reduction in Runx1 expression was observed between Mx1Cre⁺ and Mx1Cre⁻ mice indicating background excision (gray vs. light green bar; P = <0.05). Inclusion of pIpC to excise Runx1 generated a further reduction (black vs. green bar; P = <0.01). (C) Remaining splenic tissues were pulverized and analyzed by HPLC mass spectrometry for C18‐S1P. Data are expressed relative to 1 mg of pulverized spleen tissue and represent means ± SD of spleens of three mice for each set of conditions.

Figure 3

Runx1 protects lymphoma cells against dexamethasone‐mediated apoptosis. (A) p53 null lymphoma cells transduced with the pBabeRunx1 retroviral vector or the pBabePuro vector control (p/m97 shown) were plated in triplicate in the presence and absence of 1.0 μM dexamethasone and monitored for live/dead counts by trypan blue exclusion over 4 days. (B) The same cells were plated at 4 × 10⁵ per well in triplicate wells of a 12‐well plate and monitored for growth over 48 h. Ectopic Runx1 significantly reduced cell proliferation at 24 h (P < 0.01), 40 h (P < 0.01), and 48 h (P < 0.01). (C) qt‐RT‐PCR analysis of Runx1, Sgpp1, Ugcg, and Nr3c1 in pBabePuro vector control and pBabeRunx1‐expressing T lymphoma cells (p/m97 shown) grown in the presence and absence of 1.0 μM dexamethasone for 6 h. The data are means ± SD where n = 9 representing three technical replicates of each biological replicate (3) from one experiment typical of two.

Figure 4

Enforced deletion of Runx1 promotes dexamethasone‐mediated apoptosis and Sgpp1 transcription. (A) Western blotting analysis as described in Figure 1A to detect the deleted (ΔRunx1Fl) and full length (Runx1Fl) Runx1 proteins from in vitro excised in Runx1^fl/flMx1Cre⁺ 3s B lymphoma cells. (B) Paired cell lines expressing the deleted (ΔRunx1Fl) and full length (Runx1Fl) proteins after in vitro excision of Runx1^fl/flMx1Cre⁺ 3s B lymphoma cells were plated in triplicate in the presence of 1.0 μM dexamethasone and monitored for live/dead counts by trypan blue exclusion. (C) qt‐RT‐PCR analysis of steady state levels of Sgpp1 in ΔRunx1Fl and Runx1Fl 3s cells grown in the presence and absence of 1.0 μM dexamethasone for 6 h. The data are means ± SD where n = 9 representing three technical replicates of each biological replicate (3) from one experiment typical of two. (D) Runx1 schematic showing the mutated residues in the heterodimerization (T161A) and DNA‐binding (K83N) domains. qt‐RT‐PCR analysis of Sgpp1 expression in ΔRunx1Fl 3s cells transfected with full length Runx1, T161A Runx1, or K83N Runx1. Absolute levels of Sgpp1 were compared to control cultures expressing the pBabe Puro vector alone. The data were compiled as described in (C).

Figure 5

shRNA knockdown of Sgpp1 reduces dexamethasone‐mediated apoptosis. (A) qt‐RT‐PCR analysis of Sgpp1 expression in p/m97 cells stably infected with viral supernatants expressing an Sgpp1 or a non‐coding (NC) control shRNA sequence. Cells were grown for 6 h in the presence of 1.0 μM dexamethasone prior to RNA extraction and qt‐RT‐PCR analysis. The data were calculated as described in Fig. 4C. (B) The same cells were plated in triplicate and grown for 36 h in the presence (Dex) and absence (MtOH) of 1.0 μM dexamethasone and monitored for live/dead counts by trypan blue exclusion. Knockdown of Sgpp1 had no effect on cell viability under control conditions but gave a significant reduction in cell death in the presence of 1.0 μM dexamethasone. (C) Interplay between Runx1 and dexamethasone on the expression of sphingolipid metabolism enzymes involved in the synthesis and breakdown of sphingosine and ceramide and their potential contributions to cell death and survival [Bianchini et al., 2006; Kilbey et al., 2010].