Fig 2: Lack of GPR44 activates KRAS-MAPK signaling in AML cells(A) Scheme for sequencing analysis. Bone marrow cells were isolated from mice secondarily transplanted with 1° WT or Gpr44-/- AML donor cells and LICs were flow cytometrically sorted following RBC lysis, Lin- selection, and Sca-1 and c-Kit staining. LICs (4 × 105) were used for RNA sequencing. Differential gene expression analysis, IPA, and GSEA were performed to compare WT and Gpr44-/- LICs (n = 3 in each group).(B) Heatmap for top 100 most differentially regulated genes in WT and Gpr44-/- LICs.(C) Volcano plot for differential gene expression analysis between WT and Gpr44-/- LICs(D) Heatmap of indicated pathways in WT and Gpr44-/- LICs as analyzed by IPA. Z scores were plotted.(E) Expression of Kras assessed by qPCR analysis in WT or Gpr44-/- AML cells isolated from bone marrow (left) and spleen (right) of 2° AML-recipient mice. Data were normalized to WT AML cells and 18S rRNA expression (n = 4–8 biological replicates).(F) Western blot showing the expression of KRAS in WT or Gpr44-/- AML cells isolated from 2° AML-recipient mice.(G) Western blot showing the expression of MAPK signaling pathway components including P-C-RAF, RAF-1, P-MEK3/6, MEK1/2, P-ERK, ERK1/2, P-C-JUN,and ß-actin in WT or Gpr44-/- AML cells isolated from 2° AML-recipient mice (n = 5–8 biological replicates in each group).(H) Western blot showing the expression of P53 in WT or Gpr44-/- AML cells isolated from 2° AML-recipient mice. (F and H) Densitometry was done by normalizing to WT AML cells and relative to ß-actin (n = 3–5 biological replicates in each group).Data shown are mean ± SEM per group; *p < 0.05, **p < 0.01.

Fig 3: GPR44 activation decreases the severity of leukemia(A) CD45.2-recipient mice were maintained on Se-A diet for 4 weeks before retro-orbital transplantation with 1° CD45.1 WT AML donor cells; at 1 week post transplantation, mice were treated i.p. daily with PBS or 0.2 mg/kg CyPGs (DK-PGD2, 15d-PGJ2, ?12-PGJ2, and ?12-PGJ3) for 2 weeks and euthanized after treatment; blood, bone marrow, and spleen were sampled (n = 7–8 in each group).(B) CBC analysis of AML mice at the endpoint in (A).(C) Representative image of spleens isolated from AML mice in (A).(D) Spleen weights (mg) of AML mice in (A).(E and F) Counts of LICs (CD45.1+Lin-Sca-1-c-Kit+, see also Figure S1J) in the bone marrow (E) and spleen (F) of AML mice in (A).(G) 1°-recipient mice were maintained on Se-A diet for 4 weeks before sublethal irradiation (4.75 Gys) and retro-orbital transplantation with GFP BCR-ABL-transduced HSCs; DK-PGD2 (0.2 mg/kg) or PBS control was given daily i.p. starting when peripheral blood WBC is over 5 K/µL for 2 weeks post transplantation. Mice were euthanized after treatment and blood, bone marrow, and spleen were sampled (n = 7–9 in each group).(H) CBC analysis of CML mice in (G).(I) Spleen weights (mg) of CML mice in (G).(J) Weight gain of Se-A CML mice treated with or without DK-PGD2 was expressed as the ratio of after and before treatment (n = 4 in each group).(K and L) Counts of CML cells (GFP+) in the Lin- population in the bone marrow (K) and spleen (L) of CML mice in (G).(M and N) Counts of LICs (GFP+Lin-Sca-1+c-Kit+, see also Figure S3D) in the bone marrow (M) and spleen (N) of CML mice in (G).Data shown are mean ± SEM per group; each dot represents a mouse; *p < 0.05, **p < 0.01.

Fig 4: Lack of GPR44 in AML cells results in aggressive disease(A) Western blot showing the expression of GPR44 and ß-actin in purified 1° WT and Gpr44-/- AML cells. GPR44 antibody detects glycosylated (63 kDa) and unglycosylated (33 and 50 kDa) and forms.(B) CD45.2- and Ai14TdTomato-recipient mice were maintained on Se-S diet for 4 weeks before transplantation until the endpoint. 2° transplantation was done retro-orbitally with 1° CD45.1 WT or CD45.2 Gpr44-/- AML donor cells to CD45.2= and Ai14TdTomato-recipient mice, respectively; 3 weeks later, mice were euthanized; blood, bone marrow, and spleen were sampled (n = 8 in each group).(C) CBC analysis of Se-S AML mice in (B).(D and E) Counts of AML cells in the Lin- population in the bone marrow (D) and spleen (E) of Se-S AML mice in (B).(F and G) Counts of LICs (WT: CD45.1+Lin-Sca-1-c-Kit+, see also Figure S1J; Gpr44-/-: RFP-Lin-Sca-1-c-Kit+, see also Figure S3K) in the bone marrow (F) and spleen (G) of Se-S AML mice in (B).(H) Survival curve of recipient mice with competitive 2° transplantation of WT (4 × 105), WT + Gpr44-/- (2 × 105: 2 × 105), or Gpr44-/- (4 × 105) AML donor cells (n = 6–9 in each group).(I) Progression of WBCs in the peripheral blood of recipient mice secondarily transplanted with WT or Gpr44-/- AML donor cells (n = 8–9 in each group).(J) Purified 1 WT and Gpr44-/- AML cells were plated in methylcellulose medium (2,500 cells/well, 4 replicates). CFUs were counted on day 8.(K) Representative image of colony growth from purified 1 WT and Gpr44-/- AML cells. Scale bar, 100 µm.(L) Comparison of GPR44 expression in blood cancers compared with normal subjects.(M) Comparison of GPR44 expression in AML FAB subtypes including M0, M1, M2, M3, M4, M4Eo, M5, M6, and M7.(L and M) Data were generated from the ONCOMINE database. (L) Each point represents a comparison of the study between the cancer population and normal population. (M) Each point represents a comparison of the study of one FAB subtype with the other subtypes. Plots were generated using the -log10 (p value) and the fold change in expression. p values were obtained by t test of the mean values. An absolute fold change of 1.5 (red line) or higher is considered significant.Data shown are mean ± SEM per group; *p < 0.05, **p < 0.01.

Fig 5: Inhibition of KRAS/MAPK signaling reverses GPR44 KO-associated effect in AML(A) Scheme showing the treatment of BAY293 and CCK8 assay on unpurified Gpr44-/-AML cells. Frequency of AML cells was above 95%.(B) Comparison of the viability of unpurified Gpr44-/- AML cells treated with BAY293 (0, 5, 10, and 15 µM, n = 3) for 24 or 48 h. Data were normalized to 0 µM (24 h) treatment. **p < 0.01, comparison between concentrations in the 24 and 48 h time points analyzed by Student’s t test; ##p < 0.01, comparison between time points analyzed by two-way ANOVA followed by appropriate post hoc test (Bonferroni correction).(C) Scheme showing the treatment of BAY293 and live/dead cell measurement on purified Gpr44-/- AML cells. CD45.2 Gpr44-/- AML cells were transplanted into CD45.1 recipients; purification of Gpr44-/- AML cells isolated from spleen was done by CD45.1-positive magnet selection kit.(D and E) Frequency (D) and count (E) of live cells in purified Gpr44-/- AML cells treated with 15 µM BAY293 for 48 h (n = 3).(F) Western blot showing the expression of P-MEK3/6 and P-ERK in purified 1° Gpr44-/- AML cells transduced with pLV hU6-sgRNA KRAS hUbC-dCas9-KRAB-T2a-GFP virus.(G) Western blot showing the expression of KRAS in purified 2° Gpr44-/- AML cells transduced with pLV hU6-sgRNA KRAS hUbC-dCas9-KRAB-T2a-GFP virus. Densitometry was done by normalizing to control group and relative to ß-actin (n = 3–4 in each group).(H) Survival curve of AML mice tertiarily transplanted with purified KRAS KD Gpr44-/- AML cells (n = 7).(I and J) Survival analysis of mice tertiarily transplanted PD98059-treated Gpr44-/- AML cells. Gpr44-/- AML cells were cultured ex vivo with or without 100 µM PD98059 for 24 h and then retro-orbitally transplanted into mice (n = 6 per group). Survival was followed up for 60 days post transplantation.(K) Western blot showing the expression of P-ERK and P53 in Gpr44-/- AML cells treated with PD98059 (0, 10, 50, and 100 µM) for 24, 48, and 72 h.(L and M) Representative image of colony growth from purified 2° Gpr44-/- AML cells (2,500 cells/well, n = 4 for biological replicate, n = 2 for technical replicate) treated with 100 µM PD98059 and 15 µM BAY293 in methylcellulose medium. Scale bar, 100 µm.(N) Counts of CFUs from purified 2° Gpr44-/- AML cells treated with 100 µM PD98059 and 15 µM BAY293 in methylcellulose medium. CFUs were counted on day 8.Data shown are mean ± SEM per group; *p < 0.05, ** p < 0.01.