Fig 2: Range of CD-marker expression in thymocyte populations and LSC represented by Lck-Lmo2 thymus cells. TPM values of each mRNA were computed from the DN cells of asymptomatic, pre-leukaemic Lck-Lmo2 or age-matched wild type mice for GPR56 (A), CD53 (B) and CD59 (C). Closed circles are Lck-Lmo2 values, closed squares are wild type values. Panel D summarizes the expression patterns of the surface proteins analyzed in this study, during T cell development from immature DN1 cells through to single positive CD4 or CD8, T cell receptor expressing T cells. Two key markers, CD25 and GPR56, first expressed on the LICs are arrowed where their co-expression is found and that could be employed to target the LMO2-dependent T-ALL LSC.

Fig 3: Expression levels of candidate target surface proteins in immature thymocytes and mouse or human T cell neoplasias. The RPKM values for four surface markers (y axis) for the triplicate samples of DN1, 2, 3 and 4 from wild-type (blue) or asymptomatic Lck-Lmo2 (red) compared to two thymocyte tumors in Lck-Lmo2 mice (green) or human primary T-ALL samples (black; TN1, 2 and NT1-4) (Supplementary Table S7). Panel A: CD53 values. Panel B: CD59a values. Panel C: GPR56 values; Panel D: CD7 values (NB: Cd7 could not be computed for mouse mRNA as the mouse equivalent of CD7 is equivocal).

Fig 4: CDK7/12/13 inhibitors suppress the GPR56+ LSC compartment in vitro and in vivo Gene expression of GPR56, SMO (left), and TGFb targets SRC, ZEB1, and ZEB2 (right) normalized to GAPDH in AML-491 cells determined by q-RT-PCR 4 h after treatment with THZ1 0.5 µM or LDC4297 2 µM. Unpaired t-test. Mean, SD, and individual values from three individual treatments. ***P < 0.0005, **P < 0.005, *P < 0.05.Setup of in vivo drug treatment. NSG mice were injected with 105 AML 04H112 cells. Four weeks post injection, bone marrow (BM) was analyzed for human leukemic engraftment by BM aspiration. Treatment with either THZ1 or vehicle was started in the following week as indicated. BM was analyzed again after the end of the 4-week treatment period.Overall percentage of human (huCD45+) leukemic cells in mice before and at the end of the 4-week treatment period. Individual mice and mean engraftment are shown. Unpaired t-test. ***P < 0.0005. N = 10 mice for each group. Left: Representative FACS plots showing the typical CD34lowGPR56high profile of sample 04H112 before injection and after the 4-week treatment with THZ1. Right: both LSC compartments, the CD34-GPR56+ and the CD34+GPR56+ fractions were significantly reduced in vivo in the THZ1 treatment group. Individual mice and mean engraftment are shown. Unpaired t-test. ***P < 0.0005, **P < 0.005, *P < 0.05. N = 10 mice for each group. Left: Overall percentage of human (huCD45+) leukemic cells and the fraction of CD34+GPR56+ among human cells (mean, individual mice) two weeks after drug withdrawal. Right: FACS plot visualizing that the CD34+GPR56+ is re-established upon drug removal. Unpaired t-test, ***P < 0.0005, **P < 0.005, *P < 0.05.Correlation plot showing an anti-correlation between the percentage of CD34+GPR56+ cells in AML samples and the corresponding IC50s for THZ1 determined in the respective samples. Pearson correlation. Each dot represents one sample. See Dataset EV12 for sample characteristics.FACS histogram plot (left) and summary bar graph from three replicate wells (right) showing reduction of GPR56 surface expression on the GPR56-positive AML cell line HEL only in conditions that contain the more specific CDK7 inhibitor YKL-5-124, but not those that contain only the specific CDK12/13 inhibitor THZ531. Unpaired t-test, bars, and error bars represent mean and SD of three individual treatments, ***P < 0.0005, **P < 0.005, *P < 0.05.SRF reporter assay showing dose-dependent suppression of the GPR56-CTF-induced SRF signal by the CDK7 inhibitor YKL-5-124, but not by the CDK12/13 inhibitor THZ531. Four technical replicates. One of the four individual experiments. Unpaired t-test, bars and error bars represent mean and SD, ***P < 0.0005, **P < 0.005, *P < 0.05.

Fig 5: GPR56 modulates Wnt and Hh signaling pathways Cartoon illustrating at which levels GPR56 may enhance canonical Wnt signaling (turquoise dashed arrows): at the level of ligand-receptor/co-receptor interaction, ß-catenin, downstream of ß-catenin, indirectly via activation of Rho/SRF through G?12/13. CHIR99021 is a GSK3 inhibitor that prevents degradation of ß-catenin thus enhancing Wnt signaling.Cartoon illustrating GPR56 full length (FL, left) and the C-terminal fragment (CTF, right). Amino acid changes introduced in the intracellular loops are indicated by circles and letters. GAIN: GPCR autoproteolysis-inducing domain, GPS: GPCR proteolytic site, PLL: Pentraxin/Laminin/neurexin/sex-hormone-binding-globulin-Like (Salzman et al, 2016).SuperTop reporter assay showing fold change of relative luminescence units (RLU) in Wnt3a conditioned culture media normalized to empty pcDNA3.1+ in the respective control (Ctrl) media after transfection of HEK293T cells with empty vector (violet), GPR56 FL (light turquoise), and GPR56 CTF (turquoise). Shown are means, SD, and individual values of four technical replicates performed in either complete media or 1:3 diluted Wnt3a conditioned media. Unpaired t-test, ****P < 0.0001, ***P < 0.0005, **P < 0.005.ß-catenin protein expression in K562 cells infected with shGPR56strong or shLuc negative control in presence and absence of Wnt3a (representative Western Blot (above) and quantification after normalization to GAPDH (below)). Shown is fold-change compared with shLuc in control media, bars and error bars represent mean and SD of three biological replicates, unpaired t-test.SuperTop reporter assay showing fold-change of RLU in presence of Wnt3a normalized to a scrambled shRNA (shSCR) in control media after transfection with shSCR, shRNAs against GPR56 with or without additional transfection with 1 ng or 2 ng of ß-catenin plasmid. Overexpression of ß-catenin fully rescues the SuperTop signal reduction caused by shRNAs against GPR56. Unpaired t-test, bars, and error bars represent mean and SD of four biological replicates, ****P < 0.0001, ***P < 0.0005, **P < 0.005, *P < 0.05.SuperTop reporter assay indicating fold change in RLU normalized to control media after transfection of HEK293T cells with empty vector, GPR56-FL, or 5 GPR56-CTF mutants. Unpaired t-test, bars, and error bars represent mean and SD of four biological replicates, ****P < 0.0001, **P < 0.005. Left: SMO-GFP reporter RPE cell line or RPE-1 wild-type cells were infected with shGPR56strong or shLuc control followed by administration of doxycycline or the SMO agonist SAG as indicated. GPR56 KD significantly reduces the baseline and doxycycline-induced SMO-GFP signal. Right: representative FACS histogram plots showing GFP intensity in shLuc versus shGPR56strong RPE cells with and without doxycycline. Unpaired t-test, bars, and error bars represent mean and SD of three biological replicates, ***P < 0.0005, **P < 0.005, *P < 0.05. Left: Representative immunofluorescence (IF) images showing primary cilium formation in RPE cells upon starvation after infection with shLuc (above) or shGPR56strong (below). The non-motile primary cilium is visualized using antibodies against ?-tubulin, which stains the basal body, and the ciliary membrane marker ARL13B. DAPI indicates the nucleus. Images on the right show selected cells at 3× magnification. Brightness was increased in all images to enhance visibility of cilia. Right: Percentage of RPE-1 cells with primary cilium in shLuc vs. shGPR56strong RPE-1 cells at different time points after the start of serum starvation. Unpaired t-test, bars, and error bars represent mean and SD of three biological replicate wells. ****P < 0.0001, ***P < 0.0005, **P < 0.005, *P < 0.05.