Increased baseline RASGRP1 signals enhance stem cell fitness during native hematopoiesis

Abstract

Oncogenic mutations in RAS genes, like KRAS^G12D or NRAS^G12D, trap Ras in the active state and cause myeloproliferative disorder and T cell leukemia (T-ALL) when induced in the bone marrow via Mx1CRE. The RAS exchange factor RASGRP1 is frequently overexpressed in T-ALL patients. In T-ALL cell lines overexpression of RASGRP1 increases flux through the RASGTP/RasGDP cycle. Here we expanded RASGRP1 expression surveys in pediatric T-ALL and generated a RoLoRiG mouse model crossed to Mx1CRE to determine the consequences of induced RASGRP1 overexpression in primary hematopoietic cells. RASGRP1-overexpressing, GFP-positive cells outcompeted wild type cells and dominated the peripheral blood compartment over time. RASGRP1 overexpression bestows gain-of-function colony formation properties to bone marrow progenitors in medium containing limited growth factors. RASGRP1 overexpression enhances baseline mTOR-S6 signaling in the bone marrow, but not in vitro cytokine-induced signals. In agreement with these mechanistic findings, hRASGRP1-ires-EGFP enhances fitness of stem- and progenitor- cells, but only in the context of native hematopoiesis. RASGRP1 overexpression is distinct from KRAS^G12D or NRAS^G12D, does not cause acute leukemia on its own, and leukemia virus insertion frequencies predict that RASGRP1 overexpression can effectively cooperate with lesions in many other genes to cause acute T-ALL.

Figure 1:. T-ALL patient analysis and mouse models.

(A) Regularized log (rLog, normalized by DESeq2 R package) gene expression values of RASGRP1 expression in 264 pediatric T-ALL patients. (B) Schematic of six subtypes in T-ALL. (C) rLog values of RASGRP1 expression, plotted in the distinct LMO2/LYL1-, HOXA-, TLX3-, TLX1-, NKX2–1-, and TAL1- nodes. (D) Schematic representation of possible signaling networks in the BM of normal, oncogenic Kras^G12D and hRASGRP1 overexpressing mice.

Figure 2:. In vivo accumulation of hRASGRP1-overexpressing cells in RoLoRiG mouse model

(A) Schematic representation of the RoLoRiG model with overexpression of hRASGRP1-ires-EGFP. hRASGRP1 stands for human RASGRP1. hRASGRP1-ires-EGFP construct was knocked into the Rosa26 locus. A stop codon is flanked by LoxP sites, which is excised when CRE recombinase is induced by injection of polyinosinic-polycytidylic acid (pIpC). (B) Representative Western Blot showing expression of hRASGRP1 detected with a human-specific monoclonal antibody in lymph node protein lysates of RoLoRiG^+/+ Mx1CRE⁺ mice, three months following pIpC injection. The Jurkat T-cell line was used as a positive control for hRASGRP1, TBP as loading control. (C) Percentages of GFP⁺ cells in peripheral blood of RoLoRiG^+/+ Mx1CRE⁺ mice on days 21, 62, and 94 after pIpC injection. Data are presented as mean ± SEM. Each symbol represents a single mouse in all figures (n>4 in Figure 2C). (D, E, F) GFP⁺ cells within of T-, myeloid-, and B- cells in RoLoRiG^+/+Mx1CRE⁺ mice with or without pIpC injection on days 21, 62 and 94. (G) Splenic weights from RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ mice (n>7). Figures 2G–2J are at 94 days post pIpC injection. (H) Percentage of T-, myeloid-, and B- cells in spleens of RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ mice (n=6 or more). Increased percentages of CD11b⁺ cells were accompanied with decreases in B220⁺ B cells and CD3⁺ T cells. (I) Percentages of GFP⁺ cells in spleens of RoLoRiG^+/+Mx1CRE⁺ mice with or without pIpC injection (n=6 or more). (J) Percentages of GFP⁺ cells in BM of RoLoRiG^+/+Mx1CRE⁺ mice with or without pIpC injection (n>6).

Figure 3:. hRASGRP1 overexpression in a bone marrow transfer model.

(A) RoLoRiG^+/+Mx1CRE^- or RoLoRiG^+/+Mx1CRE⁺ mice were injected with pIpC at 4 weeks of age. BM cells were freshly isolated (7–8 weeks) from donor mice (CD45.2), which were either RoLoRiG^+/+Mx1CRE^- or RoLoRiG^+/+Mx1CRE⁺. Sorted 10 000 LSK cells were co-injected i.v. with 5×10⁵ carrier cells (CD45.1) into lethally irradiated recipients (CD45.1/CD45.2). Recipient mice were bled every 4 weeks until the final analysis at the endpoint (22 weeks). BM: bone marrow; LSK: Lin⁻Sca1⁺c-kit⁺; PB: Peripheral blood; pIpC: polyinosinic-polycytidylic acid. Data are presented as mean ± SEM of n=2 independent experiments. Congenic alleles allow for accurate enumeration of cell populations that are donor-derived (CD45.2) or are radiation-resistant host cells (CD45.1/CD45.2) (Supplemental Figure S3A). (B) Spleen weight from RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ mice at endpoint (n>9). Each dot represents a single mouse. **p<0.01. (C) Percentage of T-, myeloid and B-cells in spleen of RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ mice at endpoint (n>9). **p<0.01. (D-G) Analyses of percentages of LSK, PreGM, MEP, GMP and CMP within PreGM, and CLP in the BM. (n=9 or more). For additional subsets, see Supplemental Figure S4.

Figure 4:. Cytokine-induced bone marrow colony formation and signaling.

(A) 50 000 total BM cells from either RoLoRiG^+/+Mx1CRE^- or RoLoRiG^+/+Mx1CRE⁺ mice were seeded in cytokine-rich (M3434) medium. Representative images of the different colonies are shown after 7 days of incubation (100 μm Scale bar). Original magnification 10X. CFU: clonogenic forming units. For cytokine-induced signaling assays, cells were sorted, rested, stimulated, and processed for phosphor-antibody staining. (B-C) CFU potential of RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ cells grown in cytokine-rich medium (n>15). (B) Clonogenic (CFU) potential of RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ cells grown in cytokine-rich medium (n>15). (C) As in Figure 4B but in the presence of inhibitors of Ras kinase effector pathways: GDC0941 (0.1μM), U0126 (5μM) and Rapamycin (0.1μM) (n>3). (D) Analysis of p-STAT3 (Y705) and p-STAT5 (Y694) levels in rested and GM-CSF (40ng/ml) stimulated, sorted CD11B⁺ cells from RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ mice. N = 4 independent experiments. ***p<0.001. Unstimulated RoLoRiG^+/+Mx1CRE^- samples were arbitrarily set at 100. For isotype control antibody staining see Supplemental Figure S5. (E) As in Figure 4D, assessment of p-AKT (S473), p-ERK (T202/Y204), and p-S6 (S235/S236) levels. N = 4 independent experiments. ***p<0.001, *p<0.05.

Figure 5:. RasGRP1 overexpression drives spontaneous bone marrow colony formation and S6 signaling.

(A) BM CFU analysis in cytokine-poor (M3231) medium. supplemented with either GM-CSF (GM-CFU) or IL-7 (PreB-CFU). Representative images following 7 days of incubation (100 μm Scale bar). Original magnification 10X. Schematic of baseline signaling assays in whole BM. (B) Clonogenic (CFU) potential of RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ cells grown in M3231 with 0.01 ng/ml of GM-CSF and inhibitors: GDC0941 (0.1μM), U0126 (5μM) and Rapamycin (0.1μM) (n>4). (C) Clonogenic (PreB-CFU) potential of RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ cells using 0.1, 1 and 10 ng/ml of IL-7 (n>3). (D) Analysis of baseline p-STAT3 (Y705), p-STAT5 (Y694), p-AKT (S473), p-ERK (T202/Y204), and p-S6 (S235/S236) signals in RoLoRiG^+/+Mx1CRE^- and RoLoRiG^+/+Mx1CRE⁺ total BM, immediately fixed after isolation. Mean fluorescence intensity is depicted by the numerical value. Data is representative of four independent experiments.

Figure 6:. Increased fitness in Native hematopoiesis in RoLoRiG^+/+ mice.

(A) Percentage of LSK in the BM of RoLoRiG^+/+Mx1CRE^- or RoLoRiG^+/+Mx1CRE⁺ mice (n>18). In panel 6A and all subsequent panels in figure 6, data are presented as mean ± SEM. Each symbol represents a single mouse. *p<0.05, **p<0.01, ***p<0.001. (B-F) Percentage of specific developmental stages in the BM RoLoRiG^+/+Mx1CRE^- or RoLoRiG^+/+Mx1CRE⁺ mice, analogous to the developmental gating scheme used in Figure 3. (G) Inclusion of the ires-EGFP cassette allows for specific assessment of the GFP⁺ fraction within each of the hematopoietic subsets within the BM. (H-K) Proportions of GFP⁺ cells within in RoLoRiG^+/+Mx1CRE⁺ mice with or without pIpC injection. (H. LSK; I. MPP, ST-HSC and LT-HSC cells; J. CLP cells; K. MEP, GMP and CMP cells.

Figure 7:. Co-insertion table of SL3-3 leukemia virus insertions.

SL3–3 leukemia virus insertions that resulted in T-ALL in mice were mapped against each other to understand the relationship between the top 8 SL3–3 insertion sites.