STAT5BN642H is a driver mutation for T cell neoplasia

Abstract

STAT5B is often mutated in hematopoietic malignancies. The most frequent STAT5B mutation, Asp642His (N642H), has been found in over 90 leukemia and lymphoma patients. Here, we used the Vav1 promoter to generate transgenic mouse models that expressed either human STAT5B or STAT5BN642H in the hematopoietic compartment. While STAT5B-expressing mice lacked a hematopoietic phenotype, the STAT5BN642H-expressing mice rapidly developed T cell neoplasms. Neoplasia manifested as transplantable CD8+ lymphoma or leukemia, indicating that the STAT5BN642H mutation drives cancer development. Persistent and enhanced levels of STAT5BN642H tyrosine phosphorylation in transformed CD8+ T cells led to profound changes in gene expression that were accompanied by alterations in DNA methylation at potential histone methyltransferase EZH2-binding sites. Aurora kinase genes were enriched in STAT5BN642H-expressing CD8+ T cells, which were exquisitely sensitive to JAK and Aurora kinase inhibitors. Together, our data suggest that JAK and Aurora kinase inhibitors should be further explored as potential therapeutics for lymphoma and leukemia patients with the STAT5BN642H mutation who respond poorly to conventional chemotherapy.

Keywords: Hematology; Leukemias; Lymphomas; Oncology; T cells.

Figure 1. hSTAT5B^N642H is an activating mutation.

(A) Schematic of STAT5B mutations identified in leukemia and lymphoma patients. Each dot represents 1 patient. (B) WB analysis of pY-STAT5, total STAT5 protein, and HSC70 in 293T cells that were transiently transfected with different hSTAT5B (hS5B) variants using a pMSCV-IRES-GFP vector, with or without growth hormone (GH) stimulation. (C) WB analysis of pY-STAT5, STAT5, FLAG, and HSC70 in hSTAT5B- or hSTAT5B^N642H-expressing (N642H) Ba/F3 cells with or without IL-3 stimulation. (B and C) Nontransfected and pMSCV-transfected cells are shown as controls. Data presented in B and C are representative of 3 independent experiments. Samples were run on parallel gels for B and C, and a loading control is provided for each gel.

Figure 2. Moderate Vav1-driven expression of hSTAT5B^N642H in mice leads to HSC expansion.

(A) Survival curve shows the percentages of disease-free survival of hSTAT5B^N642H (N642H) mice (n = 21) compared with that of hSTAT5B (hS5B) (n = 20) and WT (n = 10) mice. (B) WB analysis of pY-STAT5, total STAT5, and HSC70 in the LNs and spleens of WT mice and hSTAT5B^N642H- and hSTAT5B-transgenic mice. Quantification of the WB was performed using ImageJ. Data are representative of 3 independent experiments. (C) Flow cytometric analysis of the percentage of LSKs, LT-HSCs (CD150⁺CD48^–), ST-HSCs (CD150⁺CD48⁺), MPPs (CD150^–CD48⁺), (D and E) common lymphoid progenitors (lineage⁻Sca1⁺IL-7R⁺AA4⁺), MPCs (lineage⁻Sca1^–IL-7R^–c-Kit⁺), and CD3⁺ cells in the BM of WT, hSTAT5B, and hSTAT5B^N642H mice. Analyses in C–E included 7-week-old WT (n = 7), hSTAT5B (n = 5), and hSTAT5B^N642H (n = 5) mice. Data represent the mean ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001, by 1-way ANOVA with Bonferroni’s correction.

Figure 3. hSTAT5B^N642H mice suffer from an aggressive CD8⁺ T cell lymphoma.

(A) Macroscopic comparison of hSTAT5B^N642H and hSTAT5B mouse spleens and LNs with those from WT mice. Scale bars: 1 cm. (B) Modified Wright staining of blood smears from hSTAT5B^N642H (N642H), hSTAT5B (hS5B), and WT mice (original magnification, ×100). (C) WBC count using an animal blood counter (scil Vet ABC). CD8/CD4 ratios in the peripheral blood were determined using flow cytometry. Analysis included 7- to 10-week-old WT (n = 20), hSTAT5B (n = 15), and hSTAT5B^N642H (n = 20) mice. (D) CD8/CD4 T cell ratios in LNs were determined using flow cytometry. Analyses included 7-week-old WT (n = 5), hSTAT5B (n = 5), and hSTAT5B^N642H (n = 5) mice. (E) Quantification of the absolute number of CD4⁺ and CD8⁺ T cells, myeloid cells (CD11b⁺Gr1⁺), and B cells (CD19⁺) in spleens from hSTAT5B^N642H- and hSTAT5B-transgenic mice and WT mice. Analyses included 7-week-old WT (n = 13), hSTAT5B (n = 6), and hSTAT5B^N642H (n = 6 and 11) mice. (F) CD3⁺CD8⁺ splenic cells were analyzed by flow cytometry for their expression of CD25. Analyses included 8-week-old WT (n = 8), hSTAT5B (n = 9), and (n = 6) hSTAT5B^N642H mice. (G) CD3⁺CD8⁺ splenic cells were further analyzed for CD62L and CD44 expression. Analyses included WT (n = 8), hSTAT5B (n = 5), and hSTAT5B^N642H (n = 5) mice at 8 weeks of age. Data represent the mean ± SD. n ≥ 6. **P < 0.01 and ***P < 0.001, by 1-way ANOVA with Bonferroni’s correction.

Figure 4. Highly proliferative T cells infiltrate into the peripheral organs of hSTAT5B^N642H mice.

Histological analysis using H&E, CD3, and Ki67 staining of the lungs of 8- to 10-week-old hSTAT5B, hSTAT5B^N642H, and WT mice. Data are a representative of 3 independent experiments. Scale bar: 100 μm. Original magnification, ×20 and ×40 (insets).

Figure 5. hSTAT5B^N642H CD8⁺ T cells are the cancer-initiating cells.

(A) Percentage of disease-free survival following hSTAT5B^N642H whole BMT into 8-week-old NSG recipient mice compared with WT BMT. (B) Macroscopic view of LNs and spleen from a hSTAT5B^N642H BMT recipient mouse compared with those from a WT BMT recipient mouse. Scale bar: 1 cm. (C) Flow cytometric analysis shows the quantity of CD3⁺ cells and CD8/CD4 T cell ratio in the spleens of BMT recipient mice. (D) Histological analysis of CD3⁺ cells from the lungs of NSG recipient mice after hSTAT5B^N642H or WT BMT. Scale bar: 100 μm. Original magnification, ×20 and ×40 (insets). (E) Percentage of disease-free survival after hSTAT5B^N642H or WT CD8⁺ T cell transplantation into nonirradiated 8-week-old Ly5.1/CD45.1 recipient mice. (F) Flow cytometric analysis shows the quantity of splenic CD3⁺CD8⁺ cells in CD8⁺ T cell–transplanted mice. (G) Spleen versus BW ratios of WT and hSTAT5B^N642H CD8⁺ T cell–transplanted Ly5.1/CD45.1 mice. (A–C) n = 4 WT mice and n = 5 hSTAT5B^N642H mice; (E–G) n = 6 WT mice and n = 5 hSTAT5B^N642H mice. Data represent the mean ± SD. *P < 0.05, **P < 0.01, and ***P < 0.001, by unpaired, 2-tailed Student’s t test.

Figure 6. hSTAT5B^N642H-driven diseased T cells can be treated with JAK inhibitors.

(A) WB analysis of pY-STAT5 levels in isolated and cultivated LN T cells from hSTAT5B^N642H, hSTAT5B, and WT mice after IL-2 removal. (B) Dose-response curve of WT, hSTAT5B^N642H, and hSTAT5B T cells 72 hours after ruxolitinib treatment, analyzed using CellTiter-Glo (CTG) assay. IC₅₀ values were determined using GraphPad Prism. Error bars indicate the mean ± SEM. DMSO (100% viability) and 10 μM bortezomib (0% viability) on each plate served as controls. (C) WB of hSTAT5B^N642H, hSTAT5B, and WT T cell cultures after 5 hours of treatment with ruxolitinib or tofacitinib, analyzed for pY-STAT5. (D) Macroscopic view of LNs and spleens from CD8⁺ T cell–transplanted mice treated with ruxolitinib compared with vehicle controls. CD8⁺ T cell–recipient mice were treated with ruxolitinib at the dosage of 45 mg/kg twice a day for 30 days. (E) Quantification of spleen versus BW ratio of vehicle- and ruxolitinib-treated hSTAT5B^N642H CD8⁺ cell–transplanted mice. (F) WBC counts of vehicle- and ruxolitinib-treated hSTAT5B^N642H CD8⁺ cell–transplanted mice, measured using a scil Vet ABC animal blood counter. Flow cytometric analysis of CD25 expression in peripheral blood CD8⁺ T cells. MFI, mean fluorescence intensity. (G) Histological analysis of CD3⁺ cells in the lungs of recipient mice after treatment with ruxolitinib compared with the vehicle-treated group. Scale bar: 100 μm. Original magnification, ×20 and ×40 (insets). n = 5 vehicle-treated mice and n = 4 ruxolitinib-treated mice. Data represent the mean ± SD. n ≥ 6. *P < 0.05, by unpaired, 2-tailed Student’s t test. Data presented in A–C are representative of 3 independent experiments.

Figure 7. hSTAT5B^N642H provokes substantial changes in gene expression, accompanied by specific changes in DNA methylation of CD8⁺ T cells.

(A) Heatmap showing Z scores of rlog-transformed and library size–normalized counts of genes upregulated (red) or downregulated (blue) in hSTAT5B or hSTAT5B^N642H and WT CD8⁺ T cells (FDR-adjusted P < 0.05). Analyses included 13-week-old WT (n = 5), hSTAT5B (n = 4), and hSTAT5B^N642H (n = 5) mice. Each column in the heatmap represents data from CD8⁺ T cells from 1 mouse of a given genotype, and each row represents data for a given gene. (B) Enrichment blot of the CD8⁺ T cell lymphoma expression signature. Barcode blot indicates the position of the gene in the gene set. Red and blue colors represent, respectively, positive and negative Pearson’s correlations with hSTAT5B^N642H CD8⁺ T cells. The gene set was obtained from published gene signature cytotoxic T cells (43, 44). (C) Top enriched gene sets are the results of GSEA including E2F target, G₂M checkpoint, MYC target, and cell-cycle progression in hSTAT5B^N642H CD8⁺ T cells. P values in B and C were determined by Kolmogorov-Smirnov test. (D) Scatterplot contrasting the mean DNA methylation levels in WT and hSTAT5B^N642H-mutant T cells in all CGIs covered in at least 1 sample per genotype (n = 15,209). The density of data points in each plot region is indicated by color intensity, and CGIs with lower DNA methylation in WT (n = 770) or hSTAT5B^N642H (n = 610) cells are indicated by black and red crosses, respectively (absolute difference ≥5 percentage points, n = 2 per genotype). Analyses included 13-week-old mice. NES, normalized enrichment score.

Figure 8. hSTAT5B^N642H-driven DNA methylation changes accompanied by enhanced DNA-binding activity of STAT5 result in the induction of Aurora kinase B.

(A) Region set enrichment analysis testing CGIs with lower DNA methylation in hSTAT5B^N642H cells than in WT cells (top) or lower DNA methylation in WT cells than in hSTAT5B^N642H cells (bottom). Enrichment was determined using LOLA (51). Each dot represents 1 ChIP-seq experiment for a given transcription factor from the CODEX database. The vertical dashed line represents the significance threshold (FDR-adjusted P ≤ 0.05). (B) Enrichment blot of EZH2 target genes in HSCs, together with their methylation states of EZH2-bound and EZH2-unbound CGIs 100 kb up- and downstream of the transcriptional start sites (TSSs). Barcode blot indicates the position of the gene in the gene set. Red and blue colors represent, respectively, positive and negative Pearson’s correlations with hSTAT5B^N642H CD8⁺ T cells. The gene set was obtained from the MSigDB (72). Black circles indicate CGIs overlapping with EZH2-binding sites. p.p., percentage points. n = 2 per genotype. ChIP with anti-STAT5 (C) or anti-EZH2 (D) in CD8⁺ T cells isolated from WT (n = 7), hSTAT5B (n = 7), or hSTAT5B^N642H (n = 4) mice. Binding of STAT5 to the Cis and Ccnd2 promoters or binding of EZH2 to the promoter regions of Cdkn2A and Ccnd2 served as positive controls. Horizontal dotted line indicates the threshold for nonspecific binding. (E) ChIP with anti-STAT5, anti-EZH2, or IgG in STAT5B^N642H-expressing CD8⁺ T cells, followed by WB analysis. IB, immunoblot. Data presented in C–E are representative of 2 independent experiments. Error bars indicate the mean ± SD.

Figure 9. hSTAT5B^N642H-driven diseased T cells are sensitive to Aurora kinase B inhibition.

(A) WB analysis of p-AURKB, total AURKB, and HSC70 in LNs from WT and hSTAT5B^N642H- and hSTAT5B-transgenic mice. WB quantification (bar graph) was performed using ImageJ. (B) Dose-response curves of WT, hSTAT5B^N642H, or hSTAT5B T cells in response to AT9283 after 72 hours of treatment, analyzed using a CTG assay. IC₅₀ values were determined using GraphPad Prism. Error bars indicate the mean ± SEM. DMSO (100% viability) and 10 μM bortezomib (0% viability) on each plate served as controls. n = 6 per genotype. (C and D) WB of hSTAT5B^N642H, hSTAT5B, and WT T cell cultures after 5 hours of treatment with AT9283 or AZD1152, for determination of pY-STAT5, total STAT5, p-AURKB, total AURKB, p-H3 (Ser10), and total H3 levels. Data presented in A–D are representative of 3 independent experiments.