CREB is a critical regulator of normal hematopoiesis and leukemogenesis

Abstract

The cAMP-responsive element binding protein (CREB) is a 43-kDa nuclear transcription factor that regulates cell growth, memory, and glucose homeostasis. We showed previously that CREB is amplified in myeloid leukemia blasts and expressed at higher levels in leukemia stem cells from patients with myeloid leukemia. CREB transgenic mice develop myeloproliferative disease after 1 year, but not leukemia, suggesting that CREB contributes to but is not sufficient for leukemogenesis. Here, we show that CREB is most highly expressed in lineage negative hematopoietic stem cells (HSCs). To understand the role of CREB in hematopoietic progenitors and leukemia cells, we examined the effects of RNA interference (RNAi) to knock down CREB expression in vitro and in vivo. Transduction of primary HSCs or myeloid leukemia cells with lentiviral CREB shRNAs resulted in decreased proliferation of stem cells, cell- cycle abnormalities, and inhibition of CREB transcription. Mice that received transplants of bone marrow transduced with CREB shRNA had decreased committed progenitors compared with control mice. Mice injected with Ba/F3 cells expressing either Bcr-Abl wild-type or T315I mutation with CREB shRNA had delayed leukemic infiltration by bioluminescence imaging and prolonged median survival. Our results suggest that CREB is critical for normal myelopoiesis and leukemia cell proliferation.

Figure 1

CREB expression in mouse and human hematopoietic progenitors. (A) CREB expression levels in lin⁻ and lin⁺ cells as measured by q-PCR. CREB expression was higher (2.6-fold) in the lin⁻ population than the lin⁺ population (P = .01). (B) Gating strategy for cell sorting of lin⁻ cells to isolate HSC (Ba), CMP (Bb), GMP (Bc), MEP (Bd), CLP (Be), LT-HSC (Bf), ST-HSC (Bg), and MPP (Bh). (C) cDNA from HSC CMPs, CLPs, GMPs, and MEPs were confirmed by expression of SCL, c-Mpl, Pu-1, Aiolos, and CEBPα. (D) CREB expression levels in HSC, CMP, GMP, MEP, and differentiated cells. Compared with mature cells, CREB was 2.5- to 4-fold higher in HSC, CMP, GMP, and MEP. (E) CREB expression in LT-HSC, ST-HSC, and MPP. CREB was slightly higher (4- vs 3-fold) in the MPP fraction compared with the LT-HSC and ST-HSC populations, but this was not found to be statistically significant (P = .06). CREB expression in human lin⁻ and lin⁺ cells isolated from cord blood (F) or peripheral blood (G) stem cells. CREB expression was higher for both cell types in the lin⁻ population than the lin⁺ population (P = .01). (H and I) CREB expression in human bone marrow cells. CREB was also expressed at higher levels in CD34⁺ CD38⁻ cells (b) than CD34⁺ CD38⁺ cells (a). All experiments were performed in triplicate. Error bars represent SE.

Figure 2

CREB is critical for normal myelopoiesis in vitro. (A) (i) Western blot analysis demonstrating knockdown of CREB approaching 80% compared with control cells. (ii) Total numbers of CFU-GM colonies after 21 days in methylcellulose for murine hematopoietic cells. (iii) Flow cytometric analysis of murine bone marrow transduced with CREB shRNA or control lentivirus and sorted for GFP⁺ fraction, cultured in methylcellulose over 21 days. CREB knockdown cells had a lower fraction of mature granulocyte and monocytes compared with control cells. (B) (i) Western blot analysis demonstrating knockdown of CREB up to 65% compared with control cells. (ii) Total number of CFU-GM colonies after 21 days in methylcellulose for human peripheral blood stem cells. (iii) Flow cytometry analysis of transduced CD34⁺ human peripheral blood stem cells cultured in methylcellulose over 21 days. All experiments were performed in triplicate. Error bars in Aii,Bii represent SE.

Figure 3

CREB shRNA induces apoptosis in HSCs and bone marrow progenitor cells. (A) Transduced human peripheral blood cells plated in methylcellulose for 3 weeks and stained with monoclonal antibodies for CD34 and CD38 expression. Cells were stained with PI to assess cell death. (B) Murine bone marrow cells were transduced at a density of 10⁶ cells/mL with lentivirus expressing CREB, scrambled, or vector shRNA at a multiplicity of infection (MOI) of 100. After 2 days of culturing in media containing cytokines (mIL-3, 10ng/mL; mSCF, 25 ng/mL; and hIL-6, 10 ng/mL), cells were sorted using flow cytometry for GFP expression. Sorted calls were cultured in cytokine containing media for 5 days and stained for annexin-V and PI. All experiments were performed in triplicate. (C) Western blot analysis with lysates from mouse BM cells (10⁶) transduced with CREB shRNA, scrambled, luciferase, and vector control lentivirus. Immunoblots were probed with anti-PARP or β-tubulin antisera. Error bars in panels A and B represent SE.

Figure 4

CREB is critical for myelopoiesis in vivo. Bone marrow from CD45.1 mice transplanted into CD45.2 mice were analyzed with lineage-specific and hematopoietic stem-cell markers using FACS analysis at 5 and 12 weeks after transplantation. (A) Myeloid engraftment as measured by staining of bone marrow cells from transplant-recipient mice at 5 weeks. (B) Myeloid engraftment as measured by FACs staining of bone marrow cells at 12 weeks. At least 5 mice in each group were analyzed. Experiments were performed in triplicate and repeated twice. Error bars represent SE.

Figure 5

CREB is essential for leukemia cell proliferation and survival. (A) Human (K562, TF-1) leukemia cells were transduced with a lentivirus expressing no shRNA, CREB shRNA-1, CREB shRNA-2, or luciferase shRNA at a multiplicity of infection (MOI) of approximately 100. Wild-type cells were also used as a control. Western blot analyses were performed with CREB, phospho-CREB, and β-tubulin antisera. (B) Five micrograms of total RNA were extracted from transduced leukemia cells, and q-PCR was performed to determine CREB expression. CREB was knocked down by up to 75% relative to control shRNA (vector) in human myeloid leukemia cells. (C) Trypan blue exclusion method was performed in triplicate to assess growth and survival of transduced leukemia cells. CREB knocked-down cells demonstrated diminished proliferation and viability 72 hours after transduction. (D) K562 cells were transduced and cultured for 48 hours before harvesting total RNA. Parental K562 cells were cultured in the presence of interferon-2α (100 units/mL) for 48 hours as a positive control. Quantitative reverse transcription-PCR was performed in triplicate with primers specific for CREB, actin, and OAS-1. (E) Luciferase reporter assays in human TF-1 leukemia cells transduced with CREB or control shRNAs. Decreased transcriptional activity was observed in CREB knocked-down cells and repeated in triplicate. (F) Cell-cycle analysis of CREB knocked-down TF-1 cells after synchronization by serum starvation overnight and stimulated for 12 hours with GM-CSF revealed decreased percentage of cells in S-phase. Experiment was performed in triplicate. Error bars in panels B-F represent SE.

Figure 6

CREB inhibits progression of leukemia in vivo. (A) Western blot analysis with CREB and tubulin antisera, demonstrating 2-fold increase in expression of CREB in T315I mutant of Bcr-Abl in murine pro-B lymphocyte line (Ba/F3) compared with wild-type Ba/F3 cells. (B) Western blot analysis after lentiviral transduction with CREB shRNAs demonstrating 90% inhibition. (C) Quantitative reverse transcription-PCR showing diminished CREB mRNA levels in transduced Ba/F3 T315I cells. (D) Trypan blue exclusion method performed in triplicate shows diminished growth after transduction with CREB shRNA compared with empty vector. (E) Bioluminescence imaging of SCID mice injected with 10⁶ cells transduced with CREB shRNA or CREB scrambled shRNA lentivirus. Mice were imaged at days 7 and 14. Tumor burden is lower in CREB shRNA–injected mice. (F) Kaplan-Meier survival analysis of mice injected with 5 × 10⁵ cells showing longer survival with CREB knockdown (n = 9) compared with scrambled shRNA (n = 9). All deaths were due to leukemia, except for a day 7 handling event in the scrambled group treated as a censored observation. Error bars in panels C and D represent SE.

Figure 7

Expression of cyclins A and D in leukemia cells. TF-1 (A) or K562 (B) myeloid leukemia cells were transduced with control and CREB shRNA lentivirus and synchronized. At 12 hours, 5 μg of total RNA was isolated for q-PCR by SyberGreen method. Cyclin A1- and D1-specific primers were used, and expression was normalized to the house keeping gene β-actin. Experiments were performed in triplicate. Error bars represent SE.