CAMKs support development of acute myeloid leukemia

C57 BL/6 and NOD/SCID-IL2RG (NSG) mice were purchased from the UT Southwestern Medical Center animal breeding core facility or Jackson Labs. The PirB TM mice [14] and CaMKIV-null mice [10] were purchased from MMRRC and Jackson Labs, respectively. Mice were maintained at the UT Southwestern Medical Center animal facility. All animal experiments were performed with the approval of UT Southwestern Committee on Animal Care and University of Missouri Committee on Animal Care.

Antibodies for flow cytometry, anti-Kit-APC, anti-Mac-1-APC, anti-Gr-1-PE, anti-CD3-APC, and anti-B220-PE, were purchased from BioLegend and used as described [1]. The manufacturers and catalog numbers for other antibodies and reagents are as follows: anti-PirB-PE, R&D Systems, FAB2754P; pCAMKI, Santa Cruz, sc-28438; anti-pCAMKII, Abcam, ab32678; anti-pCAMKIV, Santa Cruz Biotechnology, sc-28443-R; anti-CAMKI, Abcam, ab68234; anti-CAMKII, Cell Signaling, 4436; anti-CAMKIV, Cell Signaling, 4032; anti-pCREB, Cell Signaling, 9198S; anti-CREB, Cell Signaling, 9197S; anti-actin, Sigma Aldrich, A2066; STO-609, Sigma Aldrich, S1318; KN93, Sigma Aldrich, K1385; The PCR primer sequences were as follows: hCAMKI forward: CGGAGGACA TTAGAGACA, reverse: CTCGTCATAGAAGGGAGG-3; hCAMKIV forward: GATGAAAGAGGCGATCAG, reverse: TAGGCCCTCCTCTAGTTC. PirB forward: GAG AATCACCAGACACATGC, PirB reverse: CTGCCCTCATGTCTTAACTT, mCAMKIV forward: AAGCAGGCGGAAGACATTAGG, CAMKIV reverse: AGTTTCTGAGTCCTCTTGTCCT.

For retrovirus production, human embryonic kidney 293T cells were grown in DMEM with 10% fetal bovine serum (FBS) and transfected by MSCV-MLL-AF9-IRES-YFP, MSCV-CAMKIV-IRES-GFP, MSCV-CAMKI-IRES-GFP, or MSCV-CREB-IRES-GFP encoding plasmids and pCL-ECO to produce retroviruses. The infection of Linage-negative cells with retrovirus was performed as described previously [1, 15, 16]. Briefly, we infected Lin⁻ cells with retroviral supernatant in the presence of 8 μg/mL polybrene. After infection, these cells were incubated overnight in medium with 10% FBS, 20 ng/mL SCF, 10 ng/mL IL-3, and 10 ng/mL IL-6. Infected cells (300,000) were transplanted into lethally irradiated (1000 rad) WT mice by retro-orbital injection. For secondary transplantation, YFP⁺ bone marrow (BM) cells from primary transplanted mice were sorted and transplanted into mice with 200,000 normal BM cells as competitors.

For lentivirus production, the lentiviral vector Pll3.7 was used to express shRNAs to target CAMKI (sense: TGCCAGAGAATCTGCTGTACTATTCAAGAGATAGTA CAGCAGATTCTCTGGCTTTTTTC; antisense: GAAAAAAGCCAGAGAATC TGCTGTACTATCTCTTGAATAGTACAGCAGATTCTCTGGCA) and CAMKIV (sense: TGATATTACAGTGAGCGAGATGTTCAAGAGACATCTCGCTCACTGTAATATCTTTTTTGGAAC; antisense: TCGAGTTCCAAAAAAGATATTACAGTGAGCGAGATGTCTCTTGAACATCTCGCTCACTGTAATATCA). The lentivirus plasmid, pSPAX2, and pMD2.G (4:3:1) were transfected using PolyJet™ In Vitro DNA Transfection Reagent (SignaGen Laboratories) into 293T cells. The resulting virus supernatant was collected 48–72 h later.

For Crisper-Cas9 system, we used pCW-Cas9 (gift from Eric Lander & David Sabatini (Addgene plasmid # 50661) [17]) to make stable MV4-11-Cas9 cell line, and designed three gRNA targeting the CAMK1 (gRNA1-GCTACGACTTCCGAGATGTTC; gRNA2-GGAGCTCTTTGACCGTATTGG; gRNA3-GATCCCGGTGTACAATGCCC) and one scramble gRNA (GAACGACTAGTTAGGCGTGTA).

We performed flow cytometry, immunohistochemistry, and cytospin as described previously [1, 15, 16]. For flow cytometry analysis of AML cells, peripheral blood or BM cells were stained with anti-Mac-1-APC, anti-Gr-1-PE, anti-CD3-APC, anti-B220-PE, or anti-Kit-PE monoclonal antibodies (BioLegend). For analysis of apoptosis, indicated AML cells were stained with PE-conjugated anti-annexin V and 7-AAD (BD Pharmingen) according to the manufacturer’s instructions.

Human leukemia cell lines were cultured in RPMI supplemented with 10% FBS at 37 °C in 5% CO₂ and the normal level of O₂. Human leukemia cell lines were from ATCC. Xenografts were performed essentially as we described [11, 18]. Briefly, adult NSG mice (6–8 weeks old) were sub-lethally irradiated with 250 cGy total body irradiation prior to transplantation. ShRNA-infected cells were resuspended in 200 μl PBS containing 1% FBS at a final dose of 1 × 10⁶ human GFP⁺ viable cells per mouse for retro-orbital injection. One to 4 months after transplantation, the peripheral blood (PB), BM, spleen, and liver were assessed for engraftment by flow cytometry.

Five hundred YFP⁺Mac-1⁺Kit⁺ BM cells from AML mice were plated in colony-forming unit (CFU) medium (M3534, Stem Cell Technologies) for CFU-GM assays, according to the manufacturer’s protocols [1, 18]. After 7 days, 2000 cells from three dishes were isolated for secondary plating.

Data were obtained from the TCGA AML database (https://​tcga-data.​nci.​nih.​gov/​docs/​publications/​tcga/​; accessed September 5, 2013), and levels were normalized to GADPH expression. Patients were divided into two groups based on whether their expression was above or below the median level of a gene.

Cell lysates (100 μg samples) were separated by electrophoresis on a 4–12% SDS-polyacrylamide gel (BioRad), and the proteins were transferred onto a nitrocellulose membrane. The membrane was probed with indicated primary antibody for 1 hour at room temperature and then incubated with horseradish peroxidase-conjugated secondary antibody, which was detected with the chemiluminescence SuperSignal kit (Pierce).

Data are expressed as mean ± SEM. For all experiments except determination of survival, data were analyzed by Student’s t test and differences were considered statistically significant if p < 0.05. The survival rates of the two groups were analyzed using a log-rank test, and differences were considered statistically significant if p < 0.05.

Because LILRBs are highly expressed by monocytic AML (M5) cells [1], we used a retroviral MLL-AF9 transplantation mouse M5 AML model [19, 20] to investigate the relationship between PirB signaling and CAMKs in AML development. PirBTM mice [14], in which four exons encoding the transmembrane domain and part of the intracellular domain of PirB are deleted, were used to provide PirB-defective cells. PirBTM and control wild-type (WT) cells infected with MSCV-MLL-AF9-IRES-YFP retrovirus were transplanted to establish AML mice. We sought to determine whether the CAMK family decreased expression/activities in the PirB-defective MLL-AF9 AML mouse model. Compared to WT controls, PirBTM cells from MLL-AF9 AML mice had significantly decreased phosphorylation of CAMKI, CAMKII, and CAMKIV (Sun et al. [22]) (Fig. 1a), suggesting that CAMK activities are regulated by the PirB signaling pathway.

To test whether CAMK activities are essential to AML cells, we treated primary mouse MLL-AF9 AML cells with the CAMKK inhibitor STO-609 and CAMK inhibitor KN93. Both inhibitors diminished colonies in WT AML cells instead of PirB-deficient AML cells (Fig. 1b, c). These data show that the kinase activity of CAMK regulated by PirB signaling is critical for the growth of AML cells.

We used gain-of-function approaches to further determine the role of CAMKI and CAMKIV in PirB signaling in AML cells. We introduced retroviruses encoding CAMKI or CAMKIV into PirB-defective AML cells to study whether CAMKs could rescue PirB defects in AML development. Mice transplanted with MLL-AF9-transduced PirBTM cells developed AML more slowly and survived longer than those transplanted with PirBTM cells that overexpressed CAMKI or CAMKIV (Fig. 1d). The PirBTM leukemia cells that overexpress CAMK exhibited accelerated development as demonstrated by the twofold increase in leukemia cell infiltration in the peripheral blood. This was detected using flow cytometry of these mice, compared to mice transplanted with PirBTM cells that did not overexpress CAMK, at 28 days post-transplantation (Fig. 1e). The in vitro CFU assay also showed both CAMKI and CAMKIV can increase PirBTM AML cells’ colony-forming ability (Fig. 1h). Interestingly, in all of the rescue assays (Fig. 1d, e, h), neither the kinase-inactive CAMKI mutant (K49E) nor the CAMKIV mutant (K75M) was capable of supporting PirBTM leukemia development. We also performed the rescue assays in WT leukemia cells. Though CAMK overexpression did not affect WT leukemia development, the kinase-inactive CAMK1 mutant and CAMKIV mutant were able to play dormant negative roles in leukemia development in vivo and in vitro (Fig. 1f, g, i). These results demonstrate that CAMK, depending on their kinase activity, can rescue PirB defects in AML development, supporting our hypothesis that CAMKs are downstream mediators of PirB signaling.

To gain a deeper understanding of the mechanism by which CAMKs support AML development, we sought to examine AML development in genetic CAMK deletion model. While CAMKI and CAMKII have multiple isoforms, CAMKIV exists as a single form. The availability of the Camk4-null mice [10] allowed us to focus on the role of CAMKIV as a representative of the CAMK family in the mouse AML model. It is clear that PirB and CAMKIV are highly expressed by hematopoietic progenitors and MLL-AF9 AML stem cell-enriched Mac-1⁺Kit⁺ population [11, 20, 21] (Additional file 1: Figure S2a-c).

The mice transplanted with MLL-AF9-transduced Camk4-null cells developed AML significantly more slowly than controls in both primary and secondary transplantation (Fig. 2a). The delayed development of the Camk4-null AML was also evident from the significantly decreased liver and spleen sizes and the numbers of white blood cells in circulation in the mice transplanted with Camk4-null cells compared to controls (Fig. 2b). The infiltration of Camk4-null myeloid leukemia cells into the liver and spleen was significantly decreased (Fig. 2c). There were more mature and differentiated hematopoietic cells especially the Gr-1⁺ myeloid cells in mice that received Camk4-null cells than in those given control WT cells, based on cytospin and flow cytometry analyses (Fig. 2d–f). These results demonstrated that a lack of CAMKIV results in slower AML development.

We further analyzed whether CAMKIV affects AML stem cell (AML-SC) activity. Primary transplants with Camk4-null cells did not have significant differences in YFP⁺Mac1⁺Kit⁺ cells from mice transplanted with WT cells (Fig. 3a). In drastic contrast, the YFP⁺Mac1⁺Kit⁺ cells in secondarily transplanted mice decreased from 53 to 30% in BM and from 11 to 1% in peripheral blood (Fig. 3b). A serial plating CFU assay was performed to test the self-renewal of AML-SCs in vitro. The deficiency of CAMKIV decreased CFUs in both primary plating and secondary plating, with dramatically lower numbers of colonies in secondary plating (Fig. 3c, d). Accordingly, CAMKIV was capable of rescuing the defects of PirBTM AML cells (Fig. 1h).

We also measured the apoptosis of Camk4-null and control AML BM cells. Overall, the MLL-AF9 BM AML cells had very low levels of early and late apoptosis (1.2 and 0.1%, Fig. 3e). Levels of apoptosis were dramatically increased in the Camk4-null AML BM cells (4.6 and 0.9%, Fig. 3e). These results suggest that CAMKIV supports the survival of AML cells. This is consistent with the report that CAMKIV is a critical survival factor for HSCs [10].

To quantitate how CAMKIV deficiency affects the AML-SC frequency, we performed transplantations by limiting dilution of sorted YFP⁺ WT and Camk4-null MLL-AF9 BM cells, collected from primary recipients. Leukemia development, as measured by survival ratio and latency days, is summarized in Fig. 3f. The frequency of AML-SCs was assessed and plotted as the percentages of non-leukemia-developing mice. Strikingly, the frequency of functional AML-SCs in Camk4-null primary MLL-AF9 AML model mice was only 1/12 (= 24/295) of that in control WT AML mice. This limiting dilution transplantation assay confirms that CAMKIV supports the activity of AML-SCs.

Previously, we showed that the activation of LILRB2 induced CAMK activation. Similarly, based on our in silico analysis of human samples, the expression of LILRB2 and several CAMKs correlate inversely with the overall survival of AML patients (Fig. 4a–d), and CAMK1 is highly expressed by M5 AML cells (Fig. 4e) as LILRB2 [1]. These results suggest that CAMKs may play important roles in AML development. The activation of CAMKIV in human myeloid cells including the monoblastic KASUMI-1 cells and MLL-rearranged MV4-11 cells is greater than that in lymphoblastic leukemia cells such as KASUMI-2 and SUP-B15 cells (Additional file 1: Figure S3). The knockdown of CAMK1 or CAMK4 in AML cell lines including MV4-11 cells and KASUMI-1 led to significant inhibition of cell growth (Fig. 2g–j), indicating that CAMKs are functionally essential for human leukemia cells. By performing rescue assays (Fig. 4k), and Crisper assays (Additional file1: Figure S4) in MV4-11 cells, we proved the specific effect of CAMKI deficiency on AML cell growth.

Next, we tested the in vivo effect of knockdown of CAMK1 or CAMK4 expression in MV4-11 leukemia cells in transplanted NOD/SCID-IL2RG (NSG) mice. Both Camk1 or Camk4 knockdown significantly prolonged the survival of xenografted mice (Fig. 5a) and greatly inhibited leukemia development as determined by analysis of knockdown cells (Fig. 5b), human leukemic hCD45⁺ cells (Fig. 5c), and spleen size (Fig. 5d).

CREB was previously reported to be a possible downstream target of CAMKIV [22]. We found that PirBTM AML cells have decreased phosphorylation of both CAMKIV and CREB (Fig. 1a, Fig. 6a). Therefore, we hypothesized that CREB is a downstream mediator of PirB signaling in AML cells. To test this hypothesis, we introduced retrovirally expressed WT or inactive mutant CREB (S129A) into PirBTM AML cells. WT but not S129A CREB could rescue the defective phenotype of PirB TM AML cells in vitro and in vivo (Fig. 6b–d). These results suggest that CREB acts downstream of PirB-mediated signaling to support AML development.

To test the relationship between LILRB2 and CAMKs, we transfected Flag-tagged LILRB2 and Myc-tagged CAMKI or CAMKIV into 293T cells, then performed co-immunoprecipitation and Western blotting assays. Both CAMKI and CAMKIV bind LILRB2 in transfected 293T cells (Fig. 6e). Importantly, our bidirectional pull-down experiments showed that endogenous LILRB2 and CAMK1 interact with each other in MV4-11 cells (Fig. 6f). These data further support the conclusion that CAMKs play a crucial role in LILRB2-mediated signaling in AML cells (Fig. 6g).