Human AML cell lines HL-60, HEL and THP-1 were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China), and cultured in RPMI-1640 medium supplemented with 7% fetal bovine serum (FBS; both from Gibco; Thermo Fisher Scientific, Inc.) at 37°C in 5% CO₂. A total of 61 newly diagnosed AML patients and 6 AML patients in complete remission (CR) in the First Affiliated Hospital of Wenzhou Medical University from March to December 2018 according to the FAB classification system were enrolled in this study (Table SI). Eleven age-matched healthy donors were enrolled as controls. Bone marrow mononuclear cells (BMMNCs) were isolated using Ficoll-Hypaque (Haoyang Institute of Biotechnology, China). In addition, BMMNCs of the newly diagnosed AML patients were not used for the following experiments until the percentage of CD45^dimSSC^dim cells was more than 70% using flow cytometric analyses. All primary blast cells were cultured in RPMI-1640 medium supplemented with 20% FBS at 37°C in 5% CO₂.

This study was approved by the Institutional Review Board of the First Affiliated Hospital of Wenzhou Medical University, and informed consent was obtained from all participants in accordance with the Declaration of Helsinki protocol.

The relationship of CSNK1A1 expression and the prognosis in patients with AML was analyzed using UALCAN (22), an interactive web-portal to perform to in-depth analyses of TCGA gene expression data. UALCAN is publicly available at http://ualcan.path.uab.edu.

The total RNA was extracted from cells using Trizol reagent (Life Technologies, USA) and reverse transcribed into cDNA with following procedure: 42°C for 30 min, 70°C for 15 min, and then cDNA was amplified and quantified according to the procedure: 95°C for 30 sec, followed by 40 cycles at 95°C for 5 sec and at 60°C for 34 sec by ABI Prism 7500 using SYBR Green PCR Master Mix (Takara) with primer pairs. All results were normalized against GAPDH, RQ=2^−ΔΔCq (23). Primer sequences are listed in Table SII.

Four separate short hairpin RNA (shRNA) constructs targeting CSNK1A1 (shCSNK1A1-A, shCSNK1A1-B, shCSNK1A1-C, and shCSNK1A1-D) and one control shRNA construct targeting lactose gene were synthesized (Hanyin Biotechnology Co., Ltd.). The sequences of the target sequences for shRNAs and synthetic oligo information are shown in Table SIII. Each sequence pair was cloned into the LV3 shuttle plasmid with a RSV and CMV and H1 three promoter-driven GFP expression cassette, respectively. Each lentiviral expression plasmid and three corresponding packaging plasmids were co-transfected into 293T cells with RNAi-Mate reagent (GenePharma), respectively. Supernatants were collected 72 h after transfection, and the final virus titer in the supernatant was 1.3×10⁹ TU/ml for each lentivirus. AML cells were infected with the lentivirus in the presence of 10 µg/ml polybrene, and 72 h after infection, 3 µg/ml puromycin was used to screen those successfully transfected cells. The efficiency of gene silencing was examined using RT-PCR and western blot analysis.

The cells were harvested, and lysed with RIPA lysis buffer (Beyotime) containing PMSF and protease inhibitor cocktail (Beyotime). The protein was boiled and subjected to western blot analysis with different antibodies, including p53 [Cell Signaling Technology, Inc. (CST), #2527], SQSTM1/p62 (CST, #8025), p-AMPK (CST, #4184), AMPK (CST, #2532), mTOR (CST, #4517), p-mTOR (CST, #5536), PARP (CST, #9542), LC3-I/II (CST, #4108), GAPDH (CST, #2118), CK1α (Abcam, ab206652), ATG-7 (CST, #2631), at 1/1,000 dilution for incubation overnight at 4°C, respectively. For Co-IP, the cells were lysed with RIPA lysis buffer containing PMSF on ice for 0.5 h. Protein concentration was determined using a BCA protein assay kit. Five hundred micrograms of protein per sample was used and coated with 20 µl protein A/G-agarose beads and 1 µg mouse-control IgG with protein to eliminate non-specific proteins for 1 h, and then was centrifugated at 2,400 × g for 5 min. The supernatants were incubated with beads coated with 10 µg/ml mouse-source anti-MDM2 or 1 µg control IgG overnight at 4°C. Beads were washed three times with lysis buffer and boiled for 5 min. The bead-protein mixture was subjected to western blotting with antibodies against mouse-source anti-MDM2 (#AF208, Affinity, at 1/1,000 dilution), rabbit-source anti-p53 (#2527, CST, at 1/1,000 dilution), anti-CK1α (ab206652, Abcam, at 1/1,000 dilution), respectively. Optical densities of the bands were scanned and analyzed with FluorChem E and AlphaView SA software (Version: 3.4.0.0) (both from ProteinSimple).

Apoptosis assay was measured using the Annexin V-FITC/PI apoptosis kit and Annexin V-APC/7-AAD apoptosis kit (MultiSciences Biotech Co., Ltd.) as described previously (24). The cell viability assay was measured using the Cell Counting Kit-8 (CCK-8, Dojindo, Japan) according to the manufacturer's protocol.

The cells were cultured in methylcellulose medium (MethoCult™ H4034, Stemcell Technologies) according to the manufacturer's protocol. Visible colonies, defined to consist of at least 60 cells were counted. The colonies were photographed by microscope (BX51, Olympus, Japan) and the total cells were counted after rinsing with PBS.

Cell masses were fixed with 2.5% glutaraldehyde and were cut into 1×1×1 mm pieces. Then, the cell masses were cleaned by 0.1 M phosphoric acid bleach, dehydrated by 50, 70, 90 and 100% ethanol and 100% epoxy for 15 min, embedded with epoxypropane + buried solution (2:1) for 2 h, epoxypropane + burden (1:2) for 3 h, pure buried liquid overnight and pure buried liquid for 4 h, sliced by slim slicer (70 nm), and stained by 3% uranium acetate-lead citrate. Finally, the mitochondrial membrane and typical autophagic vacuoles of each cell sample were observed under ×20,000 magnification using a transmission electron microscope (Jeol JEM 1230, Japan).

Two-tailed unpaired Student's t-tests with 95% confidence interval (CI) were used to analyze the data involving direct comparison of an experimental group with a control group. One-way ANOVA tests with Dunnett's method for multiple comparisons with 95% CI were used to analyze the data involving 2 or more test groups and a control group. One-way ANOVA with Tukey test was used to perform multiple comparisons for up to 3 groups. P-value <0.05 was considered to indicate a statistically significant difference. All experiments were performed with three replicates, unless stated.

Mounting evidence indicates that β-catenin is overexpressed and promotes disease recurrence contributing to a poor prognosis in AML (25), which motivated us to investigate whether CK1α, a negative regulator of β-catenin, also has a similar role. CK1α mRNA was significantly higher in BMMNCs from the newly diagnosed AML patients than those from healthy donors or AML patients who achieved CR (Fig. 1A). There was no difference in the level of CK1α mRNA among the different types of FAB (Fig. 1B). Furthermore, the patients with a high level of CK1α mRNA showed a poorer outcome than those with a low or medium level of CK1α mRNA (P=0.034, Fig. 1C). These findings suggested that CK1α may be overexpressed and confer a poor prognosis in AML. All patient information is listed in Table SI.

To investigate the role of CK1α in AML cells, we decreased CK1α using pharmacological or genetic inhibition. We found that D4476 significantly reduced the cell viability of three AML cell lines HL-60, HEL and THP-1 in a time- and dose-dependent manner (Fig. 2A), while only a dose-dependent manner was shown after treatment with D4476 lower than 10 µM in patient blast cells. We further investigated whether D4476 affects the colony formation ability of AML cells, and found that D4476 at 40 µM dramatically impaired the colony formation, as indicated with colony number and total cell number, in both HL-60 and HEL cells (Fig. 2B). Notably, there were many small cell clusters, but did not reach the standard of a colony in the D4476 treatment group in these two cell lines (Fig. 2B). Meanwhile, similar results were also observed in these two AML cell lines with genetic inhibition of CK1α using lentiviral-mediated shRNA (Figs. 2C and S1). Importantly, we found that the colony size was decreased in both the D4476 treatment group and the shCSNK1A1 group (Fig. 2B and C), suggesting that pharmacological or genetic inhibition of CK1α affects the colony formation mainly through decreasing the size of the colony in AML cells.

We found that D4476 dramatically induced apoptosis in a dose-dependent manner in three AML cell lines tested as well as patient blast cells (Fig. 3A). As expected, D4476 treatment promoted the cleavage of PARP, a classical marker of apoptosis, and decreased the expression of survivin, a negative marker protein of apoptosis, in HL-60, THP-1 and HEL cells as well as in patient blast cells (Fig. 3A). Similarly, the genetic inhibition of CK1α also induced a moderate apoptosis in the three AML cell lines HL-60, THP-1 and HEL (Fig. 3B). Furthermore, combination with cytarabine (Ara-c) and D4476 or gene silencing of CK1α had a synergistic effect on the induction of apoptosis in THP-1 and HEL cells (Fig. S2). In addition, we found that pharmacological or genetic inhibition of CK1α had almost no effect on cell cycle distribution of HL-60 or HEL cells (Fig. S3).

Autophagy is a cellular process activated in response to various forms of stresses such as nutrient deprivation, hypoxia and endoplasmic reticulum (ER) stress. Apoptosis triggered by ER stress is often accompanied by increased autophagy. Apoptosis induced by CK1α inhibition prompted us to investigate whether CK1α inhibition also activates autophagy. We used TEM to observe autophagosomes and found that pharmacological or genetic inhibition of CK1α promoted the autophagy flux of HL-60 and HEL cells with increased autophagosomes (Fig. 4A). We found that D4476 treatment upregulated the expression of ATG-7 and LC3-II in all three AML cell lines tested as well as patient blast cells (Fig. 4B), which further confirmed the occurrence of autophagy. SQSTM1/p62, a substrate of LC3-II, is usually reduced when autophagy occurs. However, we found that D4476 treatment concomitantly increased the levels of SQSTM1/p62 and LC3-II protein in HL-60 or HEL cells in a dose-dependent manner within 48 h (Fig. S4). It has been reported that in nutrient-rich conditions, β-catenin can inhibit autophagy and the generation of SQSTM1/p62 mRNA by regulating TCF4/LEF, but when autophagy occurs, β-catenin can be degraded by autophagy and p62 becomes derepressed and dramatically increased (26). We therefore determined the mRNA expression of SQSTM1/p62 and found that D4476 treatment significantly increased the level of SQSTM1/p62 mRNA, suggesting that enhanced autophagy reduced the inhibition of β-catenin on p62, resulting in greater p62 generation than degradation, ultimately leading to the accumulation of p62 (Fig. 4C). Furthermore, combination with Ara-c and D4476 resulted in the upregulation of autophagic flux in THP-1 and HEL cells (Fig. S2A).

Once confirmed, autophagy and apoptosis occurring together in the D4476-treated AML cells, the underlying relationship aroused our curiosity. Firstly, we combined Spautin-1 with D4476 to unravel the veil of mystery. Spautin-1 aggravated the cell growth inhibition induced by D4476 on all three AML cell lines tested as well as in the patient blast cells (Fig. 5A). Furthermore, we determined the apoptosis when the autophagy was inhibited. Spautin-1 alone did not induce apoptosis in AML cells, and Spautin-1 potently aggravated D4476-induced apoptosis on HEL cells but merely moderately exacerbated D4476-induced apoptosis in HL-60 cells (Fig. 5B). These data suggested that D4476 combined with Spautin-1 exerted a synergistic antitumor effect on AML cells.

A number of signaling pathways including PI3K/Akt/mTOR, p53/AMPK/mTOR, and MAPKs, are implicated in autophagic cell survival and cell death (27,28). We found that CK1α had a relation with MDM2 and p53 (Fig. 5C). It has been recognized that MDM2 is a negative regulator of p53 (29,30) and p53 regulates AMPK/mTOR signaling to activate autophagy (31). We found that CK1α inhibition by D4476 increased the expression of p53 and MDM2, promoted the phosphorylation of AMPK, and inhibited the phosphorylation of mTOR in a dose-dependent manner in HL-60 and HEL cells (Fig. 5D). Furthermore, we found that D4476 did not affect the level of MDM2 when CK1α was genetically inhibited in HEL cells (Fig. 5E), indicating that D4476 may influence MDM2 through CK1α. These data suggested that CK1α may regulate p53 through MDM2, and inhibit autophagy by the AMPK/mTOR signaling pathway downstream of p53.