The CML imatinib-resistant cell line K562/G01 was a kind gift from Professor Zhenlun Gu (Suzhou University, China). The CML cell line, KCL22 and K562, and acute myeloid leukemia (AML) cell lines, HL60 and NB4, were purchased from the Cell Bank of Shanghai Institute of Cell Biology, Chinese Academy of Science (Shanghai, China) and stored at our laboratory. All cell lines were maintained in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (FBS) (HyClone; GE Healthcare) and 1% penicillin-streptomycin (Beyotime Institute of Biotechnology) at 37°C in a 5% CO₂ atmosphere.

The reagents and standard protocols used for the extraction of total RNA (RNAiso Plus), RNA reverse transcription into cDNA (PrimeScript™ RT reagent Kit), and RT-qPCR (SYBR^® Premix Ex Taq™ II) were obtained from Takara Bio. Inc. The thermocycling conditions were as follows: Initial denaturing step (95°C, 3 min), followed by 40 cycles of denaturing (95°C, 10 sec), annealing (55°C, 30 sec) and extension (72°C, 30 sec). ACTB was used as an internal reference gene. The primers used for RT-qPCR are listed in Table I. Relative expression levels of mRNA were calculated using the 2^−ΔΔCq method (28).

Western blot analysis was performed according to a standard protocol, as described previously (29). The following primary antibodies were used: Anti-ACTB (cat. no. TA09 purchased from ZSGB-BIO/now OriGene Technologies, Inc.), anti-ABCC2 (anti-MRP2) (cat. no. ab172630 purchased from Abcam, Inc.). Moreover, anti-BCR-ABL1 (cat. no. 2862), anti-p-BCR-ABL1 (cat. no. 2864), anti-CCND1 (cat. no. 2922), anti-CTNNB1 (cat. no. 9562), anti-PARP1 (cat. no. 9532), anti-STAT5 (cat. no. 25656), anti-p-STAT5 (cat. no. 4322) and anti-TCF7 (cat. no. 2203) were purchased from Cell Signaling Technology, Inc. (CST). The antibodies were used at a dilution of 1:1,000, except for anti-ACTB (1:2,000).

One scrambled negative control and two independent TCF7-targeting short hairpin RNAs (shRNAs) were cloned into the lentiviral vector GV248 (GeneChem, Shanghai, China) at the AgeI and EcoRI sites. The shRNA sequences are provided in Table II. K562/G01 and K562 cells in logarithmic growth phase were plated into 96-well plates (5,000 cells per well) and infected for 24 h with 50 IFU/ml lentivirus and 10 µg/ml polybrene, and then replaced with the normal medium and cultured for 48 h. Next, puromycin was added to the plates at a final concentration of 2.0 µg/ml for reverse selection of the stable cell lines. Medium containing puromycin was replaced every 3 days. Transfection efficiency was monitored by inverted fluorescence microscopy and flow cytometry. After stable cell lines were produced, normal medium was used.

Cells were smeared across a gelatin-coated slide to form a cell monolayer, and then the cell smear was fixed with methanol at −20°C for 20 min. The cell membranes were permeabilized using 1% Triton X100-PBS at 37°C for 15 min. After washing the fixed slides three times in PBS, non-specific antigens were blocked with 10% goat serum at 37°C for 1 h. The anti-TCF7 primary antibody (cat. no. 2203; CST) was diluted using 10% goat serum to 1:400, applied to the slide to cover the cell smear, and incubated overnight at 4°C. The slides were washed three times in 400 µl of wash buffer (0.1% BSA in 1X PBS). The secondary antibody (goat anti-rabbit IgG (H+L) cross-adsorbed secondary antibody, cyanine 3; cat. no. A10520; Invitrogen; Thermo Fisher Scientific, Inc.) was diluted to 1:1,000, and 500 µl was added to the smear and incubated at room temperature for 1 h. Slides were rinsed twice in 500 µl of wash buffer, and the nuclei were stained using DAPI (Beyotime Institute of Biotechnology) diluted with PBS to 1:1,000 for 5 min. Slides were rinsed thrice with PBS and once with water. Finally, the smears with a drop of 70% glycerin were covered with coverglasses. The expression and distribution of fluorescence were observed using a fluorescence microscope (magnification, ×1,000; Nikon Corporation).

Cells were plated into 96-well flat-bottomed plates with 2×10³ cells per well and treated with or without imatinib at the indicated concentrations. After cell culture for 12, 24, 48, 72, 96 and 120 h, cell viability was determined using a CCK-8 kit (Solarbio, Inc.). For the colony formation assay, cells were seeded into 24-well flat-bottomed plates with 200 cells per well and grown in semi-solid medium containing 1.35% methylcellulose. After 9 days, the colonies were counted using an inverted fluorescence microscope (magnification, ×40; Nikon Corporation).

Cell cycle, apoptosis, cell counts, and GFP fluorescence were detected using flow cytometry (FCM). Cells were collected after treatment with or without imatinib at the indicated concentrations. To examine cell cycle dynamics, cells were subjected to serum starvation for 64 h to obtain synchronized cells, following which the serum supply was restored. Cell cycle status was monitored at 0, 8, 16, 24, and 32 h. Cell cycle profiling was delineated by the FL2 fluorescence generated by the binding of propidium iodide (PI) to DNA, and the percentages of cells in different phases of the cell cycle were analyzed by FlowJo VX.0.7 software (FlowJo LLC). To detect cell apoptosis, the cells were double-labeled with Annexin V-APC and DAPI and measured by FCM according to the manufacturer's protocol. Given that flow cytometry records the volume of fluid and the fluorescence parameters of particles simultaneously, if the sample is thoroughly mixed, an accurate cell count and GFP fluorescence can be obtained.

Total RNAs from four groups of K562/G01 cells with scramble, imatinib, TCF7_KD, and TCF7_KD+imatinib treatment were extracted using an RNeasy kit (Qiagen, Inc.), and treated with DNase I (Qiagen Inc.). Imatinib was used at 1 µM in K562/G01 cells. Shanghai Lifegenes Biotechnology performed RNA quantification, quality appraisal, library preparation, and sequencing. Raw data (raw reads) of fastq format were firstly processed through in-house perl scripts. HTSeq v0.6.1 (https://htseq.readthedocs.io/en/master/) was used to count the read number mapped to each gene. Gene fragment per kilobase of exon per million reads (FPKMs) were computed by summing the FPKMs of transcripts in each gene group. Gene set enrichment analysis (GSEA) (30) software v4.0.3 was used to analyze RNA-seq data. Cytoscape software v3.6.0 was used to visualize the GSEA reasults (31). The cut-offs of differentially expressed genes (DEGs) were set as |log₂ (fold change)| >0.5 and FPKM >0.3, and consequently 1,034 DEGs were obtained. The Gene Ontology (GO) enrichment analysis of DEGs was executed using the R package clusterProfiler v3.11.1 (32). All sequencing data were used in principal component analysis (PCA). Gene expression heatmaps and PCA were performed using the web tool ClustVis (https://biit.cs.ut.ee/clustvis/) (33). Three-dimensional plots were produced using the R package lattice v0.20-38. Venn diagrams were calculated and drawn using a web tool (http://bioinformatics.psb.ugent.be/webtools/Venn/). The visualization of RNA-seq and chromatin immunoprecipitation sequence (ChIP-seq) data were performed using Integrative Genomics Viewer software v2.6.3 (http://software.broadinstitute.org/software/igv/) (34).

Chromatin immunoprecipitation kit (cat. no. 9005) was purchased from CST, and anti-TCF7 (cat. no. bs1987) was purchased from Bioword, Inc. The ChIP experiment was performed according to the manufacturer's instructions. Immunoprecipitated DNA fragments were purified by phenol extraction and then quantified by qPCR. The primer sequences are listed in Table III.

Results of column charts and line charts are presented as the mean ± standard deviation and were analyzed by GraphPad (Prism 5) (GraphPad Software, Inc.). Each experiment was performed at least three times. Statistical analysis were performed using the Student's t-test or one-way analysis of variance (ANOVA) with Tukey's post hoc test. Statistical significance levels were as follows: *P<0.05, **P<0.01, ***P<0.001 (as shown in the figure legends with the respective symbols).

In the chronic phase, BCR-ABL1 is recognized as an effective target for CML treatment, but other targets need to be explored when resistance develops. We first investigated gene expression microarray datasets GSE47927 (35) and GSE4170 (36), and results of the analysis revealed that the Wnt/β-catenin signaling pathway was activated in all phases of CML and in the event of imatinib resistance (Fig. 1A). TCF7 expression was higher in blast crisis and imatinib-resistant samples, when compared with chronic phase and imatinib-sensitive samples, respectively. (Fig. 1B and C). Futhermore, we analyzed dataset GSE76312 (37) and the results were consistent with the results in GSE47927 and GSE4170 (Fig. S1E). To assess the expression of TCF7 in leukemia cell lines, we tested CML cell lines (K562, K562/G01 and KCL22), and acute myeloid leukemia (AML) cell lines (HL60 and NB4) using RT-qPCR and western blot analyses. The results showed that TCF7 was significantly overexpressed in blast crisis (K562, KCL22) and imatinib-resistent (K562/G01) CML cell lines (Fig. 1D and E). The K562/G01 cell line is evolved from the K562 cell line by long-term treatment with imatinib, and it has the characteristic of imatinib resistance.

To investigate whether the expression of TCF7 in imatinib-resistant cells was affected by the activity of BCR-ABL1, K562/G01 cells were treated with imatinib at concentrations of 1 and 5 µM, and the mRNA expression levels of BCR-ABL1, TCF7, and CCND1 were detected using RT-qPCR. Next, the activation status of BCR-ABL1, STAT5, and the protein expression levels of BCR-ABL1, STAT5, PARP, CCND1, and TCF7 were detected using western blot analysis. In response to imatinib, the results of RT-qPCR showed that there was no significant change in the mRNA expression of TCF7 and BCR-ABL1 except CCND1 (Fig. 1F). Consistent with this, the results of western blot analysis showed that the expression of CCND1 was gradually decreased while the expression of BCR-ABL1, STAT5, and TCF7 did not change. In addition, when cells were exposed to a high concentration of imatinib (5 µM), PARP1 began cleaving into fragments, and the activity of BCR-ABL1 and its downstream target STAT5, in the form of phosphorylated (p)-BCR-ABL1 and p-STAT5, were significantly inhibited (Fig. 1G). These results indicate that BCR-ABL1 activity has no significant effect on the regulation of TCF7. In addition, changes in cell morphology suggested that increasing drug concentrations led to increased death of imatinib-resistant cells to a certain extent (Fig. 1H).

K562 and K562/G01 cells were transduced with two LV-TCF7-RNAi recombinant lentiviruses and one LV-Scramble lentivirus, respectively. After puromycin treatment, stably transduced cells were obtained. The results of flow cytometry indicated that transduction efficiency was close to 100% (Fig. 2A). Next, The RT-qPCR results showed that the knockdown efficiencies of the designed shRNAs were all greater than 80% (Fig. 2B). Furthermore, the western blot results confirmed the results of the RT-qPCR (Fig. 2C). In addition, immunofluorescence assays results visualized TCF7 expression changes and nuclear localization in K562/G01 cells (Fig. 2D).

To test the effect of TCF7 knockdown on the proliferation of K562 and K562/G01 cells, we first performed a cell viability test. The results showed that the cell viability of K562/G01cells was significantly inhibited in the TCF7_KD groups compared with the Scramble groups (Fig. 3A). Interestingly, in contrast to K562 cells, the inhibition of cell viability caused by TCF7 knockdown was more pronounced in the K562/G01 cells (Fig. 3B). In addition, cell count results showed a lower cell counts in the TCF7_KD groups of K562/G01 cells (Fig. 3F).

We further performed cell cycle assays. The results showed that compared with the Scramble group, the TCF7_KD groups consisted of a higher proportion of G0/G1 phase cells, and less S+G2/M phase cells in the K562/G01 cells but not in the K562 cells (Fig. 3C and D). In addition, the TCF7_KD groups showed a significant increase in the number of sub-G1 phase cells (Fig. 3C). The serum starvation release test showed that after restoring serum to the serum-free medium, cells in the TCF7_KD group re-entered the cell cycle more slowly (Fig. 3E). The above results suggest that TCF7 knockdown led to an inhibition of proliferation, particularly in the CML imatinib-resistant cells.

Compared with the parental K562 cells, K562/G01 cells exhibit significant resistance to imatinib, and previous reports have shown that TCF7 may affect the drug resistance of tumor cells (26,27). Therefore, we investigated whether TCF7 knockdown can increase imatinib sensitivity in CML cells. The results of the drug sensitivity test showed that, in K562/G01 cells, the half maximal inhibitory concentration (IC₅₀) value of the Scramble group was 8.3 µM, while the IC₅₀ value of the TCF7_KD group was 4.7 µM (Fig. 4A and B). In comparison, no significant change in imatinib sensitivity was observed in K562 cells (Fig. 4C and D). Cell viability and GFP-positive cell count results showed that, in K562/G01 cells, cell proliferation in the TCF7_KD group was inhibited while imatinib concentration increased from 0.5 to 1.0 µM, but not in the Scramble group (Fig. 4E and F). In addition, the colony formation assay showed that TCF7 knockdown combined with imatinib could significantly inhibit the colony formation rate of K562/G01 cells (Fig. 4G and H). The above results showed that TCF7 knockdown can increase imatinib sensitivity in imatinib-resistant cells and that TCF7 knockdown combined with imatinib can inhibit imatinib-resistant cells more effectively.

To investigate why TCF7 knockdown affects proliferation and drug resistance of K562/G01 cells, we obtained RNA-seq data (GSE152220) from the Scramble, TCF7_KD, Scramble+Imatinib, and TCF7_KD+Imatinib groups. PCA result showed that the combination group underwent more intervention on the transcriptome (Fig. 5A). GO enrichment analysis showed that differentially expressed genes (DEGs) in the TCF7_KD group were particularly enriched in the term of leukocyte proliferation (Fig. 5B). These results explain our findings that TCF7 knockdown can affect proliferation of K562/G01 cells. Next, the GO Chord plot lists the core genes such as ABCC2 (Fig. 5C). ABCC2 is a member of the ATP-binding cassette (ABC) transporter superfamily, and this family is often associated with multidrug resistance of tumors (38,39).

Given the critical role of the ABC transporter family in chemotherapy resistance of tumor cells (40,41), we used RNA-seq data to analyze the changes in the ABC transporter signaling pathway in CML cells. Using single-cell RNA-seq data of 1,062 BCR-ABL1⁺ LSCs from the GSE76312 dataset (37) for TCF7 single-gene GSEA analysis, it was found that the expression of TCF7 was positively correlated with the gene expression of ABC transporter signaling pathway (Fig. S1A). We subsequently set the Scramble group as the control and compared it with the TCF7_KD, Scramble+Imatinib, and TCF7_KD+Imatinib groups using GSEA analysis. The results showed that although the ABC transporter signaling pathway was upregulated in K562/G01 cells following imatinib treatment, TCF7 knockdown caused its expression to be downregulated. When imatinib was used after TCF7 knockdown, the upregulated trend of the ABC transporter signaling pathway was neutralized (Fig. S1A).

Furthermore, we analyzed the Wnt (Fig. S1B) and Wnt/β-catenin signaling pathway (Fig. 6A), and the results showed that the trend in the changes in each group was consistent with the changes in ABC transporters. The expression levels of core enrichment genes representing ABC transporters and Wnt signaling pathway in the TCF7_KD group are displayed in the heatmap (Fig. 6B). Next, the expression levels of ABCC2 and CCND1 were verified by RT-qPCR and western blot analysis. The results confirmed that the expression levels of ABCC2 and CCND1 were decreased when TCF7 was silenced in the CML imatinib-resistant cells (Fig. 6C and D). In addition, the expression level of CTNNB1, a key protein of the canonical Wnt signaling pathway, was also decreased (Fig. 6C).

As a transcription factor, TCF7 promotes the transcription of many genes by binding to motifs. TCF7 target genes were calculated from ChIP-seq data in the GTRD database (42) and collated into a gene set, named TCF7_targets. We then analyzed the single-cell RNA-seq data of BCR-ABL1⁺ LSCs in the dataset GSE76312 (37) and our four groups of RNA-seq dataset GSE152220. The results showed that the expression of TCF7 was positively correlated with TCF7_targets gene set in BCR-ABL1⁺ LSCs cells (Fig. S1C). On the other hand, TCF7 knockdown resulted in its downregulation in K562/G01 cells (Fig. S1C).

ABCC2 was identified by screening a intersection of four gene, namely TCF7_targets, TCF7 correlated core enrichment genes, TCF7_KD downregulated genes, and ABC transporters (Fig. S1D). Integrative genomics viewer (IGV) was used to visualize the processed ChIP-seq data ENCFF476IUK, and it was found that TCF7 had a binding peak in the promoter region of ABCC2 (Fig. 6E). In addition, by integrating our RNA-seq data into IGV, it can be seen intuitively that the transcription level of ABCC2 was lower in the TCF7_KD group compared with the Scramble group (Fig. 6E). Furthermore, ChIP-qPCR results showed that TCF7 was recruited to the promoter region of ABCC2 in K562/G01 cells (Fig. 6F). These results indicate that TCF7 is a direct transcriptional regulator of ABCC2 in K562/G01 cells. In summary, the roles of TCF7 and imatinib in CML imatinib-resistant cells are shown in a graphical abstract (Fig. 6G).