8-Hydroxydaidzein Downregulates JAK/STAT, MMP, Oxidative Phosphorylation, and PI3K/AKT Pathways in K562 Cells

A metabolite isolated from fermented soybean, 8-hydroxydaidzein (8-OHD, 7,8,4′-trihydroxyisoflavone, NSC-678112), is widely used in ethnopharmacological research due to its anti-proliferative and anti-inflammatory effects. We reported previously that 8-OHD provoked reactive oxygen species (ROS) overproduction, and induced autophagy, apoptosis, breakpoint cluster region-Abelson murine leukemia viral oncogene (BCR-ABL) degradation, and differentiation in K562 human chronic myeloid leukemia (CML) cells. However, how 8-OHD regulates metabolism, the extracellular matrix during invasion and metastasis, and survival signaling pathways in CML remains largely unexplored. High-throughput technologies have been widely used to discover the therapeutic targets and pathways of drugs. Bioinformatics analysis of 8-OHD-downregulated differentially expressed genes (DEGs) revealed that Janus kinase/signal transducer and activator of transcription (JAK/STAT), matrix metalloproteinases (MMPs), c-Myc, phosphoinositide 3-kinase (PI3K)/AKT, and oxidative phosphorylation (OXPHOS) metabolic pathways were significantly altered by 8-OHD treatment. Western blot analyses validated that 8-OHD significantly downregulated cytosolic JAK2 and the expression and phosphorylation of STAT3 dose- and time-dependently in K562 cells. Zymography and transwell assays also confirmed that K562-secreted MMP9 and invasion activities were dose-dependently inhibited by 8-OHD after 24 h of treatment. RT-qPCR analyses verified that 8-OHD repressed metastasis and OXPHOS-related genes. In combination with DisGeNET, it was found that 8-OHD’s downregulation of PI3K/AKT is crucial for controlling CML development. A STRING protein–protein interaction analysis further revealed that AKT and MYC are hub proteins for cancer progression. Western blotting revealed that AKT phosphorylation and nuclear MYC expression were significantly inhibited by 8-OHD. Collectively, this systematic investigation revealed that 8-OHD exerts anti-CML effects by downregulating JAK/STAT, PI3K/AKT, MMP, and OXPHOS pathways, and MYC expression. These results could shed new light on the development of 8-OHD for CML therapy.


Introduction
Leukemia is a hematologic disorder, which can be classified into four main types, including acute lymphocytic leukemia (ALL), chronic lymphocytic leukemia (CLL), acute myeloid leukemia (AML), and chronic myeloid leukemia (CML), according to the speed and affects the activity of multiple transcriptional regulators which are involved in cell proliferation and survival [20]. The PI3K/AKT/mammalian target of rapamycin (mTOR) pathway plays a critical role in TKI-resistance in CML [21]. It was reported that NVP-BEZ235, a dual PI3K/mTOR inhibitor, effectively inhibited proliferation, and promoted apoptosis and autophagy in K562 cells [22].
MYC is a proto-oncogene and encodes a nuclear phosphoprotein that is involved in cell-cycle progression, apoptosis, and cellular transformation. It was proposed that ABL activates MYC through indirectly augmenting the acetylation of MYC and directly phosphorylating MYC [23]. Ubiquitination of MYC is considered an essential regulatory step in CML [24]. It was reported that a higher MYC expression level is associated with the progression of CML into a blastic crisis and poorer prognoses [25,26]. Recently, various types of potential direct and indirect MYC inhibitors have been developed to treat hematologic malignancies [25].
High-throughput technology uses an extensive holistic approach to detect differences in thousands of gene expressions from functional genomics and biological systems [41][42][43][44][45]. In this study, a microarray analysis was employed to investigate differentially expressed genes (DEGs) in 8-OHD-treated K562 cells, a model cell line derived from a female CML patient in blast crisis [46]. The gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were then analyzed [47][48][49]. In our previous study, we investigated the anti-CML effects of 8-OHD, focusing on cell-cycle regulation, cell apoptosis, autophagy, differentiation, and the altered level of BCR-ABL in K562 cells [39]. However, there are still many gaps in our knowledge of the roles of 8-OHD in CML, especially in terms of metabolism, the ECM during invasion and metastasis, and survival signaling pathways. Therefore, in the present study, we combined a bioinformatics analysis with experimental validation to investigate how 8-OHD regulates the JAK/STAT, MMP, and OXPHOS pathways. Furthermore, two hub genes, AKT and MYC, with high degrees of connectivity to biological process were found. These results shed new light on potential therapeutic targets for CML treatment. The isolation of 8-OHD from soybeans fermented with Aspergillus oryzae was conducted, from which the nuclear magnetic resonance (NMR) spectral data and purification were reported previously [33,34,37]. The chemical structure of 8-OHD is presented in Figure 1a.

Cell Culture
K562 cells were purchased from the Bioresource Collection and Research Center (Hsinchu, Taiwan) and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), 1% nonessential amino acids (NEAAs), 100 units/mL of penicillin, and 100 µg/mL of streptomycin (Thermo Fisher Scientific, Rockford, IL, USA) in a 5% CO 2 incubator at 37 • C.  (c) Heatmap of enriched gene ontology (GO) terms associated with 50 and 100 µM 8-OHD differentially expressed genes (DEGs). GO processes listed in the upper half are terms of upregulated DEGs. The lower processes are GO terms of downregulated DEGs. Only some of the representative GO terms are shown. (d) Venn diagrams demonstrating the numbers of common downregulated DEG between those treated with 50 µM 8-OHD (H5) and 100 µM 8-OHD (H1) compared to that of vehicle (V), based on the threshold of multiples of change of >1.5 or <−1.5, and p < 0.05. The 'ˆ' symbol represents the intersection of two sets.

Microarray Analysis
Cells were treated with vehicle (0.1% DMSO) or 8-OHD (50 and 100 µM) for 48 h. RNA was extracted from K562 cells with a Illustra RNA Spin Mini RNA Isolation Kit (GE Healthcare, Wauwatosa, WI, USA). A microarray analysis was performed as previously described [50,51]. All raw data were processed using the CLC Genomics Workbench v10.1 according to our previously established pipeline [52][53][54][55][56][57]. DEGs related to the treatment with 8-OHD were screened, and those with multiples of change values of >1.5 or <−1.5 and p < 0.05 were selected. DAVID (v6.8) is a well-known web-based database with dominant gene functional classification, and various embedded biological process and pathway annotations. Gene ontology (GO) and KEGG analyses cluster targeted genes into different subgroups according to biological functions, signaling pathways, or diseases by an agglomeration algorithm method [52][53][54][55][56][57], and p < 0.05 was used as the cut-off criterion [58].

Pathway Enrichment Analysis
Molecular functions and the pathway database from the MetaCore platform (GeneGo, St. Joseph, MI, USA) were further used to explore potential signaling pathways and process networks modulated by 8-OHD. GO terms and a heatmap were used to summarize downregulated DEGs and pathways as we previously described [59][60][61].
The Search Tool for the Retrieval of Interacting Proteins (STRING) database contains a huge number of protein-protein networks from 5090 organisms, 24.6 million proteins, and more than 2000 million interactions [63]. In the present study, we used the STRING database (v11.5) to analyze protein-protein interacting (PPI) networks [63]. A Markov clustering algorithm (MCL) with inflation factor = 2 was employed to find cluster structure in graphs by a mathematical bootstrapping procedure [64].

Western Blot Analysis
Whole-cell lysates and nuclear extracts were prepared from cultured K562 cells respectively using RIPA lysis buffer and nuclear extraction kit (Cayman Chemical, Ann Arbor, MI, USA). Protein concentrations were measured by the Bradford assay (Bio-Rad Laboratories, Hercules, CA, USA).
Equal amounts of protein were loaded into the wells of 5~12% sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE). Following electrophoretic separation, proteins were further transferred to polyvinylidene difluoride (PVDF) membranes, which were then blocked with freshly made buffer (5% skim milk in phosphate-buffered saline (PBS) with 0.05% Tween 20, pH 7.4) and incubated with specific primary antibodies ( Table 1) overnight at 4 • C. After rinsing, membranes were incubated with properly diluted horseradish peroxidase (HRP)-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA) for 1 h. The signal on the membrane was developed by enhanced chemiluminescence detection (GE Healthcare). RNA was extracted from K562 cells using an Illustra RNA Spin Mini RNA Isolation Kit (GE Healthcare). A High-Capacity cDNA Archive kit (Thermo Fisher Scientific) was used for cDNA synthesis. A qPCR was performed with Power SYBR Green PCR Master Mix (Thermo Fisher Scientific) in a 20 µL total volume that contained 0.4 µM of each primer ( Table 2). The PCR program consisted of pre-incubation for 2 min at 95 • C, followed by 40 cycles at 94 • C for 15 s and 60 • C for 60 s (ABI StepOne Real-Time PCR System). A melting curve was performed to verify the amplification specificity. The relative mRNA expression was normalized to GAPDH expression and then calculated by the comparative Ct method. The primer pairs used in real-time PCR are given in Table 2.

Transwell Cell Invasion Assay
The in-vitro invasiveness of K562 cells was assayed using Matrigel ® -coated Falcon 24-well transwell cell culture chambers (Boyden chambers) with 8-µm-pore-size polyethylene terephthalate filter inserts (BD Biosciences, Bedford, MA, USA). The chamber filters were coated with 50 mg/mL of Matrigel ® matrix overnight at 37 • C. K562 cells were treated with vehicle (0.1% DMSO) or 8-OHD (25 and 50 µM) for 24 h and collected by centrifugation to eliminate the medium. A 0.2-mL cell suspension containing 4 × 10 5 treated cells in serum-free medium was added to the Matrigel ® -coated filter inserts. Next, 750 µL of serum-containing medium was placed as a chemoattractant in 24-well plates. A light microscope was used to count invasive cells that had migrated into the lower chamber after K562 cells were cultured at 37 o C for 4 and 24 h [65,66]. Table 2. The primer pairs used in real-time PCR.

Statistical Analysis
All experiments were repeated at least three times, and the values were expressed as the mean ± standard deviation (SD). Results were analyzed using a one-way analysis of variance (ANOVA) with Dunnett's post-hoc test, and p < 0.05 was considered statistically significant.

Analysis of Microarray Data and Determination of DEGs
K562 cells were treated with vehicle (0.1% DMSO) or 8-OHD (50 and 100 µM) for 48 h, and transcript expressions were analyzed using the Human OneArray system, which contains 25,765 known genes. We first performed sample-level quality control using a principal component analysis (PCA) and hierarchical clustering [67]. Figure 1b shows that PC1 and PC2 accounted for more than 85% of the data variance, and two duplicates of each treatment were closely clustered, with each treatment cleanly separated.
Next, the intensity data were pooled and calculated to identify DEGs based on the threshold of multiples of change of >1.5 or <−1.5, and p < 0.05. Correlations of genetic signatures between control samples and 8-OHD treatment conditions were analyzed by unsupervised hierarchical clustering. Figure 1c shows hierarchical clustering of GO terms, associated with DEGs of 8-OHD-treated samples (H1 and H5), revealing that H1 and H5 could be clustered into a group. The regulated genes for each GO term are listed in Supplementary Table S1. We intended to investigate the pathways and networks associated with 8-OHD-downregulated genes in this study; therefore, 1910 commonly downregulated DEGs were selected for further investigation (Figure 1d).

Enrichment of Downregulated Signaling Pathways Using MetaCore
We used the MetaCore bioinformatics suite to analyze pathways enriched by the 1910 downregulated DEGs, and the top 15 pathways are shown in Figure 2a. Downregulated genes for the top 50 pathways are listed in Supplementary Table S2. The JAK/STAT pathway is crucial in transferring extracellular signals to cells and initiating gene expressions involved in cell proliferation, differentiation, survival, and developmental processes [68,69]. STAT3 and STAT5A/STAT5B are of particular interest because they are downstream effectors of tyrosine kinase oncogenes and contribute to the development of hematopoietic malignancies through both canonical and non-canonical pathways [6,69]. We found that STAT3 and STAT5 appeared in nine out of the top 15 enriched pathways. These included "Development_Transcription regulation of granulocyte development", "Prolactin/JAK2 signaling in breast cancer", "IL-6 signaling in colorectal cancer", "Cell cycle progression in prostate cancer", "Regulation of microRNAs in colorectal cancer", "Immune response_IL-4-induced regulators of cell growth, survival, differentiation and metabolism", "Immune response_M-CSF-receptor signaling pathway", "c-Myc in multiple myeloma" and "Devel-opment_Thrombopoietin signaling via JAK-STAT pathway".
Rajabi et al., 2020 also reported that 8-OHD triggered apoptosis in breast cancer stemlike cells through inhibiting the interleukin (IL)-6-mediated JAK2/STAT3 pathway [73]. STAT3 activation was found to be correlated with TKI resistance in CML [74]. Recently, transcriptome and reverse-phase protein arrays with a STRING analysis of differentially active proteins within TKI-persistent K562 cells revealed that STAT3 and AKT1 are central nodes [75]. We therefore further verified the involvement of the JAK2/STAT3 pathway in 8-OHD-treated K562 cells by Western blotting. Figure 3a shows that cytosolic JAK2 expression was dose-and time-dependently downregulated. Cytosolic STAT3 expression had profoundly decreased after 48 h in a dose-dependent manner and was correlated with transcriptomic data (Figure 2c). Furthermore, STAT3 phosphorylation was accordingly reduced (Figure 3a). Phosphorylated (p)-STAT3 is translocated into nuclei and serves as a transcription factor in various gene expressions. Figure 3b shows that nuclear pan-and p-STAT3 dosedependently decreased. This indicates that STAT3-regulated pathways were downregulated by 8-OHD in K562 cells.

8-OHD Repressed MMPs Expression and Activities
It was reported that BCR-ABL + cells secrete angiogenic factors, such as MMPs and vascular endothelial growth factor (VEGF), and stimulate angiogenesis [76].      [77]. It was reported that MMP14 is the target of BCR-ABL/ABL interactor 1 (Abi 1) signaling, and induces increased motility and invasiveness leukemic cells [78]. MMP15 is highly expressed in AML patients and was correlated with poor overall survival [79]. Downregulation of MMP14 and MMP15 by 8-OHD may indicate that it possesses anti-metastatic activity (Figure 4c,d).
Son of Sevenless 1 (SOS1), a guanine nucleotide exchange factor (GEF), is a dual activator of the small GTPases RAC and RAS. RAS activates MEK/ERK and PI3K/AKT pathways, which are essential for cell survival and proliferation [83]. The SOS1-RAC axis is critical for CML transformation and the leukemogenic potential of BCR-ABL [84]. Recently, it was found that SOS1 is one of the essential genes which have higher expression and encode proteins that engage in PPIs in K562 cells [85]. Selective and potent small-molecule inhibitors of SOS1 were developed to block RAS activity in tumor cell lines [86]. We found herein that 8-OHD could dose-dependently downregulate SOS1 transcription (Figure 4f).
Most MMPs are secreted as inactive forms and activated by cleavage of extracellular proteinases. On the other hand, MMP11 is processed intracellularly and secreted in its active form. MMP11 plays a vital role in the progression of epithelial malignancies, such as colon cancer, breast carcinoma, esophageal cancer, basal cell carcinoma, and others; but there are no reports of its role in CML progression [87]. We found that 8-OHD dosedependently repressed MMP11 expression. However, the detailed underlying mechanism by which 8-OHD downregulates the above gene expressions needs further study.
Most studies of myeloproliferative malignancies focused on expressions or activities of MMP9 and MMP2, which are secreted gelatinases [76,[88][89][90]. Neither of these two MMPs was downregulated in the microarray assay; we therefore analyzed MMP activity in cultured medium by gelatin zymography. We found that K562-secreted MMP9 activity (96 kDa) could be dose-dependently inhibited by 8-OHD (Figure 5a). We previously reported that 100 µM 8-OHD exerted about a 30% inhibitory effect on K562 proliferation after 24 h of treatment, while 25 and 50 µM, 8-OHD respectively, showed less than 10% and 20% inhibition [39]. To explore the effects of 8-OHD on cellular invasion, K562 cells were treated with the vehicle, or 25 or 50 µM 8-OHD for 24 h. Then treated cells were collected by centrifugation, washed and resuspended in serum-free medium. Equal amounts of cells were seeded in the Matrigel ® -coated inserts while the lower chambers were filled with medium containing serum as a chemoattractant. Figure 5b shows that the invasive ability of 8-OHD-treated K562 cells through an ECM-coated membrane was significantly lower than that of the vehicle-treated cells, in a dose-dependent manner, in both the 4-and 24-h assays (p < 0.01). In conclusion, 8-OHD represses MMP gene expression and activity, and inhibits K562 cell invasion.

8-OHD Repressed OXPHOS
It is known that enrichment of DEGs by the multiple of change may discard many useful genes. We thus re-selected target genes commonly dose-dependently downregulated by 50 and 100 µM of 8-OHD using normalized data. An entire set of 5042 downregulated genes was selected and subjected to functional annotation and clusterization using the DAVID bioinformatics tools. The top KEGG pathways are shown in Figure 6a. It was found that "metabolic pathways" were the most significantly associated, followed by "oxidative phosphorylation" and three neurodegenerative diseases: Parkinson's, Huntington's, and Alzheimer's diseases. These results agree with our finding in Figure 1c that "mitochon-dria" was the top GO term enriched by 8-OHD-downregulated DEGs (p = 1.81 × 10 −33 ) (Supplementary Table S1).  Using a global mRNA microarray analysis combined with an Ingenuity Pathway Analysis, Flis et al. found that CML stem cells display enhanced expression of OXPHOS genes [91]. Furthermore, disrupting the abnormal increase in OXPHOS in TKI-resistant CML stem cells can prevent disease relapse [92]. We previously reported that OXPHOS (GO:0006119) was significantly associated with 100 µM 8-OHD-downregulated DEGs [39]. Here, we further demonstrated the effects of 8-OHD on KEGG "oxidative phosphorylation" pathway as shown in Figure 6b. Expressions of most genes in complex I (NADH dehydrogenase) were repressed in response to 8-OHD. Expression patterns of some significantly downregulated genes are further shown in a heatmap in Figure 6c.

PI3K/AKT Pathway Is Resposible for K562 Cell Cycle Arrest
Genes associated with ML were collected from the DisGeNET database (http://www. disgenet.org/; accessed on 1 November 2021), which led to a collection of 385 pathological targets of ML (C0023470). Subsequently, the intersection of ML-associated genes and 8-OHD-downregulated 5042 genes was regarded as putative targets of 8-OHD against ML. As shown in Figure 8a, 112 genes were obtained in this manner. The top 15 KEGG pathways associated with these 112 targeted genes are shown in Figure 8b. Detailed information and downregulated genes involved in the top 25 KEGG pathways are shown in Supplementary  Table S3. "Chronic myeloid leukemia" (hsa05220) and "PI3K-Akt signaling" (hsa04151) are the top 2 and 4 on the list and the 8-OHD-downregulated genes involved in these two pathways are indicated by red marks in Figure 8c,d. It was found that PI3K/AKT signaling and BCR-ABL-induced gene expressions of the SOS-RAC axis, STAT5, and MYC were downregulated by 8-OHD in the CML pathway (hsa05220) (Figure 8c). We previously reported that 8-OHD caused K562 cell-cycle arrest at the S phase by inhibiting p21Cip1 expression and upregulating CDK6 and cyclin D2 expressions [39]. The current data further demonstrated that the PI3K/AKT pathway is responsible for the upstream signaling of 8-OHD-induced cell cycle arrest (Figure 8d).

AKT and MYC Are Hub Proteins Downregulated by 8-OHD
We then used the STRING database to analyze potential PPI networks for the 112 gene targets. Markov clustering algorithm (MCL) was selected to classify proteins into different clusters; the resulting network was divided into six clusters that contained most of the genes associated with cancer progression and metastasis (Figure 9a). The proteins enriched in AML and CML are respectively shown in blue and red. There were nine common proteins, including AKT1, MYC, NRAS, KRAS, PIK3CB, SOS1, MAPK3, STAT5A, and STAT5B. It clearly showed that AKT1 and MYC are two hub proteins, with 45 and 46 edges, respectively. A Western blot analysis of AKT activation and MYC expression was employed to reconfirm the bioinformatics data. Figure 9b shows that phosphorylation of AKT was pronouncedly inhibited by 8-OHD in a time-and dose-dependent manner. It was reported that inhibitors of PI3K/AKT/mTOR pathway can enhance CML apoptosis and autophagy and increase CML sensitivity to TKIs [22]. In combination with our previous finding that 8-OHD can induce apoptosis and autophagy [39], the current study further demonstrated that these effects are related to downregulation of the PI3K/AKT/mTOR pathway.   Figure 9c shows that the nuclear c-Myc level was decreased by 8-OHD. The inhibition profile was less effective than that of AKT and was completely correlated with that of BCR-ABL in our previous report [39]. This result supports 8-OHD being able to repress both ABL and MYC oncogene expressions and inhibit uncontrolled CML proliferation.

Conclusions
The bioinformatics data revealed that 8-OHD downregulated JAK/STAT, PI3K/AKT, MMP, and OXPHOS pathways in K562 cells. Validated data from Western blotting revealed that cytosolic JAK2 expression and phosphorylation of STAT3 decreased in dose-dependent manners in 8-OHD-treated K562 cells. Meanwhile, the repression of MMP14, MMP15, VEGFA, SOS1, and MMP11 by 8-OHD may indicate that it possesses anti-metastatic activity. Gelatin zymography and the transwell assay further confirmed that K562-secreted MMP-9 and invasive activities were inhibited by 8-OHD. In addition, RT-qPCR data revealed that 8-OHD dose-dependently repressed OXPHOS-and energy metabolism-related genes, including NDUFB10, NDUFV1, NDUFS7, NDUFB8, SDHB, CYC1, COX5A, and ATP5D. In combination with DisGeNET, we found that 8-OHD-downregulated PI3K/AKT is a key pathway for CML development. The PPI network also indicated that AKT1 and MYC are two hub proteins for cancer progression and metastasis, and the Western blot analysis of AKT activation and MYC expression was employed to reconfirm these prediction data.
Collectively, the combination of bioinformatics analyses and experimental validation has shed new light on the development of 8-OHD for CML therapy.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10.3390/ biomedicines9121907/s1, Table S1: Biological processes in the Gene Ontology (GO) analysis of up-and downregulated genes, associated with differentially expressed genes (DEGs) of 8-OHD-treated samples (H1 and H5); Table S2: Pathway and network associated with downregulated DEGs of 8-OHD-treated samples (H1 and H5) via the MetaCore platform; Table S3: Top KEGG pathways associated with putative targets of 8-OHD against myeloid leukemia (ML); Figure S1: Western blotting analysis of BAX expression in response to 8-OHD treatment.