Pharmacological Inhibition of PIP4K2 Potentiates Venetoclax-Induced Apoptosis in Acute Myeloid Leukemia

Abstract

Phosphatidylinositol-5-phosphate 4-kinase type 2 (PIP4K2) protein family members (PIP4K2A, PIP4K2B, and PIP4K2C) participate in the generation of PIP4,5P₂, which acts as a secondary messenger in signal transduction, a substrate for metabolic processes, and has structural functions. In patients with acute myeloid leukemia (AML), high PIP4K2A and PIP4K2C levels are independent markers of a worse prognosis. Recently, our research group reported that THZ-P1-2 (PIP4K2 pan-inhibitor) exhibits anti-leukemic activity by disrupting mitochondrial homeostasis and autophagy in AML models. In the present study, we characterized the expression of PIP4K2 in the myeloid compartment of hematopoietic cells, as well as in AML cell lines and clinical samples with different genetic abnormalities. In ex vivo assays, PIP4K2 expression levels were related to sensitivity and resistance to several antileukemia drugs and highlighted the association between high PIP4K2A levels and resistance to venetoclax. The combination of THZ-P1-2 and venetoclax showed potentiating effects in reducing viability and inducing apoptosis in AML cells. A combined treatment differentially modulated multiple genes, including TAp73, BCL2, MCL1, and BCL2A1. In summary, our study identified the correlation between the expression of PIP4K2 and the response to antineoplastic agents in ex vivo assays in AML and exposed vulnerabilities that may be exploited in combined therapies, which could result in better therapeutic responses.

1. Introduction

Acute myeloid leukemia (AML) is an aggressive hematological neoplasm characterized by the proliferation, survival, and accumulation of blasts in the bone marrow, peripheral blood, and other organs, which suppresses normal hematopoiesis [1,2]. Several factors influence the prognosis of AML, including clinical and laboratory characteristics, as well as the choice and response to treatment, among others [3]. In recent years, significant advances in the understanding of the cellular and molecular biology of AML have been made, paving avenues for the identification and approval of new drugs for the treatment of the disease [4,5].

Recently, our research group identified the fact that phosphatidylinositol-5-phosphate 4-kinase type 2 (PIP4K2) inhibitors could be a new class of drugs with antileukemic potential [6]. PIP4K2 (PIP4K2A, PIP4K2B, and PIP4K2C) are lipid kinases that phosphorylate phosphatidylinositol (PI)5 P at position four of the inositol ring and generate PI4,5P₂, which acts as a secondary messenger in signal transduction, a substrate for metabolic processes, and has structural functions [7,8]. In experimental models of AML, PIP4K2A was identified as a key protein for the maintenance of leukemia-initiating cells, acting on proliferation, clonogenicity, and survival [9]. A high PIP4K2A or PIP4K2C negatively impacts survival outcomes in a univariate analysis, but the combination of the expression of PIP4K2A and PIP4K2C was an independent prognostic factor in patients with AML [10]. THZ-P1-2 is a pan-inhibitor of PIP4K2 and selectively reduces viability through the disruption of mitochondrial homeostasis and changes in the autophagic flux [6].

The objective of this study was to comprehensively characterize the expression profile of PIP4K2 in normal and malignant hematopoietic cells, including myeloid cell lines. Furthermore, our investigation aimed to elucidate the role of PIP4K2 in the response to antineoplastic drugs within AML models. Consequently, we endeavor to propose promising pharmacological combinations incorporating the PIP4K2 inhibitor, THZ-P1-2, for combined therapeutic strategies.

2. Results

2.1. Expression of PIP4K2 Genes in Normal and Leukemic Hematopoiesis

The transcript levels of PIP4K2A, PIP4K2B, and PIP4K2C were characterized in different myeloid cell subpopulations. PIP4K2A expression was higher in erythroid (ERY) and megakaryocytic (MEGA) cells compared to common myeloid progenitors (CMP), granulocyte-macrophage progenitors (GMP), megakaryocyte-erythrocyte progenitors (MEP), metamyelocytes (META), neutrophils (NEU), eosinophils (EOS), basophils (BAS), and monocytes (MONO). PIP4K2B expression was higher in hematopoietic stem cells (HSC). PIP4K2C expression was lower in CMP and ERY cells (Figure 1A–D). In patients with AML, PIP4K2A and PIP4K2C expression was higher in complex karyotype groups, whereas PIP4K2B expression was higher in del(5q)/5q- and t(8;21)+others groups (Figure 1E). In the myeloid leukemia cell lines, PIP4K2A expression was higher in K562, KU812, HEL, and SET2 cells, while the expression of PIP4K2B and PIP4K2C was similar among the evaluated cells, highlighting a high gene and protein expression of PIP4K2C in K562 cells (Figure 1F–G).

2.2. PIP4K2 mRNA Levels Are Associated with Drug Sensitivity in Silico Analysis from Ex Vivo Assays Data in Acute Myeloid Leukemia

Next, we evaluated the association of the PIP4K2 genes with the response to antineoplastic agents evaluated in the primary samples from AML patients. Using the Beat AML study [11,12] dataset, the transcriptional levels of PIP4K2A were positively correlated with the resistance to seven drugs (selumetinib, pelitinib, ABT-737, cytarabine, RAF265, venetoclax, and cediranib) and negatively correlated with four (JNJ-7706621, saracatinib, bosutinib, and vargetef). PIP4K2B expression was positively correlated with the resistance to five drugs (vargatef, gefitinib, quizartinib, vandetanib, and LY-333531) and negatively correlated with four (selumetinib, GDC-0879, PHT-427, and nilotinib). PIP4K2C expression was positively correlated with the resistance to twelve drugs (nilotinib, SNS-032, ibrutinib, rapamycin, bortezomib, YM-155, flavopiridol, elesclomol, AT7519, selumetinib, VX-745, and MK-2206) and negatively correlated with seventeen (erlotinib, neratinib, dovitinib, AZD1152-HQPA, vandetanib, MLN8054, NVP-TAE684, RAF265, KU-55933, GSK-1838705A, alisertib, KW-2449, midostaurin, vatalanib, gefitinib, afatinib, and tozasertib; Figure 2A). By analyzing the molecular targets of the identified drugs, it was identified that a high frequency of molecular targets is mutually exclusive between the PIP4K2 genes associated with drug sensitivity and resistance. ABL1 was identified as a molecular target associated with sensitivity, related to the three genes, while EGFR was identified as a target associated with resistance, related to the PIP4K2 genes (Figure 2B).

2.3. THZ-P1-2, a PIP4K2 Pan-Inhibitor, Augments Venetoclax-Induced Apoptosis

Among the drugs identified as being associated with resistance in the presence of high levels of PIP4K2A expression, venetoclax draws attention. Venetoclax is a selective BCL2 inhibitor that was recently introduced into AML therapy and has been successful in the treatment of patients who were ineligible for intensive chemotherapy [13]. THZ-P1-2 serves as a pan-inhibitor for PIP4K2 using covalently binding to a specific cysteine residue situated within a disordered loop near the ATP-binding site of the kinase domain across all three PIP4K2 isoforms. This binding leads to the irreversible inhibition of their enzyme activity [14]. Therefore, the combination of venetoclax and THZ-P1-2 was tested in multiple cellular models of AML. In Kasumi-1, NB4, and U-937 cells, THZ-P1-2 showed potentiating effects on the reduction in cell viability caused by venetoclax (Figure 3A). These findings were confirmed through the analysis of apoptosis using flow cytometry (Figure 3B) and through the detection of molecular markers of cell death (increase in cleaved PARP1 and expression of γH2AX; Figure 3C).

2.4. Differential Gene Expression upon THZ-P1-2 and/or Venetoclax Exposure in Kasumi-1 Cells

Finally, to better understand the molecular events triggered by THZ-P1-2 in the presence or absence of venetoclax, a panel of genes associated with cell cycle progression, apoptosis, DNA damage, and autophagy were evaluated. In the Kasumi-1 cells, a total of thirteen and nine genes were modulated using THZ-P1-2 and venetoclax monotherapies, respectively. The combination of THZ-P1-2 and venetoclax modulated nine genes, highlighting the upregulation of TAp73 as a target significantly modulated only in the combined treatment (Figure 4A and Table S1). In a network analysis, the genes differentially expressed upon THZ-P1-2 treatment were associated with apoptotic mitochondrial changes, a positive regulation of cysteine-type endopeptidase activity, a regulation of the response to endoplasmic reticulum stress, DNA damage response (signal transduction by p53 class mediator), and cell cycle arrest. The venetoclax treatment was associated with apoptotic mitochondrial changes, a regulation of the response to endoplasmic reticulum stress, signal transduction in the absence of ligand, phagophore assembly site, and a cellular response to external stimulus. The combined treatment, THZ-P1-2 plus venetoclax, was associated with apoptotic mitochondrial changes, signal transduction in the absence of ligand, mitochondrial transport, regulation of autophagy, and the regulation of cysteine-type endopeptidase activity (Figure 4B). Corroborating our molecular findings, the combination of THZ-P1-2 and venetoclax elicited a more pronounced degree of mitochondrial damage when compared to the monotherapies (Figure 4C).

3. Discussion

Here, we characterized PIP4K2 expression in normal and malignant myeloid cells. The high expression of PIP4K2A found in erythroid cells agrees with the initial description of this protein in the literature, reporting its abundance in erythrocytes [15], and with more recent findings showing that it participates in the process of erythroid differentiation and hemoglobin production [16,17,18]. Along the same line, we described high levels of PIP4K2A in myeloid cell lines committed to the erythro-megakaryocytic compartment (K-562, KU812, SET2, and HEL). Regarding PIP4K2B, our data suggested that it is highly expressed in hematopoietic stem cells, which deserve future study. In primary cells from AML patients, the highest levels of PIP4K2A and PIP4K2C were observed in unfavorable cytogenetic risk groups, which agrees with previous studies by our group in two independent cohorts [10,19]. Of note, both PIP4K2A and PIP4K2C have been previously associated with unfavorable clinical outcomes in AML [10].

The analysis of the association between PIP4K2 genes and sensitivity to antineoplastic drugs in ex vivo assays in AML identified targets of clinical importance in hematological malignancies, including ABL1, EGFR, VEGFR, AKT, ERK, BCL2, and FLT3 [20,21,22,23,24]. In the present study, we chose to evaluate the combination of THZ-P1-2 and venetoclax in different AML models. Venetoclax is a selective BCL2 inhibitor that has gained prominence in the treatment of lymphoid and myeloid malignancies, and its use has become the standard of care for patients with AML ineligible for intensive chemotherapy, being active in all spectra of risk groups, inducing a complete hematological remission with complete (CHR) or incomplete blood count recovery (Cri) in more than 65% of patients [25,26,27,28,29,30]. In addition, the rate of deaths in the first 30 or 60 days was lower than that reported with intensive chemotherapy; this observation represents an advantage in the context of constraint resources where these values exceed 20% [31]. Nevertheless, the median duration of the response was 17.8 months in the Vial A trial [25]. In the present study, THZ-P1-2 potentiates apoptosis induced using venetoclax in multiple AML models with different degrees of sensitivity to the drug. In a previous preliminary study with OCI-AML3 cells, similar results were observed, corroborating the current experimental findings [6].

The molecular analysis indicates that the combination of THZ-P1-2 and venetoclax induces increased levels of PARP1 and γH2AX, markers of apoptosis and DNA damage [32,33]. The TP73 gene was significantly upregulated only in the combined treatment. TP73 was the first TP53 homologous gene described [34]. The TP73 transcript evaluated in this study corresponds to the TAp73 isoform, which functionally performs functions similar to the p53 protein [35]. Furthermore, the combined treatment of THZ-P1-2 and venetoclax suppressed antiapoptotic genes, such as MCL1 and BCL2, induced in THZ-P1-2 monotherapy, which could also explain the better cell death rates. The MCL1 gene, a member of the BCL2 family, deserves attention as its relevance has been consolidated by multiple research groups as a potent mechanism of resistance to venetoclax [36,37,38]. The treatment with THZ-P1-2 alone or combination therapy suppressed the expression of BCL2A1, another antiapoptotic member of the BCL2 family associated with venetoclax resistance in AML [39,40]. Our findings suggest that, among other effects, THZ-P1-2 may act as a mitocan, which is a diverse group of anti-cancer compounds that target mitochondria. They disrupt energy production leading to the enhanced generation of reactive oxygen species along with the activation of the intrinsic pathway of apoptosis [41]. In the context of overcoming resistance to venetoclax, or improving its effects in AML, preclinical studies have shown promising results in studies analyzing molecules targeting mitochondrial respiration or structure, which include those disrupting the mitochondrial electron transport chain, mitochondrial translation, mitochondrial DNA replication, or mitochondrial protease ClpP [42].

4. Materials and Methods

4.1. Expression Data in Normal and Malignant Hematopoietic Cells

PIP4K2A (probe 205570_at), PIP4K2B (probe 201080_at), and PIP4K2C (probe 218942_at) mRNA expression data from the different hematopoietic cell populations were obtained from the GSE24759 dataset using the GEO2R platform (https://www.ncbi.nlm.nih.gov/geo/geo2r (accessed on 2 August 2023)). The gene’s expression values were obtained from cDNA microarray experiments using the Affymetrix HUG133 plus 2.0 arrays system and the data were crossed using specific identification numbers. Graphical visualizations of PIP4K2A, PIP4K2B, and PIP4K2C mRNA levels in different molecular subtypes of AML (TCGA AML study [43]) were obtained from BloodSpot (https://servers.binf.ku.dk/bloodspot/ (accessed on 2 August 2023)).

4.2. Cell Lines and Inhibitors

OCI-AML3, Kasumi-1, MOLM-13, and MV4-11 cells were obtained from Deutsche Sammlung von Mikroorganismen and Zellkulturen (DSMZ, Braunschweig, Germany). THP-1 and HL-60 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). NB4 cells were kindly provided by Prof. Pier Paolo Pandolfi (Beth Israel Deaconess Medical Center, Havard Medical School, Boston, MA, USA). U-937, K-562, KU812, and HEL cells were kindly provided by Prof. Sara Teresinha Olalla Saad (University of Campinas, Campinas, Brazil). SET-2 cells were kindly provided by Prof. Fabíola Attié de Castro (University of São Paulo, Ribeirão Preto, Brazil). Cells were cultured in appropriate media (RPMI-1640, IMDM, or alpha-MEM), supplemented with 10 or 20% of fetal bovine serum (FBS) according to the ATCC’s and DSMZ’s recommendations, plus 1% of penicillin/streptomycin, and maintained at 5% of CO₂ and 37°C. THZ-P1-2 was obtained from MedChemExpress (Monmouth Junction, NJ, USA) and diluted to 25 mM in dimethyl sulfoxide (Me₂SO₄; DMSO). Venetoclax (ABT-199) was obtained from TargetMol (Target Molecule Corp. Boston, MA, USA) and diluted to a 50 mM stock solution in DMSO.

4.3. Primary Cells

CD34⁺ progenitors were separated on MIDI-MACS immunoaffinity columns (Miltenyi Biotec, Auburn, CA, USA) from the peripheral blood of a healthy donor. Informed consent was obtained from the healthy donor and the Ethics Committee of the Institute of Biomedical Sciences of the University of São Paulo approved this study (Protocol number: 4423074; CAAE: 39510920.1.0000.5467).

4.4. Quantitative RT-PCR (qRT-PCR)

RNA was extracted using the TRIzol reagent (Thermo Fisher Scientific, Cleveland, OH, USA). cDNA was synthesized from 1 μg of RNA using a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Quantitative PCR (qPCR) was performed using a QuantStudio 3 Real-Time PCR System in conjunction with a SybrGreen System (Thermo Fisher Scientific) to assess the expression of PIP4K2, cell cycle-, apoptosis-, DNA-damage-, and autophagy-related genes (Table S2). HPRT1 and ACTB were used as reference genes. Relative quantification values were calculated using the 2^-ΔΔCT equation [44]. A negative ‘No Template Control’ was included for each primer pair. The data were illustrated using the multiple experiment viewer (MeV) 4.9.0 software [45].

4.5. Western Blotting

Protein extraction was performed using a buffer containing 100 mM of Tris (pH 7.6), 1% of Triton X-100, 2 mM of PMSF, 10 mM of Na₃VO₄, 100 mM of NaF, 10 mM of Na₄P₂O₇, and 4 mM of EDTA. Equal amounts of protein (30 μg) from the samples were subsequently subjected to SDS-PAGE in an electrophoresis device, followed by the electrotransfer of the proteins to nitrocellulose membranes. The membranes were blocked with 5% of non-fat dry milk and incubated with specific primary antibodies diluted in a blocking buffer, followed by the secondary antibodies conjugated to horseradish peroxidase (HRP). A Western blot analysis was performed using a SuperSignal^TM West Dura Extended Duration substrate system (Thermo Fisher Scientific) and G: BOX Chemi XX6 gel document system (Syngene Cambridge, Cambridge, UK). Antibodies directed against PIP4K2A (#5527), PIP4K2B (#9694), PARP1 (#9542), γH2AX (#9718), and α-tubulin (#2144) were obtained from Cell Signaling Technology (Danvers, MA, USA). Antibody directed against PIP4K2C (PA5-15289) were obtained from Thermo Fisher Scientific. Cropped gels retained important bands, but whole gel images are available in the Supplementary Materials.

4.6. Drug Sensitivity Prediction

To explore the correlation between drug response and the PIP4K2A, PIP4K2B, and PIP4K2C expression, the AUC values resulting from 165 drugs tested ex vivo assays with a Beat AML study (n = 520) were used [11,12]. For a better interpretation of the results obtained, the molecular targets (proteins/genes) of the identified drugs were listed and analyzed using Venn diagrams (http://bioinformatics.psb.ugent.be/ (accessed on 2 August 2023)).

4.7. Cell Viability Assay

In total, 2 × 10⁴ cells per well were seeded in a 96-well plate in the appropriate medium in the presence of vehicle or different concentrations of THZ-P1-2 (1.6, 3.2, or 6.4 µM) alone or in combination with venetoclax (0.3, 0.6, 1.2, or 2.5 µM) for 48 h. Next, 10 μL of methylthiazoletetrazolium (MTT, Sigma-Aldrich, St. Louis, MO, USA) solution (5 mg/mL) was added and incubated at 37 °C and 5% of CO₂ for 4 h. The reaction was stopped using 100 μL of 0.1 N HCl in anhydrous isopropanol. Cell viability was evaluated by measuring the absorbance at 570 nm. The data were illustrated using the multiple experiment viewer (MeV) 4.9.0 software [45].

4.8. Apoptosis Assay

In total, 1 × 10⁵ cells per well were seeded in a 24-well plate in an FBS-supplemented medium in the presence of vehicle or THZ-P1-2 (3.2 or 6.4 μM) alonea or in combination with venetoclax (0.6, 1.2, or 2.5 μM) for 48h. Next, the cells were washed with ice-cold phosphate-buffered saline (PBS) and resuspended in a binding buffer containing 1 μg/mL of propidium iodide (PI) and 1 μg/mL of APC-labeled annexin V (BD Pharmingen, San Diego, CA, USA). All the specimens were analyzed using flow cytometry (FACSCalibur) after incubation for 15 min at room temperature in a light-protected area. Ten thousand events were recorded for each sample.

4.9. Mitochondrial Membrane Potential Evaluation

In total, 2 × 10⁵ Kasumi-1 cells per well were seeded in a 24-well plate in 20% of FBS RPMI in the presence of vehicle, THZ-P1-2 (3.2 μM), and/or venetoclax (1.2 μM) for 24 h. Cells were then washed with PBS and resuspended in a buffer containing 5 μg/mL of JC-1 (BD), and 10,000 events were acquired using flow cytometry (FACSCalibur) after incubation for 15 min at 37°C and 5% of CO₂ in a light-protected area. The events were analyzed using the FlowJo software v.X.0.7 (Treestar, Inc., Mesa, AZ, USA).

4.10. Statistical Analysis

Statistical analyses were performed using GraphPad Prism 8 v.8.0.2 (GraphPad Software, Inc. San Diego, CA, USA). ANOVA, Bonferroni post-test, and Student’s t-test were used for comparisons as appropriate. Gene networks were constructed using the GeneMANIA database (https://genemania.org/ (accessed on 2 August 2023)). All the p-values were two-sided with a significance level of 5%.

5. Conclusions

Our study characterized the expression of PIP4K2 in the myeloid compartment of hematopoietic cells, as well as in different molecular subsets of AML. The identification of the correlation between the expression of PIP4K2 and the response to antineoplastic agents in ex vivo assays in AML exposed vulnerabilities that can be exploited in combined therapies, which could result in better therapeutic responses.