Abstract PO3-25-08: Artificial Intelligence-based screening of small molecule inhibitors of phosphoenolpyruvate carboxykinase 2 (PCK2), a novel cancer therapeutic target

Abstract Background: Metabolic rewiring is a hallmark of cancer that is often mediated by alterations in metabolic isozyme expression. Isozymes that are overexpressed in cancer and catalyze rate limiting steps in key metabolic processes are potential therapeutic targets. We previously identified phosphoenolpyruvate carboxykinase 2 (PCK2) as cancer dominant isoform, required for maintaining high metabolic activity and proliferation of lung, prostate, and breast cancer (BC) cells in vitro and in vivo. No PCK2 specific small molecule inhibitors exist. Methods: mRNA expression levels were assessed in the TCGA data. PCK2 MISSION shRNAs (Millipore) were used to knock-down expression of PCK2 in BC cells. Alterations in metabolic flux caused by PCK2 down-regulation were examined via 13C-labeling and mass spectrometry using a SCIEX 5500 QTRAP and SelexION instrument for mobility separation of metabolites. The AtomNetTM (Atomwise Inc.) deep convolutional neural network structure-based drug design software was used to identify candidate small molecule PCK2 inhibitors which were tested in PCK1 and PCK2 in vitro enzyme activity assays. The lead compound was assessed for drug-like properties by QickPropV4.0.001w, and its growth inhibitory activity was tested in MDA-MB-321 and BT-549 cells. Results: Analysis of mRNA levels in TNBC samples revealed a significant decrease in PCK1 expression compared to normal tissues, while PCK2 expression remained elevated or even increased. Quantification of mRNA levels in TNBC cell lines confirmed the higher expression of PCK2 compared to PCK1. To assess functional significance of PCK2 expression, we downregulated PCK2 using shRNA constructs in TNBC cell lines. Western blot analysis confirmed efficient knockdown of PCK2 protein without affecting PCK1 expression. Clonogenic assays demonstrated that PCK2-depleted BT-549 and MDA-MB-231 cells exhibited impaired proliferation, with reduced colony number and size compared to control cells. Metabolic analysis of PCK2-depleted TNBC cells demonstrated decreased levels of glycolytic and TCA intermediates and reduced flux through pyruvate carboxylase, indicating altered metabolic pathways upon PCK2 depletion. To identify potential inhibitors of PCK2, virtual screening was conducted using compound structures. We used a PCK2 protein structure homology model and designated the 3-mercaptopicolinic acid-binding site of the PCK1 structure as the site of interest to perform virtual drug screening of 7 million diverse compounds. The first screen yielded 86 structures to be tested in vitro, of which 5 demonstrated 16% PCK2 inhibition. The top hit (IC50=2.4µM) was used to refine a second round of in silico screen that yielded 87 second-round candidates that were assessed for PCK2 inhibition in vitro. 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate emerged as the lead compound. QickProp predicted favorable absorption, distribution, metabolism, and excretion properties, but relatively low cell permeability (apparent Caco-2 and MDCK2 cell permeabilities 18 nm/sec and 8 nm/sec, respectively). The specificity of 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate was tested in PCK1 and PCK2 enzymatic assays and revealed IC50 of 500nM on PCK1 and an IC50 of 3 nM on PCK2. We next tested the cell growth inhibitory effect of 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate on parental MDA-MB-231 and BT-549 cells and their PCK2 knockdown clones. Cell growth and viability significantly decreased in the MDA-MD-231 and BT-549 parental cells (the IC50 for MDA-MB-231 was 173 μM) and in shRNA non-targeting control cells, but not in the PCK2 depleted clones, demonstrating PCK2-dependent inhibitory effect. Conclusion: We identified 3-(3,4-dihydroxyphenyl)-2-hydroxypropanoate as a high affinity selective PCK2 enzyme inhibitor. This compound also has significant growth inhibitory activity in BC cell lines in vitro and represents a novel therapeutic lead compound. Citation Format: Alejandro Rios Hoyo, Naing Lin Shan, Hao-Kuen Lin, Gerson Espinoza Campos, Mostafa Ahmed, Sheila Umlauf, Rebecca Cardone, Yulia V. Surovtseva, Vignesh Gunasekharan, Lajos pusztai. Artificial Intelligence-based screening of small molecule inhibitors of phosphoenolpyruvate carboxykinase 2 (PCK2), a novel cancer therapeutic target abstract. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO3-25-08.