PKM splice-switching ASOs induce upregulation of dual-specificity phosphatases and dephosphorylation of ERK1/2 in hepatocellular carcinoma

Pyruvate kinase M isoform 2 (PKM2) is preferentially expressed in nearly all cancers. It primarily functions as the last enzyme in glycolysis, but has other reported non-canonical functions, including recruiting transcription factors to oncogenes, and phosphorylating proteins. We previously described antisense oligonucleotides that disrupt alternative splicing of PKM pre-mRNA (PKM-ASOs), resulting in PKM2-to-PKM1 isoform switching in hepatocellular carcinoma (HCC), which reduces HCC growth. PKM1 has higher enzymatic activity than PKM2, which potentially drives metabolism away from macromolecule synthesis, and may explain decreased HCC growth upon PKM-ASO treatment. As PKM-ASOs also reduce PKM2 levels, how PKM splice-switching inhibits HCC cell proliferation was unclear. Here, we characterized the individual consequences of altering PKM1 or PKM2 protein levels in HCC, and observed that reducing PKM2 alone was sufficient to decrease HCC cell proliferation, whereas overexpressing PKM1 had no effect. Moreover, increasing PK activity via a small-molecule PKM2 activator had no effect on HCC cell proliferation, suggesting that PKM-ASOs affect PKM2’s non-metabolic functions. Transcriptomic and RT-qPCR analyses of HCC cells treated with PKM-ASO or PKM2-siRNA revealed upregulation of dual-specificity phosphatase 2 (DUSP2) and other related DUSPs, which act directly on ERK1/2 in the MAPK signaling pathway. Luciferase reporter assays demonstrated that PKM-ASO treatment activated the DUSP2 promoter, which correlated with decreased ERK1/2 phosphorylation. Lenvatinib is a second-line HCC therapy that indirectly reduces ERK1/2 phosphorylation, and combined treatment with PKM-ASOs inhibited proliferation of HCC cells more than either treatment alone. In summary, our results reveal a mechanism by which PKM-ASOs affect PKM2 dependency in HCC. To investigate changes in transcription after altering PKM1 or PKM2 protein levels in HCC cells, we generated a dox-inducible SNU449-PKM1OE cell line and treated with either 0.25 ug/ml dox, 5 nM of siPKM2, or 60 nM of a PKM2-to-PKM1 splice switching antisense oligonucleotide (PKM-ASO) for 48 hr. We included a lipofectamine-only control, and an siCtrl. We then performed bulk RNA-Seq analysis. There were 3 biological replicates for each group. We then used the RNA-seq data to perform comparative gene expression analysis between the treatment groups.