NF-κB-Dependent Activation of the Proteasome Components, PSMD1 and PSMD3, As a Mechanism of Resistance to Imatinib

Tyrosine kinase inhibitors (TKIs) targeting BCR-ABL1 are remarkably effective therapies in chronic myeloid leukemia (CML). Despite clinical success, TKIs do not target the CML leukemic stem cell (LSC), and the majority of patients must be treated for life to maintain remission. Our previous work has shown that BCR-ABL1-independent resistance is driven by STAT3 in CML stem/progenitor cells (Eiring et al. Leukemia 2015). Unexpectedly, RNA-sequencing on TKI-resistant K562 cells (K562-R) versus parental controls (K562-S) revealed that resistance is not associated with STAT3-mediated transcription, but is rather reminiscent of TNFa signaling via NF-κB (p=0.024). Nucleocytoplasmic fractionation confirmed these findings, demonstrating higher levels of phospho-NF-κB in the nucleus of CD34⁺ cells from TKI-resistant patients (n=3) compared to newly diagnosed CML patients (n=2) or normal individuals (n=2). Surprisingly, ELISA results revealed that K562-R cells do not produce autonomous TNFa, but they do produce IL-6 (p<0.01). These data suggest that NF-κB may be driving the gene expression signature of BCR-ABL1-independent resistance, and implicate non-canonical functions for STAT3.

To better understand the mechanism by which NF-κB drives resistance, we correlated our RNA sequencing data with gene expression profiles of CML patients not responding to imatinib (McWeeney et al. Blood 2010), identifying 36 genes commonly dysregulated in both TKI-resistant cell lines and patient samples. Of the 30 upregulated genes, 21 had p65-NF-κB bound to their promoter regions via ChIP in hematopoietic cells (UCSC Genome Brower). Two of these genes are members of the ubiquitin proteasome system, including PSMD1 and PSMD3, both of which were implicated as hits in a previously published shRNA library screen for BCR-ABL1-independent resistance (Khorashad et al. Blood 2015). PSMD1 and PSMD3 are non-ATPase subunits of the 19S regulatory complex in the 26S proteasome, likely involved in proteasome substrate recognition and binding. In breast cancer, PSMD1 was shown to regulate cell growth by inducing p53 degradation (Okumura et al. 2018), whereas PSMD3 was shown to protect HER2 from degradation (Fararjeh et al. 2019). qRT-PCR confirmed upregulation of PSMD1 and PSMD3 by 3-fold and 6-fold, respectively, in K562-R cells versus parental controls in the presence of imatinib. Interestingly, according to data from The Cancer Genome Atlas (TGCA), higher levels of PSMD1 and PSMD3 mRNA correlates with a worse prognosis in acute myeloid leukemia (PSMD1, p=0.0138; PSMD3, p=0.0229). We hypothesized that PSMD1 and PSMD3 upregulation contributes to NF-κB activation and TKI resistance.

We used doxycycline-inducible shRNAs to assess the function of PSMD1 and PSMD3 in CML cell survival and TKI response. Induction of knockdown (100 ng/mL doxycycline, 72h) resulted in a reduction of PSMD1 and PSMD3 mRNA and protein by ~73% and ~77%, respectively, in K562-R cells. Importantly, immunoblot analysis revealed that knockdown of either PSMD1 or PSMD3 in TKI-resistant K562-R cells resulted in a significant reduction of phospho-NF-κB (p65), suggesting that upregulation of these proteins promotes NF-κB activation. Reduced phospho-NF-κB (p65) correlated with phenotypic effects, including reduced colony formation, increased response to TKIs as assessed in MTS assays, and increased apoptosis in both the presence and absence of imatinib. Our results suggest that NF-κB activation in TKI resistance depends on the proteasome components, PSMD1 and PSMD3, forming a positive feedback loop potentiating NF-κB signaling. Our data also suggest that specific targeting of the ubiquitin proteasome system through either PSMD1 or PSMD3 may be a novel strategy to restore TKI sensitivity in patients with BCR-ABL1-independent TKI resistance. Future studies will address the non-canonical functions of STAT3 in TKI resistance.