Effective targeting of primitive AML CD 34 + cells by the second‐generation proteasome inhibitor carfilzomib

The ubiquitin-proteasome system plays an essential role in protein homeostasis of eukaryotic cells and is important for the activation of the pro-survival transcription factor nuclear factor (NF)-KB. (Crawford & Irvine, 2013). Elevated proteasome expression and activity have been described in leukaemia (Kumatori et al, 1990; Ma et al, 2009). Moreover, NFjB activity was shown to be increased in stem cell-enriched acute myeloid leukaemia (AML) subpopulations (Guzman et al, 2001; Kagoya et al, 2014), which might be related to the increased activity of various components upstream of NF-jB, such as IRAK1 and TAK1 (MAP3K7) (Rhyasen et al, 2013; Bosman et al, 2014). Inhibition of NF-jB by proteasomal inhibition selectively induced apoptosis in AML CD34 cells (Guzman et al, 2001). Therefore, proteasome inhibition may be a promising treatment strategy to target stem cellenriched AML cell populations. The first-in-class reversible proteasome inhibitor bortezomib reduces NF-jB activity, but mainly targets the more mature CD34 AML cell fraction (Bosman et al, 2013). Here, we studied the effectiveness of the second-generation irreversible proteasome inhibitors carfilzomib and oprozomib in comparison with bortezomib on AML stem cell survival (detailed in the Data S1). First, we investigated the sensitivity of patient-derived AML CD34 cells to the proteasome inhibitors (n = 20, Table SI). After a 24-h incubation, bortezomib and carfilzomib showed similar cytotoxic effects in most of the samples (69% 19% survival, P < 0 0001; and 56% 20% survival, P < 0 0001, respectively), whereas AML cells were relatively resistant to oprozomib (93% 9% survival, P = 0 020; Fig 1A). The survival reduction was preceded by a decrease in chymotrypsin-like proteasome activity (97%, 98% and 55% reduction after 4 h incubation with 20 nM bortezomib, carfilzomib and oprozomib, respectively; Fig S1). Similar results were observed in the AML cell lines HL60, OCIAML3, THP1 and MOLM13 (data not shown). To examine whether the AML stem cell-enriched cell fractions are affected upon addition of carfilzomib and oprozomib, AML CD34 cells (n = 10) were cultured on an MS5 stromal layer in limiting dilution for 5 weeks to determine the long-term culture-initiating cell (LTC-IC) frequency. Importantly, upon a single 24-h treatment with carfilzomib, the LTC-IC frequency was reduced to 47% of the untreated control ( 24%; P = 0 003), while the LTC-IC frequency was not affected by bortezomib and oprozomib (97% 36%, P = 0 44; and 89% 39%, P = 0 21, respectively; Fig 1B). Higher carfilzomib-sensitivity was associated with a larger decrease in proteasome activity (data not shown). Staining of the quiescent HoechstPyroninYCD34CD38 cells after 24 h incubation, resulted in a slight but significant reduction of this stem cell-enriched AML cell fraction upon incubation with carfilzomib (41% quiescent CD34CD38 cells vs. 52% in untreated control; P = 0 03), but not with bortezomib or oprozomib (Fig 1C, Fig S2). The difference between carfilzomib and bortezomib could be related to the irreversible and more specific binding of carfilzomib to the proteasome (Kortuem & Stewart, 2013), which results in a decreased viability in long-term assays due to prolonged proteasome inhibition. Indeed, we observed prolonged proteasomal inhibition of carfilzomib compared to bortezomib in HL60 cells transduced with destabilized enhanced green fluorescent protein 2 (d2EGFP), which is rapidly degraded by the proteasome. Upon incubation with carfilzomib or bortezomib for 8 h, a similar accumulation of d2EGFP was induced (Fig S3), suggesting that both compounds equally inhibited the proteasome activity. Sixteen hours after removal of the inhibitors, the mean fluorescence intensity (MFI) declined in bortezomib-treated cells but not in carfilzomib-treated cells (relative MFI 1 27 vs. 1 78; P = 0 007), suggesting that carfilzomib had a longer-lasting inhibitory effect on proteasome activity (Fig S3). In contrast to AML CD34 cells, the LTC-IC frequency of normal CD34 cells was not significantly decreased by carfilzomib (91% 19% of the untreated control (P = 0 11); Fig 1D). Also, quiescent normal bone marrow (NBM) CD34CD38 cell frequency was not significantly altered upon exposure to carfilzomib (55% 30% vs. 64% 25% in controls; Fig 1E). Furthermore, the colony-forming potential of NBM cells was not significantly affected by carfilzomib (68% 51% colonies in the carfilzomib group compared to controls normalized to 100% (P = 0 053), data not shown). The different sensitivity of AML CD34 and normal CD34 cells might be explained by difference in proteasomeand NF-jB activity, which results in higher dependency of stem cell-enriched AML CD34 cells on these pathways (Kumatori et al, 1990; Guzman et al, 2001). By measuring the chymotrypsin-like activity of AML CD34 (n = 6) and NBM CD34 cells (n = 8), we observed a trend towards increased proteasome activity in AML CD34 cells (Fig 1F). Higher proteasome activity in AML cells might be the result of higher levels of proteasome complexes due to increased expression Correspondence