Ara-C is the key standard agent used in the treatment of AML. Pediatric patients with DS-AMKL are highly responsive to Ara-C in comparison to non DS-AMKL [
31]. In the present study, this type of response was seen in the DS-AMKL cell line CMK cells compared to the non-DS-AMKL CMS cell line. However, the DS-AMKL cell line CMY was 25-fold more resistant to Ara-C compared to CMK cells, confirming previous reports [
7]. Increased sensitivity to Ara-C was reported to be due to increased expression of the enzyme cystathionine-[β]-synthase (CBS) [
32], and to the presence of somatic mutations in the transcription factor gene
GATA-1 in leukemia cells of patients with DS-AMKL [
33]. The combination of trisomy 21 and the
GATA-1 mutations increases the amount of active Ara-C metabolites present in DS-AMKL cells, thereby enhancing its cytotoxicity [
34]. However, the mechanisms of Ara-C resistance in the DS-AMKL CMY cells are not known. In this study, the presence of a high percentage of ALDH
bright cells in this cell line may contribute to its Ara-C resistance. ALDH is highly expressed in LSCs [
35], which are thought to contribute to drug resistance and relapse in AML. Therefore, in order to prevent disease recurrence, this subpopulation of stem cells must be eliminated. BTZ was previously shown to induce apoptosis in the “stem cell like” blasts from AML patients [
9], so it was used in the present study to target the ALDH-positive “stem-like” cells in the Ara-C resistant CMY cell line, in comparison to the ALDH inhibitor DSF/Cu
2+. While approximately 1% of the ALDH
bright cell population of CMY cells remained resistant to BTZ, they were very sensitive to DSF/Cu
2+, indicating that DSF/Cu
2+, but not BTZ, is equally cytotoxic to both the AML proliferating cells, as well as to the ALDH
bright “stem-like” cells. This is in agreement with previous studies demonstrating that DSF targets ALDH-positive cancer stem cells in breast cancer [
36], in glioblastoma [
29,
37], and in Non-Small Cell Lung Cancer (NSCLC) [
38].
To further understand the mechanism of BTZ resistance in the DS-AMKL cell lines, we generated BTZ-resistant variants of both CMY and CMK cell lines and detected a novel A362C mutation in the
PSMB5 exon 2 in the CMY-BR cells, causing a change from glutamine to proline (Q62P). We confirmed that this mutation caused BTZ resistance in transformed HEK293A cells. However, this mutation did not completely abrogate the CT-like proteasome activity, and was accompanied by an upregulation of the β5 subunit on the mRNA and protein levels. This is consistent with previous studies showing that cells with acquired BTZ resistance have upregulated mutant β5 subunit, which serves as a compensatory mechanism to retain sufficient CT-like proteasome activity [
39]. To find compounds that are clinically used in the treatment of DS-AMKL and would overcome BTZ resistance in this model, we tested the second-generation proteasome inhibitor CFZ, Ara-C, VP-16, daunorubicin, in comparison to MG 132 and DSF/Cu
2+. Interestingly, the CMY-BR cells became more resistant to Ara-C at higher concentrations. For Ara-C to function as an anti-tumor agent, it is sequentially phosphorylated, eventually to Ara-C triphosphate (Ara-CTP), which is incorporated into DNA strands during the S phase of the cell cycle, thereby inhibiting DNA synthesis and causing S-phase arrest [
40]. In the present study, Ara-C induced S-phase arrest in the parent CMY cell line, confirming previous studies in AML cells [
9]. However, in the CMY-BR variant, only G1 arrest was observed. BTZ was previously reported to decrease the levels of CCND1, CDK4 and CDK2 (required for G1/S transition), while it increases those of p27 and p21 (inhibitors of CDK2) [
9,
41]. Cells would subsequently accumulate in G1, as observed in the present study. G1 arrest, here, may protect the cells from Ara-C effects, thus supporting previous findings using MCL cell lines, in which G1 arrest by abemacilib (a CDK4/6 inhibitor) protected the cells from Ara-C damage [
42]. Our results may also explain the failure of combined BTZ and Ara-C therapy to show substantial improvement in the OS in pediatric patients with relapsed/refractory or secondary AML [
43]. Our results show that the BTZ resistant variants were cross-resistant to CFZ and MG 132 but sensitive to DSF/Cu
2+, which suggests that DSF/Cu
2+ interferes with the proteasome through a different mechanism than CFZ and BTZ. DSF/Cu
2+ induced ubiquitination, apoptosis and PARP cleavage similar to BTZ, but it did not inhibit the CT-like activity. This is in contrast to previous studies using breast cancer cell lines [
27] and cultured glioma stem cells [
37] that showed that DSF/Cu
2+, and Cu
2+ alone, but not DSF, caused inhibition of CT-like activity, suggesting that the proteasome effect of DSF/Cu
2+ is due to copper [
44]. Our results support the notion that unlike BTZ, DSF/Cu
2+ complexes target the 26S proteasome rather than the 20S [
45]. A likely target could be the JAB1/MPN/Mov34 metalloenzyme (JAMM) domain of the POH1 subunit within the lid of the 19S proteasome, as was proposed in [
46]. POH1, a member of the JAMM domain deubiquitinases, is necessary for activity of the 26S proteasome [
47] and for cell viability [
48]. Inhibition of the 26S proteasome-associated deubiquitinases result in apoptosis and in vivo inhibition of tumor progression [
49]. Furthermore, proteasome inhibition leads to the accumulation of misfolded proteins and possible toxic protein aggregates, which typically induces the unfolded protein response and heat shock protein activation [
50]. Other reported mechanisms of DSF/Cu
2+ include the induction of reactive oxygen species, leading to pro-apoptotic JNK activation [
51], as well as the inhibition of the HER2 pathway in breast cancer [
36].