Previously it was shown that autophagy contributes to crizotinib resistance in ALK-positive anaplastic large cell lymphoma (ALK + ALCL). We asked if autophagy is equally important in two distinct subsets of ALK + ALCL, namely Reporter Unresponsive (RU) and Reporter Responsive (RR), of which RR cells display stem-like properties. Autophagic flux was assessed with a fluorescence tagged LC3 reporter and immunoblots to detect endogenous LC3 alongside chloroquine, an autophagy inhibitor. The stem-like RR cells displayed significantly higher autophagic response upon crizotinib treatment. Their exaggerated autophagic response is cytoprotective against crizotinib, as inhibition of autophagy using chloroquine or shRNA against BECN1 or ATG7 led to a decrease in their viability. In contrast, autophagy inhibition in RU resulted in minimal changes. Since the differential protein expression of MYC is a regulator of the RU/RR dichotomy and is higher in RR cells, we asked if MYC regulates the autophagy-mediated cytoprotective effect. Inhibition of MYC in RR cells using shRNA significantly blunted crizotinib-induced autophagic response and effectively suppressed this cytoprotective effect. In conclusion, stem-like RR cells respond with rapid and intense autophagic flux which manifests with crizotinib resistance. For the first time, we have highlighted the direct role of MYC in regulating autophagy and its associated chemoresistance phenotype in ALK + ALCL stem-like cells.

Transcription factors Sox2 and Oct4 are essential in maintaining the pluripotency of embryonic stem cells and conferring stemness in cancer stem-like (CSL) cells. SORE6, an in-vitro reporter system, was designed to quantify the transcription activity of Sox2/Oct4 and identify CSL cells in non-hematologic cancers. Using SORE6, we identified and enriched CSL cells in ALK-positive anaplastic large cell lymphoma (ALK + ALCL). Two ALK + ALCL cell lines, SupM2 and UCONN-L2, contained approximately 20% of SORE6+ cells, which were purified based on their expression of green fluorescent protein. We then performed functional studies using single-cell clones derived from SORE6− and SORE6+ cells. Compared to SORE6− cells, SORE6+ cells were significantly more chemoresistant and clonogenic in colony-formation assays. Sox2/Oct4 are directly involved in conferring these CSL properties, since the shRNA knockdown of Sox2 in SORE6+ significantly lowered their chemoresistance, while enforced expression of Sox2/Oct4 in SORE6− cells produced opposite effects. Using Western blots, we found that the expression and subcellular localization of Sox2/Oct4 were similar between SORE6− and SORE6+ cells. However, in SORE6+ but not SORE6− cells, Sox2 and Oct4 abundantly bound to a probe containing the SORE6 consensus sequence. c-Myc, previously shown to regulate cancer stemness in ALK + ALCL, regulated the SORE6 activity. In conclusion, SORE6 is useful in identifying/enriching CSL cells in ALK + ALCL.

Many patients diagnosed with acute myeloid leukemia (AML) relapse within two years of the initial remission. The biology of AML relapse is incompletely understood, although cancer stem-like (CSL) cells have been hypothesized to be important. To test this hypothesis, we employed SORE6, a reporter designed to detect the transcriptional activity of the embryonic stem cell proteins Oct4 and Sox2, to identify/purify CSL cells in two FLT3-mutated AML cell lines. Both cell lines contained ~10% of SORE6⁺ cells in the steady state. Compared to SORE6⁻ cells, SORE6⁺ cells exhibited more characteristics of CSL cells, with significantly higher chemoresistance and rates of spheroid formation. SORE6⁺ cells had substantially higher expression of Myc and FLT3 proteins, which are drivers of SORE6 activity. Using a mixture of SORE6⁻/SORE6⁺ cells that were molecularly barcoded, we generated an in vitro study model for AML relapse. Specifically, after ‘in vitro remission’ induced by Ara-C, both cell lines regenerated after 13 ± 3 days. Barcode analysis revealed that most of the regenerated cells were derived from the original SORE6⁺ cells. Regenerated cells exhibited more CSL features than did the original SORE6⁺ cells, even though a proportion of them lost SORE6 activity. In bone marrow samples from a patient cohort, we found that relapsed blasts expressed significantly higher levels of Myc, a surrogate marker of SORE6 activity, compared to pre-treatment blasts. To conclude, using our in vitro model, we have provided evidence that CSL cells contribute to AML relapse.

We recently demonstrated that a small subset of cells in FLT3-mutated acute myeloid leukemia (AML) cell lines exhibit SORE6 reporter activity and cancer stem-like features including chemoresistance. To study why SORE6⁺ cells are more chemoresistant than SORE6⁻ cells, we hypothesized that these cells carry higher autophagy, a mechanism linked to chemoresistance. We found that cytarabine (Ara-C) induced a substantially higher protein level of LC3B-II in SORE6⁺ compared to SORE6⁻ cells. Similar observations were made using a fluorescence signal-based autophagy assay. Furthermore, chloroquine (an autophagy inhibitor) sensitized SORE6⁺ but not SORE6⁻ cells to Ara-C. To decipher the molecular mechanisms underlying the high autophagic flux in SORE6⁺ cells, we employed an autophagy oligonucleotide array comparing gene expression between SORE6⁺ and SORE6⁻ cells before and after Ara-C treatment. ULK2 was the most differentially expressed gene between the two cell subsets. To demonstrate the role of ULK2 in conferring higher chemoresistance in SORE6⁺ cells, we treated the two cell subsets with a ULK1/2 inhibitor, MRT68921. MRT68921 significantly sensitized SORE6⁺ but not SORE6⁻ cells to Ara-C. Using our in vitro model for AML relapse, we found that regenerated AML cells contained higher ULK2 expression compared to pretreated cells. Importantly, inhibition of ULK2 using MRT68921 prevented in vitro AML relapse. Lastly, using pretreatment and relapsed AML patient bone marrow samples, we found that ULK2 expression was higher in relapsed AML. To conclude, our results supported the importance of autophagy in the relapse of FLT3-mutated AML and highlighted ULK2 in this context.

In ALK-positive anaplastic large cell lymphoma (ALK+ALCL), a small subset of cancer stem-like (or RR) cells characterized by high Myc expression have been identified. We hypothesize that NPM-ALK contributes to their high Myc expression. While transfection of NPM-ALK into HEK293 cells effectively increased Myc by inhibiting its proteasomal degradation (PD-Myc), this effect was dramatically attenuated when the full-length NPM1 (FL-NPM1) was downregulated using shRNA, highlighting the importance of the NPM-ALK:FL-NPM1 heterodimers in this context. Consistent with this concept, immunoprecipitation experiments showed that the heterodimers are abundant only in RR cells, in which the half-life of Myc is substantially longer than the bulk cells. Fbw7γ, a key player in PD-Myc, is sequestered by the heterodimers in RR cells, and this finding correlates with a Myc phosphorylation pattern indicative of ineffective PD-Myc. Using confocal microscopy and immunofluorescence staining, we found that the fusion signal between ALK and FL-NPM1, characteristic of the heterodimers, correlates with the Myc level in ALK+ALCL cells from cell lines and patient samples. To conclude, our findings have revealed a novel oncogenic function of NPM-ALK in the nucleus. Specifically, the NPM-ALK:FL-NPM1 heterodimers increase cancer stemness by blocking PD-Myc and promoting Myc accumulation in the cancer stem-like cell subset.

Mantle cell lymphoma (MCL) is an aggressive B-cell non-Hodgkin lymphoma characterized by the t(11;14)(q13;q32) cytogenetic abnormality and cyclin D1 overexpression. We have found evidence that Forkhead box M1 (FOXM1), a transcription factor with oncogenic potential, is important in the pathogenesis of MCL. Relatively high levels of FOXM1 proteins were detectable in all six MCL cell lines examined. By immunohistochemistry, we consistently found a subset of FOXM1-positive cells in MCL tumors. Analysis of two Gene Expression Omnibus (GEO) datasets from MCL patients showed that elevated FOXM1 levels significantly correlate with a worse clinical outcome. In MCL cell lines, inhibition of FOXM1 using thiostrepton or shRNA effectively triggered apoptosis and significantly reduced cell growth. FOXM1 forms a positive feedback loop with NFκB in MCL cells. Specifically, inhibition of FOXM1 dramatically decreased the protein level/transcription activity of p65, while enforced FOXM1 expression upregulated p65 and downregulated IκBα, a key NFκB inhibitor. Conversely, curcumin-mediated NFκB inhibition decreased the protein level/DNA binding of FOXM1, while transduction of a constitutively active IKKα construct into MCL cells significantly dampened the inhibitory effects of thiostrepton. Confocal microscopy revealed that FOXM1 and p65 colocalize with each other. In conclusion, FOXM1 and NFκB work collaboratively in promoting the growth and drug resistance of MCL, and FOXM1 may be a potentially useful therapeutic target.