MUC1‐C drives myeloid leukaemogenesis and resistance to treatment by a survivin‐mediated mechanism

Abstract Acute myeloid leukaemia (AML) is an aggressive haematological malignancy with an unmet need for improved therapies. Responses to standard cytotoxic therapy in AML are often transient because of the emergence of chemotherapy‐resistant disease. The MUC1‐C oncoprotein governs critical pathways of tumorigenesis, including self‐renewal and survival, and is aberrantly expressed in AML blasts and leukaemia stem cells (LSCs). However, a role for MUC1‐C in linking leukaemogenesis and resistance to treatment has not been described. In this study, we demonstrate that MUC1‐C overexpression is associated with increased leukaemia initiating capacity in an NSG mouse model. In concert with those results, MUC1‐C silencing in multiple AML cell lines significantly reduced the establishment of AML in vivo. In addition, targeting MUC1‐C with silencing or pharmacologic inhibition with GO‐203 led to a decrease in active β‐catenin levels and, in‐turn, down‐regulation of survivin, a critical mediator of leukaemia cell survival. Targeting MUC1‐C was also associated with increased sensitivity of AML cells to Cytarabine (Ara‐C) treatment by a survivin‐dependent mechanism. Notably, low MUC1 and survivin gene expression were associated with better clinical outcomes in patients with AML. These findings emphasize the importance of MUC1‐C to myeloid leukaemogenesis and resistance to treatment by driving survivin expression. Our findings also highlight the potential translational relevance of combining GO‐203 with Ara‐C for the treatment of patients with AML.


| INTRODUCTION
Acute myeloid leukaemia (AML) is a lethal haematological malignancy characterized by the emergence of a clonal population of primitive myeloid cells that exhibit a pattern of dysregulated growth and self-renewal. 1 Responses to cytotoxic agents, such as Cytarabine (Ara-C) and daunorubicin, are often observed in patients with AML; however, disease relapse is common because of the emergence of chemotherapy-resistant disease. 2 In this context, progression of AML is associated with genetic and epigenetic changes that promote aggressiveness and resistant disease.
There is a significant need to identify critical pathways that govern leukaemic progression and offer potential targets for novel therapeutics.
Mucin 1 (MUC1) is a heterodimeric protein that is aberrantly expressed in cancer cells, including AML blasts. The oncogenic MUC1-C subunit drives critical hallmarks of malignant cells, including cell proliferation, resistance to apoptosis, self-renewal, and tissue invasion. [3][4][5][6] Intriguingly, we have demonstrated that MUC1 is uniquely expressed by AML stem cells as compared to normal haematopoietic stem cells. 7 Primary AML cells expressing high levels of MUC1 efficiently induce leukaemic engraftment in a xenogeneic murine model, whereas MUC1 low expressing cells isolated from bone marrow of patients with active AML engraft normal haematopoietic elements.
The oncogenic function of the transmembrane MUC1-C subunit is dependent on the formation of homodimers, which are required for translocation to the nucleus and interactions with downstream effectors. [8][9][10][11] The MUC1-C cytoplasmic domain is phosphorylated by c-Src and receptor tyrosine kinases, and interacts with effectors, such as b-catenin and NF-jB, that have been linked to transformation. In AML, MUC1-C associates with the b-catenin/TCF4 complex, which regulates cell proliferation and differentiation. 5,12 Accumulation of b-catenin in the cytoplasm promotes its translocation to the nucleus as a cofactor for transcription factors of the T-cell factor (TCF) family and activates the transcription of Wnt/b-catenin target genes. MUC1-C facilitates the nuclear translocation of dephosphorylated active b-catenin that is necessary for inducing the expression of cyclin D1, MYC and survivin, a negative regulator of apoptosis. [13][14][15] Survivin also plays a role in the proliferation and survival of leukaemia induced by the internal tandem duplication of FLT3. 16 Moreover, survivin is highly expressed in AML progenitor cells and is predictive of poor clinical outcomes in patients with AML. 17 In this study, we demonstrate that MUC1-C signalling is critical for leukaemia progression and sensitivity to the cytotoxic agent Ara-C by a survivin-mediated mechanism. These findings emphasize the importance of MUC1-C as a target in AML and support targeting of MUC1-C with GO-203 in combination with Ara-C for the treatment of patients with AML. shRNA (MUC1shRNA; Sigma) or with a scrambled control shRNA vector (CshRNA; Sigma). 18 Alternatively, MUC1 knockdown was achieved using CRISPR/Cas9 technology as described. 19 For overexpression of MUC1-C, cells were transduced with lentiviruses expressing pHR-CMV-GFP (vector) or MUC1-C (MUC1-C). Primary cells were transduced with help of Transdux transduction reagent (System Biosciences Cat# LV850A-1) using a previously concentrated viral particles, which were obtained with help of Lenti-X concentrator (clontech Cat #631231). For qRT-PCR analysis the total RNA was isolated from above transduced cells, using RNAeasy kit (Qiagen Cat#74104) and cDNA synthesis was done with 2 lg of total RNA using High Capacity RNA transcription kit (Thermo Fisher Cat#4368814). The cDNA was diluted further for Real Time PCR using SYBR green master mix Bone marrow aspirates samples were obtained from patients with newly diagnosed AML as per an institutionally approved protocol. Mononuclear cells were isolated by ficoll density centrifugation.

| AML patient derived cells and cell lines
For assessment of active b-catenin, CD34+ cells were isolated using the MiniMacs CD34 cell isolation kit (Miltenyi Biotec). The bulk AML cell population was used for in vivo experiments. Cells were treated with the MUC1-C inhibitor, GO-203, and as a control, the CP-2 peptide. 8

| Quantitative RT PCR
Quantitative real-time (RT)-PCR was performed on cDNA synthesized from total cell RNA using the Thermoscript RT-PCR system (Invitrogen). The SYBR green qPCR assay (Applied Biosystems) was used with diluted cDNA. The samples were amplified with the ABI Prism 7000 Sequence Detector (Applied Biosystems). Forward and

Key Points
• MUC1-C is essential for AML establishment in an NSG mouse model.

| Leukaemia engraftment in NSG mice
Bone marrow derived AML cells and AML cell lines were inoculated retro-orbitally into sublethally irradiated (300 rads) NOD-SCID IL2Rgammanull (NSG; 6 week old female) mice (Jackson Laboratories). After sacrifice, bone marrow and spleen cells were harvested and red blood cells (RBC) were removed using RBC lysis buffer (Sigma). Human AML engraftment was detected by staining cells with PE-conjugated anti-hCD45 and, as a control, FITC-conjugated antimouse mCD45. In certain experiments, the cells were also analysed for hCD34, hCD11C, hCD19 or hCD20 by multichannel flow cytometry using CellQuest, Diva or Kaluza software. In order to confirm AML blast morphology, cytospins were made from BM cells. The cells were then fixed in methanol and stained using the standard Wright Giemsa protocol. The cells were visualized with contrast light microscopy (Olympus AX70 microscope) using an oil immersion objective lens (9100).

| Immunoblot analysis
Cell lysates were prepared as described. 20

| Analysis of intracellular protein expression by flow cytometry
For b-catenin expression -isolated leukaemia CD34+ progenitors were permeabilized with a saponin-based reagent (eBioscience).
The cells were then stained with purified anti-active b-catenin (Millipore) for 1 hour followed by secondary labelling of the cells with  In addition, dead cells were detected by addition of 0.1 mg/mL propidium iodide (PI) and apoptotic cells were detected by Annexin V (FITC) apoptosis detection kit (BD Biosciences) using flow cytometry.

| Microarray gene expression data
Gene expression and clinical data were analysed for previously expression. Overall survival of both low and high expression groups were examined using "survival" package using R version 3.3.1.

| Statistical analysis
Data of two tested groups were compared using the Student's t-test. P-Values less than. 05 were considered significant.

| MUC1-C overexpression leads to increased leukaemogenicity in NSG mice
To investigate the role of MUC1-C in leukaemia induction in vivo,

| MUC1 silencing leads to reduced nuclear translocation of b-catenin and decreased survivin expression
To evaluate the mechanism by which MUC1-C-mediated signalling promotes leukaemogenicity, we assessed the effect of silencing F I G U R E 1 MUC1-C overexpression leads to increased leukaemogenicity in NSG mice. MUC1-C was overexpressed in MOLM14 cells. A, The cells were harvested and lysates were immunoblotted for the expression of MUC1-C using anti-CT2 monoclonal antibody. MCF7 cells were used as positive control. The cells were then inoculated into sublethally irradiated NSG mice at a low dose of 1000 cells/mouse. 21 d post inoculation the mice were killed and BM cells were isolated and analysed for human CD45 expression. B, The results are expressed as percentages of hCD45+ leukaemia cells in the BM of individual mice inoculated with AML cells with MUC1-C overexpression and control vector. The horizontal bar represents the mean percentage of hCD45+ cells (n = 3; P < .05). C, MUC1-C was overexpressed in AML cells obtained from a BM aspirate of a patient with AML. The cells were then inoculated into sub-lethally irradiated NSG mice 5 9 10 5 cells/mouse. Ninety days post inoculation the mice were killed and BM cells were isolated and analysed for human CD45 expression. The results are expressed as percentages of hCD45+ leukaemia cells in the BM of individual mice inoculated with AML cells with MUC1-C overexpression and control vector. The horizontal bar represents the mean percentage of hCD45+ cells (n = 3; P < .05). D, Representative FACS plots of mice inoculated with MUC1-C overexpressed and control patient derived AML cells. E, Cytospins were prepared from bone marrow cells isolated from mice inoculated with MUC1-C overexpressed AML cells and control cells. The cytospins were then stained using standard Giemsa staining protocol. BM morphology of representative mice is shown  These results were confirmed using THP1/MUC1shRNA cells ( Figure 5E, Figure S2D). Inoculation of mice with MUC1shRNA/Vector cells did not lead to leukaemia induction. However, in experiments with the MUC1shRNA/surviving cells, 3/6 mice showed AML engraftment with mean levels of 2% blast involvement ( Figure 5F, G).

AML cells leads to enhanced leukaemia induction
These findings indicate that MUC1 silencing leads to decreases in leukaemogenicity, at least in part by survivin downregulation.

| MUC1-C inhibition leads to increased susceptibility to cytarabine via down-regulation of survivin
We further sought to determine whether targeting MUC1-C with the resultant decrease in survivin levels would confer increased susceptibility to Ara-C. In this way, MOLM-14 cells were treated with increasing doses of Ara-C alone, GO-203 alone or the combination of both agents. Assessment of cell viability after 48 hours demonstrated dose-dependent cytotoxicity for each agent alone ( Figure 6A, B).
Interestingly, exposure of MOLM-14 cells to the combination of GO-203 and Ara-C for 48 hours showed a statistically significant reduction in cell viability compared to that obtained with each agent alone ( Figure 6C). These findings were further confirmed using Annexin/PI staining ( Figure 6D). Similar results were obtained with MV4-11 AML cells ( Figure S1A-C), indicating that the GO-203/Ara-C combination is synergistic. To further assess this interaction, we FACS plots of BM cells analysed for AML engraftment of representative mice. MUC1-C was silenced in THP1 AML cells using lentiviral infection with a shRNA hairpin sequence against MUC1-C. Subsequently, GFP tagged survivin or control vector genes were overexpressed in these cells using lentiviral transduction. GFP+ cells were then isolated using flow cytometric sorting and (E) underwent Western blot analysis for survivin expression. b-actin was used as a control. Subsequently, 10 9 10 3 cells/mouse were inoculated into NSG mice using retro-orbital injections for 21 d. Bone marrow cells were then analysed for AML engraftment as described above.  Figure 6G).

| Low MUC1 and survivin expression in human AML is associated with increased overall survival
We hypothesized that increased MUC1 expression on human AML contributes to pathogenesis and predicted that AML with lower expression of MUC1 would be associated with better clinical outcomes. Consistent with this hypothesis, analysis of previously  Figure 1C). Our studies have also assessed the role of MUC1-C mediated signalling on the b-catenin/WNT pathway and its role in inducing survivin, a downstream effector critical for leukaemia biology. 17 The MUC1-C cytoplasmic domain has been shown to interact with bcatenin in several carcinoma cell, but not AML, models. 12 MUC1-C associates with the b-catenin/T-cell factor 4 (TCF4) complex, which regulates cell differentiation, proliferation and apoptosis. 5 Accumulation of b-catenin in the cytoplasm favours its translocation to the nucleus as a cofactor for TCF family transcription factors and thereby activates the transcription of Wnt/b-catenin target genes. 44 The MUC1-C oncoprotein is known to facilitate the nuclear translocation of active b-catenin to the nucleus, necessary for downstream signalling pathways. 12,45,46 Here, we have demonstrated that silencing MUC1 in AML cells leads to reduced translocation of b-catenin to the nucleus. Furthermore, targeting MUC1-C with GO-203 led to F I G U R E 7 Low MUC1 and survivin Expression in Human AML is Associated with Increased Overall Survival. Overall survival of 260 AML patients with diverse cytogenetic and molecular abnormalities was assessed. A, Patients were stratified into low MUC1 and high MUC1 expression groups based on an optimal threshold (234 low; 26 high) determined by microarray analysis from an independent dataset. The significance measures are based on univariate analysis (P = .04). B, Patients were also stratified into low survivin (BIRC5) and high survivin (BIRC5) expression groups based on an optimal threshold (231 low; 29 high) determined by microarray analysis from an independent dataset. The significance measures are based on univariate analysis (P = .01). C, Overall survival was assessed in patients with concomitant low expression of MUC1 and survivin compared to the remainder of AML patients in this group. The threshold was determined based on overlap of previously described single gene expression stratification (52 Low MUC1/Low BIRC5; 208 other MUC1/BIDC5 expressers). The significance measures are based on univariate analysis (0.001) significantly decreased levels of active b-catenin in primary CD34+ AML cells.
Survivin (BIRC5), a member of the inhibitor of apoptosis protein (IAP) gene family, inhibits apoptosis, enhances proliferation and promotes angiogenesis. [14][15][16]30,47 Survivin is highly expressed in AML progenitor cells and is predictive of poor clinical outcomes in patients with AML. 17 Survivin is thus a critical target in AML that is regulated, at least in part, by the b-catenin/TCF4 complex. 48