ALOX5 exhibits anti-tumor and drug-sensitizing effects in MLL-rearranged leukemia

MLL-rearranged acute myeloid leukemia (AML) remains a fatal disease with a high rate of relapse and therapeutic failure due to chemotherapy resistance. In analysis of our Affymetrix microarray profiling and chromatin immunoprecipitation (ChIP) assays, we found that ALOX5 is especially down-regulated in MLL-rearranged AML, via transcription repression mediated by Polycomb repressive complex 2 (PRC2). Colony forming/replating and bone marrow transplantation (BMT) assays showed that Alox5 exhibited a moderate anti-tumor effect both in vitro and in vivo. Strikingly, leukemic cells with Alox5 overexpression showed a significantly higher sensitivity to the standard chemotherapeutic agents, i.e., doxorubicin (DOX) and cytarabine (Ara-C). The drug-sensitizing role of Alox5 was further confirmed in human and murine MLL-rearranged AML cell models in vitro, as well as in the in vivo MLL-rearranged AML BMT model coupled with treatment of “5 + 3” (i.e. DOX plus Ara-C) regimen. Stat and K-Ras signaling pathways were negatively correlated with Alox5 overexpression in MLL-AF9-leukemic blast cells; inhibition of the above signaling pathways mimicked the drug-sensitizing effect of ALOX5 in AML cells. Collectively, our work shows that ALOX5 plays a moderate anti-tumor role and functions as a drug sensitizer, with a therapeutic potential, in MLL-rearranged AML.

another factor involved in both leukemogenesis and therapeutic response 16 . K-RAS mutations have been reported in many AML cases and are associated with prognosis and chemotherapy resistance 17,18 . It holds great potential to cure AML by either targeting these factors directly or targeting a common upstream regulator that controls all these factors in AML.
The arachidonate 5-lipoxygenase gene (ALOX5) encodes a non-heme iron-containing enzyme of the lipoxygenase family. It catalyzes the production of leukotrienes (LTs) and reactive oxygen species (ROS) from arachidonic acid 19,20 . ALOX5 is known to be involved in various physiological and pathological processes, including oxidative stress response, inflammation and cancer [21][22][23][24][25] . It was reported previously that loss of Alox5 impairs the function of leukemic stem cells (LSCs) in BCR-ABL-induced chronic myelogenous leukemia (CML) 26 . To our surprise, our gene profiling reveals that ALOX5 expression is particularly down-regulated in MLL-rearranged AML. Both in vitro and in vivo studies were carried out to investigate the effects and underlying mechanism of ALOX5 in AML pathogenesis and drug response.
To understand the mechanisms of the repression of ALOX5 in MLL-rearranged AML, we conducted chromatin immunoprecipitation (ChIP) assays. We found no significant enrichment of MLL fusion proteins at the promoter region 32 of the ALOX5 locus ( Fig. 1e-g). It is known that gene silencing mediated by the Polycomb repressive complex 2 (PRC2) and cofactors, e.g. EZH2 and SIN3A, is essential for MLL-rearranged AML 33 . Here we show a significant enrichment of EZH2 and SIN3A at the ALOX5 promoter region. Histone H3 lysine 27 tri-methylation (i.e., H3K27me3), a marker for repressive transcription associated with EZH2 34,35 , also exhibited high enrichement at these genomic locus ( Fig. 1e-g). Therefore, it is highly likely that the PRC2 complex mediates the transcriptional repression of ALOX5 in MLL-rearranged leukemia.
Anti-tumor effect of Alox5 in MLL-rearranged AML. To assess the pathological role of Alox5 in MLL-rearranged AML, we cloned Alox5 CDS into MSCV-PIG retroviral vector, and then co-transduced MSCVneo-MLL-AF9 (MA9) with MSCV-PIG-Alox5 (Alox5) or MSCV-PIG (Ctrl) into mouse BM progenitor cells for in vitro colony-forming/replating assays. We showed that along with the increased number of passages, Alox5 overexpression showed a more significant degree of repression on MLL-AF9-induced colony forming (Fig. 2a). We also analyzed cell viability of those retrovirus transduced progenitor cells and showed that forced expression of Alox5 significantly suppressed cell viability (Fig. 2b). Knockdown of ALOX5 with siRNA did not show significant alterations in the viability of MONOMAC-6/t(9;11) cells (Supplementary Figs 1a-c and 3a,b).
In order to determine the in vivo effect of Alox5 in leukemogenesis, we performed primary BM transplantation (BMT) assay first and found that forced expression of Alox5 showed no significant influence on overall survival (medium overall survival: 108 days for the MA9 + Ctrl group vs. 96 days for MA9 + Alox5; P = 0.556, log-rank test) (Fig. 2c). We then isolated the leukemic blast cells from primary MLL-AF9-leukemic mice and transduced the cells with Alox5 or control retrovirus, and performed secondary BMT assays. Notably, Alox5 overexpression significantly delayed the progression of MLL-AF9 AML in secondary BMT recipients (median overall survival, 42.5 days (MA9_Alox5) vs. 37 days (MA9_Ctrl); P = 0.018, log-rank test) (Fig. 2d). While all the leukemic mice died from AML, overexpression of Alox5 resulted in a significant decrease in peripheral white blood cell count (Fig. 2e) and spleen size (Fig. 2f). Moreover, through flow cytometry analysis, we found that although the Alox5 overexpressing group and the control group have similar degrees of engraftment of MLL-AF9 donor cells in BM at their end points, the former has a significantly lower population of the Mac-1 + /Gr-1 + leukemic blast cells than the latter (Fig. 2g,h). Tissue staining showed that, compared with the control group, the BM cells in the Alox5 overexpressing group were more differentiated, consistent with the flow cytometry results, and the infiltration of leukemic cells into spleen, liver, and peripheral blood was less severe (Fig. 2i). Thus, our data indicated that Alox5 exhibited a moderate anti-tumor effect in the maintenance of MLL-rearranged AML and restrains leukemic cell infiltration.
Forced expression of Alox5 sensitizes MLL-rearranged AML cells to standard chemotherapy. The anti-tumor role of Alox5 implies a therapeutic potential in treating AML. Since restoration of Alox5 expression/function alone showed only moderate inhibitory effect on AML progression, we sought to investigate whether Alox5 restoration could facilitate chemotherapeutic response and thus yield a more effective therapeutic effect. We first used lentivirus to overexpress human ALOX5 in MONOMAC-6/t(9;11) AML cells, and treated the transduced cells with doxorubicin (DOX) or cytarabine (Ara-C). Our results showed that overexpression of ALOX5 significantly sensitized MONOMAC-6 cells to DOX and especially Ara-C treatment (Fig. 3a-c; Supplementary Fig. 3c,d).
Similarly, in mouse BM progenitor cells transduced with MLL-AF9, co-expression of Alox5 remarkably enhanced the inhibitory effect of DOX or Ara-C on cell viability and colony forming capacity (Fig. 4a-c). We further tested the drug sensitizing effect of Alox5 in vivo. We transplanted MLL-AF9 AML cells with forced expression of Alox5 (i.e., MA9_Alox5) or without (i.e., MA9) into secondary recipient mice, and after the onset of leukemia, the recipient mice were treated with or without DOX + Ara-C (i.e. "5 + 3" regimen 5 ). As shown in Fig. 4d, while DOX + Ara-C treatment alone only moderately (though statistically significantly) improved survival in mice carrying MLL-AF9 AML, forced expression of Alox5 dramatically improved the response of MLL-AF9 AML to the DOX + Ara-C treatment (Fig. 4d). As a result, amazingly over 70% of mice in the MA9_ Alox5 + DOX/Ara-C group survived over 200 days, while all mice in the other groups died from AML within 70 days (Fig. 4d). The infiltration of leukemic cells into the spleen, liver, and peripheral blood was almost completely blocked in the MA9_Alox5 + DOX/Ara-C group of mice (Fig. 4e). Therefore, restoration of Alox5 holds great potential in improving chemotherapeutic response in MLL-rearranged AML patients.
Signaling pathways correlated with Alox5 overexpression in AML. To delineate the potential molecular mechanism underlying the anti-tumor and drug-sensitizing effects of Alox5, we performed RNA sequencing (RNA-seq) of two pairs of mouse BM leukemic blast cells collected from the MA9_Ctrl and MA9_ Alox5 mice in secondary BMT assays shown in Fig. 2d. Through gene set enrichment analysis (GSEA) 36 , we  found that Il-2/Stat5 signaling and K-Ras pathway were significantly suppressed in MA9_Alox5 AML cells as compared with MA9_Ctrl AML cells (Fig. 5a,b). Stat5 and K-Ras function as critical oncogenes in both solid tumors and leukemia, and both STAT5 and K-RAS signaling pathways were known to be closely associated with oncogenesis and drug sensitivity 15,16,37 . c-MYC is a downstream target gene shared by STAT and K-RAS signaling pathways 38,39 . We analyzed two independent AML patient databases and found a significant negative correlation between the expression levels of ALOX5 and c-MYC (Fig. 5c,d). In BM leukemic blast cells of MLL-AF9 leukemic mice, overexpression of Alox5 suppressed the expression of c-Myc and Flt3, an upstream key regulator of both Stat5 and K-Ras signaling 40 (Fig. 5e,f; Supplementary Fig. 3e-g). In order to determine whether repression of STAT5 and/or K-RAS pathways could mimic the drug sensitizing effect of ALOX5, we treated MONOMAC-6 cells with a combination of DOX or Ara-C, together with control, STAT5 inhibitor sc-355979 41 , and/or RAS signaling inhibitor Tipifarnib 42 . Results showed that a combined treatment of both sc-355979 and Tipifarnib significantly sensitized the cells to chemotherapy (Fig. 5g,h). In a data set composed of 82 AML samples (including 26 MLL-rearranged AMLs) 43 , those bearing K-RAS mutations (n = 5) are associated with a lower ALOX5 level ( Supplementary Fig. 2a), indicating a potential positive feedback between K-RAS signaling and ALOX5. Noticeably, the only AML sample with mutated K-RAS that has relatively higher ALOX5 level also bears FLT3 mutation. The remarkable suppression of ALOX5 in MLL-rearranged AML bearing K-RAS mutation was further verified in human CD34 + derived MLL-AF9 cell lines ( Supplementary Fig. 2b). Therefore, our results suggest that Alox5 exerts its anti-tumor and drug sensitizing effects in MLL-rearranged AML through suppressing the Stat and K-Ras signaling pathways.

Discussion
In contrast to the oncogenic role of ALOX5 reported previously in CML 26 , here we show that ALOX5, suppressed by PRC2 at the transcription level, exhibits a moderate anti-tumor effect both in vitro and in vivo in MLL-rearranged AML, in which ALOX5 expression is especially repressed. More interestingly, we show that restoration of Alox5 expression can substantially increase the sensitivity of MLL-rearranged AML cells to standard chemotherapeutic agents such as DOX and Ara-C both in vitro and in vivo, as well as the underlying mechanism through suppressing the STAT and K-RAS oncogenic signaling pathways.
Interestingly, ALOX5 was largely known as an oncogene in solid tumors, e.g. prostate cancer and pancreas cancer 24,25 . However, its role remains vague in leukemia. It was reported that Alox5 deficiency resulted in a significant reduction of LSCs in BM, and thus largely prolonged survival of BCR-ABL-induced CML mice 26 . Nonetheless, there was no significant influence of Alox5 deficiency on normal hematopoietic stem cells (HSCs) or on the induction of lymphoid leukemia by BCR-ABL 26 . Another study showed that Alox5 deficiency inhibited in vitro colony-forming capacity of BM progenitor cells induced by some AML oncogenic fusion genes such as AML1-ETO9a, a potent oncogenic isoform of AML1-ETO resulting from t(8;21) 44 , but had no significant effects on AML1-ETO9a-induced leukemogenesis in vivo 32 . Those findings together with our present results suggest that Alox5 likely plays distinct roles at different lineages (myeloid and lymphoid) or different stages (i.e. induction and maintenance) of various hematopoietic malignancies.
One of the major challenges in AML therapy is the high frequency of occurrence of therapeutic resistance and the subsequent relapse 7 . Thus far, a variety of factors, such as JAK/STAT signaling and RAS pathway, have been identified to be closely related with chemotherapy response 9,13,16 . Abnormal activation of JAK/STAT signaling is found in most de novo AML 13 . K-RAS is a proto-oncogene essential for many solid tumors and hematopoietic malignancies 16 . Although RAS (i.e. N-RAS or K-RAS) gene mutations were only found in no more than 10% of AML cases, activation of RAS pathway by mutations in upstream receptors, e.g. FLT3 and c-KIT, or downstream effectors, broadly exist in AML 45 . A set of inhibitors have been developed to target JAK/STAT signaling (e.g. OPB-31121 and pacritinib) or RAS pathway (e.g., Selumetinib), but their efficacy in clinical trials to treat hematopoietic malignancies was still limited [46][47][48] . Our data suggest that restoration of ALOX5 expression/function could suppress both JAK/STAT and RAS signaling pathways simultaneously, and thus represents an alternative strategy, other than individual small-molecule inhibitors of JAK/STAT and RAS signaling pathways, to target these two critical oncogenic pathways in treating AML. Especially, restoration of ALOX5 expression/function in combination with the standard chemotherapy represents a potentially more effective therapeutic strategy for curing AML, at least MLL-rearranged AML. Thus, it would be interesting to identify small-molecule compound(s), such as natural product(s), that can specifically induce endogenous expression of ALOX5 expression in AML cells, and such compound(s) can be applied together with standard chemotherapy to treat MLL-rearranged AML.
Cell viability assays. Cells were seeded into 96-well plates at the concentration of 10,000 cells/ well in triplicates and MTT (Promega, Madison, WI) was used to assess cell proliferation and viability following the manufacturer's instructions.

Chromatin immunoprecipitation (ChIP).
ChIP assay was conducted as described previously 50  Western blotting. Cells were washed twice with ice-cold phosphate-buffered saline (PBS) and ruptured with RIPA buffer (Pierce, Rockford, IL) containing 5 mM EDTA, PMSF, cocktail inhibitor, and phosphatase inhibitor cocktail. Cell extracts were microcentrifuged for 20 min at 10,000 g and supernatants were collected. Cell lysates (20 μl) were resolved by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked for 1 hour with 5% skim milk in Tris-buffered saline containing 0.1% Tween 20 and incubated overnight at 4 °C with anti-ALOX5 antibody (Cell Signaling Technology Inc., Danvers, MA) or anti-ACTIN antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA). Membranes were washed 30 min with Tris-buffered saline containing 0.1% Tween-20, incubated for 1 hour with appropriate secondary antibodies conjugated to horseradish peroxidase, and developed using chemiluminescence substrates.
In vitro colony forming and replating assays (CFAs). CFAs were conducted as described previously with some modifications 29,50,51 . Briefly, retrovirus vectors were co-transfected with pCL-Eco packaging vector (IMGENEX, San Diego, CA) into HEK293T cells using Effectene Transfection Reagent (Qiagen, Valencia, CA) to produce retrovirus. BM cells were harvested from a cohort of 4-to 6-week-old B6.SJL (CD45.1) donor mice after five days of 5-fluorouracil (5-FU) treatment, and primitive hematopoietic progenitor cells were enriched with Mouse Lineage Cell Depletion Kit (Miltenyi Biotec Inc., Auburn, CA). An aliquot of enriched hematopoietic progenitor cells was added to retroviral supernatant together with polybrene in cell culture plates, which were centrifuged at 2,000 g for 2 hours at 32 °C (i.e., "spinoculation" [27][28][29] ) and then the medium was replaced with fresh media and incubated for 20 hours at 37 °C. Next day, the same procedure was repeated once.
Then, on the day following the second spinoculation, an equivalent of 2.0 × 10 4 cells were plated into a 35 mm Petri dish in 1.5 ml of Methocult M3230 methylcellulose medium (Stem Cell Technologies Inc, Vancouver, Canada) containing 10 ng/ml each of murine recombinant IL-3, IL-6, and granulocyte-macrophage colony-stimulating factor (GM-CSF), and 30 ng/ml of murine recombinant SCF (R&D Systems, Minneapolis, MN), along with 1.0 mg/ml of G418 and 2 μg/ml of puromycin. For each transduction, there were two duplicate dishes. Cultures were incubated at 37 °C in a humidified atmosphere of 5% CO 2 in air. The colonies were replated every 7 days under the same conditions. The colony-forming/replating assays were repeated 3 times.
Peripheral blood (PB) cells were collected from the transplanted recipient mice monthly via tail bleeding and at any time that the mice showed signs of systemic illness to obtain the complete blood counts with white blood cell (WBC) differentials and the blood smears for the presence of immature or abnormal hematopoietic cells. The engraftment was assessed by flow cytometry analysis of CD45.1 in PB. Leukemic mice were euthanized by CO 2 inhalation if they showed signs of systemic illness to collect specimens as controls for further analyses. The spleen, liver and thymus were weighed. Cells were obtained from PB, BM, spleen, and liver for flow cytometric analysis. Blood smear and BM cytospin slides were stained with Wright-Giemsa. For histological analysis, portions of the spleen and liver were fixed in formalin, sectioned, embedded in paraffin, and stained with hematoxylin and eosin (H&E).
Gene set analyses. Gene Set Enrichment Analysis (GSEA) 36 was used to analyze the signal pathway enrichment in different groups of samples. "Hallmark gene sets" obtained from MsigDB (The Molecular Signatures Database) were used as the "gene sets database" input. Vertebrate homology resource from the Mouse Genome Database (MGD) 55 was extracted to convert between homologous human and mouse gene symbols.
Statistical software. The gene/exon array data analyses and qPCR data analyses were conducted by use of Partek Genomics Suite (Partek Inc, St. Louis, MI), TIGR Mutiple Array Viewer software package (TMeV version 4.6; TIGR, Rockville, MA) 56 , and/or Bioconductor R packages. The t-test, Kaplan-Meier method, and log-rank test, etc. were performed with WinSTAT (R. Fitch Software), GraphPad Prism version 5.0 (GraphPad Software, San Diego, CA), and/or Partek Genomics Suite (Partek Inc, St. Louis, MI). The P-values less than 0.05 were considered as statistically significant. Significance analysis of microarrays (SAM) 57 , embedded in the TMeV package (TIGR, Rockville, MA), was used to identify the genes that are significantly (q < 0.05; false discovery rate, FDR < 0.05) dysregulated in human AML samples relative to the normal controls.
Data availability. Data referenced in this study are available in The Gene Expression Omnibus. The Affymetrix exon array data and the microarray data are available under accession codes code GSE34184 and GSE30285. The mouse RNA sequencing data is available under accession code GSE94840. The RNA sequencing results of the AML sample cohort analyzed for the correlation of ALOX5 levels and K-RAS mutations is under accession code GSE62190.