Multiomics of azacitidine-treated AML cells reveals variable and convergent targets that remodel the cell-surface proteome Kevin K. Leunga, Aaron Nguyenb, Tao Shic, Lin Tangc, Xiaochun Nid, Laure Escoubetc, Kyle J. MacBethb, Jorge DiMartinob, and James A. Wellsa,1 aDepartment of Pharmaceutical Chemistry, University of California, San Francisco, CA 94143; bEpigenetics Thematic Center of Excellence, Celgene Corporation, San Francisco, CA 94158; cDepartment of Informatics and Predictive Sciences, Celgene Corporation, San Diego, CA 92121; and dDepartment of Informatics and Predictive Sciences, Celgene Corporation, Cambridge, MA 02140 Contributed by James A. Wells, November 19, 2018 (sent for review August 23, 2018; reviewed by Rebekah Gundry, Neil L. Kelleher, and Bernd Wollscheid) Myelodysplastic syndromes (MDS) and acute myeloid leukemia of DNA methyltransferases, leading to loss of methylation in (AML) are diseases of abnormal hematopoietic differentiation newly synthesized DNA (10, 11). It was recently shown that AZA with aberrant epigenetic alterations. Azacitidine (AZA) is a DNA treatment of cervical (12, 13) and colorectal (14) cancer cells methyltransferase inhibitor widely used to treat MDS and AML, can induce interferon responses through reactivation of endoge- yet the impact of AZA on the cell-surface proteome has not been nous retroviruses. This phenomenon, termed viral mimicry, is defined. To identify potential therapeutic targets for use in com- thought to induce antitumor effects by activating and engaging bination with AZA in AML patients, we investigated the effects the immune system. of AZA treatment on four AML cell lines representing different Although AZA treatment has demonstrated clinical benefit in stages of differentiation. The effect of AZA treatment on these AML patients, additional therapeutic options are needed (8, 9). cell lines was characterized at three levels: the DNA methylome, Our group has recently generated antibodies toward potential the transcriptome, and the cell-surface proteome. Untreated AML targets in RAS-driven cancers, and there is significant interest cell lines showed substantial overlap at all three omics levels; in identifying surface protein targets for antibody-derived thera- however, while AZA treatment globally reduced DNA methyla- peutic strategies in combination with AZA for the treatment of SYSTEMS BIOLOGY tion in all cell lines, changes in the transcriptome and surface AML (15). Currently, there are numerous antibody-based ther- proteome were subtle and differed among the cell lines. Tran- apeutics in development for AML patients, targeting about a scriptome analysis identified five commonly up-regulated coding dozen cell-surface proteins, but it is not clear if AZA changes genes upon AZA treatment in all four cell lines, TRPM4 being the only gene encoding a surface protein, and surface proteome Significance analysis found no commonly regulated proteins. Gene set enrich- ment analysis of differentially regulated RNA and surface proteins showed a decrease in metabolic pathways and an increase in Acute myeloid leukemia (AML) is a heterogeneous disease immune defense response pathways. As such, AZA treatment led commonly treated with azacitidine (AZA), but additional ther- to diverse effects at the individual gene and protein levels but apeutic strategy is needed. We found that AZA treatment converged to common responses at the pathway level. Given in four AML cell lines had diverse effects at the individ- the heterogeneous responses in the four cell lines, we discuss ual gene and protein level, and these changes converged to potential therapeutic strategies for AML in combination with AZA. common responses at the pathway level. The most promi- nent responses were the down-regulation of metabolism and up-regulation of immune defense. Given the heterogeneous AML j azacitidine j target discovery j multiomics j surface proteomics responses in the four cell lines, we discuss serval potential therapeutic strategies for AML in combinations with AZA. In yelodysplastic syndromes (MDS) and acute myeloid addition, the surface proteomics experiment has identified the Mleukemia (AML) are hematopoietic malignancies that are greatest number of surface proteins for these cell lines to date genetically and epigenetically diverse in nature. As myeloid lin- and represents a valuable resource to others who use these eage cells differentiate from their hematopoietic stem/progenitor cell lines as AML models. cells, aberrant epigenetic changes can occur at any differen- tiation stage, driving cells into cancerous phenotypes (1). As Author contributions: K.K.L., A.N., L.E., K.J.M., J.D., and J.A.W. designed research; K.K.L. and A.N. performed research; K.K.L., T.S., L.T., and X.N. analyzed data; and K.K.L., A.N., such, AML is routinely classified according to hematopoietic T.S., L.T., and J.A.W. wrote the paper.y lineages by cell morphology or by cytometry using sparse sur- Reviewers: R.G., Medical College of Wisconsin; N.L.K., Northwestern University; and B.W., face markers (2, 3). Among many epigenetic changes that occur ETH Zurich.¨ y in MDS and AML, the best-characterized change is the DNA Conflict of interest statement: This study was funded by the Celgene Corporation. A.N., methylation of cytosine bases in CpG islands (4). In fact, a T.S., L.T., X.N., L.E., K.J.M., and J.D. are employees of Celgene Corporation. K.K.L. and hallmark of epigenetic changes in AML is the redistribution of J.A.W. received research funding from Celgene Corporation but no personal financial methylated CpG dinucleotides with loss of methylation across gain or equity. y intergenic regions, primarily transposable elements and repeats, Published under the PNAS license.y and gain of aberrant methylation near the promoters of a num- Data deposition: Proteomics data have been deposited to ProteomeXchange Consortium ber of genes, including well-known tumor suppressors such as (proteomecentral.proteomexchange.org) via the MassIVE partner repository (dataset identifier PXD011298). Methylome and transcriptome datasets have been deposited to p16INK4a (5). As such, it is believed that these diseases are more the National Center for Biotechnology Information Gene Expression Omnibus database, sensitive to hypomethylating agents such as DNA methyltrans- https://www.ncbi.nlm.nih.gov/geo (SuperSeries GSE123211, accession nos. GSE123140 ferase inhibitors (DMNTi) (6, 7). One such DMNTi, azacitidine and GSE123207).y (AZA), has been efficaciously used for over a decade to treat 1 To whom correspondence should be addressed. Email: [email protected] MDS and AML (8, 9). At high doses, AZA induces rapid This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. DNA damage and is cytotoxic; at lower doses, AZA induces 1073/pnas.1813666116/-/DCSupplemental.y DNA hypomethylation by covalent trapping and degradation www.pnas.org/cgi/doi/10.1073/pnas.1813666116 PNAS Latest Articles j 1 of 6 Downloaded by guest on October 1, 2021 the expression levels of these proteins (16). Furthermore, over A 3 KG1a B ) 15 ongoing clinical trials are investigating the combination of HL60 5 HNT34 AZA and checkpoint inhibitors in various leukemias and solid AML193 2 tumors, since AZA induces checkpoint inhibitory molecules on 1 0 methylation (10 Density of hyper- and hypo- Intersection set size both tumor and immune cells (7, 17). To identify cell-surface 1 markers, cell-surface capture proteomics has recently emerged HL60 as a highly sensitive target discovery technology and has been KG1a 0123456 HNT34 used to define a large number of common and distinct mark- 0.0 0.2 0.4 0.6 0.8 1.0 Beta value AML193 C 21012345 ers in AML (18–20). Taken together, a broader understanding Methylation set size KG1a in each cell line (105) HL60 of how AZA treatment remodels the cell-surface proteome in HNT34 AML193 E 0 ) AML cells could aid in identifying surface protein targets for 5 antibody-based therapy, leading to unique immunotherapies for 1 use in combination with AZA. Density Using a multiomics approach, we characterized four AML 2 de-methylation (10 Intersection set size of cell lines, representing different stages of differentiation, and 01234 0.0 0.2 0.4 0.6 0.8 1.0 KG1a studied the changes in DNA methylation, RNA expression, and Beta value surface proteome induced by AZA treatment. Across the four D AML193 KG1a HNT34 cell lines, AZA reduced DNA methylation in nearly all of the HL60 HNT34 HL60 AML193 6 4 2 0 hypermethylated CpG sites probed, but surprisingly the changes De-methylation set size -0.5 0.0 0.5 in each cell line(105) in gene expression and surface protein expression were few and Change in beta value after AZA treatment diverse. Transcriptome analysis identified only one gene encod- Fig. 1. AZA treatment drives global DNA demethylation among all four ing a surface protein that is commonly up-regulated in all four AML cell lines. (A) Vehicle-treated cell lines have a bimodal distribution of cell lines, and surface proteomics analysis did not identify any genome-wide beta values (kernel density estimation). (B) Vehicle-treated commonly regulated proteins. Despite little overlap, functional cells share a high proportion of hypermethylated and hypomethylated sites. analysis revealed some common responses among the four cell Overlapping hypermethylated sites (red, beta values >0.8) and hypomethy- lines—down-regulation of genes and proteins in metabolism and lated sites (blue, beta values <0.2) are indicated by upward and downward up-regulation of genes in immune response. Collectively, our bars, respectively, in the vertical bar graph. The specific overlapping groups study detailed the distinct impact of AZA treatment in four AML are indicated by the black solid points below the bar graph. Total hyperme- cell line at the individual gene level and illustrated that functional thylated and hypomethylated sites found in each cell line are indicated in networks are commonly regulated. the horizontal bar graph. (C) AZA-treated cells have decreased hypermethy- lated beta values indicating DNA demethylation.