Identifying Novel Associations for Iron-Related Genes in High-Grade Ovarian Cancer Abigail Descoteaux*1, José W. Velázquez*2, Anna Konstorum3, Reinhard Laubenbacher3,4 1Vassar College, 2University of Puerto Rico at Cayey, 3Center for Quantitative Medicine, UConn Health, 4The Jackson Laboratory for Genomic Medicine *These authors contributed equally to this work
Introduction Methods Testable Hypothesis § Iron can gain and lose electrons, making it enzymatically Microarray gene expression Adjacency matrix Community detection in a weighted, KEGG pathways Tumor cells promote an inflammatory microenvironment via [1] data from HGSOC clinical • Top 10,000 most variable genes undirected network • Manually curated biological pathways useful in cell replication, metabolism, and growth • Bicor derived from Pearson but • ORA: ask whether any known pathways • Order Statistics Local Optimization Method (OSLOM) recruitment of tumor-associated macrophages, that in turn samples less sensitive to outliers are significantly over-represented in the • Assembled unsupervised with respect to biology by locally § Cancer cells sequester iron by altering the expression of • The Cancer Genome Atlas (TCGA) • Correlation cut-off at 0.45 genes within the community secrete IL4 and CCL4 to promote an iron-influx phenotype in optimizing the statistical significance of clusters and using [1] • Affymetrix HGU133A (n=488) (FDR corrected, α=0.01) Pathway 2 several regulators of iron metabolism (Figure 1) • Tothill et al. [2] this as a fitness score ovarian cancer tumor cells. • • HGU133plus2 (n=217) Accounts for edge weight, overlapping communities, and Pathway 1 hierarchies a. Normal epithelial cell Fe3+ Fe3+ Pathway 3 Tumor-associated
3+ 3+ Tumor cells Fe Fe macrophages Hepcidin (HAMP) samples genes Biweight midcorrelation Over-representation (bicor) Hypothesis CSF1R Ferroportin OSLOM analysis (ORA) CCR1[8] (SLC40A1) generation and [7] genes genes 1. CSF1 secretion validation
IL6 expression[10] 2. Monocyte recruitment, differentiation, and CCR1[12] Labile iron pool Gene of interest Pathway 4 proliferation[8,9] Pathway 5 Fe3+ Fe2+ Fe2+ Fe3+ HAMP Fe2+ 4. HAMP Fe3+ expression expression 2+ Fe2+ Fe Ferritin [11] (FTL & FTH1) ? 3. IL4 & CCL4 secretion[13] Transferrin receptor (TFRC) mod HAMP tothill IL4R Neigbors at a distance of 1 or less from TFRC tothill Results IL4R Validation: TFRC & RRM2, p53 pathway Unique results: HAMP & Immune pathways b. Cancer cell Hepcidin (HAMP) 1° Neighbors (Tothill) Top 20 1° Neighbors (Tothill) Ferroportin * # * DOCK2 (SLC40A1) Cytokine−Cytokine receptor interaction signallingMARCH1 pathway genes MYO1Fand HAMP Future Work RRM2
CDKN3 E2F8 GPR65 C1orf162 Topics for experimental investigation of changes in iron levels in MEF2C APBB1IP * #CCR1 3+ Degradation HMOX1 epithelial cells: Fe * § Fe3+3+ TFRC Fe CCNE2 # HAMP ADAP2 § Interrupt IL4 in macrophages or IL4R * TNFRSF1B Labile iron pool HCK MILR1 CENPM 2+ 3+ Fe 2+ § Fe Fe Disrupt CCL4/other CCR1 ligand or block CCR1 3+ p53 genes and TFRC Fe2+ Fe CYTH4 Fe2+ # Cytokine−Cytokine receptor interaction signallingOLR1 pathway genes and HAMP TNFRSF11B 3+ TK1 CENPI CD14 Fe Fe2+ Fe2+ CD300A § Over-express CSF1 or CSF1R 2+ Fe CX3CR1 # Fe2+ Ferritin TNFSF13B Fe2+ CSF1R TBXAS1 Fe2+ (FTL & FTH1) * = p53 signaling ARHGDIB Fe3+ ANLN TGFB1 # = pyrimidine metabolism * = chemokineTNFSF8 signaling Fe3+ Topics for computational investigation: § = cell cycle # = cytokine-cytokine receptorTNFRSF11A interactionCSF2RA IFNGR1 OSMR Transferrin receptor § Repeat analysis for subtypes of ovarian cancer to compare (TFRC) FAS ACVR2B p53 signaling pathway IL4R Cytokine-cytokine receptor interaction pathway CSF1R TNFRSF11BIL17RB pathways perturbed by different cell types CCND2 CCR1 Figure 1. Iron metabolism in a) normal TNFRSF1B IL10RA CX3CR1 TNFSF13B epithelial and b) cancer cells. CDK6 CCNE1
TGFB1 HAMP Adapted from [1]. CHEK2 TNFSF8 CCNB2 CDK1 IL10 TNFRSF11A CSF2RA CCNB1 IFNGR1 OSMR § Increase in metabolically available iron may promote tumor SESN3 CHEK1 FAS Acknowledgements ACVR2B IL4R CCNE2 CSF1R growth via production of reactive oxygen species and RRM2 IL17RB Funded by National Science Foundation Award #1460967. CCR1 [1] CDK2 TNFRSF1B oxidative DNA damage IL10RA We would like to thank Dr. Frank Torti and Dr. Suzy Torti and their lab at the UConn TFRC Health Center for their collaboration. § HAMP While some pro-oncogenic actions of iron have been CCND1 elucidated, there are still many questions regarding how iron IL10 [1] NDP can contribute to cancer progression Connections to iron and cancer:[3] Connections to iron and cancer: § Catalyzes rate-limiting step of References § Tumor cells secrete CSF1 to recruit monocytes, which [1] Torti SV, Torti FM. Nat Rev Cancer. 2013;13(5):342–55. RRM1 dNTP biosynthesis (Figure 2) + [6-9] [2] Tothill RW, Tinker AV, George J, Brown R, Fox SB, Lade S, et al. Clin Cancer Res. 2008;14(16):5198–208. Y-O differentiate into tumor-associated macrophages (TAMs) RRM2 [3] Aye Y, Li M, Long MJC, Weiss RS. Oncogene. 2015;34(16):2011–21. [13] Objectives § dNTP imbalance can disrupt Fe Fe Fe Fe § TAMs secrete IL4 and CCL4 [4] Grivennikov SI, Greten FR, Karin M. Cell. 2010;140(6):883–99. [11] [5] Marques O, Porto G, Rêma A, Faria F, Cruz Paula A, Gomez-Lazaro M, et al. BMC Cancer. 2016;16:187. 1. To identify correlations between iron-related genes and DNA repair and replication § In macrophages, IL4R increases HAMP expression [6] Tagliani E, Shi C, Nancy P, Tay C, Pamer E, & Erlebacher A. Journal Of Exp. Medicine. 2011;208(9):1901-1916. § RRM2 overexpression correlated [10] [7] Zhu Y et al. Cancer Research. 2014;74(18):5057-5069. other genes/pathways involved in ovarian cancer. § CCR1 activates a signaling cascade to increase IL6 expression [8] Wang C, Sun B, Tang Y, Zhuang H, & Cao W. Journal of Cancer Research and Clinical Oncology. 2008;135(5):695-701. with high tumor grade and poor dNDP [9] Jenkins S, Ruckerl D, Thomas GD, Hewitson JP, et al. The Journal of Experimental Medicine. 2013;210(11):2477. 2. To generate testable hypotheses about how iron-related § IL6 is associated with increased HAMP expression [10] Lee MMK, Chui RKS, Tam IYS, Lau AHY, Wong YH. The Journal of Immunology. 2012;189(11):5266-5276. genes promote cancer cell survival in high-grade serous overall survival in ovarian cancer [11] Haldar K, & Mohandas N. Hematology. 2009(1), 87-93. dNTP [12] Olson T & Ley K. American Journal of Physiology - Regulatory, Integrative & Comparative Physiology. 2002;283(1):R7-R28. ovarian cancer patients. patients Figure 2. RNR in dNTP biosynthesis [13] Bankaitis K, & Fingleton B (2015). Clinical & Experimental Metastasis. 2015;32(8):847-856.