a, Cell lines were screened using a retrovirally-delivered, doxycycline-inducible, shRNA library to identify genes required for cell growth or survival, as described10. Depletion of cells bearing an IRF4-targeted shRNA in shRNA-induced versus uninduced cells is plotted; error bars represent the s.d. of triplicate measurements. b, Expression of the IRF4 coding region rescues myeloma cells from lethality of an shRNA targeting the IRF4 3'UTR (see text for details). c, An IRF4 shRNA is toxic to myeloma but not lymphoma cell lines. A vector for constitutive expression of IRF4 shRNA was transduced into cell lines, and viability of shRNA+ cells was monitored. In (b) and (c), cells expressing shRNA were monitored by flow cytometry for a co-expressed GFP marker and data were normalized to the % of GFP+ cells at day 2 post infection.

a, Venn diagram depicting IRF4 target genes and the overlap between the myeloma, plasma cell, and activated B cell gene expression programs. Of the 308 IRF4 target genes (Supplemental Fig. 3), 262 were well-measured on Affymetrix gene expression arrays. 101 were more highly expressed in primary myeloma samples than primary mature B cells (>1.4-fold, red circle), 67 were more highly expressed in primary plasma cells than mature B cells (>1.4-fold, green circle), and 81 are induced between 1 hr and 24 hr following activation of primary human B cells by anti-IgM crosslinking (>2-fold, yellow circle). red: direct IRF4 targets by ChIP, *: direct MYC targets. b, Representative conventional ChIP assays for genes identified as IRF4 targets by both gene expression profiling and ChIP-CHIP. Individual ChIP assays were performed on chromatin from the KMS12 myeloma line and the OCI-Ly19 lymphoma line using either anti-IRF4 or control antibodies. The ChIP signal is given in arbitrary relative units calculated from quantitative PCR data, based on the relative abundance of the indicated gene in the immunoprecipitated DNA versus input DNA. Error bars are s.d. from triplicate measurements.

a, Knockdown of IRF4 decreases MYC mRNA expression. The SKMM1 myeloma line was transduced with IRF4 or MYC shRNAs, and gene expression was measured by quantitative RT-PCR after 4 days of shRNA induction, normalized to the signal from uninduced cells. b, Binding of IRF4 to the MYC promoter. ChIP was performed as in Figure 2, comparing the myeloma line KMS12 to the lymphoma line OCI-LY19, for regions of the MYC promoter (as indicated relative to the transcriptional start site) or a control locus, CYP2E1. c, IRF4 knockdown decreases IRF4 binding to the MYC promoter. ChIP was performed using KMS12 cells with or without shRNA induction for 4 days. d, Activation of human blood B cells leads to IRF4 binding to the MYC promoter. ChIP assays were performed using purified peripheral human blood B cells, either unstimulated or activated with P/I for the indicated times. e, Genetic deficiency of IRF4 impairs MYC induction during lymphocyte activation. Quantitative RT-PCR was performed on RNA from resting splenic B cells of IRF4-deficient or wild type mice, either unstimulated or activated with P/I for the indicated times. Values were normalized to B2M expression. f, IRF4 transactivates the MYC promoter. The OCI-Ly7 lymphoma line was transiently transfected with a GFP expression vector driven by the human MYC promoter, either alone, with an IRF4 expression vector, or with an empty vector control. Flow cytometry for GFP fluorescence is shown, with error bars indicating s.d. of triplicate measurements.

a, Lethality of a MYC shRNA for cell lines expressing MYC. Cell lines were transduced with a MYC shRNA vector and the fraction of shRNA+ (GFP+) cells was monitored over time. All lines express MYC, except U266, which expresses MYCL1. b, Expression of the MYC coding region rescues H929 myeloma cells from lethality of an shRNA targeting the MYC 3'UTR. c, MYC knockdown downregulates MYC direct target genes and IRF4 target genes. KMS12 myeloma cells were induced for MYC shRNA expression for 4 days and profiled for gene expression changes. Each experiment utilized a different MYC shRNA. Exemplar array elements are shown (reduced by >1.3-fold in both experiments), for known MYC direct targets17 and IRF4 targets (this work). d, MYC knockdown decreases IRF4 mRNA expression. Shown are quantitative RT-PCR measurements of MYC and IRF4 mRNA levels in KMS12 myeloma cells, with or without induction of MYC shRNA. Error bars indicate s.d. of triplicate measurements. e, MYC binds to the IRF4 locus. ChIP of MYC binding to the IRF4 first intron in myeloma cells expressing MYC (KMS12, H929), but not in the myeloma line U266 that lacks MYC expression. f, MYC binding to the IRF4 locus is induced in activated human B cells. ChIP of MYC binding to the IRF4 first intron in human blood B cells, either unstimulated or activated with P/I for 6 hr. g, MYC and IRF4 are more highly expressed in primary myeloma patient samples than in normal human bone marrow plasma cells. Previously published gene expression profiling data4 was analyzed for mRNA expression of MYC, IRF4, and a control gene, UBC.

a, IRF4 and MYC form a positive autoregulatory loop during normal B cell activation and in multiple myeloma. Genetic abnormalities of MYC upregulate its expression in myeloma, thereby augmenting IRF4 expression. In normal plasma cells, Blimp-1 represses MYC, but this control circuit may be abrogated in myeloma. b, IRF4 as a master regulator of the myeloma phenotype. IRF4 controls a myeloma-specific gene expression program that fuses the IRF4 regulatory programs from activated B cells and plasma cells. IRF4 direct targets regulate many essential cellular processes, causing myeloma cells to be addicted to IRF4.