Dinah S. Singer

A community effort to assess and improve drug sensitivity prediction algorithms

Predicting the best treatment strategy from genomic information is a core goal of precision medicine. Here we focus on predicting drug response based on a cohort of genomic, epigenomic and proteomic profiling data sets measured in human breast cancer cell lines. Through a collaborative effort between the National Cancer Institute (NCI) and the Dialogue on Reverse Engineering Assessment and Methods (DREAM) project, we analyzed a total of 44 drug sensitivity prediction algorithms. The top-performing approaches modeled nonlinear relationships and incorporated biological pathway information. We found that gene expression microarrays consistently provided the best predictive power of the individual profiling data sets; however, performance was increased by including multiple, independent data sets. We discuss the innovations underlying the top-performing methodology, Bayesian multitask MKL, and we provide detailed descriptions of all methods. This study establishes benchmarks for drug sensitivity prediction and identifies approaches that can be leveraged for the development of new methods.

BRD4 is an atypical kinase that phosphorylates Serine2 of the RNA Polymerase II carboxy-terminal domain

The bromodomain protein, BRD4, has been identified recently as a therapeutic target in acute myeloid leukemia, multiple myeloma, Burkitt’s lymphoma, NUT midline carcinoma, colon cancer, and inflammatory disease; its loss is a prognostic signature for metastatic breast cancer. BRD4 also contributes to regulation of both cell cycle and transcription of oncogenes, HIV, and human papilloma virus (HPV). Despite its role in a broad range of biological processes, the precise molecular mechanism of BRD4 function remains unknown. We report that BRD4 is an atypical kinase that binds to the carboxyl-terminal domain (CTD) of RNA polymerase II and directly phosphorylates its serine 2 (Ser2) sites both in vitro and in vivo under conditions where other CTD kinases are inactive. Phosphorylation of the CTD Ser2 is inhibited in vivo by a BRD4 inhibitor that blocks its binding to chromatin. Our finding that BRD4 is an RNA polymerase II CTD Ser2 kinase implicates it as a regulator of eukaryotic transcription.

Foxp3 Transcription Factor Is Proapoptotic and Lethal to Developing Regulatory T Cells unless Counterbalanced by Cytokine Survival Signals

Immune tolerance requires regulatory T (Treg) cells to prevent autoimmune disease, with the transcription factor Foxp3 functioning as the critical regulator of Treg cell development and function. We report here that Foxp3 was lethal to developing Treg cells in the thymus because it induced a unique proapoptotic protein signature (Puma⁺⁺⁺p-Bim⁺⁺p-JNK⁺⁺DUSP6⁻) and repressed expression of prosurvival Bcl-2 molecules. However, Foxp3 lethality was prevented by common gamma chain (γc)-dependent cytokine signals that were present in the thymus in limiting amounts sufficient to support only ∼1 million Treg cells. Consequently, most newly arising Treg cells in the thymus were deprived of this signal and underwent Foxp3-induced death, with Foxp3⁺CD25⁻ Treg precursor cells being the most susceptible. Thus, we identify Foxp3 as a proapoptotic protein that requires developing Treg cells to compete with one another for limiting amounts of γc-dependent survival signals in the thymus.

A community computational challenge to predict the activity of pairs of compounds

Recent therapeutic successes have renewed interest in drug combinations, but experimental screening approaches are costly and often identify only small numbers of synergistic combinations. The DREAM consortium launched an open challenge to foster the development of in silico methods to computationally rank 91 compound pairs, from the most synergistic to the most antagonistic, based on gene-expression profiles of human B cells treated with individual compounds at multiple time points and concentrations. Using scoring metrics based on experimental dose-response curves, we assessed 32 methods (31 community-generated approaches and SynGen), four of which performed significantly better than random guessing. We highlight similarities between the methods. Although the accuracy of predictions was not optimal, we find that computational prediction of compound-pair activity is possible, and that community challenges can be useful to advance the field of in silico compound-synergy prediction.

Transcriptional coactivator, CIITA, is an acetyltransferase that bypasses a promoter requirement for TAFII250

The CIITA coactivator is essential for transcriptional activation of MHC class II genes and mediates enhanced MHC class I transcription. We now report that CIITA contains an intrinsic acetyltransferase (AT) activity that maps to a region within the N-terminal segment of CIITA, between amino acids 94 and 132. The AT activity is regulated by the C-terminal GTP-binding domain and is stimulated by GTP. CIITA-mediated transactivation depends on the AT activity. Further, we report that, although constitutive MHC class I transcription depends on TAF(II)250, CIITA activates the promoter in the absence of functional TAF(II)250.

A Spontaneously Arising Mutation in the DLAARN Motif of Murine ZAP-70 Abrogates Kinase Activity and Arrests Thymocyte Development

Development of immature CD4+ CD8+ thymocytes into functionally mature CD4+ and CD8+ T cells is driven by selection events that require signals transduced through the T cell antigen receptor (TCR). Transduction of TCR signals in the thymus involves tyrosine phosphorylation of the protein tyrosine kinase ZAP-70 by p56(lck) and results in induction of ZAP-70 enzymatic activity. We have identified a novel, spontaneously arising point mutation within a highly conserved motif (DLAARN) in the kinase domain of murine ZAP-70 that uncouples tyrosine phosphorylation of ZAP-70 from induction of ZAP-70 kinase activity. Mice homozygous for this mutation are devoid of mature T cells because thymocyte development is arrested at the CD4+ CD8+ stage of differentiation. The developmental arrest is due to the inability of CD4+ CD8+ thymocytes to propagate TCR signals in the absence of ZAP-70 kinase activity despite tyrosine phosphorylation of TCR-associated ZAP-70 molecules.

Regulation of Major Histocompatibility Complex Class I Gene Expression in Thyroid Cells ROLE OF THE cAMP RESPONSE ELEMENT-LIKE SEQUENCE

The major histocompatibility complex (MHC) class I gene cAMP response element (CRE)-like site, −107 to −100 base pairs, is a critical component of a previously unrecognized silencer, −127 to −90 bp, important for thyrotropin (TSH)/cAMP-mediated repression in thyrocytes. TSH/cAMP induced-silencer activity is associated with the formation of novel complexes with the 38-base pair silencer, whose appearance requires the CRE and involves ubiquitous and thyroid-specific proteins as follows: the CRE-binding protein, a Y-box protein termed thyrotropin receptor (TSHR) suppressor element protein-1 (TSEP-1); thyroid transcription factor-1 (TTF-1); and Pax-8. TTF-1 is an enhancer of class I promoter activity; Pax-8 and TSEP-1 are suppressors. TSH/cAMP decreases TTF-1 complex formation with the silencer, thereby decreasing maximal class I expression; TSH/cAMP enhance TSEP-1 and Pax-8 complex formation in association with their repressive actions. Oligonucleotides that bind TSEP-1, not Pax-8, prevent formation of the TSH/cAMP-induced complexes associated with TSH-induced class I suppression, i.e. TSEP-1 appears to be the dominant repressor factor associated with TSH/cAMP-decreased class I activity and formation of the novel complexes. TSEP-1, TTF-1, and/or Pax-8 are involved in TSH/cAMP-induced negative regulation of the TSH receptor gene in thyrocytes, suppression of MHC class II, and up-regulation of thyroglobulin. TSH/cAMP coordinate regulation of common transcription factors may, therefore, be the basis for self-tolerance and the absence of autoimmunity in the face of TSHR-mediated increases in gene products that are important for thyroid growth and function but are able to act as autoantigens.

HIV Tat represses transcription through Spl-Like elements in the basal promoter

MHC class I genes are potently repressed by HIV Tat, which transactivates the HIV LTR. Tat represses class I transcription by binding to complexes associated with a novel promoter element, consisting of Sp1-like DNA binding sites. Transcription by other Sp1-dependent promoters, such as MDR1 and the minimal SV40 promoters, is also repressed by Tat, whereas the human beta-actin promoter is neither activated by Sp1 nor repressed by Tat. Tat repression can be overcome by a strong enhancer element. Thus, the SV40 72 bp enhancer element confers protection from Tat-mediated repression on both the minimal SV40 promoter and the class I promoter. Surprisingly, Tat can activate the class I promoter in the presence of both the HIV TAR element and a strong upstream enhancer. These data demonstrate that Tat differentially affects Sp1-responsive promoters, depending on promoter architecture.

BRD4 is a histone acetyltransferase that evicts nucleosomes from chromatin

Bromodomain protein 4 (BRD4) is a chromatin-binding protein implicated in cancer and autoimmune diseases that functions as a scaffold for transcription factors at promoters and super-enhancers. Although chromatin decompaction and transcriptional activation of target genes are associated with BRD4 binding, the mechanisms involved are unknown. We report that BRD4 is a histone acetyltransferase (HAT) that acetylates histones H3 and H4 with a pattern distinct from those of other HATs. Both mouse and human BRD4 have intrinsic HAT activity. Importantly, BRD4 acetylates H3 K122, a residue critical for nucleosome stability, thus resulting in nucleosome eviction and chromatin decompaction. Nucleosome clearance by BRD4 occurs genome wide, including at its targets MYC, FOS and AURKB (Aurora B kinase), resulting in increased transcription. These findings suggest a model wherein BRD4 actively links chromatin structure and transcription: it mediates chromatin decompaction by acetylating and evicting nucleosomes at target genes, thereby activating transcription.

Characterization of a new subfamily of class I genes in the H-2 complex of the mouse

A previously undescribed subfamily of mouse class I MHC genes, consisting of two to three members, has been identified. The structure and organization of one of these, Mb1, has been determined. Mb1, consists of five exons with open reading frames and potentially encodes a class I-like transmembrane protein. In the genome, Mb1 is linked to the H-2 complex, mapping telomeric to Qa. However, this gene has low (ca. 60%) nucleotide identity with other class I sequences and is no more related to mouse class I genes than to class I genes from other species. Mb1 transcripts have not been found in a variety of adult tissues or cell lines, suggesting that, if Mb1 is expressed, its expression is highly regulated. From DNA sequence identity and intron-exon organization, Mb1 appears to be a primordial gene which antedates mouse speciation and which has evolved independently of the rest of the class I gene family. Examination of various species of wild mice demonstrates the presence of a discrete Mb1 subfamily over long evolutionary periods of time.