Cancer is widely perceived as a heterogeneous group of disorders with markedly different biological properties, which are caused by a series of clonally selected genetic changes in key tumour-suppressor genes and oncogenes. However, recent data suggest that cancer has a fundamentally common basis that is grounded in a polyclonal epigenetic disruption of stem/progenitor cells, mediated by 'tumour-progenitor genes'. Furthermore, tumour cell heterogeneity is due in part to epigenetic variation in progenitor cells, and epigenetic plasticity together with genetic lesions drives tumour progression. This crucial early role for epigenetic alterations in cancer is in addition to epigenetic alterations that can substitute for genetic variation later in tumour progression. Therefore, non-neoplastic but epigenetically disrupted stem/progenitor cells might be a crucial target for cancer risk assessment and chemoprevention.

Accumulating evidence converges on the possibility that chromosomes interact with each other to regulate transcription in trans. To systematically explore the epigenetic dimension of such interactions, we devised a strategy termed circular chromosome conformation capture (4C). This approach involves a circularization step that enables high-throughput screening of physical interactions between chromosomes without a preconceived idea of the interacting partners. Here we identify 114 unique sequences from all autosomes, several of which interact primarily with the maternally inherited H19 imprinting control region. Imprinted domains were strongly overrepresented in the library of 4C sequences, further highlighting the epigenetic nature of these interactions. Moreover, we found that the direct interaction between differentially methylated regions was linked to epigenetic regulation of transcription in trans. Finally, the patterns of interactions specific to the maternal H19 imprinting control region underwent reprogramming during in vitro maturation of embryonic stem cells. These observations shed new light on development, cancer epigenetics and the evolution of imprinting.

Nuclear receptors regulate metabolic pathways in response to changes in the environment by appropriate alterations in gene expression of key metabolic enzymes. Here, a computational search approach based on iteratively built hidden Markov models of nuclear receptors was used to identify a human nuclear receptor, termed hPAR, that is expressed in liver and intestines. hPAR was found to be efficiently activated by pregnanes and by clinically used drugs including rifampicin, an antibiotic known to selectively induce human but not murine CYP3A expression. The CYP3A drugmetabolizing enzymes are expressed in gut and liver in response to environmental chemicals and clinically used drugs. Interestingly, hPAR is not activated by pregnenolone 16␣-carbonitrile, which is a potent inducer of murine CYP3A genes and an activator of the mouse receptor PXR.1. Furthermore, hPAR was found to bind to and trans-activate through a conserved regulatory sequence present in human but not murine CYP3A genes. These results provide evidence that hPAR and PXR.1 may represent orthologous genes from different species that have evolved to regulate overlapping target genes in response to pharmacologically distinct CYP3A activators, and have potential implications for the in vitro identification of drug interactions important to humans.The recent progress made in determining the full genomic sequences of model organisms as well as the rapid accumulation of sequence data from the human genome has opened up new possibilities to determine the functional organization of genomes by computational approaches (1). Multiple alignments of members of different protein families followed by homology searching are powerful methods to infer gene function from sequence data as well as to identify novel genes within a given gene family. Hidden Markov models (HMMs) are a general statistical modeling technique that can be used as formal, fully probabilistic forms of sequence profiles (2, 3), describing the consensus of a set of sequences. This approach toward the identification and functional characterization of novel genes is particularly amenable to evolutionary conserved gene families consisting of a large number of orthologs and paralogs. The nuclear receptors constitute one such large gene family that is structurally and functionally conserved and represented within different metazoan phylae from cnidarians to vertebrates (4). These receptors are conditionally regulated transcription factors that exert their effects by interacting with small lipophilic ligands followed by sequence-specific binding of the receptor to DNA sequences called hormone response elements (HREs). Binding of the receptor to DNA results in changes in gene expression of specific target genes (5).Steroid hormones were the first group of small, lipophilic molecules identified as nuclear receptor ligands. Today, the number of substances known to regulate the activity of this group of receptors is represented by a large and chemically diverse group of molecules including retinoids, vi...

No abstract