PPARγ maintains ERBB2-positive breast cancer stem cells.

Overexpression of the adverse prognostic marker ERBB2 occurs in 30% of breast cancers and is associated with aggressive disease and poor outcomes. Our recent findings have shown that NR1D1 and the peroxisome proliferator-activated receptor-γ (PPARγ)-binding protein (PBP) act through a common pathway in upregulating several genes in the de novo fatty acid synthesis network, which is highly active in ERBB2-positive breast cancer cells. NR1D1 and PBP are functionally related to PPARγ, a well-established positive regulator of adipogenesis and lipid storage. Here, we report that inhibition of the PPARγ pathway reduces the aldehyde dehydrogenase (ALDH)-positive population in ERBB2-positive breast cancer cells. Results from in vitro tumorsphere formation assays demonstrate that the PPARγ antagonists GW9662 and T0070907 decrease tumorsphere formation in ERBB2-positive cells, but not other breast cells. We show that the mechanism by which GW9662 treatment causes a reduction in ALDH-positive population cells is partially due to ROS, as it can be rescued by treatment with N-acetyl-cysteine. Furthermore, global gene expression analyses show that GW9662 treatment suppresses the expression of several lipogenic genes, including ACLY, MIG12, FASN and NR1D1, and the stem-cell related genes KLF4 and ALDH in BT474 cells. Antagonist treatment also decreases the level of acetylation in histone 3 and histone 4 in BT474 cells, compared with MCF7 cells. In vivo, GW9662 pre-treatment inhibits the tumor-seeding ability of BT474 cells. Together, these results show that the PPARγ pathway is critical for the cancer stem cell properties of ERBB2-positive breast cancer cells.

RNAi applications in target validation.

The emergence of systems biology is certain to transform the identification and validation of therapeutic targets in modern drug discovery. A relatively recent systems biology approach is functional genomics, which identifies the molecular mechanisms responsible for a specific phenotype by interrogating the activity of all of an organism's genes. Initially undertaken in model organisms such as Caenorhabditis elegans, Saccharomyces cerevisiae, and Drosophila melanogaster, functional genomics has now moved into the realm of mammalian cells both in vitro and in vivo due to the development of RNA interference. RNA interference is a conserved biological process that has evolved to specifically and efficiently silence genes. Genome-wide screens using RNA interference have proven powerful in elucidating components of functionally related pathways and have therefore become integral for the development of new and improved therapeutic targets. This article provides an overview of many of the systems biology approaches taken, using RNA interference, in order to demonstrate how it may be used today for drug discovery and tomorrow as a targeted therapy.

New tools for protein linkage mapping and general two-hybrid screening.

The two-hybrid system has proved to be a facile method for detecting and analyzing protein-protein interactions. An expanded application of this system, protein linkage mapping, provides a means of identifying interactions on a global scale and should prove a powerful tool in analyzing whole genomes as their sequences become available. To overcome some of the inherent difficulties in such a large-scale approach, we have constructed a set of new strains and vectors that will allow for more efficient screening. The strains contain a GAL1-URA3 reporter for positive and negative selection, as well as a UAS(G)-lacZ reporter. The strains are of opposite mating types, permitting libraries present in one strain to be easily screened against a second library in the companion strain. We also constructed a family of CEN-based vectors for expression of both Gal4 DNA-binding and activation domain fusions. These plasmids include a hemagglutinin epitope tag and different polylinkers to increase the ease of subcloning. CEN-based vectors are maintained at 1-2 copies per cell, limiting the number of individual cells containing multiple plasmids that can confuse further analyses, and ensuring that fusions are not expressed at toxic levels. Using these vectors, both homo- and heterodimeric interactions resulted in up to 10-fold higher reporter gene transcription than obtained with 2micro;-based plasmids, despite significantly lower protein levels. In addition to protein linkage mapping, these reagents should be generally useful in standard two-hybrid applications.

Messenger RNA editing of the human serotonin 5-HT2C receptor.

RNA encoding the rat serotonin 5-HT2C receptor undergoes editing whereby one to four adenosines are converted to inosines. This conversion can change up to three codons out of a stretch of five in the second intracellular loop of the receptor. RNA editing of the rat 5-HT2C receptor that changes all three codons was shown previously to alter intracellular signaling by 5-HT without changing its receptor-binding affinity. We analyzed 5-HT2C receptor editing in human brain and hypothalamic RNA samples and confirmed that all four adenosine editing sites observed in rat were also present in human samples. Additionally, we identified a novel editing site in the middle edited codon that extends the repertoire of 5-HT2C receptors by six additional protein isoforms. We observed that editing reduces both the binding affinity and functional potency of agonists for recombinant human 5-HT2C receptor isoforms. This effect on binding affinity was proportional to the agonist's intrinsic activity, with full agonists most affected, and antagonists showing no effect. These data suggest that RNA editing may alter coupling energetics within the ternary complex, thereby altering agonist binding affinities, G protein coupling, and functional responses. RNA editing may thus provide a novel mechanism for regulating 5-HT synaptic signaling and plasticity.

High-affinity agonist binding correlates with efficacy (intrinsic activity) at the human serotonin 5-HT2A and 5-HT2C receptors: evidence favoring the ternary complex and two-state models of agonist action.

Many modern models of receptor-G protein function assume that there is a direct relationship between high-affinity agonist binding and efficacy. The validity of this assumption has been recently questioned for the serotonin 5-HT2A receptor. We examined the intrinsic activities of various ligands in activating phosphoinositide hydrolysis and measured their respective binding affinities to the high- and low-affinity states of the 5-HT2C (VNV isoform) and 5-HT(2A) receptors. Ligand binding affinities for the high-affinity state of the receptors were determined using 1-(4-[125I]iodo-2,5-dimethoxyphenyl)2-aminopropane, whereas [3H]mesulergine and N-[3H]methylspiperone were used, in the presence of excess guanine nucleotide [guanosine 5'-O-(3-thiotriphosphate)], to define binding to the low-affinity state of the 5-HT2C and 5-HT2A receptors, respectively. Antagonists labeled the high- and low-affinity states of each receptor with comparable affinities. Previously identified inverse agonists of the 5-HT2C receptor behaved as silent antagonists in our systems even when the receptor was overexpressed at a relatively high density. In contrast, the ability of agonists to bind differentially to the high- and low-affinity states of the 5-HT2A and 5-HT2C receptors was highly correlated (r2 = 0.86 and 0.96, respectively) with their intrinsic activities. These data suggest that high-affinity agonist states can account for agonist efficacy at human 5-HT2A or 5-HT2C receptors without the need for considering additional transition or active states of the receptor-ligand complex. The procedure described herein may expedite drug discovery efforts by predicting intrinsic activities of ligands solely from ligand binding assays.

Pharmacological and biochemical characterization of a recombinant human galanin GALR1 receptor: agonist character of chimeric galanin peptides.

The galanin neuropeptide system is widely distributed throughout the brain and periphery and is thought to play a role in feeding, pain and reproduction. To evaluate the human galanin receptor 1 as a potential therapeutic target, we fully characterized its interaction with several galanin-like peptides. The human galanin receptor 1 receptor was stably expressed using an episomal system in human embryonic kidney 293E cells. Saturation isotherms using 125I-human galanin revealed two distinct populations of receptor affinity states in membranes and whole cells with picomolar and nanomolar affinities at the high- and low affinity states, respectively. A scintillation proximity assay revealed that 125I-human galanin binding in membranes reached steady-state within 2 to 2.5 hr; however, only 50% of galanin radiolabel dissociated from the receptors by excess galanin or guanosine 5'-O-3-thiotriphosphate even after 20 hr. In contrast, galanin binding in whole cells was completely reversible within 1 hr. Competition binding assays showed that galanin-like peptides bound with picomolar affinities in membranes and whole cells. These peptides behaved as full agonists as determined by the inhibition of forskolin-stimulated cyclic 3'5'-adenosine monophosphate production and the stimulation of guanosine 5'-O-(3-[35S]thiotriphosphate binding. The agonist profile of M40, a representative chimeric peptide, was found not to be the result of receptor reserve because receptor inactivation by partial alkylation experiments confirmed its full intrinsic efficacy under conditions of a "zero" reserve state. These observations suggest that the antagonist effects in vivo of M40, and perhaps other chimeric peptides, are not mediated via direct interactions with the galanin receptor 1 receptor.

14-3-3 proteins associate with cdc25 phosphatases.

The cdc25 phosphatases play key roles in cell cycle progression by activating cyclin-dependent kinases. Two members of the 14-3-3 protein family have been isolated in a yeast two-hybrid screen designed to identify proteins that interact with the human cdc25A and cdc25B phosphatases. Genes encoding the human homolog of the 14-3-3 epsilon protein and the previously described 14-3-3 beta protein have been isolated in this screening. 14-3-3 proteins constitute a family of well-conserved eukaryotic proteins that were originally isolated in mammalian brain preparations and that possess diverse biochemical activities related to signal transduction. We present evidence that indicates that cdc25 and 14-3-3 proteins physically interact both in vitro and in vivo. 14-3-3 protein does not, however, affect the phosphatase activity of cdc25A. Raf-1, which is known to bind 14-3-3 proteins, has recently been shown to associate with cdc25A and to stimulate its phosphatase activity. 14-3-3 protein, however, has no effect on the cdc25A-kinase activity of Raf-1. Instead, 14-3-3 may facilitate the association of cdc25 with Raf-1 in vivo, participating in the linkage between mitogenic signaling and the cell cycle machinery.

Interactions between gene products involved in divalent cation transport in Saccharomyces cerevisiae.

The COT1 and ZRC1 genes of Saccharomyces cerevisiae are structurally related dosage-dependent suppressors of metal toxicity. COT1 confers increased tolerance to high levels of cobalt; ZRC1 confers increased tolerance to high levels of zinc. The two genes are not linked and have been mapped; COT1 to chromosome XV and ZRC1 to chromosome XIII. Phenotypes related to metal homeostasis have been examined in strains with varied COT1 and ZRC1 gene doses. Overexpression of COT1 confers tolerance to moderately toxic levels of zinc and ZRC1 confers tolerance to moderately toxic levels of cobalt. Strains that carry null alleles at both loci are viable. The metal-hypersensitive phenotypes of mutations in either gene are largely unaffected by changes in dosage of the other. COT1 and ZRC1 function independently in conferring tolerance to their respective metals, yet the uptake of cobalt ions by yeast cells is dependent on the gene dosage of ZRC1 as well as of COT1. Strains that overexpress ZRC1 have increased uptake of cobalt ions, while ZRC1 null mutants exhibit decreased cobalt uptake. The defects in cobalt uptake due to mutations at COT1 and ZRC1 are additive, suggesting that the two genes are responsible for the majority of cobalt and zinc uptake in yeast cells. The function of either gene product seems to be more important in metal homeostasis than is the GRR1 gene product, which is also involved in metal metabolism. Mutations in the GRR1 gene have no effect on the cobalt-related phenotypes of strains that have altered gene dosage of either COT1 or ZRC1.

Saccharomyces cerevisiae mutants sensitive to the antimalarial and antiarrhythmic drug, quinidine.

Mutations at three loci in Saccharomyces cerevisiae have been shown to confer increased sensitivity to the antimalarial and antiarrhythmic alkaloid, quinidine. Two of these groups are composed of strains carrying recessive mutations, the other group contains two dominant alleles. The largest complementation group has been designated QDS1, for increased quinidine-sensitivity. Exposure of qds1 cells to lethal concentrations of quinidine results in a novel small-budded terminal morphology in about 70% of the cells in the culture. Strains which carry qds1 alleles share other pleiotropic phenotypes. qds1 mutants are incapable of mating as alpha but not a cells, due to a defect in alpha-factor production. Homozygous diploid qds1 strains cannot sporulate. Genetic evidence indicates that QDS1 is allelic to KEX2, a precursor processing protease. Loss of QDS1/KEX2 function results in quinidine sensitivity.

The COT2 gene is required for glucose-dependent divalent cation transport in Saccharomyces cerevisiae.

Eleven cobalt-tolerant mutants were found to belong to a single complementation group, cot2. In addition to cobalt, the cot2 mutants were found to tolerate increased levels of the divalent cations Zn2+, Mn2+, and Ni2+ as well. All of the cot2 mutants exhibited a wiener-shaped cellular morphology that was exacerbated by the carbon and nitrogen source but was unaffected by metals. The rate of glucose-dependent transport of cobalt into cells was reduced in strains that carry mutations in the COT2 gene. COT2 is not essential for growth. Strains that carry a COT2 allele conferring complete loss of function are viable and exhibit phenotypes similar to those of spontaneous cot2 mutations. The sequence of the COT2 gene shows that it is identical to GRR1, which encodes a protein required for glucose repression. The glucose dependence of the transport defect implies that cot2 mutations affect the link between glucose metabolism and divalent cation active transport.

COT1, a gene involved in cobalt accumulation in Saccharomyces cerevisiae.

The COT1 gene of Saccharomyces cerevisiae has been isolated as a dosage-dependent suppressor of cobalt toxicity. Overexpression of the COT1 gene confers increased tolerance to cobalt and rhodium ions but not other divalent cations. Strains containing null alleles of COT1 are viable yet more sensitive to cobalt than are wild-type strains. Transcription of COT1 responds minimally to the extracellular cobalt concentration. Addition of cobalt ions to growth media results in a twofold increase in COT1 mRNA abundance. The gene encodes a 48-kDa protein which is found in mitochondrial membrane fractions of cells. The protein contains six possible membrane-spanning domains and several potential metal-binding amino acid residues. The COT1 protein shares 60% identity with the ZRC1 gene product, which confers resistance to zinc and cadmium ions. Cobalt transport studies indicate that the COT1 product is involved in the uptake of cobalt ions yet is not solely responsible for it. The increased tolerance of strains containing multiple copies of the COT1 gene is probably due to increased compartmentalization or sequestration of the ion within mitochondria.

Identification of a putative structural gene for cathepsin D in Caenorhabditis elegans.

Mutants of Caenorhabditis elegans having about 10% of wild-type activity of the aspartyl protease cathepsin D have been isolated by screening. Mutant homozygotes have normal growth rates and no obvious morphological or developmental abnormalities. The mutant gene (cad-1) has been mapped to the right extremity of linkage group II. Heterozygous animals (cad-1/+) show intermediate enzyme levels and animals heterozygous for chromosomal deficiencies of the right extremity of linkage group II have 50% of wild-type activity. Cathepsin D purified from a mutant strain has a lower activity per unit mass of pure enzyme. These data suggest that cad-1 is a structural gene for cathepsin D.