FGF6 and FGF9 regulate UCP1 expression independent of brown adipogenesis.

Uncoupling protein-1 (UCP1) plays a central role in energy dissipation in brown adipose tissue (BAT). Using high-throughput library screening of secreted peptides, we identify two fibroblast growth factors (FGF), FGF6 and FGF9, as potent inducers of UCP1 expression in adipocytes and preadipocytes. Surprisingly, this occurs through a mechanism independent of adipogenesis and involves FGF receptor-3 (FGFR3), prostaglandin-E2 and interaction between estrogen receptor-related alpha, flightless-1 (FLII) and leucine-rich-repeat-(in FLII)-interacting-protein-1 as a regulatory complex for UCP1 transcription. Physiologically, FGF6/9 expression in adipose is upregulated by exercise and cold in mice, and FGF9/FGFR3 expression in human neck fat is significantly associated with UCP1 expression. Loss of FGF9 impairs BAT thermogenesis. In vivo administration of FGF9 increases UCP1 expression and thermogenic capacity. Thus, FGF6 and FGF9 are adipokines that can regulate UCP1 through a transcriptional network that is dissociated from brown adipogenesis, and act to modulate systemic energy metabolism.

Blockade of nonhormonal fibroblast growth factors by FP-1039 inhibits growth of multiple types of cancer.

The fibroblast growth factor (FGF) pathway promotes tumor growth and angiogenesis in many solid tumors. Although there has long been interest in FGF pathway inhibitors, development has been complicated: An effective FGF inhibitor must block the activity of multiple mitogenic FGF ligands but must spare the metabolic hormone FGFs (FGF-19, FGF-21, and FGF-23) to avoid unacceptable toxicity. To achieve these design requirements, we engineered a soluble FGF receptor 1 Fc fusion protein, FP-1039. FP-1039 binds tightly to all of the mitogenic FGF ligands, inhibits FGF-stimulated cell proliferation in vitro, blocks FGF- and vascular endothelial growth factor (VEGF)-induced angiogenesis in vivo, and inhibits in vivo growth of a broad range of tumor types. FP-1039 antitumor response is positively correlated with RNA levels of FGF2, FGF18, FGFR1c, FGFR3c, and ETV4; models with genetic aberrations in the FGF pathway, including FGFR1-amplified lung cancer and FGFR2-mutated endometrial cancer, are particularly sensitive to FP-1039-mediated tumor inhibition. FP-1039 does not appreciably bind the hormonal FGFs, because these ligands require a cell surface co-receptor, klotho or ß-klotho, for high-affinity binding and signaling. Serum calcium and phosphate levels, which are regulated by FGF-23, are not altered by administration of FP-1039. By selectively blocking nonhormonal FGFs, FP-1039 treatment confers antitumor efficacy without the toxicities associated with other FGF pathway inhibitors.

The mechanism of shared but distinct CSF-1R signaling by the non-homologous cytokines IL-34 and CSF-1.

Interleukin-34 (IL-34) and colony stimulating factor-1 (CSF-1) both signal through the CSF-1R receptor tyrosine kinase, but they have no sequence homology, and their functions and signaling activities are not identical. We report the crystal structures of mouse IL-34 alone and in complex with the N-terminal three immunoglobulin-like domains (D1-D3) of mouse CSF-1R. IL-34 is structurally related to other helical hematopoietic cytokines, but contains two additional helices integrally associated with the four shared helices. The non-covalently linked IL-34 homodimer recruits two copies of CSF-1R on the sides of the helical bundles, with an overall shape similar to the CSF-1:CSF-1R complex, but the flexible linker between CSF-1R D2 and D3 allows these domains to clamp IL-34 and CSF-1 at different angles. Functional dissection of the IL-34:CSF-1R interface indicates that the hydrophobic interactions, rather than the salt bridge network, dominate the biological activity of IL-34. To degenerately recognize two ligands with completely different surfaces, CSF-1R apparently takes advantage of different subsets of a chemically inert surface that can be tuned to fit different ligand shapes. Differentiated signaling between IL-34 and CSF-1 is likely achieved by the relative thermodynamic independence of IL-34 vs. negative cooperativity of CSF-1 at the receptor-recognition sites, in combination with the difference in hydrophobicity which dictates a more stable IL-34:CSF-1R complex compared to the CSF-1:CSF-1R complex.

Functional overlap but differential expression of CSF-1 and IL-34 in their CSF-1 receptor-mediated regulation of myeloid cells.

CSF-1 is broadly expressed and regulates macrophage and osteoclast development. The action and expression of IL-34, a novel CSF-1R ligand, were investigated in the mouse. As expected, huIL-34 stimulated macrophage proliferation via the huCSF-1R, equivalently to huCSF-1, but was much less active at stimulating mouse macrophage proliferation than huCSF-1. Like muCSF-1, muIL-34 and a muIL-34 isoform lacking Q81 stimulated mouse macrophage proliferation, CSF-1R tyrosine phosphorylation, and signaling and synergized with other cytokines to generate macrophages and osteoclasts from cultured progenitors. However, they respectively possessed twofold and fivefold lower affinities for the CSF-1R and correspondingly, lower activities than muCSF-1. Furthermore, muIL-34, when transgenically expressed in a CSF-1-dependent manner in vivo, rescued the bone, osteoclast, tissue macrophage, and fertility defects of Csf1(op)/(op) mice, suggesting similar regulation of CSF-1R-expressing cells by IL-34 and CSF-1. Whole-mount IL34 in situ hybridization and CSF-1 reporter expression revealed that IL34 mRNA was strongly expressed in the embryonic brain at E11.5, prior to the expression of Csf1 mRNA. QRT-PCR revealed that compared with Csf1 mRNA, IL34 mRNA levels were lower in pregnant uterus and in cultured osteoblasts, higher in most regions of the brain and heart, and not compensatorily increased in Csf1(op/op) mouse tissues. Thus, the different spatiotemporal expression of IL-34 and CSF-1 allows for complementary activation of the CSF-1R in developing and adult tissues.

Controlling feature selection in random forests of decision trees using a genetic algorithm: classification of class I MHC peptides.

Feature selection is an important challenge in many classification problems, especially if the number of features greatly exceeds the number of examples available. We have developed a procedure--GenForest--which controls feature selection in random forests of decision trees by using a genetic algorithm. This approach was tested through our entry into the Comparative Evaluation of Prediction Algorithms 2006 (CoEPrA) competition (accessible online at: http://www.coepra.org). CoEPrA was a modeling competition organized to provide an objective testing for various classification and regression algorithms via the process of blind prediction. In the competition GenForest ranked 10/23, 5/16 and 9/16 on CoEPrA classification problems 1, 3 and 4, respectively, which involved the classification of type I MHC nonapeptides i.e. peptides containing nine amino acids. These problems each involved the classification of different sets of nonapeptides. Associated with each amino acid was a set of 643 features for a total of 5787 features per peptide. The method, its application to the CoEPrA datasets, and its performance in the competition are described.