New insights into the natural course of knee osteoarthritis: early regulation of cytokines and growth factors, with emphasis on sex-dependent angiogenesis and tissue remodeling. A pilot study.

OBJECTIVE: This study was conducted to identify cytokine profiles associated with radiographic phenotypes of knee osteoarthritis (rKOA) with a focus on early stage of the disease. METHODS: The pilot population study involved 60 middle-aged patients (mean age 50 ± 7.3y.). Standardized weight-bearing anteroposterior and axial radiographs were used to assess rKOA severity in tibiofemoral (TFJ) of patellofemoral joint (PFJ) by grading system (grades 0-3). Luminex (xMAP®) technology was used to simultaneously assess 60 biomarkers (BMs). RESULTS: Several pathways of angiogenic (CXCL10/IP-10, FGF1/2, PDGF-AA/BB, ANG1, RANTES), tissue remodeling/fibrosis (MMP1/3, TIMP2/3/4, TGFß), and fat tissue (leptin) BMs associated with rKOA severity already in very early phase (grade 1). We identified several sets of cytokines as key markers of early knee osteoarthritis (KOA) predicting radiographic features in logistic-regression models (AUC = 0.80-0.97). Marked sex-specificity of rKOA course was detected: upregulation of angiogenesis dominated in females, whereas the activation of tissue remodeling was dominant in males. Several of these shifts, e.g., decrease of CXCL10/IP-10, took place only in grade 1 KOA and disappeared or reversed in later stages. OA of different knee-joint compartments has distinct profiles of cytokines. A broad list of BMs (TIMP2/3/4, MMP1/3, TGFß1/2, vWF-A2, sE-selectin and leptin) associated with OA in the PFJ. CONCLUSION: Our results demonstrate that substantial and time-limited shifts in the angiogenic and TIMP/MMP systems occur in the early stage of KOA. Our study findings highlight the sex-, grade- and compartment-dependent shifts in above processes. The data may contribute to the individualized prevention of KOA in the future.

Inhibition of Fgf signaling in short bowel syndrome increases weight loss and epithelial proliferation.

BACKGROUND: Signaling by fibroblast growth factor is critical for epithelial proliferation, differentiation, and the development of many organs, including the intestine. Fibroblast growth factor 10 and fibroblast growth factor 2c are upregulated after massive bowel resection during intestinal adaptation. This pathway is conserved highly. We hypothesized that inhibition of fibroblast growth factor signaling would impair intestinal adaptation in the zebrafish model of short bowel syndrome and allow insight into the negative regulation of this pathway. METHODS: Short bowel syndrome equivalent to a high jejunostomy was generated in adult male hsp70:dnfgfr1-GFP zebrafish, wildtype fish exposed to tyrosine-kinase inhibitor, and wildtype fish in absence of tyrosine-kinase inhibitor. Heat shock in hsp70:dnfgfr1-GFP fish decreases fgf 1 expression. Parameters including weight, proliferation, and differentiation were evaluated after harvest in experimental and control groups. RESULTS: Although short bowel syndrome zebrafish lost more weight relative to sham zebrafish in both groups, heat shock fish with short bowel syndrome lost more weight compared with non-heat shock fish with short bowel syndrome. In the non-heat shock controls, the villus epithelial perimeter increased in short bowel syndrome compared with sham fish, but this did not occur in heat shock fish. Non-heat shock fish with short bowel syndrome fish had significantly increased Bromodeoxyuridine(+) proliferative cells per hemivillus compared with non-heat shock-sham, while heat shock-short bowel syndrome had a more substantial increase in Bromodeoxyuridine(+) cells compared with HS-sham. Non-heat shock-short bowel syndrome demonstrated a significantly increased percentage of Alcian blue(+) goblet cells per hemivillus compared with non-heat shock-sham, while the heat shock-short bowel syndrome demonstrated decreased Alcian blue(+) cells compared with non-heat shock-short bowel syndrome. In contrast, SU5402 inhibited epithelial proliferation while increasing weight loss. CONCLUSION: Inhibition of fibroblast growth factor-1 signaling in short bowel syndrome decreases epithelial adaptation, increases Bromodeoxyuridine-labeled cells at 2 weeks, and exacerbates weight loss while decreasing epithelial goblet cells.

FGF1 Mediates Overnutrition-Induced Compensatory ß-Cell Differentiation.

Increased insulin demand resulting from insulin resistance and/or overnutrition induces a compensatory increase in ß-cell mass. The physiological factors responsible for the compensation have not been fully characterized. In zebrafish, overnutrition rapidly induces compensatory ß-cell differentiation through triggering the release of a paracrine signal from persistently activated ß-cells. We identified Fgf1 signaling as a key component of the overnutrition-induced ß-cell differentiation signal in a small molecule screen. Fgf1 was confirmed as the overnutrition-induced ß-cell differentiation signal, as inactivation of fgf1 abolished the compensatory ß-cell differentiation. Furthermore, expression of human FGF1 solely in ß-cells in fgf1(-/-) animals rescued the compensatory response, indicating that ß-cells can be the source of FGF1. Additionally, constitutive secretion of FGF1 with an exogenous signal peptide increased ß-cell number in the absence of overnutrition. These results demonstrate that fgf1 is necessary and FGF1 expression in ß-cells is sufficient for the compensatory ß-cell differentiation. We further show that FGF1 is secreted during prolonged activation of cultured mammalian ß-cells and that endoplasmic reticulum stress acts upstream of FGF1 release. Thus, the recently discovered antidiabetes function of FGF1 may act partially through increasing ß-cell differentiation.

FGF1-mediated cardiomyocyte cell cycle reentry depends on the interaction of FGFR-1 and Fn14.

Fibroblast growth factors (FGFs) signal through FGF receptors (FGFRs) mediating a broad range of cellular functions during embryonic development, as well as disease and regeneration during adulthood. Thus, it is important to understand the underlying molecular mechanisms that modulate this system. Here, we show that FGFR-1 can interact with the TNF receptor superfamily member fibroblast growth factor-inducible molecule 14 (Fn14) resulting in cardiomyocyte cell cycle reentry. FGF1-induced cell cycle reentry in neonatal cardiomyocytes could be blocked by Fn14 inhibition, while TWEAK-induced cell cycle activation was inhibited by blocking FGFR-1 signaling. In addition, costimulation experiments revealed a synergistic effect of FGF1 and TWEAK in regard to cardiomyocyte cell cycle induction via PI3K/Akt signaling. Overexpression of Fn14 with either FGFR-1 long [FGFR-1(L)] or FGFR-1 short [FGFR-1(S)] isoforms resulted after FGF1/TWEAK stimulation in cell cycle reentry of >40% adult cardiomyocytes. Finally, coimmunoprecipitation and proximity ligation assays indicated that endogenous FGFR-1 and Fn14 interact with each other in cardiomyocytes. This interaction was strongly enhanced in the presence of their corresponding ligands, FGF1 and TWEAK. Taken together, our data suggest that FGFR-1/Fn14 interaction may represent a novel endogenous mechanism to modulate the action of these receptors and their ligands and to control cardiomyocyte cell cycle reentry.

Oncogenic FGFR3 gene fusions in bladder cancer.

FGF receptor 3 (FGFR3) is activated by mutation or over-expression in many bladder cancers. Here, we identify an additional mechanism of activation via chromosomal re-arrangement to generate constitutively activated fusion genes. FGFR3-transforming acid coiled coil 3 (TACC3) fusions resulting from 4p16.3 re-arrangements and a t(4;7) that generates a FGFR3-BAI1-associated protein 2-like 1 (BAIAP2L1) fusion were identified in 4 of 43 bladder tumour cell lines and 2 of 32 selected tissue samples including the tumour from which one of the cell lines was derived. These are highly activated and transform NIH-3T3 cells. The FGFR3 component is identical in all cases and lacks the final exon that includes the phospholipase C gamma 1 (PLCγ1) binding site. Expression of the fusions in immortalized normal human urothelial cells (NHUC) induced activation of the mitogen-activated protein kinase pathway but not PLCγ1. A protein with loss of the terminal region alone was not as highly activated as the fusion proteins, indicating that the fusion partners are essential. The TACC3 fusions retain the TACC domain that mediates microtubule binding and the BAIAP2L1 fusion retains the IRSp53/MIM domain (IMD) that mediates actin binding and Rac interaction. As urothelial cell lines with FGFR3 fusions are extremely sensitive to FGFR-selective agents, the presence of a fusion gene may aid in selection of patients for FGFR-targeted therapy.

Transgenic expression of nonclassically secreted FGF suppresses kidney repair.

FGF1 is a signal peptide-less nonclassically released growth factor that is involved in angiogenesis, tissue repair, inflammation, and carcinogenesis. The effects of nonclassical FGF export in vivo are not sufficiently studied. We produced transgenic mice expressing FGF1 in endothelial cells (EC), which allowed the detection of FGF1 export to the vasculature, and studied the efficiency of postischemic kidney repair in these animals. Although FGF1 transgenic mice had a normal phenotype with unperturbed kidney structure, they showed a severely inhibited kidney repair after unilateral ischemia/reperfusion. This was manifested by a strong decrease of postischemic kidney size and weight, whereas the undamaged contralateral kidney exhibited an enhanced compensatory size increase. In addition, the postischemic kidneys of transgenic mice were characterized by hyperplasia of interstitial cells, paucity of epithelial tubular structures, increase of the areas occupied by connective tissue, and neutrophil and macrophage infiltration. The continuous treatment of transgenic mice with the cell membrane stabilizer, taurine, inhibited nonclassical FGF1 export and significantly rescued postischemic kidney repair. It was also found that similar to EC, the transgenic expression of FGF1 in monocytes and macrophages suppresses kidney repair. We suggest that nonclassical export may be used as a target for the treatment of pathologies involving signal peptide-less FGFs.

Acidic and basic fibroblast growth factors involved in cardiac angiogenesis following infarction.

Acidic and basic fibroblast growth factors (FGF-1/FGF-2) promote angiogenesis in cancer. Angiogenesis is integral to cardiac repair following myocardial infarction (MI). The potential regulation of FGF-1/FGF-2 in cardiac angiogenesis postMI remains unexplored. Herein, we examined the temporal and spatial expression of FGF-1/FGF-2 and FGF receptors (FGFR) in the infarcted rat heart at days 1, 3, 7, and 14 postMI. FGF-1/-2 gene and protein expression, cells expressing FGF-1/-2 and FGFR expression were examined by quantitative in situ hybridization, RT-PCR; western blot, immunohistochemistry and quantitative in vitro autoradiography. Compared to the normal heart, we found that in the border zone and infarcted myocardium 1) FGF-1 gene expression was increased in the first week postMI and returned to control levels at week 2; FGF-1 protein levels were, however, largely reduced at day 1, then elevated at day 3 peaked at day 7 and declined at day 14; and cells expressing FGF-1 were primarily inflammatory cells; 2) FGF-2 gene expression was significantly elevated from day 1 to day 14; the increase in FGF-2 protein level was most evident at day 7 and cells expressing FGF-2 were primarily endothelial cells; 3) FGFR expression started to increase at day 3 and remained elevated thereafter; and 4) FGF-1/FGF-2 and FGFR expression remained unchanged in the noninfarcted myocardium. Thus, FGF-1/FGF-2 and FGFR expression are enhanced in the infarcted myocardium in the early stage after MI, which is spatially and temporally coincident with angiogenesis, suggesting that FGF-1/FGF-2 are involved in regulating cardiac angiogenesis and repair.

Early growth response-1 induction by fibroblast growth factor-1 via increase of mitogen-activated protein kinase and inhibition of protein kinase B in hippocampal neurons.

BACKGROUND AND PURPOSE: The transcription factor early growth response-1 (Egr-1) and the acidic fibroblast growth factor (FGF-1) are involved in many regulatory processes, including hippocampus-associated learning and memory. However, the intracellular signalling mechanisms regulating Egr-1 in hippocampal cells are not entirely understood. EXPERIMENTAL APPROACH: We used primary mouse hippocampal neurons and the mouse hippocampal neuronal cell line HT22 to investigate how FGF-1 transiently induces Egr-1 protein. This was accomplished by a range of techniques including Western blotting, immunofluorescence, specific protein kinase inhibitors and transfectable constitutively active protein kinase constructs. KEY RESULTS: Protein kinase B (PKB) and mitogen-activated protein kinase (MAPK) were both initially phosphorylated and activated by FGF-1 treatment, but when phosphorylated MAPK reached maximal activation, phosphorylated PKB was at its lowest levels, suggesting an interaction between MAPK kinase (MEK-1/2) and phosphatidyl inositol-3-kinase (PI3K) during Egr-1 induction. Interestingly, pharmacological inhibition of MEK-1/2 resulted in a robust increase in the phosphorylation of PKB, which was repressed in the presence of increasing doses of a PI3K inhibitor. FGF-1-mediated Egr-1 induction was impaired by inhibition of MEK-1/2, but not of PI3K. However, elevated levels of PKB, induced by transfection of constitutively active PKB (myrAkt) into hippocampal neuronal HT22 cells, led to reduced levels of Egr-1 after FGF-1 application. CONCLUSIONS AND IMPLICATIONS: Our data indicate a contribution of inactive (dephosphorylated) PKB to FGF-1-mediated induction of Egr-1, and strongly suggest a functionally and pharmacologically interesting cross-talk between MEK-1/2 and PI3K signalling in hippocampal neurons after FGF-1 stimulation that may play a role in hippocampal synaptic plasticity.

A novel fibroblast growth factor-1 (FGF1) mutant that acts as an FGF antagonist.

BACKGROUND: Crosstalk between integrins and FGF receptors has been implicated in FGF signaling, but the specifics of the crosstalk are unclear. We recently discovered that 1) FGF1 directly binds to integrin alphavbeta3, 2) the integrin-binding site and FGF receptor (FGFR) binding site are distinct, and 3) the integrin-binding-defective FGF1 mutant (R50E) is defective in inducing FGF signaling although R50E still binds to FGFR and heparin and induces transient ERK1/2 activation. PRINCIPAL FINDINGS: We tested if excess R50E affect DNA synthesis and cell survival induced by WT FGF1 in BaF3 mouse pro-B cells expressing human FGFR1. R50E suppressed DNA synthesis and cell proliferation induced by WT FGF1. We tested if WT FGF1 and R50E generate integrin-FGF1-FGFR ternary complex. WT FGF1 induced ternary complex formation (integrin-FGF-FGFR1) and recruitment of SHP-2 to the complex in NIH 3T3 cells and human umbilical endothelial cells, but R50E was defective in these functions. It has been reported that sustained ERK1/2 activation is integrin-dependent and crucial to cell cycle entry upon FGF stimulation. We thus determined the time-course of ERK1/2 activation induced by WT FGF1 and R50E. We found that WT FGF1 induced sustained activation of ERK1/2, but R50E was defective in this function. CONCLUSIONS/SIGNIFICANCE: Our results suggest that 1) R50E is a dominant-negative mutant, 2) Ternary complex formation is involved in FGF signaling, 3) The defect of R50E to bind to integrin may be directly related to the antagonistic action of R50E. Taken together, these results suggest that R50E has potential as a therapeutic in cancer.

Fibroblast growth factor-1 within the ventral tegmental area participates in motor sensitizing effects of morphine.

Drug addiction is viewed as a form of neural plasticity, and neurotrophic factors have been implicated in many forms of plasticity in the adult nervous system. Here we show that the fibroblast growth factor-1 (FGF-1), that is expressed on dopamine and GABA neurons of the ventral tegmental area (VTA), is involved in the sensitizing effects of morphine. The receptor FGFR-1 is expressed on VTA astrocytes, as well as dopamine and GABA neurons. FGF-1 or anti-FGF-1 infusions into the VTA during the induction (not expression) phase of sensitization advanced or blocked morphine's activating motor effects respectively, in a dose-dependent manner. Infusions into the adjacent substantia nigra, whose neurons also express FGF-1 and FGFR-1, did not modify normal morphine-induced sensitization. Biochemical traits related to morphine's sensitizing effects were altered by intra-VTA anti-FGF-1 because morphine-induced upregulation of both tyrosine hydroxylase (TH) and N-methyl d-aspartate glutamate receptor 1 (NMDAR1) in the VTA was blocked after anti-FGF-1. Changes in the activation state of VTA calcium/calmodulin-dependent kinase type II seem to participate in FGF-1-induced effects as well. We conclude that the FGF-1 system of the ventral tegmental area is required for biochemical and behavioral sensitization to this drug.

Protein engineering of a fibroblast growth factor-1 fusion protein with cell adhesive activity.

Fibroblast growth factor-1 (FGF1) is one of the most potent angiogenic growth factors, and also plays an important role in regulating cellular functions including cell proliferation, motility, differentiation, survival, and tissue regeneration processes. Here we described a novel fusion protein that was designed by combining the cell adhesion sequence from fibronectin with FGF1. The F1-Fn fusion protein functions as a minimized protein that directs integrin-dependent cell adhesion and stimulates cellular responses including cell proliferation and differentiation. Moreover, our results indicate that Fn-mediated signaling synergizes with signals from FGF1 in promoting cellular adhesion, proliferation, and differentiation in MG63 cells.

Involvement of acidic fibroblast growth factor in spinal cord injury repair processes revealed by a proteomics approach.

Acidic fibroblast growth factor (aFGF; also known as FGF-1) is a potent neurotrophic factor that affects neuronal survival in the injured spinal cord. However, the pathological changes that occur with spinal cord injury (SCI) and the attribution to aFGF of a neuroprotective effect during SCI are still elusive. In this study, we demonstrated that rat SCI, when treated with aFGF, showed significant functional recovery as indicated by the Basso, Beattie, and Bresnahan locomotor rating scale and the combined behavior score (p < 0.01-0.001). Furthermore proteomics and bioinformatics approaches were adapted to investigate changes in the global protein profile of the damaged spinal cord tissue when experimental rats were treated either with or without aFGF at 24 h after injury. We found that 51 protein spots, resolvable by two-dimensional PAGE, had significant differential expression. Using hierarchical clustering analysis, these proteins were categorized into five major expression patterns. Noticeably proteins involved in the process of secondary injury, such as astrocyte activation (glial fibrillary acidic protein), inflammation (S100B), and scar formation (keratan sulfate proteoglycan lumican), which lead to the blocking of injured spinal cord regeneration, were down-regulated in the contusive spinal cord after treatment with aFGF. We propose that aFGF might initiate a series of biological processes to prevent or attenuate secondary injury and that this, in turn, leads to an improvement in functional recovery. Moreover the quantitative expression level of these proteins was verified by quantitative real time PCR. Furthermore we identified various potential neuroprotective protein factors that are induced by aFGF and may be involved in the spinal cord repair processes of SCI rats. Thus, our results could have a remarkable impact on clinical developments in the area of spinal cord injury therapy.

FGF-1: from biology through engineering to potential medical applications.

Human fibroblast growth factor 1 (FGF-1) is one of the best characterized members of the FGF superfamily. FGF-1 is a powerful mitogen exhibiting strong action on numerous different cell types. It plays a role in various stages of development and morphogenesis, as well as in angiogenesis and wound healing processes. Engineering of FGFs can bring many advantages. Design and construction of different mutants can contribute to a better understanding of the mechanism of action of protein growth factors. Moreover, application of FGFs as recombinant polypeptides in the treatment of wound and fracture healing, cardiovascular diseases and neurodegenerative diseases seems to be a rational medical approach. However, low thermal stability and high sensitivity to proteases limit the potential pharmaceutical use of wild-type FGFs. Thus, advanced protein design techniques and recombinant protein production can help to obtain new variants of FGFs with radically increased thermodynamic stability, prolonged half-life and improved proteolytic resistance. Such studies can provide a good starting point to convert short-lived and/or sensitive growth factors to effective therapeutic proteins.

Could chronic wounds not heal due to too low local copper levels?

Copper is an essential trace element involved in numerous human physiological and metabolic processes. It plays a key role in many of the processes that together comprise wound healing, including induction of endothelial growth factor, angiogenesis and expression and stabilization of extracellular skin proteins. We hypothesize that in individuals with diabetic ulcers, decubitus, peripheral vascular, or other wounds which might have compromised circulation to the wound site, that part of the incapacity of the wounds to heal is due to low local copper levels. Contamination of wounds is also an important factor causing impaired wound healing. Importantly, copper has potent broad biocidal properties. In contrast, the risk of adverse skin reactions due to exposure to copper is extremely low. We thus hypothesize that introducing copper into wound dressings would not only reduce the risk of wound and dressing contamination, as silver does but, more importantly, would stimulate faster wound repair directly. This would be done by the release of copper from the wound dressings directly into the wound site inducing angiogenesis and skin regeneration.

Microarray analyses of transdifferentiated mesenchymal stem cells.

The molecular events associated with the age-related gain of fatty tissue in human bone marrow are still largely unknown. Besides enhanced adipogenic differentiation of mesenchymal stem cells (MSCs), transdifferentiation of osteoblast progenitors may contribute to bone-related diseases like osteopenia. Transdifferentiation of MSC-derived osteoblast progenitors into adipocytes and vice versa has previously been proven feasible in our cell culture system. Here, we focus on mRNA species that are regulated during transdifferentiation and represent possible control factors for the initiation of transdifferentiation. Microarray analyses comparing transdifferentiated cells with normally differentiated cells exhibited large numbers of reproducibly regulated genes for both, adipogenic and osteogenic transdifferentiation. To evaluate the relevance of individual genes, we designed a scoring scheme to rank genes according to reproducibility, regulation level, and reciprocity between the different transdifferentiation directions. Thereby, members of several signaling pathways like FGF, IGF, and Wnt signaling showed explicitly differential expression patterns. Additional bioinformatic analysis of microarray analyses allowed us to identify potential key factors associated with transdifferentiation of adipocytes and osteoblasts, respectively. Fibroblast growth factor 1 (FGF1) was scored as one of several lead candidate gene products to modulate the transdifferentiation process and is shown here to exert inhibitory effects on adipogenic commitment and differentiation.

fgf1 is required for normal differentiation of erythrocytes in zebrafish primitive hematopoiesis.

Hematopoiesis in vertebrate development involves an embryonic, primitive wave and a later, definitive wave in which embryonic blood cells are replaced with adult blood cells. We here show that zebrafish fgf1 is involved in vivo in primitive hematopoiesis. Fibroblast growth factor-1 (FGF1) morpholino knockdown leads to abnormal accumulation of blood cells in the posterior intermediate cell mass at 32 hr postfertilization. Expression of the erythroid markers gata1 and ika, normally diminishing in differentiating erythrocytes at this stage, is maintained at abnormally high levels in primitive blood cells. The onset of erythrocyte differentiation as assessed by o-dianisidine staining is severely delayed. Most fgf1 morphants later recover to wild-type appearance, and primitive erythrocytes eventually differentiate. Zebrafish fgf1 is syntenic to human FGF1, which maps to a critically deleted region in human del(5q) syndrome posing an increased risk of leukemia to patients. As its knockdown in zebrafish changes expression of gata1, a gene involved in hematopoietic stem cell decisions, FGF1 should be considered to play a role in the pathogenesis of del(5q) syndrome.

Three faces of mortalin: a housekeeper, guardian and killer.

Mortalin was first cloned as a mortality factor that existed in the cytoplasmic fractions of normal, but not in immortal, mouse fibroblasts. A decade of efforts have expanded its persona from a house keeper protein involved in mitochondrial import, energy generation and chaperoning of misfolded proteins, to a guardian of stress that has multiple binding partners and to a killer protein that contributes to carcinogenesis on one hand and to old age disorders on the other. Being proved to be an attractive target for cancer therapy, it also warrants attention from the perspectives of management of old age diseases and healthy aging.

Sustained low levels of fibroblast growth factor-1 promote persistent microvascular network formation.

BACKGROUND: Therapeutic neovascularization using high growth factor concentrations may lead to transient vessel formation and abnormal microvascular structure. The goal of this study was to quantify temporal and concentration effects of fibroblast growth factor-1 (FGF-1) on the persistence and morphology of microvascular networks. METHODS: Endothelial cells were incubated in suspension culture forming aggregates that were embedded in fibrin glue (FG) and stimulated with varying concentrations of FGF-1 with of heparin. Capillary networks formed were quantified for 21 days. RESULTS: High FGF-1 concentrations resulted in rapid and intense sprout formation, with excessive branching. At later times, these vessels regressed, with cellular debris in former vessel locations. At later times, the 1-ng/mL group surpassed the high concentration groups with continuous sprout growth and complete FG vascularization by 23 days. CONCLUSION: Sustained low levels of FGF-1 maintained a persistent microvascular network response, whereas higher levels resulted in abnormal phenotype followed by vessel regression.

Forced expression of hepatocyte-specific fibroblast growth factor 21 delays initiation of chemically induced hepatocarcinogenesis.

Inappropriate fibroblast growth factor (FGF) signaling is involved in most tissue-specific pathologies including cancer. Previously we showed that inappropriate expression and chronic activity of FGF receptor (FGFR) 1 in hepatocytes accelerated diethylnitrosamine (DEN)-initiated hepatocarcinogenesis. Here we showed that although widely expressed FGF1 and FGF2 are frequently upregulated in hepatocellular carcinoma (HCC), germline deletion of both FGF1 and FGF2 had no effect on DEN-initiated hepatocarcinogenesis. Thus overexpression of FGF1 or FGF2 may be a consequence rather than contributor to hepatoma progression. FGF21 is the first of 22 homologues whose expression has been reported to be preferentially in the liver. We showed that similar to FGF1 and FGF2, FGF21 mRNA was upregulated in neoplastic and regenerating liver after partial hepatectomy (PH) and CCl4 administration. In situ hybridization analysis confirmed that in contrast to FGF1 and FGF2, expression of FGF21 mRNA was limited to hepatocytes. Forced overexpression of FGF21 in hepatocytes by gene targeting had no apparent impact on normal liver development and compensatory response to injury. Surprisingly, overexpression of FGF21 delayed the appearance of DEN-induced liver tumors. At 8 and 10 mo, only 10% and 30% of transgenic mice, respectively, developed adenomas compared to 50% (all adenomas) and 80% (60% adenoma/20% HCC) in the wild-type (WT) mice. However, the incidence and burden of HCC at 10 mo and later was equal in the FGF21 transgenic and WT mice. We propose that FGF21 may delay development of adenomas through activation of resident hepatocyte FGFR4 at early times, but counteracts the delay by acceleration of progression to HCC through interaction with ectopic FGFR1 once it appears in hepatoma cells. This indicates a dual function of FGF21 that may reflect changes in FGFR isotype during progression of differentiated hepatoma cells.