Reconstitution of a pentameric complex of dimeric transforming growth factor beta ligand and a type I, II, III receptor in baculoviral-infected insect cells

Two transmembrane serine-threonine kinases (type I and II receptors), a membrane-anchored proteoglycan (type III), and a homodimeric ligand participate in the transforming growth factor beta type one (TGF beta 1) signal transduction complex. The expression of recombinant receptors in insect cells co-infected with up to three recombinant baculoviruses was employed to study interactions among the ectodomains of the three types of receptors and the TGF beta 1 ligand in absence of uncontrollable extrinsic factors in mammalian cells. Multi-subunit complexes were assembled in intact cells and purified on glutathione-conjugated beads for analysis by tagging one of the subunits with glutathione S-transferase (GST). Intrinsic ligand-independent interactions were observed among receptor subunits as follows: type III-III, type I-I, type III-I, and type II-I. The homeotypic complex of type II-II receptors and the heterotypic type III-II interaction was ligand dependent. The type I, but not the type III, subunit displaced about 50% of the type II component in either ligand-dependent homomeric type II-type II complexes or heteromeric type III-type II complexes to form type II-I or type III-II-I oligomers, respectively. The type II subunit displaced type I subunits in oligomers of the type I subunit. Specificity of type I receptors may result from differential affinity for the type II receptor rather than specificity for ligand. A monomeric subunit of the TGF beta 1 ligand bound concurrently to type III and type II or type III and type I receptors, but failed to concurrently bind to the type II and type I subunits. The binding of TGF beta 1 to the type I kinase subunit appears to require an intact disulfide-linked ligand dimer in the absence of a type III subunit. The combined results suggest a pentameric TGF beta signal transduction complex in which one unit each of the type III, type II, and type I components is assembled around the two subunits of the dimeric TGF beta ligand. An immobilized GST-tagged subunit of the receptor complex was utilized to assemble multi-subunit complexes in vitro and to study the phosphorylation events among subunits in the absence of extrinsic cell-derived kinases. The results revealed that (a) a low level of ligand-independent autophosphorylation occurs in the type I kinase; (b) a high level of autophosphorylation occurs in the type II kinase; (c) both the type III and type I subunits are trans-phosphorylated by the type II subunit; and (d) the presence of both type I and II kinases complexed with the type III subunit and dimeric TGF beta 1 ligand in a pentameric complex causes maximum phosphorylation of all three receptor subunits.

Androgen-dependent expression of fibroblast growth factor-1 in submaxillary gland of mouse

We have purified a 16,000 dalton protein that stimulates growth of human umbilical cord vein-derived endothelial cells (HUV-EC) from mouse submaxillary glands by using heparin-Sepharose affinity and C4 reverse phase chromatography. The purified molecule was identified as an FGF-1 on the basis of its biological activities, its affinity for heparin and its N-terminal amino-acid sequence. The concentrations of FGF-1 in the submaxillary gland of male or testosterone-treated female mice were about 12 times those of untreated females or castrated males. The 2.3 and 4.1 kb FGF-1 mRNAs were expressed in the glands of male mice older than 4 weeks but not in the glands of female mice. These results suggest that FGF-1 may have important functions for growth, differentiation and development of mouse submaxillary glands, and it may act as an endocrine hormone.

Immunohistochemical study of overexpression of fibroblast growth factor-1 (FGF-1), FGF-2, and FGF receptor-1 in human malignant salivary gland tumours

Fibroblast growth factor-1 (FGF-1) and FGF-2 are broad spectrum mitogens. The expression of FGF-1, FGF-2, and their receptor, FGF receptor-1 (FGFR-1), was examined in malignant salivary gland tumours and normal salivary glands, using immunohistochemical methods. In seven cases of adenoid cystic carcinoma (ACC), both duct-like cells and modified myoepithelial cells were apparently immunopositive for FGF-1, FGF-2, and FGFR-1. In five cases of mucoepidermoid carcinoma (MC), all three types of tumour cells including epidermoid cells, and intermediate cells expressed immunoreactive FGF-1, FGF-2, and FGFR-1. In these malignant salivary gland tumours, increased expression of FGFR-1 correlated with the intensity of both FGF-1 and FGF-2 immunoreactivity. In contrast to malignant salivary gland tumours, eight cases of normal salivary gland showed negative immunostaining for FGF-1, FGF-2, and FGFR-1 while four cases were weakly immunoreactive for FGF and its receptor. These results demonstrate that malignant salivary gland tumours overexpress FGF-1, FGF-2, and FGFR-1 compared with normal salivary glands and suggest that these growth factors may play an important role in facilitating neoplastic progression in human salivary glands.

Expression of fibroblast growth factor-1 (FGF-1), FGF-2 and FGF receptor-1 in a human salivary-gland adenocarcinoma cell line: evidence of growth

Fibroblast growth factor-1 (FGF-1) and FGF-2 are heparin-binding polype ptides which express potent mitogenic properties in neoplastic cells. In the present study, we have examined the contribution of endogenous FGF-1 and FGF-2 to the autocrine growth of HSY human salivary-gland adenocarcinoma cells in vitro. Using specific monoclonal antibodies against FGF-1 and FGF-2, immunohistochemical analysis of HSY cells revealed strong expression of both FGF-1 and FGF-2 in the cytoplasm and nucleus. Consistent with these data, 2 molecular mass species of FGF-1 (16 and 18 kDa) and 3 FGF-2 (18, 24 and 27 kDa) were identified in HSY cells by Western-blot analysis. Scatchard analysis of FGF binding sites on HSY cells indicated the presence of 23,000 [125I]FGF-1 binding sites/cells with a dissociation constant (KD) of 178 pM and 13,000 [125I]FGF-2 binding sites/cell with a KD of 102 pM. In addition, HSY cells were shown to express the mRNA for FGF receptor-1 (FGFR-1) by reverse transcription-polymerase chain reaction (RT-PCR), confirming the existence of high-affinity FGF binding sites. The influence of endogenous FGF-1 and FGF-2 on HSY cell growth was evaluated by suppressing the expression and activity of FGF by using anti-sense oligonucleotides and neutralizing antibodies. The addition of 50 micron FGF-1-specific anti-sense oligonucleotides to HSY cells resulted in a 61% inhibition of cell growth, while 50 microM FGF-2-specific anti-sense oligonucleotides resulted in a 76% inhibition. These effects were dose-dependent and specific, since sense oligonucleotides were ineffective in inhibiting HSY cell growth at the same concentration. Furthermore, HSY cell growth was suppressed in the presence of anti-FGF-1 or anti-FGF-2 neutralizing antibody, resulting in a 58% inhibition at 8 micromilligrams/ml. Our observations suggest that FGF-1 and FGF-2 may act as autocrine regulators by interacting with FGF receptors on HSY cells.

Nuclear localization of a complex of fibroblast growth factor(FGF)-1 and an NH2-terminal fragment of FGF receptor isoforms R4 and R1alpha in human liver cells

FGF ligands and FGF receptor 1 (FGFR1) appear associated with the nucleus in addition to their extracellular and transmembrane locations. After receptor-dependent internalization in liver cells, radiolabeled 16-kDa FGF-1 appears in a 40-kDa covalent complex with a cellular protein. In this report, we show that in a human hepatoma cell line, HepG2, which expresses both FGFR4 and FGFR1, the 40-kDa complex cross-reacts with antibodies against the ectodomain of both types of receptors. In addition to antibody against FGF-1, a polyclonal antiserum against the three immunoglobulin (Ig)-like loop ectodomain of FGFR4 and a monoclonal antibody to a 19-residue sequence in the NH2-terminus of the NH2-terminal Ig Loop I of the three loop splice variant of FGFR1 (FGFR1alpha) reacts with the complex. A monoclonal antibody against an epitope in FGFR1 downstream of the inter-loop I/II sequence which reacts with intact FGFR1 failed to cross-react with the 40-kDa complex. Cell fractionations and indirect immunofluorescent localization revealed that the 40-kDa complex associates with the particulate fraction of cells, particularly the nucleus and associated cytoskeletal elements. We propose that the NH2-terminal Ig-loop of the three loop isoforms of FGFR, which are generally associated inversely with cell growth, may play a role at or in the nucleus in addition to modification of affinity of the FGFR ectodomain for heparan sulfate and FGF ligand.

Expression of CD44 splice variants in bladder cancer

BACKGROUND

Recently, it was noted that CD44 participates in a wide variety of cell-cell or cell-matrix interactions including tumor metastasis and the altered expression of CD44 splice variants is helpful in the diagnosis of colon and breast cancer. The purpose of the study was to investigate altered expression of CD44 in bladder cancer.

METHODS

Thirty-three surgical specimens of 22 patients and seven bladder cancer cell-lines were analyzed by various methods including RT-PCR-Southern blotting. Northern blotting, and flow-cytometric analysis.

RESULTS

In RT-PCR analysis, the hematopoietic variant and numerous additional variants were detected in all cell-lines and almost all surgical specimens. In three out of seven cell-lines of bladder cancer (42.9%), especially numerous larger bands (ladder pattern) were detected using RT-PCR. These three cell-lines expressed a smaller quantity of CD44 mRNAs and CD44 molecules than the other four cell-lines. In eighteen out of 22 (81.8%) bladder cancer specimens, we detected larger variants that exceed 1500 base pairs, compared with the detection in four out of eleven (36.4%) visually normal mucosa specimens (P < 0.05). However, there was no correlation between the expression of the larger variants and clinicopathological features.

CONCLUSION

The quantitative and qualitative change of CD44 variants were demonstrated in bladder cancer.

High-level expression of the CD44 variant sharing exon v10 in renal cancer

To examine whether renal cell carcinoma displays altered CD44 expression we performed reverse transcription-polymerase chain reaction (RT-PCR) analysis of CD44 in 38 specimens from renal cancer, normal kidney and metastases of 19 patients and 6 renal cancer cell lines. To detect the CD44 variants, we utilized the RT-PCR Southern blot method. One out of 19 (5.3%) renal cancer specimens expressed a larger molecular weight band than 1 kb by RT-PCR analysis, in contrast to previous findings in colon and breast cancer. The band patterns in RT-PCR were different in 14/17 (82.4%) cases between normal kidney and tumors, and a band of about 700 bp was especially marked in 12/17 (70.6%) tumor specimens and 4/6 (66.7%) cell lines. By cloning and sequencing of the 700 bp band, we found that this variant is identical to the CD44 variant sharing only exon v10. Examination by Northern blot analysis has revealed that all tumors express a higher level of CD44 mRNA than paired normal kidneys. These findings suggested that the CD44 variants sharing exon v10 play some role in renal cancer.

Frequency of mutations of insulin receptor gene in Japanese patients with NIDDM

To examine the prevalence of abnormalities in the insulin receptor structure gene in Japanese with non-insulin-dependent diabetes mellitus (NIDDM), a population of 51 patients with NIDDM was screened for mutations in this gene. Patient genomic DNAs of both alleles corresponding to 22 exons of the gene were amplified by polymerase chain reaction (PCR). The PCR products on pUC19 were sequenced. Three patients with heterozygous missense mutation Thr831-->Ala831 in exon 13 and one patient with heterozygous missense mutation Tyr1334-->Cys1334 in exon 22 of the beta-subunits were identified. Linkage analysis of one of the families plus statistical studies showed that the mutation Thr831-->Ala831 is possibly responsible for the onset of NIDDM. In COS cells transiently expressing both mutant receptor cDNAs and a cDNA of a M(r) 85,000 regulatory subunit of phosphatidylinositol 3-kinase (PI 3-kinase), the mutation Tyr1334-->Cys1334 impaired binding of the receptor with the M(r) 85,000 subunit of PI 3-kinase, but linkage analysis of the family showed that the mutation did not cosegregate with NIDDM in the pedigree. Therefore, one missense mutation (Thr831-->Ala831) in the insulin receptor, as found in three patients, is possibly involved in the etiology of a subset of the 51 NIDDM patients.

[Expression of CD44 variant form in human renal cell carcinoma]

To examine whether renal cell carcinoma display altered CD44 expression we performed reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of CD44 in the tissues resected from 19 patients with renal cell carcinoma and 6 renal cancer cell lines. To detect the CD44 variants, we utilized the RT-PCR Southern blot method reported by Matsumura et al. In 12 of 17 (70.6%) cases, about a 700 base pairs band was emphasized in cancerous tissues compared with normal kidney. Moreover, we found that this isoform is the CD44 variant sharing only exon v10. Examination by Northern blot analysis has revealed that all tumors express a higher level of CD44 variants sharing exon v10. Our findings suggest that this variant form plays some roles in renal cell carcinoma.

Alternately spliced NH2-terminal immunoglobulin-like Loop I in the ectodomain of the fibroblast growth factor (FGF) receptor 1 lowers affinity for both heparin and FGF-1

Alternate splicing of a single exon encoding an NH2-terminal immunoglobulin (Ig) disulfide loop in the ectodomain of the fibroblast growth factor receptor (FGFR) types 1 and 2 results in alpha and beta isoforms that exhibit 3- and 2-Ig loops, respectively. Previously we demonstrated that alternately spliced Loop I has no independent ligand binding activity but is sufficiently interactive with the ligand- and heparin-binding site formed by Loops II and III to lower affinity for the same fibroblast growth factor (FGF) ligand. Here we show that a lower affinity of FGFR1 alpha for heparin parallels the lower affinity for FGF-1. A mutant of FGFR1 alpha in which the sequence between Loops I and II was deleted exhibits high affinity for both FGF-1 and heparin and other properties of the FGFR1 beta isoform, which include resistance to degradation by trypsin and display of specific antibody epitopes. This suggests that the interloop sequence facilitates the interaction of Loop I with Loops II and III. Lack of expression of both exons coding for Loop I and the sequence between Loops I and II in the FGFR2 gene characterizes rat prostate tumor cells, which exhibit a loss of the low affinity class of FGF receptors. Although the exon coding for the sequence between Loops I and II is alternately spliced in the FGFR2 beta isoform, coordinate expression with the exon coding for Loop I results in the functional differences between the FGFR alpha and FGFR beta variants.

Ligand-specific structural domains in the fibroblast growth factor receptor

Two tandem immunoglobulin-like disulfide loops (Loops II and III) linked by a short connecting sequence in the ectodomain of the fibroblast growth factor receptor kinase compose the binding sites for glycosaminoglycan and fibroblast growth factor (FGF) ligands. Alternate splicing of exons IIIb and IIIc coding for the COOH-terminal half of Loop III confers high affinity for FGF-7 or FGF-2, respectively, on the fibroblast growth factor receptor ectodomain without effect on the binding of FGF-1. Here we show that a 139-amino acid fragment composed of Loop II, the inter-Loop II/III sequence, and a short segment of the NH2 terminus of Loop III is sufficient and near the minimal requirement for binding of FGF-1, FGF-2, and FGF-7. Extension of the fragment by five additional highly conserved residues (SD(P/A)QP) within a distinct constitutive structural domain (fl1) in Loop III restricts the binding of FGF-7 without effect on FGF-1 and FGF-2. Since the presence of exon IIIc in the full-length ectodomain does not change this ligand binding profile, we suggest that alternately spliced exon IIIc plays no active role in binding of the three ligands. In contrast, exon IIIb actively abrogates the restriction on the binding of FGF-7 and concurrently lowers the affinity for FGF-2.

Variation in equation coefficients in the gravimetric method to determine brain water content

The gravimetric method to determine the tissue water content is widely used for studying brain edema because the tissue water content can be calculated from only the density of the wet tissue using a gradient column. This method has been used in only a few studies of cerebellar water content. We estimated the water content of the cerebellar cortex by the gravimetric method and compared the results with those measured by the conventional drying-weighing method, and found a significant difference. We therefore propose modified coefficients for the equation for the water content of the cerebellar cortex as follows: % tissue water content = 456.13(1/wet tissue density) - 356.13 This equation and the drying-weighing method were used to investigate normal and edematous cerebellar cortices after compression with an epidural balloon. The modified equation provided results consistent with the drying-weighing method. Differences in the total experimental system including the breed and pre-experimental condition of the cats may be factors in variations in measured values. We recommend that investigators establish equation coefficients valid for individual experimental systems.

The role of insulin in activation of two enhancers in the mouse GLUT1 gene

We identified two enhancer elements of the mouse GLUT1 gene responsive to serum, growth factor, and oncogenes; the first enhancer element (enhancer-1) is located 2.7 kilobases upstream of the cap site of the gene, and the second one (enhancer-2) is located in the second intron of the gene (Murakami, T., Nishiyama, T., Shirotani, T., Shinohara, Y., Kan, M., Ishii, K., Kanai, F., Nakazuru, S., and Ebina, Y. (1992) J. Biol. Chem. 267, 9300-9306). In the present work, we describe the role of insulin in activation of these two enhancers. NIH/3T3 HIR3.5 cells, which express a large number of insulin receptors, were stably transformed by hybrid genes containing the enhancer(s) and promoter of GLUT1 gene and the coding region of chloramphenicol acetyltransferase (CAT) gene as a reporter gene. In stable transformants of the reporter gene without the enhancers, the CAT mRNA was not induced by insulin; however, in clones containing the reporter gene with enhancer-1, the CAT mRNA was induced by insulin at 30 min and reached a maximum at 1 h. In clones transfected by the reporter gene with enhancer-2, the CAT mRNA was induced at 1 h and reached a maximum at 3 h. To determine the early response element to insulin in enhancer-1, transformants of hybrid reporter genes containing truncated or mutated enhancer-1 were examined. The homologous sequence with the serum response element in enhancer-1 is essential for an early response to insulin.

Inositolhexakisphosphate (InsP6): an antagonist of fibroblast growth factor receptor binding and activity

Fibroblast growth factors (FGF), which have been implicated in tumor cell growth and angiogenesis, have biological activities that appear to be mediated by both heparinlike extracellular matrix sites and transmembrane tyrosine kinase receptor sites. In the present study, we demonstrated that inositolhexakisphosphate (InsP6) inhibits basic FGF (bFGF) binding to heparin. Our spectrofluorometric analyses demonstrated that InsP6 not only bound to bFGF, presumably within the bFGF heparin-binding domain, but also protected bFGF from degradation by trypsin. Also, InsP6 inhibited the cellular binding of bFGF and other fibroblast growth factor family members such as acidic FGF (aFGF) and K-FGF in a saturable and dose-dependent manner. Furthermore, concentrations as low as 100 microM InsP6 inhibited bFGF-induced DNA synthesis in AKR-2B fibroblasts, as well as the growth of bFGF- and K-FGF-transfected NIH/3T3 cells. Together, these results indicate that InsP6 may serve as a useful antagonist of FGF activity.

Release of fibroblast growth factor-1 by human squamous cell carcinoma correlates with autocrine cell growth

A squamous cell carcinoma cell line Nakata proliferated in serum-free culture and was not responsive to exogenous fibroblast growth factor-1 (FGF-1). Immunostaining revealed that Nakata cells expressed FGF-1 in their cytoplasms and nuclei. Two molecular mass species of FGF-1 (16 and 18 kDa) were identified in cell extracts by Western blot. These cells also expressed high-affinity FGF-1 binding sites (Kd = 360 pM, 28,000 sites/cell). The results of cross-linking with [125I]FGF-1 demonstrated the presence of two bands with molecular masses of 160 and 140 kDa. The addition of FGF-1 specific antisense oligonucleotides at 25 microM to Nakata cells resulted in an 82% inhibition in cell growth and suppressed FGF-1 expression. This effect was dose-dependent and specific, because sense oligonucleotides were ineffective in inhibiting cell growth. In addition, Nakata cell growth was suppressed by an anti-FGF-1 neutralizing antibody, which resulted in a 52% inhibition at 8 micrograms/ml. These results demonstrate that Nakata cells produce FGF-1, and indicate that this growth factor acts in an autocrine manner by interacting with FGF-1 binding sites on Nakata cells.

Heparan sulfate fibroblast growth factor receptor complex: structure-function relationships

Splice variations in genes coding for the transmembrane FGF receptor (FGFR) result in isoforms that vary in the ectodomain, intracellular juxtamembrane domain, and the intracellular kinase domain. An analysis of biochemical functions of distinct recombinant isoforms expressed in baculoviral-infected insect cells allowed generation of models for function of splice variants in both the ecto- and intracellular domains. A structural model for the ectodomain of the FGFR is proposed as follows. Alternately-spliced immunoglobulin-like disulfide Loop I, which is not required for ligand-binding, is sufficiently interactive with the base FGF binding site formed by Loops II and III to modify ligand affinity and affect interaction of the receptor with heparan sulfate cofactor. The NH2-terminal domain of Loop II, which is highly conserved across all isoforms, exhibits a 19-residue heparin-binding domain which is obligatory for FGF binding. Heparin protects a 30-kDa ligand-binding fragment from proteolysis that is composed of Loop II, the inter-Loop II/III sequence, and the NH2-terminus of Loop III. This suggests that the high-affinity FGF receptor complex is an intimate ternary complex of transmembrane tyrosine kinase, heparan sulfate glycosaminoglycan, and FGF, each of which have interactive binding domains for the other and may contribute to specificity of the FGFR complex. Although Ig Loop II, the inter-Loop II/III sequence, and the NH2-terminus of Loop III with heparan sulfate form the base FGF binding site, mutually exclusive alternate splicing of two exons coding for the COOH-terminal half of Loop III determines which specific members of the FGF ligand family bind with high affinity to the base site. A kinase- and tyrosine phosphorylation site-defective splice variant, FGFR type 2, acts as a dominant-negative suppressor of phosphorylation of specifically tyr-653 in the catalytic domain of the kinase, with less effect on phosphorylation of tyr-766 in the COOH-terminal tail. We propose that phosphorylation of tyr-766, which is required for interaction of phospholipase C gamma 1 (PLC gamma 1) with the receptor, may occur by a cis-intramolecular mechanism within FGFR monomers, while phosphorylation of tyr-653, which is required for phosphorylation of PLC gamma 1, may occur by a trans-intermolecular mechanism between monomers within kinase homodimers. From the combined results, we propose a model whereby increasing concentrations of FGF may control FGF-mediated signal transduction by heterodimerization of different FGFR monomers. Different monomers arise by regulated combinatorial alternate splicing that alters both the extracellular and intracellular domains.

Genomic structure and cloned cDNAs predict that four variants in the kinase domain of serine/threonine kinase receptors arise by alternative splicing and poly(A) addition

Heterodimers of types I and II serine/threonine kinase receptor monomers compose the active receptor complex for ligands of the transforming growth factor beta family. Here we show that the genomic organization of coding sequences for the intracellular domain of a widely expressed type I serine/threonine kinase receptor is similar to that of the activin type II receptor gene. The genomic structure and cDNA clones indicate that poly(A) addition to alternative exons at each of three carboxyl-terminal coding exon-intron junctions may be a common feature of both type I and II receptor genes. The predicted products are monomers truncated at kinase subdomains VII, IX, and X which vary in kinase activity and potential serine, threonine, and tyrosine phosphorylation sites. These results suggest that combinations of variants that affect the signal-transducing intracellular kinase domain of both type I and II receptor monomers within the transforming growth factor beta ligand family may add to the heterogeneity of biological effects of individual ligands in the family.

[Serum levels and gene expressions of interleukin-1 beta (IL-1 beta), interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-alpha) in human renal cell carcinomas]

We examined serum levels of interleukin-1 beta (IL-1 beta), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-alpha) in 32 patients with renal cell carcinoma. Serum IL-1 beta was elevated in only one patient (3.1%), IL-6 in 17 (53.1%) and TNF-alpha in 6 (18.8%). Positive rate of serum IL-6 increased with progression of the stage (stage I and II: stage III and IV = 14.3%: 88.9%, p < 0.0001) and was higher for rapid growth type (slow and intermediate growth type: rapid growth type = 36.4%: 90.0%, p < 0.005). However, increased serum levels of IL-1 beta and TNF-alpha were not correlated with stage, grade, or growth type. Gene expressions of IL-1 beta, IL-6, IL-6 receptor and TNF-alpha in the tissues resected from 15 patients with renal cell carcinoma and 3 cell lines established from renal cancers were evaluated by Northern blot analysis. Gene expression of IL-1 beta and TNF-alpha were not detected in any of the tissues or cell cultures, but highly enhanced expressions of IL-6 were revealed in two tumor specimens (13.3%, 1 rapid and 1 slow growth type) and 2 of 3 cell lines. Expressions of IL-6 receptor in tissues and cells were weak. These results suggest that IL-6 plays an important role in the progression of renal cell carcinoma, and that measurement of IL-6 in patients with renal cell carcinoma may be useful clinically.

Studies on the biological characterization and mitogenic interactions between hepatic stimulator substance and acidic fibroblast growth factor

During liver regeneration, hepatic stimulator substance (HSS) and acidic fibroblast growth factor (FGF-1) are produced in the liver. These growth factors may be involved in liver growth control but an understanding of their regulatory interactions is limited. To further characterize the mitogenic activity of HSS, we compared its effects with FGF-1 in cells of hepatocyte, non-parenchymal liver epithelial and non-hepatic lineages. Our studies with these cell types demonstrated differences in the mitogenic specificities of HSS and FGF-1. Whereas exposure of primary hepatocytes to epidermal growth factor and HSS synergistically increased DNA synthesis, simultaneous exposure to HSS and FGF-1 resulted in no such effect. Receptor-binding assays showed that HSS did not compete with FGF-1 in binding to FGF-1 receptors on rat primary hepatocytes. Additional immunoblot analysis demonstrated no cross-reactivity between FGF-1 antibodies and HSS. Distinct mitogenic and immunologic properties of HSS and FGF-1 should facilitate further analysis of liver regeneration and hepatic oncogenesis.

Monoclonal antibodies against heparin-binding growth factor-1: neutralization of biological activity and recognition of specific amino acid sequence

A panel of three monoclonal antibodies against heparin-binding growth factor-1 (HBGF-1) was obtained. These antibodies, Ab-47 alpha, Ab-15, Ab-29, were able to recognize HBGF-1 but not HBGF-2. One of the antibodies, Ab-47 alpha, was identified as a HBGF-1 neutralizing antibody on the basis of its ability to inhibit the binding of [125I]HBGF-1 to receptors on HepG-2 cells and the proliferation of fetal bovine heart endothelial cells induced by HBGF-1. Ab-15 reacted with truncated HBGF-1(Mr = 16,000), intact HBGF-1(Mr = 18,000) and mutant HBGF-1U which lacks a putative nuclear translocation sequence (amino acid residues 21 to 27 of HBGF-1). Ab-29 reacted with only truncated HBGF-1 and was thought to recognize the putative nuclear translocation sequence of HBGF-1. The three monoclonal antibodies did not inhibit the binding of [125I]HBGF-1 to heparin. These data indicate that each monoclonal antibody recognizes a distinct epitope of HBGF-1 and identifies them as useful reagents for evaluating functional domains and biological roles of HBGF-1.

Control of fibroblast growth factor receptor kinase signal transduction by heterodimerization of combinatorial splice variants

A differentiated liver cell (HepG2), which exhibits a dose-dependent growth-stimulatory and growth-inhibitory response to heparin-binding fibroblast growth factor type 1 (FGF-1), displays high- and low-affinity receptor phenotypes and expresses specific combinatorial splice variants alpha 1, beta 1, and alpha 2 of the FGF receptor (FGF-R) gene (flg). The extracellular domains of the alpha and beta variants consist of three and two immunoglobulin loops, respectively, while the intracellular variants consist of a tyrosine kinase (type 1) isoform and a kinase-defective (type 2) isoform. The type 2 isoform is also devoid of the two major intracellular tyrosine autophosphorylation sites (Tyr-653 and Tyr-766) in the type 1 kinase. An analysis of ligand affinity, dimerization, autophosphorylation, and interaction with src homology region 2 (SH2) substrates of the recombinant alpha 1, beta 1, and alpha 2 isoforms was carried out to determine whether dimerization of the combinatorial splice variants might explain the dose-dependent opposite mitogenic effects of FGF. Scatchard analysis indicated that the alpha and beta isoforms exhibit low and high affinity for ligand, respectively. The three combinatorial splice variants dimerized in all combinations. FGF enhanced dimerization and kinase activity, as assessed by receptor autophosphorylation. Phosphopeptide analysis revealed that phosphorylation of Tyr-653 was reduced relative to phosphorylation of Tyr-766 in the type 1 kinase component of heterodimers of the type 1 and type 2 isoforms. The SH2 domain substrate, phospholipase C gamma 1 (PLC gamma 1), associated with the phosphorylated type 1-type 2 heterodimers but was phosphorylated only in preparations containing the type 1 kinase homodimer. The results suggest that phosphorylation of Tyr-653 within the kinase catalytic domain, but not Tyr-766 in the COOH-terminal domain, may be stringently dependent on a trans intermolecular mechanism within FGF-R kinase homodimers. Although phosphotyrosine 766 is sufficient for interaction of PLC gamma 1 and other SH2 substrates with the FGF-R kinase, phosphorylation and presumably activation of substrates require the kinase homodimer and phosphorylation of Tyr-653. We propose that complexes of phosphotyrosine 766 kinase monomers and SH2 domain signal transducers may constitute unactivated presignal complexes whose active or inactive fate depends on homodimerization with a kinase or heterodimerization with a kinase-defective monomer, respectively. The results suggest a mechanism for control of signal transduction by different concentrations of ligand through heterodimerization of combinatorial splice variants from the same receptor gene.

A widely expressed transmembrane serine/threonine kinase that does not bind activin, inhibin, transforming growth factor beta, or bone morphogenic factor

Molecular cloning of complementary DNAs (cDNA) whose expression products bind activin and transforming growth factor beta (TGF-beta 1 and -beta 2) suggests that transmembrane serine/threonine kinases constitute a new class of signaling molecules. A human liver cell cDNA which codes for a new serine/threonine kinase receptor (SKR1) was identified using degenerate oligonucleotide primers complementary to coding sequence for mouse activin and Caenorhabditis elegans daf-1 serine/threonine receptor kinase subdomains VI and VIII in the polymerase chain reaction. The deduced 509-amino acid product consisted of a cysteine-rich extracellular domain and a cytoplasmic serine/threonine kinase domain which are 10-20 and 40% homologous to the respective domains in the activin and transforming growth factor beta receptor kinases. Cells overexpressing SKR1 exhibited no increase in binding of activin, inhibin, TGF-beta 1, TGF-beta 2, or bone morphogenic factor type 2B. Except for its absence in bone and spleen, SKR1 exhibits a tissue expression pattern similar to the TGF-beta receptor II gene. Similarly, SKR1 is expressed in normal parenchymal cells, endothelial cells, fibroblasts, and tumor-derived epithelial cells. The expression pattern and lack of binding to prototypic members of the TGF-beta 1-5 branch of the TGF-beta superfamily suggests that SKR1 is potentially a receptor for a new member of the TGF-beta branch of the ligand superfamily.