Inhibition of TNF-alpha-induced sickle RBC retention in retina by a VLA-4 antagonist.

PURPOSE: Patients with sickle cell disease have elevated circulating levels of cytokines including tumor necrosis factor (TNF) alpha. TNF-alpha stimulates expression by endothelial cells of adhesion molecules, including vascular cell adhesion molecule (VCAM) 1. Others have demonstrated that VLA-4 (alpha(4)beta(1)), a ligand for VCAM-1 or fibronectin, is present on a fraction of sickle reticulocytes. The intent of this study was to determine, using a rat model, if TNF-alpha increases retention of sickle erythrocytes in retina and if that retention can be inhibited. METHODS: TNF-alpha was given intraperitoneally to rats 5 hours before IV administration of FITC-labeled, density-separated sickle erythrocytes. After 5 minutes, rats were exsanguinated, and retinas were excised and incubated for ADPase activity, permitting the determination of the number and location of retained cells. RESULTS: TNF-alpha caused a three- to fourfold increase in retention of sickle erythrocytes in retinal capillaries (P < 0.05) but not of normal human erythrocytes. Preincubation of sickle erythrocytes with TBC772, a peptide that blocks the binding of alpha(4)beta(1) and alpha(4)beta(7), or a monoclonal antibody against VLA-4 (19H8), significantly inhibited the TNF-alpha-induced retention (P < or = 0.02), whereas a control cyclic peptide and antibody had no effect. IV TBC772 also inhibited sickle erythrocyte retention (P = 0.01). Two intravenously administered anti-fibronectin antibodies inhibited sickle cell retention as well, but an anti-rat VCAM-1 antibody did not inhibit retention. CONCLUSIONS: The authors conclude that TNF-alpha stimulates retention of sickle erythrocytes in the retinal vasculature. This increased retention can be blocked by a VLA-4 antagonist, suggesting that the cells retained after cytokine stimulation are reticulocytes. The counter-receptor for VLA-4 in this rat retina model appears to be fibronectin and not VCAM-1, based on data obtained using antibodies against these molecules.

A novel protein with homology to the junctional adhesion molecule. Characterization of leukocyte interactions.

We have cloned a novel cDNA belonging to the Ig superfamily that shows 44% similarity to the junctional adhesion molecule (JAM) and maps to chromosome 21q21.2. The open reading frame of JAM2 predicts a 34-kDa type I integral membrane protein that features two Ig-like folds and three N-linked glycosylation sites in the extracellular domain. A single protein kinase C phosphorylation consensus site and a PDZ-binding motif are present in the short intracellular tail. Heterologous expression of JAM2 in Chinese hamster ovary cells defined a 48-kDa protein that localizes predominantly to the intercellular borders. Northern blot analysis showed that JAM2 is preferentially expressed in the heart. JAM2 homotypic interactions were demonstrated by the ability of JAM2-Fc to capture JAM2-expressing Chinese hamster ovary cells. We further showed that JAM2, but not JAM1, is capable of adhering to the HSB and HPB-ALL lymphocyte cell lines. Neutralizing mouse anti-JAM2 polyclonal antibodies provided evidence against homotypic interactions in this assay. Biotinylation of HSB cell membranes revealed a 43-kDa counter-receptor that precipitates specifically with JAM2-Fc. These characteristics of JAM2 led us to hypothesize a role for this novel protein in adhesion events associated with cardiac inflammatory conditions.

Novel synthetic inhibitors of selectin-mediated cell adhesion: synthesis of 1,6-bis[3-(3-carboxymethylphenyl)-4-(2-alpha-D- mannopyranosyloxy)phenyl]hexane (TBC1269).

Reports of a high-affinity ligand for E-selectin, sialyl di-Lewis(x) (sLe(x)Le(x), 1), motivated us to incorporate modifications to previously reported biphenyl-based inhibitors that would provide additional interactions with the protein. These compounds were assayed for the ability to inhibit the binding of sialyl Lewis(x) (sLe(x), 2) bearing HL-60 cells to E-, P-, and L-selectin fusion proteins. We report that dimeric or trimeric compounds containing multiple components of simple nonoligosaccharide selectin antagonists inhibit sLe(x)-dependent binding with significantly enhanced potency over the monomeric compound. The enhanced potency is consistent with additional binding interactions within a single selectin lectin domain; however, multivalent interaction with multiple lectin domains as a possible alternative cannot be ruled out. Compound 15e (TBC1269) showed optimal in vitro activity from this class of antagonists and is currently under development for use in the treatment of asthma.

Novel inhibitors of alpha 4 beta 1 integrin receptor interactions through library synthesis and screening.

A library of 2302 small molecule beta-turn mimetics was screened for inhibition of the alpha 4 beta 1 integrin-CS1 splice variant binding interaction. Preliminary data revealed several active ligands, and validation with purified material culminated in the identification of some of the first small molecule ligands (1, IC50 = 5 microM, and 2, IC50 = 8 microM) to be reported for this class of integrins.

Angiogenesis induced by tumor necrosis factor-agr; is mediated by alpha4 integrins.

Tumor necrosis factor-alpha (TNF-alpha) and fibroblast growth factor-2 (FGF-2 or bFGF) are potent stimulators of angiogenesis. TNF- alpha, but not FGF-2, can induce the expression of vascular cell adhesion molecule-1 (VCAM-1) on the surface of endothelial cells. The soluble form of VCAM-1 has recently been demonstrated to function as an angiogenic mediator. Here we demonstrate that monoclonal antibodies directed against VCAM-1 or its alpha4 integrin counter- receptor inhibited TNF-alpha-induced endothelial cell migration in vitro. Angiogenesis induced in vivo in rat corneas by TNF-alpha was inhibited by a neutralizing antibody directed against the rat alpha4 integrin subunit. A peptide antagonist of the a4 integrins blocked TNF-alpha-induced endothelial cell migration in vitro and angiogenesis in rat corneas in vivo. No inhibition by the antibodies or peptide antagonist was observed either in vitro or in vivo when FGF-2 was used as the stimulus. The peptide antagonist did not inhibit TNF-a binding to its receptor nor did it block the function of alphavbeta3, an integrin previously implicated in TNF-a and FGF- 2 mediated angiogenesis. These results demonstrate that angiogenic processes induced by TNF-alpha are mediated in part by agr;4 integrins possibly by a mechanism involving the induction of soluble VCAM-1.

Regulation of human T lymphocyte coactivation with an alpha4 integrin antagonist peptide.

The cyclic hexapeptide CWLDVC (TBC 772) is an antagonist of alpha4 integrins and a potent inhibitor of lymphocyte interactions with fibronectin, vascular cell adhesion molecule-1, and muscosal vascular addressin cell adhesion molecule-1 (MAdCAM-1). As such, peptide TBC 772 effectively inhibits the activation of freshly isolated human T lymphocytes stimulated with purified vascular cell adhesion molecule-1 coimmobilized with anti-CD3 mAb. The influence of peptide binding on distinct sites of the alpha4beta1 complex was determined by flow cytometry and cellular adhesion assays employing a panel of mAbs. Binding of the alpha4-specific mAb L25 and the beta1-specific mAb 33B6 was not altered by the peptide; however, binding of mAb 19H8, which is specific for a combinatorial epitope of alpha4beta1, was dramatically inhibited. Treatment of lymphocytes with the peptide caused an increase in a ligand-induced epitope on beta1 integrin defined by mAb 15/7. In T cell activation studies using coimmobilized anti-CD3 mAb and the anti-integrin mAbs, the peptide had broader inhibitory activity, suppressing costimulation induced by all the integrin mAbs. The peptide was not generally toxic and was integrin selective in its suppressive activity, as coactivation by ligation of CD3 in conjunction with CD28 or CD26 was not affected. These results suggest that the antagonist peptide CWLDVC can effectively neutralize integrin coactivation systems by a mechanism independent of competitive binding.

A cyclic hexapeptide is a potent antagonist of alpha 4 integrins.

The alpha 4 integrins mediate leukocyte adhesion to specific counter-receptors, including vascular cell adhesion molecule-1 (VCAM-1), the fibronectin splice variant containing connecting segment 1 (CS1), and mucosal addressin cell adhesion molecule-1. A series of cyclized peptides based on the LDV sequence of CS1 were synthesized and assayed for inhibition of alpha 4 integrin binding. The most potent peptide, C*WLDVC* (where * indicates disulfide-linked residues), inhibited alpha 4 beta 1-dependent binding of lymphocytes to VCAM-1 and CS1 with half-maximal inhibition achieved at 1 to 3 microM of peptide. The peptide proved more potent when the lymphocytes were activated with 1 mM MnCl2; half-maximal inhibition was reached at 0.4 and 0.05 microM for VCAM-1 and CS1, respectively. This represents a 100- to 800-fold increase in potency over a linear CS1 peptide in these same assays. C*WLDVC* also inhibited alpha 4 beta 7-dependent lymphocyte binding to the ligands mucosal addressin cell adhesion molecule-1, VCAM-1 and CS1. Immunoprecipitation of radiolabeled integrin indicated that the peptide could bind alpha 4 beta 1 and alpha 4 beta 7 directly and elute alpha 4 beta 1 from a CS1-conjugated agarose resin. The peptide showed selectivity for alpha 4 integrins in that it effectively inhibited alpha 4 beta 1-dependent, but not alpha 5 beta 1-dependent, binding of cells to intact fibronectin. Due to its small size and potency, C*WLDVC* may serve as a useful tool for the study of alpha 4 integrin biology and the development of small molecule therapeutics.

Human skin mast cell carboxypeptidase: functional characterization, cDNA cloning, and genealogy.

We functionally characterized human skin mast cell carboxypeptidase A (MC-CPA), and explored its evolutionary relationship to other carboxypeptidases to understand further the structural basis for the substrate preferences of this enzyme. Purified human skin MC-CPA displayed more activity than did bovine pancreatic carboxypeptidase A (CPA) against carboxyl-terminal leucine residues, about equal activity with phenylalanine and tyrosine residues, and no activity with tryptophan or alanine. To correlate kinetic data with structure, we isolated and sequenced a cDNA encoding MC-CPA from human skin, and directly sequenced 30% of the purified protein. These sequences agreed with that of human lung MC-CPA, and further support the evidence for a single MC-CPA gene in humans. Four amino acid replacements, resulting in a net positive change in non-hydrogen atoms in the S1' subsite of MC-CPA, were associated with less alteration in substrate specificity, relative to bovine CPA, than might be expected from studies using rat CPA1 and CPA2. We noted two consensus N-linked glycosylation sites in human MC-CPA that are not found in rat and mouse MC-CPA, or in bovine CPA; that at least one of these sites is glycosylated in vivo was verified by N-glycosidase F treatment, lentil lectin binding, and Concanavalin A-Sepharose chromatography. Evolutionary trees constructed from the known carboxypeptidase sequences suggested that MC-CPA most likely evolved from a carboxypeptidase B-like enzyme, independent of the pancreatic CPA. Thus, in the carboxypeptidase gene family, MC-CPA displays a unique genealogy and several amino acid replacements in its S1' binding pocket that result in substrate specificity quite similar to bovine CPA.

Structure, chromosomal assignment, and deduced amino acid sequence of a human gene for mast cell chymase.

A gene encoding human chymase was cloned and sequenced. The protein-coding exons reveal a preproenzyme with a 19-amino acid signal peptide, an acidic 2-amino acid propeptide, and a 226-amino acid catalytic domain. The mature enzyme is predicted to be cationic (net charge of +13) and to be modified by N-glycosylation at two sites. The amino acid sequence is identical to the 35 residues of NH2-terminal amino acid sequence reported for human skin chymase and is identical to 29 of 31 residues of NH2-terminal and internal amino acid sequence reported for human heart chymase. The full predicted sequence of the catalytic domain reveals a high level of sequence identity to dog mast cell chymase (83%) and a lower level of identity to the sequences of rodent chymases (58-62%). In the phase and placement of introns, the organization of this human chymase gene is similar to that of several other granule-associated leukocyte serine proteases, including rat chymase II, lymphocyte granzymes, and neutrophil cathespin G and elastase. However, the gene organization differs from that of mast cell tryptase, providing additional evidence that the major mast cell serine proteases are separated by substantial evolutionary distance. Amplification of chymase gene-specific fragments from hamster/human hybrid cell line DNA suggests localization of the chymase gene to human chromosome 14. High stringency hybridization of chymase DNA to a human genomic DNA blot suggests the possibility of more than one human chymase gene. Evidence that the chymase gene is expressed in human tissues was obtained by the amplification of chymase-specific DNA from skin and placental cDNA libraries.

Dog mast cell chymase: molecular cloning and characterization.

We cloned and characterized a cDNA coding for the complete amino acid sequence of dog mast cell chymase. The cDNA was identified by screening a dog mastocytoma cDNA library with an oligonucleotide probe based on the amino acid sequence of a fragment of dog mastocytoma chymase. The deduced amino acid sequence reveals a putative 21-residue prepropeptide followed by a catalytic domain of 228 residues. The primary structure of the preproenzyme shares features with rat mucosal mast cell chymase (RMCP II), several lymphocyte-associated proteases, and neutrophil cathepsin G. The common characteristics include an apparent activation peptide terminating in glutamic acid, strict conservation of an octapeptide (residues 9-16) in the N-terminal portion of the catalytic domain, and the presence of only six cysteines available for intramolecular disulfide bond formation. However, dog chymase differs in being modified by N-glycosylation. Although the dog chymase catalytic domain exhibits a similar level of sequence identity when compared with both RMCP II and the rat connective tissue mast cell chymase RMCP I (58% and 61%, respectively), the dog enzyme most closely resembles RMCP I in its high predicted net charge (+16) and in the presence of serine at the base of its putative primary substrate binding pocket. The dog chymase differs strikingly from dog mast cell tryptase in the preprosequence and in the structure of the catalytic domain. Therefore, chymase appears not to be closely related to tryptase and may not share a mechanism of activation, even though both enzymes are packaged and released together.

Human mast cell tryptase: multiple cDNAs and genes reveal a multigene serine protease family.

Three different cDNAs and a gene encoding human skin mast cell tryptase have been cloned and sequenced in their entirety. The deduced amino acid sequences reveal a 30-amino acid prepropeptide followed by a 245-amino acid catalytic domain. The C-terminal undecapeptide of the human preprosequence is identical in dog tryptase and appears to be part of a prosequence unique among serine proteases. The differences among the three human tryptase catalytic domains include the loss of a consensus N-glycosylation site in one cDNA, which may explain some of the heterogeneity in size and susceptibility to deglycosylation seen in tryptase preparations. All three tryptase cDNAs are distinct from a recently reported cDNA obtained from a human lung mast cell library. A skin tryptase cDNA was used to isolate a human tryptase gene, the exons of which match one of the skin-derived cDNAs. The organization of the approximately 1.8-kilobase-pair tryptase gene is unique and is not closely related to that of any other mast cell or leukocyte serine protease. The 5' regulatory regions of the gene share features with those of other serine proteases, including mast cell chymase, but are unusual in being separated from the protein-coding sequence by an intron. High-stringency hybridization of a human genomic DNA blot with a fragment of the tryptase gene confirms the presence of multiple tryptase genes. These findings provide genetic evidence that human mast cell tryptases are the products of a multigene family.

Molecular cloning of dog mast cell tryptase and a related protease: structural evidence of a unique mode of serine protease activation.

Mast cell tryptase is a secretory granule associated serine protease with trypsin-like specificity released extracellularly during mast cell degranulation. To determine the full primary structure of the catalytic domain and precursor forms of tryptase and to gain insight into its mode of activation, we cloned cDNAs coding for the complete amino acid sequence of dog mast cell tryptase and a second, possibly related, serine protease. Using RNA from dog mastocytoma cells, we constructed a cDNA library in lambda gt 10. Screening of the library with an oligonucleotide probe based on the N-terminal sequence of tryptase purified from the same cell source allowed us to isolate and sequence overlapping clones coding for dog mast cell tryptase. The tryptase sequence includes the essential residues of the catalytic triad and an aspartic acid at the base of the putative substrate binding pocket that confers P1 Arg and Lys specificity on tryptic serine proteases. The apparent N-terminal signal/activation peptide terminates in a glycine. A glycine in this position has not been observed previously in serine proteases and suggests a novel mode of activation. Additional screening of the library with a trypsinogen cDNA led to the isolation and sequencing of a full-length clone apparently coding for the complete sequence of a second tryptic serine protease (DMP) which is only 53.4% identical with the dog tryptase sequence but which contains an apparent signal/activation peptide also terminating in a glycine. Thus, the proteases encoded by these cloned cDNAs may share a common mode of activation from N-terminally extended precursors.

Site-directed alteration of serine 82 causes nonproductive chain cleavage in prohistidine decarboxylase.

Prohistidine decarboxylase from Lactobacillus 30a normally autoactivates by cleavage of the Ser-81-Ser-82 peptide bond, converting Ser-82 to a pyruvoyl moiety which serves as the enzymatic cofactor. We have used site-directed methods to make two conservative mutations, converting Ser-82 to cysteine (S82C) and threonine (S82T). Both mutant proteins autoactivate, although dramatically (20- to 80-fold), more slowly than wild type. Roughly 55% of the mutant protein in each case undergoes a nonproductive chain cleavage which does not result in cofactor production. This finding suggests that an important feature in the enzyme's evolution has been development of an activation scheme which minimizes nonproductive side reactions. Catalytic constants are also affected by the mutations, particularly kcat which drops 8-fold in S82C and 450-fold in S82T. In addition, the S82T protein activates to produce a novel alpha-ketobutyroyl cofactor.

Expression and characterization of Lactobacillus 30a histidine decarboxylase in Escherichia coli.

The genes coding for histidine decarboxylase from a wild-type strain and an autoactivation mutant strain of Lactobacillus 30a have been cloned and expressed in Escherichia coli. The mutant protein, G58D, has a single Asp for Gly substitution at position 58. The cloned genes were placed under control of the beta-galactosidase promoter and the products are natural length, not fusion proteins. The enzyme kinetics of the proteins isolated from E. coli are comparable to those isolated from Lactobacillus 30a. At pH 4.8 the Km of wild-type enzyme is 0.4 mM and the kcat = 2800 min-1; the corresponding values for G58D are 0.5 mM and 2750 min-1. The wild-type and G58D have autoactivation half-times of 21 and 9 h respectively under pseudophysiological conditions of 150 mM K+ and pH 7.0. At pH 7.6 and 0.8 M K+ the half-times are 4.9 and 2.9 h. The relatively slow rate of autoactivation for purified protein and the differences in cellular and non-cellular activation rates, coupled with the fact that wild-type protein is readily activated in wild-type Lactobacillus 30a but poorly activated in E. coli, suggest that wild-type Lactobacillus 30a contains a factor, possibly an enzyme, that enhances the activation rate.

Cloning and nucleotide sequence of wild type and a mutant histidine decarboxylase from Lactobacillus 30a.

Prohistidine decarboxylase from Lactobacillus 30a is a protein that autoactivates to histidine decarboxylase by cleaving its peptide chain between serines 81 and 82 and converting Ser-82 to a pyruvoyl moiety. The pyruvoyl group serves as the prosthetic group for the decarboxylation reaction. We have cloned and determined the nucleotide sequence of the gene for this enzyme from a wild type strain and from a mutant with altered autoactivation properties. The nucleotide sequence modifies the previously determined amino acid sequence of the protein. A tripeptide missed in the chemical sequence is inserted, and three other amino acids show conservative changes. The activation mutant shows a single change of Gly-58 to an Asp. Sequence analysis up- and downstream from the gene suggests that histidine decarboxylase is part of a polycistronic message, and that the transcriptional promotor region is strongly homologous to those of other Gram-positive organisms.