A novel anti-inflammatory drug, SDZ ASM 981, for the treatment of skin diseases: in vitro pharmacology.

SDZ ASM 981, a novel ascomycin macrolactam derivative, has high anti-inflammatory activity in animal models of allergic contact dermatitis and shows clinical efficacy in atopic dermatitis, allergic contact dermatitis and psoriasis, after topical application. Here we report on the in vitro activities of this promising new drug. SDZ ASM 981 inhibits the proliferation of human T cells after antigen-specific or non-specific stimulation. It downregulates the production of Th1 [interleukin (IL)-2, interferon-gamma] and Th2 (IL-4, IL-10) type cytokines after antigen-specific stimulation of a human T-helper cell clone isolated from the skin of an atopic dermatitis patient. SDZ ASM 981 inhibits the phorbol myristate acetate/phytohaemagglutinin-stimulated transcription of a reporter gene coupled to the human IL-2 promoter in the human T-cell line Jurkat and the IgE/antigen-mediated transcription of a reporter gene coupled to the human tumour necrosis factor (TNF)-alpha promoter in the murine mast-cell line CPII. It does not, however, affect the human TNF-alpha promoter controlled transcription of a reporter gene in a murine dendritic cell line (DC18 RGA) after stimulation via the FcgammaRIII receptor. SDZ ASM 981 also prevents the release of preformed pro-inflammatory mediators from mast cells, as shown in the murine cell line CPII after stimulation with IgE/antigen. In summary, these results demonstrate that SDZ ASM 981 is a specific inhibitor of the production of pro-inflammatory cytokines from T cells and mast cells in vitro.

SDZ 281-977: a modified partial structure of lavendustin A that exerts potent and selective antiproliferative activities in vitro and in vivo.

The chemical derivatization of biologically active microbial metabolites continues to be a promising approach to the identification of new drugs. We recently synthesized the novel antiproliferative compound SDZ 281-977, 5-[2-(2,5-dimethoxy-phenyl)ethyl]-2-hydroxy-benzoic acid methylester, a derivative of the EGF receptor tyrosine kinase inhibitor lavendustin A. Here we report on our studies of the anticancer efficacy and the mode of action of SDZ 281-977. The growth of both the human pancreatic tumor cells MIA PaCa-2 and the human vulvar carcinoma cells A431 was inhibited in the low micromolar range. Tumors from these cells were induced in nude mice and were shown to respond to orally or intravenously administered SDZ 281-977. In contrast, no antitumor effect was detected in rats bearing dimethylbenzanthracene-induced mammary tumors. Studies in mice indicated that SDZ 281-977 was neither immunosuppressive nor hematosuppressive at doses effectively inhibiting tumor growth. Surprisingly, the mode of action of SDZ 281-977 apparently does not involve inhibition of EGF receptor tryosine kinase, because, in contrast to lavendustin A, SDZ 281-977 failed to inhibit this enzyme in a cell-free assay. The mechanism of the antiproliferative effect can be explained on a cellular level by the ability of the compound to arrest cells in mitosis. SDZ 281-977 is thus the first example of an antimitotic agent derived from the potent tyrosine kinase inhibitor lavendustin A. The therapeutic potential of SDZ 281-977 is enhanced by the fact that it is not subject to multidrug resistance, because tumor cells expressing the multidrug resistance phenotype were as sensitive to SDZ 281-977 as their nonresistant counterparts. In conclusion, SDZ 281-977 represents a novel lavendustin A derivative with potent antiproliferative properties in vitro and in vivo that may be explained on the basis of its antimitotic effects. SDZ 281-977 may be a candidate drug for the treatment of selected cancers, including those expressing the multidrug resistance phenotype.

Novel antiproliferative agents derived from lavendustin A.

The active partial structure of the potent tyrosine kinase inhibitor lavendustin A was derivatized in the search for novel agents against cellular proliferation. The antiproliferative potential of the new derivatives was determined using the human keratinocyte cell line HaCaT as the primary test system. Whereas the lavendustin A partial structure is ineffective in inhibiting cell proliferation, esterification of its carboxylic acid function leads to measurable antiproliferative activity. Additional O-methylation of the 2,5-dihydroxyphenyl moiety yields activity in the micromolar range. Further substantial increases in activity are achieved by replacing the nitrogen with oxygen and carbon within the 2,5-dimethoxyphenyl series (but not within the 2,5-dihydroxyphenyl analogs) leading to 5-[2-(2,5-dimethoxyphenyl) ethyl]-2-hydroxybenzoic acid methyl ester (13) as the most potent analog identified to date. These increases in antiproliferative activity are paralleled, however, by the disappearance of activity against the epidermal growth factor receptor-associated tyrosine kinase, suggesting another mechanism of action.

ICAM-1 expression in a spontaneously transformed human keratinocyte cell line: characterization by a simple cell-ELISA assay.

Interferon gamma (IFN-gamma) is known to induce ICAM-1 on keratinocytes (KC) in vitro, and its expression in vivo is correlated with epidermal T-cell infiltration in various dermatoses. However, the mechanisms for this cytokine-mediated ICAM-1 expression are essentially unknown. We investigated the induction of ICAM-1 by IFN-gamma in HaCaT cells, a spontaneously transformed human KC cell line, using an immunoperoxidase-ELISA with the monoclonal antibody (MoAb) R6.5. HaCaT cells constitutively expressed low levels of ICAM-1, which were upregulated by IFN-gamma. The kinetics and dose response were similar to those published for primary KC, regardless of whether the HaCaT cells were cultured in low- or high-calcium medium. ICAM-1 expression was increased significantly at 4 h with 500 U/ml IFN-gamma, and reached a plateau (approximately 5 x greater than constitutive) by 24 h. At concentrations greater than 10 U/ml for 24 h, IFN-gamma induced ICAM-1 expression in a dose-dependent fashion (half maximal at 100 U/ml). TNF-alpha alone, and in synergistic combination with IFN-gamma, also upregulated the expression of HaCaT ICAM-1. IFN-gamma treatment of HaCaT cells increased the level of ICAM-1 mRNA and enhanced (approximately 3x) the adherence of fluorescently labeled (calcein) human T lymphoblasts, as determined by Northern blotting and an in vitro adhesion assay, respectively. Our findings suggest that HaCaT cells, in conjunction with a simple immunoperoxidase cell-ELISA, provide a reliable system for studying pharmacologic modulation of ICAM-1 on KC.

Determination of the transbilayer distribution of fluorescent lipid analogues by nonradiative fluorescence resonance energy transfer.

We measured the nonradiative fluorescence resonance energy transfer between 7-nitro-2,1,3-benzoxadiazol-4-yl (NBD) labeled lipids (amine labeled phosphatidylethanolamine or acyl chain labeled phosphatidylcholine) and rhodamine labeled lipids in large unilamellar dioleoylphosphatidylcholine vesicles. Two new rhodamine labeled lipid analogues, one a derivative of monolauroylphosphatidylethanolamine and the other of sphingosylphosphorylcholine, were found to exchange through the aqueous phase between vesicle populations but not to be capable of rapid transbilayer movement between leaflets. Energy transfer from NBD to rhodamine was measured using liposomes with symmetric or asymmetric distributions of these new rhodamine labeled lipid analogues to determine the relative contributions of energy transfer between donor and acceptor fluorophores in the same (cis) and opposite (trans) leaflets. Since the characteristic R0 values for energy transfer ranged from 47 to 73 A in all cases, significant contributions from both cis and trans energy transfer were observed. Therefore, neither of these probes acts strictly as a half-bilayer quencher of NBD lipid fluorescence. The dependence of transfer efficiency on acceptor density was fitted to a theoretical treatment of energy transfer to determine the distances of closest approach for cis and trans transfer. These parameters set limits on the positions of the fluorescent groups relative to the bilayer center, 20-31 A for NBD and 31-55 A for rhodamine, and provide a basis for future use of these analogues in measurements of transbilayer distribution and transport.

Fluorescent probes of electrostatic potential 1 nm from the membrane surface.

We measured the electrostatic potential 1 nm from the surface of charged phospholipid bilayer membranes to test the predictions of the Gouy-Chapman theory. Fluorescent probes (anthraniloyl, 5-(dimethylamino)naphthalene-1-sulfonyl, Lucifer yellow) were attached covalently to the sialic acid residue of the ganglioside galactosyl-N-acetylgalactosaminyl(N-acetylneuraminyl)galactosylglucosylc eramide (GM1). These fluorescent gangliosides were incorporated into neutral [phosphatidylcholine (PC)] or charged [phosphatidylserine (PS)] phospholipid bilayers, and the fluorescence was quenched with the cations thallium and 4-amino-2,2,6,6-tetramethylpiperidine-1-oxyl (tempamine). We calculated the electrostatic potential at the chromophore from the quenching ratio using the Boltzmann relation: the average potential was -30 mV for PS bilayers in 0.1 M NaNO3. We assume the chromophore is 1 nm from the surface because X-ray diffraction measurements demonstrate that the sialic acid residue of GM1 is 1 nm from the surface of a PC/GM1 bilayer [McDaniel, R. V., & McIntosh, T. J. (1986) Biophys. J. 49, 94-96]. We also used thallium and tempamine to quench the fluorescence of chromophores located at the surface of the PS membranes; in 0.1 M NaNO3 the average surface potential was -80 mV, which agrees with other measurements. The Gouy-Chapman theory predicts that the potential 1 nm from a membrane with a surface potential of -80 mV is -24 mV; this prediction agrees qualitatively with the experimental results obtained with fluorescent gangliosides.

An experimental test of the discreteness-of-charge effect in positive and negative lipid bilayers.

The electrostatic properties of charged bilayers and the bilayer component of biological membranes are often described theoretically by assuming the charge is smeared uniformly over the surface. This is one of the fundamental assumptions in the Gouy-Chapman-Stern (GCS) theory. However, the average distance between the charged phospholipids in a typical biological membrane is 2-3 nm, which is 2-3 times the Debye length in a 0.1 M salt solution. Existing discreteness-of-charge theories predict significant deviations from the GCS theory for the adsorption of ions to such membranes. We considered the predictions of the simplest discreteness-of-charge theory [Nelson, A. P., & McQuarrie, D. A. (1975) J. Theor. Biol. 55, 13-27], in which the charges are assumed to be fixed in a square lattice and the potential is described by the linearized Poisson-Boltzmann relation. This theory predicts deviations that are larger for counterions than for co-ions and much larger for divalent than for monovalent counterions. We tested these predictions by measuring the adsorption of a fluorescent monovalent anion and a paramagnetic divalent cation to both positive and negative membranes, which we demonstrated experimentally had the same average surface potential. All our experimental results with probes, including those obtained on membranes in the gel rather than in the liquid-crystalline state, agreed with the predictions of the GCS theory rather than with the discreteness-of-charge theory. A simple calculation indicates that the agreement between the experimental results and the predictions of the GCS theory could be due to the finite size of the lipids.

An experimental test of new theoretical models for the electrokinetic properties of biological membranes. The effect of UO2++ and tetracaine on the electrophoretic mobility of bilayer membranes and human erythrocytes.

For a large smooth particle with charges at the surface, the electrophoretic mobility is proportional to the zeta potential, which is related to the charge density by the Gouy-Chapman theory of the diffuse double layer. This classical model adequately describes the dependence of the electrophoretic mobility of phospholipid vesicles on charge density and salt concentration, but it is not applicable to most biological cells, for which new theoretical models have been developed. We tested these new models experimentally by measuring the effect of UO2++ on the electrophoretic mobility of model membranes and human erythrocytes in 0.15 M NaCl at pH 5. We used UO2++ for these studies because it should adsorb specifically to the bilayer surface of the erythrocyte and should not change the density of fixed charges in the glycocalyx. Our experiments demonstrate that it forms high-affinity complexes with the phosphate groups of several phospholipids in a bilayer but does not bind significantly to sialic acid residues. As observed previously, UO2++ adsorbs strongly to egg phosphatidylcholine (PC) vesicles: 0.1 mM UO2++ changes the zeta potential of PC vesicles from 0 to +40 mV. It also has a large effect on the electrophoretic mobility of vesicles formed from mixtures of PC and the negative phospholipid phosphatidylserine (PS): 0.1 mM UO2++ changes the zeta potential of PC/PS vesicles (10 mol % PS) from -13 to +37 mV. In contrast, UO2++ has only a small effect on the electrophoretic mobility of either vesicles formed from mixtures of PC and the negative ganglioside GM1 or erythrocytes: 0.1 mM UO2++ changes the apparent zeta potential of PC/GM1 vesicles (17 mol % GM1) from -11 to +5 mV and the apparent zeta potential of erythrocytes from -12 to -4 mV. The new theoretical models suggest why UO2++ has a small effect on PC/GM1 vesicles and erythrocytes. First, large groups (e.g., sugar moieties) protruding from the surface of the PC/GM1 vesicles and erythrocytes exert hydrodynamic drag. Second, charges at the surface of a particle (e.g., adsorbed UO2++) exert a smaller effect on the mobility than charges located some distance from the surface (e.g., sialic acid residues).

Electrokinetic and electrostatic properties of bilayers containing gangliosides GM1, GD1a, or GT1. Comparison with a nonlinear theory.

We formed vesicles from mixtures of egg phosphatidylcholine (PC) and the gangliosides GM1, GD1a, or GT1 to model the electrokinetic properties of biological membranes. The electrophoretic mobilities of the vesicles are similar in NaCl, CsCl, and TMACl solutions, suggesting that monovalent cations do not bind significantly to these gangliosides. If we assume the sialic acid groups on the gangliosides are located some distance from the surface of the vesicle and the sugar moieties exert hydrodynamic drag, we can describe the mobility data in 1, 10, and 100 mM monovalent salt solutions with a combination of the Navier-Stokes and nonlinear Poisson-Boltzmann equations. The values we assume for the thickness of the ganglioside head group and the location of the charge affect the theoretical predictions markedly, but the Stokes radius of each sugar and the location of the hydrodynamic shear plane do not. We obtain a reasonable fit to the mobility data by assuming that all ganglioside head groups project 2.5 nm from the bilayer and all fixed charges are in a plane 1 nm from the bilayer surface. We tested the latter assumption by estimating the surface potentials of PC/ganglioside bilayers using four techniques: we made 31P nuclear magnetic resonance, fluorescence, electron spin resonance, and conductance measurements. The results are qualitatively consistent with our assumption.

Bilayer membranes containing the ganglioside GM1: models for electrostatic potentials adjacent to biological membranes.

Although the Gouy-Chapman-Stern theory of the aqueous diffuse double layer describes well the electrostatic potential adjacent to negatively charged phospholipid bilayer membranes, it does not describe adequately the zeta potential of biological membranes: the zeta potential of an erythrocyte is about half the value predicted from the theory by using the known density of negatively charged sialic acid residues. To investigate the factors responsible for this low electrophoretic mobility, we formed membranes from mixtures of the zwitterionic lipid phosphatidylcholine, PC, and the glycolipid galactosyl-N-acetylgalactosaminyl(N-acetylneuraminyl) -galactosylglucosylceramide, GM1. This glycolipid differs from phospholipids in two respects. First, the negative charge on GM1 is located about 1 nm from the surface, which tends to increase the electrophoretic mobility of vesicles. Second, the head group of GM1 contains five sugar groups that exert a hydrodynamic drag, which tends to decrease the mobility of the vesicles. In a decimolar monovalent salt solution, where the Debye length is about 1 nm, the electrophoretic mobility of the PC-GM1 vesicles is about half the mobility of PC-phosphatidylserine or PC-phosphatidylglycerol vesicles of equivalent composition. In addition, conductance measurements with planar bilayer membranes as well as 31P nuclear magnetic resonance and fluorescence measurements with sonicated vesicles indicate that the potential at the surface of PC-GM1 membranes is about half the value measured for PC-phosphatidylserine membranes in a 0.1 M monovalent salt solution.