Thiourea enhances mapping of the proteome from murine white adipose tissue.

Adipose tissue imposes problems in two-dimensional (2-D) analysis due to its extremely high content of fat. To improve protein separation detergents and chaotropes were varied in the IEF step. The most important factor for obtaining distinct spots in the 2-D gel was whether thiourea was included or not. Many high molecular weight spots became resolved by using thiourea, while no spots disappeared or showed inferior characteristics, thus approximately twice as many spots were possible to quantify. Hydrophobic indices were compared for a set of proteins that gave rise to sharper spots with proteins that were not improved on the use of thiourea. The comparison did not give any statistically significant difference between the two groups of proteins. One of the effects obtained by inclusion of thiourea was that the dominating protein, serum albumin, appeared as more condensed spots allowing other minor proteins to be detected. This work resulted in a protocol which greatly enhances the resolution of proteins in adipose tissue. A 2-D map of mouse white adipose tissue from epididymal fat pads was constructed in which 140 spots were identified by mass spectrometry. This work lays the ground for our further studies on white adipose tissue in metabolic diseases such as obesity and dyslipidemia.

Rosiglitazone (BRL49653), a PPARgamma-selective agonist, causes peroxisome proliferator-like liver effects in obese mice.

The PPAR (peroxisome proliferator activated receptor) transcription factors are ligand-activated nuclear receptors that regulate genes involved in lipid metabolism and homeostasis. PPARalpha is preferentially expressed in liver and PPARgamma preferentially in adipose tissue. Activation of PPARalpha leads to peroxisome proliferation and increased beta-oxidation of fatty acids in rodents. PPARgamma-activation leads to adipocyte differentiation and improved insulin signaling of mature adipocytes. Both PPAR receptors are believed to be functional targets for treatment of hyperlipidemia in man. We have treated obese diabetic mice (ob/ob), which have highly elevated levels of plasma triglycerides, glucose and insulin, for 1 week with WY14,643 (180 micromol/kg/day), a selective PPARalpha agonist, or rosiglitazone (BRL49653; 2.5 micromol/kg/day), a selective PPARgamma agonist. The doses used produce a similar therapeutic effect in both treatment groups (lowering of triglycerides and glucose). High resolution two-dimensional gel electrophoresis of livers showed that WY14,643 and rosiglitazone both produced changes in expression pattern of many proteins involved in peroxisomal fatty acid beta-oxidation. However, similar experiments performed in lean mice showed significant up-regulation of these proteins only with WY14,643 treatment. Furthermore, the proteins up-regulated by the drugs in obese mice had a higher basal expression in obese controls compared to the lean littermates. Liver PPARgamma mRNA levels were determined and we observed that PPARgamma2 mRNA levels were elevated in obese mice compared to lean littermates. As PPARalpha and PPARgamma recognize similar DNA response elements, it is likely that the effects of rosiglitazone on PPARalpha responsive genes in livers of the ob/ob mice are mediated by PPARgamma2.

A proteome analysis of livers from obese (ob/ob) mice treated with the peroxisome proliferator WY14,643.

The PPAR (peroxisome proliferator activated receptor) transcription factors are ligand-activated receptors which regulate genes involved in lipid metabolism and homeostasis. PPARalpha is preferentially expressed in the liver and PPARgamma preferentially in adipose tissue. Activation of PPARalpha leads to peroxisome proliferation in rodents and increased beta-oxidation of fatty acids. PPARgamma-activation leads to adipocyte differentiation and improved insulin signaling of mature adipocytes. Both of these PPAR receptors are potential targets for treatment of dyslipidemia in man. Studies by others using a proteomics approach have characterized the effects of PPARalpha agonists in livers from lean healthy mice. However, we wanted to map the effects of a therapeutic dose of a PPARalpha agonist in a disease model of insulin resistance and diabetes, the obese diabetic ob/ob mouse, by proteomics. Therefore, ob/ob mice, which have highly elevated levels of plasma triglycerides, glucose and insulin, were treated for one week with WY14,643 (180 micromol/kg/day), a well-characterized selective PPARalpha agonist. Plasma triglycerides, glucose and insulin levels were determined and we found significant therapeutic effects on triglycerides and glucose levels. The liver protein compositions were investigated by high-resolution two-dimensional gel electrophoresis which showed that WY14,643 produced up-regulation of at least 16 spots. These were identified by mass spectrometry and 14 spots were found to be components of the peroxisomal fatty acid metabolism. Thus, WY14,643 at a therapeutic dose, caused induction of peroxisomal fatty acid beta-oxidation in obese diabetic mice.

Interaction of estramustine with tubulin isotypes.

The interaction of the antimitotic agent estramustine with bovine microtubule proteins and purified tubulin was investigated. Direct photoaffinity labeling of microtubule protein with [14C]estramustine resulted in the labeling of both alpha- and beta-tubulin, and this was inhibited with unlabeled estramustine in a dose-dependent manner. [14C]Estramustine was incorporated into both the soluble and polymerized forms of tubulin. The affinity constant for estramustine binding to tubulin was determined by equilibrium dialysis to be 23 +/- 5 mM. Estramustine did not affect [3H]vinblastine binding, and vinblastine had no effect on direct labeling with [14C]estramustine. Both rhizoxin and paclitaxel decreased the covalent labeling of tubulin with [14C]estramustine in a dose-dependent fashion and were noncompetitive inhibitors of the binding of estramustine to tubulin. The binding of colchicine to tubulin was not inhibited by estramustine as detected by fluorescence and DEAE filter assays. The estramustine binding site on tubulin is therefore distinct from that of colchicine and vinblastine and may at least partially overlap with the binding site for paclitaxel. In both bovine brain microtubules and cytoskeletal proteins from human prostatic carcinoma cells, the incorporation of [14C]estramustine into the beta III isotype of tubulin was found to occur with a reduced efficiency compared to that of the other beta-tubulin isotypes and alpha-tubulin. Since this isotype is overexpressed in estramustine resistant human prostate carcinoma cells, these results indicate that beta III-tubulin may play a role in the response to the effects of estramustine.

Differential expression of uPA in an aggressive (DU 145) and a nonaggressive (1013L) human prostate cancer xenograft.

Prostate cancer has a slow growing noninvasive phase, but, in general, is invasive on diagnosis. An initial step in the invasion of surrounding normal tissue is the activity of proteolytic enzymes such as components of the plasminogen activator system (PA). In cell culture, the primary human prostate cancer cell line 1013L expressed no urokinase type-PA (uPA), while DU 145, a cell line derived from a metastatic lesion, expressed high levels of uPA. The DU 145 cells grew easily as xenografts but the establishment of 1013L in the SCID mice was possible only with the aid of a gelatin sponge (Spongostan). The latency period was 42-64 days, followed by a slow growth phase before a fast growth phase occurred. This fast growth phase was characterized by rapid degeneration of tumor tissue, while high proliferation occurred around the blood vessels. On serial transplantation of tumor material, the growth pattern was similar. Furthermore, the 1013L tumor was encapsulated by connective tissue and no invasiveness could be detected. We found that 1013L tumor homogenates had hardly detectable levels of uPA, i.e., 300-fold lower than we found in the invasive prostate xenograft DU 145. In addition, no expression of uPA was found in the plasma of 1013L tumor-bearing mice whilst uPA antigen was detected in the plasma of DU 145 tumor-bearing mice. In conclusion, the 1013L cell line, which exhibits a nonaggressive pattern, could be a good model for studying progression of prostate cancer to a more aggressive phenotype in vivo and in vitro.

Estramustine depolymerizes microtubules by binding to tubulin.

To investigate the mechanism of action of the antineoplastic drug estramustine, we compared its effects on human prostate cancer cells with those of vinblastine. At their respective concentrations that result in 50% inhibition of clonogenic growth, both drugs caused an accumulation of cells blocked at mitosis and similar dose- and time-dependent depolymerization of interphase microtubules. Also, colcemid-resistant and colcemid-hypersensitive Chinese hamster ovary cells with tubulin mutations were collaterally cross-resistant or -sensitive to estramustine. Thus, the cytotoxicity of estramustine is due to its microtubule depolymerization properties. This could be caused by interaction with tubulin and/or with microtubule-associated proteins (MAPs). Previous investigations have shown that high concentrations of estramustine phosphate can inhibit microtubule polymerization in vitro by binding to MAPs. However, estramustine phosphate is the clinical prodrug to estramustine, the intracellular active compound. In this study, we investigated the effects of estramustine on the binding of MAPs to taxol-stabilized microtubules in vivo. In contrast to previous reports, no effect of estramustine on the binding of MAPs to microtubules was found. Furthermore, we found that polymerization of purified tubulin could be inhibited by estramustine in vitro. Taken together, these results demonstrate that estramustine causes depolymerization of microtubules by direct interaction with tubulin.

Characterization of a 58 kDa cis-Golgi protein in pancreatic exocrine cells.

We have studied the biochemical characteristics and localization of a 58 kDa cis-Golgi marker protein (p58) in rat pancreatic exocrine cells. The protein remained associated with membranes after extraction at alkaline pH and was largely resistant to proteases, added to intact microsomes. By electrophoresis p58 could be resolved into two bands which in two-dimensional gels separated into several charge variants around pI 5.5. This size and charge heterogeneity of p58 did not appear to be due to acylation, glycosylation or phosphorylation. In non-reduced gels p58 migrated as two kinetically related, high relative molecular mass forms, apparently corresponding to disulfide-linked homo-dimers and -hexamers. Immuno-electron microscopy localized p58 to both the fenestrated cis-Golgi cisternae and small Golgi vesicles or buds as well as large, pleiomorphic structures, scattered throughout the cells and associated with distinct smooth ER (endoplasmic reticulum) clusters. These findings correlated with cell fractionation results showing the concentration of p58 in two microsomal subfractions, banding at intermediate densities between the rough ER and trans-Golgi in sucrose gradients. Our results indicate that p58 is a major component of pre- and cis-Golgi elements and could be part of the transport machinery that operates in these membranes. Together with results obtained with other cell types, these observations suggest that the peripheral smooth ER clusters are involved in the early stages of the secretory pathway in the pancreatic acinar cells.

The endoplasmic reticulum retention signal of the E3/19K protein of adenovirus-2 is microtubule binding.

The signal for retention in the endoplasmic reticulum of the E3/19K protein of adenovirus type 2 is located within the carboxyl-terminal cytoplasmic extension. A synthetic peptide corresponding to this sequence showed affinity for beta-tubulin, could promote tubulin polymerization in vitro, and bound to taxol-polymerized microtubules. When compared with the microtubule binding sequences from two microtubule-associated proteins (MAPs; MAP2 and tau), we found similarities suggesting that the cytoplasmic tail might bind to tubulin/microtubules in a MAPs-like fashion. A synthetic peptide corresponding to the cytoplasmic tail of an E3/19K deletion mutant not retained in the endoplasmic reticulum was also tested. It had the same net charge but did not promote tubulin polymerization in vitro nor did it show measurable affinity for tubulin or microtubules. This indicates that binding to microtubules is important for retention of the E3/19K protein in the endoplasmic reticulum.

Down-regulation of MHC class I antigens is not a general mechanism for the increased tumorigenicity caused by c-myc amplification.

Previous studies have shown that increased expression of oncogenes from the myc-family can down-regulate the level of MHC class I antigens in tumor cells. This has suggested a mechanism by which amplification/overexpression of myc-genes may contribute to the malignancy development of certain tumors. Earlier published data from the murine SEWA tumor, a polyomavirus-induced osteosarcoma, have correlated the degree of tumorigenicity of different sublines to their level of c-myc amplification. Here I present a quantitative and qualitative analysis of MHC class I antigens from five sublines of this tumor system. No differences could be found, between sublines with different degrees of tumorigenicity, regarding MHC class I antigen expression. Thus, in the SEWA tumor, the enhancement of the tumorigenicity caused by c-myc amplification is not mediated through down-regulation of MHC class I antigens.

Human arterial smooth muscle cells in culture. Effects of platelet-derived growth factor and heparin on growth in vitro.

Human arterial smooth muscle cells (hASMC) were cultured from explants of the inner media of uterine arteries obtained at hysterectomy. The presence of alpha-actin and smooth muscle-specific actin isoforms and the microscopic appearance of the cells in secondary culture established their smooth muscle origin. The hASMC were diploid and had no signs of transformation. Plasma-derived serum failed to stimulate their proliferation in vitro. Their rate of proliferation was, however, proportional to the concentration of whole blood serum in the medium. Anti-PDGF IgG at high concentrations inhibited the stimulatory effect of whole blood serum on cell proliferation. This suggests that hASMC depend on exogenous PDGF for their growth. In PDS or bovine serum albumin cell numbers remained constant for 7 days in culture and the thymidine index was below 1% per 24 h. When reexposed to whole blood serum these cells started to proliferate within 2 days. This indicates that hASMC when deprived of PDGF enter a quiescent state that is fully reversible upon rexposure to the mitogen. Heparin is a powerful growth inhibitor for SMC. In our system, heparin caused a dose-dependent inhibition of cell proliferation despite optimal concentrations of whole blood serum. This inhibition was reversible upon withdrawal of heparin. At heparin concentrations which caused a half-maximal inhibition it was also competed for by increasing concentrations of whole blood serum. Quiescent hASMC expressed the PDGF receptor on their surface as judged from immunofluorescence with a monoclonal antibody. This was true irrespective of whether growth arrest was achieved by serum depletion or by the addition of heparin to serum-containing medium. Cells growing in the presence of whole blood serum did not, however, express the receptor antigen. These observations suggest that heparin may interfere with PDGF or with its binding and further processing at the level of the cell-surface receptor.

Characterization of multidrug resistance in SEWA mouse tumor cells: increased glutathione transferase activity and reversal of resistance with verapamil.

The glutathione transferase (EC 2.5.1.18) activity of multidrug-resistant and drug-sensitive SEWA mouse tumor cells was determined. Increase in activity was found in all multidrug-resistant sublines. Electrophoretic separation of cytosolic proteins and immunodetection, using antisera against Class Alpha, Class Mu and Class Pi glutathione transferase, revealed a two-fold increase of the Class Pi transferase in the resistant cells. The role of glutathione transferase of Class Pi in the multidrug resistance and its distinction from the over-produced 21,000 dalton protein p21 is discussed. Reversal of the resistance, using the calcium antagonist verapamil, and different sensitivity to verapamil between different sublines are also reported.

Changes in Protein Composition of Three Bacterial Isolates from Marine Waters during Short Periods of Energy and Nutrient Deprivation.

Two-dimensional gel electrophoresis analysis of and total cell protein determination for three bacterial isolates from marine waters at the onset and after 24 h of energy and nutrient deprivation demonstrated that the three isolates exhibited different pathways of starvation survival. Two strains appeared to synthesize new proteins during starvation.

Pleiotropic drug resistance, protein overexpression, and cytogenetic signs of gene amplification in mouse tumor cells.

We have isolated pleiotropic drug resistant mouse tumor cell lines with resistance-associated double minute chromosomes (DM) or homogeneously staining regions (HSR). In resistant cells, two-dimensional protein electrophoresis revealed elevated levels of a 21,000 dalton protein. Two resistant lines, with cytogenetically unstable (DM) and stable (HSR) gene amplification, were grown in the absence of drug for several weeks. The resistance and the protein overexpression was correlated to the presence of either HSR or DM. This suggests that the 21,000 dalton protein is a gene amplification product, and its importance for the resistance is discussed.

Pleiotropic drug resistance and gene amplification in a SEWA mouse tumor cell line. Complex relations revealed by drug uptake data, and lipid and protein analysis.

SEWA mouse lines resistant to actinomycin D (AMD) or vincristine (VCR) exhibit the pleiotropic drug resistance (PDR) phenotype, and express a low-MW protein (p21) and numerous double minutes (DM). In drug uptake studies these lines were compared with the non-resistant parental line and with a methotrexate (MTX)-resistant line, not exhibiting PDR. On treatment with labelled AMD or VCR the two PDR lines displayed a highly reduced intracellular content of drug, whereas uptake of MTX was unchanged. Uptake of AMD was shown to be temperature-dependent. The MTX-resistant line did not exhibit any significant change in AMD or VCR uptake. Other workers have emphasized the role of a high-MW glycoprotein in the development of PDR. A search for a similar glycoprotein in our cells was unsuccessful. Since all indications point to membrane factors being important in the development of PDR, the lines were also subjected to lipid analysis. Compared with control cells distinct differences were detected in the lipid composition of all resistant lines (including the MTX-resistant line). In the course of our experiments, the DM in our most AMD-resistant line were replaced by two homogeneously staining regions (HSR). Simultaneously, the overproduction of p21 ceased, but the PDR phenotype persisted. This event tends to implicate a minor role for the p21 protein in PDR, but similar transitions from DM to HSR in other AMD-resistant SEWA lines were not accompanied by a decrease in p21 over-production. Our data point to a complex genetic control of multi-drug resistance.

Resistance to actinomycin D and to vincristine induced in a SEWA mouse tumor cell line with concomitant appearance of double minutes and a low molecular weight protein.

By increasing stepwise the drug concentration of the medium, we have induced, in a cell line of the murine SEWA tumor, resistance to actinomycin D (AMD) and vincristine (VCR) 30-50 times above the normal. In both types of resistant cells, we have revealed, by two-dimensional gel electrophoresis, a protein (MW 21K , pI 5) not found in control cells. Large fractions of the resistant cells contained double minutes (DM). In AMD-resistant cells, correlation was demonstrated between number of DM and degree of resistance. Back in AMD-free medium, resistant cells lost both the DM and the 21K protein. Cross-resistance prevailed between AMD and VCR. Cells resistant to AMD and VCR showed erratic resistance to methotrexate (MTX), but no significant resistance existed in the reverse direction.