Expression of constitutive and inducible HSP70 and HSP47 is enhanced in cells persistently spread on OPN1 or collagen

Cellular Environment and Phenotypic Heterogeneity: How Data-Driven Modeling Finds the Smoking Gun

The cellular environment modifies cellular phenotypes, in particular, the stress response phenotype, which easily exhibits high phenotypic heterogeneity due to the common characteristics of its regulatory networks. The aim of this work is to quantify and interpret the impact of collagen type I, a major component of the cellular environment, on the phenotypic heterogeneity of the cellular response. Our approach combines in an original way the monitoring of the response of a single cell and the mathematical modeling of the network. After a detailed statistical description of the phenotypic heterogeneity of the cellular response, the mathematical modeling explains how the observed changes can be explained by an induced increase in the average expression of a central protein of the regulatory network. The predictions of the data-driven model are fully consistent with the biochemical measurements performed. The framework presented here is also a new general methodology to study phenotypic heterogeneity, although we focus here on the response to proteotoxic stress in HeLa cells.

Expression of heat shock protein 47 in the fibrous tissue adjacent to mouse tumour subjected to photodynamic therapy

The reaction of normal fibrous tissue adjacent to tumours subjected to photodynamic therapy (PDT) was investigated by assessment of the immunohistochemical expression of heat shock protein 47 (HSP47) and proliferating cell nuclear antigen (PCNA), as well as by immunoblot analysis of procollagen type I. PDT was administered to NR-S1 mouse squamous cell carcinoma or normal mouse skin. Each of four mice was investigated at several time points after receiving PDT. The levels of HSP47 expression were determined by computer-assisted image analysis. The expression of procollagen type I in the fibrous tissue adjacent to the tumours was examined by immunoblot analysis at intervals of 24 and 48h after PDT. The expression of HSP47 was first detected 6h post-PDT in the tumour-bearing mice, but no such expression was observed in the normal mice. It was also revealed that, after PDT, the fibroblast PCNA labeling indices at 24, 48, and 72h were significantly higher in both the tumour-bearing and the normal mice than in the control animals that did not receive PDT. Furthermore, procollagen type I was detected in the fibrous tissue adjacent to the tumours at 24 and 48h after PDT, but was not detected in the fibrous tissue adjacent to tumours of mice that did not receive PDT. Therefore, the present results suggest that PDT enhances the synthesis of collagen type I in the fibrous tissue adjacent to NR-S1 squamous cell carcinoma in mice, which contributes to the resultant encapsulation of such tumours.

Matrix remodeling after stroke. De novo expression of matrix proteins and integrin receptors

Following an ischemic insult to the central nervous system a reorganization of cells and tissue takes place as the surrounding cells attempt to limit the injury, repair the damage, and restore normal architecture of the brain. This tissue remodeling requires de novo synthesis of genes and proteins which enables cells to actively change their relationship with the existing extracellular matrix and with other cells to reorganize the damaged tissue. We have identified two key molecular components of the matrix remodeling process after focal ischemia: osteopontin (OPN) and its integrin receptor alpha v beta 3 (alpha v beta 3). OPN is initially expressed by activated macrophages and microglia in the periinfarct region (24-48 hr) and at later times (5-15 days) in the core infarct. After focal stroke the alpha v beta 3 was upregulated by astrocytes in the periinfarct region. Spatial and temporal analyses demonstrated that at 5 days after injury the alpha v beta 3-positive astrocytes were at a distance from the osteopontin-expressing macrophages; by 15 days the alpha v beta 3-expressing astrocytes were localized within an osteopontin-rich matrix. In vitro OPN was shown to induce migration of astrocytes in a Boyden chamber system. These data suggest that OPN derived from microglia at the infarct border zone (and possible macrophages in the infarct core) may serve as an "astrokine" (suggested term for astrocyte chemoattractant) to organize the astrocyte scar after focal stroke. Our data demonstrate profound changes in brain matrix remodeling after focal ischemic stroke, including the synthesis and release of matrix proteins alien to the normal brain, the expression of integrin receptors that ligate these proteins, and possibly a novel function for microglial-derived OPN in astrocyte migration after focal ischemia that may drive glial activation, organization, and repair functions.

Differential processing of osteopontin characterizes the proliferative vascular smooth muscle cell phenotype induced by allylamine

Repeated cycles of vascular injury by allylamine induce vascular lesions similar to those seen in atherosclerotic vessels, or following balloon catheterization. Vascular (aortic) smooth muscle cells harvested from allylamine-treated animals (i.e., allylamine cells) acquire a proliferative advantage relative to control counterparts that is associated with differential secretion and extracellular matrix sequestration of several proteins. In the present study, we have characterized two of these proteins (M(r) 52 and 36 kDa; pl 5.6 and 5.2, respectively) and their putative role in the expression of a proliferative phenotype. Because the physical properties of these proteins were comparable to those of osteopontin (OPN) and its thrombin-generated fragment(s), initial experiments were conducted to examine the expression and processing of OPN in this cell system. OPN mRNA expression was enhanced during early G1 cell cycle progression in allylamine cells relative to control counterparts. However, comparable amounts of OPN (M(r) 56, 52, and 50 kDa) were detected by Western analysis in media conditioned by both cell types using the OP-199 or B77-Rat1 antibodies to OPN. Allylamine cells, however, produced increased amounts of a 36 kDa protein recognized by the OP-199 antibody. Incubation of conditioned media from [35S]methionine-labeled allylamine cells with thrombin decreased the intensity of the 52 kDa protein, while increasing the intensity of a 36 kDa protein. RT-PCR analysis demonstrated expression of a 1.2 kb OPN band in both cell types consistent with the predicted size of OPN mRNA, suggesting that the 36 kDa fragment recognized by OP-199 in allylamine cells was likely not due to altered splicing of the OPN transcript. To determine if OPN and/or the 36 kDa fragment played a central role in the proliferative capacity of allylamine cells, the effect of an antibody to an alpha v integin subunit was examined. An antibody to the alpha v subunit, but not alpha 4, nullified the proliferative advantage of allylamine cells relative to control counterparts, suggesting that integrin-mediated signaling is a key feature of the proliferative phenotype of allylamine cells. We conclude that enhanced proteolytic cleavage of OPN may characterize the modulation of vascular SMCs to a more proliferative phenotype following chemical injury by allylamine.

Expression of the collagen-related heat shock protein HSP47 in fibroblasts treated with hyperthermia or photodynamic therapy

Heat shock protein (HSP) 47 is associated with collagen type I metabolism, both constitutively and after stress-inflicted injury. It has been claimed that, in contrast to hyperthermia (HT), photodynamic therapy (PDT) does not damage collagen, as measured at the level of tissue. We have studied HSP47 expression in normal murine skin fibroblasts (3T6) treated with hyperthermia or photodynamic therapy (PDT) mediated by three different photosensitizers: (1) haematoporphyrin ester (HpE), (2) meta tetra hydroxyphenyl chlorin (mTHPC) and (3) riboflavin (RB). Riboflavin is not an established photosensitizer for PDT and was chosen here because it is known to provoke collagen damage. The applied doses of the treatments were isoeffective in terms of 3T6 clonogenic cell survival. Analysis, at both transcriptional and translational levels, revealed HSP47 elevation after hyperthermia and after PDT with RB. PDT sensitized by HpE and mTHPC did not significantly alter HSP47 expression. These observations are consistent with our hypothesis that this collagen chaperone is up-regulated by laser-mediated modalities known to damage collagen (i.e. HT and RB PDT) but not by more conventional PDT treatments. Additionally, unexpected significant up-regulation of HSP47 was detected after illumination alone (no photosensitizer) of 3T6 cells at 653 nm laser light, but not at 630 nm.

Differential expression of heat shock 70 proteins in primary cultures from rat cerebellum

While a number of studies have described the heat shock response in established cell lines and in primary cultures of cells derived from the nervous system, there has been no systematic analysis comparing expression and localization of the inducible heat shock 70 (hsp70) proteins and the constitutively synthesized members of the family (hsc70) in neurons and glia. In the present communication, we utilized specific probes to compare the expression of hsp70 and hsc70 mRNAs and proteins in two types of primary cultures, astroglial and neuro-astroglial, from postnatal rat cerebellum. Conditions were adjusted to maintain physiological numbers of microglia in both types of culture, and cultures were analyzed at a number of different time points following a precisely defined heat shock. The northern, in situ hybridization and immunohistochemical analyses resulted in a number of novel observations concerning the nature of the heat shock response in these neuronal and glial cells. In postnatal day 4-5 cultures, hsp70 mRNA levels were elevated for at least 10 h in both types of culture, but in situ hybridization analysis showed no evidence for hsp70 mRNAs in neurons. Microglia were the only cell type in which hsp70 was detected in non-stressed cultures and this cell type contained the highest concentrations of hsp70 proteins in stressed cultures. Hsc70 mRNA levels were also increased after heat shock, but the increase was more transient. Hsc70 mRNAs and proteins were present in all cell types, again with the highest concentrations being present in microglia. Hsc70 mRNAs and proteins were localized in the cytoplasm at all time points examined, with hsc70 protein also being localized in nucleoli. Hsp70 mRNAs and proteins were diffusely localized over nuclei of astrocytes, as well as of most microglia. Hsp70, but not hsc70, was localized on chromosomes in glia once they had resumed cell division after heat shock, suggesting a role for hsp70 either in targeting damaged chromosomal proteins or in cell division. Some cytoplasmic hsp70 was observed in astrocytes of the mixed neuro-astroglial cultures and a delayed hsp70 immunoreactivity was observed in granule neurons in these cultures, suggesting either that translation of low levels of hsp70 mRNAs was more efficient in neurons, or that glial-neuronal translocation of hsp70 proteins had taken place. These results suggest that metabolism and functions of different heat shock protein family members may not always be identical and that care must be taken in extrapolation of results from one cell type to another.

Hsp47 and cyclophilin B traverse the endoplasmic reticulum with procollagen into pre-Golgi intermediate vesicles. A role for Hsp47 and cyclophilin B in the export of procollagen from the endoplasmic reticulum

Hsp47 and cyclophilin B (CyPB) are residents of the endoplasmic reticulum (ER). Both of these proteins are closely associated with polysome-associated alpha 1(I) procollagen chains. Hsp47 possesses chaperone properties early during the translation of procollagen while the cis/trans-isomerase properties of CyPB facilitate procollagen folding. In this report, we further investigate the interaction of these proteins with procollagen I during export from the ER. To inhibit vesicular budding and retain procollagen within the ER, cells were treated with the heterotrimeric G protein inhibitor mastoparan or calphostin C, a specific inhibitor of diacylglycerol/phorbol ester binding proteins. To arrest procollagen in pre-Golgi intermediate vesicles, cells were treated with guanosine 5'-3-O-(thio)triphosphate. Pulse-chase experiments of cells labeled with [35S]methionine followed by immunoprecipitation during the chase period with anti-procollagen, anti-Hsp47, and anti-CyPB antibodies were performed to reveal the relationship between Hsp47/CyPB/procollagen I. The distribution of procollagen, Hsp47, and CyPB to the ER and/or pre-Golgi vesicles was verified by immunofluorescence. Hsp47 and CyPB remained associated with procollagen retained within the ER. Hsp47 and CyPB were also associated with procollagen exported from the ER into pre-Golgi intermediate vesicles. Treatment of cells with cyclosporin A diminished the levels of CyPB bound to procollagen and diminished the rate of Hsp47 released from procollagen and the rate of procollagen secretion, suggesting that Hsp47 release from procollagen may be driven by helix formation. Also, these studies suggest that Hsp47 may resemble protein disulfide isomerase and possess both chaperone and anti-chaperone properties. During translation, high levels of Hsp47 are seen to limit protein aggregation and facilitate chain registration. Later, Hsp47 and/or CyPB and protein disulfide isomerase act as anti-chaperones and provide the basis for concentration of procollagen for ER export.

Regulation and function of osteopontin in ras-transformed cells

Transformation of nontumorigenic NIH 3T3 fibroblasts with an activated ras oncogene produces malignant cells that can metastasize. This induction of malignant behavior is due to changes in gene expression induced by Ras-mediated signal transduction. Osteopontin expression is induced in response to Ras, due to increased osteopontin transcription mediated by Ras-responsive regions in the osteopontin gene promoter. The increased expression of osteopontin contributes functionally to the malignant ability of the cells. Ras-transformed cells that express antisense osteopontin RNA show markedly reduced ability to form tumors and to metastasize in experimental animals. Increased osteopontin expression in human tumors thus may also contribute to increased malignancy. Site-directed mutagenesis of recombinant osteopontin protein indicates that an intact RGD sequence is required for cell adhesion and induction of chemotaxis, consistent with the idea that integrin-mediated signal transduction is a consequence of osteopontin binding to cells. Osteopontin may contribute to malignancy by inducing responses in host and/or tumor cells.

Site-directed mutagenesis of the arginine-glycine-aspartic acid sequence in osteopontin destroys cell adhesion and migration functions

Osteopontin (OPN) is a secreted calcium-binding phosphoprotein produced in a variety of normal and pathological contexts, including tissue mineralization and cancer. OPN contains a conserved RGD (arg-gly-asp) amino acid sequence that has been implicated in binding of OPN to cell surface integrins. To determine whether the RGD sequence in OPN is required for adhesive and chemotactic functions, we have introduced two site-directed mutations in the RGD site of the mouse OPN cDNA, in which the RGD sequence was either deleted or mutated to RGE (arg-gly-glu). In order to test the effect of these mutations on OPN function, we expressed control and mutated mouse OPN in E. coli as recombinant glutathione-S-transferase (GST)-OPN fusion proteins. Control mouse GST-OPN was functional in cell adhesion assays, supporting attachment and spreading of mouse (malignant PAP2 ras-transformed NIH 3T3, and, to a lesser extent, normal NIH 3T3 fibroblasts) and human (MDA-MB-435 breast cancer, and normal gingival fibroblast) cells. In contrast, neither of the RGD-mutated OPN proteins ("delRGD" or "RGE") supported adhesion of any of the cell lines, even when used at high concentrations or for long assay times. GRGDS (gly-arg-gly-asp-ser) peptides inhibited cell adhesion to intact GST-OPN, as well as to fibronectin and vitronectin. In chemotaxis assays, GST-OPN promoted directed cell migration of both malignant (PAP2, MDA-MB-435) and normal (gingival fibroblast, and NIH 3T3) cells, while RGD-mutated OPN proteins did not. Together these results suggest that the conserved RGD sequence in OPN is required for the majority of the protein's cell attachment and migration-stimulating functions.

Overcoming obstacles to metastasis--defenses against host defenses: osteopontin (OPN) as a shield against attack by cytotoxic host cells

Osteopontin (OPN) serves both a cell attachment function and a cell signalling function via the alpha v beta 3 integrin. In its cell attachment capacity it can promote attachment of both osteoclasts to bone hydroxyapatite and various other cell types to basement membrane/extracellular matrix. In its cell signalling capacity it initiates a signal transduction cascade that includes changes in the intracellular calcium ion levels and the tyrosine phosphorylation status of several proteins including paxillin. Effects on gene expression include suppression of the induction of nitric oxide synthase by inflammatory mediators. OPN can also reduce cell oxidant levels and inhibit the killing of tumor cells by activated macrophages and endothelial cells. We hypothesize that those cancer cells that produce OPN at elevated levels can suppress the oxidative burst, inhibit NO production, and thus protect themselves from killing by specific host cell types.

Dynamic variations in the expression of type I collagen and its molecular chaperone Hsp47 in cells of the mouse dental follicle during tooth eruption

Tooth eruption is a precisely timed and sequenced event that brings the tooth from within bone into a functional position in the mouth. Every part of the developing tooth has been theoretically implicated as a primary factor in this process, but it now appears that eruption is multifactorial, with the dental follicle and type I collagen playing an important part. Immunological probes were used here to investigate in vivo and in vitro the temporal and spatial expression of type I collagen and its molecular chaperone Hsp47 in the dental follicle during eruption. Mandibles were dissected from 2-, 5-, 9- and 11-day-old neonatal mice and fixed in 95% ethanol overnight. Sections of 7 microns were obtained and reacted with antibodies directed against type I collagen. Dental follicles were isolated from 2-, 5-, 9- and 11-day-old neonates and cells were grown in culture for 8 days. Slides were then reacted with antibodies directed against type I collagen and Hsp47. The production of type I collagen and Hsp47 in the follicle varied with the stage of dental development and eruption. There was a progressive decrease of type I collagen in the coronal part of the follicle, leading to an arrest of its production in these areas. These findings support the notion that cells of the coronal portion of the dental follicle stop producing type I collagen as a prerequisite to the initiation of tooth eruption and that this phenotype persists in vitro.

Recombinant GST-human osteopontin fusion protein is functional in RGD-dependent cell adhesion

Osteopontin (OPN) is a secreted phosphoprotein expressed by many tumor cells, as well as a limited set of normal cells. Native OPN has been shown to support cell adhesion in an RGD-peptide-inhibitable fashion. Here we expressed human OPN in E. coli as a recombinant fusion protein with glutathione-S-transferase (GST). We report that the GST-OPN fusion protein has functional activity. PAP2 (ras-transformed, metastatic murine NIH 3T3) and MDA-MB-435 human mammary carcinoma cells bound to GST-OPN in an in vitro cell adhesion assay nearly as well as to native bovine OPN. Adhesion to the recombinant fusion protein was blocked by addition of GRGDS peptide, suggesting that the cells adhere to the recombinant and native OPN proteins by similar, integrin-mediated mechanisms. Adhesion to both sources of OPN also was inhibited by thrombin treatment of the protein. Thrombin cleaves GST from OPN in the fusion protein, and also cleaves internally in OPN, adjacent to the RGD sequence of the protein. Our results suggest that (a) thrombin cleavage of native OPN may be a natural regulator of OPN function, and (b) the majority of OPN cell binding activity is mediated by the RGD sequence in the protein backbone, with little or no requirement for post-translational modifications that occur in native OPN for adhesive function as measured here.

Hsp 47 is localized to regions of type I collagen production in developing murine femurs and molars

To determine whether the proposed molecular chaperone Hsp47 is associated with the production of heterotrimeric procollagen, the distribution of anti-Hsp47 and anti-collagen antibodies were examined in developing murine femurs and molars of 22-23-day CD-1 mice. In addition, the expression of Hsp47, and collagen mRNAs were assessed by in situ hybridization using oligonucleotide probes. These studies revealed that Hsp47 was developmentally expressed and produced in regions that are coincident with type I collagen. Hsp47 was not localized in cartilaginous zones of developing femurs or in the regions of developing molars producing type III collagen. These results support the hypothesis that Hsp47 is necessary for the assurance of type I procollagen and is not expressed with other homotrimeric procollagen molecules.

Lead inhibits secretion of osteonectin/SPARC without significantly altering collagen or Hsp47 production in osteoblast-like ROS 17/2.8 cells

In an effort to better understand the consequences of lead (Pb2+) on skeletal growth, the effects of Pb2+ were investigated using ROS 17/2.8 bone-like cells in vitro. These studies revealed that Pb2+ (4.5 x 10(-6) M -4.5 x 10(-7) M) has little or no effect on cell shape except when added immediately following seeding of the cells. However, proliferation of ROS cells was inhibited, in the absence of serum, at concentrations of 4.5 x 10(-6) M Pb2+. Protein production was generally increased, however, the major structural protein of bone, type I collagen, production was only slightly altered. Following treatment of ROS cells with Pb2+, intracellular levels of the calcium-binding protein osteonectin/SPARC were increased. Osteonectin/SPARC secretion into the media was delayed or inhibited. Coincident with retention of osteonectin/SPARC there was a decrease in the levels of osteonectin/SPARC mRNA as determined by Northern analysis. These studies suggest that processes associated with osteonectin/SPARC translation and secretion are sensitive to Pb2+.