Automatic classification of human epithelial type 2 cell indirect immunofluorescence images using cell pyramid matching

This paper describes a novel system for automatic classification of images obtained from Anti-Nuclear Antibody (ANA) pathology tests on Human Epithelial type 2 (HEp-2) cells using the Indirect Immunofluorescence (IIF) protocol. The IIF protocol on HEp-2 cells has been the hallmark method to identify the presence of ANAs, due to its high sensitivity and the large range of antigens that can be detected. However, it suffers from numerous shortcomings, such as being subjective as well as time and labour intensive. Computer Aided Diagnostic (CAD) systems have been developed to address these problems, which automatically classify a HEp-2 cell image into one of its known patterns (eg. speckled, homogeneous). Most of the existing CAD systems use handpicked features to represent a HEp-2 cell image, which may only work in limited scenarios. We propose a novel automatic cell image classification method termed Cell Pyramid Matching (CPM), which is comprised of regional histograms of visual words coupled with the Multiple Kernel Learning framework. We present a study of several variations of generating histograms and show the efficacy of the system on two publicly available datasets: the ICPR HEp-2 cell classification contest dataset and the SNPHEp-2 dataset.

Collagen induced MMP-2 activation in human breast cancer

Matrix metalloproteinase-2 (MMP-2), a zymogen requiring proteolytic activation for catalytic activity, has been implicated broadly in the invasion and metastasis of many cancer model systems, including human breast cancer (HBC). MMP-2 has been immunolocalized to carcinomatous human breast, where the degree of activation of MMP-2 correlates well with tumor grade and patient prognosis. Using Matrigel assays, we have stratified HBC cell lines for invasiveness in vitro, and compared this to their potential for metastatic spread in nude mice. HBC cell lines expressing the mesenchymal marker protein vimentin were found to be highly invasive in vitro, and tended to form metastases in nude mice. We have further discovered that culture on collagen-I gels (Vitrogen(TM): Vg) induces MMP-2-activator in highly invasive but not poorly invasive HBC cell lines. As seen for other MMP-2-activator inducing regimens, this induction requires protein synthesis and an intact MMP-2 hemopexin-like domain, appears to be mediated by a cell surface activity, and can be inhibited by metalloproteinase inhibitors. The induction is highly specific to collagen I, and is not seen with thin coatings of collagen I, collagen IV, laminin, or fibronectin, or with 3-dimensional gels of laminin, Matrigel, or gelatin. This review focuses on collagen I and MMP- 2, their localization and source in HBC, and their relationship(s) to MMP-2 activation and HBC metastasis. The relevance of collagen I in activation of MMP-2 in vivo is discussed in terms of stromal cell: tumor cell interaction for collagen I deposition, MMP-2 production and MMP-2-activation. Such cooperativity may exist in vivo for MMP-2 participation in HBC dissemination. A more complete understanding of the regulation of MMP-2-activator by type I collagen may provide new avenues for improved diagnosis and prognosis of human breast cancer.

Pro-matrix metalloproteinase-2 transfection increases orthotopic primary growth and experimental metastasis of MDA-MB-231 human breast cancer cells in nude mice

The ability to activate pro-matrix metalloproteinase (pro-MMP)-2 via membrane type-MMP is a hallmark of human breast cancer cell lines that show increased invasiveness, suggesting that MMP-2 contributes to human breast cancer progression. To investigate this, we have stably transfected pro-MMP-2 into the human breast cancer cell line MDA-MB-231, which lacks MMP-2 expression but does express its cell surface activator, membrane type 1-MMP. Multiple clones were derived and shown to produce pro-MMP-2 and to activate it in response to concanavalin A. In vitro analysis showed that the pro-MMP-2-transfected clones exhibited an increased invasive potential in Boyden chamber and Matrigel outgrowth assays, compared with the parental cells or those transfected with vector only. When inoculated into the mammary fat pad of nude mice, each of the MMP-2-tranfected clones grew faster than each of the vector controls tested. After intracardiac inoculation into nude mice, pro-MMP-2-transfected clones showed a significant increase in the incidence of metastasis to brain, liver, bone, and kidney compared with the vector control clones but not lung. Increased tumor burden was seen in the primary site and in lung metastases, and a trend toward increased burden was seen in bone, however, no change was seen in brain, liver, or kidney. This data supports a role for MMP-2 in breast cancer progression, both in the growth of primary tumors and in their spread to distant organs. MMP-2 may be a useful target for breast cancer therapy when refinement of MMP inhibitors provides for MMP-specific agents.

Tyrosine phosphorylation mediates ConA-induced membrane type 1-matrix metalloproteinase expression and matrix metalloproteinase-2 activation in MDA-MB-231 human breast carcinoma cells

ConA-induced cell surface activation of pro-matrix metalloproteinase-2 (pro-MMP-2) by MDA-MB-231 human breast cancer cells is apparently mediated by up-regulation of membrane type 1 MMP (MT1-MMP) through transcriptional and posttranscriptional mechanisms. Here, we have explored the respective roles of cell surface clustering and protein tyrosine phosphorylation in the ConA- induction effects. Treatment with succinyl-ConA, a variant lacking significant clusterability, partially stimulated MT1-MMP mRNA and protein levels but did not induce MMP-2 activation, suggesting that clustering contributes to the transcriptional regulation by ConA but appears to be critical for the nontranscriptional component. We further found that genistein, an inhibitor of tyrosine phosphorylation, blocked ConA-induced pro-MMP-2 activation and ConA-induced MT1-MMP mRNA level in a dose-dependent manner, implicating tyrosine phosphorylation in the transcriptional aspect. This was confirmed by the dose-dependent promotion of pro-MMP-2 activation by sodium orthovanadate in the presence of suboptimal concentrations of ConA (7.5 μg/ml), with optimal effects seen at 25 μg/g orthovanadate. Genistein did not inhibit the ConA potentiation of MMP-2 activation in MCF-7 cells, in which transfected MT1-MMP is driven by a heterologous promoter, supporting the major implication of phosphotyrosine in the transcriptional component of ConA regulation. These data describe a major signaling event upstream of MT1- MMP induction by ConA and set the stage for further analysis of the nontranscriptional component.

Complex regulation of membrane-type matrix metalloproteinase expression and matrix metalloproteinase-2 activation by concanavalin A in MDA-MB-231 human breast cancer cells

Matrix Metalloproteinase-2 (MMP-2) is secreted as a zymogen, the activation of which has been associated with metastatic progression in human breast cancer (HBC). Concanavalin A (Con A) has been found to induce activation of MMP-2 in invasive HBC cell lines. Con A effects on the expression of mRNA for membrane-type matrix metalloproteinase (MT-MMP), a newly described cell surface-associated MMP, showed a close temporal correlation with induction of MMP-2 activation. It is surprising that MT-MMP mRNA is constitutively present in the uninduced MDA-MB-231 cell, despite a lack of MMP-2 activation. We have used actinomycin D to demonstrate a partial requirement for de novo gene expression in the induction of MMP-2 activation by Con A in MDA-MB-231 HBC cells. Furthermore, this transcriptional response to Con A appeared to require the continued presence of Con A for its manifestation. The nontranscriptional component of the Con A induction manifests rapidly, is quite substantial, and persists strongly despite actinomycin D abrogation of both constitutive and Con A-induced MT-MMP. Cycloheximide analyses suggest that protein synthesis may be involved in this rapid transcription-independent response. These studies suggest that Con A induces MMP-2-activation in part by up-regulation of MT-MMP expression but has a more complicated mode of action, involving additional nontranscriptional effects, which apparently require protein synthesis.

MT1-MMP correlates with MMP-2 activation potential seen after epithelial to mesenchymal transition in human breast carcinoma cells

We have previously reported that human breast carcinoma (HBC) cell lines expressing the mesenchymal intermediate filament protein vimentin (VIM+) are highly invasive in vitro, and highly metastatic in nude mice when compared to their VIM- counterparts. Since only VIM+ cell lines can be induced to activate matrix metalloproteinase-2 (MMP-2) upon stimulation with Concanavalin A (Con A), we have examined here membrane type 1 MMP (MT1-MMP), a cell surface activator of MMP-2. Northern analysis reveals baseline expression of MT1-MMP in five of the six VIM+ cell lines studied (MDA-MB-231, MDA-MB-435, BT-549, Hs578T, MCF-7(ADR)), each of which showed variable activation of exogenous MMP-2 after treatment with Con A. In contrast, the four VIM-, poorly invasive HBC cell lines studied (MCF-7, T47D, MDA-MB 468, ZR-75-1) lacked baseline MT1-MMP mRNA expression, and showed no induction of either MT1-MMP expression or MMP-2-activation with Con A. Such differential MT1-MMP expression was confirmed in vivo using in situ hybridization analysis of nude mouse tumor xenografts of representative cell lines. Western analysis of the MDA-MB-231 cells revealed baseline membrane expression of a 60 kDa species, which was strongly induced by Con A treatment along with a weaker band co-migrating with that from MT1-MMP-transfected COS-1 cells (63 kDa), presumably representing latent MT1-MMP. MT1-MMP immunofluorescence strongly decorated Con A-stimulated MDA-MB-231 cells in a manner consistent with membranous staining, but did not decorate the unstimulated MDA-MB-231 cells or MCF-7 cells under either condition. Collectively, the results suggest the constitutive production of active MT1-MMP which is unavailable for either MMP-2 activation or immuno-decoration until Con A treatment. Since VIM expression arises by virtue of the so-called epithelial to mesenchymal transition (EMT) in invasive embryonic epithelia, we propose that this represents a major metastasis mechanism in breast carcinomas. MT1-MMP on the surface of such 'fibroblastoid' carcinoma cells may mediate a paracrine loop for the utilization of stromally produced MMP-2, and contribute to the poorer survival associated with VIM+ breast carcinomas.

Hepatocyte growth factor stimulates invasion across reconstituted basement membranes by a new human small intestinal cell line

Hepatocyte growth factor/scatter factor (HGF/SF) is a protein growth factor whose pleiotropic effects on epithelial cells include the stimulation of motility, mitosis and tubulogenesis. These responses are mediated by the cell surface tyrosine kinase receptor c-met. Because both the cytokine and receptor are found in the gastrointestinal tract, we have studied the effects of HGF/SF on transformed gut epithelial cells which express c-met. Here we describe the response of a new transformed human jejunal epithelioid cell line (HIE-7) to HGF/SF. Morphologically HIE-7 cells are immature. Their epithelial lineage was confirmed by reactivity with the epithelial specific antibodies AE1/AE3, Cam 5.2, Ber-EP4 and anti-EMA and is consistent with their expression of c-met mRNA and protein. In addition, electron microscopic analysis revealed the presence of primitive junctions and rudimentary microvilli, but features of polarization were absent. When grown on reconstituted basement membranes, HIE-7 cells formed closely associated multicellular cord-like structures adjacent to acellular spaces. However, the cells did not mature structurally, form lumen-like structures or express disaccharidase mRNA, even in the presence of recombinant HGF (rHGF). On the other hand, rHGF induced HIE-7 cells to scatter and stimulated their rapid migration in a modified wound assay. To determine whether the motogenic effect caused by rHGF is associated with HIE-7 cell invasiveness across reconstituted basement membranes, a Boyden chamber chemoinvasion assay was performed. rHGF stimulated a 10-fold increase in the number of HIE-7 cells that crossed the basement membrane barrier, while only stimulating a small increase in chemotaxis across a collagen IV matrix, suggesting that the cytokine activates matrix penetration by these cells. rHGF also stimulated the invasion of basement membranes by an undifferentiated rat intestinal cell line (IEC-6) and by two human colon cancer cell lines which are poorly differentiated (DLD-1 and SW 948). In contrast, two moderately well differentiated colon cancer cell lines (Caco-2 and HT-29) did not manifest an invasive response when exposed to rHGF. These results suggest that HGF/SF may play a significant role in the invasive behavior of anaplastic and poorly differentiated gut epithelial tumors.

Invasive phenotype of MCF10A cells overexpressing c-Ha-ras and c-erbB-2 oncogenes

Infection with erbB-2 (E) of Ha-ras (H) oncogene-transfected cells has been previously shown to cooperatively induce anchorage-independent growth of the MCF10A human mammary epithelial cell line in vitro, but not to induce nude mouse tumorigenicity. Here we show that oncogene-transformed MCF10A are able to halt in the lungs of nude mice, a sign of organ colonization potential. We have therefore studied the transformants for in vitro migratory and invasive properties known to correlate with the metastatic potential of human mammary carcinoma cells in nude mice. MCF10A transfected with Ha-ras, infected with a recombinant retroviral vector containing the human c-erB-2 proto-oncogene (MCF10A-HE cells), show a higher invasive index than either the single transfectant (MCF10A-H) or MCF10A-erB-2(MCF10A-E) cells in the Boyden chamber chemotaxis and chemoinvasion assays. The MCF10A-HE cells also adopted an invasive stellate growth pattern when plated or embedded in Matrigel, in contrast to the spherical colonies formed by the single transformants MCF10A-H, MCF10A-E, and the parental cells. Dot-blot analysis of gelatinase A and TIMP-2 mRNA levels revealed increasing gelatinase A mRNA levels (HE > E > H > MCF10A) and reduced TIMP-2 expression in both single and double transformants. Furthermore, MCF10A-HE cells show more MMP-2 activity than parental MCF10A cells or the single transformants. CD44 analysis revealed differential isoform banding for the MCF10A-HE cells compared to parental cells, MCF10A-H and MCF10A-E, accompanied by increased binding of hyaluronan by the double transformants. Our results indicate that erB-2 and Ha-ras co-expression can induce a more aggressive phenotype in vitro, representative of the malignancy of mammary carcinomas.

Induction of tumorigenicity by galectin-3 in a nontumorigenic human breast-carcinoma cell-line

The human galectin-3 is a galactoside-binding protein of 31 kDa which functions as a receptor for glycoproteins containing poly N-acetyllactosamine side chains and as a substrate for matrix metalloproteinases-2 and -9. We studied its expression by flow cytofluorimetry, Western, Northern and Southern analyses, in five cultured human breast carcinoma cell lines previously characterized as non-tumorigenic, poorly metastatic or metastatic in nude mice. The expression of galectin-3 correlated with the reported tumorigenicity of the cells. The introduction of recombinant galectin-3 into the null expressing non-tumorigenic BT-549 cells resulted in the acquisition of anchorage-independent growth properties in alland tumorigenicity in 3/4 sense transfected cell crones. The data indicate a relationship between galectin-3 expression and malignancy of human breast carcinoma cell lines.

Association of MMP-2 activation potential with metastatic progression in human breast cancer cell lines independent of MMP-2 production

Background: Expression of matrix metalloproteinase-2 (MMP-2), the 72-kd type IV collagenase/gelatinase, by cancer cells has been implicated in metastasis through cancer cell invasion of basement membranes mediated by degradation of collagen IV. However, the abundance of this latent proenzyme in normal tissues and fluids suggests that MMP-2 proenzyme utilization is limited by its physiological activation rather than expression alone. We previously reported activation of this proenzyme by normal and malignant fibroblastoid cells cultured on collagen I (vitrogen) gels. Purpose: Our purposes in this study were 1) to determine whether MMP-2 activation is restricted to the more invasive human breast cancer cell lines and 2) to localize the activating mechanism. Methods: Zymography was used to monitor MMP-2 activation through detection of latent MMP-2 (72 kd) and mature species of smaller molecular weight (59 or 62 kd). Human breast cancer cell lines cultured on plastic, vitrogen, and other matrices were thus screened for MMP- 2 activation. Collagen I-cultured cells were exposed to cycloheximide, a protein synthesis inhibitor, or to protease inhibitors to determine the nature of the MMP-2-activating mechanism. Triton X-114 (TX-114) detergent extracts from cells cultured on collagen I or plastic were incubated with latent MMP-2 and analyzed by zymography to localize the MMP-2 activator. Results: MMP-2 activation was only induced by collagen I culture in the more aggressive, highly invasive estrogen receptor-negative, vimentin-positive human breast cancer cell lines (Hs578T, MDA-MB-436, BT549, MDA-MB-231, MDA- MB-435, MCF-7(ADR)) and was independent of MMP-2 production. MMP-2 activation was detected in cells cultured on collagen I gels but not in those cultured on gelatin gels, Matrigel, or thin layers of collagen I or IV, gelatin, or fibronectin. Collagen-induced activation was specific for the enzyme species MMP-2, since MMP-9, the 92-kd type IV collagenase/gelatinase, was not activatable under similar conditions. MMP-2 activation was inhibited by cycloheximide and was sensitive to a metalloproteinase inhibitor but not to aspartyl, serine, or cysteinyl protease inhibitors. MMP-2 activation was detected in the hydrophobic, plasma membrane-enriched, TX-114 extracts from invasive collagen I-cultured cells. Conclusion: Collagen I-induced MMP-2 activation is restricted to highly invasive estrogen receptor-negative, vimentin-positive human breast cancer cell lines, is independent of MMP-2 production, and is associated with metastatic potential. Our findings are consistent with plasma membrane localization of the activator. Implications: The MMP-2 activation mechanism may represent a new target for diagnosis, prognosis, and treatment of human breast cancer.

The type I collagen induction of MT1-MMP-mediated MMP-2 activation is repressed by αVβ3 integrin in human breast cancer cells

The influence of αVβ3 integrin on MT1-MMP functionality was studied in human breast cancer cells of differing β3 integrin status. Overexpression of β3 integrin caused increased cell surface expression of αV integrin and increased cellular adhesion to extracellular matrix (ECM) substrates in BT-549, MDA-MB-231 and MCF-7 cells. β3 integrin expression also enhanced the migration of breast cancer cells on ECM substrates and enhanced collagen gel contraction. In vivo, αVβ3 cooperated with MT1-MMP to increase the growth of MCF-7 cells after orthotopic inoculation in immunocompromised mice, but had no influence on in vitro proliferation. Despite these stimulatory effects, overexpression of β3 integrin suppressed the type I collagen (Col I) induced MMP-2 activation in all breast cancer cell lines analyzed. This was also evident in extracts from the MCF-7 tumors in vivo, where MMP-2 activation was stimulated by MT1-MMP transfection, but attenuated with β3 integrin expression. Although our studies confirm important biological effects of αVβ3 integrin on enhancing cell adhesion and migration, ECM remodeling and tumor growth, β3 integrin caused reduced MMP-2 activation in response to Col I in vitro, which appears to be physiologically relevant, as it was also seen in tumor xenografts in vivo. The reduction of MMP-2 activation (and thus MT1-MMP activity) by αVβ3 in response to Col I may be important in scenarios where cells which are activated for matrix degradation need to preserve some pericellular collagen, perhaps as a substrate for cell adhesion and migration, thus maintaining a balanced level of proteolysis required for efficient tumor growth.

Modelling the dynamics of Plasmodium falciparum histidine-rich protein 2 in human malaria to better understand malaria rapid diagnostic test performance

BACKGROUND: Effective diagnosis of malaria is a major component of case management. Rapid diagnostic tests (RDTs) based on Plasmodium falciparumhistidine-rich protein 2 (PfHRP2) are popular for diagnosis of this most virulent malaria infection. However, concerns have been raised about the longevity of the PfHRP2 antigenaemia following curative treatment in endemic regions. METHODS: A model of PfHRP2 production and decay was developed to mimic the kinetics of PfHRP2 antigenaemia during infections. Data from two human infection studies was used to fit the model, and to investigate PfHRP2 kinetics. Four malaria RDTs were assessed in the laboratory to determine the minimum detectable concentration of PfHRP2. RESULTS: Fitting of the PfHRP2 dynamics model indicated that in malaria naive hosts, P. falciparum parasites of the 3D7 strain produce 1.4 x 10(-)(1)(3) g of PfHRP2 per parasite per replication cycle. The four RDTs had minimum detection thresholds between 6.9 and 27.8 ng/mL. Combining these detection thresholds with the kinetics of PfHRP2, it is predicted that as few as 8 parasites/muL may be required to maintain a positive RDT in a chronic infection. CONCLUSIONS: The results of the model indicate that good quality PfHRP2-based RDTs should be able to detect parasites on the first day of symptoms, and that the persistence of the antigen will cause the tests to remain positive for at least seven days after treatment. The duration of a positive test result following curative treatment is dependent on the duration and density of parasitaemia prior to treatment and the presence and affinity of anti-PfHRP2 antibodies.

Conjugation of carcinogens by 6 class glutathione S-transferases : mechanisms and relevance to variations in human risk

Conjugation of chemicals with glutathione (GSH) can lead to decreased or increased toxicity. A genetic deficiency in the GSH S-transferase μ class gene M1 has been hypothesized to lead to greater risk of lung cancer in smokers. Recently a gene deletion polymorphism involving the human θ enzyme T1 has been described; the enzyme is present in erythrocytes and can be readily assayed. A rat θ class enzyme, 5-5, has structural and catalytic similarity and the protein was expressed in the Salmonella typhimurium tester strain TA1535. Expression of the cDNA vector increased the mutagenicity of ethylene dibromide and several methylene dihalides. Mutations resulting from the known GSH S-transferase substrate 1,2-epoxy-3-(4′nitrophenoxy)propane were decreased in the presence of the transferase. Expression of transferase 5-5 increased mutations when 1,2,3,4-diepoxybutane (butadiene diepoxide), 4-bromo-1,2-epoxybutane, or 1,3-dichloracetone were added. The latter compound is a model for the putative 1,2-dibromo-3-chloropropane oxidation product 1-bromo-3-chloroacetone. These genotoxicity and genotyping assays may be of use in further studies of the roles of GSH S-transferase θ enzymes in bioactivation and detoxication and any changes in risk due to polymorphism.

Human glutathione S-transferase T1-1 enhances mutagenicity of 1,2-dibromoethane, dibromomethane and 1,2,3,4-diepoxybutane in Salmonella typhimurium

The rat theta class glutathione S-transferase (GST) 5-5 has been shown to affect the mutagenicity of halogenated alkanes and epoxides. In Salmonella typhimurium TA1535 expressing the rat GST5-5 the number of revertants was increased compared to the control strain by CH2Br2, ethylene dibromide (EDB) and 1,2,3,4-diepoxybutane (BDE); in contrast, mutagenicity of 1,2-epoxy-3-(4'-nitrophenoxy)propane (EPNP) was reduced. S.typhimurium TA1535 cells were transformed with an expression plasmid carrying the cDNA of the human theta ortholog GST1-1 either in sense or antisense orientation, the latter being the control. These transformed bacteria were utilized for mutagenicity assays. Mutagenicity of EDB, BDE, CH2Br2, epibromohydrin and 1,3-dichloroacetone was higher in the S.typhimurium TA1535 expressing GSTT1-1 than in the control strain. The expression of active enzyme did not affect the mutagenicity of 1,2-epoxy-3-butene or propylene oxide, GSTT1-1 expression reduced the mutagenicity of EPNP. Glutathione S-transferase 5-5 and GSTT1-1 modulate genotoxicity of several industrially important chemicals in the same way. Polymorphism of the GSTT1 locus in humans may therefore cause differences in cancer susceptibility between the two phenotypes.