Artificial Neural Networks in Image Processing for Early Detection of Breast Cancer.

Medical imaging techniques have widely been in use in the diagnosis and detection of breast cancer. The drawback of applying these techniques is the large time consumption in the manual diagnosis of each image pattern by a professional radiologist. Automated classifiers could substantially upgrade the diagnosis process, in terms of both accuracy and time requirement by distinguishing benign and malignant patterns automatically. Neural network (NN) plays an important role in this respect, especially in the application of breast cancer detection. Despite the large number of publications that describe the utilization of NN in various medical techniques, only a few reviews are available that guide the development of these algorithms to enhance the detection techniques with respect to specificity and sensitivity. The purpose of this review is to analyze the contents of recently published literature with special attention to techniques and states of the art of NN in medical imaging. We discuss the usage of NN in four different medical imaging applications to show that NN is not restricted to few areas of medicine. Types of NN used, along with the various types of feeding data, have been reviewed. We also address hybrid NN adaptation in breast cancer detection.

Binding of a 50-kD Protein to a U-Rich Sequence in an mRNA Encoding a Proline-Rich Protein That Is Destabilized by Fungal Elicitor.

The mRNA encoding the bean proline-rich protein PvPRP1 has been shown previously to be destabilized in elicitor-treated cells. In this study, we identified a 50-kD protein in cellular extracts that binds specifically to the PvPRP1 mRNA by UV cross-linking assays. Using 32P-labeled RNAs transcribed in vitro from a series of 5[prime] deleted PvPRP1 cDNA clones, we demonstrated that the PvPRP1 mRNA binding protein (PRP-BP) binds to a 27-nucleotide U-rich (~60%) domain in the 3[prime] untranslated region. Poly(U) and, to a lesser extent, poly(A-U) competed for the PRP-BP binding activity. PRP-BP activity is redox regulated in vitro, as shown by the effects of sulfhydryl-modifying reagents on the RNA binding activity. Treatment of cellular extracts with the reducing agents DTT and [beta]-mercaptoethanol increased binding activity, whereas treatment with the oxidizing agent diamide and the alkylating agent N-ethylmaleimide inhibited binding. In extracts from elicitor-treated cells, PRP-BP activity increased approximately fivefold prior to rapid PvPRP1 mRNA degradation. The increase in PRP-BP activity was apparently due to post-translational regulation because control and elicitor-treated cell extracts supplemented with DTT showed high comparable levels of RNA binding activity. The kinetics of PRP-BP activation after elicitor treatment and its capacity for redox regulation in vitro suggested that PRP-BP may function in the elicitor-induced destabilization of PvPRP1 mRNA.

Fungal Elicitor-Induced Bean Proline-Rich Protein mRNA Down-Regulation Is Due to Destabilization That Is Transcription and Translation Dependent.

In bean cells treated with fungal elicitor, the transcripts of PvPRP1, a gene encoding a proline-rich protein, decreased to ~6% of the original level within 4 hr. The apparent mRNA half-life during the period of rapid degradation was ~45 min. The rate of PvPRP1 gene transcription remained constant over this period, as determined by nuclear run-off assays, indicating a decrease in mRNA stability. By using actinomycin D to block transcription, the half-life of PvPRP1 mRNA in unelicited cells was estimated to be ~60 hr. In cells treated with actinomycin D followed by the addition of elicitor, the PvPRP1 mRNA half-life was ~18 hr, whereas cells treated with these reagents in reciprocal order exhibited a half-life of ~6 hr. The protein synthesis inhibitors emetine and anisomycin also inhibited the rate of PvPRP1 mRNA degradation in elicited cells. Based on these data, we concluded that the rapid decrease in the PvPRP1 mRNA level in elicited cells is due to destabilization, which is dependent on new RNA and protein synthesis.

Purification, characterization, and cDNA cloning of an NADPH-cytochrome P450 reductase from mung bean.

We report here the isolation and deduced amino acid sequence of the flavoprotein, NADPH-cytochrome P450 (cytochrome c) reductase (EC 1.6.2.4), associated with the microsomal fraction of etiolated mung bean seedlings (Vigna radiata var. Berken). An 1150-fold purification of the plant reductase was achieved, and SDS/PAGE showed a predominant protein band with an apparent molecular mass of approximately 82 kDa. The purified plant NADPH-P450 reductase gave a positive reaction as a glycoprotein, exhibited a typical flavoprotein visible absorbance spectrum, and contained almost equimolar quantities of FAD and FMN per mole of enzyme. Specific antibodies revealed the presence of unique epitopes distinguishing the plant and mammalian flavoproteins as demonstrated by Western blot analyses and inhibition studies. Peptide fragments from the purified plant NADPH-P450 reductase were sequenced, and degenerate primers were used in PCR amplification reactions. Overlapping cDNA clones were sequenced, and the deduced amino acid sequence of the mung bean NADPH-P450 reductase was compared with equivalent enzymes from mammalian species. Although common flavin and NADPH-binding sites are recognizable, there is only approximately 38% amino acid sequence identity. Surprisingly, the purified mung bean NADPH-P450 reductase can substitute for purified rat NADPH-P450 reductase in the reconstitution of the mammalian P450-catalyzed 17 alpha-hydroxylation of pregnenolone or progesterone.

Developmental regulation and phytochrome-mediated induction of mRNAs encoding a proline-rich protein, glycine-rich proteins, and hydroxyproline-rich glycoproteins in Phaseolus vulgaris L.

We have studied developmental and light regulation of mRNAs encoding a putative cell wall proline-rich protein (PvPRP1), cell wall glycine-rich proteins (GRPs), and cell wall hydroxyproline-rich glycoproteins (HRGPs) in bean (Phaseolus vulgaris). Light increases the levels of these mRNAs 2- to 150-fold in highly spatially regulated patterns during seedling development. These mRNA changes include differential regulation of transcripts derived from the GRP and HRGP multigene families. In 6-day-old light-grown seedlings, the PvPRP1 and GRP1.0 mRNAs were most abundant in the apical region of hypocotyls, epicotyls, and roots. In contrast, several HRGP transcripts were most abundant in the mature region of hypocotyls and roots in light-grown seedlings. When etiolated 6-day-old seedlings were illuminated with white light for 8 hr, maximal accumulation of PvPRP1 and GRP1.0 mRNAs occurred in the apical hook, whereas HRGP and GRP1.8 mRNAs accumulated in the mature region of hypocotyls. Etiolated seedlings subjected to a pulse of red light accumulated PvPRP1, GRP, and HRGP mRNAs in the hypocotyls. Far-red light inhibited red light induction of these mRNAs, indicating a phytochrome-mediated process. The possible roles of PRPs, GRPs, and HRGPs in cell differentiation and photomorphogenesis are discussed.

cDNA cloning, structure and expression of a novel pathogenesis-related protein in bean.

We have isolated cDNA clones spanning the full length of a transcript (PvPR3) encoding a novel pathogenesis-related (PR) protein in bean (Phaseolus vulgaris). The PvPR3 transcript accumulates gradually over 24 hr in elicitor-treated cell suspensions. This pattern of expression is distinct from those of previously reported elicitor-induced transcripts in bean. Specifically, transcripts encoding two recently described acidic bean PR proteins, phenylpropanoid pathway enzymes, accumulate to maximal levels by 4-8 hr, while hydroxyproline-rich glycoprotein mRNA accumulation is delayed by several hours. The PvPR3 mRNA also accumulates after wounding of hypocotyls with kinetics comparable to those of mRNA encoding phenylpropanoid pathway mRNAs. PvPR3 appears to exist as a single gene within a family of approximately 15 related genes in the bean genome. The PvPR3 protein deduced from the cDNA sequences (14,950 Da pI = 10.0) lacks a putative signal peptide suggesting a cytosolic localization. Amino acid sequence comparisons with databases revealed that PvPR3 represents a new class of PR proteins without significant sequence homology to previously characterized PR proteins or other proteins.

Negative and positive regulation of a novel proline-rich protein mRNA by fungal elicitor and wounding.

The structure and expression of a cDNA clone (PvPRP1) isolated from a cDNA library prepared from bean (Phaseolus vulgaris) cells treated with fungal elicitor have been characterized. Sequence analysis of the 1.1 kb insert revealed a complete open reading frame which encodes a 32 kDa protein. The protein resembles other proline-rich proteins in plants but possesses several unique features: (i) the N-terminal half of the protein is proline rich and contains three identical repeats of Pro-Val-His-Pro-Pro-Val-Lys-Pro-Pro-Val and six related repeats; (ii) the proline-rich region contains two tracts of six histidine residues; and (iii) the C-terminal half is low in proline and lacks repeats. Genomic blotting experiments suggest the presence of a single PvPRP1 gene as well as more distantly related genes within the bean genome. A dramatic decrease in PvPRP1 mRNA levels occurs within 2 h of elicitor treatment of cell cultures. The PvPRP1 mRNA is present at a moderate level in hypocotyls. Upon wounding, the mRNA level initially decreases over 5 h and then accumulates over 25 h to levels which are higher than the basal level in unwounded hypocotyls. Based on the similarity to other proline-rich proteins with repeated motifs, including the presence of a putative signal peptide, it is likely that the PvPRP1 protein is targetted to the cell wall. The expression of the PvPRP1 gene appears to be integrated with the remodeling of the plant cell wall during the defense response.

Chalcone isomerase cDNA cloning and mRNA induction by fungal elicitor, wounding and infection.

The environmentally regulated synthesis of phenylpropanoid natural products was studied by examining the expression of the gene encoding chalcone isomerase (CHI). This enzyme catalyzes a step common to the synthesis of flavonoid pigments and isoflavonoid phytoalexins. A lambdagt11 library was constructed using mRNA from cell cultures of bean (Phaseolus vulgaris L.) treated with fungal elicitor. Two positive clones were obtained by screening 10 recombinants with an antiserum to purified bean CHI. The identity of the cloned sequences was confirmed by hybrid-select translation and the production of antigenic polypeptides from transcripts synthesized in vitro. Addition of elicitor to cell cultures resulted in the rapid accumulation of CHI mRNA, with maximum levels achieved 3-4 h after elicitation. CHI mRNA also accumulated during the natural infection of hypocotyls with the fungal pathogen Colletotrichum lindemuthianum, and in mechanically wounded hypocotyls. The kinetics of accumulation of CHI mRNA in response to these environmental signals were strikingly similar to those of mRNAs encoding two other phenylpropanoid pathway enzymes, phenylalanine ammonialyase and chalcone synthase. In contrast to the multi-gene families encoding these two enzymes, chalcone isomerase is encoded by a single gene which is regulated by several environmental stimuli.

A secreted factor and cyclic AMP jointly regulate cell-type-specific gene expression in Dictyostelium discoideum.

We are studying cell differentiation in Dictyostelium discoideum by examining the regulation of genes that are preferentially expressed in different cell types. A system has been established in which prestalk- and prespore-cell-specific genes are expressed in single cells in response to culture conditions. We confirm our previous results showing that cyclic AMP induces prestalk genes and now show that it is also required for prespore gene induction. The expression of both classes of genes is additionally dependent on the presence of a factor(s) secreted by developing cells which we call conditioned medium factor(s). An assay for conditioned medium factor(s) shows that it is detectable within 2.5 h after the onset of development. Conditioned medium factor(s) also promotes the expression of genes induced early in development, but has no detectable effect on the expression of actin genes and a gene expressed maximally in vegetative cells. In the presence of conditioned medium factor(s), exogenous cyclic AMP at the onset of starvation fails to induce the prespore and prestalk genes. The addition of cyclic AMP between 2 and 12 h of starvation results in rapid prestalk gene expression, whereas prespore genes are induced at an invarient time (approximately 18 h after the onset of starvation). These data suggest that cyclic AMP and conditioned medium factor(s) are sufficient for prestalk gene induction, whereas an additional parameter(s) is involved in the control of prespore gene induction. In contrast to several previous studies, we show that multicellularity is not essential for the expression of either prespore or prestalk genes. These data indicate that prespore and prestalk genes have cell-type-specific as well as shared regulatory factors.

Developmental regulation of a Dictyostelium gene encoding a protein homologous to mammalian ras protein.

We have cloned, sequenced, and examined the regulation of a Dictyostelium gene encoding a protein homologous to mammalian ras proteins. The Dictyostelium, yeast, and mammalian proteins have homologous N-terminal regions and less conserved C-terminal regions. We have used DNA probes and a polyclonal antibody to examine the differential accumulation of ras RNA and protein through development. The gene encodes two mRNAs (0.9 and 1.2 kb) that are differentially expressed. The 1.2 kb RNA is found in vegetative cells and disappears rapidly upon initiation of development. Later, both RNAs accumulate preferentially in prestalk cells. The level of the Dd-ras protein remains constant until early culmination and then decreases. Like other prestalk genes, Dd-ras can be induced with cAMP in the absence of cell contact. When aggregated cells are dissociated, both mRNAs decrease. Upon addition of cAMP, the 1.2 kb mRNA reaccumulates at a higher level than that in normal developing cells. The presence of the Dd-ras protein in vegetative cells corroborates other reports suggesting a possible function during cell growth. The sustained level of Dd-ras protein in prestalk cells suggests an additional role during differentiation.

Induction and modulation of cell-type-specific gene expression in Dictyostelium.

We have identified genes that are expressed preferentially in either prestalk or prespore cells in Dictyostelium. The prestalk mRNAs are detectable at 7.5 hr prior to the completion of cell aggregation, while the prespore mRNAs are not detectable until approximately 15 hr of development. Exogenous cAMP in the absence of sustained cell contact is sufficient to induce prestalk-specific gene expression, while multicellularity is required for the induction of prespore-specific genes. A gene expressed equally in both cell types, which has the same developmental kinetics as the prestalk genes, is induced in shaking culture in the absence of either cAMP or stable cell associations. Dissociation of aggregates results in the rapid loss of prespore- and prestalk-specific mRNAs, and these can be induced to reaccumulate with the addition of cAMP. We conclude that there are substantial differences in the timing and requirements for tissue-specific gene expression in Dictyostelium.