Validation procedures in radiological diagnostic models. Neural network and logistic regression

The objective of this paper is to compare the performance of twopredictive radiological models, logistic regression (LR) and neural network (NN), with five different resampling methods. One hundred and sixty-seven patients with proven calvarial lesions as the only known disease were enrolled. Clinical and CT data were used for LR and NN models. Both models were developed with cross validation, leave-one-out and three different bootstrap algorithms. The final results of each model were compared with error rate and the area under receiver operating characteristic curves (Az). The neural network obtained statistically higher Az than LR with cross validation. The remaining resampling validation methods did not reveal statistically significant differences between LR and NN rules. The neural network classifier performs better than the one based on logistic regression. This advantage is well detected by three-fold cross-validation, but remains unnoticed when leave-one-out or bootstrap algorithms are used.

The dexamethasone-induced inhibition of proliferation, migration, and invasion in glioma cell lines is antagonized by macrophage migration inhibitory factor (MIF) and can be enhanced by specific MIF inhibitors.

Glioblastomas (GBMs) are the most frequent and malignant brain tumors in adults. Glucocorticoids (GCs) are routinely used in the treatment of GBMs for their capacity to reduce the tumor-associated edema. Few in vitro studies have suggested that GCs inhibit the migration and invasion of GBM cells through the induction of MAPK phosphatase 1 (MKP-1). Macrophage migration inhibitory factor (MIF), an endogenous GC antagonist is up-regulated in GBMs. Recently, MIF has been involved in tumor growth and migration/invasion and specific MIF inhibitors have been developed on their capacity to block its enzymatic tautomerase activity site. In this study, we characterized several glioma cell lines for their MIF production. U373 MG cells were selected for their very low endogenous levels of MIF. We showed that dexamethasone inhibits the migration and invasion of U373 MG cells, through a glucocorticoid receptor (GR)- dependent inhibition of the ERK1/2 MAPK pathway. Oppositely, we found that exogenous MIF increases U373 MG migration and invasion through the stimulation of the ERK1/2 MAP kinase pathway and that this activation is CD74 independent. Finally, we used the Hs 683 glioma cells that are resistant to GCs and produce high levels of endogenous MIF, and showed that the specific MIF inhibitor ISO-1 could restore dexamethasone sensitivity in these cells. Collectively, our results indicate an intricate pathway between MIF expression and GC resistance. They suggest that MIF inhibitors could increase the response of GBMs to corticotherapy.

Identification of optimal structural connectivity using functional connectivity and neural modeling.

The complex network dynamics that arise from the interaction of the brain's structural and functional architectures give rise to mental function. Theoretical models demonstrate that the structure-function relation is maximal when the global network dynamics operate at a critical point of state transition. In the present work, we used a dynamic mean-field neural model to fit empirical structural connectivity (SC) and functional connectivity (FC) data acquired in humans and macaques and developed a new iterative-fitting algorithm to optimize the SC matrix based on the FC matrix. A dramatic improvement of the fitting of the matrices was obtained with the addition of a small number of anatomical links, particularly cross-hemispheric connections, and reweighting of existing connections. We suggest that the notion of a critical working point, where the structure-function interplay is maximal, may provide a new way to link behavior and cognition, and a new perspective to understand recovery of function in clinical conditions.

The role of migration and domestic transmission in the spread of HIV-1 non-B subtypes in Switzerland.

BACKGROUND: By analyzing human immunodeficiency virus type 1 (HIV-1) pol sequences from the Swiss HIV Cohort Study (SHCS), we explored whether the prevalence of non-B subtypes reflects domestic transmission or migration patterns. METHODS: Swiss non-B sequences and sequences collected abroad were pooled to construct maximum likelihood trees, which were analyzed for Swiss-specific subepidemics, (subtrees including ≥80% Swiss sequences, bootstrap >70%; macroscale analysis) or evidence for domestic transmission (sequence pairs with genetic distance <1.5%, bootstrap ≥98%; microscale analysis). RESULTS: Of 8287 SHCS participants, 1732 (21%) were infected with non-B subtypes, of which A (n = 328), C (n = 272), CRF01_AE (n = 258), and CRF02_AG (n = 285) were studied further. The macroscale analysis revealed that 21% (A), 16% (C), 24% (CRF01_AE), and 28% (CRF02_AG) belonged to Swiss-specific subepidemics. The microscale analysis identified 26 possible transmission pairs: 3 (12%) including only homosexual Swiss men of white ethnicity; 3 (12%) including homosexual white men from Switzerland and partners from foreign countries; and 10 (38%) involving heterosexual white Swiss men and females of different nationality and predominantly nonwhite ethnicity. CONCLUSIONS: Of all non-B infections diagnosed in Switzerland, <25% could be prevented by domestic interventions. Awareness should be raised among immigrants and Swiss individuals with partners from high prevalence countries to contain the spread of non-B subtypes.

Incremental neural networks for function approximation

A new strategy for incremental building of multilayer feedforward neural networks is proposed in the context of approximation of functions from R-p to R-q using noisy data. A stopping criterion based on the properties of the noise is also proposed. Experimental results for both artificial and real data are performed and two alternatives of the proposed construction strategy are compared.

Macrophage migration inhibitory factor is involved in a positive feedback loop increasing aromatase expression in endometriosis.

Immune-endocrine interplay may play a major role in the pathogenesis of endometriosis. In the present study, we have investigated the interaction between macrophage migration inhibitory factor (MIF), a major pro-inflammatory and growth-promoting factor markedly expressed in active endometriotic lesions, and estradiol (E(2)) in ectopic endometrial cells. Our data showed a significant increase of MIF protein secretion and mRNA expression in endometriotic cells in response to E(2). MIF production was blocked by Fulvestrant, an estrogen receptor (ER) antagonist, and induced by ERα and ERβ selective agonists propyl-pyrazole-triol (PPT) and diarylpropionrile (DPN), respectively, thus demonstrating a specific receptor-mediated effect. Cell transfection with MIF promoter construct showed that E(2) significantly stimulates MIF promoter activity. Interestingly, our data further revealed that MIF reciprocally stimulates aromatase protein and mRNA expression via a posttranscriptional mRNA stabilization mechanism, that E(2) itself can upregulate aromatase expression, and that inhibition of endogenous MIF, using MIF specific siRNA, significantly inhibits E(2)-induced aromatase. Thus, the present study revealed the existence of a local positive feedback loop by which estrogen acts directly on ectopic endometrial cells to upregulate the expression of MIF, which, in turn, displays the capability of inducing the expression of aromatase, the key and rate-limiting enzyme for estrogen synthesis. Such interplay may have a considerable impact on the development of endometriosis.

Targeting the PI3K p110alpha isoform inhibits medulloblastoma proliferation, chemoresistance, and migration.

PURPOSE: The phosphoinositide 3-kinase (PI3K)/Akt pathway is frequently activated in human cancer and plays a crucial role in medulloblastoma biology. We were interested in gaining further insight into the potential of targeting PI3K/Akt signaling as a novel antiproliferative approach in medulloblastoma. EXPERIMENTAL DESIGN: The expression pattern and functions of class I(A) PI3K isoforms were investigated in medulloblastoma tumour samples and cell lines. Effects on cell survival and downstream signaling were analyzed following down-regulation of p110alpha, p110beta, or p110delta by means of RNA interference or inhibition with isoform-specific PI3K inhibitors. RESULTS: Overexpression of the catalytic p110alpha isoform was detected in a panel of primary medulloblastoma samples and cell lines compared with normal brain tissue. Down-regulation of p110alpha expression by RNA interference impaired the growth of medulloblastoma cells, induced apoptosis, and led to decreased migratory capacity of the cells. This effect was selective, because RNA interference targeting of p110beta or p110delta did not result in a comparable impairment of DAOY cell survival. Isoform-specific p110alpha inhibitors also impaired medulloblastoma cell proliferation and sensitized the cells to chemotherapy. Medulloblastoma cells treated with p110alpha inhibitors further displayed reduced activation of Akt and the ribosomal protein S6 kinase in response to stimulation with hepatocyte growth factor and insulin-like growth factor-I. CONCLUSIONS: Together, our data reveal a novel function of p110alpha in medulloblastoma growth and survival.

Construction and electrophoretic migration of single-stranded DNA knots and catenanes.

In recent years there has been growing interest in the question of how the particular topology of polymeric chains affects their overall dimensions and physical behavior. The majority of relevant studies are based on numerical simulation methods or analytical treatment; however, both these approaches depend on various assumptions and simplifications. Experimental verification is clearly needed but was hampered by practical difficulties in obtaining preparative amounts of knotted or catenated polymers with predefined topology and precisely set chain length. We introduce here an efficient method of production of various single-stranded DNA knots and catenanes that have the same global chain length. We also characterize electrophoretic migration of the produced single-stranded DNA knots and catenanes with increasing complexity.

Desenvolvimento e análise de uma rede neural artificial para estimativa da erosividade da chuva para o Estado de São Paulo

O conhecimento do valor da erosividade da chuva (R) de determinada localidade é fundamental para a estimativa das perdas de solo feitas a partir da Equação Universal de Perdas de Solo, sendo, portanto, de grande importância no planejamento conservacionista. A fim de obter estimativas do valor de R para localidades onde este é desconhecido, desenvolveu-se uma rede neural artificial (RNA) e analisou-se a acurácia desta com o método de interpolação "Inverso de uma Potência da Distância" (ID). Comparando a RNA desenvolvida com o método de interpolação ID, verificou-se que a primeira apresentou menor erro relativo médio na estimativa de R e melhor índice de confiança, classificado como "Ótimo", podendo, portanto, ser utilizada no planejamento de uso, manejo e conservação do solo no Estado de São Paulo.

NSAIDs inhibit alpha V beta 3 integrin-mediated and Cdc42/Rac-dependent endothelial-cell spreading, migration and angiogenesis.

Cyclooxygenase-2 (COX-2), a key enzyme in arachidonic acid metabolism, is overexpressed in many cancers. Inhibition of COX-2 by nonsteroidal anti-inflammatory drugs (NSAIDs) reduces the risk of cancer development in humans and suppresses tumor growth in animal models. The anti-cancer effect of NSAIDs seems to involve suppression of tumor angiogenesis, but the underlying mechanism is not completely understood. Integrin alpha V beta 3 is an adhesion receptor critically involved in mediating tumor angiogenesis. Here we show that inhibition of endothelial-cell COX-2 by NSAIDs suppresses alpha V beta 3-dependent activation of the small GTPases Cdc42 and Rac, resulting in inhibition of endothelial-cell spreading and migration in vitro and suppression of fibroblast growth factor-2-induced angiogenesis in vivo. These results establish a novel functional link between COX-2, integrin alpha V beta 3 and Cdc42-/Rac-dependent endothelial-cell migration. Moreover, they provide a rationale to the understanding of the anti-angiogenic activity of NSAIDs.

Podocalyxin is a novel polysialylated neural adhesion protein with multiple roles in neural development and synapse formation

Neural development and plasticity are regulated by neural adhesion proteins, including the polysialylated form of NCAM (PSA-NCAM). Podocalyxin (PC) is a renal PSA-containing protein that has been reported to function as an anti-adhesin in kidney podocytes. Here we show that PC is widely expressed in neurons during neural development. Neural PC interacts with the ERM protein family, and with NHERF1/2 and RhoA/G. Experiments in vitro and phenotypic analyses of podxl-deficient mice indicate that PC is involved in neurite growth, branching and axonal fasciculation, and that PC loss-of-function reduces the number of synapses in the CNS and in the neuromuscular system. We also show that whereas some of the brain PC functions require PSA, others depend on PC per se. Our results show that PC, the second highly sialylated neural adhesion protein, plays multiple roles in neural development.