New algorithms for multi-class cancer diagnosis using tumor gene expression signatures.

MOTIVATION: The increasing use of DNA microarray-based tumor gene expression profiles for cancer diagnosis requires mathematical methods with high accuracy for solving clustering, feature selection and classification problems of gene expression data. RESULTS: New algorithms are developed for solving clustering, feature selection and classification problems of gene expression data. The clustering algorithm is based on optimization techniques and allows the calculation of clusters step-by-step. This approach allows us to find as many clusters as a data set contains with respect to some tolerance. Feature selection is crucial for a gene expression database. Our feature selection algorithm is based on calculating overlaps of different genes. The database used, contains over 16 000 genes and this number is considerably reduced by feature selection. We propose a classification algorithm where each tissue sample is considered as the center of a cluster which is a ball. The results of numerical experiments confirm that the classification algorithm in combination with the feature selection algorithm perform slightly better than the published results for multi-class classifiers based on support vector machines for this data set. AVAILABILITY: Available on request from the authors.

ST14A cells have properties of a medium-size spiny neuron.

The ST14A cell line was previously derived from embryonic day 14 rat striatal primordia by retroviral transduction of the temperature-sensitive SV40 large T antigen. We showed that cell division and expression of nestin persists at 33 degrees C, the permissive temperature, whereas cell division ceases, nestin expression decreases, and MAP2 expression increases at the nonpermissive temperature of 39 degrees C. In this study, we further characterized the cells and found that they express other general and subtype-specific neuronal characteristics. ST14A cells express enolase and beta III-tubulin. Furthermore, they express the striatal marker DARPP-32, which is up-regulated upon differentiation of the cells by growth in serum-free medium. Stimulation with dopamine, the D2-dopamine receptor agonist quinpirole, or the D1-dopamine receptor agonist SKF82958 results in phosphorylation of CREB. Treatment of the cells with a mixture of reagents which stimulate the MAPK and adenylyl cyclase pathways radically changes the morphology of the ST14A cells. The cells develop numerous neurite-like appearing processes which stain with beta III-tubulin. Moreover, under these conditions, intracellular injection of rectangular depolarizing current stimuli elicits overshooting action potentials with a relatively fast depolarization rate when starting from a strongly hyperpolarized membrane potential. Taken together, these data imply that the ST14A cell line displays some of the characteristics of a medium-size spiny neuron subtype and provides a new tool to elucidate the pathways and molecules involved in medium-size spiny neuron differentiation and disease.

Expression of the striatal DARPP-32/ARPP-21 phenotype in GABAergic neurons requires neurotrophins in vivo and in vitro.

The medium spiny neuron (MSN) is the major output neuron of the caudate nucleus and uses GABA as its primary neurotransmitter. A majority of MSNs coexpress DARPP-32 and ARPP-21, two dopamine and cyclic AMP-regulated phosphoproteins, and most of the matrix neurons express calbindin. DARPP-32 is the most commonly used MSN marker, but previous attempts to express this gene in vitro have failed. In this study we found that DARPP-32 is expressed in <12% of E13- or E17-derived striatal neurons when they are grown in defined media at high or low density in serum, dopamine, or Neurobasal/N2 (Life Technologies), and ARPP-21 is expressed in <1%. The percentage increases to 25% for DARPP-32 and 10% for ARPP-21 when the same cells are grown in Neurobasal/B27 (Life Technologies) for 7 d. After growth in Neurobasal/B27 plus brain-derived neurotrophic factor (BDNF) for 7 d, E13-derived MSNs are 53.7% DARPP-32-positive and 29. 0% ARPP-21-positive; E17-derived MSNs are 66.8% DARPP-32-positive and 51.5% ARPP-21-positive. The percentage of calbindin-positive neurons also is increased under these conditions. Finally, ARPP-21 expression is reduced in mice with a targeted deletion of the BDNF gene. We conclude that BDNF is required for the maturation of a large subset of patch and matrix MSNs in vivo and in vitro. In addition, we introduce a culture system in which highly differentiated MSNs may be generated, maintained, and studied.

Islet injury induces neurotrophin expression in pancreatic cells and reactive gliosis of peri-islet Schwann cells.

Pancreatic islets are enveloped by a sheath of Schwann cells, the glial cells of the peripheral nervous system (PNS). The fact that Schwann cells of the PNS become reactive and express nerve growth factor (NGF) and other growth factors following axotomy suggested the possibility that peri-islet Schwann cells could become activated by islet injury. To test this hypothesis, we examined two animal models of islet injury. The first model was mice and rats injected with streptozotocin (SZ), a specific beta-cell toxin. The second model was NOD mice, a strain in which beta cells are deleted by an autoimmune process. We found that peri-islet Schwann cells became reactive following islet injury and began to express increased levels of NGF and the neurotrophin receptor p75. Lesions to the pancreas also markedly induced NGF expression by exocrine and endocrine cells. Neurotrophin expression was not unique to adult tissues since pancreatic cells transiently expressed p75, the NGF receptor Trk A, and NGF during development. These observations suggest that NGF could play an important role in pancreas during embryogenesis and in processes leading to repair following islet injury in adults.

Brain-derived neurotrophic factor regulates maturation of the DARPP-32 phenotype in striatal medium spiny neurons: studies in vivo and in vitro.

The medium spiny neuron is the predominant striatal neuronal subtype. The striatum, a participant in motor and cognitive functions, is a site of pathophysiology in prevalent neuropsychiatric diseases and is the target of many currently utilized pharmacologic agents. DARPP-32, a dopamine and cyclic AMP-regulated phosphoprotein, is a widely-used marker of mature striatal medium-sized neurons, but the molecules regulating DARPP-32 transcription have not been identified. We show that a null mutation in the mouse brain-derived neurotrophic factor gene leads to decreased DARPP-32 immunoreactivity in striatal medium spiny neurons at birth and postnatal day 10. Striatal DARPP-32 messenger RNA and protein are decreased relative to wild-type littermate controls. In densely plated (1 x 10(6) cells/cm2) primary cultures derived from the ganglionic eminences, addition of brain-derived neurotrophic factor (100 ng/ml) to defined media results in a greater than 3-fold increase in the number of DARPP-32-immunopositive cells after 12 h and greater than 4-fold (P<0.005) after 24 h. The increase in DARPP-32-immunopositivity is abolished by the addition of 2 microg/ml actinomycin D without a significant effect on cell viability. These data suggest that brain-derived neurotrophic factor directly or indirectly regulates DARPP-32 transcription in medium spiny neurons. This is the first demonstration of transcriptional regulation of DARPP-32, and the first evidence of a forebrain abnormality in a newborn neurotrophin "knockout" mouse.

Enhanced serum response element binding activity correlates with down-regulation of c-fos mRNA expression in the rat brain following repeated cortical lesions.

Repeated lesions of rat cerebral cortex result in transient peaks in the level of the c-fos transcript, but after the second lesion, this peak is substantially diminished. Using this lesion paradigm, we have analyzed the participation of the c-fos promoter elements SRE and DSE in the regulation of c-fos transcription. Following a single lesion, SRE/DSE binding activity peaked at 2 h, subsequent to the maximal levels of c-fos mRNA and parallel to the peak of c-Fos protein. After a second lesion (reinduction), 4 h following the initial lesion, SRE/DSE binding activity peaked after only 30 min and was significantly higher than following the first lesion. Once again, this peak occurred after the peak of c-fos mRNA expression and parallel with the second peak of c-Fos protein expression. These results suggested that the SRE and DSE promoter elements participated in the induction and down-regulation of c-fos transcription in vivo and suggested the possible involvement of Fos protein in its own regulation. The ability of Fos/Fra proteins to participate in a transcriptional complex was confirmed in gel-shift experiments with an AP-1 element, and the biphasic trend of binding activity was observed. Supershift experiments were performed to directly determine whether Fos protein was participating in SRE and/or DSE transcriptional complexes. No alterations in the position or intensity of the shifted band were observed using Fos/Fra antiserum suggesting that Fos/Fra proteins could be involved in c-fos down-regulation through mechanisms other than direct participation in the SRE/DSE transcription complex.

GAP-43 mRNA expression in early development of human nervous system.

The temporal and spatial distribution of GAP-43 mRNA in early human development, from 6 to 23 gestational weeks (g.w.), was examined by in situ hybridization histochemistry. GAP-43 mRNA was expressed as early as 6 g.w. in all regions of developing nervous system, the spinal cord, brainstem, cerebellum, diencephalic and telencephalic regions. Although the pronounced level of expression persisted during the entire examined period, the intensity of expression varied along the spatial axis over time. Analysis at the cellular level revealed that early on in development (6 g.w.) GAP-43 mRNA was expressed in the entire neuroblast population. With the onset of differentiation, at 13-23 g.w., GAP-43 mRNA expression had switched to the neurons that are in the process outgrowth. The highest level of GAP-43 mRNA expression was localized in the regions consisting of differentiating neurons, such as the cortical plate and intermediate zone of the telencephalic wall, and several delineated subcortical and thalamic nuclei. The spatial and temporal pattern of GAP-43 mRNA expression obtained suggests a possible dual role of GAP-43 in the development of the human nervous system: in the embryonic brain it could be involved in fundamental processes underlying cell proliferation; in the fetal brain its expression is specifically correlated with differentiation and the outgrowth of axons.

ARPP-21: murine gene structure and promoter identification of a neuronal phosphoprotein enriched in the limbic striatum.

ARPP-21 (cAMP-regulated phosphoprotein, Mr = 21,000 on sodium dodecyl sulfate polyacrylamide gel electrophoresis) is a phosphoprotein highly enriched in concentration in the neurons of the limbic striatum. It is likely a third messenger in the intracellular cascade of events following neuronal stimulation by first-messenger activators of the adenylate cyclase system, including dopamine via the D1 receptor. ARPP-21 expression is restricted to telencephalic post-mitotic, post-migrational neurons, and its precise pattern of temporal and spatial expression makes it an attractive candidate for the study of transcriptional regulation of neuronal maturation. To define genomic regions likely to contain functional promoter elements, we isolated the murine ARPP-21 gene. Primer extension and T2 RNase protection analyses identified multiple transcription start sites, but 1.3 kb of 5'-flanking DNA revealed few consensus transcription factor binding sequences. A series of transient transfection assays in clonal cell lines which do not express ARPP-21 identified a basal promoter active in both neuronal and non-neuronal lines. Expression in all lines was decreased by the inclusion of regions further upstream, and extinguished by the inclusion of the first intron. Further analyses are likely to reveal cell specific regulatory sequences.

Desensitization of c-fos mRNA expression in rat brain following cortical lesions.

The effect of cortical lesions on the c-fos mRNA expression in adult rat brain was studied using in situ hybridization and Northern blot analysis. It was observed that the single cortical lesion evoked a rapid increase in the content of c-fos mRNA in whole brain. However, when rats received a second and third cortical injury 4 and 8 h after the first one, there was reduced elevation of c-fos mRNA after the second (60%) and third (75%) injuries in comparison with those seen after a single cortical lesion. These results represent the desensitization period during which the induction of c-fos mRNA by repeated cortical lesions was decreased. On the contrary, when animals were treated with a fourth cortical lesion 24 h after the first injury, with recovery for 16 h after the third one, the levels of c-fos mRNA were again elevated to a degree almost comparable to that seen after the single injury alone. The anatomical distribution of c-fos mRNA-expressing cells after cortical lesions revealed a large increase of hybridization to a heterogeneous population of hippocampal neurons including small cells of stratum granulosum in the dentate gyrus and larger cells within hippocampal stratum pyramidale. A dramatic labeling of neurons with c-fos riboprobe in frontoparietal and piriform cortex was also observed but with predominantly localized hybridization to c-fos mRNA on the damaged side of the cortex following the first and second cortical injury. However, hippocampal distribution of c-fos mRNA-containing cells was bilateral as a result of polysynaptic potentials evoked by cortical lesions.

Temporal and spatial preferences of c-fos mRNA expression in the rat brain following cortical lesion.

The expression of the proto-oncogene c-fos is increased in neuronal cells by a number of stimuli and the usefulness of this gene as a marker of neuronal activity has been demonstrated. The temporal and spatial expression of c-fos mRNA following the induction of a unilateral cortical lesion have been investigated in the rat brain by Northern blot analysis and in situ hybridization histochemistry. It was observed that the lesion evoked a rapid increase (20-fold) in the content of c-fos mRNA in the ipsilateral cortex, whereas in the contralateral cortex c-fos mRNA expression was more modest (7-fold). In the whole hippocampus a large and very rapid increase (17-fold) of c-fos mRNA expression was detected. The effect of a cortical lesion on Ca2+ uptake and membrane potential was also investigated. Using synaptosomes as a model system, we have provided evidence that Ca2+ entry via membrane depolarization increases in coordination with c-fos gene expression in neuronal cells. The principal conclusions from this study are that cortical lesions induce transient expression of the c-fos gene in specific neuronal cells of the rat brain.