The PFKFB3 splice variant UBI2K4 is downregulated in high‐grade astrocytomas and impedes the growth of U87 glioblastoma cells

Aims: Fructose‐2,6‐bisphosphate, a key regulator of glycolysis, is synthesized and degraded by four different isozymes of 6‐phosphofructo‐2‐kinase/fructose‐2,6‐bisphosphatase (PFKFB1–4). The PFKFB3 isozyme is upregulated in several human tumours. Six alternatively spliced variants of PFKFB3 mRNA are known in humans (UBI2K1–6). Here, we studied the role of the PFKFB3 splice variants in human astrocytic gliomas. Methods: We analysed the PFKFB3 splice variants in 48 astrocytic gliomas by RT‐PCR and real‐time PCR. The effect of transient and stable overexpression of the PFKFB3 isoforms was studied in U87 glioblastoma cells by MTT, cell counting, clone formation assay and metabolic measurements. Results: UBI2K5 and UBI2K6 are the predominant splice variants in rapidly proliferating high‐grade astrocytomas while the expression of UBI2K3 and UBI2K4 is mainly restricted to low‐grade astrocytomas and nonneoplastic brain tissue. Overexpression of UBI2K5 or UBI2K6 in the U87 glioblastoma cell line enhances the glycolytic flux but does not affect cell growth. In contrast, overexpression of UBI2K4 reduces cell viability and anchorage‐independent growth of U87 cells. The UBI2K4 mRNA level is downregulated in astrocytic gliomas with increasing malignancy grade. Moreover, the UBI2K4 mRNA level correlates with growth rate of several human cancer cell lines derived from different tissue types. Conclusions: Our results demonstrate that the splice variant UBI2K4 impedes the tumour cell growth and might serve as a tumour suppressor in astrocytic tumours.     

Multistep expression and assembly of neuronal nicotinic receptors is both host-cell- and receptor-subtype-dependent

We tested the hypothesis that the folding, assembly and insertion of neuronal nicotinic receptors are critically dependent on the host cell line. We used recombinant adenoviruses encoding either the rat alpha7, alpha4 or beta2 subunits in which expression of the subunit is controlled by a tetracycline-dependent promoter to screen five cell lines (GH4C1, SH-EP1, CV1, SN-56, and CHO-CAR). All five lines do not express detectable nicotinic receptor but do express receptor for human adenovirus, and all expressed mRNA for alpha7, alpha4 and beta2 subunits when infected with viruses. Each cell line expressed varying levels of alpha4beta2 receptors that bound [3H]cytisine, but only the GH4C1 and SH-EP1 cell lines expressed either surface or internal alpha7 receptors that bound [125I]alpha-bungarotoxin ([125I]alpha-BGT). All five cell lines expressed a 60 kDa protein immunoblotted by anti-alpha7 antibodies when infected with the alpha7 virus, presumably representing unassembled alpha7 subunits. In addition, GH4C1 cells expressed over 10-fold more surface alpha7 receptor than SH-EP1 cells, even though the total alpha7 receptor in the two cell lines was similar. Sedimentation experiments indicate that SH-EP1 cells only partially assemble alpha7 receptors compared with GH4C1 cells and control alpha7 from rat brain. These data suggest that not only is surface alpha7 receptor expression a multistep process, but that each step may involve cell-specific assembly factors.     

Neuronal distribution and functional characterization of the calcium channel alpha2delta-2 subunit

The auxiliary calcium channel alpha2delta subunit comprises a family of three genes, alpha2delta-1 to 3, which are expressed in a tissue-specific manner. alpha2delta-2 mRNA is found in the heart, skeletal muscle, brain, kidney, liver and pancreas. We report here for the first time the identification and functional characterization of alpha2delta-2 splice variants and their mRNA distribution in the mouse brain. The splice variants differ in the alpha2 and delta protein by eight and three amino acid residues, respectively, and are differentially expressed in cardiac tissue and human medullary thyroid carcinoma (hMTC) cells. In situ hybridization of mouse brain sections revealed the highest expression of alpha2delta-2 mRNA in the Purkinje cell layer of the cerebellum, habenulae and septal nuclei, and a lower expression in the cerebral cortex, olfactory bulb, thalamic and hypothalamic nuclei, as well as the inferior and superior colliculus. As the in situ data did not suggest a specific colocalization with any alpha1 subunit, coexpression studies of alpha2delta-2 were carried out either with the high-voltage-gated calcium channels, alpha1C, alpha1E or alpha1A, or with the low-voltage-gated calcium channel, alpha1G. Coexpression of alpha2delta-2 increased the current density, shifted the voltage dependence of channel activation and inactivation of alpha1C, alpha1E and alpha1A subunits in a hyperpolarizing direction, and accelerated the decay and shifted the steady-state inactivation of the alpha1G current.     

Identification of a novel urokinase receptor splice variant and its prognostic relevance in breast cancer

The cellular receptor for urokinase-type plasminogen activator, uPAR, plays a central role in both cell surface-associated proteolysis and cellular adhesion. In the present study, we systematically searched for splice variants of uPAR mRNA in human cells and tumor tissues by qualitative RT-PCR using specific primers for uPAR exons 1 and 6. Beside the wild-type (wt) uPAR mRNA and the previously described splice variant lacking exon 5 (uPAR-del5), a novel splice variant lacking both exons 4 and 5 (uPAR-del4/5) was found predominantly in various cancer cell lines. To elucidate whether alternatively spliced uPAR mRNA may be translated and post-translationally processed, we generated stably transfected Chinese hamster ovary cells, which harbor expression plasmids of wt uPAR and various uPAR variants including uPAR-del5 and uPAR-del4/5. By ELISA, flow cytofluorometry, and Western blot analysis, we confirmed synthesis and secretion of wt uPAR and the uPAR variants by the use of domain-specific monoclonal antibodies against uPAR. For quantification of uPAR mRNA variants, we established two highly sensitive real-time RT-PCR assays based on LightCycler technology. Study of their expression in a representative set of breast cancer tissues indicated that the novel mRNA variant uPAR-del4/5 was expressed very frequently and independently of uPAR mRNA variants covering exon 4 (uPAR-wt and uPAR-del5). Higher uPAR-del4/5 expression was significantly associated with shorter disease-free survival (p = 0.0004) of breast cancer patients. These results suggest that uPAR-del4/5 mRNA may serve as a novel prognostic marker in breast cancer.     

Cloning, transport properties, and differential localization of two splice variants of GLT-1 in the rat CNS: implications for CNS glutamate homeostasis

At least two splice variants of GLT-1 are expressed by rat brain astrocytes, albeit in different membrane domains. There is at present only limited data available as to the spatial relationship of such variants relative to the location of synapses and their functional properties. We have characterized the transport properties of GLT-1v in a heterologous expression system and conclude that its transport properties are similar to those of the originally described form of GLT-1, namely GLT-1alpha. We demonstrate that GLT-1alpha is localized to glial processes, some of which are interposed between multiple synapse types, including GABAergic synapses, whereas GLT-1v is expressed by astrocytic processes, at sites not interposed between synapses. Both splice variants can be expressed by a single astrocyte, but such expression is not uniform over the surface of the astrocytes. Neither splice variant of GLT-1 is evident in brain neurons, but both are abundantly expressed in some retinal neurons. We conclude that GLT-1v may not be involved in shaping the kinetics of synaptic signaling in the brain, but may be critical in preventing spillover of glutamate between adjacent synapses, thereby regulating intersynaptic glutamatergic and GABAergic transmission. Furthermore, GLT-1v may be crucial in ensuring that low levels of glutamate are maintained at extrasynaptic locations, especially in pathological conditions such as ischemia, motor neurone disease, and epilepsy.     

Expression of neurexin Ialpha splice variants in sympathetic neurons: selective changes during differentiation and in response to neurotrophins

Neurexins are a surprisingly diverse group of alternatively spliced proteins possibly involved in neural cell recognition processes. We find neurexin Ialpha and its splice variants highly conserved between mammals and birds. In vivo, neurexin Ialpha is expressed in sympathetic neurons during target innervation and relative expression levels of splice variants change with development. In vitro, no such changes are observed in the absence of growth factors, indicating that interactions with the environment are required to modify the splicing pattern. Specific alterations in splice variant expression are induced in vitro by neurotrophins. Expression patterns of splice variants in vivo and neurotrophin-induced regulation without changes in cell composition in vitro demonstrate that neurexin splice variant expression varies during differentiation of individual neurons. Our data suggest that changes in neurexin splice variants contribute to alterations of neuronal cell surface properties during target innervation.     

Tenascin mRNA isoforms in the developing mouse brain

The extracellular matrix glycoprotein tenascin is expressed in the developing mouse cerebellum as a group of four protein species of different molecular weights. The difference is most likely due to alternative splicing which is known to occurr in tenascin mRNA isoforms that would account for this heterogeneity, tenascin splice variants were isolated from mouse brain by the polymerase chain reaction (PCR). In agreement with Northern blot analysis, amplification by PCR revealed a general decrease in tenascin mRNA expression during development from embryonic and early postnatal to adult stages. This decrease was more pronounced for isoforms of high molecular weight compared to those of low molecular weight. In accord with the observations at the protein level, four splice variants were found to be predominantly expressed, containing insertions of either six, five, or one fibronection type III repeat, or comprising no insertion. In addition, a minor splice variant with an insertion of four fibronectin type III repeats was isolated. Three of the isolated mRNA splice variants have not yet been described for mouse tenascin. Among them, an isoform containing six alternatively spliced repeats was found to include a novel fibronectin type III repeat. The sequence of this repeat displays 96.7% similarity to a corresponding type III repeat in human tenascin, revealing a strict evolutionary conservation between tenascin molecules from different species in the region of alternative splicing. Southern blot analysis of the amplified mRNA isoforms showed that the novel mouse type III repeat is confined to splice variants with an insertion of six fibronectin type III repeats. Furthermore, in situ hybridization on sections from mouse embryos indicated that tenascin-specific mRNAs containing the novel type III repeat are predominantly expressed in the central nervous system. © 1994 Wiley-Liss, Inc.     

Expression and identification of a new splice variant of neuroglycan C, a transmembrane chondroitin sulfate proteoglycan, in the human brain

Neuroglycan C (NGC) is a transmembrane chondroitin sulfate proteoglycan with an EGF module. We studied the expression of NGC in the human brain, mainly in the hippocampus, and confirmed some observations by conducting experiments using rat brain. In humans, NGC mRNA was expressed exclusively in the brain, especially in the immature brain. The telencephalon, including the hippocampus and neocortex, showed strong mRNA expression. NGC was immunolocalized to neuropils in the hippocampus and neocortex of the adult rat. RT-PCR experiments showed that four splice variants (NGC-I, -II, -III, and -IV) were expressed in the adult human hippocampus. By Western blotting, the expression as proteins of all splice variants except NGC-II was confirmed in the adult rat hippocampus. NGC-IV, which was first found in the present study, had the shortest cytoplasmic domain among the four variants. NGC-IV mRNA was expressed by neurons, but not by astrocytes, in culture prepared from the fetal rat hippocampus, suggesting that NGC-IV plays a role specific to neurons. In addition, the human NGC gene, which is registered as CSPG5, comprised six exons and was approximately 19 kb in size. In exon 2, a single nucleotide polymorphism resulting in Val188Gly in the NGC ectodomain was observed.     

Distribution of two splice variants of the glutamate transporter GLT-1 in rat brain and pituitary

We have performed immunocytochemistry on rat brains using a highly specific antiserum directed against the originally described form of the glutamate transporter GLT-1 (referred to hereafter as GLT-1alpha), and another against a C-terminal splice variant of this protein, GLT-1B. Both forms of GLT-1 were abundant in rat brain, especially in regions such as the hippocampus and cerebral cortex, and macroscopic examination of sections suggested that both forms were generally regionally coexistent. However, disparities were evident; GLT-1alpha was present in the intermediate lobe of the pituitary gland, whereas GLT-1B was absent. Similar marked disparities were also noted in the external capsule, where GLT1A labeling was abundant but GLT-1B was only occasionally encountered. Conversely, GLT-1B was more extensively distributed, relative to GLT-1alpha, in areas such as the deep cerebellar nuclei. In most regions, such as the olfactory bulbs, both splice variants were present but differences were evident in their distribution. In cerebral cortex, patches were evident where GLT-1B was absent, whereas no such patches were evident for GLT-1alpha. At high resolution, other discrepancies were evident; double-labeling of areas such as hippocampus indicated that the two splice variants may either be differentially expressed by closely apposed glial elements or that the two splice variants may be differentially targeted to distinct membrane domains of individual glial cells.     

Expression of prostaglandin H synthase-2 in human brain tumors

Numerous studies have demonstrated that prostaglandin H synthase-2 (PHS-2) is involved in gastrointestinal carcinogenesis, and that nonsteroidal anti-inflammatory drugs (NSAIDs), which inhibit PHS, can reduce the risk of colon cancer. In brain tumors, elevated prostaglandin production and its correlation to anaplastic grade of gliomas have been demonstrated. To determine whether the increased prostaglandin production is due to enhanced expression of PHS-2 and whether the up-regulation of PHS-2 has any correlation to histopathological findings in brain tumors, we evaluated the profile of PHS expression in several human glioma cell lines and surgical specimens from patients with various types of brain tumors. In glioma cell lines, five out of six cell lines showed constitutive expression of PHS-2, whereas PHS-1 was weakly expressed in all of them. All surgical specimens, except an ependymoma, which expressed both isozymes equally, expressed PHS-2 mRNA predominantly. Immunohistochemistry of various types of brain tumors, including six glioblastomas, nine astrocytomas, six meningiomas, five medulloblastomas, four craniopharyngiomas, three ependymomas, three neurinomas, two oligodendrogliomas, two malignant lymphomas, two dysembryoplastic neuroepitherial tumors and one metastatic brain tumor showed PHS-2 staining in most cases. In gliomas, astrocytomas (grade 2 and 3) were strongly stained, but the staining intensity of glioblastomas was relatively weak. Meningiomas and a metastatic brain tumor were also strongly stained. Our data thus suggest that most brain tumors express PHS-2, which may also play a role in tumorigenesis in the brain.     

Differential activities in adhesion and neurite growth of fibronectin type III repeats in the PTP-δ extracellular domain

The full-length extracellular domain (ECD) of protein tyrosine phosphatase delta (PTP-delta) functions as a ligand to promote cell adhesion and neurite outgrowth; this ECD contains three immunoglobulin (Ig) repeats and eight fibronectin type III (FN III) repeats. However, it is not known which regions of the ECD regulate its ligand functions. Therefore, we constructed and expressed a fusion protein of the PTP-delta ECD lacking FN III repeats 4-8, and tested this protein for neuronal adhesion and neurite-promoting ability. Compared to the full-length isoform, the truncated ECD was poorer at promoting adhesion, but a more potent promoter of neurite growth. The results suggest that distal FN III repeats of PTP-delta are important in adhesive functions, but dispensable for neurite outgrowth promotion. As the predominant isoform of PTP-delta during neural development (type D) also lacks distal FN III repeats, the functional properties we observe may be relevant to periods of axon extension, suggesting that splice variants of receptor PTPs play distinct roles in neural development.     

Astrocytes express specific variants of CaM KII delta and gamma, but not alpha and beta, that determine their cellular localizations

Multiple isoforms of type II Ca(2+)-calmodulin-dependent kinase (CaM KII) are composed of two major neuron-specific subunits, designated alpha and beta, and two less well-characterized subunits that are also expressed in non-neuronal tissues, designated delta and gamma. Regulated expression of these 4 gene products, and several variants produced by alternative splicing, shows temporal and regional specificity and influences intracellular targeting. We used immunoblotting and RT-PCR to analyze subunit and variant expression and distribution in cultured cerebellar astrocytes and neurons, and whole cerebellar cortex from rodent brain. The data indicate that: (i) astrocytes express a single splice variant of delta, namely delta(2); (ii) like neurons, astrocytes express two forms of CaM KII gamma; gamma(B) and gamma(A); (iii) these CaM KII variants are enriched in the supernate fraction in astrocytes, and the particulate fraction in neurons; (iv) unlike neurons, astrocytes do not express detectable levels of alpha or beta subunits or their respective splice variants. The results indicate that neurons and astrocytes express distinct CaM KII subunits and variants that localize to distinct subcellular compartments and, by inference, exert distinct cellular functions.     

Evidence that a single nucleotide polymorphism in the promoter of the G protein receptor kinase 3 gene is associated with bipolar disorder

In a genome-wide linkage survey, we have previously shown evidence suggesting that the chromosome 22q12 region contains a susceptibility locus for bipolar disorder (BPD). Two independent family sets yielded lod scores suggestive of linkage at markers in this region near the gene G protein receptor kinase 3 (GRK3). GRK3 is an excellent candidate risk gene for BPD since GRK3 is expressed widely in the brain, and since GRKs play key roles in the homologous desensitization of G protein-coupled receptor signaling. We have also previously shown GRK3 expression to be induced by amphetamine in an animal model of mania using microarray-based expression profiling. To identify possible functional mutations in GRK3, we sequenced the putative promoter region, all 21 exons, and intronic sequence flanking each exon, in 14-22 individuals with BPD. We found six sequence variants in the 5'-UTR/promoter region, but no coding or obvious splice variants. Transmission disequilibrium analyses of one set of 153 families indicated that two of the 5'-UTR/promoter variants are associated with BPD in families of northern European Caucasian ancestry. A supportive trend towards association to one of these two variants (P-5) was then subsequently obtained in an independent sample of 237 families. In the combined sample, the P-5 variant had an estimated allele frequency of 3% in bipolar subjects, and displayed a transmission to non-transmission ratio of 26 : 7.7 (chi(2)=9.6, one-sided P value=0.0019). Altogether, these data support the hypothesis that a dysregulation in GRK3 expression alters signaling desensitization, and thereby predisposes to the development of BPD.