Jamal Nasir

Expression profiling in cancer using gene chips

Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin’s lymphoma, accounting for at least 25 000 cases annually. Clinical presentation of DLBCL is variable and only 40% of patients respond well to chemotherapy. Now, using gene chip technology, Alizadeh and colleagues provide a molecular characterisation of DLBCL, which enables a tidy classification of this clinically heterogeneous condition reflecting differences in ontogeny, clinical presentation and survival in response to treatment. DNA microarray technology can provide a snapshot of gene expression in a given tissue, allowing the expression of tens of thousands of genes to be monitored on a grid not much more than the size of a postage stamp. In practice, large numbers of cDNA clones are spotted on a grid, which is hybridised with fluorescently labelled probes derived from mRNA from a given sample. A control probe, usually derived from normal tissue and labelled with a different dye, is also included. The colour and intensity of the fluorescence indicate whether a gene is down-regulated or over-represented in the experimental sample. Alizadeh and colleagues produced a ‘lymphochip’ with 17 856 cDNA clones derived from a variety of sources and representing genes known to be expressed in lymphoid cells or known to play an important role in immunology or cancer. The three most common type of adult lymphoid malignancies were examined: DLBCL, follicular lymphoma (FL) and chronic lymphocytic leukaemia (CLL). Ninety-six separate samples were analysed, including normal lymphocyte subpopulations as well as lymphoid cell lines propagated under a range of ‘activation’ conditions to provide a framework for the interpretation of the results based on approximately 1.8 million readings from 128 lymphochips. A hierarchical clustering algorithm was used to group genes according to expression patterns across all samples and, similarly, tumour and cell samples were grouped according to similarities in gene expression patterns. The data were represented in matrix format with each column corresponding to a sample and each row a cDNA (Fig. 1). The expression level of each gene relative to median expression levels across all samples is represented by red if expression is greater than the mean and by green if it is less. The intensity of the colour gives an impression of the level of expression relative to the mean. Distinct clones representing the same gene were found clustered in adjacent rows and where different tissues from the same patient were analysed, they were found clustered in adjacent columns, showing experimental noise is minimum. Overall, large sets of genes show coordinated expression patterns in related biological samples to produce a unique ‘signature’, which might represent the cell type in which the genes are expressed (for example, T-cells) or the biological process in which they are thought to function (for example, proliferation). One of the clearest distinctions between the three B-cell malignancies arose with respect to cell proliferation signature, which is associated with a diverse variety of genes including those involved in cell-cycle control, cell-cycle checkpoint, DNA synthesis and DNA replication. FL and CLL were found next to resting B cells, reflecting their relatively low proliferation rate. On the other hand, DLBCL displayed higher expression of genes associated with the cell proliferation signature, although there was some variation between individual samples, which is consistent with observed variability in proliferation index in DLBCL. DLBCL could be further divided into two distinct subtypes, with separate expression profiles, corresponding to different stages in B-cell development. The group expressing genes characteristic of germinal centre B cells (GC B-like DLBCL) expressed cell surface markers such as CD10 and CD38 plus a variety of genes previously associated with translocation events including BCL-6, BCL7A and LMO2. The DLBCLs expressing genes normally induced during in 6itro activation of peripheral blood B cells (activated B-like DLBCL)

To be or not to be an aggregate.

Establishing the molecular basis for disorders associated with expanded triplet repeats within genes continues to be a thriving if not a fashionable area of research, attracting some of the biggest names in molecular biology. In particular, disorders associated with polyglutamine encoding, long CAG repeat tracts, have attracted considerable attention. These include Huntington disease (HD), dentatorubral pallidoluysian atrophy and spinocerebellar ataxia (SCA) types 1, 2, 3, 6, and 7. Each of these disorders is associated with a distinctive pattern of neurodegenerative changes. Yet, they seem to be united by a common pathological mechanism that acts at the protein level, and involves a ‘gain of function’ due to the presence of excessively long polyglutamine tracts within their respective proteins. Furthermore, polyglutamine diseases appear to be associated with intranuclear inclusions comprising aggregates of protein filaments. Based on evidence for proteolytic cleavage and accumulation of polyglutamine containing truncated proteins within these aggregates, a coherent molecular model has emerged. It is thought that cleavage of mutant protein within the cytoplasm promotes the entry of cleaved products into the nucleus where they tend to aggregate, become ubiquitinated, and produce inclusions which culminate in cell death. However, a bold and remarkable series of experiments by Klement et al. and Saudou et al. are beginning to question the role of aggregates in disease (see Fig. 1). Ataxin-1, the protein product of the SCA-1 gene, contains a nuclear localization signal (NLS). In order to assess the importance of nuclear localization, Klement et al. generated transgenic mice expressing full length ataxin-1 with 77 CAG repeats, but a mutated NLS. The presence of the mutated NLS resulted in a redistribution of ataxin1 to the cytoplasm. Although the mice express atrophin-1 at similar levels in the cell to a previously established SCA-1 mouse model, there were no signs of nuclear inclusions or phenotypic effects, due to the absence of the protein in the nucleus. In a follow-up experiment, Klement et al. removed the self-association domain of ataxin-1 to generate transgenic mice expressing mutant protein. In this study, no aggregates were found within the nucleus, but the mice showed pathological changes as early as 3 weeks, and went on to develop ataxia, indicating that aggregate formation is not a requirement for disease. However, the 122 amino acid deletion is relatively large and it is impossible to predict whether the deletion itself might have an impact on the phenotype, or whether the presence of a protein containing an extended polyglutamine tract within Purkinje cells is able to confer toxicity by utilizing an alternative cellular pathway that does not involve aggregate formation. Saudou et al. describe an in 6itro model using a striatal cell line expressing a variety of HD truncated constructs with both normal length (17 CAG) as well as expanded (68 CAG) repeats. This in 6itro model recapitulates many of the features of HD. As expected, polyglutamine induced intranuclear aggregate formation and apoptotic cell death were observed. Furthermore, only the medium spiny enkephalin-positive neurons succumbed to cell death, but neurons in the hippocampus were spared. Saudou et al. extended their study to explore the effect of a variety of anti-apoptotic reagents and other compounds as potential therapeutic agents. Cell death was blocked with the small molecule, Ac-DEVD-CHO, a caspase-3 inhibitor, and by co-transfection with the BclXL, a anti-apoptotic gene. Addition of brain derived growth factors, BDNF and CNTF, also suppressed cell death. Using a nuclear export signal Saudou et al. were also able to transport the HD protein out of the nucleus and suppress the formation of nuclear inclusions and the induction of cell death, thus supporting the observation of Kelment

Antibody therapy for Alzheimer's disease

Alzheimer’s disease (AD) is a leading cause of dementia and costs the USA alone a staggering 40 billion dollars annually in terms of nursing homes and ancillary care. This disease of insidious onset and progressive deterioration affects cognitive functions, resulting in gradual loss of memory and intellectual decline. Up to one in ten people over the age of 80 are likely to suffer from AD, for which there is no effective cure. AD is associated with characteristic neuropathological changes, involving loss of neurons in the hippocampus and the neocortex, accumulation of intracellular protein deposits termed neurofibrillary tangles and extracellular protein deposits known as amyloid or senile plaques, surrounded by misshapen nerve terminals (dystrophic neurites). A major constituent of these plaques is the 42 amino acid amyloid-b peptide (Ab42) produced by cleavage of the b-amyloid precursor protein (APP). The role of APP is supported by genetic studies, which have identified mutations in this gene leading to increased production of Ab42. Schenk et al. have taken advantage of a transgenic mouse model which over-expresses Ab42, to test whether immunisation with the peptide can alleviate symptoms of the disease. Mice were treated at 6 weeks of age with synthetic human Ab42 (see Fig. 1), before the onset of significant plaque pathology. The control group was immunised with PBS buffer. Mice received monthly immunisations over an 11-month period. Eight of the nine mice immunised with Ab42 had serum antibody titres against Ab42 of greater than 1:10 000. Cross-reactivity of the antisera against the mouse amyloid-b peptide was also observed, although the titres were an order of magnitude lower. Mice were sacrificed at 13 months and quantitative immunohistochemistry was deployed to determine the extent of amyloid-b burden, the prevalence of neuritic plaques, astroglyosis and microglyosis. Quantitative image analysis revealed no detectable amyloid-b deposits in the brains of seven out of nine mice, including one with the lowest anti Ab titre. Similarly, there was almost a complete reduction in dystrophic neurites and astrocytosis, as measured using an APP specific monoclonal antibody and GFAP immunohistochemistry, respectively. The authors claim the process of plaque formation has essentially been disrupted, although the production of Ab peptide is apparently not affected. However, it is not clear whether plaque formation is inhibited by preventing the deposition of Ab42 and/or enhanced clearing of it through monocytic or microglial cells. The absence of neuritic and gliotic changes would suggest that Ab42 mice never developed neurodegenerative lesions that typify the progression of AD-like pathology in these mice. The absence of astrocytes also suggests preventing b-amyloidosis does not in itself cause damage to the neuropil. The authors also sought to investigate whether immunisation would be effective against older mice, which had already received a significant Ab plaque burden. For this purpose, they took 11month-old mice with Ab deposition of similar extent to that which is seen in AD. These mice were immunised repeatedly, as before, and similar titre responses were noted. Half of the group was sacrificed at 15 months (after 4 months treatment) and the remainder sacrificed at 18 months. Compared to the control group, the amyloid-b burden was significantly reduced at both 15 months (96% reduction) and 18 months (99% reduction). These data clearly demonstrate that pre-existing amyloidb deposits were cleared. The progression of neuritic pathology was also significantly reduced in the frontal cortex of immunised animals at both 15 months (84% reduction) and 18 months (55% reduction). In addition, reactive astrocytosis was reduced in the retrosplenial cortex at both 15 months (down 56%) and 18 months (down 34%). Clearly, immunisation with the Ab peptide is effective in both preventing and clearing amyloid plaques. However, a major question mark remains as to whether or not this is accompanied by allevi-