Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning.

Diffuse large B-cell lymphoma (DLBCL), the most common lymphoid malignancy in adults, is curable in less than 50% of patients. Prognostic models based on pre-treatment characteristics, such as the International Prognostic Index (IPI), are currently used to predict outcome in DLBCL. However, clinical outcome models identify neither the molecular basis of clinical heterogeneity, nor specific therapeutic targets. We analyzed the expression of 6,817 genes in diagnostic tumor specimens from DLBCL patients who received cyclophosphamide, adriamycin, vincristine and prednisone (CHOP)-based chemotherapy, and applied a supervised learning prediction method to identify cured versus fatal or refractory disease. The algorithm classified two categories of patients with very different five-year overall survival rates (70% versus 12%). The model also effectively delineated patients within specific IPI risk categories who were likely to be cured or to die of their disease. Genes implicated in DLBCL outcome included some that regulate responses to B-cell-receptor signaling, critical serine/threonine phosphorylation pathways and apoptosis. Our data indicate that supervised learning classification techniques can predict outcome in DLBCL and identify rational targets for intervention.

Combined array-comparative genomic hybridization and single-nucleotide polymorphism-loss of heterozygosity analysis reveals complex changes and multiple forms of chromosomal instability in colorectal cancers.

Cancers with chromosomal instability (CIN) are held to be aneuploid/polyploid with multiple large-scale gains/deletions, but the processes underlying CIN are unclear and different types of CIN might exist. We investigated colorectal cancer cell lines using array-comparative genomic hybridization (CGH) for copy number changes and single-copy number polymorphism (SNP) microarrays for allelic loss (LOH). Many array-based CGH changes were not found by LOH because they did not cause true reduction-to-homozygosity. Conversely, many regions of SNP-LOH occurred in the absence of copy number change, comprising an average per cell line of 2 chromosomes with complete LOH; 1-2 terminal regions of LOH (mitotic recombination); and 1 interstitial region of LOH. SNP-LOH detected many novel changes, representing possible locations of uncharacterized tumor suppressor loci. Microsatellite unstable (MSI+) lines infrequently showed gains/deletions or whole-chromosome LOH, but their near-diploid karyotypes concealed mitotic recombination frequencies similar to those of MSI- lines. We analyzed p53 and chromosome 18q (SMAD4) in detail, including mutation screening. Almost all MSI- lines showed LOH and/or deletion of p53 and 18q; some near-triploid lines had acquired three independent changes at these loci. We found consistent results in primary colorectal cancers. Overall, the distributions of mitotic recombination and whole-chromosome LOH in the MSI- cell lines differed significantly from random, with some lines having much higher than expected levels of these changes. Moreover, lines with more LOH changes had significantly fewer copy number changes. These data suggest that CIN is not synonymous with copy number change and some cancers have a specific tendency to whole-chromosome deletion and regain or to mitotic recombination.

Isolation and characterization of murine Cds (CDP-diacylglycerol synthase) 1 and 2.

Phototransduction in Drosophila is a phosphoinositide-mediated signalling pathway. Phosphatidylinositol 4,5-bisphosphate (PIP2) plays a central role in this process, and its levels are tightly regulated. A photoreceptor-specific form of the enzyme CDP-diacylglycerol synthase (CDS), which catalyzes the formation of CDP-diacylglycerol from phosphatidic acid, is a key regulator of the amount of PIP2 available for signalling. cds mutants develop light-induced retinal degeneration. We report here the isolation and characterization of two murine genes encoding this enzyme, Cds1 and Cds2. The genes encode proteins that are 73% identical and 92% similar but exhibit very different expression patterns. Cds1 shows a very restricted expression pattern but is expressed in the inner segments of the photoreceptors whilst Cds2 shows a ubiquitous pattern of expression. Using fluorescent in situ hybridization we have mapped Cds1 and Cds2 to chromosomes 5E3 and 2G1 respectively. These are regions of synteny with the corresponding human gene localization (4q21 and 20p13). Transient transfection experiments with epitope tagged proteins have also demonstrated that both are associated with the endoplasmic reticulum.

Large-scale analysis of the relationship between CYP11A promoter variation, polycystic ovarian syndrome, and serum testosterone.

CYP11A, the gene encoding p450scc, a key enzyme in steroid biosynthesis, is a strong biological candidate for polycystic ovary syndrome (PCOS) susceptibility. Four of the five published studies that have examined CYP11A for evidence of linkage and/or association have reported significant relationships with polycystic ovary (PCO) status and/or serum testosterone levels. However, study sizes have been modest, and the current study aimed to reevaluate these findings using significantly larger clinical resources. A pair of CYP11A promoter microsatellites, including the pentanucleotide (D15S520) previously implicated in trait susceptibility, were genotyped in 371 PCOS patients of United Kingdom origin, using both case-control and family-based association methods, and in 1589 women from a population-based birth cohort from Finland characterized for PCO symptomatology and testosterone levels. Although nominally significant differences in allele and genotype frequencies at both loci were observed in the United Kingdom case-control study (for example, an excess of the pentanucleotide four-repeat allele in cases, P = 0.005), these findings were not substantiated in the other analyses, and no discernable relationship was seen between variation at these loci and serum testosterone levels. These studies indicate that the strength of, and indeed the existence of, associations between CYP11A promoter variation and androgen-related phenotypes has been substantially overestimated in previous studies.

Candidate genes that determine response to salt in the stroke-prone spontaneously hypertensive rat: congenic analysis.

The existence of blood pressure quantitative trait loci exaggerated by salt on rat chromosome 2 has been confirmed previously using congenic strains derived from stroke-prone spontaneously hypertensive rats (SHRSP) and Wistar-Kyoto (WKY) rats. This study aimed to dissect the implicated chromosome 2 region and to identify candidate genes based on microarray expression profiling and real-time PCR. A marker-assisted breeding strategy generated congenic strains SP.WKYGla2a (D2Rat13-D2Rat157), SP.WKYGla2c* (D2Wox9-D2Mgh12), and SP.WKYGla2k (D2Mit21-D2Rat157) using SHRSP as the recipient and WKY as the donor strain. The SP.WKYGla2k strain contains a 10-cM congenic interval, which is encompassed within the larger (64-cM) SP.WKYGla2a congenic region. Salt-loaded systolic blood pressure, measured by radiotelemetry, was significantly lower in the SP.WKYGla2a and SP.WKYGla2k strains compared with SHRSP. Salt sensitivity in SP.WKYGla2c* was not significantly different from SHRSP. Exclusion mapping identified a 6-Mbp region harboring genes responsible for salt-sensitive blood pressure regulation. Microarray expression profiling was carried out in whole homogenized kidneys from parental and SP.WKYGla2a strains. Examination of expression data within the minimal congenic interval identified the positional candidates Edg1 and Vcam1, demonstrating significantly elevated renal RNA expression levels in the SHRSP compared with WKY and SP.WKYGla2a congenic strains. These results were confirmed by quantitative real-time PCR. DNA sequencing identified SNPs in both Edg1 and Vcam1 between SHRSP and WKY rats. In conclusion, we have identified a suggestive minimal interval encompassing a 6-Mbp region on rat chromosome 2. This region contains several physiological candidate genes for salt-sensitive hypertension in the SHRSP, including Edg1 and Vcam1, which are differentially expressed and lie on common and functionally important pathways.

Refining molecular analysis in the pathways of colorectal carcinogenesis.

BACKGROUND & AIMS: In the stepwise model, specific genetic and epigenetic changes accumulate as colorectal adenomas progress to carcinomas (CRCs). CRCs also acquire global phenotypes, particularly microsatellite instability (MSI) and aneuploidy/polyploidy (chromosomal instability, CIN). Few changes specific to MSI-low or CIN+ cancers have been established. METHODS: We investigated 100 CRCs for: mutations and loss of heterozygosity (LOH) where appropriate, of APC, K-ras, BRAF, SMAD4, and p53; deletion on 5q around APC and 18q around SMAD4; total chromosomal-scale losses and gains; MSI; and CIN. RESULTS: As expected, CIN- cancers had fewer chromosomal changes overall than CIN+ lesions, but after correcting for this, 5q deletions alone predicted CIN+ status. 5q deletions were not, however, significantly associated with APC mutations, which were equally frequent in CIN+ and CIN- tumors. We therefore found no evidence to show that mutant APC promotes CIN. p53 mutations/LOH were more common in CIN+ than CIN- lesions, and all chromosomal amplifications were in CIN+ tumors. CIN- cancers could be subdivided according to the total number of chromosomal-scale changes into CIN-low and CIN-stable groups; 18q deletion was the best predictor, being present in nearly all CIN-low lesions and almost no CIN-stable tumors. MSI-low was not associated with CIN, any specific mutation, a mutational signature, or clinicopathologic characteristic. CONCLUSIONS: Overall, the components of the stepwise model (APC, K-ras, and p53 mutations, plus 18q LOH) tended to co-occur randomly. We propose an updated version of this model comprising 4 pathways of CRC pathogenesis, on the basis of 5q/18q deletions, MSI (high/low), and CIN (high/low/stable).

Prediction of central nervous system embryonal tumour outcome based on gene expression.

Embryonal tumours of the central nervous system (CNS) represent a heterogeneous group of tumours about which little is known biologically, and whose diagnosis, on the basis of morphologic appearance alone, is controversial. Medulloblastomas, for example, are the most common malignant brain tumour of childhood, but their pathogenesis is unknown, their relationship to other embryonal CNS tumours is debated, and patients' response to therapy is difficult to predict. We approached these problems by developing a classification system based on DNA microarray gene expression data derived from 99 patient samples. Here we demonstrate that medulloblastomas are molecularly distinct from other brain tumours including primitive neuroectodermal tumours (PNETs), atypical teratoid/rhabdoid tumours (AT/RTs) and malignant gliomas. Previously unrecognized evidence supporting the derivation of medulloblastomas from cerebellar granule cells through activation of the Sonic Hedgehog (SHH) pathway was also revealed. We show further that the clinical outcome of children with medulloblastomas is highly predictable on the basis of the gene expression profiles of their tumours at diagnosis.