Affymetrix SNP Arrays Application Note II II Copy-neutral Loss of Heterozygosity in Cancer Loss of heterozygosity (LOH) is a com - Introduction oncogenes, unmasking mutated tumor sup - mon contributor to tumorigenesis, lead - pressor genes or contributing to patho - ing to the loss of a wild-type allele and Cancer samples can exhibit chaotic genicity as a result of altered gene expres - the unmasking of a recessive mutation. genomes. By definition, cancer results from sion due to imprinting. Scans of genomic copy number (CN) an accumulation of genetic alterations that can reveal LOH due to hemizygous dele - lead a cell population from initiation Copy-neutral LOH events cannot be tions, but LOH can also occur independ - through promotion and then progression. detected when scanning cancer genomes for ently of CN change, where one chromo - These genetic alterations include subtle CN alone, but they can be detected when some or chromosomal region has been changes, such as small gains and losses and viewing CN in parallel with LOH or viewing duplicated and its homologue has been nucleotide substitutions, and more conspic - allele-specific CN. Although the majority of deleted. When LOH occurs without CN uous alterations, such as changes in chro - cancer genome screens have focused on CN change, it is commonly termed copy- mosomal copy number (CN), translocations alone, recent studies combining CN and neutral LOH. and high-level amplifications. These are the LOH detection in a single experiment, using cause and effect of impairments in cell cycle Affymetrix Genome-Wide Human SNP The Affymetrix ® Genome-Wide Human regulation leading to errors in replication, Arrays, have revealed a growing list of cancer SNP Array 6.0 provides industry-leading recombination and cell division. types that present frequent and recurring CN detection with more than 1.8 million copy-neutral LOH. markers, including more than 900,000 Cancer research on the DNA level has his - SNPs for LOH identification.The ability torically emphasized CN profiling through This Application Note presents a collec - of Affymetrix SNP arrays to combine CN cytogenetic techniques and focused molecu - tion of recent articles demonstrating the and LOH detection for the identification lar analyses, such as microsatellite PCR. Yet importance and relevance of copy-neutral of copy-neutral LOH is changing the par - significant changes in the genome can occur LOH in cancer. adigm for analyzing chromosomal without changes in chromosomal CN, and changes in different cancer types. the scope of molecular analyses has not Publications Without LOH detection, traditional bacte - offered a feasible approach to genome-wide rial artificial chromosome (BAC) or com - observations of these events. LOH AND FOLLICULAR LYMPHOMA parative genomic hybridization (CGH) arrays from other commercial providers Copy-neutral loss of heterozygosity Follicular lymphoma (FL) is a common type only provide half of the picture. (LOH) represents one example of a genom - of non-Hodgkin’s lymphoma that originates ic abnormality in which no net change in from B-cell lymphocytes and is mostly exclu - This Application Note describes the fre - CN occurs, yet the abnormality can con - sive to adults. The majority of FL cases pres - quency and relevance of copy-neutral tribute to tumorigenesis. Copy-neutral ent a t(14;18) translocation resulting in the LOH in a variety of cancer samples, and LOH can occur due to duplication of one constitutive over-expres sion of an altered Bcl- presents a review of recent publications chromosome segment along with loss of 2 that blocks apoptosis. While knowledge of in which Affymetrix SNP arrays were the corresponding homologous region, so genomic changes beyond this translocation used to simultaneously study CN and that the cell retains two copies derived has been limited, and prognosis based on LOH.These publications identify signifi - from one parental source and no copies genomics is not yet available for this patient cant and common chromosomal aberra - derived from the other parental source. The cohort, a few CN abnormalities have been tions that cannot be identified when acquired homozygosity can contribute to associated with the disease, including del6q, examining CN alone. tumorigenesis by activating potential del1p32-36, +7, +12 and +X. AFFYMETRIX ® PRODUCT FAMILY > ARRAYS Charles W. Ross 1 and colleagues used Figure 1: Each row represents one sample; the x-axis represents chromosome position. (A) Affymetrix SNP arrays 2 to simultaneously LOH is indicated by blue for 46 samples, along a stretch of chromosome 1p. (B) Copy number detect CN and genotype in 46 FL samples is indicated on a scale from red (gain) to blue (loss) for 58 samples. CN for 46 of these samples with paired CD3+ T cell controls. This was is aligned horizontally with LOH predictions from the same sample. Red arrows = samples pre - senting copy-neutral LOH; black arrows = samples presenting LOH with CN loss. the first time that researchers have examined FL samples simultaneously for CN and LOH on a genome-wide level. A number of regions were consistently affected by gains and dele - tions, as previously seen, but this experiment also identified a number of regions presenting LOH where CN was unaffected. On chromosome 1p, for example, a region containing p73 and CASP9 exhibit - ed a high frequency of LOH with and with - out deletion (Figure 1). In total, LOH at chromosome 1 was observed in 50 percent of the samples, becoming the second most frequent genetic lesion ever described in FL. The majority of these LOH events were examples of copy-neutral LOH, with 15 samples exhibiting no CN change and eight samples harboring deletions. By detecting CN alone, only the eight events, 17 percent of all samples, would have been detected, and the significance of this region would have been underestimated. Interestingly, chromosome 1p LOH was the only genetic lesion significantly over- represented in FL grade 2 as compared with ALLELIC IMBALANCE AND MYELOPROLIFER - tumor heterogeneity, where CN or LOH grade 1 samples, suggesting the potential ATIVE DISEASE events may only affect a minor population differential importance of LOH, including within the whole tumor. Because genotyp - copy-neutral LOH, between these grades. Myeloproliferative disease (MPD) is charac - ing, by definition, assigns a single genotype terized by excess production of cells in the across the whole tumor sample, it was not Chromosome 6p also exhibited a high bone marrow. This group of diseases is effective at detecting LOH in minor popula - recurrence of copy-neutral LOH, with 30 divided into four main groups: chronic tions of these MPD samples. percent of FL samples presenting LOH in myelogenous leukemia (CML), which con - the affected area. Eighty-six percent of these tains the Philadelphia chromosome, and Instead, the authors took advantage of events were without CN change, indicating three diseases without this translocation. allele-specific CN detection, which pro - a prevalence of 6p copy-neutral LOH. These include Polycythemia vera (PV), vides a CN value for separate alleles at a Interestingly, copy-neutral LOH at this site Essential thrombocytosis (ET), and given SNP. When 100 percent of cells have co-occurred with LOH at chromosome 1p. Myelofibrosis (MF). A clonal mutation of LOH, allele-specific CN values would be JAK2 tyrosine kinase (V617F) occurs with “0 and 2” for the two SNP alleles. During The authors comment, “Copy-neutral high frequency in patients with PV, ET and retention of heterozygosity, allele-specific LOH is not detectable using either con - MF, suggesting a common pathogenesis for CN values would be “1 and 1” for het - ventional cytogenetics or array-CGH and, the diseases that is negative for the erozygous SNPs. The utility of allele-spe - therefore, has not been previously Philadelphia chromosome. cific CN in detecting LOH of mixed sam - described in FL.” The novelty of these ples is that CN values can fall between findings, therefore, is rooted in the fact Go Yamamoto 3 and coauthors used integers, such that a tumor in which 50 that adequate means to detect these types Affymetrix SNP arrays 4 to detect CN and percent of the cells exhibited LOH would of changes had simply not been applied to LOH across a sampling of MPD cases. The display allele-specific CN values of “0.5 this disease type. MPD samples were characterized by high and 1.5.” II II 2 In Figure 2, Yamamoto, et al. demonstrate Figure 2: Allelic imbalance in AML samples (A-C) and MPD samples (D-F) were detected. both the sensitivity of allele-specific CN to For each image, the top panel of blue dots represents raw and unsmoothed CN values across the mixed populations and the significance of chromosome. The second panel showing a dark blue line is smoothed CN. Below the cytoband, copy-neutral LOH events identified by this green notches represent heterozygous SNP genotype calls and pink notches represent conflicting genotype calls between paired tumor and normal samples. On the bottom panel, red and green lines method. In Figures 2A and 2B, the same pri - represent smoothed allele-specific CN for each of two SNP alleles. The blue bar at the bottom indi - mary acute myeloid leukemia (AML) speci - cates regions of LOH while the pink bar at the bottom indicates a gain. (A) AML sample compared men is studied using paired (Figure 2A) or to paired normal displays copy-neutral LOH on chromosome 17p in a portion of the tumor population. unpaired (Figure 2B) references. It is clear that (B) The same AML sample displays the same copy-neutral LOH but in an unpaired analysis to anony - the whole of chromosome 17 remains diploid, mous reference samples. (C) Residual tumor component is identified in the bone marrow sample that had been used as the paired reference in part A.