Identification of Hodgkin and Reed-Sternberg cell-specific genes by gene expression profiling

Abstract

Hodgkin lymphoma (HL) is a malignancy of unknown pathogenesis. The malignant Hodgkin and Reed/Sternberg (HRS) cells derive from germinal center B cells (or rarely, T cells) but have a heterogeneous and largely uncharacterized phenotype. Using microarrays, we compared the gene expression profile of four HL cell lines with profiles of the main B cell subsets and B cell non-HLs to find out whether HRS cells, despite their described heterogeneity, show a distinct gene expression, to study their relationship to other normal and malignant B cells, and to identify genes aberrantly or overexpressed by HRS cells. The HL lines indeed clustered as a distinct entity, irrespective of their B or T cell derivation, and their gene expression was most similar to that of EBV-transformed B cells and cell lines derived from diffuse large cell lymphomas showing features of in vitro-activated B cells. Twenty-seven genes, most of which were previously unknown to be expressed by HRS cells, showed aberrant expression specifically in these cells, e.g., the transcription factors GATA-3, ABF1, EAR3, and Nrf3. For five genes, expression in primary HRS cells was confirmed. The newly identified HL-specific genes may play important roles in the pathogenesis of HL, potentially represent novel diagnostic markers, and can be considered for therapeutic targeting.

Figure 1

The gene expression profile of HL lines is related to that of LCL and ABC-type DLCL. Dendrogram showing the hierarchical clustering of gene expression data (see Methods for algorithm and criteria) generated from 23 transformed B cell lines derived from HL, DLCL, BL, and EBV-transformed peripheral blood B cells (LCL). Cell lines are color-coded according to their cellular origin: DLCL, red; BL, blue; HL, green; LCL, violet. If known, the EBV status is indicated in brackets. ABC- or GC-type DLCL-subtypes are indicated. The corresponding matrix is shown in supplementary Figure 1 (http://www.jci.org/cgi/content/full/111/4/529/DC1).

Figure 2

Identification of genes differentially expressed between HL lines, LCL, and ABC-type DLCL lines versus BL and GC-type DLCL. The gene expression profiles of four HL, five LCL, and two ABC-type DLCL lines that clustered together in the unsupervised analysis (see Figure 1) were compared with those generated from six BL and six DLCL (non–ABC-type) cell lines by supervised hierarchical clustering using Genes@Work (see Methods). Columns represent individual cell lines, and rows correspond to genes. Color changes within a row indicate expression relative to the average of the sample population. Values are quantified by the scale bar that visualizes the difference in the ζ-score (expression difference/SD) relative to the mean. Genes are ranked based on the z-score (mean expression difference of the respective gene between phenotype and control group/SD). The support value for supervised analysis was chosen as n = n₀ – 1, where n₀ is the number of cells in the given phenotype set, allowing for one unclustered sample per pattern in the phenotype set. The corresponding expression data for BCL-2 and BCL-6 were obtained by relaxing the criteria in the supervised clustering and are therefore shown separately at the bottom. Gene names and cell lines are indicated.

Figure 3

Identification of genes specifically expressed or downregulated in HL cell lines. Supervised cluster analysis using Genes@Work. Gene expression profiles of four HL cell lines were compared with four normal B cell subsets (five each of naive B cells, memory B cells, CBs and CCs), five LCL, seven DLCL cases, seven DLCL lines, four BL cases, eight BL lines, six FL, ten B-CLL. Matrices and gene ranking are as in Figure 2. The support value for supervised analysis was chosen as n = n₀, where n₀ is the number of cells in the given phenotype set, allowing for no unclustered sample per pattern in the phenotype set. Gene names are indicated. Twenty-seven distinct genes are significantly upregulated in the HL lines, and 45 distinct genes are downregulated (for one upregulated [IPL] and seven downregulated [Igα, Syk, Lck, CD20, PLCγ2, CD45, PTPN7] genes), each represented by two probes on the array; only the more significant probe is shown).