Global gene expression profiling of cHL samples and control samples. (A) The hierarchic clustering of all tissue samples analyzed using cDNA microarrays. The resulting dendrogram includes 63 lymph node tissue samples from cHL patients (white boxes), 5 from histiocyte/T cell–rich B-cell lymphoma (H/TCRBCL) patients (orange boxes), and 5 from benign lymphadenitis patients (green boxes). This clustering is based on the mRNA expression levels of 10 394 probe sets retained after a filtering process that removed probe sets with low and poorly measured expression as defined by an expression value inferior to 100 units in all 73 samples. Three major clusters of samples are observed and designated as I, II, and III. The heterogeneity of cHL samples is obvious. H/TCRBCL and benign lymphadenitis samples are grouped in cluster III and seem more homogeneous at the transcriptional level than cHL samples, although that could be due to their comparatively small number. (B) Hierarchic clustering restricted to cHL samples (n = 63) and based on 6229 probe sets with significant variation in mRNA expression levels across these samples. Each row represents a gene and each column represents a sample. The 2 separated color matrixes on the right correspond to the expression profile of control tissue samples (5 benign lymphadenitis, 5 H/TCRBCL samples, and 5 cell lines from left to right). Because these control samples are not considered in the clustering of cHL samples, genes are in the same order than in the major left matrix. The expression level of each gene in a single sample is relative to its median abundance across the 63 cHL tissue samples and is depicted according to a color scale (log₂ scale) shown at the bottom. Red and green indicate expression levels, respectively, above and below the median. The magnitude of deviation from the median is represented by the color saturation. The dendrogram of cHL tissue samples (above matrix) represent overall similarities in gene expression profiles, whereas colored bars on the right indicate the locations of 11 gene clusters of interest. Black dashes on the right represent the 501 probe sets associated with clinical outcome according to the univariate Cox analysis (“Outcome” GES). A zoomed view of panel B is shown in panel C, highlighting the dendrogram and gene clusters. In the dendrogram (C, top) of cHL tissue samples, 2 large groups are evidenced by clustering and delimited by an orange vertical line. Below the dendrogram, relevant characteristics of cHL tissue samples are represented according to a color ladder (gray when unclassifiable or unavailable): age (adult, white; children, black), clinical outcome (favorable, white; unfavorable, purple), Ann Arbor stage (I-II, white; III-V, black), histologic type (nodular sclerosis, blue; mixed cellularity, white), and EBV tumoral status (negative, white; positive, black). The control samples are colored as in panel A. The 5 cell lines are HUVEC (endothelial cells), HFFB (fibroblastic cells), L-428, KM-H2, and L-1236 (Reed-Sternberg cells) (from left to right). Expanded views of selected gene clusters corresponding to relevant cell types/function are named from top to bottom: “Reed-Sternberg cells” (dark blue bar), “interferon pathway and antiviral response (1)” (orange bar), “apoptosis” (2 clusters; light gray and black bars), “cell cycle” (dark gray bar), “B cells” (pink bar), “interferon pathway and antiviral response (2)” (red bar), “plasma cells” (green bar), “extracellular matrix” (light blue bar), and “histiocytes/T cells/innate immune response” (brown bar). The “cell metabolism” cluster (yellow bar in B) is not zoomed in panel C. The most relevant genes included in these clusters are indicated on the right by their EntrezGene symbol.³⁹