ETS1 encoding a transcription factor involved in B-cell differentiation is recurrently deleted and down-regulated in classical Hodgkin's lymphoma

The neoplastic B-cell derived Hodgkin and Reed-Sternberg cells (HRS) in classical Hodgkin's lymphoma (cHL) undergo a substantial loss of B-cell identity. It has been suggested that epigenetic mechanisms such as DNA methylation are involved in this cHL-associated reprogramming. 1 In a recent array-based DNA-methylation study using 27K methylation arrays, we identified 209 protein encoding genes hypermethylated in cHL but not in other germinal center derived B-cell lymphomas as compared to normal B cells. These genes, exclusively hypermethylated in cHL, are enriched for functions such as regulation of B-cell or T-cell activation.2

DNA methylation is an epigenetic mechanism associated with gene inactivation that can not only cause gene silencing de novo, but can also be acquired as a consequence of upstream gene downregulation3 and in cHL, a loss of a master transcription factor could lead to the methylation of genes involved in B-cell identity. We, therefore, tested the 209 genes hypermethylated exclusively in cHL that we had identified in our previous study for transcription factor binding sites.2 The most significant enrichment was seen for the transcription factors EBF1/OLF1 and E2A (E12/E47) which have both been previously shown to be involved in B-cell regulation and to be deregulated in cHL (Table 1).4-5 Binding sites for another key B-cell transcription factor, ETS1, ranked third among the group of genes hypermethylated exclusively in cHL. This led us to analyze the ETS1 gene for mutations, copy number alterations and expression both on mRNA and protein level in cHL cell lines and primary biopsies.

TABLE 1.Top 10 transcription factor motifs significantly over-represented in the group of genes hypermethylated in cHL.

We collected published gene expression microarray data for the L428, HDLM2, KMH2 and L1236 cHL cell lines as compared to normal B-cell entities (5 centroblasts, 5 centrocytes, 5 naive B cells, 5 memory B cells)6 and analyzed the expression of the transcription factors that were identified (Table 1). ETS1 was strongly downregulated with a 7.23-fold reduced expression in cHL cell lines (t-test P<0.001) (Figure 1A). This prompted us to analyze whether the observed downregulation of ETS1 is also present in primary cHL and results in reduced protein expression. In reactive tonsils, the nuclei of germinal center, mantle zone and parafollicular cells expressed the ETS1 protein as shown by immunohistochemistry whereas the squamous epithelium remained negative. In 8 of 17 (47%) evaluable cHL, the HRS cells were completely devoid of ETS1 protein expression whereas expression was confirmed in the nuclei of bystander cells (Figure 1B). In addition, 7 of 17 (41%) cHL presented a heterogeneous pattern of expression with both positive and negative HRS cells for ETS1 protein. Only in 2 of 17 (12%) cases did all HRS cells retain expression of ETS1 on a protein level.

Figure 1.(A) Significant downregulation of the ETS1 gene in cHL cell lines as compared to normal B-cell entities (5 × centroblasts, 5 × centrocytes, 5 × naive B cells, 5 × memory B cells) (Affymetrix U95 microarray tag 1518_at) prepared using (http://www.vertex42.com/ExcelTemplates/box-whisker-plot.html) and based on published profiles.6 (B) cHL case with complete absence of ETS1 protein expression in HRS cells (arrows) and positively stained bystander cells.

As ETS1 was not among the hypermethylated genes in cHL and, therefore, is itself not silenced by gene methylation, 2 we performed mutation screening for the entire coding sequence of this gene in 7 cHL cell lines. In the L540 cell line, an 11 bp heterozygous intronic deletion (NM_001143820.1:c.1123+5_1123+15delGCGCCCTCC CG; chr11:128,350,071-128,350,081; hg19) was identified. This deletion cannot be attributed to a known copy number variation as only a single nucleotide polymorphism (SNP) (chr11:128350071) is annotated for this location according to the University of California Santa Cruz (UCSC) Genome hg19 (http://genome.ucsc.edu) and the 1000 Genome (http://browser.1000genomes.org) data-bases. As the deletion is located 5 bp from the 3' splicing site of exon 8, we used RT-PCR for exons 6-10 to detect potential changes in splicing but this mutation was not seen to exert any effect. These findings suggest that the deletion identified does not affect ETS1 protein functionality. Neither mutations in the coding exons nor the adjacent intronic sites were identified in any other cell line.

Next, we performed a copy number analysis by fluorescence in situ hybridization (FISH) and identified heterozygous deletions of the ETS1 locus in 2 of 7 (29%) cHL cell lines, KMH2 and SUPHD1. Expression of the ETS1 transcription factor in KMH2 is strongly down-regulated in the gene expression microarray data.6 Interestingly, for SUPHD1, that was not included in this microarray study, we failed to amplify a transcript by real-time polymerase chain reaction (RT-PCR) indicating complete silencing of ETS1 in this cell line. The size of PCR products obtained for the other 6 cell lines was as expected.

In primary cHL biopsies, 10 of 16 (63%) informative cases analyzed by FICTION showed losses of the ETS1 locus, including one case with homozygous deletion (in 93% of HRS nuclei) and 2 cases with homozygous deletion in a sub-population of the HRS cells (in 28% and 30% of HRS nuclei, respectively). In 7 of 8 cases with ETS1 deletion, where FICTION and immunohistochemical data were available, complete or partial loss of ETS1 protein expression was observed. Only in one of 8 cases was no reduction in ETS1 protein expression observed showing good correlation between genomic loss and ETS1 downregulation. FISH analysis of 19 cytogenetic suspensions identified only 3 of 19 (16%) primary cases with deletion. This probably reflects the bias that arises from the difficulty of identifying HRS cells in cases not analyzed by FICTION.

Recently, non-synonymous mutations of ETS17 and copy number gains8 probably resulting in activation of the gene were identified in diffuse large B-cell lymphomas. In contrast, the heterozygous deletions of the ETS1 locus in HRS cells reported here can result in haploinsufficiency causing reduced expression of ETS1 protein and, consequently, of its target genes. Haploinsufficiency has already been described in hematologic malignancies; the PAX5 transcription factor is lost in approximately 30% of cases of acute lymphoblastic leukemia.9 It has previously been shown that ETS1 functions either as transcriptional activator or repressor, and it was reported to have oncogenic potential.10 Furthermore, several authors reported the involvement of ETS1 in hematopoiesis, including impaired B-cell development in mice with defective ETS1.11-12

According to these data, we propose a model in which the epigenetic downregulation of B-cell specific genes in cHL might be mediated by primary deregulation of upstream regulators of the B-cell program. We suggest ETS1 is a player in this process and recurrent deletions and loss of expression of this transcription factor in cHL might contribute to the escape potential and survival of HRS cells.

Acknowledgements

this study has been supported by the Deutsche Krebshilfe through project n. 107748 to RS, the network project “Molecular Mechanisms in Malignant Lymphomas” subproject M2 (70-3173-Tr3 to RS and JIM-S), the Ministerio de Ciencia e Innovación (SAF2009-08663 to JIM-S), the Wilhelm Sander Stiftung (2005.168.2 to RS) and the Kinderkrebsinitiative Buchholz/Holm-Seppensen (infrastructure to RS and WK). JIM-S is a Ramon y Cajal researcher of the Ministerio de Ciencia e Innovación. MG is supported by an FEBS Long-Term Fellowship and Support for International Mobility of Scientists fellowships of the Polish Ministry of Science and Higher Education. The authors thank Dr. Ralf Küppers for critical reading of the manuscript and the technical staff of the laboratories involved for their excellent assistance.

Footnotes

• The information provided by the authors about contributions from persons listed as authors and in acknowledgments is available with the full text of this paper at www.haematologica.org.
• Financial and other disclosures provided by the authors using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are also available at www.haematologica.org.