We have shown that EWS/FLI is responsible for the expression of BCL11B in Ewing sarcoma cells. The dramatic decrease in BCL11B expression levels with EWS/FLI knock-down suggests that BCL11B is not normally expressed in the Ewing sarcoma cell of origin. BCL11B is, however, necessary for the proper development of many organs including the brain 
, skin 
, and teeth 
, as well as for terminal differentiation of T-cells 
. This allows for the possibility that in the likely primitive Ewing sarcoma cell of origin, the BCL11B promoter has an open chromatin structure that permits the strong activation of BCL11B following the formation of the translocation. At this time we consider BCL11B an indirect target of EWS/FLI as EWS/FLI did not show binding to the BCL11B promoter in our ChIP-chip data set 
. Furthermore previously defined EWS/FLI binding sites – GGAA repeats and ETS consensus sites – are absent from the BCL11B promoter/enhancer region. There are however variant ETS sites in BCL11B’s promoter/enhancer region that may allow for EWS/FLI binding and warrant further investigation.
This EWS/FLI-dependent up-regulation of BCL11B is necessary for the maintenance of the transformed phenotype in Ewing sarcoma cell lines in vitro
. This oncogenic capacity is in stark contrast to its tumor inhibitory function in leukemia and lymphoma – the cancers in which BCL11B have mainly been studied 
. Mutations or deletions of the BCL11B
gene are found in 9–16% of human T-ALL 
where it is thought to be a haploinsufficient tumor suppressor. Loss of one allele via the involvement of BCL11B in translocations in this malignancy as well as distinct mutations often found in the zinc finger region 
may contribute to the oncogenic process in part by preventing differentiation. Moreover in a mouse model of thymic lymphoma, spontaneous homozygous deletions and point mutations occurred in Bcl11b. This group also found that ectopic expression of Bcl11b in HeLa cells suppressed cell growth 
. BCL11B has not been widely studied in the context of solid malignancies and is not expressed in many 
The differing function of BCL11B in Ewing sarcoma does not appear to be due to a unique transcriptional activity in this tumor: BCL11B acts mainly as a transcriptional repressor in Ewing sarcoma, and thus acts similarly to what has been demonstrated in other cellular contexts. However, BCL11B regulates a unique set of genes in the A673 cellular background in comparison to other cell types in which it has been studied. For example, BCL11B has been implicated in cell cycle progression by directly repressing the cell cycle inhibitors p21WAF1 
and p57KIP2 
in microglial cells and SK-N-MC (which were originally characterized as a neuroblastoma cell line, but are in fact a Ewing cell line) cells, respectively. We were able to confirm downregulation of p57KIP2 transcript levels by BCL11B in SK-N-MC cells by qRT-PCR; however this was not observed in A673 or TC71 cells (ETW unpublished observation). At this time, we do not understand the differences among these cellular contexts that account for this discrepancy. Nonetheless, this discrepancy across Ewing sarcoma cell lines suggests that inhibition of p57KIP2 is not a central feature of BCL11B function in this tumor type. In developing T-cells BCL11B represses genes that allow for a more primitive state thus contributing to the differentiation process 
. These classes of genes were not observed in our genome wide analysis of BCL11B regulated genes in A673 cells. Our data suggests that BCL11B contributes to the overall EWS/FLI repressed gene signature in Ewing sarcoma, and that the repression of a subset of these genes may be necessary for the transformed phenotype. EWS/FLI directly represses approximately 5% of the total EWS/FLI repressed genes 
. This allows for a model where EWS/FLI up-regulates the expression of transcriptional repressors, such as BCL11B, which then indirectly account for the repression of the remaining 95% of the EWS/FLI down-regulated genes.
BCL11B facilitates transcriptional repression by recruiting a variety of chromatin modifying enzymes to the promoters of genes. BCL11B physically interacts with the histone methyltransferase SUV39H1 
, the histone demethylase LSD1 
, histone deacetylases HDAC1 and HDAC2 
, the class III histone deacetylase, SIRT1 
, as well as the NuRD co-repressor complex 
. BCL11B likely utilizes these interactions to mediate repression in both a cell-type specific and a promoter-specific fashion. We have shown that the chromatin remodeling activity of the NuRD co-repressor complex participates in the repression of the four BCL11B repressed genes investigated in this study. Our data further suggest that vorinostat-repressible HDAC activity is involved in this repression. These data are consistent with vorinostat inhibition of NuRD-associated HDACs, but inhibition of non-NuRD associated HDACs may also play a role. We have also demonstrated that the LSD1 inhibitor, HCI-2509, increases the expression of these genes. The mechanism for HCI-2509 is unclear due to the fact that BCL11B levels are also somewhat reduced. This small molecule has recently been shown to de-repress some EWS/FLI directly repressed target genes and show specific toxicity for Ewing sarcoma cells 
. The data presented here further demonstrate that HCI-2509 is able to reverse at least a portion of the EWS/FLI regulated gene signature which provides a possible mechanism for the toxicity to Ewing sarcoma cells.
We have not investigated the direct or indirect nature of this BCL11B mediated repression. In vitro
studies have identified a GC-rich BCL11B binding site 
; however, a genome-wide analysis of BCL11B binding in vivo
failed to confirm the in vitro
findings. Thus, a bona fide
BCL11B consensus site remains elusive. Without this information we were unable to inspect the promoters of our BCL11B regulated genes to identify potential direct targets. We did however perform motif enrichment analysis using the MEME suite 
on the BCL11B repressed gene list to identify any enriched sequence motifs. This analysis failed to show any significant sequence enrichment (ETW, unpublished observations). This may be expected due to the mixed set of directly and indirectly regulated genes.
We have shown that the re-expression of the BCL11B repressed gene, SPRY1, reduces the transformation potential of Ewing sarcoma cells. This further implicates BCL11B mediated repression as an important contributor to the repressed gene signature in Ewing sarcoma cells. At this time the mechanism involved in SPRY1’s ability to reduce transformation remains unclear, although it does not appear to be acting in its classic role by inhibiting RAS/MAPK signaling. Spry1 is also known to inhibit phospholipase C (PLC) activation, and during Xenopus mesoderm development, Xtsprouty inhibits the PLC pathway while still allowing for RAS/ERK signaling. SPRY1 may be impinging on an alternate growth factor signaling pathway or perform a novel function in the context of a Ewing sarcoma cell.
A recurring theme in Ewing sarcoma is alteration in the expression levels of important developmental genes. Various pathways involved in proper development and differentiation have been disrupted by EWS/FLI in Ewing sarcoma – sonic hedgehog 
, transforming growth factor beta (TGFB) 
, and WNT 
, among others. The timely expression of BCL11B during the development of many cell types is crucial for proper differentiation. Here we have shown the aberrant expression of BCL11B in Ewing sarcoma cell lines represses a subset of the EWS/FLI repressed gene signature and contributes to the transformed phenotype.