OCT2 (Octamer Binding Protein 2): An Essential Role in Germinal Center B Cell Differentiation and the “Achilles Heel” of Derived Lymphomas.

Abstract 362

To identify genes requiered for the proliferation and survival of Diffuse Large B Cell Lymphomas (DLBCL) we conducted an “Achilles Heel” RNA interference screen in cell lines model of ABC (Activated B cell-like) and GCB (Germinal Center B-cell like) DLBCL subtypes. One of the most toxic small hairpins RNAs (shRNAs) in this screen targeted Oct2, encoding a POU domain transcriptional activator. Unlike Oct1, which is constitutively expressed in many cell types, Oct2 is primarily lymphoid restricted. It was identified by virtue of its ability to bind the highly conserved DNA octamer motif (ATGCAAAT) within immunoglobulin (Ig) genes promoters. The B cell specific co-activator OCA-B interacts with the POU domain of the octamer binding proteins enhancing their transactivation potential.

Although Oct2 and OCA-B are not essential for Ig transcription, they are required for germinal center (GC) B cell differentiation.

To understand the massive apoptotic cell death of DLBCL cells following shRNA Oct2 induction we investigated the genetic pathways controlled by Oct2. We profiled gene expression changes in DLBCL cell lines after knocking down Oct2 and merged this data set with data from genome-wide assessment of Oct2 and OCA-B binding sites, coupling chromatin immunoprecipitation (ChiP) with high-throughput sequencing technologies (ChIP-Seq).

ChIP-Seq uncovered an extensive network of Oct2 target genes in DLBCL cells. More than 60% of the Oct2 target genes also showed OCA-B biding. This Oct2/OCA-B overlaping set of targets was enriched for genes selectively expressed in pan-B cells and GC B cells.

We found that Oct2/OCA-B lie upstream many of the main transcription factors known to play an essential role in inducing and mantaining the GC stage of B cell development such as BCL6, MTA3, PU.1, IRF8, SpiB and OCA-B, among others. Oct2/OCA-B target these genes both in DLBCL cells and in normal human primary centroblasts.

Strikingly, among Oct-2 downstream effectors, BCL6 cDNA was enough to rescue both ABC and GCB DLBCL cells from the Oct2 shRNA lethal effect.

The Oct2/OCA-B binding of BCL6 promoter was confirmed in vivo by single locus ChIP in different GCB-DLBCL cell lines as well as in primary centroblasts isolated from human tonsils. Gel shifts experiments showed Oct2 binding to more than one non canonical octamer motif within the BCL6 promoter. Furthermore, computational analysis of the BCL6 promoter region bound by Oct2, showed PU.1 binding sites. Knocked down of PU.1 decreased Oct2 enrichment and viceversa suggesting cooperative Oct2/PU.1 biding to BCL6 pomoter.

ChIP-Seq findings opened an entire and exciting new chapter in the Oct2 biology field. Pou transcription factors were suppossed to regulate the activity of octamer containing promoters. Interestingly, Oct2 binds to and control the expression of many GC specific genes that do not harbor a “canonical octamer” motif.

Eventhough Oct2 is expressed throughout the different stages of B cell maturation, both mRNA and protein levels are enhanced in centroblasts as compared to pre-GC B cells. We found Oct2 capable of inducing its own expression as well as Oct-1 and OCA-B. This autoregulatory circuit might partially account for the Oct2 predominant role in GC specific genes expression control.

By array cGH, high level amplification of Oct2 and OCA-B was found in less than 10% of non Hodgkin lymphoma patients samples. Nonetheless, these lymphoma cells become addicted to the Oct2 controlled network that sustain cell survival and proliferation, wich turns Oct2 into an attractive therapeutic target for non Hodking lymphoma patients. This critical Oct2 DLBCL cells dependency, is an example of “non oncogene addiction” that we recently defined based on our RNA interference screening. Oct-2 controls a network of regulatory relationships that sustain both normal GC B cells and malignant counterparts.

In summary, we showed that Oct2 and OCA-B lie upstream of BCL6, one of the critical regulators of germinal center B cell differentiation. This suggests that Oct2-directed therapy should kill the same DLBCLs as BCL6-directed therapy.

Furthermore, all GC and Post-GC B cells that were tested requiere Oct-2 for survival, indicating that Oct2-directed therapy might have a broder activity spectrum than the BCL6-directed therapy.

The Oct2/OCA-B binding interface would be amenable to attack with potential manageable toxicity, since this interaction is exclusively required in GC B cells.