Abstract
A significant fraction of B cell non-Hodgkin lymphomas (B-NHL) of germinal center origin carry heterozygous missense mutations in FOXO1, a member of the FOXO family of transcription factors. FOXO1 is a central component of the PI3K signaling cascade engaged by the B cell receptor and is essential for B cell homeostasis and survival (Dengler et al, Nat Immunol 2008; Srinivasan et al, Cell 2009; Lin et al, Nat Immunol 2010). In response to PI3K activation, AKT phosphorylates FOXO1 leading to its nuclear-cytoplasmic translocation and inactivation. Missense mutations of the FOXO1 gene are detectable in germinal center (GC)-derived B-NHL, including ~12% of Burkitt Lymphoma (BL) and ~9% of Diffuse Large B Cell Lymphoma (DLBCL) cases (Schmitz et al, Nature 2012; Trinh et al, Blood 2013; Pasqualucci et al, Cell Rep 2014). The role of FOXO1 in normal GC development as well as the contribution of its mutations to lymphomagenesis is unclear.
We show that FOXO1 expression is restricted to the dark zone of GCs, where its nuclear localization is detectable in most B cells. Mice carrying the conditional inactivation of FOXO1 in GC B cells display normal GC in number and size. However, these GCs lack phenotypically defined (CXCR4hi/CD86lo) dark zones and are entirely composed by light zone B cells (CXCR4lo/CD86hi). FOXO1-/- GC B cells express AICDA and carry a normal number of mutations in their immunonoglobulin genes, but do not undergo affinity maturation, resulting in severely impaired antigen responses.
In order to identify the biological program controlled by FOXO1 in GC B cells, we identified candidate transcriptional target genes by integrating ChIP-seq and gene expression data. These analyses showed that that the establishment of the dark zone fate relies on a FOXO1-dependent transcriptional network that is enriched for genes involved in immune signaling cascades triggered by the B cell receptor and by a variety of cytokines controlling GC polarity. Notably, a majority of these target genes are co-bound and co-regulated, in a FOXO1-dependent manner, by BCL6, a well characterized GC master regulator.
To assess the role of BL- and DLBCL-associated mutations, we first investigated the subcellular localization of FOXO1 mutant proteins by transfecting wild type and mutant GFP-tagged FOXO1 alleles into HeLa cells. As previously shown (Trinh et al, Blood 2013), this analysis showed that mutant FOXO1 proteins, but not the wild-type one, readily localize in the nucleus. Analogously, immunofluorescence analysis of BL and DLBCL samples showed the presence of nuclear FOXO1 in all tumors carrying mutations in the FOXO1 gene. However, nuclear localization was also detectable in virtually all cases carrying normal FOXO1 genes. Accordingly, in vitro experiments testing the ability of normal and mutated FOXO1 proteins to respond to various signals activating the PI3K pathway in multiple BL and DLBCL cell lines, failed to display a correlation between the presence of mutations and responsiveness to these signals. Taken together, these results suggest that other mechanisms in addition to direct gene mutation are responsible for the constitutive nuclear localization of FOXO1 in tumors. We are now examining the consequences of FOXO1 missense mutations in vivo, by reconstituting FOXO1-/- GC B cells with FOXO1 mutants using bone marrow chimeras.
Disclosures
No relevant conflicts of interest to declare.
β-Catenin Activation in Hepatocellular Cancer: Implications in Biology and Therapy
