The untold stories of the speech gene, the FOXP2 cancer gene

Abstract

FOXP2 encodes a transcription factor involved in speech and language acquisition. Growing evidence now suggests that dysregulated FOXP2 activity may also be instrumental in human oncogenesis, along the lines of other cardinal developmental transcription factors such as DLX5 and DLX6 [1-4]. Several FOXP familymembers are directly involved during cancer initiation, maintenance and progression in the adult [5-8]. This may comprise either a pro-oncogenic activity or a deficient tumor-suppressor role, depending upon cell types and associated signaling pathways. While FOXP2 is expressed in numerous cell types, its expression has been found to be down-regulated in breast cancer [9], hepatocellular carcinoma [8] and gastric cancer biopsies [10]. Conversely, overexpressed FOXP2 has been reported in multiple myelomas, MGUS (Monoclonal Gammopathy of Undetermined Significance), several subtypes of lymphomas [5,11], as well as in neuroblastomas [12] and ERG fusion-negative prostate cancers [13]. According to functional evidences reported in breast cancer [9] and survey of recent transcriptomic and proteomic analyses of different tumor biopsies, we postulate that FOXP2 dysregulation may play a main role throughout cancer initiation and progression. In some cancer conditions, FOXP2 levels are now considered as a critical diagnostic marker of neoplastic cells, and in many situations, they even bear strong prognostic value [5]. Whether FOXP2 may further become a therapeutic target is an actively explored lead. Knowledge reviewed here may help improve our understanding of FOXP2 roles during oncogenesis and provide cues for diagnostic, prognostic and therapeutic analyses.

Keywords: FOXP2 factor; cancer; language; oncogene; prognosis.

Figure 1. DNA

HUMAN FOXP2 LOCUS (NCBI ID: 93986; Atlas ID: 40633). Location of the FOXP2 locus on the long arm of chromosome 7 band 7q31.1 (fragment, ENSEMBL coordinates GRCh38:CM000669.2). FOXP2 is located within a 5 Mb-large region of fragile genomic hotspots involving the highlighted neighboring genes.

Figure 2. RNA

HUMAN FOXP2 pre-mRNA structure. It encodes for 17 exons (blue segments), with exon 1 being the first in the predominant FOXP2 isoform. Bottom part: location of mutational variants assembled from different sources [19,131,176,177], including DECIPHER database.

Figure 3. PROTEIN

Human FOXP2 main protein isoform (Uniprot O15409). Structural and functional domains are highlighted. The Forkhead P2 domain harbors two nuclear localization signals (‘NLS’) [70]. Two major mutated variants are indicated above, with R328X interrupting the protein and R553H altering its subcellular localization (KE family verbal dyspraxia mutation). The two human lineage-specific aminoacids N303 and S325 are indicated in blue. The sumoylation site (K674) is indicated. A Q204Q substitution observed in multiple cancers is discussed in Figure 5.

Figure 4. Onco-diagnostic relevance of FOXP2 expression level, comparing immunophenotyping in cancer biopsies (top pannel) with mechanistic data (bottom pannel)

Compared regulations: red writing= different results; green writing= identical results.

Figure 5. PROTEIN

Hypothetic scenarios of oncogenic events involving a known regulatory element embedded in the fifth exon of FOXP2 encoding the polyQ40 stretch. This proposal stems from the observation that six different cancer types, with downregulated FOXP2 expression, share an identical point mutation at 7:114,629,945 (Q204) in the fifth exon of FOXP2. This scheme explores a few of the putative functional consequences of this mutation, considering the observation that this position belongs to a validated promoter. On the one hand the mutation may hinder the fixation of an important transcription factor to this promoter. MYOD appears compatible with this site. On the other hand, this mutation may lead to the creation of a new binding site consensus for factors which normally do not bind this promoter. We represent here two compatible candidates: SOX5 (5′-TWWCAAAG-3′), and ABL1 (5’- AA/CAACAAA/C -3’). Binding of these two factors may have long-range consequences, including for instance the activation of TWIST1 by SOX5. Transcriptomic data suggest this latter scenario may prove true at least for the breast cancer [9,178].

Figure 6. A putative FOXP2-dependent pro-oncogenic/tumor suppressor regulatory network

This scheme illustrates how diverse activation pathways may converge to convert a typical cell from a pre-oncogenic to an oncogenic state through abnormal FOXP2 expression and activity. Genes and factors indicated here have been observed in numerous but distinct cancer types detailed in the main text and should not be considered as collectively acting throughout all steps of the oncogenic progression. The illustration of the FOXP2 structure is from Wikipedia.