Embryonic stem cell factor FOXD3 (Genesis) defects in gastrointestinal stromal tumors

Abstract

Gastrointestinal stromal tumors (GISTs) are mesenchymal neoplasms, believed to originate from the interstitial cells of Cajal (ICC), often caused by overexpression of tyrosine kinase receptors (TKR) KIT or PDGFRA. Here, we present evidence that the embryonic stem cell factor FOXD3, first identified as 'Genesis' and involved in both gastrointestinal and neural crest cell development, is implicated in GIST pathogenesis; its involvement is investigated both in vitro and in zebrafish and a mouse model of FOXD3 deficiency. Samples from a total of 58 patients with wild-type GISTs were used for molecular analyses, including Sanger sequencing, comparative genomic hybridization, and methylation analysis. Immunohistochemistry and western blot evaluation were used to assess FOXD3 expression. Additionally, we conducted in vitro functional studies in tissue samples and in transfected cells to confirm the pathogenicity of the identified genetic variants. Germline partially inactivating FOXD3 sequence variants (p.R54H and p.Ala88_Gly91del) were found in patients with isolated GISTs. Chromosome 1p loss was the most frequent chromosomal abnormality identified in tumors. In vitro experiments demonstrate the impairment of FOXD3 in the presence of those variants. Animal studies showed disruption of the GI neural network and changes in the number and distribution in the ICC. FOXD3 suppresses KIT expression in human cells; its inactivation led to an increase in ICC in zebrafish, as well as mice, providing evidence for a functional link between FOXD3 defects and KIT overexpression leading to GIST formation.

Keywords: Carney triad; Carney–Stratakis syndrome; FOXD3; GIST; interstitial cells of Cajal.

Fig 1.. Molecular analysis of FOXD3, including CGH, sanger sequencing and methylation analysis

(A) CGH array showing somatic loss of the 1p locus harboring the FOXD3 in a Gastrointestinal stromal tumor (GIST) from a patient with Carney triad (CT); (B) Top: Sanger sequencing showing genetic variants in FOXD3; from left to right: Germline heterozygous variant c.161G>A (DNA from blood), LOH-consistent appearance of c.161G>A (*) in DNA from matched tumor tissue, germline deletion (p.Ala88_Gly91del, del12), germline variant p.V96L; bottom: conservation of the corresponding FOXD3 loci across species; (C) Left: DNA methylation analysis (range of DNA methylation from 0 to 100% is indicated by yellow-blue color gradient, gray indicates missing information) of CpG units within the FOXD3 upstream and promoter region in (i) normal abdominal tissue; (ii) tumor DNA from wild-type GIST; (iii) tumor DNA from GIST and PGL samples of CT patients; (iv) germline DNA from patients with wild-type GIST and CT (CT patient indicated with *); right: schematic representation of the identified genetic variants along the FOXD3 gene; CpG sites are indicated.

Fig 2.. Down regulation of FOXD3

(A) Top: Western blot showing decreased FOXD3 protein expression in tumors with 1p31 deletion (*) or FOXD3 promoter methylation (#) compared to samples without known FOXD3 defects; bottom: band densitometry quantification of each sample; bars represent the ratio FOXD3 / GAPDH; (B) decreased FOXD3 immunoreactivity in a tumor with 1p31 deletion (right) compared to a GIST caused by a PDGFRA mutation (left); bar indicates 10 μm; (C) Pigmented skin lesions in a patient with CT whose tumor showed methylation.

Fig 3.. In vitro tests for evaluation of FOXD3 variants activity and relation between FOXD3 and c-KIT expression

(A) Decreased transcriptional transactivation of the identified FOXD3 sequence variants; *p <0.001; (B) Top: siRNA knockdown of the FOXD3 resulted in an upregulation of c-KIT expression in both HEK293 (left) and NTERA-2 (right) cell lines; bottom: band densitometry quantification of each sample in each cell line; for each sample the black bars represent the ratio FOXD3 / GAPDH and the grey bars represent the ratio c-KIT / GAPDH;

Fig 4.. foxd3^zdf10/zdf10 mutant zebrafish embryos and gastrointestinal immunohistochemistry analysis

Foxd3 mutant zebrafish (B,G) embryos show hyperpigmentation (white brackets and black arrows) compared to wild-type (A,F). Foxd3 mutant zebrafish embryos show increased numbers of Cajal-like cells and decreased neural innervations (H-J) compared to wild-type (C-E), bar indicates 10 μm.

Fig 5.. Analysis of the number of Interstitial Cells of Cajal and thickness in the muscularis of the ileum in Foxd3^+/− mutant and wild type 12-month old mice

(A) Interstitial cells of Cajal (ICC) identified by c-Kit immunostaining; right: Foxd3^+/− mutant showing a higher c-Kit staining in the inner circular muscularis and a lower c-Kit staining in the outer longitudinal muscularis when compared with a wild type 12-month old mice (left); bar indicates 20 μm; (B) c-Kit positive cells were counted in random fields from the ileum (n=3 wild type and n=5 Foxd3^+/− mutant mice); (C) Thickness of the muscularis was measured in random fields (10–20 measurements/field) from the ileum (n=3 wild type and n=5 Foxd3^+/− mutant mice). Data represented as mean ± SEM. * p <0.0001.