Candidate Functional Promoter Variant in the FOXD3 Melanoblast Developmental Regulator Gene in Autosomal Dominant Vitiligo

To the Editor: Generalized vitiligo is an acquired disorder in which white patches of skin and hair result from autoimmune loss of melanocytes (Hahn and Nordlund, 2000), often associated with other autoimmune disorders (Alkhateeb et al, 2003). Most cases occur in a sporadic family pattern suggesting polygenic, multifactorial inheritance. However, we previously described a striking family in which a somewhat unusual vitiligo phenotype, characterized by progressively coalescent diffuse depigmentation and relatively early disease onset, segregated as an apparent autosomal dominant with incomplete penetrance (Alkhateeb et al, 2002). By genetic linkage analysis of this family, we mapped a vitiligo susceptibility locus, AIS1, to a 7.4-Mb interval of chromosome 1p31 (Alkhateeb et al, 2002). The AIS1 interval contains only ∼27 genes, including an obvious biological candidate, FOXD3, which encodes a forkhead transcription factor that is a primary regulator of melanoblast differentiation in the embryonic neural crest (Sutton et al, 1996; Hromas et al, 1999; Dottori et al, 2001; Kos et al, 2001), and perhaps also for some endodermal lineages (Guo et al, 2002). FOXD3 is expressed in all embryonic neural crest lineages except late-emigrating melanoblasts (Sutton et al, 1996; Hromas et al, 1999; Dottori et al, 2001; Kos et al, 2001). Inhibition of FOXD3 expression expands the melanoblast lineage and persistent FOXD3 overexpression results in developmental failure of melanoblasts (Dottori et al, 2001). FOXD3 mutations that result in transcriptional upregulation thus might interfere with melanoblast differentiation, thereby predisposing to vitiligo. To search for AIS1, we prioritized genes in the AIS1 interval based on their known functions and/or involvement in other diseases, and sequenced the coding and promoter regions of the eight high-priority (FOXD3, JUN, NFIA) and intermediate-priority genes (PPAP2B, MPRP1, FLJ10986, CYP2J2, BBP, ITGB3BP) in the proband of the study family and an unrelated control. The other genes in the AIS1 interval, mostly prioritized lower because their known functions or disease phenotypes argued against involvement in vitiligo or autoimmunity, were not studied. We detected a total of 25 sequence variants, of which 15 were heterozygous in the proband and thus potentially consistent with dominant inheritance. Family analysis showed that only two of these single-nucleotide polymorphisms (SNPs) (FOX D3−639G>T and NFIA [IVS]3 +80G>T) co-segregated with vitiligo in the study family and were not common variants. NFIA IVS3 +80G>T is located deep within an intron and does not create a potential cryptic splice site, whereas FOXD3−639G>T is located in the transcriptional promoter, and thus seemed much more likely to be of potential functional significance. FOXD3−639G>T does not appear in dbSNP, and we did not detect it in 100 unrelated Caucasian controls and 96 additional unrelated Caucasian vitiligo probands. FOXD3−639G>T thus appears to be a private variant. 5′ RACE (rapid amplification of cDNA ends) analysis of human FOXD3 mRNA showed that the −639G>T substitution is located 305 nt upstream of the major FOXD3 mRNA 5′ start and 519 nt upstream of a minor 5′ start (Fig 1). This region of the FOXD3 promoter has been remarkably conserved between humans, mice, rats, dogs, and zebrafish, with 64% sequence identity over the 171-bp segment immediately downstream from the −639G>T variant. Although −639G>T is located adjacent to high-likelihood Rc, XF1, and ZF5 motifs, the variant neither alters these motifs nor creates any known transcriptional motifs. Alignment of orthologous FOXD3 promoter sequences. Promoter and 5′-untranslated sequences of the human, mouse, rat, dog, and zebrafish (Danio) FOXD3 orthologues are shown (translation initiation codon underlined). Nucleotides identical in all five species are highlighted in red; the position of the human −639G>T substitution is highlighted in light blue. No significant sequence identity was evident upstream of this region. Green arrow overline indicates the major 5′ terminus of human FOXD3 mRNA at nt −344; orange arrow overline indicates the minor mRNA 5′ terminus at nt −130. Note that DanioFOXD3 mRNA apparently initiates further 5′, approximately at a position indicated by an asterisk underline. The 5′ terminus of human FOXD3 mRNA was determined by 5′-RACE using the GenRacer Kit (Invitrogen, Carlsbad, California), 1 μg Tera1 total RNA, and gene-specific primers 5′-CAACACCGCCCTTGCAGC-3′ and 5′-ACATGTCGCTGGCGCTGC-3′. To measure the effect of the FOXD3−639G>T variant on the transcriptional activity of the cis FOXD3 promoter, we transfected Tera1 teratocarcinoma cells (which express endogenous FOXD3 mRNA; Sutton et al, 1996) and primary human melanocytes (which do not express endogenous FOXD3; not shown) with transcriptional reporter constructs containing the FOXD3−639G wild-type versus−639T variant promoters (nt −927 to −261 relative to the translational initiation codon) driving transcription of a firefly luciferase reporter gene. As shown in Fig 2, the transcriptional activity of the −639T variant promoter was significantly elevated (activity 150%±26; n=18; p<0.0001 compared with the −639G wild-type promoter) in transfected Tera1 cells, whereas the transcriptional activity of the −639T variant promoter was not elevated (activity 95%±21; n=9; p=0.7 compared with the −639G wild-type promoter) in transfected melanocytes. The transcriptional activity of the exogenous −639T variant promoter is thus significantly greater than that of the −639G wild-type promoter in transfected cells that are permissive for expression of endogenous FOXD3, but not in cells that do not express endogenous FOXD3. These results are thus consistent with the hypothesis that the −639T substitution results in elevated FOXD3 transcription in neural crest melanoblast precursors, potentially suppressing or altering the differentiation profile of the melanoblast lineage. Transcriptional activity of FOXD3−639T and −639G promoters in transiently transfected cells. Transfected Tera1 teratocarcinoma cells (A). Transfected primary human epidermal melanocytes (B). Luciferase activity of the wild-type −639G promoter was set as 100%. Luciferase values are the mean of 18 replicates for Tera1 cells and 9 replicates for melanocytes; error bars represent standard deviations. HEMn-LP (Cascade Biologics, Portland, Oregon) are primary epidermal melanocytes from a lightly pigmented normal individual, cultured in Medium 154 containing Human Melanocyte Growth Supplement and penicillin, streptomycin, and amphotericin B (Cascade Biologics). Tera1 is a testicular teratocarcinoma cell line (ATCC) cultured in McCoy's 5A medium containing 15% FBS, 100 U per mL penicillin, and 100 μg per mL streptomycin. 667-bp segments spanning the entire FOXD3 promoter through the major 5′ mRNA terminus at nt −344 of both the −639G and −639T alleles were amplified by PCR from DNA of the proband using primers that create MluI and BglII restriction sites: 5′-TTTGCTACGCGTAACACATATGGGAACCTCGTTG-3′ and 5′-CGCTAGATCTAGGGCGCTCGCTACCGCTCCCG-3′. PCR products were ligated to pCR2.1-TOPO (Invitrogen) and used to transform TOP10F' E. coli. Plasmids containing −639G and −639T inserts were identified by sequencing; the corresponding MluI–BglII fragments were purified and ligated to pGL3-basic (Promega, Madison, Wisconsin) and used to transform Escherichia coli XL1-blue (Stratagene, La Jolla, California), and plasmids containing both the wild-type (pGL3-FOX/-639G) and mutant (pGL3-FOX/-639T) promoters were identified and re-sequenced. Tera1 cells were plated at 1 × 105 cells per 9.6 cm2 well and were transfected after 24 h using 1 μg pGL3-basic, pGL3-FOX/-639G, pGL3-FOX/-639T, or pGL3-control plasmids, plus 20 ng pRL-SV40 (Renilla luciferase) plasmid as an internal control for transfection efficiency, using 3 μL FuGENE transfection reagent (Roche, Basel, Switzerland). For melanocytes, 5 × 105 log-phase cells were mixed with 100 μL nucleofector solution (Amaxa, Gaithersburg, Maryland), and 3 μg of pGL3-basic, pGL3-FOX/-639G, or pGL3-FOX/-639T, plus 60 ng pRL-SV40 plasmid were electroporated using the nucleofector electroporator (Amaxa) and the U24-specific nucleofector program, and were then replated in 9.6 cm2 wells. Cells were lysed and assayed for luciferase expression after 30–42 h for Tera1 cells and ∼16 h for melanocytes, which yielded optimal luciferase expression. The dual-luciferase reporter system (Promega) was used to simultaneously measure firefly and Renilla luciferase activities utilizing a Monolight 3010 luminometer (Macam Photometrics, Livingston, Scotland, UK), and the ratio of firefly and Renilla luciferase activities was determined. Data were analyzed using Student's t test. To determine whether functional variation in FOXD3 might occur in other multiplex vitiligo families, we sequenced 3680 bp spanning the entire FOXD3 gene and promoter (nt −1492 to +2188) in two affected members of each of six small Caucasian multiplex vitiligo families, and we used DHPLC to screen for variation in the FOXD3 promoter region in probands of another 91 small Caucasian vitiligo families and 96 unrelated Caucasian controls. Sequence variants detected in vitiligo probands were then tested for co-segregation with vitiligo in the families in which they occurred. We thereby detected 13 additional FOXD3 SNP, all relatively uncommon and none that consistently co-segregated with vitiligo in the small families in which they were found. Whereas the FOXD3−639G>T promoter substitution may be the cause of AIS1-dependent vitiligo in the original study family, variation in the FOXD3 structural gene or its immediate vicinity clearly is not a frequent genetic risk factor for vitiligo. Indeed, a number of atypical features of the study family argue that this family is exceptional, perhaps unique. First, vitiligo occurs as an apparent autosomal dominant trait in this very large family, with many individuals affected. Second, the vitiligo phenotype is atypical, comprising slowly progressive diffuse depigmentation, although invariably diagnosed as “vitiligo” by patients' dermatologists. Third, the age of disease onset in this family is relatively early, 12.9±10.1 vs 22.4±16.0 y among vitiligo patients in general (Alkhateeb et al, 2003). These atypical characteristics argue that vitiligo in this family may involve a quite uncommon pathogenic mechanism. FOXD3 is thus a strong candidate gene for vitiligo susceptibility in this family, with autosomal dominant vitiligo linked to chromosome 1p. Nevertheless, identification of FOXD3 mutations in additional AIS1-linked families, and more extensive functional studies, will be required to definitively establish the role of FOXD3 in AIS1-linked autoimmune disease. These studies were approved by the Colorado Multiple Institutional Review Board (COMIRB), and were performed in accord with Declaration of Helsinki principles. Written informed consent was obtained from all study participants prior to their participation. We thank the National Vitiligo Foundation (Tyler, Texas) and the Vitiligo Society (London, UK) for their participation. This work was supported by grants AR45584 and AI46374 from the National Institutes of Health.