Mosaic Activating Mutations in GNA11 and GNAQ Are Associated with Phakomatosis Pigmentovascularis and Extensive Dermal Melanocytosis

1Mongolian blue spots (or, more appropriately, dermal melanocytosis) are common birthmarks, seen in up to 95% of African neonates and approximately 10% of white Caucasians. As a result they are easily overlooked as a possible sign of underlying genetic disease. Classically they are flat, dark-blue lesions with indistinct edges overlying the buttocks and lower back, which undergo spontaneous resolution over a period of years. However, the appearances and behavior are not always classic. Some dermal melanocytosis involve unusual sites, cover large areas of the body surface, are more deeply pigmented, are better defined, and persist rather than resolve. These more atypical patterns of dermal melanocytosis can be associated with anomalies such as colocalized cleft lip, ocular melanocytosis, ocular melanoma, lysosomal storage disorders, or vascular birthmarks. This latter association then falls under the diagnostic group of phakomatosis pigmentovascularis (PPV) (for clinical examples see Figure 1). Phakomatosis pigmentovascularis is a group of sporadic disorders of unknown frequency, defined by the co-occurrence of pigmentary and vascular birthmarks and subclassified clinically by the exact cutaneous phenotypes (Supplementary Table S1 online). Whether the subtypes are distinct disorders or variable expressions of a single disorder has not been clear as the molecular and developmental pathogenesis of PPV has been unknown. The hypothesis of non-allelic twin spotting was previously expounded to explain the co-occurrence of two disparate birthmarks but was recently retracted by the author because of lack of supporting molecular evidence. Extracutaneous associations of PPV can be severe, including scleral or intraocular melanocytosis, glaucoma, intracerebral vascular or other malformations leading to seizures and cognitive delay, overgrowth of soft tissues or limbs, and melanoma. The latter can arise in choroid or conjunctiva, with melanocytoma of the optic disk also described. Without an understanding of the genetic pathogenesis of the disease, targeted therapies for the congenital and neoplastic aspects of the disease have been impossible.

2When considering the genetic basis of extensive dermal melanocytosis and of PPV, we hypothesized that these conditions could be the result of a postzygotic mutation in a member of the G-protein nucleotide binding protein alpha subunit family. This hypothesis was generated by the rare concurrent description of PPV with Sturge-Weber syndrome (SWS), a vascular disorder with no pigmentary phenotype recently found to be the result of postzygotic mosaicism for activating mutations in GNAQ. We also hypothesized that an identical mutation could be present in both types of birthmarks in PPV, secondary to a single mutation in a pluripotent progenitor cell.

3Missense GNA11 or GNAQ mutations were found in 8 of 11 patients tested. In each case, these mutations were detected in affected skin (and in one case ocular tissue) at very low levels and were undetectable in blood, indicating postzygotic mosaicism (Table 1). When levels of mutant allele <1% were detected in blood, these could not be distinguished from background noise despite the substantial depth of coverage (Supplementary Table S2 online) and were therefore called as wild type (WT). In the three patients with extensive dermal melanocytosis, we identified mutations in GNAQ in two, one c.548G>A, p.R183Q, and one c.626A>C, Q209P. In the eight patients with PPV, we identified somatic GNA11 mutations in four, three at position c.547C>T, p.R183C, and one novel mutation in the same codon c.547C>A, p.R183S. and GNAQ mutations in two, both at position c.548G>A, p.R183Q (Figure 2). Importantly there was conservation of mutation between pigmentary and vascular lesions in each patient for whom more than one skin biopsy was available (Table 1 and Supplementary Table S2). Measured percentage of mosaicism in skin samples was lowest at 1.5% using next-generation sequencing mutant allele count as a percentage of the total number of reads (Supplementary Table S2). None of these mutations in GNA11 and GNAQ is described in the largest current population database.

4For the in vitro transfection experiments, HEK293 cells were transfected with vector, WT GNA11, or mutant GNA11 constructs encoding R183C and Q209L. Figure 3 shows results of western blots for Flag, total GNA11, and the phosphorylated and total p38, JNK, ERK and AKT in total protein lysates from these cells. Analysis of the ratio of phosphorylated to total p38, JNK, ERK and AKT normalized by the actin loading control (from two replicates, pooled results of WT compared to those of mutant) demonstrated increased phosphorylation and therefore activation of the p38 MAPK pathway by the R183C mutant and of ERK, p38 MAPK, and JNK pathways by Q209L, in keeping with previous results for Q209L. These results were significant at the 0.05 level using a one-tailed t test.

5GNA11R183C- and GNA11Q209L-expressing zebrafish exhibited dark cutaneous patches of melanocytes by 1 month. When the mutant transgenic zebrafish reached adulthood at 3 months, nearly all injected fish (n = 17/18 for R183C and n = 16/16 for Q209L) developed large and clearly visible pigmentary lesions (Figure 4a and b), recapitulating the human phenotype of dermal melanocytosis. Wild-type GNA11 expression in melanocytes also produced mosaic animals with some pigmentary lesions (n = 10/27); however, these were usually smaller and less dark, suggesting that the melanocyte lineage is highly sensitive to GNA11-coupled receptor signaling. Histopathology revealed many extra melanocytes in both the epidermis (over the scales, not shown) and the dermis in the pigmentary lesion of mutant GNA11R183C injected fish, with occasional involvement of the underlying muscle (Figure 4c).

6We sought to differentiate a subgroup of the common birthmark dermal melanocytosis in infants who are at risk of extra-cutaneous complications, by elucidating the underlying genetic cause. We show here that extensive dermal melanocytosis and PPV are genetic conditions associated with postzygotic mutations in genes encoding Gα subunits of heterotrimeric G proteins. Specifically, postzygotic mutations in GNAQ have been found in extensive dermal melanocytosis, and mutations in GNA11 or GNAQ have been found in PPV types I, II (cesioflammea), IV, and V (cesiomarmorata), as well as the unclassifiable achromico-melano-marmorata type. Type IV is considered to be exceptionally rare and was not tested because no samples were available. Type III (spilorosea) is distinguished clinically from all other types of PPV by the absence of dermal melanocytosis, and because only one patient from this group was tested (and was WT), we cannot yet conclude anything about the genetic etiology of this type. Of note, the percentage mosaicism was often very low, the result of very few mutant cells within a skin biopsy of a completely macular birthmark, thus emphasizing the need for adequately sensitive detection techniques in this type of genetic disease. A possible alternative explanation for our findings is that these patients are somehow genetically predisposed to somatic mutations in the same genes, and this is particularly relevant for those in whom we have only been able to obtain one sample. Although this cannot be entirely excluded, it is a much less convincing explanation of the results than postzygotic mosaicism, as we discovered different mutations in different patients, but the same mutation was present in more than one location when two samples were obtained from one individual. As candidate gene sequencing was performed in this study, we cannot exclude the presence of further mutation in other genes. Further exome sequencing could be performed to look for secondary mutations.

7True embryonic mosaicism has not yet been described for GNA11 mutations. Isolated somatic GNA11 mutations have been found in uveal melanoma and single blue nevi (a lesion distinct from dermal melanocytosis both clinically and histologically). In contrast to our findings in PPV, the prevalence of codon 183 mutations in single melanomas and blue nevi was much lower than of codon 209 mutations. Postzygotic mosaicism for GNAQ mutations has been described as the major cause of SWS, a purely vascular disorder with no pigmentary phenotype. Somatic mutations in GNAQ are frequent in uveal melanoma and single blue nevi. Therefore, our findings not only describe GNA11 mosaicism but also extend the phenotypic spectrum of human GNAQ mosaicism from SWS (vascular only), through to PPV (vascular and pigmentary), to extensive dermal melanocytosis (pigmentary only). Exactly how the same mutation can generate these three distinct phenotypes is not yet clear.

8Codon 183 in both GNA11 and GNAQ is located within the guanosine triphosphate binding region of the human Gα subunit, a region required for hydrolysis of guanosine triphosphate to guanosine diphosphate and an essential step for inactivation of G-protein coupled receptor signaling. Therefore, these mutations lead to decreased function of guanosine triphosphatase and to constitutive activation of downstream effector pathways. We have shown that the most common mutation at codon 183 of GNA11 activates the downstream p38 MAPK pathway in human cells, whereas the codon 209 mutant (not found in this cohort in GNA11 but in one patient in GNAQ) activates p38 MAPK, JNK and ERK pathways, broadly supporting previous findings of effects of mutations in both GNA11 and GNAQ.

9Of interest, different mutations in GNA11 in the germline leading to altered sensitivity of cells to extracellular calcium concentrations have been described in autosomal dominant hypoparathyroidism and hypocalciuric hypercalcemia, respectively. One of these mutations (GNA11R60L) has been characterized as less activating than GNA11Q209L. This pattern of less severe mutations being supportable in the germline, whereas more severe mutations can survive only by mosaicism, once again confirms Happle’s established theory. Patients with PPV and SWS have not been documented to have abnormalities of calcium homeostasis; however, intravascular calcification can be a feature of neurological abnormalities in PPV, which could hypothetically involve localized mosaic calcium imbalances. This is an area that merits future careful research.

10A causal link between our genetic findings and mechanism is strongly supported by the mosaic zebrafish models demonstrated here, in which expression of GNA11R183C or GNA11Q209L leads to ectopic and increased numbers of melanocytes in the epidermal and dermal layers, recapitulating the human phenotype of dermal melanocytosis. Overexpression of WT GNA11 in melanocytes also leads to a low level of pigmentary mosaicism in zebrafish, underscoring the importance of close regulation of G-protein coupled receptor signaling in the melanocyte lineage. No other mosaic animal models of GNA11 gain-of-function mutations exist; however, further conformation of our findings comes from germline hypermorphic alleles of GNA11, which lead to increased cutaneous pigmentation in the murine Dsk phenotype.

11We found no evidence for the now-retracted theory of twin spotting or didymosis. Rather, we have shown that a single somatic mutation in GNA11 or GNAQ is responsible for both types of cutaneous lesion in PPV, mirroring closely the pathogenesis of the disparate birthmarks seen in HRAS mosaicism. We propose that extensive dermal melanocytosis and PPV are members of an expanding group of heterotrimeric G-protein alpha subunit gene mosaic conditions, joining McCune-Albright syndrome, caused by mosaicism for gain-of-function mutations in GNAS, and SWS, caused by mosaic gain-of-function mutations in GNAQ. We also propose that GNAQ mosaicism forms a spectrum from SWS, through PPV, to extensive dermal melanocytosis. This knowledge will allow accurate clinical molecular diagnosis of a subset of extensive dermal melanocytosis and of PPV, leading to identification of neonates at risk for serious complications associated with these birthmarks. Importantly, it will also pave the way for therapeutic options in these children.

12All human and animal studies were approved by the authors' institutions' research ethics review boards. Declaration of Helsinki protocols were followed, and all patients gave written informed consent. Eleven patients from four international centers were recruited with written consent for genetic research, approved by the local research ethics committees. Patients were classified by clinical phenotype: three with extensive or atypical dermal melanocytosis (which was defined as involving areas other than only the lumbosacral lesion) and eight with PPV, with representatives of subtypes I, II (cesioflammea), III (spilorosea), and V (cesiomarmorata) (Supplementary Table S1 and Figure 1), and also the unclassifiable achromico-melano-marmorata type. Blood samples and skin biopsies were taken from each patient, and in those patients with PPV, skin was biopsied from both the vascular and pigmentary birthmarks when possible.

13Deoxyribonucleic acid (DNA) was extracted from whole blood and directly from fresh skin lesion samples using the DNeasy Blood and Tissue Kit (Qiagen, Düsseldorf, Germany) and from paraffin embedded tissue using the RecoverAll total nucleic acid extraction kit for formalin-fixed paraffin-embedded tissue (Life Technologies, Carlsbad, CA). To maximize detection of mutant alleles at low percentage mosaicism, we designed restriction enzyme digests of the normal allele at hotspots codons 183 and 209 of GNA11 and 183 of GNAQ using validated methods and Sanger sequencing. Primer sequences and restriction enzymes for each hotspot are given in Supplementary Table S3 (online). Touchdown polymerase chain reaction programs were used throughout, with 35 cycles for the first polymerase chain reaction and 25 for the second.

14We performed targeted deep sequencing of the regions spanning mutations previously identified by selective amplification of mutant alleles in all samples for which DNA quality was sufficient (as determined by the presence of a band above 10 kb size on agarose gel). In subjects with no identified mutation, all coding regions of GNAQ and GNA11 were screened using the same method. Targeted regions were amplified using custom intronic primers and standard long-range polymerase chain reaction protocols. Polymerase chain reaction products were purified and libraries were prepared using the Nextera XT DNA Sample Preparation kit (Illumina, Cambridge, UK). Samples were then pooled and sequenced on a MiSeq instrument (Illumina) according to the manufacturer’s recommendations for paired-end 150-bp reads. In-depth sequencing was performed to achieve a sequencing depth of at least 1,000 reads for all targeted coding bases and splice junctions. Identification of candidate variants was performed as described previously. Briefly, all targeted bases were systematically screened to count all sites with at least one read not matching the reference sequence, using a base-quality threshold of 30. Candidate postzygotic variants were confirmed by at least one independent experiment in all DNA samples available from the patient.

15To test the causality and model the effects of the mutation, we injected zebrafish embryos with WT human GNA11, GNA11R183C, or GNA11Q209L expressed from the melanocyte mitfa promoter and grew the genetically mosaic animals to adulthood. All zebrafish work was performed in accordance with United Kingdom Home Office Animals (Scientific Procedures) Act (1986) and approved by the University of Edinburgh Ethical Review Committee.

16The human GNA11 cDNA clone was purchased from Origene (purchased from NM_002067.1 precloned into an untagged pCMV6-XL4 vector, catalog number SC303115). Mutant GNA11Q209L (626a> t) was generated by site-directed genesis polymerase chain reaction with primers (forward: 5′-GATGTGGGGGGCCTGCGGTCGGAGCGGAGG-3′, reverse: 5′-CCTCCGCTCCGACCGCAGGCCCCCCACATC-3′). Mutant GNA11R183C (547c>t) was also generated with primers (forward: 5′-CGTGCTGCGGGTCTGCGTGCCCACCACCG-3′, reverse: 5′-CGGTGGTGGGCACGCAGACCCGCAGCACG-3′). The WT and mutant GNA11 cDNAs were cloned together with the zebrafish mitfa gene promoter into the pDestTol2CG2 expression vector using the Tol2kit gateway cloning method, resulting in mitfa-GNA11, mitfa-GNA11Q209L and mitfa-GNA11R183C constructs. For selection purposes, these constructs contain an additional green fluorescent protein gene expressed from the heart cml promoter. Two nanoliters of mixed mitfa-GNA11 plasmid DNA and Tol2 mRNA (62.5 ng/μl and 70 ng/μl, respectively) was injected into one-cell stage zebrafish embryos. Zebrafish embryos with green fluorescent protein transgenic marker in the heart were selected and grown to adulthood.

17Adult zebrafish were anesthetized in 50 mg/L tricaine solution and ×1 images from both sides of the fish were taken under a stereomicroscope. Dark patches of melanocyte lesions with areas larger than that of one scale were counted. Adult zebrafish were sacrificed by immersion in tricaine solution as directed by Home Office Schedule 1 methods. Zebrafish were then dissected in half transversely to increase penetration of the fixative and fixed in 4% freshly prepared paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) at 4 °C for 3 days. Samples were washed in phosphate buffered saline and decalcified in 0.5 M EDTA at pH 8 for 5 days before storage in 70% ethanol. Samples were embedded in paraffin wax and sectioned at 5-μm thickness. For hematoxylin and eosin staining, slides were deparaffinized in xylene twice for 5 minutes and then rehydrated through graded alcohol solutions (100%, 90%, 70%, 50% and 30% for 3 minutes each) and stopped in water. Slides were stained with Mayer hematoxylin and eosin after the standard hematoxylin and eosin staining procedure.

18To characterize the functional effect of mutations on downstream signaling pathways, HEK293T cells were transfected with vector, WT GNA11, or mutant GNA11 constructs encoding R183C and Q209L. The Q209L was used as a positive comparator because it is a more studied mutation in GNA11. Flag-tagged GNA11 WT, R183C, and Q209L constructs and vector alone control (pDEST26, Invitrogen, Carlsbad, CA) were transfected into HEK293 cells using lipofectamine 2000 (Invitrogen), according to the manufacturer’s instructions. Cells were cultured for another 24 hours, and protein lysates were prepared by boiling in a denaturing SDS buffer (2% 2-mercaptoethanol, 2% SDS, 10mM Tris pH 7.5) for 10 minutes. Levels of GNA11 effector activation were assessed by western blot using the following antibodies at the following concentrations: Rabbit anti–p-Serine473 Akt (1/500; Cell Signaling Technology, Danvers, MA), rabbit anti-total Akt (1/1,000, Cell Signaling Technology), rabbit anti-phospho T180/Y182 p38 MAPK (1/500; Cell Signaling Technology), rabbit anti-Total p38 MAPK (1/500, Cell Signaling Technology), mouse anti-phospho T183/Y185 SAPK/JNK (1/500; Cell Signaling Technology), rabbit anti-total SAPK/JNK (1/500; Cell Signaling Technology), pT202/Y204 ERK (1/500; Cell Signaling Technology), and rabbit anti-total ERK 1,2 (1/1,000; Cell Signaling Technology), mouse anti-Flag (DYKDDDDK tag) (1/500; Cell Signaling Technology), rabbit anti-GNA11 (1/500; GeneTex, Irvine, CA), and mouse anti-GAPDH (1/2,000; Millipore, Watford, UK).

19Primary antibody incubations were in TBST (100 mM Tris HCl, 0.2 M NaCl, 0.1% Tween-20 v/v) containing 5% bovine serum albumin (Sigma, Gillingham, United Kingdom) or 5% skimmed milk powder either overnight at 4 oC or for 1–2 hours at room temperature, whereas secondary antibody incubations were performed in 5% skimmed milk powder for 1 hour at room temperature. The concentrations used were swine anti-rabbit HRP (DakoCytomation, Carpinteria, CA) 1:3,000 and rabbit anti-mouse HRP (DakoCytomation) 1:2,000. Protein was visualized using Luminol reagent (Santa Cruz Biotechnologies, Dallas, TX). Densitometry of bands was performed using thresholded images in the ImageJ suite (imagej.nih.gov/ij/).

20Veronica Kinsler: http://orcid.org/0000-0001-6256-327X

21Pierre Vabres: http://orcid.org/0000-0001-8693-3183

22The authors state no conflicts of interest.