ARAF recurrent mutation causes central conducting lymphatic anomaly treatable with a MEK inhibitor

Dong Li1, Michael E. March1, Alvaro Gutierrez-Uzquiza1,12, Charlly Kao1, Christoph Seiler2, Erin Pinto3, Leticia S. Matsuoka1, Mark R. Battig1, Elizabeth J. Bhoj1, Tara L. Wenger4, Lifeng Tian1, Nora Robinson1, Tiancheng Wang1, Yichuan Liu1, Brant M. Weinstein5, Matthew Swift6, Hyun Min Jung5, Courtney N. Kaminski1, Rosetta Chiavacci1, Jonathan A. Perkins7, Michael A. Levine8,9, Patrick M. A. Sleiman1,9, Patricia J. Hicks9, Janet T. Strausbaugh9, Jean B. Belasco9,10, Yoav Dori3,9 and Hakon Hakonarson   1,9,11*

Summary & Introduction

The treatment of lymphatic anomaly, a rare devastating disease spectrum of mostly unknown etiologies, depends on the patient manifestations1. Identifying the causal genes will allow for developing affordable therapies in keeping with precision medicine implementation2. Here we identified a recurrent gain-of-function ARAF mutation (c.640T>C:p.S214P) in a 12-year-old boy with advanced anomalous lymphatic disease unresponsive to conventional sirolimus therapy and in another, unrelated, adult patient. The mutation led to loss of a conserved phosphorylation site. Cells transduced with ARAF-S214P showed elevated ERK1/2 activity, enhanced lymphangiogenic capacity, and disassembly of actin skeleton and VE-cadherin junctions, which were rescued using the MEK inhibitor trametinib. The functional relevance of the mutation was also validated by recreating a lymphatic phenotype in a zebrafish model, with rescue of the anomalous phenotype using a MEK inhibitor. Subsequent therapy of the lead proband with a MEK inhibitor led to dramatic clinical improvement, with remodeling of the patient’s lymphatic system with resolution of the lymphatic edema, marked improvement in his pulmonary function tests, cessation of supplemental oxygen requirements and near normalization of daily activities. Our results provide a representative demonstration of how knowledge of genetic classification and mechanistic understanding guides biologically based medical treatments, which in our instance was life-saving.
Although recent studies have demonstrated the benefit of sirolimus in the treatment of generalized lymphatic anomaly (GLA) and central conducting lymphatic anomaly (CCLA)3–5, the absence of clear clinical distinctions between these entities, due to their rarity and overlapping of diagnostic criteria, has hampered the development of innovative therapies6–9. GLA is defined as multifocal lymphatic anomaly that has multiple areas of micro/macrocystic lymphatic malformation and often involves bone destruction9–11. CCLA, on the other hand, describes dysfunction of the thoracic duct (TD) or cisterna chyli, leading to a retrograde flux of lymphatic fluid or abnormal drainage of lymphatic fluid1,12,13. Both conditions can manifest with chylothorax, effusions, chylous ascites or lymphedema. The overlap of these apparently disparate disorders suggests that a common pathway rather than a common gene is responsible for the various clinical syndromes, and implies that the distinction between entities may be artificial. Here we report the use of wholeexome sequencing (WES) to identify a recurrent missense mutation in ARAF as the basis for a severely advanced lymphatic disease characterized by a complex lymphatic anomaly in two unrelated patients. Our results provide a representative demonstration of how genetic classification presents a way to categorize complex medical disorders, thereby guiding biologically based medical treatments, which in our instance was life-saving.
The first tier of WES analyses of the known lymphatic anomaly-associated genes was unrevealing, including mutation analysis of AKT1, PIK3CA, KRAS, HRAS, NRAS, BRAF, RAF1, PTPN11, SHOC2, CBL, RIT1 and SOS1. Subsequent gene prioritization revealed a novel X chromosomal ARAF mutation, c.640T>C:p. S214P, in both patient P1, a male with CCLA (Fig. 1a,c–e; see Methods for a detailed clinical description), and patient P2, a female diagnosed with lymphangiomatosis in 2012 before the establishment of the 2015 International Society for the Study of Vascular Anomalies classification. The mutation affects a conserved phosphorylation site, which putatively resulted in a gain-of-function (GoF) effect as the residue Ser 214 is a paralogous regulatory site in its homologous protein C-RAF (also known as RAF1) for inhibition by 14-3-3 proteins. This missense mutation was absent from 1000 Genomes Project, ESP6500SI, ExAC v0.3, gnomAD v2.1 or additional exome-sequencing data from more than 5,000 samples that we had in our in-house database. Sanger sequencing of bloodderived DNA from P1 and both parents confirmed that this X-linked ARAF mutation occurred as a somatic heterozygous event as shown in the male patient (Fig. 1f). Sanger sequencing of the ARAF mutation in P2, her unaffected daughter and mother confirmed the mutation was present only in P2 (Fig. 1f). The father was unavailable for sequencing; however, as her father had no reported respiratory symptoms it remains likely that the ARAF mutation arose as a de novo or somatic mutation in P2. Patient P2 was lost to follow-up and we were informed later that she died in 2017 from complications of her lymphatic disease, five years after her diagnosis.
The Ser 214 residue, which is one of the 14-3-3 binding sites in conserved region 2 (CR2)14, in ARAF is highly conserved across vertebrate species, as well as within the RAF proteins, suggesting that it may serve an essential role in the function of these kinases (Fig. 1g). The binding of 14-3-3 proteins to phosphorylated Ser 214 of ARAF would prevent recruitment of ARAF protein to the plasma membrane by activated Ras15. Previous studies showed that the mutations in the ARAF-S214 paralogous residue Ser 259 in C-RAF impaired binding of 14-3-3 proteins, leading to plasma membrane localization and inducing ERK/MEK signaling16. As shown in Fig. 2a, HEK293T cells transfected with ARAF-S214P showed reduced co-immunoprecipitation of 14-3-3 proteins, and in turn significantly greater activation of ERK1/2, as measured by increased phosphorylation, compared with HEK293T cells expressing wildtype (WT) ARAF (Fig. 2a,b). Phosphorylation of AKT, p70S6K, mTOR and p38 (another family of MAP kinases) was not altered by ARAF-S214P (Fig. 2b). Similar results were obtained in HeLa cells (Extended Data Fig. 1) and in primary human dermal lymphatic endothelial cells (HDLECs) (Fig. 2c). This marked overactivation was also present even in the absence of cytokines or growth factors (Extended Data Fig. 1).
HDLECs expressing ARAF-S214P manifest enhanced lymphangiogenic capacity compared with HDLECs expressing ARAF-WT, as measured by the number of sprouts and the sprout length in the three-dimensional lymphatic spheroid sprouting assay conducted in the absence of vascular endothelial growth factor C (VEGFC) (Fig. 2d). The MEK inhibitor trametinib rescued the increased sprouting in the mutant (Fig. 2d). We then performed a morphological analysis of the endothelial adherens junctions of primary HDLECs expressing ARAF-S214P. As shown by immunofluorescence microscopy, ARAF-S214P expression caused a significant absence of VE-cadherin accumulation between adjacent cells suggesting increased VE-cadherin internalization (Fig. 2e, yellow arrowheads, and Extended Data Fig. 2a). Additionally, expression of ARAF-S214P altered actin organization (Extended Data Fig. 2b), with mutant-expressing cells possessing fewer discrete F-actin filaments within the cell body (Extended Data Fig. 2c). We then examined the ability of MEK1/2 inhibitors to reverse these abnormalities. The MEK inhibitor trametinib, at a concentration of 100 nM, rescued the loss of VE-cadherin from cell–cell junctions observed in HDLECs expressing ARAF-S214P with an almost complete restoration of the cell monolayer integrity and a recovery of the normal appearance of VE-cadherin at junctions and actin filaments (Fig. 2e and Extended Data Fig. 2). Although ARAF-S214P clearly activates ERK in HDLECs, and ERK activation is typically associated with cell proliferation, we did not observe any measurable differences in proliferation between ARAF-WT- and -S214P-expressing HDLECs across two independent retroviral transductions (Fig. 2f and Extended Data Fig. 3).
Analysis of lymphatic development in zebrafish was performed in the Tg(mrc1a:egfp)y251 transgenic line17, where all lymphatic endothelial cells are labeled with EGFP. ARAF expression was targeted to lymphatic vessels with the mrc1a promoter, and ARAF-expressing cells were marked by mCherry expression. ARAF-S214P expression induced dilated lymphatic vessels in different locations (Extended Data Fig. 4), and most consistently we observed dilation of the trunk TD (Fig. 2g). Expression of ARAF-WT, in contrast, had no effect on lymphatic morphology (Fig. 2h). Expression of ARAF-S214P induces p-ERK in zebrafish (Extended Data Fig. 5). To determine whether MEK signaling inhibitors can reverse the anomalies, we treated mrc1a:ARAFS214P larvae with cobimetinib from 3 d post fertilization (dpf), when the lymphatic progenitor cells sprout to form the TD17. We analyzed body segments (somites) with ARAF expression in the TD at 7 dpf and found a significant rescue of duct morphology by cobimetinib (Fig. 2i,j). Meanwhile, we treated WT Tg(mrc1a:egfp)y251 larvae with cobimetinib, and found that they tolerated the drug well (Extended Data Fig. 6).
In view of our demonstration that the ARAF mutation led to a GoF effect in P1 that was unresponsive to sirolimus and that MEK inhibitors could rescue the lymphatic phenotype in both transduced endothelial cells and in a transgenic zebrafish model, we sought Institutional Review Board clearance to use MEK inhibitor therapy in P1. Trametinib (Mekinist), a Food and Drug Administration (FDA)-approved MEK inhibitor, was subsequently used off-label in this 12-year-old patient following comprehensive baseline evaluation. We used a starting dose of 1 mg d−1 of trametinib and began observing improvement in pulmonary function testing within 2 months of therapy (Fig. 3a). Moreover, there were significant reductions in lymphatic fluid retention and supplemental oxygen requirements after three months of treatment, and he was able to wean to room air with improved levels of physical activity and without any adverse events being observed from trametinib. At 12 months of therapy, his pulmonary function tests showed near doubling of his total lung capacity (TLC) and his forced expiratory volume in 1 s (FEV1) improved from 23% to 42% predicted (Fig. 3a). Electrolytes (low Na and K) normalized and his magnetic resonance imaging scan showed lymphatic remodeling with restructuring of his lymphatic system (Fig. 3b–f), a remarkable recovery in an individual who was frequently hospitalized before initiation of this lymphatic disease unresponsive to sirolimus therapy. HDLECs genetically guided therapy (Fig. 3f). transduced with the mutant ARAF showed elevated ERK1/2
In sum, we performed WES for two unrelated patients with activity, enhanced lymphangiogenic capacity, and disassembly of lymphatic anomaly and identified a recurrent GoF mutation in actin skeleton and VE-cadherin junctions, which were rescued the ARAF gene, including in a 12-year-old male with an advanced using the MEK inhibitor trametinib. Sprouting was observed in ARAF-S214P-expressing HDLECs in the absence of VEGFC (a potent lymphangiogenic factor)18. Under the same conditions, sprouting was absent in cells expressing ARAF-WT. This suggests that the ARAF mutant is mimicking the stimulatory behavior of VEGFC or inducing the expression of VEGFC by the HDLECs, which is necessary for endothelial cell sprouting, as seen in many stromal cell types19–21. Further experiments would be required to distinguish these possibilities. We reproduced the anomalous lymphatic phenotype, which is attributed to a GoF mutation in ARAF, in a zebrafish model observing rescue of the phenotype using MEK inhibitor therapy. Remarkably, therapy of the lead proband with the ARAF mutation using trametinib resulted in dramatic improvement in patient symptoms, with remodeling of his dilated and torturous lymphatic vasculature, resolution of the lymphatic edema and resumption of regular daily activities within 12 months of therapy.
From ongoing patient recruitment, we investigated additional lymphatic anomaly patients, including patients with Noonan (or Noonan-related) syndrome, Gorham–Stout disease, kaposiform lymphangiomatosis (KLA), lymphangiectasia and CCLA. On sequencing 43 additional patients, we identified 7 additional mutations in KRAS, BRAF, RASA1, PTPN11 and SOS1 (Table 1), suggesting that the RAS–MAPK signaling is a common pathway responsible for the various clinical lymphatic disease manifestations. Indeed, it has been increasingly acknowledged that the RAS–MAPK pathway plays a key role in the signaling of lymphangiogenesis21–23. Reviewing the literature, we identified more than 50 patients who have mutations in KRAS, HRAS, BRAF, RAF1, PTPN11, SHOC2, CBL, RIT1 and SOS1, and present clinical features of Noonan or Noonan-related syndromes with lymphatic defects, including pleural effusion, pericardial effusions, chylothorax, hydrops, lymphangiectasis and lymphedema24–36. While our work was in progress, a recurrent NRAS variant was implicated in GLA37 and also in KLA38, lending further support for the shared genetic etiology between these disease entities and the importance of mutations in the RAS–MAPK pathway in lymphatic anomalies.
The widespread prevalence of mutations in RASopathies in human cancer has been recognized for decades. A close scrutiny of the ARAF mutation we uncovered, using the cBioPortal39 database (n = 71,857 subjects and queried on February 6, 2019), reveals 2 patients with the same exact mutation in ARAF. Interestingly, they both have concurrent TP53 mutations, which are considered as oncogenic drivers. Different mutations at this residue (S214T, S214A, S214Y, S214C and S214F), three of which have been shown to result in elevated MEK/ERK phosphorylation40, were also observed in ten patients with different types of cancer. However, nine out of ten patients have co-occurring oncogenic mutations in TP53, GNAS, AKT2, APC, EGFR, ATM, CHEK2, KIT or U2AF1, raising the possibility that these oncogenic drivers may be responsible for the excessive proliferation in cancer cells. The lead proband with the ARAF mutation has dilated lymphatic vessels but the lesion shows no increase in size over years of follow-up. Thus, these data are consistent with our observation that the ARAF mutation we uncovered may not drive increased proliferation in lymphatic endothelial cells in vitro.
Regarding the prevalence of mutation-positive lymphatic anomalies, among 11 centers in the USA forming a lymphatic anomaly consortium to facilitate multi-center clinical trials for this group of lymphatic anomalies, including but not limited to GLA, Gorham– Stout disease, CCLA, KLA, Klippel–Trenaunay syndrome and kaposiform hemangioendothelioma, there are more than 3,000 patients recruited with moderate to severe disease course, and the number of new patients per year is about 300 combined. Based on the current molecular diagnostic yield (20%), we anticipate that about 20% of them will have defects in the RAS–MAPK pathway, suggesting that a few thousand patients overall in the USA may benefit from MEK inhibitor therapy. Thus, our work exemplifies how genetic discoveries can impact disease classification and uncover novel biological and life-saving treatments as represented here in a patient with lymphatic anomaly of a previously unknown etiology, a realization of a precision medicine approach.

Methods

Patients. After obtaining approval from the Institutional Review Board at The Children’s Hospital of Philadelphia (CHOP) and written informed consent, blood specimens from the lead proband (P1) and his parents were obtained for sequencing analysis. The proband had severe accumulation of lymphatic fluid in his chest, pericardium, abdomen, lower extremities and genitalia and was being followed and treated at the Center for Lymphatic Imaging and Interventions at CHOP. An unrelated second adult patient (P2) was recruited through the Patient Registry of the Lymphangiomatosis & Gorham’s Disease Alliance (LGDA), together with available family members.
Birth and family history for P1 were unremarkable except for a capillary malformation on the left side of his abdomen and his childhood growth and development milestones were normal. At age 10 years, he developed swelling of his lower abdomen, thighs, scrotum and penis. Two months later, he presented to a local hospital with shortness of breath and exercise intolerance. A chest radiograph demonstrated cardiomegaly and echocardiogram revealed a large pericardial effusion. Pericardiocentesis was performed with drainage of 1 l of chylous fluid. Despite institution of total parenteral nutrition, the drainage continued and he was transferred to CHOP for further management. At CHOP, his initial evaluation included dynamic contrast-enhanced magnetic resonance lymphangiography that demonstrated large pericardial effusion and antegrade flow in dilated lumbar and retroperitoneal networks into a dilated and tortuous TD coursing towards the innominate vein on the left (Fig. 1a,c,d). An image of an unaffected person is shown in Fig. 1b as a reference. In addition, there was retrograde lymphatic flow into the liver, mesentery, penis and scrotum, and from the distal TD there was retrograde flow into the mediastinum and pericardium (Fig. 1c–e). He underwent placement of a stent in the distal TD and Lipiodol embolization with the aim of stopping the abnormal mediastinal and pericardial lymphatic effusion. He was discharged after a month with a stable pericardial effusion but then presented shortly thereafter in respiratory distress due to large fluid re-accumulation that necessitated an increasing requirement for supplemental oxygen (up to 5 l by nasal cannula). He was started on sirolimus and was dosed based on trough levels tolerated well on a dose of 2.5 mg per day, which resulted in a median trough level of 11.8 between May and November 2016 (range 6.8–16.2 μg dl−1). Over the course of 1.5 years, he underwent multiple percutaneous interventional and surgical lymphatic procedures, including repetitive thoracentesis and pleural drains, multiple percutaneous lymphatic embolizations, bilateral surgical pleurodesis twice, surgical lymphovenous anastomosis in his thighs, abdomen and retroperitoneum and, due to worsening penile and scrotal edema, surgical ligation and embolization of groin lymph channels. Despite multiple attempts to control his pericardial effusions, his penile, scrotal, lower extremity and lower abdominal lymphedema worsened and his condition continued to deteriorate to the point that consideration of palliative care was discussed. The last procedure was performed and sirolimus was discontinued five months before trametinib began in March 2017 (Fig. 3a).
Patient P2, an unrelated adult female, was diagnosed with lymphangiomatosis at the age of 31 in 2012 before the International Society for the Study of Vascular Anomalies classification was established in 2015. She had extensive symptoms for many years before her diagnosis with prominent pulmonary involvement and required multiple pleurocentesis procedures before pleurodesis. She had widespread involvement of her gastrointestinal tract, requiring a specialized fat-restricted diet and medium-chain triglyceride oil supplementation with intermittent total parenteral nutrition. She underwent computed tomography and magnetic resonance imaging after persistent unexplained symptoms, which were consistent with lymphangiomatosis affecting her kidneys, liver, spleen and lungs. A liver biopsy confirmed the diagnosis of lymphangiomatosis. She was additionally treated with albuterol and diuretics and used a motorized scooter because of fatigue and dyspnea. She was never confirmed to have bone involvement. As the patient was recruited from the Lymphangiomatosis & Gorham’s Disease Alliance and was not local, she was lost to follow-up and was not available for a trial of other therapies (we later learned she had died in 2017 from complications related to her underlying lymphatic disorder).

WES and bioinformatics analysis. We examined missense, nonsense, splicealtering and coding indels matching either the dominant or recessive inheritance models in the exome data. Results were filtered to exclude variants with the following factors: synonymous variants; variants in known pseudogenes; variants with a minor allele frequency (MAF) greater than 0.5% in either the 1000 Genomes Project or the 6,503 exomes from the National Heart, Lung, and Blood Institute Exome Sequencing Project (ESP6500SI); variants previously identified in controls by our in-house exome variant database. Subsequent gene prioritization was performed on the basis of deleterious prediction and biological relevance by referring to the Online Mendelian Inheritance in Man database.

Expression and characterization of ARAF mutation in mammalian cell lines. HEK293T and HeLa cells were obtained from the American Type Culture Collection and grown at 37 °C in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Primary adult HDLECs were obtained from Promocell, and were cultured in Endothelial Cell Growth Medium MV 2 (Promocell) according to the manufacturer’s directions. The full-length ARAF cDNA obtained from Addgene (plasmid no. 23725)41 was amplified from the original vector and cloned as a BamHI/XhoI fragment into the pcDNA3.1 vector that contains two copies of the FLAG tag (DYKDDDDK), followed by two STREP tags (WSHPQFEK). The S214P mutation was introduced by site-directed mutagenesis using the Q5 mutagenesis kit from NEB following the manufacturer’s instructions. Transfections in HEK293T and HeLa were performed using Fugene HD (Promega), with 3 μg DNA (empty vector, WT ARAF (ARAF-WT) or ARAF mutant (ARAF-S214P)) and 9 μl of the transfection reagent, according to the manufacturer’s protocols. At 36–48 h after transfection, cells were washed twice with ice-cold phosphate-buffered saline (PBS) and lysed on ice using a freshly prepared ice-cold cell lysis buffer containing 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 50 mM β-glycerophosphate, 10% glycerol (w/v), 1% NP-40 (w/v), 1 mM EDTA, 2 mM NaVO4 and a complete, EDTA-free protease inhibitor cocktail (Roche Applied Science) at 20 μl per millilitre of lysis buffer. After clearing the cell lysates by centrifugation, the supernatants were collected and used for western blotting or immunoprecipitation with Anti-FLAG M2 Affinity Gel (cat. no. A2220, Sigma) followed by western blotting. Immunoprecipitates and lysates were run on NuPAGE 4–12% Bis-Tris gels (Thermo Fisher Scientific) and blotted with primary antibodies including anti-phospho-p70S6K-Thr389 (cat. no. 9205S, Cell Signaling Technology; 1:1,000), anti-phospho-mTOR Ser2448 (cat. no. 5536P, Cell Signaling Technology; 1:1,000), anti-FLAG (cat. no. F3165, Sigma; 1:4,000), antiphospho-p38 Thr180/Tyr182 (cat. no. 4511, Cell Signaling Technology; 1:1,000), anti-PAN-14-3-3 (cat. no. sc-629, Santa Cruz Biotechnology; 1:500), anti-phosphoAkt-Ser473 (cat. no. 4060, Cell Signaling Technology; 1:1,000), anti-phosphop44/42-(Erk1/2)-Thr202/Tyr204 (cat. no. 4376, Cell Signaling Technology; 1:1,000) or anti-β-actin (cat. no. sc-69879, Santa Cruz Biotechnology; 1:1,000) antibodies. The ARAF sequence, from pCDNA3.1-F2S2-ARAF-WT or -S214P constructs as previously indicated, was cut with BamHI/XhoI and introduced into the BglII/ XhoI sites of a modified version of the pMSCV plasmid that contains aminoterminal FLAG and HA tags. Viral production was performed using Fugene, with 8 μg of total DNA (pMSCV-ARAF-WT or -S214P together with envelope and packaging plasmids) and 18 μl of the transfection reagent in HEK293T. After 72 h, viral supernatant was collected and filtered. HDLECs were infected by replacing the cell culture medium with the viral supernatant, supplemented with 8 μg ml−1 Polybrene and filtered through a 0.45 μm filter. Cells were spinfected at 650g for 90 min, and subsequently cultured for 6 h at which point the viral supernatant was replaced by standard culture medium. Transduced HDLECs were cultured for 48 h before use in experiments. Transduction efficiencies observed by HA staining were between 40% and 60%.

Immunofluorescence staining and western blotting of HDLECs. Round (12 mm) coverslips (VWR) were coated with 0.1% gelatin in water for 10 min in 24-well plates (Corning), and then air-dried for 15 min. Transduced HDLECs were plated at 100,000 cells per well in 0.5 ml of culture medium in the presence or absence of trametinib for 48 h. Cells were washed in warm serum-free Dulbecco’s modified Eagle’s medium and fixed in 4% paraformaldehyde. Fixed cells were washed twice with PBS and twice with 0.1% BSA in PBS. Cells were permeablized and blocked by incubation with 10% normal donkey serum (Jackson Immunoresearch) and 0.3% Triton X-100 (Sigma Aldrich) in PBS. VE-cadherin antibody (Thermo Fisher Scientific) was diluted (final concentration: 2 μg ml−1) in 0.01% normal donkey serum, 0.1% BSA and 0.3% Triton X-100 in PBS, and staining was performed for 1 h. Coverslips were washed twice with 0.1% BSA in PBS. Goat-anti-rabbit Alexa546 (Thermo Fisher Scientific; final concentration: 8 μg ml−1) and phalloidin Alexa350 (Thermo Fisher Scientific; final concentration: 5 units ml−1) were diluted in 0.01% normal donkey serum, 0.1% BSA and 0.3% Triton X-100 in PBS, and staining was performed for 1 h. When used, HA-Tag (6E2) mouse antibody (cat. no. 2367, Cell Signaling Technology) was diluted 1:100 in 0.1% BSA and 0.3% Triton X-100 in PBS, and staining was performed for 1 h. Coverslips were washed twice with 0.1% BSA in PBS and twice with PBS. Coverslips were dipped in water to remove residual salts, and mounted to slides using Prolong Gold antifade reagent (Thermo Fisher Scientific). Image acquisition was performed on a Leica DM6000 motorized upright microscope with a Photometrics HQ2 high-resolution monochrome CCD (charge-coupled device) camera using LAS AF software (Leica Microsystems). Z-stacks were acquired at ×10 magnification. Images were further processed in the Fiji software package42. Brightness and contrast adjustments were made. Identical brightness and contrast settings were applied to all images. Fluorescence values were measured in regions of interest (ROIs) drawn to contain entire individual cells, or in ROIs drawn to contain the entire cell body but exclude the cell–cell junction. From those measured values, a value for the plasma membrane was derived (total cell − intracellular), and the ratio of plasma membrane to intracellular values was derived and plotted. Additionally, the length and width of cells were measured with the line tool and ROI manager. For both analyses, five clearly ARAF-expressing cells, as determined by HA staining, were analyzed per ×10 field. Five ×10 fields were acquired per condition per experiment. Experiments were conducted with cells from 3 independent thaws and transductions of HDLECs, for a total of 75 cells per condition. For western blotting of HDLECs with trametinib, 20,000 transduced HDLECs cells were plated into 96-well plates in the presence of increasing amounts of trametinib. Cells were cultured for 24 h in the presence of the drug, and then lysed with 40 mM HEPES pH 7.5, 120 mM NaCl, 0.3% CHAPS, 50 mM NaF, 1.5 mM NaVO3 and a protease inhibitor cocktail. Lysates were cleared by centrifugation at 20,000g for 5 min at 4 °C. Proteins were separated on 4–12% NuPAGE Bis-Tris gels. Blotting was performed using the antibodies described above.

Three-dimensional lymphatic spheroid sprouting assay. Multicellular spheroids for the lymphatic sprouting assay were initiated by seeding 7,500 HDLECs expressing ARAF-WT or ARAF-S214P into wells of a 96-well plate that were precoated with 1.5% agarose. Under these conditions, all of the HDLECs would aggregate into a single spheroid by 24 h. After formation, each spheroid was transferred into a gelling solution comprised of type I collagen (cat. no. 354236, Corning; final concentration = 1.5 mg ml−1; pH neutralized with NaOH) and trametinib at the indicated concentrations, which was then allowed to polymerize at 37 °C. Once solidified, Endothelial Cell Growth Medium MV 2 (without VEGFC) containing trametinib at the appropriate concentration was added onto the collagen gels. After 2 days of incubation, z-stack images with a step size of ~8.5 μm were taken of the embedded spheroids using an EVOS FL Auto Imaging System (Thermo Fisher Scientific). The numbers and lengths of capillary-like sprouts growing from each spheroid were measured using the software ImageJ (https://imagej.nih.gov/ij/).

MTT proliferation assay with transduced HDLECs. Proliferation of transduced HDLECs was measured using Cell Proliferation Kit I (MTT) from Roche Applied Science. Briefly, at 2 d post retroviral transduction, ARAF-WT- and -S214Pexpressing HDLECs were collected, counted and replated into flat-bottom 96-well plates at 10,000 cells per well in 100 μl of medium. At the indicated times after plating, 10 μl of the MTT was added to the appropriate wells, and incubated for 4 h at 37 °C. A 100 μl volume of the solubilization reagent was added followed by overnight incubation at 37 °C. Absorbance at 550 nm and 700 nm was measured on a Spectramax i3 Multi Mode plate reader (Molecular Devices), and A550 nm–A700 nm was calculated. A time point of 4 h after plating was included as an approximate measure of cells loaded into the experiment with minimal proliferation.

Transgenic expression of human ARAF in zebrafish. All procedures using zebrafish were approved by the Institutional Animal Care and Use Committee of CHOP (IAC 001154) and were in accordance with the Guide for the Care and Use of Laboratory Animals by the National Institutes of Health. Human mutant and WT ARAF cDNAs were cloned without stop codons into the pDONR221 vector; a zebrafish-adapted kozak sequence (GCAAACATGG) was used43. Expression constructs were assembled using a Tol2 backbone vector including a gateway cloning cassette44,45. Constructs were co-injected with Tol2 messenger RNA46. ARAF was expressed in vein and lymphatic vessels using the zebrafish mrc1a promoter, and expression was visualized by mCherry linked to ARAF by an autocatalytic V2a protein cleavage site. For imaging, larvae were mounted in lowmelting agarose, and multiple Z-images were taken with a Zeiss LSM710 confocal microscope using a ×20 lens. Confocal z-stacks of images were superimposed using Zeiss Zen software’s maximum intensity projection function. To analyze dilation of the TD, body segments separated by intersegmental lymphatic vessels with expression of the transgene in the TD were selected. Morphology was scored as normal (WT), moderate dilation (TD expanded but separate from the PCV) or severe dilation (TD and PCV not distinguishable in Z-projections). Images were compiled in ImageJ (Fiji). Each experiment was performed 3 times, and a total of 40 animals were analyzed.

Inhibitory drug treatment in zebrafish. Drug treatments were performed in 6-well plates with up to 20 larvae per group. Cobimetinib was diluted in embryo medium containing 0.01 M Tris pH 7.2 and 0.1% DMSO. Cobimetinib was used at 1 μM. p-ERK antibody staining in zebrafish. Fish were injected as described above and larvae with prominent WT or mutant ARAF/mcherry expression were selected for analysis. Larvae were fixed overnight in a 4% paraformaldehyde solution in PBS with Tween-20 (PBST). Larvae were washed with PBST and incubated in 2% Triton X-100 for 24 h at 4 °C. Then, larvae were blocked in 10% bovine serum and stained with phospho-ERK T202/Y204 antibody (cat. no. 9101, Cell Signaling Technology, 1:200) overnight at 4 °C, washed with PBST and stained with Alexa Fluor 488 goat anti-rabbit secondary antibody (cat. no. A11008, Thermo Fisher Scientific, 1:400).

Statistics. For all of the cell-based assays, significance Naporafenib was assessed by unpaired, two-tailed Student’s t-tests for comparison of two groups. Statistical analysis was performed with GraphPad Prism 7.0d software. The data are represented as box-and-whisker plots with boxes ranging from the 25th to 75th percentile, whiskers from the minimum to maximum and the median as the center, or as dot plots with bar graphs for mean ± s.e.m., as indicated. For all of the assays performed on HDLECs, three independent experiments were performed with independent transductions of HDLECs, except for the proliferation study, where no statistical analysis was performed. For the 14-3-3 protein association assay, three independent experiments were performed with independent transfection of HEK293T cells, while other results for HEK293T cells represent six independent experiments. All of the zebrafish-related assays were performed in three independent experiments and tested by unpaired, one-tailed Student’s t-tests for comparison of two groups.

References

1. Trenor, C. C. 3rd & Chaudry, G. Complex lymphatic anomalies. Semin. Pediatr. Surg. 23, 186–190 (2014).
2. Collins, F. S. & Varmus, H. A new initiative on precision medicine. N. Engl. J. Med. 372, 793–795 (2015).
3. Adams, D. M. et al. Efficacy and safety of sirolimus in the treatment of complicated vascular anomalies. Pediatrics 137, e20153257 (2016).
4. Hammill, A. M. et al. Sirolimus for the treatment of complicated vascular anomalies in children. Pediatr. Blood Cancer 57, 1018–1024 (2011).
5. McCormick, A., Rosenberg, S., Trier, K. & Balest, A. A case of a central conducting lymphatic anomaly responsive to sirolimus. Pediatrics 137, e20152694 (2016).
6. Hilliard, R. I., McKendry, J. B. & Phillips, M. J. Congenital abnormalities of the lymphatic system: a new clinical classification. Pediatrics 86, 988–994 (1990).
7. Levine, C. Primary disorders of the lymphatic vessels—a unified concept. J. Pediatr. Surg. 24, 233–240 (1989).
8. Smeltzer, D. M., Stickler, G. B. & Fleming, R. E. Primary lymphatic dysplasia in children: chylothorax, chylous ascites, and generalized lymphatic dysplasia. Eur. J. Pediatr. 145, 286–292 (1986).
9. Wassef, M. et al. Vascular anomalies classification: recommendations from the International Society for the Study of Vascular Anomalies. Pediatrics 136, e203–e214 (2015).
10. Chen, W., Adams, D., Patel, M., Gupta, A. & Dasgupta, R. Generalized lymphatic malformation with chylothorax: long-term management of a highly morbid condition in a pediatric patient. J. Pediatr. Surg. 48, e9–e12 (2013).
11. Lala, S. et al. Gorham–Stout disease and generalized lymphatic anomaly—clinical, radiologic, and histologic differentiation. Skeletal Radiol. 42, 917–924 (2013).
12. Clemens, R. K., Pfammatter, T., Meier, T. O., Alomari, A. I. & Amann-Vesti, B. R. Combined and complex vascular malformations. Vasa 44, 92–105 (2015).
13. Li, D. et al. Pathogenic variant in EPHB4 results in central conducting lymphatic anomaly. Hum. Mol. Genet. 27, 3233–3245 (2018).
14. Wellbrock, C., Karasarides, M. & Marais, R. The RAF proteins take centre stage. Nat. Rev. Mol. Cell Biol. 5, 875–885 (2004).
15. Lavoie, H. & Therrien, M. Regulation of RAF protein kinases in ERK signalling. Nat. Rev. Mol. Cell Biol. 16, 281–298 (2015).
16. Molzan, M. et al. Impaired binding of 14-3-3 to C-RAF in Noonan syndrome suggests new approaches in diseases with increased Ras signaling. Mol. Cell Biol. 30, 4698–4711 (2010).
17. Jung, H. M. et al. Development of the larval lymphatic system in zebrafish. Development 144, 2070–2081 (2017).
18. Karkkainen, M. J. et al. Vascular endothelial growth factor C is required for sprouting of the first lymphatic vessels from embryonic veins. Nat. Immunol. 5, 74–80 (2004).
19. Carmeliet, P. & Jain, R. K. Molecular mechanisms and clinical applications of angiogenesis. Nature 473, 298–307 (2011).
20. Karaman, S., Leppanen, V. M. & Alitalo, K. Vascular endothelial growth factor signaling in development and disease. Development 145, dev151019 (2018).
21. Potente, M. & Makinen, T. Vascular heterogeneity and specialization in development and disease. Nat. Rev. Mol. Cell Biol. 18, 477–494 (2017).
22. Coso, S., Bovay, E. & Petrova, T. V. Pressing the right buttons: signaling in lymphangiogenesis. Blood 123, 2614–2624 (2014).
23. Brouillard, P., Boon, L. & Vikkula, M. Genetics of lymphatic anomalies. J. Clin. Invest. 124, 898–904 (2014).
24. Bulow, L. et al. Hydrops, fetal pleural effusions and chylothorax in three patients with CBL mutations. Am. J. Med. Genet. A 167A, 394–399 (2015).
25. Gargano, G. et al. Hydrops fetalis in a preterm newborn heterozygous for the c.4A>G SHOC2 mutation. Am. J. Med. Genet. A 164A, 1015–1020 (2014).
26. Gos, M. et al. Contribution of RIT1 mutations to the pathogenesis of Noonan syndrome: four new cases and further evidence of heterogeneity. Am. J. Med. Genet. A 164A, 2310–2316 (2014).
27. Hanson, H. L. et al. Germline CBL mutation associated with a Noonan-like syndrome with primary lymphedema and teratoma associated with acquired uniparental isodisomy of chromosome 11q23. Am. J. Med. Genet. A 164A, 1003–1009 (2014).
28. Milosavljevic, D. et al. Two cases of RIT1 associated Noonan syndrome: further delineation of the clinical phenotype and review of the literature. Am. J. Med. Genet. A 170, 1874–1880 (2016).
29. Koenighofer, M. et al. Mutations in RIT1 cause Noonan syndrome – additional functional evidence and expanding the clinical phenotype. Clin. Genet. 89, 359–366 (2016).
30. Lee, K. A. et al. PTPN11 analysis for the prenatal diagnosis of Noonan syndrome in fetuses with abnormal ultrasound findings. Clin. Genet. 75, 190–194 (2009).
31. Croonen, E. A. et al. Prenatal diagnostic testing of the Noonan syndrome genes in fetuses with abnormal ultrasound findings. Eur. J. Hum. Genet. 21, 936–942 (2013).
32. Joyce, S. et al. The lymphatic phenotype in Noonan and cardiofaciocutaneous syndrome. Eur. J. Hum. Genet. 24, 690–696 (2016).
33. Yaoita, M. et al. Spectrum of mutations and genotype–phenotype analysis in Noonan syndrome patients with RIT1 mutations. Hum. Genet. 135, 209–222 (2016).
34. Lo, I. F. et al. Severe neonatal manifestations of Costello syndrome. J. Med. Genet. 45, 167–171 (2008).
35. Ebrahimi-Fakhari, D. et al. Congenital chylothorax as the initial presentation of PTPN11-associated Noonan syndrome. J. Pediatr. 185, 248–248.e1 (2017).
36. Morcaldi, G. et al. Lymphodysplasia and Kras mutation: a case report and literature review. Lymphology 48, 121–127 (2015).
37. Manevitz-Mendelson, E. et al. Somatic NRAS mutation in patient with generalized lymphatic anomaly. Angiogenesis 21, 287–298 (2018).
38. Barclay S. F. et al. A somatic activating NRAS variant associated with kaposiform lymphangiomatosis. Genet. Med. https://doi.org/10.1038/ s41436-018-0390-0 (2018).
39. Gao, J. et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci. Signal. 6, pl1 (2013).
40. Imielinski, M. et al. Oncogenic and sorafenib-sensitive ARAF mutations in lung adenocarcinoma. J. Clin. Invest. 124, 1582–1586 (2014).
41. Johannessen, C. M. et al. COT drives resistance to RAF inhibition through MAP kinase pathway reactivation. Nature 468, 968–972 (2010).
42. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
43. Grzegorski, S. J., Chiari, E. F., Robbins, A., Kish, P. E. & Kahana, A. Natural variability of Kozak sequences correlates with function in a zebrafish model. PLoS One 9, e108475 (2014).
44. Kwan, K. M. et al. The Tol2kit: a multisite gateway-based construction kit for Tol2 transposon transgenesis constructs. Dev. Dyn. 236, 3088–3099 (2007).
45. Villefranc, J. A., Amigo, J. & Lawson, N. D. Gateway compatible vectors for analysis of gene function in the zebrafish. Dev. Dyn. 236, 3077–3087 (2007).
46. Kawakami, K. & Shima, A. Identification of the Tol2 transposase of the medaka fish Oryzias latipes that catalyzes excision of a nonautonomous Tol2 element in zebrafish Danio rerio. Gene 240, 239–244 (1999).