Human mutations in the SOX18 gene give rise to Hypotrichosis-Lymphedema-Telangiectasia and Renal syndrome (HLTRS) (Irrthum et al., 2003; Moalem et al., 2015). Most disease-causing SOX18 mutations lead to a premature truncation at the C-terminus due to an early stop codon, producing defective protein products that act in a dominant negative fashion. The mutated SOX18 protein can still bind to chromatin but fails to recruit its protein partner(s) to transactivate target genes, and thereby disrupts the endogenous function of SOX18. Further, the dominant-negative mutant prevents other SOX transcription factors to act redundantly and rescue the molecular pathway (Hosking et al., 2009). From birth to adolescence HLTRS patients develop symptoms such as lymphedema and abnormalities to hair and nails for which very limited effective treatment options exist. Patients suffering from HLTRS display a certain level of variability in their phenotypic characteristics with distinct clinical manifestations (see for review Valenzuela et al., 2018) that likely depend on the specific type of SOX18 mutation and how it interferes with SOX18 transcriptional regulation.

A recent report of an HLTRS patient described symptoms so mild that the initial diagnosis failed to identify this syndrome, mostly due to the absence of lymphedema up until late adolescence (Wünnemann et al., 2016). Eventually, the patient was subjected to whole-exome sequencing, which led to the discovery of a premature stop-codon in SOX18, p.Q161*; with such a mutation the patient would be expected to present a severe phenotype. A search for a potential cause of the milder course of the syndrome in this patient revealed that she had been treated with high doses propranolol since a young age due to a significant dilation of the thoracic artery causing a high blood pressure. The milder phenotype observed in this HLTRS patient coincident with the course of the propranolol treatment raised the possibility that this drug has a SOX18-dependent molecular mode of action in addition to its β-blocking activity.

The main indications for propranolol therapy are hypertension, ischemic heart disease, arrhythmia, migraine and essential tremor. Sensitivity to propranolol does not always correlate with the expression levels of the β-adrenergic receptors (Wolter et al., 2014; MacDonald et al., 2006; Chang et al., 2015; Powe et al., 2010). Propranolol use in obstructive hypertrophic cardiomyopathy revealed its beneficial effects in infantile hemangioma (Léauté-Labrèze et al., 2008), raising the possibility that β-adrenergic receptor signaling, and thus by inference, propranolol, modulates VEGF levels (Chen et al., 2013; Ozeki et al., 2013; Lavine et al., 2013). Conclusive proof of such a mode of action is still lacking. Further, this drug is a racemic mixture with R(+) and S(-) enantiomers in equal proportions. The S(-) form displays most of the anti-β-adrenergic activity, whereas the R(+) enantiomer is carried over during drug synthesis. This makes the mode of action of propranolol even more elusive since therapeutic outcomes may result from a combination of effects from the two enantiomers.

In this study, we report an additional case of HLTRS successfully treated with propranolol and describe a R(+) enantiomer-specific mode of action that acts through interference of SOX18 transcriptional activity and thereby functions as an anti-angiogenic agent whereas the S(-) enantiomer does not. This finding paves the way for new therapeutic applications using SOX18 transcription factor as a molecular target in the context of vascular diseases.

Based on the observation that a previous case of HLTRS may have benefited from a long-term exposure to propranolol (Wünnemann et al., 2016), we considered propranolol therapy for another HLTRS patient (Bastaki et al., 2016) in an attempt to mitigate the severity of his symptoms. The patient was diagnosed with HLTRS at 11 months of age, after presenting with alopecia, edema and telangiectasia (Bastaki et al., 2016). Exome sequencing revealed that this patient had a truncation mutation in SOX18 (SOX18 c.492_505dup). In September 2015, at 17 months of age, he developed progressive pericardial effusion over 3 weeks, which required pericardiocentesis. Fluid analysis showed transudate. Two weeks after the procedure, pericardial effusion re-accumulated progressively over 6 weeks, despite strong anti-inflammatory regimens including 3 weeks of oral ibuprofen and 10 days of oral prednisone. Past clinical experience in HLTRS suggested the progression of this pericardial effusion was intractable. The effusion was large enough to cause hemodynamic compromise and to necessitate another pericardiocentesis. After obtaining parental consent, oral propranolol was introduced and increased gradually every 2–3 weeks from 0.8 mg/kg/day in three divided doses to reach a maximum of 4.1 mg/kg/day, as tolerated clinically (pulmonary wheezing). We observed spontaneous and rapid resolution of the pericardial effusion after reaching a dose threshold of 3 mg/kg/day. Bronchial side effects limited the maximal dosing in this patient to 4 mg/kg/day (Figure 1A and B). The positive outcome in this second HLTRS patient harboring SOX18 mutation suggested propranolol might act via an alternative pathway than the well characterized β-adrenergic receptor blockade.

Figure 1 with 1 supplement see all

Download asset Open asset

We next investigated the potential of propranolol to exert β-adrenergic independent effects in cancer cell lines and in human endothelial cells (Figure 1C–D). Survival of primary placentally-derived endothelial colony forming cells (ECFC) was measured in presence of increasing concentrations (5–500 μM) of propranolol and specific β1 (Atenolol) and β2 (ICI 118,551) adrenergic-receptor blockers. Results showed that pharmacological inhibition of specific β-adrenergic receptors did not replicate the marked effect propranolol had in the ECFC survival assay (Figure 1C).

To further explore β-adrenergic independent effects of propranolol we knocked down ADRB2 in a propranolol-sensitive angiosarcoma cell line (ISO-HAS, Figure 1—figure supplement 1A). The cytotoxic activity of propranolol (Figure 1D) and its enhancement of vinblastine cytotoxicity (Figure 1E) were unaltered when ADRB2 was silenced. These results show that the anti-survival and sensitization to chemotherapy activities of propranolol are independent of β-adrenergic receptor expression. Of note, ADRB1 is expressed at low levels in the angiosarcoma cell line (not shown) but is expressed at higher levels in the SHEP neuroblastoma cell line, where its expression can be reduced by siRNA (Figure 1—figure supplement 1B). Similar results were obtained when using a non-endothelial cell type (SHEP neuroblastoma cell line) whereby knock-down of ADRB1 and ADRB2 alone or in combination did not affect the ability of propranolol to block the proliferative response and its ability to enhance response to chemo-therapies (Figure 1—figure supplement 1C and 1D).

The observation that adrenergic receptors are not required for the cellular response to propranolol combined with the clinical benefit of propranolol treatment in the context of HLTRS suggests that this FDA approved drug may act by directly modulating SOX18 protein activity .

Although SOX18 is involved in a variety of patho-physiological processes (Olbromski et al., 2018), HLTRS is a rare and severe vascular disorder. In order to further investigate propranolol in this disease scenario, we used a mouse model of HLTRS, SOX18 Ragged Opossum (RaOp) (Pennisi et al., 2000; Slee, 1957). This mouse mutant is considered a murine counterpart of human HLTRS syndrome since it displays the same range of cardio-vascular and hair follicle defects and exhibits the same type of dominant-negative mutation in the Sox18 gene (Pennisi et al., 2000; Downes et al., 2009; François et al., 2008; Villani et al., 2017). The Ragged mouse is characterized by pathologic corneal neo-vascularisation (CNV, Figure 2—figure supplement 1A-C) (François and Ramchandran, 2012). This phenotypic hallmark offers an indirect biological readout in vivo for aberrant SOX18 function in the control of vessel outgrowth. Propranolol was tested on CNV in the RaOp mouse model to validate its therapeutic potential. Because of early postnatal onset of HLTRS defects (between postnatal day (P)8–22), oral treatment with propranolol (25 mg/kg/day), or vehicle PBS, was initiated at P8 and continued daily through to P28. Both wild type and mutant littermates were subjected to identical treatment schemes in order to safeguard consistent maternal feeding and grooming.

Treatment with propranolol (21 days) did not affect the gross morphological appearance of either wild type or RaOp pups, nor did it significantly affect the weight of the RaOp animals (Figure 1—figure supplement 2A and B, n = 5–10 mice per group). We did not observe any obvious side effects, consistent with the safety profile of propranolol established over the past decades.

Propranolol treatment of RaOp mice, however, led to an almost complete rescue of the CNV phenotype (Figure 2A). The corneas from propranolol-treated mutant animals were devoid of blood vessels (Figure 2A, Endomucin- and ERG-positive cells) in five out of five mice, and one mutant pup only had a mild penetration of blood vessels into the cornea (Figure 2A). These pre-clinical data indicate that propranolol treatment is an effective therapeutic strategy to block aberrant vessel growth in a mouse model of HLTRS and is indicative of on-target engagement on SOX18 protein in vivo.

Figure 2 with 2 supplements see all

Download asset Open asset

These in vivo data prompted us to assess the transcriptional activity of Sox18 in presence of propranolol. Sox18 was overexpressed in COS-7 cells, along with a synthetic construct containing a luciferase reporter gene fused to the Vcam1 promoter fragment, a direct target of Sox18 (Hosking et al., 2004). In this assay, we interfere with wild type Sox18 function using the RaOp dominant-negative protein. We found that RaOp completely abolished the activity of the wild type protein at ratios as low as 1:30 RaOp:Sox18 WT (Figure 2B). We next tested whether propranolol could rescue the dominant-negative inhibitory effect of RaOp reproduced in this cell-based system. Propranolol restored SOX18 wild type functionality in a dose-dependent manner; it significantly increased Vcam1 reporter promoter activity to base line levels at 15 µM (p-value≤0.001) (Figure 2B).

Lastly, to demonstrate the molecular mode of action of propranolol on SOX18 protein partner recruitment, we used an ALPHAScreen assay to measure pairwise protein-protein interaction (PPI). SOX18 homodimer and SOX18/RaOp (Q161*) heterodimer formation was measured in absence or presence of propranolol (Figure 2C). Propranolol showed a mild efficacy at disrupting SOX18 homodimer and was able to interfere more effectively with the assembly of the non functional SOX18/RaOp protein complex. An effector of the NOTCH signaling pathway (RBPJ) was recently shown to act as a SOX18 protein partner (Overman et al., 2017; Fontaine et al., 2017). In addition to disrupting SOX18 homodimer assembly (Figure 2C), we found that propranolol inhibits the association of SOX18 with RBPJ at low micro-molar range (5 μM) (Figure 2—figure supplement 2C). Taken together these in vivo pre-clinical data and in vitro results strongly suggest that one mechanism by which propranolol mediates its anti-angiogenic effects is via direct interference with a SOX18-dependent transcriptional activation.

We have previously identified, developed and validated a small compound, Sm4, that disrupts SOX18 homodimer formation in turn affecting SOX18 ability to recruit RBPJ. Further, we have shown potential therapeutic applications in managing vascular over-growth in pre-clinical model of breast cancer metastasis (Overman et al., 2017). When using Sm4 as a reference, propranolol was equally effective at disrupting SOX18/RBPJ PPI however not as effective at blocking SOX18 homodimer formation (Figure 2—figure supplement 2C). Altogether, these results suggest that the same molecular mode of action of propranolol accounts for both the rescue of dominant-negative SOX18 loss of function by disruption of the non-functional SOX18/RaOp (Q161*) complex, and the disruption of functional SOX18 wild type PPIs, namely SOX18/RBPJ and SOX18/SOX18 dimers. In the context of the Ragged heterozygous mutation dominant negative mutation, the rescue process corresponds to the inhibition by propranolol of an inhibitor (RaOp) which in turn results to the partial restoration of the SOX18 wild type active form.

This novel mode of action of propranolol, independent of its anti-adrenergic effects, prompted us to examine its mechanism of action in infantile hemangiomas, the most common vascular tumor in infancy. Although the molecular mode of action of propranolol in infantile hemangioma therapy is largely unknown, it is the mainstay of treatment for infantile hemangioma. (Léauté-Labrèze et al., 2008; Léauté-Labrèze et al., 2015a; Léauté-Labrèze et al., 2015b) A variety of propranolol-induced cellular effects have been observed in infantile hemangioma-derived stem cells (HemSCs), hemangioma endothelial cells (HemECs) and hemangioma pericytes (Lamy et al., 2010; Lee et al., 2014; Tu et al., 2013; Wong et al., 2012). To learn if SOX18 is involved, we first explored the expression of SOX18 in vivo in hemangioma tissue sections from three patients and in vitro in different cell populations derived from infantile hemangiomas: SOX18 protein was observed in a sup-population of CD31-positive endothelial cells (Figure 3A, arrows) and SOX18 mRNA was expressed in ECFC (positive control) and in HemEC, whereas HemSC showed low expression and hemangioma pericytes were completely devoid of SOX18 transcripts (Figure 3—figure supplement 1).

Figure 3 with 4 supplements see all

Download asset Open asset

Figure 3—source data 1

Supplemental table for Figure 3D (p-values).

https://doi.org/10.7554/eLife.43026.012

Download elife-43026-fig3-data1-v1.docx

To interrogate whether propranolol inhibits SOX18 in this setting (Figure 3B), we induced patient derived HemSCs to differentiate into HemECs using VEGF-B and low serum media (Khan et al., 2008). CDH5 (a pan-endothelial marker), SOX18, and ADBR2 mRNAs were increased in VEGF-B-treated HemSC (Figure 3C) further showing a process by which differentiation of stem cells results in endothelial cells expressing SOX18. This was extended to other genes such as CD31, NOTCH1, PLXND1 and VEGFR1 (Table 1 and Figure 3—source data 1), each of which are expressed in infantile hemangioma endothelium in vivo (Boscolo et al., 2011; Nakayama et al., 2015; Wu et al., 2010), further confirming the successful differentiation of HemSCs into HemECs (Figure 3D and Figure 3—figure supplement 2A). Using this differentiation assay as a read-out, we next evaluated the capacity of propranolol to inhibit the differentiation of HemSCs isolated from four infantile hemangioma patients. The pharmacologic treatments of HemSC included: propranolol, Sm4 (SOX18 small molecule inhibitor), aspirin (a negative control [Overman et al., 2017; Fontaine et al., 2017]) compared to vehicle control. Of note, SOX18 has been shown to directly up-regulate Notch1 transcription via an intronic enhancer during arterial specification in vertebrates (Chiang et al., 2017), hence transcriptional output of this target gene indirectly reflects SOX18 activity. Propranolol significantly decreased VEGF-B induced expression of all five endothelial markers in HemSC to levels also achieved by Sm4 as compared to controls (see p-values in Figure 3—source data 1).

Propranolol exists as a 1:1 mixture of the S(-) and R(+) enantiomers. The S(-) form displays potent anti-β-adrenergic activity whereas the R(+) enantiomer presents strongly reduced antagonist activity towards β2 adrenergic receptors on peripheral arteries (Stoschitzky et al., 1995). To test whether HemSC were affected by propranolol in a stereoselective manner, we assessed VEGF-B-induced HemSC to HemEC differentiation from two infantile hemangioma in the presence of separated enantiomers.

The R(+) enantiomer recapitulated propranolol and Sm4-induced inhibition on all markers, whereas the S(-) enantiomer showed only weak or no inhibition on hemangioma markers (Figure 3D and Figure 3—source data 1). These results are consistent with a SOX18-dependent, β-adrenergic receptor independent blockade of hemangioma endothelial differentiation, which strongly suggests that blockade of SOX18 activity is the main mode of action of propranolol in this setting. Further, analysis of VEGF-R2 expression and phosphorylation levels in ECFC and HemSC to EC differentiation assay ruled out that the propranolol racemic mixture, the R(+) enantiomer or Sm4 interfere with the VEGF signaling pathway (Figure 3—figure supplement 2B-D).

As a final demonstration of the therapeutic utility of the R(+) enantiomer we performed a corneal neo-vascularisation rescue assay using the Ragged Opossum mutant mouse model. In this in vivo setting both propranolol and the R(+) enantiomer prevented aberrant vessel growth in the corneal tissue of the heterozygous mutant animals. Of note the S(-) enantiomer was also active at disrupting corneal vessel outgrowth, suggesting that conformational changes that occur in the SOX18 mutated protein may attenuate the discrepancy in enantiomeric activities previously observed in the hemangioma endothelial differentiation model (Figure 3—figure supplement 3).

In order to investigate whether propranolol treatment might improve the morphogenesis defects of other vascular beds, we analysed the patterning of the lymphatic vasculature in the ear skin tissue of the adult RaOp animals. As previously reported the RaOp mouse presented with a highly disorganized lymphatic vascular network (François et al., 2008) (Figure 3—figure supplement 4, top graph). In the skin, neither the propranolol racemic mixture nor its enantiomers rescued the lymphatic vascular phenotype caused by the RaOp mutation (Figure 3—figure supplement 4); this is most likely due to the post-natal timing of the treatment: lymphatic vascular defects are acquired in the RaOp mutant during the initial steps of embryonic lymphangiogenesis (early to mid-organogenesis), which precedes the drug treatment window.

In summary, we suggest the R(+) enantiomer of propranolol is a potent selective inhibitor of SOX18 activity and it is sufficient to achieve the effects of propranolol racemic mixture in the context of infantile hemangioma and other SOX18-dependent vascular diseases.

Despite being the front line therapy for infantile hemangioma, approximately 20% of the patient population fails to respond to propranolol treatment and/or experience tumor regrowth when the medication is discontinued too early (Bagazgoitia et al., 2011; Xiao et al., 2013). Our finding opens up new treatment options using enantiopure R(+) propranolol instead of the racemic mixture. This could potentially allow a dose adjustment to achieve optimal benefit for the patients while mitigating adverse events that may result from long-term propranolol treatment, particularly in the refractory population of patients that currently do not respond to the highest dose of propranolol. In addition, the side effects associated with long-term exposure to propranolol could be greatly reduced if β-adrenergic receptors blockage would be minimized.

This discovery is of broad interest because propranolol, comprised of 2 enantiomers, has been used clinically for decades. Clinical trials of the R(+) versus the S(-) form have been performed and deemed safe with full toxicology data available (Stensrud and Sjaastad, 1976). Pharmacokinetic studies suggest that propranolol therapy results in an accumulation of the S(-) enantiomer following drug metabolism, which could be problematic if one is interested in the activity of the R(+) form (Mehvar and Brocks, 2001). Indeed toxicology studies further suggest that propranolol in its racemic form is more toxic than either enantiomer alone (Stoschitzky et al., 1992). This study presents indirect evidence that in a HLTRS patient treated with propranolol, SOX18 wild type function is at least partly restored. The observation that the S(-) enantiomer also has the capability to rescue the corneal neo-vascularisation phenotype in a pre-clinical model of HTLRS suggests that the effects of propranolol as a racemic mixture are likely to be a combination of the blockade of beta-adrenergic receptor activity and SOX18 RaOp mutant protein inhibition. Of importance, in the hemangioma HemSC-EC differentiation assay the R(+) enantiomer was as efficient as propranolol and more efficient than the S (-) enantiomer at blocking this differentiation process. This may suggest that the SOX18 wild type protein blockade displays some enantio-selectivity, whereas conformational changes in the SOX18/RaOp complex caused by the RaOp mutation sensitize the dominant-negative protein complex to both R(+) and S(-) enantiomers.

The studies in human patients with HLTRS and the murine RaOP model suggest that there may be critical time windows in which treatment with R(+) propranolol would be most efficient as an anti-angiogenic compound via a blockade of SOX18 transcriptional activity.

The low number of patients with HLTRS – less than 10 documented survivors have been reported – makes it difficult to establish genotype-phenotype relationships in greater detail. In contrast, the unambiguous phenotypic rescue by propranolol racemic mixture in two human patients and by the R(+) enantiomer in the Ragged Opossum mouse model is strongly supported by our in vitro mechanistic results. Our findings outline that the R(+) enantiomer is a SOX18 small molecule inhibitor with weak anti-beta adrenergic activity whereas the S(-) molecule is both a beta-blocker and a weak SOX18 inhibitor which introduces variability in net efficacy of propranolol treatment.

The strong monogenic effect of human SOX18 mutations and their rescue with propranolol allowed us to identify its mode of action. We suggest the dosing and duration of propranolol therapy may be adjusted by using R(+) propranolol in potential future trials of HLTRS and infantile hemangioma.

All data generated or analysed during this study are included in the manuscript and supporting files.

1. 1. Bagazgoitia L
   2. Hernández-Martín A
   3. Torrelo A

   (2011) Recurrence of infantile hemangiomas treated with propranolol

   Pediatric Dermatology 28:658–662.

   https://doi.org/10.1111/j.1525-1470.2011.01644.x

   • PubMed
   • Google Scholar
2. 1. Bastaki F
   2. Mohamed M
   3. Nair P
   4. Saif F
   5. Tawfiq N
   6. Al-Ali MT
   7. Brandau O
   8. Hamzeh AR

   (2016) A novel SOX18 mutation uncovered in jordanian patient with hypotrichosis-lymphedema-telangiectasia syndrome by whole exome sequencing

   Molecular and Cellular Probes 30:18–21.

   https://doi.org/10.1016/j.mcp.2015.11.005

   • PubMed
   • Google Scholar
3. 1. Boscolo E
   2. Mulliken JB
   3. Bischoff J

   (2011) VEGFR-1 mediates endothelial differentiation and formation of blood vessels in a murine model of infantile hemangioma

   The American Journal of Pathology 179:2266–2277.

   https://doi.org/10.1016/j.ajpath.2011.07.040

   • PubMed
   • Google Scholar
4. 1. Cao DX
   2. Wu GH
   3. Yang ZA
   4. Zhang B
   5. Jiang Y
   6. Han YS
   7. He GD
   8. Zhuang QL
   9. Wang YF
   10. Huang ZL
   11. Xi QL

   (2010) Role of beta1-adrenoceptor in increased lipolysis in cancer cachexia

   Cancer Science 101:1639–1645.

   https://doi.org/10.1111/j.1349-7006.2010.01582.x

   • PubMed
   • Google Scholar
5. 1. Chang PY
   2. Huang WY
   3. Lin CL
   4. Huang TC
   5. Wu YY
   6. Chen JH
   7. Kao CH

   (2015) Propranolol reduces cancer risk: a Population-Based cohort study

   Medicine 94:e1097.

   https://doi.org/10.1097/MD.0000000000001097

   • PubMed
   • Google Scholar
6. 1. Chen XD
   2. Ma G
   3. Huang JL
   4. Chen H
   5. Jin YB
   6. Ye XX
   7. Hu XJ
   8. Lin XX

   (2013) Serum-level changes of vascular endothelial growth factor in children with infantile hemangioma after oral propranolol therapy

   Pediatric Dermatology 30:549–553.

   https://doi.org/10.1111/pde.12192

   • PubMed
   • Google Scholar
7. 1. Chiang IK
   2. Fritzsche M
   3. Pichol-Thievend C
   4. Neal A
   5. Holmes K
   6. Lagendijk A
   7. Overman J
   8. D'Angelo D
   9. Omini A
   10. Hermkens D
   11. Lesieur E
   12. Liu K
   13. Ratnayaka I
   14. Corada M
   15. Bou-Gharios G
   16. Carroll J
   17. Dejana E
   18. Schulte-Merker S
   19. Hogan B
   20. Beltrame M
   21. De Val S
   22. Francois M

   (2017) SoxF factors induce Notch1 expression via direct transcriptional regulation during early arterial development

   Development 144:2629–2639.

   https://doi.org/10.1242/dev.146241

   • PubMed
   • Google Scholar
8. 1. Chou TC
   2. Talalay P

   (1984) Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors

   Advances in Enzyme Regulation 22:27–55.

   https://doi.org/10.1016/0065-2571(84)90007-4

   • PubMed
   • Google Scholar
9. 1. Downes M
   2. François M
   3. Ferguson C
   4. Parton RG
   5. Koopman P

   (2009) Vascular defects in a mouse model of hypotrichosis-lymphedema-telangiectasia syndrome indicate a role for SOX18 in blood vessel maturation

   Human Molecular Genetics 18:2839–2850.

   https://doi.org/10.1093/hmg/ddp219

   • PubMed
   • Google Scholar
10. 1. Fontaine F
    2. Overman J
    3. Moustaqil M
    4. Mamidyala S
    5. Salim A
    6. Narasimhan K
    7. Prokoph N
    8. Robertson AAB
    9. Lua L
    10. Alexandrov K
    11. Koopman P
    12. Capon RJ
    13. Sierecki E
    14. Gambin Y
    15. Jauch R
    16. Cooper MA
    17. Zuegg J
    18. Francois M

    (2017) Small-Molecule inhibitors of the SOX18 transcription factor

    Cell Chemical Biology 24:346–359.

    https://doi.org/10.1016/j.chembiol.2017.01.003

    • PubMed
    • Google Scholar
11. 1. François M
    2. Caprini A
    3. Hosking B
    4. Orsenigo F
    5. Wilhelm D
    6. Browne C
    7. Paavonen K
    8. Karnezis T
    9. Shayan R
    10. Downes M
    11. Davidson T
    12. Tutt D
    13. Cheah KS
    14. Stacker SA
    15. Muscat GE
    16. Achen MG
    17. Dejana E
    18. Koopman P

    (2008) Sox18 induces development of the lymphatic vasculature in mice

    Nature 456:643–647.

    https://doi.org/10.1038/nature07391

    • PubMed
    • Google Scholar
12. 1. François M
    2. Ramchandran R

    (2012) Studies on Axenfeld-Rieger syndrome patients and mice reveal Foxc1's role in corneal neovascularization

    PNAS 109:1818–1819.

    https://doi.org/10.1073/pnas.1119291109

    • PubMed
    • Google Scholar
13. 1. Gagoski D
    2. Mureev S
    3. Giles N
    4. Johnston W
    5. Dahmer-Heath M
    6. Škalamera D
    7. Gonda TJ
    8. Alexandrov K

    (2015) Gateway-compatible vectors for high-throughput protein expression in pro- and eukaryotic cell-free systems

    Journal of Biotechnology 195:1–7.

    https://doi.org/10.1016/j.jbiotec.2014.12.006

    • PubMed
    • Google Scholar
14. 1. Hosking BM
    2. Wang SC
    3. Downes M
    4. Koopman P
    5. Muscat GE

    (2004) The VCAM-1 gene that encodes the vascular cell adhesion molecule is a target of the Sry-related high mobility group box gene, Sox18

    Journal of Biological Chemistry 279:5314–5322.

    https://doi.org/10.1074/jbc.M308512200

    • PubMed
    • Google Scholar
15. 1. Hosking B
    2. François M
    3. Wilhelm D
    4. Orsenigo F
    5. Caprini A
    6. Svingen T
    7. Tutt D
    8. Davidson T
    9. Browne C
    10. Dejana E
    11. Koopman P

    (2009) Sox7 and Sox17 are strain-specific modifiers of the lymphangiogenic defects caused by Sox18 dysfunction in mice

    Development 136:2385–2391.

    https://doi.org/10.1242/dev.034827

    • PubMed
    • Google Scholar
16. 1. Irrthum A
    2. Devriendt K
    3. Chitayat D
    4. Matthijs G
    5. Glade C
    6. Steijlen PM
    7. Fryns JP
    8. Van Steensel MA
    9. Vikkula M

    (2003) Mutations in the transcription factor gene SOX18 underlie recessive and dominant forms of hypotrichosis-lymphedema-telangiectasia

    The American Journal of Human Genetics 72:1470–1478.

    https://doi.org/10.1086/375614

    • PubMed
    • Google Scholar
17. 1. Khan ZA
    2. Boscolo E
    3. Picard A
    4. Psutka S
    5. Melero-Martin JM
    6. Bartch TC
    7. Mulliken JB
    8. Bischoff J

    (2008) Multipotential stem cells recapitulate human infantile hemangioma in immunodeficient mice

    Journal of Clinical Investigation 118:2592–2599.

    https://doi.org/10.1172/JCI33493

    • PubMed
    • Google Scholar
18. 1. Kovtun O
    2. Mureev S
    3. Jung W
    4. Kubala MH
    5. Johnston W
    6. Alexandrov K

    (2011) Leishmania cell-free protein expression system

    Methods 55:58–64.

    https://doi.org/10.1016/j.ymeth.2011.06.006

    • PubMed
    • Google Scholar
19. 1. Lamy S
    2. Lachambre MP
    3. Lord-Dufour S
    4. Béliveau R

    (2010) Propranolol suppresses angiogenesis in vitro: inhibition of proliferation, migration, and differentiation of endothelial cells

    Vascular Pharmacology 53:200–208.

    https://doi.org/10.1016/j.vph.2010.08.002

    • PubMed
    • Google Scholar
20. 1. Lavine JA
    2. Sang Y
    3. Wang S
    4. Ip MS
    5. Sheibani N

    (2013) Attenuation of choroidal neovascularization by β(2)-adrenoreceptor antagonism

    JAMA Ophthalmology 131:376–382.

    https://doi.org/10.1001/jamaophthalmol.2013.1476

    • PubMed
    • Google Scholar
21. 1. Léauté-Labrèze C
    2. Dumas de la Roque E
    3. Hubiche T
    4. Boralevi F
    5. Thambo JB
    6. Taïeb A

    (2008) Propranolol for severe hemangiomas of infancy

    New England Journal of Medicine 358:2649–2651.

    https://doi.org/10.1056/NEJMc0708819

    • PubMed
    • Google Scholar
22. 1. Léauté-Labrèze C
    2. Hoeger P
    3. Mazereeuw-Hautier J
    4. Guibaud L
    5. Baselga E
    6. Posiunas G
    7. Phillips RJ
    8. Caceres H
    9. Lopez Gutierrez JC
    10. Ballona R
    11. Friedlander SF
    12. Powell J
    13. Perek D
    14. Metz B
    15. Barbarot S
    16. Maruani A
    17. Szalai ZZ
    18. Krol A
    19. Boccara O
    20. Foelster-Holst R
    21. Febrer Bosch MI
    22. Su J
    23. Buckova H
    24. Torrelo A
    25. Cambazard F
    26. Grantzow R
    27. Wargon O
    28. Wyrzykowski D
    29. Roessler J
    30. Bernabeu-Wittel J
    31. Valencia AM
    32. Przewratil P
    33. Glick S
    34. Pope E
    35. Birchall N
    36. Benjamin L
    37. Mancini AJ
    38. Vabres P
    39. Souteyrand P
    40. Frieden IJ
    41. Berul CI
    42. Mehta CR
    43. Prey S
    44. Boralevi F
    45. Morgan CC
    46. Heritier S
    47. Delarue A
    48. Voisard JJ

    (2015a) A randomized, controlled trial of oral propranolol in infantile hemangioma

    New England Journal of Medicine 372:735–746.

    https://doi.org/10.1056/NEJMoa1404710

    • PubMed
    • Google Scholar
23. 1. Léauté-Labrèze C
    2. Voisard JJ
    3. Moore N

    (2015b) Oral propranolol for infantile hemangioma

    The New England Journal of Medicine 373:284–285.

    https://doi.org/10.1056/NEJMc1503811

    • PubMed
    • Google Scholar
24. 1. Lee D
    2. Boscolo E
    3. Durham JT
    4. Mulliken JB
    5. Herman IM
    6. Bischoff J

    (2014) Propranolol targets the contractility of infantile haemangioma-derived pericytes

    British Journal of Dermatology 171:1129–1137.

    https://doi.org/10.1111/bjd.13048

    • PubMed
    • Google Scholar
25. 1. MacDonald ML
    2. Lamerdin J
    3. Owens S
    4. Keon BH
    5. Bilter GK
    6. Shang Z
    7. Huang Z
    8. Yu H
    9. Dias J
    10. Minami T
    11. Michnick SW
    12. Westwick JK

    (2006) Identifying off-target effects and hidden phenotypes of drugs in human cells

    Nature Chemical Biology 2:329–337.

    https://doi.org/10.1038/nchembio790

    • PubMed
    • Google Scholar
26. 1. Mehvar R
    2. Brocks DR

    (2001)

    Stereospecific pharmacokinetics and pharmacodynamics of beta-adrenergic blockers in humans

    Journal of Pharmacy & Pharmaceutical Sciences : A Publication of the Canadian Society for Pharmaceutical Sciences, Societe Canadienne Des Sciences Pharmaceutiques 4:185–200.

    • PubMed
    • Google Scholar
27. 1. Moalem S
    2. Brouillard P
    3. Kuypers D
    4. Legius E
    5. Harvey E
    6. Taylor G
    7. Francois M
    8. Vikkula M
    9. Chitayat D

    (2015) Hypotrichosis-lymphedema-telangiectasia-renal defect associated with a truncating mutation in the SOX18 gene

    Clinical Genetics 87:378–382.

    https://doi.org/10.1111/cge.12388

    • PubMed
    • Google Scholar
28. 1. Mureev S
    2. Kovtun O
    3. Nguyen UT
    4. Alexandrov K

    (2009) Species-independent translational leaders facilitate cell-free expression

    Nature Biotechnology 27:747–752.

    https://doi.org/10.1038/nbt.1556

    • PubMed
    • Google Scholar
29. 1. Nakayama H
    2. Huang L
    3. Kelly RP
    4. Oudenaarden CR
    5. Dagher A
    6. Hofmann NA
    7. Moses MA
    8. Bischoff J
    9. Klagsbrun M

    (2015) Infantile hemangioma-derived stem cells and endothelial cells are inhibited by class 3 semaphorins

    Biochemical and Biophysical Research Communications 464:126–132.

    https://doi.org/10.1016/j.bbrc.2015.06.087

    • PubMed
    • Google Scholar
30. 1. Olbromski M
    2. Podhorska-Okołów M
    3. Dzięgiel P

    (2018) Role of the SOX18 protein in neoplastic processes

    Oncology Letters 16:1383–1389.

    https://doi.org/10.3892/ol.2018.8819

    • PubMed
    • Google Scholar
31. 1. Overman J
    2. Fontaine F
    3. Moustaqil M
    4. Mittal D
    5. Sierecki E
    6. Sacilotto N
    7. Zuegg J
    8. Robertson AA
    9. Holmes K
    10. Salim AA
    11. Mamidyala S
    12. Butler MS
    13. Robinson AS
    14. Lesieur E
    15. Johnston W
    16. Alexandrov K
    17. Black BL
    18. Hogan BM
    19. De Val S
    20. Capon RJ
    21. Carroll JS
    22. Bailey TL
    23. Koopman P
    24. Jauch R
    25. Smyth MJ
    26. Cooper MA
    27. Gambin Y
    28. Francois M

    (2017) Pharmacological targeting of the transcription factor SOX18 delays breast cancer in mice

    eLife 6:e21221.

    https://doi.org/10.7554/eLife.21221

    • PubMed
    • Google Scholar
32. 1. Ozeki M
    2. Kanda K
    3. Kawamoto N
    4. Ohnishi H
    5. Fujino A
    6. Hirayama M
    7. Kato Z
    8. Azuma E
    9. Fukao T
    10. Kondo N

    (2013) Propranolol as an alternative treatment option for pediatric lymphatic malformation

    The Tohoku Journal of Experimental Medicine 229:61–66.

    https://doi.org/10.1620/tjem.229.61

    • PubMed
    • Google Scholar
33. 1. Pasquier E
    2. Street J
    3. Pouchy C
    4. Carre M
    5. Gifford AJ
    6. Murray J
    7. Norris MD
    8. Trahair T
    9. Andre N
    10. Kavallaris M

    (2013) β-blockers increase response to chemotherapy via direct antitumour and anti-angiogenic mechanisms in neuroblastoma

    British Journal of Cancer 108:2485–2494.

    https://doi.org/10.1038/bjc.2013.205

    • PubMed
    • Google Scholar
34. 1. Pennisi D
    2. Gardner J
    3. Chambers D
    4. Hosking B
    5. Peters J
    6. Muscat G
    7. Abbott C
    8. Koopman P

    (2000) Mutations in Sox18 underlie cardiovascular and hair follicle defects in ragged mice

    Nature Genetics 24:434–437.

    https://doi.org/10.1038/74301

    • PubMed
    • Google Scholar
35. 1. Powe DG
    2. Voss MJ
    3. Zänker KS
    4. Habashy HO
    5. Green AR
    6. Ellis IO
    7. Entschladen F

    (2010) Beta-blocker drug therapy reduces secondary cancer formation in breast cancer and improves cancer specific survival

    Oncotarget 1:628–638.

    https://doi.org/10.18632/oncotarget.101009

    • PubMed
    • Google Scholar
36. 1. Sierecki E
    2. Giles N
    3. Polinkovsky M
    4. Moustaqil M
    5. Alexandrov K
    6. Gambin Y

    (2013) A cell-free approach to accelerate the study of protein-protein interactions in vitro

    Interface Focus 3:20130018.

    https://doi.org/10.1098/rsfs.2013.0018

    • PubMed
    • Google Scholar
37. 1. Sierecki E
    2. Stevers LM
    3. Giles N
    4. Polinkovsky ME
    5. Moustaqil M
    6. Mureev S
    7. Johnston WA
    8. Dahmer-Heath M
    9. Skalamera D
    10. Gonda TJ
    11. Gabrielli B
    12. Collins BM
    13. Alexandrov K
    14. Gambin Y

    (2014) Rapid mapping of interactions between human SNX-BAR proteins measured in vitro by AlphaScreen and single-molecule spectroscopy

    Molecular & Cellular Proteomics 13:2233–2245.

    https://doi.org/10.1074/mcp.M113.037275

    • PubMed
    • Google Scholar
38. 1. Slee J

    (1957) The morphology and development of ragged— a mutant affecting the skin and hair of the house mouse II. Genetics, Embryology and Gross Juvenile Morphology

    Journal of Genetics 55:570–584.

    https://doi.org/10.1007/BF02984073

    • Google Scholar
39. 1. Stensrud P
    2. Sjaastad O

    (1976) Short-term clinical trial of phopranolol in racemic form (Inderal), D-propranolol and placebo in migraine

    Acta Neurologica Scandinavica 53:229–232.

    https://doi.org/10.1111/j.1600-0404.1976.tb04342.x

    • PubMed
    • Google Scholar
40. 1. Stoschitzky K
    2. Lindner W
    3. Egginger G
    4. Brunner F
    5. Obermayer-Pietsch B
    6. Passath A
    7. Klein W

    (1992) Racemic (R,S)-propranolol versus half-dosed optically pure (S)-propranolol in humans at steady state: hemodynamic effects, plasma concentrations, and influence on thyroid hormone levels

    Clinical Pharmacology and Therapeutics 51:445–453.

    https://doi.org/10.1038/clpt.1992.45

    • PubMed
    • Google Scholar
41. 1. Stoschitzky K
    2. Lindner W
    3. Kiowski W

    (1995) Stereoselective vascular effects of the (R)- and (S)-enantiomers of propranolol and atenolol

    Journal of Cardiovascular Pharmacology 25:268–272.

    https://doi.org/10.1097/00005344-199502000-00012

    • PubMed
    • Google Scholar
42. 1. Tu JB
    2. Ma RZ
    3. Dong Q
    4. Jiang F
    5. Hu XY
    6. Li QY
    7. Pattar P
    8. Zhang H

    (2013) Induction of apoptosis in infantile hemangioma endothelial cells by propranolol

    Experimental and Therapeutic Medicine 6:574–578.

    https://doi.org/10.3892/etm.2013.1159

    • PubMed
    • Google Scholar
43. 1. Valenzuela I
    2. Fernández-Alvarez P
    3. Plaja A
    4. Ariceta G
    5. Sabaté-Rotés A
    6. García-Arumí E
    7. Vendrell T
    8. Tizzano E

    (2018) Further delineation of the SOX18-related hypotrichosis, lymphedema, telangiectasia syndrome (HTLS)

    European Journal of Medical Genetics 61:269–272.

    https://doi.org/10.1016/j.ejmg.2018.01.001

    • PubMed
    • Google Scholar
44. 1. Villani R
    2. Hodgson S
    3. Legrand J
    4. Greaney J
    5. Wong HY
    6. Pichol-Thievend C
    7. Adolphe C
    8. Wainwight B
    9. Francois M
    10. Khosrotehrani K

    (2017) Dominant-negative Sox18 function inhibits dermal papilla maturation and differentiation in all murine hair types

    Development 144:1887–1895.

    https://doi.org/10.1242/dev.143917

    • PubMed
    • Google Scholar
45. 1. Willems E
    2. Leyns L
    3. Vandesompele J

    (2008) Standardization of real-time PCR gene expression data from independent biological replicates

    Analytical Biochemistry 379:127–129.

    https://doi.org/10.1016/j.ab.2008.04.036

    • PubMed
    • Google Scholar
46. 1. Wolter JK
    2. Wolter NE
    3. Blanch A
    4. Partridge T
    5. Cheng L
    6. Morgenstern DA
    7. Podkowa M
    8. Kaplan DR
    9. Irwin MS

    (2014) Anti-tumor activity of the beta-adrenergic receptor antagonist propranolol in neuroblastoma

    Oncotarget 5:161–172.

    https://doi.org/10.18632/oncotarget.1083

    • PubMed
    • Google Scholar
47. 1. Wong A
    2. Hardy KL
    3. Kitajewski AM
    4. Shawber CJ
    5. Kitajewski JK
    6. Wu JK

    (2012) Propranolol accelerates adipogenesis in hemangioma stem cells and causes apoptosis of hemangioma endothelial cells

    Plastic and Reconstructive Surgery 130:1012–1021.

    https://doi.org/10.1097/PRS.0b013e318267d3db

    • PubMed
    • Google Scholar
48. 1. Wu JK
    2. Adepoju O
    3. De Silva D
    4. Baribault K
    5. Boscolo E
    6. Bischoff J
    7. Kitajewski J

    (2010) A switch in notch gene expression parallels stem cell to endothelial transition in infantile hemangioma

    Angiogenesis 13:15–23.

    https://doi.org/10.1007/s10456-009-9161-5

    • PubMed
    • Google Scholar
49. 1. Wünnemann F
    2. Kokta V
    3. Leclerc S
    4. Thibeault M
    5. McCuaig C
    6. Hatami A
    7. Stheneur C
    8. Grenier JC
    9. Awadalla P
    10. Mitchell GA
    11. Andelfinger G
    12. Preuss C

    (2016) Aortic dilatation associated with a de novo mutation in the SOX18 Gene: Expanding the Clinical Spectrum of Hypotrichosis-Lymphedema-Telangiectasia Syndrome

    Canadian Journal of Cardiology 32:135.e1–13135.e1.

    https://doi.org/10.1016/j.cjca.2015.04.004

    • PubMed
    • Google Scholar
50. 1. Xiao Q
    2. Li Q
    3. Zhang B
    4. Yu W

    (2013) Propranolol therapy of infantile hemangiomas: efficacy, adverse effects, and recurrence

    Pediatric Surgery International 29:575–581.

    https://doi.org/10.1007/s00383-013-3283-y

    • PubMed
    • Google Scholar

Thank you for submitting your article "Propranolol in HLTRS and infantile hemangioma: Stereoselective blockage of the SOX18 transcription factor activity" for consideration by eLife. Your article has been reviewed by three peer reviewers, including Gou Young Koh as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Sean Morrison as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity:); Pia Ostergaard (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

While reviewers 1 and 2 are positive, the reviewers have raised some issues, which are constructive and insightful.

Essential revisions:

We encourage the authors to do additional experiments to address the comments raised by reviewer 1 (comments 4, 5 and 7) and reviewer 3 (comment 2). The remaining comments could be addressed by changes to the text, such as through further discussion.

Reviewer #1:

This study by Overman et al. suggests a new mechanistic insight of propranolol in that R(+)-propranolol enantiomer acts as a small molecule inhibitor of SOX18. They also demonstrate the therapeutic potential of the R(+)-propranolol enantiomer in a patient-derived in vitro model of infantile hemangioma. The hypothesis tested in the study is intriguing and might be clinically relevant with potential therapeutic implications for pathologies involving SOX18 mutations and/or aberrant vascular growth, such as infantile hemangioma. Therefore, I would like to support the publication of this work after addressing the following points.

1) The authors suggest that the R(+)-propranolol enantiomer directly inhibits the transcriptional activity of SOX18 and thereby functions as an anti-angiogenic agent. However, the authors also demonstrate that propranolol restores SOX18 functionality in a dose-dependent manner, which is demonstrated by Vcam1 (SOX18 target) promoter luciferase assay (Figure 2B). This particular aspect seems counterintuitive. Considering that SOX18 mutations in hypotrichosis-lymphedema-telangiectasia and renal syndrome (HLTRS) act in a dominant-negative manner by impairing SOX18 function, please clearly address whether and how the R(+)-propranolol enantiomer restores or inhibits SOX18 function.

2) Could the authors check the protein expression of SOX18 in infantile hemangioma tissue samples or in cell populations derived from infantile hemangioma before and/or after treatment with propranolol to test its beneficial effect in these circumstances?

3) The authors demonstrated that the propranolol treatment in a patient with HLTRS dramatically improved pericardial effusion (Figure 1A-B). This change might have resulted secondarily from the process caused by SOX18 mutation (possibly including lymphedema, and/or renal syndrome), as the primary pericardial effusion itself is a relatively rare disease entity (Imazio et al., Nature Review Cardiology, 2009). Does the HLTRS patient have improvements in symptoms other than pericardial effusion, which might be directly related to HLTRS (e.g. telangiectasia, lymphedema) after propranolol treatment rather than pericardial effusion alone? Furthermore, if the mechanism of action of propranolol in the patient with SOX18 mutation is independent of anti-β-adrenergic activity as the authors suggested, how was the heart function of the patient, including tamponade physiology, improved? Is there any direct or indirect evidence to suggest that SOX18 function is restored or at least altered with propranolol treatment in this patient? These issues might need to be discussed in the Discussion section.

4) The authors suggest that propranolol inhibits SOX18 activity during VEGF-B-induced differentiation of hemangioma-derived stem cells (HemSC) into the hemangioma endothelial cells (HemECs). Could the authors verify that there is no change in SOX18 activity with β-adrenergic agonist treatment to support their claim that the effect of propranolol on SOX18 is independent of a β-adrenergic mechanism?

5) Does propranolol rescue only the aberrant vessel growth in Sox18 RaOp mouse cornea? Or also in other tissues such as skin, as the hypotrichosis and infantile hemangioma usually involves and grows on the skin surface.

6) In Figure 2—figure supplement 1B, the authors demonstrate the lymphatic vessel outgrowth (Neuropillin-2 and Prox1) in corneal whole-mount staining of Sox18 RaOp mice. This data is not mentioned in the main text although it is demonstrated in the supplemental figure and its legend. Why do the cornea lymphatic vessels outgrow in Sox18 RaOp mice?

7) Does the R(+)-propranolol enantiomer modulate VEGFR2 levels or its phosphorylation upon VEGF-A stimulation? This should be easily testable as the authors already have the cell lines, and this might partly explain and support the regression of infantile hemangioma by propranolol treatment.

Reviewer #2:

I have read the manuscript from Overman et al. with great interest. It is very interesting to get an insight into how some of these genetic defects can possibly be overcome using various forms of drugs. Overall, I think this manuscript represents a solid piece of work and I do not have any major comments except that it would be nice if the discussion had been a bit more comprehensive. For example, the authors mention that the propranolol increased the mean weight of the wild type mice, but not of the RaOp mice. Why is that?

Reviewer #3:

The key message of this study is that the block of Sox18 is a main 'mechanism of action' of propranolol in treating HLTRS and infantile hemangioma. Several experiments support this conclusion, but insufficiently.

1) Abnormal SOX18 protein derived from mutation in a patient can affect downstream broadly. Thus, it is difficult to judge targets of propranolol in a HLTRS patient. The RaOp mouse model has same issue, too.

2) Although both HLTRS and RaOp have mutations in Sox18, phenotypes focused in this study are different: pericardial edema in HLTRS and corneal angiogenesis in RaOp. They seem to be far away for comparison. In addition, the dose of propranolol for a HLTLS patient was 4 mg/kg/day while that for RaOp mice was 25 mg/kg/day (with an increasing potential of non-selective effects).

3) Use of the R(+) enantiomer of propranolol was suggested to preclude β-adrenergic-dependent blocking activity. However, the R(+) enantiomer was not used in the assessment of corneal neo-vascularisation.

4) Similarly, differentiation of hemangioma stem cells into endothelial cells relies on many factors including all SoxF members but not Sox18 alone. Downregulated genes by propranolol are representative endothelial markers rather than Sox18 targets (Figure 3C). Altogether, there may be a possibility that propranolol plays a role in a Sox18-independent manner. Additional experiments using other non-selective β-adrenergic blockers such as carvedilol may support the conclusions.

5) Some in vitro experiments appear to lack a physiologic link to in vivo and patient data.

– There was no link to Sox18 function in cell assays shown in Figures 1C and ID.

– The cell systems used in Figures 2B and 2C may be different from lymphatic or hemangioma cells.

– In the present manuscript, only numbers are presented without additional help on ALPHAScreen assay and it is very difficult to figure out how reliable the data of protein-protein interactions in the bar graph are.

– Sox18 protein increased luciferase activity by two-fold in the assay in Figure 2B. This extent of enhanced activity, in general, is not enough to prove binding of transcription factors to their target regions, raising a concern whether the Vcam-1 promoter region used in Figure 2B is a convincing target region of Sox18 in this context.

Essential revisions:

We encourage the authors to do additional experiments to address the comments raised by reviewer 1 (comments 4, 5 and 7) and reviewer 3 (comment 2). The remaining comments could be addressed by changes to the text, such as through further discussion.

Reviewer #1:

This study by Overman et al. suggests a new mechanistic insight of propranolol in that R(+)-propranolol enantiomer acts as a small molecule inhibitor of SOX18. They also demonstrate the therapeutic potential of the R(+)-propranolol enantiomer in a patient-derived in vitro model of infantile hemangioma. The hypothesis tested in the study is intriguing and might be clinically relevant with potential therapeutic implications for pathologies involving SOX18 mutations and/or aberrant vascular growth, such as infantile hemangioma. Therefore, I would like to support the publication of this work after addressing the following points.

1) The authors suggest that the R(+)-propranolol enantiomer directly inhibits the transcriptional activity of SOX18 and thereby functions as an anti-angiogenic agent. However, the authors also demonstrate that propranolol restores SOX18 functionality in a dose-dependent manner, which is demonstrated by Vcam1 (SOX18 target) promoter luciferase assay (Figure 2B). This particular aspect seems counterintuitive. Considering that SOX18 mutations in hypotrichosis-lymphedema-telangiectasia and renal syndrome (HLTRS) act in a dominant-negative manner by impairing SOX18 function, please clearly address whether and how the R(+)-propranolol enantiomer restores or inhibits SOX18 function.

In this study we show that propranolol has the ability to disrupt both SOX18 protein-protein interactions and SOX18-SOX18mut (RaOp) protein-protein interactions (i.e. SOX18 homodimer and SOX18/SOX18RaOp heterodimer formation Figure 2C). In the context of a transcription-based assay (Figure 2B) SOX18 RaOp mutant protein is a potent repressor of SOX18 activity (even up to a 30wt/1RaOp ratio), this causes the loss of Vcam1 transactivation (Figure 2B). SOX18 RaOp mutant protein acts by “poisoning functional SOX18 protein complexes” by forming SOX18-RaOp dimers. SOX18 function is partly restored by propranolol because propranolol disrupts SOX18RaOP protein recruitment and interaction with SOX18 WT. The non “poisoned” SOX18 is then able to transactivate Vcam1.

In other words the process corresponds to the inhibition (propranolol) of an inhibitor (RaOp) which in turn results to the restoration of SOX18 active form. Please find in Author response image 1 a schematic of the proposed partial agonist mechanism.

Author response image 1

Download asset Open asset

2) Could the authors check the protein expression of SOX18 in infantile hemangioma tissue samples or in cell populations derived from infantile hemangioma before and/or after treatment with propranolol to test its beneficial effect in these circumstances?

We have assessed SOX18 expression in 3 different human infantile hemangioma tissue samples from patients that did not received propranolol treatment, as patients treated with propranolol do not typically undergo surgery. We detected expression of the SOX18 protein in the pathological blood vasculature by immunofluorescence. The SOX18 protein was detected in a large subset of CD31-positive endothelial cells, in accordance with its role during hemangioma stem cell to EC differentiation process. This new result is now incorporated into Figure 3 as Figure 3A and reported in the text (Results section).

We next examined the effect of Sm4 and propranolol on SOX18 gene expression levels in the differentiated HemSCs (n=4). As shown in Author response image 2 Panel A, SOX18 increased upon VEGF-B-induced endothelial differentiation (p=0.0026), consistent with results shown in Figure 3B; inclusion of Sm4 or propranolol during the differentiation assay decreased levels of SOX18 transcript, measured by qPCR. This result is expected because both inhibitors block HemSC-EC differentiation (Figure 3C and Figure 3—figure supplement 2) and therefore decrease the total number of mature ECs. We normalised the data relative to a pan-endothelial cell marker such as CD31 (Panel B), to account for the bias introduced by the lack of endothelial cells differentiation. Using CD31 as a reference gene we show that SOX18 levels were constant, with a small but significant increase when propranolol was included in the differentiation media.

Author response image 2

Download asset Open asset

3) The authors demonstrated that the propranolol treatment in a patient with HLTRS dramatically improved pericardial effusion (Figure 1A-B). This change might have resulted secondarily from the process caused by SOX18 mutation (possibly including lymphedema, and/or renal syndrome), as the primary pericardial effusion itself is a relatively rare disease entity (Imazio et al., Nature Review Cardiology, 2009). Does the HLTRS patient have improvements in symptoms other than pericardial effusion, which might be directly related to HLTRS (e.g. telangiectasia, lymphedema) after propranolol treatment rather than pericardial effusion alone?

The patient had mild improvement in both the telangectasia (especially face and lips) and lymphedema (evidenced by reducing the dose of the diuretic Furosemide). However, the dramatic improvement was in the complete resolution of the pericardial effusion which was sustained even after stopping the propranolol. The exposure to propranolol was rather short, only 16 months of treatment were necessary to improve the pericardial effusion.

It is important to bear in mind that most of HLTRS symptoms are acquired during embryonic development when SOX18 is at play during the molecular process of endothelial cell specification and hair stem cell differentiation. It is therefore not expected to correct symptoms acquired during early development especially when the molecular target is not at play post-natally.

Furthermore, if the mechanism of action of propranolol in the patient with SOX18 mutation is independent of anti-β-adrenergic activity as the authors suggested, how was the heart function of the patient, including tamponade physiology, improved?

The systolic cardiac function (ejection fraction) was normal even during the time when there was large pericardial effusion. The diastolic function however was decreased due to the effusion. Nevertheless, the diastolic dysfunction observed did not reach the degree of tamponade.

Is there any direct or indirect evidence to suggest that SOX18 function is restored or at least altered with propranolol treatment in this patient? These issues might need to be discussed in the Discussion section.

So far the only indirect evidence that SOX18 function is restored has been reported in Wunnemann et al., 2015 whereby after years of propranolol treatment (from a young age to 13yo) the HTLRS patient did not develop alopecia and display normal hair growth.

We agree with the comment. With the current amount of scientific evidence we cannot conclude that in HLTRS the effects are mediated via a partial restoration of SOX18 function only. This could be due to an additive or synergistic effect with β-adrenergic blockade. This is now addressed in the Discussion.

4) The authors suggest that propranolol inhibits SOX18 activity during VEGF-B-induced differentiation of hemangioma-derived stem cells (HemSC) into the hemangioma endothelial cells (HemECs). Could the authors verify that there is no change in SOX18 activity with β-adrenergic agonist treatment to support their claim that the effect of propranolol on SOX18 is independent of a β-adrenergic mechanism?

We tested whether or not inclusion of isoproterenol, a non-selective β-adrenergic receptor agonist, during the VEGF-B-induced differentiation would alter SOX18 activity or increase endothelial differentiation. We measured NOTCH1, a SOX18 target gene, and VE-cadherin, an EC marker, and SOX18 in VEGF-B treated HemSCs cells by qPCR. In four biological replicates with HemSC isolated from three different infantile hemangioma specimens, we did not detect a significant increase in VE-cadherin, NOTCH1 or SOX18 in isoproterenol-treated cells compared to DMSO (vehicle)-treated cells. Propranolol caused a significant decreased in VE-cadherin, NOTCH1 and SOX18 transcripts; p values <0.05 (consistent with Figure 3C and Figure 3—figure supplement 2.) This suggests SOX18 activity is not modulated by a β-adrenergic pathway. Isoproterenol in the absence of VEGF-B-induced differentiation had no effect on VE-cadherin, NOTCH1, SOX18 levels (Author response image 3).

Author response image 3

Download asset Open asset

5) Does propranolol rescue only the aberrant vessel growth in Sox18 RaOp mouse cornea? Or also in other tissues such as skin, as the hypotrichosis and infantile hemangioma usually involves and grows on the skin surface.

We have performed an additional in vivo experiment to assess whether propranolol, the R-enantiomer, or the S-enantiomer impact aberrant vessel patterning in skin tissues. Animals were treated at 25mg/Kg/day (oral gavage) from P1 to P28. Ear skin tissues were collected and whole mount immunofluorescence staining was performed for CD31, Podoplanin and ERG. Samples were imaged by confocal microscopy and vessel morphogenesis was quantified using IMARIS software.

As previously published, we observed that the Ragged Opossum heterozygous mutant mice display severe lymphatic vascular defects (Francois et al., 2008) with major morphogenesis defects of the dermal lymphatic vasculature. The loss of a significant portion of vessels with valve structure (pre-collectors and collectors) revealed a dramatic phenotype of the mature lymphatic vascular network in the RaOp animals (Figure 3—figure supplement 4, top graph).

There was no phenotypic rescue corresponding to a significant improvement of the lymphatic vasculature mis-patterning observed for any of the drug treatments in the mutant animals. This result is in accordance with the fact the drug treatment occurs at post-natal stages whereas the dermal lymphatic vascular defects are acquired from early to mid-stages of organogenesis in the embryo. This result is now indicated in the manuscript.

For transparency of the review process we have appended in Author response image 4 all the images we have used for quantification of the skin lymphatic vascular experiment.

Author response image 4

Download asset Open asset

6) In Figure 2—figure supplement 1B, the authors demonstrate the lymphatic vessel outgrowth (Neuropillin-2 and Prox1) in corneal whole-mount staining of Sox18 RaOp mice. This data is not mentioned in the main text although it is demonstrated in the supplemental figure and its legend. Why do the cornea lymphatic vessels outgrow in Sox18 RaOp mice?

The molecular mechanism that underpins vascular outgrowth in the cornea for both blood and lymphatic vessels remains unclear. Since most causes of corneal neo-vascularisation (CNV) are indeed VEGF-dependent, and some are successfully treated with anti-VEGF based therapies, we decided to focus on the potential contribution of augmented VEGF signalling to the invasion of blood vessel into the Sox18^RaOp corneas. In the context of CNV, Flt1 is of particular interest, as the balance of distinct isoforms of this protein has been shown to be of key importance in maintaining corneal avascularity (Ambati et al., 2006). The pro- or anti-angiogenic activity of FLT1 is highly dependent on the particular splice variant that is produced (Kearney et al., 2004, Ferrara et al., 2003, Claesson-Welsh, 2016). The shorter isoform sFLT is released into the extracellular matrix to sequester endogenous pro-angiogenic VEGF-A, preventing the activation of FLT1 and VEGFR2 receptors and any downstream cellular response. Deficiency of sFLT in the cornea has been shown to be causative for CNV (Ambati et al., 2006). Further, we have recently shown that FLT1 is a potential direct target gene of SOX18 dimer complex (Moustaqil et al., NAR 2018). This observation prompted us to investigate FLT1 expression levels in the RaOp mutant mouse.

In order to identify whether VEGF pathway is dysregulated in the tissue of the mutant animals we have performed qPCR analysis for VEGF receptors and soluble receptors on the whole cornea of wild type and mutant animals at different time points (8weeks, P12weeks, P16weeks). See Author response image 5. To distinguish between the different splice variants that give rise to either the full length FLT1 receptors or soluble FLT1 (sFLT1), we designed primers to specifically amplify the following regions:

‘Flt1 (extra)’, extracellular domain that is present in both full length FLT1 and sFLT1;

‘Flt1 (intra)’, intracellular domain that is only present in full length FLT1;

‘sFlt1’, unique 3’end of exon 12 and 3’UTR only present on the sFlt1 transcript

Author response image 5

Download asset Open asset

The most obvious changes in Flt1 levels were observed at 8 weeks, in agreement with the expression pattern of Pecam (Figure 2—figure supplement 1C). At this stage, the amount of total Flt1 (Flt1 extra) in the Sox18^RaOp corneas was 4.6 fold higher than in the wild type (p-value < 0.05). This is partly explained by the fact that these corneas are covered in blood vessels, and the increased amount of endothelial cells contributes to the induction of Flt1 over the whole tissue. However, the increase in Pecam transcript levels – representing the increase in endothelial cells – is much lower than that of Flt (2 fold). In order to compensate for the increase in endothelial cells, we normalized the Flt1 transcript levels to Pecam levels, which makes it evident that total Flt1 increased disproportionality to the amount endothelial cells (2.0 fold, p-value < 0.05).

Although total Flt1 is upregulated in the corneas of SOX18^RaOp mice, sFlt1 levels are unchanged regardless of the amount of vascular endothelial cells. This suggests that the ratio between the amounts of full length FLT1 and sFLT1 in the cornea shifted in favour of the full-length isoform. Transcript levels for Vegfr2 were not significantly different between wild types and mutants at 8 weeks. Since FLT1, VEGFR2 and sFLT1 all compete for the same pool of VEGF-A, the shift in Flt1 isoforms could explain the corneal phenotype. However, whether exactly this is causative for the penetration of vessels into the cornea of Sox18^RaOp mice, and whether SOX18 is directly involved in regulating the balance between Flt1/sFlt1, is uncertain at this stage.

Of note FOXC1 levels remains unchanged; we tested for this gene expression level since loss of this transcription factor function has been shown to cause corneal neo-vascularisation (Axenfeld-Rieger Syndrome) (Mears et al. Am j Hum Genet 1998)

7) Does the R(+)-propranolol enantiomer modulate VEGFR2 levels or its phosphorylation upon VEGF-A stimulation? This should be easily testable as the authors already have the cell lines, and this might partly explain and support the regression of infantile hemangioma by propranolol treatment.

We address this question with three experiments.

In panel A, normal human ECFCs were pre-treated for 1 hour with propranolol, R+ enantiomer or S-enantiomer (each tested at 5μM) and then stimulated for 5 minutes with VEGF-A. Cells were lysed and analysed for phosphorylated VEGFR2 and total VEGFR2 by Western blotting. The drug treatments had no effect on levels of VEGFR2 or levels of VEGF-A-stimulated phosphorylation. The same results were seen when drug pre-treatment was carried out for 16 hours (not shown).

In Panel B, HemSC were induced to undergo endothelial differentiation with VEGF-B for 5 days, then pre-treated with propranolol, R+ enantiomer or S-enantiomer for 1 hour and next stimulated with VEGF-A for 5 minutes. Western blotting for VEGFR2 shows the drug treatments had no effect on VEGFR2 levels; the same results were seen with HemSC isolated from two different IH.

In Panel C, the same experiment was performed and VEGFR2 was immunoprecipitated to increase detection of phosphorylated VEGFR2. Drug pre-treatment had no effect on phosphor-VEGFR2 levels. The lack of VEGF-A-induced phosphorylation of VEGFR2 is consistent with previous reports wherein VEGFR2 is primarily intracellular in HemSC (Khan et al., 2008) and is constitutively phosphorylated in HemEC (Jinnin et al., Nat Med 2008). In summary, we did not observe any changes to VEGFR2 protein levels or its phosphorylation upon R-enantiomer, propranolol or S-enantiomer treatment. This is now shown in the manuscript as Figure 3—figure supplement 2B-D.

Reviewer #2:

I have read the manuscript from Overman et al. with great interest. It is very interesting to get an insight into how some of these genetic defects can possibly be overcome using various forms of drugs. Overall, I think this manuscript represents a solid piece of work and I do not have any major comments except that it would be nice if the discussion had been a bit more comprehensive. For example, the authors mention that the propranolol increased the mean weight of the wild type mice, but not of the RaOp mice. Why is that?

One of the phenotypes of the Ragged Opossum mutant is a complete lack of sub-cutaneous fat (Francois et al unpublished). The absence of adipose tissue makes it impossible for adipocytes to respond to a beta-adrenergic treatment.

It is for this particular reason that we think the Ragged Opossum is not gaining weight under propranolol treatment. The adipose tissue phenotype, currently under investigation, is beyond the scope of this study.

Reviewer #3:

The key message of this study is that the block of Sox18 is a main 'mechanism of action' of propranolol in treating HLTRS and infantile hemangioma. Several experiments support this conclusion, but insufficiently.

1) Abnormal SOX18 protein derived from mutation in a patient can affect downstream broadly. Thus, it is difficult to judge targets of propranolol in a HLTRS patient. The RaOp mouse model has same issue, too.

This concern is directed to the in vivo data analysis of propranolol treatment. On the other hand, our data from in vitro and cellular assays shows strongly that propranolol mode of action is mediated via a SOX18-dependent mechanism.

The recent description of the molecular function of the SOX18 homodimer in the context of endothelial cell differentiation (Moustaqil et al. NAR 2018) is an important piece of information. Here we show that the RaOp mutant protein has the capability to recruit SOX18 wild type protein and therefore interfere with the endothelial specific signature of SOX18 wild type homodimer.

Further, in this study we show that use of specific β adrenergic blockers or loss of β-adrenergic receptors does not affect ECFC or cancer cell line survival. As a final demonstration of a role of SOX18-mediated mode of action in vivo we now show that treatment with the R-enantiomer of propranolol (low to no anti-β adrenergic activity) is able to rescue the corneal neo-vascularisation phenotype.

2) Although both HLTRS and RaOp have mutations in Sox18, phenotypes focused in this study are different: pericardial edema in HLTRS and corneal angiogenesis in RaOp. They seem to be far away for comparison. In addition, the dose of propranolol for a HLTLS patient was 4 mg/kg/day while that for RaOp mice was 25 mg/kg/day (with an increasing potential of non-selective effects).

The point of using the Ragged Opossum mouse model experiment is to show on target engagement by rescuing a phenotype which relies on the activity of the SOX18 mutant protein. We agree that the readout in human is different, but in this particular case the decision to pharmacologically manage a new HLTRS case with propranolol was guided by the results obtained from pre-clinical data. This provided a therapeutic option where none was available before.

The choice of the dosage was originally guided by our experience in vivo in mice using Sm4 (SOX18 small molecule inhibitor) which we have shown is a potent anti-metastatic compound at 25mg/kg/day in pre-clinical model of breast cancer (Overman et al eLife 2017).

Of note we have shown efficacy of Sm4 at lower concentration in a dose dependent manner (from 5 to 50mg/Kg/day, unpublished data). This suggests that the patient treated with low dose of propranolol (4mg/Kg/day) has received a dose of small molecule inhibitor that is able to block SOX18 activity.

In the Ragged Opossum in vivo experiment we aimed at a maximum tolerated dose to get the best rescuing effect possible in turn translating into an obvious phenotypic rescue. The concentration scouting is a great suggestion but is not necessary as all in vitro data show concentration dependence and that the main goal of the rescuing approach is to show selective on-target engagement.

3) Use of the R(+) enantiomer of propranolol was suggested to preclude β-adrenergic-dependent blocking activity. However, the R(+) enantiomer was not used in the assessment of corneal neo-vascularisation.

We have now assessed the effect of the R(+) enantiomer in vivo in a rescue experiment of the corneal neo-vascularisation phenotype in the Ragged Opossum (RaOp) mouse model. These results are now included in the manuscript as a supplemental figure to Figure 3.

In short, the R(+) enantiomer has the ability to rescue the corneal neo-vascularisation phenotype of the RaOp heterozygous mouse. Further, we have also tested the S(-) enantiomer in the same model system. Interestingly it seems that this compound has the potential to rescue the phenotype albeit to a lesser extent. This results suggests that both R(+) and S(-) enantiomers have the capability to inhibit the SOX18 RaOp mutant protein activity. This is most likely by disturbing the SOX18/SOX18RaOp or the RBPJ/SOX18RaOp protein-protein interactions as shown by our in vitro assays. Of importance, in the hemangioma hemSC-EC differentiation assay the R(+) enantiomer was as efficient as propranolol in halting this differentiation process, whereas the S(-) enantiomer showed a weak inhibitory response, suggesting an enantiomer selective blockade of the SOX18 wild type protein mediated by the R(+) enantiomer.

Because of the weak activity of the S(-)enantiomer, we have revised the title of our manuscript to reflect accurately the key finding of this study: “R-propranolol is a small molecule inhibitor of the SOX18 transcription factor in a rare vascular syndrome and hemangioma.”

4) Similarly, differentiation of hemangioma stem cells into endothelial cells relies on many factors including all SoxF members but not Sox18 alone. Downregulated genes by propranolol are representative endothelial markers rather than Sox18 targets (Figure 3C). Altogether, there may be a possibility that propranolol plays a role in a Sox18-independent manner. Additional experiments using other non-selective β-adrenergic blockers such as carvedilol may support the conclusions.

We did not test additional non-selective β-adrenergic blockers.

5) Some in vitro experiments appear to lack a physiologic link to in vivo and patient data.

– There was no link to Sox18 function in cell assays shown in Figures 1C and ID.

The point of experiments shown in Figure 1C and 1D is to show that propranolol acts has a mode of action which is at least in part independent to β-adrenergic receptors. The link of propranolol to SOX18 is provided by other assay such as the in vitro transcriptional report assay (Figure 2B) and ALPHAScreen assay (Figure 2C).

– The cell systems used in Figures 2B and 2C may be different from lymphatic or hemangioma cells.

These in vitro assays are synthetic and designed to show propranolol on target engagement on SOX18 activity.

We actually use a cell based assay where SOX18 is not expressed and add exogenous SOX18 to demonstrate specific inhibition of this TF by propranolol.

– In the present manuscript, only numbers are presented without additional help on ALPHAScreen assay and it is very difficult to figure out how reliable the data of protein-protein interactions in the bar graph are.

We have now provided more information on the statistical analysis of the ALPHAScreen data analysis in the text of the corresponding figure legend.

– Sox18 protein increased luciferase activity by two-fold in the assay in Figure 2B. This extent of enhanced activity, in general, is not enough to prove binding of transcription factors to their target regions, raising a concern whether the Vcam-1 promoter region used in Figure 2B is a convincing target region of Sox18 in this context.

Although we agree that this assay efficacy window is not optimum, this result is highly reproducible and quite robust. The transactivation of the Vcam1 promoter fragment and the regulation of Vcam1 by SOX18 has been previously characterised and published, please refer to Hosking et al., 2004.

Author details

1. Jeroen Overman

   Institute for Molecular Bioscience, The University of Queensland, Brisbane, Australia

   Contribution

   Conceptualization, Formal analysis, Supervision, Funding acquisition, Investigation, Writing—original draft, Project administration, Writing—review and editing

   Contributed equally with

   Frank Fontaine

   Competing interests

   No competing interests declared

2. Frank Fontaine

   Institute for Molecular Bioscience, The University of Queensland, Brisbane, Australia

   Contribution

   Conceptualization, Formal analysis, Investigation, Methodology, Writing—review and editing

   Contributed equally with

   Jeroen Overman

   Competing interests

   No competing interests declared

3. Jill Wylie-Sears

   Vascular Biology Program, Department of Surgery, Boston Children's Hospital, Harvard Medical School, Boston, United States

   Contribution

   Conceptualization, Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

4. Mehdi Moustaqil

   Single Molecule Science, Lowy Cancer Research Centre, The University of New South Wales, Sydney, Australia

   Contribution

   Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

5. Lan Huang

   Vascular Biology Program, Department of Surgery, Boston Children's Hospital, Harvard Medical School, Boston, United States

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

6. Marie Meurer

   Centre de Recherche en Cancérologie de Marseille (CRCM Marseille Cancer Research Centre), Inserm UMR1068, CNRS UMR7258, Aix-Marseille University UM105, Marseille, France

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

7. Ivy Kim Chiang

   Institute for Molecular Bioscience, The University of Queensland, Brisbane, Australia

   Contribution

   Formal analysis, Visualization

   Competing interests

   No competing interests declared

8. Emmanuelle Lesieur

   Institute for Molecular Bioscience, The University of Queensland, Brisbane, Australia

   Contribution

   Formal analysis, Visualization

   Competing interests

   No competing interests declared

9. Jatin Patel

   Translational Research Institute, Diamantina Institute, The University of Queensland, Brisbane, Australia

   Contribution

   Formal analysis, Investigation, Writing—original draft

   Competing interests

   No competing interests declared

10. Johannes Zuegg

    Institute for Molecular Bioscience, The University of Queensland, Brisbane, Australia

    Contribution

    Resources, Formal analysis, Investigation

    Competing interests

    No competing interests declared

11. Eddy Pasquier

    Centre de Recherche en Cancérologie de Marseille (CRCM Marseille Cancer Research Centre), Inserm UMR1068, CNRS UMR7258, Aix-Marseille University UM105, Marseille, France

    Contribution

    Formal analysis, Investigation, Writing—original draft

    Competing interests

    No competing interests declared

12. Emma Sierecki

    Single Molecule Science, Lowy Cancer Research Centre, The University of New South Wales, Sydney, Australia

    Contribution

    Resources, Formal analysis, Supervision, Methodology

    Competing interests

    No competing interests declared

13. Yann Gambin

    Single Molecule Science, Lowy Cancer Research Centre, The University of New South Wales, Sydney, Australia

    Contribution

    Resources, Supervision, Methodology

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-7378-8976

14. Mohamed Hamdan

    Dubai Healthcare City, Dubai, United Arab Emirates

    Contribution

    Resources, Supervision, Investigation, Methodology, Writing—original draft, Project administration

    Competing interests

    No competing interests declared

15. Kiarash Khosrotehrani

    Translational Research Institute, Diamantina Institute, The University of Queensland, Brisbane, Australia

    Contribution

    Supervision, Funding acquisition, Investigation, Writing—original draft, Project administration

    Competing interests

    No competing interests declared

16. Gregor Andelfinger

    Department of Pediatrics, University of Montreal, Ste-Justine University Hospital Centre, Montréal, Canada

    Contribution

    Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing

    Competing interests

    No competing interests declared

17. Joyce Bischoff

    Vascular Biology Program, Department of Surgery, Boston Children's Hospital, Harvard Medical School, Boston, United States

    Contribution

    Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    joyce.bischoff@childrens.harvard.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-6367-1974

18. Mathias Francois

    Institute for Molecular Bioscience, The University of Queensland, Brisbane, Australia

    Contribution

    Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing

    For correspondence

    m.francois@imb.uq.edu.au

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-9846-6882

• 2,606

  Page views

• 282

  Downloads

• 28

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, PubMed Central.