Sox18 Genetically Interacts With VegfC to Regulate Lymphangiogenesis in ZebrafishSignificance
Objective—Lymphangiogenesis is regulated by transcription factors and by growth factor pathways, but their interplay has not been extensively studied so far. We addressed this issue in zebrafish.
Approach and Results—Mutations in the transcription factor–coding gene SOX18 and in VEGFR3 cause lymphedema, and the VEGFR3/Flt4 ligand VEGFC plays an evolutionarily conserved role in lymphangiogenesis. Here, we report a strong genetic interaction between Sox18 and VegfC in the early phases of lymphatic development in zebrafish. Knockdown of sox18 selectively impaired lymphatic sprouting from the cardinal vein and resulted in defective lymphatic thoracic duct formation. Sox18 and the related protein Sox7 play redundant roles in arteriovenous differentiation. We used a novel transgenic line that enables inducible expression of a dominant-negative mutant form of mouse Sox18 protein. Our data led us to conclude that Sox18 is crucially involved in lymphangiogenesis after arteriovenous differentiation. Combined partial knockdown of sox18 and vegfc, using subcritical doses of specific morpholinos, revealed a synergistic interaction in both venous and lymphatic sprouting from the cardinal vein and greatly impaired thoracic duct formation.
Conclusions—This interaction suggests a previously unappreciated crosstalk between the growth factor and transcription factor pathways that regulate lymphangiogenesis in development and disease.
The lymphatic system is a major component of the vertebrate vasculature and plays key roles in tissue fluid homeostasis, fat absorption, and the immune response.1 Lymphatic vessel function and dysfunction contribute to the progression of several pathological conditions, such as tumor metastasis, lymphedema, obesity, and inflammation.2 Despite its relevance for human health and disease, lymphangiogenesis has been far less studied than blood vessel angiogenesis. Hereditary (primary) lymphedema is a rare genetic disorder commonly caused by inherited mutations in genes that regulate crucial pathways during the development of the lymphovascular system. Several genes have been identified in lymphedema, including VEGFR3 (Milroy disease),3,4 FOXC2 (lymphedema distichiasis syndrome),5 HGF/MET,6 GJC2,7 PTPN14,8 GATA2,9 and KIF11.10
Hypotrichosis–lymphedema–telangiectasia is a human syndrome displaying mutations in the transcription factor SOX18 and is characterized by the association of heritable alopecia, lymphedema, and vascular malformations.11 The spontaneous ragged (Ra) mutant mice represent the murine counterpart of the hypotrichosis–lymphedema–telangiectasia syndrome.12,13 The most severe Ra mutation, ragged-opossum (RaOp), and the Sox18 null mice, in a pure C57BL/6 (B6) background, showed gross subcutaneous edema at 13.5 days after coitum and died after 14.5 days after coitum, precluding study of lymphatic physiology beyond that stage.14 Recent findings in mice demonstrated that Sox18 directly activates the transcription of Prox1, a master regulator of lymphatic endothelial cells (LECs) specification, by binding to its proximal promoter.14
In mammals, according to the currently widely accepted model, lymphatic vessels arise by direct sprouting of precursors from the cardinal veins to give rise to the early lymph sacs, which remodel and sprout to give rise to the entire lymphatic vasculature.15,16 Very recently, it has been shown that a combination of cellular processes that involve ballooning from the cardinal vein and migration as single cells are necessary to establish the lymphatic vascular plexus.17,18 A venous origin has been shown also for the recently discovered zebrafish lymphatic system,19,20 that shares key molecular regulators (including the vascular endothelial growth factor VEGFC and its receptor VEGFR3/Flt4) with the mammalian system. Recently Ccbe1 (Collagen and calcium-binding EGF domains-1), identified in zebrafish as an important regulator of lymphangioblast budding,21 has been found to be mutated in human patients affected by Hennekam Lymphangiectasia Lymphedema syndrome,22,23 clearly demonstrating that the zebrafish model system could directly aid in defining the molecular basis of human lymphangiopathies.
In zebrafish, developmental angiogenesis occurs in 2 different waves: during the first wave, primary sprouts from the dorsal aorta (DA) give rise to intersomitic vessels (ISVs) from ≈22 hours postfertilization (hpf); during the second wave, half of the sprouts from the vein will convert arterial ISVs into venous ISVs (vISVs), whereas the other half gives rise to a pool of lymphatic precursors at the horizontal myoseptum (HMS),19,24 the parachordal lymphangioblasts (PLs), from ≈32 hpf.21,25 The lymphatic thoracic duct (TD), the main lymphatic vessel described in zebrafish,19,20 is then generated by the ventral migration of these PLs.
We decided to specifically address the role played by sox18 in zebrafish lymphatic development and the interplay between sox18 and the central lymphatic growth factor vegfc. We have previously shown that sox18 and the closely related sox7 gene play redundant roles in arteriovenous differentiation: their simultaneous partial knockdown impairs particularly the acquisition of a full venous identity, thus pointing to a potential role in lymphatic differentiation.26 We now show that the 2 genes are differentially expressed in the posterior cardinal vein (PCV): at stages crucial for early lymphatic development, only sox18 expression is clearly detectable in the PCV. The knockdown of sox18 specifically affects lymphatic development: the number of sprouts from the vein and of PLs at the myoseptum are significantly impaired, however the number of vISVs is largely unaffected. The inducible expression of a dominant-negative Sox18–RaOp mutant protein in zebrafish embryos causes impairment of lymphatic precursor sprouting from the vein at stages after arteriovenous differentiation, dissociating the phenotype from earlier potential arteriovenous defects.
Significantly, TD defects are synergistically induced by the coinjection of subcritical doses of sox18 and vegfc morpholinos. The simultaneous partial knockdown of sox18 and vegfc reduces the number of PLs, but also vISVs, thus suggesting that both venous and lymphatic sprouting are coregulated by vegfc and sox18. These data support the key importance of Sox18 in early phases of lymphatic development and, for the first time, indicate that a strong genetic interaction exists between Sox18 and VegfC in this process.
Materials and Methods
Materials and Methods are available in the online-only Supplement.
sox18, but not sox7, Is Expressed in the Cardinal Vein During Lymphatic Precursor Sprouting
Sox18 and Sox7 play redundant roles in arteriovenous differentiation of endothelial cells in zebrafish.26–28 Simultaneous partial knockdown of both genes, but not single partial knockdowns, causes multiple fusions between the major axial vessels (the DA and the PCV) because of an incomplete acquisition of arteriovenous identity by endothelial cells. In particular, venous endothelial cell differentiation is more impaired than arterial differentiation in sox18/sox7 double morphants.26
We found that both sox18 and sox7 are expressed in the developing axial and ISVs, and in the developing head vasculature at 29 hpf (Figure 1Aa, 1Aa’, 1Ab, and 1Ab’). However, at later stages of development, the 2 genes are differentially expressed in the PCV: at 36 hpf, only sox18 expression is clearly detectable in the PCV by in situ hybridization (ISH), whereas both sox18 and sox7 are expressed in the DA (Figure 1Ac, 1Ac’, 1Ad and 1Ad’). Given the venous origin of LECs, we decided to evaluate the role played by Sox18 and Sox7 in the early phases of lymphatic development.
Knockdown of sox18 Affects TD Formation
The analysis of TD formation is commonly used to study lymphatic development in zebrafish (Materials and Methods in the online-only Data Supplement). To knockdown sox18 or sox7, we injected splice-blocking morpholinos (sox18-MO2 and sox7-MO2) and translation-blocking morpholinos (sox18-MO1 and sox7-MO1)26 into tg(fli1a:EGFP)y1 embryos, where both blood and lymphatic vessels are labeled.19 Larvae were subdivided into phenotypic classes of increasing severity, ranging from fully formed to completely absent TD, to account for the variability of the lymphatic defects (Materials and Methods in the online-only Data Supplement).
We optimized the dose of sox18-MOs to produce relevant defects in TD formation, while minimizing morphological or circulatory defects that would interfere with TD analysis (data not shown). The injection of sox18-MO2 at 1pmol/embryo specifically impairs TD formation at 5 dpf (Figure 1B and 1C; Table I and Figure I in the online-only Data Supplement). In contrast, most control larvae showed fully formed TD (Figure 1C; Table I and Figure I in the online-only Data Supplement). Similar defects in TD formation resulted from the injection of an independent morpholino targeting sox18 (sox18-MO1) although with lower penetrance (Figure IIA and IIB in the online-only Data Supplement). Moreover, sox18 RNA rescues the lymphatic phenotype of sox18 morphants in a dose-dependent manner (Figure 1B and 1C; Table I in the online-only Data Supplement), supporting the specificity of these phenotypes (Figure 1C; Table I in the online-only Data Supplement).
We next analyzed the effects of sox7 knockdown on lymphatic development. Using 2 different morpholinos at the maximal doses allowing robust blood circulation, we found that the knockdown of sox7 caused only minor defects in TD formation (Figure IIC and IID in the online-only Data Supplement and data not shown). The coinjection of even low doses of sox18- and sox7-MOs blocks blood circulation in the trunk, due to impaired arteriovenous differentiation and arteriovenous shunt formation26; therefore, we could not fully investigate whether knocking down sox7 exacerbates the lymphatic phenotype of sox18 morphants.
When dealing with a subfamily of SOX genes, it is important to check for background-dependent effects, like strain-specific compensatory upregulation of other Sox family members when one is knocked down or out.14,29 We analyzed the expression levels of sox7 in sox18 morphants at various developmental stages (18–20 somites, 24 hpf and 36 hpf), before or around early lymphatic development, and we did not observe any significant changes (Figure IIIa–IIIf in the online-only Data Supplement).
Our data thus point to an important role of Sox18 in lymphatic development, whereas Sox7 does not seem to play such a prominent role in this process and has not been analyzed any further in this study.
Knockdown of sox18 Impairs the Sprouting of Lymphatic Precursors From the Vein, but Does Not Alter vISV Angiogenesis
We next sought to better characterize early steps of lymphatic development in sox18 morphants. In the secondary wave of zebrafish angiogenesis, half of the sprouts from the vein anastomose with and convert arterial derived ISVs into vISVs, whereas the other half gives rise to a pool of lymphatic precursors at the myoseptum, the PLs (also called lymphangioblasts or parachordal chain cells).19,21,24,25,30 This process involves several molecular players and is known to be controlled by VEGF-C/VEGFR3 signaling.24,25,31
Knockdown of sox18 caused a reduction of ≈40% in the total number of sprouts from the vein scored at 1.5 dpf, which is compatible with an effect limited to lymphangiogenic sprouting (Figure 2A). Knockdown of vegfc caused, instead, a much more drastic reduction (Figure 2A), pointing to an overall impairment of secondary sprouting, as already reported in literature.31 These data indicate considerably more specificity to the sox18 morphants phenotype we observe here.
Lymphatic precursors, originating from the PCV, are transiently residing at the HMS before migrating ventrally or dorsally to give rise to TD and other trunk lymphatic vessels.19,21 We directly scored PLs at the HMS in circulating sox18 morphants and found a significant decrease at 56 hpf (5.1+0.5 versus 8.3+0.2 in controls, Figure 2B).
We next scored vISVs in circulating control embryos and sox18 morphants at 2.5 dpf. In 3 independent experiments, we found that sox18 knockdown did not significantly alter vISV numbers (Figure 2C). Additionally, in separate experiments, we scored for a/v ISVs at 2.5 dpf and kept morphants for further TD scoring at 5 dpf. This enabled us to calculate, a posteriori, the number of a/v ISVs at 2.5 dpf in sox18 morphants showing different degrees of lymphatic defects at 5 dpf. We found that sox18 morphants with more severe TD defects and those either unaffected or with minor TD defects showed comparable vISV numbers (Figure IVA and IVB in the online-only Data Supplement).
These analyses confirm that sox18 knockdown, at the morpholino dose we chose to avoid circulatory defects, impairs lymphatic development, without significantly altering the venous component of the secondary angiogenic wave.
Heat-Shock Inducible Overexpression of Mouse Sox18 Ragged Opossum Inhibits Zebrafish PL Development Postarteriovenous Segregation
Sox18 plays a role in early arteriovenous differentiation and, theoretically, a ubiquitous knockdown across all developmental stages (morpholino approach) could induce PL and TD defects secondarily to a mild arteriovenous defect that we might not be able to score at a basic morphological level.
To inhibit Sox18 activity at stages subsequent to arteriovenous differentiation, we generated a transgenic line for the temporally inducible inhibition of its transcriptional activity. RaOp is the strongest of the 4 known ragged mutant alleles, all coding for Sox18 mutant proteins with an intact DNA–binding domain but compromised transactivation ability.32–34 Ragged mutant proteins act in a dominant-negative fashion, preventing the binding of redundant SoxF factors (ie, Sox7 and Sox17) to Sox18 target genes.
Sox18RaOp dominantly interferes with Sox18-, 7-, and 17-regulated transcription in mouse embryonic lymphangiogenesis.14,29 We took advantage of this mutant allele and cloned the mouse Sox18RaOp cDNA, fused in frame with the mCherry coding sequence, under the control of the hsp70l promoter (Figure 3A). We used Tol2-mediated transgenesis35 to generate a stable zebrafish tg(hsp70l:Sox18RaOp mCherry) line.
We performed staged heat shocks and observed the nuclear accumulation of mCherry protein by 3 to 4 hours (but not 2 hours) after heat-shock treatment (data not shown) using confocal microscopy. The tg(hsp70l:Sox18RaOp mCherry) line was crossed to tg(fli1a:EGFP)y1 or tg(fli1a:EGFP)y1;tg(flt1enh:RFP). Therefore, half of the GFP+ progeny carries the inducible transgene, and the other half of the GFP+ progeny serves as nontransgenic controls, alongside non-heat-shocked transgenic controls. Heat shock was performed at 24, 29, 36, 48, and 72 hpf, and embryos were separated based on mCherry expression at 3 to 4 hours after heat shock.
Heat-shock induction of Sox18RaOp at 24 hpf led to significant cardiovascular defects and a general (not lymph-) edema phenotype by 5 dpf (Figure VA in the online-only Data Supplement), attributable to circulatory defects, thus further revealing an ongoing role of SoxF proteins in cardiovascular development, as reported in the literature for other organisms.36 Heat shock at 29 hpf did not cause cardiac edema (Figure 3C) and did not interfere with major axial blood circulation through the DA and PCV, however the circulation through ISVs was abnormal at 2.5 dpf. Embryos heat shocked at later time points showed normal circulation.
Heat shock at 29 hpf led to a near complete loss of PLs at the HMS at 56 hpf, whereas 36 and 48 hpf heat shocks led to milder reductions in PL numbers (Figure 3G and 3J; Table IV in the online-only Data Supplement). Later heat shock at 72 hpf had no effect on later TD formation (data not shown). Non-heat-shocked transgenic animals and heat-shocked nontransgenic controls showed no phenotype (Figure 3B, 3D–3F, and data not shown).
Heat shock at 29 hpf also led to a highly significant reduction in the number of vISVs with respect to control embryos (Figure 3E, 3F, 3H, 3I, and 3K). Such a reduction was not detectable in sox18 morphants (Figure 2C, and Figure IVA and IVB in the online-only Data Supplement). This prompted us to analyze the expression level of vegfc in the Sox18RaOp induced transgenic embryos, because Vegfc/Vegfr3 signaling is crucial for venous and lymphatic sprouting.21,31 Heat shock at 29 hpf resulted in a reduction of vegfc expression in these embryos (Figure VC in the online-only Data Supplement). This reduction may contribute to the severe phenotypes observed on overexpression of the dominant-negative Sox18 mutant protein, but these may also be due to a more generally impaired input of SoxF proteins.
These experiments show that dramatic PL defects were observed when Sox18RaOp mCherry overexpression was induced at 29 hpf, which leads to the expression of nuclear mCherry protein by 32 to 33 hpf. The milder phenotype with 36 hpf heat shocks suggests a critical period for SoxF activity in lymphatic development between 32 and 40 hpf, during the period of PL sprouting from the cardinal vein.
sox18 and vegfc Genetically Interact in Zebrafish TD Formation
To gain insight into the molecular events governing the early phases of lymphatic development, we decided to analyze the interplay between sox18 and vegfc, a growth factor crucial for this process.19,20,31
We reproduced the lymphatic phenotype associated with knockdown of vegfc and then constructed a dose–response curve by injecting several doses of vegfc-MO to identify a critical range (Figure VI in the online-only Data Supplement). This led us to define a subcritical dose of vegfc-MO (0.06 pmoles/embryo) to be used in coinjection experiments along with a subcritical dose of sox18-MO2 (0.5 pmoles/embryo).
The single injection of these low doses of sox18 and vegfc morpholinos caused no gross morphological or lymphatic abnormalities (Figure 4; Figure VII in the online-only Data Supplement). On the contrary, coinjection of subcritical doses led to severe defects in TD development (Figure 4): almost 70% of coinjected larvae showed a total absence of TD or the presence of only 10% to 30% TD+ segments (Figure 4B; Figure VII in the online-only Data Supplement). The synergistic effect of the coinjection can also be obtained even when cutting by half the subcritical doses of sox18-MO2 and vegfc-MO (0.25 and 0.03 pmoles/embryo, respectively; Figure VIIIA and VIIIC in the online-only Data Supplement), and TD formation is drastically affected also by coinjecting a low dose of vegfc-MO (0.06 pmoles/embryo, Figure 4) with a subcritical dose of an independent sox18 morpholino (sox18-MO1, 0.5 pmoles/embryo), that does not largely affect TD formation when injected on its own (Figure VIIIB in the online-only Data Supplement).
Moreover, synergistic defects in TD formation were also obtained by simultaneous partial knockdown of sox18 and of the VegfC receptor gene flt4 (Figure IX in the online-only Data Supplement), coinjecting low doses of sox18-MO2 (0.5 pmoles/embryo) and flt4-MO (0.06 pmoles/embryo).
To determine whether Sox18 and VegfC cross-regulate at the mRNA level, we analyzed by ISH whether the mRNA levels of sox18 or vegfc are perturbed by the knockdown of vegfc or sox18, respectively. sox18 transcripts did not show any significant reduction in vegfc morphants nor did vegfc expression show changes in sox18 morphants (Figure X in the online-only Data Supplement). In addition, we found that the simultaneous partial knockdown of sox18 and vegfc does not alter sox7 expression (Figure III in the online-only Data Supplement). These data suggest that the interactions observed here are not occurring at the level of embryonic transcription of these genes.
The overexpression of a dominant-negative Sox18 mutant protein resulted in a reduction of vegfc expression, whereas the knockdown of sox18 did not produce detectable changes in the vegfc ISH signal. These data might imply other SoxF proteins in the regulation of vegfc. Alternatively, they might imply that a more pronounced reduction in Sox18 than the one caused by knockdown is needed to produce an alteration in vegfc levels.
To further address the molecular basis of the Sox18/VegfC interplay, we coinjected vegfc RNA while knocking down sox18 by morpholino injection: overexpression of vegfc partially rescued TD formation defects in sox18 morphants (Figure XIA in the online-only Data Supplement). The reverse experiment, namely sox18 RNA injection in vegfc morphants, did not cause any amelioration of the TD phenotype (Figure XIB in the online-only Data Supplement).
Taken together, our data for the first time suggest a relationship between the growth factor pathways that specifically regulate lymphangiogenesis (VegfC/Vegfr3 signaling) and the transcriptional pathways that modulate lymphangiogenesis.
Sox18 and VegfC Cooperate in Both PL and vISV Sprouting From the Cardinal Vein
We next investigated the specific population of venous-derived cells impaired in these double morphants. In full knockdown scenarios, Sox18 primarily regulates PL sprouting, but VegfC regulates both PL and vISV sprouting.
We scored total sprouts from the vein, PLs, and vISVs in embryos coinjected with subcritical doses of sox18 and vegfc MOs. Combined partial knockdown caused a reduction of >50% in the total number of sprouts from the vein at 1.5 dpf (Figure 5A) and a marked loss of PLs at the HMS at 56 hpf (Figure 5B). Furthermore, we scored vISVs at 2.5 dpf and found a synergistic interaction in vISV development: the subcritical doses of sox18- and vegfc-MOs caused a statistically significant reduction in vISVs only when coinjected (Figure 5C; Figure IVC in the online-only Data Supplement).
To test how robust these observations are, we decided to examine the interaction with independent molecular markers of the vasculature. We performed ISHs with the pan-endothelial marker cdh5 and some venous specific markers, such as dab2, ephB4, and flt4 around 29 hpf. Hybridization signals for these molecular probes were comparable in sox18 morphants and in combined partial sox18 and vegfc morphants with respect to controls, suggesting that blood endothelial cells were largely unaffected (Figure 6A; Figure XIIA in the online-only Data Supplement).
Next, we examined venous and lymphatic precursor sprouting using the lyve1 marker21,37 in ISHs at 2 dpf. lyve1+ sprouts from the PCV were clearly visible in controls (Figure 6Bb, white arrows) but severely reduced or absent in sox18 morphants and in combined partial sox18 and vegfc morphants (Figure 6Bd, 6Be, and 6Bg). We subdivided morphants to better describe their phenotypes in terms of presence/absence and length of lyve1+ sprouts (Figure 6B). Normal lyve1+ sprouts were detectable in most control embryos, but in only ≈10% of sox18 morphants and 5% of combined partial sox18-vegfc morphants (Figure 6Bh). Interestingly, among sox18 morphants, the prevalent phenotype was that of embryos with reduced numbers of lyve1+ sprouts, accounting for almost 40% (Figure 6Bd and 6Bh), but in combined partial sox18-vegfc morphants, the complete absence of lyve1+ sprouts (asterisk) prevailed, characterizing ≈40% of the embryos (Figure 6Bg and 6Bh). These data give an alternative confirmation of the genetic interaction and could be considered indicative of a combined impairment in lymphatic differentiation and in secondary sprouting in the combined partial sox18-vegfc knockdown but not in single sox18 knockdown scenarios.
In the past few years, zebrafish has emerged as a very potent system to study lymphangiogenesis.38,39 Overall, the zebrafish lymphatic system shares several morphological, functional, and molecular characteristics with mammals. Since the initial description of the zebrafish lymphatic system in 2006,19,20 a handful of molecular players have been shown to be evolutionarily conserved, but much remains to be elucidated and we are far from a complete picture of the degree of conservation of molecular pathways from zebrafish to human lymphangiogenesis. This prompted us to study the role of sox18 in zebrafish lymphatic development, because SOX18 mutations are associated with lymphedema in patients affected by the hypotrichosis–lymphedema–telangiectasia syndrome, and studies in mouse placed Sox18 very high in the hierarchy of transcription factors governing LEC differentiation.40
Sox18 belongs to the Sox F group of Sry-related high mobility group box transcription factors, also comprising the closely related Sox7 and Sox17 proteins. Sox proteins of the same subfamily tend to be biochemically interchangeable in vitro, and the relevance of individual Sox genes for a specific process is often linked to their differential expression in vivo.14,29
Ours and other groups have reported that sox7 and sox18 are coexpressed in angioblasts and endothelial cells of the forming vasculature, and that they play redundant roles in arteriovenous differentiation.26–28 We show here that sox7 stops being expressed earlier than sox18 in the axial vein, whereas both genes are still expressed in the DA, and that Sox18 specifically regulates lymphatic development.
In mice, Sox18 acts in concert with CoupTFII to drive the transcription of Prox1.14,41 The polarized expression of Sox18 precedes that of Prox1 in a subset of cells along the dorsolateral aspect of the cardinal vein at 9 days after coitum,14 whereas no polarized expression of CoupTFII has been reported so far. Remarkably, we have no evidence of a polarized expression of sox18 within the PCV, when secondary sprouts are arising from the dorsal aspect of the axial vein in zebrafish.
Although quite prominent and statistically highly significant, the degree of impairment in TD formation we observe in sox18 morphants versus control embryos is less striking than that observable when vegfc is fully knocked down. Our analysis is limited to circulating morphants without gross morphological abnormalities, and this sets an upper limit to the dose of sox18 morpholino we use. Hence, the strong but not full impairment in TD formation in sox18 morphants could be the result of a submaximal dose of MO in these experiments.
Lymphatic precursors sprout from the vein at ≈1.5 dpf and account for approximately half of the total sprouts,19,25,31 the other half consisting of venous sprouts that will connect to arterial ISVs and convert them into venous ISVs. Our results point to a specific impairment of lymphangiogenic sprouting from the cardinal vein in sox18 morphants, whereas secondary venous angiogenic sprouting is not substantially perturbed. Importantly, several pan-endothelial (cdh5) or venous markers (dab2, ephB4, flt4/vegfr3) are unaffected in sox18 morphants, whereas the venous/lymphatic endothelial marker lyve1 is altered: lyve1+ sprouts are either shorter or reduced in number, up to totally absent in the most severely affected sox18 morphants, providing an independent molecular assay for our morphological observations. Notably, PL defects in sox18 morphants do not seem to be secondary to overt venous differentiation problems. These data together imply a specific role of Sox18 in the early phases of lymphatic differentiation and sprouting in zebrafish.
Injection of an uncaged morpholino at very early stages of embryo development leads to a constitutive knockdown of gene function. We used a complementary approach, based on an inducible overexpression of a dominant-negative mutant form of mouse Sox18 in a newly developed stable transgenic line, to interfere with zebrafish SoxF proteins function in a temporally regulated way. A series of heat shocks enabled us to conclude that Sox18 and, possibly, other SoxF proteins function at the time of lymphatic precursor emergence from the PCV and regulate PL sprouting post-arteriovenous differentiation.
Although many signaling pathways have been implicated in lymphangiogenesis,2,42 it has been pointed out that most of their effects may be secondary to the induction of VEGF-C/D in a variety of cell types.1 VEGF-C/VEGFR3 signaling has an established and evolutionarily conserved role in lymphatic development.3,19,20,31,43 The current literature holds that the specification of LEC fate and the sprouting of LECs from the PCV are regulated independently.1,44,45 Our findings that Sox18 and VegfC show strong genetic interaction in zebrafish lymphatic development challenge this model and mandate a careful mechanistic analysis of this interaction in vertebrate model systems. Specifically, we show that embryos coinjected with subcritical doses of morpholinos against sox18 and vegfc, which produce little or no effect when injected separately, display severe PL and TD defects. This observation seems to be highly specific because the trunk vascular tree does not show abnormalities at the morphological or molecular marker levels. Interestingly, whereas the single knockdown of sox18 does not perturb venous sprouts but only PLs, double partial knockdown of sox18 and vegfc impacts more generally on all secondary angiogenesis from the vein, thus possibly revealing a combined role of both genes in endothelial cell migration. Notably, cell culture data revealed a role for Sox18 in controlling cell migration.46
The molecular mechanisms underlying the Sox18–VEGF-C crosstalk remain to be elucidated. Our data exclude a simple cross-regulation at transcripts level between sox18 and vegfc. Knockdown of sox18 does not alter vegfc (nor flt4/vegfr3) ISH signals, and sox18 hybridization signals are not affected in vegfc morphants. A possibility exists that VEGF-C/VEGFR3 signaling is implicated in modulating Sox18 transcriptional activity by inducing a post-translational modification. Sox18 has been shown to bind to and activate its target genes in vitro only on stimulation with VEGF-C (M.F., personal communication).14 In cultured mouse LECs, stimulation with VEGF-C does not alter Sox18 mRNA level or the activity of a 5-kb fragment of Sox18 promoter (M.F., personal communication), whereas modulating the transcriptional activity of SOX18 protein. These observations may point to a post-translational modification mechanism, which remains to be studied. Interestingly, vegfc overexpression ameliorates TD formation in embryos where sox18 is knocked down, possibly suggesting that the transcriptional activity of the residual Sox18 protein is positively modulated by enhanced VegfC/Vegfr3 signaling. Whatever the mechanism may be, it is clear from the data presented here that the strong Sox18–VEGF-C interplay in lymphangiogenesis is evolutionarily conserved and points to a novel molecular mechanism in lymphangiogenesis that remains to be further investigated.
Sox18 expression is not required for the maintenance of the lymphatic identity in mammals, whereas under pathological conditions, such as tumor growth, Sox18 is critical for tumor-induced angiogenesis and lymphangiogenesis.46,47 Both Sox18 and VEGF-C/VEGFR3 are promising targets for inhibition of tumor lymphangiogenesis.47,48 Our findings uncover a novel interplay between a key transcription factor and one of the most potent lymphangiogenic growth factors, hence opening new potential therapeutic avenues.
Note added in proof: A mutation in VEGFC has just been reported in a patient affected by Milroy-like primary lymphedema.49
We thank Maria V. Flores for sending a lyve1 plasmid; Neil Bower and Giuseppina Caretti for their help in quantitative reverse transcriptase-polymerase chain reaction primer design and data analysis; and Giuseppe Brunetti for his help in fish husbandry. Imaging at Institute for Molecular Bioscience was performed through the Dynamic Biology Imaging Facility of the Australian Cancer Research Foundation.
Sources of Funding
We acknowledge financial support by Regione Lombardia (grant SAL-01 to M. Beltrame) and by Fondazione Cariplo (grant 2011-0555 to M. Beltrame). B.M. Hogan was supported by an Australian Research Council Future Fellowship (FT100100165).
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.112.300254/-/DC1.
- Received August 3, 2012.
- Accepted March 5, 2013.
- © 2013 American Heart Association, Inc.
- Karkkainen MJ,
- Haiko P,
- Sainio K,
- Partanen J,
- Taipale J,
- Petrova TV,
- Jeltsch M,
- Jackson DG,
- Talikka M,
- Rauvala H,
- Betsholtz C,
- Alitalo K
- Kazenwadel J,
- Secker GA,
- Liu YJ,
- et al
- Carter TC,
- Phillips RJS
- Oliver G,
- Srinivasan RS
- Connell F,
- Kalidas K,
- Ostergaard P,
- Brice G,
- Homfray T,
- Roberts L,
- Bunyan DJ,
- Mitton S,
- Mansour S,
- Mortimer P,
- Jeffery S
- Isogai S,
- Lawson ND,
- Torrealday S,
- Horiguchi M,
- Weinstein BM
- Bussmann J,
- Bos FL,
- Urasaki A,
- Kawakami K,
- Duckers HJ,
- Schulte-Merker S
- Cermenati S,
- Moleri S,
- Cimbro S,
- Corti P,
- Del Giacco L,
- Amodeo R,
- Dejana E,
- Koopman P,
- Cotelli F,
- Beltrame M
- Herpers R,
- van de Kamp E,
- Duckers HJ,
- Schulte-Merker S
- Hosking B,
- François M,
- Wilhelm D,
- Orsenigo F,
- Caprini A,
- Svingen T,
- Tutt D,
- Davidson T,
- Browne C,
- Dejana E,
- Koopman P
- Lim AH,
- Suli A,
- Yaniv K,
- Weinstein B,
- Li DY,
- Chien CB
- Hogan BM,
- Herpers R,
- Witte M,
- Heloterä H,
- Alitalo K,
- Duckers HJ,
- Schulte-Merker S
- Hosking BM,
- Wang SC,
- Downes M,
- Koopman P,
- Muscat GE
- Francois M,
- Harvey NL,
- Hogan BM
- Srinivasan RS,
- Geng X,
- Yang Y,
- Wang Y,
- Mukatira S,
- Studer M,
- Porto MP,
- Lagutin O,
- Oliver G
- Young N,
- Hahn CN,
- Poh A,
- Dong C,
- Wilhelm D,
- Olsson J,
- Muscat GE,
- Parsons P,
- Gamble JR,
- Koopman P
- Duong T,
- Proulx ST,
- Luciani P,
- Leroux JC,
- Detmar M,
- Koopman P,
- Francois M
- Gordon K,
- Schulte D,
- Brice G,
- Simpson MA,
- Roukens MG,
- van I,
- mpel A,
- Connell F,
- Kalidas K,
- Jeffery S,
- Mortimer PS,
- Mansour S,
- Schulte-Merker S,
- Ostergaard P
This work reveals for the first time a conserved role for the transcription factor Sox18 in lymphatic development in a non-mammalian organism, thus strengthening the use of zebrafish as a potent system to study the molecular network at the basis of lymphangiogenesis. Our data reinforce the notion that Sox18 controls the early phases of lymphangiogenesis.
Transcription factors and growth factor pathways have been implicated in lymphangiogenesis, but their interplay has not yet been extensively studied. The genetic interaction we observe points to a so far poorly characterized role of the VEGFC growth factor pathway in the modulation of the activity of a key transcription factor that regulates lymphangiogenesis.