Lymphatic fate determination: playing RAF with ERK.

The mammalian circulation is composed of three distinct vascular networks: arterial, venous and lymphatic. Their distinct identities are established by endothelial cell fate specification that occurs during embryonic development. While arteries and veins form at the early stage of embryonic development, lymphatic vessels are thought to originate from veins at a later time point. In mouse, this process begins at E9.5, when a subset of cardinal vein endothelial cells in its dorsolateral portion begins expressing SOX18 transcription factor. SOX18, in turn, induces the expression of PROX1, a master controller of lymphatic endothelial cells fate at E10.5. These SOX18+/PROX1+ cells then migrate laterally from the cardinal vein, starting at E11.5, to form jugular lymphatic sacs at E12.5 that later condense to form a lumenized peripheral longitudinal lymphatic vessel, which will eventually become a thoracic duct. Once delaminated from the cardinal veins, these cells soon become mature lymphatic endothelial cells and express a full array of lymphatic markers., At later stages of development, SOX18 expression is shut off both in cardinal veins and in jugular lymph sacs, while PROX1 expression becomes largely restricted to the lymphatic vasculature.

Recent studies have demonstrated a critical role played by VEGF-C in the induction of migration of SOX18+/PROX1+ lymphatic endothelial progenitors from cardinal veins to form jugular lymphatic sacs and in promotion of growth of the primitive lymphatic vasculature. However, the key initiating event in this process, polarized activation of SOX18 expression, remained a mystery. SOX18 belongs to group F family of the Sox family transcription factors that also includes SOX 7 and SOX17. It induces expression of a number of endothelial proteins, including aforementioned PROX1, claudin-5 and VCAM-1 among others. However, to date, there is a paucity of information regarding molecular mechanisms regulating SOX18 expression.

In a recent study, we demonstrated that introduction of a mutant RAF1 gene (RAF1) that cannot be shut down by phosphorylation of its Ser259 site into endothelial cells leads to constitutive activation of ERK. Remarkably, this results in induction of SOX18 and PROX1 expression both in vitro and in vivo. Unlike the normal pattern of SOX18 expression, when the transcription factor is restricted to the dorsolateral portion of cardinal veins, its expression in RAF1 endothelial-specific transgenic mice is observed throughout the entire cardinal vein. Moreover, PROX1 is also strikingly induced both in venous and arterial endothelium. The latter is never seen under normal circumstances. COUP-IIF, a venous and lymphatic endothelium-specific transcription factor, is thought to be required for PROX1 expression. However, these data indicate that ERK signaling is capable of inducing PROX1 expression in the absence of COUP-TFII. It also induces expression of other lymphatic endothelial markers, such as VEGFR3, Podoplanin and LYVE1, demonstrating a broad induction of lymphatic fate determination.

In developing mice, this increased commitment of venous endothelial cells to the lymphatic fate results in excessive conversion of cardinal vein endothelium to the lymphatic endothelium with a massive expansion in the size of jugular lymphatic sacs and, subsequently, formation of enlarged and malformed lymphatic vessels, while veins are dramatically reduced in diameter. The resulting abnormalities, termed lymphangiectasias, are highly reminiscent of pathology seen in patients with Noonan syndrome and similar “Rasopathies” that are caused by increased ERK activation. This phenotype could be rescued by inhibition of ERK signaling, indicating that excessive ERK activation-induced lymphatic endothelial fate specification is the basis of lymphatic abnormalities seen in these patients.

These results clearly demonstrate that ERK activation induces SOX18 expression. However, ERK activation here was induced not in a conventional way, for example, by activation of a receptor tyrosine kinase, but by removal of an inhibitory effect of RAF1 Ser259 phosphorylation. Previously we have demonstrated that under normal conditions AKT inhibits ERK activation by phosphorylating RAF1 on Ser259. It should be noted, however, that AKT is not the only kinase capable of phosphorylating this Ser site in RAF1. Other known candidates include protein kinase A (PKA) and protein kinase Cα (PKCα).

Given the polarized nature of SOX18 expression in cardinal veins and the temporarily restricted pattern of its expression, it is tempting to speculate that ERK activation in these cells occurs due to a spatially localized signal that transiently inhibits a kinase that phosphorylates RAF1 Ser259, thereby allowing ERK activation and induction of SOX18 expression to take place. This, in turn, allows induction of PROX1 and subsequent specification of lymphatic fate (Fig. 1). The failure of this process to take place or its shorter than normal duration will result either in total or partial failure of lymphatic development, while excessive activation of ERK signaling will lead to lymphangiectasia and other features of human RASopathies syndromes.

Figure 1. Schematic model of RAF1/ERK signaling-induced lymphatic endothelial cell (LEC) fate specification in mouse. At around E9.5, the inhibitory effect of AKT or other kinases on RAF1 is released in response to a certain unknown signal in the dorsolateral portion of the cardinal vein, leading to activation of RAF1 and downstream ERK signaling. This, in turn, induces SOX18 and, subsequently, PROX1 expression. The SOX18+/PROX1+ cells then move out from cardinal vein and migrate dorsolaterally at around E11.5 to form jugular lymph sacs by E14.5. Once delaminated from the vein, these cells shut down SOX18 expression and start expressing mature lymphatic markers such as VEGFR3 and Podoplanin.

In summary, our study revealed a novel basis for the lymphatic fate specification and described a molecular basis for lymphangiectasia and other lymphatic abnormalities seen in various “RASopathy” syndromes, which may be of therapeutic benefit for treatment of certain diseases of lymphatic circulation.