Nonvenous Origin of Dermal Lymphatic VasculatureNovelty and Significance
Rationale: The formation of the blood vasculature is achieved via 2 fundamentally different mechanisms, de novo formation of vessels from endothelial progenitors (vasculogenesis) and sprouting of vessels from pre-existing ones (angiogenesis). In contrast, mammalian lymphatic vasculature is thought to form exclusively by sprouting from embryonic veins (lymphangiogenesis). Alternative nonvenous sources of lymphatic endothelial cells have been suggested in chicken and Xenopus, but it is unclear whether they exist in mammals.
Objective: We aimed to clarify the origin of the murine dermal lymphatic vasculature.
Methods and Results: We performed lineage tracing experiments and analyzed mutants lacking the Prox1 transcription factor, a master regulator of lymphatic endothelial cell identity, in Tie2 lineage venous–derived lymphatic endothelial cells. We show that, contrary to current dogma, a significant part of the dermal lymphatic vasculature forms independently of sprouting from veins. Although lymphatic vessels of cervical and thoracic skin develop via sprouting from venous-derived lymph sacs, vessels of lumbar and dorsal midline skin form via assembly of non–Tie2-lineage cells into clusters and vessels through a process defined as lymphvasculogenesis.
Conclusions: Our results demonstrate a significant contribution of nonvenous-derived cells to the dermal lymphatic vasculature. Demonstration of a previously unknown lymphatic endothelial cell progenitor population will now allow further characterization of their origin, identity, and functions during normal lymphatic development and in pathology, as well as their potential therapeutic use for lymphatic regeneration.
- developmental biology
- endothelial cells
- endothelial progenitor cells
- lymphatic vessels
Lymphatic vasculature was traditionally considered a passive drainage system responsible for removal of fluid, molecules, and cells from tissues. However, emerging evidence shows active roles of lymphatic vessels in inflammation, immunity, lipid metabolism, blood pressure regulation, and metastasis, and consequent involvement in common diseases such as autoimmune diseases, atherosclerosis, and cancer.1 Despite these recent discoveries, our knowledge about the mechanism regulating lymphatic vessel formation and function is limited.
According to a widely accepted theory, originally proposed by Florence Sabin in the early 20th century, lymphatic vessels form during embryogenesis by sprouting from the veins. An alternative theory by Huntigton and McClure suggested that lymphatic vessels develop from mesenchymal lymphangioblasts. Recent studies using molecular biological and real-time in vivo imaging techniques provide support for the concept of transdifferentiation of venous into lymphatic endothelial cells (LECs) and suggest veins as the sole origin of the entire mammalian lymphatic vasculature.2,3 Alternative nonvenous sources of LECs have been suggested in chicken and Xenopus,4,5 but whether they exist in mammals is unclear.
Editorial, see p 1630
In This Issue, see p 1629
Here, we investigated the development of murine dermal lymphatic vessels that are thought to form by sprouting from venous-derived primitive lymphatic vessels, the peripheral longitudinal lymphatic vessel, and primordial thoracic duct, also referred to as jugular lymph sacs (JLS).2,6,7 We provide genetic lineage tracing data and functional evidence to demonstrate that, contrary to current dogma, a significant part of the dermal lymphatic vasculature forms independent of Tie2 lineage venous–derived LECs.
Detailed Methods section is available in the Online Data Supplement.
Region-Specific Differences in the Development of the Lymphatic Vessels in the Skin
To visualize dermal lymphatic vessel formation, we used the Vegfr3-lacZ reporter mice (Vegfr3lz). Consistent with previous data, we observed that the first lymphatic vessel sprouts reaching the skin emanated from the JLS at embryonic day (E) 12.5 (Figure 1A; Online Figure IA). Concomitant with the extension of sprouts dorsally, rapid emergence of dermal lymphatic vessels was observed on the lateral side of the embryo between E12.5 and E13.5 (Figure 1A and 1B; Online Figure IA and IB). Whole-mount analysis of the skin showed the presence of scattered cells and discontinuous vessel networks that were not connected to JLS in the lateral skin at lumbar region (Figure 1B). Analysis of skin at E15.5 further revealed isolated clusters of Vegfr3lz-positive cells in the dorsal midline (Figure IC), where vessels from contralateral sides anastamose by E17.5 (Online Figure IB). Notably, dermal lymphatic vessels at the lumbar region seemed to develop independently from subcutaneous lymphatic vessels until E17.5 when connections between the 2 networks were formed (Online Figure IB). The latter developed along major arteries and veins, whereas superficial dermal lymphatic vessels did not show an apparent alignment with blood vessels (Online Figure IB, data not shown). These results suggest that lymphatic vessels in different regions of the skin (cervical versus lumbar, dermis versus subcutis) develop via different mechanisms. In particular, emergence of LEC clusters without connection to vessel sprouts from the JLS suggests a novel mechanism of vessel formation and potentially a different cellular origin.
We next analyzed the expression of Prox1, the first known marker of differentiated LECs,8 and the established lymphatic markers Nrp2, LYVE-1, and VEGFR3 in the dermal vasculature. Immunofluorescence analysis of lumbar skin from E13.5 Prox1-green fluorescent protein (GFP) embryos confirmed the presence of a discontinuous network and isolated clusters of GFP and Nrp2-positive LECs (Figure 1D). Some clusters were interconnected via long membrane protrusions, whereas others were isolated (Figure 1D; Online Movies I and II). Immunofluorescence for Prox1 similarly highlighted the Nrp2+ LEC clusters (Figure 1D). Surprisingly, most LEC clusters appeared LYVE-1−, whereas developing vessels showed a weak LYVE-1 immunoreactivity (Figure 1E). Analysis of Vegfr3-GFP reporter embryos allowed visualization of isolated GFP+ LECs already at E11.5 (Figure 1F).
Formation of Lymphatic Vessels in the Lumbar Skin Is Independent of Tie2 Lineage Venous–derived LECs
To investigate the origin of dermal lymphatic vessels, we used Cre/loxP-based lineage tracing. A transgenic mouse line expressing Cre recombinase under the control of the (blood) endothelial/hematopoietic-specific Tie2 promoter was crossed with the R26-mTmG double reporter line to allow irreversible marking of Tie2-expressing cells and their descendants with GFP, whereas all other cell types express the red fluorescent protein Tomato (Figure 2A). To first assess the efficiency of Cre-mediated recombination in venous LYVE-1+ LEC progenitors and LECs forming the primordial thoracic duct/peripheral longitudinal lymphatic vessel, we performed fluorescence-activated cell sorting (FACS) analysis of ECs from Tie2-Cre;R26-mTmG embryos at E9, E10, and E11 (Figure 2A). At E9 ≥89% of the LYVE-1+ ECs expressed GFP, increasing to ≥95% at E10 and E11, indicating efficient targeting by the Tie2-Cre transgene (Figure 2B; Online Figure IIA). LYVE-1− ECs, including LEC progenitors generated from the superficial venous plexus,6 also showed highly efficient Tie2-Cre–mediated recombination at E11 (98.8±0.2% [n=4]; Online Figure IIB). A major proportion of ECs were also Tomato+ at E9 and E10 (Figure 2B), because of perdurance of Tomato protein after Cre recombination that results in double-marker expression.9 Immunofluorescence analysis of E10.5 Tie2-Cre;R26-mTmG embryos further revealed GFP+ cardinal veins (Figure 2C), which provide a source of LECs.2 Consistent with this, JLSs of E12.5 Tie2-Cre;R26-mTmG embryos also showed efficient Cre recombination (Figure 2D; Online Figure IIC).
We next assessed the contribution of venous-derived cells to dermal lymphatic vessels at E12.5, when the first LEC clusters emerge, and at E15.5, when lymphatic sprouts reach the dorsal midline area (Figure 2E). If derived by sprouting from Tie2-lineage blood ECs (BECs), lymphatic clusters and vessels were expected to express GFP. Whole-mount analysis of the skin unexpectedly revealed GFP− LECs (Figure 2F and 2G). FACS analysis of E13 Tie2-Cre;R26-mTmG skin demonstrated that a significant proportion of dermal LYVE-1+/podoplanin+ LECs (33.5±6.7%; n=8) were indeed GFP− and thus not derived from Tie2 lineage cells (Figure 2H). GFP− LECs were instead Tomato+, thus confirming reporter gene expression but lack of recombination in these cells and their progenitors. A large proportion (35.7±3.5%; n=8) of dermal LECs coexpressed Tomato and GFP (Figure 2H), demonstrating recent upregulation of Tie2. The proportion of GFP+ LECs increased at later stages of development, suggesting progressive induction of Tie2 in the developing vessels (Online Figure IID and data not shown). Importantly, analysis of venous-derived LECs at E11 (Figure 2B; Online Figure IIB) and blood endothelial cells at E13 (Figure 2H), at the stage of LEC cluster emergence, showed efficient targeting by the Tie2 transgene (98.1±0.4% at E11; n=3 and 99.4±0.7% at E13; n=8), thus excluding these cells as the origin of dermal LECs.
We next sought functional evidence for the nonvenous origin of dermal lymphatic vessels by deleting Prox1, the master regulator of LEC fate,10 in blood endothelia using a conditional Prox1flox allele (Online Figure IIIA–IIID) in combination with the Tie2-Cre transgene. If lymphatic vessels were formed entirely by transdifferentiation of Tie2 lineage ECs, as stated by the current dogma, no or few lymphatic vessels were expected to form in the Prox1flox/flox;Tie2-Cre embryos. In agreement with previous data,2 we indeed found that E14.5 Prox1flox/flox;Tie2-Cre embryos showed subcutaneous edema and failure of lymphatic vessel formation in the cervical skin (Figure 3A; Online Figure IVA). However, we observed blood-filled dermal lymphatic vessels and isolated LEC clusters in the lumbar skin (Figure 3A). Most LEC clusters were not targeted by the Tie2-Cre transgene and thus expressed Prox1 in the mutant embryos (Online Figure IVB). These data demonstrate that the formation of lymphatic vessels in the lumbar skin is independent of Tie2 lineage venous–derived LECs. Most lymphatic structures were however lost in Prox1flox/flox;Tie2-Cre skin by E17.5 (Online Figure IVC), likely because of progressive induction of Tie2 in the dermal lymphatic vasculature (Online Figure IID, data not shown) and consequent loss of Prox1 that is required for lymphatic vessel maintenance.11
Tracing of Prox1-Expressing Cells Suggests Continuous LEC Differentiation During Dermal Lymphatic Vessel Formation
To further investigate the origin of the dermal lymphatic vasculature, we used a tamoxifen-inducible Cre line to allow genetic labeling of cells expressing Prox1 (Online Figure V). In contrast to a previous report,2 we found that injection of 4-OHT to pregnant females at E10.5 or E11.5 led to a nearly complete absence of GFP+ LECs in the skin (Online Figure VIA and VIB) and JLS (data not shown). It was possible that cells were not labeled because of continuous LEC differentiation between E10 and E122 and a short <24-hour time-window of 4-hydroxytamoxifen (4-OHT) activity in our study, in comparison with the longer period of activity after administration with tamoxifen in the previous study2 (Online Figure VIIA and VIIB). When 4-OHT was instead administered at E12.5 (Online Figure VIC) or E13.5 (data not shown), JLS was well labeled (Online Figure VIC). If dermal lymphatic vessels were derived through continuous migration and proliferation of LECs sprouting from the GFP+ JLS, they would be expected to express GFP particularly at the tips of the sprouts. Surprisingly, however, dermal lymphatic vasculature showed a high proportion of GFP− cells at the distal end of the vasculature (Figure 3B; Online Figure VIB). In addition, 4-OHT administration at E14.5 or E15.5 resulted in isolated GFP− LEC clusters at the midline, despite efficient recombination in the rest of the vasculature (Figure 3B; Online Figure VIB). Low GFP labeling selectively at vessel tips and isolated LEC clusters suggests incorporation of newly differentiated cells into the growing vessels at the vascular front.
In addition to visualization of lymphatic vessels, we observed a small population of scattered GFP+ cells in the Prox1-CreERT2;R26-mTmG embryos that were also positive for LYVE-1 and markers of the macrophage lineage (Online Figure VIIIA and VIIIB). Lineage tracing using Vav-Cre mice however excluded definitive hematopoietic lineage cells as a source of LECs (Online Figure VIIIC), as previously reported.12
Taken together, our data demonstrate the existence of a population of LEC progenitors that is distinct from venous-derived LECs and contributes to the formation of the dermal lymphatic vasculature (Figure 3C).
This study demonstrates that a large part of the superficial dermal lymphatic vasculature does not form via transdifferentiation and sprouting of Tie2 lineage venous ECs, which are currently thought to be the sole origin of the mammalian lymphatic vasculature. We found that lymphatic vessels of the dorsal midline and lumbar regions of the skin instead form from nonvenous-derived progenitors through a lymphvasculogenesis process involving assembly of LECs into clusters and their further coalescence to continuous vessel networks. The Prox1-CreERT2 lineage tracing defined E12.5–E14.5 as the critical time-window when nonvenous-derived LEC progenitors most actively incorporate into dermal lymphatic vessels, which is in agreement with the rapid emergence of these vessels at around E13.5.
The embryonic origin of lymphatic vessels has been controversial until recently. Genetic lineage tracing experiments in mouse and real-time imaging in zebrafish confirmed Sabin’s theory on the venous origin of lymphatic vessels.2,3 These experiments did not, however, exclude the existence of alternative sources of LECs. Earlier observations interestingly showed that avian lymphatic vasculature has a dual origin, with JLSs originating from veins and superficial dermal lymphatic vessels from an unidentified nonvenous-derived precursors of mesodermal origin.5 Together with our findings, this suggests evolutionarily conserved origins of LECs and mechanisms of lymphatic vessel formation.
Given the distinct origins, molecular mechanisms regulating lymphatic vessel formation in different regions of the skin are likely different. To understand these mechanisms, it is now of critical importance to first clarify the cell of origin of dermal LECs and identify their potential contribution to lymphatic vessels in other organs. Here, we excluded Tie2 lineage endothelial/hematopoietic cells and Vav lineage definitive hematopoietic cells as the source of dermal LECs. Interestingly, our recent study identified hemogenic endothelium-derived cells as an alternative nonvenous origin of lymphatic vessels in the mesentery.13 However, the origin of nonvenous-derived LECs in the skin and mesentery seems to be different because, unlike dermal LECs, mesenteric LECs are derived from Tie2-lineage cells (unpublished data). Additional cell-type–specific lineage tracing experiments are required to identify the cellular source of nonvenous-derived LEC progenitors, which will allow further studies on their potential therapeutic use for lymphatic regeneration in disease in the future.
We thank Dimitris Kioussis (National Institute for Medical Research, London) for providing Vav-Cre mice, transgenic services at London Research Institute (LRI) for help with establishing mouse lines, staff at LRI and Uppsala University animal units for animal husbandry and Henrik Ortsäter for help with mice. Anna Caldwell at King’s College London is acknowledged for mass spectrometry analysis of 4-OHT in sera. We thank the Light Microscopy unit at LRI and BioVis at Uppsala University for advice and help with experiments.
Sources of Funding
This study was supported by Cancer Research UK (I. Martinez-Corral, L. Stanczuk, F. Tatin, and T. Makinen), EMBO Young Investigator Programme, Swedish Research Council and the Kjell and Märta Beijer Foundation (T. Makinen), Fundación Alfonso Martin Escudero (I. Martinez-Corral), Howard Hughes Medical Institute (K. Kizhatil, S.W.M. John), EY11721 (S.W.M. John), European Research Council (ERC-2010-AdG-268804) and the Leducq Foundation (11CVD03; K. Alitalo), and Ministry of Science and Innovation of Spain (grants BIO2009-09488 and SAF2010-18765; S.O.). S.W.M. John is an investigator of the Howard Hughes Medical Institute.
In February 2015, the average time from submission to first decision for all original research papers submitted to Circulation Research was 13.9 days.
Brief UltraRapid Communications are designed to be a format for manuscripts that are of outstanding interest to the readership, report definitive observations, but have a relatively narrow scope. Less comprehensive than Regular Articles but still scientifically rigorous, BURCs present seminal findings that have the potential to open up new avenues of research. A decision on BURCs is rendered within 7 days of submission.
Current address (F.T.): I2MC INSERM UMR 1048, Toulouse Cedex, France.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.116.306170/-/DC1.
- Nonstandard Abbreviations and Acronyms
- embryonic day
- Green fluorescent protein
- Jugular lymph sac
- Lymphatic endothelial cell
- Received February 5, 2015.
- Revision received February 23, 2015.
- Accepted March 2, 2015.
- © 2015 American Heart Association, Inc.
- Srinivasan RS,
- Dillard ME,
- Lagutin OV,
- Lin FJ,
- Tsai S,
- Tsai MJ,
- Samokhvalov IM,
- Oliver G.
- Hägerling R,
- Pollmann C,
- Andreas M,
- Schmidt C,
- Nurmi H,
- Adams RH,
- Alitalo K,
- Andresen V,
- Schulte-Merker S,
- Kiefer F.
- Yang Y,
- García-Verdugo JM,
- Soriano-Navarro M,
- Srinivasan RS,
- Scallan JP,
- Singh MK,
- Epstein JA,
- Oliver G.
- Wigle JT,
- Harvey N,
- Detmar M,
- Lagutina I,
- Grosveld G,
- Gunn MD,
- Jackson DG,
- Oliver G.
- Johnson NC,
- Dillard ME,
- Baluk P,
- McDonald DM,
- Harvey NL,
- Frase SL,
- Oliver G.
- Gordon EJ,
- Rao S,
- Pollard JW,
- Nutt SL,
- Lang RA,
- Harvey NL.
- Stanczuk L,
- Martinez-Corral I,
- Ulvmar MH,
- Zhang Y,
- Laviña B,
- Fruttiger M,
- Adams RH,
- Saur D,
- Betsholtz C,
- Ortega S,
- Alitalo K,
- Graupera M,
- Makinen T
Novelty and Significance
What Is Known?
The mammalian lymphatic vasculature forms by sprouting from embryonic veins.
What New Information Does This Article Contribute?
Part of the mammalian dermal lymphatic vasculature originates from an alternative nonvenous source.
Nonvenous-derived lymphatic endothelial progenitors form vessels through a novel process defined as lymphvasculogenesis.
Endothelial cells forming the lymphatic vasculature have been described to originate through transdifferentiation and sprouting of venous endothelial cells. Alternative nonvenous sources of lymphatic endothelial cells have been suggested, although never demonstrated to exist in mammals. In this study, we reinvestigated the origin of the lymphatic vasculature by fate mapping. We found that lymphatic vessels in different regions of the skin have different origins and mechanisms of development. Vessels in the neck region form from veins through a sprouting process, as demonstrated in previous studies. However, in the lumbar region of the skin, lymphatic vessels develop from an alternative nonvenous source via a different process that we define as lymphvasculogenesis. Our data provide fundamental novel insight into the mechanism of lymphatic vessel formation by demonstrating the existence of a previously unknown lymphatic endothelial cell progenitor. Identification and characterization of the nonvenous lymphatic endothelial cell progenitors will allow further studies on their potential therapeutic use for lymphatic regeneration in disease in the future.